Professional Documents
Culture Documents
f-. '
LINEAR
CIRCUITS
EDITION
Raymond A. DeCarlo
Purdue University
Pen-Min Lin
Purdue University
Kendall Hunt
p u b l i s h i n g
c o m p a n y
n
o
Kendall Hunft
p u b l i s h i n g
c o m p a n y
www.kendallhunt.cpm
Send all inquiries to:
4050 Westmark Drive
Dubuque, lA 52004-1840
Copyright 2001, 2009 Raymond A. DeCarlo and Pen-Min Lin
Copyright 1995 Prentice-Hall, Inc.
ISBN 978-0-7575-6499-4
All rights reserved. No part of this publication may be reproduced,
stored in a retrieval system, or transmitted, in any form or by any means,
electronic, mechanical, photocopying, recording, or otherwise,
without the prior written permission of the copyright owner.
r^
TABLE OF CO N TEN TS
Preface......................................................................................................................................................................vii
Chapter 1 Charge, Current, Voltage and Ohms Law ............................................................................ 1
Chapter 2 Kirchhoffs Current & Voltage Laws and Series-Parallel Resistive C ircu its..............51
Chapter 3 Nodal and Loop Analyses....................................................................................................... 107
Chapter 4 T he Operational Amplifier..................................................................................................... 155
Chapter 5 * Linearity, Superposition, and Source Transform ation................................................... 191
Chapter 6 Thevenin, Norton, and Maximum Power Transfer Theorems.................................... 227
Chapter 7 Inductors and C apacitors....................................................................................................... 269
Chapter 8 First Order RL and RC Circuits...........................................................................................321
Chapter 9 Second Order Linear Circuits................................................................................................379
Chapter 10 Sinusoidal Steady State Analysis by Phasor Methods .................................................431
Chapter 11 Sinusoidal State State Power Calculations.......................................................................499
Chapter 12 Laplace Transform Analysis L Basics................................................................................. 543
Chapter 13 Laplace Transform Analysis II: Circuit Applications................................................... 603
Chapter 14 Laplace Transform Analysis III; Transfer Function Applications.............................683
Chapter 15 * Time Domain Circuit Response Computations: The Convolution M ethod...... 763
Chapter 16 Band-Pass Circuits and Resonance....................................................................................811
Chapter 17 * Magnetically Coupled Circuits and Transformers........................................................ 883
Chapter 18 Tw o-Ports...................................................................................................................................959
Chapter 19 Principles o f Basic Filtering ............................................................................................. 1031
Chapter 20 Brief Introduction to Fourier Series .............................................................................. 1085
In d ex................................................................................................................................................................... 1119
o
n
p ,
0
o
n
o
0
o
n
o
o
0
0
0
n
O
n
^
o
PREFACE
For the last several decades, EE/ECE departments o f US universities have typically required two
semesters o f linear circuits during the sophomore year for EE majors and one semester for other
engineering majors. Over the same time period discrete time system concepts and computer engi
neering principles have become required fare for EE undergraduates. Thus we continue to use
Laplace transforms as a vehicle for understanding basic concepts such as impedance, admittance,
fdtering, and magnetic circuits. Further, software programs such as PSpice, MATLAB and its tool
boxes, Mathematica, Maple, and a host o f other tools have streamlined the computational drudg
ery o f engineering analysis and design. MATLAB remains a working tool in this 3'''^ edition o f
Linear Circuits.
In addition to a continuing extensive use o f MATLAB, we have removed much o f the more com
plex material from the book and rewritten much o f the remaining book in an attempt to make the
text and the examples more illustrative and accessible. More importantly, many o f the more diffi
cult homework exercises have been replaced with more routine problems often with numerical
answers or checks.
Our hope is that we have made the text more readable and understandable by todays engineering
undergraduates.
Introduce and investigate three basic electrical quantities: charge, current, and voltage,
and the conventions for their reference directions.
2.
3.
4.
5.
6.
7.
8.
an electric motor that runs the compressor in an air conditioner or the pump in a dish
washer;
a microwave oven;
a radio, TV, or stereo;
an iPod;
a car heater.
In this text, we define and analyze common circuit elements and describe their interaction. Our
aim is to create a modular framework for analyzing circuit behavior, while simultaneously devel
oping a set o f tools essential for circuit design. These skills are, o f course, crucial to every electri
cal engineer. But they also have broad applicability in other fields. For instance, disciplines such
as bioengineering and mechanical engineering have similar patterns o f analysis and often utilize
circuit analogies.
W H A T IS A C IR C U IT ?
A circuit is an energy or signal/information processor. Each circuit consists o f interconnections o f
simple circuit elements, or devices. Each circuit element can, in turn, be thought o f as an ener
gy or signal/information processor. For example, a circuit element called a source produces a
voltage or a current signal. This signal may serve as a power source for the circuit, or it may rep
resent information. Information in the form o f voltage or current signals can be processed by the
circuit to produce new signals or new/different information. In a radio transmitter, electricity
powers the circuits that convert pictures, voices, or music (that is, information) into electromag
netic energy. This energy then radi
ates into the atmosphere or into
space from a transmitting antenna.
A satellite in space can pick up this
electromagnetic energy and trans
mit it to locations all over the
world. Similarly, a T V reception
antenna or a satellite dish can pick
up and direct this energy to a T V
set. T h e T V contains circuits
(Figure 1.1) that reconvert the
information within the received
signal back into pictures with
sound.
FIG U RE 1.1 Cathode ray tube with surrounding circuitry for
converting electrical signals into pictures.
C?
AN SW ER; 333,3 9 1 ,5 9 7
Exercise. Sketch the time-dependent charge profile q{t) = 3 (l-^ ^ 0 C, ? > 0, present on a metal
plate. M ATLAB is a good tool for such sketches.
A conductor refers to a material in which electrons can move to neighboring atoms with relative
ease. Metals, carbon, and acids are common conductors. Copper wire is probably the most com
mon conductor. An ideal conductor offers zero resistance to electron movement. Wires are
assumed to be ideal conductors, unless otherwise indicated.
Insulators oppose electron movement. Common insulators include dry air, dry wood, ceramic,
glass, and plastic. An ideal insulator offers infinite opposition to electron movement.
C U R R EN T
Current refers to the net flow o f charge across any cross section o f a conductor. T he net move
ment o f 1 coulomb (1 C) o f charge through a cross section o f a conductor in 1 second (1 sec)
produces an electric current o f 1 ampere (1 A). The ampere is the basic unit o f electric current
and equals 1 C/s.
The direction o f current flow is taken by convention as opposite to the direction o f electron flow,
as illustrated in Figure 1.2. This is because early in the history o f electricity, scientists erroneously
believed that current was the movement o f only positive charges, as illustrated in Figure 1.3. In
metallic conductors, current consists solely o f the movement o f electrons. However, as our under
standing o f device physics advanced, scientists learned that in ionized gases, in electrolytic solu
tions, and in some semiconductor materials, movement o f positive charges constitutes part or all
o f the total current flow.
One Ampere
of Current "
One
;
;
Cloud o f \
se co n d ^ ....... |---- 6.24x10 1
later
i
;
k electrons
Boundary
FIG U RE 1.2 A cloud o f negative charge moves past a cross section of an ideal conductor from right
to left. By convention, the positive current direction is taken as left to right.
One Ampere
of Current
One
Coulom b
One
of positive
'second
later
charge
Boundary
FIGURE 1.3 In the late nineteenth cenmry, current was thought to be the movement of a positive charge
past a cross section of a conduaor, giving rise to the conventional reference direction of positive current flow.
Both Figures 1.2 and 1.3 depict a current o f 1 A flowing from left to right. In circuit analysis, we
do not distinguish between these two cases: each is represented symbolically, as in Figure 1.4(a).
The arrowhead serves as a reference for determining the true direction o f the current. A positive
value o f current means the current flows in the same direction as the arrow. A current o f negative
value implies flow is in the opposite direction o f the arrow. For example, in both Figures 1.4a and
b, a current o f 1 A flows from left to right.
1A
-1A
>
<
(a)
(b)
FIG U RE 1.4 1 A of current flows from left to right through a general circuit element.
In Figure 1.4, the current is constant. The wall socket in a typical home is a source o f alternating
current, which changes its sign periodically, as we will describe shortly. In addition, a current direc
tion may not be known a priori. These situations require the notion o f a negative current.
E X A M P L E 1.1.
Figure 1.5 shows a slab o f material in which the following is true:
1. Positive charge carriers move from left to right at the rate o f 0.2 C/s.
2. Negative charge carriers move from right to left at the rate o f 0.48 C/s.
Given these conditions,
a) Find
and /^;
b)
Connecting
o
0
Connecting
wire
wire
Sem iconductor iVlaterial
a)
The current from left to right, due to the movement o f the positive charges, is 0.2 A. The
current from left to right, due to the movement o f the negative charges, is 0.48 A.
Therefore, /^, the total current from left to right, is 0.2 + 0.48 = 0.68 A. Since ly is the
current from right to left, its value is then -0 .6 8 A.
b)
Exercise. In Example 1.1, suppose positive-charge carriers move from right to left at the rate o f 0.5
C/s, and negative carriers move from left to right at the rate o f 0.4 C/s. Find
and
AN SW ER: /, = - 0 .9 A; ^ = 0.9 A
If a net charge
crosses a boundary in a short time frame o f At (in seconds), then the approxi
Aq
At
( 1 . 1)
where I, in this case, is a constant. The instantaneous (time-dependent) current flow is the limit
ing case o f Equation 1.1, i.e.,
dq{t)
dt
( 1. 2)
Here q{t) is the amount o f charge that has crossed the boundary in the time interval [tQ, t] . The
equivalent integral counterpart o f Equation 1.2 is
q{t) = J i{r)dr
(1.3)
E X A M P L E 1.2
The charge crossing a boundary in a wire is given in Figure 1.6(a) for ? > 0. Plot the current i{t)
through the wire.
(a)
(b)
FIG U RE 1.6 (a) Charge crossing a hypothetical boundary; (b) current flow
associated with the charge plot o f (a).
S o lutio n
As per Equation 1.2, the current is the time derivative o f q{t). The slopes o f the straight-Une seg
ments o f q{f} in Figure 1.6(a) determine the piecewise constant current plotted in Figure 1.6(b).
l-cos(co?)
Exercise. The charge crossing a boundary in a wire varies as q[t) = ---------------- C, for t >Q.
Compute the current flow.
A N SW ER: sin(cof) A, for f > 0
C, for t > 0 .
A, for f > 0
E X A M PLE 1.3
Find q{t), the charge transported through a cross section o f a conductor over [0, f], and also the total
charge Q transported, if the current dirough the conductor is given by die waveform o f Figure 1.7(a).
-l-*-t(se c)
FIG U RE 1.7 (a) Square-wave current signal; (b) q{t) equal to the integral of i{t) given in (a).
S o lutio n
Thus, q{t) is the running area under the i{t) versus t curve. Since i{t) is piecevv'ise constant, the
integral is piecewise linear because the area either increases or decreases linearly with time, as
shown in Figure 1.7(b). Since q{t) is constant for ^ > 3, the total charge transported is Q = q{5) =
3 C.
Exercise. If the current flow through a cross section o f conductor is i{t) = cos(120jtf) A for ? > 0
and 0 otherwise, find q{t) for t>Qi.
AN SW ER: q{t)
120jt
C for r > 0
Exercise. Suppose the current through a cross section o f conductor is given in Figure 1.8. Find
q{t) for t > 0 .
FIGURE 1.8
AN SW ER; q(t) =
C for 0 <
T Y P ES OF C U R R EN T
There are two very important current types: direct current (do) and alternating current (ac).
Constant current (i.e., dqldt = / is constant) is called direct current, which is illustrated graphi
cally m Figure 1.9(a). Figure 1.9(b) shows an alternating current, generally meaning a sinusoidal
waveform, i.e., current o f the form y4sin(w? + ()>), where A is the peak magnitude, co is the angu
lar frequency, and (|) is the phase angle o f the sine wave. W ith alternating current, the instanta
neous value o f the waveform changes periodically through negative and positive values, i.e., the
direction o f the current flow changes regularly as indicated by the + and - values in Figure 1.9(b).
Household current is ac.
Lastly, Figure 1.9(c) shows a current that is neither dc nor ac, but that nevertheless will appear in
later circuit analyses. There are many other types o f waveforms. Interestingly, currents inside com
puters, C D players, TV s, and other entertainment devices are typically neither dc nor ac.
i(t) (A)
t(sec)
-H ----------------------1 -
-I-----
3
(a)
F IG U R E 1.9 (a) Direct current, or dc; i{t) = Iq\ (b) alternating current, or ac;
10
Because the value o f an ac waveform changes with time, ac is measured in different ways. Suppose
the instantaneous value o f the current at time t is A!sin(ci)i- + (j>). The term peak value refers to K
in K sin(co? + (j)). The peak-to-peak value is 2K. Another measure o f the alternating current,
indicative o f its heating effect, is the root mean square (rms), or effective value. The rms or effec
tive value is related to the peak value by the formula
rms =
(i.4)
A derivation o f Equation 1.4 with an explanation o f its meaning will be given in Chapter 11.
A special instrument called an ammeter measures current. Some ammeters read the peak value,
whereas some others read the rms value. One type o f ammeter, based on the interaction between
the current and a permanent magnet, reads the average value o f a current. From calculus, Fave!
the average value o f any function y(^), over the time interval [0, 7] is given by
(1.5)
For a general ac waveform, the average value is zero. However, ac signals are often rectified, i.e.,
converted to their absolute values, in power-supply circuits. For such circuits, the average value o f
the rectified signal is important. From Equation 1.5, the average value o f the absolute value o f an
ac waveform over one complete cycle with T = 2jt/co, is
K ^
2.K
Average Value = ^\s,m{wt)\dt = ----- J
0
2K -cos{(ot)
T
(O
0.5T
sin(cot)clt
^ 0
2,K
= 0.636K
jt
( 1.6)
Exercise. Suppose i{t) - 169.7 sin(50jtr) A. Find the peak value, the peak-to-peak value, the rms
value o f i{t), and the average value o f
AN SW ER: 169.7, 339.4, 120, and 107.93 A, respectively
3. VO LTAG E
W hat causes current to flow? An analogous question might be. W hat causes water to flow in a pipe
or a hose? W ithout pressure from either a pump or gravity, water in a pipe is still. Pressure from
a water tower, a pressured bug sprayer tank, or a pump on a fire truck will force water flow In
electrical circuits, the pressure that forces electrons to flow, i.e., produces a current in a wire or
a device, is called voltage. Strictly speaking, water flows from a point o f higher pressure say,
p o in ts to a point o f lower pressure say, point 5 along a pipe. Between the two points and
B, there is said to be a pressure drop. In electrical circuits, a voltage drop from point A to point B
11
along a conductor will force current to flow from point A to point B; there is said to be a voltage
drop from point A to point B in such cases.
Gravity forces the water to flow from a higher elevation to a lower elevation. An analogous phe
nomenon occurs in an electric field, as illustrated in Figure 1.10(a). Figure 1.10(a) shows two con
ducting plates separated by a vacuum. O n the top plate is a fixed amount o f positive static charge.
On the bottom plate is an equal amount o f negative static charge. Suppose a small positive charge
were placed between the plates. This small charge would experience a force directed toward the
negatively charged bottom plate. Part o f the force is due to repulsion by the positive charges on
the top plate, and part is due to the attraction by the negative charges on the bottom plate. This
repulsion and attraction marks the presence o f an electric field produced by the opposite sets o f
static charges on the plates.
The electric field indicated in Figure 1.10 sets up an electric pressure or voltage drop from the
top plate to the bottom plate, which forces positive charges to flow downhill in the way that
water flows from a water tower to your faucet. Unlike water flow, negative charges are forced
uphill from the negatively charged bottom plate to the positively charged top plate. As men
tioned in the previous section, this constitutes a net current flow caused by the bilateral flow o f
positive and negative charges. The point is that current flow is induced by an electric pressure
called a voltage drop.
A 0
Positive
negative
charge, q
Electric Field
Force on
Electric Field
charge q
charge
Negative
B charge,-q
B
0
Force on
(a)
(b)
FIGURE 1.10 (a) Positive charge in a (uniform) electric field; (b) negative charge in a uniform elearic field.
As mentioned, in Figure 1.10, the positive charge ^ at ^ tends to move toward B. We say, quali
tatively, that point A in the electric field is at a higher potential than point B. Equivalently, point
5 is at a lower potential than point A. An analogy is now evident: a positive charge in an electric
field falls from a higher potential point to a lower potential point, just as a ball falls from a high
er elevation to a lower elevation in a gravitational field.
Note, however, that if we turn the whole setup o f Figure 1.10(a) upside down, the positive charge
q still moves from point A to point B, an upward spatial movement. Similarly, if a negative charge
- q is placed at B, as in Figure 1.10(b), then the negative charge experiences an upward-pulling
force, moving from the lower potential, point B, to the higher potential, point A.
n
12
-----------------------------------------------------------------------------------------------------------------------
Again, consider Figure 1. 10(a). As the charge q moves from point ^ toward B, it picks up veloci
ty and gains kinetic energy. Just before q hits the bottom plate, the kinetic energy gained equals
the (constant) force acting on q multiplied by the distance traveled in the direction o f the force. The
kinetic energy is proportional to q and to
The missing proportionality constant in this relationship is defined as the potential difference or
voltage between A and B, The term voltage is synonymous with potential difference.
Mathematically,
,
. ,
voltage = potential difference =
energy converted
magnitude of charge
( 1.8)
The standard unit for measuring potential difference or voltage is the volt (V). According to
Equation 1.8, i f 1 joule {]) o f energy is convertedfrom one form to another when moving 1 C o f charge
from point K to point B, then the potential difference, or voltage, between A and B w i VTIn equation
form, with standard units of V, J, and C, we have
1V = 1 ^
(1.9)
The use of terms such as elevation diflFerence, energy converted, potential difference, or
voltage implies that they all have positive values. If the word difference is changed to drop
(or to rise), then potential drop and elevation drop have either positive or negative values, as the
case may be. The following four statements illustrate this point in the context of Figure 1.10:
The voltage between (or across) A a n d 5 is 2 V.
The voltage between (or across) B and A is 2 V.
' The voltage drop from A to B is 2W.
This discussion describes the phenomena of voltage. Voltage causes current flow. But what pro
duces voltage or electric pressure? Voltage can be generated by chemical action, as in batteries. In
a battery, chemical action causes an excess of positive charge to reside at a terminal marked with
a plus sign and an equal amount of negative charge to reside at a terminal marked with a negative
sign. When a device such as a headlight is connected between the terminals, the voltage causes a
current to flow through the headlight, heating up the tiny wire and making it Ught up. Another
source of voltage/current is an electric generator in which mechanical energy used to rotate the
shaft of the generator is converted to electrical energy using properties of electro-magnetic fields.
All types of circuit analysis require knowledge of the potential difference between two points, say
^
^
A and B, and specifically whether point A or point 5 is at a higher potential. To this end, we speak
of the voltage drop from point A to point B, conveniently denoted by a double-subscript, as Vj^.
If the value of
is positive, then point ^ is at a higher potential than point B. On the other
hand, if
is negative, then point 5 is at a higher potential than point A. Since
stands for
the voltage drop from point B to point A,
The double-subscript convention is one o f three methods commonly used to unambiguously specify
a voltage drop. Using this convention requires labeling all points o f interest with letters or integers so
that
KiO ^12
^13
on two points, together with a variable or numerical labeling o f the voltage drop from the point
marked + to the point marked - . Figure 1.11 illustrates this second convention, where Vq denotes the
voltage drop from A (marked +) to B (marked - ) . If Vq is positive, then ^4 is at a higher potential than
B. O n the other hand, if Vq is negative, then 5 is at a higher potential than A. The value o f Vq, togeth
er with the markings + and
this. For a general circuit element, the (+, ) markings that is, the reference directions can be
assigned arbitrarily. A third method for specifying a voltage drop, using a single subscript, will be dis
cussed in Chapter 2.
B
-I-
FIGURE 1.11 The + and - markings establish a reference direction for voltage drop. For accuracy,
always place the (+, - ) markings reasonably close to the circuit element to avoid uncertainty.
The following example illustrates the use o f the double subscript and the (+, - ) markings for des
ignating voltage drops.
E X A M P L E 1.4
Figure 1.12 shows a circuit consisting o f four general circuit elements, with voltage drops as indi
cated. Suppose we know that
= 4 V, and
V^q and
CD-
-I-
3V
FIG U RE 1.12 Arbitrary circuit elements for exploring the use of (+, - ) for specifying a voltage drop.
14
S o lutio n
T he meaiiing o f the double subscript notation and the (+, - ) markings for a voltage imply that
'DA
^ 5 C = 3I V
^CZ> = - ^ Z )C = -(-2 ) = 2 V
and Vp.^-
A N S W E R :- 3 V ; - 2 V
Exercise. T he convention o f the (+, - ) markings is commonly used as described. Figure 1.13 shows
an old 12-V automobile battery whose (+, - ) markings cannot be seen because o f the corrosion o f
the terminals. A digital voltmeter (DVM ) is connected across the terminals, as shown. The display
reads -1 2 V. Figure out the (+, - ) marking o f the battery terminals.
A N SW ER: left terminal,
right terminal, +
DVM
12V
battery
FIG U RE 1.13 Digital voltmeter connected to a 12-V (car) battery whose plus and
minus markings have corroded away.
One final note: As with current, there are different types o f voltages dc voltage, ac voltage, and
general voltage waveforms. Figure 1.9, with the vertical axis relabeled as v{t), illustrates different
voltage types.
4. C IR CU IT ELEM EN TS
Circuits consist o f interconnections o f circuit elements. The most basic circuit element has two
terminals, and is called a two-terminal circuit element, as illustrated in Figure 1.14. A circuit eie-
15
ment called a source provides either voltage, current, or both. The battery is a very common
source, providing nearly constant voltage and the usually small current needed to operate small
electronic devices. Car batteries, for example, are typically 12 volts and can produce large currents
during starting. The wall outlet in a home can be thought o f as a 110 -volt ac source. Figure 1.14(a)
shows a (battery) voltage
across a general undefined circuit element. A current z(r) flows
through the element. Recall from our earlier intuitive discussion that voltage is analogous to water
pressure: pressure causes water to flow through pipes; voltage causes current to flow through cir
cuit elements. Total water into a pipe equals total water out o f the pipe. Analogously, the current
entering a two-terminal device must, by definition, equal the current leaving the two-terminal device.
Current
The capacitor
The inductor
For a resistor, the amount o f current flow depends on a property called resistance; the smaller the
resistance, the larger the current flow for a fixed voltage across the resistor. A small-diameter pipe
offers more resistance to water flow than a large-diameter pipe. Similarly, different types o f con
ductors offer different resistances to current flow. A conductor that is designed to have a specific
resistance is called a resistor. If the device is an ideal resistor, then v(f) = Ri{i), where i? is a con
stant o f resistance. More on this shortly.
The circuit elements called the capacitor and the inductor will be described later in the text. Also,
future chapters will describe the operational amplifier and the transformer that are circuit elements
having more than two terminals.
16
> 0 and
charge moves through circuit element 1 from A to B. In falling from a higher potential, point A, to
a lower potential, point B, the charge loses electric potential energy. The lost potential energy is con
verted within element 1 into some other form o f energy heat or light being two o f several possibil
ities. According to Equation 1.8, the amount o f energy converted {absorbed by the element) is
y. T) >Q. The power absorbedhj element 1 is, by definition, the rate at which it converts or absorbs
energy. This rate equals
^a b (^ab ^ T)
^Vab I a b > 0 .
T
17
W ith respect to Figure 1.15, for constant (direct) voltages and currents, we arrive at a very simple
relationship:
P\-V ab I ab
where
0 -1 0 )
is the power (in W ) absorbed by the circuit element. Consequently, the energy, W , (in
W^=P\xT
(1-11)
Now, let us reconsider Figure 1.15. One can think o f-/ ^ g as flowing from A w B through the
remainder o f the circuit. In this case,
| = ^a ^AB
circuit
element 1. As such, the remainder o f the circuit is said to generate electric energy. By definition,
the electric power generated by the remainder o f the circuit is the rate at which it generates elec
tric energy. From Equation 1.8, this rate equals
--------- ---------- - ^ ab ^ab
Observe that the rate at which the remainder o f the circuit generates power precisely equals the
rate at which circuit element 1 absorbs power. This equality is called the principle o f conserva
tion o f power: total power generated equals total power absorbed. Equivalently, the sum o f the
powers absorbed by all the circuit elements must add to zero,
+ Pq = y^gl^B
^AB^^^AB^ ~
In general, whenever a two-terminal general circuit element is labeled according to the passive sign
convention, as in Figure 1.15, then P =
whereas
P = V^b ^ab ^
is connected. As a general convention, non-source circuit elements are labeled according to the
passive sign convention. Usually, sources are labeled with the current leaving the terminal labeled
with +. For such labeling o f sources, if the product o f the source voltage and the current leaving
the + terminal is positive, then the source is delivering power to the network.
V AB
FIGURE 1.16
Exercise. Compute the power absorbed by each o f the elements in Figure 1.17.
-1A
_____________
-2A
_____________
2A
>
< Z3
10V
10V
10V
(a)
(b)
(c)
FIGURE 1.17
AN SW ER: (a) 10 W; (b) - 2 0 W; (c) 20 W
As mentioned, power is the rate o f change o f work per unit o f time. T he ability to determine the
power absorbed by each circuit element is highly important because using a circuit element or
some device beyond its power-handling capability could damage the device, cause a fire, or result
in a serious disaster. This is why households use circuit breakers to make sure electrical wiring is
not overloaded.
Exercise. In Figure 1.18, a car heater is attached to a 12-volt D C voltage source. How much power
can the car heater absorb before the 20 -amp fuse blows.
20 Amp Fuse
FIGURE 1.18 Car heater connected to a 12-volt car battery through a 20-amp fuse.
A N SW ER: 240 watts
19
As mentioned earlier, the calculated value o f absorbed power P may be negative. If the absorbed
power P is negative, then the circuit element actually generates power or, equivalently, delivers
power to the remainder o f the circuit. In any circuit, some elements will have positive absorbed
powers, whereas some others will have negative absorbed powers. If one adds up the absorbed
powers o f ALL elements, the sum is zero! This is a universal property called conservation o f
power.
The 2"*^ edition o f this text contains a rigorous proof o f this principle. For the present, we will
simply use it to solve various problems. The following example will help clarify the sign conven
tions and illustrate the principle o f conservation o f power.
E X A M P L E 1.5
Light bulbs come in all sorts o f shapes, sizes, and wattages. W a t t l e measures the power consumed
by a bulb. Typical wattages include 15, 25, 40, 60, 75, and 100 W. Power consumptions differ
because the current required to light a higher-wattage (and brighter) bulb is larger for a fixed out
let voltage: a higher-wattage bulb converts more electric energy into light energy. In Figure 1.19,
the source delivers 215 watts o f power. W hat is the wattage o f the unlabeled bulb?
7?
100V
watts
watts
watts
From conservation o f power, the total power delivered by the battery equals the total power
absorbed by all the bulbs. Therefore, the power absorbed by the unknown bulb is
215 - 4 0 - 100 = 75 watts
20
10A
Electroplating
Apparatus
S o lutio n
10 = 2 2 0 0 W, or 2.2 kW
Step 2. According to Equation 1.11, the energy consumed per 12-h period is
2.2
12 = 26.4 kWh
Exercise. Suppose the electroplating apparatus o f Example 1.6 draws 12 A D C at the same volt
age. W hat is the cost o f operation for a single 12-h day? W hat is the cost o f operating for a 20
workday month?
AN SW ER: $3,168; $63.36
E X A M PLE 1.7
Each box in the circuit o f Figure 1.21 is a two-terminal element. Compute the power absorbed by
each circuit element. W hich elements are delivering power? Verify the conservation o f power prin
ciple for this circuit.
21
a)
For element 1 P i = 4 X 1 = 4 W
b)
For element 2 P l = 8 x 2 = 1 6 W
c)
For element 3 ^ 3 = 10 X 1 = 10 W
d)
For element 4
e)
For element 5 P 5 = 2
For element 6 Pe = 1 0 X ( - 2 ) = - 2 0 W
14
x
(-1)=-14W
2 = 4W
Step 2 . Verify conservation o f power. Since P 4 and Pg are negative, element 4 delivers 14 W, and
element 6 delivers 20 W o f power. T he remaining four elements absorb power. Observe that the
sum o f the six absorbed powers, 4 + 16 + 10 - 14 + 4 - 2 0 = 0, as expected from the principle o f
conservation o f power. Equivalently, the total positive generated power, (14 + 20) = 34 W, equals
the total positive absorbed power, (4 + 16 + 10 + 4) = 34 W.
FIG U RE 1.22
AN SW ER: 8 W, 20 W, - 2 8 W; element 3 equivalently delivers 28 W
22
Exercise. In Figure 1.22, suppose the current 2 A were changed to - 4 A. W hat is the new power
absorbed by element 3?
A N SW ER: 56 watts
If the power absorbed by a circuit element is positive, the exact nature o f the element determines the
type o f energy conversion that takes place. For example, a circuit element called a resistor (to be dis
cussed shortly) converts electric energy into heat. If the circuit element is a battery that is being
charged, then electric energy is converted into chemical energy within the battery. If the circuit ele
ment is a dc motor turning a fan, then electrical energy is converted into mechanical energy.
N O N -D C PO W ER A N D EN ER G Y C A LC U LA TIO N S
Consider Figure 1.23, where i{t) is an arbitrary time-varying current entering a general two-ter
minal circuit element, and v{t) is the time-varying voltage across the element. Because voltage and
current are functions o f time, the power p{t) = v{t)i{t) is also a function o f time. For any specific
value o f ^ = ?j, the value p{t^) indicates the power absorbed by the element at that particular
time hence, the terminology instantaneous power for p{t).
i(t)
FIGURE 1.23 Calculation of absorbed power for time-varying voltages and currents for circuit ele
ments labeled with the passive sign convention; here, power is p{t) = v{t)i{t).
Equation 1.12 extends Equation 1.10 in the obvious way.
p{t) = v{t)i{t)
( 1. 12)
i.e., the instantaneous (absorbed) power p{t), in W, is the product o f the voltage v{t), in V, and
the current i{t), in A, with labeling according to the passive sign convention. This product also
makes sense from a dimensional point o f view;
,
volts X amps =
joules
coulombs
joules
; - x
=
coulomb
second
second
Knowing the power p{t) absorbed by a circuit element as a function o f t allows one to compute
the energy W{tQ, t) absorbed by the element during the time interval [^q, t > Iq], W[tQ, t) (J) is the
integral o f p{t) (W) with respect to t over [?q, t], i.e..
23
W(to,t)^ r p ir ) d r
where
the lower limit o f the integral, could possibly be -oo. For the dc case, p{t) = P (a con
t
t
W(tQ,t) = f p ( r ) d T = P f d r = P(t-tQ) = P x T
where T = t - t^, as given in Equation 1.11. If, in Equation 1.13, tg = -oo, then W (-co, t) becomes
a function only o f t which, for convenience, is denoted by
t
W{t)= f p ( r ) d r
L
(1.14)
W{t) = W{00, t), in joules, represents the total energy absorbed by the circuit element from the
beginning o f time to the present time rwhen p{t) is in watts.
Exercise, a) Suppose the power absorbed by a circuit element over [0,oo) is p{i) =
W (0, oo). b) Now suppose the absorbed power o f the circuit element is
p{t) =
j >0
watts. Find
for t > 0 .
A N SW ER: 4 J; (4+t) J
Since energy is the integral o f power, power is the rate o f change (derivative) o f energy.
Differentiating both sides o f Equation 1.14 yields the expected equation for instantaneous power.
dW(t)
v m o = P ( o = ^
(1.15a)
Exercise. Suppose that for t > 0 , the work done by an electronic device satisfies W{t) = 10(1
J- If
the voltage supplied by the device is 10 V, then for t > 0, find the power and current supplied by the
device, assuming standard labeling, i.e., the passive sign convention.
AN SW ER: p{t) = \0e- watts; i(f) = e' A
24
FIGURE 1.24 (a) Current and (b) voltage profdes with respect to t for circuit o f Figure 1.23.
S o lutio n
A simple graphical multiplication o f Figures 1.24(a) and (b) yields the sketch o f the curves in
instantaneous power shown in Figure 1.25(a). From Equation 1.13 with
= 0, we have, for 0 <
t< %
,2
p(T)dr - J c/t =
0
25
(b)
FIGURE 1.25 (a) Profile of the instantaneous power p{t) = v{t)i{t) for the current and voltage wave
forms of Figure 1.24; (b) associated profde of energy versus time.
26
alized in circuit analysis by the ideal voltage source symbol shown in Figure 1.26(a), a circle with
a reference inside. The symbol is more commonly referred to as independent voltage source.
FIGURE 1.26 Equivalent representations of ideal voltage source attached to a hypothetical circuit.
The waveform or signal v{t) in Figure 1.26 represents the voltage produced by the source at each
time t. The plus and minus (+, ), on the source define a reference polarity. T he reference polari
ty is a labeling or reference frame for standardized voltage measurement. T he reference polarity
does not mean that v(t) is positive. Rather, the reference polarity (+, - ) means that the voltage drop
from + to - is v{t), whatever its value/sign. Finally, the voltage source is ideal because it maintains
the given voltage v{t), regardless o f the current drawn from the source by the attached circuit.
voltage (V)
V,
1(A)
(b)
FIGURE 1.27 (a) Ideal battery representation of ideal voltage source; (b) v-i characteristic of ideal battery.
Figure 1.2 7 (a) shows a source symbol for an ideal battery. The voltage drop from the long-dash
side to the short-dash side is Vg, with Vjj > 0. In commercial products, the terminal marked with
a + sign corresponds to the long-dash side o f Figure 1.27(a). An ideal battery produces a constant
voltage under all operating conditions, i.e., regardless o f current drawn from an attached circuit
or circuit element, as indicated by the v-i characteristic o f Figure 1.27(b). Real batteries are not
ideal but approximate the ideal case over a manufacturer-specified range o f current requirements.
Practical sources (i.e., non-ideal); voltage sources, such as commercial dc and ac generators; and
real batteries deviate from the ideal in many respects. One important respect is that the terminal
voltage depends on the current delivered by the source. The most common generators convert
mechanical energy into electrical energy, while batteries convert chemical energy into electrical
27
energy. There are two general battery categories: nonrechargeable and rechargeable. A discussion
o f the dramatically advancing battery technology is beyond the scope o f this text.
Besides batteries and ideal voltage sources, devices called ideal or independent current sources
maintain fixed current waveforms into a circuit, as illustrated in Figure 1.28. T he symbol o f an
ideal current source is a circle with an arrow inside, indicating a reference current direction. An
ideal current source produces and maintains the current i{t) under all operating conditions. O f
course, the current i{t) flowing from the source can be a constant (dc), sinusoidal (ac), or any other
time-varying function.
FIGURE 1.28 Equivalent ideal current sources whose current i{t) is maintained
under all operating conditions o f the circuit.
In nature, lightning is an example o f an approximately ideal current source. W hen lightning
strikes a lightning rod, the path to the ground is almost a short circuit, and very little voltage is
developed between the top o f the rod and the ground. However, if lightning strikes a tree, the path
o f the current to the ground is impeded by the trunk o f the tree. A large voltage then develops
from the top o f the tree to the ground.
Independent sources have conventional labeling, as shown in Figure 1.29, which is different from
that o f the passive sign convention. Here the source delivers power if p{t) = v{t)i{t) > 0 and would
absorb power i f p{t) =
< 0. A complicated circuit called a battery charger can deliver ener
gy to a drained car battery. T he car battery, although usually a source delivering power, exempli
fies a source absorbing power from the charger.
28
Another type o f ideal source is a dependent source. A dependent source or a controlled source
produces a current or voltage that depends on a current through or voltage across some other ele
ment in the circuit. Such sources model real-world devices that are used in real circuits. In the text,
the symbol for a dependent source is a diamond. If a appears inside the diamond, it is a depend
ent voltage source, as illustrated in Figure 1.30. If an arrow appears inside the diamond, it is a
dependent current source, as illustrated in Figure 1.31. In Figure 1.30, the voltage across the dia
mond-shaped source, v{t), depends either on a current, labeled
through some other circuit
device, or on the voltage
across it. If the voltage across the source depends on the voltage v^,
i.e., v{t) = p
then the source is called a voltage-controlled voltage source (VCVS). If the volt
age across the source depends on the current z^, i.e., v{t) =
then the source is called a cur
rent-controlled voltage source (CCVS).
FIGURE 1.30 The right element is a voltage-controlled voltage source (VCVS) if v{t) =
(p is here
dimensionless), or a current-controlled voltage source (CCVS) if v(t) = r^i (r^ here has units of ohm).
a)
If a V CV S (Figure 1.30) with p = 0.4 were associated with the controlled-source branch,
fmd vit).
b)
v{t).
ANSW ER: a) 2 V; b) 1.5 V
There is dual terminology for dependent current sources. The configuration o f Figure 1.31 shows
a voltage-controlled current source (VCCS), i.e., i{t) = g^v^, or a current-controlled current
source (CCCS), for which i{t) =
r~\
29
i(t) =
or
Pi
Q V
or
Pi:
-o
FIG U RE 1.31 The right element is a voltage-controlled current source (VCCS) if i{t) =
(g^ has
units o f siemens) or a current-controlled current source (CCCS) if i{t) = |3/^ ((3 is dimensionless).
Source voltages or currents are called excitations, inputs, or input signals. A constant voltage will nor
mally be denoted by an uppercase letter, such as V, Vq,
cally be denoted by /, /g, /p
and so on. The units are volts, amperes, and so on. Smaller and larger
quantities are expressed by the use o f prefixes, as defined in Standard Engineering Notation Table 1.1.
a)
If a VCCS (Figure 1.31) with ^^ = 0.1 S were associated with the controlled-source branch, find
b)
If a CCCS (Figure 1.31) with P = 0.5 were associated with the controlled-source branch, find i{i).
i{i).
AN SW ER; a) 0.5 A; b) 0.25 A
Name
Prefix
Value
femto
10-15
pico
10-12
nano
10-9
micro
10-6
milli
10-3
kilo
103
mega
lO^
gga
109
tera
1012
30
I
+
R
V
(a)
FIG U RE 1.32 (a) Symbol for a resistor with reference voltage polarity and current direction
consistent with the passive sign convention; (b) resistor connected to an ideal battery.
In 1827, Ohm observed that for a connection like that o f Figure 1.32(b), the direct current
through the conductor/resistor is proportional to the voltage across the conductor/resistor, i.e., I
= V. Inserting a proportionality constant, one can write
or, equivalently,
1 = V GV
^
(1 .1 6a)
V = R1
The proportionality constant R is the resistance o f the conductor in ohms. The resistance R meas
ures the degree to which the device impedes current flow. For conductors/resistors, the ohm (Q)
is the basic unit o f resistance. A two-terminal device has a 1-Q resistance i f a 1-V excitation causes
1-A o f current to flow. In Equation 1.16(a), the proportionality constant is the reciprocal o f R, i.e.,
G = HR, which is called the conductance o f the device. T he unit for conductance according to
the International System o f Units (SI) system is the siemen, S. In the United States, the older term
for the unit o f conductance is the mho ^5, that is, ohm spelled backward, which is still widely
used. In this text, we try to adhere to the SI system. If a device or wire has zero resistance {R = 0)
or infinite conductance {G = t), it is termed a short circuit. On the other hand, if a device or
wire has infinite resistance (zero conductance), it is called an open circuit. Technically speaking,
31
a resistor means a real physical device, with resistance being the essential property o f the device.
In most o f the literature on electronic circuits, resistor and resistance are used synonymously, and
we will continue this practice.
O H M 'S LAW
Ohms law, as observed for constant voltages and currents, is given by Equation 1.16(b), with
its equivalent form in Equation 1.16(a). However, it is true for all time-dependent waveforms
exciting a linear resistor. Thus, we can generalize Equation 1.16 as
v (0 = W
or
(1.17b)
according to Figure 1.33, whose voltage-current labeling is consistent with the passive sign con
vention.
i(t)
AO
+
^
v(t)
----- OB
-
FIG U R E 1.33
If either the voltage or the current direction is reversed, but not both, then Ohms law becomes
v(t) = -Ri{t). As an aid in writing the correct v-i relationship for a resistor. Ohms law is stated
here in words:
For a resistor connected between terminals A and B, the voltage drop from A to B is equal to the
resistance multiplied by the current flowing from A to B through the resistor.
Exercise. Find the resistance R for each o f the resistor configurations in Figure 1.34.
AN SW ER: (a) 12 Q ; (b) 3 Q; (c) 6 Q
-1A
+
R
12V
(a)
4A
-
R
12V
-2A
- -
(b)
R
12V
(c)
FIG U RE 1.34
Once the voltage and the current associated with a resistor are known, the power absorbed by the
resistor is easily calculated. Assuming the passive sign convention, then combining Equation 1.12 for
32
power and Ohms law (Equation 1.17), the instantaneous absorbed power is
9
(1.18a)
Exercise. Find the power absorbed by each o f the resistors in Figure 1.35.
80
+
12V
(a)
4A
-
R
100
(b)
90
-
12V
(c)
FIGURE 1.35
AN SW ER: (a) 18 W; (b) 160 W; (c) 16 W
Equations 1.18(a) and (b) bring out a very important property; a resistor always absorbs power,
dissipating it as heat. Intuitively speaking, electrons that flow through the resistor collide with
other particles along the way. The process resembles the action in a pinball game: the pinball suecessively collides with various pegs as it rolls from a higher to a lower elevation. W ith each colli
sion, part o f the electrons kinetic energy is converted into heat as the voltage pressure continues
to reaccelerate the electron.
Electrical energy that is converted to heat or used to overcome friction is usually called a loss. Such
losses are termed /-squared-i? {f-R) losses because o f the form o f Equation 1.18. On the other
hand, a stoves heating element purposely converts to heat as much electric energy as possible, in
which case, the P-R loss is desirable. This heating effect also proves useful as the basis for the oper
ation o f fuses. A fuse is a short piece o f inexpensive conductor with a very low resistance and a
predetermined current-carrying capacity. When inserted in a circuit, it carries the current o f the
equipment or appliances it must protect. W hen the current rises above the fuse rating, the gener
ated heat melts the conducting metal inside the fuse, opening the circuit and preventing damage
to the more-expensive appliance. Oversized fuses or solid-wire jumpers circumvent safe fuse oper
ation by permitting unsafe operation at overload currents, with consequent electrical damage to
the appliance that may cause overheating and fire.
Resistance o f a conductor depends on the material and its geometrical structure. For a specific
temperature, R is proportional to the length I o f a conductor and inversely proportional to its
cross-sectional area A,
R= p^
(1.19)
33
0.94
Chromium
1.8
Tin
6.7
Copper
1.00
Zinc
3.4
Carbon
2 .4 X 10^
Gold
1.4
Nickel
5.1
Aluminum
1.6
The resistivities o f aluminum and nickel wire relative to copper are 1.6 and 5.1, respectively.
Hence, 100 feet o f aluminum/nickel wire has a resistance o f
(aluminum) 1.6 x 0.4094 = 0.655 Q
(nickel) 5.1 X 0.4094 = 2.088 Q
Given a 10-A current flowing through 100 feet o f copper, aluminum, and nickel wire, Ohms law
implies
(copper) V = /?/ = 0 .4 0 9 4 x 10 = 4 .0 9 4 V
(aluminum) V = RI = 0 .655 x 10 = 6.55 V
(nickel) V = RI = 2 .088 x 10 - 2 0.88 V
Finally, from Equation 1.18(b), the absorbed power given off as heat is
(copper) P = V I ^ R I - = 0 .4 0 9 4 x 100 = 4 0 .9 4 W
(aluminum) p = VI = RI^^ 0 .655 x 100 = 6 5.5 W
(nickel) P = VI = RI~ = 2 .088 x 100 = 208.8 W
Notice that every 100 feet o f 16 AWG aluminum wire would absorb 65.5 - 4 0.9 = 2 4 .6 W more
power than copper. And nickel wire absorbs even more power:
^ ^
208.8
4 0.94
times more power than copper per unit length. This absorbed power, given off as heat, is why
nickel wire is used for heating elements in toasters and ovens.
34
Exercise, (a) If a constant current o f 10 A flows through 1,000 feet o f (16 AWG) copper wire, how
many watts o f heat are generated by the wire?
(b) If the wire o f part (a) were changed to (16 AWG) aluminum, how many watts o f heat would
be generated?
AN SW ER; (a) 409.4 watts; (b) 65 5 .0 4 watts
Temperature also affects resistance. For example, light bulbs have a cold resistance and a hot
resistance o f more importance during lighting. For most metallic conductors, resistance increases
with increasing temperature except carbon, which has a decrease in resistance as temperature
rises. Since resistors absorb power dissipated as heat, they should have adequate physical dimen
sions to better radiate the heat or there must be some external cooling to prevent overheating.
EXA M PLE 1.10
T he hot resistance o f a light bulb is 120 Q. Find the current through and the power absorbed by
the bulb if it is connected across a constant 90-V source, as illustrated in Figure 1.36.
-O 90V
90V
R =120Q
-O -
V
90
/ = - = ----- = 0.75 A
R
120
120 = 67.5 W
Step 3 . C/?eck conservation o f power. T he power delivered by the source is 90 x 0.75 = 67.5 W.
Therefore, the power delivered by the source equals the power absorbed by the resistor. This ver
ifies conservation o f power for the circuit.
Exercise. In Example 1.10, suppose the battery voltage is cut in half to 60 V. W hat is the power
absorbed by the lamp? W hat is the power delivered by the battery? Repeat with the battery volt
age changed to 120 V.
AN SW ER; 30 watts; 120 watts
35
The following example illustrates power consumption for a parallel connection o f light bulbs.
EXA M PLE 1.11
Figure 1.37 shows four automobile halogen Hght bulbs connected in parallel across a 12-V bat
tery. Find the following;
(a) The effective hot resistance o f each bulb
(b) T he total power delivered by the source
(c) After 700 hours o f operation, the current supplied by the source drops to 11.417 A.
Discover which light bulb has burned out.
ijt )
27 watts
35 watts
50 watts
60 watts
12V
12^
Rxiw -
27
144
^50W =
50
= 5.33D
= 2 .88 Q
144
^ 3 5 W = ^ - 4 .1 1 4 Q
144
- ' 60
= 2.4 Q
(b) The power delivered by the source equals the sum o f the powers consumed by each bulb,
which is 172 W.
(c) Since the current supplied by the source has dropped to 11.417 A, then the power delivered
by the source drops to P;urcenew ~
lier-delivered power o f 172 watts. Hence, the 35-watt bulb has gone dark.
Exercise. Repeat Example 1.11 (a) with the battery voltage changed to 48 V and a new set o f light
bulbs whose operating voltage is 48 V.
AN SW ER;
85.333 Q; R^c,^= 65.83 Q; R^q^ = 4 6 .0 8 Q; R(^^^= 38.4 Q..
E X A M PLE 1.12
W hen connected to a 120-volt source, halogen light bulb number 1 uses 40 watts o f power. W hen
similarly connected, halogen light bulb 2 uses 60 watts o f power.
(a)
(b)
36
(c)
b u ib l
120V
120 V ^ = i-
bulb 2
(a)
(b)
FIGURE 1.38 Series connection o f two light bulbs and equivalent resistive circuit model.
S o lu tio n
Step 1. Find the hot resistances. The hot resistances o f each bulb are given by
120^
Vt
40
bulb\
= 360 and R 2 =
120 "
= 240
Step 2. Find the current through each bulb, the power absorbed by each bulb, and the power delivered
by the source. The circuit o f Figure 1.38(a) has the equivalent representation in terms o f resistanc
es in Figure 1.38(b). By definition, in a nvo-terminal circuit element, the current entering each
resistor equals the current leaving. Therefore, the current through each resistor in the series con
nection is the same, and is denoted /. So the new power dissipated by each bulb/resistor is
^l,new
To calculate these values, we need to know I. By conservation o f power, the power delivered by
the source is the sum o f the absorbed powers, i.e.,
^ sou rce ~
X I P\^new
^ 1,new ~
+ -^2^
;==o.2A
600
Hence,
= 360 X 0 .2 ^ = 14.4 W ,
= 24 W
37
Although involved, the solution o f this problem uses the definition o f a two-terminal circuit ele
ment and conservation o f power to arrive at the result in a roundabout way. In Chapter 2, we can
more directly arrive at the answers by using Kirchhoffs voltage and current laws.
A potential problem with series connections o f light bulbs is circuit failure. If one bulb burns out,
i.e., the filament in the bulb open-circuits, then all other lights are extinguished. Parallel circuits
continue to operate in the presence o f open-circuit failures and are easier to fix: only the unlit bulb
must be replaced.
v(V )
= 0, the voltage source looks like a short circuit because the current flow, generated
by the remaining circuit, will induce no voltage across the source. For now, we must be content
with this brief discussion. Chapter 2 will reiterate and expand on these ideas.
38
+
V ( + )
V-
/ Circuit
-I
(a)
(b)
= 0, the current
source looks like an open circuit because no current will flow, regardless o f any voltage generated
by the rest o f the circuit. Again, we must be content with this brief discussion until Chapter 2 reit
erates and expands on the ideas.
+
Y
/ Circuit /
(a)
FIG U RE 1.41 (a) Constant source
(b)
attached to circuit;
9. SUM M ARY
Building on a simplified physics o f charge (coulombs), electric fields, and charge movement, this
chapter set forth the notions o f current, i{t) or / for dc, and voltage, v{t) or V for constant volt
ages. A rigorous treatment would require field theory and quantum electronics. More specifically,
the notions o f current, current direction, voltage, and voltage polarity, a two-terminal circuit ele
ment (the current entering equals the current leaving), the passive sign convention, power con
sumption [pit) = v{t)i{t) assuming the passive sign convention], and dissipated energy (the inte
gral o f power) were all defined. In general, we can say that every circuit element does one o f the
following:
Absorbs energy
Stores energy
Delivers energy, or
Converts energy from one form to another
39
T he chapter subsequently introduced ideal independent and dependent voltage and current
sources: the voltage-controlled voltage source (VCVS), the current-controlled voltage source
(CC VS), the voltage-controlled current source (VCCS), and the current-controlled current source
(C C C S). A dependent source produces a voltage or current proportional to a voltage across or a
current through some other element o f the circuit. The various types o f dependent sources are
summarized in Table 1.4.
TABLE 1.4 Summary of the Four Possible Dependent Sources.
VCVS
(Voltage-Controlled
Voltage Source, p is dimensionless)
ccvs
(Current-Controlled
Current Source,
is in
ohms)
V CC S
(Voltage-Controlled
Voltage Source,
is in S)
CCCS
(Current-Controlled
Current Source, P is
dimensionless)
-I-
40
The chapter keynoted a special two-terminal element, called a resistor, whose terminal voltage and
current satisfied Ohms law, v(t) = Ri{t), where v{t) is the voltage in volts, R is the resistance in
ohms, and i{t) is the current in amperes. The resistor, as defined in this chapter, is a passive ele
ment, meaning that it always absorbs power,/>{;) = v{t)i{t) = i?-{t)IR = Rp-{t) > 0 since R>Q. This
absorbed power is dissipated as heat. Hence, the (passive) resistor models the heating elements in
a stove or toaster oven quite well. In addition, the resistor models the hot resistance o f a light bulb.
Throughout the text, the resistor will often represent a fixed electrical load. In a later chapter, we
will discover that it is possible to construct a device with a negative resistance, R<Q, which can
generate power. However, such a device is rather complex to build and requires such things as the
operational amplifier covered in Chapter 4.
The various quantities defined and used throughout the chapter have various units. The quanti
ties and their units are summarized as follows:
TABLE 1.5 Summary of Units
Charge
Current
Coulomb
Ampere (A)
(C)
Voltage
Resistance
Conductance
Volt (V)
Ohm (Q)
S (Siemens)
mhof3
Power
w a tt
= volt X
am p
Energy
Joule (J)
Throughout this chapter, a number o f examples illustrated the various concepts that were intro
duced. Some simple resisrive circuits were analyzed. To analyze more complex circuits, one needs
Kirchhoffs voltage and current laws, which specify how circuit elements interact in a complex cir
cuit. These basic laws o f circuit theory are set forth in the next chapter.
41
'w '
Dependent (controlled) voltage source: a voltage source whose output voltage depends on the
voltage or current of some other element in the circuit.
Independent (ideal) voltage source: an ideal device whose terminal voltage is a prescribed func
tion of time, e.g., {2 cos{t) + 12}V, no matter what current goes through the device.
Instantaneous power: the value of p{t) =
at a particular time instant.
Insulator: a material that opposes easy electron movement.
Mho: historical unit of conductance equal to the reciprocal of an ohm.
Ohm: unit of resistance. One ohm equals the ratio of IV to lA.
Ohms law: for a linear conductor, the current through the conductor at any time t is proportional
to the voltage across the conductor at the same time.
Resistivity: the resistance of a conductor is proportional to its length and inversely proportional
to its cross-sectional area. The proportionality constant p is called the resistivity of the
material. The resistivity of copper at 2 0 ^C is 1.7 x 10~^ ohm-meters.
Resistor: physical device that obeys Ohms law. There are commercially available nonlinear resis
tors that do not obey Ohms law. Resistors convert electric energy into heat.
Root mean square (rms) or eflfective value: measure of ac current, which is related to the peak
value by the formula rms = 0.7071if, where K sin(o)^ + (|)) is the ac waveform.
Short circuit: connection of zero resistance or infinite conductance.
Siemens: unit of conductance (formerly, mho) or inverse ohms.
v-i characteristic: graphical or functional representation of a memoryless circuit element.
Voltage (potential difference): positive charge, without obstruction, will move from a higher
potential point to a lower potential point, accompanied by a conversion of energy.
Voltage is measured in volts; 1 volt between two points A and B means that the energy
converted when moving 1 Coulomb of charge between A and B is 1 joule.
42
PROBLEMS
C H A R G E A N D C U R R EN T
PRO BLEM S
(c)
(d)
10 ^^ electrons.
If this number o f electrons moves uni
formly from the left end o f a wire to
the right in 1 ms (milli second), what
current flows through the wire?
How many electrons must pass a given
point in 1 minute to produce a current
o f 10 Amperes?
If the charge profile across the cross-sec
tion o f a conductor from left: to riglit is
given by q{t) = t+ 0.2e'5^- 0.2 C for t >
0 and zero for ? < 0 , plot the profile o f the
current that flows across the botmdary. In
(c)
(d)
is
(i)
(ii)
(a)
(b)
4. (a)
(b)
43
(b)
5. (a)
(b)
i(t) (A)
2--
-t (secs)
1- -
Figure P I.7
-
2- -
Figure P I.5
6 sec, in Coulombs.
i (amps)
0.1
-t (sec)
10
-0.1--
t-*-t (sec)
Figure P I .6
Figure P I .8
44
9. (a)
12.(a)
2k
T = -
(b)
(0
5A
+ 3V -
2V -
+ 5V -
-cz>
4A
+
4V
il'
7A
3A
0 '
Figure P I. 12
(i)
If element A generates 28
power, find Va (ii) Find the power absorbed by
ment B.
(iii) I f element C generates 6
power, find Vq
(iv) I f element D absorbs 2 7
power, find //).
6 V
- 4 V
W
ele
W
W
(b)
2A
____________
2A
(a)
___________
20 V
lO V
(i)
(ii)
__ .
3A
2A
C _ 3
20 V
(iii)
Figure PI. 11
A N SW ERS (b): (i) -40 W; (ii) 20 W; (iii) 60 W
6V
1A
3A
lOV
circuit
element 1
circuit
6V element 2
16v Q ) 2 A
lOV
Figure P I. 13
45
14. (a)
+ 80V -
-25A
P 1.14a below.
Show that the algebraic sum o f the
absorbed powers is zero. Be careftil o f sign.
(c)
=2 +
for t
> 0.
9A
Figure P I. 15
14A|
circuit
15V
element 1
4A
5A ( ^ ^ 2 0 V
lOA
+ 10V-
circuit
element 2
5V
lOV
- 5V +
5A
(a)
circuit
element 2
-v,(t) +
2A
Figure P I. 14a
30
CHECK: Re
(c)
4V
(b)
Figure P 1.14b
15. In the circuit o f Figure P I. 15, there are
three independent sources and five ordinary
resistors.
(a)
Determine which o f the circuit ele
(b)
25V
ti
Figure P I. 16
C H E C K : (b) 125 W, 1 2 5 tJ
17. Suppose energy cost in Indiana is 10 cents
per kwh.
(a)
How much does it cost to run a 100watt T V set 8 hours per day for 30
days?
(b)
How many 100-watt light bulbs run
for 6 hours a day are needed to use
$9.00 o f energy every 30 days?
A N SW ER: (a) 8 cents per day; $2.40 per
month; (b) 5 bulbs
46
RESISTA N CE
18. Using Equation 1.19 and Table 1.2 , find
the resistance o f a nickel ribbon having these
dimensions:
length:
width:
thickness:
19. (a)
(b)
(c)
40 m
1.5 cm
0.1 cm
Compute the resistance o f800 feet o f 14gauge copper wire (2.575 Q /1000 ft).
Repeat (a) for 200 feet o f 14-gauge
nickel wire.
If one end o f the copper wire is soldered
to one end o f the nickel wire, find the
total resistance o f the 1000 feet o f wire.
Can you justify your answer?
Figure P I.21
22. For the circuit o f Figure P 1.22,
= 1Q , and
the
instantaneous
0.5 s.
(b)
= 2 seconds?
(c)
Figure P I.22
23. The switch S in Figure P I .23 is assumed to
be ideal, i.e., it behaves as a short circuit when
closed, and as an open circuit when open.
Suppose the switch is repeatedly closed for 1 ms
and opened for 1ms.
(a) W hat is the average value o f z(^)?
(b) W hat is the average power delivered
by the source?
Kt)>'
Figure P I.23
A N SW ERS: 0.25 mA. 1.25m W
47
(a)
CH ECK:
= 1.875 m W
Figure P I.25
C H E C K : Average current is 1.8 mA.
26 (a)
93.75 Q
= 62.5 Q ,
(a)
(b)
^2Q ^
(a)
(b)
(a)
(b)
Figure P I.27
48
20
+ V, -
3n
40
''" 6
60 <
>50
Figure P I . 3 0
C H E C K : (a)
Figure P I.3 2 b .
(c)
(a)
(b)
Ko = 2 0 V
(d)
watts
(c)
deliv
12V
Figure PI.32a
12V
Figure PI.32b
200
Figure P1.31.
C H E C K : /q = 6 A.
32. (a)
to the load?
W hat current does the battery supply
to the load?
49
(c)
Figure PI.35b
D EP EN D EN T SO U RC E PRO BLEM S
36. Consider the circuit in Figure P I .36.
(a)
-A /V ^
R,
If 1/ = 6 V, find
I'.
(b)
Figure P I.34
C H EC K : 1 A, 20 Q or ?????
R, = 20n
(b)
(c)
given
that
the
unknown
Figure P I.36
C H EC K : (a)
solution for I J
(b) 1/ = 4 V
= 60 V,
= 3.6 watts;
Figure P1.37
38. Consider the circuit o f Figure P I .38.
(a)
Determine an expression for
F igure P I . 3 5 a
and
V
G y = ^
Vin terms o f R^, R2 , a , and V-^^.
(b)
50
(c)
= 4.
Given your answer to (b), determine
the power gain, which is the ratio of
the power delivered to
divided by
= 1 0 V.
(a)
(b)
Gv =
Gp =
^out
^
Pin
(c)
(d)
Figure P I.38
- 100 mA,
50 Q,
10 Q, and
100 Q.
(a)
(b)
put current.
Find the current gain,
G/ =
0.1V.
0.21
v,^
Rb
Gy =
(c)
10
G p = ^
P-
Findthepowerabsorbedbyeachresistor.
^ 3=
= 1000
Vin
2001
1000
2kO
2kn
Figure P I.39
40. In problem 39, suppose R^= 1 k Q, 7?2=
C ty
CHECK; 4 . 1 mA
C H EC K : i?,= 5 k Q
C H A P
KirchhofFs Current & Voltage
Laws and Series-Parallel Resistive Circuits
A CAR HEATER FAN SPEED -CO N TRO L APPLICATION
One use o f resistors in electronic circuits is to control current flow, just as dams control water flow
along rivers. Ohms law, V = RI, gauges the ability o f resistors to control this current flow: for a
fixed voltage, high values o f resistance lead to small currents, whereas low values o f resistance lead
to higher currents. This property underlies the adjustment o f the blower (fan) speed for ventila
tion in a typical car, as represented in the following diagram.
In this diagram, three resistors are connected in series, and their connecting points are attached to
a switch. As we will learn in this chapter, the resistance o f a series connection is the sum o f the
resistances. So with the switch in the low position, the 12-V car battery sees three resistors in series
with the motor. Th e series connection o f three resistors represents a large resistance and heavily
restricts the current through the motor. W ith less current, there is less power, and the fan motor
speed is slow. W hen the switch moves to the Med-1 position, a resistor is bypassed, producing less
52
Chapter 2 K irchhofFs Current & Voltage Laws and Series-Parallel Resistive Circuits
resistance in the series circuit and allowing more current to flow. More current flow increases the
fan motor speed. Each successive switch position removes resistance from the circuit, and the fan
motor speed increases accordingly.
Analysis o f such practical circuits builds on the principles set forth in this chapter.
CHAPTER O U TLIN E
1.
2.
3.
4.
5.
6.
7.
8.
9.
CHAPTER OBJECTIVES
Define and utilize Kirchhoffs current law (KCL), which governs the distribution o f cur
rents into or out o f a node.
Define and utilize Kirchhoffs voltage law (KVL), which governs the distribution o f volt
ages in a circuit.
Introduce series and parallel resistive circuits.
Develop a voltage division formula that specifies how voltages distribute across series con
nections o f resistors.
Develop a current division formula that specifies how currents distribute through a par
allel connection o f resistors.
'
Show that a series connection o f resistors has an equivalent resistance equal to the sum o f
the resistances in the series connection.
Show that a parallel connection o f resistors has an equivalent conductance equal to the
sum o f the conductances in the parallel connection.
Explore the calculation o f the equivalent resistance/conductance o f a series-parallel con
nection o f resistances, i.e., a circuit having a mixed connection o f series and parallel con
nections o f resistors.
Explore the calculation o f voltages, currents, and power in a series-parallel connection o f
resistances.
Revisit the notion o f a dependent source and use a V C C S to model an amplifier circuit.
Describe a practical battery source and look at a general practical source model.
Chapter 2 K irchh offs Current & Voltage Laws and Series-Parallel Resistive Circuits
53
FIG U RE 2.1
must equal the exit current. Similarly, in a parallel connection, such as Figure 2.2b, the same volt
age, in this case, Vj^, appears across every circuit element.
+
node
d-
0
-1-
node 1
Vr
-
node 2
(a)
(b)
FIG U RE 2.2 (a) Series connection of resistors with the property that each resistor carries
the same current; (b) parallel connection of resistors with the property that
the same voltage appears across each resistor.
Sources interconnected with circuit elements produce currents through the elements and voltages
across the elements. For example, a voltage source connected across Figure 2.2a would generate a
current 2^ and the voltages
tion o f voltages around loops o f circuit elements, as shown in Figure 2.2a. Similarly, a current
source connected across the circuit o f Figure 2.2b would produce the voltage
and the currents
Z] through
K irchhoffs current law (KCL) governs the flow o f currents into and out o f a com-
54
Chapter 2 KirchhofPs Current & Voltage Laws and Series-Parallel Resistive Circuits
mon connection point or node, as in the top and bottom connections o f Figure 2.2b. This chap
ter sets forth precise statements o f these laws and illustrates their application.
A proper statement o f KVL and KCL requires the additional notion o f branch. A branch o f a cir
cuit is a generic name for a two-terminal circuit element and is denoted by a line segment, as in
Figure 2.3. T he endpoints o f a branch (the terminals o f the circuit element) are called nodes, as
in Figure 2.3a. Ordinarily, however, node means a common connection point o f two or more cir
cuit elements (branches), as shown in Figure 2.3b.
node A
(a)
(b)
FIG U RE 2.3 (a) Single branch representing a circuit element with terminals labeled as
nodes A and B; (b) interconnection of branches (circuit elements) with common
connection points labeled as nodes A through D.
The voltage polarity and current direction for the branches in Figures 2.2 and 2.3 are labeled in accor
dance with the passive sign convention; the arrowhead on a branch denotes the reference current direc
tion, which is from plus to minus. Recall that the + to - does not mean that the voltage is always posi
tive if measured from the plus-sign to the minus-sign. In general, reference directions can be assigned
arbitrarily. The conventional assignment o f voltage polarity and current direction to voltage and current
soiurces is given in Figure 2.4, which is different from the passive sign convention. Note that with these
conventional assignments, the (instantaneous) power delivered by a source is/^/^) =
power absorbed by a source is
Circuit
Circuit
' 'J O
(a)
th^
= -pjeff)-
(b)
FIG U RE 2.4 Conventional labeling o f (a) voltage, and (b) current sources.
Chapter 2 K irchh offs Current & Voltage Laws and Series-Parallel Resistive Circuits
55
time.
S tatem ent 2 : Equivalently, the algebraic sum o f the currents leaving a node is zero for every
instant o f time.
The two statements o f KCL are equivalent because the negative o f the sum o f the currents enter
ing a node corresponds to the sum o f the currents leaving the node. Further, from physics we know
that charge is neither created nor destroyed. Thus, the charge transported into the node must equal
the charge leaving the node because charge cannot accumulate at a node. KCL expresses the con
servation o f charge law in terms o f branch currents. Moreover, KCL specifies how branch currents
interact at a node, regardless o f the type o f element connected to the node.
Referring to Figure 2.3b, KCL at nodey4 requires that i^{t) +
- i^ii) = 0 for all t. KCL at node
Finally, KCL at node D requires that 25(f) =
B requires that
E X A M P L E 2.1
For the node shown in Figure 2.5, find
56
Chapter 2 K irchh offs Current & Voltage Laws and Series-Parallel Resistive Circuits
S o lutio n
By KCL, the sum o f the currents entering the node must be zero. Hence, the current z^(^) =
9cos(2r) - 3cos(2?) - cos(2r) - 2cos(2z) = 3cos(2^) A.
Exercise. 1. Suppose the current through the voltage source in Figure 2.5 is changed to - 2 cos(2?).
Find
AN SW ER: - 4 cos(2^) A.
2. Three branches connect at a node. All branch currents have reference directions leaving the
node. If /j = /2 = 2 A, then find ly
A N SW ER; - 4 A
Two implications o f KCL are o f immediate interest. First, as a general rule, KCL forbids the series
connection o f current sources. Figure 2.6a shows an invalid connection o f two arbitrary current
sources i,(?) and i^i), where z,(?)
the other hand, a parallel connection o f two current sources can be combined to form an equiva
lent source, as in Figure 2.6b, where
= i^{t) + i2 {t).
(a)
-O
(b)
ijii)- Avoid this violation o f KCL; (b) equivalent representation o f a parallel connection
o f rwo current sources in which
= i]{t) +
A second immediate consequence o f KCL is that a current source supplying zero current [i{t) = 0
in Figure 2.7] is equivalent to an open circuit because the current through an open circuit is zero.
An open circuit has infinite resistance, or zero conductance. This means that a current source has
infinite internal resistance. From another angle, a constant current source is represented by a ver
tical line in the iv plane (see Figure 1.4 lb ). The slope o f the vertical line, which is infinite, deter
mines the internal resistance o f the source.
Chapter 2 K irchh offs Current & Voltage Laws and Series-Parallel Resistive Circuits
57
-O
+
v(t)
-O
FIG U RE 2.7 Ideal current source with i{t) = 0 is an open circuit.
A typical application o f KCL is given in the following example.
E X A M P L E 2.2
In the parallel resistive circuit o f Figure 2.8, the voltage across each resistor is 6 cos(z) V. Find the
current through each resistor and the current,
'm
6cos(t) V
IQ
<20
<3Q
S o lu tio n
By Ohm s law,
/^ l ( 0 = 6 c o s ( 0 A
6 c o s (/)
'R l W -------- = 3 c o s ( 0
, 3 , - 5 i 2 5 W = 2 c o s ( ,) A
By KCL,
Exercise. 1. In Figure 2.8, suppose the source voltage is changed to a constant, labeled
Iin in terms o f V-^.
AN SW ER:
Find
58
Chapter 2 KirchhofFs Current & Voltage Laws and Series-Parallel Resistive Circuits
2. Suppose the source voltage in the circuit o f Figure 2.8 were changed to - 1 2 cos(2?) V. Find
AN SW ER: - 2 2 cos(2?) A
Kirchhoffs current law holds for closed curves or surfaces, called Gaussian curves or surfaces. A
Gaussian curve or surface is a closed curve (such as a circle in a plane) or a closed surface (such as
a sphere or ellipsoid in three dimensions). A Gaussian curve or surface has a well-defined inside and
outside. Figure 2.9 illustrates the idea o f a Gaussian curve for three (planar) situations.
.............
Two
Terminal
Circuit
Elem ent /
i,(t)
(a)
(b)
i2 {t), which is precisely the definition o f a two-terminal circuit element. For the three-terminal
device o f Figure 2.9b, KCL for Gaussian curves implies that i^{t) = i^{t) + i2 {t). Finally, for Figure
2.9c, i^ i^ + if^ = 0. From these illustrations, one might imagine that the use o f Gaussian surfaces
might simplify or provide a short cut to certain branch current computations. T he general state
ment o f KCL for Gaussian surfaces is next followed by an example that demonstrates its use for
computing branch currents.
59
Chapter 2 K irchh offs Current & Voltage Laws and Series-Parallel Resistive Circuits
E X A M P L E 2 .3
This example shows how the use o f a Gaussian curve or surface can sometimes simplify a calcula
tion. Figure 2.10 portrays a complicated circuit whose branch currents and voltages are not solv
able by methods learned so far. Our objective is to find the current
FIG U RE 2.10 Circuit for Example 2.3, showing a Gaussian surface to compute
directly.
S o lutio n
Using KCL for the indicated Gaussian curve, - 1 .1 5 + / ^ - 0.3 + 0.95 = 0. Equivalently, /^ = 1.15
+ 0.3 - 0.95 = 0.5 A.
In the next chapter, circuits such as the one in Figure 2 .10 are analyzed using a technique called
nodal analysis.
Exercise. 1. Draw a Gaussian surface on the circuit in Figure 2.10 that is different from the sur
face given but still allows one to compute /^.
AN SW ER; One choice is a circle enclosing the bottom node.
60
Chapter 2 KirchhofFs Current & Voltage Laws and Series-Parallel Resistive Circuits
2. Draw an appropriate Gaussian curve to find / in the graphical circuit representation in Figure
2 . 11.
AN SW ER: 2 A
voltage.
61
Chapter 2 K irchhofFs Current & Voltage Laws and Series-Parallel Resistive Circuits
V.
BD
FIG U RE 2.12 Circuit diagram illustrating (i) three closed paths {A-B-C-D-E-A)\
(ii) the concept of node voltages with respect to a given reference node E,
(iii) the concept of branch voltages
and Vj^-,
and
The concept o f a closed path and the concept o f a node voltage allow us to state our first two ver
sions o f Kirchhoffs voltage law.
instant o f time.
Statem ent 2 : For any pair o f nodes j and k, the voltage drop Vjj^ from node j to node k is
given by
at every instant o f time, where Vj is the voltage at node j with respect to the reference and
is the voltage at node ^ with respect to reference. Herey and k stand for arbitrary node indices.
For example, in Figure 2.12, j, k can be any o f the nodes A, B, C, D, or E.
62
Chapter 2 KirchhofFs Current & Voltage Laws and Series-Parallel Resistive Circuits
Exercise. 1. Find Vy^g, VgQ and V^^-for the circuit o f Figure 2.1 3 in which we have introduced
the ground symbol at node E to identify the reference node.
AN SW ERS:
= - 3 V, Vg(^=2\ V,
= 18 V
2. Again, with reference to Figure 2.13, find the node voltages V^, Vg, Vq and
AN SW ERS: 2 V, 5 V, - 1 6 V, - 6 V
3. In Figure 2.13, suppose the branch labeled 6 V is now labeled - 1 2 V. Find
AN SW ER: - 3 V
FIG U RE 2.13
A third concept needed for two further equivalent statements o f KVL is that o f a closed node
sequence. A closed node sequence is a finite sequence o f nodes that begins and ends at the same
node. A closed node sequence generalizes the notion o f a closed path. Finally, we define the notion
o f a connected circuit. In a connected circuit, each node can be reached from any other node by
some path through the circuit elements. Figures 2 .12 and 2.14 show connected circuits. However,
in Figure 2.14, the sequence o f nodes A-B-C-D-E-A is a closed node sequence but not a closed
path because there is no circuit element between nodes B and C.
+ 2.5V -
+ 10V D
+
V,
F IG U R E 2 .1 4 Simple dependent source circuit for illustrating the concepts o f a connected circuit and a closed node sequence.
63
Chapter 2 KirchhofFs Current & Voltage Laws and Series-Parallel Resistive Circuits
age drop
GP
Vab +
Referring to Figure 2.14, for the closed node sequence E-A-B-E, V-^ = 10 =
+ Vg = 2.5 + Vg and Vg = 7.5 V. Now, consider the closed node sequence E-C-D-E. For this
sequence,
^CD
= 20 V. Finally,
consider the closed node sequence, E-B-C-E, which is not a closed path because there is no cir
cuit element between nodes B and C. Nevertheless, by statement 4 o f KVL, - Vg+ Vb + ^ c ~ ^
or equivalently that
= Vg -
= 7.5 - 30 = - 22.5 V.
cos( 1 2 0 tc?),
Vg^ =
18
= 20 V. Find Vg, Vq
cos( 1 2 0 ti?)
and
3 2 cos(1207t^). Find
at ? =
0 .5 s.
(c) Find
when v^ = 1 0 0 V,
= - 1 0 V and
25V.
SC RA M BLED A N SW ER: 85 V, - 1 3 V, - 7 0 V
FIG U RE 2.15 Circuit with nodes labeled A through E. Node E is taken as the reference node.
64
Chapter 2 K irchhoff s Current & Voltage Laws and Series-Parallel Resistive Circuits
Two further implications o f the KVL are o f immediate interest. First, as a general rule, KVL for
bids the parallel connection o f two voltage sources say,
and V2 (^) for which Vj(?)
as illustrated in Figure 2.16a. O n the other hand, two voltage sources in series can be combined
to form a single source, as illustrated in Figure 2.16b, where
FIGURE 2.16 (a) An improper connecdon of voltage sources when v^{t) ^ i>2 (i);
(b) an equivalent representation of two voltage sources connected in series in which
= Vj{i) +
-O
-o
+
ov
ov
Chapter 2 K irchh offs Current & Voltage Laws and Series-Parallel Resistive Circuits
65
wire, that gives o ff an intense light when hot. In a circuits perspective, the filament acts as a resis
tor and has an equivalent hot resistance. The series connection o f bulbs can be modeled by a series
connection o f resistors, with each resistor paired with a specific bulb. Computing the voltage
across each light (a very important type o f calculation) would then be equivalent to finding the
voltage across each o f the resistors in the equivalent circuit model. It is quite common to model
electrical loads, such as a light, by resistors.
E X A M P L E 2 .4
Figure 2.18a shows a voltage source v-JJ) connected to three resistors in series. The objectives o f
this example are to compute
the
ijt )
(a)
(b)
FIGURE 2.18 (a) Three series resistors connected across a voltage source. By the
definition o f a two-terminal resistor or by the KCL, the current through each resistor is
(b) equivalent resistance
So l u t io n
Step 1. Express the voltage across each resistor in terms o f the input current. For the circuit o f Figure
2.18a, the current through each resistor is i^JJ) by KCL. From Ohms law, the voltage across each
resistor is
fo t j = 1, 2, 3.
Step 2. Express v-J^t) in terms ofi-JJ), solvefor i-^<^t), and then compute an expression for v^t) in terms
o f the Rj and v-J^t). By KVL, the source voltage equals the sum o f the resistor voltages, i.e..
(2 . 1)
where we have substituted Rjii{t) = Vj{t). Dividing Equation 2.1 by (i?j + Rj + -^3) yields
66
Chapter 2 K irchh offs Current & Voltage Laws and Series-Parallel Resistive Circuits
V j{t) =
= --- ---------5
,2 2^)
Equation 2.2a is a volt^e division formula for a three-resistor series circuit. This formula imphes
that if a resistance R. is small relative to the other resistances in the series circuit, then only a small
portion o f the source voltage develops across it. O n the other hand, if a resistance R. is large rela
tive to the other resistances, then a larger portion o f the source voltage will develop across it. One
concludes that the voltage distributes around a loop o f resistors in proportion to the value o f each
resistance. The proportion is simply the ratio o f the branch resistance R- to the total series resist
ance.
Step 3. Compute the equivalent resistance R^^ seen by the voltage source. The equivalent resistance
seen by the voltage source for a resistive circuit is implicidy defined by Ohm s law, i.e., Vi^i) =
^eqhri^^nonzero currents, the equivalent resistance is defined as
Figure 2.18b illustrates the idea o f the equivalence. By Equation 2.1, v^^i) = R^,^ij{t) = (7?^ +
+
implies that the equivalent resistance is R^^ = R-^ + R^ + Ry This means that from the
perspective o f the voltage source, the series connection o f resistors is equivalent to a single resistor
o f value equal to the sum o f the resistances. A formal discussion o f equivalent resistance and its
generalization (the Thevenin resistance) is taken up in Chapter 6 .
and
Example 2.4 suggests some generalizations. Consider Figure 2.19. The first is that the equivalent
resistance R^^ seen by the source is the sum o f the resistors. This means that resistances in series add,
i.e., resistors in series can be combined into a single resistor whose resistance is the sum o f the indi
vidual resistances.
Req = R \ + R 2+ "' + Rn
Further, since vi^t) = Rjii{t), a general voltage division formula can be derived as
Ri
R\ +
+ ' + Rn
(2 .2 b)
fory = 1, ... , n.
67
Chapter 2 K irchh offs Current & Voltage Laws and Series-Parallel Resistive Circuits
Exercise. In Figure 2.19, suppose each resistor has value Rq. Find the equivalent resistance seen by
the source and the voltage across each resistor in terms o f the source voltages.
A N SW ER: nR^^, v jt)ln
So l u t io n
From the preceding discussion, R^^ is defined by Ohms law, i.e., v-J^t) = Rg^i-Js)Step 1. Express v-^ in terms o f the remaining branch voltages. From KVL,
(2.3)
68
Chapter 2 K irchh offs Current & Voltage Laws and Series-Parallel Resistive Circuits
and ^2 in
terms of i^^, observe that i^^ is the current through each resistor (KCL or definition o f two-terminal circuit
element) and use Ohms law:
Therefore,
20// = R e q iin
= 20
Notice that the dependent source increases the resistance o f the two series resistors by 4 Q.
Dependent sources can increase or decrease the resistance o f the circuit. W ith dependent sources,
it is even possible to make the equivalent resistance negative.
Step 3. Find the power absorbed by the 14-Q. resistor. To find the power absorbed by the l4-2 resis
tor when
= 2 V, first compute i^^ via Ohm s law; i^^ = v J R
=
= 0 .0 1 x 1 4 = 0.14 W.
Exercise. Suppose the dependent source in the circuit o f Figure 2 .20 has its value changed to 2[v-^
+ V2 ). Find R^q.
AN SW ER: 48 Q
-o
ijt )
-I-
- f-
v jt )
ijt )
v Jt)
-O
(a)
(b)
Chapter 2 KirchhofFs Current & Voltage Laws and Series-Parallel Resistive Circuits
69
So l u t io n
Step 1. Find expressions for ij{t) in terms o f v-J^t). The variable that Hnks the branch current ij{t) to
the input current i^J^t) is the voltage
which by KVL appears across each resistor. Since v-J^t)
appears across each o f the resistors, Ohms law implies that each resistor current is
ij{t) = ^
^
(2.4)
(j\ + Cj2
R]
Rn
^eq
^3
R^
(2.5)
Step 3. Compute ijyi) in terms ofi-J^t). To obtain a relationship between i-J,t) and ij{t), substitute
Equation 2.5 into Equation 2.4 to obtain
1
G,
+
+
Rl R2
R3
G, ,
G 1 + G 2 + G3
(2 .6)
Equation 2.6 is called a current division formula. It says that currents distribute through the
branches o f a parallel resistive circuit in proportion to the conductance o f the particular branch
G. relative to the total conductance o f the circuit G^^ = G^ + G2 + Gy The greater the conduc
tance, i.e., the smaller the resistance, the larger the proportion o f current flow through the associ
ated branch.
Step 4. Compute the equivalent resistance
R.
1
+
R,
R2
R3
G 1 + G2 + G 3
G ,,
70
Chapter 2 KirchhofFs Current & Voltage Laws and Series-Parallel Resistive Circuits
= 1 Q,
= 0-5
and
hn ={G\ + G2 + G 3 )v, = G Vj
in addition to implying that G
o f the parallel circuit, one can further interpret this to mean that conductances in parallel add to
form equivalent conductances. This parallels the property that resistors in series add to form equiv
alent resistances. O n the other hand, resistances in parallel do not add, and conductances in series do
not add. We can conclude that from the perspective o f the source, the parallel circuit o f Figure
2.21a has the equivalent representations given in Figure 2.21b.
These ideas generalize to n resistors in parallel, as illustrated in Figure 2.22. In particular, the
equivalent resistance R
Ren = - 1
R,
R2
R.,
(2.7)
^eq =
+ *^2 + +
equivalent conductance. Further, the current through each
branch satisfies the general current division formula
R,
G/
1 1
+ ---- -!- +
R
^1
^2
rj A
A-j-+ rG
G
+. r
G2 +
G, .
~ Gr^ q
(2 .8 )
o
-Iijt )
( f
V,
(t)
-o
FIG U RE 2.22 Parallel connection o f n resistors driven by current source.
Chapter 2 K irchh offs Current & Voltage Laws and Series-Parallel Resistive Circuits
Exercise. Consider Figure 2.22. Suppose ten 10 Q resistors are in parallel. Find
71
E X A M P L E 2 .7
Consider the circuit o f Figure 2.23 exhibiting a current source driving two parallel resistors. Show
that
Ry
-O
FIG U RE 2.23 Two resistors in parallel driven by a current source.
So l u t io n
Step 1. Find the equivalent resistance seen by the current source. From Equation 2.7, with = 2, it
follows that
R^R2
Ri ^ Ri
This formula, called the product over sum rule, is quite useful in many calculations.
Step 2. Find i^{t) and i^it)- From Equation 2.8, with w = 2, it follows that
R^
Gi
X 7X
Ry
and
/2 (0 =
Gi 4- G 2
Ri
,__ Ri + R'^
R2
72
Chapter 2 K irchh offs Current & Voltage Laws and Series-Parallel Resistive Circuits
= 12 A and
Find Rj so that
= 10
= 4 A.
AN SW ER: /?, = 20 Q
E X A M P L E 2 .8
the current ijit) through i?2> and the
5e
^ 0
t< 0
y r i,(t)
^2
ijt )
U i3 (t)
SG 3
l|i,( t )
<G ,
So l u t io n
Step 1. Compute the equivalent conductance and equivalent resistance o f the circuit. Since conduc
tances in parallel add,
^ e q - G j + G 2 + G 3 + G 4 = 0.25 S
and
Step 2 . Compute v^JJ). From Ohms law, the voltage across the current source is
2 0 e~ 'V t^ 0
0
t< 0
Step 3. Compute the current i2 (t). Using the current division formula o f Equation 2.8 yields
3e~'A t > 0
Cjeq
0.25
r<0
73
Chapter 2 K irchhoffs Current & Voltage Laws and Series-Parallel Resistive Circuits
Step 4. Compute the power absorbed by Rj- To compute the power absorbed by
P 2 (0 = v,(O x i2 {t) = 1'2 (0 r ^2 =
for t >0,
Exercises. 1. For the circuit o f Figure 2.24, find i/^(t) and the power absorbed by i?4 .
A N SW ER: 0.6^-^ A, 12
2. In the circuit o f Figure 2.24, suppose each conductance is doubled and i-J^t) = 100 mA. Find
the voltages
V2 , the
power absorbed by the 6 kQ resistor, and the power delivered by the source for the circuit o f
Figure 2.25, when
= 20 mA.
74
Chapter 2 KirchhofFs Current & Voltage Laws and Series-Parallel Resistive Circuits
So l u t io n
Step 1. Compute
To compute
8 x 4.8
Kq\ = ------=
1+ 4 .8
38.4
= 3 kQ
12.e
and
^^^2
1 ^ 6
12
R, = IkQ. /!R,. =
2 +6
= 1.5 kQ
= Reqhn = 20 X 1.5 = 30 V
Step 3. Compute Vj and V^- By voltage division.
'
Req-i
= >5 V and Vj = - ^ v ; = ?
resistor.
^ (V 2 )^ ^ 1 0 0
6000
6000
1
60
^source ~ ^inhn 30 x 0 .0 2 = 0 .6 W
Exercise. 1. W hat is the current through the 2 k2 resistor firom top to bottom?
AN SW ER: 50 mA
2 . In Example 2.9, suppose the resistance o f each resistor is doubled. Find the new R^ and the
power delivered by the source.
AN SW ER: 3 k n , 1.2 watts
This example points out a very interesting fact: finding the equivalent resistance o f a series-paral
lel connection o f resistors requires only two types o f arithmetic operations no matter the network
complexity: adding two numbers and taking the reciprocal o f a number. A hand calculator easily
executes both operations. Such is not the case with a non-series-parallel network. To find the
equivalent resistance o f a non-series-parallel network, one usually must write simultaneous equa
tions and evaluate determinants, a topic detailed in Chapter 3 .
Chapter 2 KirchhofFs Current & Voltage Laws and Series-Parallel Resistive Circuits
75
It is then important to recognize when a problem belongs to the series-parallel category in order to take
advantage o f the simple arithmetic operations. In the previous series-parallel examples, one and only
one independent source was specified on the circuit diagram. This is part o f the definition o f a seriesparallel network. The independent source must be indicated, or, equivalendy, the pair o f input termi
nals to which the source is connected must be specified. The specification o f the input terminals deter
mines whether or not a network is series-parallel. The following example illustrates the effect of differ
ent input terminal designations on the computation o f equivalent resistance.
E X A M P L E 2 .1 0
For the circuit o f Figure 2.26a, determine whether or not the network is series-parallel as seen
from each o f the following terminal pairs:
1.
2.
C a s e l;( A , B)
Case 2; (A, C)
3.
Case 3: (C, D)
If the answer is affirmative, give an expression and compute the numerical value for the equivalent
resistance, using the notation // (double slash) for combining resistances in parallel, i.e.,
and
are
is in parallel with
means
(a)
>R,
A<
R1
D'
(0
FIG U RE 2.26 (a) From terminals (C, D) the network is not series-parallel. However, from terminals
{A, B) the network is a series-parallel one. (b) Redrawing of the network of (a) as seen from terminals
(A, Q; the resulting network is series-parallel.(c) Non-series-parallel network seen from (C, D).
76
Chapter 2 K irchh offs Current & Voltage Laws and Series-Parallel Resistive Circuits
So l u t io n
Case 1. Find equivalent resistance seen at (A, B). W ith an independent source connected to nodes
A and B, the source sees a series-parallel network. By inspection o f Figure 2.26a, the equivalent
resistance is
i?eq = ^ i //[(^2 +
^
I
Case 2. Find equivalent resistance seen at (A, Q. W ith (A, Q as the input terminal pair, the net
work is again series-parallel. This is made apparent by redrawing the network, as shown in Figure
2.26b, from which
i?eq =
+ [(i ?4 +
Case 3. Find equivalent resistance seen at (C, U). W ith (C, D) as the input terminal pair, the net
work is not series-parallel, as can be garnered from Figure 2.26c. T he calculation o f
for this
is changed to 40 Q. Find
In electrical engineering laboratories, a student often uses a meter to measure voltages associated
with a piece o f electronic equipment. In older laboratories, or when using an inexpensive meter,
the voltage reading will sometimes differ from what the student calculated or expected to meas
ure. Typically, this results from the loading effect o f the meter. Using the concept o f series-paral
lel resistances, the following example explores the phenomenon o f loading.
E X A M P L E 2 .il
Suppose the circuit in Figure 2.27a is part o f a laboratory experiment to verify voltage division. In
this experiment, you calculate the expected voltage Vq and then measure the circuit voltage using
an inexpensive voltmeter.
(a)
(b)
V q.You
this range, the meter is represented by a 10-kQ resistance, i.e., 10 kD = full-scale reading
meter sensitivity = 10 V x 1 kQ/V. W hat voltage will the meter read?
(c)
^0
^ 0 -10-V scale. This better-quality meter is represented by a 2 0 0 -k tl resist
ance. W hat new voltage will the meter read?
77
Chapter 2 K irchh offs Current & Voltage Laws and Series-Parallel Resistive Circuits
" h
10 kQ
r-O
+
15V^
10 kO
20 ko
f-O
-1-
IS v X
>10 kO
20 kO
lOkQ
r-O-----1-
20 kQ
-o
FIG U RE 2.27 Three circuits for exploring the effect of loading on a circuit: (a) circuit for validating
voltage division; (b) circuit o f (a) with an attached voltmeter having an internal resistance o f 10 k2;
(c) circuit o f (a) with an attached voltmeter having an internal resistance of 200 k^2.
So l u t io n
(a) Voltage division on the circuit o f Figure 2.27a yields
20
Vo =
(b)
-15 = 10 V
O n the 0-10-V range, the voltmeter internal resistance between the probes is 10 kD, as
stated. This represents a 1G-Id2 load connected in parallel with the 20-kQ resistance, as
shown in Figure 2.27b. The voltage Kg will now change because the 15-V source no
longer sees 10 ld2 in series with 20 kO. Rather, the source sees 10 kQ in series with 6 .67
k ii = 20 kX2//10 kQ. By voltage division,
6 .67
-15 = 6 V
Vo = 10 -H6 .67
This is a 4 0 % deviation from the true answer, V q = 10 V, as calculated in part (a).
(c)
O n the 0 -10-V range with the better voltmeter, the internal resistance between the probes
is 200 kD. As before, this represents a 200-kQ load connected in parallel with the 2 0 -k 0
resistance, as shown in Figure 2.27c. 20 kQ//200 klQ = 18.18 kO. By voltage division,
this yields
18.18
Vb =
-15 = 9 .6 7 7 V
10 + 18.18
78
Chapter 2 KirchhofFs Current & Voltage Laws and Series-Parallel Resistive Circuits
2. Th e circuit o f Figure 2.28 shows a voltage divider whose voltage Kq is to be measured by a volt
meter having an internal resistance o f 80 kO. Find Kq without the meter attached, and then find
the value o f Vq measured by the meter.
20 V
1.
2.
(VCVS)
(VCCS)
3.
4.
An arrow inside the diamond indicates a controlled current source having the reference current
direction given by the arrow. A inside the diamond specifies a controlled voltage source, with
the reference voltage polarity given by the sign. A parameter value completes the specification
o f a linear controlled source. In Figure 2.2 9 the (constant) parameters are fx, g^, r^, and |3. These
symbols are common to many electronic circuit texts and have useful physical interpretations to
practicing engineers and technicians. For consistency, a ^^-type controlled source is a V C C S and
a jO,-type source is a V C V S, and so on.
79
Chapter 2 K irchh offs Current & Voltage Laws and Series-Parallel Resistive Circuits
O
+
a ................... o
(a) VCVS or |j-type
O
Pi,
>r
6-
(c )C C V S o rr -type
<;
o - ......... .........o
(d) CCCS or P-type
in Figure 2.29c and d, the controlling current Zj is shown to flow through a short circuit. (Strictly
speaking, neither an open circuit nor a short circuit is a circuit element.) In a real circuit, the cur
rent may be flowing through an actual circuit element, such as a resistor or even a source.
In Figure 2.29b, once the controlling voltage v-^ is known, the right-hand source behaves as an
independent current source o f value
is called a trans
fer conductance, or transconductance. The other controlled sources have a similar interpretation.
The parameter
has the unit o f resistance, ohms, and is called a transfer resistance. The param
80
Chapter 2 K irchh offs Current & Voltage Laws and Series-Parallel Resistive Circuits
vcvs
VCCS
ccvs
cccs
Unit
Appellation
dimensionless
Voltage gain
Parameter
siemens
Sm
Transfer
conductance
ohm
Transfer resistance
dimensionless
Current gain
Figure 2.29 portrays each controlled source as a four-terminal device. In practical circuits, the
great majority o f controlled sources have one terminal or node in common, making them threeterminal devices. The dashed lines joining the two bottom nodes in Figure 2 .29 suggest this quite
common configuration.
The controlled sources as defined in Figure 2.29 have linear v-i relationships. Controlled sources
may also have a nonlinear v-i relationship. In such cases, the element will be called a nonlinear
controlled source. This text deals only with linear controlled sources.
The next few examples describe some o f the unique features o f controlled sources.
FIG U RE 2.30
AN SW ER: 4 V, 0.05 A, 1.6 W, 0.05 W
E X A M P L E 2 .1 2
This example analyzes the circuit o f Figure 2.31. The independent voltage source in series with
the 3-Q. resistor represents a practical source discussed at greater length later in this chapter. The
circuit within the box o f Figure 2.30 approximates a simplified amplifier circuit by a V C C S. The
8 -Q resistor is considered a load and could, for example, model a loudspeaker. Two important
quantities o f an amplifier circuit are voltage gain and power gain, which are computed here along
with various other quantities.
(a)
(b)
Chapter 2 K irchh offs Current & Voltage Laws and Series-Parallel Resistive Circuits
(c)
Compute /out'
(d)
Compute
(e)
(f)
(g)
(h)
Compute
81
(i)
FIG U RE 2.31 Practical source (ideal independent voltage source in series with a resistor)
driving a simplified VCCS approximation of an amplifier circuit loaded by an 8 -Q resistor.
So l u t io n
(a)
(b)
(c)
By Ohms law, / =
= 0.8 mA.
To compute
one must first compute K j. Here one can use Ohms law directly, since
= 3 + 47 = 50 Q.
we know /-^, or one can use voltage division. Doing the calculation by voltage division,
V = 4 0 X 10'^ = 3 7 .6 X 10'^ V
50
Using this value o f
0.125 + 0 .0 6 2 5
(d)
(e)
8 = 16V
=
= 400
0.0 4
82
Chapter 2 K irchhofFs Current & Voltage Laws and Series-Parallel Resistive Circuits
(f)
p.^ =
(g)
= 47
(h)
(i)
Pout is simply the product o f voltage and current delivered to the load
The resulting amplifier power gain is the ratio o f the power absorbed by the 8 -Q load to
the power delivered to the amplifier, P-^,
Pin
= 1.064 X 10^
3 0 ,0 8
Exercise. Suppose the 8-f2 load resistor in Figure 2.30 is changed to 16 Q. Compute
and the power gain.
AN SW ER: 1.5 A, 24 V, 1.197 X 10<^
The analysis in Example 2.12 required only KCL, KVL, and simple voltage divider and/or cur
rent divider formulas. More complicated linear circuits necessitate a more systematic approach. To
see this need, add a resistor between the top o f the 47-Q resistor and the top o f the dependent
current source in Figure 2.31. The methods o f solution used in the example immediately break
down because the circuit is no longer series-parallel; hence, one cannot use voltage division to
compute V j. Chapter 3 will explain more systematic methods called nodal and loop analysis.
Unlike a passive element such as a resistor, which always dissipates power as heat, a controlled
source may generate power as computed in part (g) o f Example 2 . 12, or may dissipate power in
other cases. Since a controlled source has the potential o f generating power, it is called an active
element.
In Example 2.12, the practical voltage source delivers 30.08 pW o f power to the circuit, which is
easy to accept because the source could have been a small battery On the other hand, the con
trolled source generates 48 W. This seems a litde puzzling. Where does the power come from?
W hy not purchase a controlled source at a local electronics store and use it to power, say, a lamp?
Here it is important to recognize that a controlled source is not a stand-alone component picked
o ff the shelf like a resistor. A controlled source is usually constructed from one or more semicon
ductor devices and requires a dc power supply for its operation. The power delivered by the con
trolled source actually comes from the power supply. Here, we use the controlled source to math
ematically model an amplifier and facilitate analysis o f the circuit.
Chapter 2 K irchhoff s Current & Voltage Laws and Series-Parallel Resistive Circuits
83
W ith simple series-parallel connections o f resistors, the equivalent resistance is always positive.
When controlled sources are present, a strange result may happen, as illustrated in the next exam
ple.
E X A M P L E 2 .1 3
Find the equivalent resistance
= 2.
for the circuit o f Figure 2.32 when (a) p = 0.5 and (b) p
O^
FIG U RE 2.32 Calculation of
So l u t io n
W ith p unspecified, we can apply KVL to the single loop, noting that
Consequently, (1 - p)V^ =
= V^. Here,
and
i-x
For p = 0.5, R^q = 2R, which means that the dependent source acts like a resistor o f R Q.. In this
case, it absorbs power. O n the other hand, for p = 2, R,^ = -R , a negative equivalent resistance. In
this case, the dependent source acts like a -2R-Q. resistor and, in fact, delivers power to the inde
pendent source. An important conclusion can be drawn from this example: in the study o f linear
circuit analysis, controlled sources allow the possibility o f negative resistances. Since a negative
resistance generates power, it is also an active element.
Exercise. In Figure 2.32, find the values o f p so that R^q - 0 .5 R and R^q - 2R.
A N S W E R :-!, 0.5
84
Chapter 2 K irchh offs Current & Voltage Laws and Series-Parallel Resistive Circuits
for the following three values of^^: 0.5 mS, 1 mS, and 2 mS.
FIG U RE 2.33
AN SW ER: 2 kl, open circuit, -1 kQ
>
Vs
(a)
(b)
(c)
Chapter 2 K irchh offs Current & Voltage Laws and Series-Parallel Resistive Circuits
85
E X A M P L E 2 .1 4
This example shows the effect o f the internal resistance o f a battery on the terminal voltage.
Suppose a nickel-cadmium battery has an open circuit terminal voltage o f 6 volts. W hen con
nected across a 2-Q. resistor, the voltage drops to 5.97 V. Find the internal resistance o f the bat
tery.
So l u t io n
Figure 2.35 illustrates the situation. Here, the dashed box represents the battery model with inter
nal resistance R^. In Figure 2.35a, no load is connected to the battery. Hence, no current flows
through the internal resistance, in which case, the terminal battery voltage is 6 V.
(a)
(b)
FIG U RE 2.35 Battery model with internal resistance; (a) open circuited (Is = 0); and
(b) connected to a 2 -0 load.
Figure 2.35b shows the battery connected to the 2-Q resistive load. The measured voltage is 5.97
V. By KVL, the voltage across the internal resistance,
law, the current through
is
= 0 .0 1 0 0 5 Q
Exercise. In Example 2.14, suppose the internal resistance is known to be R^ = 0.005 Q and
although the load resistance is unknown, the load current is 4 A. W hat is the voltage across the
load resistance, and what is the load resistance?
AN SW ER; 5.98 V and 1.495 ^
Chapter 2 KirchhofFs Current & Voltage Laws and Series-Parallel Resistive Circuits
86
be other effects. However, for our purposes, a more realistic model o f a voltage source contains a
series internal resistance, as illustrated in Figure 2.36a.
L .(t)
(b)
FIG U RE 2.36 (a) A non-ideal voltage source as an ideal voltage source with an internal series resist
ance; (b) a non-ideal current source as an ideal current source with a parallel internal resistance.
Ideal current sources have infinite internal resistance. Real current sources have a finite, typically
large, internal resistance. Figure 2.36b depicts a more realistic current source model where the
internal resistance is in parallel with the ideal current source.
In the case o f constant voltage and current sources, ideal and non-ideal source models have a
graphical interpretation. The i-v (current-voltage) characteristic o f an ideal constant voltage
source {v^{t) = 1^) is a horizontal straight line. This means that the voltage supplied by the source
is fixed for all possible current loads. An ideal constant current source (z^(z) =
has a vertical
straight line characteristic, which means that the current is constant for all possible voltages across
the source. Figure 2.3 7 illustrates these relationships graphically.
Vout
(a)
FIG U RE 2.37 v-i characteristics o f (a) an ideal constant voltage source, and
(b) an ideal constant current source.
Chapter 2 KirchhofFs Current & Voltage Laws and Series-Parallel Resistive Circuits
87
The non-ideal case is quite different. Because o f the internal resistance a non-ideal constant volt
age source i-v characteristic satisfies the linear relationship
^out
^s^out
Ks'
(2 . 10)
= l/R^,
hut ^s'^oul
( 2 . 11 )
Equations 2.10 and 2.11 are illustrated by the graphs in Figure 2.38 when v^{t) =
ideal voltage source and i^{t) =
for the non-ideal current source. For a voltage source, if the value
o f R; is very small in comparison with potential load resistances, as ordinarily expected, then the
hne in Figure 2.38a approximates a horizontal line, the ideal case. O n the other hand, for a cur
rent source, the line in Figure 2.38b approximates a vertical line whenever
er than a potential load resistance. This would then approximate the ideal current source case.
FIGURE 2.38 v-i characteristics of (a) non-ideal constant voltage source, and (b) non-ideal current source.
In a similar way, non-ideal dependent voltage sources are a connection o f an ideal dependent
source with a series resistance. A non-ideal dependent current source is a connection o f an ideal
dependent current source with a parallel resistance.
E X A M PLE 2 .1 6 Figure 2.39 shows the measured voltages o f a dc power supply found in an old
laboratory. Assuming a non-ideal model o f Figure 2.38a, find
the power supply.
88
Chapter 2 K irch h offs Current & Voltage Laws and Series-Parallel Resistive Circuits
Vout (V)
' '
FIG U RE 2.39 Graph of measured voltages and currents for a dc power supply.
So l u t io n
= -R^
= 0,
SUM M ARY
This chapter has presented the essential building blocks o f linear lumped circuit theory, beginning
with the two fundamental laws for interconnected circuit elements: KVL and KCL. KVL states
that for lumped circuits, the algebraic sum o f the voltages around any closed node sequence o f a
circuit is zero. Similarly, KCL says that for lumped circuits, the algebraic sum o f the currents enter
ing (or leaving) a node is zero.
These laws in conjunction with Ohms law allowed us to develop a voltage division and a current
division formula. The voltage division formula applies to series-resistive circuits driven by a volt
age source. The voltage developed across each resistor was found to be proportional to the resist
ance o f the particular element relative to the equivalent resistance seen by the source. For exam
ple, in a two-resistor series circuit, Rj in series with
we found that
Vi =
7^1 + /?2
T he current division formula applies to parallel-resistive circuits driven by a current source. Here,
the current through each resistor with conductance Gi was found to be proportional to G/ divid
ed by the equivalent conductance seen by the source. Since conductance is the reciprocal o f resist
ance, the idea can also be expressed in terms o f the resistances o f the circuit. For example, in a tworesistor parallel circuit, Rj is parallel with /?2>
/i =
G,
G 1 + G2
R, +Rn
In deriving the voltage division formula, we learned that the resistances o f a series connection o f
resistors may be added together to obtain an equivalent resistance, prompting the phrase resistors
in series add. Analogously, the derivation o f the current division formula for parallel circuits led
Chapter 2 Kirchhofifs Current & Voltage Laws and Series-Parallel Resistive Circuits
US
89
to conclude that a parallel connection o f resistors has an equivalent conductance equal to the
sum o f conductances. This is sometimes expressed in terms o f resistances as the inverse o f the sum
o f reciprocal, i.e., for n resistors in parallel,
p .
7^1
which leads to the very special formula for two resistors in parallel,
= G^ + ... + G^.
Dependent (controlled) current source: a current source whose output current depends on the
voltage or current o f some other element in the circuit.
Dependent (controlled) voltage source: a voltage source whose output voltage depends on the
voltage or current o f some other element in the circuit.
Gaussian surface: a closed curve in the plane or a closed surface in three dimensions. A Gaussian
surface has a well-defined inside and outside.
Kirchhoff'* current law (KCL): the algebraic sum o f the currents entering a node o f a circuit
consisting o f lumped elements is zero for every instant o f time. In general, for lumped
circuits, the algebraic sum o f the currents entering (leaving) a Gaussian surface is zero at
every instant o f time.
90
Chapter 2 KirchhoflFs Current & Volt:e Laws and Series-Parallel Resistive Circuits
KirchhoF$ voltage law (KVL); for lumped circuits, the algebraic sum of the voltage drops
around any closed path in a network is zero at every instant of time. In general, for
lumped connected circuits, the algebraic sum of the node-to-node voltages for any closed
node sequence is zero for every instant of time.
Node: the common connection point between each element; in general, a node is a connection
point of one or more circuit elements.
Node voltage: the voltage drop from a given node to the reference node.
Parallel circuit: a side-by-side connection of two-terminal circuit elements whose top terminals
are wired together and whose bottom terminals are wired together.
o
n
n
n
91
Chapter 2 K irchh off s Current & Voltage Laws and Series-Parallel Resistive Circuits
PROBLEMS
KIR C H H O FF'S C U R R EN T LAW
1. (a)
Figure P2.3
1^ = 3 A, and
(b)
= 4 A.
Repeat part (a) when l 2 = I^ = 1^ = 2 A.
A N SW ER: ( b ) 4 A
4. (a)
(b)
in the circuit o f
Figure P2.1
A N SW ERS: (b) 0, 2 A
2.
(a)
Suppose
(b)
Figure P2.4
A N SW ER: (a) 6 A
100 mA.
5. (a)
(b)
Find Kj and
unspecified
represents
an
40V
lOOV
Figure P2.2
+ V, -
3. (a)
V,
+
VI
Reference Node
5A
Reference Node
(a)
lO O v C ^
+
40V
(b)
Figure P2.5
(c)
6 . (a)
92
Chapter 2 K irchh offs Current & Voltage Laws and Series-Parallel Resistive Circuits
(b)
Vg, Vq and
reference
indicated
node
by
the
ground symbol.
(c)
Compute
50 V
K C L A N D KVL
9. (a)
Figure P2.6
7. (a)
(b)
and V^.
circuit element.
(b)
K l ^B
respect
to the reference node indicated by
in
iUV
6^
4A
30V
3A
10
(c)
Compute
and
(a)
(b)
Figure P2.9
AN SW ERS: ( b ) - 1 7 V 14 V
10.
8 . (a)
(b)
93
Chapter 2 KirchhofFs Current & Voltage Laws and Series-Parallel Resistive Circuits
11.
(b)
a = 0.25, determine
delivered to Rj^.
Figure P2.14
A N SW ERS: (b) 80 Q, 20 W
15.
which
25V
(a)
(b)
T he power absorbed by
(c)
watts.
The power absorbed by R^ is 13.44
13.44 watts.
Figure P 2 .ll
Find the currents and voltages /^, V^, 1^,
12.
and
is 13.44
watts.
Figure P2.15
SC RA M BLED A N SW ERS: 21, 16, 336, 56
Figure P2.12
If /? = 5 Q, find Vjf
Find the value o f R when
pose
= 40 V.
120V
4on
50V
20V
Figure P2.16
+
V.
AN SW ER: (b) 25 Q
200
,4on
Figure P2.13
C H E C K : (b) P^= 160 watts
I4.(a)
the voltage V2 ,
the power delivered by each source.
94
Chapter 2 KirchhofFs Current & Voltage Laws and Series-Parallel Resistive Circuits
R
>J .
10 k n ,
6kO
Figure P 2.17
AN SW ER: (c) 12.5 watts and - 7 .5 watts
'+
V,
(a)
Figure P2.20
in the cir
21.
500
50 V
6x 10"
ANSWERS: (b) Vi =
3V
'in
4 8 V 4 kQ
i 6xl 0' -i-/?-i-6axl0
The circuit o f in Figure P2.21 is a blower
0.8 A
<----
Figure P2.18
19. (a)
(b)
Ignition
Switch
JUA
and
12V
Chassis .
Ground
44 V
'r
3R
10V
(a)
source.
Figure P2.19
C H EC K : (b) K^,= 4 V
20. (a)
(i)
mW. Compute
(b)
, V-^, and R.
= 40 V.
in terms o f a and R. If
95
Chapter 2 KirchhofFs Current & Voltage Laws and Series-Parallel Resistive Circuits
(b)
determine:
(i)
C H EC K : 10 ohms, 40 ohms
(a)
(b)
(d)
value o f R.
(i)
(t)
^^^lOOV
= 0 .2 5 Q ,
1.25 a
Figure P2.24
(ii)
C H E C K : (b) 8 </?< 15
W
(b)
(iii) 26%
(i)K^5=0> ^s c = 3 -4 3 V ,
1-43
= 7.14 V
(ii) 5.71 A
(iii) 59.5%
(c)
(i)
=0.
150 Q
2o on
Vcn = 2V,
Vmotor = 10 V
300n'
lo o n
(ii) B A
(d)
(iii) 83.3%
0.4 Q
Figure P2.25
26. In the circuit o f Figure P2.26, suppose V2 =
60 V. Find
the source.
6on'
'' 6
4o n
9o n
1 800
Figure P2.26
SCRA M BLED AN SW ERS: 3 360, 840, 4
96
Chapter 2 Kirchhofifs Current & Voltage Laws and Series-Parallel Resistive Circuits
lOon
500
0.8 A
20V
C H EC K : (a) 45 <
(c) 8 0 < i? ^ < 125
200
Figure P2.27
SC RA M BLED A N SW ERS: 59.5, 9, 8, 49.
0.5, 45, 18
(a)
28.
(b)
in terms o f a
and 4
For the circuit o f Figure P2.28 with the
(c)
Currents /j through
Voltages Vj through
independent
source?
How
much
current source
500 0
2000
Figure P2.30
C H EC K : 4 = 0.05 A
Figure P2.28
C H E C K : ^2 = 12 V, /^ = 1 mA,
= 60m W
V O LTA G E A N D C U R R EN T
D IVISIO N
31. Consider the circuit o f Figure P2.31 in
(a)
which
(b)
Voltage drop V2
(c)
(a)
ance, R
= 30 V and
= 20 V. Find
and
(b)
Switch
closed.
(c)
Switch
closed.
97
Chapter 2 K irchh offs Current & Voltage Laws and Series-Parallel Resistive Circuits
(a)
= 120 V
Suppose
Figure P2.31
= 100 V and
1/2 = 80 V.
32. Construct a series voltage divider circuit
wiiose total resistance is 2400 Q as illustrated
in Figure 2.32.
(a)
(b)
^s'
(c)
6on
R2 , and R^.
Figure P2.34
C H EC K ; (b) 60 Q, 240 Q
Figure P2.32
Rc + Rci
33. For the circuit o f Figure P 2.33, suppose
= 48 V.
(a)
Find
(b)
about
Find
If
(c)
terminals. Then
6Q
(d)
Figure P2.33
C H E C K : (b) 25 <
< 30 V
AN SW ER: (d) 3 V
=6
98
Chapter 2 KirchhofFs Current & Voltage Laws and Series-Parallel Resistive Circuits
lokn
60kn
Figure P2.39
lokn
C H EC K :
= 6.4 watts
Figure P2.36
SC RA M BLED A N S W E R S :- 5 , 10, 15
37. Consider the circuit o f Figure P 2.37 in
-o
i',
(b)
4kn
'
Kq' h
6kn
9kn
J 8 kO
-O Figure P2.40
C H E C K : V.^ = 360 V
41. Find /[ and I 2 for the circuit o f figure P2.41
when
= 10 mA, 1^^ = ^
and
= 14 mA.
Figure P2.37
SC RA M BLED AN SW ERS: (a) 50, 0 . 5 ,
5, 50, 20
' 7
r '0f
>
Y
30mS>
Figure P2.41
42. For the circuit o f Figure P2.42, find the cur
rents /p /2, ly and
-O V
'
when
= 300 mA.
j"
3oon> 6 o o n /l
I >i 2on
-o Figure P2.38
i 2on >4on
CH ECK:
= 40/, and
in
= 10 mA.
Figure P2.42
SC RA M BLED AN SW ERS: - 5 0 , 40, 80 (in
mA)
N.
Chapter 2 KirchhofFs Current & Voltage Laws and Series-Parallel Resistive Circuits
99
3000
3000
5000,
" V
5000 .
1.5kO<
J 7500
i6kn
lkO<
Ik0<
Figure P2.45
(b)
SC RA M BLED AN SW ERS: (b) 29.63, 675,
0.1185
46.
Find
15kn
Figure P2.43
A N SW ERS: 0.5R , 5 kQ, 2.6 kQ
44. Find the value o f R-^ for each o f the
circuits o f Figure P2.44.
O-
Figure P2.46
SC RA M BLED A N SW ERS: 60 kn ,
1.5R
(b)
the power
when
80 V.
C H E C K : Answers are the same.
100
Chapter 2 KirchhofFs Current & Voltage Laws and Series-Parallel Resistive Circuits
>5000
2kO
(e)
Figure P2.49
statements is true?
> K ql
(a)
(b)
Kill = ^eq2
(d)
There
is no
general
relationship
(b)
Suppose
= 320 V,
= 256 V, R^ =
R^ = 800 n . Find R,
and the
resulting R^^.
Suppose
= 320 V, V^= 192 V, R^ =
400 Q, R^ = 800 Q. Find R,
and
the resulting R
C ircuit 1
Figure P2.50
SC RA M BLED AN SW ERS: 500, 1000, 1600,
400, 170.67, 128
C ircuit 2
Figure P2.48
and C, respectively.
49. For each o f the circuits o f Figure P2.49,
find the value o f R that makes
= 1000 Q,.
(b)
O-
5000
>3kO
O-
7500
O-
be
1 .2 kn
52Sn
o-
(c)
(a)
O-
(c)
(d)
Chapter 2 K irchhoff s Current & Voltage Laws and Series-Parallel Resistive Circuits
(a)
101
(b)
Figure P2.52.
(a)
Find the equivalent resistance
=1Q
and
= 100 Q. This type of
problem is useful for represent
ing series and shunt conductance
(b)
in transmission lines.
(b)
Find
Figure P2.53
- 0 --------------->
>^1
= 40 Q. The switch
= 330
closes at ? = 5
and cal
Figure P2.52
N U M ERICA L AN SW ERS: 10.512, 14.177
Figure P2.54
53.
which
= 0.5 ^. Suppose each AA-bulb rep
resents a 12-watt fluorescent bulb at approxi
mately 12 volts, having an internal resistance o f
12 a
value o f Zj(^).
Find ijit), max[ i 2 {t)], and the average
value o f i2 {t).
102
Chapter 2 K irch h offs Current & Voltage Laws and Series-Parallel Resistive Circuits
i,(t)
9kn
3kO <
6kn<
3cost(2t)V
2kn
12V
out'
=5
If
Find
up, find
pointing
ignition
model for
car battery
with engine
not running
Chasis
ground
Figure P2.58
CH ECK:
= 900 watts
model for a
12V battery
15Q
Figure P2.57
AN SW ER: 0.0376 Q
(a)
103
Chapter 2 KirchhofFs Current & Voltage Laws and Series-Parallel Resistive Circuits
(b)
v(V)
Find
(d)
Find
v (V)
60
40 -
20 -
^--------- 1-----------
0.5
1 V3
>i(A)
i(A)
(b)
power
con
is the power
Load
(c)
Figure P2.62
AN SW ER: 6.25 mS
when
= 1 kQ, R2 =
gm = 0-2 mS.
104
Chapter 2 K irchh offs Current & Voltage Laws and Series-Parallel Resistive Circuits
MATLAB PROBLEMS
(a) Find the output voltage,
the output
C H EC K :
Figure P2.63
= 10 kO.
in terms o f
Figure 2.66
Then find
(b)
is the absorbed
Figure P2.64
C H EC K : 0.75 S.
(a)
l/-^.
2kn
SmV,
8kQ'
2kn<
5kO
r I,
should
insert
the
proper
number or formula.
8kO<
Figure P2.65
RL = 8:1:64;
% This command generates an array o f
PL = RL .* I L . ^ 2 ;
Chapter 2 K irchh offs Current & Voltage Laws and Series-Parallel Resistive Circuits
plot(RL, PL)
Gb = Ga + 1/Rl;
grid
% Plot IL in mA
Rb = 1/Gb;
% Continue these additions and reciprocals until
plot(RL, IL *]0 0 0 )
obtaining Req.
grid
% typing grid adds a grid to your
Geq = 1/Req;
IRc = 20*Gc/Geq;
V Rb = IRc*Rb;
% Now write down the MATLAB expression for
finding Vout.
AN SW ERS: (a) 3 Q; (b) 24 V
Find R^^_
(b)
Find Vouf
R^= 180
--R =20
R =90
20A
105
R,= 150
R =40
Figure P2.67
R =80
106
Chapter 2 K irchh offs Current & Voltage Laws and Series-Parallel Resistive Circuits
for the
on of
K uf
-Cr
200 mA
---
2kO
-- ---------1,2kO
3kO
^3,2kn
(D
IkO
2.2kO
3kO
1.6kn
-
Figure P2.68
A N SW ERS: 591.2 Q, 8.869 V
69. Use MATLAB to find R^^,
and /j for
300n
300
2on
ion
eon
i3on
10V
'4
4on
i5on
2oon >
135Q
> V
500*
-OFigure P2.69
70. Use MATLAB to find Ri in the circuit
o f Figure 2P.70.
3on
2on
8000
100
6000
600
1300
1500
100V
4000 <
400
1350
-o
Figure P2.70
A N SW ER: 50.53 II, 133.8 mA
5000
< 500
C H A P
Nodal and Loop Analyses
108
culation. These software packages use a node equation method over the mesh equation approach.
One o f several reasons is that a node is easily identifiable, whereas a set o f proper meshes is diffi
cult for a computer to recognize.
For resistive networks driven by current sources, writing node equations is straightforward.
Certain difficulties arise in writing node equations for circuits containing independent and
dependent voltage sources. During the 1970s, a modification o f the conventional node method
by a research group at IBM resulted in the modified nodal analysis (MNA) technique. W ith the
M N A method, the formulation o f network equations, even in the presence o f voltage sources and
all types o f dependent sources, becomes very systematic.
This chapter discusses the writing and solution o f equations to find pertinent voltages and cur
rents for linear resistive networks.
CHAPTER O U TLIN E
1.
2.
3.
4.
5.
6.
7.
8.
CH APTER O BJECTIVES
1.
2.
Describe and illustrate the method o f node analysis for the computation o f node voltages
in a circuit. Knowledge o f the node voltages o f a circuit allows one to compute all the
branch voltages and, thus, with knowledge o f the element values, all the branch currents.
Define the notion o f a mesh or loop current and describe and illustrate the method o f
mesh or, more generally, loop analysis for the computation o f loop currents in a circuit.
Knowledge o f all the loop currents o f a circuit allows one to compute all the branch cur
rents. Thus, in conjunction with the knowledge o f the branch element information, one
can compute all the branch voltages.
3.
Formulate the node analysis and loop analysis equations as matrix equations and use
matrix methods in their solution emphasizing the use o f existing software for the gener
al solution.
4.
Describe and illustrate the modified nodal approach to circuit analysis. This method
underlies the general software algorithms available for computer simulation o f circuits.
109
chapter deals almost exclusively with dc, the uppercase notation for voltages and currents is com
monplace.)
FIG U RE 3.1. Diagram of a circuit with labeled node voltages, V^, Vg, Vq V^,
with respect to the given reference node.
KVL states that every branch voltage is the difference o f the node voltages present at the terminals
o f the branch: for circuits in this text and all pairs o f nodes, j and k, the voltage drop
from
n o d ej to node k, is
at every instant o f time, where VJ- is the voltage at node j with respect to the reference and
is
the voltage at node k with respect to reference. Here, j and k stand for arbitrary indices and could
be any o f the nodes. A, B, C, or D , in Figure 3.1. These statements mean that knowledge o f all
node voltages in conjunction with device information paints a rather complete picture o f the cir
cuits behavior. This chapter develops techniques for a systematic construction o f equations that
characterize a circuits behavior.
One last introductory point: Throughout this chapter and in many subsequent chapters, software
programs such as MATLAB facilitate calculations. Constructing sets o f equations that character
110
ize the voltages and currents in a circuit is often a challenge. Solving such sets o f equations with
out the use o f software tools presents a much greater challenge. Yet facilitated by MATLAB or
equivalent, the calculations reduce to a hit o f the return key. MATLAB and the circuit simulation
program called PSpice or Spice (utilized in Chapter 4) are but two o f the many modern and
important software tools available to engineers.
111
90
FIG U RE 3.2. Simple resistive circuit showing three closed paths (dotted lines) that represent
three loop currents, /j, Ij, and 1^; the branch current
which is a difference
o f the two loop currents through the resistor.
Using a fluid flow analogy, one can think o f loop currents as fluid circulating through closed sec
tions o f pipe. The fluid in different closed paths may share a segment o f pipe. This segment is anal
ogous to a branch o f a circuit on which two or more loop currents are incident. The net current
in the branch is analogous to the net fluid flow. Note that each branch current can be expressed
as a sum o f loop currents with due regard to direction. For example, in Figure 3.2, the branch cur
rent 7^3 = ^\ -
Using loop currents, element resistance values, and source values, it is possible
by KVL and Ohms law to express the sum o f the voltages around each loop in terms o f the loop
currents. For example, the first loop, labeled
^ = 9 / i + 3(/i -/ 2 )+ 6 (/ ^ -/ 3 )
We will explore this concept more thoroughly in Section 5. Here we see that loop analysis builds
on KVL, whereas node analysis builds on KCL.
112
be computed from the set o f nodal equations. Each nodal equation will sum the currents leaving
a node. Each current in the sum will be expressed in terms o f dependent or independent current
sources or branch conductances and node voltages. The set o f these equations will have a solution
that yields all the pertinent node voltages o f the circuit. Examples 3.1 and 3.2 illustrate the basic
techniques o f nodal analysis.
EX A M P L E 3 .1 .
The circuit o f Figure 3.3a contains an independent voltage source, an independent current source,
and five resistances whose conductances in S are G j through Gy The nodes other than the refer
ence are labeled with the node voltages V^,
c. T he analysis o f this circuit illustrates the process o f nodal analysis to find the node voltages V^,
y,,andK =
FIG U RE 3.3A. Resistive circuit for Example 3.1. Note that node voltage
is specified by the voltage source.
So l u t io n .
Step 1. Consider node c. A voltage source ties node c to the reference node. Hence, the node volt
age
is fixed at V-^, i.e.,
Because
it is not necessary to apply KCL to this node
unless the current through the voltage source is required, for example, when determining the
power delivered by the source.
Step 2. Sum the currents leaving node a. From KCL, the sum o f the currents leaving node a is
zero. As per the partial circuit in Figure 3.3b, this requires that
and
and moving the source values to the right side o f the equa
113
(3.1)
Step 3. Sum the currents leaving node b. Applying KCL to node b, reproduced in Figure 3.3c,
yields the equation
G 2 ^y b- y a) ^G ,V b^ G, {V ,- VJ = Q
After regrouping terms, one obtains our second nodal equation:
G4(v ,- v
G 3V,
G .(V ,-V )
FIG U RE 3.3C
-G ^ V ^ + {G ^ + G ,+ G ,)V ,= G,V.
(3.2)
Step 4 . Write set o f nodal equations in matrix form. Equations 3.1 and 3.2 in matrix form are
Gi + G
2+
G5
-G 2
-G 2
G 2 -I- G 3 -I- G 4
(3.3)
Matrix equations organize relevant data into a unified framework. Because many calculators do
matrix arithmetic, because o f the widespread availability o f matrix software packages such as
MATLAB, and because equation solution techniques in circuits, systems, and control heavily uti
lize matrix methods, the matrix equation formulation has widespread and critical importance.
Step 5 . Solve the matrix equation 3.3: For this part, suppose that the conductance values in S are
Gj = 0.2, G2 = 0.2, G3 = 0.3, G 4 = 0.1, G 5 = 0.4, that = 2.8 A, and that
= 24 V. After sub
stitution, equation 3.3 simpUfies to
114
0.8
-0 .2 - K,'
- 0 .2
0 .6
12.4'
(3.4)
2.4
Solving using the inverse matrix method leads to the node voltages (in volts):
0.8
- 0 .2
- 0 .2 '
0.6
-1
T 2 .4 -
2.4
0 .6
0.4 4 0.2
0 .2 ' 12.4'
18-
0.8
10
2.4
Alternately, one could have solved equation 3.4 via MATLAB, its equivalent, or the age-old hand
method o f adding and subtracting equations. For example, in MATLAB
M =[0.8 -0.2;-0.2 0.6];
b= [12.4 2.4]';
>>NodeV = M\b
NodeV =
1.8000e+01
l.OOOOe+01
% O R EQU IVA LEN TLY
NodeV = inv(M )*b
NodeV =
18
10
Step 6. Compute
18 - 10 = 8 V.
Exercises. 1. Utilize the solution o f Example 3.1 to compute the current leaving and the power
delivered by the independent voltage source.
AN SW ER: 3.8 A and 91.2 watts
2. Referring to Figure 3.3a and the values set forth in Step 5 o f Example 3.1, suppose the value o f
is cut in half, the value o f V-^ is 24 V, and the value o f each o f the conductances is also cut in
half W hat are the new values o f the node voltages?
AN SW ER: All node voltages are the same.
3. By what single factor must the values o f
and V-^ be
in Figure 3.4.
F IG U R E 3.4.
115
EX A M PLE 3.2.
Consider the circuit o f Figure 3.5a. Similar to Example 3.1, the objective is to find the node volt
ages V^,Vf^, and
replaced the independent voltage source o f Figure 3.3a. This change unfreezes the constraint on
the value o f
present in the circuit Figure 3.3a. There will result three nodal equations in the
three unknowns
, and V^.
0.4 U
, and V^.
So l u t io n .
Step 1. Sum currents leaving node a. This step is the same as Step 2 o f Example 3.1. By inspec
tion o f node a,
0 . 2 + 0.2(K^ - V^) + 0.4(V^ - V J - 2 = 0
which upon regrouping terms yields
(3.5)
Step 2. Sum currents leaving node b. This step is the same as Step 3 o f Example 3.1. Again, by
inspection,
0.2
- VJ + 0.3
Simplification yields
-0 .2 7 ^ + 0.6V ^ ^ -0.1K ^ = 0
(3.6)
o. 4 ( v - v :
(3.7)
116
Step 4. Write equations 3 .5 -3 .7 as a matrix equation and solve. The matrix form o f our nodal
equations 3 .5 -3 .7 is
0.8
- 0.2
- 0 . 4 ' v;'
- 0.2
0.6
- 0.1
- 0 .4
- 0 .1
0.5
'T
(3.8a)
= 0
Solving equation 3.8 using MATLAB or equivalent, using a calculator that does matrix operations,
or solving via some form o f row reduction, one obtains the solution (in volts)
0.8
= - 0.2
Vc
- 0 .4
- 0.2
- 0 .4 '
0.6
- 0.1
-1
'2 .9
1.4
2.6
8.4'
- 0.1
0 = 1.4
2.4
1.6
0 = 4 .4
0.5
2.6
1.6
4 .4
(3.8b)
9.6
Specifically, in MATLAB
M = [0.8 -0.2 -0.4;-0.2 0.6 -0.1;
-0.4 - 0.1 0.5];
>>b = [2 0 1]';
NodeV = M\b
NodeV =
8.4000e+00
4.4000e+00
9.6000e+00
Exercises. 1. Suppose the values o f the current sources in Figure 3.5a are doubled. W hat are the
new values o f the node voltages? Hint: Consider the effect on equation 3.8.
AN SW ER: All node voltages are doubled.
2 . Suppose the conductances in the circuit o f Figure 3.5a are cut in half, i.e., the resistances are
doubled. W hat are the new node voltages?
A N SW ER: Node voltages are doubled.
3. Suppose the conductances in the circuit o f Figure 3.5a are cut in half W hat happens to the
magnitudes o f the branch currents? Hint: Express the branch current in terms o f the branch con
ductance and its terminal node voltages.
AN SW ER: The magnitudes o f the branch currents
stay the same.
I
4. Find two node equations characterizing the cir
cuit o f Figure 3.6.
AN SW ER: (G j + G^) -V~G^^Vy =
and
FIG U RE 3.6
117
The matrices in equations 3.3, 3.4, and 3.8a are symmetric. A symmetric matrix, say A, is one
- 0.2
- 0 .4 '
A = - 0.2
0.6
- 0.1
- 0 .4
- 0.1
0.5
present in the circuit., as in Examples 3.1 and 3.2, the coefficient matrix o f the node equations (as exemplified in equations 3.3, 3.4, and 3.8) is always sym
metric, provided the equations are written in the natural order.
When only resistances, independent current sources, and grounded independent voltage sources are
present in the circuit, the value o f the entries in the coefficient matrix o f the nodal equations can be
computed by inspection. The 1-1 entry o f the matrix is the sum o f the conductances at node a (or
1); the 2 - 2 entry is the sum o f the conductances at node b (or 2). In general, the i-i entry o f the
coefficient matrix is the sum o f the conductances incident at node i. Further, the 1 -2 entry o f the
matrix is the negative o f the sum o f the conductances between nodes a and b (or between nodes 1
and 2), and the 2 -1 entry has the same value. In Example 3.2, the 1 -2 entry o f - 0 .2 S is the nega
tive o f the sum o f the conductances between nodes a and b; the 1 -3 entry o f - 0 .4 S is the negative
o f the sum o f the conductances between nodes a and c (or between 1 and 3, if the nodes were so
numbered). Thus, whenever the circuit contains no dependent sources, the node equations can be
written by inspection. Further, if independent voltage sources are absent, then the right-hand side o f
the nodal matrix equation can also be written by inspection: the i-th entry is simply the sum o f the
independent source currents injected into the ith node at which KCL is applied.
W hen controlled sources are present in the circuit, the resultant nodal matrix is generally not sym
metric, as illustrated in the following two examples.
the input signal should have a relatively small magnitude so that a LIN EA R circuit will adequately
represent the amplifier. Similarly, low-frequency means that the frequency o f any sinusoidal input
must be relatively low for the (resistive) circuit model o f the amplifier to remain valid.
The amplifier circuit model contains a current-controlled current source (CC CS) and a voltagecontrolled current source (VCCS). These two dependent sources have currents that depend on
other circuit parameters and require some special handling when constructing node equations.
Our objective is to set forth the methodology for writing the node equations when dependent cur
rent sources are part o f the circuit and to compute the magnitude o f the voltage gain, |
V-^ |
= \y 2 |y^n\
Note that the source voltage, V-^, specifies the voltage at the node at the bottom o f Gp hence, a
nodal equation at this node is unnecessary. Nodal equations must be written at the remaining
118
often a matter o f personal preference. In this example, we have chosen 1, 2, and 3 as node labels,
in contrast to the previous two examples, where we used a, b, and c.)
G,
So l u t io n .
Step 1. Sum the currents leaving node 1. Summing the current leaving node 1 leads to
(^1 -
+ ^2 (^1 - ^ 3) + ^3 (^1 -
+ P
(3.9)
Step 2. Substitute fo r i^ in equation 3 .9 and simplify. In equation 3.9, |3 i^ accounts for the effect
o f the C C C S at node 1 and is not given in terms o f the circuit node voltages. To specify this term
in terms o f the circuit node voltages, observe that in Figure 3.7, i^ is the current from node 2 to
node 3 through G^. Hence,
(3.10)
Substituting equation 3.10 into 3.9, again grouping like terms, one obtains the first nodal equation,
(3.11)
+ G,{V^ - K3) +
= 0
( 3 . 12)
119
Step 4, Specify
Hence,
gm^. = & j y x - y , ) - g n y x - g r r y ,
Substituting
(3-13)
o f equation 3.13 into equation 3.12 leads to our second nodal equation,
^ 3)^ , + (G 3 + G4 + G ,)V ^ -{G , + ^ J K 3 = 0
(3.14)
Step 5. Sum the currents leaving node 3. Applying KCL to node 3 yields,
G 2 ( V ,- V , ) ^ G , i V , - V , ) ^ G , V , - p i ^ - g ^ v ^ = 0
(3.15a)
= G , { V , - V ,) + G , { V , - V ,) + G ^ V ^ - ^ G , { V , - V , ) - g ^ { V , - V , )
Grouping like terms leads to our third equation in the three unknowns V^, V^, and V^:
-(G i + kJ V
x-
(3.15b)
Step 6. Put nodal equations in matrix form. The three nodal equations 3.11, 3.14, and 3.15b
have the matrix form
Gi + G 2 + G 3
Sm - G
PG^ G 3
~ ^ 2
G 3 + Gg + G 4
^2 8 m
~^A -
-G 4 -
G 2 + G 4 + fiG^ + G 5 + g
y,'
G.V^v/
^2 =
^^3
Step 7. Substitute values and solve. Suppose that the various circuit conductances have the fol
lowing values in [xS: G j = 1,000, G j = 2.0, G j = 1.0, G^ = 10, G j = 2 0 ,100, and Gg = 200 .
Suppose further that
beta*G 4-G 3
G 3+G 6+G 4
-G 4-beta*G 4
b = [G l* V in
NodeV = M\b
NodeV =
2 . 0000 e +00
-l.OOOOe+02
l.OOOOe+00
0]';
-G 2-beta*G 4;
-G4-gm ;
G2+G4+beta*G4+G5+gm ];
120
in which case,
V,-
V^2 = - 1 0 0
V3
Step 8. Compute the voltage gain. The voltage gain o f the amplifier is given by
Kut
V2
-1 0 0
Vin
Vin
2.1
= 4 7 .6 2
Exercises. 1. Suppose V-^ in the circuit o f Figure 3 .7 is doubled. W hat are the new node voltages?
Hint: Consider the matrix equation o f Step 6.
A N SW ER: Node voltages are doubled.
2. Suppose all conductances in the circuit o f Figure 3.7 are cut in half (resistances are doubled)
and (3 is held constant. How must^^ change for the node voltages to remain at their same values?
AN SW ER:
must double.
Realistic problems do not permit hand solutions. For hand solutions, the smallest number o f
equations is generally desired. For matrix solutions using software packages such as MATLAB,
more variables with more equations may often be easier to construct and may often result in more
reliable numerical calculations. This can be illustrated using the equations o f Example 3.3. All the
pertinent basic equations o f the circuit o f Figure 3 .7 can be written down as follows: from equa
tions 3.9 and 3.10 we have
and
However, in contrast to the example, we do not substitute 3.10 into 3.9 to obtain 3.11. Rather,
we just let them be two independent equations. Further, from equations 3.12, 3.13, and 3.15a,
we have
- G 3 K, + (G 3 + G4 + G ,) K2 - G 4 K3 + g^v^ = 0
and
By not substituting for and v^, we avoid unnecessary hand calculation, and if there is an error,
it is easier to find. The resulting equations have the matrix form where i and
now appear as
additional unknowns, easily handled by a software program:
121
G] 4- G 2 + G 3
-G 3
-G 2
13
G4
-G 4
-I
-G 3
G 3 -1- G 4 -HGg
-G 4
8m
-1
-1
-G 2
-G 4
G2 + G 4 + G5
-/?
~Sm
V,-
^2
=
iy
0
0
0
As a general rule, we would reorder the equations so that rows 1, 3, and 5 came first, as they cor
respond to the three nodal equations at Kj, V2 , and
equations for i and
Exercise. Solve the above matrix equation in MATLAB or equivalent, using the numbers o f
A and - 1 V.
Example 3.3 to verify that = -1 .0 1 x 10^-3 A
Matrix methods as used in the above examples and in the ones to follow necessitate the power o f
a calculator or a software program such as MATLAB for easy solution. Such programs permit a
straightforward calculation o f the required answers and are not prone to arithmetic errors.
The next example illustrates how to write node equations for circuits containing a voltage-con
trolled voltage source (VCVS) grounded to the reference node. The analysis o f CCVSs grounded
to the reference node is similar. The more challenging analysis o f circuits containing floating
dependent or independent voltage sources is taken up in the next section.
EX A M P L E 3 .4 .
The circuit o f Figure 3.8 models a poor operational amplifier circuit' in which the output voltage
1/^^ = V2 approximates
For the analysis, let |i = 70. The adjective poor arises because ^
V. =
So l u t io n .
The circuit contains two nodes labeled
and
by voltage sources. The goal o f our analysis is to find these node voltages by writing two equations
in these voltages and solving. As is commonly the case, resistances are in ohms and will be con
verted to conductances in S for convenience in writing the node equations.
/~N,
122
Step 1. Compute conductance values in S. Conductances are the reciprocal o f resistances, i.e., G-
= ^IRj- Hence,
= 3.3 3 3 3 3 lQ-5, G^ = 10^, G^ = 0.01, and G^ = lO'^
G j = 2.0 10-3,
Step 2 . Write a node equation at node 1. Summing the currents leaving node 1 yields
^3 (^^2 - ^l) + ^5 ^2 + G4 {I /2 +
Kj) = 0
+ (G 3 + G4+ ^ 5)^2 = 0
(3.17)
Step 4. Write equations 3 .1 6 and 3 .1 7 in matrix form and solve. In matrix form
3.3333
-1
-V f
0 .6 9 9 9
O .O Ill
.^2.
2'
0
Using the formula for the inverse o f a 2 x 2 matrix (interchange the diagonal entries, change the
sign on the o ff diagonals, and divide by the determinant), one obtains
Vf
0.0111
0 .0 3 0 1 2 6
.^2.
0.7369
-0 .6 9 9 9
3.3333
-1 .8 9 9 6
^^ 2
Exercises. 1. Write MATLAB code to solve the above example. Check that your code works. Hint:
See Example 3.3.
2. If R2 is changed to 100 k2 in Example 3.4, show that V2 = -1 .9 0 6 3 V.
123
l^y. This additional current becomes an unknown in a set o f nodal equations generated by apply
ing KCL to each node. At this point, further explanation is best done by an example, but the con
cept is similar to the discussion following Example 3.3.
E X A M P L E 3.5.
Find the node voltages V^, Vy,
FIG U RE 3.9. Resistive circuit containing a floating voltage source for the given reference;
generally, the reference node may be chosen arbitrarily.
So l u t io n .
Step 1. Write a node equation at node a. Summing the currents leaving node a yields
8 + 0.15
+ 3 + 0.2 ( K ^ -K .) = 0
(3.18)
124
- 3 + 0 .1 5 ( K ^ - K ; + 0.05
or equivalently,
- 0 .1 5
(0 .1 5 + 0.05) n - / ^ ^ = 3
(3.19)
{V^- K J = 0
or equivalently.
- 0 .2 K ,+ 0.45K^ + /,, = 25
(3.20)
Step 4. Write the node voltage relationship fo r the terminal nodes o f the floating voltage source,
i.e., between the voltages
and
= 440, i.e.,
K ^ = 440
(3.21)
Step 5. Write thefour equations 3.18, 3.19, 3.20, and 3.21 in matrixform and solve. In matrix form,
0.35
- 0 .1 5
- 0 .2
-1 1
- 0 .1 5
0.2
-1
- 0 .2
0.45
25
-1
hb
(3.22)
440
Because o f the extra variable, the equations become too large for hand calculation. Hence, we use
MATLAB as follows:
>>M = [0.35 - 0 .1 5 - 0 .2 0 ;-0 .1 5 0.2 0 - 1 ;
- 0 .2 0 0.45 1 ; 0 - 1 1 0];
b = [-11 3 25 440]';
x = M\b
X =
-9.0000e+ 01
-3 .1 0 0 0 e + 0 2
1.3000e+02
-5.1500 e+ 01
Hence,
= - 90 V,
= - 3 1 0 V,
= 130 V,
= -5 1 .5 A
12 5
In a conventional nodal analysis, all unknowns are node voltages. Here we have the additional
unknown current,
and
2. For Example 3.5, compute the power absorbed by the 0.15 S resistor.
3. Compute the power delivered by the floating voltage source.
AN SW ERS in random order: 220 V, 22.6 6 kw, 310 V, 440 V, 7260 watts
The next example investigates a circuit having floating independent and dependent voltage
sources. By convention, the reference node o f this circuit, figure 3.10, and all subsequent circuits,
will be the bottom node o f the circuit unless stated otherwise.
E X A M P L E 3.6.
The circuit o f Figure 3.10 contains a floating independent and a floating dependent voltage
source. Find the node voltages V^, Vy, V^, and the unknown currents
and
FIG U RE 3.10. Resistive circuit containing a floating dependent voltage source and a floating
independent voltage when node d is chosen as the reference node.
So l u t io n .
Step 1. Sum currents leaving node a. Here,
126
Equivalently,
100
(3.23)
100
100
500
Equivalently,
(3.24)
Step 3. Sum currents leaving node c. Here,
1
(3.25)
800
Step 4. Write an equation relating the terminal voltages o f the independent voltage source. Here,
K ,-n = 3 0
(3.26)
Step 5. Write an equation relating the terminal voltages o f the dependent voltage source. Here,
40
K ,- K = 4 0 ., =
Equivalently,
(3.27)
0.6V;, + 0 .4 y ^ - V ^ = 0
Step 6. Write equations 3.23 through 3 .2 7 in matrix form an d solve in MATLAB. Combining
the above equations into a matrix produces
0 .03
-
0.01
-0 .0 1
0.012
-1
0.0 0 1 2 5
-1
^ac
30
Icb,
-1
0 .6
0.4
________
-1
2.2'
0
V'i
Again, this matrix equation is too large for hand computation. Hence in MATLAB,
M = [0.03 - 0 .0 1 0 1 0;
-
0.01 0 .012 0 0
0 0 0.00125 - 1 1;
0 - 1 1 0 0;
0.6 0 . 4 - 1 0 0];
1;
127
l.OOOOe+02
5.0000e+01
8.0000e+01
-3 .0 0 0 0 e -0 1
-4 .0 0 0 0 e -0 1
Hence,
100
50
=
80
^ac
- 0 .3
Icb_
- 0 .4
Step 4. Compute the power delivered by the 3 0 Vsource. The power delivered by the 30-V source is
^^./ = - 3 0 / ,^ = 3 0 x 0.4 = 1 2 W
Step 5. Compute the power delivered by the dependent source. The power delivered by the depend
ent source is
Pdel =
- 4 0
^ ( - 0 . 3 ) = 0.12(V^ - V^) = 6 W
100
Q resistor.
A N SW ERS in random order: 8 watts, 20 V, - 8 0 V
2. Suppose the two independent voltage source values in Example 3.6 are doubled. W hat are the
new node voltages? W hat are the new branch currents?
A N SW ERS: Node voltages are doubled and branch currents are doubled.
3. Suppose all resistances in the circuit o f Figure 3.10 are doubled and the value o f the parameter
on the dependent source is also doubled. W hat are the new branch currents?
AN SW ER: All branch currents are cut in half.
The above example increases the number o f unknowns beyond the node voltages to include the
two currents through the floating voltage sources. However, we could have included additional
currents to the set o f equations making the dimension even higher. W ith a tool like MATLAB,
this poses no difficulty. However, it does make hand computation a challenge. For example, we
128
ber o f equations. By adding addirional unknowns we would simplify the writing o f the individual
node equations but increase the dimension o f the matrix equation. Specifically, the node equation
at a becomes
_
t i
50
IV
50
2.2
0.002
-1
0.00 1 2 5
-1
-1
-1
-4 0
-1
-100
0
^ac
30
Ic h
0
0
This completes our discussion o f the standard nodal equation method o f circuit analysis. T he next
section takes up a discussion o f an alternative analysis method entided loop analysis.
5. LOOP ANALYSIS
Loop analysis is a second general analysis technique for computing the voltages and currents in a
circuit. Mesh analysis is a special type o f loop analysis for planar circuits, i.e., circuits that can
be drawn on a plane without branch crossings. For planar circuits, loops can be chosen as mesh
es, as illustrated in Figure 3.2, or as in 3.11 below. Associated with each loop is a loop current.
Loop currents circulate around closed paths (loops) in the circuit. Similarly, for planar circuits,
the term mesh current is used traditionally for loop current. By KVL, the sum o f the voltages
across each branch in a loop is zero. By expressing each o f these branch voltages in terms o f the
designated loop currents, one can write an equation in the loop currents for each designated loop
in the circuit. For branches that are often common to two or more designated loops, the branch
current equals the net flow o f the loop currents incident on the branch. Writing an equation for
each loop produces a set o f equations called loop equations. If sufficient independent loops are
defined, one can solve the loop equations for the loop currents. Once the loop currents are known,
we can easily compute the branch currents and then the branch voltages in the circuit. Then we
can compute any other quantities o f interest, such as power absorbed, power delivered, voltage
gain, etc.
EXA M PLE 3.7.
Consider the planar circuit o f Figure 3.11 with the three specified loops, which are also called
meshes. Denote the loop currents for each loop by /j, 1^, and /j. The objective is to write three
equations in the currents /j,
and /j using KVL and solve these equations for their values. Then
we will compute the power absorbed by the 2 -Q resistor marked with the voltage v. Suppose the
source voltages are
= 4 0 V andV^2 = 20 V.
129
So l u t io n .
Step 1. Write a KVL equation based on loop 1 by summing voltages around this loop. Summing
the voltages around loop 1 using Ohms law and the defined loop currents produces
K., = 40 = /i + 4 (/; - /j) + Va + ih - ^3) = 6 /, - 4 /2 - 7 3 + 20
(3.25a)
Here, observe that the 4 Q resistor is incident on two loops; the net current flowing from top to
bottom, i.e., with respect to the direction o f loop 1, is /j - l 2 - The idea is analogous to a pair o f
distinct water pipes that share a common length. The common length is analogous to the 4 -Q
resistor. The flow rate in each pipe is analogous to the currents /] and I 2 , which in fact, are rates
at which charge flows past a cross sectional area o f the conductor. It follows that the net flow
through the common length o f pipe with respect to the direction o f loop 1 is the difference in the
net flow rates o f pipes 1 and 2, respectively. This is precisely the meaning o f /j - /2 . A similar
explanation can be made for the 1 -Q resistor common to loops 1 and 3 for which the net flow
rate with respect to the direction o f loop 1 is 7j - ly
Simplifying equation 3.25a yields
6/1 - AI^ - 73 = 20
(3.25b)
Step 2. Write a KVL equation based on loop 2 by summing the voltages around this loop. Applying
Ohms law and KVL to loop 2 produces
0 = 4(72 - 7,) + 272 + 2(^2 -
(^.26)
h'
Notice that with respect to the direction of loop 2, the net flow rate through the 4 Q resistor is
Step 3. Finally, write a KVL equation based on bop 3. Stmiming the voltages around loop 3 yields
(3.27)
130
Step 4 . Write eqtiations 3.25b, 3.26, an d 3 .2 7 in matrix form an d solve. Writing the above three
loop equations in matrix form yields
'6
-4
-r
-4
-2
-1
-2
20'
(3.28)
20
Solving Equation 3.28 by the matrix inverse method (by a numerical algorithm or by Cramers
rule) yields the loop currents in amps as
h'
-4
-r
-1
0 .35
'2 0 '
= -4
-2
-1
-2
20
= 0 .2 2 5
0 .225
0 .2
20'
0 .2 8 7 5
0 .2
0 .2
0 .4
20
0 .2
11
=
8.5
12
Step 5. Compute the power consumed by the 2 Q resistor. Knowledge o f the loop currents makes
it possible to compute all voltages and currents in the circuit. For our purpose, the voltage
/ 2 = 24.5 watts.
Observe that there are no dependent current or voltage sources in the circuit. Similar to the nodal
analysis case, whenever dependent sources are absent and the equations are written in the natural
order, the loop (or mesh) equations are symmetric, as illustrated by the coefficient matrix o f equa
tion 3.28 where, for example, the 1 -2 and 2 -1 entries coincide, as do the 1 -3 and 3 -1 entries,
etc. Also, the value o f all entries can be computed by inspection. The 1-1 entry o f the matrix is
the sum o f the resistances in loop 1; the 2 - 2 entry is the sum o f the resistances in loop 2, etc. In
general, the i- i entry is the sum o f the resistances in loop i. T he 1 -2 entry o f the matrix is 'L{R^
(the large sigma means summation), where each
131
sign otherwise. Further, if independent current sources are absent, then the right-hand side o f the
loop equations can also be written by inspection. The i-th entry is simply the net voltage o f the
sources in the ith loop that tends to deliver a current in the direction o f the loop current.
Exercises. 1. Use the inspection rules described above to write two mesh equations for the circuit
o f Figure 3.12, when both mesh currents are assigned clockwise direction.
FIG U RE 3.12.
2. Use the inspection rules described above to write two mesh equations for the circuit o f Figure
3.12, when the left mesh current is clockwise and the right mesh current is counterclockwise.
3. Use the inspection rules described above to determine the right-hand side o f the mesh equation
for the circuit o f Figure 3.13.
AN SW ERS: 8, 0, 10
A simplifying reduction to the set o f loop equations occurs if an independent current source coin
cides with a single loop current. The analysis becomes simpler because that loop current is no
longer an unknown; rather it is equal to the value o f the source current if their directions coin
cide, or to the negative value if their directions are opposing. Because the associated loop current
is known, there are fewer loop equations to write and solve. One would apply KVL to such a loop
only if it were necessary to compute the voltage across the independent current source, which
might be necessary for determining the power delivered by the source. Th e entire situation is anal
ogous to an independent voltage source tied between a node and the reference in nodal analysis.
The following example illustrates the details o f this discussion.
132
FIG U RE 3.14. A resistive circuit containing an independent current source on the perimeter
o f loop 3 forcing /^ = 8 A.
So l u t io n .
Step 1. So/ve fo r
by inspection. Because
branch containing the independent 8 A current source, /j = 8 A. This phenomena is similar to the
fact that in nodal analysis, the node voltage o f a grounded voltage source is fixed at the voltage
source value.
Step 2. Write a KVL equation fo r loop 1 by summing voltages around this loop. Summing the
voltages around loop 1 using Ohms law and the designated loop currents produces
28 = / j + 4 ( /j - /2) + 12 + ( /j - 8) = 6 /, - 4 /2 + 4
Hence,
6 /, - 4 /2 = 24
(3.29)
Step 3. Write a KVL equation fo r loop 2 by summing the voltages around this loop. Applying KVL
and Ohms law to loop 2 produces
0 = 4 (/2- /i) + 2 /2 + 2 (/2 - 8 ) = - 4 /, + 8 /2 - 16
Equivalently,
- 4 /1 - 8 /2 = 16
(3.30)
133
Step 4 . Write above loop equations in matrix form and solve. The matrix form o f equations 3.29
and 3.30 is
' 6 -A
'i r
24
-4
l2
16
6 -4
I2
-4
-1
24 '
16
32
' 24'
4 6 16
'8
's'
6
Pdel = -
= - (8
8) = - 64 watts
Exercise. In the circuit o f Figure 3.15, nvo o f the three mesh currents coincide with independent
source currents. By writing and solving just one mesh equation, find /j.
A N SW ER: 3 A
Not only do independent current sources constrain loop currents, but dependent currents sources
do also. This situation is illustrated in Example 3.9.
E X A M PLE 3 .9 .
This example illustrates the writing o f loop equations for a simplified small signal equivalent cir
cuit, Figure 3.16, o f a two-stage amplifier that contains a current-controlled current source
(C C C S) and a current-controlled voltage source (CC VS). This process extends the techniques o f
Examples 3.7 and 3.8 to find some important characteristics o f the amplifier. Specifically, find
(a) the input resistance seen by the source, i.e.,
= v-Jij^,
134
j and
FIG U RE 3.16. Small signal equivalent circuit for a two-stage amplifier. Signals in amplifiers are
usually time dependent, so we adopt the lowercase notation for voltages and currents.
So l u t io n .
The circuit o f Figure 3.16 contains three loop or mesh currents. The direction o f the loops is a
user-chosen preference. For convenience, we have chosen mesh current z'2 to be consistent with
the direction o f the arrow in the dependent current source. Because this dependent current source
lies on the perimeter o f the circuit, it constrains the value o f /'2> i-e-) ?2 ^ P ^b-
the control
ling current,
, z'2 = . ih ~ ^ h'
relationship implies that the mesh current o f loop 2
depends directly on the mesh current o f loop 1. This observation allows us to skip constructing a
mesh equation for loop 2. Only equations for loops 1 and 3 are needed, thereby reducing the
number o f simultaneous equations from three (because there are three loops) to two.
Step 1. Apply KVL to loop/mesh 1. Here, by KVL and the observation that
^in = h h +
('1 + P ' 1) =
+ (1 + p)
12
= P zp
i\
in
in
(3.31)
Step 2. Apply KVL to loop/mesh 3. In this case, observe that z^ = (z^ + z'g) = - (|3 z'j + Zj). By BCVL,
'3 -
K + K ('3 + ' 2) =
h +
K +
ph +
'3 = 0
(3.32)
135
Step 3. Write equations 3.31 and 3.32 in matrix form an d solve. The matrix form o f these equa
tions
IS
/?^ + (P + l ) ^ ,
^in
'3
(3.33)
V.
Because the solution is desired in terms o f the hteral variables, we solve equation 3.33 using
Cramers rule, which utilizes determinants. In this task, first define
Using the notation A for the determinant, Cramers rule provides the solution for i^ according to
the formula
det
^in
R^ + R ^+r
(3.34)
I =
Applying Cramers rule for the solution o f Zj, yields
det
Step 4. Compute
(3.35)
in terms o f v-^ and then the voltage gain vjv^^. As per the circuit o f Figure
Vo =
V ; = r
---V.
in
P/?
I d
r= y
Step 4 . Compute v. To compute the voltage across the dependent current source, apply KVL to
mesh 2 to obtain
n
h +
136
FIG U RE 3.17.
c' '1
To see the importance o f the calculations o f the amplifier circuit o f Example 3.9, suppose two
amplifiers are available for use with a non-ideal voltage source. The non-ideal voltage source is
modeled by an ideal one-volt source in series with a 100 Q source resistance. Suppose amplifier 1
has a voltage gain, vjv^^ = 1 0 and
particular application, although it has a lower voltage gain than amplifier 2. Hence, Example 3.9
illustrates the need to know both the voltage gain and the input resistance to determine the out
put voltage in practical applications. Further, using the literal solution to the example allows us to
apply the formulas to different sets o f parameter values without repeating the complete analysis.
In the previous two examples, there were current sources on the perimeter o f the circuit. Such cur
rent sources were incident to only one loop. It often happens that independent and dependent
current sources can be common to two or more loops. When this happens, a situation analogous
to floating voltage sources in nodal analysis occurs. To handle such cases, many texts define some
thing called a supermesh and write a special loop equation for this supermesh. Supermeshes often
confuse the beginner. There is an easier way.
Example 3.10 below illustrates how to write loop equations when current sources are common
to two or more loops. In such cases, we introduce auxiliary voltage variables across current
sources common to two or more loops. The resulting set o f simultaneous equations will contain
not only the loop currents as unknowns, but also the auxiliary voltages as unknowns. Because the
resulting set o f equations contains both loop currents and additional (auxiliary) voltage variables,
the equations are called modified loop equations. T he process o f writing modified loop equa
tions is extremely systematic and straightforward. Further, it allows us to avoid explaining the very
confusing concept o f a supermesh. On the other hand, the presence o f auxiliary voltage variables
increases the number o f unknowns, i.e., the number o f simultaneous equations increases.
Because o f the availability o f software packages such as MATLAB, M ATH EM ATICA, and
MAPLE, this increased dimension is not a hindrance.
137
E X A M P L E 3 .1 0 .
Consider the circuit o f Figure 3-18 in which
pendent current source is common to loops 1 and 3 and a voltage-controlled current source is
common to loops 1 and 2. Find values for the loop currents 7p Ij, l y and the power delivered by
each independent source.
200Q
So l u t io n .
To begin the solution, we introduce two auxiliary voltage variables Vj and Vj associated with the
current sources common to two (or multiple) loops. The purpose o f these variables is to facilitate
the application o f KVL for constructing the loop equations. This will require that we obtain three
KVL equations, one for each loop, and two constraint equations, one for each current source.
Step 1. Apply KVL to loop 1. By a clear-cut applicanon o f KVL,
28 = 2007] - V] - V2
(3.36)
Step 2. Apply KVL and Ohms law to loop 2. Again applying KVL and Ohms law to loop 2, we
obtain 100 7j + 200 (Jr^ 7^) + V2 = 0. After grouping like terms.
3 0 0 7 , - 20 0 /3 + V2 = 0
(3.37)
Step 3. Apply KVL to loop 3. Applying KVL to loop 3 yields 150 7^ + Vj + 200 (/j - Tj) = 0.
Equivalendy,
- 200/2 + 350/3 + V| = 0
Step 4. Write a constraint equation determined by the independent current source. Here, loops 1
and 3 are incident on the independent current source so that
0 .0 6 = / , -/ 3
(3.39)
138
Step 5. Write a constraint equation determined by the dependent current source. In a straightfor
ward manner, we have
/ , - / , = 0 .0 2 V, = 0 .02 r2 0 0 (/3 -
12
) = 4 7 , - 47,
After simplification,
/i + 3/2 - 4 /3 - 0
(3.40)
Step 6. Write equations 3 .3 6 to 3.40 in matrix form an d solve. The matrix form o f these equations is
0
-1
-r
300
-2 0 0
-2 0 0
350
-1
''I
0 .0 6
-4
V2
28
0
=
(3.41)
Solving equation 3.41 by the matrix inverse method or by an available software package yields the
solution (currents in A and voltages in V) given by equation 3 .42 below:
200
-1
300
-2 0 0
-2 0 0
350
-1
-4
Li
,''2.
-r -1 28
0 300 -200 0 1
0 -200 350 1 0
1 0 -1 0 0
1 3 -4 0 0];
b = [28 0 0 0.06 0 ];
LoopIplus= M\b
Looplplus =
l.OOOOe-01
2,0000e-02
4,0000e-02
-l.OOOOe+01
2.0000e+00
0.1
0 .02
= 0 .0 4
0 .06
-To
(3.42)
139
Step 7. Compute the powers delivered by the independent sources. First, the power delivered by the
independent voltage source is
Pl-source =
V^ = - 0 .6 watts
This last value indicates that the independent current source actually absorbs power from the cir
cuit.
Exercise. For the circuit o f Figure 3.19, write the modified loop equations having two unknowns
/j and V, following the procedure described in Example 3.10. Solve the equations and find the
power absorbed by the 2 -Q resistor.
4A
AN SW ER: 18 watts
One final point before closing our discussion o f loop analysis. Loops can be chosen in different
ways. Cleverly choosing loops can sometimes simplify the solution o f the associated equations. For
example, by choosing a loop that passes through a current source so that no other loop is com
mon to the source, the loop current is automatically specified by that current source.
6. SUM M ARY
This chapter introduced the technique o f nodal analysis. Nodal analysis is a technique for writing a set
o f equations whose solution yields all node voltages in a circuit. With knowledge o f all the node volt
ages and all the element values, one can compute all branch voltages and currents. As mentioned, when
ever there are no dependent sources present, the coefficient matrix o f the node equations is always sym
metric. Hence, whenever dependent sources are absent, it is possible to write the nodal equation coef
ficient matrix by inspection. Further, if independent voltage sources are absent, then the right-hand side
o f the matrix form o f the nodal equations can also be written by inspection: the entry is simply the sum
o f the independent source currents injected into the node at which KCL is applied. When VCCSs are
present, the steps for writing nodal equations are the same as illustrated in Example 3.3. Generally, in
such cases, the resultant coefficient matrix is not symmetric.
140
W hen floating dependent or independent voltage sources are present with respect to a given ref
erence node, we introduce new current variable through these floating sources as unknowns. The
node equations then incorporate these additional unknown currents, as was illustrated in
Examples 3.5 and 3.6. This method increases the number o f equations but simplifies the con
struction o f the individual equations. W ith a tool like MATLAB to compute solutions, there is no
difficulty, although hand computation may become more difficult. This concept is the basis o f the
modified nodal analysis method used in circuit simulation programs like SPICE.
Loop/mesh analysis, an approach dual to nodal analysis, was introduced in Section 5. Mesh analy
sis is a special case o f loop analysis for planar circuits when the loops are chosen to be the obvious
meshes, similar in geometry to a fish net. In loop analysis, one sums the voltages around a loop or
mesh to zero. Each o f the branch voltages in the loop is expressed as a product o f resistances and
(fictitious) loop currents that circulate through the branch resistance, as illustrated in Figures 3.10,
3.14, and 3.16. The branch current o f the circuit are equal to the net flow o f the loop currents
incident on a particular branch, meaning that each branch current is expressible as a sum o f loop
currents. The desired set o f loop equations is produced by summing the voltages around each loop,
expressing these voltages either as source values or as resistances times loop currents. One solves
the loop equations for the loop currents. Once the loop currents are known, we can then compute
the individual branch currents and then the branch voltages, and thus any other pertinent current,
voltage, or power. Whenever there are no dependent sources present, the coefficient matrix o f the
loop equations is always symmetric. Whenever dependent sources are absent, it is possible to eas
ily write the loop matrix by inspection.
As the size o f an arbitrary circuit grows larger, there are two good reasons for choosing the nodal
method over the loop method; (i) the number o f nodal equations is usually smaller than the num
ber o f loop equations, and (ii) the formulation o f nodal equations for computer solution is easier
than methods based on loop equations. Writing nodal equations is particularly easy if the circuit
contains only resistances, independent current sources, and VCCSs for short, an R
I
g net
work. For an
network, one simply applies KVL to every node (except the reference node)
and obtains a set o f node equations directly. For floating independent or dependent voltage
sources, the task is more complex. Examples 3.5 and 3.6 illustrate cases where, besides the node
voltages, additional unknown auxiliary currents are added. By adding additional auxiliary variables
to the formulation o f the nodal equations, we described the concept behind the modified nodal
analysis (MNA) method. The MNA method retains the simplicity o f the nodal method while
removing its limitations and is the most commonly used method in present-day computer-aided
circuit analysis programs.
141
Gaussian surface: a closed curve or a closed surface surrounding two or more nodes.
Linear matrix equation: an equation of the form Ax =b, where A i s z n x n matrix, x is an n-vector of unknowns, and b is an n-vector of constants.
Loop (closed path): a contiguous sequence of branches that begins and ends on the same node
and touches no node more than once.
Loop analysis: an organized method of circuit analysis for computing loop currents in a circuit.
Knowledge of the loop currents allows one to compute the individual element currents
and, consequently, the element voltages.
Mesh analysis: the special case of loop analysis for planar circuits in which the loops are chosen
to be the meshes.
Nodal analysis: an organized method of circuit analysis built around KCL for computing all node
voltages of a circuit.
Node voltage: the voltage drop from a given node to a reference node.
Symmetric matrix: a matrix whose transpose is itself I f ^ =
is a x matrix whose i- j entry
is
then A is symmetric if a - = a-.
142
PROBLEMS
^ 2-
= 2R.
Figure P3.3
4 Ksi-
Figure P3.1
A N SW ER: K, = l.5V^^
reference node
Figure P3.4
value o f
50 A
>0.020
7 0.20
''load <
v
I
Live
Battery
0.02Q
V
12V
(b)
(c)
Figure P3.2
Starter
Motor
a b
c d
'd -b'
ad - be
-c a
143
(d)
Find 1 ^ ,7 ^ and
(e)
reference node d
Figure P3.5
= 8
V* Further,
= 5 kQ,
R l = R^ = R^ = 20 k2, and R^ = 10
kfl. Find the node voltages,
and
mA, V^2 -
(b)
resistor
- Vg is a measure of how
= 50
resistor is a
cut in half
(b)
Assuming
----------^/S/^-------
variables
and Vq
= 40 kQ at 0C, put
Figure P3.6
(d)
voltages as
0.4, G2 = 2, G3 = 3, and
analysis to find
= 5. Use nodal
varies from 15 to 65
144
II.
= 0.01 S, G^ =
Figure P3.8
= 500 Q. Suppose
=
G,
G,
'
Figure P 3 .1 1
12. In the circuit o f Figure P 3.12,
= 20 kQ, T?3 = 20 kQ,
= 5 mA. Find
Figure P3.9
1,7,0.6,7
= 30 V,
Figure P3.12
(b)
(c)
(d)
(c)
= 15
, Vg the power delivered by
, and the power absorbed by R^ .
Compute I 2 , the current through Rj
^2 = 10 kQ, T?3 = 10 kQ, and
mS, find
ion
100
(e)
ANSWERS IN RANDOM O R D E R
= 5 kQ,
= 0.55 x 103, 7-
ion V.
ion.
Reference node
Figure P 3 .1 0
Figure P 3 .13
145
(c)
(e)
= 20 Q, i?2 = 10 Q,
= 10 Q, and I^=
= 4 Q,
A. Note that in
0.1 S,
and Vq when
= 0.1 A, 7^2 =
0-2 A,
= 7 mS;
= 2 mS,
=
500 Q ; i?2 = 333.33 2; and /?3 = 1 fl.
Determine Kq.
ages
(d)
, and
Suppose
for
. 9.iVa
'4
Figure P3.14
Figure P3.17
= 0 ; node voltage C is
(b)
and
Find
"
and the
= 9
(e)
-------- --------------
Figure P3.15
16. Redo problem 15 with
\
J
= 60V .
0.25 S and
which
60 ^2.
- 20 Q, /?2 ~ 20 Q,
= 2 0 Q,
= 30 Q, R^ =
= 12 V, and
146
and
Vy,
(b)
W rite
modified
nodal
equations,
'A-.'
.^0.
(c)
,a n d .
Solve the equations for the node voltages Kg and Vq and the current
through the 50 V source. C H ECK :
Vg = 50 V and
Figure P3.20
= 100 V.
(d)
(e)
0 .0 1 9
- 0 .0 0 5 '
-0 .0 0 1
0 .0 0 5 5
- 0 .0 0 2
-0 .0 0 5
- 0 .0 0 2
0 .1 0 7
0) given by
0
=
0
O .n/
0
=
0
G A.
Figure P 3 .2 4
147
which l/j = 2 50 V,
: 20 Q,
(a)
(b)
= 5 0 Q,
= 5 0 V,
= 50 Q ,
= 40 Q , and R^ = 10 Q.
0.004
-0 .0 0 1
-0 .0 0 2
-0 .0 0 1
0.001
-1
W rite
-0 .0 0 2
0.004
-1
-1
nodal
equations,
0 T 4 '
O'
yB
0
^CB
to resistances.
= 0.02
= 10 Q.,
=
= 100 Q, R^ = 100 Q,
= 50 Q ,
(b)
(c)
(a)
Determine Vq
(b)
(c)
nodal equations.
(d)
^AB and 4 .
(e)
(0
sources.
AN SW ERS (D) IN R A N D O M O R D E R : 12,
0.24, 9.6, 0.16
the sources.
Figure P 3 .2 6
in
148
terms o f
= 4 kQ,
30. For the circuit o f Figure P3.30,
= 100 Q,
mS,fj. = 4 S,
= 1
= 4/3 kQ,
= 0.75
= 160 V, and
= 40 mA.
= 300
(a)
Specify V^.
2 A,
= 2 A, Vj2 ~
nodal analysis
to find all node voltages, the current
, the
(b)
= 20 Q,
= 20 Q, G4 = 0.09 S,
(c)
(d)
(a)
Determine Vq
(b)
(e)
and
(f)
(c)
(d)
(e)
Figure P3.30
using
W rite
= 100 V,
modified
1^2
nodal
= 1 A, and
equations
(b)
(c)
the equivalent
Reference node
Figure P3.31
Figure P3.33
AN SW ERS T O (C) IN RA N D O M O R D E R :
40, -2 5 ,3 8 .7 5 ;/ ?
= 12 Q .
149
SIN G LE LOOP-EQUATION
PROBLEM S
34. In the circuit o f Figure P3.34,
= 50 Q, and
= 0.5. If
= 400 Q,
= 50 V, find 4 ,
sources.
R,
= 2 0 0 Q, R^ = 300 Q, R2 = 500
Figure P3.37
C H EC K : /j = 100 mA.
(b)
C H EC K S: R^^ = 2 50 Q, and
= 8 watts.
= 4 00 2,
= 56 V,
= 100 mA
independent source.
12.5 watts.
= 200 V
Figure P3.38
Figure P3.36
C H EC K : 7, = 4 mA.
150
(b)
(c)
Suppose
= 2 50 Q,
= 5 0 0 Q,
20 Q and R2 = 80
and 13 . Suppose
5 00 Q , plot R
s 13 s
= 250 Q and
as a function o f
,0
es o f the battery?
Pi,
0 '
C H EC K : P.s\
Figure P3.39
C H EC K : Power absorbed by resistors is 15
watts and R^^ > 4 50 Q .
= 100
/j and
(b)
(c)
Figure 3.42
Figure P3.40
150
-4 0
-1 0 0 '
-4 0
140
= -2 0
-1 0 0
150
20
h'
TOO
151
(b)
(c)
CHECKS: V^2 = ^0 V,
= 40 Q.
-8 0
- 1 0 \h'
-3 0
130
-5 0
-1 0
-5 0
70
[*3.
n
=
-V2
0
Figure P3.45
Find
*
46. Reconsider Problem 3.45. Let
= 590 V
(c)
45. Figure P 3.45a shows an electric locomotive
(d)
for
, ^2
H
Determine the two motor currents.
Determine the power delivered by
each of the 660 V sources.
660VVf
R
R
Figure P3.46
0.15 Q. Using m e ^ analy4 7 . Reconsider the Problem 3.5 and the circuit
sis find
(a)
l=
the currents
and I 2
152
node voltages
which
= 40 V and V^ 2 = 20 V. Write a set o f
three loop equations by inspection. Refer to
n\
> and
= 9 kQ,
= 9 kQ,
= 18 kQ,
=
and /2.
= 3000 Q, and
20 mA.
(a)
(b)
(c)
^BO ^CD
K iD'
Find the power delivered by each o f
IQ
the sources.
parameters.
(b)
and
Q,
= 40 Q,
A, and
= 20 Q,
V.
= 100
= 80 Q,
(c)
(d)
(c)
Find
(d)
r^
and V^.
pendent sources.
(e)
r )
n
Figure P 3 .4 9
Figure P 3.51
153
K2 =
V-
(b)
= 60 f i,
80 Q,
1^2
= 80 Q,
= 1 A, and V^j =
40 V.
(c)
Compute
(d)
and
(b)
Figure P3.53
(c)
30 3,
0.25 A, and
= 40 Q, 7?2 =
= 20 2,
= 10 2, 7^2 =
= 60 V.
>mv,-
Q,
1 + ??'
??
-1
R3
????
-1
(b)
If
y s -R ih n
A'
=
0
0
Q, 7?4 = 100 2,
= 100
mined entries.
'7?2 + ? ?
= 400
= 150 V, / = 0.5
154
C H EC K : /j = 0.4 A and
= 50 V.
(b)
If
= 4 0 0 ^ ^ ^ 2 = 2 0 0 V,i?i = 3 0 Q ,
= 20 Q,
= 270 Q,
= 80 Q, T?5
(d)
R3
R.
-VESA
's <-----N/N^
Rs
Figure P3.58
CHECKS: V, = 265 V and
watts.
= -3 9 7 .5
C H A P
'
"
L in D lif ie r
The Operational Amplifier
Amplification o f voice allows announcers at sports events to convey their comments on the playby-play action to the crowd. At concerts, high-powered amplifiers project a singers voice and the
instrumental music into a crowded auditorium. Electronic amplifiers make this possible. One o f
the simplest and most common amplifiers is the operational amplifier, the subject o f this chapter.
The word operational, though, suggests a purpose beyond simple amplification. Often one must
sum signals to produce a new signal, or take the difference o f two signals. Sometimes one must
decide whether one dc signal is larger than another. The operational amplifier is operational pre
cisely because it can be configured to do these things and many other tasks, as we will see later in
the text.
2.
3.
4.
SECTIO N HEADIN GS
1.
2.
3.
4.
5.
Introduction
The Idealized Operational Amplifier: Definition and Circuit Analysis
The Design of General Summing Amplifiers
Saturation and the Active Region of the Op Amp
Summary
Terms and Concepts
Problems
156
1. IN TRO D U CTIO N
Chapters 1 and 2 defined and discussed independent and dependent voltage and current sources.
Chapter 3 investigated the nodal and loop analysis o f resistive circuits containing such sources.
Ofi:en, dependent sources supply energy and power to a circuit, making them so-called active ele
ments. O n the other hand, resistors are passive elements because they only absorb energy. Circuit
models o f real amplifiers (see Examples 3.3 and 3.4 with associated Figures 3 .7 and 3.8, respec
tively) contain controlled sources that underlie their analysis and performance evaluation. Indeed,
the VCVS is the core component o f the operational amplifier (op amp), the main focus o f this
chapter. Thus, the op amp is an active circuit element whose analysis is done with the techniques
o f Chapters 1 through 3.
A real op amp is a semiconductor device consisting o f nearly two dozen transistors and a dozen
resistors sealed in a package from which a small number o f terminals protrude, as shown in Figure
4.1(a). Despite its apparent internal complexity, advances in integrated circuit manufacturing
technology have made the op amp only slightly more expensive than a single discrete transistor.
Its simplicity, utility, reliability, and low cost have made the op amp an essential basic building
block in communication, control, and the instrumentation circuits that can be found in all under
graduate EE laboratories.
Top View
Balance 1 [
TO-5
Dual-in-line
Inverting
input
Noninverting ^ [
input
E- 4 [
3 6 Output
H 5 Balance
(a)
Inverting
E+-
input
V-
Output
Reference
node
Noninverting
E-
input
(c)
(d)
FIG U RE 4.1 (a) Typical op amp packages; (b) typical terminal arrangement of an op amp package;
(c) dual power supply notation; (d) essential terminals for circuit analysis.
Figure 4.1 (b) shows a typical arrangement o f terminals for a dual-in-line op amp package. Th e ter
minal markings and the symbol shown in Figure 4.1(b) do not appear on the actual device, but
157
are included here for reference. In Figure 4.1(b), the terminal labeled N C (no connection) is not
used. The E+ and E- terminals (Figure 4.1(b)) are connected to a dual power supply, illustrated in
Figure 4.1(c), where
adequate voltage is required for proper operation. The three terminals in Figure 4.1(b) marked
inverting input, non-inverting input, and output interact with a surrounding circuit, and
correspond to V,
and Vq in Figure 4.1(d). The two terminals labeled balance or ojf-set have
importance only when the op am is part o f a larger circuit: resistors o f appropriate values are con
nected to these terminals to make sure the output voltage is zero when the input voltage is zero.
This balancing process is best discussed in a laboratory session.
This chapter sketches the basic properties o f the op amp: just enough to understand some o f the
interesting applications. The ideal op amp model and the saturation model are described. Using
these models and the principles o f analysis covered in Chapters 1 through 3, we then analyze the
behavior o f some widely used op amp configurations. These application examples hint at the
importance o f the op amp and furnish motivation for the study o f electronic circuits.
Several o f the examples include a SPIC E simulanon o f the circuit being analyzed. SPIC E is a
sophisticated circuit simulation program. Behind the user-interface, SPIC E uses complex models
o f the real operational amplifier. Our purpose in using SPIC E simulation is to verify or test the
theoretical analysis set forth in the examples. W hat we show is that the simplified theoretical
analysis provides a very good approximation to the actual circuit behavior represented in the
SPIC E simulation results. Industrial circuit designers often use SPIC E to visualize the expected
behavior o f very complex circuits. Later chapters cover some o f the more complex op amp appli
cations.
FIG U RE 4.2
One possibility for analyzing op amp circuits is to represent the op amp by one o f the simplified
models shown in Figure 4.3 that do not account for saturation effects. The first model o f Figure
4.3(a) consists o f an input resistor,
an output resistor,
O f practical import is the idealization o f this model (Figure 4.3(a)) to the one o f Figure 4.3(b) by
158
(1) letting R-^ become infinite, setting up an open circuit condition at the input terminals; (2) let
ting
become zero, making the output voltage o f the op amp equal to that o f the V C V S; and
(3) letting the gain A approach infinity. These conditions are idealizations because (1) with R-^
infinite, there is no loading to a circuit attached to the input; (2) with R^^^ = 0, the full output
voltage appears across any circuit connected to the output; and {5) A ^
leads to a simplifica
tion o f the associated analysis. These conditions, stated below as equation 4.1, define the so-called
ideal operational amplifier;
-> 00 (infinite gain)
R-^
Rout
(4.1a)
(4.1b)
(4.1c)
Rest of Circuit
Rout
+
+
V
A(v - V
(a)
FIG U RE 4.3
To see how this idealization simplifies op amp circuit analysis, consider an equivalent set o f conditions for the ideal op amp, called the virtual short circuit model:
(4.2a)
(4.2b)
v^ = v_
(4.2c)
From Figure 4.3(b), the conditions that = 0 and i_ = Q follow directly from the open circuit con
dition at the input terminals. The condition that
= v_ (hence, the term virtual short circuit)
will be discussed later, but occurs because A ^ co, forcing {v^ v j
0.
159
The recommended way to analyze circuits containing op amps is to replace any ideal op amp by the
model o f Figure 4.3{b), the virtual short circuit model o f equation 4.2. Th e following examples illus
trate the use o f the virtual short circuit model.
EX A M P L E 4 .1 . This example investigates the inverting am plifier o f Figure 4.4, which is used in
a wide range o f commercial circuits. The objective is to compute
Rjs and
in terms o f
Rf
So l u t io n
Step 1. Compute
R
Step 3. Compute iy. Again, since v_ = 0, the voltage across Rf is
if = R
Step 4. Relate the currents i-^ and ip and substitute the results o f Steps 2 and 3. From KCL, i^^ i_
+ ir= 0. From the properties o f the ideal op amp, i_ = 0, in which case, ir= -i-^. This imphes that
R*
V
Hence, the voltage gain relationship o f the inverting op am circuit is
Rf
V ou t= --^ ^ in
Rin
(4.3)
Equation 4.3 shows that the input and output voltages are always o f opposite polarity, hence the
name inverting am plifier. One also observes that by choosing proper values for Rjr and R^ a volt
age gain o f any magnitude is possible, in theory. In practice, other factors limit the range o f obtain
able gains.
160
Exercises. 1. Find
lOOkO
25kQ
-s/ W '
50 mV
lookn
= - 200 mV.
A few remarks are in order. Op amp configurations in which one o f the input terminals is ground
ed, as is the non-inverting terminal in Figure 4.4, are said to operate in the single-ended mode. The
input terminal can be grounded directly or through a resistor, as in Exercise 2 above. Also, since
v_ =
or v_-v^ = 0, the terminals are virtually short circuited even though there is no hard-wired
direct connection between them. This condition is called a virtual short circuit. Further, if one
o f the terminals is grounded, then the other terminal is said to be virtually grounded, as is the
case in Figures 4.4, 4.5, and 4.6. Specifically, in Figure 4.4, there is a virtual ground at the invert
ing input terminal.
The next example continues the investigation o f the ideal inverting amplifier for the two-input,
single-output op amp circuit o f Figure 4.7. The solution again makes use o f the virtual ground
and virtual short circuit properties o f the ideal op amp.
161
and v^2 Rf
+
V
FIG U RE 4.7 Inverting (ideal) amplifier with rwo inputs for which
Rf
Rj
v,=-^v,|V ,2.
is
R^
Step 2. Compute ij-. Again, since v_ = 0, the voltage across Rj- is
'/ = R
/
Step 3. Relate the currents /j, ^2
H i_ + ijr = ^- From the properties o f the ideal op amp, i_ = 0, in which case, ijr = -(/j +
Hence,
Rf
Rj
(4.4)
162
= 25 kD and
and ^2 so that
= 50 kD.
''out
V.
So l u t io n
Step 1. Compute
and v_. Since the + terminal is connected to the input voltage source,
= v-^.
is
Zj, has reference direction different from the passive sign convention. Hence, from Ohms law,
lf =
From O hm s law,
Vout - Vin
Step 4 . Relate the currents z'j and ip and substitute the results o f Steps 1 and 2. From KCL, z'j - z_ +
Zyr= 0. From the ideal op amp property o f equation 4.2, z_ = 0, forcing z^= -Zj. This implies that
Rf
/?,
163
Rf\
(4.5)
1+ ^
R i)
/
R \
From equation 4.5, the voltage gain is greater than 1, i.e., i 1 + - i - i > 1, and
I
}
have the same polarity; the circuit is naturally called a non-inverting amplifier.
and
and
always
EXA M PLE 4 .4 . This example analyzes the ideal general di^Ference amplifier circuit o f Figure
4.9. We show that
Kf
^s 2 - - r ^ s \
R\
In a basic difference amplifier, the output is the difference o f two input voltages. For the gener
al difference amplifier o f this example, the output is a difference o f the scaled input voltages,
and a^.
G 2 { v + - v , 2 ) + GgV+=Q
164
Solving for
leads to
(4.6a)
Step 2. Write a node equation at the inverting input terminal o f the op amp. Recall v_ = v^. T he sum
o f the currents leaving the - node satisfies
G i ( v+ - v, ] ) + G ^ ( v+ - v,) = 0
Thus,
Cl
^out =
1+
(4.6b)
Step 3. Combine Steps 1 and 2. Substituting equation 4.6a into 4.6b yields
(4.7a)
G t + Go
or, in terms o f resistances,
[R g + R2 j
(4.7b)
^s2
out
= K {v , 2 -V si)
with K = I and for an arbitrary K > 0 , Rj- = KR^ and R^ = KR2 fits the bi
bill.
A N SW ER:
w ith
165
and
=
= v_. From the proper
= v-^.
Exercise. Compute the power delivered by the source in Figure 4 .1 0 and the power delivered to
the load R^.
^
AN SW ER: 0 and
R,
The circuit o f Figure 4.10 is called an isolation or buffer amplifier, because no current is drawn
from the source
maintaining
However, the op amp does supply current (and power) directly to the load by
under the condition that
Figure 4.11 shows a SPIC E simulation that verifies the behavior arrived at in Example 4.5. Here
a dc voltage sweep, Q <v-^< 12 V, was input to a highly accurate SP IC E model o f a Burr Brown
741 connected to 10 V power supply. Observe in Figure 4.11 that the output follows the input
up to the 10-volt value, after which, the output remains at 10 V despite increased input values.
This non-ideal phenomenon, called saturation, is due to the power supply voltage level and is dis
cussed in Section 4.
166
(V)
+0 .0006+000
+2 .000
+4.000
+6.000
V2
+8.000
+10.000
+12.000
Exercise. Find
for the circuit o f Figure 4.12, the power supplied by the source
=v,.0.
1 2 x 10-
and the
167
^out -
+ 2 'a 2 ) +
(4.8)
where the constants ay>0 and |3>0. The inverting and non-inverting amplifier configurations
(Examples 4.1, 4.2, and 4.3), as well as the difference amplifier configuration o f Example 4.4, are
special cases o f equation 4.8. W ith a little cleverness, it is possible to design by inspection an op
amp circuit whose input-output characteristic is precisely equation 4.8. The op amp circuit o f
Figure 4.13 having the four inputs
V^2 > ^hv ^bl accomplishes this. The circuit looks ordi
nary except for the presence o f one additional conductance, AG, incident on the inverting termi
nal o f the op amp. T he dashed lines in Figure 4.13 are present because this conductance may or
may not be needed. Computation o f the values o f AG and
below.
b2 0 s / \ / V
G
168
Design Step 2. Prototype design continued: Computation ofG^ andhr tS.G so that the total conduc
tance incident at the inverting terminal o f the op amp equals the total conductance incident at the non
inverting terminal.
To achieve this equality, recall that in design Step 1, Cy= 1 S,
S, and
S. Define
a numerical quantity
^ = (1 + a i +
) - (^1 +
Similarly,
multiply all resistances (divide all conductances) incident at the non-inverting terminal o f the op
amp by
It is permissible to choose
So l u t io n
Step 1. Prototype design. Using Figure 4.13, choose Gy- = 1 S, G^j = 7 S, G^2 = 3 S, G^, = 2 S, and
G^2 = 4 S .
Step 2. Equalization o f total conductances at inverting and non-inverting terminals. Since 8 = (1 +
7 + 3) - (2 + 4) = 5 > 0, set AG = 0 and G^ = 8 = 5 S. The circuit in Figure 4.14(a) exemplifies
the prototype design.
Step 3. Scaling. To have practical element values, let us choose
to a design with resistances Rjr= 100 kQ, 7?^, = 14.28 kQ,
25 kD and R = 20 kQ.
= 50 kQ, R^^ =
169
14.28 kQ
V3, o------
lOOkO
25 kQ
(a)
(b)
FIGURE 4.14. (a) Prototype design of equation 4.9; (b) final design
after scaling with
= lO^.
So l u t io n
Step 1. Prototype design. Again, using Figure 4.13, choose
(4.10)
S, and G^2 = 5 S.
Step 2. Equalization o f total conductances at inverting and non-inverting terminals. 8 = (1 + 2 + 4)
- (7 + 5) = - 5 <
0; set C = 1 S, AG = |8| + G = 5 + 1 = 6 S. This prototype design is given in
A
o
Figure 4.15(a).
Step 3. Scaling. To have practical element values, let us again choose
ing leads to a design with resistances Rjr = 100 kQ, R^j = 50 k fl, R ^2 = 25 kQ,
= 14.28 kQ,
R^j = 20 kQ,
- 100 kQ, AR = 16.67 kQ. The final design is set forth in Figure 4.15(b).
16.67kn
50 kQ
V3, o------v X / X . -
lOOkQ
25 kQ
14.28 kQ
lOOkQ
'^b2 O--20 kQ
(a)
(b)
FIGURE 4.15 (a) Prototype design of equation 4.10; (b) final design after scaling with
= 10^.
170
o f one resistor.
AN SW ER: In prototype design, AG = 5 mho.
2. Design a difference amplifier so that
- v^-^, with
= 10 kD.
At this point, the reader may wonder how this simple procedure is derived. The derivation o f this
procedure is beyond the scope o f the Hght edition^ Th e interested reader is directed to the 2nd
edition o f this text.
Exercise. 1. Find
AN SW ER:
in terms o f
_
' out -
G,
G,
Gj
G3
,v2
2. Find t',,
AN SW ER:
- 7z>2 +
6K U
(a)
FIG U RE 4 .17
171
= v_ and
inverting amplifier o f Example 4.1, this led to the very simple gain formula,
'^out
_
Rh,
Thus, as the input voltage increases, the output voltage increases proportionately. For real circuits,
this proportional relationship holds only when
that is associat
ed with the power supply voltage. Intuitively speaking, an op amp cannot generate an output volt
age beyond that o f its power supply voltage, typically less than or equal to 15 V. W hen the V^^limit is reached, further increases in the magnitude o f v-^ produce no change in the value o f
This behavior is called saturation.
To explain this saturation behavior, we refer to Figure 4 .1 8 . In Figure 4.18,/'(v^ - v j represents
a nonlinear controlled voltage source, as opposed to the linear relationship A(v^ - p_), shown in
Figure 4.4(a). However, because the op amp functions more or less linearly until reaching its sat
uration limits, we can approximate f(v^ - v_) by the three-segment piecewise linear relationship
shown in Figure 4.19(a), wherein the saturation effects are captured by segments II and III. One
observes that when
. If
and when
mV; if A is infinite, as in Figure 4.19(b), then saturation occurs when \v^ > 0.
FIG U RE 4.18 Practical op amp model with a nonlinear controlled voltage source.
The linear r e l a t i o n s h i p , - v^=A{v^ - v J , holds for segment 1 in Figure 4.19(a), which is said
to be the linear region or active region o f the op amp, denoted by
172
Typical values o f finite A range from 10'^ to 10*^. The active region is the ordinary region o f oper
ation. In the active region, the op amp provides a very high (open loop) voltage gain A, the slope
o f segment I. The phrase open loop gain means that there is no connection through a wire, a
resistor, or some other device back to the input terminals.
Models o f the three operating regions o f the op amp are summarized in Table 4.1.
A f(v^-v)
V
Positive
Saturation
/
Active /
Region
= V -V
Negative
Saturation
-V
(b)
FIG U RE 4.19 A piecewise linear (three-segment) curve for the op amp that specifies the active
and positive/negative saturation regions of operation: (a) finite gain A, and (b) (ideal) infinite gain A.
TABLE 4.1 Operating Regions o f the Op Amp with Associated Models
N ame
of
R e g io n D efining E quations
Vcl =
Active
f(Vd)
A
and
*^sar
II
Positive
saturation
Vsa,
A
and
sat
III
Negative
saturation
<and
I dealized C ircuit M od el
173
The use o f a three-segment curve in Figure 4 .1 9 is different from the techniques o f earlier chap
ters. The operating point,
input is small, one reasonably assumes the operation is in the active region, segment I. However,
when the input magnitude is large, one must guess and check to determine the appropriate oper
ating region. For example, should the guess be incorrect, then the model for one o f the other
regions must be used and the analysis repeated until a valid solution (and operating region) is
obtained. The following example illustrates the approach.
E X A M P L E 4 .8 . The purpose o f this example is to illustrate that an op amp may operate in any
o f three regions and also to illustrate that the determination o f the region o f operation using the
guess and check method. Recall the inverting am plifier o f Figure 4.5. Suppose
= 50 kQ,
V; (b)
= 10 kQ and
= 4 V; and (c)
= 15 V. Find
and
is infinite,
= 0.5
So l u t io n
(a) Assume the op amp operates in its active region. From equation 4.3 in Example 4.1, the out
put voltage is
vout
= - ^ v , m, = - ^JQx 0 . 5 = - 2 . 5 V
Since |-2.5| <
= - 2 .5 V and
v^ = 0 are valid.
(b) W ith v-^ = 4 V, assuming operadon in the active region,
^
50
= -----= -------------- x 4 = - 2 0 V.
10
Rl
invalid, but does suggest operation in the negative saturation region. The negative
saturation model o f Table 4.1 yields the circuit o f Figure 4 .20 in which
= - 1 5 V.
R'
50
= 25 V
This result suggests that the op amp is really operating in the positive saturation region. Using the
positive saturation model o f Table 4.1, Figure 4.21 shows the proper circuit configuration with
174
^out ~ 15 V. As in the previous case, by writing and solving a single node equation at the invert
= 1.667 V. In this case and in case
0, as we were not in the active region o f operation, and it was necessary to change
< 4 V,
is an adequate input to demonstrate the saturation effects. In the SPIC E simulation, an accurate
model for a 741 op amp manufactured by Burr Brown was used. T he resulting dc transfer curve
is shown below in Figure 4.22.
Lin/Decarlo E xI-D C Transfer-4
Output voltage
(V)
-4.000
- 1.000
+ 0 .0006+000
Vin
+
1.000
+3.000
+4.000
V(IVM)
From this curve, one can see that the op amp saturates for input voltages v-^ such that
> 3,
and the op amp operates in its linear region whenever \v-^ < 3. As hoped, the simplified three-seg
ment model in Table 4.1 yields very good results in all regions o f operation relative to the realis
tic SPIC E simulation.
O ne can conclude from the above example that for the purpose o f faithfully amplifying an input
signal, the input should not be so large as to drive the op amp into saturation. Driving an op amp
into saturation distorts the output signal relative to the input. O n the other hand, for some spe
cial applications, such as the com parator, saturation is precisely the property to be utilized. Figure
17 5
4.23 shows two comparator circuits. A com parator circuit compares tlie input voltage v-^ with a
reference voltage Vygjr{or some multiple o f
one for
EXA M PLE 4 .9 . For the com parator circuits shown in Figure 4.23, each op amp has infinite gain
and a saturation voltage
= 1 5 V"^:
v -^ relationship for the comparator o f Figure 4.23(a).
(a)
Find the
(b)
=
Vin O------s / W 2 0 kQ
''re f O --------
= -2 0 V
80 kn
1 .
1 .
(a)
(b)
FIG U RE 4 .23 Two comparator circuits that are used to determine when an input
voltage is above or below a reference voltage.
S
o l u t io n
(a) For
^out ^ ~^sat = - 1 5 V. Similarly, for v-^ < - 5 V, the voltage v^ = v^ - v_ = -5 - v-^ > 0, and hence
^out = ^sat = 15 V.
(b) By the fact that no current flows into the input terminals o f Figure 4.23(b), using nodal analy
sis, we have that
V
- V
20x10^
80x10^
m which case,
- 4 + 0.8v,
For
- v_ = 0 - {-A +
= 15 V.
176
To verify this analysis, the circuit o f Figure 4.23(b) was simulated in B2Spice using a Burr Brown
741 op amp model. T he results o f the simulation are given in Figure 4.24. The theoretical analy
sis based on the simplified models o f Table 4 .1 shows a very good match with the more realistic
SPIC E simulation results.
example4.8-DC Transfer-6
Output voltage
(V )
-1 0 0
+2.000
+3.000
V(IVM)
amp is in positive saturation. Then find the range o f v^ 2 fof negative saturation.
AN SW ER:
when
< -^ V, and
=-
when
V , C3------- v N / \ - -
75 kO
-o
+
25 kQ
1 .
FIG U RE 4.25
177
5. SUM M ARY
This chapter has introduced the operational amplifier and a number o f practical circuits that uti
lize this new device. These circuits include the inverting and non-inverting amplifiers, the buffer
amplifier, the difference amplifier, and the general summing amplifier. W ith regard to the gener
al summing amplifier, a simple design algorithm is described and exemplified. The analysis o f
these circuits builds on the definition o f an ideal op amp, meaning that, when properly config
ured, no current enters the input terminals and the voltage across the input terminals is zero; these
properties are referred to as the virtual short circuit model o f the op amp, i.e., the ideal op amp
has infinite input resistance, zero output resistance, and an infinite internal gain, A. (See equations
4.1 and 4.2.) Practically speaking, the gain A, is not infinite, but ranges between lO'^ and 10*^.
After exploring properties o f the ideal op amp, we discussed the phenomena o f output voltage sat
uration. By introducing output saturation, the ideal model o f the op amp gives way to a more real
istic one, characterized by three regions o f operation, each having its own ideal model, as set
forth in Table 4.1. In practical design and applications, output saturation is either to be avoided
or utilized to some advantage, as in the case o f the comparator circuit studied in Example 4.8. For
a faithful amplification o f an input signal, saturation is to be avoided.
<
and
^out ^
appropriate constant k, often taken as 1.
General summing amplifier: an op amp circuit having the input-output relationship
constant a - and p ..
Ideal op amp: An operational amplifier with infinite input resistance and infinite open-loop gain.
-
+ ^ n ^ a r) +
+ - + ^ m ^ bn )
p o s itiy c
Inverting amplifier: An operational amplifier connected to provide a negative voltage gain at dc.
Linear active region: In the op amp output vs. input transfer characteristic, the region where the
curve is essentially a straight line through the origin is called the linear active region.
Non-inverting amplifier: An operational amplifier connected to provide a positive voltage gain
at dc.
Open-loop gain: The ratio o f the output voltage (loaded, but without any feedback connection)
to the voltage across the two input terminals o f an op amp. The slope, p, o f the straight
line in the active region o f an op amp is the open loop gain under no load condition.
When a load
+ R^, where R^ is
178
Operational amplifier (abbreviated op amp): A multi-stage amplifier with very high voltage
gain (exceeding 10"^) used as a single circuit element.
Passive elements: a circuit element that cannot deliver net power to a circuit such as a resistor.
Saturation regions: In the op amp output vs. input transfer characteristic, the region where the
curve is essentially a horizontal line is called the saturation region. There are two such
regions: one for positive input voltage, and the other for negative input voltage.
SPICE: Acronym for Simulation Program with Integrated Circuit Emphasis. It is a very sophisti
cated software tool for simulating electronic circuit behavior.
Virtual ground: When an ideal op amp has one of its input terminals grounded, and is operat
ing in the active region, then the other input terminal is also held at the ground potential because of the virtual short effect (see below). Such a condition is called a virtual
Virtual short circuit: When an ideal op amp is operating in the active region, the voltage across
the two input terminals is zero, even though the two terminals are not hard-wired togeth
er. Such a condition is called a virtual short circuit (in contrast to a physical short circuit).
Voltage follower: A voltage-controlled voltage source with gain equal to 1, often utilized to sep
arate stages of amplification in a multi-stage amplifier device.
^The circuit proposed in this section is a modification of one proposed in W. J. Kerwin, L. P. Huesman, and R.
W. Newcomb, State-Variable Synthesis for Insensitive Integrated Circuit Transfer Functions, IEEE Jr.
of Solid State Circuits, Vol. SC-2, pp. 87-92, Sept. 1967. The modification consists of an additional
resistor, which greatly simplifies the design calculations and was published by P. M. Lin as Simple
Design Procedure for a General Summer, Electron. Eng., vol. 57, no. 708, pp. 37-38, Dec. 1985.
'
^ See Linear Circuit Analysis by DeCarlo and Lin, 2nd edition, New York: Oxford University Press, 2002.
^ Any of the SPICE or PSPICE software programs available by a variety of vendors will suffice to obtain the indi
cated curve.
^ An op amp and a comparator as seen in a parts catalog are essentially the same, except that the comparator device
has a modified output stage that makes it compatible with digital circuits.
179
Problems
ANALYSIS USING IDEAL OP
AM P M O D EL
1. Consider the inverting amplifier circuit o f
Figure P 4 .1, in which
= 4 V.
(a)
If
= 2 kD, find i ?2
dehvered to
(b)
= 100
Figure 4.3
is 4 W.
so
= 40 Q.
(b)
Figure P 4.1
Check: (b) 1500 <R^< 2000, and
= 10 m W
(b)
load.
+
Rs
= 2 kH
' l >r
(b)
Figure P4.4
SCRAMBLED ANSWERS: (a) 0.256, 3.2, 0.08
5. Figure P4.5 contains three circuits that
explore loading and the elimination o f loading
effects using either a dependent source or an
equivalent buffering op amp circuit.
(a)
For the circuit o f Figure P4.5(a), com
pute
and
in terms o f v^. Observe
Figure P4.2
Check: (b) 1500 <
< 3000
3. For the circuit o f Figure P4.3, find the volt
age gains,
G,
andG , =
^ Vin
in terms o f the literal resistor values.
V,
180
7. (a)
<,v,
320 n<
(c)
12 kD,
= 4 kQ, Rj = 1 k2,
=
= 1.5
-o
+
V r + ^
as a function
80 n
80 n
240
(a)
80 n
son
-o
+
240 n
320 n<
(b)
80 0
son
-v \ ^ -
-o
+
240 n
320 n
Figure P4.5
AN SW ER: (a)-V /
3
0 5V
(b)
= -2t^^j 8. (a)
">^a-
= 20. If
= 600 mV.
so that
(c)
out
20
Finally, suppose Rj =
and find
out
= 200 mV and
out
20
181
-0 -,
sn
and v^2 -
(a)
Find
(b)
If
= 250 m V and v^ 2 - 500 mV,
find the power delivered to the 1 kQ
in terms o f
load resistance.
Figure P4.8
A N SW ERS (in random order): 14 kO, 8 kf2,
10 k n , 18
lOkn
lokn
lokn
5kn
2kQ
Find
and ^2 so that
1 kn<
= -lOr^^j -
'
20^,2(b)
- 2 0 0 m V and
Figure P 4.11
C H EC K : (b) Pj^ = 12.25 m W
12.
(a)
For
the
circuit
o f Figure
= 2 V and
1 kn
resistor.
(b)
= 4 V and
Figure P4.9
= 0.1 watt
(c)
(b)
3R
= 1 V.
Suppose 6R^ = i ?2
v^ = 2Y . Find
and R2 so that the power delivered
to
= 1.25 k il is 2 watts.
Figure P4.12
13. (a)
Figure P 4.10
o f R so that
= 10 V.
182
(b)
= 2.5 V.
12R
r e
R,
-V S/V
2R
5R
-O
+
3R _ d
(a)
Figure P 4 .15
C H EC K : (a)
12R
2R
rO
5R
terms o f
in
V, v^2
2.5R
(b)
Figure P4.13
AN SW ERS; (a) 6R-, (b) 3R
14. For the circuit in Figure P4.14, the input
voltages are
= 2 V,
= '1-5 V, and
=2
V.
(a)
Find
(b)
2R
0.5 R
+ Av^2 -
1.5R
-o+
0.75R
0 - .
in terms o f R so that
and
Figure P4.14
C H EC K :
i = 0.9 mW,
= 0-20667
watts
15. For the circuit o f Figure P4.15, the input volt
age
= 5 V, and the input voltage
(a)
(b)
If
= 8 R and Rj - Ry find
If R^ = 8R and R^ = ARy find
(c)
If /?2 =
that
find
= lOV.
in terms o f R so
CH ECK:
Figure 4.17
= 8R
= 8f^j +
183
Time in s
AN SW ERS:
= AR, R^ =
(0
R, = O m
Figure P4.19
NON -IDEAL OP
AM P-SATURATIO N EFFECTS
19. The op amp in Figure 4.19(a) has
V,
=15
-O+
given in
Figure 4.19(b).
for 0 < /< 6 s for v(J) in
(b)
Plot
(c)
Figure 4.19(c).
Figure P4.20
(a)
= 5 Y
R,
Find
-o (b)
and
If
= -2 -5 V and
= 15 V, find
so that no amplifier saturates.
(a)
4R
1.5R
2R
-O
Figure P4.21
AN SW ER: (a) 15 V; (b)
= 3.75 V
22. For the circuits o f Figure P 4.22, suppose R^
(b)
= 40 k n , and R^ = 120 Q.
(a)
184
(b)
(b)
compute
in terms o f v-^^.
load
(c)
in terms o f
(a)
(a)
(b)
Figure P4.24
a.
(a)
(b)
(b)
= 40 Q , and Rj = Rl = 120 Q.
(a)
Figure P4.25
AN SW ERS: (a)
3.7037 x l O 'V .
185
= 0.003i^y; (b) Pl =
{Vs^,
compute
is
and
Notice that
3 2 -Q resistor.
(c)
Your
Figure P4.27
8Q
80
-o
+
32 0 <
(c)
Figure P4.26
AN SW ERS: (a) 0.6665 V^, 0.5 V^; (b) and (c)
0.8 V^, 0.6
186
80 kO
A N SW ER:
32. Find the
comparator
Figure P4.28
29. (a)
Find the
1
1.5
of
Figure
P 4.32.
versus
relationship for
Vin >
, then
and when
Vin <
R,
then
V
= - V^sa f
'^out
(b)
^in<-^^ref , then
Now suppose
^ref ~
characteristic if
= R2 = 100 kQ and
^out versus v-^
= 15 V. Verify
Figure P4.32
G EN ERA L SUM M IN G
AM PLIFIER (ID EA L OP AM P
M O D EL)
3 3 .(a)
Figure P4.29
AN SW ERS: (b)
= 15 V if
< 2 V, and
(b)
Rf
Rf
R,
Ro
Suppose
so that
25 kO. Find
, i?2>
is the negative o f the
average o f
, V2 , and Vy
= 10 V when
< 1 V,
V l-
Figure P4.33
AN SW ER: (b) R. = 75 kO
187
relationship
relationship
+H i
- 3^.2 +
+H i
Two different designs are to be produced for
comparison and selection:
1:
. = 50 kQ.
iqq
^
39. Generalizing the topology o f Figure 4.13,
design an op amp circuit to have the input-output relationship
^^h\
^'^bl
(a)
(b)
= -4v^j + 2z;^j +
n
/ r^
^/=
4 1
relationship
^out ^
~ '^'^al ^
^'^bl
^out
^out =
- ">^al + ^^b\ +
hi
188
h .
SIM ULATION OF
C O N TR O LLED SO URCES
USING OP AM PS
44. Design an op amp circuit to simulate the
grounded VCVS in Figure P 4.44 when p > 1.
Hint: Consider the non-inverting amplifier o f
Example 4.3.
Figure P4.42
43. T he circuit o f Figure P4.43 is another mod
ification o f the basic inverting amplifier to
obtain a variable gain amplifier. Show that as
the sliding contact o f the potentiometer is
moved between the two extreme positions, the
range o f achievable voltage gain is
R
U a = 1, H
where
:-a -
H -
Rq
Rf
R,
Rp
R0n // /Ii\f
Rf
example, Rj^ may be replaced by an LED (lightemitting diode), as shown in Figure P4.49.
Then by turning the knob o f the 10-kf2 poten
tiometer, one can control the brightness o f the
LED. The current through the load is supplied
by the op amp. The potentiometer, which con
trols the brightness o f the LED , uses a low-voltage part o f the circuit. Find the magnitude o f
the LED current if the potentiometer is set at
(a)
189
4.
SECTIO N HEADIN GS
1.
2.
3.
4.
5.
6.
Introduction
Linearity
Linearity Revisited: Superposition and Proportionality
Source Transformations
Equivalent Networks
Summary
192
7.
8.
1. IN TRO D U CTIO N
Chapter 3 covered nodal and loop/mesh analyses. Node voltage or loop current calculation pro
ceeds by constructing a set o f simultaneous node or loop equations and solving them by hand, by
MATLAB, or with some equivalent software package. Few o f us will attempt a paper-and-pencil
solution o f four equations in four unknowns. Yet, MATLAB, Mathematica, or some other com
putational software program, can easily and reliably crunch numbers, relieving us o f tedious hand
calculations. Nevertheless, manual analysis in some form remains important for a deeper under
standing or insight into a circuits behavior, as well as a way to check the validity o f a program
output.
Experience teaches us that manual analysis is ordinarily practical only for small circuits.
Fortunately, the network theorems studied in this chapter and the next can often reduce seemingly
complex circuits to simpler ones amenable to manual analysis. They also provide shortcuts for
computing outputs and allow us to obtain deeper insights into a circuits behavior.
This chapter talks about linearity and superposition, which are motivated by the following ques
tions: What is the effect on the circuit output (voltage or current) o f a single independent voltage
source, say
acting alone. Acting alone means that the independent source, Vj^, has a nonzero
value, while all other independent sources are set to zero. A deactivated voltage source acts as short
circuit (see Chapter 2), and a deactivated current source acts as an open circuit (again, see Chapter
2). Is there a shortcut to computing the response if Vj^ is doubled in value?
To answer the above questions and others, our discussion begins with the important property o f
linearity. Linearity relates the values o f independent sources to a circuit output with a very com
pact equation. This equation defines the effect o f any independent source on a circuit output.
After studying linearity, we discuss two special consequences called superposition and propor
tionality. Each o f these concepts helps reduce manual computation o f responses, and each provides
insight into circuit behavior. Next, we state the source transformation theorem and show how this
method can reduce a complex circuit to a more simple form. Finally, we set forth the notion o f an
equivalent two-terminal network and then outline a proof o f the source transformation theorem.
193
2. LIN EARITY
This section investigates the circuit property o f linearity, which we introduce with a motivating
example.
E X A M P L E 5 .1 . For the circuit o f Figure 5.1, find the outputs
current /^j, and the source voltage V^2 -
jgQ .sz
I.
60 0
+
V
120Q
So l u t io n
Step 1. Find the voltage Vg. A node equation at the top o f the current source is
Vb , V B - V s l ^ r
120
60
Solving for Vg yields
= 4 0 7 ,1 + - V ,2
Here, Vg appears as a constant times /^j, plus another constant times K^2>^ so-called linear com
bination.
Step 2. Find the current I
40 / , i + - V , 2 - V ,2
linear combination.
satisfies
180 '
194
Exercise. 1. In Example 5.1, suppose the 60 Q resistor is changed to 120 Q. Find the outputs
and Vg in terms o f the sources, /^j and V^2 AN SW ER: Vg = 60/ ,, + 0 .5 V ,2 and
= 0 .5 / ,, -
and V^2-
(j
V..
= 0.251/, + 0 .5 V^2
In the above example and exercises, the desired output voltage or current was a so-called linear
combination o f the independent source values. This is, in fact, a quite general phenomena, as indi
cated by the linearity theorem below.
LIN EA R ITY TH EO R EM
For all practical linear resistive circuits, as per Figure 5.3, any output voltage,
or any cur
rent, ig, can be related linearly to the independent source values, as in the following equa
tions:
= ^i^s\ + - +
^nfsm
(5.1a)
= l Kl + - +
+ + ^mhm
(5.1b)
or
195
Linear Circuit
containing
no
independent
sources.
FIG U RE 5.3. A linear circuit driven by n independent voltage sources and m independent current
sources with outputs of
and
A rigorous proof o f the linearity theorem entails solving a set o f modified nodal or loop equations
using matrix algebra and is beyond the scope o f this text.
EXAM PLE 5.2. For the circuit o f Figure 5.4, our objective in this example is to express
ear combination o f /^j, Iq, and
lin
)v .
FIG U RE 5.4
So l u t io n
Step 1. Write nodal equation at A. For node A,
(5 .2 )
196
or equivalently,
(5.3)
Step 3. Write equations 5.2 and 5.3 in matrix form. In matrix form, the nodal equations are
G ^ -g^
G
8m
2+
/.I
(5.4)
J s l + G ^ V ,,
Vou t,
and
yields
-1
-V
Vou t,
'G l- g m
8m
G 2
h i
G2 + G 3
{G :-gJ(G 2+ G ,)
-8 m
s\
G, - I
fs2+G3V,3
It follows that
Vout--
8m
(G i-g ,)(G 2 + G 3 )
hl+-
Gi-g.
(G ,-g ,)(G
jGl-8m)G3
2+
G3)
(G i-g ,)(G
2+
-V.s3
3)
(5 3 )
if G j = 1 S, G2 = 2 S, G3 = 3 S, and
Sm^ouf Compute
= 5 S.
if G j = 1 S, G2 = 2 S, G3 = 3 S, and
to
+ ^0 ^)-
= 5 S.
^ ^.'i + ^ h i + J ^v3
197
exciting
source is removed and replaced by a short circuit using a jumper cable, i.e.,
rent source is excited by a power supply producing a constant current o f
(a)
(b)
Find
when
= 10 V and ^'.2 = 0.5 A, i.e., complete the third row o f Table 5.1.
i^ 2 (amperes)
Vout (volts)
0.2
10
10
0.5
(volts)
So l u t io n
From the linearity equation 5.1(a),
^out =
+hhl
for appropriate ttj and ^2 - From the data in rows 1 and 2 o f Table 5.1,
4 = a , x 5 + P2><0 = 5a, ^
U j = 0.8
and
10 = a , X 0 + p2 X 0.2 = O.2 P 2 => P 2 = 50
in which case,
+ 5 0 i,2
So from row 3 o f Table 5.1, if i',] = 10 V and z'^2 = 0.5 A, we have that
0 .8 X 10 + 5 0 X 0 .5 = 33 V.
(5.7)
198
AN SW ER: 60 V
2. Suppose the data in row 1, column 3, o f Table 5.1 is changed to 10 V. Find y^^^when
V and
= 50
= 0.4 A.
AN SW ER: 120 V
Comparing the development o f equation 5.7 in Example 5.3 with equation 5.1 suggests that the
coefficients Qj and P 2 can be defined as ratios:
1 =
''out
and 132=
I,-,v2
Example 5.3 and these equations suggest the algorithm for finding the coefficients in equation 5.1
by setting all inputs to zero except the input associated with the desired coefficient. This approach
is sometimes impractical. It is not always possible to set an independent source voltage or current
source to zero: imagine turning off a generator for downtown Manhattan to obtain a coefficient.
T he following example illustrates an alternate approach.
E X A M P L E 5 .4 . Consider Figure 5.6, which has two inputs
and
output
Table 5.2
lists measurement data taken in a laboratory. Row 1 ofTable 5.2 lists the nominal operating con
ditions o f the circuit. Rows 2 and 3 illustrate measurements in which one source has its value only
slightly changed (although the change may be arbitrary) while keeping the other source value the
same. From the linearity theorem, we know
find
to complete row 4 ofTable 5.2.
+ ^2^s2- Compute
199
i^ 2 (amperes)
hut (amps)
0.25
-1
5+0.1
0.25
- 1 .0 3
0.25+ 0.05
- 0 .9
15
0.5
????
(volts)
So l u t io n
From rows 1 and 2 o f Table 5.2,
- 1 = ttj X 5 + P2 X 0.25
(5.8a)
(5.8b)
and
(5.8c)
Pa = 2
Equation 5.1 for the given data has the linear form
(5.9)
Hence, for row 4 o f Table 5.2, we have that
2,, = - 0 . 3 x 1 5 + 2 x 0 . 5 = - 3 .5 V
200
i^2 (niA)
20
100
15
22
100
15.9
20
110
15.6
28
80
???
AN SW ER: 17.4 A
As a final comment on linearity, we note that by simply using the data o f rows 1 and 2 o f Table
5.2, one can solve for the coefficients by solving simultaneous equations. Specifically, using the
data o f rows 1 and 2 ofTable 5.2, we have the following matrix equation
5
0 .2 5 ' i
5.1
0.25
.^1.
-1
-1 .0 3
5.1
.A .
0 .2 5 '
0.25
-1
-1
- 1 .0 3
"
= -4 0
'0 .2 5
-5 .1
- 0 .2 5 ' -1
5
-0.3'
-1 .0 3
Exercise. Find the unknown entry in Table 5.4 after finding a , and (3j in the equation
+ Pl^.2TABLE 5.4. Two Sets of Measurements of a Linear Circuit
(volts)
AN SW ER:
ia (itiA)
^out
10
100
15
20
100
20
30
150
???
= 0.5z^,i + 100/^2
201
y=
( 5 . 10)
+ ... + au
whether it be voltage or current. A special consequence o f the linearity principle is the superpo
sition property. Equation 5.10 says that the total response jy is the sum o f the responses '"aju".
Each a-u is the response o f the circuit to u- acting alone, i.e., when all other independent
sources are set to zero. Although implied by linearity, this property is so important that we single
it out.
and
with output
as shown in
Rj = 2 Q , R 2 = 2.5
So l u t io n
Step 1. Fini/ the contribution to
and R^ = 10 Q.
W ith V^2
= 0, the equivalent circuit is shown in Figure 5.8(a). Here, the 2.5 Q and 10 Q, resistors are in par
allel, yielding an equivalent resistance o f 2 = 2.5 x 10/12.5 f i. By voltage division,
2 +2
202
+
Vo^ut
due to
= 0;
= 0.
W ith
= 0,
the equivalent circuit is shown in Figure 5.8(b). Here, the 2 Q and 10 O resistors are in parallel,
yielding an equivalent resistance o f 5/3 = 2 x 10/12 Q. By voltage division,
2.5 + 3
Step 3. Compute
^ out
= 0 5 K , ^+ 0 4K .,
Pr3 =
(youtT
= 0 . 1(0.5
Note that the total power, Pj^^, is not the sum o f the powers due to each source acting alone
because o f the presence o f the cross product term. Hence, in general, superposition does not apply
to the calculation o f power.
For dc circuit analysis, the principle o f superposition does NOT apply to power calculations.
0.5K^, + 0.25
by the
203
The next example adds a controlled source to the circuit o f Figure 5.7 and repeats the superposi
tion analysis.
=5
-^3 = 10
= 0-2
superposition theorem to
q V.A
So l u t io n
Step 1. Compute the contribution due only to
= ^out~
Setting
204
Step 3. Using superposition, add up the contributions due to each independent source acting
alone.
^out =
+ 0-4 ^',2
= /?2 = ^3 = ^ ^
Sm ~
(5.11)
S.
= 0.25i',i + 0.5t^,2
The above examples used voltage division and superposition to compute an output voltage due to
two independent voltage sources.
E X A M PLE 5 .7 . This example illustrates the principle o f superposition for the three-input op
amp circuit o f Figure 5. 12 . Show that
is the contribution o f
+ 2 . 5 K^2 +
205
So l u t io n
Step 1. Find the contribution to
due only to
Denote this output by
W ith V^ 2 = ^ 3
= 0, the circuit o f Figure 5.12 reduces to that o f Figure 5.13(a). The properties o f an ideal op amp
2R
(a)
FIG U RE 5.13
Step 2. Find the contribution to
age division,
^ -V ,2 = 0 .5 V ,2
R+R
Hence, from Example 4.3,
due only to
acting
alone is
Vout =
out
out
out
^3 ^
(5 . 1 2 )
^ouf
AN SW ER: 2 V
2. Now suppose
A N SW ER: - 1 2 V
= 8 V,
= ^s3 =
saturates at |
= 12 V; compute
206
The above examples have generated the linearity formula, equation 5.1, using superposition, i.e., the
response o f a circuit is the sum o f the responses due to each source acting alone. The technique is equiv
alent to that described in Example 5.3. However, superposition alone is not equivalent to linearity.
Linearity is equivalent to the properties o f superposirion AND proportionality, which is now stated.
However, for dc analysis, the proportionality property does N O T apply for power calculations.
- V'.., V I , * V i ,
. K,
If
and
* 2-5^2 ^ 2-5^3
. V . 0, then V ^ . - 4 ( V ) . - 4 K ) -
Exercises. For certain nonlinear circuits, the principle o f superposition may be satisfied, but pro
portionality not satisfied, or vice versa. This exercise explores these distinctions.
1.
= a,Wjj +
A very interesting and significant application o f the proportionality property occurs in the analy
sis o f a resistive ladder network. A resistive ladder netw ork is one having the patterned structure
shown in Figure 5.14, where each box represents a resistor.
-H
v.Q
207
Define K =
-----------
^
L
"
R =100
'4 W R4 = 6 Q
R =5Q
1.
tS
V =50V
-------- '
^
V.
V,
R3 = 5 0
R j= 1 0 Q
t .
V,
R, = 1 0 Q
So l u t io n
Assume Vj = 1 V. Repeatedly apply Ohms law, KCL, and KVL as follows: (Q, V and A are used
throughout):
(Ohms law)
/2 = / i =
V2 = /?2 / 2 =
V3 =
/3
/2
1 0
(KCL)
0 .1
0 .1
+ V2 =
(Ohms law)
(KVL)
(Ohms law)
= ^ = 0.4
^3
+ / 3 = 0 .1 + 0.4 = 0.5
V4 = R4 / 4 =
Vg =
\/3
X 0.5 = 3
(Ohms law)
(KVL)
+ y4 = 5
(Ohms law)
7, = "^ = 0 .5
^5
^6 = 7 4 + 7 5 = 0.5 + 0.5 =
V6 = 7?67e=5
y, = y5 + y6 = io
(KCL)
(KCL)
(Ohms law)
(KVL)
- 5 V,
= W V,
= 15 V, and V5 = 25 V.
208
In the solution given above, we have separated the expressions into calculation blocks to empha
size the repetitive pattern. For example, the expressions in block #3 are simply obtained from
block #2 by increasing all subscripts by 2. When the ladder network has more elements, the
sequence o f expressions contains more blocks, each o f which entails two additions and two mul
tiplications. This method then allows us to straightforwardly solve ladder networks o f any size
without writing or solving simultaneous equations.
Exercise. In Example 5.7, change all resistances to 2 Q and find V^. Would it make any differ
ence in the voltage Kj if all the resistors were changed to R ohms?
50
AN SW ER: Vi = = 3.85 V, and no difference.
'
13
4.5 V
FIG U RE 5.16 (a) Three voltage sources in series; (b) equivalent single voltage source.
s,eq
209
V =RL
5 kO
6kQ
50 V
210
So l u t io n
Step 1. Substitute all series V^- R combinations by their parallel
case, /^ = . Applying the source transformation theorem four times results in Figure 5.20.
10 mA
10 mA
FIG U RE 5.20 Circuit equivalent to that o f Figure 5.19 by source transformation theorem.
Step 2. Combine the parallel resistances and the parallel current sources.
To the left o f point A are two independent current sources and two resistors, all in parallel.
Similarly, to the right o f B are two current sources and two resistors in parallel.
Combining current sources and resistors to the left o f A results in a single current source o f 5 mA
directed upward and an equivalent resistance o f 4 kD. To the right o f B, the current sources can
cel each other out, and the equivalent resistance is 2 kXl. T he resulting simplified circuit is shown
in Figure 5.21.
5 mA
10 mA
--A
20 V
4kQ
B
20 V
Step 4 . Find
Thus,
211
20 + 20
4+4+2
= 4 mA
= 4 0 0 0 X (0 .0 0 4 )2 ^ <54
= ? and
= ? in Figure 5.23(b).
AN SW ER: 25 V, 500 Q
f^series
-o Circuit
V.
Circuit
r +
B
O
B
-O
(b)
(a)
FIG U RE 5.23
2. For the circuit o f Figure 5.24(a), /^j = 50 mA and R^ = 500 Q , while I ^ 2 = 1 5 0 mA and R^ =
300 Q . Convert the two parallel 1^R combination to a single series V^R combination, where
= ? and
= ? in Figure 5.24(b).
AN SW ER: - 2 0 V, 800 Q
A
ho-
-o Circuit
Circuit
V
B
-O -
B
-O
(b)
(a)
FIG U RE 5.24
3. Consider the circuit in Figure 5.25(a). Using a source transformation and resistance combina
tions, determine the values o f
212
100Q
100Q
FIGURE 5.25
AN SW ER: 150 LX 50 niA
FIGURE. 5.26 Examples of 2-terminal networks, i.e., networks in which only two terminals are
available for connection to other networks.
Observe that networks N j and N 2 in Figure 5.26(a) and (b) have the same terminal characteris
tics: at the terminals o f N j, the v - i characteristic is
v = 2i+ 10
213
i= - - 5
2
The two equations are identical. We then say that a pair o f 2-terminal networks are equivalent if
they have the same terminal characteristics. Therefore,
and N 2 are equivalent.
Now, observe that networks
and
are also equivalent to A^j and A^2- To see this, note that
for N y
V
v-1 5
6/ = 3 v - 3 0
v = 2i + 10
And for N^, first observe that i = \0i^ from KCL, in which case
10 + 20/'a ^
v = 2 i + 10
as was to be shown.
Because equivalent 2-terminal networks have the same terminal v i characteristic, if one network
is interchanged with its equivalent, all currents and voltages outside the box remain the same as
illustrated in Figure 5.27; i.e., all voltages and currents in the rest o f the circuit are the same as
before.
A
o
N
N.1
V
J
Rest of
Circuit
-O B
FIG U RE 5.27 The networks denoted N^, i = 1,2, are equivalent when the v-i values at the terminals
are identical; logically then, all voltages and currents inside the Rest o f Circuit remain the same.
These examples allow us to justify the source transformation theorem as follows. Both 2-terminal
networks in Figure 5.18 have the same v - i relationship: v = Ri +
v = Ri + Rlj, = Ri +
and
.
Therefore, the two networks o f Figure 5.18 are equivalent, and the source transformation is a valid
analysis technique.
214
6. SUM M ARY
This chapter covers the notions o f Unearity, superposition, proportionaHty, and source transfor
mations. Linearity states that for any hnear resistive circuit, any output voltage or current, denot
ed as y, is related to the independent sources by the formula j =
+ ... +
are the voltage and current values o f the independent sources, and
ate constants. O nce values for the
where
through
through
are appropri
are known, one can compute the output for any (new) set o f
input values without having to resolve the circuit equations, a tremendous savings in time and
effort. A special consequence o f linearity is the widely used principle o f superposition.
Superposition means that in any linear resistive circuit containing more than one independent
source, any output (voltage or current) can be calculated by adding together the contributions due
to each independent source acting alone with the remaining independent source values set to zero.
Practically speaking, this is the customary path to computing the coefficients, a^, in the linearity
formula.
Proportionality, another consequence o f linearity, means that if a single input is scaled by a constant,
with the other inputs set to zero, then the output is scaled by the same constant. This property led
to a clever technique for analyzing ladder networks without writing simultaneous equations.
Since
power
is
proportional
to
the
square
o f a voltage
or
current, P = =
for dc resistive circuits, the principle o f linearity and its consequences, superposition and propor
tionality, D O N O T APPLY for power calculations.
Using the notion o f an equivalent 2-terminal network, the chapter set forth the theorems on
source transformations for source-resistor combinations: a 2-terminal network consisting o f a
series connection o f an independent voltage source
= V^IR, in parallel
with R, as illustrated in Figure 5.18. These transformations, applied multiple times to a circuit,
often simplify the analysis o f a circuit.
215
substituted for the other without affecting the voltages and currents in any attached net
work.
and Uj, each acting alone, be y-^ and y 2 . U2 are applied simultaneously, the response is =
and
0 2
+ 2^2- Linearity implies both superposition and proportionality, and vice versa.
Linear resistive element: a 2-terminal circuit element whose terminal voltage and current rela
tionships is described by Ohms law.
Linear resistive circuit/network: a network consisting o f linear resistive elements, independent
voltage and current sources, op amps, and controlled sources.
Proportionality property: when an input to a linear resistive network is acting alone, multiply
ing the input by a constant, K, implies that the response is multiplied by K.
Source transformation: a 2-terminal network consisting o f an independent voltage source in
series with a resistance is equivalent to another 2-terminal network consisting o f an inde
pendent current source in parallel with a resistance o f the same value.
Superposition property: when a number o f inputs are applied to a linear resistive network simul
taneously, the response is the sum o f the responses due to each input acting alone.
216
Problems
(a)
LIN EARITY
(b)
linear relationship v^ 2 =
= a j V^j +
^l^sT
rnay be expressed
Compute
ttj and
(b)
(c)
find v^Jt).
Redo part (a), but this time express
ttj and P 2 in terms o f the hterals
G] = and G t = .
Figure P5.4
5. Consider the circuit o f Figure P5.5.
(a)
(b)
If
(c)
and the input sources V^j and l^j= 20 V and ^ = 0.5 A, find 1/
'^(t)
Figure P5.1
A N SW ER: (b) 8 cos(lO^) - 8 V
2. For the circuit o f Figure P5-2,
(a)
(b)
find Vg in terms o f
and Gy and
*^2
find Ig in terms o f
and Gy
CH ECK:
= 0 .2 5 ^ 1 +????/,2
Figure P5.2
3. For the circuit o f Figure P5.3,
Figure P5.6
(a)
(b)
. ^2>
^3-
^oia
the two independent sources.
Hint: Write two loop equations.
-t"
(b)
Figure P5.3
If
= 20 V and 7^2 = 0.125 A, com
pute the power delivered by the
dependent source.
217
G,
<+ i;
I .Q
>R,
'
Figure P5.7
C H EC K : (a)
Figure P5.9
Ry = 18 Q.
=9
and R^= 18 D.
(a)
^3 = 18
Ra = 36 Q,
= 20 cos(2 i)
= 10cos(2^) mA
= 2 cos(2z-) V. Find the
and
= 0
Now compute
when
V________________^
Figure P5.10
11. Again, consider the configuration o f Figure
= 0 with v^2 ^t) = 10 V, then
P5.10. If
= -2cos(2?) V.
(a)
(b)
10cos(2z-) V, find
20cos(5^) V, find
C H EC K : (b) v^Jt) = 108cos(5^) V
Figure P5.8
CH ECK:
27
- - Vi + ?? *4 -
= 15 V
=10
V when
= 3 A and v^ 2 = 5 V.
(a)
Determine the linear relationship
(b)
Find
AN SW ER: 7.5 V
218
(a)
(b)
(a)
(b)
= 20 V and
= 500 mA
Resistive Circuit
witli
Dependent
Sources
V. . -
Figure P 5.I3
Figure P5.15
Table P5.13
I',, (volts)
z'^2 (amperes)
0.4
-1
10
Table P5.16
Km
(b)
tionship
+ p 2i,2 Given the equation found in part (a),
compute
when
= 5 A.
AN SW ER: (b) 90 watts
= 15 V and z^2
Case 1
30
-1
11.5
Case 2
40
-1
13
Case 3
30
2.2
-1
11.6
Case 4
30
- 0 .9
11.9
Case 5
40
10
(a)
ments.
Case 1
50
-2
Case 2
Case 3
^^.3 (V)
1
(b)
Table P5.15
(mA) 1^.2 (V)
(V)
-1 3
219
Table P5.17
h\
Table P5.18
^ 2 (V)
^",3 (V)
Kut (V)
i,4 (mA)
-out (V)
Case 1
30
-1
11.5
Case 2
-2 0
27
10
Case 3
-1 0
-3
-1 4
Case 4
40
10
10
???
>
(a)
(b)
R2 = 50 Q, 'sX = 12 V, and
(a)
1^2
using
Find
= 200 Q,
mA.
superposition.
due to
to z,2 act-
is adjustable. In
out
ing alone.
AN SW ER: 2.4 V, 2.4 V, 4.8 V
(b)
Find
and i^ 2
If
= 10
determine
, R^-,
12 V and j2 = ^ x 60 mA,
using the proportional
^out =
+ ^2^.2 +
\'s\
'ih-i
acting alone,
ing alone.
We have used the fact here that the term (a,v^
+ ^2^j2
constant because the associat
ed source values are constant. Thus,
+
^2^i2
^^
some K. Hence, one can
use the data from the first two rows o f Table
P5.18 to solve for
and K.
Figure P5.19
20. Consider the circuit o f Figure P5.20 in
which
= 20 2 , R2 = 60 Q, R^ = 20 Q.
(a)
Find the coefficients o f the linear rela
Linear resistive
network with
dependent
sources
tionship
+ a 2 i, 2 + ^3^,3 by
superposition. Specifically, first find
a
Figure P5.18
220
(b)
(c)
Find
power delivered to
V,
(d)
(a)
= 2 A, and
and
when
(b)
due to
Ri
= 4 A.
and the
= 100
due to
is doubled, i^2
is halved.
1
AN SW ER: (a)
5
2
7
= -''.v i- '' m = - ',v2
Figure P5.22
Figure P5.20
fi by superposition.
Repeat part (a), but express your
0 -2 1
(b)
20 Cl, /?2 = 50 Q,
(a)
= 100 Q and
= 0.02 S.
due to
l lR- ,i = 1 , 2 , 3 , 4.
(c)
Find
R^ when
= 100 V and v^ 2 = 50 V.
= 60 V;
= 5 watts
AN SW ER:
(d)
= 0.5 x 100 V
and v^ 2 = 10 x 2 V.
Figure P5.21
Figure P5.23
24. For the circuit o f Figure P 5.24, find the
contribution to
- 0
by
221
i8on
225 n
20V
Figure P5.27
0.1 A
28. (a)
^900 0
MATLAB program.
Figure P5.24
25. For the circuit shown in Figure P5.25,
160 V. Find
(b)
If it is known that
(c)
= 1V
Figure P5.28
Hint: Assume
= 1 A and
1n
1n
1n
III
2n
1n
2n
2n
2n
due only
to
(b) Find the contribution to
due only
by superposition.
r2n
Figure 5.26
2 7 .(a)
Figure P5.29
=
(b)
5 Q,
= 20 Q, and
= 5 Q..
If it is known that
= 175 volts, find
(c)
(d)
to I/j.
Find the contribution to
(c)
to
Find
by superposition.
due only
222
Figure P5.30
Figure P5.32
Find
(c)
33. (a)
(c)
= 5 V, V^2
= 2 V, determine
r^
34. (a)
Figure P5.31
Kd-
between V^^^^and
32. Consider the circuit in Figure P5.32.
(a)
due only
(b)
to Ki
Find the contribution to
due only
to
(c)
Find
(d)
If
= ^3 = 0 . 5 ^ = 5 kQ
and V;i = 2V;2 = 4 V, find the
(b)
= 0.25 V,
= 2 V, determine
by superposition.
load.
n
Figure P 5 .3 4
(b)
(c)
to 1/2.
Find
(d)
223
(a)
= Q.25R^= R2 = Ry
Find the contribution to
due only
(b)
to
Find the contribution to
due only
(c)
to
Find
which
due only
by superposition.
by superposition.
(e)
Figure P5.37
38. Consider the circuit in Figure P5.38.
Find the contribution to Vout due only
(a)
to l/i(b)
Figure P5.35
(c)
36. Consider the circuit in Figure P 5.36 in
(b)
= Q.25Rjr.
Find the
contribution to
only to
Find the contribution to
(c)
(d)
which
(a)
(d)
due
due
Figure P5.38
when Rjr=
224
200 n
0
Figure P5.39
AN SW ER: 6 V source in series with 12 Q resis
40 0
80
40 n
20 n
son
Figure P5.41
AN SW ER: 13.5 and 9.1125
42. Use source transformations on the circuit o f
tor
need
(b)
Figure P5.42
A N SW ER: 28 V
Figure P5.40
AN SW ER: (a) 48 V in series with 3.2 kQ; (b)
135 mW
is 16 watts.
40 A
sn
2A
<
50<
= 1n
V ^ appears.
(b)
Find
Find the power dissipated in the 20 Q
resistor.
(c)
is n
Figure P5.43
44. Apply source transformations to the circuit
shown in Figure P 5.44. Then write two nodal
equations to find
and V2 .
Figure P5.45
AN SW ER: 2.8 V, - 0 .4 V
225
CHAPTER OBJECTIVES
1.
2.
3.
4.
Define and construct the Thevenin and Norton equivalent circuits for passive networks.
Define and construct the Thevenin and Norton equivalent circuits for active networks
containing dependent sources or op amps.
Illustrate several different techniques for constructing the Thevenin and Norton equiva
lent circuits.
Investigate maximum power transfer to a load using Thevenin or Norton equivalents.
228
SECTIO N H EADIN GS
1. Introduction
2. Thevenin and N orton Equivalent Circuits
3. A General Approach to Finding Thevenin
4. Thevenin and N orton Equivalent Circuits
5. Thevenin and N orton Equivalent Circuits
6 . Thevenin and N orton Equivalent Circuits
7. M aximum Power Transfer Theorem
8 . Summary
9. Terms and Concepts
10. Problems
1. IN TRO D U CTIO N
Practicing electrical engineers often want to know the power absorbed by one particular load. The
load may be a large machine in a factory or a lighting network in the electrical engineering build
ing. Simple resistances often represent such loads. Usually the load varies over time in which dif
ferent resistances are used at different times to represent the load. W hat is the effect o f this load
variation on the absorbed power and on the current drawn by the load? To simplify analysis, the
rest o f the linear network (exclusive o f the load) is replaced by a simple equivalent circuit consist
ing o f just one resistance and one independent source.
For our purposes, a (resistive) load is a two-terminal network defined in Chapter 1, meaning that
the current entering one o f the terminals equals the current leaving the other. More generally, a
two-terminal network is any circuit for which there are only two terminals available for connec
tion to other networks. (See Figure 6.1.) T he important question for our work in this chapter is:
How does one characterize a two-terminal networks As is shown in Figure 6.1(a), there is a voltage
v{t) across the terminals and a current i(t) entering one terminal and leaving the other. T he rela
tionship between the voltage v{t) and the current i{t) characterizes the two-terminal network. For
example, if v{t) = Ri{t), we would recognize the terminal network as an equivalent resistance R.
Or, if v{t) = Ri{t) + Vq, we might recognized this equation as that o f a resistance in series with a
voltage source. In fact, this equation could be represented as graph, e.g. Figure 6.1(b).
This leads to our next question: When are two 2-terminal networks equivalent'^ As developed in
Chapter 5, two 2 -terminal networks are said to be equivalent when their terminal v-i characteris
tics are the same. O f particular interest for this chapter is an equivalent network consisting o f a
voltage source in series with a resistance, called the Thevenin equivalent network, and a current
source in parallel with a resistance, called a Norton equivalent network. Figure 6.1c shows a
Thevenin equivalent for a linear resistive circuit.
N-
229
i(t)
2-terminal
Linear
o
+
Resistive
Network
V(t)
FIG U RE 6.1. (a) a 2-terminal linear network with terminal voltage v{t) and current i{t)\
(b) graphical representation of the equation v{t) =
+ Vq,
(c) Thevenin equivalent network
This chapter investigates the replacement o f a network N by its Thevenin equivalent or its Norton
equivalent. The first section describes the Thevenin and Norton equivalent theorems for passive net
works, those containing only independent sources and resistors. Following that, we generalize the
statements to include active networks. However, because op amps have peculiar properties, Thevenin
and Norton equivalents o f circuits with op amps are explored exclusively in Section 4. Following
this, in Section 5, we describe how to obtain a Thevenin or Norton equivalent from measured data
without having to know anything about the internal circuit structure. This is particularly useful
when one has equipment such as a power supply but no schematic diagram o f the internal circuit
ry. Unfortunately, not all linear devices have a well-defined Thevenin or Norton equivalent. The
homework exercises illustrate a few cases. Section 6 explores the problem o f maximum power trans
fer to a load in the context o f the Thevenin equivalent circuit, which ends the chapter.
230
experiment, while N, complicated in its own right, remains unchanged. To simplify repeated cal
culations, N is replaced by its Thevenin equivalent, as illustrated in Figure 6.2(b). The more sim
ple Thevenin equivalent consists o f a single voltage source,
in series with a single resistance.
Rthf
------ o-----+
Resistances
and
independent
Sources
Arbitrary
Networi<
r-\
(a)
i, = 0
-A
NResistances
and
independent
Sources
-o
-I-
-o
NReslstances
with independent
Sources
-o
R..
Deactivated
(c)
(d)
231
Th e voltage,
called the open-circuit voltage, is what appears across the 2 terminals o f N. R^j^, called the
Thevenin equivalent resistance, is the equivalent resistance o f N when all independent sources
are deactivated. Figure 6.2(c) shows the appropriate polarity for v^J^t),
In the above theorem, for almost all means there are exceptions. For example, an independent
current source does not have a Thevenin equivalent. More generally, any two-terminal network
characterized by i{t) = constant does not have a Thevenin equivalent. This leads us to suggest that
there ought to be an equivalent current source formulation o f an equivalent network. From
Chapter 5, the source transformation theorem tells us that the Thevenin equivalent o f Figure
6.2(b) when
(c)
FIG U RE 6.3 (a) Arbitrary 2-terminal linear network o f resistors and independent sources;
(b) Norton equivalent circuit; (c) circuit for computing
with
computed, as per Figure 6.2(d).
A single voltage source does not have a Norton equivalent, and as mentioned a single current
source does not have a Thevenin equivalent. Both Thevenin and Norton equivalents exist for a 2terminal linear circuit when R^^
lents exist for the same network, the source transformation theorem and Ohms law imply that
thhc'kt)
(6.1a)
232
4^ 0, then
and when
r,
th -
(6.1b)
circuits.
E X A M P L E 6 .1 . For the circuit o f Figure 6.4, using literals, find the open circuit voltage,
short circuit current,
200 Q ,
the
Then, if R^ = 50 Q,
-O
-O
40 n
160V
40 0
4A
-o
-O
(c)
(b)
FIG U RE 6.4. (a) Resistive 2-terminal network; (b) Thevenin equivalent; (c) Norton equivalent.
S
o l u t io n
Step 1. Find
^1^2
~ R\+R2
^1 + ^2
100 + 4 0 x 2 = 160 V
Step 2. Find i^^. As per Figure 6.4, with terminals A and B shorted ^together, all the current from
flows through the short circuit. From superposition, isc = 's2'^----- Substituting numbers
into this formula yields
i^^ = 2 + 0.02
100 = 4 A
233
Step 3. Find
Replacing
^ 40
R, +R2
Step 4 . Determine the Thevenin and Norton equivalent circuits. The Thevenin equivalent circuit fol
lows from Steps 1 and 3 and is illustrated in Figure 6.4(b). The Norton equivalent circuit follows
from Steps 2 and 3 and is illustrated in Figure 6.4(c). We also note that
R ,, =
V =155 =
he
40 Q
as expected.
It is important to note that for many circuits, especially when the deactivated circuit is a seriesparallel connection o f resistances, one can obtain the Thevenin equivalent by a series o f source
transformations.
AN SW ER; 80 Q, 240 V, 3 A
follows easily from Equation 6.1. In some cases, the choice o f which two to find first either
increases or decreases the amount o f calculation. The following exercises illustrate this point.
Exercises. 1. For the circuit o f Figure 6.5, R^^= 200 Q, R2 = 50 O., R^= 10 Q,
= 50 V. Find R^f^, i^^ and v^^ in this order.
A N SW ERS: 8 Q, 1.5 A, 12 V
2. For the circuit o f Figure 6.5 with the same values as in Exercise 1, find v^^, i^^, and
this order.
3. For the circuit o f Figure 6.5, find the Thevenin equivalent circuit using a series o f source trans
formations.
The next example illustrates the computation o f the Thevenin and Norton equivalent circuits
using loop analysis.
234
E X A M PLE 6 .2 . Find the Thevenin and Norton equivalent circuits seen at the terminals A-B for
the circuit depicted in Figure 6.6, where
= 100 V and
= 3.2 A. We show that
= 4 0 0 Q,
= 200 V, and
= 0.5 A.
1500
500 Q
FIGURE 6.6 Two-source circuit for Example 6.2 with loop currents shown;
So l u t io n
Step 1. Compute
To compute R^^^, we set all source values to zero. Each voltage source becomes
a short, and each current source becomes an open. This leads to the circuit o f Figure 6.7. Here, we
have a 500 Q in series with 100 Q, yielding 600 Q. Since this 600 Q resistance is in parallel with
400 2, the resulting equivalent resistance is 240 Q. Hence,
500 0
1500
FIG U RE 6.7 The circuit of Figure 6.6 with all independent sources deactivated.
Step 2. Compute an expression for
A-B is N O T present. Hence,
its voltage drop is zero. (One ofi:en says that the 150 Q resistor is dangling.) Thus, from KVL we
have
= '^oozj + ioz;2
(6 .2)
23^
loop 1,
^s\ =
+ 100/^2= 1000/jand
in which case,
+ 160 = 200 V
Step 4 . Construct the Thevenin and Norton equivalent ciraiits. Equation 6.3 with
(6.3)
= 400 Q yields
the Thevenin equivalent o f Figure 6.8(a). Further, from the source transformation theorem,
(6.4)
+ 100(/j - /^2)
in which case.
v^\ + 100/^2 = 1000/j - 400/^^ = 420 V
Around loop 2 we have
560/;^-400/, = 10/^2 = 32 V
In matrix form, the pertinent equations are
1000
-4 0 0 '
-4 0 0
560
h
Jsc
420'
32
2M^
Thus,
' i\
1000
-4 0 0 '
j.sc
-4 0 0
560
-1
420
0.62'
32
0.5
Vo,.
th
Exercises. 1. Suppose all source values in the circuit o f Figure 6.6 are doubled. What is the new
v j Does
change?
A N S W E R : /, =-)()() V. no
2. Suppose all resistances in the circuit o f Figure 6.6 are multiplied by 4 and the independent cur
rent source is changed to 0.6 A. Find
and
Hint: For
is in ohms, so if the resistances are multiplied by four, what is the new value?
AN SW ER: r
= 160 V. A',
// 4 x 400
3. A 400 L resistor is connected in series with terminal A o f the circuit o f Figure 6.6. Find the
I V
ANSWLR:
im
V, A',,,
- .2S A
4. A 400 Q resistor is connected across terminals A and B o f the circuit o f Figure 6.6. Find the
V
" I V
ANSV('-UR:
In the above two examples, deactivation o f all independent sources led to a series-parallel network.
Calculation o f
Nortons theorems.
Norton equivalent consists o f a single resistance R^f^. For a series-parallel net%vork, R^f^ can be
computed by straightforward resistance combinations.
23'
The terminal v-i characteristics o f the network N and its Thevenin equivalent must be the same.
Consider that the v-i characteristic at A-B o f the Thevenin equivalent o f N is
(6.5)
while the Norton equivalent o f N as per Figure 6.10(c) has the v-i relationship
'a
- 1
^th
(6 .6 )
- ^sc -
These relationships tell us that if we have a linear net\vork and assume there is a voltage
across
its terminals and a current /^j entering the network, as shown in Figure 6.10(a), then obtaining an
equation o f the form
(6.7)
or o f the form
(6.8)
Vj n
and
or to march
^th
an easier approach for non-simple circuits, as the next two examples illustrate.
Remaining
Network
^
A
------oLinear
Network
A8
- o .......+
B
(a)
Remaining
Remaining
Network
(c)
I'lG U R M 6.10 (a) Nervvork N attached to an unknown network;
(b) theTheveinin equivalent of N attached to the unknown nervvork; (c) the
Norton equivalent of N attached to the unknown network.
23.S
E X A M PLE 6 .3 . This example revisits Example 6.2 using the new approach. Again, we find the
Thevenin and Norton equivalent circuits seen at the terminals A-B for the circuit depicted in
Figure 6. I I , where
= 100 V and
= 3.2 A. Our goal is to find the v-i characteristic at the ter
minals A-B.
FIGURE 6.11 Two-source circuit for Example 6.2 with loop currents
shown;
= 100 V and ip = 3.2 A.
So l u t io n
Step 1, Consider i^ loop. Around the loop for i^, we have
^AB ~ 560/^ + 400/j + 10/^2 = 560/^ + 400/j + 32
(6.9)
42 0 - 400/^
(6 . 10)
1000
R,h = 4 0 0 Q,\-
= 0.5 A.
th
(6.11)
ly )
EXAM PLE 6.4. For the circuit o f Figure 6. ] 2, find theThevenin equivalent o f the 2-terminaJ Network
N defined by the dashed line box. We show that
= 9.6 V,
= 4.4 Q., and = 2.1818 A.
and
So l u t io n
Our objective is to compute the relationship o f the form o f equation 6.7 using Nodal analysis and
then match coefficients with equation 6.5 to obtain R^j^ and
~^
Step 1. Write nodal equations. For writing the equations o f this circuit, the reader might first
review Example 3.2. Alternately, using the inspection method, the matrix nodal equations are
- 0 .2
0.8
- 0 .4 ^1
- 0 .2
0.6
-0.1
- 0 .4
-0 .1
0.5
h\
' 1
\'2
\>2
( 6. 12)
js l + U .
^'ab
det(A/) = det
- 0 .2
0.2
0.6
- 0 .4
-0 .1
- 0 .4 '
-
0.1
= 0.:
0.5
- 0 .2
det - 0 .2
0 .6
h\
0
- 0 .4
-0 .1
s2 + U
^ab -
d et(M )
- 0 .2
O'
det - 0 .2
0.6
- 0 .4
-0 .1
d et(M )
4 . 4/4
0.8
- 0 .2
0.8
- 0 .2
O'
det - 0 .2
0.6
det - 0 .2
0.6
- 0 .4
-0 .1
- 0 .4
-0 .1
. /. -1------"
d et(M )
+ 2 . 6/^1 + 4 .4 /^ 2 = 4.4/,^ + 9 .6
/ . 4- -----
d et(M )
( 6 .1 3 )
240
or equivalent.
Step 3. Match coefficients o f equations 6.13 and 6.5. Matching coefficients o f equation 6.13 with
equation 6.5, we obtain
V,
= 4.4 a and
2.1818 A
Exercises. 1. If the independent current sources in the circuit o f Figure 6.11 arc set to zero, find
the Thevenin equivalent circuit.
AN.SW'j-^R: The I hcvcnin ci|uiv.ilcnt .onsist.s oi a single rc.sisn)r.
2. Find
= 4.4 12.
A N S W I- R ; -48 \
3. A 4.4 d resistor is connected in series with terminal A o f the circuit o f Figure 6.12. Find the
V
ANSW'l-.R:
. ).(, \.
=1
4. A 4.4 Q resistor is connected across terminals A and B o f the circuit o f Figure 6.12. Find the
ANS\V1;R:
- 2 .1SIS A.
= 2.2 f l and
M .S V
At this point, we end our development in this section with an example that shows how to com
pute a Thevenin equivalent from measured, e.g., in a laboratory setting where there is a power sup
ply with an adjustable voltage.
EX A M PLE 6 .5 . Consider Figure 6.13, which show^s the Thevenin equivalent o f an unknow'n network N attached to a variable voltage,
Variable
Voltage
Power
Supply
oc
B
MCURI-. 6.13 Thevenin equivalent of an unknown network N
connected to a variable voltage power supply.
2/f
TABLE 6.1
/^I (mA)
-
(V)
10
24
20
40
o l u t io n
yields
24 = 0.01,;,+ V
from row 1 o f Table 6.1, and
40 = 0 . 0 2 ,^ * V
from row 2 o f Table 6.1. In matrix form,
0.01
0.02
^th
24
40
Solving produces
^r/i
'V>(Hence, R,,, = 1600
1
-1 24
r -I 24
= -100
-0.02 0.01 40
40
0.02 1
0.01
and
1600'
8
= 8 V.
Thus, one can use the technique o f Example 6.5 to determine Thevenin equivalent circuits in the
laboratory.
242
called the N orton equivalent (Figure 6.14(c)). In most cases, both the Thevenin and Norton
equivalent circuits exist. Computation o f y^^is characterized by Figure 6.14(a), computation
of
/^Neq
NResistances,
independent
and
R..
dependent
sources
(a)
(b)
N-
(c)
NResistances,
Independent
sources
deactivated
independent
and
(d)
(e)
-o
dependent
sources
EXA M PLE 6.6 . Find the Thevenin equivalent circuit for the 2-terminal network (marked by
dashed line box) in Figure 6.15(a) using the method o f Section 3. (The dependent source acts as
a voltage amplifier.)
B
O(b)
FIGURE 6.15. (a) circuit with terminal voltage i/^^and input current /^;
(b)
/?, = /?,//
SO L u n o s
Step 1. Since there are no independent internal sources, the Thevenin equivalent consists o f a single
resistance,
Step 3. Match coefficients with equation 6.6. Matching coefficients implies that
= ((7j + (p +
R^R
j A2
{ 1
J_ \
R2+(M + ^)R\
d 4.
R^
(6.14)
'(A^ + 1)
and _.^2
and Rj = 4 kf2.
Then
iJ + \
Exercises. 1. For the above example, suppose p = 99, R^ = 500 Q, and /?2 = 1
A N SW ER: 10 Q
P*rid R^jy
244
2. For the circuit o f Figure 6.16, find the Thevenin equivalent resistance by obtaining
of
in terms
'AB-
0 and
^ih -
('jth
E X A M PLE 6 .7 . Find the Thevenin and Norton equivalent circuits seen at the terminals A-B in
Figure 6.17 when = 50 mA. Our computations will proceed using loop analysis to find the ter
minal v-i characteristic A-B.
5001
o l u t io n
and
= 0.05 A.
Step 1. Write a set o f loop equations for the circuit o f Figure 6.17.
For loop 1, vvc obtain
0 = 1000(/j + i^) + 1000(/, - i^) - m i ^
which simplifies to
- 5 = 2000/, - 1500/;^
For loop A, we have
Jn
Step 2. Write the loop equations in matrix form and solve. Writing the loop equations in matrix
form yields
2000
- 1 5 0 0 ' 'i\
-1 0 0 0
-50'
1100
del
IA=
det
-5 0 '
det
-1000
[2 0 0 0
-1 0 0 0
-1 5 0 0
-= -5 0 -
2000
-1 0 0 0
det
AB
700x10-
2000
()
-1 0 0 0
700x10-
1100
50
2
------- 1* AR-----700
700
(6.15)
or equivalently,
(6.16)
'/IS = 350/;^ + 25
Step 3. Match coejficietits o f equation 6.15 with equation 6.6 or equation 6.16 with equation 6.5 to
obtain
50
1
1
= = A, G,i, = =
S. R, = 350 Q, and v,,,. = 25 V
700
14
700
350
A.
350
S.
R^f^, and
= 350 a a n d i ^ = 2.5 V
2. In Example 6.7, if /^= 5 mA and the 100 Q resistor is replaced by a short circuit, find
and
ANSWKR:
S, and
= - ^ = 0 .0 1 A
3. Find the Norton equivalent at the terminals A-B o f the circuit o f P'igure 6.18 when mA.
AN SW ER:
= 200 LI and
= 0.1 25 A
8 0 0 I.
200 0
--- O
<
I
<
^ 800 0
800 Q
----- (
B
FIG U R H 6 .1 8 Modification o f the circuit o f Figure 6 .1 7 .
2-i6
S o lution
Step 1. Find
2 0 -h 5
' Cfi Step 2. Find v^g. By inspection.
Step 3. Match coefficients with equation 6.5. Matching coefficients we obsen'e that
R.i. = 16 Q. V r =
, and
=
16
Our next example illustrates how to construct a negative resistance using an ideal op amp.
EXA M PLE 6 .9 . Find the Thevenin equivalent seen at the terminals A-B for the (ideal) op amp
circuit o f Figure 6.20.
lO k O
S o lution
VcB -
5 ,
^AB - Vc
10x10^
lOxlO-"'
10x10^
Exercise. For the circuit o f Figure 6.21, find the Thevenin equivalent circuit at A-B.
AN SW ER:
= 0 and R^,, = -R
:-i
Our third example constructs a Thevenin equivalent of the standard inverting op amp configura
tion with a terminal resistance. However, we will consider both the ideal and non-ideal cases.
EXA M PLE 6 .1 0 . Find the Thevenin equivalent seen at the terminals A-B for the op amp circuit
o f figure 6.22 when
(a) w'hen the op amp is assumed ideal, and
(b)
when the op amp has a saturation voltage,
= 15 V.
S o lution
Step 1. Fiuci the Theveniti equivalent seen at the terminals A-B assuming an ideal op amp. I'h e prop
erties of an ideal op amp imply that
On the other hand,
= 0. In flict,
Thevenin equivalent seen at the terminals A-B consists only o f a voltage source o f value
=5
uration. When the non-ideal op amp operates in its linear region, the Thevenin equivalent by Step
I is a voltage source having value v^^^ = - 5 V^. When, |-5 V} >
= 15 V , or equivalently, when
I
> 3 V, then the op amp saturates at 15 V. Specifically, when K > 3 V, then v^^ = - \5V and
when K, < - 3 V, then v^^ = 15 V. The Thevenin equivalent for an op amp with output saturation
is summarized in Figure 6.23, where v^^. takes on three separate values depending on the region o f
operation o f the amplifier.
1 5 V fo rV ^ < -3 V
- 5 V f o r |V j< 3 V
1 5 V fo rv / > 3 V
FIGURE 6.23 Thevenin equivalent at output terminals of an inverting amplifier (Figure 6.22)
with non-ideal op amp.
249
This cnd-s our investigation of I hevenin equivalents o f op amp circuits. There arc many more
interesting examples that are beyond the scope o f this text.
all functions o f Rj. The main objective o f this section is to show that for fixed R^j^, maximum
power is transferred to the load when R^ = R^f^. We illustrate this assertion with an example that
shows the power delivered to Rj as a function o f Rj. Throughout this section, it is assumed that
all resistances are non-nesative.
FIG URE 6.24 Thcvcnin equivalent o f network N connectcd to a variable load, Rj.
E XA M PLE 6 .1 1 . For the circuit o f Figure 6.24, suppose that R^j^ = 20 Q and
= 20 V, Plot the
(6.17)
IL =
^ (2 0 )'
To obtain the plot w'e use the following MA'FLAB code, resulting in the plot o f Figure 6.25.
voc = 20; Rth = 20;
RL = 0 :0.25:100;
PL = RL .* voc^2 ./ ((RL + Rth) .^2);
>>plot(RL,PL)
grid
250
RL in Ohms
FIG URE 6.25 Plot of power delivered to the load in Figure 6.24 as a function of Rj.
= R^j^ = 20
In a neighborhood o f R^ =
20 ^2, the curve remains fairly flat. At R^ = 40 Q. and Rj = \0 Q, the curve shows that about 88%
o f maximum power is transferred.
This experimentally observed fact, that maximum power transfer occurs when R^ =
plays an
important role when matching speaker resistances to the output resistance o f a stereo ampli
fier or when trying to get as much power as possible out o f an antenna and into a receiver.
rL.max
A verification o f the maximum power transfer theorem proceeds using differential calculus. From
equation 6.17, the power absorbed by the load is
2S1
R
Pl =
L____
VV? o c
[^L + ^th)
Following the standard procedure o f calculus for determining a maximum/minimum, we com
pute the derivative o f w i t h respect to Rj, set to zero, and solve for R^.
dPL ^ d
cJRl
(IRl
Rr
-9 .
oc
Rl v I c
= y2
/
\3
from which R^ = R^^^ and R^ = ^ are the only possible solutions. But, if R^ = oc, then
= 0.
Hence, because equation 6.17 is positive for /?^ > 0, /?^ = R^j^ produces maximum power,/>^, deliv
ered to the load.
Further, substituting
Rfll
P L ,m a x
^n/(6.18)
.......N ................................
600 O
<
V -/
300 0
Thevenin
equivalent
(b)
FIGURE 6.26 (a) A network N connected to a load /?/-, (b) Thevenin equivalent of N connected to
2^2
SO L U T IO N
urrent
the independent voltage source becomes a short and the independent cur,
Step 1. To compute
= 200
source becomes an open. Finding the equivalent: resistance seen at the terminals produces
Hence, maximum power is transferred when
Step 2.
= 200
may be computed by any o f the methods discussed throi4ghoi4t this chapter. For example, by
repeated source transformations, the network N reduces to its Thevenin equivalent shown in Figure
6.26(b) with
Plugging
PU
vi
(1 6 )
4R,th
800
= 320 mW
Exercise. Suppose the 400 Q resistor in Figure 6.26(a) is changed to 100 Q. Find
R^j^, and pj
EXA M PLE 6 .1 3 . This example shows that the Thevenin equivalent cannot be used to calculate
power consumption within the network N it represents. For this demonstration, consider the net
work N given in Figure 6.27(a) with its Thevenin equivalent given in Figure 6.27(b). Compute
the power loss within the actual network N and within its Thevenin equivalent. We show that
these are different.
N.
-O1n
1A
20
3V
--Thevenin equivalent -(a)
(b)
Pm
- 1 X 1- -
2^3
This means that theTIicvenin equivalent is not, in general, representative o f pov/er relationships
within the network, i.e., the losses that are dissipated as heat, for example.
When a network N is a voltage source in series with a resistance R^, and hence is its own Thevenin
equivalent, one may ask about maximum power transfer when
fixed, assuming
is also fixed. The following example is an experiment for investigating this sit
uation.
E X A M P L E 6 .1 3 . For the circuit o f Figure 6.28, suppose /?^ = 20 Q and
delivered to the load as a function o f R^ along with the power loss,
SO L U T IO N
The power delivered to the load R^ is
R,
(R
+ Rs )
0
vr =
20 X 400
(20+ / ?,)
To obtain the plots, we use the following MATLAB code, resulting in the plot o f Figure 6.29.
vs = 20; RL = 20;
..Rs = 0:.25:50;
>>PL = RL .* vsA2 ./ ((RL + Rs) .^2);
plot(Rs,PL)
grid
hold
Ploss = Rs .* vs'^2 ./ ((RL + Rs) .'^2);
plot(Rs,Ploss,b)
25-1
= 0. Observe that
niaximizep^.
The proof for the maximum power transfer theorem given earlier considers Rj as the independ
ent variable and sets
dRi
approach whose derivation is simpler mathematically, but is more meaningful for applications in
the sense that the load can be a general 2-terminal linear network, N^, instead o f a single resistor.
For this alternate derivation, refer to Figure 6.30. We ask the question. What v-i characteristic
FIGURE 6.30 The Thcvcnin equivalent of a network N connected to a loading ncrvvork N^.
= y i= y l^ - V
R,th
we differentiate
lIPL
dv
^oc R,i,
V = 0.5v
(6.19a)
at which value
(6.19b)
ih
th
If
consists o f a single
^ L - ~ - ^ih
I
At V= 0.5v , the corresponding maximum power is
( 6 .20 )
PLmax = V X /= -
4/?,th
S o lution
= 0.5 A. Thus,
^th
V= \ = 0 .2 / +
= 0.\ + Vj
=o
Vf = 0 .9
= 1 V and from
2S 6
resistor that
8. SUM M ARY
This chapter has set forth a powerful strateg)' for analyzing complex networks by replacing portions o f the nerwork by their simpler Thevenin and Norton equivalents. The Thevenin and
Norton theorems assure us that almost any 2-terminal linear nerwork, no matter the number o f
internal elements, is equivalent to a simple nerwork consisting of an independent source either in
series with or in parallel with a resistance. O f course, an independent current source does not have
a Thevenin equivalent, and an independent voltage source does not have a Norton equivalent.
\4ore generally, there are some circuits that have one but not the other. Further, some circuits have
neither.
'
rhe chapter has illustrated various techniques for constructing the Thevenin and Norton equiv-
ing series and parallel resistances as learned in Chapter 2. If the resultant nerwork is not seriesparallel, then one should use the main technique set forth in this chapter, which is to find the vi characteristic o f the terminals. This technique is valid for all circuit t\'pes.
With the ideas o f a Thevenin and Norton equivalent circuit, we then investigated the problem of
transferring power to a load. When R^f^ is fixed, maximum power is transferred when R^ is adjust
ed to be R^j^. If
25'
ance.
: the open circuit voltage o f a 2-terminal network N when no load is connected.
' For a generalization o f this condition to the case where the controlling voltage or current is out
side o f N, see the article by Peter Aronheim entitled Frequenc)' Domain Methods in The Circuits
258
Prob ems
= 12 V,
= 60 Q, /?2 = 60 Q, and R^ = 40 Q.
0.4 A,
TH EVEN IN /N O R TO N FOR
PASSIVE CIRCU ITS
using C/ = .
R.
Figure P 6.1
Figure P6.4
= 8 k n , Ri = 6 kQ,
= 60 V, i^ 2 - ^9 mA and
5kQ
-o-r
Figure P6.2
-O -v
= 30 V, and /p = 10 mA.
across
M ATLAB
A-B.
Plot
is
using
Figure P6.5
C H E C K : 200 V, 10 kQ
6. For the circuit o f Figure P6.6,
R.
^(l)
' 6
Figure P 6 .6
Figure P6.3
ANSWF.R: /?;/; =2 kil, isr = 20 mA, // . = 40 V
= 48
- 6
= 18 kH,
CH ECK:
= 2 mA
2S9
= 3.6 kQ.
= 32 k n .
R,
' 6
8R
6R
Figure P6.9
C H ECK S: (a) 900 Q, 30 V; (c) 250 mW, (d) 1
Figure P6.7
C H EC K : R.th = 10 kD
Figure P6.10
= 2R,
V()
= +
= 204 V
Figure P 6 .11
A N S W E R : 5 2 V. 2 4 k X 2 mA
260
1 kn
2kO
general method.
(b) If a load resistance
is connected to
is max
15. (a)
Find a so that
o f Figure P6.15.
(b)
100 o
100 Q
R^^^ = -1 kQ.
100 Q
200 Q
priately.
C H EC K : (b) a = 4000 Q
i.
Figure P 6.12
ANSW FR: For (a),
------O
= 1(H) LI
1 kO
1 kO
2kn
TH EVEN IN /N O R TO N FOR
A CTIVE CIRCU ITS
13.(a) Find the value o f
so that the
Figure P6.15
- 2 5 0 Q. C H EC K : (b) 6.25 mS
(b)
200 n
1 kn
2k O
800 n
A N SW FR:
= -1 kl^ .
Figure P 6.13
= 1000 uS
Figure P 6 .16
17. For the circuit shown in Figure P6.17, find
the Norton equivalent circuit.
200 n
= -0 .0 0 1
A
7/l
18kQ
Figure P 6 .17
ANSWFR: - 6 0 0
/ = 0
261
and
= 20 V,
100 0
-----1----- -
loon
mine R^fj.
'300 0
' 0.01
0>
800 n
ANSW I-R;
Figure P 6,18
= 0,
= 250 12
terminals,
plot
the
AN SW ER:
Figure P6.21
and
= 60
= 18 V
power
value o f
alent circuit.
i.j
50 0
V. <v 50 O
'
50 0
bv
Figure P 6.19
A X 'S W I- .R :/'DC=6/<. / w' = U Mf, P,//
Figure P6.22
W SW I'R; R^,, = 100 12.
= 50/,
P6.23.
(a) Find the Thevenin equivalent.
= 0.2 A.
' 0
1 kO
'400 0
200i
Figure P6.20
AN SW FR: r^. = 60 \. A,;, = SOO LI
2(>:
' 6
Figure P6.23
= 20 mA.
AA
400 n
2000i
io o v (^ ^
0.1 A
ANSWI-:R:
= 12.S V,
= 650 Q
Figure P6.24
25. Find the Norton equivalent for the circuit
shown in Figure P6.25 when = 30 niA,
=
0.04 S, /?j = 100 Q and
= 400 n .
C H EC K :
Figure P6.25
0.09 A
Figure P6.28
= 8 mA
= 60 Q,
= 2.
sumed by Rj.
CH ECK:
OP AM P PROBLEM S
2 9 .(a) Find the Thevenin equivalent seen at the
terminals A-B for the op amp circuit o f
Figure P6.29.
263
C H EC K :
= 5000 Q
3 2 .(a) Find
the
Thevenin
and
Norton
Figure P6.29
C H EC K : /?,/,=/?3
30. (a) Find the Thevenin and Norton equiva
lents seen at the terminals A-B for the op amp
circuit o f Figure P6.30.
Figure P6.32
AN SW ER: ( b ) = A*,,
=0
3 3 .(a) Find
the
Thevenin
and
Norton
= 4 V
20 kQ
= 0.9 watts
50 kO
in terms o f
15 kn
2 64
N
Linear resistive
network with
dependent sources
and fixed
independent
sources
Power
Supply
Figure P6.36
Tabic P6.36
3 5 .(a) Find the Thevenin equivalent seen at the
terminals A-C for the op amp circuit of
/^l (mA)
12
= 12 V.
C H E C K : ( b )P ,,,= 2 m W
Figure P6.35
C H EC K :
/W =
144
2800
10
54
40
66
C H E C K : (b)
.5625 W
= 6 mW
be changed
Table P6.38
(niA )
What i s />?
2
10
10
j>
26S
Linear resistive
network with
dependent sources
and fixed
independent
sources
C H EC K :
Figure P6.38
= 10.667 mW
and a signal
R, (2)
200
>
1200
>
^''o-
31.25 m\V
= R j - R^
^m
Table P6.40
(pA)
R, (M Q)
2
0.4
>
10
>
N
Linear resistive
network with
dependent sources
and fixed
independent
sources
Figure P6.40
=4 V
'Fhe experimental
266
= co,
= q, where the
R ,l,+ R ,
1+
Rm/
'0
ANSWT.R: 0.5 Q, 4 A
46. Repeat Problem 45 for the measurement
curve shown in Figure P6,46, Then determine
the value of a load resistor for maximum power
transfer and compute
.
Figure P6.42
43. The Thevenin equivalent o f a linear resistive
network containing no independent sources is to
be found experimentally using the method o f
Problem 41. fhe voltmeter has an internal resist
ance R^^^= 20 k n . The dc signal generator has an
internal resistance, R^ = 2 kI2. The following
measurements are taken: (i) with R = 0,
is
0.2 V
0.1 mA
0.7 V
10.1 mA
and
dated voltage
267
the load.
R
scale)
' 6
Figure P6.50
(a)
Figure P6.47
ANSWl-.R: (a)
= 0.
= 2 kl2; (h)
R-, = 480
and
e
2 mA
12 kO
8kn
24 V
6kfi
Figure P6.48
C H EC K : 7.5 watts
Figure P 6.51
49. For the circuit o f Figure P6.49, R^ = 900 Q,
R j= 180
/?3 = 50 Q,
= 60 mA, and v^ 2 =
for maximum power
for maximum
transfer and
40 V
Figure P6.49
lon
50. For the circuits o f Figure P6.50,
= 10 V
Figure P 6 .5 2
268
>^ v.,(V)
80 -
needed for
<
--- 0 +
a
o b
40 -
(a)
>
1
0.2
1
1
0.4
i(A)
^
r
(b)
Figure P6.56
57. (a)
Figure P6.53
resistor
which
by roots([aO al a2]).
(b)
son
15Q
30 V
15
in 2 Q
increments.
^ (b
Load
(a)
R.
30 Q
(b
+ V
15Q
.s v Q
30 V
Figure P6.57
(b)
C H EC K : (a)
Figure P6.54
10 watts
= 2S mW
C H A P
Inductors and Caoacitors
REGULATION
RECTIFICATION
O
O
SMOOTHING
This process o f converting ac to dc has three stages: First, the ac waveform is rectified into its
absolute value. Then a smoothing operation takes place that reduces the variation in the voltage
to a reasonable but still unacceptable level. This first level o f smoothing is nccessar) becausc the
voltage regulator is a precision subcircuit that requires a fairly constant voltage for its proper oper
270
ation. The partially smoothed waveform is fed into a voltage regulator, which limits the voltage
oscillation between critical levels even when the load drawn by any connected device (e.g., your
computer) varies in the course o f its operation.
As mentioned, the rectified sine wave is smoothed before entering the voltage regulator. A crude
smoothing can be accomplished with a capacitor, a device studied in this chapter. Intuitively,
capacitors resist voltage changes and are designed to steady the voltage at a constant level. In this
chapter, we will study the capacitor and investigate a simplified smoothing operation for a power
supply.
CH APTER O BJECTIVES
1.
Define the notion o f inductance and introduce the inductor, whose terminal voltage is
proportional to the time derivative o f the current through it.
2.
3.
Investigate the ability o f an inductor to store energy and the computation o f the equiva
lent inductance o f series-parallel connections.
Define the notion o f capacitance and introduce the capacitor, whose current is propor
tional to the time derivative o f its terminal voltage.
4.
Investigate the ability o f a capacitor to store energy and the computation o f the equiva
lent capacitance o f series-parallel connections.
5.
CHAPTER O U TLIN E
1.
2.
Introduction
The Inductor
3.
The Capacitor
4.
5.
6.
7.
8.
1. IN TRO D U CTIO N
This chapter introduces two new circuit elements, the linear inductor and the linear capacitor,
hereafter referred to as an inductor and a capacitor. The inductor, shown in Figure 7.3, is a device
whose voltage is proportional to the time rate o f change o f its current with a constant o f propor
tionality I , called the inductance o f the device, i.e.
as set forth in equation 7.1. The unit o f the inductance Z., is the henry, denoted by H.
Macroscopically, inductance measures the magnitude o f the voltage induced by a change in the
current through the inductor.
The capacitor, shown in Figure 7.15, is a device whose current is proportional to the time rate o f
change o f its voltage, i.e.,
ic(0= C
dvcO)
(it
as set forth in equation 7.5. Here, the constant o f proportionality, C, is the capacitance o f the
device with unit farad, denoted by F. Capacitance measures the devices ability to produce a cur
rent from changes in the voltage across it.
By adding the inductor and the capacitor to the previously studied devices (the resistor, inde
pendent and dependent sources, etc.), one discovers an entire panorama o f possible circuit
responses, to be explored in the next four chapters. Together, these devices allow one to design
radios, transmitters, televisions, stereos, tape decks, and other electronic equipment. In this chap
ter, our goal is to understand the basic operation o f inductors and capacitors.
2. TH E IN DU CTO R
SomePhysics
In Figure 7.1, a changing current flowing from point A to point B through an ideal conductor
induces a voltage
pendently observed the same phenomenon at about 1831. The induced voltage,
to be proportional to the rate o f change o f current, i.e.,
was found
= .
dt
(a)
(b)
(c)
(d)
(e)
FIGURK 7.2 (a) Ramp currcnt inputs to iincoilcd and coilcd wire, (b) Six feet of #22 wire attached
to a current generator, (c) Voltage responses to ramp current inputs of uncoiled wire.
(d) Six feet o f #22 wire coiled into 45 turns 1 long and 1 in diameter.
(e) Voltage responses to ramp current inputs o f coiled wire.
273
Now suppose the wire is coiled into a qrlinder 1 in diameter and 1 long, as in Figure 7.2d. Apply
the same ramp currents o f Figure 7.2a to the coiled wire. This time, the measured responses are as
shown in Figure 7.2e. These responses have the same shape as those of Figure 7.2c, except for the
offsets o f 30 mV and 60 mV, respectively. These offiet voltages are proportional to the derivatives
of the input currents, i.e.,
Offset -
for k = 1 , 2 , where L is the proportionality constant, called the inductance o f the coil. Since the
derivative o f i s lO'^ A/sec, and the derivative o f
the coil can be computed as
3x10
Offset
0.03
= j: = ^
10^
ini^t)
0.06
of
,
.
hennes
2x10^*
dt
As mentioned earlier, the heniy, equal to 1 volt-sec/amp and abbreviated H, is the unit of induc
tance. Also, from the above experiment, one concludes that the inductance o f a cylindrical coil o f
wire is much greater than the inductance of a straight piece of wire, which in the above experi
ment was not measurable by our apparatus.
The physics of the preceding interaction is governed by Maxwells equations, which describe the
interaction between electric and magnetic fields. A time-varying current flow through a wire creates a time-varying magnetic field around the wire. The magnetic field in turn sets up a time-varying electric field, i.e., an electric potential or voltage. One can verify the presence o f this magnetic field by bringing a compass close to a wire carrying a current. The magnetic field surrounding
the wire will cause the compass needle to deflect. Physically speaking, a changing current causes a
change in the storage o f energy in the magnetic field surrounding the conductor. The energy trans
ferred to the magnetic field requires work and, hence, power. Because power is the product of volt
age and current, it follows that there is an induced voltage between the ends of the conductor.
W hat is even more interesting is that if a second wire is immersed in the changing magnetic field
of the first wire, a voltage will be induced between the ends of the second wire. A proper (mathematical) explanation of this phenomenon is left to a fields course. For our purposes, three fects
are important: (1) energy storage occurs, (2) the induced voltage is proportional to the derivative
o f the current, and (3) the constant of proportionality is called the inductance of the coil and is
denoted by L.
As mentioned, a straight wire has a very small inductance, whereas a cylindrical coil o f the same
length o f wire has a much greater inductance. This inductance can be increased many times over,
possibly several thousand times, simply by putting an iron bar in the center o f a cylindrical coil.
Alas, the calculation o f inductance is the proper subject of more advanced texts, e.g., on field the
ory or transmission line theory. Nevertheless, there are empirical formulas for estimating the
inductance of a single-layer air-core coil as described in the homework exercises.
274
dt
i,(t)
h/Y Y V
+ V jt) FIG U RE 7.3 The inductor and its differential voltage-current relationship
as per the passive sign convention.
0.5H
S o lution
From equation 7.1, direct differentiation o f the inductor current /^(/) leads to
y^(t) = 0 . 5 - " '
dt
^ = 0 .5 (-2 f)e-''
t> 0 .
The differential equation 7.1 has a dual integral relationship. Safely supposing that at / =
inductor had not yet been manufactured, one can take
= 0, in which case
the
275
(7.2)
L-fh
The time
represents an initial time tiiat is o f interest or significance, e.g., the rime when a switch
specifies the initial current flowing through the inductor at ^q. This quantity,
sums up the
entire past history o f the voltage excitation across the inductor. Because o f this, the inductor is said
to have memory.
EXA M PLE 7 .2
For the circuit o f Figure 7.5a, determine /^(O) and ij{t) for / > 0 when Vj{t) =
V as plotted in
Figure 7.5b.
\(t)
ijt)
,(t,
L = 0.5H
(b)
(a)
FICJURE 7.5 (a) Simple inductor driven by a voltage source, (b) Source waveform Vj{t).
(c) Resulting inductor current
1~ G
S olution
T,
= z j l .
1
i / o " " ' ' = z + i
It follows that
2. Repeat Example 7.2 with L = H and with v,{t) = cos(27ir) V for r > -0 .2 5 sec and zero
I
.
4jt
otherwise.
ANSWl'.R: //(()) = 2 A, /y(/) = 2 + .sinUni) A for r > 0.
E X A M PLE 7 .3
Consider the circuit o f Figure 7.6a with voltage excitation v^{t) shown in Figure 7.6b. Find the
inductor current /^(r) for f > 0, assuming that /^(O) = 0.
i,(t)
,(t )
L = 0.5H
(a)
FIGURE 7.6 (a) Voltage source driving inductor, (h) Square wave excitation !(/).
S o lution
It is necessary to apply equation 7.2 to each interval, [0, 1], [1, 2], ... , [;;, n + 1], .... For this we
need to first specify the initial conditions for each interval.
Step 1. Compute i^\). From equation 7.2,
Step 3. Compute the iuitial condition for the interval [n,u + 1] for u even. Again from equation 7.2,
with t = n and n even, we liave
1
/I-1
+j
/j
f
ri-1
y i(r)d T = J
v7 (tV /t = 2
n-l
since n - 1 is even.
Step 5. Compute ij^t) over [n,n + 1] ivith n even. If n is even, then the value o f the inductor cur
rent over the interval [;/,;/ + 1] is
l^Jn
Observe that i^it) = 2 t - In A is the equation of a straight line having slope +2 and^-intercept -In .
Step 6. Compute i^(t) over [n,n + 1] with n odd. If n is odd, then for the inter\'al [n,n + 1], the
inductor current is
/^(/) = //(/2) + ( dT = i i i n ) - 2{t - n ) = 2 - 2{t - n) A
[^J It
Here, i^{t) = 2 + 2n - 2t is the equation o f a straight line, with slope - 2 and )'-intercept 2+2.
Step 7. Piece segfnentsfivm steps 5 and 6 together. Thus the segments computed in steps 5 and 6 inter
cept the /-axis at the same points. Figure 7.7 sketches the resulting triangular response for /> 0.
FIGURE 7.7 Triangular shape o f inductor current for the square wave voltage excitation of
Figure 7.6b applied to the circuit of Figure 7.6a.
27H
Exercises. (All time is in seconds.) 1. Again consider the circuit o f Figure 7.6a. Compute iyr (?) for
(i) 0 < r < 1, (ii) 1 < t < 3 , and (iii) 3 <t For the waveform o f Figure 7.8a, assuming i^(0) = 0.
2. Again consider the circuit o f Figure 7.6a. Compute i^(^) for (i) 0 < r < 1, (ii) 1 < ^ < 3, (iii) 3 <
( < 4, and (iv) 4 < /, for the waveform o f Figure 7.8b, assuming i^(0) = 0.
It is important to recognize that the square wave voltage input o f Figure 7.6b is discontinuous but
the current waveform o f Figure 7.7 is continuous. Integration (computation o f area) is a smooth
ing operation: it smoothes simple discontinuities. This means that the inductor current is a con
tinuous function o f t, even for discontinuous inductor voltages, provided that the voltages are
bounded. A voltage or current is bounded if the absolute value o f the excitation remains smaller
than some fixed finite constant for all time. Thus, equation 7.2 leads to the continuity property
o f the inductor: if the voltage Vf{t) across an inductor is bounded over the time interval /] < t <
then /^(^o) =
^
The notation
and + on /q is used to dis
tinguish the moments immediately before and after /q- For example, in Figure 7.9, t = 2 shows a
discontinuity o f
be seen as the limiting value o f z^^(r) when approaching r - 2 from the right, whereas Vf{2~) can
be seen as the limiting value o f v^{t) when approaching t
279
\ {2 )
(U
Lu
T3
C
rtJ
01
CT>
TO
*->
o
>
u
O
c
Time (seconds)
F IG U R E 7.9 A possible discontinuous voltage v^{t) appearing across an inductor of 1 H,
and the resulting continuous inductor current.
PotverandEnergy
Rccall that the instantaneous power absorbed by a devicc is the product o f the voltage across and
the current through the device assuming the passive sign convention. For an inductor,
dt
Plit)=\'L{t)ilU) =
where
w atts.
Since energy (absorbed or delivered) is the integral o f the instantaneous power over a given time
interval, it follows that the net energy
(h
/
(7.3)
/(/,)-//^(/q ) joules.
time interval o f length T, the net stored energ)' in the inductor is zero because
= /^(/-q + 7)
280
forces equation 7.3 ro zero. To further illustrate this propert)', consider Figure 7.10a, which shows
a 0.1 H inductor driven by a periodic current /^(^) = sin(27tr) V. This current signal has a funda
mental period T = I, i.e., the smallest 7 'over which the signal repeats itselh From equation 7.3,
VV^(OJ) =
pi iOdt = I L/7(l)
L/7(0) = 0
However, we can interpret this result in terms o f the waveform o f pi{t). First note that the volt
age across the inductor in Figure 7.10a is Vjit) = 0.27Tcos(27if) V. Hence, the instantaneous power
is pf{t) =
regions o f Figure 7.10b in which the area under the power curve has equal parts ot positive and
negative area. This means that all the energ)' stored by the inductor over the part o f the cycle o f
positive power is delivered back to the circuit over the portion of the cycle when the power is neg
ative. Fhis is true for all periodic signals over any period. Because no energ)' is dissipated, and
because energy is only stored and returned to the circuit, the (ideal) inductor is said to be a loss
less device.
IlW
v jt)
sln(27T) A
0.1 H
= 0.2n cos(2n) V
(a)
P lW
(b)
FIGUllE 7.10 (a) Inductor excited by periodic current, (b) Plot of the power absorbed by the inductor.
It is convenient to define the instantaneous stored energy in an inductor as
(7 .4 )
281
for all t. Equation 7.4 can be viewed as a special case o f equation 7.3 in which r,) = -oo and /^(-oo)
= 0. Thus, equation 7.4 can be interpreted as the change in stored energ)' in the inductor over the
inten'al (^x>, t].
E XA M PLE 7 .4
Find the instantaneous energy stored in each inductor o f the circuit o f Figure 7.11 a for the source
waveform given in Figure 7.1 lb. In Figure 7.1 lb , note that ij^t) = 0 for r < 0.
FIGURE 7.11 (a) Series inductors excitcd by a source current, (b) Graph o f the source current.
S olution
for all t. Since i^{t) = 2r A for 0 < r < 1 and i^{t) = 2 A for /> 1, equa
tion 7.3 or 7.4 immediately yields the instantaneous stored energies (in J) as plotted in Figure 7.12:
0^/<l
1
4r
0:sr<l
1s
> t
> t
(a)
(b)
FIGURE 7.12 (a) Encrg)' (in J) stored in inductor Z.,. (b) Energy (in J) stored in inductor Ly
282
Exercises. 1. For the circuit o f Figure 7.1 la, find analytic expressions for the instantaneous stored
energ)' for the current excitation in Figure 7.13a for r > 0.
2. Repeat Exercise 1 for Figure 7.13b.
0 .2 5 r
UV/) =
{) ^ / < 2
t-
()s/<l
0 .2 5 (9 - 6/+ / )
0 s /< 2
2s/
4/
0 s /< 1
(9-fv + r )
I s / <3
3</
L^ = 1H
FIGURE 7.14 Parallel inductive circuit with switch in which v^{t) = cos(t) V for /> 0 and 0 otherwise.
283
S olution
Step 1. Since no voltage is applied to either inductor for / < 0, /,(0) = 0. Further, no voltage
appears across the second inductor until r > 1. Hence, /^(l) = 0.
Step 2. Equation 7.2 implies that, for 0 < r < 1,
( s ( 0 = /i(/) =
L|
v ,(T )r/ T =
-jL|
/ ^ c o s (t)Jt =
sin(/) A
Step 3. At ^ = 1, the switch closes. T he r\vo inductors are then in parallel, and the source voltage
appears across each. Hence, by equation 7.2,
co s (t
cos(T)r/T = sin(/) - s i n ( l ) A
From the KCL, the input current ij^t) = /,(/) + ijit) = 2sin(/) - sin(I) A for / > 1.
Step 4. Compute the energy stored in the inductors over the interval [0, t]. From equation 7.3, it fol
lows that for 0 < r < 1,
->((), t) = 0.
Step 5. Compute the energy stored in the inductors over the interval [0, f] for 1 < t. Again from equa
tion 7.3, for 1 < t,
sin^(l)] joules.
28-
3. THE CAPACITO R
DefinitiojisandProperties
D EFIN ITIO N OF T H E C A PA CITO R
Like the inductor, the capacitor, denoted by Figure 7.15a, is an energy storage device.
Physically, one can think o f a capacitor as two metal plates separated by some insulating mate
rial (called a dielectric) such as air, as illustrated in Figure 7.15b. Placing a voltage across the
plates o f the capacitor will cause positive charge to accumulate on the top plate and an equal
amount o f negative charge on the bottom plate. This generates an electric field between the
plates that stores energy. Hence, for a capacitor.
(7.5)
clt
d\
where q{() is the accumulated charge on the top plate, which is proportional to the voltage
V({t) across the plates; thus q{i) = Cv^t), with proportionality constant C denoting capaci
tance and having the unit o f Farad (F). One Farad equals 1 amp-sec/volt. The capacitance C
is a measure o f the capacitors potential to store energ)' in an electric field.
ic(t) >r
^
V ,(t)
(a)
+ -h
-i- + 4-
4- + + -I--1- +
A- + + + + + +.
(b)
FIG URE 7.15 (a) The symbol for the capacitor with conventional voltage and current direc
tions. (b) Illustration o f electric field between plates of a parallel-plate capacitor.
Modern-day capacitors take on all sorts o f shapes and sizes and materials. In keeping with craditio n , the parallel-platc concept remains the customar)' perspective. Calculating the capacitance o f
t^vo arbitrarily shaped conducting surfaces separated by a dielectric is, in general, ver)- difficult
Fortunately, the ordinary capacitor o f a practical circuit is o f the parallel-plate variety, with the
plates separated by a thin dielectric. The two plates are often rolled into a tubula, ruu, .,>1
complete structure is sealed.
EXA M PLE 7.6
For the capacitor circuit o f Figure 7 . 16a, compute i^t) when v j t ) = r5""'sin(1000r) V for / > 0.
28S
ic(t)
v Jt )
2mF
(a)
Time in milli-seconds
(b)
i^{t) =
sin(1000/) +
dt
cos(l()00/) A
ANS\V1-:R: - 0 . 0 5 , A.
2. Repeat Exercise 1 with
= e~-^^ cos( 100/) V for f > 0 but plot over die time interval [0, 0.15 sec].
A.\S\V1-:K: -.- -^ q 0 .0 5 COS. KJOr) r 0.2 sin(lOOr)] A.
The differential relationship o f equation 7.5 has the equivalent integral form
' t < '> =
^ f J c W d T + I j ; " ic W d T
(7.6)
C '^0
where
=0
because the capacitor was not manufactured at t = -oo. The time /q represents an initial time o f
interest or significance, e.g., the time when the capacitor is first used in a circuit. The quantit)'
286
specifies the initial voltage across the capacitor at ^q. This initial voltage,
past history o f the current excitation into the capacitor. Because o f this, the capacitor, like the
inductor, is said to have memory.
EXA M PLE 7 .7
Suppose a current source with sawtooth waveform
0.5 F capacitor (zero initial voltage) as in the circuit o f Figure 7.17a. Compute and plot the volt
age across the capacitor.
i.(t) A/v^(t)V
(a)
(b)
The input waveform is periodic in that it repeats itself every 2 sec. Therefore, the solution will pro
ceed on a segment-by-segment basis.
Step 1. Consider the interval 0 < f < 2. For this interval ij^t) = {It - 2) A. With
= 0, it fol
age over the interval 2 < /< 4 is simply a right-shifted version o f the voltage over the first inter
val. Right-shifting is achieved by replacing t with t - 2 . In other words,
...
Lastly, obser\^e that the voltage across the capacitor, as illustrated in Figure 7.17b, is continuous
despite the discontinuity o f the capacitor current. Again, this follows because the capacitor volt
age is the integral (a smoothing operation) o f the capacitor current supplied by the source.
287
Exercise. Consider the capacitor circuit o f Figure 7.18. Suppose the current source is i^{t) = e~ A
for /> 0 and
and the
+
I (t) = e-'u(t) 0
20
M O
0.5 F
v,(t)
FIGURE 7.18 Scries RC circuit driven by a currcnt source for accompanying exercise.
AN SW ERS: v^^t) = 3 -
for t> 0,
It is important to emphasize that the sawtooth current input depicted in Figure 7.17b is a dis
continuous function, but the associated voltage waveform is continuous because integration
(equation 7.6) is a smoothing operation. This means that the capacitor voltage is a continuous
function o f t even for discontinuous capacitor currents, provided they are bounded. This obser
vation leads to the continuity property o f the capacitor: if the current i(^t) through a capacitor
is bounded over the time interv'al
< ^ < ^2> then the voltage across the capacitor is continuous
At the macroscopic level, there appear to be some exceptions to the continuit}' propert}' o f the
capacitor voltage, e.g., when two charged capacitors or one charged and one uncharged capacitor
are instantaneously connected in parallel. In such cases, KVL takes precedence and will force an
instantaneous equality in the capacitor voltages, subject to the principle o f conservation o f
charge, to be discussed shortly. Another example is w'hen capacitors and some independent volt
age sources form a loop. When any o f the voltage sources has an instantaneous jump, so will the
other capacitor voltages. Upon closer examination, however, we see that there is really no excep
tion to the stated continuity rule: it can be shown that in all o f the cases where the capacitor volt
age jumps instantaneously, an impulse current flows in the circuit. Physically, an impulse cur
rent is one that is ver)' large (infinite from an ideal viewpoint) and o f very short duration. The cur
rent is not bounded, and consequently, the capacitor voltage may jump instantaneously. This
jump does not violate the rule, which presumes that the currcnt is bounded.
Relatio7ishipofChargetoCapacitorVoltageandCurrent
We have defined the capacitance o f a two-terminal device strictly from its terminal voltage-current
relationship the differential equation 7.5 and the integral equation 7.6, which is now repeated:
v 'c ( 0 = V c(fo) + ^ f ^
C 'M)
Physically speaking, the integral o f i(^t) over [/q, t\ represents the amount o f charge passing
through the top wire in Figure 7.19 over [rQ, r].
288
ic(t)
/ ''+ + +q + + /
/+ + +
^
/ ++++++
A- + + + + + +y
-q
FIGURE 7.19 Capacitor cxcitcd by a currciu.
Bccausc o f the insulating' material (the dielectric), this charge cannot pass through to the other
plate. Instead, a charge o f +q{t) is stored on the top plate, as shown in Figure 7.19. By KCL, if
to be deposited on the bottom plate. The positive and negative charges on these two
plate. For a linear capacitor, the only t)'pe studied in this text, the value o f V(^t) is proportional
to the charge
qit) = C\U)
(7.7)
where q{t) is in coulombs, Cis in farads, and t^i^) is in volts. Thus, equation 7.6 has the following phys
ical interpretation: the first term,
ic(T )d r.
represents the additional charge transferred to the capacitor during the interval [r,j, /]. Dividing
this integral by Cgives the additional voltage attained by the capacitor during [^q, ^]. Therefore,
the sum o f these rwo terms, i.e., equation 7.6, is the voltage o f the capacitor at r. Since q(/) =
it follows direcdy that
=
(7.8)
(It
cl!
ThePj'hicipleofConservationofCharge
It is important in terms of modern trends in circuit applications to further investigate the rela
tionship o f charge to capacitor voltages and currents. The principle o f conservation o f charge
requires that the total charge tramferred into a junction {or out o f a junctiori) be zero.~ This is a direct
consequence of KCL. To exemplify, consider the junction o f four capacitors shown in Figure 7.20.
289
v,(t)
+ V ,(t) -
i,(t)
i3(t)
- V3(t) +
v,(t)
+
l4(t)
/ ^ (m
'2
+'3
M^
(7.9)
where qj^{t) is the charge transferred to capacitor k. By equation 7.6, at ever}' instant of time,
qi{t) = C-v.{t)
(7.10)
which defines the relationship between transported charge, capacitance, and the voltage across the
capacitor. Hence, from equations 7.9 and 7.10, at every instant of time,
C^v^it) + C2 V2 U) +
+ C^v^{t) = 0
This simple equation relates voltages, capacitances, and charge transport. The following example
provides an application o f these ideas.
EXA M PLE 7 .8
This example shows that under idealized conditions, capacitor voltages can change instanta
neously. Consider the circuit o f Figure 7.21 , in which ^q(0~) = 1 V and
and V(^{t) for f > 0.
= 0 V. Find
290
t=0
C l = 1F
C2
Cl
C2 = 1F
FIGURE 7.21 Two parallel capacitors connectcd by a switch.
S o lution
transferred between the capacitors to equalize the voltages. According to the principle o f conser
vation o f charge, the total charge before and after the transfer is the same. Thus, conservation o f
charge requires that
v - c , ( 0 ^ ) - l l + C 2 [ i c i ( 0 ^ ) - 0
= Vq^{0*) = 0.5 V.
Exercises. 1. In Example 7.8, make C, = 0.75 F and C2 = 0.25 F, and compute ^/^(O'^).
AN SW ER: /Y-,(0") = 0.75 V.
2. In Example 7.8, sufipose ^q ( 0 " ) =10 V and Vqj,{Q~) = - 8 V. Also let C, = 0.75 F and C-, = 0.25
F. Compute
ANSW ER: /.v.,(0^) = 5.5 V.
Example 7.8 is illustrative o f a charge transport that is germane to switched capacitor circuits,
which are o f fundamental importance in the industrial world.
EnergyStorageinaCapacitor
As with all devices, the energy stored or utilized in a capacitor is the integral o f the power absorbed
by the capacitor. The net energ}' entering the capacitor over the interval [/q, /J is
Jff,
2 )1
chciT )]
= c r '' (
(It
Vcih)
Oo)
dT =
(7.11)
= -C
for C in farads,
in volts, and energy in joules (J). From equation 7. 11, the change in energy
stored in the capacitor over the inter\'al [rQ, r j depends only on the values o f the capacitor volt
ages at times /q and
i.e., on
+ T) for some r > 0, then over any time interval [t, / + 7], the change in
+ 7) =
to zero. Analogous to the inductor, for all periodic voltages, the capacitor stores energy and then
returns it to the circuit and is thus called a lossless device.
As with the inductor, it is convenient to define the instantaneous stored energy in a capacitor as
Wc{t) = l^Cvc{t)
(7.12)
which is really the integral o f power over the interval (x , /], assuming that all voltages and cur
rents are zero at r = - x .
E X A M PLE 7 .9
Consider the circuit o f Figure 7.22, in which
= 0. It is known that
for f > 0, the source current is i^{t) and the voltage across the capacitor is
V for ^ > 0. Compute (i) the energ)', in joules, stored in the capacitor for
/ > 0, (ii)
and (iii)
(i) Since
VV^^(0,/) = lc v J ( / ) = 8 C / ?
\-e
RC
<li
RC
) - 4/? \ - e
2 )2
(iii) To find
we first compute
\ -e
Thus
1- e-
+ 4^
E X A M PLE 7 .1 0
= 4 A
ijt )
For the circuit o f Figure 7.23a, it is known that the voltage across
the capacitor is
+
N
5mF
k
-
T im e t in seconds
(b)
F IG U R E
S o lution
Step 1. Compute
dt
= 0 .2 co s 2t + -
6}
= 20 sin(7r/6) = 10 V, we obtain
2t + - - 0 . 2 5 J
6}
Plots o f pf4f) i^nd V\^^0, t) arc given in Figure 7.23b. Notice that WT^O, /) can be negative,
because W^oo, 0) = 0.25 joules, meaning that at r = 0, there is an initial stored energy that can
be returned to the circuit at a later time . Figure 7.23b substantiates this.
Sei'iesInductors
Just as resistors in series combine to form an equivalent resistance, inductors in series combine to
form an equivalent inductance. As it turns out, series inductances combine in the same way as
series resistances.
E X A M PLE 7 .1 1 .
Compute the equivalent inductance o f the series connection o f three inductors illustrated in
Figure 7.24. Then find the voltages
Leq
+
V ..
+
Leq
Vl2
+
V .,
Q-
o+
^eq
o -
(a)
FICIJRH
Leq
(b)
29-4
S o lution
First we must answer the question o f what it means to be an equivalent inductance. Earlier, we
defined the inductor in terms o f its terminal voltage-current relationship. Two 2-terminal induc
tor circuits have the same inductance if each circuit has the same terminal voltage-current rela
tionship as defined in equation 7.1.
Step 1. The voltage labeled
il
equivalent inductance,
O ur goal is to express
Step 2. Find
(7.13)
dt
dt
it follow's that
^'Uq - (^1 + ^2 + ^ 3 )
dt
Hence, the series inductors o f figure 7.24a can be replaced by a single inductor with inductance
= -^-1 + ^2 +
Finally, since
= Lj
dt
= Lj ^
dt
and
_
dt
= { L, +L^ +
~
'
dt
, it follows that
Lj
(L| 4- Zy-> -l- L-^)
in terms o f
3
= 5 mH, and
= 1 mH, find
295
Extension o f the formulas in the above example to n inductors is fairly clear, and we state the
results without rigorous proof: the formula for series inductances is
(7.14a)
and the formula for voltage division o f series inductances is
^'IJ =
(7.14b)
L\ + Lo + ... + Lfj
InductorsinParallel
The same basic question as with inductors in series arises with a parallel connection o f inductors:
what is the equivalent inductance? Rather than derive the general formula, let us consider the case
o f three inductors in parallel, as illustrated in Figure 7.25a.
E X A M PLE 7 .1 2
For this example our goal is to show that the equivalent inductance o f the circuit o f Figure 7.25a
is given by the reciprocal o f the sum-of-reciprocals formula,
- ~\
i
T
+ +
(7.15)
Ly
Leq
O +
L3
L2,
^eq
Leq
L,
L,
o-
o (a)
(b)
FIG URE 7.25 (a) Parallel connection of three inductors, (b) Equivalent inductance.
So l u t io n
Once again, equation 7.13 defines the relationship for the equivalent inductance:
^U-q
The goal is to construct
in terms o f Z ,,
^eq
and
+ 'L 3
2%
(Hl\ ^
dt
dt
ill
dt
di,
Step 2. Find Ul in terms o f a n d L^,. From equation 7.1, For each inductor
dt
dt
Li
di,
dt
( \
Ln
1
n
+ +
Li-q
L]
L t^)
L'S
This has the form o f equation 7.13, which implies equation 7.15, i.e.,
Le, =
+ ---- +
Z/j
L~i L,'^
lb generate a currcnt division formula we first note that
= ^/1 =
1 '
1 '
= J y i J c i T ) d T = J \ ' , ^ { T ) d T and
^ _-TT
'
= / v^,^(tV/t
^ _nr.
Thus
(/)
Exercises. 1. If, in Flxample 7.12, Z.j = 2.5 mH, Z., = 5 mFl, and
ANSWq-R:
= 1 mH, find
= 0.625 mH.
The above arguments easily generalize. Suppose there are u inductors, /,,, Z-,, ... , Z.,^, connected in
parallel. Then the equivalent inductance is given by the reciprocal o f the sum-of-reciprocals formula.
(7.16a)
ii.jn
~i-------- [
(7.16b)
Exercise. For two inductors Zj and L-, in parallel, show chat the equivalent inductancc satisfies the
formula
(7.17)
Series-ParallelCombinations
This subsection examines series-parallel connections o f inductors. This allows us to use the for
mulas developed above in an iterative way.
EXA M PLE 7 .1 3
Find the equivalent inductance,
S o l u t io n
Step 1. In the circuit o f Figure 7.26, several inductors are enclosed by an ellipse. Let
denote
the equivalent inductance o f this combination. Observe that the series inductance o f the 5/6 H
and 0.5 H inductors equals 4/3 FI. This inductance is in parallel with a 1 H and a 4 Fi induc
tance. Hence,
, =
I
1 I
3= 1
!----- !---1
Step 2. The equivalent circuit at this point is given by Figure 7.27. This figure consists o f a series
combination o f a 1.5 FI and a 0.5 FI inductor connectcd in parallel with a 6 H inductor. It fol
lows that
29S
-------------+ 0 .5 + 1.5 6
= -= 1 .5
Exercise. In Example 7.13, suppose the 5/6 H and 0.5 H inductors are both changed to 0.4 H
inductors. Find L o f the circuit.
AN SW ER: 1.443 H.
CapacitorsinSeines
Capacitors in series have capacitances that combine according to the same formula for combining
resistances or inductances in parallel. Similarly, capacitances in parallel combine in the same way
that resistances or inductances in series combine. This means that the equivalent capacitance o f a
parallel combination o f capacitors is the sum o f the individual capacitances, and the equivalent
capacitance o f a series combination o f capacitances satisfies the reciprocal o f the sum-of-reciprocals rule. These ideas are illustrated in the examples to follow.
EXAM PLE 7 .1 4
Compute the equivalent capacitance,
2)9
o-
Cl
'C2
o+
C3
eq
Q-
(a)
(b)
FIGURI^ 7.28 (a) Series combination of three capacitors, (b) Equivalent capacitance,
So l u t io n
The equivalent capacitance denoted in Figure 7.28b is defined implicitly by the current-voltage
terminal conditions according to equation 7.5, i.e.,
; - r
dt
Our goal is to express this same terminal v-i relationship in terms o f the capacitances, C j, C 2, and
Ck - Q
But, by KCL, i^ =
dt
Hence,
dt
C,
Differentiating this expression with respect to time and using the result o f step 1 yields
dv/^
d v f^ ]
d v 'c '')
d v (--i
\ .
^ = + + = \ + + \ir
dt
dt
dt
dt
l,C|
Cj
c J
3D0
Step 3. Compute
\
iQ -
[C l
dt
H--------h --
Cl
dt
C^j
It follows that
Q ,=
C,
C3
= /q =
1 '
1 '
'
''CA-(0 = J ick^T)dr = J ic{T)dT and C,^ Vci 0 = f ic(r)dT
^ -rr-
^ rri
Thus
''a - ( 0 = f ic(r)dT = -
C2
C3
= 16 pF. Compute
in terms of
.ANSWER: /V;: = 0 .1 6 /y ;
Generalizing the result o f Example 7.14, we may say that capacitors in series satisfy the reciprocal
o f the sum-of-reciprocals rule. Thus, for n capacitors C j, C2, ... , C^, connected in series, the
equivalent capacitance is
1
- 1 -------------- r
+ + ... +
C,
G
C
(7.18a)
------- i---------
+ --- +...
C,
Co
C.,
( 7 .1 8 b )
3 01
Exercise. Show that if two capacitors Cj and C-y are connected in series, then
(7.19)
c ,+ c .
CapacitorsinParallel
If rvvo capacitors are connected in parallel as in Figure 7.29a, there results an equivalent capaci
tance
'c
Cl
o
+
'C2
eq
a-
a(a)
(b)
FIGURE 7.29 (a) Parallel combination of two capacitors, (b) Equivalent capacitance,
Since the voltage
Hence,
^eq -
^2
One surmises from the above example that, in general, capacitors in parallel have capacitances that
add. And, indeed, this is the case: if there are n capacitors C j, C2, ... ,
lent capacitance is
C
= C, + C2 + ... + C
(7.20a)
Cf,
Ceq
Q + Q
Q
C] + C j + + C
'c
( 7 .2 0 b )
302
Series-ParallelCombinations
Wc round out our discussion o f capacitance by considering a simple series-parallel interconnection.
EX A M PLE 7 .1 5
Consider the circuit o f Figure 7.30. Compute the equivalent capacitance,
0.45 mF
0.6 mF
o l u t io n
Step 1. Combine series capacitances. Observe that the rwo series capacitances o f 0.5 mF and 0.5
mF combine to make a 0.25 mF capacitance.
Step 2. Combine parallel capacitances. First, as a result o f step 1, the three capacitances, 0.3 mF,
0.25 mF, and 0.45 mF, add to an equivalent capacitance o f 1 mF. Further, the two parallel capac
itances, 0.3 mF and 0.6 mF, at the bottom o f the circuit, add to make a 0.9 mF capacitance. The
new equivalent circuit is shown in Figure 7.31.
a-
II mF
1.125 mF
0.9 mF
a-
^ eq-
1 1
-------- + - + ----1.125
1 0.9
Exercise. Suppose the two 0.5 mF capacitors in Figure 7.30 are changed to 2.5 mF capacitors.
Find the new
A N SW ER: 0.4 mF.
303
FIGURE 7.32 Simple power supply with capacitive smoothing for low- power applications.
Consider, for example, the circuit shown in Figure 7.32. The four (ideal) diodes are arranged in a
configuration called a fidl-wave bridge rectifier circuit. An ideal diode allows current to pass only
in the direction o f the arrow. The diode configuration ensures that i^{t) remains positive, regard
less o f the sign o f the source current. Specifically, the diodes ensure that /j(f) =
Using the
integral relationship (equation 7.6) o f the capacitor voltage and current, it follows that
V cit)= yc(h)) + ^ f
ici-^)ch = VcUo) + ^ f
[M t)|-/o(t)]^/t
(7.21)
Because o f the difference |/j(r)| - i^it) inside the integrand o f the integral, i^{t) tends to increase
the capacitor voltage, whereas i^^{t) tends to decrease the capacitor voltage. Further, because the
diodes are assumed ideal, it follows that
(7.22)
biased. But this is impossible, because an ideal diode behaves like a short circuit when forward
biased. The consequence is that V(4,t) will be 12 V whenever |?>'^(^)| is 12 V. This occurs every 1/120
o f a second. Thus, the rectifier output will recharge the capacitor every 1/120 o f a second. Between
charging times, the current, i(){t), will tend to discharge the capacitor and diminish its voltage.
304
The design problem for the capacitive smoothing circuit is to select a value for C that guarantees
that v^t) is sufficiently smooth to ensure proper operation of the voltage regulator. Here, suffi
ciently smooth means that the maximum and minimum voltages differ by less than a prescribed
amount. To be specific, suppose that i>(\t) must remain between 8 V and 12 V. Recall that i^{t)
tends to increase the capacitor voltage, while
fairly constant. The value for /(,(^) is obtained from the specification sheet o f the voltage regulator.
Suppose this value is a constant 1 A. It remains to select C so as to ensure that V({t) remains above
8 V between charging times. From equation 7.21, it is necessary that
CJio
Now we need consider only values for t between 0 and 1/120, because the capacitor will recharge
and the process will repeat itself every 1/120 o f a second. Thus, because i^{t) will only increase the
capacitor voltage, to ensure that v^{t) remains above 8 V, it is sufficient to require that
A 2,100 |.iF capacitor satisfies this requirement. A method for computing the capacitor voltage
waveform is described in Chapter 22 o f o f 2"^ edition. However, using SPIC E or one o f the other
available circuit simulation programs, one can generate a plot o f the time-varying capacitor volt
age produced by this circuit, as shown in Figure 7.33. In the figure, it is seen that the capacitor
voltage varies between 12 and 9.02 V, which is smaller than the allowed variation o f (12 - 8) V.
Two factors contribute to this conservative design: (1) we used C = 2,100 uF instead o f the cal
culated value, C = 2,083 uF, and (2) the increase in the capacitor voltage due the charging current
is is not included in the calculation.
volts
The preceding brief introduction made several simplifying assumptions to clarify the basic use o f
a capacitor as a smoothing or filtering device. Practical power supply design is a challenging field.
A complete design would need to consider many other issues, some o f which are the nonzero
resistance o f the source, the non-ideal nature o f the diodes, the current-handling abilit)' o f the
components, protection o f the components from high-voltage transients, and heat-sinking o f the
components.
6. SUM M ARY
This chapter has introduced the notions o f a capacitor and an inductor, each o f which is a lossless
energ}' storage device whose voltage and current satisfy a differential equation. The inductor has a
voltage proportional to the derivative o f the current through it; the constant o f proportionalit}^ is
the inductance L. T he capacitor has a current proportional to the derivative o f the voltage across
it; the constant o f proportionalit}' is the capacitance C. It is interesting to observe that the roles of
voltage and current in the capacitor are the reverse o f their roles in the inductor. Because o f this
reversal, the capacitor and the inductor are said to be dual devices.
That the (ideal) inductor and the (ideal) capacitor are lossless energy storage devices means that
they can store energ)- and deliver it back to the circuit, but they can never dissipate energ)^ as does
a resistor. The inductor stores energy in a surrounding magnetic field, whereas the capacitor stores
energy in an electric field between its conducting surfaces. Unlike energ)' in a resistor, the energy
stored in an inductor over an interval [r^, fj] is dependent only on the inductance L and the val
ues o f the inductor current //(/^()) and //(/^j). Likewise, the energ)' stored in a capacitor over an
interval [r^, /,] is dependent only on the capacitance C and the values o f the capacitor voltage
and
Both the inductor and the capacitor have memor)'. The inductor has memory because at a partic
ular time Tq, the inductor current depends on the past histor}' o f the voltage across the inductor.
The capacitor has a voltage at, say, time
itor. The concept o f memory stems from the fact that the inductor current is proportional to the
integral o f the voltage across the inductor and the capacitor voltage is proportional to the integral
o f the current through the capacitor. This integral relationship gives rise to the important proper
ties o f the continuity of the inductor current and the continuit)' o f the capacitor voltage under
bounded excitations.
rhe dual capacitor and inductor relationships are set forth in Table 7.1.
Finally, we investigated the smoothing action o f a capacitor in a power supply.
306
lABLE 7.1. Summary of the Dual Relationships ot the Capacitor and Inductor
ic )
icit) = C
d\'c )
dt
V[{t) = L
dilit)
dt
LJ k>
In particular, if r, <
across an inductor
when
Coulom b: quantit)' o f charge that, in 1 second, passes through any cross section o f a conductor
maintaining a constant 1 A current flow.
Dielectric: an insulating material often used between two conducting surfaces to form a capacitor.
Farad: a me;Lsure o f capacitance in which a charge o f 1 coulomb produces a 1 V potential difference.
Faradays law o f induction: law' asserting that, for a coil of wire sufficiently distant from any mag
netic material, such as iron, the voltage induced across the coil by a time-varying current
is proportional to the time derivative o f the current; the constant of proportionality,
307
denoted Z, is die inductance o f the coil. Faradays law is usually stated in terms of flux
and flux linkages, which are discussed in physics texts.
H enry: the unit o f inductance; equal to 1 V-sec/amp.
Inductance: property of a conductor and its local environment (a coil with an air core or iron
core) that relates the time derivative o f a current through the conductor to an induced
voltage across the ends o f the conductor.
Inductor (linear): a two-terminal device whose voltage is proportional to the time derivative of
the current through it.
Instantaneous powen p(t) = */(/)/(<), in watts when v{t) is in volts and i{t) in amps.
Lossless device: device in which energy can only be stored and retrieved and never dissipated.
Lossy device: a device, such as a resistor (with positive R), that dissipates energy as some form o f
heat or as work.
Maxwells equations: a set o f mathematical equations governing the properties o f electric and
magnetic Beids and their interaction.
M emory: property o f a device whose voltage or current at a particular time depends on the past
operational history o f the device; e.g., the current through an inductor at time /q depends
on the history o f the voltage excitation across the inductor for t< /q.
Unbounded voltage or current: a voltage or current whose value approaches infinity as it nears
some instant o f time, possibly r = oo.
Voltage r^ u la to r: circuit that produces a voltage having only a small variation between set lim
its for a given range o f load variation from a fairly smooth input signal.
^The word stored emphasizes that the energy in the inductor is not dissipated as heat and can be recovered by
the circuit, whereas the word absorbed is used to mean that the energy cannot be returned to the circuit. In a
resistor, energy absorbed is dissipated as heat.
^ More generally, conservation of charge says that the total charge transferred into a Gaussian sur&ce (or out of a
Gaussian surface) is zero.
31)8
Problems
for 0 <
lOv ft)
o f tiirns)~
4 X
18 {dkimeter)+ 40 (lengths)
Figure P7.4
2mH
0.2 mH
t
I
-1
1
6
L(A)
-2 L(t)
t(s)
(a)
Figure P7.2
-2
(a)
, 0.8 mH
(b)
(t)
6 ijt)
v,{t)
0.6 mH
0.75 mH
1.5 mH
i.JA )
(b)
Figure P7.5
6. (a) Find and plot for 0 < r < 6 sec the induc
Figure P 7.3
3 0 ')
P7.8b, and
Figure P7 .6 b.
P7.8c.
i,(t)
vJt)
(a)
(b)
Figure P7.6
7. Repeat Problem 6 for
(a)
u(t), and
Figure P7.8
9. Consider the circuit in Figure P7.9 in which
Z,
= 0 .2 H , Z2 = 0 .5 H , and
310
v,(t)
% iQ
ic(t)
Figure P7.12
Figure P7.9
and
for 0 <
Figure P 7.10
(b)
Figure P7.13
pF capacitor to 100 V.
= 100[1 +
31
15. Suppose
and
and
as t
00.
O '
w
O '
Figure P7.16
(b)
0
Figure P7.15
voltage,
given
as sketched in Figures
P7.16B and c.
= 10 V. Find
0
(b) W hat is the instantaneous power deliv
ered by the dependent source?
i jt )
(a)
0
0
0
Figure P 7 .1 7
312
in Figure P 7 .19.
(a) Find and sketch
Figure P7.21
msec.
(b) What is the instantaneous power deliv
Figure P7.22
Figure P7.20.
and 2 ms < t.
(b) Compute the energy stored in each
capacitor as a function o f t for the same
time intervals.
Figure P7.20
t = 2 ms
for
for the
parallel combination.
Figure P7.23
M IX ED C A PA CITO R AN D
IN D U C TO R PRO BLEM S
24. Consider the circuit o f Figure P7.24,
313
rent waveform
(a) Compute
and 0 otherwise.
(a) Compute and sketch I'lit),
and
L.
Figure P7.27
28. For the circuit o f Figure P7.28, compute
and
Figure P7.24
as a function of
+
" 0
c:
\|>
-5
L ,2
L,
Figure P7.28
SERIES-PA RALLEL IN D U C TO R S
29. In the circuit o f Figure P7.29, all inductors
are initially relaxed at /^= 0 and /.j = 6 mH, L-,
Figure P7.25
26.(a) In the circuit o f Figure P7.26, (i = 10, C,
= 38.5 mH,
= 22 mH. A voltage
200re~' mV is applied for r > 0. Find,
Vjj(t), and
Challenge: Find
initially
uncharged.
If
vj^t) =
for / > 0.
Figure P 7 .2 6
314
(mA)
= 36
= 120cos(1000r) mA.
mH, and
(a) Find L
and
(b) Find
(c) Plot the instantaneous power deliv
ered by the source for 0 < r < 14
msec.
/
L.
-200
--
Figure P7.32
(!)
Figure P7.30
C H EC K :
12 mH, - 1 .4 4 sin(lOOOr) V, 90
equivalent inductance,
Vj^, and
cos(lOOOr) mA.
tion D.
31. For the circuit o f Figure P7.31, Z., = 260
mH,
10^"^ tiiA.
(a) Find
(b) Compute
(c) Com pute
stored in
instantaneous
energy
as a function o f t.
C H EC K S: 52 mH. 0.2
0.8 i j t ) ,
for r > 0.
31
ijt )
____rvY V
v X
L.
L,
1-6
L,
i(t)
f r r \ ____ T Y Y \
(a)
i-N/YYA___ TYYV
v(t)(V)
>k
16'
16-
L.
(b)
rOA___ TYYV
Figure P7.34
35. Find
L,
(b)
Figure P7.36
SCRA M BLED ANSWERS: 0.1, 0.08, 0.6 (in 11)
36. Find
= 40
= 180
316
38. Find L
P7.38.
4 mH
AN.SWER: C//I
for
. all / values.
40. Like Problem 39, this is a conceptual prob
lem and requires no calculations for the answer.
lOmH
1 mH
5 mH'
36 mH
Bo(a)
/YYV
7mH
/
2.4 mH
(h)
< /v,2-
1.2 mH
^eq\
0.6 mH
relationship depends
on the value o f L
Explain your reasoning.
(b)
Figure P7.38
ANSWHR: (a) 13 m il: (b) 2 in 11
39. This is a conceptual problem
and requires no calculations for
the answer. Consider circuits 1
and 2 o f Figure P7.39. All induc
tors are 1 H except the one labeled
and
presence
^eq\ ^ ^eql(b)
< K ,l(d) There is no general relationship between
and
Any relationship depends
on the value o f L.
Explain your reasoning.
Circuit 1
Circuit 2
Figure P7.40
ANSW ER: /.
SERIES-PARALLEL CAPACITORS
41. (a) Find the indicated equivalent capacitance
for the circuit o f Figure P 7.4la where C,
= 4 pF, C 2 = 3 pF, C3 = 2 pF, Q = 4 pE
31'
-C,-
Vw/'
C. c,
<!>
^ 1
Figure P7.43
44. For the circuit of Figure P 7.44, Cj = 8 mF,
C2 = 6 mF, C3 = 12 mF, and
(b)
Figure P7.41
CHECKS: 6 pF, 66 ^F
If
if if
if
c,
c.
c,
Figure P7.44
= 48 pF, Cg
= 24 pF, Cg = 24 pF.
oc
C -L
c.
C,
(a)
^eql
(b)
Figure P7.42
^ ^eq2 '
(W
on the value of C.
C2 = C5 = 40 pF, C3 = Cg = 2 0 pF.
(c) Repeat parts (a) and (b) for Cj = 12 pF, Cj
318
v jt)
c,
V,
Figure P7.48
C H EC K ;
= 6(1 -
Figure P7.46
ANSWMR: C ; <
v jt)
C. hi Figure
Figure P7.49
are 2 V,
= 5 mF, Cj = 20 mF,
(a) Find
(b) Find
= 4 mF, Q = 80 mF,
and
for r > 0 .
(b)
(a)
ANSWHK: (a)
Figure P7.47
= 8 mF
v,(t)
Figure P7.50
31')
WYiLit)
v^(/) = -
at
Figure P7.51
M ISCELLAN EO U S
52. Find and sketch
redundant as far as
is concerned?
er v^alue
0.5 F
0.2
+
pkv,
y+\
__
^ . 4 f^ . 2
-_ y
f
depends on the
i/y
0.4 F
1-^4
2 T f)
(a)
V ,(t)
a con
Figure P7.52
0.5H
(a)
C H EC K : v,,,(/) = :^ v ,(/ )
53. Using the circuit given in Figure 7.32,
select a capacitor value to filter the voltage
for a regulator requiring 14 V <
Use
< 20 V.
C H E C K : C > 1.667 mR
54. When driving a car into a left-hand turn
lane, one often sees a large circular or hexago
nal cut in the concrete. Embedded in these
cuts is a coil o f wire. When your car (contain
ing a large percentage o f iron) passes over this
coil, its inductance changes. This change o f
(b)
Figure P7.54
L(t)
Exponential Approximation
Ideally, the sawtooth voltage increases linearly with time until reaching a threshold where it imme
diately drops to zero, which reinitiates the process. The threshold voltage corresponds to a fixed
unit o f time. The linear voltage increase then acts as an electronic second hand, ticking o ff the
322
smaller units o f time. In practice, the linear increase in voltage is approximated by the linear part
o f an exponential response o f an RC circuit. ^X1^en the voltage across the capacitor reaches a cer
tain threshold, an electronic switch changes the equivalent circuit seen by the capacitor, allowing
the capacitor to discharge ver)^ quickly, i.e., the capacitor voltage drops to zero almost instanta
neously. Once the voltage nears zero, the electronic switch reinstates the earlier circuit structure,
causing the capacitor to charge up again. The process repeats itself indefinitely.
CHAPTER O BJECTIVES
1.
2.
4.
current) o f first-order RL and RC circuits without sources and with constant excitations.
Interpret the solution form o f the differential equation model in terms o f the circuit time
constant and the initial and final values o f the capacitor voltage or inductor current.
Develop techniques to handle s\vitching and piecewise constant excitations within first-
5.
3.
SECTIO N HEADIN GS
1.
2.
Introduction
Some Mathematical Preliminaries
3.
Problems
1. IN TRO D U CTIO N
Our study prior to Chapter 7 focused exclusively on resistive circuits. Recall that all nodal equa
tions and loop equations for resistive circuits lead to (algebraic) matrix equations whose solution
yields node voltages and loop currents, respectively. Chapter 7 then introduced the capacitor and
the inductor. Interconnections o f sources, resistors, capacitors, and inductors lead to new and fas
cinating circuit behaviors. How? Inductors and capacitors have differential or integral voltage-cur
rent relationships. Interconnecting resistors and capacitors or resistors and inductors leads to cir
cuits that must satisfy both algebraic (KVL, KCL, and Ohms law) and differential or integral rela
tionships for L and C values. When only one inductor or one capacitor is present along with resis
tors and sources, these relationships lead to first-order RL and RC circuits. When the sources are
dc, such circuits have vohages and currents o f the form A + Be~^ for constants A, B, and X. The
main purpose o f this chapter is to develop techniques for computing the exponential responses o f
first-order RC and RL circuits driven by dc sources. A simple example serves to explain some of
these points.
In the series RC circuit o f Figure 8.1, suppose an initial voltage
where 0~ designates the instant immediately before zero. Often vve distinguish among 0, 0, and
O'*' when switching occurs or when discontinuities o f excitation functions occur at r = 0.
v(t)
v,(t)
vp) = Ri(ir) +
Since iciO = C
(8.1)
at
+Vc(/)
dt
RC
RC
subject to the initial condition Vf^Qr). This equation says that the derivative o f the capacitor volt
age plus MRC times the capacitor voltage equals MRC times the source voltage. The equation
enforces constraints on the capacitor voltage, its derivative, and the source voltage, and is differ
ent from the algebraic node or loop equations studied earlier. The terminology first-order differ
ential equation applies because only the first derivative appears. Equation 8.2 is linear because it
comes from a linear circuit. Our goal is to find capacitor voltage waveforms that satisfy the con
straints imposed by the differential equation 8.2.
Exercise. For the circuit o f Figure 8.1, show that the capacitor current i({t) satisfies a differential
equation o f the form
cti.it)
dt
1 .
RC
1 dv,{t)
dt
32 4
Our scope in this ciiaptcr is limited to circuits containing one inductor or one capacitor equiv
alently, first-order RL or RC circuits. W ithin this category we further constrain our investigation
to circuits with no sources but nonzero initial conditions, circuits driven by constant (dc) sources,
circuits driven by piecewise constant sources, and circuits containing switches. First-order circuits
driven by arbitrary source excitations are covered in later chapters using the Laplace transform
method.
1
0
/ < ()
The unit step function is a universally used function and will appear many times in the remain
der o f this text. MATLAB code for specifying the step function is
function f = ustep(t)
t = t + le-12;
f = (sign(t)+l)*0.5;
With the unit step so defined, v^{t) - 2e~"'u(t) V, and both relations are plotted in Figure 8.2.
V.
32S
Further, if v<^t) = l e
er representation because the shifted unit step function, //(/- /q), means
Plots o f v^{t) = 2e~^u{t - ^q) and u{t - t^) are given in Figure 8.3 for
= 0.5.
Exercise. Plot //(/) and (/q t). Hint: For what values o f t are the functions zero and for what
values are they 1?
A working model ot a physical system underlies an engineers ability to methodically anai)'ze,
design, or modify its behavior. Linear circuits are physical systems that have differential equation
models. The RL and RC circuits investigated in this chapter have differential equation models o f
the form
dx{t)
dt
-Y(/(,) = .Vo
(8.3a)
or, equivalently.
dxU)
dt
(8.3b)
326
{a/^it) + ajfjU),
+ ajXjU))
( c ^ ' - " /( x U k
Jk)
(8.4)
This means that the expression on the right-hand side o f the equal sign (1) satisfies the differen
tial equations 8.3 [its derivative equals K times itself plus/r)], and (2) it satisfies the correct ini
tial condition, xit^) = x^^. A simple example illustrates this point.
E XA M PLE 8.1.
Compute and verify the solution o f equation 8.3a using equation 8.4.
SO L U T IO N
Suppose in equation 8.3a., J{t) = u{t - 1), a shifted unit step function, X = - 1 ,
10, in which case
cit
\)(k=
+ I
= 1, and at(1) =
32'
+ 1] does indeed satisfy the differential equation, observe that for ? > 1,
dx{t) _ d
dt
+ I = -,v(/) + 1
= -9 e
~~dt
Further, at ^ = 1, [9f
dx(t)
_>j
------ h e x i l ) = e
dt
f/,\
f{ t )
(8.5)
By the product rule for differentiation, the sum on the left equals
d_
dt
e""x{t)
dt
One can integrate both sides o f equation
f -dr
J'o
8 .6
from
Tq
to t as follows:
dr = e
(8.7)
JI q
Bringing the term e'^Ox{tQ) to the right-hand side o f equation 8.7 and multiplying through by
results in the solution to the differential equation 8.3a or 8.3b, given by equation 8.4. This
completes the derivation o f the very powerful formula o f equation 8.4.
There are four points to remember about the preceding discussion: (1) circuits have behaviors
modeled by differential equations such as equations 8.3; (2) the solution to a first-order differen
tial equation is a waveform (also called a signal or response) satisfying equation 8.4; (3) the for
mula o f equation 8.4 works for all continuous and piecewise continuous time functionsy(/); and
(4) a solution to a differential equation means that the waveform satisfies the given differential
equation with the proper initial condition.
328
Exercise. Show that the hinction .v(/) = (1 0. ^or r > 0, is a solution to the difFerential equation
^
dt
= -.v(/) + //(/) with initial condition a-(0) = 0 by showing that x{t) satisfies the difFerential
+
^
yf
v^(t)
R ^
ijt )
\ fi)
^R(t) <
-
S v jt )
(a)
(b)
FIGURK 8.4
Our first goal is to derive differential equation models for the RL and RC circuits o f Figures 8.4a
and 8.4b, respectively. We do this in parallel.
(1) At the top node o f Figure 8.4a, KCL implies
= v^t)
ifiU) = -iiit)
(2) However,
(2) However,
Vi it)
.d\'c{t)
n it ) R icit) RCdt
L d ii U)
equation model
equation model
dii it)
dV( {t) _
(8.8a)
dr
with
RC
v'c(/)
(8-81 )
^
dt
= hc{t) = - - m
(8.9)
i.e., the derivative of x(t) is a constant, X = -1/x, times itself. Applying equation 8.4 to equation
8.9 implies that both equations 8 .8a and 8 .8 b have solutions given by
where x is a special constant called the time constant o f the circuit. Equation 8.10 means that the
responses for
W
1^(0 =e
-^ '-'0 )
vc(0 = e
- ^ ' - 0)
(8.11)
vc(/q)
where the time constant o f the RL circuit is T = and the time constant o f the RC circuit is x
= RC.
The time constant of the circuit is the time it takes for the source-free circuit response to drop to
e~^ = 0.368 o f its initial value. Roughly speaking, the response value must drop to a little over onethird of its initial value. This is a good rule of thumb for approximate calculations involving decaying exponentials.
The mathematics that underlie the solution to the differential equation 8.9 given in equation 8.10
is nothing more than elementary calculus. To see this, consider the exponential solution form
(8. 12)
where K 'ls an arbitrary constant. The fiinaion
This is precisely what equation 8.9 requires. Therefore equation 8.12 satisfies the differ-
ential equation 8.9 and is said to be a solution. To completely specify x{t) it only remains to iden
tify the proper value of K from the initial condition. Evaluating x(r) at ^
yields
Mt) = /Cf-Vr
in which case
Substituting this value o f i n t o equation 8.12 produces the solution given in equation 8.10,
which is adapted to specific RL and R C circuits in equations 8 . 11. Figure 8.5 plots equation 8.12
for arbitrary K and x > 0 . This plot proves instructive for understanding how the response decays
as a function o f the time constant.
330
Time
decays to 0.368
Linear
\
>f
Resistive
Circuit
i,(t)
\
Linear
Resistive
Circuit
L
p
No
Sources
-N V jt)
No
Sources
i,(t)
+
v,(t)
331
These facts imply that the general formulas for computing the responses o f undriven RL and RC
circuits have the structures
(8.13)
VrU) = e
The difference between equations 8.11 and 8.13 is that in equations 8.13 R^f^ is the Thevenin
equivalent resistance seen by the inductor or capacitor.
t = 0.4sec
S
20Q
5Q
+
v,(t)
8H
o l u t io n
Step 1. With switch S open, compute the response for 0 < /< 0.4 sec. From the continuity property
o f the inductor current, /^(O'*) = ^^(0) = 10 A. Using equation 8.13,
'1 ( 0 = c
allel 20 Q and 5 ^ combination. According to equation 8.13, the response for t>tQ = 0.4 sec is
,- ,( 0 = /
Step 3. Write the complete response as a single expression using step fitnctions:
0.4)A
(8.14)
332
Step 4. Plot the complete response. To plot this using MATLAB, we use the following m-file along
with the code given earlier for the unit step function:
t = 0:0.005:1.4;
iL= 10*exp(-2.5*t) .* (ustep(t).* ustep(t - 0.4)) + 3.679*exp(-0.5*(t-0.4)) .* ustep(t - 0.4);
plot(t,iL)
grid
Using this code, Figure 8.8 illustrates the complete response, showing the two different time con
stants. The 0.4 sec time constant has a much faster rate o f decay than the lengthy 2 sec time con
stant.
a
E
<
Time (seconds)
FIGURE 8.8 Sketch of response i^{t) for Example 8.2.
.IL l
yi(f) = e
(0 .4 "') =
Step 6. Compute the energy dissipated in the 5 O. resistor over the interval [0.4, oo). The power
absorbed by the 5
= 43.31
PfaU/)A/ = 4 3 .3 12 j;^ ^
= 4 3 .3 12 J
Exercises. I. Plot v^it) using the above m-file, ustep, and the appropriate .code.
2. Repeat the calculations o f Example 8.2 with the 8 H inductor changed to 8 mH and a switch
closing time o f 0.4 ms.
A N SW ER: ij{t) = l{)r-^'^"'//(f)//(0.4 x lO"-^ - r) +
- 0.4 x Ur-^) A
E X A M PLE 8.3
Find V(^t) for r > 0 for the circuit o f Figure 8.9 given that y^^^O) = 9 V.
So l u t io n
Because there is a switch that changes position at r = 1 sec, there are two time intervals to consider.
Step 1. Compute the response forO < t < 1. Over this time interv\il, the equivalent circuit is a par
allel /?C circuit, as shown in Figure 8.10a.
334
0.1 F
0.1 F
80
3Q
V jt)
+
V jt)
(a)
(b)
FIG URE 8.10 Equivalent circuits for Figure 8.9: (a) 0 :s r < 1 and (b) 1 s t.
------- 1
v c(t)= e
Vc(O^) =9^- 5^V
Step 2. Compute the response for r > 1. Figure 8.10b depicts the pertinent equivalent circuit.
Observe that
= 2.58 V and
=3
-St-to)
V c (0 = f
= 1,
v .c (^ ) = 2 .5 8 e
Step 3. Use step functions to specify the complete response. By using the shifted unit step function,
the two expressions obtained previously can be combined into a single expression:
t-\
V cir)=9e~^-'^'[u{t)-u{f-\)] + 2.5Se
- \) V
Step 4. Obtain a plot o f the response. Using MATLAB and code similar to that used in Example
8.2, the plot in Figure 8 . 1 1 w'as obtained. Here the part o f the response with the 0.3 sec time con
stant shows a greater rate o f decay than the longer 0.8 sec time constant.
Time (seconds)
F IG U R E 8 .1 1 Response,
- t)-
2. Suppose that in Example 8.3 the switch moves to the 4.5 ^ resistor at r = 0.5 sec instead oF 1
sec. Compute the vakie V(\t) at f = 1.2 sec.
ANSW ER: 0.4671 V
For all o f these examples x > 0 and the response is a decaying exponential. Intuitively, the response
decays because the resistor dissipates as heat the energy initially stored in the inductor or capaci
tor. One o f the homework exercises will ask the student to show that the total energ)' dissipated
in the resistor from
negative resistance supplies energ)' to the circuit and the source-free response will grow exponen
tially. This is illustrated in the next example.
E X A M PLE 8 .4
Find Vf^t) for the circuit o f Figure 8.12, assuming that^^^^ = 0.75 S and i^(;(0) = 10 V.
------ o -----q V
0.25F
4Q
+
v,(t)
<
0 -----0.25F
-2 Q
v,(t)
-o-
-o-
So l u t io n
It is straightforward to show that theThevenin equivalent seen by the capacitor is a negative resist
ance,
vc{0^)=\0e^'u{t) V
Because o f the negative resistance, this response grows exponentially, as shown in Figure 8.13. A
circuit having a response that increases without bound is said to be unstable. Practically speaking,
an unstable circuit will destroy itself or exhibit a nonlinear phenomenon that clamps the voltage
at a finite value, as in the case o f saturation in an op amp.
336
Tim e (seconds)
FIGURE 8.13 Plot of unbounded voltage response due to presence of negative resistance.
Circuits with such responses arc said to be unstable.
ic{0-^) = 5e-'ii{t)A
Linear
Resistive
Circuit
with
i^Ct)
Linear
Resistive
Circuit
with
Constant
Sources
Constant
Sources
(a)
(b)
\{t)
FIGURE 8.14 (a) Driven first-order RL circuit, (b) Driven first-order RC circuit, (c) Thevenin
equivalent representation oF (a), (d) Thevenin equivalent representation of (b).
Given these basic circuit representations and initial conditions at Tq, what is the structure o f a dif
ferential equation mode! that governs their voltage and current behavior for t>tQ' The first objec
tive is to derive the differential equation models characterizing each circuits voltage and current
responses. It is convenient to use ij{t) as the desired response for constructing the differential
equation for the series RL circuit (Figure 8 .l4 c ), whereas for the series RC circuit (Figure 8 .l4 d ),
is the more convenient variable.
(1) The circuit mode! for the inductor is
vl O )= L
dv(^(t)
dilit)
ic(t)= C
dt
dt
ic (t) -
d iijt)
dt
R.
dvc(r)
+
(8-> 5a)
^oc R,ill
dt
_
-^ ^ 'C (0 + -
v o c (8 .1 5 b )
338
since
Exercise. Construct differential equation models for the parallel RL and RC circuits o f Figure
8 . 15 . Note that these circuits are Norton equivalents o f those in Figure 8.14a and Figure 8.14b.
Again choose /^(r) as the response for the RL circuit and v^^t) as the response for the RC circuit.
(constant)
(constant)
dll (/)
.\NSWI-RS: V di
V/(
I.
//(/)+
L
A simple application o f basic circuit principles has led to the two differential equation models of
equations 8.15. The next important question is: What do these t%vo differential equation models
tell us about the behavior o f each circuit? Equivalently, how do we find a solution to the equa
tions? Observe that both differential equations 8.15 have the same structure:'
dx(t)
dt
(8.16a)
circuits and F = vJiR^i^j for the RC case. This equation is valid for t>
here with/^y) = F, presents the general formula for solving the differential equation 8.16a:
) + ( e^^^'-f^Fdcj
A-(/)=
where A =
(8.16b)
'h)
is a natural frequenc)' o f the circuit, and where we have emphasized the use o f
initial condition is continuous, i.e., x{t^p = x(fQ+), because F h a . constant (non-impulsive) forc
ing function. A straightforward evaluation o f the integral o f equation 8.16b yields
t
x(/) = e
^ \v(fo)+ F e
( I-In
f e
Jif)
dq = e
^ 'x (t^ )+ F T
- e
(8.16c)
339
(t-tn
x(
0=
f t
+ U 4 )-
ft
(8.17)
= -r
which is valid for t > /q. After some interpretation, this formula will serve as a basis for comput
ing the response to RL and RC circuits driven by constant sources. A homework exercise will ask
for a different and direct derivation of this formula.
At this point it is helpful to interpret the quantity i r in equation 8.17. For RL circuits, when x(r)
= /*(/), equation 8.15a implies that
For RC cir
cuits when ;c(/) = V({t), equation 8.15b implies that F = vJR ^ C , x = Rf^,C, and hence Fz =
This interpretation is valid for both positive and negative values of x. If x > 0, then
t-tQ y
jr(oo) = lim x(t) = lim
/*00
t~ * 0 0
Fx + \ x {t^ )-F x y
isc =
= Ft =
for RL case
Rih
(8.18)
for R C case
O f course,
ferential equation. Since the constant x(/) = F i satisfies the differential equation 8.16, / r is an
equilibrium state of the differential equation 8.16.
Whenever X > 0, equation 8.18 implies that the formula (equation 8.17) for the solution o f equa
tion 8.1 6 given constant or dc excitation becomes
.
a: ( 0
_ ^-'o
(8.19a)
= j :(>) + U ( ^ ^ ) - J c( oo) e
iL (0 = / z .( ) + k ( 4 ) - ' L ( )
o
(8.19b)
t-t
K/.C
(8.19c)
Note that x > 0 is true whenever R^j^ > 0, C > 0, and Z, > 0, i.e., the circuit is said to be passive.
This allows us to state a nice physical interpretation of equation 8.19a:
elapsed time
w
w
's . ;
x (/ q ).
340
Elapsed Time
I'iG U RE 8.16 Graphical interpretation o f equation 8.19a for the case xico) > x{tQ).
Exercise. Redo the curve o f Figure 8.16 for the case x{<x)) <
The initial value
computed from initial conditions and possibly the value o f the source
v^(t)
1^(1-)=1 A
.Vtl
S o l u t io n
Step 1. Determine the circuits differential equation model. Since the circuit o f Figure 8.17 is a driv
en series RL circuit, equation 8 .1 5a implies that the differential equation model o f the circuit valid
for r > 1 is
cl'iAt)
10
/^(/)=/^(oo)+^/^(r)
Here the presence o f u{t - 1) emphasizes that the response is valid only for / > 1.
Step 3. Compute i^i^X)) and set forth the fin al expression for /^(r). Since x = 0.4 > 0, we replace the
inductor in the circuit o f Figure 8.17 with a short circuit to compute i^^. = //(oo) = 2 A. It follows
that
[2 + (1 -
- 1) = (2 -
- 1) A
Step 4 . Plot i^it). One cannot presume that the response is zero for r < 1. Hence, using MATLAB
or the equivalent, one can construct the graph o f i^{t) for / > 1 as given in Figure 8.18.
c
0;
u
O
tj
3
o
c.w
Time (sec)
FIGURE 8.18 Plot of /^(r) for Example 8.5.
Step 5. Compute v^it). Given the expression for the inductor current in step 3, it follows that for
t> 1,
cliLit)
Vi{t) = Ldt
/-I
u{t -\^ ) = 5 e
* ^ / (r -l'* ')V
342
Exercises. 1. Verify that in Example 8.5 v^it) can be obtained without differentiation by
Note that we have used the differential equation 8.16 (or equations 8.15) to obtain the solution
form o f equation 8.19. However, when using equations 8.19, it is not necessary to reconstruct the
differential equation of the RL or RC circuit. Specifically, we need only compute xit^), x{cc), and
the time constant x = LIR^i^ or
The method described for computing final values can also be used to find the initial values o f
and i^ at f =
if dc excitations have been applied to the circuit for a long time before t = ?q. The
next example illustrates this technique and extends the preceding discussion.
EXA M PLE 8.6
The source in the circuit o f Figure 8.19 furnishes a 12 V excitation for / < 0 and a 24 V excita
tion for r < 0, denoted by yy^(r) = [\2u{-t) + lAu{t)] V. The switch in the circuit closes at / = 10
sec. First determine the value o f the capacitor voltage at r = 0", which by continuity equals
Next determine
+
v,(t)
So l u t io n
Step 1. Compute initial capacitor voltage. For r < 0, the 12 V excitation has been applied for a long
time. Therefore, at r = 0 , the capacitor has reached its final value and looks like an open circuit
to the source. Hence the entire source voltage o f 12 V appears across the capacitor at r = 0~, i.e.,
^(40~) =
Step 2. Use equation 8.19c to obtain v^it) fb r Q < t< 10 sec. Equation 8 .1 9c requires only that we
know V(iQ*) (step 1), x, and Vf^co). For 0 < t< 10 sec, r = R^C= 3 sec. It is important to realize
here that for 0 < r < 10 sec the circuit behaves as if the switch were not present. Hence, the com
putation o f v^(co) proceeds as if no switching would take place at r = 10 sec. Here v^^ = v^^oo) =
24 V. Hence, for 0 < ^ < 10, equation 8.19c implies
.vi3
= 2 4 + (1 2 -2 4 )c^ - ^ = 2 4 - 1 2 ^
(8.20)
8.20 and using the continuity property o f the capacitor voltage yields
= 2 4 - 12f>-io/3 = 2 3 .57 V
Step 4. Find V(^t) for / > 10. For r > 10, the resistive part o f the circuit can be replaced by its
Thevenin equivalent, which yields Figure 8.20.
Rtn = 2 k O
+
V(t)
FIGURE 8.20 Circuit equivalent to that of Figure 8.19 for / > 10.
Here, equation 8.19c applies again. The value for y^^co), however, is now 8 V and the new time
constant is
Step 5. Set forth the complete response using step functions. Using step functions, the response V(\()
for f > 0 is
v ^ {t)= 2 A -\ le
+ S + \5.51e^'
Step 6. Plot V(it). Plotting i^(^t) yields the graph of Figure 8.21.
3-m
o
OJ
O)
Q.
Time (seconds)
for r > 0 .
Exercise. Suppose the switch in Example 8.6 opens again at r = 20 sec. Find v^it) at r = 25 sec.
ANSW ER: 20.98 \
EXA M PLE 8 .7 The circuit o f Figure 8.22a has a capacitor voltage given by the cur\'e in Figure
8.22b. We note that ^f^O.l) = 7.057 V. Find, Kq, y^^O), the time constant r = RC, the exact value
of j/(j(Q.25), and the value o f C i f /? = 100^2.
------ O-
v.(t)= V u (t)
v,(t)
-o(a)
345
01
IB
I
Q.
fO
U
T im e (s)
(b)
lu t io n
) - \(-(0C) e ^
Hencc as / -> oo, v^{t) -> v^co) = 10 V. Since the capacitor looks like an open circuit at / = oo, Vq
=
Simplifying yields
-
T =
0.1
/ 1 0 - 7 .0 5 7
= 0.
8
Therefore C = 0.01 F and M 0 .2 5 ) = 9..34.33 V.
When switching occurs frequently, or the excitation changes its constant level frequently, then
hand analysis, as in lixample 8.6, becomes very tedious. For such problems a SPIC E simulation
(or the equivalent) proves useful and saves time. The next example u.ses SPIC E to compute the
waveform o f a simple RC circuit whose input excitation is a square wave. Like the previous exam
ple, the solution is broken down into time intervals such that during each time interval inputs are
constant. Because no switching occurs, the time constants for all time intervals are the same. In
applying equation 8.19c, the quantities that vary from one time interval to the next are the initial
values and final values.
346
Plot
Find the initial value and the final value in equation 8.19 when t is very large, for exam
ple, at the beginning and end o f the interval 1 < t< 1.01 sec. Plot the v^t) wave for this
interval using MATLAB or the equivalent.
V (t) (V)
> t (msec)
10
20
30
(b)
FIG U RE 8.23 (a) Series RC circuit excited by the 50 Hz square wave o f (b).
So
lu t io n
Part (a)
Doing a SPICE or equivalent simulation gives rise to the response curve shown in Figure 8.24, over
which the square wave input is superimposed. Observe that the response V(\t) has an approximate tri
angular shape. What is happening is that from zero to 10 msec, the circuit sees a step and hence the
capacitor voltage rises toward one volt. At 10 n:isec, the square goes to zero for 10 msec. The capacitor
then discharges its stored energ)' through the resistor, causing a decrease in its voltage value. The decrease
does not go to zero, however. So when the square wave again is at 1 volt the capacitor voltage begins to
rise again and achieves a slighdy higher value at f = 30 msec compared to f = 10 msec. In fact, one notices
in Figure 8.24 that the peak and minimum values are increasing slighdy as time increases. Eventually the
peak and minimum values will reach their respective fixed viilues, c;illed steady-state values. To find these
values, a simulation program could require a very lengdiy simulation interv'al, which often proves
impraaical. The steady-state values can be computed analytically as in part (b).
10
20
30
40
50
t(m se c)
FIG URE 8.24 Response of circuit of Figure 8.23a calculated using SPICE. For reference, the input
square wave excitation is superimposed on the plot.
_____________________________________________________
Part (b)
Let Tq = mT, where 7'= 20 msec is the period o f the square wave and m is some large integer. Then,
0.57-
+ 0.571 = 1 + {v^t^) - 1) /
Further, in steady state, V(it^ + T) =
= 1+
- 1]^-
(8.21)
0.5T
+ 71 =
+ 0 .5 7 ) ^ ' KC .
+ 0.571^'-' =
(8.22a)
(8.22b)
^ "
1+e
= 0-2689 V
It follows that
r/J/o + 0 . 5 7 ) = i^(3<ro)e> = 0 .7 3 1 1 V
An examination o f the response in Figure 8.24 shows that the minimum and peak values are
approaching the steady-state values o f 0.26 8 9 V and 0.7311 V, respectively.
Exercise. Based on the response in Figure 8.24, roughly sketch the capacitor current,
At
r\
348
which is also a linear equation. To see linearity in this i-v relationship, suppose voltages
and
Let
_^ dvc2
, >c2 - c
a - c
and
i.e.,
C3 = (^-^(^Cl+^C2)
However, the linearity o f the derivative implies the property of superposition;
.
ic=C-
_ dvf
^
+ C = ic i + t 2
dt
dt
By the same arguments, the current due to the input excitation t/Q - ^\^c\
+
On the other hand, suppose two separate currents / q and
^2^Cl
^C3 ^l^Cl
^Ci (0 =
(<:) d t , vc2 (0 =
ic2 (T^) ^
+ <?2^C2 would
be a voltage,
vc3 (0 = ^ f_ Ja \ ic \ (T ) + a2ic2('^)] d r
= ai
\ 7 S ' - J c 2 W ^
r-
efFect o f the initial condition can be viewed as the effect o f some input that turns o ff at the time
the initial condition is specified. This means that when using superpositioti on a circuit, one first
looks at the effect o f each independent source on a circuit having no initial conditions. Then one
sets all independent sources to zero and computes the response due to each initial condition with
all other initial conditions set to zero. The sum o f all the responses to each o f the independent
sources plus the individual initial condition responses yields the complete circuit response, by the
principle o f superposition. A rigorous justification o f this principle is given in a later chapter using
the Laplace transform method.
The following example illustrates the application o f these ideas.
E XA M PLE 8.9
The linear circuit o f Figure 8.25 has two source excitations applied at r = 0, as indicated by the
presence o f the step functions. The initial condition on the inductor current is
= -1 A.
So
lu t io n
Because the circuit is linear, having a linear differential equation, superposition is but one o f sev
eral methods for obtaining the solution. An alternative approach is to find the Thevenin equiva
lent circuit seen by the inductor. As we will see, the superposition approach sometimes has an
advantage over the Thevenin approach.
Superposition must be carefully applied, however. First one computes the response due only to the
initial condition with the sources set to zero. Second, one computes the response due to Vj with
all initial conditions and all other sources set to zero. Third, one computes the response due to /j
with all initial conditions and all other sources set to zero. Finally, one adds these three responses
together to obtain the complete circuit response.
Step 1. Compute the part o f the circuit response due only to the initial condition, with all independ
ent sources set to zero. With both sources set to zero, there results the equivalent circuit given by
Figure 8.26. The Thevenin equivalent resistance is
= 4 Q, resulting from the parallel combi
nation o f
and R->. Figure 8.26 depicts the equivalent undriven RL circuit having response
352
An approach based on the Thevcnin equivalent circuit seen by the inductor would allow one to
quickly compute the complete response, but not in a way that identifies the contributions due to
each o f the individual sources. Answ^ers to the preceding three questions would have required
repeated solutions to the circuit equations. However, if one keeps the source values in literal form,
then the Thevenin equivalent approach would be as efficient.
Under these conditions it is possible to define the natural and forced responses as follows: (1) the
natural response is that portion of the complete response that has the same exponents as the zeroinput response, and (2) the forced response is that portion of the complete response that has the
same exponent as the input excitation provided the input excitation has exponents different
from that of the zero-input response.
This decomposition is important for rwo reasons. First, it agrees with the classical method o f solv
ing linear ordinary differential equations with constant coefficients where the natural response cor
responds to the com plem entary function and the forced response correspontls to the particular
integral. Students fresh from a course in linear differential equations will feel quite at home with
these concepts. The second reason is that the forced response is easily calculated for dc inputs. For
general systems this type o f decomposition is not used.
{I- to)
x{l) = X^,+
~ r~
(8.23)
For T negative or positive equation 8.23 is identical to equation 8.17 with Pi = X^. Further, T is
the circuit time constant and X is that voltage or current of interest computed under the condi-
3 S3
tion that the inductor is replaced by a short circuit for the RL case or the capacitor is replaced by
an open circuit for the RC case.
How do we justify the form o f equation 8.23 for all variables? We invoke the linearity theorem o f
Chapter 5 and the source substitution theorem o f Chapter 6 o f 2"^ edition. Suppose in a firstorder RC circuit we have found V(\t). By the source substitution theorem, the capacitor can then
be replaced by a voltage source whose voltage is the computed V(it). This new circuit consists o f
constant independent sources, one independent source o f value
ent sources. By linearity, any voltage or current in the circuit has the form
x(r) =
for appropriate ATj and Kj- By equation 8.17,
Mr) = K , + ^6^
for appropriate
"
and
-X^.
^rid
Note that equation 8.23 requires that the initial value be evaluated at ^
instead o f r =
This
is because only the inductor currents and the capacitor voltages are guaranteed to be continuous
from one instant to the next for constant input excitations. The capacitor current and the induc
tor voltage as well as other circuit voltages and currents may not behave continuously.
E XA M PLE 8 .1 0
This example illustrates the application o f equation 8.23. For the circuit o f Figure 8.29,
-18//(-r) +
6kQ
3kQ
2kQ
v jt )
0.5 mP
354
SO L U T IO N
Step 1. Compute /y(0^). To obtain /y(0*), we first compute V(\Q~) =
V has excited the circuit for a long time, the capacitor looks like an open circuit. By voltage divi
sion,
j = ( - 1 8 ) = - 6 V
Thus at r = 0*, the equivalent circuit is as shown in Figure 8.30.
= 1.75 mA
Step 2. Find the circuit time constant and the equilibrium value o f ijj,t). From Figure 8.29, the
equivalent resistance seen by the capacitor is
Further, since t><d,X^ = ij{'^), which is computed when the capacitor is replaced by an open cir
cuit. In this case.
9 x 1 0 -'
= 1 mA
Step 3. Apply equation 8.23. Using equation 8.23, we have, for r > 0,
= 1 + 0.75^'
-0.5/
mA
Exercise. In Flxample 8.10, find /^O ), /(^O"^), and i^^t) for r > 0 using equation 8.23 directly.
AN SW ERS: 0, 2.25 mA. 2.25^-** "^ mA
straightforward arithmetic shows that /y(0) = - 2 mA. Since /y(0'^) = 1.75 niA, the input current
is discontinuous at f = 0, unlike the capacitor voltage, which is continuous at f = 0. This empha
sizes the need to use x{t^) in equation 8 .23.
3S5
In several o f the examples o f sections 3 and 4, the circuits contain switches that operate at pre
scribed time instants. In some electronic circuits, the switch is a semiconductor device whose
on/off state is determined by the value o f a controlling voltage somewhere else in the circuit. If the
controlling voltage is below a certain level, the electronic switch is off; if the voltage moves above
a fixed level, the electronic switch is on. The time it takes for a controlling voltage to rise (or fall)
from one level to another is very important because timing is as critical in electronic circuits as is
scheduling for large organizations. For first-order linear networks with constant excitations, cal
culation o f the time for a voltage or current to rise (or fall) from one level to another is straight
forward because all waveforms are exponential functions, as per equation 8.19. The situation is
illustrated in Figure 8.31.
c
0)
k_
I
D
u
<U
cn
(Z
4->
o
>
FIGURE 8.31 First-order response showing a rise from the voltage or current level A'j to the voltage
or current level A'2, for which the elapsed time is h - ty
In equation 8.19a, let .v(rj) =
and .v(^^) = X , be the two levels o f interest. A straightforward
manipulation o f equation 8.19a leads to the elapsed tim e formula for first-order circuits.
h -ri
[^1 --V(oo)
(8.24)
^2 - -v(x)
E X A M PLE 8.11
This example uses the elapsed time formula o f equation 8,24 for the circuit o f Figure 8.32. The
switch in this circuit is used to produce r^vo different final capacitor voltages. When the switch
is open, the final capacitor voltage is 12 V. When the switch closes, the final capacitor voltage, by
V-division, changes to 4 V. Thus the switch causes the capacitor to charge and discharge repeat
edly. For our purposes we show that the choice o f resistances produces an approximate triangular
waveform.
For the purposes o f this example, suppose the switch in Figure 8.32 is controlled electronically so
that it closes when
SNvitchings.
rises to 9 V and opens when V(^ falls to 5 V. Find and plot v^^t) for several
FIGURE 8.32 Switched driven RC circuit used to generate an approximate triangular waveform.
So
lu t io n
= 12 V. From equa
tion 8 .1 9c,
t
VcU) = 12 1 - e 3
- 0 = 3//?
Now, for
19-12^
= 3 X 1.386 = 4 .1 5 9 s
= 9, and
= 4. .^gain using
equation 8.19c,
- t =/n
Finally, for the time interval
- /n{5) = 1.61 s
\5-4/
, x= 3 sec,
JJz Itl
v c(/) = 1 2 - 7 e
- th =3/n
rh e w aveform o f
v^t)
9-12/
^ / ; (1\
\3
= 2.54 s
is plotted in Figure 8 .3 3 .
= 4.16 sec,
evident Irom Figure 8.33. Practically speaking, the first c)'cie o f a periodic, approximately trian
gular waveform occurs in the time interv^al [t^, r j , and the period is
period
2.5 4 + 1.61
4. 15
= 0.241 Hz
The waveform in figure 8.33 is approximately triangular. This is due the fact that two time con
stants, 1 and 3 s, have the same order o f magnitude. If we select the resistances so that the charg
ing time constant is much larger than the discharging time constant, then the capacitance voltage
waveform will look more like a sawtooth waveform. Sawtooth waveforms arc used to drive the
horizontal sweep of the electronic beam in an oscilloscope or a T V picture rube.
3^H
EXA M PLE 8 .1 2
Compute the response
1.
FIGURE 8.34 Differentiating op amp circuit.
So
lu t io n
Observe that
Chapter 4. Also,
capacitor,
= ,,( ,) = - / c (0 = - R C ^
=- R C ^
(8.25)
Since the output is a negative constant (user chosen) times the derivative o f the input, the circuit
is called a differentiator.
Exercise. Suppose y/(^) = cos(250^). Find R for the circuit o f Figure 8.34 so that
= sin(250/)
V a n d C = 1 pF.
AN SW FR: 4 kH
EXA M PLE 8 .1 3
Compute the response
= 0.
1.
So
lu t io n
= v-^{t)IR by the virtual short circuit property o f the ideal op amp. Also, i^^t)
Observe rhat
3S9
Hence, from these equalities and the integral i>-i relationship o f a capacitor,
(8.26)
Since the output is a negative constant (user chosen) times the integral o f the input, the circuit is
called an integrator.
Exercise. Suppose
= cos(250r) V. Find R for the circuit o f Figure 8.35 so that
sin(250r) V and C = 1 pF.
A N SW ER: 4 kD
EXA M PLE 8 .1 4
This example considers the so-called leaky integrator circuit o f Figure 8.36, which contains an
ideal op amp. The input for all time is v^{t) =
capacitor. Given C and Rj, the resistance /?| is chosen to achieve a dc gain o f 10. The objective is
to compute the response
assuming
v jt )
1
FIG U RE 8.36 Leaky integrator op amp circuit in which v^{t) = -5u{t) V.
So
lu t io n
Because there is only one capacitor, the circuit o f Figure 8.36 is a first-order linear circuit. Because
the inverting terminal o f the op amp is at virtual ground,
an equivalent resistance R^i^ = /?2- Hence x = /?,C = 10 sec > 0, Equation 8.19 implies that
360
(8.27)
Because the voltage source v^{t) = 0 for r < 0,
the source voltage is constant, the capacitor looks like an open circuit at r = co having final value
.. ( co) = - A ( . 5 ) = 5() V
Entering numbers into equation 8.27 yields
= 50 + (0
= 5 0 (l
ii{t) V
A plot o f the op amp output voltage appears in Figure 8.37 along with that o f an ideal integrator.
O ne observes that the more realistic leaky integrator circuit approximates an ideal integrator only
for 0 < r < 0 .1 5 t before the error induced by the feedback resistor R-^ becomes noticeable. Such
integrators need to be reinitialized periodically by resetting the capacitor voltage to zero.
OJ
Ol
ra
>
4-'
D
CL
4-1
D
Q.
E
<
Q.
O
Time (s)
361
Leaky lntegrator-Transient-0
Time (s)
Observe rhar the response approximaces the ideal up to about 0.1 5t = 1.5 sec, which corroborates
our analysis using the ideal op amp. Note, however, that the simulation accounts for saturation
present in practical op amps but absent from the ideal.
possibly exist.
4kO
Capacitance
(a)
(b)
F IG U R E 8.39 (a) Incorrectly connectcd Inverting amplifier, (b) Circuit model, including a stray
362
So l u t io n
C'hapter 4 op amp models contain only resistors and controlled sources. One way to explain the
situation described in this example is to postulate a small stray capacitance, C = 1 pF, across the
input terminals. In fact, this is a more accurate circuit model and is shown in Figure 8.38b. This
means that the response will be o f the form o f equation 8.23.
The first quantit)' to compute is the circuit time constant t =
'
4000
Hence,
= 1 =
/,
1 -1 0 ^
resistance seen by C is
R^,, = 1000
= lOOOll ( - 2.5) = - 2 .5 0 6 Q.
- 2 .5 0 6
The negative time constant spells instabilit)'. The complete response may be written directly with
the use o f equation 8.23, where x{t) =
= lO-le V
and
363
Kg =
with C open-circuited
/
-2 5
\ j
= 1 X ------ 10' = -2 5 .0 6 3 V
V
1000-2.5/
= -2 5 .0 6 3 + (lO'^E + 2 5 .0 6 3 )e
0.339xl0^r
(8.28)
For any small positive initial capacitor voltage E, equation 8.28 implies that the output would
increase exponentially, had the op amp been ideal. Because this particular real op amp saturates at
15 V, the output more or less instantaneously saturates at 15 V, the phenomenon observed by the
student. Had the initial capacitance voltage been sufficiently negative,
-2 5 .0 6 3
< ------- j ---- V,
10-^
equation 8.28 implies that v
would saturate at - 1 5 V.
9. SUMMARY
This chapter has explored the behavior o f first-order RL and RC circuits (1) without sources for
given ICs, (2) for constant excitations (dc), (3) for piccewise constant excitations, and (4) with
switching under constant excitations. In general, first-order RL and RC circuits have only one
capacitor or one inductor present, although there are special conditions when more than one
inductor or capacitor can be present. Our discussion has presumed only one capacitor or one
inductor is present in the circuit.
Using a first-order constant-coefficient linear differential equation model o f the circuit, the chap
ter sets forth rwo t}'pes o f exponential responses, the source-free response and the response when
constant independent sources are present. The source-free responses for the RL and RC circuits
have the exponential forms
ii(0= e
where
iiito)
vcit)=e
seen by the capacitor. For an RL circuit, the time constant T= L/R^i^, where R^^^now is theThevenin
equivalent resistance seen by the inductor.
When independent sources are present in the circuit, the response o f a first-order RC or RL cir
cuit has the general form
I-10
x{t) = ,v(3c) + [,y(/o ) -
lie-
36-i
<^<>nsvdn{
the interpretation as discussed in section 7, with equation 8.23 identifying the form o f any volt
age or current in the circuit:
The time constants of- a circuit can be changed by switching within the circuit. By changing time
constants in a circuit, one can generate different t)'pes o f waveforms such as the triangular wave
form o f Figure 8.32. As mentioned at the beginning o f the chapter, wave shaping is an important
application o f circuit design. When inductors, resistors, and capacitors are present in the same cir
cuit, many other wave shapes can be generated. RLC circuits arc the topic o f the next chapter and
allow even greater freedom in waveform construction.
As a final application o f the concepts o f this chapter, we looked at the leaky integrator op amp cir
cuit. Integrators are present in a host o f signal processing and control applications. Unfortunately,
ideal integrators do not exist in practice. The leaky integrator circuit o f Figure 8.35 provides a rea
sonable model o f an ideal integrator.
365
Natural iirequency of a circuit: natural mode of oscillation of the circuit. For a first-order circuit having a response proportional to
it is the coefficient X in the exponent.
Natural response: that portion of the complete response that has the same exponents as the zeroinput response.
Passive RLC circuit: circuit consisting of resistors, inductors, and capacitors that can only store
and/or dissipate energy.
Sawtooth waveform: triangular waveform resembling the teeth on a saw blade and typically used
to drive the horizontal sweep of the electronic beam in an oscilloscope or a T V picture
tube.
Source-free response: response of a circuit in which sources are either absent or set to zero.
Step response: response, for ^> 0, of a relaxed single-input circuit to a unit step, i.e., a constant
excitation of unit amplitude.
Stray capacitance: small capacitance always present between a conductor and ground. It usually
can be ignored, but as Example 8.14 shows, it can critically affect the response of a cir
cuit.
Superposition: in linear RC and RL circuits, the complete response is the superposition of the
relaxed circuit responses due to each source with all other sources set to zero, plus the
responses to each initial condition when all other initial conditions are set to zero and all
independent sources are set to zero.
Time constant: in a source-free first-order circuit, the time it takes for the circuit response to drop
to e~^ = 0.368 of its initial value. Roughly speaking, the response value must drop to a
litde over one-third of its initial value or rise to within one-third of its final value. For RL
circuits x = LIR^f^ and for RC circuits x =R^/jC.
Unit step (unction: function denoted (/) whose value is 1 for f > 0 and 0 for ^< 0.
Unstable response: response whose magnitude increases without bound as t increases. The time
constant for first-order circuits is negative for an unstable response.
Zero-input response: response in which all sources are set to zero.
Zero-state response: response to a specified input signal or set of input signals given that the initial conditions are all set to zero.
w
w
^ It happens that all variables in a first-order RL or R C circuit satisfy a differential equation o f the same form. The
interpretation o f the solution is somewhat different fi-om what follows. A detailed explanation o f the general solu
tion is presented in section 7.
366
Problems
i,(t)
Plot your
with
Figure P8.2
3 . In Figure P8.1, suppose R = 25 kH and
v^O) = 20 V.
(a) Find C so that v^O.25) = 2.7 0 6 7 V.
(b) Given your answer to part (a), find
C H EC K ; C = 5
pF
2.7 0 6 7 mA.
(b) Given your answer to part (a), find
C H E C K ; Z.= 1.25 H
5. The response o f an undriven parallel RC cir
cuit is plotted in Figure P8.5. Find the time
ic(t)
. v,(t)
Figure P8.1
2. Consider the RL circuit o f Figure P8.2 in
which R = 50 Q and Z, = 0.1 mH.
(a) If the energy stored in the inductor at t
= 0 is 2 [ij, find /^(O) and
Plot your
Figure P8.5
36"
> 0. Plot
(X
constant is positive.
Figure P8.6
R.
7. In the circuit o f Figure P8.7, suppose /?j = 5
= 20 V,
kQ, /?2 = 20 k n , C = 50 pF H,
and the switch opens at r = 0.4 sec.
R,
0.25 mF
Figure P8.9
C H E C K : (c) a > - 3
10. In the circuit o f Figure P8.10, /?, = 100 Q,
= 20
P = 2 0 0 ,1 = 0.5 H, and /'^(O) = 250
i{t)
,Vc(t)
for r > 0.
Figure P8.7
Consider the circuit o f Figure P8.8.
(a) Find the value o f
R,
ijt)
Figure P8.10
C H EC K :
= -2 5 n
ijt )
1 kf)
80 mH
(a) With
CX
Figure P8.8
= 500 mV.
368
t=4RC
i,(t)
R.
1=0
.4R
\4R
R,
Figure P 8 .1 1
Figure P8.14
has
= V^ii{-t) V. Find
and
V(^t) for ^ > 0 in terms o f
R, and C, where
R is in Q and C in F. Sketch v^^t) for 0 < f < 3x,
given by
.4R
Figure P8.15
Figure P8.12
13. In Figure P8.13, the current excitation is
given by
4R
and
and
RESPONSE OF DRIVEN
CIRCUITS
18. Consider the RC circuit o f Figure P8.18 in
which R = 10 k n and C = 0.4 niF.
Figure P8.13
(a) If
is in V,
= 0 and
V, find
= 2i)ti(t) V,
369
where
= -1 0 V
and V^2 = 20 V.
(a) Find ^(;^0'*^) and V(it) for / > 0.
(b) Sketch v^^t) for 0 < f < 5x, where x is the
circuit time constant for r > 0.
(c) Identify the zero-input response (f > 0)
and the zero-state response (r > 0) for the
yoj
'.w O
Figure P8.18
19. Consider the RL circuit o f Figure P 8.19.
Suppose /? = 100 Q, Z = 0.2 H.
(a) If /^(O) = 0 and
= 20(^) V, find
i.
'j (D
Kl = -1 0 V
= 20(r) V,
and 1^2 = 20 V.
(a) Find /^(O^) and ijit) for t> 0.
(b) Sketch i^{t) for 0 < r < 4x, where x is the
circuit time constant for t> 0.
(c) Identify the zero-input response (t > 0)
and the zero-state response (^ > 0) for the
answer computed in part (a).
(d) Now compute Vj{t) for ? > 0 assuming
the switch opens at ^ = 0,0 4 sec. Plot
your result for 0 < f < 0 .2 sec.
I'
'<b O
Figure P8.21
Figure PS. 19
'Vw>
370
Assume C = 0.25 mE
and C so that the circuit
(b) Find
+
;vc(t)
'.( s O
r\
(b)
Figure P8.23
(a)
= 25(/) mA
for t >
r>
0.
Time in seconds
(b)
Figure P8.22
using linearity.
(a) If Z = 2 H , find
1 kn
Vu(t)
r>
i,(t)
d)
Figure P8.24
25. In Figure P 8.25 Ri = 2 0 0 Q, R2 = 6 0 0 Q,
(a)
T?3 = 650Q , Z = 20 H,
= -1 0 0 (-r ) +
50u{t) V, and
= 5 0 ( ? - 0.5) V. Compute
i^it) for ^> 0. Plot your answer using MATLAB
or its equivalent for 0 < r < 8x.
371
Figure P8.25
26. Repeat Problem 25, except compute VjjJ)
for t>Q.
27. In Figure P 8.27
= 2 0 0 Q, /?2 = 6 0 0 Q,
= 8 5 0 a . C = 2.5 mF, v^^{t) = - 5 0 M +
100W V, and
Figure P8.30
= - 5 0 ( f - 5 ) V. Compute
at t = 0 . Suppose
= 6R, R2 = 30i?, R^ = 20R,
and
= Vqu{T - /). In terms of Vq, R, C,
and T = 6RC,
(a) Find V(^0~) and t/(;;(0'^).
(b) Find the Thevenin equivalent seen by
the capacitance for 0 < ^ < T.
(c) Using the Thevenin equivalent found in
pan (b), find an expression for Vf^t)
Figure P8.27
20 msec
2kO
'. w ( D
8kO
=0
1.6 kn
+
Vc(t).
S jjF
20 V
+
.Vc(t)
vjt)
Figure P8.29
Figure P8.31
30. Consider the circuit o f Figure P 8.30 in
24u{t) V.
372
of Figure
v,(t)
5
t(Msec)
2.5
(b)
--10V
----------(a)
IMegO
(b)
t(msec)
40
--V.0
(a)
V ({0 ) = 0
as follows:
400
(b)
5kO
C = 0.1 F
IV
V,
Figure P 8 .3 3
Figure P 8 .3 6
lOlv.
373
OP AMP CIRCUITS
expression for
in terms of
Vo,{t) when
= 100sin(20r)//(r) mV
assuming that
= 0 in each case.
v.(t)
R.
O v/W
+
Vjt)
A .
Figure P8.37
KR
Figure P8.41
M ISCELLAN EOUS
Figure P8.38
= \QOe~~ii{t)
mV.
for r > 0.
and
r
C, + Co
Prove this equivalence using the integral rela
tionship o f a capacitor to show that the i-v ter
minal conditions are the same for both
N,.
Figure P 8 .4 0
and
374
C .d - v ,
N1
(a)
(b)
Figure P8.42
43. As mentioned in Problem 42, although
most of the first-order circuits considered in the
text have only one inductor or one capacitor, it
is possible to have a first-order circuit contain
ing more than one energy storage element.
Consider the situation depicted in Figures
P8.43a and b. Here
and
given. The networks
and A/'j are equivalent
under two conditions:
^eq -
(a)
(b)
Figure P8.43
Figure P8.45
C, - t
V,
Figure P 8 .4 4
r>
375
and K 2 . Find
for appropriate
'JO
and
K2 in terms of K^.
iA n i'
Figure P8.47
= 220W mV,
= 4 4 mA, and
rect?
11 mA.
a.
capacitors at ^ = 0^?
'> ^ '
( 1)
Figure P8.49
where
capacitors.
Pi.
(b) With K2 determined, evaluate x(/) at f =
. Your
result should be
Figure P8.50
(a) Suppose vj^i) = K^u{t) V. Show that for t
W
dv
It
= -K
iv + K 2
x Uq
) - F t
(2)
3 "6
APPLICATIONS
52. An approximate sawtooth waveform can be
drops to 60 V.
= 10 V,
for one
for four
switchings.
(b) Plot
/:
slope
= 1m U
/
Figure P8.52
slope
negative resistance
region
= 0.5 mU
60
(b)
N fc i
v(V)
-----------
90
(0
Figure P8.53
m is based on a hypothetical
= 0. At r =
ates as follows;
the device
is switched from
age).
377
0 < ^<
denote
Tj < t.
K kf
tori
^1>
^ 2> d
This is to prevent the substitution of
Compute
parts for
^ston>
^store
down an expres
sion for the power absorbed at time t.
Call the results
and PRstore
Figure P 8 .5 4
the
378
1.5V
lO kn
E
Potentionmeter
crude transistor
circuit model
Figure P8.55
56. The circuit o f Figure P 8.56 suggests a way of generating a sustained sinusoidal oscillation. All
op amps are assumed to be ideal. Capacitors, C = 0.1 fiF are uncharged at f = 0. The first two op
amps are differentiators and the last is an inverting amplifier.
(a) With switch S at position A and v^{t) = sin(lOOOr) V, find vj^t),
and
for r > 0.
(b) If at a later instant switch S is quickly moved to position B, what would you expect
to be?
Figure P8.56
s B s m
380
food
food
conducting plate
conducting plate
(b)
(a)
food
0
Reversal of the polarity o f the applied voltage at a low frequency can be easily achieved with the
circuit elements studied in earlier chapters: the resistor, the capacitor, and the inductor. However,
the friction-induced heat production is inefficient at low frequency, l b produce a useful amount
o f heat for cooking purposes, ver)' high frequencies must be used. The t}'pical frequency used in
a microwave oven is 2.45 gigahertz, i.e., the water molecules reverse their orientations 2 x 2.45 x
10^ times per second. At such a high frequency, capacitors and inductors are quite different in
their behavior from their conventional forms. For example, the LC circuit becomes a resonant
cavit} and the connecting wire becomes a waveguide. These microwave components will be
studied in a future field theory course. T he theory studied in this chapter will enable us to under
stand the low-frequenc}' version o f the phenomenon, i.e., how a connected inductor and capaci
tor can produce oscillator)' voltage and current waveforms.
CHAPTER OBJECTIVES
1.
Investigate the voltage-current interactions that occur when an ideal inductor is con
nected to an ideal capacitor with initial stored charge.
2.
Use a second-order differential equation for modeling the series RLC and parallel RLC
circuits.
3.
Learn to solve a second-order differential equation circuit model by first finding the nat
ural frequencies o f the circuit, then looking up the general solution form, and finally
determining the associated arbitrary constants.
3<S I
4.
5.
damped responses.
Investigate and understand the underlying principles o f various oscillator circuits.
SECTION HEADINGS
1.
Introduction
2.
3.
4.
5.
Oscillator Application
6.
Summary
7.
8.
Problems
1. IN TRO DUCTION
The previous chapter developed techniques for computing the responses of first-order linear net
works, either without sources or with dc (constant) sources, having first-order linear differential
equation models. Recall that the source-free response contains only real exponential terms.
This chapter focuses on second-order linear networks having second-order linear differential equa
tion models. Usually, but not always, a second-order net\vork contains t^vo energ)^ storage ele
ments, either {L, Q , (C Q , or {L, L). Second-order circuits have a wide variety o f response wave
forms: exponentials
sinusoids, and exponentially growing sinusoids, among others. Tables 9.1 and 9.2 catalogue the
various response types. W ith no sources or with constant-value sources, some straightforward
extensions o f the solution methods o f Chapter 8 are sufficient to compute the various responses.
The behavior of second-order circuits is a microcosm o f the behavior o f higher-order circuits and
systems. Many higher-order systems can be broken down into cascades o f second-order systems or
sums o f second-order systems. This suggests that our exploration o f second-order circuits can
build a core knowledge base for understanding the behavior o f higher-order, more complex phys
ical systems.
Many introductory texts discuss only parallel and series RLC circuits, stating separate formulas for
the responses o f each. Our approach seeks a unified treatment. To this end, we formulate a basic
second-order differential equation circuit model. The associated solution techniques become
applicable to any second-order linear nersvork and, for that matter, to second-order mechanical
systems.
An oscillator circuit (section 5) motivates our study o f second-order linear networks. The chapter
contains several other practical examples illustrative o f the wide variety o f second-order circuit
382
applications. Some advanced applications pertinent to higher-level courses include low-pass, highpass, and bandpass filtering (covered later in the text); dc motor analysis; position control; and
many others. Most important, the concepts presented in this chapter are common to a host o f
engineering problems and disciplines. Hence, the techniques and concepts described here will
prove useful time and time again.
EX A M PLE 9 .1 . The goal o f this example is to develop a differential equation model o f the cir
cuit in Figure 9.1b. In Figure 9.1a, with the switch S in position A, the voltage source charges the
capacitor to
(a)
(b)
FICiURF 9.1 (a) A voltage source charges a capacitor, (b) An LC second-order linear network in
which the cnerg) stored in the capacitor in part (a) is passed back and forth to the inductor.
S
o l u t io n
Step 1. Write down the terminal i-v relationship for the capacitor and itiductor; then apply KCL and
KVL, respectively. Using the i-v relationships for L and C (see Chapter 7) in conjunction with KCL
and KVL, it follow's that
ih C
_ >c _
dt
and
L_
V
J V
C definition
KCL
di
(9.1a)
(9 .1 b )
L definition
KVL
383
Step 2. Obtain a dijferential equation in the capacitor voltage, V(j^t). For this, first differentiate
equation 9.1a to obtain
1 dii^
dt
C dt
= - VC
dt
LC ^
(9.2)
Equation 9.2 is a second-order linear differential equation circuit model o f Figure 9.1b in terms o f
the unknown capacitor voltage, v^t). Equation 9.2 stipulates that the second-order derivative
o f the unknown function, v^it), must equal the function itself multiplied by a negative constant,-----
LC
Step 3. Obtain a differential equation in the current, ij{t). An alternative circuit model in i^ is obtained
by first differentiating equation 9.1b and substituting equation 9.1a into the result to produce
Equation 9.3 has precisely the same form as 9.2: the second-order derivative o f the unknown
function, ij{t) , equals the function itself multiplied by a negative constant, -----. This similarity
suggests a similarity o f solutions, which we shall pursue further.
Exercise. Fill in the details o f the derivation o f equation 9.3 from 9.1.
Our next goal is to construct the waveforms V(^t) and ij{t), which are the solutions o f the differ
ential equations 9.2 and 9.3. Although differential equations are not usually part o f the common
background o f students in a beginning course on circuits, the solutions o f equations 9.2 and 9.3
do not demand this background. Some elementary knowledge o f differential calculus is sufficient.
Specifically, recall the differential properties o f the sine and cosine functions:
dt
dt
d2
^y2
7 sin((or-f-0) = - 0) sin(co/+ 0 ) and 7 cos(o)/+ 0 ) = - c a cos(co/+ 0 )
dt-
dt-
In both cases, the second derivative equals the function itself multiplied by a negative constant.
This is precisely the propert}' required by equations 9.2 and 9.3. Thus one reasonably assumes that
the solutions o f equations 9.2 and 9.3 have the general forms
V(^t)
and
K cos(ojr
+ 0)
(9 .4 a )
38^
/^(r) = /^cos(tor+0)
(9.4b)
7'hese forms are general because the cosine function can be replaced by the sine function with a
proper change in the phase angle. Specifically, we note that A'sin((i)/ + (j)) = A'cos(co/ + (j) - 0.5ti)
= K cos((or + 0) with 0 = ([) - ().5ti. Computing values for (o, A', and 0 specifies the solutions to
equations 9.2 and 9.3.
E X A M PLE 9.2. Find A"and B for the capacitor voltage in equation 9.4a.
So
lu t io n
dvrit)
(It
= -A c o sin (O )/+ 0 )
/'o S'!
Step 2. Dijferentiate n second time. Differentiating equation 9.5 (the second derivative o f 9.4)
yields
d~Vr
">
dt~
___
(9.6)
Step 3. Match the coefficients o f equation 9.6 with those o f 9.2 to specify O). Under this matching,
' o1r CO =
(0 ^ = -----
LC
7)
j= =
fo
fiC
^ ^
Equation 9.7 specifies co, the angular frequency o f oscillation, in rad/sec, o f the capacitor voltage.
Step 4. Compute K an d f) in equation 9.4a. These two constants depend on the initial conditions
as follows: when the switch is at position A, the capacitor is charged up to I^q volts and the induc
tor current is zero; immediately after the switch moves to position B, i.e., at r = 0+, the continu-
it>' properry o f the capacitor voltage ensures that y^^O*) = Vj, and^the continuity property o f the
inductor current ensures that //(O^) = 0. The initial value,
^ , is now calculated from
equation 9.1a as
^/r(--(0 ) i( ^ ( 0 )
------------ = ----------- = ------------- = ()
dt
= K c o m = v;,
(9.8a)
and
</v'r(0" )
^
= -A o )s m (e ) = 0
(9 g y
From equation 9.8b, 0 = 0. Consequently from 9.8a, K = V^^. Hence the capacitor voltage, i.e.,
the solution o f the second-order differential equation 9.2, is
Vc(f) = VqCOS
[J
lc
(9.9)
38S
As per equation 9.1a, one can obtain ij{t) directly by differentiating equation 9.9 and multiply
ing by C. However, one could aLso solve equation 9.3 by repeating the above steps to arrive at
the same answer.
Exercise. Assuming that /^(r) = K coslcor + 6), solve for (o, K, and 0 in terms o f the initial condi
tions, and show that
Several very interesting and significant facts about this parallel ZCcircuit and the solution method
are apparent:
(1)
For the source-free LC circuit o f Figure 9.1, the voltage and current responses are sinusoidal
waveforms with an angular frequency equal to _L _ . Since the amplitude o f sinusoidal
oscillations remains constant (i.e., does not
VZc
rhe frequenq^ (o, depends on the values o f L and C only, while the amplitude K and the
phase angle 0 depend on L, C, and the initial values o f the capacitor voltage and inductor
current.
(3)
Although the instantaneous energ}" stored in the capacitor, \V^{t), and the instantaneous
energy stored in the inductor,
investigated in a homework exercise.) Physically there is a continuous exchange o f the energ)' stored in the magnetic field o f the inductor and that stored in the electric field of the
capacitor, with no net energy loss. This is analogous to a frictionless hanging mass-spring
system: because o f the absence o f friction, the up-and-down motion of the mass never
stops; in such a mechanical system there is a continuous interchange between potential and
kinetic energy.
Figure 9.1 shows what is, in theory, the simplest circuit that generates sinusoidal waveforms. Such
an electronic circuit is an (idealized) oscillator circuit. Oscillator circuits play an important role
in many communication and instrumentation systems.
386
EXA M PLE 9 .3 . For the series and parallel RLC circuits shown in Figure 9.2, develop two sec
ond-order differential equation models (one in
and one in
+
V
''c = V''l
= V
''r
(b)
FIGURE 9.2 (a) Series RLC circuit, (b) Parallel RLC circuit. Passive sign convention is assumed as usual.
S
o l u t io n
Step 2. Choose
as circuit variable
and iQ in terms o f
^ fL(ii
1
Ril + L + / ^ ( t V t = 0
dt
C
oo
'
\f^ I f
,
dv/"
+
vrix)dx + C ^ = 0
R
L ^ ^
dt
divide by C to obtain
d^il
Rdi^
1 .
^
^ + ------ ^ + ----- 1, = 0
dr
L dt
LC ^
d^Vf
dt-
d\>r
RC dt
1
H-------= 0
LC
in
d ir
Rin -f" L ------- 1= 0
^
dt
^
arid i^ in
dt
38
= C dv^^dt,
L dij/dt,
LC yields
LC yields
R dvr
L dt
d t
llL
dr
LC
RC dt
LC
Each circuit has two second-order differential equation models, one each for
= 0
and
as the
unknown quantity.
or the voltage
to satisfy cer
tain differential constraints. All four (differential) equations have the general form
d^x
dx
~ + h ----- h (.'.V = 0
dtdt
(9 . 10)
ond derivative o f the function x{t) plus h times the first derivative o f x{t) plus c times x(t) itself
adds to zero at all times, t. Unlike the example o f section 2, where a sinusoidal solution was easi
ly predicted, the present differential equation requires a more careful mathematical analysis. Recall
from elementary calculus that the derivative o f an exponential is an exponential. Thus the first
and second derivatives o f an exponential are proportional to the original exponential. This sug
gests postulating a solution o f the form at(^) = Ke^ where we make no a priori assumptions about
s. If it is truly a solution, it must satisfy equation 9.10. Under what conditions will x{t) = Ke^ sat
isfy equation 9.10?
E X A M PLE 9.4. Determine conditions under which the postulated solution x{t) = Ke^ satisfies
equation 9.10.
S
o l u t io n
rfV '
dr
+ bK
<h
+ cK e = Ke" (,v^ + hs + c) = 0
^
(9.11)
Step 2. Interpret equation 9.11. For nontrivial solutions, K\s nonzero. The function e^ is always
different from zero. Hence the quadratic in s on the right side o f equation 9.11 must be zero. This
necessarily constrains $ to be a root of
r + bs + c =0
(9 .1 2 )
3<S8
Step 3. Solve equation 9.12. From rhe quadratic formula, rhe roots o f equation 9.12 are
.V|, .v. =
-b
- 4c
(9.13)
C O N C LU SIO N : x(/) = Ke^' satisfies equation 9.10 provided s tatces on values given by equation
9.13. Equation 9.10 does not constrain K\ however, the initial conditions will.
Equations such as 9.12 whose solution is given by equation 9.13 are a common characteristic o f
second-order networks. Hence, equation 9.12 is called the characteristic equation o f the secondorder linear circuit. The associated roots, equation 9.13, are called the natural frequencies o f the
circuit. These are the natural or intrinsic frequencies o f the circuit response and are akin to the
natural frequencies o f oscillations o f a pendulum (for small swings) or o f a bouncing ball.
From elementary algebra, a quadratic equation (the above characteristic equation) can have dis
tinct roots or equal roots. Distinct roots can be real or complex. Thus
tinct real roots, two distinct conjugate complex roots, or two repeated (equal) roots, depending
on whether the discriminant, Ip- - 4r, is greater than, less than, or equal to zero. This trifold
grouping separates the solution o f equation 9.10 into three categories, listed below as cases 1, 2,
and 3:
Case 1. Real and distinct roots, i.e., b~ -A c> 0. If the roots are real and distinct, then for arbitrar}'
and K2 , both
constants
a:,6^"''
and
xif) = X2it)=K2e-'^-'
satisfy the second-order linear differential equation 9.10, i.e., are solutions to the differential equation. Since equation 9.10 is a
X2 (t) is also a solution, a fact easily verified by direct substitution. Therefore, whenever
s-y, the
xit)=K^e^^' + K2e^^-
The constants
(9.14)
and K-, depend on the initial conditions o f the differential equation, which
depend on the initial capacitor voltages and inductor currents. For example, if a.*(0'^) and a''(0'*')
are known, then from equation 9.14,
.v (0^ )=
A y '*'-h
and
.v-(O^) = ^
(It
= .s'l
1= 0 ^
+ stK')
389
and
and $2 are negative, the response given by equation 9 .1 4 decays to zero for large f and the cir
and $2 , o f the characteristic equation are distinct hut complex, i.e., iP- -A c <Q.
Since Jj ^
9.14, i.e.,
=
with complex
and
$2
given by
+K2e"^^
_______
-b
. y l^ c -b ^
^2 = Y J ------ ------- =
,
J^ d
(9.15)
y l^ c-b ^
where a = bH and co^ = ------ ^ S i n c e s^ and S2 are conjugates, so are d^i^and e^2^in equation 9.14.
For x(t) to be real, the constants /Cj and A'j in equation 9.1 4 must also be complex conjugates,
i.e.,
+ j sin y
the two terms in equation 9.14 combine to yield a real time function:
= e
response
x(r) = e~^^ [A cos((o/) + B sin((0/ )]
where ^4 =
+^
= 2 Re[^j] =
+ A^2 and 5
(9.16)
stants and where Re[ ] denotes the real part and Im[ ] denotes the imaginary part. The solution
expressed in equation 9.16 is completed by specifying ^4 and B. As before, A and B depend on the
initial conditions, x(0^) and xXO"*^) as follows:
jc(0'*) =
[Acos(tOjf) + Bsin((0^r)])^_^^ = A
390
jr'(0'*) =
= - o A + (o^B
and
which are easily solved for A and B.
Making use of a standard trigonometric identity, the general solution of equation 9.16 has the
equivalent form
x{t) =
(9.17a)
where
K = ylA^ + B^ ,Q = tan"
f-B \
(9.17b)
KA)
and the quadrant of 0 is determined by the signs o f - B and A. In MATLAB, one uses the command
atan2(-BvA) to obtain the angle in the proper quadrant. Note that the response waveforms have
oscillations with angular frequency O)^ These oscillations are bounded by the envelope Ke~^*. If
Re[jj] = - a < 0, the amplitude of the oscillations decays to zero and the response is said to be underdamped. If Re[jj] = - a > 0, the amplitude of the oscillations grows to infinity.
C ased. The roots are real and eqml, i.e.,lP' - 4 c = 0 . When the two roots of the characteristic equa
tion are equal, equation 9.1 4 does not represent the general solution form because if
two terms collapse into a single term. However, the general solution for
the
is
x(t) = {K^+K2t)e^^^
(This is investigated in a homework exercise.) Calculation of
and
in equation 9.18 is
straightforward:
x{0-) =
and
i(0+) =
Substituting the value o f
If^l = i2 is negative, the response decays to zero and is said to be critically damped. Critically
damped defines the boundary between overdamped and underdamped. This means that with a
slight change in circuit parameters, the response would almost always change to either over
damped or underdamped.
T he discussion o f the three cases is summarized in Table 9.1.
391
d~x
dx
dr
dt
^ + h + cx - 0
having characteristic equation
+ r = (j - 5 j)(j -
= 0, where
- h yjb" 4 c
^2=
-----------
+ Kie^-
where
40^) =
characteristic equa
, e = tan"^
(-B \
A
= Sj and Ir - Ac
x{0*) =
Figure 9.3 displays the various response forms described above for the case where Re[xj] and Re[j-,]
are negative or zero. Because o f their similarity, it is not possible to distinguish between the over
damped and the critically damped responses by merely looking at the waveforms. Both types o f
response may have at most one zero-crossing.
392
> t
FIGURE 9.3 (-'tncric waveforms corresponding to the four cases o f damping: (a) undamped (sinu
soidal) response, (b) undcrdamped (exponentially decaying oscillatory) response, (c) overdamped
(exponentially decaying) response, and (d) critically damped (exponentially decaying) response.
The terms undamped, underdamped, overdamped, and critically damped stem from an
intuitive notion o f damping. The sourcc-frce response o f an undamped second-order linear sys
tem, whether electrical or mechanical, has an oscillatory response (waveform) o f constant ampli
tude. Damping, due to system elements that consume cnerg\', means a monotonic decrease in the
amplitude o f oscillation. In electrical circuits, resistances produce the damping effect. In mechan
ical systems, friction causes damping. When the amount o f damping is just enough to prevent
oscillation, the system is critically damped. Less damping corresponds to the underdamped case,
where oscillation is present but eventually dies out. A greater amount o f damping corresponds to
the overdamped case, where the waveform is non-oscillatory, and a ver\ small perturbation o f any
circuit parameter will not cause oscillations to occur.
In summary, once the roots o f the characteristic equation are found and the expression for the
general solution selected from the above cases or Table 9.1, it remains to find the constants A', and
3^)3
K j (or A and 5 ) from the initial conditions on the circuit. In the above development,
and Kj
(or A and B) are given in terms of a:(0^) and x'(O^). Since jf(/) represents either a capacitor voltage
or an inductor current, its value at f = 0"^ is usually given, or can be determined from the past his
tory o f the circuit. (See Example 9.5.) The value o f x'(O^), on the other hand, is often unknown
and must be calculated. If >:(/) =
xX0*) = v'c{0*) = ^ ^ .
^
If x{t) =
, or an unknown
or y^(O^), we construct an auxiliary resistive circuit valid at r = 0^. Since the initial
values,
and /^(O^), are known, we replace (each) capacitor in the original circuit by an inde
pendent current source o f value /(0'*^). Here the current /(;^0^) retains its original direction and
the voltage ^^(0^) retains its original polarity. After the replacements, the (new) circuit is resistive.
Values for
and y^(O^) follow by applying any of the standard methods of resistive circuit
analysis learned earlier. This allows us to specify x(0^) and x'(O^) in terms of the initial conditions
on the circuit. Two equations in the two unknowns A'j and K2 (or A and B) result. Example 9.5
and, in particular. Figure 9.4c illustrate this procedure.
39-'i
E XA M PLE 9 .5 . In the circuit o f Figure 9.4a, the 1 pF capacitor is assumed to be ideal, and the
inductor is modeled by a 10 mH ideal inductor in series with a 20 Q resistor to account for the
resistance o f the coiled wire. Suppose the switch S in Figure 9.4a has been in position A for a long
time. The capacitor becomes charged to 10 V. Then the switch moves to position B at f = 0. Find
and plot V(\t) for r > 0 for the following three cases; (1)
= 405
180 ti.
:L=10mH
lO O
+
10V
C=1
R=200
Practical Inductor
mF
(a)
i,(0*) = 0
L=10mH
+
C=1
mF
(b)
FKJURH 9.4 (a) Discharge of a capacitor through a practical inductor in series with a resistance /?,.
(b) Kquivalent circuit for t > 0. (c) Equivalent circuit at / = 0^ for calculating
in which the
inductor has been replaced by an independent current source of value //(O^) and the capacitor by an
independent voltage source of value
SoL u rioN
- 10 V by continuit)' o f the capacitor voltage; the circuit now becomes a series RLC, for t
> 0, with / ?= / ?!+ R-y as shown in Figure 9.4b. The first step in the calculation o f the circuit
response is to find a second-order differential equation in the unknown
R dvf'
*- + -------^ +
dt~
L dt
\r = 0
LC ^
(9.19)
( 9 .2 0 )
v c(0 =
+ K2e^~ =
(9.21)
(9.22b)
by the continuity o f the inductor current. Solving equations 9.22a and b after substituting the
above values yields
AT, = 10.667 and K-, = - 0 .6 6 7
Step 3. Set forth the solution for V(^t). For / > 0,
v^t) = 10.667^-2500^-0.667e-^0'V
This function is plotted in Figure 9.5.
Exercise. You may verify' the above answer with the Student Edition o f MATLAB (version 4.0 or
later) by typing the command: y = dsolve(D 2y+42500*D y + leS^y = 0,y(0) = 10, Dy(0) = O).
Case 2: /?, = 0 or /? = 20 Q
Step 1. If /?2 = 0> then from equation 9.20, the characteristic equation is
+ 2,000^ + 10^ = 0.
Since tr - Ac = -3 9 6 ,0 0 0 ,0 0 0 < 0, the roots are complex. From the quadratic root formula,
= - 1 0 0 0 + y9950 = - a + yo)^ and Sj = - 1 0 0 0 -y '9950 = - a
From Table 9.1, the underdamped response form is
v^t) =
cos(9950f) + B sin(9950f)]
(9.23)
396
Step 2. Fiiid A and B. Ir remains to determine A and B in equation 9.23. From equation 9.23
and its derivative,
i / ^ 0 ^ ) = 1 0 = /I
(9 .2 4 a )
and
(9.24b)
A = 10 and
= = 1.005
CO,/
Exercise. You may verify the above answer with the Student Edition oF MATLAB (version 4.0 or
later) by typing the command: y = dsolve(D 2y+2000*D y + le8*y = 0,y(0) = 10, Dy(0) = O).
Case 3: ^2 = 180
Step 1. If
=180
or R= 200 Q
then the characteristic equation from 9.20 is
+ 2 0 ,0 0 0 j + 10^ = 0, whose
roots are
= -lO '*, implying a critically damped response. From Table 9.1, the general criti
cally damped response form is, for r > 0,
v^(/) = (/r, + K2t)e^^' = (A", + K2t)e~^^'^
(9-25)
Step 2. Find /Tj and Kj. From equation 9.25, its derivative, and the known initial conditions
from cases 1 and 2,
v^iO^) = 10 = A",
(9.26a)
and
^yO ") =
(9.26b)
39'
FIGURU 9.5 Waveforms of vj,t) in Example 9.5 for three different degrees of damping. Critical
damping represents the boundary between the overdamped condition and the oscillatory behavior of
underdamping.
Exercise. Verify the answer calculated in Example 9.5 using the Student Edition o f MATLAB and
the desolve command.
On a practical note, commercially available resistors come in standard values each with an associ
ated tolerance. Tolerances vary from 1% (precision resistor) to as much as 20% . Further,
because o f heating action over a long period o f time, resistance values change. Given the above
example, in which the type o f response depends on the resistance, one can imagine the care need
ed in the design o f such circuits: without consideration o f precision and long-term heating effects,
a desired critically damped response could easily become oscillator)'.
Not all second-order circuits arc RLC. Some are only RC but with two capacitors and some are
RL with two inductors. Passive RC or RL circuits cannot have an oscillatory response. The proof
o f this assertion can be found in texts on passive network synthesis. However, with controlled
sources a second-order RC or RL circuit can have an oscillatory response that is not characeristic
o f a first-order circuit, but o f a second- or higher-order circuit. The example below illustrates the
analysis o f a second-order RC circuit containing controlled sources that has an oscillator)^
response.
398
EX A M PLE 9 .6 . This example illustrates the analysis o f the second-order RC circuit shown in
Figure 9.6. The objective is to find
= 10 V and
= 0.
'Cl
1 kn
I
and
mF
FIGURE 9.6 Sccond-order RC circuit with controlled sources that has an oscillatory response.
S o l u t io n
Step 1. Write a dijferential equation in V(^^{t). From the properties o f a capacitor and KCL at the
left node,
di
Multiplying through by 10*^ yields
d\
^^ 10
dt
(9.27a)
, Q3
dt
dt~
To obtain a differential equation in
(9.27b)
dt
106 C l _
_ -0 .
dt
or equivalently,
dt
= -1 0 ^ Vc,
(9.28)
dt
*2
'2
dt^
dt
(9.29)
Step 2. Determine the characteristic equation, its roots, and the form o f the response. The differential
equation 9.29 has characteristic equation
+ 1 0 = 0
From the quadratic formula, the complex roots are
= - 5 0 0 + 799,998.75 = - a y W
From Table 9.1, complex roots imply an underdamped response o f the form
dt
(9.31)
Differentiating equation 9.30, evaluating at r = 0, and equating the result with equation 9.31
produces
dl
"
Exercise. Construct a parallel RLC circuit to have the same second-order difFerential equation
model as 9.29. Note that there is no unique solution.
C H EC K : /?C= 10-^ and ZC = \0~^^
It is important to observe here that the design o f Example 9.6 achieves a second-order RLC
response without the use of an inductor, which is important for integrated circuit technolog)'.
clf-
+ / ; ^ + c-.v = /(/)
dt
(9.32)
wherey{r) is a scaled sum o f the inputs and/or their first-order derivatives. Ordinarily one might
expecty(f) to be the value o f the input. A homework problem illustrates that j{t) can depend not
only on the source input, but also on the derivatives o f the input. For general circuits, those not
reducible to parallel or series RLC circuits, constructing equation 9.32 can be a challenge. Further,
the solution o f 9.32 for arbitrary inputs and initial conditions is no less challenging but is best
obtained via the Laplace transform method, which is a topic studied in a second circuits course.
However, when the input excitations are constant,/r) = F, the solution to 9.32 is a straightfor
ward modification o f the source-free solution, as explained in the remainder o f this section.
Since the expressions o f Table 9.1 satisfy the homogeneous diflferential equation 9.10, the gen
eral solution to equation 9.32 follows by adding a constant
in Table 9.1. Specifically, the general solution o f the driven differential equation
tl~x
dx
- + h ^ cx = F
(9 . 33 )
(9.34)
dr
dt
where-v^j(f) is the solution to the homogeneous equation 9.10 (equivalently, equation 9.33 with F
= 0). Recall that the form o f x j t ) is determined by the roots of the characteristic equation r + bs
+ r = (i - ^])(j - ^2) = 0, given by the quadratic formula
h yjh~ - 4 c
i |-) = -------------9
9
*0 1
To verify that the structure o f equation 9.34 is a solution to 9.33 and to compute the value o^Xp sub
stitute the structure given by equation 9.34 into 9.33. Since
9.10, it contributes zero to the left-hand side. What remains is cXp= F. Therefore,
= which
is independent o f the roots o f the characteristic equation. However, if Re[j|] and Re[^2] < 0, then
tends to zero for large t. Hence x{t) tends to Xp for large t. Consequently Xp is termed the
final value o f the response.
Because o f the trifold structure o f
9.34 once again breaks dow'n into three distinct cases. We summarize this trifold structure for the
constant-input case in Table 9.2.
TABLE 9.2 General Solutions for Constant-Source Second-Order Networks
-z- + h + cx = F
dr
dt
h yjb^ - 4 c
-7 = ---- --------------
2
2
Case 1. Real and distinct roots, i.e.,lP' 4 r > 0:
x {t)=
+ Xp
F
with Xp = . Further,
c
x(0'^) =
x{0^) = A + X p x(0^) = - g A +
K
V
Case 3. The roots are real and equal, i.e., s, = s, and Ir -A c= Q . The solution
form is
.v(/) = (/ r ,+ ^ 2 0 ^ '
where Xp = , and
x{Q^) = K^+Xp
402
The interpretation Xp = F/c. is a mathematical one. When the differential equation describes a lin
ear circuit with constant inputs, there is a physical interpretation o f Xp and a circuit theoretic
= Xp
method for computing its value, even without writing the differential equation. Since
= a constant satisfies the differential equation 9.33, it is also a constant solution to the circuit.
Hence Xp\s either a constant capacitor voltage or a constant inductor current. If a capacitor volt
age is constant, its current is zero; this is interpreted as an open circuit. Similarly, if an inductor
current is constant, its voltage is zero; this is interpreted as a short circuit. Therefore, Xp is an
appropriate (capacitor) voltage or (inductor) current obtained when the capacitor (or capacitors)
are open-circuited and the inductor (or inductors) arc short-circuited. The value o f Xp can be
obtained by analyzing the resistive network resulting when all capacitors are open-circuited and
all inductors are short-circuited. Recall that if Re[^j] and Re[^2] < 0, then x{t) tends to the constant
value Xp Physically speaking, then,
equals either
Once the proper general solution structure is ascertained from Table 9.2 and the constant Xp is
found, the parameters
and
source-free case. The following example illustrates the procedure for a parallel RLC circuit.
whose initial conditions satisfy /)(0) = 0 and y^^O) = 0. This simply means that the current source
turns on with a value of 1 amp at r = 0 and maintains this constant current excitation for all time.
The objective is to find the inductor current,
So l u t io n
Because the circuit is a parallel RLC, the characteristic equation is
RC
LC
(9.35)
For all positive values o f R, the roots of the circuits characteristic equation have negative real parts.
Thus for large t or ideally at t = oo, the inductor looks like a short circuit and the capacitor like
an open circuit. Hence, for all cases o f this example, Xp = i[{^) = 1 A; note that y^co) = 0 because
the inductor looks like a short at t = <x>.
2 = - 1 .0
lO'^ 7 I .9975
105 = - a yco^
10*^^ = 0.
'103
/^(r) =
Since 0 =
elicit)
0=
dt
^
= -0/4 + (HjB
1 = 0^
From derivative o f
expression for ij^t)
+ 4 x 10^5 + 4 x lO' = 0.
j, 2 = - 2 . 0 x 105
indicating a critically damped response o f the form (Table 9.2)
+ 1, then
= - 1 . Further,
dt
10^ and S2 ~ ~
+ 1A
40 4
+ Xp. Furtiier,
dt
t=(f
1
Equivalently, -K^
~ Solving these two equations yields K\ = ---- and K-, = .
Therefore the actual response for r > 0 is
^
^
Figure 9.9 displays a graph o f the response for each o f the three cases.
Exercises. 1. Show that for t > 0 , the differential equation for the circuit o f Example 9.7 with R
= 500 Q is
clr
+ 2 x l 0 ' ^
dt
+ 4xl0',',.<n = 4 x l 0 JO
'
2. Use MATLAB s dsolve command to verify the solution obtained for case 1 in Example 9.7.
In a linear circuit or system, the response to a step input often indicates the quality o f the system
performance. The problem o f measuring a batter}' voltage using a voltmeter is illustrative o f this
indicator. Here the battery dc voltage is the input and the output is the meter pointer position.
Connecting the meter probes to the batter)'^ terminals amounts to applying a step input to the
voltmeter circuit that drives a second-order mechanical system consisting o f a spring and mass
with friction. Naturally, one would like the pointer to settle on the proper voltage reading quick
ly. If the mechanical system is underdamped, then the pointer oscillates (undesirably) for a short
time before resting at its final position. On the other hand, if the mechanical system is over
damped, the pointer will not oscillate but may take a long time to reach its final resting point.
This also is undesirable. A near critically damped response is the most desirable one: the pointer
will come to rest at the proper voltage as quickly as possible without being oscillatory, and small
changes in the mechanical system will not make it oscillatory.
In the next example, we reverse the process o f analysis and ask what the original circuit parame
ters are given a plot o f the response that might have been taken in a laboratory.
E X A M PLE 9 .8 . Consider the circuit o f Figure 9.10, which shows the response,
of a
(relaxed) series RLC circuit to the voltage input i/y^(r) = 10(r) V. In laboratory, you have meas
ured the capacitor voltage values (approximately). If the response has the form v^it) = Ke~^^
cos(ojy + 0) + vYp find
H. Your lab instructor has told you that to^and a are integers.
(a)
___________,
406
(b)
= 10 u{t) V.
TABLE 9.3
Tim e (sec)
0.316
0 .5 2 3 6
0.839
1.5708
vcKt) (V)
10
13.509
10
10.432
First crossing
First
Second crossing
Second
o f 10 V
peak
o f 10 V
peak
S olution
Step 1. FindXp. By inspection, the curve is settling out at X p - 10 V.
Step 2. Find
40
Now obscr\'e that the first two crossings o f v^^t) = 10 occur at r = 0 .3 1 6 sec,
0.839 sec (Table 9.3). This means that a full k radians is traversed by tlie cosine over [0.316,
0.839] , which is a half c>de or half period. So the period o f the cosine is 7 = 2(0 .8 3 9 - 0.316)
2 ti
= 6.007 = 6 rad/sec.
Step 3. Find O . From Table 9.3, we know that two successive peaks occur at
= 0.523 sec
(9.36)
COS(Ci),y/l + 0 )
C0.s((0^/r2 + 0 )
Thus
cos((0 ,//| +Q) V c i t j ) - ^ ^,-a(/2-/,)
Xf,-
cos(co^y/'> + 0 )
(9.37)
Equation 9.37 simplifies because two adjacent positive peaks must be 2 tu radians apart, i.e.,
(co/2
+ ^) =
Xc d o - ^ F
3 .509
(9.38)
Solving leads to a = 2.
Step 4. Find 0 and K. At the first crossing o f 10 V, we have
0 =
0.316
0 . 3 1 6 + B)
Thus 6 X 0.316+ 0 must equal 0.5?! or 1.5tt radians. We also know that since
=0 =K
cos(0) + 10, we must have K cos(0) = - 1 0 . Since A"> 0 (always by convention), the value o f
cos(0) must be negative. This means 0.57t < 0 < 1.5Tt. So therefore, it must be that at the first cross
ing
o f 10 V
D
6 X 0 .3 1 6 + 0 =
or 0 = 2.1864 rad.
Therefore K = ^ = 10.553.
cos(0)
Step 5. Find R and C. We know that the characteristic equation o f the series RLC circuit must be
Therefore
f?
1
02
^
^
s~ H- . V.V +H-------=
2Rss +H = (5' + 2)*' + 6 " = s~ + 4.V + 40
------- = .v + IR
L
LC
C
/? = 2 n and C = 0 .0 5 F
-H)K
In the previous examples one obsen'e that the characteristic equations are independent o f the
source values. I'his is a general property of linear circuits with constant parameters. Hence when
constructing the characteristic equation we may without loss of generality set independent source
values to zero; i.e., independent voltage sources become short circuits and independent current
sources become open circuits. With this operation, some circuits that appear to be non-series/parallel, become series/parallel. This allows us to easily compute the characteristic equation and then
use Table 9.2 and physical reasoning to obtain the solution without having to construct the dif
ferential equation explicitly. The following example illustrates this procedure for a pseudo-parallel/.series RLC. The example will also illustrate the computation o f initial conditions due to past
excitations and the computation o f the complete response w'hen the input changes its dc level.
EXA M PLE 9.9 . The circuit of Figure 9. l i b is driven by the input o f Figure 9.11 a, i.e., vj^t)
-60//(-r) + G^u{t) + 60//(/ - 1) V. Our goal is to find the response
for / > 0.
FIG URE 9.11 (a) Input cxcitation whose dc level changes at r = 0 and t = 1 second, (b) A pseudoparallel RLC circuit; i.e.,when the voltage source is replaced by a short, the circuit reduces to a paral
lel RLC whose characteristic equation is p- +
RC
s+
= 0.
LC
So l u t io n
Step 1. Analysis at 0. Here the circuit has been excited by a constant - 6 0 V level for a long
time. Therefore at ^ = 0", the capacitor looks like an open circuit and the inductor a short cir
cuit. Because the inductor looks like a short, the entire - 6 0 V appears across the 6 i l resistor,
making
Step 2. Analysis at 0^. By the continuity o f the capacitor voltage and the inductor current, the
equivalent circuit at 0"^ is given in Figure 9.12.
409
= -6 0 V
^ (0 *)= i,(0 )
= -10A
FICJURE 9.12 Equivalent circuit for analysis at 0^; the capacitor is replaced by a voltage source
o f value
= - 60/6 = - 1 0 A, and
RC
LC
where R = 2 LI is the parallel combination o f 6 Q and 3 i i . The characteristic roots are s^i = - 2 ,
which correspond to a critically damped response o f the form (Table 9.2)
V(4,t) = (/f,
Kjt)
exp(j,
t) + Xf:
Step 4. Find constants in the response form for 0 < /< 1. The input is constant for 0 < ^ < 1, but
changes its value to 120 V at r = 1 sec. However, the circuit does not know the input is going to
change, and so its response behaves as if the input were to remain at 60 V for all time: the circuit
cannot anticipate the future, and thus its response over 0 < r < 1 behaves as if no further switch
ing were going to occur. If no further switching were to occur and if the input remained at 60 V,
then in Figure 9.10b for large t the capacitor is an open circuit and the inductor is a short circuit;
hence Xp= 60 V. Under these same conditions we find
step 2,
c!t
H en ce,
K-,
/={)'
= 320 =
+ A'2 = 240 + ^2
V({t)
= (-1 2 0 +
m)e~^ +
60 V
410
il{t) = (-2 0 +
+ 10 A
Step 5. Analysis at t = 1. Although the circuit does not know the input will change at r = 1 sec,
we do and we must prepare for the analysis for f > 1. To do this we must evaluate the initial con
ditions at / = 1 and then use the continuity o f the capacitor voltage and inductor current to
obtain the initial conditions at / =
= 54.59 V
and/^d") = /^ (r) = 10 A.
Step 6. Analysis att= I
This step mimics step 2 for r = P . The capacitor is replaced again by an inde
pendent voltage source and the itiductor by an independent current source as shown in Figure 9.13.
Here v^{V) = 1 2 0 - 54.59 = 65.41 Vand
=-in\ +
+ //(O'*') =
= - 9 .0 9 8 + 21.8 + 10 = 22.71 A.
^ 120 - 54.59 ^
I
>
Step 7. Computation o f the responsefor t> 1. Because the characteristic equation is independent o f
the input excitation, the form o f the response is almost the same as in step 3, except for the
replacement o f t by (/ - 1); this substitution follows by the time invariance (constant parameter
values) o f the circuit. Thus, for r > 1,
f^^oo) = 120
that
=
dt
w hich makes
1 8 1 .7 = .ViA:, + K2 = 1 3 0 .8 + K 2
/=r
= 5 0 .9 . T h u s, for / > 1,
v^t) = [ - 6 5 .4 1
+ 5 0 . 9 ( r - l)]^>-( ' -
+ 120 V
(9 .3 9 )
41
Time in seconds
FIGURE 9.14 Complete analytical response of the capacitor voltage for 0 < r < 3 sec.
Exercise. Fill in the details for the computation o f i^it) = (-2 0 + 20/)^^ + 10 A for 0 < r < I and
then compute /y(/) for 2.5 > ^ > 1. Also, compute /^(/) for t> 2.5.
Despite the idea illustrated in Example 9.9, many second-order RLC circuits are not reducible to
series or parallel RLC circuits when the independent sources are set to zero. Furthermore, when a
dependent sourcc is present, the circuit is generally not reducible to a series or parallel RLC. In
such cases one ordinarily uses a systematic methodolog)' to compute the circuits differential equa
tion and, subsequently, the characteristic equation. This systematic procedure is described in more
advanced texts and in the second edition o f this text. Nevertheless, for some situations one can use
the earlier method s integro-differential equations, which must be differentiated again to eliminate
the integral. This is illustrated in the example o f the next section.
il2
5. OSCILLATOR APPLICATION
An imporrant difference between first-order and second-order linear networks is the possibilit)' o f
oscillatory responses in the latter. In some applications sinusoidal oscillations are intended
responses, while in other applications oscillations arc undesirable. This section presents an exam
ple o f a Wien bridge oscillator circuit.
The goal is to build a circuit that generates a pure sinusoidal voltage waveform at a specified fre
quency. In theory, as per section 2 of this chapter, this is achievable by discharging a capacitor
through an inductor. In practice, both capacitor and inductor have losses. Losses cause the oscil
lation amplitude to decay eventually to zero. For sustained sinusoidal oscillations, some active
element such as a controlled source or op amp must replenish the lost energy. Note that these
active elements require a dc power supply for their operation. Ultimately the dc power supply
replenishes the power losses due to various resistances in the circuit.
E XA M PLE 9 .1 0 . Figure 9.15 shows a Wien bridge oscillator constructed with an op amp as the
active clement. Find the condition on the circuit parameters R^,
R.
(b)
FIGURE 9.15 (a) Wien bridge oscillator, (b) Equivalent circuit.
S o l u t io n
From the principles described in Chapter 4, the non-inverting amplifier enclosed in the dashed box of
Figure 9.15a is equivalent to a voltage-controlled voltage source with a gain equal to {2Rj- + RJ}IRr= 3.
(See Chapter 4.) Replacing the dashed lx)x with this equivalent yields the simplified circuit o f Figure
9.15b. Using the simplified circuit, the first task is to derive die differential equation model of the circuit.
Step 1. Write a single-loop equation.
(9 .4 0 a )
413
dt
(9.40b)
7?1 dt
/?,C
V2
(9.41a)
/?,
dt^
l<2
dt
Step 3. Substitute equations 9.41 into equation 9.40b. Substituting as indicated yields
dt~
/?2 dt
R^C ^
dt
Rj y
- ^ =0
/?, dt
(9.42a)
d Vf
1 dv->
1 dv-y
- + ----------- ^ + ---------- ^ +
dt~
Ro_C dt
/?,C dt
R^R^C-
R^C dt
which simplifies to
dvo
d-V2
dt~
RjC
Vo
(9.42b)
R\C}
Step 4 . Compute the characteristic equation and determine the conditions for sustained oscillations.
The resulting characteristic equation is
s~ + hs + c = s~ +
\ R ,C
= 0
R ,C )
(9.43)
For sustained sinusoidal oscillations to occur, the roots must be purely imaginary. Thus the coef
ficient o f s must be zero, i.e.,
b=
R2C
/?,c
/e|/?2C
=0
(9.44)
Thus the condition for sustained sinusoidal oscillations reduces to /?, = R-,.
Step 6. Find the frequency o f oscillation. Under the condition R^ = Rj, the roots o f the character
istic equation are
/?,C
We conclude that the frequenq^ o f oscillation (in rad/sec) is
(Oo =
(9 .4 5 )
/?,C
A14
response is an exponentially growing sinusoid. For the oscillations to start, the value of/?| should
be designed to be slightly smaller than Rj- Then the value for b in equation 9.58 will be negative,
producing an exponentially growing sinusoidal response. If all circuit parameters are truly con
stant, the amplitude o f oscillation would theoretically grow to infinit)'. In real oscillator circuits,
such growth is limited to a finite amplitude by saturation effects or nonlinearities that clamp the
response when the voltage swing grows large. The resulting waveform then only approximates a
pure sine wave. The analysis o f this nonlinear effect is beyond the scope o f this book. However,
the next example illustrates the growing oscillation when
EXA M PLE 9 .1 1 . The circuit o f Figure 9.16a is a B2 Spice schematic for the Wein bridge oscilla
tor o f Figure 9.15. The op amp is a 741 with
= 0. Observe that
= 10
= 15 V. Suppose that
(0) = 10 m V and
put voltage labeled IVout should be a growing sinusoid. The output response o f Figure 9.16b
shows this growth and the saturation effects induced by the op amp. The waveform is not a pure
sinusoid due to these saturation effects. Also note that the frequency o f oscillation is approximately
16 Hz, which is consistent with equation 9.45, i.e..
=
2k
C2
R1
(a)
2 k ^R^R2 C
16.3 Hz
15
Time(s)
(b)
FIGURE 9.16 (a) Schematic diagram of Wein bridge oscillator, (b) Voltage response showing grow
ing oscillation clamped at 15 V due to saturation effects of op amp.
restore itself to the equilibrium level. In practice this equilibrium level only approximates the spec
ified value due to imperfections in the circuit parameter values. The resulting waveform is almost
a pure sinusoid.
4 16
6. SUMMARY
'I'his chapter has explored the differential equation modeling and response computation of^ sec
ond-order linear circuits having either no input or constant input excitation. Such second-order
circuits contain at least two dynamic elements, either an LC, CC, or LL combination. Secondorder circuits may also contain active elements such as op amps. In contrast to first-order circuits,
second-order linear circuits allow for the possibilit)^ of damped and undamped sinusoidal oscilla
tions.
Analysis o f second-order linear circuits has two phases. Pha.se 1 entails the formulation o f the sec
ond-order differential equation circuit model. For simple I C parallel RLC, or series RLC, the cir
cuit model can be found by inspection.
Phase 2 o f the development centers on the solution o f the second-order differential equation
model o f the circuit. I'he first step here is to compute the (quadratic) characteristic equation and
then solve for the two roots. The roots o f the characteristic equation determine the t}'pe o f
response. The three t)'pes o f roots for a quadratic real distinct, real identical, and complex
specify the three response types of overdamped, critically damped, and underdamped, respective
ly. These three types o f responses characterize all second-order linear differential equation models,
be they o f electrical circuits, mechanical systems, or electro-mechanical systems.
Since sinusoidal waveforms are germane to many electrical systems, this chapter presented an
oscillator circuit that generates a sinusoidal waveform. O f the many types of oscillator circuits, we
chose one containing an RC circuit built around an op amp, avoiding the use o f an inductor.
41
,v"(f) +
tic equation.
Characteristic roots: roots o f the characteristic equation, also called the natural frequencies o f the
linear circuit.
Critically damped circuit: a second-order linear circuit having characteristic roots that are real
and identical. The source-free response o f such a circuit has a non-oscillatory waveform,
but is on the verge o f becoming oscillator)'.
Damped oscillation frequency: in an underdamped second-order linear circuit, the source-free
respon.se has the form [Kc^^' cos(o)y + ()}. The angular frequency
lation frequenc)', which is the magnitude o f the imaginary part o f the characteristic roots.
Homogeneous differential equation: a differential equation in which there are no forcing terms.
For example, x"{t) + bx\t) + cx{t) = 0.
Natural frequencies: the characteristic roots.
Oscillator circuit: an electronic circuit designed to produce sinusoidal voltage or current wave
forms.
Overdamped circuit: a second-order linear circuit having a characteristic equation whose charac
teristic roots are real and distinct.
Scaled sum o f waveforms: let
...
+ ... +
+ cx{t) = /(r).
Source fi-ee: there are no independent sources, or all independent sources have zero values.
Step function: a function equal to zero for r < 0 and equal to 1 for r > 0.
Step response: the response o f a circuit to a step function input when all capacitor voltages and
inductor currents are initially zero.
Undamped circuit: a second-order linear circuit where the characteristic roots are purely imagi
nary and the unforced response is purely sinusoidal.
Underdamped circuit: a second-order linear circuit whose characteristic roots are complex with
nonzero real part.
' T h e notations
and KO arc used interchangeably in the literature to den ote the first derivative ol v{t).
418
PROBLEMS
TH EO RY RELATED
site signs.
(b)
(c)
cos(tt)0 = - W cos(0)/)
>
7 sin((0 / ) = - 0 )
crossing.
sin(coO
dt^
eral form
d^\>C
dr
LC
6. (a)
underdamped
v^t) =
and
W^{t) is constant.
the
+ 5.7)
How many cycles o f ringing occur in
and
form
is
response is given by
where
+ 2as +
form
x{t)
+ to^ with
=
volts is applied to a
x{t) =
419
are zero at / = 0.
t =0
itor.
Figure P9.7
Show that for f > 0
Figure P9.9
1
.V l
]
c
',
and
yc(t) = Vo 1 -
COS
(a)
Compute
, //(0~), and
(b)
(c)
Jlc\
previous problem.
f > 0.
(d)
Figure P9.8
i,(t)
,v (t)
v(t) I
Figure P9.10
11.
vU)
and
' l ( 0 = / o 1 - cos
(a)
Compute
/^(O"), and
(b)
;,( 0 -) .
Compute the energ}' stored in the
inductor and the capacitor at r= 0.
(c)
420
(d)
i,(t)
v(t)
(a) (t>0)
Figure P 9.11
12. Reconsider the circuit o f Figure P 9 .ll
= 50(-/) V, R = 25
Q, C = 0.8 mF, and Z, = 1 H. Repeat Problem
under the conditions
11.
Figure P9.15
/(^0),
t=0/\
L
25 0
_L
100 0
ing.
25 mV
(b)
Figure P9.14
15. Consider the circuit o f Figure P9.15, in
which
= 2 QV, R^ = 2 Cl, Rq = %Q., R = 2 Q,
Z. = 0.5 H, and C = 62.5 mF.
(a)
Figure P9.16
B at r = 0. Find
y^^O'*^), /^0~),
tion B at t = 0. Suppose
0.5 a
(a)
(b)
= 100 mV, R =
/. = 1 H, and C = 0.01 F.
Find
and
for f > 0.
421
(a)
L and C
(b)
Figure P9.17
O '
current.
0,5u{-t) A.
(a)
I f /? = 20 Q , Z = 1 H, and C = 8 mF,
find and plot V(^t) and
(b)
for f > 0.
ijt)|
Figure P 9.18
19. In Figure P 9.19. ^^<0") = 25 V, /^(O-) = 50
mA, /? = 2 kQ, Z = 0.1 H, and C = 0.1 ;/E The
switch closes at f = 0.
(a)
(b)
Compute /^(/),
(c)
(d)
and V({t).
|
|
Figure P9.20
Why?
/YYV<
L
for f > 0.
(a)
(b)
t=0
Figure P9.19
(c)
'n^
= (20 - \200t)e~^^* mV
422
i jt )
/ Y W
Figure P9.21
Figure P9.24
source-free circuit.
ically damped.
(b)
0.1 H
mV
and
/^(O)
= -6 0
mA.
Compute V(\t).
(c)
40 0
0.25 mF
+
sVc(t)
F'igure P9.25
,_ r Y Y \ =
(b)
/?///?, = 0.8
(a)
and A = 0.04 H.
(b)
(c)
Suppose
= 160 mV and /^(O) =
- 3 0 mA. Compute
Determine the first time at which the
V(^t) =
(a)
(b)
cos(8r + B)
A and 0.
24. In the circuit o f Figure P9.24, /? = 20 Q and
(c)
general
response
under this condition.
Then determine the source-free
response when V(^0) = 10 and /^(O) =
0.
form
of
the
source-free
423
//(O'*'),
in two steps:
29.
and
which
= -10w(-/) + 20(f) V, 7? = 20 Q,
Compute
/^(0), and
(b)
(c)
R=
R = 0.8 Q
ijO
L= 1H
R = i s n
C = 1/36
(b)
Figure P9.27 Parallel and series RLC circuits subjccc to resistor degradation over time.
+ 25u{t) V, /? = 25
424
ijt)
fY Y \ h ^
Figure P9.32
(a)
(b)
(c)
i,(t)
(d)
vJt)
Figure P9.35
fY Y \ k
L
+
,Vc(t)
(t)
Figure P9.36
35. For the circuit of Figure P9.35,
=
-0.5M + 2u{t) A.
37. For the circuit of Figure P9.37, R^= 5 0.^
(a) I f /? = 2 Q, Z = 1 H, and C = 8 mF, = 2 0 a , C = 2.5 mF, L = 0.25 H, and v j t ) =
find and plot the zero-input, zero- 20(^) - 20u(t-7) V, where T = 0.25 sec.
state, and complete responses of v^{t)
(a) Find the zero-input, zero-state, and
and ij\t) for / > 0. Identify the tran
complete responses of v^t) for f > 0.
sient and steady-state parts of the
Plot the complete response for 0 < r <
complete response.
0.25 sec.
(b) Repeat part (a) for R= 22.5 Q.
(b)
li')
complete responses o f
200 n .
for t > 0.
= 50
and Rj =
0.25 sec.
Q, I = 4 H, and C
(b)
for
= 50 Q, /?j = 200
Q, C = 0.05 mF, L = 0.5 H , and v j t ) =
- 5 0 (- r ) + 50k(?) - 50(r - 7) V, where T =
38. Repeat Problem 37 for
I V ,(t)
Figure P9.43
4 mF.
(a)
/^(0 ) = 10 mA.
(a)
(b)
= 40(-^) V.
for t> 0 . Plot
(b)
(f) =
response,
t> 0 .
where /?j| = 16
R^2 = 32 Q, I = 0.4
H, and C = 4 mF.
w
(a)
'n- ^
Vw>
i M = - 0 .5 ( - / ) A and v,y{t) =
24(-^) V. Find the response,
for
and Rj
426
r> 0.
= 10 V in Figure
Find
(b)
.,(0 ^ ).
Compute
t'(-(0"),
/^(O"^),
/^(0^),
and
< 5 msec.
(c)
Compute
Compute
answer for
< 5 msec.
(e)
Compute
for t >
< 5 msec.
(d)
<t
for
< 5 msec.
(0
If
(d)
= - 2 0 (- f) + 20u(r) V, find
for / > 0.
-5 ti(r-
for t > 0
= 4 Cl, Rj = 4 Q,
1 = 0.2 H, a n d C = 0 .2 F .
10V
(a)
Find
(b)
(c)
and
^1/
(d)
(e)
(a)
(b)
(g)
(c)
(d)
(h)
(f)
(e)
(f)
t> 0.
t> 0.
= 2 2, /?2 = 2 Q, Z = 0 .4 H, and C =
0.1 E
(a)
(b)
C = 0.6 2 5 mF with
= 300u(f) mA, Vq =
50 V, and /?3 = 20 Q.
Find V(iO~) and y^^O'*).
(a)
equation.
(b)
(c)
(d)
If
^>
0, Hint: After finding the initial volt
(e)
(f)
for
If
for ^ > 0.
(d)
If
= -2u{-t) + lu{t) - l u { t - 1) A,
for ^ > 0.
'S->
ijt)
d)
Figure P9.57
58. Repeat Problem 57, except find
V^
(b)
(c)
59. Repeat Problem 57 with
W
= 80 2, R2
2 0 a , L = 10 H , and C = 1/240 E
(d)
\2H
(e)
the derivative o f
response
hand side.
40 fi
40 0
Figure P9.62
GENERAL SECOND-ORDER
CIRCUITS
P9.65.
(a)
(b)
If
= R^ = 50 k n .
(a )
(b)
10 .
(c)
P9.66.
(a)
Compute
=
0.
= u{t) A.
is the
429
(a)
(b)
0.5 0
2F
0.5 n
0.5 0
2F
Figure P9.67
Figure P9.69
68. In the circuit of Figure P 9.68, the voltagecontrolled voltage source has a gain A > 0 . Find
the ranges of
damped,
(b)
(c)
critically
-R .
V,F
120
Figure P9.70
so
'<N->
o
. The fol
sin(C0f + 0)
/?, = 500 + 1 0 0 ( 7 ^ -0 .0 1 )
(a)
(b)
(c)
430
and
C=0.1pF.
(a)
(b)
If
simulation
(e.g.,
(a)
(b )
432
Because o f the small capacitances, on the order o f picofarads, associated with the quartz
(diaphragm) capacitor, the bridge circuit is driven by an ac source and is called an ac bridge.
Driving the bridge by an ac source moves its analysis outside the realm of the dc and step response
techniques studied in earlier chapters. New methods o f analysis, such as phasor analysis, are nec
essary. Phasor methods, the primary focus o f this chapter, allow us to analrze capacitive and induc
tive circuits excited by sinusoidal (ac) inputs. In particular, phasor techniques permit us to anal)'ze
an ac bridge circuit. Although the analysis ot the pressure sensor shown here is beyond the scope
o f this text, the chapter will end with a simplified pressure sensor circuit based on the one shown.
CHAPTER OBJECTIVES
1.
Review and elaborate on the basic arithmetic and essential properties o f complex num
bers pertinent to sinusoidal steady-state analysis o f circuits.
2.
3.
4.
Develop two complementary techniques for computing the response o f simple RL, RC,
and RLC circuits excited by sinusoidal inputs and modeled by differential equations.
Define the notion o f a (complex) phasor for representing sinusoidal currents and voltages
in a circuit.
Using the notion ol phasor, introduce the notions of impedance, admittance, and a gen
eralized Ohms law lor two-terminal circuit elements having phasor currents and voltages.
5.
Utilizing the methods o f nodal and loop analysis and the nerwork theorems o f Chapters
5 and 6, analyze passive and op amp circuits by the phasor method.
6.
Introduce the notion o f frequency response for linear circuits, i.e., investigate the behav
ior o f a circuit driven by a sinusoid as its frequency ranges over a given band.
SECTION HEADINGS
1.
2.
3.
Introduction
Brief Review o f Complex Numbers
Naive Technique for Computing the Sinusoidal Steady State
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
Problems
-133
1. INTRO DUCTION
Perhaps you have experienced the bouncing motion of a car with broken shock absorbers or
watched the (mechanical) oscillations o f a sw'inging pendulum. These motions reflect the sinu
soidal and damped sinusoidal oscillations in
tic equation, as detailed in Chapter 9. In this chapter we allow sources with sinusoidal forcing
functions (such as such as
cos(ov + 0) or perhaps
sinusoidal responses regardless o f the root locations o f the characteristic equation. A sinusoidal
voltage source models the voltage from the ubiquitous wall outlet.
If one hooks up an oscilloscope to measure a voltage in a linear circuit driven by sources with sinu
soidal values, the voltage may not look sinusoidal at first. However, if the circuit is stable, after a
sufficiently long period o f time the screen o f the scope will trace our a sinusoidal waveform. (Here
stable means that any zero-input response consists o f decaying exponentials or exponentially
decreasing sinusoids.) The eventual sinusoidal behavior is not immediately apparent because at
startup, stable circuits exhibit a transient response. Transient means that the circuit response is
transitioning for example, from an initial voltage or current value to another constant value.
Flickering lights during a thunderstorm illustrate the phenomenon o f transient behavior: light
ning may have struck a transmission line or pole, causing the power system to waver briefly from
its nominal behavior.
Because sinusoidal excitations and sinusoidal responses are so common, their study falls under the
heading o f sinusoidal steady-state (SSS) analysis. Here sinusoidal means that source excita
tions have the form
we take
+ 0) =
sin(to/
cos(tof + 0 - 7t/2). Steady state mean that all transient behavior o f the stable circuit has
died out, i.e., decayed to zero. Observe that every sinusoidal waveform is periodic with angular
argument (to/ + 0). In terms o f angle, each cycle o f the waveform traverses 2 n radians. In terms of
time, each cycle covers a time interval o f T = 27r/o) seconds, called the period o f the waveform.
The number o f cycles contained in 1 second is called the frequency o f the sinusoidal waveform
and is denoted by / T h e unit for/is the herrz (Hz), meaning cycles per second. The quantity
OJ, which specifies the variation o f the angular argument (tor + 0) in 1 second, is called the angu
lar frequency o f the w^aveform. The unit o f (O is radians per second (rad/sec). From these defini
tio n s,/ = 1/7'= ti)/27i and o) = 2 k /
Stable circuits driven by sinusoidal excitations produce sinusoidal voltages and currents, as illus
trated in Figure 10.1. The output excitation in Figure 10.1 has the general form
to distinguish it from the input excitation,
cos((0/ + (j))
434
CO
FlGURl'l 10.1 Graphical illustration ot steady-state sinusoidal linear circuit behavior. (K^and
could just as well be
and
or any combination thereof.) Note that 0 and (}) are often different and
cos((j)) cos(tor) -
= A cos(a)f) + B sin(tor)
where A =
obtains
cos(({)) and B =
sin(i|)) sin(cor)
( 10 . 1)
V, =
4 3 ^
n ------ ^
+ B-
(10.2a)
(10.2b)
In using equation 10.2b it is important to adjust the resulting angle for the proper quadrant o f
the complex plane. Equations 10.1 and 10.2 turn out to be useful in developing a conceptually
simple, although naive, technique for computing the steady-state response using a differential
equation model o f the circuit, as explained in section 3.
Sinusoidal steady-state (SSS) analysis o f circuits draws its importance from several areas. The
analysis o f power systems normally occurs in the steady state where voltages and currents are sinu
soidal. Music is a rhythmic blend o f different notes. Mathematically, a musical (voltage) signal can
be decomposed into a sum o f sinusoidal voltages o f different frequencies. The analysis o f a sound
system typically builds around the steady-state behavior o f the microphone, the amplifier, and the
loudspeakers driven by sinusoidal excitations whose frequency varies from around 40 Hz to 20
kHz. Indeed, almost any form o f speech or music transmission requires an understanding o f
steady-state circuit behavior. There are many other areas o f applicability.
This chapter will introduce three techniques for computing the SSS response. The first two, some
what naive, approaches map out a natural motivation and path to the third, ver)' powerful tech
nique o f phasor analysis. Phasor analysis builds on the arithmetic o f complex numbers and the
basic circuit principles studied thus far. To set the stage for phasor analysis, section 2 reviews the
necessary basics o f complex number arithmetic. O f course, the student is assumed to have stud
ied complex numbers in high school and in prerequisite calculus courses.
part o f 2 j, denoted hy a = Re[z,]. The real number b is the im aginary part o f z^, denoted by b
= Im[ 2 j]. It is simple to verify that
= Re ^1 ^ i L i L
2
and
b = Im -1
where
2y
|2 ,|, satisfies
|z|P =
+ Ip-
36
The number Z| =
tation o f 2 p called polar form or polar coordinates, follows from the simple geometry illustrated
in Figure 10.2.
and
t/ = cos(0) + ;s in ( 0 )
(10.3)
is the famous Euler identity. The Euler identity can be demonstrated by writing the Taylor series
for
and recognizing it as the sum of the Taylor series for cos(0) added to j times the Taylor
series for sin(0). Note that the symbol Z. has two meanings, depending on the context o f its use:
(1) L z means angle o f the complex number z, and (2)pZ. 0 means the complex number whose
magnitude is p and whose angle is 0. The properties o f the exponential immediately imply that
(10.4)
/T
= -1 - j and Zj = \ + j-
r ; /4S'
where ti/6 has units o f radians; for example, n rad equals 180. Find the real
43*
3. Show by direct computation that since eJ^\ = cos(0j) + j sin(0j) and eJ^i = cos( 02) + j sin(f)2),
then
= cos(0,+ 0-,) + j sin(0,+ (),) =
With these simple definitions, the product oF t^vo complex numbers z-^ = a +jb =
and z-y
~ P 1P 2
^2^
+36.87)
(b)
(c)
Compute ZjZ2 in polar coordinates.
A N SW ERS: 2.8284^'/''^", 6.3246^>-/>"''^- *-^, - 1 6 + ;8 , 17.8885^->''^-^-'''-'
Similarly, in rectangular coordinates the arithmetic for the division o f two complex numbers is
Cl _ a + jb _ {a + jb ){c - jd ) _ {a + jb ){c - jd )
Z2
c + jd
ic + j d ) { c - j d )
{ac + bd) + j ( b c - a d )
~
^2
c- + d
c -+ d ~
-i38
il- =
=-
Z2
) = -Hi-cos(6, - 0 2 ) + 7 sin(6| - G j)
P2
P2
(J.i
/ 0 . 4 , 0 . 4 4 7 2 ^ ' / "*^^'
O f particular concern in this chapter are equations involving mixed representations o f complex
numbers. For example, suppose an unknown complex number z =
{a + jb) = f + j d
Then dividing through by
+ jb yields
c + jd
a + jh
Since Vis the magnitude o f the complex number on the right-hand side o f the equal sign, it fol
lows that
+
_ V c-+ t/ ~
V ^+/r
Here we have used the fact that a complex number that is the ratio o f two other complex num
bers has a magnitude equal to the ratio o f the magnitudes. To determine the angle 0, one uses the
property that 0 equals the angle o f the complex number in the numerator minus the angle o f the
complex number in the denominator,
-1
0 = Z (c + jd ) - Z {a + jb )= tan'
tan
-ta n
-I
= 2 ,/z,.
^ a n d
Factoring Ve-^ out to the left and dividing by (-8 + j6) yields
- 8 + j6
Hence. K= 2, (}) = - 9 8 .1 3 and
v{t) =
satisfy some specific algebraic or differential equation. W hen this is the case, it is possible to use
the differential equation to find values for V and (j). The next two examples illustrate this strategy.
v{t)
satisfies the differential equation
dr
+ 2 - + 2v = 10e-'< '
dl
Since the function v{t) must satisfy the differential equation, the first step is to substitute into the
differential equation. Substituting
derivatives yields
The
2 - ( 0 " + y2 (0
-2 + y4
440
is changed to
' HI.2)
The techniques of circuit analysis in this chapter will often require complex number arithmetic.
The voltages and currents o f practical interest are always real. T he complex arithmetic is a short
cut to computing real voltages and currents. The real quantities are obtained by taking the real
part o f the complex number or complex function. The various manipulations depend on some
general properties related to the real part o f complex numbers.
Property 10.1. Re[Z| + Zj\ = Re[z,] + Rel22j.
This property has a particularly nice application to summing trigonometric waveforms. Let v^{t)
= cos(oj/ + 55) and r-y(t) = 10 sin(d)/ - 3 0 ) = 10 cosCd)/ - 120). Note that a - 9 0 shift con
verts the sine to a cosine. Hence,
?'l(r) +
+ Ret 10^-^^^'^^
= Rek>("*'"
= Relt-''^^^
to the left
= a jR e [ 2 j] + a,Re[z-,]
which is a linearity property for complex numbers with multiplication by real scalars. 'I'he next
property, which underpins the techniques of this chapter, defines how differentiation can be inter
changed with the operation Re[-].
= Re
dt
Exercise. Find Re
(10 +
cit
Our fourth property tells us the conditions for the equality o f two complex-valued time functions.
Property 10.4. For all possibly comple.x numbers A and B, Re[/lf-^^'^T =
only
A = B.
Taken together, the preceding properties imply a fifth, very important propert)'. Here note that a
complex exponential is sometimes referred to as a complex sinusoid.
Property 10.5. I he sum of any number o f (1) complex exponentials, say AjeJ^'^, or (2) derivatives
o f any order o f complex exponentials o f the same frequency co, or (3) indefinite integrals of any
order o f a complex exponential o f the same frequency O), is a complex exponential o f the same fre
quency (1).
This property is another foundation stone on which the phasor analysis o f this chapter builds.
Table 10.1 summari?,es the properties o f complex numbers.
TABLE 10.1 Summar}' of Properties o f Complex Numbers
cos((jL)r + 0) = A cos(cor) +
B sin(co^)
= V-4- +
, ({) = tan"'
Euler identit)'
Linearity property
Differentiation propert)'
Equality propert)'
dt
dt^
+ 2
= Re
Single-frequency propert)'
442
cit
+ ax{t) = K^cos(co/+ 0 )
(10.5a)
df
(10.5b)
dt
V(\t).
Property 10.5 guarantees that the sum o f any number o f cosines or derivatives o f any order o f
cosines o f the same frequenc)' OJ is a cosine o f the same frequency O). Hence, the circuit response
x{t) in equations 10.5 has a steady-state cosine form o f frequency O). Further, the scaled sum ofx(f)
and its derivatives on the left-hand side o f each differential equation 10.5 must equal V^cos(o)/+0),
the input excitation. This also implies that the steady-state circuit response, x{t), is a cosine o f the
same frequency as the input, but not necessarily the same magnitude or phase. We conclude that
^(^) =
cos(co/+(j)) = A cos(tor) + B sin(to/). The SSS response is then specified upon finding A
and B. The following example illustrates this calculation.
E X A M PLE 10.4. Let the source excitation to the circuit o f Figure 10.3 be
Compute the SSS response /jr(r).
______________
i^lt)
i,
F I G U R E 10.3 Parallel
RL circuit
= /^cos(o)^).
So
443
lu tio n
Step 1. Determine the differential equation model o f the circuit. From KCL applied to the top node
o f the circuit, i jt ) = i^(t) + //_(/) Sincc the resistor and inductor voltages coincide, the t/-i rela
tionship o f the inductor implies that the inductor current satisfies the difTerential equation
( 10.6)
which has the form o f equation 10.5a.
Step 2. Determine the form o f the response. Since the input is a cosine wave, the SSS response will
have the sinusoidal form
(10.7)
Step 3. Substitute the form o f the response {equation 10.7) into the differential equation 10.6.
Inserting equation 10.7 into the differential equation 10.6 and evaluating the derivatives yields
L '
(It
= - 0) /\sin(coO + to
RA
RB
I-t
Step 4. Group like terms and solve for A and B. Grouping like terms leads to
r
]
sin(coO = 0
(10.8)
To determine the coefficients A and B, we evaluate equation 10.8 at two distinct time instants.
Since equation 10.8 must hold at every instant o f time, it must hold at / = 0; i.e., at /= 0,
=0
or, equivalendy,
(10.9a)
In addition, equation 10.8 must hold for t = 7r/(2to), in which case
-COA + - 5 = 0
( 1 0 .9 b )
444
CO/?L/
B=
R~ + L^(0-'
/?- + Lco
Il (0 = ------ ^
R~ + L~0)~
cos(o3 0 +
R~ +
O^RLL
^
SI n(co /)
where
= - tan *
(10.10b)
is adjusted to reflect the proper quadrant o f the complex plane.
This example has illustrated a procedure for finding the SSS response o f a circuit. Step 1 is to sub
stitute an assumed sinusoidal response form, such asy4cos(cor) + Bs\n{a)t), having unspecified con
stants A and B, into the differential equation and evaluate all derivatives. Step 2 is to group like
terms, and step 3 is to compute the constants A and B. After finding A and B, one computes the
magnitude,
The next section offers an alternative approach. Using complex excitation signals o f the form
y^eJ^Mt +0)^
computes
and (j) by a more direct route.
and
cos(cor + cj)) = /I
respectively, with
out any penalt)'. To recover the actual real-valued responses, we simply take the real parts o f the
complex quantities. Again this is justified by properties 10.1 through 10.5. This process o f sub
stitution and subsequent taking o f real parts actually simplifies the calculations developed in sec
tion 3, because of the simple differential and multiplicative properties o f the exponential function.
The following example illustrates a more efficient calculation o f the steady-state response using
complex exponentials.
vn
EX A M P L E 10.5. For the series RC circuir o f Figure 10.4 let v^it) = K^cos(cof). Compute the
steady-state response
( 10. 11)
at
Step 2. Substitute complex forms o f the input and response into the differential equation. If v^{t) were
to be equal to the complex exponential
if vj^t) =
However,
terms, factoring V,e^^ out to the left, and dividing through by {jixiRC +
1), we obtain
1 + jOdRC
( 10 . 12)
Vl + ( o V c '
(10.13a)
(10.13b)
Step 4. Determine the steady-state response. Using equations 10.13 the desired response is comput
ed by taking real parts:
4 i6
\'cU) = Re
(10.14)
Ks
Vl + (0/e-C^
In deriving the relationship 10.12 from the ciifterential equation 10.11, we utilized a complex exponen
tial ftinction as the circuit input. A complex exponentid input is not a signal that cm be generated in the
laborator)'. Nc-vertheless, it is often used in advanced circuit theor)' to simplift' the derivation of many
important results, as was done in the preceding example. If one does not mind a more lengthy derivation,
then the s;imc result (equations 10.12 through 10.14) am be obtained without the flaitious complex
exjxjnentid excitation. For example, let the voltage source in Figure 10.4 represent a reiil signal source
v^t) =
cosiMt + (j)) =
/ec ( rc v...e
+ Re
= Re
Making use of properties 10.2 and 10.3, move the position o f the operator Re[] outside the first
term to obtain
Re
+ Re
= Re
(10.15)
By property 10.4, equation 10.15 holds if and only if
+ <l>) =
This is precisely the equation following equation 10.11 that leads to equations 10.12, 10.13, and
finally 10.14.
As we can see, the use o f complex exponentials does indeed lead to a more direct calculation of
the SSS response. However, this method and the method o f section 3 require a difFerential equa-
44 /
tion model o f the circuit. For circuits with multiple sources, dependent sources, and many inter
connections o f circuit elements, finding the differential equation model is ohen a nontrivial task.
In the next section we eliminate the need to find a differential equation model o f the circuit by
introducing the phasor concept.
is shorthand iov
In turn, if to is
plex number A/L^ can represent a sinu.soidal function A cos(ojf + (})), whenever (O is known.
Complex number representations that denote sinusoidal signals at a fixed frequency are called
phasors. A phasor vo\iz^c or current will be denoted by a boldface capital letter. A typical voltage
phasor is V =
25cos((Of + 4 5 ) has the phasor representation I = 25Z-45. The voltage v{t) = - 1 5 sin(tor + 3 0 )
= 15 cos(w^ + 120) has the phasor representation V = 15/-120.
As all voltages and currents satisfy KVL and KCL, respectively, one might expect phasor voltages
and currents to do likewise. This is not patendy clear. The following simple example demonstrates
why this is true for KCL.
Consider the circuit node drawn in Figure 10.5.
- ijit) +
= 10 cos(tor) - 5.043 cos(cor + 7 .5 2 ) + 8 cos(cof- 9 0 )
Using trigonometric identities or property^ 10.1 to combine terms on the right-hand side leads to
i^it) =
1 0 c o s (tO f- 6 0 )
4 <S
10Z-60O =
= I, - I 2 + I3
(10.16)
It is the properties o f complex numbers and the fact that an equation is true for all t that guaran
tee that phasors satisfy KCL. Although not general, the argument is sufficient for our present ped
agogical purpose. A similar argument implies that phasor voltages satisf}' KVL, as illustrated by
the following example.
EXA M PLE 10.6. Determine the voltage across the resistor in the circuit of Figure 10.6 using the
phasor concept.
Vj(t) = 19.68 sin(a)t
152.8)
+ . --------------
44^)
o l u t io n
Firsr note that 19.68 sinltDr+l 52.8) = 19.68 cos(co/+152.890") = 19.68 cos(cor+62.8). From KVL,
= 5Z-30 V, and
= Rc[5r>(" ' " ^0")] = 5 co${o)t + 300) V
Exercise. In Figure 10.6, suppose yj(r) = 10 cos(cor) V, v-^ (/) = 10 co s((o r- 0.5tc) V, and
V3(/)= I 0 V 2 cos(cOf - 0.25ti ) . Find the phasorV^and then
ANSWKR:
= 20 - /20.
Given that phasor voltages and currents satisfy K\^L and KCL, respectively, it is possible to devel
op phasor O hm s law-like relationships for resistors, capacitors, and inductors operating in the
SSS. This would allow us to do SSS circuit analysis with techniques similar to resistive dc analy
sis. The next section takes up this thread by introducing the notion of (phasor) impedance.
for a capacitor, and Z/(/co) for an inductor. The fact that the phasor voltage V is a
function Z(/co) times a phasor current I indicates a clear kinship with Ohms law for resistors. Indeed
the unit o f impedance is the ohm because it is the ratio of phasor voltage to phasor current. The
impedance Z(/cij) explicitly shows that the relationship is potentially frequency dependent.
450
The derivation o f these elementary impedance concepts will build on the assumption that all volt
ages and currents are complex sinusoids o f the same frequency represented by complex phasors.
This is permissible because real sinusoids can be recovered from complex sinusoids simply by tak
ing the real part. To this end consider the resistive circuit o f Figure 10.7a.
V (t) =
i,(t) =
j
0j(u)t + 6)
(b)
(a)
vjiit) =
In terms o f the phasors
= RIj^
+ 0)
= 1^-^^^ and
= /?
= Zpfjia) \j^
(10.17)
where
= Rlj^ Z.0. If
cos(u)r + 0) = Kq[
This phasor
relationship restates Ohms law for complex excitations. The distinctiveness o f phasors comes with
their application to inductors and capacitors.
Now consider the inductor circuits o f Figure 10.8. Assume the circuit o f Figure 10.8a is in the
steady state.
i jt ) =
Remainder
Remalnder
of circuit
of circuit
(a)
-I-
V = jcoL 1^
(b)
FIGURE 10.8 (a) Inductor having complex exponential voltage and current, (b)
Phasor relationship of (a).
/is:
The complex current and voltage associated with the inductor are, respectively, /^(f) =
and v^it) =
yields
j = jcoz.
Canceling out
on both sides yields
the relationship is
and
( 10. 18)
in which case the inductor Impedance is derived as Z^(/co) = p iL . The inductor impedance clear
ly depends on the value ol the radian frequency CO. Specifically, if U) = 0, then the impedance o f
the inductor is 0, i.e., in SSS the inductor looks like a short circuit to dc excitations. If co = oo, the
impedance is infinite, i.e., in the steady state the inductor looks like an open circuit to signals o f
very high frequency.
Equation 10.18 exhibits a frequency-dependent Ohms law relationship for the inductor. From the
properties o f the product o f two complex numbers, the polar form o f the voltage phasor is
\ l = (yco)I^ = (03/./,) ^ ( 0 + 9 0 )
Hence if
/^(/) =
= /, cos(tor + 0) A
then
From this relationship one sees that the voltage phase leads the current phase bv 9 0 . Equivalently,
one can say that the current lags the voltage by 9 0. This leading and lagging takes on a more con
crete meaning when one views phasors as vectors in the complex plane, as per Figure 10.9, which
shows that the voltage phasor o f the inductor always leads the current phasor by 9 0 .
452
The capacitor has a similar impedance relationship, derived as follows. Assume the circuit o f
Figure 10.1 Oa is in the steady state.
C'
Remainder
of
Remainder
circuit
of
circuit
(a)
J
(b)
FIGURF. 10.10 (a) Capacitor having complex exponential voltage and current.
(b) Phasor relationship o f (a).
The complex current and voltage associated with the capacitor are, respectively, i^^t) =
and V(^t) =
yields
Vce 7(o)r+<!>)
Canceling out
I^=yo)C V ^
or, equivalently.
- i c - ^ cO ^ ^ )Ic
ycoC
(10.19)
Equation 10.19 defines the capacitor impedance as Z(^j(a) = l/(/a)Q. if co = 0, the impedance
o f the capacitor is infinite in magnitude. This means that in SSS the capacitor looks like an open
circuit to dc signals. On the other hand, if OJ = co, then the capacitor has zero impedance and looks
like a short circuit to large frequencies.
4 S3
Vr =
70) C I c
=-
t:
(oC
- 90)
(10.20)
Equation 10.20 has a vector interpretation in the complex plane, as shown in Figure 10.11
Imaginary axis
F ' l C l J R l i 10, 1 1
Diagram of capacitor voltage and current phasors where the voltage phasor lags the
current phasor by 90.
The diagram o f Figure 1 0 .1 1 indicates that the capacitor voltage lags the capacitor current phasor
by 9 0 or that the capacitor current leads the capacitor voltage by 9 0 , which is the opposite o f
the case for the inductor.
ycoL ^
and that
V/
i r ( t ) = ^COS(Cl)/ + 0 - 90)
(OL
i(^t) =
+ ^ + 90)
iiCt)
ijt )
V ,(t) =
V ,(t) =
\J
\J
0 j((i)t + 0)
g j(( t + 0)
(a)
(b)
I-ICURE 10.12 (a) inductor driven by vohagc source, (b) Capacitor driven by voltage source.
Recall that resistance has a reciprocal counterpart, conductance. Likewise, impedance has a recip
rocal counterpart, admittance. Admittance has units o f siemens, S, as does conductance, fh e
admittance, denoted by l^yto), associated with an impedance, Z(/co), is defined by the inverse rela
tionship
K(./co) =
( 10.21)
ZOCO)
provided Z{j(M) is not equal to zero ever)^vhere. What this means is that the phasor i-v relation
ship o f a resistor, capacitor, and inductor satisfies an equation o f the form I = K(y(o)V. Hence, the
admittances o f the resistor, inductor, and capacitor are respectively given by
>"/eO'w) = - ,
r^(./co) =
./(oL
Kc-(yco) = ycoC
( 10.22)
The impedance and admittance relationships o f the resistor, capacitor, and inductor are summa
rized in table 10.2.
TABLE 10.2 Summary of Impedance and Admittance Relationships for Resistor,
Capacitor, and Inductor
Impedance
Admittance
2/?(y(0) - R
K^(;o)) =
yV-(./co) = ycoC
j(oC
JY Y V
Z Lijoi) = j(i)L
j(i)L
In the next section the notion o f impedance is applied to an arbitrar)' two-terminal network. This
generalization will allow us to consider the impedance and admittance o f interconnections o f
capacitors, inductors, resistors, and dependent sources.
Z S j^ ) = ^
= R + jX
(10.23a)
1 in
O+
Two
Terminal
Circuit
Z (j(o)orY,(j(o)
FIGURE 10.13 Two-terminal device with phasor voltage \ p h a s o r current 1^^^,
and input impcdance Zy^^(/co).
Because impedance is the ratio o f phasor voltage to phasor current, its unit is the ohm. Inverting
the relationship o f equation 10.23a defines the adm ittance o f a two-terminal device as the ratio
o f phasor current to phasor voltage, i.e..
I;
(10.23b)
cd,
As an example, the impedance o f an inductor is jwL and its admittance is \/(J(.oL). Historically,
impedance and admittance were first defined as per equation 10.23. However, with the wide
spread use and utility o f the Laplace transform (Chapter 12) in the past several decades, imped
0 6
ance and admittance have become understood as much broader and more useful concepts than
the steady-state presumptions o f equation 10.23, as set forth in Chapter 13.
In general, admittances and impedances are rational functions with real coefficients o f the com
plex variable Ju). At each d) the impedance and the admittance are generally complex numbers.
Since a complex number has a real part and an imaginary part, we can further classify the real and
imaginary parts o f an impedance or an admittance. For an impedance Z(yw) the expression
lm[Z(/co)] = X is called the reactance o f the two-terminal element, while Re[Z(/(o)] = R refers to
its resistance. 1urther, for an admittance Vijto), Im[K(/(o)j = B is called the susceptance of the
two-terminal device whereas Re[)1[yoj)] = G is referred to as the con d u aan ce. These definitions
are summari/,cd in Table 10.3.
I'ABIJ-. 10.3 Summary Definitions of Various Terms
Admittance
Impedance
V/
Y( jo) ) = ^
Z{J(}^) = = R + jX
I in
V/
= G + jB
Resistance
Reactance
Conductance
Susceptance
R = Re[Z(yw)]
X = Im[Z(;to)l
6 = R e [n / o )]
im [K(>j)]
= ^] +
o f two devices in
I.e.,
^in (
= ^ = Z, ( ./O)) + Z 2 (./(O)
*iit
(10.24)
This simple derivation has another consequence: given Zy^j(yco) = Z,(/to) + Z 2 (/co) and the fact that
Z/O )
'
Z,(./(D) + Z2(7(0)
V;.,
(10.25)
Kquations 10.24 and 10.25 are consistent with our early development o f series and parallel resist
ance.
4V
Y,(jco)
V.
V.
V,
oYJjco)
(b)
FIGURE 10.14 (a) Two impcclanccs in series, (b) Two admittances in parallel.
Exercises. 1. Duplicate the derivation ol" equation 10.23 for three impedances in .series.
2. Derive a formula for voltage division when there are three impedances in series.
v,
v,
V,
V,
since
^1 (./ ) = f
V|
Y2 ijc o ) = ^
Vo
we conclude that
(10.26)
Exercises. 1. Duplicate the derivation o f equation 10.26 for three admittances in parallel, i.e.,
show that )^y(/co) = Kj(/o)) +
2. Show that the equivalent impedance o f two devices, Z,(/co) and Z t(/co), in parallel is given by
(10.27)
3. Show that the equivalent admittance o f two devices, Kjlyco) and Y-,{ji.o), in series is given by
458
>U/co)>'2(yw)
K|aco) + K ,0 (o )
4. Show that the admittance o f two capacitors, Cj and C^, in series is yCO
(10.28)
C 1C 2
C\ + C-)
5. Show that the impedance o f two inductors, Z.j and Z.^, in parallel is /O)
Lt + Lo
Now the derivation o f equation 10.26 leads to a current division formula as follows. Since
= Kj(/co) +
and since
sion formula,
I; =
'
n-(7(0)
K ,(7co)+ r2(y(o)
I,
(10.29)
Since devices represented by impedances or admittances must satisfy KVL and KCL in terms of their
phasor voltages and currents, and since each device so represented satisfies a generalized Ohms law, i.e.,
V = Z(;o))I
or
I =
it follows that impedances can be 7nauipiilated in the same manner as resistances, and admittances in the
same manner as conductances. The voltage division formula o f equation 10.25 and the current divi
sion formula o f equation 10.27 illustrate this fact. Example 10.7 further clarifies these statements.
L|
L.
4. Find ^ in terms o f
Cj + Ct+C^
L,
for each circuit in Figure 10.15.
I
O-
+
V
(a)
(b)
FIGURE 10 . 15 . (a) Set of three parallel inductors, (b) Set of three parallel capacitors.
4S9
EXA M PLE 10.7. For the circuit o f Figure 10.16, compute the input impedance ^/(/Co) when co
= 500 rad/sec.
o l u t io n
As shown in Figure 10.16, Z^-^^(/500) can be seen as the sum o f three impedances,
+ ^2 + Zy
A = j 5 0 0 x 0 .0 0 5 -
5 00
0 .0 0 0 4 )
= - j 2 .5 Q
Step 2. Compute Z-) = MYj. From the propert)' that parallel admittances add and series imped
ances add.
Y2 = y'500 X 0.0 0 0 2 -h
10 + ( 1 0 -H7 IO)
= JO. 1 -f 0 .0 4 - jO .02 = 0 .0 4
H en ce,Z 2= 1/^2 = 5 - ; 1 0 a
Step 3. Compute Z 3 = 1/
Here
Hence, Z 3 = 5 + J5
Step 4. Compute Z-^^. Adding the three impedances together yields
Z. = Z j + Z 2 + Z 3 = - ;2 .5 + 5 - 7 IO + 5 +75 = 10 - 77.5 Q
= 12.5 Z - 3 6 .8 7 Q
y0.08
160
Calculations performed in this example are most easily done with an advanced calculator or in
M ATLAB.
For example,
in
M ATLAB
the command
for computing
is Z3
l/(sqrt(0.02)*exp(i*pi/4) - j/(50()*0.01)).
EXA M PLE 10.8. Compute the input impedance Zy^^(/a)) o f the ideal op amp circuit o f Figure
10.17.
I,
V,
The trick to solving this problem entails full use o f the ideal op amp properties discussed in
Chapter 4.
Step 1. From the properties o f an ideal op amp, from KVL, and from Ohms law,
V 2 - V , - M 3 = V,
(10.30)
This follows because the voltage across the input terminals of each ideal op amp is zero and no
current enters the + or - terminal o f each ideal op amp. This implies that
Step 2. Using the phasor voltage division formula o f equation 10.25, it follows that
R
=
T " '
R+ j(oC
or, equivalently,
v,=
1+ -
Jc^RC)
Here, o f course, because o f the idealized properties of the op amp, the voltage
the resistor R in the leftmost op amp.
(10.31)
appears across
Step 3. Writing a node equation at the inverting terminal o f the rightmost op amp yields
(10.32)
Step 4. Substituting equations 10.30 and 10.31 into equation 10.32 yields
V.
=\. +
+
jCdRC
-/?!
Equivalently,
Z i u ^ )= ^ = m ~ c
I//J
(10.33)
Equation 10.33 suggests that the op amp circuit o f Figure 10.17 can replace a grounded inductor
whose impedance is jii)L with proper choice of R and C, i.e., L = R^C. In integrated circuit tech
nology it is not possible to build a wire-wound inductor. Instead, inductors are simulated by cir
cuits such as that o f Figure 10.17.
The next section continues to develop our skill with and deepen our understanding o f the phasor
technique by computing the steady-state responses o f various circuits.
i6:
E X A M PLE 10.9. Compute the steady-state voltage V(^t) for the circuit o f Figure 10.18 when i^{t)
= cos(lOOr) A.
R=10Q
21,(t)
i ( t ) 0
C=1mF
R = 100
L = 0.1H
o l u t io n
1
Z -4 5 "
(10.34)
R
Step 2. Use equation 10.34 and voltage division on the RC part o f the cirniit to compute \ q Using
voltage division and equation 10.34, the capacitor voltage phasor is
ycoC
j(oC
Step 3. Determine V(^t). Converting the phasor
function yields
= 5
C = 0.1 F, /?2 =
463
SO LU TIO N
Step 1. FindZj^^{jl). Y,^cU'^) = +
ZrcU2) =
=-+
R
= 0 . 2 5 - ;0 .2 5 = 0 .2 5 V 2 Z - 4 5 " . Hence
j(d L
Zr,.U2) = -
= 2 v /2 Z 4 5 " = 2 + j2
= 7 g 0 9 ^ _ 3 8 .6 6
4.5-J0.5
d From step 3,
l,= ^
1 ^ 2 3 ilM iZ ! =
,5 6 ,7 Z - 3 8 .6 6 "
The next example illustrates the computation o f a Thevenin equivalent circuit with the aid o f
nodal analysis. Because impedances may be manipulated in the same manner as resistances and
admittances in the same manner as conductances, the Thevenin theorem, the source transforma
tion theorem (Chapter 5), and node and mesh analysis (Chapter 3) carr)' over directly.
46^
when a 1.2
Find theThevenin equivalent circuit, and then using theThevenin equivalent, find
Vj{t).
Step 1. Establish nodal equation. A nodal equation at the left node o f Figure 10.20 in terms o f phasors is given by
I.V = r ^ L + J^CWoc = -J^ L + y ^ o c
,/(oL
Step 2. Determine the relationship between
and
(10.36)
and
as
h = m -j)y o c
Solving for
with
= 1Z.0 yields
V^oc =
: I , = (0 .4 - y0.8)I^ = 0 .8 9 4 Z - 6 3 .4 3 V
0.5 + j
(10.38)
which is the phasor version o f Figure 10.20 with the output terminals short-circuited. Hence, the
short-circuit current phasor is
1=
4(n
Step 5. Interpret
= 0.4 Q and C = 0.3125 F. Hence, the desired Thevenin equivalent circuit (valid at O)
1.2
l.2 + ( 0 .4 - y 0 . 8 )
= (0.6 + y 0 .3 )(0 .8 9 4 Z 6 3 .4 3 )
= 0 . 6 Z - 36.87 V
Converting the load voltage phasor to its corresponding time-domain sinusoid yields
y^^(t) = 0 . 6 cos(4/^ - 3 6 . 8 7 ) V
166
j60Q
I'lG U R E 10.23 Phasor domain circuit for Example 10.12. Ail clement values indicate phasor
impedances at 100 rad/sec.
S o l u t io n
To solve this problem, it is convenient to execute a source transformation on the independent cur
rent source and to combine the impedances o f the parallel combination o f the capacitor and
inductor on the right-hand side o f the circuit. After executing these rwo manipulations, one
obtains the new circuit of Figure 10.24.
FIGURE 10.24 Phasor domain equivalent circuit to that of Figure 10.23. All element values indi
cate phasor impedances at 100 rad/sec. I denotes a phasor loop current.
For the circuit o f Figure 10.24, the indicated loop equation is
250Z.-90O = (50 - ; 2 5 ) I - 0.4(501) - ; 1 5 I = (30 - ; 4 0 ) I
Solving for I yields
I = 4 - J 3 = 5 ^ - 3 6 .8 7 A
Consequently,
467
-o-
-O
+
0.01 F
10
-o
FIG URE 10.25 RC circuit passing high-frequency content of an input signal.
S o l u t io n
= m p ))
l + iO .O lc o
/).01(0
O) =
0 and co=
oo.
1 as
CD
oo
and
468
|//(/to)|
0 as to
0 as to
oo and Z.//(/to)
9 0 as oj
0. Also, a close scrutiny o f //(/w) indicates that to = 100 rad/sec is also an important frequency.
Here H{j\00) = 0.707^^45. These values give us a prett)' good idea what the magnitude and
phase plots look like. Using a computer program. Figure 10.26a and Figure 10.26b show the exact
magnitude and phase plots. These plots are consistent with our earlier asymptotic analysis.
Frequency (rads/sec)
(a)
Frequency (rads/sec)
(b)
FIG URE 10.26 (a) Magnitude plot of frequenc)' response for Example 10.13.
(b) Phase plot of frequency response.
Do these frequency responses make sense? They should. Going back to the circuit, observe that at
to = 0, the capacitor impedance is infinite. Physically, then, in steady state, the capacitor looks like
an open circuit for dc, i.e., at zero frequenc)'. The magnitude plot bears this out. For frequencies
close
CO
zero, the capacitor approximates an open circuit and, hence, the magnitude remains small.
O n the other hand, for large frequencies, the capacitor has a very small impedance. This means
that most o f the source voltage appears across the output resistor. The gain then approximates 1,
as indicated by the magnitude plot. The frequency response o f the circuit is such that the highfrequency content o f the input signal is passed while the low-frequency content o f the input sig
nal is attenuated. Such circuits are commonly called high-pass circuits.
EXA M PLE 10.14. Investigate the frequency response o f the parallel /^Z,C circuit o f Figure 10.27.
R=10
L=0.04H
C=0.25F
LC
R
RC
. CO
jOiL
C
Inverting to obtain the input impedance yields
w
C
J_ _ ^ 2 ^ ._ o L
LC
RC
Clearly,
y'4co
I00-co-+y4co
Hence the ratio o f the output phasor to the input phasor is simply
Zy^^(yoj). Once again, co = 0 and co = oo are the first two frequencies to look at. Here Zy^,(0) =
0Z .90 and -2^,(oo) = 0Z.-90^. Also at co = 10, the impedance is real, i.e.,
0) = I . These three
points provide a rough idea o f the magnitude and phase response. Two more points are necessary
for a real .sense o f the frequency response. At what frequency or frequencies does the magnitude
drop to 0.707 o f its maximum value or when does the phase angle equal 45? This will occur
when 1100 - co^| = |4co|. This is a quadratic equation. Flowever, because o f the absolute values,
there are rvvo implicit quadratics, co" - 4co - 1 0 0 = 0 and co^ + 4co - 100 = 0. Solving using the
quadratic formula yields co = 8.2, 12.2. Since the magnitude plot is symmetric with respect to
the vertical axis (co = 0 axis), we consider only the positive values o f co. This information provides
a good idea o f the magnitude and phase plots. A computer program was used to generate the fre-
470
quenc)' response plots in Figure 10.28a (magnitude) and Figure 10.28b (phase). The magnitude
plot shows that frequencies satisfying 8,2 < O) < 12.2 are passed with little attenuation. Frequencies
outside this region are attenuated significantly. Such a characteristic is said to be o f the band-pass
type, and the corresponding circuit is a band-pass circuit.
Frequency (rads/sec)
(a)
Frequency (rads/sec)
(b)
F IG U R t 10.28 (a) Magnitude plot o f frequency response for band-pass circuit o f Figure 10.27. (b)
Phase plot of frequency response.
EX A M PLE 10.1 5 . As a final example vve consider the so-called band-reject circuit o f Figure
10.29. A band-reject circuit is the opposite o f a band-pass circuit. A band-reject circuit has a band
o f frequencies that are significantly attenuated while it passes with little to no attenuation those
frequencies outside the band. In this example our goal is to compute the magnitude and phase o f
the frequenc)' response o f the band-reject circuit o f Figure 10.29.
R =10
'
LC
t)Ul
-(0
1 0 0 - CD"
_ L _ c o 2 + /o)-5
LC
1 0 0 - 0 ) 2 +^25co
At to = 0 and
CO = o o ,
this example, to find the frequencies where |//(/to)| drops to l/ V I o f its maximum value o f 1, it
is necessary to equate the magnitudes o f the real and imaginary parts o f the denominator. This
produces two quadratics whose positive roots are to = 3 .5 0 7 8 and OJ = 28 .5 0 7 8 . At these frequen
cies the angles o f //(/w) are 45 and 45, respectively. Fhe computer-generated plots o f Figures
10.30a and 10.30b are, o f course, consistent with these quickly computed values.
472
Frequency (rads/sec)
(a)
Frequency (rads/sec)
(b)
F I G U R t 10.30 (a) Magnitude plot of frcquenc)^ response for band-reject circuit of Figure 10.29. (b)
Phase plot of frequency response.
As wc can see, a wealth o f different kinds of frequency response are obtainable by different inter
connections o f resistors, inductors, and capacitors. Historically, phasor techniques were the essen
tial tool for the analysis and design of such circuits. Nowadays, engineers ordinarily use either
M A T L A B or SPICE to obtain frequency response plots. Two examples follow where we use MATLA B , SPICE, or both to obtain the frequenc) response.
4 '3
E XA M PLE 10.16. Compute the frequency response o f the circuit o f Figure 10.31 using MATLAB and SPICE.
R=10Q
This circuit was originally analp^ed in Example 10.9. You might want to refer to that example
before proceeding.
SPICE Part. A SPIC E simulation produces the result shown in Figure 10.32.
EX1016 FreqRsp-Small Signal AC-0
+20.000
+40.000
+60.000
+80.000
Frequency (Hz)
FIGURE 10.32 SPICE plot of capacitor voltage for the circuit of Figure 10.31.
MATLAB Part. Although the analysis appears in Example 10.9, we can use MATLAB to more
easily obtain the frequency response. First define Zj(/co) = ytoA and Z-,{p)) = l/y'coC Then from
current division,
= -7 " r 7
R+
Z|(yco)
1 ,0 )
R+
Z 2 (y co )
K2( 7(0) +
21^( 700)
474
Assuming a frequency range o f 0 < co < 1000 rad/sec, the following MATLAB code will result in
a suitable magnitude frequency response plot, as shown in Figure 10.33.
L = 0.1;R = 10;C = 0.001;
w = 0:1:1000;
Z1 = j*w*L;
Y2 = jw*C;
>>IL= R./(R+Z1);
VC = 2 1L./(R *Y 2+1);
plot(w/(2*pi),abs(VC),b)
grid
xlabel(Frequency in Hz)
ylabel(Capacitor voltage (V ))
FIGURE 10.33 Magnitude plot of frequenc)^response o f capacitor voltage in the circuit of Figure
10.31. The response is of the low-pass t)'pe.
Now suppose the inductor in the circuit o f Figure 10.31 is replaced by a capacitor C, = 1 mF with
the controlling current changed to /q (0- The frequency response is easily computed with a sin
gle change to the MATLAB code, namely, Z , = 1. ./(j*w *0.001). The resulting plot shows a
band-pass characteristic, as illustrated in Figure 10.34.
47S
F IG U R E 10.34 Magnitude plot of frequency response of capacitor voltage in circuit of Figure 10.31
when inductor is replaced by a 1 mF capacitor. The response is of the band-pass type.
E X A M PLE 10.17. In Chapter 9 we investigated the Wien bridge op amp oscillator circuit,
redrawn in B2 Spice in Figure 10.35.
4" 6
Two difTerences are notable: ( !) there is a current sourcc present across the
combination,
and (2)
is now 10 kQ, as opposed to 9.5 kH in Example 9.14, forcing /?, = /?2- This means
that the characteristic equation for the circuit is
1
V*' + / ? i ' + ( = .y~ +
{RlC
R,C)
f o = - ^ = -----! = 15.92 Hz
"
2n
2 k R,C
for any initial condition on C ,. In fact one might recall that R^ < Rj causes a growing oscillation
that is limited by the saturation effects o f the op amp.
The current source, set at 1 A, is present in Figure 10.35 so that we can obtain the frequency
response cur\'e shown in Figure 10.36. In Figure 10.36 observ'e that the magnitude response peaks
at/q, as expected from the theoretical analysis. In an actual circuit, the current source would not
be present. Nevertheless, a sustained sinusoidal oscillation will occur because o f the presence o f
noise. W ithout going into the analysis, noise contains an infinite number o f frequency compo
nents, each o f which has a minute magnitude. In particular, noise contains frequency components
around / q that drive the circuit into oscillation. This is precisely what the peak in the frequency
response means: a very small (noise) voltage on Cj will cause a very large-magnitude sinusoid out
put voltage at /q. However, the presence o f nonlinearities such as saturation keep the magnitude
at an acceptable level.
+13.000
+14.000
+15.000
+16.000
+17.000
Frequency (Hz)
+18.000
+19.000
+20.000
15cos(20,000nt)V
changes as a
the pressure applied to the diaphragm and the associated capacitance. Our first task will be to
specify the relationship between the pressure applied to the diaphragm and the resulting capaci
tance. Following this, we will use nodal phasor analysis to determine the magnitude of
and finally, the relationship between pressure and the magnitude V ^ - V e
Pressure is measured in various units. Millimeters o f mercury (mm Hg) is a common standard; 1
mm Hg = 1 torr, and 760 torr = 1 atmosphere (atm), where 1 atm is the pressure o f the earths
atmosphere at sea level, which supports 76 0 mm o f mercur\' in a special measuring tube. Suppose
4 8
it has been found experimentally that the capacitance C j (in pF) varies as a function o f pressure
according to the formuhi
C 2 ( A P ) = Q ) + ^ lo g 10
= 2 6 .5 + 68 log 10
^0
(10.39)
(16i)+A P^
760
QJ
(T3
Q.
U
Change in Pressure
FIG URE 10.3H Plot of capacitancc versus pressure.
Our next task is to develop the relationship between the capacitance of^ the bridge circuit and the
magnitude o f the phasor voltage
and
= (^ 3)"^
will let Cj range as 0 < C 2 < 40 pF. Finally, co = 2h x 10"^ rad/sec. The following phasor analysis
will be done symbolically so as not to obscure the methodology.
Summing the phasor currents leaving node A leads to the phasor voltage relationship
(G j + G 2 + yujCj)v^ G-jV^ ycoCj
= Gj 15
=0
479
G JV c-
=0
G] + G 2 +
-/(oC,
-G i
G 2 + G 3 + G ,
-G j
-yco C ,
- G ,
l5 G ,'
V /
=
~^m
G i + ^ ( C | + C 2 )
Vc.
(10.40)
The matrix on the left is said to be a nodal admittance matrix. Its entries can be real or complex,
as indicated. It is nor advisable to solve such a set o f equations by hand over the range o f possible
C2 values. However, using MATLAB one can solve this matrix equation over the range 0 pF < C,
< 40 pF to produce the plot o f Figure 10.39.
c
CO
QJ
01
nj
*->
<U
cn
-o
o
cu
o
D
'E
cn
fO
10
15
20
25
30
35
40
C, in pF
V ^ as a function of capacitance.
O f course, one could measure the voltage appearing across the meter, from Figure 10.39 deter
mine the associated value o f Cj, refer to Figure 10.38 for AP, and then determine P = 760 + AP.
This is a long route. To complete our analysis, then, we need to develop the relationship between
pressure and bridge voltage. As we have the relationship between C j and APand the relationship
between C2 and |
|, it is a matter o f using equation 10.39 to derive the value o f
in
equation 10.40. This is best done with a simple MATLAB routine, which yields the plot given in
Figure 10.40.
480
>
m
>
'o
01
T3
D
'c
ro
300
400
500
600
700
900
800
1000
1100
1200
Pressure in mm Hg
I'lG U RE 10.40 Relationship herween magnitude of bridge output voltage and pressure applied to
diaphragm capacitor C-,.
An actual pressure sensor would, oFcourse, be more complex. For example, there would probably
be a difterential amplifier such as the one shown in Figure 10.41 across the terminals oFthe bridge
circuit, and this would probably drive a peak (ac) detector to determine the maximum value oF
the ac signal appearing at the output oF the dlFFerential amplifier. Fiu ther, the peak value would
probably be read by a digital voltmeter. Nevertheless, our analysis illustrates the basic principles
involved in such a measurement. OF course, one could just as easily use loop analysis to solve the
problem. This is leFt as an exercise in the problems.
kR.
-o V,
= k { i >2 - /^i).
481
= k{v-, -
, where V =
and
is used through
Phason complex number representation denoting sinusoidal signals at a fixed frequency. Bold&ce
capital letters denote phasor voltages or currents; a typical voltage phasor is V =
and a typical current phasor is I =
Polar coordinates: representation of a complex number z as p?-^, where p > 0 is the magnimde of z and
0 is the angle z makes with respea to the positive horizontal (real) axis of the complex plane.
Reactance: imaginary part of an impedance.
Real part: real part of a complex number z = a + jb for real numbers a and b, denoted by Re[], is a.
Rectangular coordinates: representation of a complex number z as coordinates in the complex
plane, i.e., zs a + jb for real numbers a and b.
Resistance: real part of a possibly complex impedance.
Sinusoidal steady-state response: response of a circuit to a sinusoidal excitation after all transient
behavior has died out. This definition presumes that the zero-input response of the cir
cuit contains only terms that have an exponential decay.
Stable circuit: circuit such that any zero-input response consists of decaying exponentials or expo
nentially decaying sinusoids.
Susceptance: imaginary part of an admittance.
Voltage division: in a series conneaion of impedances driven by a voltage source, the voltage
appearing across any one of the impedances is proportional to the ratio of the particular
impedance to the total impedance of the connection.
Zero-input response: response of the circuit when all source excitations are set to zero.
In the literature, both z and z* are used to denote the conjugate o f a complex number z. However, in matrix
arithmetic, Z* usually means the conjugate transpose o f the matrix Z. We will sometimes interchange the usage.
In MATLAB, * means multiplication and conj(Z) means conjugated. So there is some ambiguity in the usage.
^
^
483
Problems
SO LUTIO N OF DIFFEREN TIAL
EQUATIONS W ITH CO M PLEX
EXPO N EN TIALS
'
Figure P I0.3
A N SW ER S: 0 .8 sin (2500r -
3 6 .8 6 >). 80
s in (2 5 0 0 r-3 6 .8 6 )
R < v .J t )
Figure P I0.1
AN SW TR: 20 cos(400/ - jr/4) V
Figure P I0.4
AN SW ER: 1.6 cos(2500; + 36.8~) V
as a function o f
(2+j4)A
+
v(t)
V (t)
Figure P i0.2
ANS^XTR: 20 sin(250/ + 0.25k) V
Figure P i0.5
AN SW TRS: (a) 10 cos( 1OOTif - 0 .9 :" )
A. fb)
484
j20
(26+j12)V
20
(a)
Figure P i0.6
j20
(26+ jl2)V
L
20
v^(t)
(b)
e
(26 + jl2)V
Vj(t)
v,(t)
j20
20
(0
Figure P I0.7
ANSW ER; vjU) = 4 cos((./ - )0") \'
Figure P I0.9
8. Use KVL to determine the phasor voltage
10. For the circuit o f Figure P i 0.10, use KCL
4jV
(4 + ]4)V
Figure P10.8
(2-j10)A
(2+j4)A
V _y (2+
j
/ Ti
(t)
AN SW ER: - 4 V
(4 + j6) A
(4+j6)V
(5
Figure PIO.IO
-1 0 1 .3 '* A
= y8
48S
(b)
Figure P I0.14
A N SW ERS:
rad/sec.
3 0 > A. I,
(f) = 10
F, Z: = 0.1 H, R= 2.5 Q.
Find
".a/')-
phasors
r
I''
*
' r
V^,
and
"
R
JY Y V
v -
+ V. -
Figure P i0.12
13. The circuit o f Figure PI 0.13 is operating in
the sinusoidal steady state with
= 20
Figure P i0.15
C = 1 mF.
C H E C K : v ;p ) = 28.28 co s (2 5 0 0 r- 135) V or
(a)
v .p ) = 28.28 s in (2 5 0 0 f- 45") V
(b)
rent ip ) .
and
(a)
the current
(b)
= 40 sin(200^) V.
Find the phasors Ij and
Find the phasor 1^ and the associated
time function ijit).
v,(t)
Figure P I0.13
14. Find the phasor currents
Figure P I0.16
AN SW ER: (b) 1.2Z-9()'\ 1.2 sin(200r) A
and
= 0.5 H,
= 20 cos(1000r + 60) V.
V jt)
Z Jj(o )
Figure P I0.17
ANSWHR: S.831 1 cos(200r +30.96<) V
(a)
(b)
gets large?
(c)
1 mF
v jt)
lOO
0.1 H
Figure P I0.18
Figure P I0.21
A N SW FR: (a) S - p U
19. (a)
admittance is real.
+
V ,( t)
SERIES-PARALLEL
IM PED A N CE AND
A D M ITTA N CE
CA LCU LA TIO N S
20 . Consider the circuit o f Figure P I 0.20.
(a)
(b)
(c)
Figure P i0.22
If C = 0.01 F, findZ.(/100).
If Zy,^(/100) = 25j O., find the appro
priate value o f C.
which /?j = 20
/?2 = 10
Z, = 20 mH.
(a)
I f C = 0.3 mF, find K.(;-500).
(b)
(c)
If C = 0.3
487
o flK (/ 0 ))l.
(a)
Figure P I0.23
(b)
Figure P I0.25
24. Consider the circuit o f Figure P I 0.24 in
which C = 1 F. At C0= 2 rad/sec,
= 4 + jl
n.
26.
(a)
(b)
(c)
If
A N SW ER : C = 4 uF, A* = 500 Q ..
5
\'^(/) = ;= cos 500/ - 4 5 " ) V
V2
27. For a particular two-terminal device,
'jt ) Q
Z,(jw)
Figure P I0.24
AN SW ER: Z. = 2 H, /^ = 2 kQ.,
5
/(/) = - = cos 1000/ - 4 5 " ) mA
AN SW ERS: 4 Q, 0.2 H, 0 to x
25. Consider the circuits o f Figure P I 0.25 in
which R = 5 O., L = 32 mH, and C = 5 ^F.
(a)
Find
Consider
Figure
P 10.25a.
is
488
(b)
R
(c)
is 4 5 .
Figure IM0.28
-H f
C H EC K : Z.(/5000) =1 + ;2 ^
Figure P I0.31
(a)
(b)
lf/?= 1 0 Hand
10 n/2 cos(10/)
Figure P I0.29
AN SW FR: (I) = 5000 rad/sec
SERIES/PARALLEL
IMPEDANCES WITH V/l
DIVISION
^,(t)
/? =
20
Q,
Z.
H,
(b)
Figure P I0.32
and
RC
-4 5
i,(t)
= 2 A and (d =
(b)
Figure P I0.30
ANSW'F'.KS: (a)
.b; I (I
31. (a)
y20co
5+
/(I)
0.25
. 0.05 - /
(I)
5; - k !4) niA
For the circuit o f Figure P I0.31, find
the ratio
in terms of/^, C, and
CO.
0 .5
mF,
and
hS9
(a)
I^, and
v j t ) = 50 cos(4000^) mV.
when
= 2 A and OJ =
He
when
100v,(t)
v.(t)
If
= -32sin(40r) V, find
i^t) when
= 10 cos(lO'^r) mA.
Z. =
Find
If
= 10 cos{25t) V, find
V[ {t)
V({t),
''" 6
mA. Find
Figure P I0.37
AN SW TR: 0.625 mF or 0.2083 mF
38. In the circuit o f Figure PI 0.38, /?, = 50 Q,
q
= 0.625 pR
and
= 120 cos(400r)
490
phasor current
and the
values o f L and C
2Q
> j2 Q
-j2 0
NETWORK THEOREMS IN
CONJUNCTION WITH V/l
DIVISION.
(1000/-O.S3T) A
Figure P I0.44
in steady state.
Figure P i0.45
A N SW ER: 20 co s(1 0 ()0 r- 135) V
i(t)
46.
Figure P i0.42
= Rj = 2
L = 5 mH, and C =
in
j5on
-
J200Q
+ V. -
Figure P i0.46
AN SW ERS (in random order);
Fiaure P I 0.43
- 0 .8 . ./40 Q. - 1 6 cos(IO().-t O V
491
= 10 cos(2007U/)
\t) = 20 cos(2007ir +
45) V. On the other hand if i^^{t) = 0 with
/;2 (/) = 10 cos(2007if + 45) V, then v^{t) = 5
tions.
cos(200)r + 900) V
(a)
v^2 >and
(b)
Find
and
Zj
Z - ,.
Figure P I0.47
ANSWHRS: (b) r,(/) = 18.46 cos(200nr -
time function.
OA
Suppose
Again suppose
that
of
rad/sec.
W ith
the
= 0 V. The result
Find the value
ofZ,.
Now suppose R^ ^ 0. Show that the
condition o f part (a) is still valid. You
=20 a, C = 20
^iF, and
l- = 1 0 0 ^ 0
(b)
492
OA
Find
if
10 a , ^ = 5 a
= 1^^^cos(w^/), R =
and
20 Q.
when the
i jt )
-jx
Figure P I0.54
55. In the circuit o f Figure PI 0.55, assume co =
100 rad/sec, I = 40 mH, C = 5 mF, R = S Q.,
and a = A Q.. Find the Thevenin impedance
Figure P I0.51
Q, Cj = 0.2 pF,
= 20
cos(lOOf) V, find
= 1000
= 50 cos(4000f) V.
v.(t)
-OA
R,
Figure P I0.55
OB
56. (a)
in the cir
= 2 0 V 2 Z 4 5 " V, R = AQ., L = A
mH, and oj = 1000 rad/sec. Specify
the corresponding time functions.
10 cos(4000r) mA.
(b)
(0
input voltage.
493
/Y Y V
L
A and B.
'6
-OA
2R
2R
;v,(t)
Figure P i0.56
C H E C K S: 5 A, 2o v
-O B
Y.(j)
57. In the circuit o f Figure P I0.57,
= y'30 Q,
Figure P I0.59
= - ; 4 0 Q, V^, = 28 V, V^2 =
^
sinusoidal sources have been operating for a
= 40 Q., Rj =
6 ''
Figure P I0.57
z^ = y io a
z^^ = - y i o a
Calculate
^ = lo a
and
Figure P I0.58
AN SW FR: O.r. + 0.2/
59. Consider the circuit o f Figure P i 0.59 for
which C = 0.8 mF, r.(;100) = 0.01 + ;0 .0 4 S,
and v-^,(r) = 80 cos(lOOf) V.
Find R.
(a)
(b)
(c)
(d)
(e)
-
Find L.
Find
Find
At CO = 100 rad/sec, determine the
Thevenin equivalent circuit phasors
Figure P I0.60
ANSWKRS: V^. = S - /S
= -/ 0 .:5 . Z./. =
1 0 - ; 10 12
61. The circuit o f Figure P i 0.61 operates at (o
= 2 krad/sec a n d =1. 5 mS with
= 10 0 Z 0
V. Find the Thevenin and Norton equivalent
circuits seen at terminals A and B.
494
J \
l.s f
0-251
Figure P I0.61
= ;1 0 0 0 Q,
C H EC K :
'
= 25 + ;2 5 V,
= 25 + y25 mA
62. Consider the circuit o f Figure PI 0.62. If i^{t)
Figure P I0.64
and
R,
Figure P I0.62
C H EC K :
= 128 - ;3 3 7 .8 8 Cl,
1 4 0 .2 5 Z 9 5 .2 8 V
(b
R
v,(t)
Figure P i0.65
66. Inside the black box o f Figure P I0 .6 6 a
there is a two-element circuit composed o f a
resistor o f 10 ^2, capacitors, inductors, or some
combination o f these elements. A variable-fre-
Figure P i0.63
FREQUENCY RESPONSE
64. Compute the magnitude and phase func
tions o f the frequency response o f the circuit o f
Figure P I 0.64 in which L = 4 mH and C = 0.25
mF. Plot your response in MATLAB (0 < CO <
5000 rad/sec). Before sketching the responses,
cos(cor + 0) is
(b)
v,(t) =
10cos(cot)V
Black
Box
v(t)
(a)
(b)
Figure P i0.66
plots.
software program. Plot the magnitude o f V ^ V ^as a function o f Cj- Now construct a plot o f
the magnitude o f V ^ - V ^ a s a function o f pres
sure in mm Hg.
OP AMP CIRCUITS
70. (a)
Figure P I0.67
when
= sin(200r)
Compute
(c)
= cos(400r) mV,
= sin(400r) mV.
Find the phasor transfer function,
//(yco), and plot the magnitude o f the
frequency response (using iMATL^B
or the equivalent) as a function o f 03 =
496
r>
100 kfi
20 kO
80 kO
o
n
(a)
(b)
Compute
when
=
sin(400/) V for the circuit of Figure
P10.71a.
For the circuit of Figure P I 0.7 lb, find
C such that when
=sin(500/) V,
= 5 cos(500/) V. This represents
an integration of the input with gain.
Find the phasor transfer function,
and plot the magnitude of the
frequency response (using M ATLAB
or its equivalent) as a function of O) =
In fi where/is in Hz and O) in rad/sec.
71. (a)
(b)
(c)
n
o
(b)
Figure P I0.72 Leaky integrator circuits.
r s
10|j F
H e-
50 kO
He-
200 kn
+
o
+
vjt)
JL .
(a)
(b)
1kn
72. (a) If an 800 Hz sine wave of unit ampli
tude excites the leaky integrator circuit
of Figure PI0.72a, determine the
steady-state output voltage.
(b)
For the circuit of Figure P 10.72a, find
the phasor transfer function, //(/(o),
and plot the magnitude of the fi*equency response (using M ATLAB or
Figure P I0.73
its equivalent) as a function of (O =
271/where/is in Hz and O) in rad/sec.
74. For the circuit of Figure P i0.74, find the
(c) If the input to the circuit of Figure
expression for the phasor transfer function
P I 0.72b is
= cos(2000)r) V,
//(/ cd) =
Assume an ideal operational
determine the values of R and C so
amplifier. Plot the magnitude of the transfer
that v i t ) = 5cos(2000ti^ +135) V.
n
r^
o
O '
^
function as a function o f COusing M ATLAB or
^19'
0.01 mE
VO
o
o
Figure P I0.74
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
Figure P I0.75
76. For the operational amplifier circuit of
Figure P I0.76, /?j = 5 kQ, C, = 0.02 |iF, /?2 =
5 kfi, and C2 = 0.08 pF.
(a) Write two node equations and solve to
find a relationship between the output
phasor
and the input phasor
at the frequency/= 1000 Hz. Note
that the voltage from the minus termi
nal of the op amp to ground is
which equals the voltage from the plus
terminal to ground, assuming the op
amp is ideal.
(b) Repeat the calculation at/= 100 Hz
and/= 3000 Hz. What happens as the
CHAPTER OBJECTIVES
1.
2.
3.
4.
Introduce the notion o f power factor associated with a load and describe reasons and a
5.
500
SECTION HEADINGS
1.
2.
3.
4.
Introduction
Instantaneous and Average Powers
Effective Value o f a Signal and Average Power
Com plex Power and Its Com ponents: Average, Reactive, and Apparent Powers
5.
6.
7.
8. Summar}^
9. Terms and Concepts
10. Problems
1. INTRODUCTION
Chapter 1 defined the concept o f power. The following chapters were primarily devoted to the cal
culation o f voltages and currents. This does not mean that the consideration of power is o f sec
ondary importance. The very opposite is true. A homeowner pays for the energ}' used, not for volt
age and current. The integral o f power over, say, a 30-day period determines the household ener
gy consumed in a month. Hidden in the homeowners cost is an adjustment to cover the power
losses incurred in transmitting energy from the generating station to the home. Thus power con
siderations have a significant impact on everyday life.
A second reason for understanding ac power usage is safety. Each appliance, and its cord that plugs
into the wall outlet, has a maximum safe power-handling capacit)'. Misunderstanding such infor
mation and/or misusing an appliance can lead to equipment breakdown, fire, or some other lifethreatening accident.
Even for electronic equipment in which power consumption is low, such as laptops and handheld
PDAs, power consumption and, thus, battery life are important design factors. Power drainage direct
ly determines the PDAs operating time before the battery needs recharging. In fact, optimizing power
management in laptops and hybrid electric vehicles is an important research area in todays world.
In this chapter we will investigate different notions o f power in ac circuits and discuss their sig
nificance and application. The term ac circuits has a narrow meaning here. It refers to linear cir
cuits having all sinusoidal sources at the same frequency and consideration o f responses only in
steady state. The basic analysis tool is the phasor method o f Chapter 10.
501
taneous power over an interval [T*!, T^. The idea is based on the average value o f a function, say
T-,
T
h - TM r,J
Using this idea we define the average power consumed by a rwo-terminal element as shown in
Figure 11.1 over the interval [T j, Tj] as
Pave(TiJ2) =
T,
piOclt
r .-r ,
( 11. 2)
When the signal is periodic with period T, we speak o f the average power consumed by an ele
ment over the period T as
T
T
(11.3)
It is not necessary that T b e the fundamental period; the evaluation o f the integral is the same for
any integer multiple o f the fundamental period.
EXA M PLE 11.1. Compute the average power absorbed by the resistor R connected to an inde
pendent voltage source as shown in Figure 1 1.2b with the excitation shown in Figure 1 1.2a.
v.(t) f +
driving resistor R .
S{)2
So
lu t io n
Step 1. Compute the instantaneous power for 0 < f < 7q. Here
IV J
P {t) =
0 < / < 0 .5 ^
o .5 r < / < 7 ;
Step 2. Compute
Using equation 11.3 and observing that the Fundamental period is Tq, we
have
0.57n
/
>ave
p {t)d t =
rj.
TnR
6R
Exercises. 1. Suppose the sawtooth in Figure 11.2a does not drop to zero at r = 0.5 T'q , but rather
continues to increase until reaching ^ = Tq when it drops to zero and repeats. Find the average
power consumed by R.
AN SW ER:
3/^
. K?
2. Show that the average power absorbed by an R Q resistor in parallel with a Vq V dc source is
over any time interval [T'j, 7'^].
^
O f particular importance is the average power consumed by devices in the SSS assuming all exci
tations are at the same frequency, to. Consequently, all voltages and currents are sinusoids at the
sa7nefrequency. To compute the average power absorbed by a circuit element as depicted in Figure
11.1 (assuming a linear circuit), suppose v{t) =
cos(u)t + 9^^ and i(t) =
cos(a)r + 0^) . The
associated instantaneous power is
p ( f ) = v(f ) i { t ) = V, cos(co/ + e^,) X /, cos(cor + 0 ,-)
(11.4)
=
cosO , -
e,-) +
Equation 11.4 follows from the trigonometric identit)' cos(x) cos(y) = 0.5 c o s (x - y) + 0.5 cos(a' +
y). Observe that the instantaneous power o f equation 11.4 consists o f a constant term plus anoth
er component varying with time at tivice the input frequency. Figure 11.3 shows typical plots o f
/>(/), v{t), and i{t).
503
0.01
0 .005
0 .0 1 5
0.02
t in secs
FIGURE 11.3 Plots o f tit) = 10 cos(377f) A, v{t) = 2 cos(377/ + 45) V, and p{t).
Using equation 11.3 witii T = 23t/ol), and observing that the integral o f a sinusoid over any peri
od is zero, we obtain the following formula for average power in SSS:
T
_
*Pnvp =
If the two-terminal element is a resistance R, then v{t) = Ri{t) and 0^^ - 0y = 0 . It follows from
equation 11.5 that for a resistor
p
_ VmI m _ RI^m _ m
^ave,R
2
2
2R
If the two-terminal element is an inductance L, then
^aveL ~ ^
( 11.6)
c~ ^
^neans that the average power consumed by
or delivered by a capacitor or au inductor is zero. Even though an ideal capacitor and an ide;il induc
tor neither consume nor generate average power, each may absorb or deliver a large amount ot
= (/coQV^ and 0^ - 0y = - 9 0 . Hence,
504
EXA M PLE 11.2. Consider the circuit o f Figure 11.4, which consists o f a series connection o f two
(sinusoidal) voltage sourccs in parallel with a 1 Q resistor. For this investigation v^{t) =
cos(ojjr + 0|) (having fundamental period /'] = 2:t/tO|) and v-,(f) =
damental period T-, = 271/(0,). For simplicit)^ we assume that v^{t) and v-,{t) have a common peri
od o f 7 'seconds, i.e., there exist Integers ni and ;/ such that T = uT^ = niT^.
v,(t) Q
1n
< v_(t)
lu t io n
Compute the average power consumed by the 1 Q resistor. First observe that the power consumed
by the 1 Q resistor with source 1 acting alone, i.e., v-j{t) = 0, is
1
1
0
Also note that the power consumed by the 1 H resistor with source 2 acting alone, i.e., u^{t) = 0,
is
I T
, r ^
0
= v^{t) +
(^ 1 (/)
Pave = - \ V R {t)i,i{t)d t = - \
0
. By equation 11.3,
+ v'2 (r )) dr
T
^ove.l
^a\r,2 '
^ Vj (/)V2(/)f/f
0
^ave,\
^ove.l
j [ c o s ( ( ( 0 | + C O 2 )/ + (0 i + 2 ) ) + C 0 s ((0 3 , - ( O 2 ) / + (0 i - 0 2 ) ) ] ^ / ^
0
S()S
, indicating that
superposition o f average power holds. When this integral term is nonzero, superposition o f aver
age power does not hold. The next question is, under what circumstances is the integral zero and
nonzero.^ There are three cases to consider. Case 1 is when cuj
value o f the integral. In this case, the integral consists o f rwo sinusoids integrated over a common
period T. The integral of a sinusoid over any period is zero. Thus, the integral is zero and super
position o f power holds when cOj ^ O)-,.
Case 2 is when C0 | = co-, but with (Bj - O2) = knl2, k an odd integer. In this case, the integral is
again 0 . This follows because the first term o f the integrand is a sinusoid whose integral is zero
over the period T. The second term o f the integrand is a constant, cos(0j - (),) = cos{kKl2) = 0 ,
also resulting in a zero integral. Hence for case 2, superposition o f power holds.
Finally, we have case 3, for which tOj = co-, but with (6 j - 62 ) ^ kKll, k an odd integer; here
superposition ot power does not hold. The second term of the integrand is a constant, cos(0j - 62 )
0 , resulting in a nonzero integral over the period T. So
P^^^^ j +
For case 3, it is
desirable to use the phasor method o f Chapter 10 to compute the desired voltage and then use
equation 11.5 to compute average power.
Exercises. 1. In Example 11.2, suppose t/j(r) = 3cos(107rr) V and V2 {t) = 4 cos(15n:r + 0.25ti) V.
Compute T, a common period for the two sinusoids, and then compute the average power con
sumed by the 1 Q resistor.
C H EC K ; T = 0.4 sec w-ill work, and
2. In Example 11.2, suppose
age power consumed by the 1
= 12.5 watts
= 3cos(107Tf) V and
resistor.
C H EC K ; P^..=
ave 12.5 watts
3. Now suppose v^{t) = 3cos(10Tcr) V and
=4
c o s
(1 5 7 T /
Equation 11.6 resembles equation 1.18b for the dc power absorbed by a resistor connected to a
dc source. However, in equation 11.6 the factor 1/2 is present. With the introduction o f a new
concept called the efifective value o f a periodic waveform, the formulas for the average power
absorbed by a resistor can be made the same for dc, sinusoidal, or any other periodic input wave
forms.
S06
is a positive con
stant such that a dc current o f value /^-exciting the resistor causes the same amount o f average
power to be absorbed, i.e.,
(11.7a)
or
(11.7b)
R
Equation 11.7a suggests that
1
Iq+T'
I f " )
T
Hence, the mathematical definition o f the effective value of a periodic current i{t) is
/m+7
U>ff.R
(11.8a)
ef/M
(11.8b)
F e jf-
(11.8c)
Observe that the expressions under the radical sign in equations 11.8 constitute the average value
ot the square o f the signal. Hence, the expressions give rise to the alternative name for the effec
tive value, the root-m ean-square (abbreviated rms) value oiJ{t), since
mean value o f the square j{t) over one period.
SO'
Exercises. 1. Show that the average power absorbed by an R Q resistor carrying a periodic current
7?
= 2.5\/2
and
EXA M PLE 11.3. Compute the efFective value o f the periodic voltage waveform sketched in
Figure 11.5.
So
lu t io n
Therefore,
1 1
.2
4J
4J
4J
= 2.3094 V,
Exercise. Repeat the calculation o f Example 11.3 for the case where the values on the vertical axis
o f Figure 11.5 are doubled.
A N SW l'R :
as follows:
f~{t)
Fm
-
= F jC o s ( o j/+ 0 ) =
F~m
c o s (2 o j/+ 2 0 )
S08
Since by assumption to 7^ 0, the average value o f the cosine term is zero. The average value o f the
first (constant) term is itself Hence, by equation 11.8c,
F~
I itt
(11.9)
Eml
72
Thus, for a sinusoidal waveform, the effective or rms value is always 0 .707 times the maximum
value or, equivalently, the ma.ximum value divided by >/2 a basic fact well worth remembering.
The ac voltage and current ratings o f all electrical equipment, as given on the identification plate,
are rms values unless explicitly stated otherwise. For example, the household ac voltage is 110 V,
with a maximum voltage o f 1 lOx-s/^ = 1 5 6 V. A typical appliance such as a coffee maker will
have a 110 V rating, ac, at say, 900 watts. The effective values o f a few other periodic waveforms
are listed in Figure 11.6, with their derivations assigned as exercises.
Feff
=F
dc
sinusoidal
triangular
square
F.= F .
S09
Exercises. 1. Derive the formula for riie cfFectivc value o f a triangular waveform shown in Figure 11.6.
2. Compute the effeccive value of the waveform shown in Figure 11.2a.
AN SW FR:
cos(0,, - 0y) = - ^
V2 v 2
For the remainder o f the chapter, all voltage and current phasors will be taken as being effective
values unless the subscript m or
be added sometimes for emphasis, however. This practice is widely accepted in the power engi
neering literature. Omitting the subscript eff\v\ equation 11.10 yields
K / c o s (0 ^ -0 .)^ V7cos(0p
where 0^ = 0,^ - 0 y , V= 0.707
( 11.11)
Z (/co) o f the two-terminal element and is also interpreted as the angle by which the voltage phasor leads the current phasor.
EXA M PLE 11.4. Figure 11.7 shows two t)'pes o f household loads connected in parallel to a 110
V, 60 Hz source,
= 1 1 0 ^ 2 cos(120:tr) V. Lainp 1 and lamp 2 have effective hot resistances
o f 202
(a)
(b)
and 121
Lamp 1
Lamp 2
lu t io n
(a) For lamp 1, Z|(/to) = 202Z.0 Cl. Hence, Ij = Vy^Zj = 0.5446Z .0 A. From equation 11.11,
^\ave ^
z\^ = 110
5 10
11. 11,
= 110 X 0.9091 cos(0) = 100 watts
^ jJZ jj =
^Jlave =
Pave = |V/,,l|lm|co.s(0, - 0 , ) =
Observe that the sum o f the individual average powers is 250.3 watts, which equals the power
delivered by the source within the accurac)' o f our calculations, where we have rounded our
answers.
51 1
eff
o +
Two
Terminal
Veff
Device
o-
FIGURE 11.8 Two-terminal device with phasor voltage and current consistent
witli passive sign convention.
i(t) =
= Re S = Re
Also suppose
a rwo-terminal device excited by this voltage-current pair (Figure 11.8) is given by equation 11.10 as
S = \ //C /f =
j =
h r
in which case
= Re
R e[S]= R e
The curious reader may ask why a conjugate o f the current is used in the definition o f complex
power.
Suppose
Re
one
did
not
have
the
conjugate
o f the
current.
Then
= VeffI(,ffCOs{Q^,+Qj)^I\,y^, i.e., the resulting product would have no physical
meaning. Now because S is a complex number, it has an imaginary part, a magnitude, and an
angle. T he imaginary part o f S defines a quantity called the reactive power absorbed by the twoterminal device in Figure 1.18; i.e., reactivepoioer \s defined as
Q = Im [S]=
J=
14)
The unit o f reactive power, Q, is VAR, which stands for volt-amp-reactive. It follows immediately
that
where P =
. Also, the magnitude o f S is defined as the apparent power absorbed by the two-
________
^12
The Linir o f apparenr power is VA, short For volt-arnp. The iinerrclationship o f these different pow
ers is illustrated by the right triangle diagram in Figure 11.9, which is often helpful in solving
problems. Observe that the apparent power is always greater than or equal to the average poiuer,
with equality applying to the case o f a purely resistive load.
FIGURE 11.9 Relationships among complex, average, reactive, and apparent powers.
I'h e distinction among these various powers is best understood by computing the powers for some
basic circuit elements. For simplicit)', except when needed or for emphasis, from this point on we
will often drop the subscripts <^and rfw'as given in equations 1 1 .1 3 -1 1 .1 6 .
EXA M PLE 11.5. This example explores the computation o f the various powers for a simple
inductor. Given that 1^0 ) = v2/sin(CO/) in the circuit of Figure 11.10, compute
V^, S^,
Ql, the instantaneous absorbed power />/(/), and the instantaneous stored energ)' Wjit) in terms
o f L, CO, /^, and Vj. After this show that
(i)
max
Ql
(ii)
Wiit) max
Ql
(0
Remainder
of Circuit
FIG U R E. 1 1 . 1 0 Isolation o f an inductor for investigating the concept o f com plex power.
So
SI 3
lu t io n
= -jl^,
= /oZ.1^ = cdZ./^ = V^.
jVJi^ = I\ + yQ^. This implies that
= 0 and (X = V Ji- Further, the instantaneous
absorbed power is Pi ( f ) = v
=
COLy/lIi^cosUor)x J l l i sin(o)/) = V^// sin(2(0/), which
By inspection, and noting that wc again presume effective values,
= V^I
,a\ ~
~ Ql Further,
[l-c o s (2 c o / )]
[l-c o s (2 o )/ )]
-------------- ^l l -------- z-----------=
2co
2(0
[ l - c o s ( 2 (0 /)]
----------2(0
(f)
\Qi\
0)
, as was to be shown.
EXA M PLE 11.6. This example, like the previous one, investigates the concept o f reactive power,
but in the case o f a capacitor. The calculations will all be dual to those o f Example 11.5. Hence,
given that ^ ^ ( 0 = >/2V(-sin((0/) in the circuit o f Figure 11.11, compute
instantaneous absorbed power
S^^
'r
Qq the
terms o f C,
(O ,
P c (0
(ii)
lV c (0
Qc
and
Qc
(0
Remainder
of Circuit
FIGURE 11.11 Isolation of a capacitor for investigating the conccpt of complex power.
So
lu t io n
^C~
S^=
absorbed power is /;^(/) =
which
is
consistent
with
= | 0 c L Funher,
^ ,,2 [l-c o s (2 (o / )]
(oCVr
------------------ 2w
,, ,
= Vr^c
[l-c o s (2 (o / )]
-----------------2(0
= Qc
[ l - c o s ( 2 (0 /)]
-----------------2(0
^\A
Qc
These quantities,
(0
ing energy storage values in inductors and capacitors in systems where energy is to be recovered
and stored, and for modifying the power factor (to be discussed shortly) in networks with motors.
Energy storage in systems and power management are important research topics in todays world.
In Examples 11.5 and 11.6, one observes that the inductor absorbs reactive power while the capac
itor absorbs negative reactive power or, equivalently, delivers reactive power. This follows from the
definition o f complex power (equation 11.13, i.e., S =
derives from the convention that whenever the phasor current lags the phasor voltage (as with the
inductor), the device is considered to absorb reactive power, whereas if the current phasor leads
the voltage phasor (as with the capacitor), the devicc is considered to deliver reactive power.
Indeed, the overwhelming majority o f loads (toasters, ovens, hair dryers, motors, transformers,
TV s, etc.) have lagging currents.
When a t%vo-terminal element absorbs an average power
trical energ)' into other forms o f energ}' for example, heat or kinetic energy. In contrast, when a
two-terminal element absorbs reactive power Q, no energy is expended. T he energy transferred
into the two-terminal element is merely stored and later returned to the surrounding network. To
distinguish it from real (expended) power, we use VAR (volt-ampere-reactive) instead o f watt as
the unit for the reactive power Q.
EXA M PLE 11.7. This example investigates the computation o f the various powers defined above
for an /?Ccircuit. Here, consider the circuit o f Figure 11.12, where v,(/) = 100 V 2 c o s (200071:/)
V. Find the complex, average, reactive, and apparent powers absorbed by the load.
100 Q
lO kO
16nF i
Compute
Z,ry20007Cj = 100
j -I- 720007T X 16x10
10
-9
Step 2. Compute
Converting
SI^
to a phasor, we have
I,v, = ^
= 10 + j9 .8 5
= 100 V. By O hm s law,
mA
^in
Step 3. Cornpute the complex power absorbed by the load. By equation 11.12,
S =
Before doing a more complex example, we will discuss the particulars o f the principle o f conser
vation o f power in the sinusoidal steady state.
The validity o f this principle follows from KVL and KCL. This principle leads to the particular
fact that complex power is conserved in ac circuits operating in the SSS.
Note however, that the conservation principle does not hold for apparent power, i.e., for the mag
nitude o f the complex power. The following example illustrates a basic use o f the conservation law.
S I6
EXA M PLE 1 1.8. This example illustrates the application o f the principle o f conservation o f com
plex power in determining power delivered by a source and the input current to a circuit. We also
show that conser\'ation o f apparent power does not hold. Consider the circuit o f Figure 11.13.
Find the power delivered by the source and the phasor input current,
;1 6 0 VA, S 2 = 360 - ; 1 2 0 VA, S 3 = 420 + y540 VA, S 4 = 130 + >80 VA, S 5 = 40 - ; 1 0 0 VA.
/
/
lOOV
V- \
S.
\ A
FIGURE 11.13 Bridge circuit where S- represents the complex power absorbed by the element.
S o lu tio n
Hence,
= 'OOI/,, = 1 3 1 0 + p 6 QVA
= 13.1 - J5.6 A.
Exercise. Repeat the above example calculations for S j = 300 + y'400 VA, S 2 = 300 - y 4 0 0 VA, S 3
= 600 + j\ 000 VA, S^ = 60 + y80 VA, S^ = 120 -> 1 6 0 VA. What are the average and reactive pow
ers delivered by the source?
AN SW FR: S^
w ;itts
The next example illustrates the computation o f various powers through basic definitions and
application o f the principle o f conservation o f power.
EXAM PLE 11.9. Consider the circuit of Figure 11. 14, which depicts a motor connected to a commer
cial pow'er source. The motor absorbs 50 kW of average power and 37.5 kVAR o f reactive power, and has
a terminal voltage
= 230 V. Find IIj, the complex power delivered by the source, S^, and IV^I.
R,line = 0 .5 0
FICfURE 11.14 Motor absorbing 50 kW and 37.5 kVAR at a terminal voltage o f 230 V; the value of
^line
exaggerated for pedagogical purposes; electrical code requires that the size of the connecting
wire be large enough that the voltage drop is only a small percentage of the source voltage.
S o l u t io n
Step 1. FiJid the apparent power, |S^^J, absorbed by the motor. Since
S, = ^V-Q = 50t/-37.5kV A
it follows that |S,^J = 62.5 kVA.
Step 2. Find |/J. Here, |S,J =
= S, . s,,. = S, *
= 5 0 ;3 7 .5 . 3 6 .92 = 8 6 .9 2 *y 3 7 .5 kVA
h
In the above example we choose
s.
= 3 4 8.4 V
I.
Average Power
P^ye
Apparent Power
|S|
= cos(0,-0,)
(11.25)
SIS
The right-hand portion o f equation 11.25 follows directly from equation 11.13. Equation 11.25
specifies the power factor as cos(6^, Oy) , i.e., the cosine o f the difference between the angles o f
the voltage phasor V and the current phasor I. Clearly, 0 < pf < 1 . The angle
(11.26)
Since cos(x) = cos(-a:), the sign o f (0^^- 0^) is lost when only the pf is given. In order to carry the
relative phase angle information along, the common terminology is p f lagging ov p f leading. A l/igging power factor occurs when the current phasor lags the voltage phasor, i.e., 0 < (0^, - 0y) < 180 .
A leaditjg power factor occurs when the current phasor leads the voltage phasor, i.e., 0 < (0y - 0^j <
180 . Practically all t)^pes o f electrical apparatus have lagging power factors. Some typical power
factor values are listed in Table 11.1.
TABLK 11.1. Power Factors for Common Electrical Apparatus
T
ype o f
L oad
P o w e r F a c t o r ( L a g g in g )
Incandescent lighting
1.0
Fluorescent lighting
0 .5 -0 .9 5
0 .5 5 -0 .7 5 , at rated load
0 .9 -0 .9 6 , at rated load
To illustrate the idea o f leading and lagging pf, consider the circuits o f Figure 11.16. Suppose the
circuits operate at a frequency o f 400 Flz or co = 2 5 13.3 rad/sec. For the circuit o f Figure 1 1.6a,
I = (1 -jO.995) 10^ V = 1.41 10^/.44.85 V. Hence, the current phasor lags the voltage pha
sor, i.e., ( 0 J ,- 0y) = 44.8 5 and the pf is cos(44.85) = 0.709 lagging. On the other hand, for the
circuit o f Figure 11.16b, I = (1 + ;2 .5 ) 10-^ V = 2.7 103 Z 6 8 .3 V. Hence, the current phasor
leads the voltage phasor by 68.3", i.e., (0 0^^ = 6 8 .3 and the pf is cos(68.3) = 0.688 leading.
o-
O-
i0 .4 H
IH F
1 kO
1 kQ
O-
(a)
(b)
SI 9
Since the operating voltage is fixed, the line current needed to drive the load is higher than if the
load operated at a higher pf, say 0.95. Relatively speaking, a higher pf has a lower reactive power
component with correspondingly lower apparent power. Figure 11.9 helps to visualize the rela
tionships. For fixed line voltage, lower apparent power (higher pO means lower line current and
hence lower power loss in the connecting transmission line. In todays world o f energy conserva
tion, it is important to be energy efficient. The following example illustrates how improved pf on
a load can reduce line losses and thus decrease cost o f operation.
EXA M PLE 11.10. This example reconsiders Example 11.9, involving a motor connected to a
commercial power source as illustrated in Figure 11.17. The solution process will emphasize the
basic definition o f pf and the use o f voltage and current phasors. Suppose the motor absorbs 50
kW (about 67 hp) o f average power at a pf o f 0.8 lagging. The terminal voltage,
is 230 V. The
frequency o f operation is 60 Hz or co = 120 j1 . For the first part o f the example the capacitor in
Figure 11.17 is not connected to the motor. In part (c), the capacitor is connected to the motor
to improve the p f This will reduce the magnitude o f the current supplied by the source and hence
reduce the line losses.
(a)
(b)
(c)
Correct the power factor o f the combined motor-capacitor load to 0.95 lagging by choos
ing a proper value for C.
(d)
Compute the new power delivered by the source to the combined motor-capacitor load.
R,line =0.5 0
FIG U RE 11.17 Motor absorbing 50 kW and 37.5 VAR at a terminal voltage of 230 V. Again, the
value of
is exaggerated for pedagogical purposes; clectrical code requires that the size of the con
necting wire be large enough that the voltage drop is only a small percentage of the source voltage.
S o l u t io n
= 50 kW = R e[S] = |S,|cos(e -
!s,
0. 8
= 6 2 .5 k V A = \/,/,
( 1 1 .2 7 )
S20
. Ligging means that current phase lags behind voltage phase, i.e., 0 ^ -
= 0 - 0^ > 0 . Consider the diagram in Figure 11.18, which shows that the current phasor I^ la g s
the voltage phasor, i.e., the current phasor makes an angle o f - 3 6 .8 7 = cos ' (0.8) from the volt
age phasor. Hence,
= 3 6 .8 7 > 0
S m= V IDP m
and
. By definition,
s,,, = | S , k S m =
= 6 2 .2 Z 3 6 .8 7 kVA
= 5 0 + ;3 7 .5 k V A = /^t>Q
(b) Fi7ici
230
= 2 7 1 .7 4 A
, Z.1^ = - 3 6 .8 7 . Hence
s :i
and this strateg)^ is the goal o f the next part o f the example.
(c) Correct the power factor o f the combined motor-capacitor load to 0.95 biggifig- Since motors are
inductive, a properly chosen capacitor can improve the pf to 0.95 lagging. The new motor con
figuration is that o f Figure 11.17, with the capacitor connected across the motor. The proper value
o f C must be found.
Step 1. What does a p f o f 0.95 lagging require in terms ofcomplex power absorbed by the motor-capac-
itor combination'^
=
Z c o s " l( 0 .9 5 ) = 5 2 .6 3 Z 1 8 .1 9 5 = (50 + 7 1 6.4342) kVA
0.95
(11.28)
Recall that
= (50 + > 37.5) kVA
Step 2. Find a capacitor value to reduce the reactive power. For this step consult Figure 11.19.
522
Therefore,
plex power absorbed by the load is the sum o f the complex power consumed by the motor and
the reactive power of the capacitor, as illustrated in Figure 11.9, i.e.,
S f'= (50 + y i6 .4 3 ) kVA
Step 2. Compute the new
denoted
capacitor load,
Example 11.10 illustrates how adding a parallel capacitor can improve the pf o f a load. The main
motivation for improving the pf was to reduce the power loss in
However, even if
is
negligible, another strong reason exists for improving the load pf. Example 11.11 illustrates how
an improved power factor allows a single generator to run more motors. Example 11.11 will fully
utilize the principle of conservation o f complex power and the two consequences o f equation
11.25.
From equation 11.25 and the fact that S = /^ + jQ, we can express pf directly in terms o f P an d Q
as follows:
. ^
523
with a lagging pf for Q > 0 and a leading pf for Q < 0. Solving for Q from equation 11.29, we
obtain
e=
P .H r - i
P f
(11.30)
Generator
230 V ^
60 Hz
lOOkVA
Type B Motor
Type A Motor
Type B Motor
FIGURH I 1.20 A generator supplying power to one large motor and several smaller motors.
(a)
(b)
(c)
(d)
Can the generator safely supply power to one large motor and three small motors? What
are the generator current (magnitude) and the power factor o f the combined loads?
Compute the number o f small motors (besides the one large motor) that can be run
simultaneously without exceeding
If the power factor for all motors,
ing appropriate parallel capacitors
(besides the one large motor) can
tors rating?
Compute the capacitances required in part (c) for the large and the small motors.
S o lu tio n
(a) Compute the reactive power for each motor type. Using equation 11.30, the reactive power for
each type o f motor is given as
- 1
=50
-1
= 37.5 kVA
0.8^
and
Q b = P ,i, / A - >
pfg
= \
-^ 2 -
0.7
= .i . l 0 1 k V A
2 I
By the principle ot conservation o f power, the complex power (in kVA) supplied by the generator
is
V
= 65 + ;5 2 .8 = 8 3 .7 4 ^ 3 9 .
kVA
By inspection, the apparent power is 83.74 kVA, which is below the generator capacity o f 100
kVA, meaning that the generator can safely operate the large motor and three smaller motors.
The magnitude o f the generator current is 83,740/230 = 364 A. From equation 11.29, the pf o f
the combined loads is
pf =
^
P- +
= 0 .7 7 6 2
\/65 + 5 2 . 8
(b) Compute the number o f stnall motors {besides the one large motor) that can be run simultaneous
ly. When one large type A motor and n smaller t)^pe B motors are connected in parallel, the com
plex power delivered by the generator is
^gen
7 ^
100 =
10^
(11.31)
Replacing the inequalit)' sign in equation 11.31 by an equalit)^ results in the quadratic equation
51.020/r + 882.5750/7 - 6 ,0 9 3 .8 = 0
The resulting zeros are //, = 5.288 and //2 = - 22.58. The largest positive integer satisfying the
inequality 11.31 is // = 5. Thus, at most, five small motors can be run simultaneously with the
large motor without exceeding the generators capacit)'
(c) I f all power factors are corrected to 0.9 lagging, fin d the number o f small motors {besides the one
large motor) that can be run simultaneously. We essentially repeat the calculations o f part (b) with
the new given power factor o f 0.9 lagging:
and
e :r = / '.. A
V PfA
=5o
y - 1 = 2 4 . 2 1 6 kVA
V 0 .9 ^
s '" "
ficn
above inequalit}' is n = 8. Thus, eight small motors, as opposed to five in the earlier case, can be
run simultaneously with the large motor without exceeding the generators capacity.
(d) For the large motor, the capacitor must absorb a negative reactive power equal to
Q^.
(10.32)
I Q-CA I = ^CA^CA =
6 0 Q X 2302 VAR
(10.33)
Equating equations 10.32 and 10.33, we have In x 60C^|2302 = 13284. Solving produces
66 6 .1 6 X 10-<^ E
Similarly, for the smaller motors,
Qcb\ ~Vcb^Cb\ -
= In
X60Q
X 2302 VAR
Equating these two quantities and solving for C^, w'e obtain Cg = 134.35 x 10^ F. We note that
in the power industry, such capacitors are usually specified only by their kVAR rating, with no
mention o f their actual capacitive value in F.
In the above example the generator capacit}^ was given in terms o f VA, the unit o f apparent power.
The example points out the importance o f reducing reactive power to more fully utilize the power
526
capacity o f the generator. Use o f VA for generator, motor, and transformer capacit)' arises out o f
safety considerations. Most ac machiner)' operates at a specified voltage depending on the insula
tion strength. The size o f the wire and other heat transfer factors determine the maximum allow
able current o f a machine or transformer. Also, the cost and physical size o f most ac equipment
are more closely aligned to the VA rating than to other measures. Hence, the VA rating better
reflects the safe operating capacity o f ac equipment.
Another motivation for improving the power factor is economical. A power company charges a
consumer only for the actual electrical energy used. A meter measures this energy usage in units
o f kWh (kilowatt-hour). As mentioned earlier, most clectrical loads have lagging currents. As
shown in Examples 11.10 and 11.11, for a given required average power, a higher pf means lower
transmission line losses. Also, loads that operate at low pf force power companies to pursue high
er kVA ratings o f the generator equipment. Thus utilities companies encourage consumers to
operate their equipment and appliances at high pfs. Since power companies can supply more
power with the same equipment if the pf is high, they adjust their rates so that energy costs are
less with a high pf and are greater with a low pf.
and
(11.33b)
^ <e ff
max
(11.33c)
>2.
To derive the conditions o f the maximum power transfer theorem, observe that the current phasor, I, is
(11.34)
1=
{R, + RO + j(X , + XO
Thus the average power delivered to the load is
(11.35)
Here Pave
the partial derivadves ------ and ------ to zero and solve for
set
dRi
{R^ + Rl)~ +
-)12
(1 1 .36a)
dP
ax,
v "-[-1R l {X,
+ Xl )
= 0
(11.36b)
528
which is equation 11.33b. Substituting this result into the numerator of equation 11.36a yields
v3 ^
(R s + Rl T
The only physically meaningful solution here is
^ 0 and
R ,-R ,
which produces equation 11.33a. (Note that this is the condition for maximum power transfer in
purely resistive circuits.) Substituting these results into equation 11.35 produces equation 11.33c,
Ir
4/?,
(11.37)
r\
If the source is a general two-terminal linear network, then its Thevenin equivalent must be found
before application of the maximum power transfer theorem. If the source is represented by a
Norton equivalent circuit, we can use a source transformation to obtain the Thevenin form and
then apply equations 11.33.
'
As pointed out in Chapter 6, maximum power transfer is not the objective in electric power systems, as the sources usually have very low impedances. On the other hand, it is a very important
factor to be considered in the design of many communication circuits, as illustrated in the following example.
o
S29
EXA M PLE 11 .1 2 . The radio receiver shown in Figure 11.22a is connccted ro an antenna. The
antenna intercepts the electromagnetic waves from a broadcast station operating at 1 MHz. For
circuit analysis purposes, the antenna is represented by the Thevenin equivalent circuit shown in
Figure 11.22b.
(a)
(b)
receiver input
equivalent circuit
(b)
(a)
= 1\ i l and
(a)
(b)
Since the reactances in the circuit have been tuned out, the input current to the receiv
er is simply 14.6/(21 + 21) = 0.348 mA. The input impedance has a magnitude
7.
= 1070.
+ 1070- = 10 7 0 .2 n
Therefore the magnitude o f the voltage across the receiver terminals is 0.348 x 1070.2 = 3 7 2.4
mV (rms). The power transferred from the antenna to the receiver is 0.348^ x 21 = 2.54 uW.
In the preceding discussions of maximum power transfer, we have assumed that the load is
adjustable. In practice the load is often fixed, as for example, in the case o f a loudspeaker having
a4
voice coil. In such cases, one designs coupling networks consisting o f lossless passive com
ponents. These coupling networks transform the fixed load impedance into one whose conjugate
matches the fixed source impedance. This permits maximum power transfer to the load. The fol
lowing example illustrates the principle. A design procedure for some simple coupling networks
will be discussed in the second volume o f this text.
530
ed to the source o f Example 11.12 through a passive coupling network, i.e., a network that does
not generate average power, as shown in Figure 11.23.
(a)
(b)
Show that the coupling network shown in Figure 11.23 achieves this maximum voltage
across
is 0 .504 V.
(a)
From equation 11.33c, as used in Example 11.12, the available power from the source is
2.54 j.iW. If all o f the power is delivered to R^ , then the voltage,
must be
^in = yw ^ + ------------- r~
j(dC +
R,
Substituting the values OJ = 10^, L = 400.9 x 10^, C = 109.8 x 10
is conjugate-matched to the
source impedance, the maximum power o f 2.54 |aW is transferred to the coupling network. Since
the coupling network consists o f L and C, neither o f which consumes average power, the 2.54 f,iW
power must be transferred out o f the coupling network and into the load resistance. The voltage
across the load resistor, Vj, is given by
= 4 ^
= V 2 .5 4 X 10^ X 100000 = 0 .5 0 4 V
This verifies that the coupling network o f Figure 11.23 enables the largest voltage to appear across
the load resistor.
531
8. SUMMARY
Fundamental to the material in this chapter is the definition o f the effective value (rms value) o f
a periodic voltage or current waveform. For a sine wave, the effective value is the maximum value
divided by -v/2 For a general periodic voltage or current, the effective value is the value o f a dc
waveform that will produce the same amount o f heat as the periodic waveform when applied to
the same resistance. Using the definition o f the effective value o f a waveform, formulas for the
average power absorbed by a linear two-terminal network in ac steady state were set forth and
derived. Recall that for a two-termlnal element with sinusoidal voltage v{t) = ^/2 V'^ cos(co/- 0y) an
current i{t) =
the passive sign convention. Next w'e presented the definition o f complex power and its compo
nent parts, which include its real part or average power, its imaginary part or reactive power, and
its magnitude or apparent power. Various examples illustrating the calculation o f these powers
were given. Again, for a two-terminal element w'ith sinusoidal voltage v{t) and current i{t) as
above, the reactive power absorbed is defined to be Q = ^ ^ ^ s i n ( 0 y - 0/) VAR (volt-amperereactive). After introducing these different types o f power, we proved the principle o f conserva
tion o f complex power, which implies the consen^ation o f real power and the conservation o f reac
tive power. This was followed by the definition o f power factor, pf, the ratio o f average power to
apparent power, which takes on values between 0 and 1. The need for improving a low power fac
tor and a method for achieving an improved power factor were illustrated with two examples.
The maximum power transfer theorem, first studied in Chapter 6 for the resistive nervvork case,
was taken up again in this chapter for the sinusoidal steady-state case. Here, maximum power
transfer to the load requires that the load impedance be the conjugate o f theThevenin impedance
seen by the load. As pointed out earlier, the theorem has no application in electrical power sys
tems. However, for communication circuits the maximum power transfer theorem is o f extreme
importance. The power that can be extracted from the antenna o f a radio receiver is usually in the
microwatt range, a very small value. It is therefore necessary to get as much power as possible from
the antenna system. Example 11.13 illustrates this principle.
assuming the
532
Efifiecdve value (nns value): for a sine wave, the effective value is the maximum value divided by
V 2 For a general periodic voltage or current, the effective value is the value of a dc wave
form that will produce the same amount of heat as the periodic waveform when applied
to the same resistance.
Instantaneous power: the power associated with a circuit element as a function of time. The
instantaneous power absorbed by a two-terminal element is p{t) = y(/)/(/), assuming that
a passive sign convention is used.
Maximum power transfer theorem: if a variable load 7.^^ = -^j^i is connected to a fixed source
Vy having a source impedance
=7?^+/A^, then the largest average power is transferred
to the load when
is the complex conjugate of Z j , i.e.,
= R^ and AT^ =-X^.
Power factor: the ratio of average power to apparent power. The pf value lies between 0 and 1.
For a passive load, the power factor is said to be lagging when 90 >0j, - 6,->0, and lead
ing when 90 > 0^- 0^, > 0.
Real powen in ac circuits, real power means average power. It is the real pan of the complex
power.
'
Reactive power: for a two-terminal element with sinusoidal voltage v{t) = -y/2 V^cosisat + 0^)
and current i{t) = ^ /^ co s(co f + 0^, the reactive power absorbed, denoted by Q, is
defined to be Q = ^ ^ ^ sin (0 y - 0/), assuming that a passive sign convention is used.
The unit of measurement is VAR (volt-ampere-reactive).
n
o
o
n
n
o
^33
Problems
AV(t)(V)
20
INSTANTANEOUS AND
AVERAGE POWERS
t(sec)
10- -
(a)
(a)
Figure Pi 1.3 (a) Rectangular waveform, (b)
Triangular waveform.
EFFECTIVE VALUE OF
NONSINUSOIDAL SIGNALS
(b)
Figure P l l .l
AN SW FR. 0.758 watrs
4. (a)
(b)
1 0co s(10t)m A
(b)
10 |cos(10^)| niA
(c)
10 cos( 1Of) mA
(d)
3. (a)
(b)
Figure Pi 1.3.
neous power associated with each
waveform for 0 < f < 3 sec.
(c)
534
i(t)0
60 O <
= 30 Q
(c)
Figure PI 1.5
6. (a)
2Q
v(t)
(b)
R = 80
part (a).
Figure P11.4
(b)
(c)
and b.
(b)
(0
5. (a)
v^{t) = 10 + 2 cos(20r)
= 10 cos(2r) + 5 cos(4/)
= 10cos(2r) + 5 cos(4r) +
5 c o s ( 4 f - 4 5 ) V.
if the cur
AVERAGE POWER
CALCULATIONS IN
SINUSOIDAL STEADY STATE
8. Using equation
44
15
(a)
t(sec)
-4 -
(b)
Figure PI 1.8
ANSWl-.R: 25 watts
= 50 x
(b)
(c)
Zi = p O Q, and a = 49.
/Y Y V
Load
' 6
aV
Figure 1 11.1 I
ANSW'FRS: (b) in random order: - 9 0 . 100
watts; (c) 4. - 3 .9 watts
Figure P l l .9
AN SW ER: (b) 40 warts
12.
10.
= 120
R = 32 W, Z^ = ;2 0 0 Q.
Z q = y80 ri, and ^ = 4.
where
5 0 z i-9 0
/? = 6 Q,
= ;1 2 Q, and
(a)
- ;4 Q.
(a)
(b)
and deter
O nly
R absorbs
average
power.
' 6
III is
Therefore, once
known, the
average power consumed by R is
=
y?|
(d)
bl
Figure PI 1.12
watts
Figure Pi 1.10
A N SW FR; (b) ISO wans
11. For the circuit in Figure P i 1.11,
100.10 V ^ , /? = 5 n , Z^ = ;5 0 a , and ^ = 9.
(a)
Compute the current phasor I^.
14.
Consider
Figure
Pi 1.14,
where
that IV
150 V, a n d
watts. Find the magnitude o f
250
the
^36
v.(t)
Figure P 11.17
C H EC K : Complex power is 7+y6 VA
Figure P 11.14
= 4 2,
64 rad/sec,
=
= 120/ i60"
= 20 2, /?,
same.
Load
L.
COMPLEX POWER
CALCULATIONS
16. For the circuit o f Figure PI 1.16, R = 5
= 3 .jA a ,
= -yiO
and
v,(r) = 100v 2 cos(1207i:/) V.
(a)
(b)
Figure P 11.18
C H E C K : Complex power is 36 + p 2 VA
the load.
average power (in watts), and the reactive
Zc
{-
'-0
(b)
Figure P 11.16
power delivered
sources.
Z 2 is 100 + ;5 5 0 VA
by the
two
(c)
537
(a)
r-
Z.
Z.
v .Q
-I
(b)
in rectangu
Figure Pi 1.19
C H EC K : complex power delivered by source h
Z,
is 10 + 260; VA.
v .Q
CONSERVATION OF POWER
20. This problem should be done without any
phasor voltage or current computations. In the
circuit o f Figure P i 1.20,
= 2300
Figure PI 1.21
C H EC K S: P,
at 60
complex
power delivered
by
the
source.
(b)
(c)
Determine
in polar form.
(b)
(d)
(e)
Find
(b)
(c)
in
Figure PI 1.20
C H EC K S: (a) Complex power delivered by
source: 48 + y'36 kVA; (b) |lj = 26.087 A
v,(t)
C H EC K :
- 1 8 .8 2 3 5 ;
Figure P i 1.23
= 20 - 5; A^ and V , = 75.2941
538
W).
(b)
Find
=A cos(1207t^ + 0) V and
the complex power delivered by the
source.
Load
0 -"0 -
!'4<
Transmission
line resistance
'N/S/'------ OR .,
o - -oFigurePll.26
Figure P I 1.24
CH ECK: (a) 100 watts
25. As shown in Figure P I 1.25b, a capacitor is
put in parallel with a motor using average
power
= 40 kW operating at a power fector of 0.7 laggmg to boost it to a power factor
of 0.9 lading. The voltage across the parallel
motor-capacitor combination is 230Z0 V ^ .
The power relationships are shown in Figure
P I 1,25a. If the frequency of operation is CO =
120tc, compute the proper value of the capaci
tance, C (in mF).
Device
1
Device
Figure P I 1.27
CH ECK: 0.108 mF
< C<2 mF
(b)
(c)
539
current
mF.
= 0.07 Cl.
(i)
(j)
Compute V 'f":
(k)
oM o to r)
Later
Addition of
Capacitor
for p.f.
Correction
MAXIMUM POWER
TRANSFER
Figure P11.29
30. Consider the circuit shown in Figure
The phasor voltage across the motor is ^eff~
2 3 0 Z 0 V. The sinusoidal frequency is 60 Hz.
(a)
(a)
(c)
delivered
(b)
to the motor.
CO =
(e)
(f)
Compute V^.
Compute the complex power deliv
(g)
= 100
R.
v .Q
o
100 rad/sec.
(d)
C=
1 mF.
= 100 Q, R-> = 25
PI 1.30 in which
Figure PI 1.30
C H EC K : 20 + ;1 0 Q
ging. Then
(i) Compute
sou
(ii) Recall that the role of the capaci
tor is to reduce the reactive power.
540
Load
/Y Y V l
L
R,
priate.
(b)
Compute
the
load
impedance,
Figure P i 1.31
A N S W H R : R, = 16 12 and C= 0.2 mF
= 3 and
transfer.
2 ^ 0 A ^ (a)
Find the value o f the load imped
ance
power transfer.
(b)
8Q
-L
3.334 mF - J 1.667 mF
16Q
CO
(a)
equivalent
circuit.
Figure Pi 1.34
A N SW ERS: (a) 10 12, 5 ml-, 1230 wans; (b) 5
mF, 937.5 wans
541
= 100
^rms
adjusted to achieve different
goals. Assume R = GO Q., Z q = - ^ 8 0 Q.
(a)
that maximizes
(C H EC K : 24 Q .) What is the
value o f
(b)
that maximizes
Figure PI 1.37
(b)
-------------
where
-------------
= 0.1
at
CO =
Figure P l l .35
ANSWHR: (a) 48 ih . r .0 3 7 watts
(b)
maximized.
Find the values o f L and C , to achieve
(a).
Find the impedance Z^ such that Pj
(c)
Load
/
zrc,
Figure P 11.38
'" " Q
6Q
v .(t )Q
-jisn
Figure P I 1.39
542
(a)
L^ive
2(0
(c)
at
42. (a)
= 0.
impedance Z = R + jX are
and Q =
(b)
= /?|lp
where I is in
4.115, 1.27
= (j|Vp
THEORETICAL PROBLEMS
L
40.
(b)
Find
. 2
j oules.
WC,ave _
Qc
2(0
p{t)
has
peak
value
of
Find
the instantaneous energy
store in L (or, rather, in the magnetic
field) and show that W^{t) has a peak
value ot
Let
be the reactive power absorbed
bv L. Show that
Z = R + jX
o --------(a)
+
V
Y = G + jB
a
(b)
CHAPTER OBJECTIVES
1.
2.
3.
Explain and illustrate the benefits o f using the Laplace transform tool for solving circuits.
Develop a basic understanding o f the Laplace transform tool and its mathematical prop
erties.
Develop some skill in applying the Laplace transform to differential equations and cir
cuits modeled by differential equations.
SECTION HEADINGS
1.
2
3.
Introduction
Review and Summary o f Deficiencies o f Second-Order Time Domain Methods
Overview o f Laplace Transform Analysis
4.
5.
Basic Signals
The One-Sided Laplace Transform
6.
7.
8.
9.
10.
Summary
Terms and Concepts
11.
Problems
1. INTRODUCTION
I'his chaprer introduces a powerful mathematical tool for circuit analysis and design named the
Laplace transform. Later, more advanced courses will describe the design aspects. Use o f the
Laplace transform is commonplace in engineering, especially electrical engineering. A student
might ask why such a potent tool is necessary for the analysis o f basic circuits, especially since
many texts use an alternative technique called complex frequency analysis. Complex frequency
analysis does not permit general transient analysis; rather, it restricts source excitations to sinu
soids, exponentials, damped sinusoids, and dc signals. This class of signals is small and does not
begin to encompass the broad range o f excitations necessary for general circuit analysis and the
related area o f signal processing. The Laplace transform framework, on the other hand, permits
both steady-state and transient analysis of circuits in a single setting. Additionally, it affords gen
eral, rigorous definitions of impedance, transfer fimctio}!, and various response classifications perti
nent to more advanced courses on system analysis and signal processing. Introducing the Laplace
transform early allows students an entire semester to practice using the tool and learn about its
many advantages.
Section 2 describes some of the difficulties associated with the methods o f circuit analysis intro
duced in earlier chapters when applied to circuits o f order 3 or higher. Following this, we present
an overview o f Laplace transform analysis in section 3, define important basic signals in section 4,
and introduce the formal definition o f the one-sided Laplace transform in section 5. The inverse
Liplace transform and important properties o f the transform process arc introduced in sections 6
and 7, with numerous illustrative examples. Section 8 applies the technique to circuits modeled
by differential equations. Such models were developed in Chapters 8 and 9.
roots {say Aj and X-y) using, for example, the quadratic fonnula in the second-order case.
Step 3. From the location ofthe roots o f the characteristic equation, determine the form o f the solution:
or if A, = X-),
Step 4. Compute the constant D by shorting itiductors, open-circuiting capacitors, and analyzing the
age sources. The goal o f this example is to construct a differential equation model, determine the
solution form in terms o f arbitrary constants, and demonstrate the difficulties w'ith the simple
recipe o f the above algorithm by attempting to relate the arbitrary constants to the initial condi
tions.
1Q
1Q
1O
-I-
FIG URE 12.1 A cascade of three RC circuits coupled by means of dependent voltage sources. The
differential equation model of the circuit is third order.
S o l u t io n
Step 1. Construct the differential equation o f the circuit. For this task, first write a dilTerential equa
tion relating
to
Then write one relating
to
and finally, write one relating
Some straightfor^vard algebra leads to the following three differential equations:
to
at
+ v c i ( 0 = 0.5iv
( 1 2 .!)
0 . 5 - ^ + it2(/ ) = 0 .5 i'c i
(12.2)
0 . 2 5 ^ . v , ( O = 0 ,5 v
^ , ^ 3 ^
Successively substituring equation 12.1 into equation 12.2 and equation 12,3 into the result pro
duces the input-output differential equation model,
^ont
-7 ^~^out , 1 A
dt^
dr
, o
(12-4)
d,
Step 2. Compute the characteristic equation and its roots. The characteristic equation for differen
tial equation 12.4 is
^ + 7s^ + \As + ^ = {s - a) {s - b) {s - d) = {)
Vou, it) =
+ Be^
(12.5)
for r > 0.
Step 4 . Compute A, B, D, and E in equation 12.5. Using the rule o f thumb mentioned earlier, a
simple calculation yields E = 0.125
tion 12.5. Applying the recipe described earlier, we take derivatives o f equation 12.5. set t = 0, and
relate them to the circuit initial conditions:
^out (0) = ^
'^out (0) =
+ dD
Again, one dot over a variable means a first-order time derivative, and two dots denotes a second-order
time derivative. A, B, and D are computed by solving this set o f equations. The difficulty is in specifying
and
First,
tor. However, v^^^^it) is proportional to the current through it, which depends on all the initial
capacitor voltages. Further, what is the physical interpretation o f
and v^^fiO) relate to the initial capacitor voltages? The relationship is complex and lacks any mean
ingful physical interpretation. Finally, even for this simple example, computation and solution o f
the differential equation 12.4 proves tedious.
One o f the advantages o f Laplace transform analysis is that it does not destroy the physical mean
ing o f the circuit variables in the analysis process. Chapter 13 addresses how the Laplace transform
approach explicitly accounts for initial capacitor voltages and initial inductor currents.
Input Signal
\
Laplace Transfornn
of Input Signal
'' r
Laplace Transform
of C I R C U I T ^ ,
Laplace Transform
of Output Signal
S4iS
In a mathematical context, one executes the same type o f procedure on a difFerential equation
model o f a circuit and, indeed, difFerential equations in general. Figure 12.3 illustrates the idea.
Input Signal
DIFFERENTIAL'
EQUATION
Laplace Transform
of Input Signal
Laplace Transform
of DIFFERENTIAL
EQUATION
Laplace Transform
of Output Signal
FIG URE 12.3 Diagram showing flow of Laplace transform analysis for solution
o f differential equations.
The benefit o f this t)'pe o f analysis lies in its numerous uses. Some o f these uses include steadystate and transient analysis o f circuits driven by complicated as well as the usual basic signals, a
straightforward lookup table approach for computing solutions, and explicit incorporation of
capacitor and inductor initial conditions in the analysis. The forthcoming sections will flesh out
these applications.
4. BASIC SIGNALS
Several basic signals are fundamental to circuit analysis, as well as to future courses in systems
analysis. Perhaps the most common signal is the unit step function.
(r) =
1,
r>0
0,
r<0
( 12.6)
defined in Chapter 8. The bold line in Figure 12.4, resembling a step on a staircase, represents the
graph o f u{i).
u(t)
FKiURE 12.4 Graph of the unit step function. It often represents a constant voltage or current le\'el.
The unit step function has many practical uses, including the mathematical representation o f dc
voltage levels. Any t)'pe o f sustained, constant physical phenomenon, such as constant pressure,
constant heat, or the constant thrust o f a jet engine, has a step-like behavior. In the case o f jet
engine thrust, a pilot sends a command signal through the control panel to the engine requesting
a given amount of thrust. The step function models this command signal.
The shifted step, shown in Figure 12.5, models a rime-delayed unit step signal.
u(t-T)
1 --
FIGURE 12.6 Graph of flipped and shifted unit step. This function is often used to model signals
that have been on for a long time and turn ofTat time /'.
(0 / ( 0 =
1,
0</<2
0,
otherwise
I, -3</<6
(///) m =
0.
otherwise
1,
/ < -l
0,
-l< r < l
1,
; > 1
u {t
ii{ t
+ 3) - i d t - {>)
The pube function, p-^i), o f Figure 12.7 is the product o f a step and a flipped step or, equivalent
ly, the difference o f a step and a shifted step. Specifically, a pulse o f height A and width 7" is
=
Au{i)u{T -
1) =
Au{t) - A u { t - T)
(12.7)
S50
Pr(t)
>
T
FIGURH 12.7 Pulse of width T and height A.
This function is often used to model signals o f fixed magnitude and short duration.
A signal sharing a close kinship with the unit step is the ramp function r{t) depicted in Figure
12. 8.
r(t)
/(0=
w h e re t
J oo
ii(T)dx =ln{t)
( 12.8)
is s i m p l y a d u m m y v a r i a b l e o f i n t e g r a t i o n .
y t)
5S1
o l u t io n
For the signal/j (/), observe that the signal begins with a ramp with a slope o f 2. Thus we have
f\ U) = 2r(f) + ? . At ; = 7, the signal/j(^) levels off. Since the 2r{t) part o f the signal continues to
increase, the increase must be canceled by another ramp o f slope 2, but shifted to the right by T
units. Thus,/j(r) = 2r{t) - 2r(t - T).
The signal j 2(^) replicates/j(^) up to 3T . After 37', the signal drops to zero. Hence we must sub
tract a shifted step o f height 2 7 from/j(/), Thus fjit) =/j(f) - 2Tu{t - 3 7 ) = 2r{t) - 2r{t - T) -
2T u {t-5 T ).
Exercises. 1. Figure 12.10 depicts a sawtooth waveform denoted hy J{t). Sequences o f sawtooth
waveforms are used as timing signals in televisions and other electronic devices.
f(t)
A
2. For/r) o f Figure 12.10, p lo t/ l - t) and represent the ftmction in terms o f steps and ramps.
AN SW ER: /(I - /) = ;-(l - r) - >i-t) - u{-t)
EXA M PLE 12.3. Express Figure 12.11 in terms o f steps and ramps.
f(t)
> t
FIG URE 12.1 I Triangular waveform to be represented by steps and ramps.
S
o l u t io n
Observe that the signaly(f) begins with a ramp with a slope o f 1 at r = -T . Thus /(f) = r{t 7) +
?. The signal falls o ff with a linearly decreasing ramp for 0 < r < T. If we subtract r{t) the signal
would become flat for r > 0. Thus we subtract 2r{t) to obtain the linear decrease. Hence, y(r) =
^52
r{t + 7) - 2>it) + ?. For r > T, this signal, r{t + T) - 2Kr), continues to linearly decrease. Hence
for the signal to be zero for t > T, we cancel this decrease with an additional ramp. T h u s/ r) =
r{t + 7) - lr{t) + r { t - T).
Newtonian physics provides a good motivation for defining the ramp signal. Applying a constant
force to an object causes a constant acceleration having the functional form Ku{t). 'J'he integral o f
acceleration is velocity, which has the form
a ramp function.
A very common and conceptually useful signal is the (Dirac) delta function, or unit impulse
Function, implicitly defined by its relationship to the unit step as
h{cj)dq
(12.9)
, V
I-
u{t)~ ii{t - h)
h
//->{)
( 12. 10)
Strictly speaking, the derivative o f u{t) does not exist at t = 0, due to the discontinuit)' at that
point. Without delving into the mathematics, one typically interprets equations 12.9 and 12,10
as follows: define a set o f continuous differentiable functions
The derivatives,
-t
(a)
(b)
Clearly, 6^(/-) has a well-def'med area o f 1, has height I/A, and is zero outside the interval 0 s t
s A. In addition, u[t) = lim //^(r), and lim 6^ = b{t) as A - 0.
o f equation 12.10 is not mathematically rigorous, one can interpret the delta function as the limit
o f a set o f well-behaved functions. In fiict, the delta function can be viewed as the limit o f a variet)' o f different sets o f functions. A problem at the end o f the chapter explores this phenomenon.
Despite the preceding mathematics, the delta function is not a function at all, but a distribution,^
and its rigorous definition (in terms o f so-called testing functions) is left to more advanced math
ematics courses. Nevertheless, we shall still refer to it as the delta or impulse function. The stan
dard graphical illustration o f the delta function appears in Figure 12.13, which shows a pulse of
infinite iieight, zero width, and a well-defined area o f unit}', as identified by the 1 next to the
spike. Visualization of the delta function by means o f the spike in the figure will aid our under
standing, explanations, and calculations that follow.
FIGURE 12.13 Standard graphical illustration o f a unit impulse lunction having a wcll-dcfincd area
of 1. The function typically represents an energ)' transfer, large force, or large impact over a very
short time duration, as might occur when a bat hits a baseball.
The unit area property follow's from equation 12.9, i.e.,
bU)(lr=
5(rV/r = / / ( 0 ^ ) - / / ( 0 " ) = l
where 0 is infinitesimally to the left and 0"^ infinitesimally to the right o f zero. I f the area is dif
ferent from unity, a number Kalongside the spike will designate the area; i.e., the spike will be a sig
nal Kb{t).
One motivation for defining the delta function is its abilit)' to ideally represent phenomena in
nature involving relative immediate energ\' transfer (i.e., the elapsed time over which energy trans
fer takes place is very small compared to the macroscopic behavior o f the physical process). An
exploding shell inside a gun chamber causing a bullet to change its given initial velocity from zero
to some nonzero value instantaneously is an example. Another is a barter who hits a pitched ball,
instantaneously transferring the energy o f the s\vung bat to the ball. Also, the delta function pro
vides a mathematical setting for representing the sampling of a continuous signal. Suppose, for
example, that a continuous signal v{t) is to be sampled at discrete time instants t^ r,*
v{t) is to be physically measured at these time instants. The mathematical representation o f this
measuring process is given by the sifting property of the delta function.
v(//)=
(12.11
v(l)5(X
In other words, the value o f the integral is the non-impulsive part o f the (continuous) integrand,
replaced by that value o f r which makes the argument o f the impulse zero, in this case r = tj.
Verify'ing equation 12.11 depends on an application o f the definition given in equation 12.9.
Specifically, if v{t) is continuous at f = t-, then
\'(T )6 (T
1: )dX =
v (T
)5(T 5(1 -
f j )dx
rj)(lx = v(tj)
SS4
Exercises. 1. Compute the derivatives o f the signals in Figures 12.9a, 12.9b, and 12.11.
A N SW E R S;/,(/) = 2|//(/)-/ / (/ -T ) ] . / .(z) = 2[//(/) - / / ( / - 7 ')| - 276 (/ - 3 7 ').
) = //(/+
T)-2iiii)+
oo
//(/-
T)
-oo
LAPLACE TRANSFORM
Mathematically, the one-sided Laplace transform
J{t) is
( 12. 12)
where s = (5 + ./CO(y = yf-A) is a complex variable ordinarily called a complex frequency im
the signals and systems literature.
As the equation makes plain, the Laplace transform integrates out time to obtain a new func
tion, F{s), displaying the frequenc)' content o f the original time function/r). In the vernacu
lar, F{s) is the frequency domain counterpart of/ r). Analysis using Laplace transforms is often
called frequency domain analysis.
Exercises. 1. Find the Laplace transform o f a scaled Dirac delta function, Kb{t).
ANSWF.R; K
2. Find [sin(2;rr + 0 .5 jr ) 6 ( r - 0.5)].
ANSWHR: sin(l.5.T)<'-'^ ^
u{t + 1) would have the same unilateral Laplace transform). However, since ^ = 0 designates the
universal starting time o f a circuit or system, the class o f signals dealt with will usually be zero for
t < 0 and thus will have a unique (one-sided) Laplace transform. Conversely, each Laplace transform F{s) will determine a unique time fiinaion J{t) with the property that f^t) = 0 for ^ s 0.
Because o f this dual uniqueness, the one-sided Laplace transform is said to be bi-unique for signals/^) yfirh J{t) = 0 for ^ < 0.
Question 4 : Does every signalj{t) such th a tfj) = 0 fo r f < 0 have a Laplace transform?
Answer: No. For example, the function f i ) = ^ (/) does not have a Laplace transform because
the integral o f equation 12.12 does not exist for this function. To see why, one must study the
Laplace transform integral closely, i.e..
Observe that e~j^* = cosicot) -jsm{(Ot) is a complex sinusoid. As f approaches infinity, the real and
imaginary parts o f the integrand in equation 12.13 must blow up, due to the
term. Hence,
the area underneath the curve e^~^ grows to infinity, and the integral does not exist for any value
of a.
W h enever//) is piecewise continuous, a sufficient condition for the existence o f the Laplace trans
form is that
I
f.
\f{t)\<kx^^
and
(12.14)
This bound restricts the growth of a function; i.e., the fimction can
not rise more rapidly than an exponential. Such a fiinaion is said to be exponentially bounded. The
condition, however, is not necessary for existence. Specifically, the transform exists whenever the
integral exists, even if the function/f) is unbounded. W ithout belaboring the mathematical rigor
underlying the Laplace transform, we will presume throughout the book that the functions we are
dealing with are Laplace transformable.
Question 5: Why does the existence o f the Laplace transform integral depend on the value o f a , men
tioned in the answer to question 4?
Answer: If the condition in equation 12.14 is satisfied, then there is a range o f a s (recall that s =
a + yoo) over which the Laplace transform integral is convergent. This is explained in the follow
ing example.
EXA M PLE 12.4. Find the Laplace transform o f the unit step. By equation 12.12,
{u(i)\ = U(s)= \Z
O'
(12.15)
1
a +
+ 7 (0
.V
-* 0 as t
neath the curve finite. For a < 0, the Laplace transform integral will not exist for the unit step.
The smallest number Oq such that for all a > Oq the Laplace transform integral exists is called the
abscissa o f (absolute) convergence. In the case o f the unit step, the integral exists for all a > 0;
hence, Oq = 0 is the abscissa o f convergence. The region a > 0 is said to be the region o f conver
gence (RO C) of the Laplace transform o f the unit step. Figure 12.14 illustrates the R O C for the
unit step.
j (o-axis
-f->a-axis
FIGURE 12.14 Region of convergence, a > 0, o f the Laplace transform of the unit step function
(i.e., the Laplace transform integral will exist for all a > 0).
Question 6: Is the unilateral Laplace transform valid only in its region o f convergence?
Answer:
the answer is no. There is a method in the theory o f complex variables called ana
lytic continuation which, although beyond the scope o f this text, permits us to uniquely and anal)aically extend the transform to the entire complex plane.- Analytically means smoothly and also
that the extension is valid ever)^vhere except at the poles (to be discussed later) o f the transform.
Thus, the region o f convergence goes unmentioned in the standard mathematical tables o f one
sided Laplace transform pairs.
lu t io n
poo /
Ke~^ e~^dt = K
s + ci
(12.16)
The integral exists if Re[j + ^] > 0. If^7 is real, then the R O C is a > -a. As mentioned in the answer
to question 6, by analytic continuation, F{s) = M{s + a) is valid and analytic in the entire com
plex plane, except at the point s = -a . The point s =-a\s a pole o f the rational function M{s + a)
because as s approaches -a , the value o f the function becomes infinitely large.
The preceding discussion and examples set up the mathematical framework o f the Laplace trans
form method. Our eventual focus rests on its application to circuit theor)\ which builds on two
fundamental laws: Kirchhoffs voltage law (KVT) and KirchhofTs current law (KCL). KVL
requires that the voltage drops around any closed loop sum to zero, and KCL requires that the
sum o f all the currents entering a node be zero. For the Laplace transform technique to be useful,
it must distribute over such sums o f voltages and currents. Fortunately, it does.
Linearity property: The Laplace transform operation is linear. Suppose j{t) =
Then
L [ m = L [ a jl{t) +
= ^,[/i(r)] + a.L\f,{t)]
= a^F^{ s ) ^a, f , St )
(12.17)
= l
O'
This is precisely what equation 12.17 states. Hence, our curiosit)^ satisfied, we may rest peaceful
ly in the knowledge that the Laplace transform technique conforms to the basic laws o f KVL and
KCL.
E X A M PLE 12.6. Find F{s) wheny(r) = K^ti{t) +
So
and a.
lu t io n
is ]/{s + a) by equa
+- ^ ,
s +a
By analytic con
tinuation, the transform is valid in the entire complex plane except at the poles, s = 0 and s = -a.
(Henceforth we will not mention the RO C in our calculations.)
J{t) = -2u(t) +
(iil)
AN SW ERS: (/) - .
,v~ - (I~
- 2e~^^u{t), and
2
.V/- A
...
.v - (/
2
, + -----,
4
The transform integral o f equation 12.12 has various properties. These properties provide short
cuts in the transform computation o f complicated as well as simple signals. For example, the
Laplace transform o f a right shift o f the s i g n a l a l w a y s has the form e~^^F{s), T > 0. Shifts are
important for two reasons:
1.
2.
Many signals can be expressed as the sum o f simple signals and shifts o f simple signals.
Excitations o f circuits are often delayed from t = 0.
[ f { , - T l u i i - T ) ] = r 'rF{s)
(12.18)
Verification o f this property comes from a direct calculation o f the Laplace transform for the shift
ed function, i.e.,
L\S(t -T )u U - D 1 = r
fUl)e~'^e~^'^dq =
f /(qye-^ 'dq =
O'
F^s)
Observe that if T < 0, the property fails to make sense, since J { t - T )ii{t- 7) would then shift left.
Since the transform ignores information to the left o f 0 one cannot, strictly speaking, recover
J{t) from the resulting transform.
Exercises. I. Find L\J{t - T)] w hen/f) is (i) Ad(t), (ii) Au(r), and (iii) Ae~''u(t).
2. pyU) = /iu(t) - Aii{t - T).
A N SW ERS: In random order:
-----
.\----- ^
s+ a
E X A M PLE 12.7. Using the tim e shift property, find F{s) for the signal
12.15.
sketched in Figure
f(t)
3
2
1
-1
-2
3 5e~^
2e~
Direct application o f linearity and the time shift propert) yields F {s) = ------------f- -
Exercises. 1. Find the Laplace transform o f the pulse signal o f equation 12.7.
2. Find F{s) when/^) = A^u[t - T^) + A2 ^(t - T-^ + A^u{t - T^).
1
AN SW ERS:
~sl
v7i
.4
---------------------
One more property allows us to revisit the signals discussed in section 3 and take their Laplace
transforms. The new propert)^ is multiplication of//) by t. This always results in a Laplace trans
form that is the negative o f the derivative o f F{s).
M ultiplication-by-f property: Let F{s) =
Then
[tf{t)\ = - F(5)
ds
(12.19)
Verification o f this property follows by a direct application o f the Laplace transform integral to
with the observation that te~^^ = ------ In particular,
ds
S60
-St
oo
n o
j()
dt =
.d s
as
ds J
Table 12.1 lists this transform pair, as well as numerous other such pairs, without mention o f the
underlying region o f convergence. As mentioned earlier, we shall dispense with any mention of
the ROC, assuming that all functions are zero for r < 0.
EXA M PLE 12.8. Find the Laplace transform o f the ramp function, r(r) = tu(t).
S o l u t io n
M ds \ s ,
R {s)= L\r{t)\ =
( 12.20)
ds
S o l u t io n
The quickest way to obtain the answer is to apply equation 12.19. Specifically, since
^ r nt , , T
s+a
L
ds s + a
ds
{s + a ) -1
( 12.21)
F {s)= L
where v = t and dti = e
=:
te
oo
fOO
ydii = uv
oo
_
0"
O'
udv
dt. Thus,
oo
le
.V+ a
dt
The RO C is a > -a , in which case the first term on the right-hand side is zero. Thus, in this RO C ,
evaluation of the integral term implies that
F {s)'0
s+a
dt =
(s + aY
\t''e-^^u(t)] =
nl
(s +a)71+ 1
(1 2 .2 2 )
561
f(t)
From Example 12.2,/ r) = 2r(^) - 2r(t - T) - 2 T u { t - 3 7 ). Hence by linearity, the time shift prop
erty', and equations 12.15 and 12.20,
2 - 2 '^
Exercises. 1. Note that the sawtooth o f Figure 12.17 is/ f) = t[u{t) - u {t- 1)]. Suppose
= z/(r)
d.s
f(t)
2. Use equation 12.22 to compute the Laplace transforms o f/ ;) = tr{t) for the ramp function r{t)
and forjit) = p-r{t).
.s'
AN SW TRS: .
562
EX A M PLE 12 .1 1 . The circuit o f Figure 12.18a has two source excitations, /j(/) and /2W> shown
in Figures 12.18b and c. Compute
V o Jt)
(a)
(c)
FIGURE 12.18 (a) Resistive circuit driven by two current sources.
(b) Triangular signal, /,(r), in A. (c) Pulse signal, ijit), in A.
S o l u t io n
and ijit) = \.5u{t) - \ 5 u {t- 1) A. From linearity, the time shift propertys and the previously com
puted transforms.
2
h{s) = - ~
S
Step 3. Find V..Js). Since V
l + 2e
s'
and l 2 {s) =
your(s) = - - ^
+ e~^
20
l5^
S63
Step 4. As an introduction to the next section, by inspection we can compute the time function
o f the output voltage:
= 35(/) - 20r{t) - \5 u (t- 1) + 2 0 r{t- 1)
>r e
(ii)
=e
+e
+ te
+ te
2) + (t - 3).
- --------h
.v+ ^/
2. Recall that cos(cor) = ------------------ . Show that the Laplace transform ofJ{t) = cos{cot)u{t) is
. r + 0)
3. Recall that sin(CO/) = ------------------ Show that the Laplace transform o f/ r) = sin(ct)f)z^(f) is
(0
F{s) = - ------,
4
4
4. Find the time functions associated with Fi ( j) = , F) (s) = ----------^
^
(s + 2)-
= Atr"H{() , /^(/) =
4 . -
= ---------.
-v+ 4
- 4)
We end this section with Table 12.1, which lists a number o f Laplace transform pairs. Some o f
these will be developed later in the chapter and some in the homework exercises. We will refer to
this table in the next section when computing inverse transforms.
.V
KV4
Item Number
m t)
Ku{t) or K
KIs
m t)
KlP+1
\I {s+ a)
]/{s+ dp-
(OqUP- + OJ^)
cos{ci)Qt)u{t)
10
s!{P- + co^)
(Oo
s2
~>
(i + a) + coq
is + a)
11
{s + Cl)~ +
(Oq
2(0o^
12
(.v^ +toi5)
t C O s{0 )Q t)u {t)
13
is +0)o)
sin((W()/ +
14
(j>)tiit)
s~ +(0n
.vcos((t))-a)o sin(({))
0
1
+ (Oq
15
16
te
smiO)Qt)uit)
17
te
cosi(OQt)uit)
s +a
2(0
is + a)~-(OQ
((5 + rt)^ +C0n)^
2(0o
18
[(5+C/)^ +(OoJ
19
20
21
Cl cos(coor) +
2^I a ^ + B ~
Cj ~ C\Ci
sin(o)o/) n it)
(.? + ) +0)^
A + jS
C|5' + C 2
' B^
A + jB
.A)j
[s + a + ;cOo)
A j B
A -jB
{s + a - ycoo)"
/ ( ,) = r V u ) , = ^ J ^ f ( . ) e V ,
over a particular path V in the complex plane. T he path F is typically taken to be the vertical
line Oj + jio where OJ ranges from -oo to +00 and Oj is any real number greater than Oq, the
abscissa o f absolute convergence.
This integral uniquely reconstructs the time structure o f F{s) to obia.\n J{t) in whichy(r) is zero
for ^ < 0. Conceptually, the process resembles the reverse action o f a prism, to produce white
light from its spectral components. An appreciation for the power o f this integral requires a
solid background in complex variables and would not aid our purpose, the analysis o f circuits.
In fact, the evaluation o f the integral is carried out using the famous residue theorem o f com
plex variables. Further discussion is beyond the scope of this text.
Just as the Laplace transform is linear, so, too, is the inverse Laplace transform, as its integral
structure suggests, i.e.,
[fit), f(j)} is uftique, where by unique we mean the following: let F^{s) = L\f^{t)] and Fjis) =
Z[^(r)] coincide in any small open region o f the complex plane. Then F^{s) = Fjii) over their com
mon regions o f convergence, and/j(f) = f-y{t) for almost all r > 0, Almost all means except for a
small or thin set o f isolated points that are o f no engineering significance. Hence, there is a oneto-one correspondence between time functionsy(f) for whichy(f) = 0 for f < 0 and their one-sided
Laplace transforms. Linearity and this uniqueness make the Liplace transform technique a pro
ductive tool for circuit analysis.
Virtually all the transforms o f interest to us have a rational function structure; i.e., F{s) is the ratio
o f two polynomials. Rational functions may be decomposed into sums o f simple rational func
tions. These simple rational functions are called partial fractions and their sums are known as par
tial fraction expansions. Two o f the more common simple terms in partial fraction expansions
have the form K b and K!{s + a). Such simple rational functions correspond to the transforms o f
steps, exponentials, and the like. Table 12.1 lists these known inverse transforms. With the table,
direct evaluation o f the line integral in equation 12.23 becomes unnecessary. Our goal is to
describe techniques to compute the simple rational functions in a partial fraction o f F{s). Once
these are found, the transform dictionar)- in Table 12.1, in conjunction with some well-known
properties o f the Laplace transform, will allow us quickly to compute the time function y(r).
^ (^ )_
+--- + biS + bo
5" +
where m s and p^, ... ,
is - Pi )(.v - 7^2
- Pn )
+ ... +
+ l?Q, and are called the finite poles o f F{s). For the most part, rational functions are sufficient for
the study o f basic circuits. There are three cases o f partial fraction expansions to consider:
(1)
j;
(2)
(3)
the case o f complex poles. Although case (3) is a subcategory o f case (1) or (2) or both,
j\ and
F(s)=K +
(S-Pi)
{S-P2 )
++
(s-p)
then
(12.24)
where
K = lim F(s)
(12.25a)
5-400
The numbers Aj in equation 12.24 are called the residues o f the pole p- and can be computed
according to the formula
A =
(12.25b)
The rightmost equality o f equation 12.25b is valid only when the numerator factor {s - p ) has
been canceled with the factor [s - p ) in the denominator of F{s)\ othenvise, one will obtain zero
divided by zero which, in general, is undefined. As intimated earlier, this partial fraction expan
sion should enable a straightforward reconstruction o f/ r). Indeed, from Table 12.1, we immedi
ately conclude
f i t )= Kh{t) + A^e^^'uU) +
) + -!- Ae^u(t)
(12.26)
jr(^+ a )
S o l u t io n
The solution proceeds by executing a partial fraction expansion (equation 12.24) on F(s) to pro
duce the Laplace transform o f two elementar>- signals, a step and an exponential. Specifically,
F{s) = ----------- = - - h
5(5 + a )
s +a
S6 '
where Ah is the Laplace transform o f a weighted step, Au{t), and B!{s + a) is that o f a weighted
exponential,
To find A, multiply both sides by^, cancel common numerator and denom
inator factors, and evaluate the result at j = 0, to produce A = Ma. Similarly, to find B, multiply
both sides by i
cancel common numerator and denominator factors, and evaluate the result
Hence,
I ^Cl
Jit) =
3. Find a partial fraction expansion o f /r^y) =
^A +
s+a
5(i' + a)
AN SW ER: K = c . A = - , B =
(I
V ..
'>
80>v^,
By voltage division,
56H
Step 2. Construct a partial fraction expansion ofV^^^^{s). Since the numerator and denominator are
both o f degree 2,
16.v" + 2 4 i-+ 1 6
.v(.v + 2)
= K + + ------.v + 2
(12.27)
The value o f K in equation 12.27 is determined by the behavior o f F(s) at infinity (equation
12.25a), i.e..
K - lim
^ 1 6 i + 24.v+ 16
i ^OO
i(A- + 2)
Ks + /4 +
,v=()
Bs
.sT 2
.v=0
Similar!),
16.v- + 2 4 a + 16
Ii =
= -1 6
i= -2
v v , ,( o = r '
1
8
161
' 8 '
16
16 + -------------- = r ' [ Li 6 Ji + r
- r
.V .v+ 2.
s
..v + 2.
AN SW ER:
*v(.v + )
+ {a + b + c ).v + b
.v(.v+l)
A N SW l'R: /-,(/) =
+ \hn{t) + -\ce~u(t) A
ijt)|
1Q
40
F{s) =
So
-e
lu t io n
r '
.v(.v+l)
By the linearit}' o f the inverse Laplace transform,
/rt = ( l - e - O u W - ( l - r < ' - ' ) ( ; - ! )
A sketch
^](r) - [1 -
!)
F{s) =
fijs)
is-a)^d(s)
S70
tor in the denominator o f the rational function F{s), and n{s) is the numerator o f F{s). The struc
ture o f a partial fraction expansion with repeated roots is
Ak
FCv) = - ^ +
(s-af
(12.28)
are unknown constants associated with s - a, ... ,(s - a)^, respectively, and ^i^(s)
and <^(s) are whatever remains in the partial fraction expansion o f f(s). The formulas for comput
ing the y4y o f equation 12.28 are
A k=(s-arF(s)
n(s)
-*i-a
(12.29a)
f/(5)
n{s)
[d(s)
^A-l = y ( ( . v - / 'F ( ^ ) )
(12.29b)
and, in general,
1
/!
ds '
njs)''
(s-afF(s)
(12.29c)
O f these expressions, only the first looks like the case with distinct roots; the others require deriv
atives o f (.f - a)^ F{s). Computation o f high-order derivatives borders on the tedious and is prone
to error. The above formulas, equation 12.29c in particular, are included for completeness.
Computer implementation circumvents these difficulties. An example that illustrates the use o f
the preceding formulas, as well as a usefiil trick, comes next.
EXA M PLE 12.15. The goal here is to illustrate the computation
F(s) =
when
s +2
.v (.v + l)
.V
.s-
.v + 1
(.v+l)'^
(12.30)
The two easiest constants to find are A2 and Bj, as their calculation requires no differentiation.
From equation 12.29a,
A, =
s-Fis)
-I
5-1-2
-5=0
=2
i=0
and
B 2 = ( s + \r F (s )
j'-i-2
.v=-l
_ s~
.9 = - !
Finding A^ and
is more difficult, since formula 12.29b requires some differentiation.
According to equation 12.29b,
as
571
ds
5+2
_ d
(5+l)-_
y4|5 +
ds
5=0
^ B|iA-y H----------- h
(5+1)
S+\
i= 0
Observe that, on the right-hand side, it is not necessary to differentiate the terms that contain A 2 ,
and ^ 2 since these terms disappear at / = 0, as the formula for
d
ds
'
5+2
(,v+l)2
.v=0
(5+1)"
^+2
'
requires. Consequently,
= -3
(5+l)-\
i=()
Similarly,
Bx=ds
0
( 5 + l ) ^ F ( 5 )
_ d
s= - 1
ds .
+2
5
=
. A--1
^s + 2'
.5 * '
S'
= 3
. ,v=-l
Aj, and B2 were known, a simple trick allows a more direct computation o f
Bji merely evaluate equation 12.29 at j = 1 (in fact, any value o f s, excluding the poles, will do),
0.75 = - 3 + 2 + 0.25 + 0 .5 5 j
As expected, solving yields Bj = 3. Hence
-3
2
3
= + ^ + ------ +
i (.v + l)5 s + \ (5 + 1 )2
F(s) =
.v+ 2
The above result can also be found with the MATLAB command residue as follows. Let F{s) =
+ r . In MATLAB,
num = [1 2];
>>den = [ 1 2 1 0 0 ] ;
[r, p, k] = residue(num, den)
The answers from MATLAB are:
r = -3 2 3 1 (the residues associated with the poles)
p = 0 0 -1 -1
and constant
k = 0.
2 5 ^ + 2 r S 3 ^ + 35 4-2
.v^(.9+n^
F{s) = -
3s^ + \0s + 9
+ 45* + 55 + 2
A
5+ 2
+ ------ +
5+1
Compute A, B, and C
ANSWHRS: In random order: 2. 1. 2
3. Use MATLAB to find a partial l^raction expansion o f F{s) =
(Iis)
and d{s) = s{s + 1)(^ + 4). Hint: Use num = 6*poly([-2 -2 -2 2 2]) and den = poly([0 -1 -1 -4 -4]).
ANSWHR: [r,p,kj = residue(num,den) \'iclds
r = - 4 4 4 8 - 1 6 - 6 12
p = _4 .4 _| _ ] 0
,
n .r
-^8
-1 6
-6
v+l
(v + l)~
12
_______________________________ '>+ 4
(.v + 4 )
________________________
The derivative formulas o f equations 12.29 are often difficult to evaluate for complicated ration
al functions, such as
s
55^ + 955-^ + 692^^ + 2369.V- + 3715.9 + 2076
s+\
s +2
_ C ^ D\
i + 3
.y + 5
1)2
D3
(.v + 5 ) -
(.^+ 5 )'*'
two values o f j, e.g., j = 0 and s = 1, to obtain two equations in the unknowns D j and Dj.
Typically, solving the resulting two equations simultaneously is much easier than evaluating Z),
and D j directly by equations 12.29. Alternatively, one can use a software program such as MAT
LAB to compute the answers. In particular, in MATLAB:
n =[5 95 692 2369 3715 2076]
d = [l 21 176 746 1665 1825 750]
[r,p,k]=residue(n,d)
r=
-l.OOOOe+00
-l.OOOOe+00
-l.OOOOe+00
3.0000e+00
2 .0000 e+00
l.OOOOe+00
P=
-5.0000e+00
- 5 .0000 e +00
- 5 .0000 e +00
- 3 .0000 e +00
- 2 . 0000 e +00
-l.OOOOe+00
k=
573
F(s) =
n{s)
n{s)
|(.v + a ) - + ( 0 -](/(. 5)
(s + a + ju>)(s + a - j w ) d ( s )
(12.31)
Since the poles - a - j c o and - a + jco are distinct, the partial fraction expansion o f equation 12.24
is valid. Since the poles are complex conjugates o f each other, the residues o f each pole are com
plex conjugates. Therefore, it is possible to write the partial fraction expansion of f{s) as
r(s)-
.v + + yto
5 + a -y c o
(12.32)
d{s)
(12.33)
jco
With A and B known, executing a little algebra on equation 12.32 to eliminate complex numbers
results in an expression more amenable to inversion, i.e..
C\S + C->
n^is)
/?|(.v)
F{s) = ------ \2, ~ 2 + ^ 7 T = ^0(') +
(I{s)
d(s)
{s + fl) + to
(12.34)
C, =2/1
(12.35a)
(12.35b)
where
and
with A and B specified in equation 12.33. W ith Cj and Cj given by equations 12.35, it is straight
forward to show that
F'ois) =
C \S + C
{s + a)~ +oj^
=C
( C2 -C ^a \
1,
x2
(.9 + a) + 0)
to
to
{s + )"
M O = e'
C| cos(tor) +
(C2-Cici)
sin (to/) Hit)
[ to
)
+03"
(12.36)
574
Exercise. Suppose F { s ) =
. Compute/(r).
a- 2 + 4
The following example illustrates the algebra for computing C, and Cj without using complex
arithmetic.
EX A M PLE 12.16. Find/^) when
3 .r + 5 + 3
D
A + jB A - j B
D
C^s + C^
F{s) = ------------^------- = -------- + ------ =;- + ------^ = ------ + - ^ ------- (5 + 1 )(5 + 4 )
^+1
s + j2
s-jl
i +1
s +A
, ^
(12.37)
Step 1. Compute the coefficients D, C ,, and C2 in the partial fraction expansion o f equation 12.37.
First we find D by the usual techniques:
3 .r
+ 5
+ 3
.v^ + 4
= 1
s=-\
Given that D = 1, to find C , we evaluate F{s) at j = 0, in which case 0.75 = 1 + O.2 5 C2, or Cj =
- 1 . With D = 1 and C2 = - 1 , we evaluate F{s) at j = 1 to obtain 0.7 = 0.5 + 0.2(C j - 1) or, equiv
alently, Cj = 2. Thus,
V+ 1
s^ + 4
+ 4
Step 2. Compute j{t). Using Table 12.1, items 8 and 9, to compute the inverse transform yields
k =0
.y+1
^+
. A -JB
s + j2
s-jl
, l- y '0 .2 5 , l + yO.25
5+ 1
s+ jl
s-jl
(12.39)
Alternative Step 2. One must exercise caution here and note the difference between the MATLAB output and the form o f the partial fraction expansion. From equation 12.39, w = +2, A = 1,
and B = - 0 .2 5 . Again using MATLAB to obtain the form needed in item 20 o f Table 12.1,
K = 2*sqrt(A^2 + B^2)
K = 2.0616
theta = atan2(B,A)* 180/pi
theta = -1 4 .0 3 6 2
Thus
Example 12.16 illustrates not only the computation o f an inverse transform having complex poles,
but also the computation o f Cj and C , without resorting to complex arithmetic, as was needed in
equation 12.32. The trick again was to evaluate F{s) at two distinct ^-values different from the
poles o f F{s). This yields two equations that can be solved for the unknowns Cj and C ,.
5 ^ 8
Then
L[ e~^'p)] = F{s * a)
(1 2 .4 0 )
Xlc- 7 ( ' ) l =
where we have viewed the sum s + a m tlie integral as a new variable p, which leads to F{p) with
p replaced hy s + a.
EXA M PLE 12.17. Let//) = sin(wr)//(r). D efine^/) = e~" p ) = e~* s\n{iot)u{t). Suppose it is
known that
Compute G{s).
So
lu t io n
)| = L\e-"\m \ = F (s + ) =
G( j ) = |
CO
5 + (0
J{t) = e
cos{LOt)u{t).
+C0
.V+ a
ANSW I-R:
i.s +
a )-
+ (0 "
Another property o f particularly widespread applicability is the time differentiation formula. Its
utility resides not onl)' in obtaining shortcuts to transforms o f signals, but also in the solution o f
differential equations. Differential equations provide a ubiquitous setting for modeling a large
variety o f physical systems mechanical, electrical, chemical, etc. In terms o f signal computation,
recall that the velocity of a particle is the derivative o f its position as a function o f time. The accel
eration is the derivative o f the velocity. After computing the Laplace transform o f the position as
a function o f time, one finds that a differentiation formula allows direct computation o f the trans
forms of the velocity and acceleration. Also, as discussed at the very beginning o f this chapter, cir
cuits have differential equation models. For example, weighted sums o f derivatives o f the response
of- the circuit are equated to weighted sums o f derivatives o f the input signal. Therefore, a differ
entiation formula is an essential ingredient in the analysis o f circuits.
First-order tim e difiFerentiation formula: Let L\j{t)] = F{s). Then
L jfU )
(It
= sF{s)-f{Q-)
(12.41)
4/(0V' ' d t = f U ) e
y / (/ )
(It
JO'
The following examples explore some clever uses o f the first-order time difTerentiation formula
E XA M PLE 12.18.^^
Recall that 6f/) = u ( t ) . Using the sifting propert)', a direct calculation yields [(5(/)] = 1. Is
this consistent with the differentiation propert}^? Interpreting the delta function as the derivative
o f the step function and applying the differentiation formula yields
I
d
6 { l ) = - H i t ) = s u{r) - u { 0 ) = s
dt
s)
(12.42)
dt'
2s + 4
AN SW I-R: - ^-----
(.Y+ 2 r + 4
2. [sm{wt)6it)] = ?
AN SW ER: u
EXA M PLE 12 .1 9 . Suppose^r) = sin(wr)w(/) and we know (for example, from Table 12.1) that
Compute [cos(wr)//(r)] using the time differentiation formula.
F { s ) = sin((0/);/(/)
.V" + (0
So
lu t io n
(0
(0
+C0
.S - + W
^
di
(ij
-t
S~
1
(0
(!)
(I)
(0
(0
V +(0'
s' - (O'
578
EXA M PLE 12.20. Lety(r) and its derivative iiave the shapes shown in Figure 12.22. Th e goal o f
this example is to explore the relationship between the Laplace transforms o f/ r) and f'{t) in light
o f the differentiation property.
FIGURE 12.22 A pulse and its derivative for Example 12.20. Observe how the derivative o f the
pulse leads to a pair o f delta functions.
Using linearity and the shift theorem on j{t) yields
X|/(01 = | ( 0 - - 1)] = [ (f)]- X | h(/ -1)1 = - (I Applying the linearity o f the Laplace transform x.of\t) yields
l / (r)] = i : [ 5 ( 0 - 5 ( f - I ) ] = l - ^ - ^
From the differentiation formula, it must follow that L\f{t)] = sL\J{t)]. Thus,
|/'(/)l = sL U (t)] = j i d - - * ) = I demonstrating consistency.
As might be expected, the formula for the first derivative is a special case o f the more general dif
ferentiation rule:
L
cit
= s"F(s) -
(12.43)
This rule proves useKil in the solution o f general th-order difTerential equations. O f particular
use is the second-order formula:
(1 2 .4 4 )
579
The inverse o f differentiation is int^ration. The following property proves useful for quantities
related by integrals.
Integradon p r o p e r ^ Let F{s) = -[/(/)]. Then for ? > 0,
c Jo-/(9 )d? =
(12.45a)
and
F (s)
(12.45b)
As with many o f the justifications o f the properties, integration by parts plays a key role. By direct
computation (using equation 12.16),
fiq ) d q and dv = e
u=
Then
-too
rr
JO'
For the appropriate region of convergence, the first term to the right o f the equal sign reduces to
fO'
et
f(q ) d q
f( q ) d q
JO*
Since the second term to the right of the equal sign is F{s)/s, as per equation 12.45a, the proper
ty is verified.
S80
E XA M PLE 12 .2 1 . Find the Laplace transform o f the signal/f) sketched in Figure 12.23a using
the integration property.
f(t)
FIGUIIE 12.23 (a) A triangular signaly(/) for Example 12.21. (b) The derivative
S o l u t io n
Observe that the triangular waveform y{r) o f Figure 12.23a is the integral o f the square wave^^).
Since ^t) is easily represented in terms o f steps and shifted steps as
g{q)dq
.S -
EXA M PLE 12.22. This example explores the voltage-current {v-i) relationship o f a capacitor in
the frequency domain by way o f the integration property. Recall the integral form o f the voltagecurrent dynamics o f a capacitor:
1 r'
Taking the Laplace transform o f both sides and applying the integration property produces
[vcit)] =
C->
1 fO-
ic(T)dx = 7T^c(-^) +
Cs
Cs
), because
S8]
Therefore,
Cs
(12.46)
Equation 12.46 says that the voltage V(^s) is the sum of rvvo terms: a term dependent on the fre
quency domain current I^^s) and a term that looks like a step voltage source and depends on the
constant initial condition V(^0~). The quantity Z^^s) = MCs looks like a generalized resistance
generalized because it depends on the frequency variable s and a resistance because it satisfies
an Ohms law-like relationship, V^^s) = Z^^s) I({s). These analogies prompt a series-circuit inter
pretation o f equation 12.46 as depicted in Figure 12.24. An application o f this equivalent frequenc)' domain circuit to general network analysis appears in the next chapter.
\,{S)
o >+
ic(t)
O >
4-
[ ]
c ----- > VJs)
V,(t)
Cs
O 'f
O------FIG U RE 12.24 Equivalent circuit interpretarion of a capacitor in the frequency domain. This equiv
alent is arrived at by applying the integration propert)' of the Laplacc transform to the capacitor volt
age, seen as the integral of the capacitor current.
A second example interpreting the v-i characteristics o f the capacitor in the frequenc)' domain
ensues from the differentiation rule. Instead o f winding up with a series circuit, one obtains a par
allel circuit. The interpretation is thus said to be dual to the one just described.
E X A M PLE 12.23. This example has two goals: (i) Verify that equation 12.46 is consistent with
the differentiation formula interpretation o f the capacitor; (ii) Build a dual frequenc)' domain
interpretation o f the v-i characteristic o f a capacitor analogous to that o f Example 12.22.
As a first step, recall equation 12.46:
Vc(-
Cs
(12.47)
Notice that equation 12.47 is consistent with the application o f the derivative formula to i(^t) =
C[civ(Jdt]. This consistency offers some reassurance in the accuracy o f our development. The
interpretation o f equation 12.47, however, is quite different from that o f equation 12.46. In the
latter equation, the current /^j) equals the sum o f two currents, CsV^s) and -C t/^ 0). This sug
582
gests a nodal interpretation, resulting in an equivalent circuit having two parallel branches. One
branch contains a capacitor with voltage V^s). The other, parallel branch contains a current
source with amperage Cv(\Qr). The current through the capacitive branch is
where
G now acts like a generalized conductance because it multiplies a voltage, similar to Ohms law.
Q is generalized because it depends on s. Figure 12.25 presents the equivalent circuit o f the
capacitor in the frequency domain and is dual to the circuit o f Figure 12.24. Chapter 13 covers
in detail the role o f these equivalent circuits in analysis.
Ic(s)
The last elementar)' property o f the Laplace transform that we consider in this chapter is the timescaling property, also called the frequency-scaling property. Its importance is fundamental to net
work synthesis. Here, numerical problems, such as roundoff, prevent engineers from directly
designing a circuit to meet a given set o f specifications. Instead, the design engineer will normal
ize the specifications through a frequency-scaling technique. Once the normalized circuit is
designed, frequency-scaling techniques arc reapplied in an inverse fashion to obtain a circuit meet
ing the original specifications.
Time-/Frequency-scaling property: Let ^ > 0 and L\J{t)] = F{s). Then
L[J\at)\ = - F
a \ci)
(12.48)
^83
f(2t)
f(t)
A
2 --
A
2 --
i<
1
1 --
- -
> t
> t
1
FIG URE 12.26 (a) Unit impulse train, (b) Time-scaled imit impulse train. Unit impulse trains such
as these model sampling in signal-processing applications.
The time-scaled impulse train in Figure 12.26b increases the frequency at which the impulses
occur (twice as often as in the original signal). This is reflected in the Liplace transforms o f the
two signals:
'Z&U-k) = I
k=0
A-=0
\-e'
(12.49)
l - e -0 .5 5
(12.50)
Notice that what occurs at, say, Sq in equation 12.49 now occurs at 2s^^ in equation 12.50. Hence,
time scaling by numbers greater than 1 concentrates more o f the frequenc)^ contcnt o f the signal
in the higher frequency bands.
Exercise. Verify, by direct calculation, that L\J{2t)\ is given by the right side o f equation 12.50.
Several more properties o f the Laplace transform are germane to our purpose. However, these
properties have a systems flavor and are postponed until Chapter 13. We close this section by pre
senting Table 12.2, which lists the Laplace transform properties and the associated transform pairs.
Transform Pair
Linearity
L \ j{t-
Time shift
\t/lt)n(t)]=- F'{s)
as
Multiplication by t
n cr r j s )
Multiplication by t
ds"
= Rs + //)
Frequency shift
Tim e differentiation
jfO )
(it
d - f{ t )
Second-order differentiation
= sF(s)-J{0-)
= r F c v ) - 5 / ( ( r ) - / '\ ( D
dr
wth-order differentiation
T> 0
7)1 =
d''fU)
dt'
(i)X
fUl)dq
m dq
Fis)
Time integration
(ii)
Time/Frequency scaling
.-fU l)d q
Fis)
[f{at)\ = MaFista)
S8S
d'\-
+ a..V.
dl
for constants
dt'
+b
and bj might model the behavior o f a linear circuit. We may use the following steps
1.
Take the Laplace transform o f both sides o f the equation, using the appropriate deriva
2.
3.
4.
If the equation is an integro-differential equation, i.e., a mixture o f both derivatives and integrals
of the input and output signals, then we simply apply the same algorithm, except we use the inte
gral formula where appropriate. Some examples ser\-e to illuminate the procedure.
EXA M PLE 12.25. Consider the pulse current excitation o f Figure 12.27a) to the RC circuit o f
Figure 12.27b. The goals o f this example are (i) to use and illustrate Laplace transform techniques
to solve a difTerential equation derived from a simple RC circuit and (11) to find the response volt
r > 0, when V(^0~) = 1 V.
age
F IG U R E 12.27 Excitation currcnt (a) fora simple /?Ccircuit (b) for Example 12.25.
S o l u t io n
X [/ (0 ] = 0 .5
\ -e
Step 2. Find the circuits dijfereutial equation model that links the excitation current i{() to the response
d\c {t)
dt
dv({f) + v c ( f ) = 2 i(t)
dt
Step 3. Take the Laplace transform o f both sides, apply the differentiation rule to the left side, and solve
for V(i{s). Applying the Laplace transform to both sides yields
v r(0 )
.v+1
i + l
\-e ~ ^
= -----------+
^(i-i-1) .v + 1
1 _
5(.v+ l)
1
(.v+1)
Thus, with the aid o f the shift propert)' and the transform pairs o f Table 12.1, we obtain
.v(i + l)
s+\
v^^t)
S8
EXA M PLE 12.26. The goal o f this example is to compute the response, denoted here by the
given the scries RLC circuit o f
Figure 12.29. Suppose the initial conditions are /^(O-) = 1 A and V(^0~) = -2 V.
TYYY
40
1H
FIGURE 12.29 Series RLC circuit for Example 12.26. Here the current //(^) =
So
lu t io n
Step 1. Compute the Laplace transform o f the input. From the tables or by inspection, X[6(r)] = 1.
Step 2. Compute the integro-differential equation o f the circuit o f Figure 12.29. The first task is to
sum the voltages around the loop to obtain
Substituting for each o f the element voltages using the mesh current, ij{t), yields the desired integro-difFerential equation,
diin
(12.51)
Step 3. Take the Laplace transform o f both sides, substitutefor R, L, C,
solve for
Vf^Qr), and
and
W ith the aid o f the differentiation and integration formulas, taking the Laplace
(s) +
= V;(s)
L ------ 1----- ^ h U )
= K (S ) + i , / t ( 0 - ) -
V>-(0 )
produces
I
') ~
5 + 4 5 + 4
5 +2
5+ 2
( 5 + 2 )-
Step 4 . F in d iinit). Taking the inverse Laplace transform yields the desired result:
i d t ) = (2 - 2,)e-^-u(t) A
588
cli,C
dl
C oo
with C = 1 F,
L.(s) r . ( ( ) )
ANS>X^R: .v/,-.(.v) + -^----- + -^-------- = 0
^
.S
.V
dxU)
dt
y{z)dz = 2ii(t)
where x(0 ) = 2 , u{t) is the unit step function, and 8{t) is the Dirac delta function, then findATi).
AN SW ER: X{s) = V
EXA M PLE 12.27. The final example o f this chapter looks at the leaky integrator circuit o f
Figure 12.31, which contains an ideal operational amplifier (op amp).
resistance o f the capacitor. Given C and
SH)
R, = IMegO
V (t) = 5 u (t )
''ou.W
lu t io n
Step 1. Determine the dijferential equation. Since the op amp is ideal, it follows that ijr= -i^. From
O hm s law, i^ = vJR^^. On the other hand,
/?2
dt
dt
I<2C
R^C
cuit works as a simple integrator. The circuit is called a leaky integrator because /?-,C is large but
finite. Since /?, C = 1, one expects the gain constant to be 1 as well.
Step 2. Substitute values, take the Laplace transform o f both sides, a7jd solve for Vg,,f{s). Taking the
Laplace transform o f both sides, one obtains
'w ( 0 " ) +
Since
=- -
= 0, it follows that
-5 0
5( 5 + 0 . 1 )
Step 3. Invert
to obtain
.V
50
.9 + 0 .1
^92
'Problems
(a)
(b )
(d)
(e)
r > 0
(c)
r>0
f^{t) =
T>0
(a)
(h)
/ 2 W = / / - 7 ) 6 U - 7-,)
(c)
(d)
/^W = A-,6W + 7 - )
(c)
(f)
(a)
(b )
f,U)
sin(2yTr - 2jr)u{t - 1)
(c )
K^6(t-T)
/sW = c o s h (2 / )a (t-2 )^
= s in ( 2 . T / ) ( r ) -
(d )
r)
sinh(/ - 2 r , ) 6 ( f - r , )
(g)
f,{t)
needed.
(a)
2K -
-Au{t-A)
(b)
f^{t) = 2 r { t ) - l r { t -
l ) - r ( r - 3 )
At - 5)
->
(c)
(d)
t(s)
j\{t) = cosiS)5m)u{t) +
c o s (0 .5 > -r(/ -2 )) a -2 )
(a)
(e)
f,(t)
2K -
f^{t)
fJ,t)
f^{t)
f^[t)
f^{t)
= ti{t) - l i t - \ ) .
= u{t)-r{t-\)-^r{t-2)
= 2u{t) u{2 - t).
= Ar{t) u{\ - t)
= 2rit)u{\ - t) + u { t - \)r{2 - t)
>
2
(b)
A
K
f^{t) = Ktu{t-\)
(b)
(c)
(d)
f^{t) = Ktr\t-\)
f^{t) ^ K{t - \)r{t)
ir
(c)
t(s)
S93
f,(t)
2K
K
> t
>
2T
T
(d)
i k f^ft)
-2
f,(t)
K
> t
> t
-T.
-K
(e)
f(t)
(d)
^-------1-------1------ r
1
>
f,{t)
5
4K -
(f)
2K
-I---------- '----------1--------- 1 t(s)
1 2
3
4
(a)
(a)
(b)
5 94
r>
- t
r^
(0
11. For the circuit o f Figure P I 2 .11, suppose
/?, = 6 0 0 Q, Rj = 1000 Q, and
= 1500 Q.
(!)
Figure P I2,11
vJt)
f,(t)
2
>
1
(a)
2
> t
o
Figure P I2.12
(a)
t
o
r^
(e)
W )l
15. Prove the time-/frequenc)'-scaling property
by direct calculation o f the Laplace transform
integral.
16.(a)
(b)
Figure P I2.17
complex exponentials
(b)
(c)
18.(a)
Represent
sin(7cO
0<t< 2
otherwise
(b)
>
~>
S +(D
S)6
Figure P I2.18
19.(a)
22. (a)
(i)
- e~^, find
Idt
f(X )dx
(c)
as a sum o f appropri
(2)
Express
{\) [cosh(/7r)]
(2) L [sinh(^r)]
(b)
[Kt sinHat)]
(b)
=J{t + 4).
f(t)
compute
{\) [K r cosh(/7r)]
(2) [Kr
3 -
s m h ia t)]
2 -
2.V + 4
s +1
> t
Figure P I2.22
F{s) =
and
(.9+1)-
W
W)
^^W = 2 ^ , ( 2 f - 4 )
(c)
^5W = 2 ( r - 2 ) / ; - 2 )
(f)
\-e
is given as
-(s-a)
s-a
(a)
(b)
AN SW LR: (a)
.s'
24. The Laplace transform
F{s) =
(a)
-V
J{t) is given as
l-e "_ ^
cm
cit
(b)
(c)
dt
597
- t
(c)
(d)
- t
It is not necessary to have the answer be a
rational function.
27. (a)
W
O '
O
sketched
in
Figure
P12.27a.
(b)
Figure P i2,28
sketched in
O
->
o
Figure P I2.27
598
(a) F,Cv) =
s{s
a){sb)
C H E C K : One residue is a.
^
- 7.v^ + 4.V + 2
( b ) F ^ s ) = ---------------------------- ^
Figure PI 2.29
30. Let j{t) and ^t) be as sketched in Figure
P I 2.30. Find G{s) in terms o f F{s).
f(t)
20
10
C H EC K : One residue is - 2 .
. X
s 2 / + 18i-^ + 4 6 5 ^ + 4 4 5 + 12
(c) F2,{s ) = --------------------r---------- z----------(5 + l ) ^ 5 + 2 r
C H E C K : Two residues are at 2 and - 2 .
n = [2 -7 4 2];
d = conv([l -1],[1 -4 4]):
Figure P12.30
INVERSE LAPLACE
TRANSFORMS BY PARTIAL
FRACTION EXPANSION
31. Using partial fraction expansions and your
knowledge o f the Laplace transform o f simple
signals, find j{t) when F{s) equals
(a )
2 5 ^ + 1 3 r + 305 + 32
[r,p,k] =residue(n,d)
34. Find (i) the simplified partial fraction
expansion and (ii) the inverse Laplace trans
form for each o f the following functions by
hand. Show all work.
(a) Fi(.v) =
{a + b ) s + l a b
(5 + fl)(5 + b )
( a + b + c ) s ' + ( b e + 2 a b + a c ) s + a b c
s { s + a ) { s + b)
s{s^ + 6^ + 8)
C H E C K : One residue is a.
(b)
- s-6
Cv + 2 ) ( . v - - l )
(c)
cs~ + { a + 2 a c ) s +
( r +25 + 1)( j + 2)
(d ) F4(s ) =
(d)
il + c ) a ^ - a
(c) F3(.v) =
/ + 1 2 A - ^ - 2 4 r - 3 2 .y + 16
(e)
55^ + 1 4 4 5 + 2 0 4
(5+1)
(5+ 2) + 6 4
599
'w '
(a)
(b)
(c)
(d)
for the
input current
25 + 16
+16
45 + 9
/X
245-72
5^ + 4 5 + 40
25^ + 885
(5 + 4)(5^ +64)
2 5 ^ + 2 5 ^ -2 5 -6
v.(t)
+
20
80
24 0
C H EC K : Residue at j = - 2 is - 1 .5 .
(b) F2(5) =
25^ + 9 5 ^ + 1 6 5 + 1 1
(5 + i)(5 + 2 y
5^ + 4 5 ^ - 2 5 ^ - 9 5 - 3
( 5 - 1 ) 2 ( ^ + 2)2
CHECK: Two residues are at I and two are at - 1 .
4 5 ^ - 1 2 5 ^ + 325 + 16
(d) /=4(5)=r
(5 + ir+ 1
it follows that
(5 + 2)^ + 9
fj) =
(5 + 2)2 + 16
39 .
The
Laplace
[A^jf'^'cos(fi)?) +
transform
of f j )
sin(fi)r) + K^e~^^u{t) is
- 5 2 5 + 228
F {s)
(5 + 4 ) ( 5 + 0 ^ + 100
Find a, b, K^,
K y and 0).
input voltage
40. Consider the resistive circuit in Figure
P 12.40. Use Table 12.1 and the shift property
5(5 + 4 )
(b)
to find
of
600
(a)
Kv,(.v) =
IOf'"'' +
(b)
\-e'
Vi{s)=\0
- 5e'
..-4 5
-2 0 -
(b)
(c)
8kQ
-n/ V ^
8kQ
(d)
for r > 0.
1 8 kQ / 9kQ,
(e)
use superposition?
(0
(c) F^{s) =
(d) 74(' ) =
(e) F^{s) =
= 10//(r
(b) F2(s ) =
= 10//(r)
Figure P i2.40
(a) F,(5) =
= 0, find
3.v'^ + 3 0 5 - + 86.9 + 6 4
'">6
.s'* + 8 i - +20.V+ 16
-O
+
v,(t)
- 4 6 .2 5 s - 6 9 2 . 8 125
.s-'* + 14.5.V- + 169.5625s + 510.25
-2.s-^ + 23.V- - 68.V - 3265
Figure P i2.42
(h'rU)
1
1
^
+ ----- VcU) --------,ll
RC
RC
(b)
30 Q. and L = 0.2 H.
(a)
Show that the differential equation for
R. , .
1
, .
, ({) = - ViU)
L
L
= 10(r) V, valid
for t > 0.
the circuit is
di l it )
(If
= 0.
601
(a)
Find
= 10/^r) V.
- 1 0 and
R
v jt)
vJt)
-O
i,(t)
ijt)
\r
Figure P I 2.43
;r
v^(t)
= lu{t)
ANSWl-R; (a)
(c)
(e)
- \V' = 0 . ( b ) (I) = 7 =
LC ^
J lC
= 0 and /^(0 ) = 0.
Construct the differential equation o f
assuming /^(0 ) = 0.
voltage,
assuming /^(0 ) = 0.
(d)
dr
(b)
Fieurc P I2.45
(b)
show that
for t > 0.
10 -
10
c o s(0
.5 jia )
for t
l = ih
> 0 by superposition.
^
' ) Q
R.
' . ,(t)
4C=-,F
vJt)
Figure P I2.46
47. A pair o f (coupled) differential equations
Figure P i2.44
= V^. Since
and
dt
lossless.
dt
-t-^/|.v(/) = oiyit)
602
= 2,
= 2,
= 4, b^ = 3, and/r) = 2u(t).
(b)
and
then
(a)
(c)
find
kis).
(b)
ci\\
d\ 'c
(It-
RC clt
LC
''C
LC
Figure
12.1
o f Example
12.1, where we
_
Vc(A) =
RC
(c)
( .v + 5 ) v c ( 0 " ) -h v c (0 " )
I
1
s~ + ----- 5 +
1
s + -----
RC
LC
) = 0,
equations
LC
(a)
(b)
L=1H
v(t)
;
R=0.8Q I
ijt)
ditions.
If
v,(t)
(d)
for /> 0.
CHAPTER OBJECTIVES
1.
s,
y\s).
isfy a type o f O hm s law. These ideas are generalizations o f the phasor-based notions o f
impedance and admittance introduced in Chapter 10.
2.
Learn the arithmetic o f impedances and admittances in the Laplace transform domain,
which is analogous to the arithmetic o f resistances and conductances in the time domain.
3.
Apply the new concepts o f impedance and admittance to redevelop the notions o f volt
age/current division, source transformations, linearity, and Thevenin and Norton equiv
alent circuits in the /-dom ain.
4.
Define /-domain-equivalent circuits o f initialized capacitors and inductors for the pur
pose o f transient circuit analysis.
5.
6.
7.
8.
of
9.
Introduce the notion o f a switched capacitor circuit, which has an important place in
real-world filtering applications.
10.
SECTION HEADINGS
1.
Introduction
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
Summary
12.
13.
Problems
1. INTRODUCTION
Chapter 12 cultivated the Laplace transform as a mathematical tool particularly useful for circuits
modeled by differential equations. This chapter adapts the Laplace transform tool to the peculiar
needs and attributes o f circuit analysis. W ith the Laplace transform methods described in this
chapter, the intermediate step o f constructing a circuits differential equation, as was done in
Chapter 12, can be eliminated.
Available for the analysis o f resistive circuits is a wide assortment of techniques: O hm s law, volt
age and cu rren t division, nodal and loop analysis, linearit)', etc. For the sinusoidal steady-state
analysis o f RLC circuits, phasors serve as a natural generalization o f the techniques o f resistive cir
cuit analysis. The Laplace transform tool permits us to extend the sinusoidal steady-state phasor
analysis methods to a much wider setting where transient and steady-state analysis are both pos
sible for a broad range o f input excitations not amenable to phasor analysis. Recall that transient
an;ilysis is not possible with phasors.
T he keys to this generalization are the i-domain notions o f impedance and its inverse, admit
tance. Instead o f defining impedance in terms ofyoj, as in phasor analysis, we will define it in
terms o f the Laplace transform variable
s. This
j-dependent O hm s law, j-dependent voltage and current division formulas, and ^-dependent
nodal and loop analysis; in short, all o f the basic circuit analysis techniques have analogous
s-
dependent formulations. W hat is most important, however, is that with the ^-dependent formu
lation, it will be possible to define .^-dependent equivalents for circuits containing initialized
capacitors, inductors, and other linear circuit elements. These equivalent circuits make transient
analysis natural in the i-domain.
In the final section o f the chapter, we introduce the notion o f a switched capacitor circuit.
Switched capacitor circuits contain switches and capacitors, and possibly some op amps, but no
resistors or inductors. Present-day integrated circuit technolog)' allows us to build switches,
capacitors, and op amps on chips easily and inexpensively. This has fostered an important trend
in circuit design toward switched capacitor circuits. A thorough investigation o f switched capac
itor circuits is beyond the scope o f this text. Nevertheless, it is important to introduce the basic
ideas and thereby lay the foundation for more advanced courses on the topic.
v-i relationship
o f an inductor,
at
to obtain
V^is) ^ Lsliis),
assuming /^(0 )
(1 3 .1 )
an j-domain voltage,
T he units o f Z^(j) =
eralizes the concept o f a fixed resistance, and it is universally called an im pedance. This complexfrequency or ^-domain concept has no time-domain counterpart.
Although the inductor served to motivate ^-domain impedance, in general an impedance can be
defined for any two-terminal device whose input-output behavior is linear and whose parameters
do not change with time. A device whose characteristics or parameters do not change with time
is called
time invariant.
606
IMPEDANCE
T he
Figure 1 3 .1 , relates the Laplace transform o f the current, /(s), to the Laplace transform o f
the voltage, V^j), assuming that all independent sources inside the device are set to zero and
that there is no internal stored energy at ^ = 0 . Under these conditions.
V{s) = Z{s)I{s)
(13.2a)
Z(s) =
V(s)
I(s)
(13.2b)
in units o f ohms.
l(s)
Device
vis)
Z(s) orY(s)
FIG U R E 13.1 A two-terminal device having impedance
Exercise.
ANSW'HR;
Vj,j{s) = ------
and
5 + 4
Z{s) or admittance
I- (s)
^(^).
= 2. Com pute
.y" + 4
T he inverse o f resistance is conductance, and the inverse o f impcdance is admittance. For exam
ple, if we divide both sides o f equation 13.1 by
Ls,
we obtain
Ls
This suggests that
\/Ls acts
is defined as follows.
as
(1 3 .3 )
called an adm ittance, which
607
ADMITTANCE
T he
admittance,
in Figure 1 3 .1 , relates the Laplace transform o f the voltage, V(j), across the device to the
Laplace transform o f the current, /{^), through the device, assuming that all internal inde
pendent sources are set to zero and there is no internal stored energy at f = 0~. Under these
conditions,
i{s) = ns)v{s)
(13.4a)
Ijs)
(13.4b )
V{s)
in units o f S.
From equations 13.2 and 13.4 , impedance and admittance satisfy the Inverse relationship
Y(s) =
Exercise.
1
Zis)
(1 3 .5 )
16
C om pute Ky,//).
(.S-
----
arid
+ 2)
A/l(-^)="
(>v +
2 )(i + 4 )
2
ANSWER:
^+ 4
As a first step in deepening our understanding o f these notions, we compute the impedances and
admittances o f the basic circuit elements shown in Figure 13.2.
i,(t)
ic(t)
O-
O
+
-I-
v,(t)
V ,(t)
v(t)
o >
o-
(a)
FIGURi-l
(b)
1 3 .2
(c)
Rlj^is).
= Rij^it).
Laplace-transforming both
resistor is
Zf^is) = R
608
(!v(^{ t )
dt
/c ( 0 = C-
I^s) = CsV^s)
From equation 13.4 , the ad m ittan ce o f the cap acito r is
Y(is) = a
and from equation 13.5, the im pedance o f the cap acito r is
Z eis) =
1
Cs
)=
Exercises. 1. Given the integral form o f the v-i capacitor relationship, assume no initial stored
energy and take the Laplace transform o f both sides to derive the impedance o f the capacitor. This
provides an alternative, more basic means of deriving the impedance characterization.
2. Given the integral form o f the
v-i
take the Laplace transform o f both sides to derive rhe admittance o f the inductor.
Throughout the rest o f the text, whenever we refer to an impedance the unit o f Ohm is assumed,
and similarly, admittance is assumed to have the unit of siemens (S). The units for KW and
I{s)
are usually not shown, although strictly speaking they are volt-second and ampere-second, respec
tively.
609
r u le .
I{s),
to ^-domain voltages,
K(j), and because all /-domain currents must satisfy' KCL and all /-dom ain voltages must satisf)
KVL:
1.
Impedances,
Z{s), can
o f ohms.
2.
Admittances, K(/), can be manipulated just like conductances and, like conductances,
have units o f S.
This manipulation rule suggests, for example, that admittances in parallel add. T he following
example verifies this property for the case o f two admittances in parallel.
and impedance,
o f three
Yji-s]
K,(.v)+Ko(.v)+r,(.v)
(1.3.6)
An(-v)
Z Js)
FIGURE 13.3 Three general admittances, Vjis), in parallel,
having an equivalent admittance K- (/) or impcdance
Z- (s).
(>10
S o lution
W c seek the relationship
= AW
for
^3(5) =
= 1 ,2 , 3. From KCL,
>3(5))
y i(s)=
/j^(s) =
/;,(,C) = Y / : ( S ) V , ( S ) = y , ( s ) Z , ( s ) / , ; , ( s ) =
Exercises.
Z^{s)
and
Zjis),
in parallel, Z ,(^) =
Z (5 )=
C]^+C25
(C j+ C 2 )5
2 i ( ^) +
= Cj + C 2 .
3. Derive the following formula for the impedance o f two inductors in parallel:
L^ + Lo
4. A 2 |.iF and a 0 .5 uF capacitor are in parallel. Find the equivalent capacitance.
A N SW ER : 2.5
til
ANSWER: 2.^
'
.s- + i ( r "
61 1
MR,
K,(j) =
M{L$),
and
Cs,
a resistance,
1
fin (s)
L C s ^ + -s +
R
and find
terms o f I;{s).
in'
Ai\S\V1-:R: y j s ) =
I^{s)
1
^ ^
25+1
in
Find
^
l and /^(.v) =
EXAMPLE 13.2. Compute the input impedance o f the parallel RLC circuit sketched in Figure 13.4.
o
+
VJs)
Z Js) =
Y Js)
FIG U R E 1 3 .4 Parallel
RLC circuit
S o l u t io n
r +
RC
.v+'
LC
C , 2 ^ _ 1L , +
RC
w hich is the equivalent input im pedance o f a parallel
LC
RLC circuit.
(1 3 .7 )
612
Exercises.
i)-^s + 2 x 1 0 * )
1 he dual o f the parallel circuit o f Figure 13.3 is a series connection o f three impedances as shown
in Figure 13.5. T he following example verifies that impedances in series add, and simultaneous
ly develops a voltage division formula.
E X A M P L E 1 3 .3 . Com pute the equivalent impedance,
general impedances,
Z^(s), Z^is),
and
Z^{s)
and admittance,
o f three
Zj{s)
=
+ V,(s) -
V^n(.v)
(1 3 .8 )
+ V^(s)
o l u t io n
(1 3 .1 1 )
613
T he voltage division formula o f equation 13.8 follows from a modified form o f equation 1 3 .1 0 ,
and equation 13.9, to yield
Z (s )
Z ,( 5 ) + Z 2 ( s ) ...+ Z(5)
Exercises.
AN SW I-R;
C|CS
Cj + C .
2. Show that the equivalent admittance o f rwo capacitors, Cj and C-,, in series is
3. Suppose
Zj(s)
= 10 Q,
Z-,(s)
2s,
and
Z^(s) = 6^ in
Y{s) = -------- s.
C\ + Cl
V^-)W, and
.V
A N SW ER S:
Z,,(s)
= 10 + 8^. \ s ( .0 =
4.V + 5
. /-.(/) =
'
4. Suppose
terms o f
and
-7
+ 10.v + 2
A N SW ER S: Z,(.v) = ---------------------,
V
Z t^(s )
=-
=
^
Z^-^^(s)
and K^(j) in
-------------- V;(.v)
,v' + 5 .v + !
= -^i +
O f course, there are series-parallel connections of circuit elements that combine the concepts illus
trated in Examples 13.1 through 13.3, as set forth next.
EXA M PLE 13.4. Com pute the input impedance Z-J,s) o f a series connection o f t\vo pairs o f par
allel elements, as shown in Figure 13.6, in which
Then compute
in terms o f
If
= 10 Ci, C = 0.1
u{t),
find
Vjit).
= 5 O.,
and Z. = 1 H.
614
RC elements
o l u t io n
V,(s)
Z ,U ) =
10
O.l+O.l.v
.y + 1
and
Z2{s) =
Is
7+2
in which case
10(.y + 2 )+ 2 .s;(.v + 1) _ 2^ + 1Is + 20
(5 + 1 )(5 +
(s + \)(s + 2)
2)
It fo llo w s th at
1// X
(.v + l )(i + 2 )
2.V
.v(.v+l)
2.v + 12.v + 20
-V+ 2
.v + 6 .y + I 0
Finally, if
61 S
= -,
(^+1)
(-^ + 0
^2(^') = -3 ---------------= ---------- ^
.9 + 6 . 9 + 10
( 5 + 3)" +
From Table 12.1, item 19,
Exercise.
Repeat Example 13.4 with the following changes: C = 0.01 F and /?, = 10 Q.
A N SW ER S:
Z Js )
= 10
VS(.v) =
\+10
Another basic and useful circuit analysis technique is the source transform ation property,
exhibited now in terms o f impedances and admittances. The first case we will examine is the voltage-to-current source transformation, illustrated in Figure 13.8,
(b)
(a)
FIG U R E
1 3 .8
Z^[s), as shown in
Z^(s), as shown in part (b).
Often, voltage-to-current source transformations provide an altered circuit topology that is more
convenient for hand or calculator analysis. Mathematically, the goal is to change the structure o f
a voltage source in series with an impedance to a current source in parallel with an admittance
while keeping both
Vjis)
and / 2 W fixed. To justify this, one starts with Figure 13. 8a, in which
V,(.v) = ^
^ V^^is)=Z.is)l2is)
Z ,(5 )+ Z 2 (5 )
H en ce,
ZAs) ^ 0 ,
V ;(.v) =
Z2(.V)Z, is )
(V iu (s)\
(1 3 .1 2 )
616
where
Yjis)
[Zj{s)]
^ This equation identifies the parallel structure of Figure 1 3 ./b ; i.e., Figure
(a)
(b)
FIG U R E 13.9 Illustration of (a) current source to (b) equivalent voltage source transformation.
Clearly, the manipulation of impedances and admittances parallels that o f resistances and con
ductances, as suggested earlier. Indeed, for a rigorous statement o f the soiuce transformation tech
nique developed above, refer to the source transformation theorem in Chapter 5 and replace
by
Z{s),
by
and
by
Ijp).
counterparts.
This section ends with a demonstration of finding a Thevenin equivalent in the /-dom ain.
E X A M P L E 1 3 .5 . C om pute the Thevenin equivalent circuit o f Figure 13.10.
VJ s )
V Js )
v(s)
(b)
FIG U R E 1 3.10
SOLUTIO N
From the material in Chapter 6, our new concepts o f admittance and impedance, and Figure
13.10b,
(13.13a)
o r equivalently,
o s-B b )
61
(d + 1)/(^(.V)+ GV'j-^C.v)
(13.14)
( 5 ) - ( + 1).vCV/^(.v)
(a + \)sC + G
(13.15)
'
(W +D.9C + G
(a + 1).vC + G
Exercises. 1.
A N SW ER : /.^.(.v) = U +
2. Find
(fl + 1).?C + G
and
\)s(:Vjj,s)
VgJ<s) for
..
2/ / , (. s)
2 + 2.V +
FIG U RE 13.11
3. For the circuit o f Figure 13.12, use source transformations to find I^p) and Y^jj^s) for the indi
cated terminals.
= 0.2.^ + - + 0.4
0.2 F
(V Js )
2.5 Q
1 H
F IG U R E 1 3 .1 2
618
(it
= /c (0
Taking the Laplace transform and allowing for a nonzero initial condition
Cs
yields
- Cv(J,Q~) = I (is)
(1 3 .1 6 )
T he left side of equation 1 3 .1 6 is the difference o f two currents, one given by the product o f the
capacitor admittance and the capacitor voltage
{CsVf^s))
1^(5)
^
.....................................................
V,(s)
Cs
FIGllRK 13.13 Parallel form of an equivalent circuit for an initialized capacitor. Here, the capacitor
within the dotted box is relaxed while the current source
Cv(4S)~) accounts
Cs
(1 3 .1 7 )
Example 12.22 previewed this equation by taking the transform o f the integral relationship o f the
capacitor. We observe that the right-hand side o f equation 1 3 .1 7 is the sum o f two voltages, one
o f which is the product o f the capacitor impedance and the capacitor, current and the other
z/^(0)A'. Thus, the interpretation is a series circuit, as sketched in Figure 13.14.
619
V Js)
FIG URE 1 3.14 The series form of an equivalent circuit for an initialized capacitor. Here the capacitor
in the dotted box is relaxed, and the voltage source accounts for the effect of the initial condition.
Initialized inductors have similar j-domain equivalent circuits analogous to those o f the capaci
tor. W ith the voltage and current directions satisfying the passive sign convention, the differen
tial inductor current-voltage relationship is
( 13 . 18 )
Again, this equation consists o f a sum o f voltages,
Lsl^is)
FIG U R E 13.15 Series form of equivalent circuit for an initialized inductor. Here the inductor with
in the dotted box is relaxed; notice the polarity orientation of the voltage source.
To construct a parallel equivalent circuit, divide equation 1 3 .1 8
Ls
by Ls and
rearrange to obtain
(1 3 .1 9 )
T he right side o f Equation 13.19 is a sum o f currents that determines a parallel equivalent circuit,
as sketched in Figure 13 .1 6 .
620
= Au{t) V.
+ vjt) --------- (-
o l u t io n
In this example, it is convenient to replace the capacitor by its (series) 5-domain voltage source
equivalent circuit, because the capacitor is in series with the input voltage source. On the other
hand, it is convenient to replace the inductor by its (parallel) /-dom ain current source equivalent
circuit, because the desired output is the inductor voltage. This results in a three-source or multi
input circuit. O nce the equivalent circuits are in place, one can apply superposition to obtain the
answer, although there arc many other ways to solve the problem.
Using the voltage source ynodelfor the capacitor and the current source modelfor the induc
tor, draw the equivalent s-domain circuit. Using the equivalent circuits o f Figures 1 3 .1 4 and 1 3 .1 6 ,
Step 1.
V;(.v) = - .
---- = - ,and
-----= - .
621
>
O
+
1 .5 0
V,(s)
0.5s
FIG U R E 1 3 .1 8 j-domain equivalent accounting for initial conditions of the circuit of Figure 13.17.
Step 2 . Fmd the contribution to
from
S- + 3S + 2
I.5 + - + 0.55
s
Step 3. Find the contribution to
from
V l(s) =
-.y
s^ + 3s + 2
X =
.5 + - + 0.55
Step 4 . Find the contribution to Vj{s) from Z,;'^(0 ) = 1. Using O hm s law in the 5-domain,
0.5^ 1.5 + -
s/
V t{s) = -
1.5 + - + 0.55
5
-3 s - 2
^ + 3^ + 2
--------------
X =
Step 5. Su?n the three contributions and take the inverse transform.
V i(s) = v l u ) + v l { s ) + vl(.s) = -
-2
- + 35 + 2
-V+ 2
.V+ 1
in which case
Vj{t) =
2e
^hi{t)
- le
^u{t) V
ij{t) for the circuit o f Figure 1 3 .1 7 using the equivalent circuits o f Figures
/ ,(,)=
I )(.V + 2 )
622
V(^Qr)
= 1 V, /^(0 ) = 2 A, and
n{t) V.
Find
V(^t)
RLC
t>
for
0.
0.5 H
/Y Y \
-o
1.5 Q
+
v,(t)
v Jt)
1F
o l u t io n
In this example, it is convenient to replace the inductor by its (series) coniplex-frequency domain
voltage source equivalent circuit, because the inductor is in series with the input voltage source.
On the other hand, it is convenient to replace the capacitor by its (parallel) complex-frequency
domain current source equivalent circuit, because the desired output is the capacitor voltage. This
results in a three-source, or multi-input, circuit. O nce the equivalent circuits are in place, one can
combine the voltage sources and write a single node equation to find
V(As).
Using the voltage source model for the inductor and the current source ynodelfor the capaci
tor, draw the equivalent complex-frequency domain circuit. Using the voltage source equivalent for
Step 1.
the initialized inductor and the current source equivalent for the capacitor produces the circuit o f
Figure 13.20a. Combining the voltage sources and the series impedance into single terms results
in the circuit shown in Figure 13.20b.
0.5 i^(O-) = 1
/ Y Y V
1.50
_ Q
0.5 s
CvJO-) = l
(a)
FIG U R E 1 3.20 (a) Complex-frequenc) domain equivalent accounting for initial conditions o f the
circuit of Figure 13.19. (b) Circuit equivalent to part (a) with voltage sources combined.
Step 2.
623
yields
1
V c ( s ) ---------------- l + 5V c ( 5 ) = 0
1.5 + 0.5.y
Grouping terms produces
V c is )= -:
----- r+ 1
5 (0 .55'+ 1.5 )
+ 5
U . 5 + 0.5^
V(\s)
Solving for
leads to
s~ + 5s + 2
Vcis) =
5
( 5 + 0 ( 5 + 2)
Exemte a partialfraction expansion on V(^s), and take the inverse transform to obtain V(^t).
Step 3 .
5^ + 55 + 2
1
2
-2
^(7(5) = -------------------- = I----------- f-
s{s +
1) ( 5 + 2 )
.V +
+ 2
v^t)
Exercises.
le-'-2e-^~^u{i)V
A N SW ER :
2. Find
= [1 +
v^p)
I^{s)
and
Find
V(\t)
for r > 0.
= [1 0.57l4i>-'^-^'- 0 .5 7 l4 f -^ q (/) V
i^it)
for the circuit o f Example 13.7, using the equivalent circuits o f Figures
1,
EXA M PLE
1 3
( 5 ) = -------- ----------- .
( 5 + I ) ( 5 + 2)
i,{t)
^
=-
le -ii{t)
Ae--'u{t)
with classical starter, com m on in residential usage. For a fluorescent light to begin operating, there
must be a sufficient supply o f free electrons in the tube and a sufficiently high voltage between the
electrodes to allow arcing to occur. During arcing, mercury particles in the tube vaporize and give
off ultraviolet light. The ultraviolet light excites a coating o f phosphorus on the inside o f the tube
that emits light in the visible range.
For a simplified analysis, assume that all resistances are negligible and refer to Figure 1 3 .2 1 . The
source
VjJ^t)
is 120 V, 6 0 Hz, i.e., ordinary house voltage, which is too low to cause arcing inside
the fluorescent tube. Prior to arcing the gas inside the fluorescent tube acts like a very large resist
ance betvN'een the rwo electrodes. W hen the switch is turned on, the starter, a neon bulb with a
bimetallic switch inside, lights up and heats the bimetallic strip. This causes the metal to curl and
624
close the contacc. The bulb then looks like a short circuit, and a large current, limited by the
inductive ballast, flows through the heating electrodes o f the fluorescent tube, making them bet
ter able to emit electrons. During this time the neon bulb is shorted out and the bimetallic strip
cools and opens the circuit after a few seconds. At this point in time, which we will call ^ = 0 , the
inductor has an initial current
appear across the electrodes of the lamp, resulting in ignition or arcing. After the lamp ignites, the
voltage between the electrodes becomes small and is insufficient to relight the neon starter lamp.
Hence, the ac current flows between the two electrodes inside the fluorescent tube. The ballast
again serves to limit the current.
Heating
Direction
of curl when
FIGIIRK 13.21 Wiring diagram of simple fluorescent light circuit, including an inductive ballast, a
capacitor, and a starter within which is a neon bulb containing a bimctallic switch.
Suppose
L = 0 .8
H , C = 1 nF, and
due
to the initial inductor current, i.e., the zero-input response. The other com ponent, the zero-state
response, is not as important for ignition purposes. O ur strategy will be to use the ^-domain equiv
alent circuit for
L,
as illustrated in Figure 13 .2 2 .
0 .8 s
Voltage
due to
I,(S)
Li,(0)
= 0.08
intial
inductor
current
'....................................................
u- u
*
High resistance
prior to
arcing
625
Since we are assuming that all resistances are negligible and that the internal resistance (between
electrodes) o f the fluorescent lamp prior to arcing approximates infinity, voltage division in terms
o f impedances yields
+ Ls
Cs
r +
J l .25x10^
+ 1 .2 5 x 1 0
LC
J l . 2 5 x lo '
2 ,8 2 8 ^
o
. r + 1.25 X 10^
Hence, immediately prior to arcing, the capacitor voltage approximates
= - 2 ,8 2 8 sin (3 5 ,3 5 5 /) V
which is sufficiently high to induce arcing and cause the fluorescent lamp to operate.
See the homework exercises for an extension o f this analysis to the case where the ballast model
includes a resistance o f 100
Q.
TRANSFER FUNCTION
Suppose a circuit has only one active independent source and only one designated response
signal. Suppose fiirther that there is no internal stored energy at f = 0~. T he transfer func
tion o f such a circuit or system is
--------- r------ . V
(13.20)
mula for computing responses. Notice that if the input is the delta function, then
and
Y{s)
H{s).
F{s)
= 1
This means that the transfer function is the Laplace transform o f the so-
called im pulse response o f the circuit, i.e., the response due to an impulse applied at the
circuit input source when there are no initial conditions present. The idea is easily extend
ed to multiple inputs and multiple outputs to form a transfer function matrix. This exten
sion, however, is beyond the scope o f this text.
626
Exercise.
H(5 ) =
A N SW ER : ^'-"[cos(Z^/) + S sin(Z^r)]/Hf)
A transfer function, as defined by equation 1 3 .2 0 , has broad applicability to electrical and electro-mechanical systems. For example, the designated output may be a torque while the input
netiuork
driving point impedance, where the
the current source; (ii) driving point
might be voltage. However, in the context o f circuits, a transfer function is often called a
function.
admittance,
where the input is a voltage source and the output is the current leaving the voltage
source; (iii)
transfer impedance,
where the input is a current source and the voltage is across a des
transfer admittance,
output is the current through another branch in the circuit. In cases (i) and (iii), the voltage polar
ity must be consistent with the conventional labeling o f sources as set forth in Chapter 2. In gen
eral, however, we will adopt the ordinary language o f transfer function.
EXA M PLE
1 3
0 Find
^out
V :Js)
o l u t io n
There are many ways to solve this problem. O ur approach is to execute a source transformation
on the
R-L
impedance in series with the voltage source. After the source transformation, we use
Execute a source transformation to obtain three parallel branches as per Figure 13.24.
627
V Js ) r A
Y,
sT T
parallel branches, the current division formula (equation 13.9) applies, producing
>3(^)
M l
y,(A-)+r2(.s') + >3(-^'V 5 + 1
Hence,
H(s) =
Y^is)
(1 3 .2 1 )
U W + i 2 ( > ^ ) + W / s+ 1
Vinis)
Step 3 . Compute K, (^), Y2 {s), and Y^^is). Because impedances in series add, and admittance is the
inverse o f impedance (equation 13 .7 ), some straightforward algebra yields
2.5s
0 .4 5 +
, 2^1
0.4
2.5 s
H(s) =
1
5+1
2.5s
5^ +1
2.55
+ - I X---- \ 5 + 1/
5
5+1
+ 1 + 5 ( 5 + 1) +
s +l
2.5s
2.5s
(5 + 1 ) ( 5 - +
2.5s +
1)
(s +
1)(^ + 0 . 5 ) ( 5 + 2 )
.5 5 ( 5 + 1)
(i2
Exercise.
For
A N S W FR :
H{s)
//{/) =
--C
-0 .5 /
-2 /
h{t)
u(t
Zp)
and
Ip)
tively.
V,(s)
o l u t io n
Since no current enters the inputs o f an ideal op amp, I - p ) = - I p ) . Further, the voltage at the
negative op amp terminal is driven to virtual ground; hence,
V-p)
Z -p )I-p ),
and
Vg^,f{s)
Vi{s)
Zj{s)
Yj{s)
(1 3 .2 2 )
Equation 1 3 .2 2 is a verv' handy formula for computing the transfer functions and responses o f
many op amp circuits.
Exercises.
Find R so
A N SW E R :
R=
H{s)
-Ms,
Z p)
1{) kQ
Zis)
-O.S
oo.
629
3. Find the value o f C for which the transfer function o f the op amp circuit in Figure 1 3 .2 7 is
0.2 5 Q
Rj,
and Cleading to
R,
o l u t io n
Step 1.
From the given data, compute the actual transferfunction o f the circuit.
By definition o f the
transfer function,
H{s) =
L[resp(mse\
Voui{s)
(s + l)
\input\
Vi(s)
1
.y+1
5+1
( 1 3 .2 3 )
(i3 0
Step 2 . Using Figure 13.28, fin d the transferfiinction o f the circuit in terms o f R^, Rj, and C. Here,
obsen^e that Figure 1 3 .2 8 has the same topolog)^ as Figure 13 .2 6 , where
Ky.(,9)= ! =
c j+
Z /(.v )
-^
'
R2
'
i
Cs + Step 3 . Match coejfcients in equatiotis 13.23 and 13.24a to obtain the desired values ofR^, Rj, and
Cv+
Rl
O ne possible solution is
R^ =
H (s)= -
(1 3 .2 4 b )
CRj
> 0,
^2new ~
^new
magnitude scaling.
unchanged but produces more realistic values for the circuit elements.
Exercises. 1. In equation 13.24b, it is required that C = 10 uE Find appropriate values o f /?j and R-^.
AN SW ER:
R^ =
50
kLl R,
= 100
kLl
A N SW ER :
_JL
t
H,C
tl(l)
631
F{s)
^{t),
Some simple calculations verify that the transform o f the impulse response is the transfer function,
1.e.,
X[y(r)] =
H {s)m t)]
His)
Hence, the impulse response o f the circuit/system, denoted /;(/), is the inverse transform o f the
transfer function
Ht) -'[M(s)]
(13.25a)
H{s) = [h{t)]
(1 3 .2 5 b )
and conversely
Exercises.
Find
[lc^ -
function.
A N SW ER : (I
3. Suppose
- e^'^)/s
t/{t) = 2b{r-
2hU -
h{t)
2u(t) - 2u{t5).
1) - 3A(r - 3)
stant parameter values for its elements can be represented in the time domain by its impulse
response. This is shown in Chapter 15, where we define a mathematical operation called
tion
convolu
and show that the convolution o f the input function with the impulse response function
yields the zero-state circuit response. In addition to this significant theoretical result, the impulse
response is im portant for identification o f linear circuits or systems having unknown constant
parameters. Sometimes a transfer function is unavailable or a circuit diagram is lost. In such a
predicament, measuring the impulse response on an oscilloscope as the derivative o f the step
response is quite practical.
What is the step response o f a circuit? T he step response is merely the zero-state response o f the cir
f^t) to the circuit is (/), then F{s) = 1/^ and K(j)
= H[s) (1/^). By the integration propert} o f the Laplace transform, it follows that the step response
032
is the integral o f the impulse response. Conversely, the derivative o f the step response is the
impulse response. In lab, many scopes can display the derivative of a trace and hence can display
the derivative o f the step response, which is the impulse response. Alternatively, a homework prob
lem will suggest a means o f directly generating an approximate impulse response.
Exercises.
H{s)
responses?
2. If the Laplace transform o f the step response o f a circuit is given by
Y{s)
\sy{t)
= [1 - 0.5^ ^-
response?
u(t),
cus(2/ + 2 6 .5 7 ')//(r).
= b{r)
b {r -
y(t)
A
h(t)
3
A
2
--
1 H-----
(a)
(b)
FIG URE 13.29 (a) Impulse response of hypothetical circuit, (b) Response to 6(r) +
b {t-
1).
S o lu t io n
Since X [6(r)
+ 6(/- 1)]
= I +
e~\ the
response,
1).
is given by;/(^)
h{t).
o l u t io n
b{t)
is the integral o f the delta function and the ramp the integral o f the step, then the delta function
equals the second derivative o f the ramp. Hence,
the impulse response
h{t)
y\t)
b{t)
= [4 -
+ [ - 6 + 4 / + 8 ^ * -'-
(t)
y"{t)
633
= [Sf*
and
hit) = [e-^-
To see the utility o f this approach, try the alternative method o f computing
V(s)/f(s).
F(s), V(s),
and
H(s) =
As a final example, we compute a circuits step response and verify that its derivative is the impulse
response.
E X A M P L E 1 3 .1 4 . C om pute the step response o f the
RLC circuit
o f Figure 13 .3 0 .
/m
R = 40
v
L=1H
C = 0 .2 F
-o
FIG U R E 13.30
S
o l u t io n
s~ -\ --s+
L
LC
f.v + 2 ) + l
Cs
/^ + L .V +
Cs
(1 3 .2 6 )
His)
= - +
.y (.y + 2 ) + l l
-.V - 4
(.v + 2)^ + l
.V
(5 + 2 ) - + 1
{s + 2)-+\
cos(f) -
2e-~ sin(f)];K r)
(1 3 .2 7 )
fi3^
(It
= 5e
sin{t)u{t)
Thus
5
Cv + 2 ) - + 1
in which case
h{{) =
5^" ^sin(r)(f)
as expected.
Sallen and
malized low-pass filter passes frequencies below 1 rad/sec and attenuates higher frequencies. As we
will see later in the text, the 1-rad/sec frequenc)' cu toff can be changed to any desired value by
frequency-scaling the parameter values o f the circuit. (See Chapter 14 for a discussion o f frequen-
63 S
cy scaling.) The goal here is to utilize the techniques o f nodal analysis to compute the (normal
ized) transfer function o f this circuit.
o l u t io n
T he solution proceeds in several steps that utilize nodal analysis techniques in conjunction with
the properties o f an ideal op amp. Recall that for an ideal op amp, the voltage across the input ter
minals is zero and the current into any o f the input terminals is also zero. Finally, note that one
does not write a node equation at the output, which appears across a dependent voltage source
whose value depends on other voltages in the circuit.
Step 1. Find Vy. Because the voltage across the input terminals o f an ideal operational amplifier
is zero.
Step 2. 'Write a node equation at the node identified by the node voltage V^. Summing the currents
leaving the node yields
for
- V,) = 0
(J2.S + 2 ) V - ( J 2 S + \)V,=V,
(13.28)
Step 3. Write a node equation at the node identified by the node voltage Vy. By inspection, the desired
node equation is
5+1
(1 3 .2 9 )
Step 4 . Write the foregoittg tivo node equations in matrix form. In matrix form, equations 1 3 .2 8 and
13.29 combine to give
- (V 2 . + I)'
-1
-|=5+ 1
IV 2
J
K;
'^in
0
(1 3 .3 0 )
6.U>
del
'{ J lS + l )
V;
-1
(>/ 2 :i- +
-(V
.V + 1 )
del
-1
\42
K..
s- + yf2s+\
N otice that for small values o f ^ = yto (i.e., low frequencies), the magnitude o f H{s) approximates
1, and for large values o f s = JiO (i.e., high frequencies, where |/b)|
small. Since
passes low-frequency input excitations. As mentioned at the beginning o f the example, the circuit
passes low frequencies and attenuates high frequencies.
T he preceding example used matrix notation, com m on to much o f advanced circuit analysis. In
one sense, matrix notation is a shorthand way of writing n simultaneous equations: the n variables
are written only once. More generally, matrix notation and the associated matrix arithmetic allow
engineers to handle and solve large numbers o f equations in numerically efficient ways. Further,
the theory o f matrices allows one to develop insights into large circuits that would otherwise
remain hidden. Hence, many o f the examples that follow will utilize the elementary properties o f
matrix arithmetic.
T he next example uses nodal analysis to compute the response to an initialized circuit. The exam
ple combines the equivalent circuits for initialized capacitors and inductors with the technique o f
nodal analysis.
V(^t)
FIG U R E
and
v^Qi~)
v^it).
1 3 . 3 2
Given the indicated current direction of i[{t), what is the implied voltage polarit)' for
v^{t)}
63'
Step 1. Draw the s-domain ecjuivaletit circuit with an eye toward nodal analysis. Inserting the equiv
alent current source models for the initialized capacitor and inductor in Figure 1 3 .3 0 , one obtains
the 5-domain equivalent circuit shown in Figure 13 .3 3 .
V Js)
VJs)
1o
ii(O-)
,Cv,(0 )
O
1o
= 1
1F
1H
V(is)
K C L implies
that
(1 +
s)V(is)
[V(is)
Vj{s)] = 2
.9+1
-V c(^ ) +
1
=
2 1
-1
_i
.9+1
s
Step 3 . Solve the matiix equation o f step 2 for the desired voltages. Using C ram ers rule, computing
the inverse, or simultaneously solving the equations gives
2{s+\)-\
2
Vcis)-
5+ 2 . 9 + 2
[5 +1
1 1 2
s
1
5+2
(s+Vr + \
(1 3 .3 1 )
_i
C v + D -3
(.V+ 1r +1
Step 4 , Take the inverse Laplace tratisfonn to obtain time domain voltages. Breaking up equation
13.31 into its components yields
638
2(^ + 1)________ 1
Vcis) =
(5 + 1)- + !
(5+ 1)^ + 1
in which case
V(^t)
e ^[2 cos(t)
- sin(/)]/^(r)
Also,
(.^ + 1)
(A-+1)^ + 1
(.V+1)^ + 1
V^(s) =
leading to
=e
Figure 13.34 presents plots o f
V(\t)
^[cos(^) - 3sin(f)](r)
and
H G U R E 13.34 Plots of the capacitor and inductor voltages for Example 13.16.
Dual to nodal analysis is loop analysis. In loop analysis, one defines loop currents and writes KVL
equations in terms o f these loop currents. The following example illustrates the method o f loop
analysis for computing the input impedance o f a bridged-T network.
639
E X A M P L E 1 3 .1 7 . Use loop analysis to com pute the input Impedance o f the bridged-T network
illustrated in Figure 13 .3 5 .
/YY\ 1 H
2Q
20
20
0.25 F
Z Js)
o l u t io n
Define
find
= / , U) in Figure 1 3 .3 4 . Since
I-^j{s)
Step 1.
= /j(^) in terms o f
obtains
2(/, - / 3 ) + - ( / , - / 2 ) = 2
Step 2.
- - / 2 - 2 /3 =
By inspection,
- ( / 2 - / | ) + 2( / , - / 3 ) + 2 / 2 = - - / , + 4
.V
Step 3 .
Again by inspection,
/ 2 - 2 /3 = 0
-K4)73 = 0
tions are
, 2
.1
/l1
.1
-2
-2
-2
5+ 4
0
0
(viO
in terms o f
= / ] W yields
.v + 4.V + 4
=2 Q
is independent of
s, despite
constant-resistance network.
problem at the end o f the chapter shows the conditions on the elements o f a bridged-T network
for it to be a constant-resistance nerwork.
Exercise.
-2 s
\h]
-Vs:
4
2.V + - + 2
_2
-1
-2 i'
_2
25 + 4
-I
-4
h
V
M GURE 13.36
switch
shapes the behavior o f the circuit. T he switch inside an electronic circuit is a special device that
we will model simply as an ideal on /off switch. This section investigates the behavior of switching
in simple
RLC circuits,
64
sophisticated electronic circuits. O ur immediate task is to apply the Laplace transform method to
com pute the responses of switched
RLC circuits.
cedure.
u{t),
= 0.
= 20 Q ,
R = A Q., C = 0 .2 5
F, and
position A to position B at / = 1 sec and from position B to position A at ^ = 2 sec, and moves
V(^t)
for
t>
0.
v,(t)
v jt)
V(^Q~) = 0;
o l u t io n
V(-{t) over
2<. t <A,
etc.
Step 1. Compute the response for 0 ^ t < 1. Over the interval 0 :s r < 1 , the circuit o f Figure 13.37
is the simple
RC circuit
o f Figure 13.38.
20 0
+
V,(s)
0.2 5 F
< 1.
s
20
20
+ -
.V
v^t)
20e-^-^^-20e~^-^
20
642
W e note that
Step
(1") =
2 0
^0 . 1 ^ j
7 2 2
V.
the source is decoupled from the right half o f the circuit; the response then depends only on the
initial condition at r = 1 , i.e.,
^ f<
= 1 .7 2 2
V. T he goal is to compute
s r' <
where
t' = t - \. The
v^^t)
cuit that models the behavior o f the time domain circuit o f Figure 1 3 .3 7 over
form illustrated in Figure 13 .3 9 . We note that the value on the current source is
0 .4 3 0 5 .
FIG U R E
1 3 . 3 9
^ t <2.
Vc(.s) =
For 0 s ^' <
s +1
1 72^
X 0 .4 3 0 5 3 = ^
.V +
or equivalently, for
^ r<
\.722e-^'ti{t)
V(^t) = \.722e-^^-^\i{t- 1)
W e emphasize that this last equation is valid only for
: r <
. We note that
(2') = 1 .7 2 2 e ^ =
0 .6 3 3 5 .
Step 3. Compute v(it) fo r 2 ^ t <4. For this interv'al, the capacitor is initialized at /' = t - 2 = 0.
Using the parallel current-source equivalent circuit, we note that
(r')
Vjfjis)
= 8 .l 8 7 .V - ''' ( r ')
in Figure 1 3 .4 0 is
8 .1 8 7 3
5 + 0.1
(2') =
643
0 . 2 x 8 .1 8 7 3
0.25.S + 0 .0 5
U + 0 .2 )(5 + 0 .1 )
____s
20
+ -
0 .6 3 3 5
5 + 0 .2
.V
v^4~) = \63746e-^ ~ -
1 5 . 7 4 1 =
2 .8 5 4 8 V.
Step 4 . Compute the response over 4 ^ t. After the switch moves from position A to position B,
again, the source is again decoupled from the right half o f the circuit. According to Chapter 8, we
can write the answer by inspection: the solution is simply v^^t') = 2.8548^ ^ n{t). Equivalently,
f o r / > 4 , y J r ) = 2 .8 5 4 8 f - ( '- ^ \
Step 5. Combine results into a single expression and plot. I'h e combined expressions for V({i) arc
Os/ <l
1.722e"^ '"W - l )
v c(0 =
2.85486^
A plot o f the response appears in Figure 1 3 .4 1 .
\^t < 2
()-i4
FIGURK 13.41 Plot of capacitor voltage for switched circuit of Figure 13.37.
T he extension o f this method to more than three switching times is straightfoi^vard. Although the
preceding example uses an
RLC circuit,
lation o f switching transients in any linear dynamic circuit. T he following is a summary o f the
general procedure.
P rocedure for Applying Lapiace Transform M ethod to Switched C ircuits
At / = 0, a switching may or may not occur. For / > 0, denote the switching time instants
sively as
, r-, ,
successively for
fo r
succes
0 s r < sc.
Step I. C onstruct
t = 0~, which are
(see note at the end o f the procedure). This equivalent circuit is valid for the time inter
val 0 ^ r < / j .
Step 2.
Step 3.
Find the response by the Laplace transform method for the circuit o f step 1.
Evaluate the capacitor voltages and inductor currents at
t = T j.
t~).
calculated
initial
60
(2) Determine the proper form o f the input excitation(s) (if there are any), in
terms of t .
(3) I'incl the response by the Laplace transform method. Note that the time vari
able is t' . Then obtain the solution in
t by
(4) If / = w, stop. Otherwise, eN-aluate the capacitor voltages and inductor currents at
(5) Increase the subscript value / by 1 and go to the beginning o f step 5.
Note: In some situations, the first switching occurs at r = 0, but the capacitor voltages and induc
tor currents at r = 0 are not given. Instead, the problem specifies that dc and sinusoidal sources
have excited the circuit for a long time. If the network is passive i.e., if it consists o f inductors,
capacitors, or
t = 0 . The
proce
dure then is first to find the steady-state solution and then to evaluate the capacitor voltages and
inductor currents at / = 0 . It is instructive to review the dc and sinusoidal steady-state (phasor)
analysis methods studied in a first course. Recall that under certain stabilit)' conditions (to be stud
ied in Chapter 15) on the network:
1.
For dc steady-state analysis, open-circuit all capacitances and short-circuit all inductors
to find the steady-state voltages and currents.
2.
For sinusoidal steady-state analysis, use the phasor method to find the steady-state
responses.
only capacitors, switches, independent voltage sources, and possibly some operational amplifiers.
No resistors or inductors are present. One can dispense with resistors because it is possible to
approximate the effect o f a resistor with rwo switches and a capacitor. Similarly, inductors can be
approximated by circuits containing only switches, capacitors, and operational amplifiers. These
facts, coupled with the easy and relatively inexpensive fabrication ol switches, capacitors, and
operational amplifiers in M OS (metal-oxide semiconductor) technolog}^ have made switched
capacitor filters an attractive alternative to classical filters. Given this scenario, the purpose of this
section is to lay a foundation (i.e., introduce the principles) upon which switched capacitor cir
cuit design builds. More advanced courses delve into the actual analysis and design o f real-world
switched capacitor circuits.
Besides the Laplace transform approach, an alternative method for analyzing switched capacitor
networks builds on the principle o f conserv'ation o f charge.
Principle o f conservation o f charge: T he total charge transferred into a junction (or our o f
a junction) o f a circuit at any time is zero.
646
This principle is a direct consequence o f KirchhofFs current law. For example, in Figure 1 3 .4 2 ,
KCL implies that
(13.32)
I,
t yields
f
li,(r) + i 2 ( 0 + h ( 0 + U U m = 0
X
or equivalently.
EXA M PLE 1 3 .1 9 . Consider the circuit shown in Figure 13.43a. The switch S is closed at / = 0.
Just before the closing o f S, the initial conditions are known to be
C om pute the voltages
and
= 1 V and
for ^ > 0.
i(t)
^ -----
Vc, -
IF r
IF
-----
+
Vc, ^
C2
IF
IF
---- -----
(a) t < 0
(b) t > 0
= 0.
64"
l(s)
V,C2
(C )
s
Therefore,
V(y{t) = 0.5u{t)
V, and
i{t)
s s+ s
= 0.56(^) A.
M ethod 2 . Using the series voltage-source model o f a capacitor, we have the ^-domain equivalent
circuit shown in Figure 1 3.43d . Again, by inspection.
i + i '
S
'
i+ i
Conservation-of-charge approach.
total charge
and
Vci(() ) - V c 2 ( 0
and
results in
) - ------------7 -------------------C] + C 2
Since there is no external input applied, the voltages remain constant once the equilibrium con
dition has been reached. Therefore,
(vi.S
for r > 0. For the specific capacitance values given in Figure 1 3.43a, we obtain
0 .5 V for r > 0.
) = 8 V, and v a m
3 V. Find, by at least two methods, the capacitor voltages after the switch is closed. (Rework the
problem if the answers do not agree.)
A N SW ER : S \'
Computationally, the Laplace transform method is more straightforward. On the other hand, the
conservation-of-charge method is more basic and often provides better insight into what happens
to the charges stored in various capacitors. It is particularly useful for the purpose o f checking
answers obtained by other methods: the answers are correct when the conser\'ation-of-charge con
dition is met at every node.
E X A M P L E 1 3 .2 0 . The initial conditions at r = 0 o f an SC network are shown in Figure 13 .4 4 .
Switches Sj and S-, are closed at / = 0, connecting the tw'o dc voltage sources to the nersvork. Find
the node voltages for
t>Q.
lO V
o l u t io n
We first construct the.f-domain equivalent circuit using admittances. The result is shown in Figure 13.45.
fO
15
4s
3s
2s
12
+ A IQ
6 i9
4s
= 9/s and
+ 3.V
10
= 9 V for
+ I s Vt. = 4 + 1 5 + 12
t=
0~ to A= O'*, tiic voltage of the 4 F capacitor jumps from 1 V to 4 V (note that 4 = 9 - 5 ) , indi
cating that 4 X (4 - 1) = 12 coulombs o f charge have been transferred to this capacitor. The volt
age o f the 2 F capacitor jumps from 6 V to 9 V, indicating that 2 x (9 - 6) = 6 coulombs o f charge
have been transferred here. Finally, the voltage o f the 3 F capacitor changes from 5 V to - 1 V (note
that - 1 = 9 - 1 0 ) , indicating that 3 x (-1 5) = - 1 8 coulombs of charge have been transferred to
this capacitor. As a check for conservation o f charge, we have 12 + 6 + ( - 1 8 ) = 0, and the solution
is assured to be correct.
Exercise.
A N SW ER :
v^(t)
= 5 0 /J V, / > 0
T he preceding examples considered idealized circuits, i.e., no resistances were present. In zny prac
tical circuit, the connecting wires have some resistance. W hat is our interest in the analysis of an
idealized circuit? T he analysis o f an idealized circuit is much more straightforward than that of a
realistic circuit yet provides relatively accurate answers. As a case in point, reconsider Example
1 3
.2
. Suppose we insert a 0.1 Q resistance in series with ever)' capacitor. The resulting transform
analysis would produce a rational function with a cubic denominator polynomial (a third-order
network) whose factorization would require the use of a root-finding program. In sharp contrast,
the idealized circuit o f Example 13.20 was anaK'zed by writing a single first-order node equation,
making a partial fraction expansion unnecessary.
Idealizations o f circuit models sometimes lead to phenomena that defy intuitive explanations. An
interesting case is given by Example 13.20. Before S is closed, the energ)' stored in the electric field
is
0 .5
(0 .5 ) =
0 .2 5
joule. Apparently, 0 .2 5 joule o f energy has been lost. Since there is no resistance in the
circuit to dissipate the energ)', what accounts for the lost energ}'? Is energ)' not conser\'ed?
An explanation o f this paradox is as follows. Instead o f considering a zero-resistance circuit, place
a resistance R in series with all capacitances. Then analrze the circuit, and let R approach zero. The
result shows that no matter what value R takes on, the total energ)' dissipated in the resistance for
0 < / < ^ exactly equals the difference o f the total stored energies before and after the closing o f
the switch. This accounts for the apparent lost energ)'. In actuality, part o f the 0 .2 5 joule of cnerg) would be lost in the form o f radiated energ)'. However, the principles of field theor)- would be
necessar)' to explain the radiation phenomenon.
Note that for idealized SC circuits, as long as the independent voltage sources arc piecewise con
stant, all capacitor voltages are piecewise constant and all currents in the circuit are impulses. These
650
properties remain valid for more general idealized SC circuits that allow the inclusion o f VCVSs,
C C C Ss, and ideal op amps. T he reason is that the parameters characterizing these components are
instantaneous.
lossy switched capacitor circuit,
dimensionless and hence do not result in a time constant. All voltage changes are
On the other hand, if the circuit contains resistances, we have a
whose voltages arc no longer piecewise constant. The transient analysis o f a lossy SC circuit
requires the usual Laplace transform analysis.
approximate a
approximated by an
One reason for our interest in SC circuits is that a SC combination can be used to
resistor. As a result, any
RC-op
S C -o p amp circuit. A study o f the general theor)' o f such S C -o p amp circuits is beyond the level of
this book. We shall merely illustrate the approximation property with a simple integrator circuit.
E X A M P L E 13.21. Consider the RC-op amp integrator circuit shown in Figure 13.46.
R = 5kQ
C,= 1 mF
vj,t)
v{ =
= 5 V, and if
= - l ,0 0 0 ///( f ) (as long as the output has not reached the saturation level), as shown
in Figure 13.47.
t (msec)
-
10-
6SI
At r = 0, S is at position a.
2.
At t = T, S is
3.
4.
moved to position b.
T he output waveform may be determined ver)' easily by the principle o f conser\'ation o f charge as
follows: for 0 :s / <
T,
y.
and
b. Because the op amp is assumed to be ideal, the voltage across the input terminals is zero, and
so is v^. Thus, C cannot store any charge. T he charge CE previously stored on C must be trans
ferred out o f C. Since the op amp is ideal, the input impedance is infinity and the input current
is zero. Therefore, none o f the charge can flow into the op amp. Instead, the charge must be trans
ferred to the capacitor
This leads to
CE and Vgi'T') = - C /C ^
At r = 27', switch S is moved back to position a, causing C to be charged to E volts again. Since
the charge
trapped on
tion b again. At that time, another CE coulombs o f charge are transferred to Cp and
is incre
average o f the
current flowing into C^is at a constant value o f E/R. Therefore, every I T sec, the charge trans
ferred to C^is equal to 2TEIR. On the other hand, for Figure 1 3 .4 8 , the charge transferred to Cjevery I T sec is CE. Equating these rwo quantities, we have CE = 2TEIR, or RC = IT . Thus, there
is no unique combination o f C and T that will produce the approximate effect o f a resistance. A
smaller T (i.e., a higher operating frequency o f the switch) in Figure 13.48 produces a staircase
output waveform that closely hugs the ramp output o f Figure 1 3 .4 7 . For the purpose o f com
parison, the output waveform corresponding to 7 = 1 msec and C = 0 .4 pF is shown in Figure
1 3 .4 9 , together with the ramp output from the RC-op amp integrator. It is worthwhile to note
652
that each o f the circuits o f Figures 1 3 .4 6 and 1 3.48 drains the same average amount o f charge
from the voltage source and puts the same average amount of charge on the capacitor
difference is that in the former the process is
coutinuous,
Cj; The
only
in quantized steps.
2 3 4 5 6 7 8 9
19
-I I I I I I I I I It (msec)
Output of
switched capacitor
circuit
Exercise.
msec and C = 0 .0 4 pF. Also, plot the ramp output on the same graph for comparison.
As recently as two decades ago, switched capacitor circuits were considered impractical. No longer!
Advances in semiconductor technology have made the fabrication o f high-speed electronic switch
es and op amps as cheap as resistors. Furthermore, large numbers of switches, capacitors, and op
amps can be fabricated on a single chip. Consequently, switched capacitor circuits hold an impor
tant place in future signal-processing applications. Although we cannot delve into the practical
aspects o f the design of such circuits, we have at least outlined the basic principles needed for their
approximate or exact analysis.
*oiit ( '0 -
(1 3 .3 3 )
653
....
expression represents the ^-domain equivalent o f the following time domain equation assuming
zero initial conditions at ; = 0 on the variables:
'W
( 0
- 1 j j j 'V/1
+ ^
2 /,2(r)ciT
/^ V
and
are explained in
Prototype design.
S,
^2
^bl
= ^2 S.
In the derivation described later on, we conclude that a simplified design requires that the total
admittances incident on the inverting terminal equal the total admittance incident on the non
inverting terminal by proper choice o f AG^ or AG^. The proper choices are given in design step
2.
Design Step 2. Prototype design [continued): Computation o f
6S4
= 6 and
Case 2:
= - 6 and AG^ = 0.
If 6 < 0 , set
= 0.
D esign Step 3 . Scaling to achieve practical element values. Multiply all the admittances incident at
the inverting input terminal o f the op amp by a constant
K^.
= /Q,, but
(1 3 .3 4 )
S
o l u t io n
Step 1. Prototype design. Using Figure 1 3 .5 0 , choose C = 1 F. W ith all admittances in S, set
= 1 4 .2 8 kQ,
75
6SS
1 4 .2 8 kQ
lO p F
a2 '
3 3 .3 3 kO
+
5 0 kO
b2
o
2 5 kO
I O mF
2 5 kO-
(b)
FIGURE 13.51 (^) Prototype design of equation 13.34. (b) Final design after scaling with /T = AT^ = 10~5.
Exercises.
2S
6S
M GURE 13.52
656
Solving for
yields
W b2
(13.35)
Derivation Step 2. Write a node equation at the inverting input terminal of the op amp. Recall that
- V^2) Thus,
sC
sC
''al
Gi
Gno +
AG +
(13.36)
sC
n\
Derivation Step 3 . Combine steps 1 and 2 to compute the general input-output relationship.
Substituting equation 13.35 into 13.36 yields
1
/
^ a l T,
G a2 y
I I ^ a\ + ^ a l +
"^ 7 )
If we choose
and AG^ to make AG^ + ^a\ + ^al =
+ ^b\ + ^bV -e- if the total admit
tance incident on the inverting terminal is made equal to the total admittance incident on the
non-inverting terminal, then equation 13.37 simplifies to
I sC
(13.38)
sC
If we let C = 1 F (to be scaled later to a practical value), equation 13.38 further simplifies to
^out
^b\ +
^bl
(13.39)
Equation 13.39 shows that the circuit of Figure 13.50 is a general summing integrating circuit
whose gains are proportional to the admittances G^- and/or % The sign o f each gain depends on
whether the corresponding input is connected to the inverting or non-inverting terminal o f the
op amp. This completes the derivation of the input-output characteristic of the op amp circuit and
is the basis for the prototype design. The extension to more than four inputs is straightforward.
6^
It remains to justify the scaling o f step 3 in the design procedure. In step 3 , it is stated that all
admittances incident at the inverting input terminal o f the op amp can be multiplied by a (scal
ing) constant
K^.
The verification that these multiplications will not change the input-output characteristic follows
directly from equation 13 .3 7 . After inserting the scale factors, one immediately sees that they can
cel and have no effect on the overall gain.
11. SUMMARY
This chapter has presented the basic principles and techniques o f circuit analysis in the j-domain.
Impedance, admittance, Thevenin equivalents, superposition, linearity, voltage division, current
division, source transformations, and nodal and loop analysis have ^-domain forms that allow the
analysis o f complex circuits excited by a variety o f waveform types. Indeed, the simple starter cir
cuit o f a fluorescent light points to the usefulness o f Laplace transform methods in the analysis
and design o f simple, everyday electrical conveniences. Various op amp applications were also pre
sented. As subsequent chapters will illustrate, complex circuits and advanced analysis methods
build on these basic principles and techniques.
T he chapter also introduced the notion o f switched capacitor circuits. Integrated circuit technol
og)' has made such circuits easy and inexpensive to produce. Applications include speech process
ing and other t\'pes o f signal processing. Although a full-scale analysis o f such circuits is beyond
the scope o f this text, the basic principles o f their operation arc presented as a foundation for more
advanced analysis tools.
Finally the chapter introduced a general method for the design o f multi-input integrators having
both positive and negative gains. Such circuits can be used for implementing active filters and for
implementing controllers in practical situations.
65<S
Impedance: the ratio o f the Laplace transform o f the input voltage to the Laplace transform o f
the input current with the two-terminal network initially relaxed.
Steady-state analysis: analysis o f circuit behavior resulting after excitations have been on for a
long time; often refers to finding the sinusoidal or constant parts o f the response when
the circuit is excited by sinusoids or dc.
Transient analysis: analysis o f circuit behavior for a period o f time immediately after independ
ent sources have been turned on.
Voltage division: a formula for determining how voltages distribute around a series connection
o f impedances.
6S9
Problems
IMPEDANCE, ADMITTANCE,
VOLTAGE DIVISION,
CURRENT DIVISION, SOURCE
TRANSFORMATIONS,
THEVENIN AND NORTON
EQUIVALENTS
ZJs)
(a)
/Y Y V ^
L
Find
rational func
,b)
Figure P I3.2
2.5 0
0.2 F
1H
(a)
C om pute
(b)
If
(c)
If
r > 0.
(a)
(d)
0.25 H
J T
If
v jt )
0.
20e~^K(t)
V, find
ijt)
for
ijt)
0.25 F
Z(s)
v.(t) I
20
Figure P I3.3
(b,
Figure P I3.1
C H E C K S : Figure P I 3 . la:
+5
y,(5) = 0 . 4 --------- ^ F i g u r e P I 3 . lb:
( .y + l ) -+ 2 "
(as indicated)
J" + 1 6
2. Find the input impedance and admittance in
terms o f
R, L,
Z(s)
(a)
660
C = 0.01
(a)
initial conditions
(b)
initial conditions
(0
r\
Figure P I3.4
r\
) = 0. Find
and i^^t)
. Z.W.
= 5{t) and (b)
when (a)
50(1 -
n\
mA.
Z.(s)
Y Js)
Figure P I3.8
0.4 F
^m\ ~
^ml ~
= 100 12,
^2
= 1 mF.
Figure P i3.5
(a)
(b)
(c)
If v^{t) =
V, compute /j(r).
l,(s)
Figure PI3.6b.
+
JL
l,(s)
R.
V^(t)
(a)
Figure P i3.9
t vJOCV)
20-
C H EC K : Z;(5) = - 1 0 0
5-10
5 + 10
101
10. (a)
(b)
Figure P I3.6
(b)
to a step
0(11
= Lj=
1 H,
iY Y V ,
and C = 10 mF.
0.1 H
10 pF
I (s)
vjt)
Z,(s)
50 0
K,V,(s)
2,(s)
Y,(s)o-
V,{s)
(b)
Figure P I3.12
13.
Vj(s)
^ , w (b)
Z,(s)
Figure P I3.1 0
z (S)
Z,(s)
terminal as
determine
or
0.5 H
_ m
v _
Figure P I3 .13
14. Consider the circuit o f Figure P 1 3 .1 4 , in
which
= 10
= 1 mR
0.5 Q
.(i)
(a)
(b)
Find
and
(a)
Ifl(S)
(c)
2H
V, find
(b)
i,(t)
2 + 2.V +
iind
v^{t) =
Figure P I3.11
12. Find
V, find /^(r)
0.5 F
o-
ANSWF.R: (h)
o- J \ / v 'n_T Y Y \ .
20
v^i)
= 10 V.
R,
/Y Y V
L
'"Q
YJs)
Z,(s)
Figure P i 3 .1 4
Z,(s)
CH ECK:
K, = 1 0
.y -h 200.V + ? ? ? ?
(a)
662
y? 2 = 4 n , Z, = 1 H, and C =
0 .1 2 5 5 ^ + 4
(c)
Z /.(-v ) =
(d)
yM =-
2 mF.
0 .1 2 5 5 - + 4
Y.^i,s) and
(a)
Find
(b)
0 .5 5
0.5 s
in terms
and /.W .
(c)
505
zM
5 -+ 2 5 0
55
R.
(0
R,
ijs) ( t )
V.Js)
(g)
(h)
yM =
5 + 2 5 0
5 + 4
0 .2 5 5 + 4
Z ,v ,(j) = ----------+
0 .5 5
is
5^ + 16
n > ,(^ )=
45
+
'0 .2 5 5 ^ + 16
Figure P I3 . 15
TRANSFER FUNCTIONS , ,
16. Find an
RC circuit
ing:
(a)
Z:,As) = 0 .2 5
(c)
Z , ( 5 - ) = 1 0 +
+ 0 .0 2 5
Z M
('= )
s+ 2
4s~ +
o f the following:
0 .1 5 + 1 .5
V
H{s) = - ^
5+10
5+2
Z , (5 ) = 0 .2 5 + 2 0 +
^
+ 25
( ;) + 1
5 v / (r)-1 0 jJ.
5+10
5+8
your ( 0
(b)
(c)
20
5
(T)dr =
Vi{T)(h
iaut ( 0
+ 0.1
dinutit)
+ 4 2 .5
r,v,(5) = 0.025 +
18. Find an
(b)
v ^
5 + 240
V f \
2 2 5 0 V ,. , , ( 0 + l O
70
yM =
v , ( / ) + 1 0 v ,(/ ) + 2 5 .. , (/ ) =
2.V + 8
17. Find an
(a)
( A")
-----,
(5 + 2)(.y + 4 )
(c)
0 .2 5 5 + 0 .2
(a)
H{s) =
(b)
(a)
J
00
= 200V;(/)+20i:';(/)
{T )dT
663
IUU.V
2 6. (a)
.s- + 400
0
(b)
+ 40^^^^
- 0 . 1 / , ( 0 " ) + 4()
(c)
.V
2 2 , An integro-difFerential equation
= 1 0 l
.V
for an
R=
Assuming
1 kH and C = 0 .5 niF,
2u{t) - 4u (t- 2)
input
= 0 .5 F , Z = 1 H, ^ = 2 n i s
(d)
V.
R=
1 kD and C = 0 .5 mF,
as given in Figure P i3 .2 6 c .
(a)
(b)
(c)
.
%
He
c
2R
V (t)
y{t)
and
are
^out
--- r ; +
dt~
) + 2y(/) = 4 ^
at
He-
and
(a)
+ /
(b)
v.(t)
O'
He
N(s) =
(b)
v(t)
co^{t)u{t)
te~ii{t).
C onstruct the
and output
- t
(0
Figure P i 3.26
2 5. T he input to a relaxed (no initial condi
e~u{t)
V.
2 7 .(a)
(ideal)
\0e~ s'm{2t)u{t).
P I3 .2 7 a in terms o f 7?,,
(a)
op
amp
circuit
it is desired to obtain
with C = 1 IIF, then
Vl{s)
d{s)
tored form.
Figure
and C . If
H{s) =
-2
^
' ^
(b)
ol
R-,,
2te~'u{t)
V.
664
r\
(c)
29. (a)
(b)
(c)
v,(t) +
5+ 2
s+4
^^
^^
s+4
v,(t)
->t
(b)
Figure P I 3.2 7
28.(a)
(b)
(c)
(d)
30. (a)
(b)
If it is desired to obtain
//M = - 4
(c)
(s + 2)(5 + 10)
with 2 = 100 ^F, find 7?j, /?2and C,.
Given the answer to part (b), find the
zero-state response to y,(^)= -2.5e~^^
sin(2/)(/) V.
Figure P I3.30
Figure P I3.28
Gi + G t
s + ^------ ^
CHECK:
H(s) = -------- f
s+
CHECK:
H(s) =
-1
C 2R 1
s+
s+
/ ? ,C j
R 2C 2
665
3 1. (a)
(b)
C,
Suppose
R and K.
C = 1 i.i, K = 3,
3 3 . C onstruct the transfer Kinction o f the d rcuit o f Figure P 1 3 .3 3 , assuming that all op
amps are ideal, as follows:
and
R = R^
(a)
Let G| = , C t = , G t, =
(b)
R.
in Figure P 1 3 .3 1 .
(c)
ij{t)=
2 sin(l 500/)//(r).
V'i(.v)
(c)
(d)
Com pute
(c)
Com pute
(0
V'2(-v)
(a)
in terms o f
V:^{s)
and
H{s)
Vi,M)
R-y, Ry R^,
R-y, Ry
H{s) =
(c)
C p and
C-,.
Figure P i 3.33
R^ so
that
ANSW HR:
-5
-2
(.v+lOO)
(.v + 2 0 0 )
3 4
. (a)
H(s) =
Ci\
s~ + (j^s +
G -1
(b)
Vinis)
If
= 10 Q , Yj = 0.1 S, Z , = /, V, =
s, and
= 2 0 S, com pute the impulse
and step responses.
(c)
-2u{t
2u{t)
further calculations.
666
IJs)
2,(s)
+
Z.(s)
Y,(s)
VJs)
Y,(s)
9 jY
Figure P i3.34
35. (a)
Figure P13.36
Z,(s) V,(s)
< p
(b)
(c)
If Z 2 =
as follows:
(a) Compute the input impedance
_ V^is)
Zin(s) =
(b)
T>0.
VJ(j)
i(^)
Vln(s)
Figure P I3.35
36. (a)
(d)
(e)
Compute
H^is) =
(b)
(c)
in terms of
Do not substitute
your answer to part (a) into the
obtained expression.
Compute
v ,w
= ^ .a n d
^tc.
V ,( 5 )
Figure P I3.36.
Vinis)
Vi(s)
Find the impulse and step responses
and
as
products
of
your
prior
answers.
associated with each transfer
function if
= 8, Kj = 2,
IJs)
s
, _
1
z.(s) [-0
Z,{s)
Z, =
and Z 2 =
j + 5
<$>
K.IJS)
Figure P I3.37
r\
667
CHECK: Z ^ = Z i +
K 1^2
.K
n + ^L
Vout = det(MM)/det(M)
% You have now computed the symbolic
% expression for Vout in terms o f lin so you
% can now identify the transfer function as a
% symbolic ftinction o f the variable s.
% Now lets do some numerical work.
C l= 1; C2 = 2; G l = 1; R2 = 1; G3 = 2; a = 0 .5 ;
vJt)
% Re-enter M above
M = [Cl'*s+Gl 0 1; fill in the rest of your co ^ -
G is mhos; R is ohms
Figure P I3.38
(a)
(b)
cient fnatrix]
% check: the roots o f the determinant should
% be -1 and - 5 .
dt = det(M)
factor(dt)
% Re-enter MM above
% Compute the actual transfer function
Vout = det(MM)/det(M)
% Identify the transfer function
(d)
(e)
a t,H
(c)
cient matrb^
'w ^
matrii^
dt = det(M)
dt = coUea(dt)
= -8 V .
(h)
66S
E{s)
e{t) =
}i{s)ld{s)
and
-y{t).
G, W =
e{t)
x^^,j{t) = Kf^u{t)'
conditions does e{t)
00
Under what
0 as / -
(d)
XrefU)
CH ECK :
(c)
H{s) =
F ig u re P I 3 . 4 0
E{s)
00
djis)cl^(s)
from
E{s)
H{s)X^^jsy.
d{s) + n{s)
= --------
d(s)
RESPONSE CALCULATION
WITH INITIAL CONDITIONS
4 1 . For the circuit o f Figure P i 3 .4 1 ,
C = 0.1 F,
E(s)
1 + F(.v)Gi(.9)
when
_ d{s)Xr(.f{s)
Xrefis)
when
X..,(s)
E{s) =
-> Y(s)
G,(s)
V(is)
and
initialized
L = 0 .5
H,
Figure P I3.3 9
ch ec k
+
L
l-t-G](5)
d{s) + n{s)
E{s)
nj,s)ldjj)
x^^t)
- y{t).
i.e.,
e{t) =
and
=
F{s)
)tj(s)lci^s)
is
00
when
0
R1
x^^J(t) =
e{t) -*
0.4 H
A'y(^)?
0-2 v , Y
2Ci
e(0'^)
from
E{s) =
F igu re P i 3 .4 2
A N S \ V F :R :
(e)
50
e{t) -*
x^^Jit) =
^ 0 as r - X when
(d)
H{s) = E{s)IX^^,J(s).
(c)
(b)
Figure P I3.41
v,(t)
Suppose
d^s)
F{s)
669
CHECK: /q = 2.5
46. In the circuit o f Figure P I 3.46, 7? = 20 Q,
C = 0.2 F, and the capacitor is initially charged
at V(^Qr) = 10 V.
(a) If v.{t) = 20[(f) - u{t - 20)] V, find
I({s), V^s)y if^t) , and v^ij) for r > 0.
Plot V(jJ) and v-J^t) on the same graph
for 0 < f < 40 sec using MATLAB or
its equivalent.
(b) If
V, find i({t)
and V(4J^ for ^> 0. Plot v^{t) and v-JJ)
on the same graph for 0 < ? < 20 sec
using MATLAB or its equivalent.
ic(t)
+
Vc(t)
Figure P I3.46
47. Consider the circuit of Figure P I3.47, in
which R^=l k2, R^ = S kH, and C = 50 ^F.
(a) Find the transfer fiinction H{s).
(b) Compute the step and impulse
responses.
Compute the response to v-^(t) =
(c)
Figure P I3.44
(d)
45. Consider the circuit of Figure P13.45, in
which R = 200 2.
(a) Compute the transfer function in
terms o f L.
(b) If the response to the input i-JJ) =
Intuit) A is ijij) = {2.5t - 0.025 +
Vw>
(c)
(e)
(0
10r'2-5^cos(25^)W V.
Find the response to V(iO~) = 20 V.
Using the principle of superposition,
find the response to the input of pan (c)
with the initial condition o f part (d).
From the principle o f linearity, what
would the response be to the input
= 20-^^-^^cos(12.5/)(/) V and
y j^ 0-) = 1 0 V ?
+
v Jt)
lV (t)
Figure P I3.47
u (t)
C H EC K :
Figure P I3.45
H{s) =
20
2 5 + 25
670
(c)
//2 (5 ) = - ^ ^ , a n d
VcisV
Figure P I3.49
W
Given N(s),
(b)
(c)
(d)
'
g.Vc(s) 1
C = -
9
R^ = 0.5 2,
= 5 2, C2 = 0.05 F, and a = 4 .
Compute Vciis),
assuming diat
v jt ) = 36u{t) V. /^(O-) = 0. v c ^ m = - 1 8 V,
and vufiO~) = 0 V. (Hint; You must construct
the equivalent circuit in the j-domain, account
ing for initial conditions. Consider a source
transformation on v-J^t), and then draw the
equivalent circuit in the j-domain so that you
can combine sources in the front half o f the cir
cuit.)
1 L
'1
1
Figure PI3.50
Figure P I3.48
49. Consider the circuit of Figure P 13.49. Note
that the computation o f transfer functions pre
sumes no initial internal stored energy.
(a) Use a source transformation and the
current divider formula to show that
the transfer function between V-JJ)
(the input) and I(\s) (the output) is
1
s +
RC
(b)
1
LC
(c)
(d)
(e)
(0
671
(b)
.( .) 0
99
L
's ^
(c)
(d)
V^
(e)
V^
5+
1
LC
Figure P I3.52
53. Repeat Problem 52 using ij{i) as the
desired output.
54. Consider the circuit in Figure P13.54, in
which /?j = /?2 =
Suppose V(\t) is the desired
output.
(a) Find the input impedance in the form
Z /rt(j)
in terms
poles />j
negative
ances in
S-p\
S -P 2
672
r>
(b)
+
V^(t)
99
V,(s)
(c)
id)
(e)
i,(t)
RC
LC
in terms of R, L, and C
The roots of the characteristic equa
tion (denominator o f H{s)) are to be at
- 2 y4. If Z = 0.2 H, determine R and
C. Then specify the transfer function
with the proper numerical coeffi
cients.
Given the values o f part (c),
(i) Compute the impulse response of
the circuit.
(ii) Compute the step response of the
circuit.
Suppose the input is vfi) = 10aT*%(/)
A- Find the partial fraction expansion
o f V({s) using MATLABs residue
command. Now forget about the par
tial fraction expansion and instead use
MATLABs command ilaplace to
compute the time function v^it).
Note you should define H, Vc, s, t,
Vin, and vc as symbols using syms.
For example,
R = ?; L = 0.2; C = ?;
K = 1/(R"C)
syms H Vc s t Vin vc
Vin = 10/(s+ 10)^2
% H will be the transfer function defined in
terms o f s, a symbol, and R, L, C, and K.
% MATLAB will fill in the numbers.
H = ????
Vc = H^Vin
vc = ilaplace(Vc)
Figure P 13.54
r\
r\
n
o
r\
r>
r>
r\
r\
Figure P I3.55
56.(a)
(b)
r-s
n\
673
(c)
Now suppose
1 0
If v jt) = l[u{t)
(d)
0.3 H
- u { t- 0.5)]
V, y^(0") =
0.1 H
ijt) 0 1 5 0
90 n
lO Q
Figure P I3 .56
Figure P I3.58
C H E C K : (a)
{s +
4 0 0 )]
(a)
Vj{t)
= 10(1 -
e--^ii{t)
V({s)
and C , = 2 /7 F.
Q,
when
R\C,
.v +
U 2C2 ^
"JO (b)
If
|C|/?2C ,
V(^[t)
for
t>
(d)
Figure P I 3
(e)
which /?^ = 2 k n , C = 2 0
80 n , and
}d,
If
= 20
Q,
If
for
t>
0.
= 0 , t^<^'i(0 ) = 0, and /^ ^ (0 )
= 15 V, find
.5 7
V(^{t)
for r > 0.
^(^(0 ) = 15 V, find
(a)
if
= 15 V, find
Vf^{t)
for r > 0.
=5 H.
(b)
v;,,(.v)
Com pute the transfer function
(c)
Vf(.v)
Com pute the transfer function
0
Figure P i3.59
A N SW ER S: (b)
y.
Vi Js)
- ~e^^uu) V; (c,[^-0.25;^
674
Rj
=1 H,
(a)
= 1.75 a
(c)
and Z , = 7 /8 H.
Now suppose
A,
n {s) =
) = 3 V,
?>u{t) V,
R ecom pute
the
v,(t)
L jG2 ^2
+
>v+
v
-> 0 .1 H
0.5 n
= \5u{t)
If
If
(e)
If
= 0 , /^ ,( 0 -) = 0 , and
1 5 A, find
=
V, /^ ,( 0 -) = 15 A,
iu^t)
for
t>0.
/Y Y \
0,
lor / > 0.
\5u{t)
( t ) ' . ,(t)
1Q
Figure P I3.61
t^
(c)
Vjft).
^2^1/
equations
V .(t)
0.8 F
(b)
nodal
i^{Q) ) = 0
= 5u{t) A.
- +
^C|L|
and
mesh currents.
(b)
If r^iO-) =
3iiU) V, and
0,
i^(O-)
= 3 A,
i\U)
' 6
(a)
Figure P i3 .60
(b)
A:
(c)
Find l/^ W .
(d)
Find
A; (d)
7e^\u{t) .A;
6 1 .(a)
for
t> 0 .
V(^t)
- 1 0 V'c'
2\ /,-
-h i
.v + 1 0
A.
Figure P i 3.63
CH ECK:
form
K~)U{t).
675
*;C3(0) = 2 V .
(a)
(c)
(d)
(e)
^2^
Initisl
conditions. Simplijy each equation.
Put equations in matrix form.
Assuming /jr(0) = 0,
(0~) = 5 V,
W = 5<5(/) A, and
(^) = \06{t) V,
use Cramers rule to find the current
(f) and then i^ {t).
Now suppose that ii(0~) = 0,
= OV.
= OA. and () = 10(() V.
Find 1^2 W-
1 0 I.
Figure P I3.64
Figure P I3.66
tions.
Put equations in atrix form.
Using Cramers rule, find the transfer
fimction
_
H (s)^
(e)
'O
(0
(g)
I;in(s)
o f the circuit.
Find the impulse response h{t) of the
circuit.
Find the response of the circuit to
i.J^t) = -8u(t) A, assuming the initial
conditions are zero.
Find the response due only to the ini
tial condition on the capacitor. A sim
ple observation leads to the answer
direcdy.
66
= 190 a, L =20 H,
{()
and
100 V
for all
time.
L + V .(t) -
Figure 1M3.67
v jt)l
A N SW ERS:
and
/-(/)
0 .2 5
H {s)=
4,s(.v+ 1)
Figure P i3.69
1 )//(/)dcrivcd from
( 2s +\ )
2.V
t=
to position B at time
I,
1 sec, back to A at r =
/? 2
{t) = 2{t - \ + e
Cv^O-)hit) = (e- - \)u{t).
Finally
= 500
(/)
sec.
and
V \ /V
+
R,
SW ITCH IN G PROBLEM S
6 8 . The switch in the circuit o f Figure P i 3 .6 8
c,
'1
----- 0
Figure P i3.7 0
t=I
sec, back to A
= 5 0 0 Q.,
= 10//(r) +
5u{t -
1) + 5 ( / - - 2) -
20u{t)
2 0 ( f - 1) V.
7 2 . Repeat Problem 7 0 for
R^ =
500
Q. and
+
V (t)
Figure P I3.68
Rs = Ri,=
lo g ,( 2 )
Cj = 1F, and
Cj = 1 F.
1 .4 4 2 7
iX
energy dissipated in
(a)
R^ = 2 Q, R^ = 8 Q., R^
(b)
Find
and
(c)
(d)
(e)
677
i'
+
v,(t)
Figure P I3.75
32 V
c,
Figure P I3.73
74. In die circuit of Figure P13.74, Vq = 10 V,
= 4 Q,
= 3i2, /?2 = 24Q, Zj = 3 H, and
^2 = 6 H. Suppose switch S j has been closed
for a long time while S2 has been open. Ax. t =
0 S2 is closed, after which Sj is opened. At ^ =
1 sec, S j is closed and S2 is then opened.
(a) Find /j(0") and
(b) Find z2(/) and
for 0 < f < 1 sec.
Find Zj(l) and
(c)
Find
and /2(1'^).
(d)
and
for 1 < /sec.
(e) Find
vjt)l
(a)
(c)
(d)
Figure P I3.74
:\{t)
Figure P I3.76
(b)
0.
(a)
(b)
(c)
(d)
Find
and y ^ (0 ).
Draw the ^-domain equivalent circuit
valid for t >0 .
Write a set o f nodal equations in terms
of V^i W and
Solve the set of nodal equations con
structed in part (c) and determine
67H
(e)
sec,
i.e.,
find
^ ^ (1 0 ")
and
SW ITCH ED CAPACITOR
NETW ORKS
Com pute
(a)
If
v^{t)
V(^t) for
t.
vp) = 25 V and
v^it) for 2 sec < t.
Now suppose
10 V. Find
t=2 sec
Figure P I3.7 7
C H E C K : (d)
v^Jt) = [ 1,846^(-2-5')
150 mF
100 mF
v,(t)
Figure P i 3.79
21,(s)
l.w
2V Js)l
:v,u)
ijs)
2Q
Suppose
+
V fs)
(a)
= 5 7 2 mV.
for ^ > 0.
Figure P i 3.78
v-^j^t)
Part 1.
= 2 e ^ u{t) A .
(a)
r> 0.
V(4,s).
(b)
C om pute
(c)
C om pute
Part 2. \i t =
Approximate t^^^lO).
V^Js)
(d)
lu { t -
v^Jt).
new
input is given by
three
(modified) nodal
and
10) A.
W rite a set o f
R as
an undetermined
(b)
F igu re P I 3 .8 0
for
G79
t=0'
. Com pute
v\(t)
t = 2* .
(b)
lO F
20 5(t)
5F
CH ECK :
5F
5F
20u(t)V
Figure P i 3.81
=3 V
V,
3 mF
P 1 3 .8 2 .
2m F
20 V
40 8(t)
(!)
2F
6 mF
Figure P I3.85
4F
2F
1 mF
> r i4 F
4F
2m F
10u(t)V
86. In Figure P I 3 .8 6 ,
v^{t) = - 2
V,
= 0,
Figure P I 3.82
t =2
I mF
4e-
Com pute
(b)
(c)
Plot
+
v.Jt)
rlO F
lO F
lO F
f +
Figure P I3.8 6
20u(t)V
and
v^{t) = - 2
V, / = 1,
Find
0 < t < 20
680
(b)
(c)
k = 2.
v^^ reaches
kC
and make a
M ISCELLAN EO U S
8 8. Use the material on op amp integrator
design to achieve the following input-output
characteristics. In your final design, capacitors
should be 100 nF.
0 .5
(b)
(C ) V ,
0
+ - V ;b\
(0 = -
I',h\
9 0 . (Fluorescent light) Reconsider the
0 .7 5
+
K .-
s
b\
fluorescent
light
starter
circuit
hi
RC circuit
(a)
19 V
is
681
1000
0.001 pF
Li,(0)5(t) =0.088(t)
Time domain representation of
voltage due to initial inductor
current when heat sensitive
switch opens
Very high
resistance prior
to arcing
Figure P I3.90 Model of fluorescent light starter circuit that includes a ballast resistance of 100 Q.
Note that in the time domain, the effect of the initial inductor current appears as an impulse in this
model.
CHAPTER O BJECTIVES
1.
2.
Characterize the transfer function of a circuit in terms of its poles, zeros, and gain constant.
Use knowledge o f the pole locations o f a transfer ftjnction to categorize generic kinds of
responses (steps, ramps, sinusoids, exponentials, etc.) that are due to different kinds o f
terms in the partial fraction expansion o f the transfer function.
3.
Identify, categorize, define, and illustrate various classes o f circuit responses, including
fi.S-4
the zero-srate (zero initial conditions) response, the zero-input response, the step
response, the impulse response, the transient and steady-state responses, and the natural
and forced responses.
4.
Define the notion o f the frequency response o f a circuit, explore its meaning in terms of
the transfer function, and introduce the concept o f a Bode plot, which is an asymptotic
graph o f a circuits frequency response.
5.
6.
SECTIO N HEADIN GS
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
Introduction
Poles, Zeros, and the 5-Plane
Classification o f Responses
Computation o f the Sinusoidal Steady-State Response for Stable Networks and
Systems
Frequency Response
Frequency Scaling and Magnitude Scaling
Initial- and Final-Value Theorems
Bode Plots
Transfer Function Analysis o f a DC Motor
Summary
Terms and Concepts
Problems
1. IN TRO D U CTIO N
O ur experience o f using Laplace transforms to calculate responses makes clear that the pole-zero
structure o f the transfer function sets up generic kinds o f circuit behaviors: constants, ramps,
exponentials, sinusoidals, exponentially modulated sinusoids, etc. Such knowledge leads to a qual
itative understanding o f the circuits response. For example, if a pole is in the right half o f the com
plex plane, then we know that the response will grow with increasing
t.
behavior allows us to define the notion o f stabilit)' o f a circuit or system. Generally, the pole-zero
locations allow us to categorize and com pute various special t)'pes o f responses, including tran
sient, steady-state, natural, forced, step, and impulse responses. Coupling the transfer function
with the presence o f initial conditions in the circuit permits us to define two further types of
responses fundamental to both this text and advanced courses in circuits, systems, and control: the
zero-input response (due only to the initial conditions o f the circuit or system) and the zero-state
response (due only to the input excitation, assuming that all initial conditions are zero).
These time domain notions are balanced by the concept o f the frequenc)' response of the circuit
or system. Briefly, the frequenc)' response is the evaluation o f the transfer function,
Jii).
Since
H{Jw)
H{s),
for
s=
has a magnitude and phase, the frequenc)' response breaks down into a magni-
6S"S
rude response and a phase response. A technique for obtaining asymptotic (straight-line) approx
imations (called
Bode plots)
As a final introductor)^ remark, unless stated otherwise, all circuits in this chapter are linear and
have constant parameter values. Such circuits are said to be
linear
and
time invariant.
Also, for
convenience in this chapter, the symbol Z. will be used to denote either o f two things: (i) the angle
o f a complex number,
=
L{a + jb)
= arg(/z + jb) =
z^n~^{bla), with
Z{s),
}\s),
admittances
H{s)
are rational
functions of^, i.e., they are ratios o f a numerator polynomial ;/(j), divided by a denominator poly
nomial,
r/{s).
Mathematically,
(i{s)
where
s = pj
is a finite pole o f
H{s), s = zj
{ s - p^ ) (s - P2 )...{S - P)
is a finite zero o f
z-.
If pj = pj,
j,
= 0. i-e.,
pole repeated twice is second order; one repeated three times is third order, etc. The terminology is the
same for zeros. Also, transfer functions sometimes have infinite poles or infinite zeros. If m
-*
-
00
, then
m at
H{s)
n < m, H{s)
< n and s
n
of order
O ut o f all this terminolog)' comes one striking fact: transfer functions, impedances, and admit
tances are characterized by their finite poles, their finite zeros, and their gain constant.
E X A M P L E 1 4 .1 . A transfer function
H{i)
- 3 . A t x = l , / / ( j ) = 4 /3 . Fin d M ^ ).
S
o l u t io n
From equation 14.1 and the given locations o f its poles and zeros, the transfer function must have
the form
(.S -- 1 K5 -C
^ (i-H l)(5 + 3)
H{\) = 413,
l(l + 2 ) ( l - h 4 )
15
This implies that K = 2 .5 . Equation 14.2 specifies the transfer function with K = 2 .5 .
686
Exercises.
0 0 , H{s) -*
H{s).
His) = -Ms - 1 )!{s +
- 3 . Find
A N SW E R :
2. A transfer function
known that
H{0)
H{s)
1)
A N SW E R : 2
Because the essential information about transfer functions resides with the poles and zeros, a plot
of these locations in the 5-plane, called a pole-zero plot, proves informative.
o l u t io n
T he transfer function given in equation 14.2 has the pole-zero plot shown in Figure 14.1,
JW
>k
j
-o--------- 'G
-4
-3
-2
-1
...J
FIG URE 14.1 Pole-zero plot of //(j) given by equation 14.2, where the poles arc flagged by
and
the zeros by o. Since j = a + ya), the real axis is labeled o and the imaginar)' axisyw.
(1 4 .3 )
l(5 + l) -H l](5 + 2 )
which has the pole-zero plot shown in Figure 14.2.
jco
A
-j
-3
-2
X
FIG U R E 14.2 Pole-zero plot of
H{s) given
> a
-1
...J
part of the pole or zero and the /o -axis represents j times the imaginary part of the pole or zero.
68
Plots such as those in Figures 14.1 and 14.2 com m unicate much about the nature o f the imped
ance, admittance, or transfer function o f a circuit. For example, an
RC input
impedance,
satisfies the following properties: (i) all o f its poles and zeros are on the non-positive C7-axis o f the
complex plane; (ii) all o f its poles are simple (o f multiplicit)' 1) with real positive residues, i.e., the
coefficients in a partial fraction expansion are real and positive; (iii)
j = CO; and (iv) poles and zeros alternate along the a-axis. Proofs o f these assertions can be found
in texts on network synthesis.
Exercise.
C om pute the input impedance o f the circuit in Figure 14.3, and show the pole-zero plot
if / ? ,= / ? - ,= 1
Q,
s=
oc?
J_
c.
A N SW E R :
A.C ,
More commonly, pole-zero locations provide important qualitative information about the response of
the circuit. Pole locations determine the inherent, natural behavior of the circuit, and the poles are
commonly called
tmttiralfrequencies.
may be larger than the set o f poles o f the transfer function. 'Fhis is because there might have been a
pole-zero cancellation in constructing the transfer function. The canceled pole would amount to a nat
ural frequency o f the circuit that is not present in the poles o f the transfer ftmction.
T he terms in a partial fraction expansion o f the response establish the types o f behavior present in
Kh,
pj real in each case. Figure 14.4 sketches each o f the associat
ed responses. In Figure l4 .4 a , the term Kh leads to a dc response and KlJ to a polynomial response
proportional to
. In Figure l4 .4 b , the term K l{s- p-) leads to an exponential response that is
increasing if pj > 0 and decreasing if pj < 0. Finally, in Figure l4 .4 c , if pj < 0, the response curve
the response. Each term has only one o f several possible forms. Four ver)' com m on terms are
KjJ,
K/{s -
with
has a hump.
These qualitative behaviors suggest that one important application o f the transfer function is
determining the stability o f the response; i.e., under what conditions will the circuit response
remain finite for all time?
688
T ransform
(a)
T im e R esp onse
(b)
V-
Pi
(c)
(S -
Pi)
FIGURE 14.4 Response t)'pcs common to partial fraction expansion terms, (a) The term
response and
KJ^ to a polynomial
response proportional to
KJ{s
Kh leads to a dc
leads to an expo
nential response diat is increasing i f > 0 and decreiising if pj < 0. (c) If pj < 0, the curve has a hump.
68^)
In addition to the preceding response t)'pes, there is the sinusoidal response associated with terms
o f the form
/4.V+ B
-----------^------- T
f.v + a r +
> 0.
< 0.
690
Referring again to Figure 14.5 , the real part o f a pole, i.e., - a , specifies the
response. Often, the word
damping is
decay rate
o f the
lations die out. T he farther -O is to the left o f the imaginary axis, the greater the damping. If a =
O ne concludes that pole locations specify the type o f time domain behavior o f a circuit or system.
A very important type o f circuit behavior characterized by the pole locations is
stability.
H{s)
is called
imum finite height. Interpreting this definition in terms o f the poles o f the transfer function,
one discovers that a circuit or system is stable if and only if all the poles o f the transfer func
tion lie in the open left half o f the complex plane. This makes sense, because if any poles were
in the right half o f the plane, the response would contain an exponentially increasing term; if
any were on the imaginary axis with multiplicity 2 or higher, then the response would contain
an unbounded term proportional to
axis pole with multiplicity 1, excitation o f the pole by an input o f the same frequency would
yield a pole o f multiplicity 2. The corresponding response term would be proportional to
fcos(o)r + 0 ), which grows with time an unstable behavior. W hat this means is that, for exam
ple, a unit step current source in parallel with a 1 F capacitor would produce a voltage pro
portional to
tu{t). This
voltage grows without bound and would destroy the capacitor and pos
sibly the surrounding circuitry if left unchecked. Such phenomena are considered unstable.
Despite the need for stability, some circuits utilize an unstable-like response for a finite duration.
Circuits that exhibit both stable and unstable-like responses are studied in electronics courses.
A transfer function with first-order poles on the imaginary axis is sometimes called
metastable.
Such a classification has no practical or physically meaningful significance, since the ubiquitous
presence o f noise would excite the mode and cause instabilit}' o f the circuit. Moreover, in power
systems engineering, i.e., the study o f the generation and delivery o f electricit)' to homes and
industry, transfer function poles that are in the left half-plane, but close to the imaginary axis, are
highly undesirable. Such poles cause wide fluctuations in power levels. The situation is analogous
to the way a car without shock absorbers would bounce. Much work has been done on how to
move the poles that are close to the imaginary axis farther to the left. Moving these poles to the
left increases the damping in the system and maintains more stable power levels. Summarizing,
the requirement that the transfer function have no poles on the imaginar) axis is both theoreti
cally and physically meaningful.
Exercises.
1. If
H{s)
691
unbounded.
2. If H{s) =
= l/(^^ + 1), find a bounded input that will make the response unbound
ed. Use M ATLAB or some other program to plot the response for 0 ^ r s 10 s.
A N SW E R S: fa) I/- (.v) =
io r k ^
l ;( b ) V;(a) =
lor
> 1
(.V- + 1)
E X A M P L E 1 4 .3 . During a laborator}' experiment, a student tried to build an inverting amplifi
er, as shown in Figure l4 .6 a . T he student accidentally reversed the connection o f the two input
terminals and obtained the circuit o f Figure l4 .6 b . The student was gready surprised that the cir
cuit did not behave as expected. Explain this phenomenon in terms o f the stabilit)'^ theory just
developed.
R^ = 4 k n
= 4kO
(a)
(b)
FIGURE 14.6 (a) Correct wiring of op amp circuit, (b) Accidental, improper wiring of op amp circuit.
S
o l u t io n
Assume that the op amp is modeled as a voltage-controlled voltage source with a finite gain o f 10"^
and that there is a very small stray capacitance o f 1 pF across the input terminals. Figure 14.7 illus
trates the equivalent circuit model for each o f the circuits in Figure 14.6.
= 4kO
R^ = 4kQ
6 )2
P art 1:
-av
- v ;)+ ^ (- v , - v) =0
A|
A2
Y
Substituting
v.,- v,- v = o
.V + - - - - - - - +
CRi
CR| C , j
''
CR,
\
His) =
10 13
Vi(s)
R^C
1
l+ I O '
.y + ------ +
R^C
^ + 2 .501 X 10
(1 4 .4 )
12
RjC I
as expected.
C i V , , + ^ ( V ^ - V '. ) + -^ (V r f -V ' ) = 0
Kj
/<2
which produces the transfer function
H{s) =
1013
10"
Viis)
R^C
1
1 -u r
5 + -------- h
. y - 2 .4 9 9 x 1 0
12
(1 4 .5 )
T he transfer function o f equation 14.5 has a right half-plane pole, in contrast to that o f equation
14.4. This implies that the incorrectly wired circuit is unstable, which explains the students con
cern over the surprising performance o f the op amp.
A brief interpretation o f the zeros o f a transfer function ends this section. This is best done in
terms o f a simple example. Suppose
His) =
Let
v- it)
e sin(/)/^(f)
( .v + l )- + l
V;.,
(5 + l)U + 2)(^ + 3)
V, so that
Assuming that the system is initially relaxed, i.e., all initial conditions are zero, we obtain
V^,As) = H is )V :Js ) =
1
(.9 + l)(5 + 2 )C y + 3)
693
A^-.
Observe that the response dies out very quickly and does not have
any term similar to the input signal. This follows because tiie input signal has transform poles,
= -1
J,
that coincide with the zeros o f the transfer function. One can think o f the pole locations
in the transform o f the input signal as identifying frequencies that are present in the input. Hence,
the effect o f these input signal frequencies (poles) is canceled out by the transfer function zeros,
eliminating them from the circuit response.
3. CLASSIFICATIO N OF RESPONSES
In addition to the various response behaviors discussed in section 2, there are other general
response classifications. Three fundamentally important general response classifications germane
to all o f circuit and system theory are the zero-input response, the zero-state response, and the
complete response.
Zero-state response;The response o f a circuit to a specified input signal, given that the ini
tial conditions are all set to zero. Figure 14.8 illustrates this idea.
Output
Input F(s)
Relaxed Circuit
H(s)
>
V(s) = H(s)F(s)
Zero-State Response
FIGURE 14.8 Relaxed circuit having transfer function H(s) and zero-state response.
Complete response: The response o f a circuit/system to both a given set o f initial condi
tions and a given input signal. For linear circuits, the complete response equals the sum o f
the zero-input and zero-state responses.
and/jU),
K^ 2 ^t)]
is
where /T, and A'-, are arbitrary scalars. The circuits studied in this book are linear unless
otherwise stated.
T he decomposition o f the complete response into the sum o f the zero-input and zero-state
responses is important for three reasons:
1.
6 9
~^[H{s)F{s)],
F{s).
2.
3.
It illustrates a proper application o f the principle o f superposition for linear dynamic net
works having initial conditions.
T he following example illustrates point 3.
= 6u(t)
20
Va(t)
+
Vo(t);
= 3u(t)
2 0
v jt)
1F
= 6u(t)
Response o f linear resistive network o f Figure 14.9. For the resistive network o f figure 14.9,
vj^t) due to vj^t) with
= 0 is v^J^t) = l//(f), and the contribution due to
VjJ^t) with vj,t) = 0 is
= lu{t). By superposition.
P art 1:
the contribution to
For this type o f circuit there is no initial condition and the complete response consists o f only the
zero-state response, which decomposes into the superposition o f each source acting alone.
P art 2 :
Vf^set
v j t ) = {0.5e- ^ \.5)ii{t)W
Step 2 . W ith the input
Vy{t) applied,
y^(0) = 2 V, and
= {-e-^
5)u{t)
Step 3 . An
incorrect application
695
T he last answer is wrong because the response due to the initial condition has been added in twice.
Step 4 . A correct application o f superposition would entail:
(i)
(ii)
(iii)
vj^t),
Vy{t), and
By superposition, the complete response is the sum o f all three. In particular, the zero-state
response due to
is
v J t ) = \.%\-e-^u{t)V
and the zero-state response due to
is
Vobit)
= 3(1 -
e-^n{t)
le~Ui{i).
and
Further,
It is important to note that the transfer function is defined only for circuits whose input-output
to [A"j/j(r) +
is
l^^^^ce,
the transfer function model reflects the underlying linearit)^ o f the circuit.
The complete response has a second structural decomposition in terms o f the transient and steadystate responses. T he notion o f a periodic signal is intrinsic to these classifications. A s ig n a l/r) is
periodic if there exists a positive constant
T such
thaty(r)
=j{t + 7)
to r > 0 exists because our Laplace transform analysis implicitly constrains our function class to
those that are zero for t < 0.) If a signal is periodic, there are many positive constants for which
j{t) =J{t + 7) for all t> 0. For example, i f / / ) =J{t + T) for some T and for all t> 0, then it is true
for IT, 5T, etc. We define the fundam ental period, often simply called the period, o f / f ) to be
the smallest positive constant T for w h ic h /f ) = J{t + T) for all t > 0. Sinusoids are periodic sig
nals: sin(2Jtr) = sin(2Jt/^ + 2 k ) with fundamental period T= \. The square wave o f Figure 14.11 is
periodic with fundamental period 7"= 2.
f(t)
>k.
1
-1
FIGUllH 14.11 A periodic square wave with fiindamental period 7'= 2.
This notion of periodicity and, by default, non-periodicity allows us to define the transient and
steady-state responses o f a circuit.
Steady-state response: Those terms o f the complete response that satisfy the definition o f
periodicity for r > 0. This includes a constant response.
Transient response: Those terms o f the complete response that are not periodic for / > 0 , i.e.,
that do not satisfy' the definition o f a periodic function for / > 0. Note that a constant response
satisfies the definition o f a periodic function.
A circuit response may have no transient part, as illustrated by the sustained sinusoidal oscillato
ry response o f the circuit given in Figure 1 4 .1 2a. Further, the steady-state part o f the response may
be zero, as in the circuit o f Figure 1 4 .12b, where
unstable, the transient response may blow up, overwhelming the constant or periodic part o f the
complete response, as in the case o f
{e^ cos(10/)
V. Note that transient here does not mean something that diminishes in importance with time.
Most circuits have both a transient and a steady-state response. When the input is constant or peri
odic, the circuit response approaches the steady-state response asymptotically for large
t, i.e.,
as the
transient dies out, only if the circuit is stable. For such circuits, the steady state is crucial. Further,
when the input is sinusoidal, the steady-state response is easily computed via the transfer function,
//(s ), or by the phasor method. Details o f the calculation are presented in section 4.
1F
(a)
1 H
1F
1O
1 H
(b)
FlG U lll{ 14.12 (a) Unstable circuit illustrating the possibility of no transient response,
(b) Stable circuit having a zero steady-state response.
6^)'
E X A M P L E 1 4 .5 . Com puting the response o f the circuit o f Figure 1 4 .1 3 provides a simple illus
tration o f the decomposition of the complete response into tiie sum of the zero-input and zerostate responses. Also, some rearrangement o f the terms identifies the transient and steady-state
responses.
-O
R .
>
P
>
ZJs)
1
(a)
FIG U R E 14.13 /?Ccircuit for Example 14.5. (a) Time domain circuit.
(b) Frequency domain equivalent, accounting for initial condition.
S
o l u t io n
Step 1.
V(^s)
__
R
Letting
i-^^{t) =
/q/K^) and
Rl,,
Rio
.v-i-
Step 2 .
RC
= RI,
Ki t )
RC
0 , the
zero-input response is the inverse transform o f //(i)[C y^ '(0)] as per Figure 1 4 .13b. Hence, by
superposition, the complete response is
RC
u{!)
zero-state response
Step 3 .
}e
zero-input response
dc, a step function, the complete response decomposes into its transient and steady-state parts as
''c ( 0 = ( v c ( 0 " ) - / ? /o ) ^
transient response
Rl(Mn
steady-state response
698
Observe that from the above example the transient and zero-input responses are ordinarily different.
E X A M P L E 1 4 .6 . It is sometimes mistakenly said that the zero-input response contains only those
frequencies represented by poles o f the transfer function. To see the fallac)' o f this statement con
sider the y?C bridge circuit o f Figure 14 .1 4 . Com pute the zero-state and zero-input responses.
o l u t io n
H{s)
= 1 Q. Thus,
)e x p
v^^Qr) ^
0 and
C > 0. Notice that the transfer function has no poles. As a side remark, in this case, the zero-input
response is also the transient response, with the steady-state response being zero.
T he phenomenon illustrated by Example 14 .6 occurs because the symmetry o f the resistor values
precludes excitation by the current source. Moving the current source to a different position, say,
in parallel with one o f the resistors, or changing the value o f one o f the resistors to 0 .5 Q will result
in a nonzero transfer function.
RL
wave, as show^n in Figure 1 4 .1 5 . Let /^(^) denote the circuit response. Suppose
= Acos{t)u{t)
V and /y(0~) = 1 A. The objective is to isolate the transient and steady-state responses from the
zero-input and zero-state responses.
6 9 )
USO-) =
> m v
1 H
1Q
v.(t)
(a)
FIG U R E 1 4 .1 5
RL circuit
. (b).
for Example 14.7. (a) Time domain series
RL circuit,
o l u t io n
Using Figure 1 4 .15b and the principle o f superposition leads to the response
4s
T he term
the term
response
-2
(.s + l)(5 - + l)
5+1
the zero-input
2.V +2
r
+ 1
It follows that the zero-state response is -2e~^ti{t) + 2[cos(r) + sin(^)](/). Notice that both the zeroinput and the zero-state response contain a transient part, the part proportional to
e~^. A little
rear
i^{t)
-e~u{t)
+ 2[cos(r) + sin(/)]{r) A
e~^u{t)
Exercises.
1. An
RLC network
t,
H{s)
A N S W E R : 0 .7 0 7 cos(/ - 45'")
A N SW E R : A constant
//(s ) =
\l{s +
approaches what?
1). If an input
= (1
'()()
Many books on elementary circuits contain two other notions o f response: tlie
forced response.
in
Natural response: Fhe portion o f the complete response that has the same exponents as the
zero-input response.
Forced response: 'I'he portion o f the complete response that has the same exponents as the
input excitation, provided the input excitation has exponents different from those o f the
zero-input response.
It would seem natural to try to decompose the complete response into the sum of the natural and
forced responses. Unfortunately, such a decomposition applies only when the input excitation is
(i) dc, (ii) real exponential, (iii) sinusoidal, or (iv) exponentially modulated or damped sinusoidal.
Further, the exponent o f the input excitation, e.g.,
a m f{t)
e'^ti{t),
exponents appearing in the zero-inpur response. T he natural and forced responses are properly
defined only under these conditions.
T he decomposition o f a complete response into a natural response and a forced response is impor
tant for two reasons. First, it agrees with the classical method o f solving ordinary differential equa
tions having constant coefficients, where the natural response corresponds to the complementary
function and the forced response corresponds to the particular integral. Students fresh from a
course in differential equations feel quite at home with these concepts. The second reason is that
the forced response is easily calculated for any o f the special inputs dc, real exponential, sinu
soidal, or damped sinusoidal. For example, if the transfer function is
H{s)
H{a)V^
(1 4 .6 )
To justify equation 14.6, note that the Laplace transform o f the input is
V!{s a).
plete response is the sum of the zero-input and zero-state responses, we have
+iT '
Hi s ) -----s- a
a,
a.
K
H{ s) ------- = -------- + [terms corresponding to poles of H{s)]
s - (I s - a
Using the residue formula to calculate A'leads
H{s)
s-a
s- a
to K = H{a)V.
Thus,
Executing a partial
and
L'^[H{s)VI{s - a)]
has a term,
By using exacdy the same arguments, it is possible to show that if the input is a complex
exponential function
Ve, where
both Kand
simply
H{s^) \W
A complex exponential such as
tory. However, the real part, Re[V^T (or the associated imaginary part), is simply an exponential
ly modulated sinusoidal signal, as shown in Figure 1 4 .5 , and is readily generated in a laboratory.
A derivation similar to the preceding leads to the conclusion that, if the input is Re[V^^, then the
forced response is
R e[//W K ^ 1
This relationship o f the input to the forced response prompts some textbooks to define the trans
fer function
H{s)
as the ratio o f the forced response to the input, under the condition that the
Ve^. This,
applicability o f such a definition to the broad class o f inputs for which the transfer function is
most naturally defined, as covered in Chapter 13 o f the text.
H{s)
models a
stable linear
n capaci
tors and inductors. In addition, suppose there are no comm on factors in the numerator and
denominator o f H{s) and that the
degree o f the denortiinator ofH{s) is ri. (This means that the effect
H{s).) T he goal o f this section is to develop the fol
lowing formula: if H{s) satisfies the aforementioned assumptions, and the input to the circuit has
the form
A cos(to/ +
B cos(o)r
+ (}))
(14.7a)
where the
magnitude o f the
response is
B = A\H{jco)\
(14.7b )
and the phase
shift is
(j) = 0 + L H { f o )
(l4 .7 c )
'
From an input-output viewpoint, these formulas imply that the frequency response is the steadystate response o f a circuit to sinusoids o f varying frequencies. To construct this formula, suppose
H{s).
F{s)
Y\s), as
illustrated
in Figure 14.16.
F(s) ---------->1^
H(s)
Transfer Function
FIGURU 1 4 .1 6 Frequency domain representation of hypothetical circuit.
Since
H{s)
H{s) has
H{s) will
poles, p - <
is stable, all poles lie in the open left half o f the complex plane. Assume that
-(tj
Consequently,
have a partial fraction expansion containing only two types o f terms; those having real
0, and those having complex poles with negative real parts, i.e.,
A\
A-)
H{s) = -----
^ + = -+
s -p \
S -P 2
...+
/4,,,
a- < 0.
Ci.v+D|
+ --------- 7 - 7
s - p , (i- + a , r + ( P
Specifically,
T+ -"
i)"
(1 4 .8 )
/(/) =
Acoai OJ/ +
0) =
A-------------- ---------------
0.5 Aej^ 0 .5
^(v) = --------: + ---------:
.V - yco
s + jo i
Then a partial fraction expansion o f the Laplace transform o f the zero-state response, F(j) =
H{s)F{s),
A\
At
Y(s) = ----- ^
+
s - p\ s - P2
C|.v + D\
R\
R')
+ . . . + --------- ^
^----- T + . . . + ----- ^ +
(5 + a 1 j" + ( Pi j"
- yw
A--I- yco
steady-state
contribution
=
t,
the time response due to the other terms decays to zero with increasing
o f complex variables, we obtain
t.
H(s)(s - j(o
J0.5Aej^
703
0.5Ae~j^\
jw
s + jm
S = J(0
and
/?2 =
................... ( 0.5AeJ^
0.5Ae-j^\
s + j(a }
S= J(0
0.5A\H(-j(o)\e^^^^~j^^^^
^)
Consequendy, die Laplace transform o f the steady-state response when all initial conditions of the
circuit are zero is
s-
yo)
S+ jO)
In fact, this is the Laplace transform o f the actual steady-state response, provided that the zeroinput (nonzero initial conditions) response makes no additional contribution. The zero-input
response makes no contribution to the steady-state response when one or more o f the following
reasonable conditions on the circuit are met:
1 . The network has only practical passive elements, meaning that there are always stray
resistances present.
2 . The circuit may have active elements in addition to passive elements, but remains stable
in the sense that every capacitor voltage and every induaor current remains bounded for
any bounded circuit excitation.
3. The circuit contains a total o f n capacitors and induaors, and the stable transfer func
tion, H{s), has n poles.
Under conditions 1 through 3,
704
B cos(ojr
A cos(o)r
B = A\H{Jlo)\ and
//( .
H{s)
at
s = JiO.
B and
straightforward. An alternative method is to use the graphical technique o f the next section.
At this point, it is instructive to illustrate equation 14 .7 and at the same time compare it with the
phasor method studied in a first course on circuit theory.
= cos(2r) V. Find
o l u t io n
P art 1.
Phasor method.
From the principles o f phasor analysis detailed in Chapter 10, the phasor
is
1_
=0
./4 - y
1 + yO.75
= 0 .8 Z - 3 6 .9 '
t
v^{t) = 0 .8
c o s ( 2 r - 3 6 . 9 ) V
(14.9)
P art 2 .
OS
in the ^-domain is
V i-V ;-
V i-0 .5 V |
+
=0
2.V + 2
2
v,(s)
(1 4 .1 0 )
(.v + ir
+ 1)2 = 0 .8 ^ - 3 6 . 9
v^(t) = 0 .8
c o s ( 2 /- 3 6 .9 ) V
In this example, the two methods give the same answers, as expected. Since complex numbers are
easier to manipulate than rational functions, what is the motivation for such an analysis using the
transfer function
H{s)^ W hy
not stay with the phasor method? T he next example answers these
questions.
E X A M P L E 1 4 .9 . In Example 14.8 with the circuit o f Figure 1 4 .1 7 , let the value o f p be increased
to 1.5. Find y,(r) for large
S
t.
o l u t io n
P art 1.
Phasor method.
t is desired,
involving sinusoidal steady-state analysis. The phasor domain circuit o f Figure 1 4 .1 8 yields the
single node equation
V, - 1
V, - 1.5V,
= 0
t,
vAt)
= 0.8cos(2^ + 3 6 .9 ) V
A beginner who has just learned sinusoidal steady-state analysis by the phasor method might
accept this answer. Unfortunately, the answer is not the voltage
v^{t) for
His) =
%{s)
Viis)
-2s
(s-\ r
o f H{s) in the right half-plane, the circuit is unstable. As f becomes very large,
v^{t) approaches infinit)', instead o f 0 .8 cos(2r + 3 6 .9 ") V, as calculated by the
This analysis demonstrates that the unstable behavior o f a circuit cannot be determined by the
phasor method. It is desirable to know when to use a particular method in order to avoid unnecessar)' complicated calculations. The following guidelines help:
1.
W hen the stability o f the circuit has been assured by some means, and co has a specific
numerical value, the phasor method is the better method to use for computing the
response for large
t,
stability is guaranteed include those with only passive elements, such as resistors, capaci
tors, and inductors; and amplifier circuits o f well-established configurations.
2.
sL
H{s)
method is supe
H{s)
possible with the phasor method. The examples o f frequency response calculations given
in the next section clearly demonstrate this point.
3.
H{s) should
locations checked for stability. Then step 1 or step 2 should be referred to, as appropriate.
Exercises.
V :Js)
S--0.5S + 5
5 - +0.5.V + 5. 7321
j2cos{2t + 45)t^{t).
707
is given by
2. Consider the
L C circuit
o f Figure 1 4 .2 0 , in which
= 0,
= 0, and
= 100(r).
Show that the largest voltage to appear across the capacitor for r > 0 is 2 0 0 V. Hint: Show that
V(\t)
= 100(f) - 100cos(f)(f).
L= 1 H
v,(t)
C=1F
(a)
FIG U R E 1 4 .2 0 (a) Time domain
LC circuit.
5. FREQ U EN CY RESPONSE
The frequency response o f a linear stable circuit having constant parameters characterizes the
input-output behavior o f the circuit to unit magnitude sinusoids, cos(oj/), as ca varies from 0 to
00
. This extremely important concept plays a major role in the analysis and design o f circuits and
systems. In terms o f the transfer function, the frequency response o f a stable circuit is the evalua
tion o f
H{$) zi s =jw.
cy response o f a circuit corresponds to the ratio o f the output phasor to the input phasor.
From the steady-state analysis perspective o f the previous section, if an input has the form
cos(ojQr), then the steady-state response (i.e., the response for large
out) has the form
t, after
|//(/Wq)| is the magnitude o f the complex number ^(/W q), and Z.A/(/0)q) is the angle o f the com
plex number //(/cOq). Thus, //(/(o ), for 0 s to < x , defines how a linear circuit adjusts the magni
tude and phase o f an input sinusoid to produce a steady-state output sinusoid o f the same fre
quency, but possibly with a different magnitude and phase.
An example o f practical importance is the specification o f a stereo amplifier. Here one specifies the
gain, gain-magnitude
\H{jLo)\,
musical signals are composed o f sinusoids o f different frequencies. Accurate amplification o f the
music requires that all component sinusoids be amplified with equal gain.
OS
Frequency response: The frequency response of a stable circuit or system represented by a transfer
function
H{s)
is the complex-valued funaion //(/w ) for 0 s OJ < oo. The magnitude (frequency)
A complex-valued function
H{jw)
a-^ +
+ y' ^ 2
Hijoi)
is a complex
+Jb2^ =
and
a2 + p2
P2
is the
angle or phase o f
/LH( jio) -
tan
-I
the
magnitude o f the frequency response. Typically, frequency response computation requires a calcu
lator or computer. We now illustrate the idea o f frequency response with two so-called band-pass
transfer functions in which a band o f frequencies is passed with relatively little attenuation while
frequencies outside the band are significantly attenuated.
E X A M P L E 1 4 .1 0 . Consider the two transfer functions
Q.25s
and
.v^-h 0 .2 5 5 + 1
^ _________________ 0 -0 6 2 5 . v^_________________
^
"O ')
yIabcI(Magnitude response)
gtext(2nd O rder BP)
gtexr(4th O rder BP)
we obtain the magnitude response plot given in Figure 14.21,
O)
1/1
c
CL
t/i
O
)
u.
a;
-o
13
c
cn
(TJ
Frequency r/s
FIG U R E 14.21 Magnitude responses of a second-order and a fourth-order band-pass type transfer
function. Tlie fourth-order response has steeper sides (sharper cutoff) and a flatter top.
From the transfer function and by interpolation on the plots, one observes that at co = 0 the mag
nitude is zero and at co = x the magnitude is also zero; at (O = 1 the magnitude peaks, and this frequenc)' is called the center frequency. This is characteristic o f a band-pass type o f response. T he
fact that the response peaks at OJ = 1 rad/sec means that the transfer functions are normalized.
Transfer functions o f practical band-pass circuits have much higher center frequencies. Such fre
quencies can be obtained by the technique o f frequenq' scaling, taken up in the next section.
A very important system theoretic relationship is that o f the poles and zeros o f the transfer func
tion to the magnitude and phase responses. In iJie above example for
H^{s),
given by Figure 1 4 .2 2 . One immediately notices that the poles are very close to the point yl on
the imaginary axis with the magnitude response peaked. Further, the zeros at w = 0 and
where the magnitude response is zero.
co = x
are
710
Pole-zero plot
1^
0.8
:
0.6
0 .4
S. 0.2
:
0
-
0.2
-0 .4
-
0.6
:
0.8
-1
-1
-0 .5
0 .5
1.5
-1 .5
Real part
FIG U RE 1 4 .2 2 . Pole-zero plot of H^{s) in Example 14.10.
To emphasize this qualitative discussion we consider the following example.
E X A M P L E 1 4 .1 1 . Iwo circuits have transfer functions
H^{s)
shown in Figures 14 .2 3 a and b, respectively, and gain constants o f 1. Qualitatively speaking (with
out doing any com putations), what can we deduce about the magnitude response o f each circuit?
To verify our qualitative deductions we will use M ATLAB to construct the exact magnitude
response.
JCO
A
.. 12j
jO)
A.
X
- 12j
8j
8j
4j
4j
---------- >
II
-4j
-4j
-8j
- -8j
- .- 1 2 j
T
- .-1 2 j
>r
(a)
(b)
FIG U R E 1 4 .2 3 Pole-zero plots for Example 14.11. (a) H^{s) (b) Hjis).
11
o l u t io n
P art 1.
H^{s)
Qualitative analysis o f Figure 14.23a. This figure tells us that there are r\vo finite poles o f
\2j, but in the left half o f the complex plane, and that there is a finite zero at the ori
near
\H{jto)\ - 0 as ( 0 - 0
\H{j(o)\ -* 0 as (O - oo
(iii) \H{jco)\ = maximum value
(i)
(ii)
as co - 12
In the case o f point (iii), we can say in general that in some neighborhood o f to near the pole, the
transfer function peaks in magnitude.
Qualitative atialysis o f Figure 14.23b. This figure tells us that there are again rvvo finite poles
Hjis) near 12y, but in the left half o f the complex plane, and that there arc two finite zeros at
P art 2 .
of
\H{jco)\ - 0 as oj - 0
\H{jo))\ constant as to -
(iii) \H{jo})\ ~ maximum value
(i)
(ii)
oo
as co
12
al magnitude response plot for A/, (yco), 0 < co < 3 0 rad/sec. To construct the plot shown in Figure
1 4 .2 4 , we use the following M ATLAB code:
w = 0:. 1:2 0 ;
n = [1 0];
d = poly([-0.1 + 12*j -0 .1 -1 2 * )]);
h = freqs(n,d,w);
plot(w, abs(h))
grid
xlabel(Frequency in rad/sec)
ylabel(Magnitude H 1 (jw))
12
Exercises.
H^{s)
response plot for 0 < to < 3 0 rad/sec. Use iMATLAB to construct this plot. Verif)' the accuracy of
the qualitative predictions.
2. Show that the transfer function
>
.V" +
H{s) =
R=
1
s~ +
LC
.y+
RC
(1 4 .1 1 )
LC
Frequency in r/s
FIGURH 1 4 .2 5 Plot of the magnitude response of the transfer function of equation 14.11.
The example below further illustrates this relationship with a pedagogicaily useful graphical tech
nique. Mastering this technique helps concretize the meaning o f magnitude and phase and rein
forces the qualitative discussion above on using pole and zero locations to compute the magnitude
and phase.
E X A M P L E 1 4 .1 2 . 'lb better grasp the ideas o f the magnitude, |/y(/'to)|, and the phase, Z.//(ja>),
o f a frequency response, suppose a transfer function has the form
H{s) =
where Zj =
2j and
= -2j.
(s-Zi)(s-Z2)
is+]){s + \ + j ) { s + \ - j )
(1 4 .1 2 )
H{s).
~ 13
JO )
JO )
()
2j
-1
X
-J
-2j
(a)
FIG URE 14.26 (a) Pole-zero plot
of H{s) as given
o l u t io n
\H{J\)\. This
H{jio) graphically
for oj =
, i.e., to com
tor has a length that can be determined either graphically, by physically measuring the distance
with a ruler, or by the Pythagorean theorem. Figure 14 .2 6 b illustrates the idea.
Step 2 . Following from step 1, the magnitude
\j\ -y'2| x
N y i)| =
o f //(/I )
1/1 +y2|
1x3
= 0 .9 5
V I X V5 X 1
H{j\) graphically.
- z - ory ( 0 - p- can be
In Figure l4 .2 6 b ,
represented in the
form
x^\
where ip is the angle the vector makes with a horizontal line passing through its base.
For example,
{j\ - J2)
= -j\
uct o f two complex numbers is the sum o f the angles, and the angle o f the ratio of two complex
numbers is the angle of the numerator minus the angle o f the denominator. Fience, from the
angles shown in Figure 14 .2 6 b ,
LH ijl) = L { J \ - jl)
L{j\
+jl)
- a j\
+ 1) -
= _ 9 0 " + 9 0 - 4 5 - 0 - 6 3 .4 = - 1 0 8 .4
Uj\
+ 1
Uj\
+ 1 -j)
(1 4 .1 3 )
{s - P\){S - P2 )---{s- P)
Because the magnitude o f a product o f complex numbers is the product o f the magnitudes o f the
numbers, and because the magnitude o f the ratio o f two complex numbers is the ratio o f the mag
nitudes o f the numbers, the general form o f the magnitude response o f equation 14.13 is
Zx
|yto-C2
- 1>\ j M - f h
(1 4 .1 4 a )
,/ w - /^
Similarly, since the angle o f the product o f complex numbers is the sum o f the angles o f the num
bers, and since the angle o f the ratio o f two complex numbers is the difference in the angles o f the
numbers, the general form o f the phase response o f equation 1 4 .1 3 is
= [Z-(/to -
H{jn))
+ /-(yw -
[Lijco-p^)
LK\
(1 4 .1 4b)
tends to have a large magnitude for jcos near poles and a small
magnitude for jcos near zeros. As mentioned earlier, this can be used to advantage in estimating
the magnitude response and phase response o f a transfer function.
Exercise.
Draw an estimate o f the general shape o f the magnitude and phase response o f the
M i> =
5 +
Jls
-I- 1
Com pute the exact magnitude and phase at lo = 1. W hat happens to the magnitude and frequen
cy response if H{s) is changed to
H^{s)
H{sl\Q)
and
H^{s)
= / / ( / / 1 00)?
= /7 i(/]()0 )
= 0 .7 0 7 .
normalized.
tion o f a normalized design, engineers can frequency-scale to obtain realistic frequency responses
and magnitude-scale to obtain reasonable impedance levels as necessary to meet power and energ)' restrictions.
FR EQ U EN C Y SCA LIN G
E X A M P L E 1 4 .1 3 . The circuit o f Figure 1 4 .2 7 realizes the transfer function o f equation 14.11
from the previous section for /? = 1 Q , Z, = 0.1 H , and C = 1 mF. We pose the following question.
Suppose
->
----- , C
Xf
and
R remains
H{s)
FIG U R E 1 4 .2 7 Band-reject
type circuit
S O L U T IO N
Step 1.
H{s) =
R +Cs
Step 2 .
By voltage division,
LC
RC
------
LC
Ls
R
R+
R+
s +
s\
Kf
s V
LC
^ andA T
C
K
with
R unchanged,
the new and old transfer functions are related in a ver>' simple way:
\
.y
A'
I )
Step 3. Plot the magjiitude response. Rather than go directly to M ATLAB, consider that
to
Kf I
^old
to (rad/s)
0
I
1
~ ^old J
\ ^ Jf /
^ newU^f
Xf
2f
)~
s by s!K^. T he
H^[J,s)
s=
where
the term
before scaling
frequency scalhig.
In terms
o f the plot o f our specific transfer function, then, we obtain from Figure 14.25 the new plot o f
Figure 14. 28 by inspection.
Kj-
state the
2.
Kj-
KjrZ- and
for / = 1, . . . , w and
are those o f
k=
quency' scale multiplied by / y Conversely, the magnitude and phase o f //(/to ) are the
same as
Exercises.
2. Given
point 3 above.
in step 2.
K f Z \ ) " { s K fZf/i)
--------- ---------------------------U- K
- K^p)
\ //-m { s
= KlK r)
H(s) = K
(5 - /? l)( i -/> 2 )
find the new pole and zero locations if the circuit is frequency-scaled by
1000.
OAs
i* + 0.1
Figure 14 .2 9 b shows the pole-zero plot and Figure 14 .2 9 c shows the magnitude response with
peak value at
= 1 rad/sec.
(a) Frequency-scale the circuit by the factor /y = 10^ and compute the new transfer function.
(b) C om pute the new pole-zero plot and the new magnitude response curve.
/YY\
-I-
V J
>
0.1 Q >
(a)
718
JCO
FIG U R E 14.29 (a) Series resonant circuit, (b) Pole-zero plot, (c) Magnitude response.
S
o l u t io n
Kj-=
JO )
(-0.05+ j0.999)10^
X
for
Hnew(s)
o
(b)
FIG U R E 1 4 .3 0 Demonstration of the effects of frequency scaling, (a) Frequency-scaled circuit, (b)
New pole-zero plot, (c) New magnitude response curve.
10^/
10^ /
/ 5 ^
( S
+ 0.1
+1
lio^/
UO^i
(b) The ideas stated in part (a) are borne out by comparing the pole-zero plot and the magnitude
response curve for
Kj-^s
H{s)
in
H{s),
= 10^ rad/sec.
At this point we round out our discussion o f frequency scaling by relating it to the time domain
via the time/frequency scaling property o f the Laplace transform:
or, equivalently.
K fl
=H
( s \
\ j /
where H{s) is the circuit transfer function and h{t) is the circuit im pulse response.
'2 0
H(s) =
30
10
10
(5 + 2 ) ( 5 + 5 )
{s + 2)
(5 + 5)
h{t)
\Oe-">)u{t)
Kj:
o l u t io n
f-lneAs) =
(5 + 2 /r y )(5 + 5/C|)
ioa :,
(^ + 2A:/-)
(5
/:^ )
- \QKfe~^^Au{t)=Kf h[Kfi)
t In mllli-sec
(a)
h{t)
and
^21
t in milli-sec
(b)
FIG U R E 14.31 Illustration of the effect o f frequency scaling on impulse response.
(a) Original h[t). (b) h^J,t).
Observe that the plots are structurally the same. However, the time scale in Figure 14.30b is
in milliseconds and the magnitude has been scaled by
Exercises.
Kj~=
10-^.
1. Use the Laplace transform time scale propert)' to show that in general
2. In Example 14 .1 5 , suppose the poles are changed from - 2 and - 5 ro - 3 and - 6 , respectively.
Suppose further that the transfer function is scaled by
- 10^--^0//(/) and
h,
(/) = (lOOt"-^"' - I
^ouJ^in = ^-5
= 2Q.
Suppose
both resistances are made 1 000 times larger, as shown in Figure 14 .3 2 b . W hat happens to the volt
age ratio and input impedance?
lOOOQ
1O
------0
+
10000
1O
-O
(a)
(b)
F K iU R E 1 4 .3 2 (a) A normali7.ed voltage divider circuit, (b) The voltage divider of part
(a) magnitude-scaled by 1000.
S olution
By direct calculation, the voltage ratio remains unchanged, i.e.,
1000
out
\ V-in
I new
10004-1000
=0.5 =
out
\V
''in
old
magniti(de-scaled (by
The above example motivates a more general discussion. Suppose each impedance in Figure
14 .3 3 a is scaled (multiplied) by
()
4 , (,) = K,Z^ W
Similar to the case in Example 1 4 .1 6 , the input impedance is increased by the scale factor
The
If Zj is an inductance, then
mtiltipliedhy K^.
= sl{CIK^, i.e.,
Z^{s)
Ls and
dividedhy
Z^{s) =
1 /G and
723
Z,(s)
K .Z ,(s)
-o
+
V r A
J
---w
Z
Z,(s)
-o
(b)
(a)
FIG U R E 1 4 .3 3 . (a) A general voltage divider, (b) Magnitude scaling of the circuit in part (a) by
T he above discussion suggests that magnitude scaling can be achieved by multiplying resistances
and inductances by
K^.
1.
2.
3.
4.
For voltage-controlled current sources (V C C S), i.e., the^^^ typ^> divide the parameter
5.
by
Parameters for voltage-controlled voltage sources and current-controlled current sources
remain unchanged.
6.
The effect o f magnitude scaling on a transfer function is set forth below, with its verification left
for more advanced courses on circuit theory.
Effect o f m agnitude scaling on transfer functions: If
magnitude scaling has no effect on
H{s).
H{s)
work has
work has
Z{s)
I
s 2 + 0 .I ,v + l
= i-5-k 0.1 =
0.l5
v;-,,
and
Z(5)
.r-hO.I.y+1
= 1000.
/24
_JYY\
Im F
V
1000 H
V.
Znew(s)
(b)
FIG U R E 1 4 .3 4 Magnitude scaling, (a) Original network, (b) Network magnitude scaled by 1000.
So
lu t io n
Rs, L's,
Figure 14 .3 4 b .
(b) For the simple circuit o f Figure 14.34b , the new input impedance is
Ze,,.(s) =
+ 1 0 0 0 s + 1 0 0 = 1 0 0 0 ^ + 0 - 1 ^ + ^ ^ l O O O Z (s ) =
K,Z{s)
100______________ l l i
1000
.y + 0 .h + l
.V-+0.1.V + 1
= H(s)
Because the transfer function is a voltage ratio, there is no change. These results clearly illustrate
the stated effects o f magnitude scaling.
Exercises.
= 5.
and
/? 2
are magnitude-scaled by
Verify that
old
Rf,q =^m^eq
EXAM PLE 14.18. The amplifier circuit shown in Figure 14.35a consists of r\vo stages. The first
stage is a Sallen and Key low-pass active filter. The transfer function of the Sallen and Key circuit is
H2(s ) =
V
^Ollt____-'>
V,
.v +
His) =
V
^Olll _ V
* out. V
^(1^ =
...........................
y.
y y
-1
- " ^ 1
------------------------------------------------------------ {s + 1) (.v + ^ + 11 5 ' + 2 . r + 25 + 1
FIGURE 1 4 .3 5 A third-order Butterworth low-pass filter, (a) The original network, (b) The scaled
network with
K^= 2000ti
and
= 10,000;
= 3 1.4 uS.
The overall transfer function o f the two-stage amplifier is a third-order, maximally flat
(Butterworth) low-pass filter, which will be studied in Chapter 19. The magnitude response o f the
transfer function has a 3 dB down frequency at OJ = 1 rad/sec, or l/(2ji) Hz = 0 .1 5 9 Hz an
extremely non-useful audio frequency. In fact, when the 3 dB down frequenc)' is at (o = 1 rad/sec,
the magnitude response is said to be
normalized.
values are unsuitable for practical applications. However, using both magnitude and frequency
scaling, this reference or normalized amplifier circuit can be made into a very practical filter.
Suppose we wish to have the 3 dB down frequency a t^ ^ ^ = 5 0 0 Hz and the largest capacitor at
10 nF. Such a filter could be used to direct the low-frequency content o f a music signal to a woofer.
O ur goal requires that we frequency-scale the circuit by
must satisfy
10 X K )-'' =
Solving for
K jK ,
lOOOrcK,,,
yields
K^,
m = 6 .3 6 6 X lO'^
T he scaled circuit meeting the requirements is shown in Figure 14 .3 5 b , where
(5")
cs
= -^ ------------- ,with
Avj
s + fl.v + h
= 4 and
= 3 1 .4 |iS.
a = Q.\, b = Aand c = ^ 0 .
Kjr= 2,
a, b,
c.
Examples 1 4 .1 5 and 1 4 .1 6 illustrate one o f several reasons for scaling a linear nersvork. Chapter
19, on elementary filter design, will derail further applications.
nack
erence signal and the circuit output are more or less indistinguishable. To accomplish this,
design/control engineers generate an error signal,
put and
is a given reference signal. Since much o f the analysis is done in the frequency
E{s)
Y{s) -
Ordinarily, the design engineer needs to know the initial error, if(0), and the final error, f(co).
Available to engineers are the initial-value theorem and the final-value theorem, which permit the
computation o f these quantities in the frequency domain.
= F{s)
s,
Proof: The.
sF {s)-f(Q -)= L
on j{t).
y/(0
dt
(1 4 .1 5 )
Specifically,
/ 27
Applying the definition o f the Laplace transform to the right-hand side and taking limits as
lim
s-*
,V - X
(1 4 .1 6 )
lim C h O e -^ 'd t
v-oc J O
-t- lim
S-X
C h t)e-^ 'd t
J{)
where the dot over the function/(r) indicates the derivative o f the fimction. Observe that
fiOe'^'dt
lim
S -*
/(/)
Wm (e~")dt = 0
30
is continuous at r = 0,
lim
.V-*30
lim
j! r
s-*x
(1 4 .1 7 )
Hence, the left-hand side o f equation 1 4 .1 6 equals the right-hand side o f equation 14.17.
Equation 14.15 follows from equating these two sides and canceling th e /(0 ) terms in both.
.V
5i + 2
EXAMPLE 14.20.
)= lim
i->oo
.vV(^('.yJ=
lim
2-
= 2 - - = 1.8 V
5
500^ + 20
V{ s) = ----------------------------i-(5 5 -h 2 0 )(1 0 .v + l)
Find the initial velocity, y(0'^), and the initial acceleration, //(O"^).
S o l u t io n
s * x
5 0 0 ^ -f 2 0
(55 + 20)(10. v+ 1)
=0
Since acceleration is the derivative o f velocity, and the initial velocity is zero, from the time dif
ferentiation propert}'^ o f Table 12.2, assuming the velocity is continuous at f = 0,
2S
/\(.v) =
sV{s) -
500.V + 2 0
v(o^
(5.s + 2 0 )(1 0 .v + l)
20
f/(0 ) =
Urn
.V-
5 () 0 i- +
20 s
(5.V + 20X10.S + 1)
Urn
i-^oo
5.V +
20 \
s!
.y /
Exercise.
Suppose
F{s)
= (8j +
2)1{Isr
10
IO.V+-
.\N SW ER : 4
T he initial-value theorem and the examples that follow it illustrate the computation o f initial val
ues. To compute final values vve use the next theorem.
Final-Value T h eorem : Suppose
F{s) has
poles only in the open left half o f the complex plane, with
Proof:T\\Q condition
F{s)
the complex plane, with the possible exception o f a first-order pole at the origin, guarantees that
the limit on the right side o f equation 1 4.18 exists. This is because a partial fraction expansion of
F{s)
leads to a time fu n c tio n //) that is a sum o f exponentially decaying signals and at most one
= { lim
\/-X
/ ( / ) ''- / ( ( ) ) .
/
E X A M P L E 1 4 .2 1 .
v{t),
Laplace transform
V ( .) =
500^ + 20
.v(55 + 20)(10.v-h l)
29
S o l u t io n
K(y) and
A{s)
have poles that meet the conditions o f the final-value theorem. Hence,
5005+20
= 1
(55 + 2 0 X 1 0 5 + 1 )
and
5y\(.0 =
.v-()
lim
.v-^o
5 0 0 5 + 20 5
= 0
;5 5 + 2 0 )( 1 0 5 + 1)
Observe that a constant final velocity implies a zero acceleration as these expressions indicate.
E X A M P L E 1 4 .2 2 . W hat if the conditions o f the fmal-value theorem are not met? W hat would
go wrong? A simple example illustrates the problem. Let
Fis) = -
5 + 1
which corresponds
toJ{t)
= sin(r)//(r). Then
lim
sF{s) =
lim ^ 5 ----- = 0 ,
1
,v-*0 5 " +
but
lim
/{[) =
lim
t-*x
sm(t)ii{t)
is undefined, i.e., it does not exist. The theorem, however, presupposes that both limits exist.
Again, the condition of poles in the leh half complex o f the plane with at most one pole at the
origin is necessary and sufficient for both limits to exist.
Exercises.
1. If
F{s)
5y{t- 2)u{t-
2), is
1 5 t ^ - - - '( 5 - + 5 - 2 )
.v(5" + 5 5 + 6 )
t.
f{t)
approaches what
"30
8. BODE PLOTS
Section 5 described the use o f the poles and zeros o f N(s) to compute the frequency response o f a
circuit. In this regard, Hendrik Bode developed a technique for computing approximate or asymp
totic frequency response curves. These so-called Bode plots can be quickly drawn by hand. A
description o f the technique requires the introduction o f some terms widely used in the engineer
ing literature.
Let //(s) be a transfer function that is a dimensionless voltage ratio or a current ratio. As explained
in section 4 , for sinusoidal steady-state analysis, one replaces s by yoj to study the circuits magni
tude response, |//(/w)|, and phase response, Z .//(/co). For convenience, let |//(/to)| be a voltage
gain, |K,/Kj|. T he ga/;/
in dB
(decibels), denoted by
= 2 0 log,|//(/o)|
(14.19)
For convenience, whenever we write log(.v), we will mean log,Q(>;). Solving for |//(/w)| in equation
1 4 .1 9 yields the inverse relationship
(1 4 .2 0 )
Table 14.1 presents some pairs o f |/y| and
we may say that
is 2 0 dB
above
K,, or that Kj is 2 0 dB
below 2 - Similarly,
to say that
V-,
is 3
^dB
Exercise.
V2
10
100
1000
s3
^6
20
40
60
A N S W E R : H)-'
O ne of the reasons for using the dB terminology is that it simplifies the analysis and design of
multistage amplifiers. Suppose an amplifier has three stages with voltage gains equal to 2 0 , 100,
and 10, respectively. The overall voltage gain is the
product o f
sum of
the dB gains o f the individual stages. This is (2 6 + 4 0 + 2 0 ) = 8 6 dB. It is easy to justify this claim.
First,
|H| = |//,| X |//,| X |//,
731
^dB
^2,dB + ^5,dB
This summation has pronounced advantages for repetitive calculations at many frequency points,
as when plotting a magnitude response such as equation I4 .l4 a . We could, of course, convert this
equation to an equation having all terms in dB. However, with an eye toward a further simplifi
cation, it is desirable to first rewrite H{s) in a slighdy different, but equally general, form, namely
+ 1
H(s) = Ks' j
'
(14.21)
-Z 2
^ - + l'
V -P i
where \p^, p 2 >} are those poles o f H{s) that are not at the origin and {zp 2^, ...} are those zeros
of H{s) that are not at the origin; if a is positive (negative), then H{s) has a zeros (poles) at the
origin. For example, a transfer function
(^ + 5 0 ) (5 + 2 0 0 )
(5 + 8)( j + 6 0 0 )
has the equivalent form
O '
H {s) = 2.5
^50
+ 1
Uoo
(f-l ^,600
Observing that
+ 1
(14.22)
+ 1
= 10 log^^|//(70))p, setting s = /a) in equation 14.21, and noting that the mag
nitude squared o f a complex number is the imaginary part squared plus the real part squared yields
0.5
HdBi(o) = \K dB
0)
(0
dB +
dB
2
,0.5
0.5
v2
\Z2/
+ 1
dB
(14.23)
-,0.5
( co\
dB
Equation 14.23 suggests that we may compute the dB vs. (O curve for each term on the right-hand
side and graphically sum the curves to obtain the desired H jg vs. (O curve. However, each indi
vidual curve is reasonably sketched by using log(o)) instead o f 0) as the independent variable. This
amounts to using semi-log paper to plot the dB vs. (O curves. Such a plot, with a linear scale for
the dB values on the vertical axis and a logarithmic scale for O) on the horizontal axis, is called a
Bode magnitude plot, in honor o f its inventor. Similarly, a plot of LHijiiS) vs. O), with a linear
scale for Z.//(/co) and a log scale for co, is called a Bode phase plot. Note that, because of the log
arithmic scale for the (o-axis, the actual distance on the paper between (O = 1 and O) = 10 is the
same as that between o) = 0.1 and O) = 1. (See Figure 14.36a.) Note also that the (O = 0 point will
not appear on the graph, because log(co) approaches -oo as O) approaches 0.
W
"3 2
With log(oj) chosen as the independent variable, the plot of each term in equation 14.23 either is exact
ly a straight line or is a curve having two straight line asymptotes. I'his is illustrated in Figure 14.36.
FlG lIR Ii 1 4.36 Bode magnitude plots for three basic transfer functions,
(a)
H{s) = K> 0.
(b)
H{s)
sin, a > 0.
(c)
733
In Figures 15.36b and 15.36c, the rising asymptote has a slope of 20 dB/decade, which means that
along this line, an increase in frequency by a factor of 10 causes an increase in gain of 20 dB. Here,
the word decade (abbreviated dec) simply means 10 times. Another way to express the same
slope is to indicate the increase of gain in dB when the frequency is doubled, or increased by an
octave, in music terminology. It is easy to see that 2 0 dB/dec is equivalent to 6 dB/octave. In gen
eral, if a frequency, (Oj, is d decades above another frequency, coj, then, by definition, (cOj/cOj) =
lO'^. Conversely, if (0)2/t0 j) = r, then we say that (Oj is log(r) decades above (Oj.
In Figure 14.36c, the left asymptote is a horizontal line and hence has a slope of 0 dB/dec. The
point where the two asymptotes intersect is called the breakpoint, and the corresponding fre
quency is called the l>reakfrequency or comer frequency.
The derivations o f the true curves and asymptotes in Figures 14.36a and 14.36b are very simple
and are left as exercises. For figure 14.36c,
in which case
+1
+ 1
and
| H ( | ^ =
For 0)
201o g
( l) '
+1
a.
|H(j<o)|jB = 201o g [-y r] = 0
indicating that \H{j(a)\^g approaches the left asymptote in the figure. For co
|Oo))|dB = 201og ,1 - )
V\a )
(I)'
+1 as 20 log -
a,
=201og(co)-201og(a)
indicating that |^(/(o)|^^ approaches the right asymptote in the figure. The two asymptotes inter
sect at the point ( cd = a,
between the true value of |//(/co)|^^ and the value read from the asymptotic curve. The error at
twice or half the corner frequency is about 1 dB.
The following variations of the three basic Bode magnitude plots o f Figure 14.36 are easily
derived:
1.
If H{s) = {sld)^y the Bode magnitude is similar to that shown in Figure 14.36b, except
that the slope is now 2 0 dB/dec. The curve still passes through the point (co = a,
2.
If H{s) = {sla + 1 ) , the Bode magnitude is similar to that shown in figure 14.36c, except
that the right asymptote has a slope o f 2 0 dB/dec. The breakpoint is still at (o) = a,
|//(/co)|^^ = 0), and the error at the comer frequency is 3 dB. If n is negative, the right
asymptote points downward.
734
Let us now consider the Bode plot for a general transfer function
H{s).
After expressing
H{s)
in
the form of equation 1 4 .2 1 , we can draw the asymptotes for each term in equation 14.23 with the
aid o f Figure 1 4 .2 6 . The asymptotes for the
ily by graphically summing the individual asymptotes. Since the asymptote for each term in equa
tion 14.2 3 is a piecewise linear curvey the graphical sum o f all the asymptotes is also a piecewise lin
ear curve. Accordingly, it is not necessary to calculate the sum o f dB values at a large number o f
frequencies. If there are
n break frequencies,
frequencies and for the slopes o f the leftmost and the rightmost segments o f the piecewise linear
asymptote. T he following example illustrates this procedure.
H{s)
o f equation 1 4 .2 2 rewritten
below:
I s
+1
H (s)= 2.5
V50
-M
/ V200
/
+ 1
V8
Solution
H{s)
/ V6 0 0
as a product o f five factors with break frequen
H{s) =
2.5 - + 1
.8
+ 1
.5 0
.2 0 0
+1
.V
.6 0 0
+1
i}4.24)
= / / , X / / . X //3 X //4 X
Figure 14.3 7 a shows the asymptotes for the five individual terms in equation 1 4 .2 2 , and Figure
14 .3 7 b shows the asymptotes for
^^5
CO
(rad/sec)
(b)
FIGURE 14.37 Asymptote and Bode magnitude for H{s) of equation 14.22 or, equivalendy, equadon
14.24. (a) Bode plots for / / , through
Hy
through
T he slope o f the leftmost segment, i.e., the segment to the left o f co = 8, is obviously zero,
2.
3.
= 2 0 log(2.5) se 8.
H^{s)
Pj
is
8 - 2 0 X 0 .7 9 6 = - 7 .9 2
4.
There is a breakpoint
5.
There is a breakpoint
0) > 6 0 0 ,
additional 2 0 dB/dec to
resulting in a slope o f zero, we have - 7 . 9 dB for P^
at co = 6 0 0 . Since H^{s) contributes an additional 2 0 dB/dec to the
at co = 2 0 0 . Since
contributes an
P^ is
- 7 .9 2 + 2 0 X 0 .4 7 7 = 1.63
6.
tional-20
contributes an
addi
dB/dec to the slope for co > 6 0 0 , the resulting slope o f the rightmost segment
is zero.
T h e com p lete specification o f the piecewise linear asym ptote is shown in Figure 14 .3 7 b .
'3 6
Exercises.
(') =
K and
(ii) for
out _
^in
s = ja,
1 4 -
i.e., at co =
a,
H{s).
function H{s)
= 40j-/[(i- +
2){s +
2 0 )].
O nce the piecewise linear asymptote for the Bode plot has been constructed, the true curve can
be sketched approximately by noticing that the error at each corner frequency is about 3w dB (pro
vided that the two neighboring corner frequencies are more than five times larger or smaller).
Figure 14 .3 7 b shows such a rough sketch. In the pre-com puter era, the ability to draw such a curve
by hand even a crude approximation was considered valuable. Nowadays, with the wide avail
ability o f personal computers and C A D sofuvare, one might just as well get the exact plot with
out bothering to look at the straight-line asymptotes. From the perspective o f circuit analysis, the
value of the ability to construct the asymptotes for a Bode plot is greatly diminished, but the tech
nique is still important for its application in the design o f feedback control systems. Such an appli
cation utilizes both the Bode magnitude plot and the Bode pha.se plot. Some background in con
trol systems is required for one to understand the use o f the Bode technique in this kind o f appli
cation. We will relegate the discussion of the topic to a more advanced course in feedback control.
O ur objective in mentioning it in this text is twofold: (1) to introduce the definitions o f some
com m oni) used terms, such as
decibels, decade,
and
octave,
tematic procedure for adding up several piecewise linear curves to obtain a desired curve, as
described in Figure 14.37.
are permanent magnets that interact with a magnetic field produced when a current flows through
the armature winding. This interaction o f magnetic fields forces the shaft to rotate in a process o f
energ)' conversion. Here current flowing through the armature coils rotating through the magnetic
field produced by the permanent magnets encasing the coils produces a torque on the motor shaft,
which drives a load. Power is delivered from the source to the load.
The electromechanical characteristics o f the permanent magnet dc m otor have a simple circuit
like model amenable to Laplace transform analysis. The model is given in Figure 14.38 and con
sists o f an adjustable dc voltage source in series with a resistor
labeled m otor. Here,
an inductance
and a device
resents the equivalent inductance o f the wire coil. The device labeled m otor has the current
rep
i J^t)
as an input and the angular velocity co(r) o f the rotating shaft as an output. T he interaction
between the electrical part o f the model and the mechanical part o f the model occurs at the loca
tion o f this symbol. T he voltage
oj(r). Because co(/) is not a circuit variable, classical notions o f impedance, admittance, voltage
gain, etc., do not fit the problem, whereas the more general notion o f transfer function does, forc
ing us to move slightly beyond the confines o f circuit theory proper to analyze the system.
The torque
J\t)
produced on the rotating shaft by the current flowing through the armature coils
Tit)
(1 4 .2 5 )
T he mechanical rotation o f the m otor affects the electrical portion of the system by inducing a
voltage
This voltage is proportional to the m otors rotational speed, or angular velocity, co(r),
i.e.,
(1 4 .2 6 )
Since the m otor converts electrical energy to mechanical energ^^ conservation of energy dictates
that the constant o f proportionalit)' be equal to the same constant that relates torque and current
for a lossless motor. Specifically, electrical power-in must equal mechanical power-out or, in equa
tion form.
= m coit)
which forces
k.
H{s)
[o){t)]/V^{s).
For convenience,
let Q(^) denote X [o)(r)]. As a first step, sum the voltages around the loop of elements in the cir
cuit model of Figure 14.3 8 . This results in the differential equation
V (/)
at
(1 4 .2 7 )
A ssum ing zero initial con ditions, the Laplace transform of equation 1 4 .2 7 is
kQis)
(1 4 .2 8 )
738
From basic mechanics, the differential equation governing the mechanical portion o f the system
r(o = y ^ + B < o (o
at
where
is the moment o f inertia of the combined armature, rotating shaft, and load; B is the
coefficient of friction; and 7\t) is the torque produced by the motor to turn the load. Recalling
equation 14.25, T\t) =
ditions, is
(14.30)
Solving equation 14.30 for IJis), substituting the result into equation 14.28, and then solving for
Q(j) leads to the expression
k
2 V,{s)
Q(s) =
(14.31)
Equation 14.31 characterizes the pertinent dynamics of the permanent magnet dc motor and
allows one to find the angular velocity o f the motor shaft as a function of time for given inputs.
To see how the motor responds to step inputs, suppose vjit) = Ku{t). As objectives, let us find (1)
the steady-state value, or final angular velocity, o f the shaft and (2) the steady-state value o f the
armature current. The final speed o f the shaft is important because, for example, one needs a fixed
speed for the rotation o f a compact disk or a fen. The final or steady-state current is important for
determining the power needed from the source.
If Vj(t) = Ku{t), then V(^s) = KIs. It follows from equation 14.31 that
S +
S+
RgS + k
2\
(14.32)
J
Applying the final value theorem to equation 14.32 implies that
(Ovv =
lini
co(r) = ------- y K
To isolate the armature current IJJ), again apply the final-value theorem to determine the steadystate value of i {t). Combine equations 14.30 and 14.31 to obtain
1
5+
4 (^ ) =
2 's
LaJm
( 14.33 )
Equation 1 4 .3 3 allows us to find the armature current as a function o f time for a given input volt
age. As above, if the input is a step, i.e., if
v^{t) = Ku{r),
B
s + --------^
------------ t t K.
S+
\
it follows that
R,,B + k~\
(1 4 .3 4 )
O nce again, application o f the final-value theorem to this expression leads to the value o f the
steady-state armature current:
Q
hi . ss ~
^Cl
B+ k
T he preceding analysis suggests the utility o f the Laplace transform as a tool for analyzing the
dynamic behavior o f electromechanical systems. In fact, system transfer functions o f the form of
equations 14.31 and 14.33 are often starting points for further analysis. Extensions of the above
analysis can be found in the homework exercises.
10. SUMMARY
This chapter has expanded the notion o f transfer function from its definition in Cliapter 13 into
a tool for modeling not only circuits, but other practical systems, such as a dc motor. The trans
fer function characterizes circuit or system behavior by the location o f its poles and zeros. For
example, if a transfer function has a pole on the imaginary axis or in the right half-plane, the asso
ciated circuit or system is said to be
unstable,
tial condition (as in the case o f a second-order pole on the y'oj-axis) that will excite the pole and
cause the response to grow without bound. Further, the ubiquitous presence o f noise will always
excite poles on the imaginary axis and cause the response to be unstable.
This chapter categorized various types o f responses: zero-state and zero-input responses, natural
and forced responses, transient and steady-state responses, etc. Recall that the zero-state response
is the response to an input assuming zero initial conditions, which is the inverse Laplace transform
o f the product o f the transfer function and the Laplace transform o f the input excitation. Recall
that the zero-input response is due only to initial conditions on the capacitors and/or inductors o f
the circuit. The complete response for linear circuits having constant parameter values is simply
the sum o f the zero-input and zero-state responses. This decomposition generalizes to the broad
class o f linear systems studied in advanced courses. Under reasonable conditions, other decom po
sitions are possible, such as a decomposition into the natural and forced responses or transient and
steady-state responses. Other important responses are the impulse and step responses.
For stable circuits, the frequency response provides important information about the circuit.
Recall that frequency response is a plot o f the magnitude and phase o f //(/w ) as to varies from 0
to cc. T he Bode plot is a plot o f gain in dB vs. frequency represented using a log scale. In this con
text it is relatively straightforward to construct an asymptotic approximation using straight line
segments, from which the actual plot is easily sketched by hand. T he information in such a plot
tells us how a circuit behaves when excited by sinusoids o f different frequencies.
(0
Lastly, this chapter has introduced the initial- and final-value theorems, which provide a quick
means of computing the initial and final values o f a time function from knowledge o f its Laplace
transform. Such theorems have wide application in control system analysis, as evidenced in our
analysis o f the dc servo motor.
satisfies [/(r)|
all
f,
finite height.
B reak point: the point at which two asymptotes o f a Bode plot intersect.
C om p lete response: the total response o f a circuit to a given set o f initial conditions and a given
input signal.
C o rn er (break) frequency: frequency at which two asymptotes o f a Bode plot intersect.
D ecade: a frequency band whose endpoint is a factor o f 10 larger than its beginning point.
D ecibel (d B ): a log-based measure o f gain equal to 2 0 logjQ|//(yw)|.
Final-value theorem : a theorem stating the following: suppose
F{s)
half o f the complex plane, with the possible exception o f a single-order pole at ^ = 0. Then
the limit
oij{t)
as /
Forced response: the portion o f the complete response that has the same exponent as the input
excitation, provided the input excitation has exponents different from those o f the zeroinput response.
Frequency response: measure o f circuit behavior to unit magnitude sinusoids, cos(w t), as O) varies
from 0 to
00
H{s)
at ^
Frequency scaling: lor a linear passive network, dividing all inductances and capacitances by a
factor, Ayr while keeping all resistance values unchanged.
Fundam ental period o f p e rio d icy (r): the smallest positive number
T such
that J{t)
=j{t + T).
Im pulse response: assuming all initial conditions are zero, if the circuit or system input is 6(r),
then the resulting^(r) is called the
impulse response.
s,
with the degree o f the numerator less than that o f the denominator . Then /(O*") is
oo.
L in ear circuit: circuit such that lor any rwo in p u ts/j(r) an d/^ (f), whose zero-state responses are
^ j(/) andjsCO. respectively, the response to the new input
A'2 >'2 (^)]. vvhere ATj and
Kj
[Kjf^{t)
N atural response: the portion o f the complete response that has the same exponents as the zeroinput response.
Octave: a frequenq^ band whose endpoint is twice as large as its beginning point.
Op amp open-loop gain: the gain of the op amp when no feedback paths to the input terminals
are present.
Periodicy(/); function satisfying the condition that there exists a positive constant, T, such that
/ f ) =fijt+ 7) for all f > 0.*
Phase response: the angle o f the frequency response as a ftinction o f (O.
Piecewise linear curve: an unbroken curve composed of straight-line segments.
Pole (simple) o f rational function: zero of order 1 in denominator polynomial.
Poles (finite) o f rational {unction: zeros o f denominator polynomial.
Ramp function, K^)j integral o f unit step function having the form Ktu{i) for some constant K.
Rational fim cdon: ratio o f two polynomials; a polynomial is a rational function.
Stable transfer function: a transfer function for which every bounded input signal yields a
bounded response signal; i.e., all poles are in open left half o f the complex plane.
Steady-state response: that part of the complete response which either is constant or satisfies the
definition of periodicity for t> 0 .
Step response: the response of the circuit to a step function, assuming all initial conditions are
zero.
Transient response: those terms o f the complete response that are neither constant nor periodic
for ^ > 0; i.e., the transient response does not satisfy the definition of a periodic function
for f > 0.
Zero-input response: the response of a circuit to a set of initial conditions with the input set to
zero.
Zero-state response: the response of a circuit to a specified input signal, given that the initial con
ditions are all set to zero.
Zeros o f rational function: values that make the numerator polynomial zero.
O '
O '
* This (nonstandard) definition has been adapted for one-sided Laplace transform problems.
O '
742
PROBLEMS
(b)
C om p ute
the
impulse
H{s).
and
step
(b)
C om p ute
response.
(c)
If the input is
itive number
impulse
response.
Ke~u{t).
the
H{s).
given in Figure P I 4 .3 .
HiO) = 8.
(a)
H{s)
(a)
itive number
If the input is
condition.
Ke~'^u{t).
jto
Identify the
(double pole)
jO)
-2
-1
)C4j
Figure P I4.1
-q>a
-2
H{s)
given in Figure P I 4 .2 .
(a)
(b)
(c)
-A]
Figure P I4.3
function
(a)
(d)
If the input is
itive number
H{s)
given in Figure P i 4 .4 .
If the gain
H{s)
is 4 .8 at ^ = 1, find
His).
(b)
C om p ute
the
impulse
response.
Ke~^u{t).
(c)
jO)
the response.
C H ECK .- Your answer to (b) should be the
-e o
-2
:-4j
Figu re P I 4 .2
If the input is
10.9
743
(c)
steady-state responses.
jw
ik
oj
(d)
------- 1> o
-2
s+a
O-j
(b)
(c)
Z (s)
(a)
(e)
Z (s)
(b)
Figure P I4.7
jco
>
(a)
(b)
e------ ------>o
-2
2
C-4j
and then
Again leave the con
stants in these forms as literals. Does the
output remain finite as f
00?
step response.
tions:
(b)
^44
(a)
(b)
x?
(b)
^u(t)
V is applied to
it)
t > 0.
transfer
functions,
and
determine
Figure P I4.1 2
7X . T O S
,,
(b)
Is -
I^
R^C2
H(s) =
s -+ ^s +\2
^ ( /?,Ci
1.5.V- -Q .5 .V -H .2 5
H2{ s ) =
(b)
r V 2 . v + 7.5.vH-6.5
R2C2 ^ R\C2 j
If Cj =
S+
= 1 mF, find
K\C\R2C2
and
Rj
so
H-^{s) =
5-V4.3I25.V + 5.3125
2s^ + 2.5s-
- 17.V + 220.5
/ - h 7 r V 5 5 . v - + 5 6 9 5 + 520
synthesis problem.
R.) = (4 k^, 0 .5 kI2)
A N SW ER ;
2
kLl)
is
called the
He
+
v.(t)
Figu re P I 4 .1 3
or (1 k il,
STABILITY PROBLEMS
14. For each of the circuits below, compute the
indicated transfer function and determine the
range
or
R =4a,
(b)
His)
Vi s )
16. (a)
c = 0 . 2 5 F.
is given
in
Figure
P i 4 . 16a.
(c)
H{s) =
Imag Axis
Ojw,
finis)
Real Axis
------o - >
-2
-jco,
(a)
Imag Axis
(a)
Figure P I4 .1 6
17. The circuit in Figure P i 4 .1 7 is to be stabi
lized in the BIBO sense.
(a)
If y? = 2
range o f
Figure P i4.14
(b)
\/
H(s) = - ^
V
*-111
complete
cuit is stable.
range of
in the circuit in
R for
= 0 .5
S,
range o f
If
R required
for stabilit}'.
746
r-s.
(1) Find the zero-state response, the
in tiie circuit o f
Figure P i 4 . 18; then find the
range
plete response.
il.
= -2
ble, assuming
(b)
complete
R=
S. Find the
is stable.
v-^^{t) = 20e~^^^u{t)
= 0.
Is the forced
H in t:
M A T L A B s
[r,p,k]
com m and
Use
=
R.
-C
R, <
transfer function
v;-(.v)
and the range o f a for which the transfer func
tion is stable.
Figure 1P14.20
2 1 . Find the complete response for r > 0 for the
circuit o f Figure PI 4 .2 0 for each o f the follow
ing circuit conditions. (If you properly utilize
the results o f Problem 20 and the principle o f
linearity, the answers to this problem can be
written down directly.)
(a)
= 6 V,
(b)
Figure P I4 .1 9
response
(a)
response
i + io
= 50
/?, = 2 0 0 Q ,
Suppose
v^Q-)
Vj^^{t)
= 8 V.
R^^^= R^
C = 2 . 5 niF.
(b)
e~^^u{t) V
4e~~^u{t) V
(b)
response.
when
V, and ^f;(0")
RC
v-^j{t) = 10(^)
RESPONSES AND
CLASSIFICATIONS
H{s) =
5 sin(5r)w(r) V
(a)
v-^j{t) =
I 6 V > . , W = 8 f - ' M r t V
= 10 sin(5/)//{/) V, and
74
Pure
resistive
network
vJt)
sC + +
Figure P I4 .22
2 3 . In the circuit shown in Figure P I 4 .2 3 ,
= 20 a
v jt ) = 25u{t)
V(^0~) = 10 V.
C = 0 .0 5 F,
V, and
.v-f-
/?2
5-t- 2
CR.(1/
(b )
/? 2
.V+ 1.6
C/?,
H (i) =
/?1
5 a
.v +
AC H------
v,(t)
given
in Figure
V,
'' 6
V(^0~) = -1
RC circuit
R^^C.
of an undriven first-order
where x =
(d)
(e)
is
Figure P I4.2 3
(c)
Find
relationship
between
Write
v^.
sient response.
(e)
Case 1
Case 2
V(;iO )
v jt)
Vs2it)
20
\Ou{t)
\5e^^u{t)
lOuit)
dOe^^uit)
R.
R, <
(a)
100
n,
A v(t) (V)
20
Figu re P I 4 .2 4
R^^ = R^ 11 R j.
(a)
Figures
transform method.
due to
(d)
the
-t
(c)
7AH
C-,
;-'q (0 ) = 8 0 mV and
v jt ) = OAu{t)
v^{0~)
i,(t)
= 0. The input is
V.
V ,.(t)
(a)
(b)
K.U'I
(c)
(d)
(e)
response.
(0
H{s) =
is)
KM
steady-state responses.
LC
.9 -I-
- v +
R2 C}
.v +
LC
R j)
16
r +
4.V + 20
when /?, = 1 Q ,
F, and
(b)
=4
LX C =
0 .1 2 5
input
(c)
/? 2
= 0 .5 H. ^
IO(r) V.
Figure 1M4.25
one is superior.
(d)
C= 1 /40 F,
+ i>2 (() =
Q,
= 3 0 LX
= 15
il,
(e)
and /^(0).
(0
response
vc{0~)
for
= 4 V.
5u{t),
= 2 V, and /^(O") = - 2 A.
zero-input
(a)
the
and Z. = 8 H. Suppose
(b)
Find
i,(t)
of
this
one could
as
the
zero-input
Figure P I4 .2 7
response.
(d)
(e)
(a)
H\ (j)
jco
if the
(double pole)
for
the
initial
-> o
-2
capacitor voltage y ^ 0 ).
(b)
<l>-2j
Figure P 14.29
source,
H(s) =
s
accounts for the initial inductor current.
(c)
shown in Figure P I 4 .3 0
H{2)
cos{4t)u{t)
=2
V.
jO)
>k.
0 4j
(i)
= 4 V and /^(0) = 5 A;
>a
Figure P I4.3 0
ANS\\T-:R: 0
/
2H
3 1
4S <
,2 S
4v.
2F
(a)
(a)
H{s) =
2s+ 6
Vi{s)
.V--1-4.9-h i6
= 2 cos{2t
(b)
Suppose
5
=3
output
(b)
( 6 + 4 )
has
cos(3f + 4 5 ) V.
that
Figure P i4.2 8
(c)
STEADY-.STATE CALCULATION
= 7? = 1 n
P i 4 .3
H{s) =
shown in Figure
ViU)
P i 4 .2 9 H{0) = 4.
Com pute
.V -I-------5 -i-
RC
LC
750
= 20 cos(4?) V.
Figure P I4.32
V ,( t )
Figure P I4.31
Kcos{2t+ 0) when
2 sin (2t)u{t).
(b)
s+4
=
the input
16^ + 96
(a)
Vi(s)
are present.
s^ + Ss + 64
= 2 sin(2^)(^) and
= 4 cos(4?
(c)
(b)
coincide.
has output
= 25 cos(3^ + 100)
V. Find the sinusoidal input
=K
cos(fiW + 0) that gives rise to this out
put.
(c)
= 15 n , C = 0.01
O.. Suppose
voltage + to -
the
Figure P I4.33
transfer function
_
'in
hnis)
?
1
+ 5 + -----
is shown
L
LC
Find the phase and magnitude o f the
= Kcos(cot + 6),
output cosine,
99
= Z/n(^) =
At what
jco
jlO
inputs:
= 20
(a)
j3
---- -5
cos(4r + 4 5 ) V
(b)
y (/) = 4 co s(4 0 ^ ) V
(c)
v,{t) = 4cos{2t)V
-2
X -
Figure P i 4.3 4
cos{2r)u{t)
35. A linear circuit witii a transfer function
2v+ 4
H{s) = ^ =
v;
fraction
expansion,
=4
determ ine
the
+ 5^ + 6
=8
t.
for
dominant?
(a)
- 3 .7 5 j + 5
tion
H(s) =
V.
V:.
5 + 4
(b )
- ^ 5 :
5- + 2.V + :)
2.5s- 3
H2(s) = -^
s" -t- 2 i 4- 5
. .
(c)
Ht,( s )= -^
s
7 r + .v -h 4
-l- 5 ~ + 9 i + 9
inputs:
(a)
= 4 cos(2^ ) V
(b)
U ^ ) = 4 c o s (4 0 V
(d )
H^(s) = -^
75^ + 5 + 7.75
(e)
5 ' V 3 5 + 1 1 . 2 5 5 + 18.5
^ --0 -5 5 + 5
5 -h0 .5 5 + 5 .7 3 2 1
=2^2
cos(2r +
A cos{2t
+ (j)), find
A and
(j).
(f)
5^ - 5 . 5 5 - +
nin = 8*[1 0 ];
tion
n = [2.5 -3]:
V:Js)
165^ + 4 4 5 + 1 2 8
85- + 365 +
80
145-12
^6(*v) =
5- + 5 .5 5 + 1 4 5 + 12
din = [1 0 4];
K,ur(^)
=5
. s ' + r + 7 .2 5 5 4 -1 8 .5
v^^is)
14,s-2-23.V + 20
d = [l 2 51;
nn = conv(n,nin);
dd = conv(d,din);
'5 2
[r,p,k] = residue(nn,dd)
(a)
(b)
r=
V^{s)IV^{sY
-4.0000e+ 00 - 7.0000e+00i
-4.0000e+ 00 + 7.0000e+00i
4.0000e+ 00 + 4.0000e+00i
4.0000e+ 00 - 4.0000e+00i
P=
-l.OOOOe+00 + 2.0000e+00i
-1 .0 0 0 0 e + 0 0 -2 .0 0 0 0 e + 0 0 i
-7 .7 7 l6 e -l6 + 2.0000e+00i
(a)
- 7 .7 7 l6 e -l6 - 2.0000e+00i
k=D
(b)
(c)
>>K = 2*abs(r(4))
K = 1.13l4e+01
Phi = atan2(imag(r(4)),real(r(4)))* 180/pi
Phi = -4.5000e+01
is zero.
=8
for
(b )
H^(s ) = - y
H2(s) = ^
80
+4^ + 3
80
43. The circuit shown in Figure P I4.43 has zero
initial conditions, y^O") = 0, and //^(0) = 0).
80
(a)
(c )
s + 2 j + 160l5
(b)
(d)
If
5 ^ + 8 5 + 1 616
5^ + 155-^ + 8 3 5 ^ + 1 9 9 5 + 170
nomenon
has application
in the
1) V, find
2 0 0
kV.
(c)
= [6(^)
+ 6 (r - 1) + 6 (r - 2) + - ] V, find the
output
Figure P I4.4 3
4 4 . Suppose the voltage source in Problem 4 3 is
changed to
For
t >0,
ti{t)
10 sin
V.
4 5
A. Determine
Vgn,{t). W hat
Figure P I4.4 7
is the largest
FREQUENCY RESPONSE
4 8 . Draw to scale the pole-zero plot for
H{s)
Figure P i4.45
46. At r = 0, the energy stored in the
LC elements
cos{lt)ti{t)
4 9
for large t.
((.v+ l)2 + 1 6 ) ( . r - l )
His) =
H(joj)
Figure P i4 .4 6
5 0
is assumed to be ideal.
(a)
(b)
H{s).
'^ 4
jco
jw
J
X
<>8j
'l6j
-> a
---- 1-------0.2
*-8 j
-2
-1
8j
II1fV ^W
-Bj
-16j
Figure P I4.51
Figure P I4.5 0
To make the gain o f the transfer func
( .9 -
( . V - /?,)
(b)
1I
X
(a)
1
I
H{s)
H{s)
approximately.
Figure P i4.5 2
1
->
CO
value o f A'in
ANSWI-H
(.v~ + l)~
(.v + O . l r - f l
(5 - z, )
response
(m agnitude
and
Frequency-rad/s
Figure P i4.5 3
H{s).
7SS
H{s)>
( 1) K
(2) K
(3) K
(4) K
(5) K
(6 )
(^ + 0 .5 + ;5 ) ( 5 + 0 . 5 - y 5 )
filter
+ 1.414
{s + j l ) { s - j 2 )
j(5 + 0 .5 + y5)(.? + 0 . 5 -
j5)
____
the
{ s + j 2 ) { s - j 2 ) ___________________
(s+j2)(s-j2)
(7) K
5
( 5 + 0 .8 +
7 6
X5 +
7 6
steady-state
Butterw orth
phase
and
filter when
the
(5 + 0 .3 + y 4 )(5 + 0 . 3 - y 4 ) ( 5 + 0 .8 + y 6 ) ( 5 + 0 . 8 - y 6 )
5 (5
at 1000 Hz.
s { s + j 2 ) ( s - j 2 ) __________________
___________________
[X )in t
(s+j2){s-j2)
__________________
+ 1
2000:1/
\ 2 0 0 0 Ji/
cos(20007rf),
and
(iii)
cos(40007tr).
)
Do your answers to part (c)
.8 -
7 6
Remark: T he
of
A N SW E R : (5)
gain,
or the magnitude
ats= j5
s = yiOOO.
56. (a)
at
the
magnitude
and
phase
inductor,
transfer function
(in H)?
H {S ) =
1Q
- + 1 .4 1 4 5 + 1
4H
1F
(b)
Figure P I4.54
55.(a)
the
magnitude
and
2000;t
H(s) =
phase
+ 1.414
second-order
normalized
low-pass
(c)
(b)
" + 1 .4 1 4 5 + 1
+ 1
\ 2 0 0 0 jt/
OOOJ1 /
Referring to part (b), find the steadystate phase and magnitude o f the out
put o f the Butterworth filter when the
input
is
(i)
cos(5007ir),
(ii)
~S6
\H{j(S))\ vs.
transfer function
~T
0 .7 0 7 9 .S -
4 - 0 .6 4 4 9 5 +
(b )
0 .5 0 1 2 .r
H{s) =
Bode plots.
< f<
1 Hz.
(b)
logjolcol.
H{s) =
5 8 .(a)
the
m agnitude
CO
and
0 .7 0 7 9
\200 0 7 1 /
phase
low-pass
(c)
(b)
0 .5 0 1 2
Referring to part (b), find the steadystate phase and magnitude o f the out
---------------
H{s) =
200071/
normalized
\ 2 0 0 0 Jt/
input
+ 0.6449.s + 0 .7 0 7 9
is
(i)
cos(5007rr),
(ii)
responses 0 <
CO
s +2i +25' + l
0 .5 0 1 2
H{s) =
(a)
-h 0 .6 4 4 9
+ 0 .7 0 7 9
H{s)
. Verify
2000 Jt
200071/
1 Hz.
input
(d)
is
(i)
cos(5007tf),
(ii)
< f<
\NSWH
- [0:0, 01:
= 0 .5 ;
sense? Why?
Remark; The
gain,
or the magnitude
= freqsi n,d,\v);
p lot(w /(2 pi). abs(h))
If
is chosen so that
Kjr= (??);rso
S ]59.(a) Plot, using M ATLAB or the equiva
lent,
the
magnitude
and
phase
R=
10
find
th a tC = 1 5 .9 1 5 /iE W h at
U.5
H (s ) =
(a)
\3
+2
+2
(ss)
^>/
+1
s-
s+
(d)
(b)
1
RC
1
RC
rad/sec.
From
the
magnitude
'.w Q
Figure P I4.60
61. The circuit of Figure P I 4.61 has transfer
W
fiinction
H{s) =
R\Ci
1
^/?lCi
1
R i^ i
s+
^ iQ /
+
V .Jt)
'" 6
Figure P I4.61
62. The circuit of Figure P 14.62 is called an all
Figure P i4.63
Si64. A certain band-pass circuit has transfer
flmction
758
+ 1 3 9 .9 0 f - ^ )
VlOO^
^VlOO/
H{s) =
5
5^ + 2 .7 9 9 6 f )
UOO/
+ 0 .9 9 8 9 f ;^ ) ^
VlOO/^
4
+ 1 3 .0 5 l f )
3
+ 1 7 .2 7 lf )
\m J
VlOO/
+ 1 3 .0 5 1
+ 2.7991 (
) + 1
UOO/
100^
H{s) =
Vo(s)
+
RC
5+
In l
R^C
o
V.
v:-in
___ K
RxR'i C\C2
\-K\
1
+
/?2 Q
s+
y4sin(cOo^ + Cp).
^2^2/
fer function.
(b)
(c)
denominator) are ( - 2 S y R C .
L e t / ? = 5 / J t 1 .5 9 1 5 l a n d C = 0 .0 1
CO< 5 rad/sec.
---------- r ^ ( -
"
---- 1--1
R
r\
r> 0.5R
v^(
1 2C <
-----L---------- n
r\
(a)
(b)
pole-zero plot.
(c)
759
(b)
(d)
= b(t)
V.
= 5b(t-
cos{lt)n{t) V.
For large t, show that Vg,{t) ~ Kt
cos(2t + (j)) V for appropriate K and (J).
Now suppose
00
2) V.
h{t),
W(.v) =
vJt)
o f a circuit is given by
s+2
(,s + l)(.v + 3)
Find
when
(a)
8)
Figure P i4 .6 7
time by some
( 5 .v - l ) ( 4 .v - 5 ) ( 6 .v - 2 )
5 (2 5 -h l)(3 .s -F 2 )(5 i + 4 )
the Laplace
transform o f a s i g n a l / r -
2)u{t-
2) is
^_ 2 , (l(Xs - l ) ( 8 . v - 5 ) ( 1 2 . v - 2 )
^
.v(5.v+l)(3.s'-h2)(i- + 4 )
7 2 . The output,
o f a particular circuit is
4s- - 4s s-
+ 2i' + 2
Figure P i4.6 8
A N SW ER ; - 2 , 1
v^{t)
t.
for
C H E C K : for large
t,
error is 3
= r ( f - 1) V is
73. Given the following fimctions
F-{s),
find
v-^U) = u{t)
V.
2 l.v + 5
O.I.v*- +4.V + 3
F,{s) =
(b)
s{s^
+ .V " + 7 . V
H(s)=
(c)
F,is) =
.y(.v +
+ 1
+ 6 )
s~ +4.9 + 3
5s^ + 5s^ + 4.V +
iO-
4)
100
.V
+
V16
\ 400
1200
+ 1
o f various
and
for each
H(s)=
10
+ 1
V16
1200
+ 1
(a)
(i 4 .v - i r
100
s~(ls+2)
+ 1
+ 1
V400
(b )
(c )
(d )
(2 5 + 1 ) -
H(s)=
25-
25+1
V25 + 1
10-
+ 1
+ 1
\ 1200
+ I
4 5 -1 \ -'
16
100
+ 1
V 400
V 25+ I/
I'hen verify your asymptotic plots by using
o f various
below. Use only the initial-and fmal-value theorems to find /f;;(0') and
i(^^)
10
(a)
5(5
+ 1
+ 2)
H(s}=
(b )
5 (25 + 2 5 )
(c )
20
(^ -2 )
5(25 + 25)
10
\ 20
200
2400
+ 1
+1
+1
800
(c)
= 100 mH.
"^61
fixed amplitude?
(d)
amplitude?
Problem 80 with
= 5u{t) V, com
Change
(b)
to 50 Q, and recompute
time.
O '
C H A P T E R
Time Domain Circuit Responsi
Computations: The Convolution Method
AVERAGING BY A FINITE TIME INTEGRATOR CIRCUIT
Sometimes one must compute the average o f some quantity, sucii as the average vakie o f the light
intensity on a solar cell over the last fifteen minutes, or the average value o f the temperature in a
room over the last hour, or the average value o f a voltage over the last 50 milliseconds. W hen such
an average is updated continuously, it is termed a running average. The idea is that the readout o f
the device that averages these quantities always produces an updated value valid over a specified
prior time interval. If a voltage represents the value o f the quantity to be averaged, then one can
build a circuit whose output voltage is the required average. This is done by observing that an aver
age value o f a continuous time variable is simply the integral o f the variable over the proper time
interval, divided by the length o f the time interval. A device that integrates a variable over the last,
say,
T seconds
is called a finite time integrator. As an application o f the ideas o f this chapter, we will
look at a finite time integrator circuit and how it can be used to compute the average value o f a
quantit)^ The convolution concept directly leads to the required transfer function o f such a circuit.
CHAPTER OBJECTIVES
1.
2.
Using the notion of convolution, develop a technique o f time domain circuit response
computation that is the counterpart o f the transfer function approach in the frequency
domain, presented in Chapter 14. In particular, we seek to show that the convolution o f
an input excitation with the impulse response o f a circuit or system produces the zerostate circuit or system response.
3.
Develop objective 2 from r\vo angles: first, from a strict time domain viewpoint, and sec
ond, as a formal theorem relating convolution to the transfer function approach.
4.
Develop graphical and analytical methods in particular, an algebra for evaluating the
convolution o f two signals.
5.
Illustrate various properties o f convolution that are pertinent to block diagrams o f series,
parallel, and cascade interconnections o f circuits or systems modeled by transfer func
tions.
^6'i
SECTION HEADINGS
1.
Introduction
2.
3.
4.
5.
6.
7.
8.
9. Summary
10. Terms and Concepts
IL
Problems
1. INTRODUCTION
At the beginning o f Chapter 12, we claimed that circuit response computations could take place
in either the time domain or the i-dom ain. Yet, except for the solution o f some very elementary
differential equations, circuit response computations have relied almost exclusively on the Laplace
transform technique. This chapter develops and explores the time domain counterpart o f the
Laplace transform method by introducing the notion o f the convolution o f two signals to produce
a third signal. We then show that the time domain convolution o f an input excitation with the
circuits impulse response yields the zero-state circuit response. Pictorially, the idea is expressed in
Figure 15.1, which is an update ot Figure 12.3.
f(t)
Circuit Impulse
>
Response h(t)
Input Signal
y(t) = h^f
Output Signal
Transform of
Circuit Transfer
Transformed-
Function H(s)
Output Signal
Y(s) = H(s)F(s)
riCnJRF, 15.1 Diagram showing the symmetric relationship of time and hequenc)'^ domain analyses.
The upper part of the diagram asserts that the convolution, denoted
with the circuit impulse response
h{t)
h * f, of the
input signal/(f)
Justification of the diagram o f Figure 15.1 begins with the formal introduaion o f the notion of
convolution. Mathematically, the convolution o f two fiinctions h{t) andy(f), denoted hy h * f or
j^ ^ h {t-x )f{T )d T
(1 5 .1 )
The second step in verifying the diagram is proving that the convolution of the input signal with
the impulse response produces the zero-state circuit response. There are two approaches. One is
to work stricdy in the time domain and construct the actual zero-state response from the impulse
response and an arbitrary input excitation. The second approach is to prove that
is
the convolution of the signal y(/) with the impulse response h{t). Because o f the Laplace transform
development o f the last three chapters, this direction seems the most painless and will be taken up
V^
in section 3, after an introduction to the basic ideas o f convolution in section 2. Section 4 will
look at the graphical method of convolution, which helps in visually grasping the definition.
Section 5 will describe a convolution algebra, which yields a harvest o f shortcuts for evaluating the
convolution o f certain types o f signals. Section 6 looks at the computation o f circuit responses
using the convolution method, and here the averaging circuit is developed. Section 7 overviews
various properties o f convolution. The respective properties lend themselves to different structures
(e.g., parallel or cascade) for designing interconnected circuits. Section 8 describes the rudiments
o f the construaion o f the zero-state response o f a circuit working stricdy in the time domain
toward the development of a convolution integral.
Before closing this introduction, we should consider the question, why is convolution important^
One reason is that it allows engineers to directly model the input-output behavior of circuits and
general systems in the time domain, just as transfer functions model circuit behavior in the sdomain.
A second reason is that circuit diagrams are sometimes unavailable or even get lost. How would
one generate a circuit model for analysis? One way is to display the impulse response on a CRT
and approximate
O '
766
FIG UIIE
1 5 .2
A third important reason for studying convolution in the context o f circuits relates to a deficiency
o f the one-sided Laplace transform: function segments that are nonzero for f < 0 are ignored by the
one-sided Laplace transform technique. Hence, time domain convolution offers a direct means of
computing the circuit response when signals are nonzero over the entire time interval, -<
that the convolution h * /e q u a ls the co n v o lu tio n /* h, i.e., the operation o f convolution is com
mutative. In fact, this is the case. To emphasize this property, we restate equation 15.1 in its more
general form:
CONVOLUTION
T he convolution o f two sig n a ls/r) and
h{t)
the formula
v (/ ) = J
J A t - r)f{T)dT =
T)dx
(1 5 .2 )
which is well defined, provided that the integral exists. This formula emphasizes the proper
ty that convolution is a commutative operation, i.e.,
h * f = f * h.
h * { r g ) = { h- f i * g
(1 5 .3 )
and distributive,
(1 5 .4 )
Exercises.
'=t
767
- t .
Another very useful property o f convolution is the so-called tim e shift property: if
h{t) * j[t)
then
h{t-T^)^j{t-T^)=g^t-T,-T,)
(1 5 .5 )
Some simple examples serve to demonstrate the actual calculation process. These examples will
utilize the sifting p rop erty o f the delta function: if h{t) is continuous at / =
h{T
T,
then
-T)d{T)dT
6{t) .
lu t io n
By the definition o f equation 15.2 and the sifting property o f the delta function,
y{t) = f * 5 =
-T)d(r)dT
= fit
= f it )
(1 5 .6 )
Example 15.1 makes the point that 6 {t) acts like an identity' for convolution, i.e., it always returns
the function with which it is convolved. T he next example indicates that the convolution o f an
arbitrar)' continuous function y{/) with
ad(t - T)
as
a d {t-
7).
lu t io n
By the definition o f equation 15.2 and the sifting property o f the delta function.
t)^ ^ =
co m p u tin g /j
(/;)/; (r) =
+ T^)
d it-
7 , ) , / , ( / ) = u i t ) - u { t - 2 ) , a n d / 3(r) = d i t
3)
*fi*
afit - T)
(1 5 .7 )
:"6 8
Ler us a|3ply the results o f the above two examples to a simple staircase function.
EXA M PLE
1 5
. 3 . Com pute the convolution o f the functiony(/) shown in Figure 15.3a with h { t )
^ ld {t) -2 d {t -
1).
f(t)
iL
2 --
1-
H -------
(a)
FIG URE 15.3 (a) Functiony(f) for Example 15.3. (h ) 7 = / h.
So
lu t io n
By the definition of equation 15.1 and the sifting propert)' of the delta function,
v(/) = / * / / = 2 j
=
- T) d( T - l)r/r
- t ) 6 ( t ) cI t - 2 j
( 15. 8)
Graphically combining the result given in equation 15.8 yields the relation in Figure 15.3b.
Exercise. C om pute
25{t-2).
h{t)
26{t) -
AN SW 'FR: l f { t ) - l / { r - 2) =
EXA M PLE
1 5
S o lu tio n
ii(i)*u {t)= C
J -o c
u(t-T )u(T)dT =
Jo
n (t-T )d r
(1 5 .9 )
"6 9
since
T^O
|0
T< 0
We now break equation 1 5.9 into two regions: (i) t > r, in which case
tion 15.9 zero; and (ii) 0 s t s /, in which case
u{t- t)
= 1. W hen 0
to
In other words,
u{t)
u{t) = r(t),
(1 5 .1 0 )
/()
W e now apply the time shift propert}' o f convolution to the result o f Example 15.4.
E X A M P L E 1 5 .5 . Find
y = h /w h e n h{t) = ii{t)
a n d /r) =
ii{t + 1)
u{t-
15.4a and b.
u(t+ D - u(t-l)
>k
>^ u(t)
--- 1
-1
1
1
1
2
> t
^ 1
1
(b)
(a)
y(t)
h{t)
u{t +
1) - (r - 1).
So
lu t io n
y{t) = u{tY[u{t^
1) -
ii{t-
1)] =
u{tYu{t^
1) -
u{tYu{t-
1)
(1 5 .1 1 )
Now using the time shift property o f convolution and equation 1 5 .1 0 , we conclude that
y{t) = u{t) * (r
+ 1) -
1) = K / + 1) -
r(t-
1)
which is plotted in Figure 15.4c. We note that this result was achieved without any direct inte
gration.
h */^of the
two waveforms
/j(t) and/(t)
in Figure
15.5.
h(t)
(a)
(b)
y(t)
> X
{c)y = h * f
S o Lu n o N
We first note t h a t / r ) =
tive law o f convolution.
ii{t)
+ (r - 1 ) -
lu{t -
2) and
h{t)
u{t) - ii{t -
y{t)
[u{t)
= [(^) + ( / - 1) -
u{t-
1) -
l i i i t - 2)Yti{t)
ti{t-
1) -
1)]
2)]*//(/- I)
lu {t-
Now using the time shift property o f convolution and equation 1 5 .1 0 , with a further application
o f the distributive law, wc conclude that
2 r ( r - 3)] =
r{t) - M t - 2 )
2 r it-
3)
which is plotted in Figure 15.5c. We again note that this result was achieved without any direct
integration.
u{t-
h * /o f
the signals
h{t)
a n d /(r ) =
e ^[u{t)
h(t) >k
0
(a)
h{t)
lu t io n
Step 1.
yields
\{t) = h * f =
-T)f{T)dT =
i/(t
-T)]dT
Since ;/(x) - z/(x - 7 ) is nonzero only for 0 ^ x :s T, the lower and upper limits o f integration
T,
bccome 0 and
respectively:
y{t) = h * f = ^ ^
Step 2.
e~^u{T-t)clT
(1 5 .1 2 )
T, and
(iii)
^ t.
Step 3.
Evaluate the convolution integral, equation 15.12, over the given regions.
Case I: t < 0.
Z and
Here, r < 0 implies that x - f > 0 over 0 s x s 7". Hence, u(t - /) = 1 over 0 s x ^
j.
y( 0 =
Case 2: 0 ^ t < T.
- t)
i >t.
Hence, in
the region 0 ^ f < 7 , it must also be true that 0 < r s x < 7 for the integral o f equation 1 5 .1 2 to
be nonzero. Thus, the lower limit o f integration with respect to the variable x becomes
v (/) = J ^
Plot y{t).
e '^u{T-t)dT =
_^ n 7 -
=e
- e
f.
= 0 in this region.
Combining the results o f step 3 with 7 = 1 implies that y(r) has the graph
Exercise.
A N SW E R :
Ka
'//(- /) f
e ^hi{t)
with
h{t)
Ku{-t).
Ka
As claimed in the introduction, circuit analysis in the time domain by convolution and circuit
analysis in the frequenc)' domain by Laplace transformation are equivalent in terms o f zero-state
response calculations. T he purpose o f this section is to rigorize the equivalence between the time
and frequenc)' domain analysis methods by formally show'ing that
presents a time domain rendition.
Section 8
CONVOLUTION THEOREM
h{t)
Supposey(/) = 0 and
i.e., convolurion in the rime domain is equivalent to multiplication o f transforms in the fre
quency domain; or equivalently,
hit) ^j{t)
r^[H{s)F{s)]
(1 5 .1 3 b )
h[t) * J{t).
Specifically,
f /;(/)*/(Ol =
(1 5 .1 4 )
A couple of points are in order: (i) the inner integral, surrounded by parentheses, represents the
convolution
h{t) * J{t);
h{t) = 0
u{t - r)
is added as an aid to
for r < 0.
Step 2 . The goal at this point is to manipulate the integral o f equation 1 5.14 into a form that can
be identified as the product ol the Laplace transforms o f two functions, i.e., as
H{s)F{s). The
only
possible strateg)' is to interchange the order o f integration and group appropriate terms. Note that
the Re[^] must be chosen sufficiently large to ensure the existence o f the Laplace transforms of
both
h{r)
Under certain conditions that are t\'pically met, it is possible to interchange the
H{s)
and
F{s).
Interchanging the
order and regrouping the f-dependent terms inside a single parenthetical expression produces
( / ? { / ) * / ( / ) ] = J ^ _ / ( t ) ^J^_/;(/ - T) / / ( /
(15. 15)
Step 3 . Observe that the interior integral, surrounded by parentheses, in equation 15.15 is sim
ply the Laplace transform o f a time-shifted
e~^H{s)
i.e.,
[M{t -
r)] =
Substituting
= H ( . v ) f ( ,v )
This theorem asserts the equivalence o f convolution o f one-sided signals with multiplication of
their transforms in the /-dom ain. For our purposes,
h{t)
o f our circuit andy(r) the role of the input excitation. Accordingly, the convolution o f the impulse
response o f a circuit or system with an input signal, a time domain com putation, equals the
inverse transform o f the product o f the respective Laplace transforms. In other words, the diagram
o f Figure 15.1 is correct, as claimed under the conditions o f the theorem.
Exercises.
- 1)
b{t)
n{t) - u{t -
1). Find
h it -
1) - 2 r ( / - 2)
e~^u{t).
and input
h{t) * fit).
A N SW ER :
T he conditions o f the theorem are somewhat restrictive in terms o f computing circuit responses
strictly in the time domain. Specifically, it is the one-sided Laplace transform that does not rec
ognize function segments over the negative real axis hence the condition on the input excitation
t h a t / r ) = 0 for /< 0. This restriction does not lend itself to the computation o f initial conditions
and circuit responses due to input signals extending back in time to / =
volution o f an input excitation with a circuits impulse response presupposes no such restriction.
However, justification o f the computation o f zero-state responses due to input excitations extend
ing back in time to / = -oo cannot be based on the convolution theorem o f the one-sided Laplace
transform. A justification o f the time domain convolution approach to computing circuit response
is reserved for the last section o f this chapter, due to its complexit)^ Nevertheless we will use the
result as necessary, such as in the next exercise.
Exercise, T he
H{s)^
.9+ 0.2
+ 1) - 1()[1 -
')],,(/
v{t),
v(t)
J <
,1
-1
FIG U RE 15.7
simple, squarish waveforms. T he technique often leads to a more penetrating insight into the con
volution integral.
There are four key ideas in the graphical procedure: flip, shift, multiply, area. The following is a
detailed description o f the procedure.
To compute
t = T,
given by the convolution integral o f equation 15.1 or 15.2, for a specific value o f
1.
Plot
h{x)
2.
3.
4.
5.
t)
t curve.
v s.
is7(T ).
We make the following remarks with regard to the above steps:
(1)
The roles o f /;( t ) andy(T) may be interchanged because o f the commutative propert)' of
convolution, equation 15.2. In other words, we may flip and shift /?(t) instead oF /( t).
Usually we flip and shift the simpler waveform.
(2)
(3)
As r i s varied from -oo to +oo, a complete plot of^(^) is obtained. In shifting and find
ing the area, we often have to divide T in to separate inter\'als, becausc each interval may
require a different formula to com pute the area beneath the
t, h{t-T)
vs. x curve.
t moves
from
to cc,
x) moves along the x-axis from x = -oo to x = oo; a simple illustration is h(t - x) =
u{t -
x), w'hich
h{t -
h{T)j{T- r)
is sketched in Figure 15.8. Part (a) o f the figure shows (x), part (b) depicts
u{t-T),
ftuKtion
h{t-x)
u{t-
t increases.
u{-x),
u(-l) ><
>k u(x)
u(t-x)
>k
increasing t
----w -
---------------- ^ ^
(b)
(a)
(C)
increasing t
u(-T-l) A
A u (x -l)
t- 1
-1
(d)
(e)
h{t-x)
h{t-x) = u{t-x - 1),
(f)
plots of (d)
tt{x -
t)
and JU) =
2u{-t)
h{t) = u{-
f(t)
"
k.
f(T)
I
h(t-T) = u (t-1 )
h(t)
increasing t
y
(a)
(b )
lu t io n
T he graphical solution to this convolution depends on a partitioning o f the time line into special
segments over which the graphical convolution is easily done. There are two regions to consider:
(1)
-0 0
t > 0.
Step 1. Consider the region r > 0. Figure 15.9b shows that h{t - t ) J { t ) = 0 for all r > 0 and all T .
Thus the convolution integral is zero and
= 0, r > 0.
Step 2. Consider the region -oo < ^ ^ 0. From Figure 15.9b , the product h{t
=1 x 2 = 2 for
r ^ T s 0 and zero elsewhere. Consequently, the area under the nonzero portion o f the product
functions
h{t -
is
y (0 =
hit - T )f{ T ) d r =
dr = -It
Step 3. Combine the foregoing calculations into a plotted waveform. Figure 15.10 shows the func
tion j/(r) resulting from the convolution.
y(t)
Exercise.
ioj{t)
M ndj'(r) at r = 0 .5 ,
Another, more complex example will end our illustration o f the graphical convolution technique.
j{t)
h{t)
as sketched in Figure 15 .1 1 .
f(t)
T
(b)
/ \
/ \
\
/I
\
/
/
/
-1
Time (seconds)
(c)
h{t-T)
1.
2.
3.
Com pute the area under the product curve for each region determined in step 2.
Step 1.
Drauf h{t -
h{t -
a n d / x ) on the x-axis.
t.
x) audJ{T) on the T-axis for f < 0 and compute the area o f their product. Figure
15.12 shows h{t - x) andy(x) on the X-axis. From the figure, it is clear that h{t 0; hence,
x )fj)
= 0 for ? <
^ r <
-- 2T
Area
2T-t
t-2T
T.
T he shaded area o f the figure is the difference between the area o f the large triangle, defined as
Area A = 0 .5 (2 7 )^
and the area o f the smaller triangle to the left o f the vertical axis, defined as
Area B = 0 . 5 ( 2 7 -
Hence,
j/(r) = Area = Area A - Area B = 2 Tr - 0 . 5 r
for 0 :s r <
T.
Alternatively, one may use the formula for the area o f a trapezoid, i.e., the average
I T - t)t
= Area =
ITt-
0.5/^.
Step 3 .
Area
t-2T
> I
T^ t <1T.
In this figure, the shaded area, which determines j'(f), is again the difference o f two triangular
areas, specifically.
y{t) = 0 . 5 [ r - { t - I T ) ] for
- 0 . 5 [ 2 r - r]2 = 2 .5 7 ^ -
Tt
T^t<2T.
2T ^ t < 3T.
yields
- 2T
Area
> T
t-2T
F IG U R E 1 5 . 1 5 Graph o f
h{t-T)j{T)
on the x-axis
ioxlT^ t < 5 T
780
Step 5.
h{t-T )/(t)
= 0 , in which
caseyit)
= 0 for r
>37:
In sum,
0,
v(0 =
/< 0
2Tt-0.5r,
0^t<T
2.5T^-Tt,
T^[<2T.
T=
2.
Exercises. 1. Repeat the calculations o f the preceding example, except flip and shift J{t) instead of
h{t). Here, one looks zx. j{t - r) sliding through h{x). The calculations should be easier and the
result the same.
2, Find the output o f a linear and relaxed circuit with in p u t //) and impulse response
h{t)
(shown
MGURF. 15.16
3. T he impulse response o f a particular circuit is approximately measured on a scope as illustrat
ed in Figure 1 5 .1 7 . If the input to the circuit \s J{t) = u{t - 1 ) -u{t
response, y{t), at / = 3 sec and r = 3 sec.
/VNSW ER S: 2, 1
81
4. For
h{t)
h{t) j{t).
A N SW ER : 4
h(t)
-1
FIG U RE 1 5 .1 8
Sometimes the foregoing graphical techniques prove difficult to execute. Nevertheless, an under
standing o f them offers fundamental insight into the meaning o f the convolution integral. A use
ful set o f techniques for quickly evaluating convolution integrals arises from the properties o f a
convolution algebra, discussed in the next section.
5. CONVOLUTION ALGEBRA
A set o f functions, together with operations called addition and multiplication, is called an
bra,
alge
assuming certain conditions are satisfied. The set o f all functions that can be convolved with
each other also constitutes an algebra with respect to the operations o f addition (+), and convolu
tion (*). This set, together with the two operations, is called a convolution algebra. In this con
text, operations such as differentiation and integration are inverses o f each other. For example,
integrating a function and then differentiating the result returns the original function. W ithin the
convolution algebra, the co n v o lu tio n /*
the derivative o f
g is equivalent
associativity o f + is clear:/ + g = ^ + / a n d / +
ciativity o f * is equally clear: f *
g =g * / a n d
h.
In addition to obeying these laws, algebras o f functions must contain identity elements. For
+, the zero function ser\'es as the identity. For convolution, the delta function plays this role. The
delta function is an identity element because o f the sifting property as set forth in Example 15.2
and equation 15.7.
For our purposes, the interesting aspects o f a convolution algebra o f functions rests with the inter
relationship o f convolution, differentiation, and integration. To map out this kinship, we use the
following notation
/
/(t)A
(1 5 .1 6 )
and
/,< !)(,) = ^
(1 5 .1 7 )
dt
It can be shown that
/ * / 7 = / ( ) * // - I )
if /(-o o ) = 0 and
t=
(1 5 .1 8 )
t.
h{t)
Similarly,
/* /,= /(- )
(1 5 .1 9 )
u{t)
u{t
=/^'^
y =
f o r //) =
- 1 ) and
h{t) = uU) -
lu t io n
= 5(r) + 6 ( / - I)
-2 6 { t -
2) and
r{t) -
, where/^*^(/)
/H3
(b)
FIG U R E 1 5 .1 9 (a) The derivative ofy(r) =
(b) The integral of h{t) =
Since f^\t) =
+ 6{t - \) -26
u{t) - u {t -
n{t)
+ (r-l) -
2u{t -
2),
1) -
yU) =f^
W ith the picture o f g i v e n
1)
(1 5 .2 0 )
interpreted as a graphical sum o f (shifted) versions o f //~ '^ (/), as illustrated in Figure 15.20.
Equations 1 5 .1 8 and 1 5 .1 9 , as illustrated in the preceding example, are special cases o f more gen
eral formulas. Specifically, let
y =f * h- Then
yij-^ k) -J-iJ) - fj{k)
(15.21)
Chapter 15 Tim e Domain Circuit Response Com putations: riie Convolution Method
where j and
k are
-k
EXA M PLE
> 0.
t] if j < 0,
and
1 5
^t)
= :r
cos{m)u{t)
withy(f) =
r{t) - r{t-
2 ), as sketched
in Figure 15 .2 1 .
S o lution
Some preliminary thought suggests that evaluation o f the convolution integral might proceed
more smoothly via equation 1 5 .2 1 ; i.e.,
where
=
TC~
J0
cos(jTq)dq = JTsmiJTt)u(t)
and
f _n\n(jtq)dq = {\ - co^iJtt)]u{[)
J0
Differentiatingy(r) twice leads to
Hence,
f*g =f^-'> *
2)
2))\ii{t -
2)
t)
Tim e (seconds)
FIG URE 15.22 Graph of the convolution
Convolution algebra and graphical convolution lend themselves to a second application o f the
convolution technique: the computation o f circuit responses from a staircase approximation to a
circuit impulse response. If a circuit schematic is lost, such an approximation could result from a
C R T readout o f the circuit impulse response measured in a laboratory. The following example
illustrates this application.
EXA M PLE 15.12. Suppose the schematic diagram o f a very old linear circuit is lost. However,
the circuit impulse response is measured in the laboratory and approximated by the staircase wave
form o f Figure 15.23. If the input to the circuit is
at r = 0 and at r = 0.5 sec.
h t)
1.0
0 .8
--
0 .6
0.4 -
0.2
0.2
0.4
0.6
fI h H I h
0.8
1.0
1.2
H I I I I h -l1.4
1.6
1.8
2.0
'
S olution
The objective is to convolve the input with the impulse response. The technique o f convolution
algebra where one differentiates the impulse response to obtain a sum o f shifted impulse functions
and integrates the input to obtain a ramp function seems to be the most straightforward approach
for this calculation.
First, observe that
h{t) = 0.1
- 0.1) +
(15.22)
O f course, it is possible to obtain the solution just as easily in this case using the graphical method.
Simply flip the
curve and slide it through the /?(f) curve. The area under the product curve
is simply the sum o f the rectangular areas, which are easy to compute.
Exercises. 1. Compute
1 and
1.5 sec.
in equation 15.22. Hint: Use the result o f Example 15.4 or equation 15.10, and the time shift the
orem o f convolution, equation 15.5.
ANSWF.R: Replace the step functions in equation 15.22 with ramp functions.
/87
EXA M PLE 15.13. Consider the RC circuit o f Figure 15.24, whose impulse response is h{t)
find
10
IF
lu t io n
Case 1: a = 0. Here,
^'oia (7) = r
J 00
= f
J-00
h{T -
t)v ,
{T)dT = f
J oc
/Kt)v,(T - T)dr
Jo
e~'^i((T)dT = 1
- T)dx = e~^
^ W (^ ) = Jq
This also makes sense physically because at 7 = 0, the initial capacitor voltage is 1 V and the capac
itor discharges with a time constant o f 1 sec since there is no further nonzero input.
- r)dT =
- T)dx
e-'^
for 7 > 0.
AN SW l-R: I
2. For case 2 in Example 15.13, find
for T > 0.
AN SWF. R: ^
1+i/
3. In Figure 15.24, suppose the resistor has value /? > 0 and the capacitor C > 0. Show that the
impulse response o f the circuit is
h{t) = ^ e
aC
= e^'u{-t) V. Compute
1C
RC
ANSWF.RS; 1,
1
----- + a
RC
E X A M PLE 15 .1 4 . The goal o f this example is to design a circuit that computes the running aver
age o f a voltage
over the interval [/^- T, t] given a specification o f the necessary impulse
response, h{t).
S o l u t io n
A circuit that computes a running average must have the input-output relation
From our development o f convolution, such a circuit must have an impulse response h{t) satisfy
ing the relationship
Vnu,iO =
Now, h{t - t) must be a window function that captures the portion o f
Ts
t)
and /;(?).
789
>^h(t)
h(t-i)
T
1
T
-
--t-T
FIG U RE 15.25 (a) The window function h{t - r). (b) The impulse response h{t).
The circuit design problem reduces to developing a circuit that integrates the function segment
V,-,, { T ) d T +
V,. { T ) d T
=dr,
= A + r , and
(15.24)
Since both X and x are dummy variables o f integration, we may replace A by ra n d rewrite equa
tion 15.23 as
(15.25)
where Vj^^{T - 7) is simply a delayed replica o f t^/(x) and where, for practical reasons, we can
replace the lower limit of-oo by /q, the time the actual circuit turns on.
As a convenience, we will define a device called an ideal delay o f T seconds, whose input is yy,/^)
and whose output is
T). Figure 15.26 shows the ideal delay as a device having infinite input
impedance combined with a dependent voltage source whose output is a delayed version o f the
input. Such a device can be achieved by storing the values of
790
All the pieces are now available; it is merely a matter o f putting them together. Integration can
occur using an inverting ideal op amp circuit having a capacitive feedback and resistive input. The
input to this ideal op amp integrator can then scale and sum the voltages
produce the desired running average,
and
T) to
scaling to achieve the desired average, since the transfer function o f the integrator will be M{RCs).
Figure 15.27 shows a circuit that will realize the desired running average.
FIG U RE 15.27 Op amp circuit that produces a running average of the input waveform,
provided that RC = T.
For the circuit o f Figure 15.27, the input drives the first inverting op amp circuit and also feeds
the ideal delay. T he second op amp circuit is an ideal inverting integrator whose inputs are
and
mathematically irrelevant, provided that there is no loading between the circuits. The idea is illus
trated theoretically in Figure 15.28.
h,(t)
h,(t)
'9 1
(a)
h,(t)
h,(t)
(b)
FIGURE 15.28 Interchanging the order of impulse responses, in which the equivalence of parts
(a) and (b) follows from the commutativity of convolution.
Although mathematically parts (a) and (b) o f Figure 15.28 produce the same result, practical con
siderations often dictate a more careful realization process. For example, one circuit may have a
high input impedance w'hile the other has both a low input impedance and a low output imped
ance. In this case we would put the first high-input-impcdance circuit at the front end and the
other circuit at the output end.
A cascade op amp realization o f a transfer function,
H {s) =
100
(15.26)
(.y-Kl)(.9-H20)
illustrates commutativity nicely. A designer may use either Figure 15.28a or Figure 15.28b to real
ize H{s). Magnitude scaling, say by
and
20
0.05 0
(a)
92
20
0.05 Q
(b)
FIG U RE 15.29 Realization of interchanging the order oFcascaded circuits.
Exercise. Verify that each o f the circuits o f Figure 15.29 realizes the transfer function o f equation
15.26. Magnitude-scale by
= lO' to obtain more realistic parameter values. If only 1 [.iF capac
itors were available, you would need two scale factors
for the second op amp stage. What are the rwo scale factors?
AN SW ERS: Multiply each resistor by I O' and di\'ide each capacitor by 10'' to obtain the new val
ues. II only 1 uF capacitors are available, then for Figure 15.29b,
For Figure 15.29a, the two scale factors are interchanged.
= 10*^ and
= 0.5 x 10^.
ple cascades o f circuits or systems can be combined or realized in whatever order the designer
chooses. This assumes that there is no loading between the circuits or systems. Op amp circuits
called buffers or voltage followers having gains o f 1 can be used to isolate stages. On the other
hand, practical constraints may impose a condition on the realization o f a circuit for which the
mathematics o f the associative property does not account.
Finally, we consider the distributive property o f convolution: /;, * {hj +
= {h^
(/?j * h^. One interpretation o f this property is that the superposition o f the input signals
h-^) +
and
is valid. Flowever, Figure 15.30 presents rwo block diagrams with different interpretations.
Here one sees rwo possible topologies for realizing a system.
~93
y(t)
FIGURE 15.30 Two possible block diagram interpretations of the distributive law For convolution.
Exercise. In terms o f reliability, i.e., possibly continued partial operation in the face o f a circuit
failure, suggest reasons why one realization in Figure 15.30 would be superior to the other. In
terms o f minimum number o f components, suggest reasons why one realization would be better
than the other.
tem is time invariant, i.e., each circuit element is characterized by constant parameter values.
Mathematically, this means that
794
input signals y(^). intuitively speaking, time invariance means that if we shift the input in time,
then the associated zero-state response is shifted in time by a like amount. These properties under
lie the development that follows.
Rectangular Approximations to Signals
Let us define a pulse o f width A and height 1/A as 6^(r). In particular.
4A
0 < /< A
(15.27)
othenvise
0
Mgure 15.31 sketches
u(t) - u(t-A) ,
, .
6A(t) = ---- 7---- for several As
1
A.
> t
1
5^(t-kA)
kA
FIGURE 15.32
- M ) = b(^{t -
(k-Hl)A
= M.
79 S
^ v (t)
HD= 2
v(A-A)d^(r-A'A)A= ^
v{ti. )d^(t - tf^)A
k=-oa
k=-x
(15.28)
v{t)
V(/) =
v{tf, )6^ (/
tf, )A
(15.29)
k=-oo
One concludes that if v{t) is continuous, then
X
y
A-*{) A'=-00
^
- ti.)A= C
v'(/)= lim
J-oo
\iT)d{l - t ) ( I t
(15.30)
where we have interpreted the delta function as a limit o f short-duration pulses whose height is
inverse to the width so that the area is constant:
d(t)=
lim d ^ ( { )
A^O
Observe that the right-hand integral o f equation 15.30 is precisely the convolution v{t) * S(r)
C om putation o f Response for Linear Tim e-Invariant Systems
Suppose
If the
circuits impulse response satisfies smoothness conditions, i.e., if it has sufilcient continuous deriv
atives, then the circuits impulse response is the limit o f the /ja U) values as A goes to zero. In par
ticular,
//(/)= lim//^(/)= lim f
A - * ()
/?(f -
T )d^ (T )c/r
= f
A - * ()JO
h{t -T)d{T)dT,
(15.31)
follows from the convolution theorem. Equation 15.31 restates the law that the zero-state response
o f an input to a linear time-invariant circuit is the convolution o f the input with the impulse
response.
Now, by the assumption o f time invariance,
linear time-invariant circuit to
LinearTlme
8,(t)
Invariant Circuit
LinearTime
8.(t-kA)
!< = -CC
Invariant Circuit
(v(gA)8,(t-tJ
LinearTime
k = -cc
Invariant Circuit
LinearTime
v(t)
Invariant Circuit
/-oc
h
FlCirilF. 15.34 Zero-state responses to a particular linear time-invariant s)^stem, showing the framework
of the derivation. Note that the bottom condition follows because as A
0, A
di,
x, and H
j.
It is now possible to use the approximation for v{t) given in equation 15.31 to generate an approx
imation to^(^) in terms o f a summation o f terms o f the form v{kA)hi^{t - kA)A. Taking the limit
as A
0 will yield
To derive this, note tliat for each k\ v{kA) = v{ti^ is a scalar. Hence, the zero-state response to
tf^A is
the impulse response is sufficiently smooth i.e., if it has sufficient continuous derivatives then
v ( / ) = lim
v(A-A)/2 a ( / - A'a )A = r
v{T)h{t-T)ch
Thus, we conclude that for a linear time-invariant circuit, the zero-state response
(1 5.33 )
to an input
excitation v{t) is the convolution o f the input v{t) with the impulse response h{t). We will refer to
equation 15.33 as the impulse response theorem,* which says that the zero-state response o f a lin
ear time-'mvariant circuit or system to a (possibly tivo-sided) input signal is the convolution o f the input
with the impube response o f the circuit.
9 . SUMMARY
This chapter has introduced the concept o f the convolution o f two signals. The convolution can
be evaluated analytically (by direct computation o f the convolution integral) or graphically. Often,
by applying the properties o f convolution algebra, it is possible to implement shortcuts for calcu
lating the convolution o f two signals, resulting in the simplification o f the analytical calculation
or o f the graphical calculation.
Using the notion o f convolution, the chapter developed a technique o f computing circuit respons
es in the time domain. This technique is the direct counterpart o f computing the transfer func
tion in the frequency domain, the approach presented in Chapter 14. Using the convolution
approach, one can compute the zero-state response o f a circuit excited by signals that extend back
in time to -oo, something not directly possible with the one-sided Laplace transform. However,
for one-sided signals which constitute the great majority o f signals that are relevant to circuit
analysis the convolution and Laplace transform approaches are completely equivalent, as
demonstrated by the convolution theorem. The chapter gave an example o f designing a circuit to
compute a running average. In addition, it presented an application of the convolution technique
to the computation o f circuit responses for a circuit whose impulse response is approximated on
a CRT. Future courses will expand the seeds planted in this chapter. For example, convolution is
pertinent to an understanding o f radar techniques, commonly used to identify speeding motorists.
798
Ideal delay o f T seconds: waveform with inputy(/) and o u t p u t 7), a delayed replica ofy(f).
Impulse response theorem: theorem stating that the zero-state response o f a linear time-invariant circuit or system to a (possibly two-sided) input signal is the convolution of the input
with the impulse response o f the circuit.
Linearity: property whereby, if the zero-state response to the excitation f^t) is
then the zero-state response to the input excitation
for / = 1 ,2 ,
^__
* T h e derivation o f this result is, o f course, not rigorous. A rigorous justification is given as theorem 4 o f Sandbergs
Linear Maps and Impulse Responses,
pp. 2 0 1 -2 0 6 .
Problems
799
CONVOLUTION BY INTEGRAL
1. Leijit) = K^5{t - r , ) ,
plot in terms o f A', and
- t
lowing convolutions:
(a)
(b)
P )'u {t^ 2 T ^ )
J{t)'r{t^ T ^ )
(c)
r ,)- (r -4 r i)]
(d)
/ f ) M K r + 2 r , ) - r ( ^ - 2 7 ', ) ]
part (a). Then use the results for (a) along with
(e)
J{t)'u {t)
(b) j{t)-r {t)
(c) J { t ) - [ u { t ) - u { t - A ) ]
id)
J{t)-[r{t)-r{t-2)]
(e) J { t ) - [ r { t ) - r { t - 2 ) - u { t - 4 ) ]
(a)
ing convolutions.
(a )
t({r) * f i t )
(b)
u { t ^ T) * J{ t)
[u{t^ T ) - u { t - T ) ] ' P )
[uir^T^)-u{r-T,)]*J{t-T^),T^
> 0, Tj > 0,
>0
uir^ l)*[J{t)-e-^'% -T)]
(c)
(d)
(e)
(a)
J{t) - cos{m)u{t)
(b)
J{t)
- sin(;rr)(r)
7. Let f i t )
(a )
fir)* fir)
(b) T iW / z W
Now use the results for parts (a) and (b) and the
following convolutions:
(d)
u{t) * u(t - 2)
n { t - 2)
u(^)*[u(t)-;^(t-4)]
u(t^2)*lu(r)-u(r-2)]
(e)
[ ( f + 2 ) - ;v U ) ] * [ / / W - ( r - 2 ) ]
(a)
(b)
(c)
following convolutions:
(c) f i r ^ T ) * f i t - T ) , T > 0
(d)
(e)
5 (f+ 2 ).
f(t)
(/) * Hr)
Hr) r(r)
u(r)*[r(r)-r(t-4)]
r (r )* [ r ( r ) - K r - 4 ) ]
M r + 2 )-/ / W ] * l u ( r ) - u ( r - 2 ) ]
(0
(g)
>k
3 2
Figu re P i 5 .8
800
9. (a)
(b)
. v(t)
2K
K
(a)
H------h
1
3
(b)
2K
Figure P15.11
2T
>t
-T
--K
r\
(b)
u{t) * r{t).
Figure P I5.9
(b)
lO.(a)
1*1
1
/*
usmg graphical convolution ror 7 = 1.
i. v(t)
ties o f convolution.
(b )
=2
and 7 = 1.
2A
.f(t)
/N
2T
(a)
4T
(b)
2+
Figure P i 5.12
13.
> t
> t
-T
2T
2T
Figure P I 5.10
Problem 5.
(b)
1 l.(a)
f (t )
a >
>t
-T
801
2T
h{t),
of the
(a)
Figure P I5.13
b. Compute
(a)
ical convolution.)
h{t) =
(b)
tion 15.2.
(a)
(b)
yjit) = h{t) *
^<
graphical convolution.)
00.
y^{t) =
h{t)
*^ (/ ).
<
r < 00.
/(^) = u{t + 3) - (? + 1)
(c)
-o o
(d)
(e)
'" 6
Figure P I5.18
19. A particular active circuit has the transfer
function
Let h{t) = 4(/). Find^(/) = h{t) * f^it)
15.(a)
H (s) =
s+a
(b)
PI 5 .15b.
A
yit)-
f ,( t )
V (t)
>
>t
- t
-T
Figure PI 5.19
Figure P I 5.15
16. I f / / ) = K^e^ui-t), ^ > 0, and h{i) = K2 u(t),
compute ^(/) = h(t)
s+ a
s+ b
802
the circuit is
hit).
(b)
where
V,
(d)
vin = [1];
h = [0. 2, 3, 1, 1];
y = tstep*conv(vin, h)
y = [0 y 0];
t = 0:tstep:tstep (length(vin)+length(h));
= 5e^^uH) V.
plot(t,y)
grid
/
o
+
2H
5Q
0.5F
h(t)
Vou,(t)
3 --
2-
Figure P I5.20
21. The transfer function o f a particular time-
> t
H{s) =
(a)
(b)
+
.? + 1
i- + 2
Figure P I5.22
A+ 4
cuit.
(i)
(iii) define k = 0;
(iv) use
(c)
the
command
[n,d]
denominator polynomials.
(d)
?,u{t-T)y T > 0.
GRAPHICAL CONVOLUTION
22. The impulse response o f a particular circuit
a-
803
h(t)
*<
5 -hO
p(x) = vO X + vl
q(x) = ho x + h 1 X + h2
h2
r(x) = yl x^ + y2 x^ + y3 x +y4
Show that rix) = p{x)(]ix) with yl through y4
- 2 -1
-
(b)
y(t)
> t
P I5 .2 4 c is
V = [ 2 4];
h= [ 3 - 2 1];
T = 1;
tstep = T;
Figure P I5.24
y = [0 conv(v,h)*tstep 0];
% The additional beginning and
ending zeros are added to indicate
% that the initial and final values o f
the convolution are zero, due to the
% finite duration assumption,
t = 0: tstep : tstep* (length(v) +
length(h));
A N SW ERS: (a) yl = vO x hO = 6, y2 = vO x h 1
+ vl X ho = 8 , v3 = vO X h2 + vl x h 1 = - 6 . and
v4 = vl x h2 = 4
25. Repeat Problem 24 for the waveforms in
Figure PI 5.25. When using the MATLAB
code, it is necessary to account for/j(/) being
plot(t, y)
grid
v(t)
5 -vl
442
f,(t)
vO
(a)
-4..
(b)
(a)
Figure P i 5 .2 5
.so^
If the input is
output
Figure P I5.27
(c)
choice.
CONVOLUTION ALGEBRA
and
Figure P i5.26
27. A crude approximation to the impulse
response o f an active RLC circuit is given by
h{t)y as sketched in Figure Pi 5.27a. II the input
signal is v{t), as given in Figure P i5 .2 7 b , find
the response y{t) using MATLAB and the
> t
-4 . .
(b)
-T
2T
(c)
Figu re P i 5 .2 8
O '
(a)
fj(t)
in Figure PI 5.32a.
(b)
(c)
Find
> t
(or T = 2n^l(LC) .
-4 --
iYYV
Figure P i 5.29
^ouM^~
(a)
(b)
Figure P I5.32
techniques.
31. Use convolution algebra techniques to
determine y / ) = h{t) * j{t) for each of the func
tio n s //) given in Figure P I 5.31 where h{i) =
At)
>t
(a)
(a)
(b)
(c)
(d)
w
w
(a)
A(.)A
w
Vw^
Vw>
'O
> t
2
1
> t
- tse c
2n
4n
6n
(b)
Figure P I 5 .3 4
806
j{t)
P 15.35.
(a) j{t) = o r cos{ojt)u{t)
(b)
(c)
H(s)=
Vin
A
s +a
shown
in Figure P I 5.37.
(a)
Define
(/) = Vj(t - T )and compute
(b)
Compute
(c)
To compute
CONVOLUTION BY INVERSE
LAPLACE TRANSFORM
).
due to
use
<
transfer function
>t
-T
V
Figure P I5.37
s+2
Vin
shown
(t) =
V/( t
- T )
and compute
(b)
Compute
(a) f i t ) *
To compute
due to
use
(b)
Vout(^)=^out(l + T')-
- T^) * }>(t)
/)^,'(0 = [ /
f=t+T,
- r , ) * g(r - r ,
v.(t)
are cas
>t
-T
here?
Figure P I5.36
ANSWHK: (c)
= ti{t + 2) V. Would it be
= (I-
+T ) \
v.Jt)
SO
h,(t)
h,(t)
' 6
Figure P I 5.39
40. This problem shows the advantage o f the
Figure P I5.41
s
R,C
= 2tu{t).
S~ +
R\C\R2C2
(b)
H(s) =
(a)
tion expansions.
response, h(t).
(b)
v.(t)
h,(t)
h,(t)
h3(t)
Find
= 54tr^u{-t)
when
due to Vout^O~).
(c)
Figure P I5.40
from
(d)
= \Qe^^u{-t) V. Can
the input
72te~^u{t) V, find
the
complete
calculations.
C,
H(s) =
^e
and the impulse response h{t) in terms
v(t)
o f R and C.
(c)
Given
Fig u re P i 5 . 4 2
method.
Evaluate
the
MISCELLANEOUS
h-^i) = 4 ( f - 2 ), and
(a)
(b)
= k cos{,m)u{t).
j{t) = 6 (r).
2 cos(;rf)tt(r).
h,(t)
(a)
f(t)-
(b)
h,(t)
M = 6u{t).
Figure P I5.43
hjit) =
h^{t) =
and //4(f) =
K t)= y
> y(t)
f(t)
6{t-kT)
itti
be a so-called impulse train. Find L
Note that if
00
f(k T ) z -'
it=0
809
v , (0 = ^ 6 { t + kT)
A=()
is a periodic impulse train with T = 1 sec.
(a)
cc
' ,)= ' ^ d U - k T )
k=0
with r = 1. Find
resulting waveform.
(b)
for 0
<t< \.
(c)
for 1
(d)
(e)
1 -A
A -O
\.
provided
(f)
1 -A
'"6
20
IF
Figure P i 5 .5 0
(1)
-4
(1)
-3
(1)
-2
(1)
-1
(1)
-> t
Figure P I5.51
AN SW ER: i>Jr) = 1.27t'*'-^^ for r > 0
row 1
697 Hz
row 2
770 Hz
row 3
852 Hz
row 4 941 Hz
colum n 1
colum n 2
colum n 3
1209 Hz
1336 Hz
1477 Hz
keypad
Pressing any one button generates two tones, with the frequencies selected by an electronic circuit
inside the telephone set. For example, pressing the number 5 generates tones at 770 and 1336 Hz.
The row and column arrangements and the dual-tone method permit the representation o f 10
digits (0 ,... , 9) and two symbols (*, #) using only seven tones. These seven tones are divided into
two groups: the low-frequency group, from 6 9 7 to 941 Hz, and the high-frequenc)' group, from
1209 to 1477 Hz.
"
Such tones are easily produced by an LC resonant circuit. The four tones in the low-frequency
group are produced by connecting a capacitor to four different taps o f a single coil (inductor). A
similar connection generates the three tones in the high-frequenq' group. When a button is
pressed to the halRvay point, a dc current from the central office is sent through the coil in the
tank circuit. When the button is fully pressed, the dc current is interrupted. This action initiates
sinusoidal oscillations in the LC resonant tank circuit at a frequency inversely proportional to
V L C . The presence o f small resistances causes the oscillations o f the tank circuit to die out.
However, pressing the button fully also connects the tank circuit to a transistor circuit that replenishes the lost energy and sustains the oscillations.
At the central office, the equipment used to detect the presence ol the tones and to identify their
frequencies is much more sophisticated. Two filters are required, one for each ol the frequency
groups. Each filter must pass the frequencies within 2% o f their nominal values (697 to 941 Hz
for one filter and 1209 to 1477 Hz for the other) and reject the signal if the frequencies are out-
side o f 3% limits. The output tone from each filter is then proce.ssed digitally to determine its
frequency.
The concepts and methods developed in this chapter will allow us to understand the properties o f
resonant circuits and the design o f various basic t}^pes o f band-pass circuits, as used in the touchtone telephone system.
CHAPTER OBJECTIVES
1.
2.
3.
plots.
Investigate the basic band-pass transfer function and its realization as a parallel or series
5.
6.
o f resonant circuits.
Investigate general second-order transfer functions having a band-pass type o f frequency
response.
SECTION HEADINGS
1.
Introduction
2.
3.
4.
5.
6.
7.
8.
Poles
Summary
Terms and Concepts
9.
Problems
1. INTRODUCTION
How is it possible to listen to a favorite radio station by merely pushing a button or two or sim
ply turning a dial? Why do some very expensive receivers have very clear reception, while with
some cheaper models other stations chatter in the background? What circuitry inside the radio
makes this difference? The ability to clearly select a particular broadcast station depends on the
design o f an internal band-pass circuit. Such a circuit will pass signals within a narrow band o f fre
quencies while rejecting or significantly attenuating signals outside o f that band. To understand
why this is important, note that audio signals have significant frequency components up to about
3 kHz for voice and up to about 15 kHz for high-fidelity music. These frequencies are far too low
for wireless transmission. In (wireless) AM radio transmission, the audio signal modulates the
amplitude o f a carrier signal that is suitable for wireless transmission. The carrier signal is a highfrequency sinusoidal waveform bet\veen 500 kHz and 1650 kHz. The modulated waveform con
tains many frequency components centered about the carrier signal frequency, but extending over
a range o f frequencies equal to twice the highest audio frequency. For example, the radio station
WBAA, at Purdue University, has a carrier frequency o f 920 kHz and occupies a band from
approximately 915 to 925 kHz. To select WBAA from all the carrier signals received by a radio
requires a good band-pass filter to pass the frequency band o f 915 to 925 kHz while rejecting sig
nals outside this band. This chapter introduces the idea and properties o f a band-pass filter.
In its simplest form, a band-pass circuit consists o f only one capacitor, one inductor, and one resis
tor, connected either in series or in parallel. In the first half o f the text, we analyze simple RLC cir
cuits where we emphasize (1) transient behavior under dc excitation and (2) sinusoidal steady-state
behavior at a single frequency. This chapter investigates the behavior o f circuits over bands o f sinu
soidal frequencies. iMany useful results may be obtained with the phasor and impedance concepts
studied earlier. However, rapid advances in technology have made it possible to have a band-pass
circuit without any o f the usual RLC circuit components. Therefore, a study o f the band-pass
propert)' o f a transfer function H{s) dominates the material o f this chapter. The resulting analysis
is readily applicable to general linear systems, whether they be electrical, mechanical, or otherwise.
Although transfer functions underlie our approach, circuit realizations with ideal and practical
components illustrate all the basic concepts and properties.
S14
ideal band-pass (rectangular) magnitude response curve. Here, ideal means that all frequency
components o f the input signal within the range 0)j < to < (o^ are amplified with equal gain (in
magnitude), and all frequency components outside o f the range are totally eliminated from the
output. Actually, for a band-pass circuit to pass a signal with frequency components in the range
0)j < CO < 0J 2 without distortion, there is also a requirement on the phase response that is ordi
narily studied in a course on signal analysis.
(b )
S IS
(C)
istic. (b) Approximate band-pass characteristic of simple RLC circuit, (c) Band-pass characteristic of a
more complex circuit.
Unfortunately, an ideal rectangular band-pass characteristic is not realizable by a rational transfer
function or any circuit. One can, however, approximate the ideal characteristic with a simple
tuned circuit whose transfer function produces a bell-shaped magnitude response curve as illus
trated in Figure 1 6 .lb. With more complex circuits and more complex transfer functions, one can
improve the approximation as shown in Figure I6 .1 c. (How to do so is a topic to be studied in
advanced courses.) The bell-shaped curve o f Figure 16.1b has several important features. The fre
quency at which |//(/w)| reaches its maximum value,
denoted by
The two side frequencies at which |//(/w)| is 1/V 2 o f its maximum value are
called the half-power frequencies, denoted tO| and 0^2 - The term half-power comes from the
fact that if the output is a voltage across a fixed resistance, then a drop in voltage by the factor
H en ce the terminolog}^ 3 dB d o w n .
71
For obvious reasons, 0)| is called the lower half-power frequency and
frequency, and their difference
o f the band-pass circuit/transfer function. Band-pass circuits are designed so that all frequencies
o f interest fall within the pass band [(Op co^].
One way o f categorizing and comparing different band-pass circuits/transfer functions is by their
selectivity, i.e., a circuits relative capability to discriminate between frequencies inside the pass
band and signals outside the pass band. The selectivit)' is measured by the quality factor, Q, o f a
band-pass circuit/transfer function. The quality factor, Q, is the ratio o f the (geometric) center fre
quency {yj(JJiC0 2 ) to the bandwidth,
bell-shaped magnitude response (Figure 16.1 b), the center frequency and peak frequency
cide, as we will show. In this case Q =
coin
where Q denotes the pole Q, to be defined in equation 16.1. For the circuit realization o f the
transfer function, Q is sometimes denoted by
or
function |//(/o)| and, therefore, on how the transfer function H{s) is defined. Even for the same circuit,
these values are different when the output is associated with different branches or when the input is
changed from a voltage source to a currcnt source. Further, for the investigation o f the frequency-selec
tive characteristic of the circuit, the foregoing definition o f Q is most appropriate because it directly
assesses the sharpness o f the magnitude response cur\'e. As such the definition allows Q to be experi
mentally determined. In certain other applications, where only one fixed frequency is o f interest, there
is another definition o f a circuits Q based on an energ)' relationship that is inadequate for general band
pass circuit design.
(s + G p - j w j ){s + Op + jco^j)
(16.1)
+ 2apS + a l + m ; ,
r + 2ct,,.v + (o^
Qp'
"
Figure 16.2 illustrates the pole-zero plot o f the transfer function o f equation 16.1 as well as the
relationships among the various parameters. Note that there is a finite zero at the origin. Further,
because the numerator has degree strictly less than the denominator, there is also a zero at ^ = co.
In both cases the magnitude o f the transfer fijnction is zero.
SI
JO)
FIGIJRK 16.2 Pole-zero plot of a band-pass H{s) with a single zero at origin as per equation 16.1.
Our next goal is to derive the peak frequency
the half-power frequencies tOj and t02,
=|
K
(16.2a)
2 o + 7 ----------- (JO
m ax
(I)
1^1
H(Jco)
I CO., - 0 )
l\~\
I
4a+
mm
0)
(0
)
9 \(OZ, - U>~ \
\K
2a,
(16.2b)
CO
Note that since the numerator is a constant, the maximum occurs when the denominator is a min
imum, i.e., when co = (O^. Therefore we conclude that
(16.3)
COI,
Noting that Qp ~
, we further conclude
H(jco)
(16.4)
To find the half-power frequencies lOj and co-,, the maximum gain must be reduced by the factor
l/V 2 . Considering equation l6 .2 b , this occurs at those cos for which
818
(col-co-)
CO
(16.5)
= 4a
(16.6)
We denote the two positive solutions by cOj and a>2. with CO2 > o jj. Solving equation 16.6 using
the quadratic formula, we have
yJcTp+OJ-p = CO.
(
(16.7)
B,^=co2-C0^=20p=-^.
(16.8)
> 8),
(16.9)
We can deduce one more important property from the above equations. When equation 16.5
holds, the real and imaginary parts o f the denominator in equation 16.2a are equal. Hence the
denominator has an angle o f 45^^. Thus the angles o f
H(jcOi) = ----2CT. + 7
(cor-col\
CO:
--------- , find
5 ^ + 2 ^ -h 2 5 6
A N SW I'RS:
2. Suppose that
, (0,^ ,
, and a>i 2 .
= 8 and 10^ = 1000 rad/sec. Find the exact to^ ^ from equation 16.7, the
from equation 16.8, and the approximate Wj 2 from equation 16.9. Now compute the magnitude
o f the percent errors between the approximate and exact values o f tOj 2A N SW ERS: in random order; 125. 939.^^. 1064.5, 1062.5. 937.5. 6 .1833, 0 .2077
3. Given equation 16.9, show that for a high-(^ transfer function/circuit (say
> 8),
819
i.e., for the high-Q case the geometric and arithmetic means of 0)j and 0J 2 are approximately equal.
The next two questions are: (i) what is the geometric center frequency o f the magnitude response
and (ii) is it equal to to ? Here, using the square o f the geometric center frequency,
CO1 CO2 = -O p + J a j - c o l
Op +
p+<4
= - Gp + Op + ( O p = ( O p
we conclude that for the above transfer function and associated circuit realizations, the geometric
center frequency, the pole frequency, and the peak frequency coincide, i.e..
Q = Q ..- Q ,, = ^
(16.11)
Finally, the nonzero frequency for zero phase shift occurs when the transfer hmction is purely real,
i.e, when the imaginary part is zero. This occurs at CO = CO^; hence
(16.12)
= |Myto^^^)|, Q,
values) for the simple parallel RLC circuit shown in Figure 16.3a where the current source is the
input and the voltage across the input nodes is the output. Then compute and label the magni
tude response and verify that it has the bell shape o f Figure 1 6 .lb.
S20
.(b )
FICiURE 16.3 (a) A parallel /?/,C circuit, (b) Magnitude response curve.
So
lu t io n
The key to the solution o f the example is the computation o f the transfer function as follows:
1
Y(s)
^
s
C
C5 + +
Li
.V + - .y-l-
RC
LC
H{s) = K
=K
we conclude that Op = ^
, cOp = j- , and
2 R C " " ' - 1 W
iOn
RC
Q = Q, = ^
= 4^ = R
c
L
(16.13a)
2a^
( 1 6 .1 3 c )
i '- 1
= R.
li, = 2 0 p = - ^
(16.14)
Similarly from equations 16.7 and 16.9, the exact half-power frequencies are
tO\ H---------- 1- 4
IR C V \1RC}
LC
(16.15)
while the approximate half-power frequencies for the high-(^ case are
f/Ji s p = ----------,
VZC
IRC
W o s j^=H --------- .
ylLC
(16.16)
IR C
Finally, to obtain the frequency response we note the values o f the above computations, and the
fact that at
CO
= 0 and
co
tOp O),,
when R - 2.5
Z, = 0.1 H, and C = 0.1 uF.
AN SW ERS: (0,,, = 10"^ rad/sec, o , = 12.198 x lO 'l to. = 8.198 x 10\
The preceding example and exercise demonstrate that for high-Q circuits (Q > 8), there is really
no need to use the exact equation 16.15 to compute cOj and co-,, as the much simpler estimates
given by equation 16.16 are sufficiently close to the true answers. Indeed, sometimes it is con
venient to use the approximate formulas when Q > 6.
In many practical circuits, the independent source could be a voltage source in series with a resis
tor. Before applying any o f the foregoing formulas, it is necessary to transform the circuit into the
form o f Figure l6.3a by the use o f the Norton equivalent circuit studied in a first course on cir
cuits. The resistance R in Figure l6.3a then is not a physical resistor, but rather, the equivalent
resistance of several resistances in parallel. The following example illustrates this reformulation.
EXA M PLE 16.2. In the circuit o f Figure 16.4a, an independent voltage source
in series with
(c)
Find I
S22
+
V
(a)
lu t io n
The solution proceeds by using the formulas developed in Example 16.1 after replacing the prac
source by its Norton equivalent and identifying R in Example 16.1 as the parallel com
tical
bination o f
R,R,
4 0 .0 0 0 x 1 0 ,0 0 0
_
R=
=
:------ = 8 ,0 0 0 Q
Rs + R l
4 0 ,0 0 0 + 10,000
This results in the transfer function
s
C
'm
(a)
R,C
\
RC
LC
RC
Notice the new value: K =
LC
R,C
W,=(Op =
^
= 3 1 ,6 2 2 .7 7 rad/s
V 0 .0 2 x 0 .0 5 x 1 0 " ^
1
Bco =
RC
8,000x0.05x10
-6
= 2,500 rad/s
823
and
6
. ^
3 1 .6 2 2 .7 7 ^ , , ^
Z50 0
(b)
Using the approximate equations for high-Q circuits, we obtain the half-power frequen
cies as
^
+^
and
"2
(c)
R,CB^
,
/?,
Notice that we did not use equation 1 6 .13c because the input in Example 16.1 is a current source
and not the voltage source o f Figure l6.4a. Hence the maximum value is not equal to R.
This example demonstrates that putting an external resistance in parallel with the LC tank circuit
reduces the value o f R, which in turn causes a larger bandwidth and thus a lower circuit Q while
keeping the peak frequency
unaffected.
Exercise. Repeat Example 16.2 with the element values changed to R^ = 36 kQ, L = 40 mH, C =
0.25 pF, and R^ = 4 kH.
A N SW ERS: 10,000 rad/sec, 1111.11 rad/sec, 9, 9 4 44.44 rad/sec, 10,555.55 rad/sec.
The examples so far have illustrated only the analysis o f parallel IU.C circuits. In the design of a par
allel-tuned circuit, we must also pay attention to other factors, such as available component sizes,
desired voltage gain, and cost. In practice, design specifications ordinarily impose a small number of
constraints relative to the number o f circuit parameters to be determined. Consequendy, realistic
design problems usually do not have a unique answer, as illustrated in the next example.
E X A M PLE 16.3.
Design a parallel RLC circuit, as shown in Figure l6 .4 a , to have a magnitude response with^^^ =
200 kHz and a bandwidth o f 20 kHz. Only inductors in the range 1 to 5 mH are available. The
source has a resistance R^ = 50 kH.
So
lu t io n
For the circuit o f Figure l6 .4 a there is a restriction on the available inductors. Hence, we keep L
as a variable, subject to the condition that 0.001 < L < 0.005 H. Using the specified peak fre-
.S2-
LC
= 6 .3 3 x 1 0 "'^ L
C = - r r = - -r
From the specified bandwidth and equation 16.14,
= ^
x l0 U
2 ;r x 2 x lO '^
As explained in Example 16.2, R is the parallel combination ol
or
1 -_L
_L
R ~ Ri_
R, '
R[
R,
Once a specific value o f L is chosen, we can calculate successively the values of C, R, and Rj. Since
R, which is the parallel combination ot R^ and R^, must be no greater than
= 5 x lO'^ f2, the
upper limit for L is
5 0 .0 0 0
R.
^ma\ ~
1 .2 5 7 x 1 0 '
= 0 .0 0 3 9 8 H .
1.257 x 1 0 '
Numerical values corresponding to the extreme values ot L are given in Table 16.1.
TABLE 16.1
L (mH)
C (p F )
R(VQ)
Rl i ^ )
633
12.57
16.77
3.98
159
50
infinite
Table 16.1 clearly shows that there is no unique answer to the design problem. 'Fhe freedom in
choosing a value for L in the range 1 to 3.98 mH can be utilized to accommodate another design
specification, such as a value for
Exercises. 1. In Example 16.3, find the maximum and minimum values o f |//(/to^^^)|.
AN SW ER: 0.251, 1.0
2. In Example 16.3, if the bandwidth requirement is changed to 10 kHz, determine the minimum
and maximum possible values o f L.
AN SW ER: 1.0 mH and 1.99 mH
H2S
Dual to the parallel /?/.Cof Figure 16.3a is the series y?Z.Cof Figure 16.5, which has a voltage input
and a current output. Although we can use clualit)', to infer frequency response behavior, we pre
fer the direct transfer function approach.
E XA M PLE 16.4. For the series RLC circuit o f Figure 16.5, let 1^ be the desired output. Find
(a)
(b)
(c)
The bandwidth,
(d)
The circuit Q
(e)
The half-power frequencies, to, and (o-, (exact and approximate values)
(0
(g)
W, = |M;co)|,,= |//(/0|
The frequency at which the angle o f H{jio) is zero
R
V
z , ( .)
(16.17)
-)
.y + .v-l- -
C ,.'
= K S~ + l O n S + Clj:
=K
5 + ^S + W~
Q
and
'
= /,
C ircu it Q:
=Qp =
lc
(16.18)
^p^S
( 1 6 .1 9 )
Cs
826
Half-power bandwidth:
(16.20)
(0 ,2 = 0 3 ^ ,
iV
4Q -
2Qpj
2L,
-B co
i -------
^ l^s
2L,
/, /'
0 6 -2 2 )
(o,
1
(16.23)
R,
= 5 O.,
= 1 mH,
= 10 rad/scc.
The above derivations and calculations yield formulas and numbers. To add some meaning to the
concept o f circuit Q and its relationship to bandwidth, we provide a plot o f the normalized |//(yco)|
vs. (0 curve for different values o f
|//(yco)|/|//(yo))|^^^^^, while the abscissa shows the ratio to / to T h e se curves are called universal res
onance curves because they are applicable to parallel RLC circuits, to series RLC circuits, or to any
system having a transfer function o f the form o f equation 16.1. Observe that as Q increases, the
bandwidth decreases, indicating a better selectivit)^
82'
(O/Wm
FIGURE 16.6 Normalized magnitude response for equation 16.1.
To conclude this section, we present an example o f an active band-pass circuit that avoids the use
o f inductors. The band-pass circuit illustrated here is only one o f more than a dozen configura
tions in use. This example illustrates the possibility o f eliminating inductances while producing
the same kind o f frequency response as the parallel or series RLC circuit. You can learn a lot more
about these active filters in a more advanced course.
100 rad/sec. Find Q for the desired transfer function. Then realize (find values for /?p Rj,
Cp and C-y) for the normalized transfer function
H.
Ks
,{s) =
S
5 +
,
1
3.
4.
q = C2 = 1 1-iF
Verify the frequency response with a SPIC E simulation o f the real circuit.
82.S
MGURE 16.
S o lu t io n 1.
From the assumption that the operational amplifier is ideal, we have K_ = 0 (virtual ground) and
/_ = 0 (infinite input impedance). Applying KCL to the inverting input node V_, we obtain
'C , K , + - ^ = o ,
A-7
which yields
v;. = -
to obtain
V
^'ci - V
* -in
/e,
+ iC | V + iC 2 (K -V ,)= ()
^out
ou t
-s
~RxC^
+
(16.24)
1
,/?2C,
R iC iV
I^\Ri C\C2
H{s) =
V'/(^V)
+ Ba) S + 1 0 ;
Ks
2
i-" +
Ks
S + iO
We now frequency-scale s -*
with
is
_
=
?>.Computation o f element valuesfor normalized transferfiinction. Equating the circuit transfer func
tion with the normalized transfer function, we have
--------------- = 1 rad /sec
(16.25)
R\
and
H---------- --- 0.1 rad /s e c .
(16.26)
Ri R2C\C2
R[20
= I H U co, ) I =
= ^
=^
= 200
RjC^ ^ RoSi
4. Frequency and magnitude scaling. As per the problem statement, we desire
Hence we frequency-scale with Kj- = 1000. It is further required that
= 1000 rad/sec.
= 1 liF. Hence,
KfjjKj
K, =
10
X 10
= 1000
It follows that
c , = 1
50n,
iOOOR2M- 2 0 n
The ratio o f /?, to R^ is very large and unrealistic. It turns out that this circuit is best suited to lowQ transfer functions.
5. Verif' frequency response. To verify the frequency response for a realistic implementation o f the
above circuit, we consider a SPIC E simulation using the standard 740 operational amplifier as
shown in Figure 16.8. Observe that the maximum value o f the magnitude response is 200, as
expected, and that the bandwidth is about 16 Hz, which translates to about 100 rad/sec.
830
+100.000
+150.000
+200.000
Frequency (Hz)
+250.000
+300.000
Exercises. 1. In Example 16.5, find the circuit Q and the approximate half-power frequencies
CO
and
R,= 30 kU
Throughout this section we have assumed the use o f ideal inductors and capacitors. Practical
inductors have complex circuit models to account for real-world behavior. The next section takes
up an approximate analysis o f circuits containing simple models o f practical inductors and capac
itors.
8 31
a simple model o f a practical inductor for low to medium frequencies while Figure l6 .9 b is a rea
sonable model for high frequencies.
FIGURE 16.9 Two models of an inductor, (a) For low to medium frequencies, (b) For high frequencies.
The primary parameter is, o f course, the inductance L. The remaining elements, R^,
and C^,
account for undesirable yet unavoidable practical effects and are called parasitic. Since an inductor
usually consists o f a coil o f wire, R^ represents the wires resistance. Also, a capacitance is present
between adjacent turns o f wire. Hence
models this parasitic capacitance. The resistance R
accounts for the energy loss in the magnetic core material (if present) inside the coil. Complex mod
els such as Figure 16.9b, although important, if used for every inductor would unduly complicate
the analysis o f a band-pass circuit. Fortunately, for low to medium frequencies (up to a few mega
hertz), the simpler model o f Figure l6.9a suffices and hence underlies the material that follows.
Figure 16.10 shows two models o f a practical capacitor. Again, the primary parameter here is the capac
itance C; R , Rp and
are parasitics. The leakage resistance, R^, accounts for the energy loss in the
and resistance R^ are due mainly to the connecting wires o f the capacitor.
, the capacitor actually behaves as an inductor! For frequencies of up to
a few megahertz, the simpler model o f Figure 16.10a suffices, and it is used for the analyses o f this text.
(a)
FIGURE 16.10 Two models of a practical capacitor, (a) For low to medium frequencies.
(b) For high frequencies.
832
How close is a pracrical inductor (Figure l6.9a) or a practical capacitor (Figure 1 6 .10a) to the
ideal? The so-called element quality factor provides a quantitative measure. To develop the qualit)' factor, consider that each practical inductor or capacitor has an impedance
Z(;(o) = RejZ(>))} + j Im {Z(>))} = /^(oj) + >V{(o)
For both the practical inductor and capacitor, the frequenc)-dependent reactance X is the primary
parameter of concern, whereas R represents the unavoidable parasitic effect. In the ideal case, R
is not present. Hence, the ratio
provides a measure of how close the network model is to an ideal inductor or capacitor. The larg
er the ratio, the better the element behaves as an ideal inductor or capacitor, (ideally, R = 0 and
the ratio is infinite.) This suggests defining the quality factor associated with a practical inductor
or capacitor having impedance Z(yco) as
,
UCw)!
The inclusion o f co in equation 16.27 is to emphasize the fact that Qy depends on the frequency'
o f operation. The subscript Z indicates a generic impedance and may be replaced by more specif
ic descriptors such as coil or capacitor. Unlike the Q o f a circuit, which depends on the values
of the elements of the circuit and on the circuits configuration, the quality factor Qj/oi a practi
cal inductor or capacitor varies with the operating frequenc)' OJ and remains unchanged irrespec
tive o f its connection in the circuit. Any element with a finite ratio given in equation 16.27 is
termed a lossy component, as are all real-world components.
For the practical inductor model o f Figure l6.9a, equation 16.27 reduces to
R^
^
where we may synonymously denote
as
(16.28)
As mentioned, higher
implies a better-qual-
it)' coil in the sense that the energ)' loss in the component is smaller. Infinite
inductance L, which is lossless. In the audio frequency^ range,
represents a pure
as in the radio frequency range, it may exceed 100 in practical applications. Although R^ here
varies with frequency (due to the skin effect), at the level o f this text, we treat R^ as a constant inde
pendent o f (0 .
Similarly, the capacitor model o f Figure 1 6 .10a has impedance
I
Z ( = - j -----------=
"
=S =
(6.29)
S33
A higher
implies a better-qualiry capacitor, in the sense that the energ)' loss in the device is
smaller and closer to ideal. Infinite Q^-represents an idealized, lossless capacitor modeled by a pure
capacitance. In practice, Q^;;is usually much greater than Q^, meaning that
circuit performance, i.e., Q(;;is often assumed to be infinite. The reciprocal o f Q^^ is called the dis
sipation factor o f the capacitor and is denoted by cIq A lower dissipation factor means a betterquality capacitor.
The determination o f
and
o f 0) is unspecified, the analysis o f a band-pass (or tuned) circuit proceeds under the assumption
that
= Q^(cOq) and Q ^= Q^oOq), where C0q = M \[l C . We will discuss the meaning o f oJq in
section 5, on resonance.
RLC and parallel /?ZC band-pass circuits with practical sources have the more complex configura
tions o f Figures 16.1 la and b, respectively. Since they are no longer series or parallel RLC, their
transfer function is not the ideal band-pass type of equation 16.1. Hence the associated formulas
for peak frequency, bandwidth, etc. are not directly applicable.
FIG URE 16.11 (a) Model of a parallel tuned circuit using practical inductor and capacitor models,
(b) Model of a series tuned circuit using practical inductor and capacitor models.
834
Topologically speaking, the circuits o f Figure 16.11 are series-parallel because the input imped
ance seen by the source consists o f a sequence o f series connections and parallel connections o f
simple networks. Exact analysis o f such series-parallel band-pass circuits to obtain
and
is
cumbersome, especially in light o f a simpler, more efficient method widely used by engineers to
compute approximate solutions. The approximate analysis relies on the conversion between a
series circuit and an equivalent parallel circuit at a particular frequency. This conversion process
depends on the component quality factors developed in the previous subsection. See Problem 85
for the development of this equivalence. The next example illustrates the conversion process for
an inductor.
EXA M PLE 16.6. When the ideal components o f a parallel RLC circuit are modeled with a prac
tical inductor and a practical capacitor, the circuit is no longer parallel. However, by converting
the series inductor model to an equivalent parallel model, we can proceed w'ith our standard
analysis. To illustrate this conversion, consider the practical inductor model o f Figure 16.12a and
the equivalent parallel configuration in Figure 1 6 .12b. The goal o f this example is to find
L in terms o f
and
and
at a particular O).
(a)
(b)
FIG URE 16.12 Conversion o f an inductor model from (a) a series connection to
(b) a parallel connection at a fixed frequency.
So
lu t io n
Y<ko) = _____!____
For Figure 1 6 .12b,
yrU (o) = - j
Rp
Equating the real and imaginary parts o f the above t\vo admittances, we obtain
Rp =
and
ly
= Rs[' +Q h<o))
(16.30a)
83^
co"-L,
q I uo)]
where
(oL
At any particular frequency (O, if Qy (to) is sufficiently large (say Qj (o)) > 8), then equations 16.30a
and b suggest that
R., a
{(o)Rs = Q i( co ) x( oL
(16.31a)
^p-^s
(16.31b)
and
Conclusion: In a "paraller RLC with a practical inductor, we can replace the practical inductor by its
parallel counterpart valid in a neighborhood o f a single frequency and analyze the circuit to obtain
approximate values o f peak frequency, bandiuidth, etc.
the element values in the series representation and the parallel representation at 100 kHz as shown
in Figure 16,35.
A N SW ERS: l.^ = 2 mH, R^ = 25.13 LX
A similar derivation can be done for the practical capacitor model o f Figure 16.13a. The details of
the derivation are left as a homework problem. The exact conversion equations for a specific to are
(,
and
\
(16.32b)
and
1
Q c (c o )
(1 6 .3 3 b )
(a)
(b)
F-KiriU-. 16.13 Conversion of a capacitor model from (a) a parallel connection to
(b) a series connection exact at a fixed fi-equency.
Table 16.2 summarizes the various conversion formulas for both parallel to series and series to par
allel inductor and capacitor models.
I'AMI.F 16.2 Conversion of Models for Practical Inductors and Capacitors
Approximate
Original Circuit
Exact Equivalent
Circuit at (d,,
{Q, > 8
> 8)
and
Q^R.
=Q,.o)oL
R (i+ Q l)
Hf-
Q,((i)^ = o)R,C
Q;
CL
1 + 1/Q'
1
Qc(o) = (oR,C
Q,
S '-
The next example illustrates the application o f the conversion formulas to a practical circuit to
obtain approximate band-pass characteristics.
EXA M PLE 16.7. The circuit o f Figure 16.14a contains practical components that make the cir
cuit not amenable to the formulas for ideal series and parallel RLCs developed earlier. For exam
ple, the sinusoidal source is represented by an independent voltage source
internal resistance
tion factor o f
= 40
in series with an
0.01 at cOq. Here cOq represents the undamped (no resistance) natural frequen
^
' = 7 Z c -
The practical coil has an inductance o f 20 mH and Q/(cOq) = 40. The external load resistance is
/?^ = 10 k n . Our goal is to find approximate values o f
(a)
leakage
signal source
(b)
FIG URE 16.14 Approximate analysis of a high-Q circuit, (a) Original circuit,
(b) Approximate parallel /?Z,C circuit.
Solution
Step 1. Compute cOq.
= 31,623 rad /s
COq =
V 0 .0 2 x 0 .0 5 x 10"^
83R
Step 2. Find the parallel equivalent circuit values for the practical inductor. Since Q^(cOq) = 40, from
equations 16.31 and column 3 o f Table 16.1,
L r-L
Step 3. Represent the capacitor model by a parallel RC. First, the dissipation fiictor tells us that
^ ^
^^
clcCo^C
^ ------------ ^ 0 ---------- ^ ^
o jj:
3 1 6 2 3 x 0 .0 5 x 1 0 " ^
Step 4. Replace the practical source with its Norton eqidvalent and compute the eqtdvalent parallel
resistance, denoted as R. Replacing the practical source with its Norton equivalent and incorporat
ing the results o f steps 2 and 3 produces the network o f Figure 1 6 .14b. The parallel combination
o f R^, R^, ^leakage* and R^ is (using the notation // to indicate the parallel combination)
R = 40,000//25,298//63,247//10,000 = 5545 n
Step 5. Approximate analysis o f the circuit o f Figure 16.14a using Figure 16.14b. Our formulas from
the ideal parallel RLC case now apply, but the results are, of course, approximate for the circuit o f
Figure 16.14a:
0)^^^ =
1
RC
co^ = 3 1 ,6 2 3 rad/sec
=
------- = 3 ,6 0 7 rad /s ,
5 , 5 4 5 x 0 . 0 5 x 1 0 ^
3,607
and since Q^.-^ is sufficiently large.
- - ^ = 3 1 ,6 2 3 - 1 ,8 0 3 .5 = 2 9 ,8 1 9 rad/s,
0 )2
+ - ^ = 3 1 ,6 2 3 + 1,803.5 = 3 3 ,4 2 6 rad / s
Finally,
^max =
R,CB,,
839
Dictionary, in the field o f engineering, resonance refers to the phenomenon o f a vibration o f large
amplitude in a mechanical or electrical system caused by a relatively small periodic stimulus o f the
same or nearly the same period as the natural vibration period o f the system. In this section, we
shall investigate this notion o f resonance and its manifestation in RLC circuits, in both cases uti
lizing the theories studied in previous chapters. The main applications o f resonant circuits are for
the filtering and tuning purposes. Additionally, resonant circuits can be used to transform a resist
ance from one value to another value (at a single frequency) to achieve maximum power transfer.
This matching application is discussed in the last subsection.
(16.34a)
(16.34b)
Mathematically, the periodic leg movements are modeled by the periodic impulse function
x(r) = /l [ 6 ( t )
+d{t-T)-^6{t-n)^
...]
(1 6 .35 )
where T = 2 7 t/ o )T h is choice coincides with the period o f the natural response o f the system
transfer function. The build-up in magnitude can be seen quickly through superposition. The
contribution to the output, sayj/(r), due to the first Impulse at / = 0 in equation 16.35 Is
\\){t)=^AKe
sln(f/j^//)//(/)
(16 .3 6)
Figure 1 6 .15a shows the form o f this equation. Note that successive positive peaks {Kq, aV^, a^V^,
a^Vf^, ...} decrease geometrically by a = e~^P^. In fact, the waveform o f each period replicates the
waveform o f the prior period scaled by the factor a.
From time invariance, the contribution to the output ^(/) due to the second impulse at r = 7'in
equation 16.35 is simply J'q(/ - 7), as shown in Figure 16.1 5b. Similarly, for the third input
impulse zi t =1T, the response 1s7q(^- 2T), as shown In Figure 1 6 .15c.
8-10
From superposition, the response over [0, 3T) \s simply the sum
i.e., for 0 < r < 37'. This sum is illustrated in Figure I6.15cl.
a^)
F IG U R E 1 6 .1 5
In Figure l6.15cl \vc observe thar successive positive peaks arc given by
(1 +
(1 + a +
= 1
rt-1
11 + ^/+ t/ H------ ~
(16.37)
l- "
k=0
< 1 since
\-a
1-
This means that the waveform reaches a steady-state periodic response in which the positive peaks
have value
v
\-a
Exercises. 1. Consider the slightly damped second-order system with transfer function
5 +
(a) Find the first five peaks for the input o f the previous exercise.
(b) Is the system stable? Why or why not?
ANSWHRS; (a) 4, 8, 12, 16, 20. (b) The system is unstable bccausc there is a pole on the imagi
nary axis of the complex plane.
Another interesting application o f resonance is in product security in stores. The securit)' tag is an
RLC circuit with very small R. The circuit is excited by a radio wave at its resonant frequenc}' by
the security panels in front o f the exit doors. If the circuit has not been destroyed at the checkout
counter, it begins to resonate as one approaches the exit and transmits a signal back to a detecting
device at an amplitude much higher than the original transmitted signal. Fhis sets ofi an alarm.
The easiest way to understand this phenomenon is by way o f frequency response. Let us consider
the second-order transfer function
84 2
The magnitude frequency response is given in Figure 16.16. Notice the sharp peak at about 2
rad/sec, which is 10 times the dc gain o f about 1. Thus sinusoidal inputs at frequencies close to 2
rad/sec produce a steady-state response with magnitude almost 10 times larger. This is precisely
the type o f resonance phenomenon that occurs with the security tags at stores.
Frequency in rad/s
F IG U R E 16.16
/Y Y ^
L
R
V
(a)
:D
L
>R
^
(b)
F I G U R E 1 6 .1 7 Series (a) and parallel (b) reson an t circu its.
843
Using the frequency response approach o f Figure 16.16, we see that small inputs in these circuits can
produce large outputs when the conditions are right: (i) a high-Q circuit and (ii) input signal fre
quency and the peak frequency coincide. To develop a circuit-theoretic perspective on resonance we
take a more basic approach using the sinusoidal steady-state analysis ideas o f Chapter 10.
For any circuit containing one inductance L and one capacitance C, denote by (Oq the frequency
at which the two reactances
= -1/oJqC and
(16.38)
in which case
0)^ =
(16.39)
lc
In Chapter 9, on second-order RLC circuits, cOq = I I JT C is called the undamped natural fre
quency. The name stems from the fact that if all resistive elements are absent (i.e., the circuit is
undamped), then a parallel or series connection o f L and C produces a natural response o f the
form K cos((Oq^ + 0). In the jargon, the parallel LC circuit o f Figure 1 6 .17b is called a tank cir
and jXj^ =
0 and hence is equivalent to a short circuit. Similarly, if the elements L and C are connected in
parallel, then at (Oq the resulting admittance is zero and hence equivalent to an open circuit. These
properties are illustrated in Figure 16.18.
OZ(jO)o)
Z(jCOo)
O(a)
o-
O-
Y(jcOo)
Y(jcOo)
o-
O(b)
FIG URE 16.18 (a) Series- and (b) parallel-connected ZCelements with coq =
>/Zc
With the short circuit and open circuit equivalents sho\vn in Figure 16.18, we can easily deduce die fol
lowing properties for the series and parallel resonant circuits of Figure 16.17 operating a t : coq =
8 ''il
(1)
The magnitude o f the impedance seen by the voltage source in the series RLC o f Figure
16.17a is minimum, and the impedance is a pure resistance equal to R.
(2)
The magnitude o f the impedance seen by the current source in the parallel RLC o f Figure
(3)
The magnitude o f the voltage across C o r I in the series /?ZCof Figure 16.17a is Q^./;.(Wq)
The magnitude o f the current through Z, or C in the parallel RLC o f Figure 1 6 .17b is
O^.y^(ojQ) = cOq RC times the magnitude o f the source current.
> ()C
M )L
COqL
^ rO o )
Exercise. Derive property (4) using current division and property (2).
Properties (1) and (2) lead to a general definition o f resonant frequency, denoted to^.. Specifically
is the frequenq' at which the source sees an impedance or an admittance that is purely real, i.e., pure
ly resistive, despite the presence o f capacitors and inductors. For the series and parallel RLC circuits
oJ q.
Exercise. A sinusoidal voltage source at 1 MHz is applied to a series RLC circuit. If Z. = 300 uH,
/? = 5 n , and C is adjustable, what value o f C produces resonance?
A N SW ER: 84.4
p 1-
RL
10'
R, + Ri^
4 x 1 0 -^ + 10-^
0.2
The resonance condition co^ = (O^ = 1/V LC can be achieved by vary'ing one o f the three parameters
oJq, C, or L. When C or Z, or both arc adjusted to achieve resonance, the circuit is often called a tuned
circuit. The next example illustrates how this tuning can be used in a practical application.
84^
EXA M PLE 16.8. Figure 16.19 displays an amplifier model containing a V CC S with g^ = 2 mS
(milli-siemens) and
the capacitance accounts for such real-world phenomena as wiring capacitance, the device input
capacitance, and other embedded capacitances. This capacitance cannot be removed from the cir
cuit and often has deleterious effects on the amplifier performance.
FIGURE 16.19 Amplifier circuit illustrating the application of the tuned circuit conccpt to elimi
nate undesirable capacitive effccts.
The example objectives are
(a)
With the load connected directly as shown (without I ) , find the magnitude o f the out
put voltage.
(b)
If an inductance L is connected across the load to tune out the effect o f the capacitance,
find the value o f L and the resulting
M Hz is greatly increased.
So
lu t io n
(a)
3 p r = 3 9 7 . 8 Z - 8 8 .9 Q .
0 .00005 + y 2 jrx 10 X 4 0 x 1 0
Therefore, since jV'il =
^ 0-002 X 3 9 7 .8 = 0.0 7 9 6
Here the voltage gain is 0.0796/0.1 = 0.796 due to the low impedance o f C at the high operating
frequency.
(b)
By tuning out the effect o f the capacitance, this poor gain response can be eliminated.
The inductance needed to tune out the capacitance is calculatcd from equation 16.39:
4 ji
10
X 4 0 x 10
W ith a 6.33 uH inductor connected across the load, the parallel Z.C behaves like an open circuit
at 10 MHz and the load looks like a pure resistance o f 20 kQ to the amplifier. The new output
voltage magnitude is
846
I^out
1^ 0-002 X 2 0 ,0 0 0 = 4 V
Series-ParalelResonantCircuits
For parallel and series RLC circuits, co^ = cOq. For series-parallel circuits containing only one L and
one C when it exists to^ ^ (jd^j in general, but it may not exist at all. The next example illustrates
the point.
/YYY
L
Z(jo))
SO L U T IO N
Step 1. Calculate the admittance "looking into the input node pair o f the circuit o f Figure 16.20. By
the usual techniques.
(16.40)
y ? ^ + (a ),L r
Step 2. Set the imaginary part to zero. Resonance occurs when Kis real, i.e., when
\m{Y}= o ) ,C ----- ^ = 0 .
R~ -t- (oifL)
Solving for
0).
R-
1-
CR-
(16.41)
84
The rightmost term shows how the resonant frequenc)' is scaled away from the parallel or series
ideal cases where to^ = cOq.
Step 3. To obtain the values o f the admittance and impedance at resonance, substitute this value
o f (Of into equation 16.40 to obtain
Yij0^r) =
RC
RC
(16.42)
If {CR^)IL > 1, then there is no real solution for (0^ this means that the source voltage
and source current cannot be in phase at any frequency.
2.
If {CR~)IL < 1, then there is a unique nonzero resonant frequency O)^ that is strictly small
er than Wq.
3.
When O)^ exists, i.e., there is a real solution to equation 16.41, then at (O = CO^the source
sees a pure resistance, the value o f which equals {LIRQ and is greater than R.
AN SW ERS: (a) 0.6 rad/scc, 1.23 H; (b) (O^. docs not exist.
Results similar to equations 16.40 through 16.42 can be derived for the circuit o f Figure 16.21.
/YY\
Z(jco)
L
R
Exercises. 1. For the circuit o f Figure 16.21, let G = MR. Show that the input impedance is
./
G " + ( ojC )
coC
coZ-- ^------7
G~ + (coC)
1 - ^
(16.43)
CR'
Z(jo),) =
L
RC
( 1 6 .4 4 )
848
E XA M PLE 16.10. The output stage o f a certain radio transmitter is represented by a 1 MHz sinu
soidal voltage source having a fixed magnitude o f 50
shown in Figure 16.22a. A load resistance Rj models an antenna connected to the transmitter also
shown in the figure. The purpose of this example is to shov/ how a m atching network based on
the principle o f resonance can be designed to maximize the power delivered to the antenna.
(a)
(b)
(c)
Suppose that Rj^ is fixed at 20 O., but a coupling net%vork consisting o f LC elements is
absorbed
inserted between the source and load to increase the power P^ as shown in Figure 16.22b.
Choose values for L and C in the circuit so that {Pi),j,^j^ o f part (a) is again obtained.
1000
lOOO
/YYV
L
50 V
20 0
1MHz
IM Hz
(b)
(a)
FIG URE 16.22 Matching load to source using a resonant circuit, (a) Load connected direcdy to
source, (b) A coupling network designed to maximize the load power.
So
lu t io n
(a)
Ri = R^ = 100 n
and
2,500
AR,
(b)
400
= 6.25 W.
20
in which case
100 + 20
50 = 8.33 V,
3 .4 7 2 w .
20
(c)
S h9
If we can make the impedance at the input terminals o f the LC coupling network equal
to (100 + yO) LI, then maximum power will be drawn from the source. Since Z.Celements
consume zero average power, the same maximum power will be delivered to the load
resistance. The resonant circuit shown in Figure 16.22b provides a possible design. The
LCR" load circuit is the one analyzed in Example 16.9. Hence, to calculate the element
values, we use equations 16.41 and 16.42 as follows:
7
6\
1
20
(oT = 2 ji X 10 j -----------(from equation 16.41, squared)
LC
and
Z{jiD^)= 100 =
Should the source resistance be smaller than the fixed load resistance, C is moved to be
in parallel with
In that case, equations 16.43 and 16.44 are used in place o f equations
Exercise. Redesign the coupling network in Example 16.10 if the resistors are
= 300 Q and Rj
= 50 n .
AN SW ERS: 17.79 ul 1 .nui 1.18 nl-
A variation on the computation o f the resonant frequency is the design o f a circuit to achieve a
desired resonant frequency using a variable capacitor; this is the design that underlies the tuning
o f many AM radios.
E X A M PLE 16.11. Consider the series RLC o f Figure 16.23. Here, the voltage source has a fixed
magnitude |V^| and a fixedfrequency O). W ith R and L fixed, we seek the value o f the variable capac
itance C that maximizes the magnitude o f the voltage across the capacitor.
F I G U R E 1 6 .2 3 A d ju stin g
C fo r
m a x im u m o u tp u t voltage.
850
The first step is to compute the magnitude o f the voltage across the capacitor using vohage divi
sion:
I, , I
Vc
o
R -+ U)L-
1
coC/
\/
Maximizing |K j is equivalent to maximizing \V(^ or minimizing------- . To obtain this last expres
sion, we square the above expression for \V^ and rewrite it as
+ 2 ( lo ^ L C - \ W L
= 0
C=
R~ +(coL)
Zxo
+1
Although this value o f C produces a maximum capacitor voltage, the circuit is not hi resonance., as
the value o f 1/VLC is not equal to the signal source frequency o). However, given this equation
for C, for a high-Q circuit (ojZ//? > 8), the condition is practically the same as l/VTC = to.
and the corresponding value o f C for each o f the following cases: (a) Z, = 1 mH; (b)
L = 100 uH.
SO L U T IO N
(a)
For L = 1 mH,
given approximately by
C = - \ - = ^ ------ r = ( ) . l x l O ^ F ,
to -L
and I Vy
is given approxmiately by
l^ c U = Q Jn i = 2ov
0 .00 !
an d
x-6
.
^ = 0 .0 9 9 7 5 x 1 0 " F ,
5^ +(10^ X 0.001 r
n
J5 +
7 T
1 0 ^ x 0 .0 0 1 -
8S
10-' X 0 .0 9 9 7 5 X 10
10^ X 0 .0 9 9 7 5 X 10"^ J
Plainly, the approximate solutions are very close to the exact solutions.
(b)
,^ _ 6 ^
0 .8 x 1 0
F,
and
= 2 .2 3 6 V .
2
5^+
lO-'^ X 0 . 8 x 1 0 ^
iO-'^x 0 .0 0 0 1 1 0 ^ x 0 .8 x 1 0 " ^ /
Exercise. For part (b) o f Example 16.12, compute cOq = 1/ V lC . Is this value equal to the signal
frequency? Why or why not?
AN SW I:R. {.)(, = 1 .1 1 8 x 1 0 ^ rad/scc
u( A -
d(s)
s~ + 2 o pS + io~p
(16.45)
A reasonably sharp band-pass characteristic requires that H{s) have complex poles, i.e.,
<
O)^.
The finite zeros o f H{s), which are roots o f n{s) = 0, may be real or complex. The case o f complex
zeros, corresponding to more advanced filter characteristics such as the inverse Chebyshev or ellip
tic types, is beyond the scope o f this text. For practical reasons we focus on the case where H{s)
has one real zero or no zero.
When H{s) has one real zero and complex poles, then equation 16.45 reduces to
+
_
s- + 2 o ^ s + (o l
a\S + ao
+
"
SS2
A
p, = -^ p + i d
\p
s-plane
-a
(0
-e -
-a
Q =
P2 = -^ p -j d
FIG UIIE 16.24 Pole-zero plot of the transfer function H{s) of equation 16.46.
Equation 16.46 contains the usual qualit)' factor, Q , mathematically called the pole
defined as
CO,
(16.47a)
= arcsin
(16.47b)
\^P/
where the angle 0 is as shown in Figure 16.24. Relative to the pole-zero plot,
measures how
close the pole is to the/o-axis: a higher Q means a smaller 0, implying a pole closer to they'coaxis. As we already know,
is related to the circuit Q, i.e.,
the sharpness o f the response curve. For some special cases,
In Figure 16.24, another new quantity,
Chapter 15 that the impulse response o f a system characterized by equation 16.46 has the form
hit) =
r^
{/ / (.V )} =
Ke'""'" c o s ( w , y
d)
the bandwidth
Case 1. No finite zeros. When the transfer function o f equation 16.46 has no finite zero, then it
reduces to
H{s) =
K
(16.48)
.V" +
The pole-zero plot o f this H(s) differs from that o f Figure 16.24 only in that the single zero is now
absent. The series RLC circuit o f Figure 16.25a with the capacitor voltage as the output and the par
allel RLC o f Figure 16.25b with the inductor current as the output both fall into this category.
fYYVV.
(b)
(a)
and the value o f H{0). In fact, the transfer function can display
_________OOOO)-
V +
1000
VI
H{5) =
.v +
1.31
5 + (1 0 0 0 )-
then something between a low-pass and a band-pass characteristic results, as shown in Figure
16.26b. Here low frequencies are still passed, yet the characteristic has a selectivit)' propert}^ result
ing from the pole
o f 1.31. The ratio o f the maximum gain to dc gain is yfl . This means that
the peak is 3 dB above the dc gain. Now, with increasing pole Q the selectivity goes up, as does
the maximum gain, and the characteristic looks more and more like a pure band-pass. Finally, if
H{s) =
(1000)*-
1000
.v + -------- s
10
(1000)*
then the approximate band-pass characteristic o f Figure 16.26c results. Here the pole Q is 10 and
the ratio o f the maximum gain to dc gain is also 10. Although low frequencies are not attenuat-
854
ed, the characteristic is highly selective and frequencies near co^ are highly amplified, so that for
all practical purposes the characteristic is identified as band-pass.
Frequency in rad/sec
(a)
(b)
Frequency in rad/sec
(c)
FIGURK 16.26 (a) Low-pass characteristic, (b) Moderate Q^, resulting in characteristic exhibiting
both low-pass and band-pass behavior, (c) High-Q^ case, showing a marked band-pass characteristic.
8^5
and
At the peak value, the derivative o f |//(/o))| is zero. Setting the derivative of |//(/to)|
< B = 0)p
W
l - ^
^ P ^ co s(2 0 )
To compute
consider
'm
ct)p4sin (0 )c o s (0 )
Hence
K ____
^/=
2 .
in(2d)
H (0)
sin(20
0 6 .5 0 )
Similar derivations with more complex algebra yield the half-power frequencies,
<^1,2 = (Opyl cos(2d) T sin(20)
(16.51)
^(o = p ^ 2 (c o s ( 2 0 ) - 7 c o s ( 4 0 ) )
(16.52)
is greater than l/V 2 or, equivalendy, if 0 < 4 5 , the magnitude response of equation
16.48 starts fiom a nonzero value at co = 0, rises to the peak value at co = O)^, and finally decreas
es to zero as CO - 00. This behavior is illustrated in Figures 16.26b and c. If
is only slighdy
greater than 0.707, then the magnitude response is essentially that of the low-pass type, with a
small hump in the pass band.
For the high-(^ case ( ( ^ > 8), the magnitude response near co^ approximates that of a pure band
pass circuit. The preceding exact expression for (o^ reduces to O)^ a o)^ for the high-(^ case.
856
t^\Qp
s 2a^,
. and ojj 2 -
I'"*
error o f these
Exercise. For
= 8, (o^ = 1000 rad/sec, and A^= in equation 16.48, computc the exact and approx
imate values o f
and
\exact - approximate
X 100
\exacf\
AN SW liRS: Exact values arc
Approximate values are
- 8.015^.
= 8. Percentage errors are,
EXA M PLE 16.13. Consider the Sallen and K eyzcw e network o f Figure 16.27, which can be used
to realize the transfer function o f equation 16.48. As per a homework exercise in Chapter 14, the
transfer function is
1
H(,v) =
Ri R2C[C2
' nut
U ,c ,
R2 C J
R\R2<^iC2
(a)
Compute co^,
(b)
(c)
(d)
(e)
and
0)j, and co,.
FIGURE 16.27 Sallen and Key active net\vork for realizing the transfer
function o f equation 16.48.
So
8^'
lu t io n
(1 0 0 0 )2
'm
(a)
(b).
I --------V =
(JO,,, = o)p
11
Numerically evaluating N(s) at s =
5 + (1 0 0 0 )
10
= 10, and K =
= 10^.
9 9 7 .5 rad/.sec.
~^P
we obtain
n = wp^2;
d = [1 w p / Q p w p ^ 2 ];
Hm = abs(pol\^al(n,j*wm))/abs(pol)^al(d,j'wm))
Hm = 1.0013e+01
(c)
2 a ^
1 0 0 , 0 ),
co ^ -
= 9 0 0 , CO
2^
1 1 0 0 ,
all in rad/sec.
(d)
'lb obtain the magnitude response for the given transfer function, as shown in Figure
16.28, we use the following MATLAB code:
n = wp^2;
d = [1 w p / Q p w p '^ 2];
f = logspace(l ,3,600);
w = 2*pi*f;
h = freqs(n,d,w);
semilogx(f,abs(h))
grid
xlabel(Frequency in Hz)
ylabel(Magnitude H(jw))
858
Frequency in Hz
FIG URE 16.28 Magnitude response o f high-Q^ active circuit of Figure 16.27.
(e) A SPIC E simulation yields the corresponding plot in Figure 16.29.
+10.000 -H 5.849
Frequency(Hz)
8S9
A"of equa
Find approximate values o f B^, Wp and Wj.
and
= 10, K = I0 \
= 10^ rad/sec,
10-^
0,
i=- 0, and
same sign, a zero is present in the left half-plane. A second form o f the band-pass transfer func
tion o f equation 16.16 is
s +a
H{s) = K
=K
CO,
s +
Qn
s+ a
s~ + 2 0 p.v + oj
(16.53)
Figure 16.32 sketches the pole-zero plot for this transfer function. Intuitively speaking, the closer
the zero is to the origin, the more the magnitude response resembles the response o f the case with
a single zero at the origin. For the transfer function o f equation 16.53, derivations o f
of
values
and the zero-phase-shift frequency are possible. (See the homework problems.) The results
are
+ J(o )l + a -y -(2 o^ ay ,
to.., =
(16.54)
- 2o^,a
(16.55)
No exact expressions are available for the half-power frequencies and the bandwidth
case o f high
and a
For the
for
>8
(16.56)
A.
9
(
+j2oxj^^
jo3 + a
K
IS.
-)
(yco) + j 2 o i( j j , + co
co
(16.57a)
>
1+
/ c o - - c o '\
(16.57b)
2 o + ./
03
The second factor is approximately 1 for frequencies near co^ by virtue o f our assumption a OJ^.
Hence the properties o f the magnitude response reduce approximately to those o f the first factor
in equation 16.57b. These properties are those o f a single zero at the origin. Hence for high-Q^^
circuits, the relations o f equation 16.56 approximate the single zero at the origin case.
(S60
1^1
> 8 and a to , H
?Q
When exact values are desired, one must resort to SPIC E or MATLAB to obtain the frequency
response from which values can be graphically determined. SPIC E has the advantage o f not hav
ing to compute the transfer function o f the circuit; MATLAB requires this computation.
7. SUMMARY
This chapter began with a study o f simple series and parallel
its generality, we set forth a transfer function approach to the analysis and design o f these band
pass circuits. Specifically, we first investigated a second-order transfer function with a single zero
at the origin for which
Qp
'
For this transfer function we derived various formulas for determining band-pass parameters, such
as the peak frequency,
IG p
to,
a~p
These formulas have a special form for the parallel RLO.
to,,, = 0)^ =
4 lc
which is also the resonant frequency of the associated impedance; the bandwidth is found to be
^ (0 = t0 2 -0 J, = - ^
while the circuit Q is
0 =
= u . R C = ^
The resonance phenomenon of a second-order RLC circuit was then set forth from a frequency
domain perspective. One application was the design o f a matching network that produces maxi
mum power transfer, at a single frequency, from a source with fixed internal resistance to a fixed
resistance load.
iSOl
After this we cook up the consideration o f more general second-order band-pass transfer func
tions for example, those with a single zero o ff the origin,
Qp
"
quencies.
Dam ped oscillation frequency, (0^ frequency given by the condition that if the transfer func
tion o f a second-order linear circuit has complex poles at j = -a ^ jio^, then the impulse
response has the form K cos(co^f + 0). The constant (jo^ is called the dayyiped namralfre
quency.
Half-power frequencies; see lower and upper half-power frequencies.
L C resonance frequency: frequenc)' at which the reactances o f L and Chave the same magnitude;
equals M-JLC rad/sec. (Sec also tank frequency.)
Lower half-power frequency, COj: the radian frequency below the center frequency at which the
magnitude response is 0.707 times the maximum value.
M atching network: an LC net\vork that transforms a resistance
862
Chapter 16
-----------------------------------------------------------------------------------------------------------------------
Resonance frequency, 00^ the unique radian frequenq^ at which the input impedance of a twoterminal linear circuit becomes purely resistive.
Selectivity of a band-pass circuit: The circuit Q, defined as the ratio of the center frequency to
the bandwidth. A higher Q corresponds to better selectivity.
Susceptance: in sinusoidal steady-state analysis, the imaginary part of an admittance. For C, the
susceptance is coC For Z, the susceptance is -l/(coZ).
Tank circuit: the parallel connection of an inductor and a capacitor. In the idealized case (no
resistance), the total enei^ stored in a tank circuit remains constant, although there is a
continuous interchange of the energy stored in the various components.
Tank frequency, (Dq: defined as H j W in this text, regardless of the connection of the single L
and single C with other components in the circuit.
Tuned circuit: a second-order circuit containing one inductance and one capacitance, at least one
of which is adjustable to reach a condition of near resonance.
Undamped natural frequency: the natural fi-equency of a circuit consisting of lossless inductors
and capacitors. For the case of one inductor and one capacitor connected together, this
frequency is the same as the LC resonance frequency or the tank firequency and is equal
to m J l c .
Universal resonance curve: a normalized magnitude response curve of a band-pass transfer func
tion having one pair of complex poles and a single zero at the origin. The magnitude is
normalized with respect to the maximum gain, and the frequency is normalized with
respect to the center frequency.
Upper half-power frequency, (Oj: the radian frequency above the center frequency at which the
magnitude response is 0.707 times the maximum value.
r> .
n
o
* W e would like to thank Les Axelrod o f the Illinois Institute o f Technology for providing derivations that led to
this formula and those o f the bandwidth and half-power frequencies.
863
C H EC K ; poles are p, 2 =
'Problems
has
= V s 200 rad/sec,
(a)
(b)
= ^ p
i+ -
4G,-;
= 10, and
8 rad/sec.
0) 1.2 = a^; +
22,
tions.
from
, find a parallel
N(Jco) = K ^
T
i j o i ) + j l w o p+CO p
K
=
2 o -h y
^1
^2quency response using MATLAB and verify
(a)
your answers.
(b)
(c)
Find
(d)
K, H{s)y
Q,
C H EC K : /?= 0.125 n
= 100
rad/sec,
and C
Find the values o f R and L.
(b)
(c)
Find
(d)
to
= V S 2 rad/sec,
(a)
(b)
(d)
m(^)
circuit realization o f M{s).
Use SPIC E to obtain the magnitude
frequency response and compare it
with your answer in part (b).
/ = 10 mH,
S64
(a)
P I 6.10, w h e r e = 2 mS,
= 10 \dl.
(in
100 Q, and
(a)
(b)
Hz).
(c)
(c)
(d)
(0
|//(_/0))|
vs.
CO.
to an inde
to-,,
and (Oj.
Figure P I6.8
3 0 -3 0 0 , 4 0 -4 0 0
ANSWFR^.: (a) 0.20 at 1591 Hz; (b) 19.) Hz;
(c) 1601 M Hz. I 5 8 1.6 11/.; (d) 80; (0 1S.) 1 Hz
= 20 kH, and
broadcast band).
bandwidth is 20 kHz.
(b)
ZIjco)
> R
Figure P I6.9
AN SW ERS; 20 kLl. 2.438 mH, ^10.2 pF.
121.9 krad/sec. 8.2
10. 'Fhe equivalent circuit o f a radio frequenq-
Figure P 16.11
AN SW ERS; 3 0 -3 0 0 pF. 9 .'8 kLl, 7.1 6 kHz.
64.2 kHz
865
Z(s)
14.
band) as follows:
(i)
the circuit Q.
TYYY
R
Figure P i6.14
C H EC K : Q = 3.333
For
For
and C = 1 uF.
(a)
kHz
(b)
Find (o,,
(c)
C H E C K : 30 pF s C s 270 pF
(d)
(d)
(the corresponding
cast band?
(f)
and
and
quantities, in Hz, to
20 kHz.
(e)
Find
Q, and
6
Figure P i 6 .1 5
866
= 100
Q = 10, and^^ = 10
kHz.
(d)
(e)
(a)
(b)
(f)
find C and L.
C H EC K : (a) C = 1.59 uF
1 0 .1
krad/sec.
20.
Z. =
10
op amp is ideal.
(a)
niH, C =
0 .1
mF,
100
kQ,
and the
(a)
(b)
= -Zj(s)IZ,,{s) in terms
o f the circuit elements R-^^, Rp L, and
s ~ \-la p S + {H~p
Hz).
(b)
(c)
Q, and
= |M;co,)|.
Compute the value o f the half-power
(d)
(e)
(0
bandwidth
quencies
m
R.
CO,
(d)
(e)
Figure P i6.17
state.
Z = 0.25
= 3200 rad/sec.
(0
H , C = 0 . 5 u F , and/^2 = -'^20
C H EC K :
R,
867
C H EC K : Q = 100
(b)
(c)
ANSWER: (b) w, -
f r
V LC '
Figure P I6.23
24. For the circuit transfer function o f Problem
23,
''out
+
Figure P I6.21
s +^2^2/
(b)
Find
(c)
and Q.
^ 1^ 2Q ^ 2
show' that
(i) H. =
C,
Co
-hi
and (ii)
Q- ..c,
R'
^2
Hi
C,
cuit is undesirable.
arbitrary resistors.
(i)
86S
2G:
H{s) =
Model of Coil
(c)
Figure P i6.28
(a)
(b)
(c)
For
^(0'
Q-consider the case where the
and
CO,
maxi
H{s) =
C
J9 i\
\c)
(a)
= 1000 rad/sec
(a)
(b)
(c)
(d)
(e)
Suppose
= 2cos(0),^/) A. Find,
approximately,
(in volts) in steady
state.
y fiC '
869
20
(a)
(b)
(c)
(d)
CYY\
L
Coil
Figure P I6.29
= 3750 Q., L =
=2
Capacitor
C = 10 nF, and
cos(oj^^^r) mA.
Figure P I 6.33
C H EC K ; Q = 15 and
= 3 cos(co^^^^) V
L = 0.5 mH,
= 20
= 20
^
------V LC
(a)
(b)
(d)
(b)
-------
o jq
(c)
(c)
l i F.
(a)
= 20
c o s ((t) / ) \ '
R.
' 6
Figure P I6.34
C H EC K ;
100
Figure P I6.31
C H EC K ; 1 >
> 0.5
36. Consider the circuit of Figure P i 6.36.
and
= 80.
870
r>
(a)
Find
(b)
and
model at
oO
q.
(c)
series resistance by
1 series
Find the approximate circuit Q.
(d)
/\
(e)
Lossy Inductor
co^^,
cOpandcDj.
n ,0icap
-r I
n
^cotl^c
Verify that Qdr
Q coil
(c)
(d)
Qcap
, O)^,
q
+ Qcap
Lossy Capacitor
n
n
^coil^cap
~n
-I- n
WjcoiY ^cap
;c
Figure P I6.36
C H EC K : Q ,= 16
37. Again consider the circuit o f Figure PI 6 .36.
(a)
(b)
and high
at cOq,
to achieve the
desired
Q coilQ cap
(c) F
P16.38.
and high
Qcoil + Qcap
at (Oq,
*(d)
QcoilQcap
a
Q coil
Qcap
Lossy Inductor
C H EC K : (c) R^<SQ.
41. Repeat Problem40 under the conditions'^
= 1.6 MHz, a bandwidth o f 3 0 kHz, a lossy
capacitor with C = 100 pF, and
= 200.
QcoilQcap
Qdr =
.f /
^coil **^cap
Suppose you have been asked by your supervi
sor to use this circuit in a band-pass design with
= 1 MHz, a bandwidth o f 2 0 kHz, and a
8 71
desired
(c)
(d)
as given
in Figure P I 6.42.
(d)
^source
amplifier
Lossy Capacitor
coil
capacitor
Figure P I6.44
A N SW ERS: (a) 50; (b) 1 mH, 50 kQ; (c)
6 3 6 .6 Hz. 10, 15.91 kHz
:C
Figure P I6.42
C H E C K : (c) 130 k n > /?^ > 115 kQ
43. (a)
= 1.6
= 64
zeros.
44 Consider the amplifier circuit shown in
Figure P I 6.44.
Find Q o f the coil at to = HP rad/sec.
(a)
Represent the coil by a parallel RL cir
(b)
cuit that is valid for frequencies near
10^ rad/sec using column 3 o f Table
(c)
(b)
Determine
16.2.
bandwidth.
8 " 2
(cl)
______
10.000 rad/scc.
H).
)S0.0()0
coii
0.01 H
A N SW ERS: (b)
(i^ s
0.1 sin(oj,-r) V
750
10F
3000
SOOkO
48.
capacitor
Figure P I6.46
Z(j(o)
(a)
TY"YV
R
Z(jto)
(b)
Figure P i6.48
AN SW ERS: (a) 2.S niH; (b) 2.5 mi l
(b)
(c)
Q,
to,, and co,.
At resonance, fmd the approximate
(a)
the
input
voltage
is
10
sin(to^r)(r).
(d)
series.)
(b)
inductor
^
capacitor
sin(co^).
+
Z(jw)
(a)
fyy\L
(!)
i (t)
Z(j(o)
(b)
Figure P I6.49
amplifier
has a mag
load
Figure P i6.51
52. Find the resonant frequency (O^., in rad/sec,
and Z{jio) o f the circuit in Figure P i6.52 for R
= 2.8
lie '
Now compute
the resulting
zero.
(t)
ly Increased.
1 . 2 5 c d ^.
Figure P I6.52
AN SW ERS: 480 rad/sec, 35.71 LI
Figure P I6.50
is not.
(a)
(i)
If an inductance L is connected as
(ii) /^^ = 6 0 2
Rj =
80
LI
and
874
(b)
Now suppose
= 250
with co =
/Y Y V
part (a).
0.1 H
125Q
/Y Y V
Figure P i6.55
V cos(o)t)
)
Z(jw)
IO mF
in rad/sec,
Figure P I6.53
54. For the circuit o f Figure P I 6.54, R = 800
y jlc '
Now compute Z-^^{s) and plot the magnitude
and phase responses from 0.5w^ to 2to^. Verify
that at (0^, the phase angle o f
/Y Y V
is zero.
VJj(o)
Z(j(o)
Figure P i6.56
57. Consider the circuit o f Figure Pi 6.57 for R^
= 80 a , R^ = 125 n , C = 2 uF, and Z. = 20 mH.
The LC part o f the circuit can be thought o f as
a matching network when the values are prop
erly chosen.
(a)
Figure P I6.54
ANSWHR: Case I: 201)0 rad/sec, ^ cos(2000/)
V, 12.5 m\V
Find
the
resonant
frequency
CD^,
Now suppose
ered to
more power to
circuit
ues o f L and C?
o f Figure P i 6.55.
(a)
Is it possible to deliver
:..................... -
^
- J
LC
I '------L
Z(jco)
c
<
CR-
and
Z,(jw)
Z{j(0,.) =
Figure P i6.57
CR
58. A two-terminal network has input imped
ance
s~ + 4 s+ S
= 1, find co^ and
(a)
If
(b)
quenc)'.
= 1 0 0 V 2 cos(27i x lO^r) V.
a <1
CO
/Y Y \
son
v jt )
300 0
ofr.,(/-30).
Figure P i6.61
OY(s)
Figure P i6.59
is maximized. What is
^large'
^ should be connected
between the rsvo top terminals and
(b)
large
Rsmall
and
Figure P I6.60
A N SW l-RS: 17.79 uH, 1.186 pF, 8.3.^33 W
C=
(oRla/ge
Rsmall
8:^6
BAND-PASS TRANSFER
FUNCTIONS WITH NO ZEROS
OR A SINGLE ZERO OFF THE
ORIGIN
Figure P i6.62
63. The purpose o f this problem is to show that
the resonant frequenc)' (0^ depends on the
His).
(b)
(c)
Compute
the
A and B
(b)
B and C
mum gain.
(c)
A and C
(d)
40 n
10Q
Figure P I6.63
ANSWURS: 2S, I2.5\/3 = 2 1 . 2 8 . 8 6 . all in
rad/sfc
64.
The
circuit
= l/ 'Tl C
of
Figure
Pi 6.64
has
(a)
E and F
D and E
His).
(b)
(c)
With
Figure P I6.65
100
CO,,
and 0) 2-
. 0.8 H
2m F
50 0 /
Figure PI 6.64
F igu re P I 6 .6 6
H(s)=
f
UiC
____1
H{s) =
LC
]
+
/?2Q/
s+
=
(b)
= 50
kfl.
1 --
(c)
1-
0)j,
4lc\
and (Oj.
R^C^+s
\
2LC
c2 +
4 .
(e)
2LC
(b)
(d)
LC
/?l/?2C|C2
S +
/YY\
o
+
Figure P I6.69
70. Consider a transfer function with a single
zero off the origin, i.e.,
Figure P I6.67
s+2OpS+(0p
H{s) = K - 2
------j -
(a)
o ) =
(c)
m
Find the exact values o f
0)o.
(d)
o)j, and
independent variable.
(b)
set
878
P 16.71.
fer function
(a)
\/
s+
H {s) = Z;As) =
L}
[X pC
+ ''S s + 11+
L
\
LC
D \ 1
Rf,C
(b)
R.
s+
L)
(a)
(b)
Compute
0).,, =
(c)
1+-
R r
Coil
function.
sys=TF(n,d)
fer function
impulsc(sys)
pause
.
s '-
(a)
step(sys)
L}
H {s) = Z U s) =
(
-I-
1_
AN 'SW FRS: wp = 2 .9 9 6 3 e+ 0 3 ,
s+
KjLC
(b)
and that a
Compute
to^.
(d)
s 100 kHz.
(a)
(b)
(c)
What is
8~9
Z. = 1 H, C = 1 F,
and Rs = 40 ^2.
(a)
Find the transfer function H{s) =
R, = 0.05 n
(b)
(c)
Find
O btain
/Y Y V
,0
R, = 50 O
L=1 H
C =1 F
R, = 0.05 n
your answers.
Figure P I6.76
R
MISCELLANEOUS
77. Consider the idealized (tank) circuit o f
' 6
Figure P i6.75
(a)
(b)
(c )
(d)
= \/L^s,
= i//?3,
His) =
hn
Figure P I6.77
.y-hO.05
0 .1 2 0 3 8 5 + 1.009
(b)
(c)
Find
(d)
(e)
Figu re P I 6 .7 8
880
(a)
= 1 F.
(a)
(b)
mations.
among
the
various
frequencies
Figure P i6.80
Show that
R ,C [
R\
AN SW FKS; (b)
= 1, co^ = 1.0392.
=
0 .9367, (-),. = 0.6, and
= 0 .9602, all in
L)
rad/sec
.V H------
.V-l- 1+
(b)
R .) LC
C = 1 F, show that:
= 1 rad/sec.
is
82.
(Experimental
of
Q)
Q.
(a)
^p ~
= 1.2165 rad/sec.
The natural damped frequency co^y =
(b)
0.995 rad/sec.
, and
co^^^,
Hints:
1. To find
0
measurement
(c)
cuit?
rhe impulse response for the circuit is
o f the form
therefore, h{t) s
(d)
Figu re P I 6 .7 9
Plot
sin(co/).
cos((i)r).
h{t)
Ae~^cos{iOt)
using
ss;
O '
Coil
Figure P i 6.82
frequency determining
resonant circuit
where
a =
and
to.
2(2
1-
1
4(2^
(b)
Figure P i6.84
co^).
(O^ es
882
r\
YijiO) =
Find expressions for the conductance
the susceptance
in terms o f
and
and
r\
so that
r\
in terms of G^ and
condition.
>G.
iBp
JX.
r>
-------- J
o
(b)
(a)
jCsO)
andy5^ = jC^fa. Find formulas for
and R^ in
C H A P T E R
Magnetically Coupled Circuits
and Transformers
WHAT IS INSIDE THE AC ADAPTOR?
Most electronic equipment operates with dc power sources. For portable equipment, such as a
cordless phone and a cordless electric drill, batteries supply the dc power. Using non-rechargeable
batteries becomes expensive. Furthermore, replacing batteries in special equipment is a task not
easily handled by ordinary consumers. These two factors have prompted manufacturers to install
rechargeable lithium-ion batteries in portable equipment. By connecting several batteries in series,
the available dc voltage may range from 1.5 to 12 V. Whenever the battery runs low, it must be
recharged.
Recharging a battery requires a low dc voltage source (1.512V). An adaptor houses a device called
a transformer that changes the 110 V ac voltage at the household outlet to a much lower ac volt
age. The lower ac voltage is then rectified to become a dc voltage that charges the battery. Some
adaptors contain the transformer only, while others may also contain the rectifier circuit.
Typical specifications appearing on the casing o f an adaptor may be as follows:
model: AC9131
model: KX-AIO
input:
ac 1 2 0 V ,6 0 H z,6 W
input:
output:
output:
ac 3.3 V. 500 mA
ac 1 2 0 V ,6 0 H z,5 W
dc 12V, 100m A
The concepts and methods developed in this chapter will allow us to understand how a trans
former works to change the ac voltage level and also to perform some other important functions
in electronic equipment.
8iS4
CHAPTER OBJECTIVES
1.
and Z ,,
accounts for an induced voltage in each inductor due to the change o f current in the
other inductor.
2.
Develop a systematic method for writing time domain and frequency domain equations
for circuits containing mutual inductances.
3.
Understand why the mutual inductance is less than or equal to the geometric mean of
the individual self-inductances using an energy perspective.
4.
Expand the repertoire o f basic circuit elements to include ideal transformers, and learn
how to analyze circuits containing ideal transformers.
5.
Learn how to model a pair o f coupled inductors by an ideal transformer and at most two
self-inductances.
6.
Learn how a practical transformer can be modeled by an ideal transformer and some
additional RL elements.
7.
SECTION HEADINGS
1.
Introduction
2.
3.
4.
5.
6.
7.
8.
9.
10.
Ideal Transformers
Models for Practical Transformers
Coupled Inductors Modeled with an Ideal Transformer
Summary
11.
12.
1. INTRODUCTION
You may recall from your high school or grade school science class that if iron filings are sprinkled
on a piece o|- paper and a magnet is moved around beneath the paper, the iron filings move in con
cert with the magnet because the magnetic field induces a force on the iron filings. Similar to the
magnet and the iron filings, a changing current in one coil that is ver)' close to another coil induces
a voltage across the terminals o f the other coil.
Figure 17.1a shows rwo unconnected coils o f wire in close proximity. Figure 17.1b shows two
unconnected wire coils wound around a single ferromagnetic core, hi both cases, a voltage source
excites coil 1 while coil 2 is left open-circuited. Experimental evidence shows that a change in the
current z, generates a voltage Vj, called the induced voltage, across the open circuit; the induced
voltage is proportional to the rate o f change of/ j. Each pair o f coils in Figures 17.1a and 17.1b is
said to be magnetically coupled.
Coll 1
+
V ,( t )
Coil 2
(a)
(b )
FIGURE 17.1 Induced voltage in coupled coils, (a) Two coils in close proximity, (b) Two coils
wound on the same ferromagnetic core.
How does one quantitatively account for magnetic coupling? The strategv' is to introduce a new
circuit quantity called mutual inductance for coupled coils; specifically, similar to the v-i rela
tionship o f a single coil, the induced voltage satisfies the equation
V-7 = M 21
where
cli^
dt
(17.1)
> 0 is the proportionality constant callcd the mutual inductance from coil 1 to coil
2, and the sign, here , depends on the relative winding directions o f the coils. Dot markings indi
cate the relative winding directions. With reference to Figure 17.1, a dot is placed on coil 1 for
reference; if the dot on coil 2 is in position A, the sign on equation 17.1 is +, and if the dot is in
position B, the sign is - . A description of the general dot convention is presented in the next sec
tion.
The situation illustrated in Figure 17.1b is motivated by an extremely important magnetically
coupled device called a transform er, which is used to transform voltages and currents from one
level to another. In electric power systems, transformers are used to step up ac voltages from 10
kV at a generating station to over 240 kV for the purpose o f transmitting electric power efficiently
over long distances. At a customers site, such as a home, transformers step these high voltage lev
els down to 220 V or 110 V for safe, everyday uses. In addition, transformers have numerous uses
in electronic systems, including (1) stepping ac voltages up or down, (2) isolating parts o f a cir
cuit from dc voltages, and (3) providing impedance level changes to achieve maximum power
transfer berween devices, and tuning circuits to achieve a resonant behavior at a particular frequenc)'. After the basic analysis methods are set forth, some examples will illustrate these uses.
886
\>2{t) = M2\
du{t)
dt
as set forth in equation 17.1. Note, however, that coil 1 with inductance Z,, continues to act as an
inductor for which
v,(/) = Li
di\{t)
dt
There are two effects present in the circuits o f Figure 17.1: an induced effect and the usual v-i
relationship o f an inductor. Similarly, Figure 17.2 shows the reverse coupling to that o f Figure
17.1. For the circuit o f Figure 17.2, with the reference dot placed at the top o f coil 2,
v, ( 0
= M ,2
d ilit)
(17.2)
dt
where + would be used if the dot on coil 1 were in position A and - if in position B. As in figure
17.1, the dots indicate the relative directions o f the windings o f the two coils. Also as before, the
second coil continues to act as an inductor, for which
viit) Lo
d ilit)
dt
Secondary
Coil 2
v,(t)
(l )
FIGURE 17.2. Coupling from coil 2 to coil 1 (winding directions not explicidy shown).
As will be verified in a later section,
^ \ 2
>
(It
dt
887
(17.4a)
dt
(17.4b)
~ dt
Two questions remain: (i) When is the sign positive and when is the sign negative? (ii) How is the
value o f M determined (experimentally)?
Primary
+
i,(t) 0
Secondary
A
v,(t)
v,(t)
B
FIGURE 17.3. Coupled coils with current excitations present on primar}' (coil 1) and secondar}' (coil 2).
The following rule, identified with equation numbers, governs the choice o f sign lor the induced
voltage.
(17.5a)
the other coil, from the dotted terminal to the undotted terminal.
Or, equivalently,
The voltage drop across one coil, from the undotted terminal to
the dotted terminal, equals M times the derivative o f the current
(17.5b)
through the other coil, from the undotted terminal to the dotted terminal.
With reference to Figure 17.3, if the dot is in position A, all signs are positive, whereas if the dot
is in position B, the sign on M is negative. This rule gives the voltage drop due to the mutual
inductance. To obtain the total voltage drop across an inductor that is coupled to another, one
must add in the voltage drop induced by the self-inductance o f the individual coil, which depends
on whether the labeling is consistent with the passive sign convention.
8 S8
EXA M PLE 17.1. For the configurations o f Figure 17.4, determine the pair of equations that spec
ify the relationship betAveen the voltages and currents.
+
1
+
First we consider Figure 17.4a. The voltage z^,(r) and the current /, (r) as well as the voltage V2 U)
and the current ijit) satisfy the passive convention. For each coil acting alone,
cji^
(It
for k = 1 ,2 . However, the voltage induced in coil 1 by the current in coil 2 is negative relative to
the indicated polarity on v^{t) as per rule 17.5b, i.e., /-,(r) enters the dotted terminal so that-/'2W
can be viewed as entering the undotted terminal. Fience
Vt(/)= L |- + M ------ ^ = L i
(If
(it
' (li
(It
Using rule 17.5a, the same arguments apply to coil 2, in which case
~ (It
dt
~ (It
Now we consider Figure 17.4b. Here observe that neither pair (i^j, /,) nor {vj, h) satisfy the pas
sive sign convention; hence
Vk(t) = -L^
(iik
dt
for ^ = 1 ,2 . However, the voltage induced in coil 1 by the current /-,(/) satisfies rule 17.5a. Hence
dt
dt
On the other hand, the voltage induced in coil 2 by iAt) satisfies rule 17.5b. Hence
8S9
Exercise. For the configurations o f Figure 17.5, determine the pair o f equations that specify the
relationship between the voltages and currents.
M
+
V,
(a)
F IG U R E 17.5. Two more scenarios for setting up coil equations.
A N SW E R S: (a)
di]
vAr) = - L - +
di
and
di
(//
di-,
di
= -/V/ +
(If
^ ( b )
dt
,
, , di]
, di->
and v^(t) = - M - + 1.-,^^
"
dt
~ dt
EX A M P L E 17.2. This example presents the procedures for marking the dots on an unmarked
pair o f coupled inductors and for determining the value o f M. Consider the configuration o f
Figure 17.6, in which a current source is exciting terminal A ot coil 1 o f the coupled inductors
with unknown M.
DVM
Also marked
com, neg, low
increasing current
890
> q fo r all t.
clt
For example, one could set /,(/) = 10/(/) mA for 10 seconds. We know that /,(/) induces a volt
age at the terminals o f coil 2 according to
(It
where M > 0. If we put the leads o f a voltmeter across the terminals C -D as suggested in Figure
17.6, the reading will either be Vjit) > 0 or v^{t) < 0. Suppose the reading is V2 {t) < 0. If we reverse
the leads o f the voltmeter by putting them across the terminals D -C , the reading will have the
opposite sign, i.e., v-,{t) > 0.
Step 3. Reconnect the voltmeter leads until V2 {t) > 0, i.e., the reading is positive. Place a dot on the
terminal o f coil 2 for which the voltmeter lead is marked + {or "plus" or high or pos'), i.e., at the
terminal o f higher potential.
For the situation described above, the dot would be placed at terminal D. However, the dots on
coils 1 and 2 could be simultaneously moved to the opposite terminals o f each coil without chang
ing the relative information they convey. (Problem 1 confirms this statement.)
Part 2: Determining M
Again, \'2(0
v-){t) = M
implies that
(If
M=
^2(0
clidt)
dt
Compared with the ramp o f Example 17.2, a more practical input signal and measurement
scheme uses a triangular waveform, as displayed in Figure 17.7.
891
E X A M PLE 17.3. Figure 17.7 shows a circuit and two waveforms, /,(r) and V2 {t), as might be dis
played on an oscilloscope. Determine the placement o f a dot either at position C or at position
D, and the value o f the mutual inductance.
i,(A)
lu t io n
For the time interval 0 < r < 0.5 msec, /,(/) is a ramp function and Vjit) is constant. The infor
mation is similar to that given in Example 17.2 and the solution method is the same. First we place
the dot at terminal A as the current enters A. The current /, is increasing, and ^2 is positive. We
must now determine if the dot goes at terminal C or D. Since the current i^{t) enters the dotted
terminal and is increasing over 0 < r < 0.5 msec, its derivative is positive over 0 < r < 0.5 msec.
Also, the voltage iQ^t) is positive for 0 < r < 0.5 msec with the indicated polarities. Hence accord
ing to rule 17.5a, the dot goes at terminal C and
di^
dt
To determine M, consider that the measured values during 0 < r < 0.5 msec give Vj =
and
EXA M PLE 17.4. In the circuit o f Figure 17.7, suppose the dot positions are at A and C. If /jW
A, find Vjit).
= 2(1
So
lu t io n
dt
v9(/) = M = 0 .2 e ^'//(/) V
dt
Exercise, In the circuit o f Figure 17.7 with the dots in positions A and C, if /j(t) = 0.01
sin(1000?)(^) A, find v~,{i) tor r > 0.
ANSW ER:
= 0.0 Tcosi 1000.^)/,u) A
The preceding treatment o f the mutual inductance has not referred to the physical construction
o f the coils, although we have described procedures for measuring M when the coupled inductors
are assumed to be enclosed in a sealed box. However, for designing a pair ot coupled inductors, or
for a better understanding o f mutual inductance, one must relate the coil construction to the val
ues o f Z.J, Ljy and M. A rigorous study o f this problem requires a background in field theory and
magnetic circuits, which are covered in advanced texts or physics courses. Nevertheless, we set
forth here a few basic properties with reference to Figure 17.8.
M
Coil 1
Coil 2
(b)
FIGURE 17.8. (a) Coupled coils in close proximity, (b) Coils couplcd through magnetic core.
1.
In Figures 17.8a and b, the number o f turns for each o f coils 1 and 2, respectively, is TVj
and Nj. Then the self- and mutual inductances have approximately the ratio
L, :L 2 :M = n } \NI\N^N2
2.
If two coils/inductors are placed in a nonmagnetic medium (e.g., air), bringing the
inductors closer together increases the value o f M.
3.
If one inductor o f a pair is rotated, then a larger value o f M results when the axes o f the
inductors are parallel to each other. The smallest value o f M occurs when the axes are per
pendicular to each other.
4.
Changing the core on which the two inductors are wound from a nonmagnetic material
(e.g., air, plastic) to a ferromagnetic material may increase the values o f Z,j, L-,, and A/by
a factor o f several thousand.
As a final note, the development above presupposes linearity. If the rvvo inductors are placed in a
nonmagnetic medium, this holds true. If the inductors are coupled through a ferromagnetic medi
um (e.g., an iron core), then the linear relationships o f equation 17.4 hold only if both currents
are sufficiently small that the magnetic medium avoids saturation, a phenomenon discussed in
other courses or more advanced texts. Our investigations consider only the linear case.
dt
dt
dh
dh
V2(t) = M ^ + L 2 ^
dt
dt
(17.6a)
(17.6b)
V^is) =
Msl. is)
(17.7a)
(17.7b)
Equations 17.7 represent the x-domain model o f the coupled inductors. Further, if one is con
cerned w'ith the sinusoidal steady state, then replacing s by jui yields the following equations for
analysis using the phasor method:
894
(17.8a)
V j = yojz.ji, yojyv/i2
(17.8b)
V 2 = y to M j +
For equations 17.7 and 17.8, the plus sign is used when the secondar)' dot is in position A and
the negative sign otherwise. These three sets o f equations constitute the core o f the examples and
analyses presented in the remainder o f the chapter. Throughout, the j-domain method is the pre
ferred method, but each pair o f equations has its specific uses.
EXA M PLE 17 . 5 . A pair o f coupled inductors are connected in two different ways, as shown in
Figure 17.10. Find the input impedances,
inductances,
(b)
FIGURK 17.10 Impcdanccs and equivalent inductances of two series-connected inductors.
(a)
+ Z2 + 2 ^ - (b)
+ >^2 - 2 ^ -
S o lution
j and
^in (-^) =
(-^) +
^L2 (^) =
(^ 1 +
= {L ^ + L 2 + 2 M )sIi {s)
It follows that
im plying that
) I i n (^)
= I^{s)
^eq\ -
89S
+ ^2
= I^{s)
Similarly, for Figure 17.10b, with the same voltage definitions, we observe that again
= Ijis). Also note that /2W enters the undotted terminal. Applying equations 17.7, we obtain
y-m is) = ^ L \(-^) + ^ L 2 (-^ ) = (^1^- - M s ) I i { s ) + ( - M s + U s ) l i { s )
It follows that
Z,2(^>=
ImU)
= [L\ + h - 2 ) s =
implying that
Exercise. Suppose the circuits o f Figures 17.10a and b are connected in parallel. If Z,, = 40 mH,
Z,2 = 60 mH, and A/ = 25 mH, find the equivalent inductance o f the parallel connection.
AN SW ER: . r . 5 ml I
0 7 -9 )
This relationship suggests another way o f determining M and dot markings from measurements.
If an instrument for measuring self-inductance is available, we can use the instrument to measure
Z.^^j and L^^2} from equation 17.9, the difference is AM.
EXA M PLE 17.6. In the circuit o f Figure 17.11, assume that the meter resistance
is ver)' large
and looks like an open circuit to the secondary. The switch S is closed at r = 0. Find ij{t) and Vy{t).
8%
FIGURU 1 7 .11 A coLiplcd induccor circuit that might be lotind in a lab to determine dot position.
So
lu t io n
Since the secondary looks like an open circuit, /-> = 0 and K, = Msly Hence if we find /j we can
easily find K,.
Applying KYL ro the primary loop, using equation 17.7a and the fact that /, = 0 produces
sms+R)
.V
Therefore
R
/l(0 =
\ -e
Since
V/^ = yW.s7, = Ms
MZiq
_ iW/To/Li
s { L^s +R) ~ m s + R ) ~
It follows that
H(/)
h
Note that u-y(t) > 0 by virtue o f the way the meter leads were connected to the secondary. Hence
we conclude that the dot is indeed on the upper terminal of the secondary.
89'
at
= 1 msec is 3679 V. Find L^, M , and the ratio M : L^io see the
when switching with inductors occurs, a small voltage ma)' produce a dangerously iiigh voltage.
AN SW ERS: /., = 2 mi l. M = 2 H, and M : A, = 1000
An interesting application to older car ignition systems also uses an open-circuited secondary to
produce a very high voltage from a small one to fire a spark plug.
EXA M PLE 17.7. Figure 17.12a shows an automobile ignition system found on older cars while
figure 17.12b shows a simplified equivalent circuit model. Todays ignition systems use electronic
switching. Specifically, the block with the condensor (capacitor) and ignition point is replaced by
something referred to as an ignition module. The module contains a power transistor circuit to
perform the switching action electronically, with the trigger timing typically actuated by a sensor
that measures the position of the cam shaft cither optically or magnetically.
CONDENSOR
SPARK
PLUG
fL
IGNITION f
SWITCH
AL
i
IGNITION
POINT
+
---- n
IGNIT10N(f
COIL
o)
12V
DISTRIBUTOR
n---H I
t J
BATTERY
(a)
point
I mF
primary
to spark plugs
through a
distributor
secondary
(b)
F I G U R E 1 7 .1 2 (a) A n a u to m o b ile ig n itio n system , (b) S im p lified eq u ivalen t c ircu it m od el.
898
In Figure 17.12a, the ignition coil is a pair o f inductors wound on the same iron core, which cre
ates a strong coupling between the coils. The prim ary coil is connected to the battery, while the
secondary coil is connected to the spark plugs or load. The primary has a few hundred turns o f
heavy wire, the secondary about 20,000 turns o f ver)' fine wire. When the ignition point (or con
tact) opens by cam action, a voltage exceeding 20 ,0 0 0 V is induced across the secondary, causing
the spark plug to fire. The generation o f a high voltage to cause the spark plug to fire is accom
plished by a basic RLC circuit containing a switch that represents the point o f the ignition system.
Since the secondary is open-circuited, it has no effect on the solution for the primary current. Let
us do the analysis using the equivalent circuit model o f Figure 17.12b. Suppose that the switch
has been closed for a long time. Accordingly, at r = 0~, we have /j = ElR= 12 A. Using the model
for an initialized inductor given in Figure 14.18 results in the ^-domain equivalent circuit o f fig
ure 17.13.
------
t- L i
R+
sC
+ sL
(17.10)
.V + .9
R
Li
+L ,C
L~,
.v + 10()0i + 10^
(.V + 5 0 0 ) + 0 . 0 0 5 X 9 , 9 8 7 . 5
(.v + 5 0 0 )- + ( 9 ,9 8 7 .5 ^
(17.11)
cos(lO.OOOr)
(17.12a)
(1 7 .1 2 b )
899
Having obtained
we calculate V2 {t) from the basic relationship o f equation 17.6, using the
di^
(It
= 0.5 X 6
cos(10,000/) - 1
si n(10, 000/)
fourth o f a cycle o f the oscillations). This voltage is high enough to cause the spark plug to fire.
After the spark plug fires, the secondary is no longer an open circuit; the above methodology ceas
es to hold during the firing o f the spark plug.
AN SW ER. 0.036 1
In going from equation 17.12a to 17.12b, we neglected the second term in 17.12a, retaining only
the first. W ith the practical component values used in ignition circuits, this approximation is usu
ally valid. In terms o f equation 17.11, the approximation is as follows:
R \
S H------- I
H----------------
2 lJ
2L ,
2 R
S + ----- 5
+ ---------
L]
5 +
E
R
L,C
(17.13)
1 R
1
s~ + s + -----L,
L]C
/ j(/ ) ai
e
R
cos[o)jt)u (t)
where
and
^d =
\
z .,c
U tJ
I
^/z;c
')()()
/|(/) s COS
r \
Hit)
E
1
a A/----- f =
= QE
Z.|
(17.14)
Q=
is the quality factor o f the series RL\ C circuit. (See Chapter 16.) Later we will show that M!L\ is
approximately equal to the ratio o f the number o f turns o f the secondary to that o f the primary.
Equation 17.14 is a simple formula for estimating the maximum voltage that occurs at the sec
ondary. It shows that although the battery voltage is only 12 V, what appears at the secondary for
a brief moment is very much higher due to the switching action. The voltage is stepped up due to
two factors: (1) the Q o f the series RL^C circuit and (2) the turns ratio o f the ignition coil. From
equation 17.14, a smaller R produces a higher voltage across the secondar)\ But a small R causes
a larger current to flow in the primary circuit and therefore shortens the life o f the breaker point.
In practice, when the engine is running, a resistance wire or an actual resistor is placed in the pri
mary circuit to limit the amount o f current flow through the breaker point. The capacitor (con
denser) serves a similar purpose that o f protecting the breaker point by suppressing the arc that
results when the point opens.
Exercise. The ignition circuit o f Figure 17.12b has = 12 V, Z,j = 1 mH, and C = 0.01 [J.F. If the
total resistance in the primary circuit is /? = 8 Q and M/Zj = 50, estimate the maximum voltage
appearing at the open-circuited secondary when the s\vitch opens.
AN SW ER; Alniui l .s r k\'
')()!
M~s~
L'fS + Z-)
is) = Z, (.V) +
Z i
(iii) G ,2 = ^ =
L-)S + Zt
M~ s + L.\Z~)
VI
'
V\
(iv) the voltage gain Gj,| = =
Z;
Z.v
ZoM
An + 2 ]
lu t io n
'1 V /
Vi(^)
and solving.
Step 1. For the primary loop,
K, =
+ Msl^
902
In matrix notation,
L ji
Ms-
/r
My
L^s + Z 2
V,0
Step 2. Solve for /j and /j. Using the well-known formula for the inverse o f a 2 * 2 matrix, we
have
/r
L,.v
/2
A/s
-1 p
Mv
+ Z-)
L|.v(L2.v + Z 2 ) - M~s~
L-) s + Z-)
Ms
rv,
-M s
L^s
Hence
L 2 S + Z2
^1 ------------------------------- T T M
LiS{L2S + Z 2 )-M \ s-
W^.v-
L 2 S + Z2
(17.15a)
(17.15b)
/|
-M
L\(L2S
(17.16)
Z 2 ) - M~s
Part 3: Find C ^.2 = Using expression 17.16 and the fact that Vi = - ^2-^2
Vi
V2 _ -Z y l.
^v2 -
Ki
V,
conclude that
Z-yM
Z^M
m L2S + Z 2 ) - M - s
( l , L 2 - M ).v + L,Z2
(17.17)
V'l
Z,
z,- + z ,
Part 5: Find
(17.18)
- Gv^Gv^7.
G y = G ^ lG y 2 =
ZoM
(1 7 .1 9 )
903
-M v
Note that the answer can be obtained directly be writing a loop equation for the secondar\'.
The next example applies the foregoing development to a specific case and at the same time
expands the developed formulas to new cases.
, = 5 Q,
C = 0.01 F, L, = 2 H, I , = 0.5 H, and M = 0.5 H. Find (i) Z.(/10), (ii) V,(/IO), (iii) V.CylO),
(iv) Ij(/10), (v) l 2(/10), and (vi) the average power Pi{j\0) delivered to the C-R^ load.
M
V,
R,
-----------------------------+ 1
------------- ---------- +
lu t io n
First we note that co = 10 rad/sec. AJso, using the effective value for the phasor voltage o f the source
yields W = 60 Z. 0 " V. Thus all calculated voltage and current phasors are effective values.
Step 1. F
i n
Observe that
Z2(yiO ) =
CylO +
10
}_
R,
O.ly + 0.1
7 = 5 - 5 ./
1+ 7
Z,(ylO) = L|ylO- - ^
. . = Uj\{) + 2 0 M - = 5 + y2()
0.5 X ylO + 5 - ^ 7
904
= i ,S l o T 4 .
= 5 4 ^ jl2 ^ 2 1 .8 7 Z 1 2 .5 .r
ZyM
V, =
- M )./10+L|Z2
0 .5 ( 5 - 5 ./ )
- ( 5 4 + j\ 2)= 1 8 - 6 7 = 1 8 .9 7 / 1 -1 8 .4 4 ''
7.5y + 10 - IO7
u =
-M
-0.5
V, = ------- ^----- (54 + y l2 ) = - 2 . 4 - 1 . 2 7 = 2 .6 8 Z .- 1 5 3 .4 "
10
S ;
(L ,L 2 - M - ) 7 lO + L ,Z 2
1 0 - 92.57
I. =
^1
-"54+12./
Z,(ylO )
5 + y20
Step 6 . Fitid the average power delivered to the load. The complex power delivered to the load is
S(;10) = V.OlO) X [-I^l/lO)]* = (18 - 6;) X (2.4 - 1.2;) = 36 - ; 3 6
'Faking the real part yields an average power o f //(ylO) = 36 watts.
Sometimes the application o f the formulas developed in the previous rsvo examples, although
straightforward, is not the simplest route to the answer. The following example is a case in point.
EXA M PLE 17.10. Find the steady-state components o f z/j(r) and V2 {t) at the frequency 1 rad/sec
for the circuit o f Figure 17.16, in which ^^^^(t) = cos{t)u{t) V. Note that because a resistance is pres
ent, the circuit responses will contain both a transient and a steady-state component.
1Q
M = 2H
------ O
+
2F
V,
IH
5H -
-o
F I G U R E 1 7 .1 6 C irc u it fo r loo p analysis o f co u p led in d u ctor.
90S
So
lu t io n
Since only the sceady-stare responses are required, we use rlie pliasor method and write two loop
equations given that co =1 rad/sec and
= 1 Z. 0^^ V.
and
0 =
12 + ./5I 2 + y2I,
= y2I, + y4.5l2
|+y
jl
j l - *ll
y-4.5
I2
II
74.5
- 0 .5 + y'4.5
-j-
-P -
;4 .5 -
- 0 .5 + y4.5
.-j-.
It follows that
v,=v,,-i, =i-
J4.5
0. 5
- 0 .5 + ./4.5
0.5 - ./4.5
= 0 .1 1 0 4 3 Z .8 3 .6 6 "
-./9
_9
- 0 .5 + J4.5
1 - ./9
-j2
v, = - - L i , = _ _ x
jl
j2 - 0 .5 + 74.5
= 0.22086Z . - 96.34"'
Observe that
V , = -2 V ,
in which case
= 0 .2 2 0 8 6 c o s (r - 9 6 .3 4 ) V and
= 0 .1 1 0 4 3 cos(; + 8 3 .6 6 ) V
Exercise. In Example 17.7, write the simultaneous equations in matrix form and then solve by
Cramers rule. Which method is easier?
Example 17.10 is a drastic case contrived to bring up an underlying propert)': the dot markings
for coupled inductors (/.,, A,, M) determine the sign in the equation relating v-^ to eli^ldt and
the equation relating
conveyed by the dot convention. For most practical circuits, hoAvever, it is true that the voltage
drops o f coupled inductors from the dotted terminal to the undotted terminal are in phase or
nearly in phase.
906
As a practical example ro end this section, we analyze the circuitry typical o f the front end o f an
AM radio receiver.
EXA M PLE 17 .1 1 . Figure 17.17a shows circuitry typical o f the front end o f an AM radio. Figure
17.17b shows a simplified RLCM model o f this circuitr}'. Here,
= 300 ^2
together represent the antenna. Typical parameter values might be /?2 = Rj = 14,7 k fl, Z,j = 50
nH, L-, = 2450 nH, M = 350 nH, and C = 104.5 pF. Find the transfer function
H {s) =
rhen using the methods developed in Chapter 16, compute
|, Q,
coj, (JL)-,, and the maximum value and the bandwidth o f the magnitude response (i.e., the curve o f
vs. to).
antenna
RF
Amplifier
M
coupling
coils
tuning
1capacitor
SR
X
in
' 6
(b)
(a)
R = 300 Q
' 6
R = 14.7 kO
2
= 98nH
= 2612 pF
(c)
FIG U RE 17.17 Circuits for Example 17.11. (a) The original circuit, (b) A simplified circuit model
in which L^L-y = M~. (c) A design without couplcd inductors.
So
lu t io n
The first step in the solution is to observe that L^L-, = Afi. Then from equation 17.19, the trans
fer function is
Z-yM
V ,{s)
Z , + Z ,
(z,,L 2 - M ) j + L , Z 2
^m + ^l
^1
(1 7 .2 0 )
90
2 ,2
V
L2 S + Z 2
Loi + Z j
Substituting the expression for Z-^^{s) in equation 17.21 into equation 17.20, then dividing
through by Z j and simplifying, we obtain
CR^Lo
H (s) =
_y2 ^ / _ _ _ _ h ^ _ L _ \ ^ ^ _ _ L
\CRiL2
R2 C )
L2 C
4 .5568x10^ 9
5^ + 1 .3 0 2 0 X 10^^ + f 6 .2497 x 1 0 ^ ( 1 7 . 2 2 )
'
Ks
S
+ B^^.S + fO,
Ks
^2 ^
^^ ^ 2
Q
By inspection, then,
co,
= 6.2497 X 10"^ rad/s,
TU
=
P
and
= 3.5 , 0
= 4 8 ,'
XI.
w
(0, 2 =co, + 0.55(^ = 6 1 .8 5 x 1 0 ^ , 6 3 .1 5 x 1 0 ^ rad/s.
and
can be obtained
with the parallel resonant circuit o f Figure 17.17c. Following the design method described in
Example 16.3, we find the required element values to be Z, = 98 nH and C = 2612 pF. However,
the maximum voltage gain would have a much lower value, l^(;w,)| = 14,700/(14,700 + 300) =
0.98, compared with 3.5 for the coupled circuit o f Figure 17.17a.
The higher voltage gain achieved in Figure 17.17b can be explained by the concept o f nuiximum
power transfer. A routine analysis would reveal that at to = to^^, the input impedance seen by the
source is a pure resistance equal to the source resistance R^, i.e..
Z ,v ,0 , ) =
= 300 Q
extracted from the source. Since the inductors and the capacitor together form a lossless coupling
net\vork (see the homework problems), the same maximum power is transferred to the load resis
tor Rj. In Figure 17.17c, the reflected load impedance at/o^^^ is not 300
908
rhe T- and Tt-equivalent circuits char often allow the application o f series-parallel techniques to
simplify the analysis. With the common terminal the coupled inductors have only three accessible
terminals instead o f four. Such an arrangement is called a three-terminal device. The reader might
observe that if A/> Z, or M > L~,y one o f the three inductances in the T- or Tt-equivalent circuit may
have a negative value. This negative inductance appears in a mathematical model and is not the
inductance o f a physical component. Also, the equivalent circuits shown in Figures 17.18b and c
are for the specific dot locations indicated in Figure 1 7 .18a. A change o f one dot location in Figure
17.18a will result in a change in the sign in front of M in Figures 17.18b and c.
2
L ,- M
L ,- M
L^-M
A = L,L^-M2
(3)
(b)
(c)
FIGURE 17.18 (a) Coupled coils with a common terminal, (b) T-cquivalent circuit.
(c) 7t-equivalent circuit.
Verification o f the parameter values in the equivalent circuits in terms o l l j , Z.,, and M is left to the
homework problems. We now illustrate their use with a simple equivalent inductance example.
EXA M PLE 17.12. In the circuit o f Figure 17.19, all initial conditions are zero. If i-^(^t) = u { t - 1)
A, find the response,
0.5H
2H
10
IH
2.5H
3H
25+ 1
T h e re fo re
v,^it) =
~^h({t -\)
V.
= 2 H. T'hus
2.V -Hi
5 + 0.5
909
Exercises. 1. Repeat Example 17.12 with the dot on the 3 H inductor moved to the bottom.
AN SW ER: Same as in example.
2. Find the input impedance
v ^ l^ 2
2.V+ 0.5
3H
3H
0.50
o oFIGURE 17.20
A/j2 = A/t] = M by the principles o f magnetic circuits, but this is beyond the scope o f this text.
Our justification stems from the physical property that a pair o f stationary coupled coils cannot
generate average power. We will also show that the mutual inductance M has upper bound
> i.e., the mutual inductance can never exceed the geometric mean o f the self-inductances.
Justification o f Myj = ^ 2 1 ~ ^
This property is a consequence o f the principles o f electromagnetic field theory, which are beyond
the scope o f this text. To make our approach accessible to the beginning student, we build our jus
tification on the passivit)' principle for inductors.
10
, du
d ii
(17.23a)
~ dt
(17.23b)
~ dt
Let us apply /j = sin(/) and /, = cos(r) to the inductors. From equations 17.23, the terminal volt
ages are
dt
" dt
Vj = Li s in (/ )+ M p c o s ( / ) = Li c o s ( / ) - M p s in (0
and
Vo = yU^i s i n ( 0 + c o s ( 0
~ dt
~ dt
= M ^ i c o s i t ) - Losin(t)
The total instantaneous power delivered to the coupled inductors is the sum o f the powers deliv
ered to the inputs, i.e.,/>(r) = v^{t)i^{t) + V2 {t)i2 it)- Therefore
cos(f)sin(t) -
(17.24)
the average power delivered to the coupled inductors, we use the identities
sin(r) cos(f) = 0.5 sin(2f), sin^(r) = 0.5[1 - cos(2r)], and cos^(r) = 0.5[1 + cos(2/)]. It follows imme
diately that the first and the last terms in equation 17.24 make no contribution to
the terms involving
whereas
^o
=
(17.25)
12
violating the
W ith
= A/-7] = M, the average power, P^^^, is always zero for arbitrary sinusoidal excitations.
that is attainable once I j and L 2 are specified. Again we use the passivity principle and the fact
that stored energy is the integral o f the instantaneous power.
11
Consider the coupled inductors shown in Figure 17.9. The voltage-current relationships at the terminals are given by
^
v.(r) = L , 5 M &
(17.26a)
V2(/) = M ^ + t 2 ^
(17.26b)
and
^
with the upper sign (+) for the dot in position A and the lower sign ( -) for the dot in position B.
O '
Let us assume that the inductor currents are initially zero (at r = 0). In this state, there is no ener
gy stored in the system.
o
Recall that the energy stored by any device over a time period [0, 7] is the integral o f the instan
taneous power over the interval. For a pair o f coupled inductors with no initial stored energy, we
have
dt
o
= i L f (r )+ i z | (r ) j [ M (/,f + ,-,5 ),
o
^
4 L ,f ( r ) 4 i ,/ | ( r ) M j ;'3 '^ ( ,V ,)
2
2
J/,(0)/2(0)
= U ,i f ( T ) + ^ L ^ il ( T ) M i t { T M T )
O '
(17.27)
where ^(/|/2) is the total derivative o f the product /j /2 and is equal to /jc/f^ +
For the specific scenario above, we apply driving sources to the inductors to bring the currents up
VO
to /|(7) = /j and i2{1) = /j . Then, the energy delivered to the inductors during the time interval
(0, 7) is, by equation 17.27,
W (J) = ^ W \
^ hh
(17.28a)
The final integral in equation 17.27 depends only on the final values, which in the case
of equation 17.28a are 7j and /j . The exact waveforms o f /j(f) and t2 it) during
< t< T
are immaterial.
2.
The energy W{ T) delivered by the sources during 0 < f < T is not lost, but merely stored
in the system.
To grasp property 2, we may adjust the sources so that the currents are brought back from 7j and
/ j at / = r to zero at some t = T > T. Then the energy delivered by the sources to the inductors
during T < t < T may be calculated in a similar manner, to obtain
'
912
W iT') - \\\T) = - - L d }
7
(17.28b)
with the upper (-) sign for the dot in position A and the lower (+) sign For the dot in position B
in Figure 17.9. Equation 17.28b is precisely the negative ol equation 17.28a. Thus, all o f the ener
gy delivered by the sources during 0 < t < 7 has been returned to the sources during T ^ t ^ T\
For this reason, the energy given by equation 17.27 is called the stored energy. Another way of
recovering the stored energ)' is described in Problem 48. The physics o f the situation shows that
the energ)- is stored in the magnetic field produced by the currents in the inductors.
Upper Bound for M and the Coefficient o f Coupling
The energ)' W[T) must be nonnegative for arbitrar) values o f /, and /-,. Otherwise, the inductors
will ht g e n e r a t i n g during the time interval 0 < t< T , which would violate the passivity prin
ciple. To ensure a nonnegative W{T) for all /] and I 2 , the values o f Zp L j, and M must satisfy the
inequalit)'
(17.29a)
or
M ^ 7 ^ ,L 2
(17.29b)
W(T) = - l
where x = ^\l^2
U2J
[h )
L|.v
i
1 2M.X + L ^
(17.30)
term.
Consequendy, the curv'e ofy(.v) vs. a: will be a parabola opening upward. From analytic geometr)',
depending on the sign o f the discriminant D = {M~ - L^L-,), the curve may or may not intersect
they{.v) = 0 axis, as illustrated in Figure 17.21. From the figure, it is obvious that if D > 0, there
will be some current ratio that yields a negative^(a:) and hence a negative W{T), again violating
the passivity principle. Therefore, D = {M~ - L^L^) < 0, which yields equation 17.29.
f(x)
13
The degree co which M approaches its upper bound yJL\l^ is expressed by a positive number
called the coefficient o f coupling, defined as
k ^ - j= .
(17.31)
yjL\L2
= 20 H, A/ =
7) = I^= 4 A. Find:
1.
2.
S o lu tio n
1.
= 5 H,
= - 7- -------- =
0.8
V 5x20
2.
Stored energy =
= 0.5 X 5 X 2^ + 0.5 X 20 X 4^ + 8 X 2 X 4 = 234 joules
For the dot in position B,
Stored energ)' = 0.5/.j/| +
- A/Zj/j
= 16 H, A/ = 6 Fi, 7j = 4 A, and 7, = - 2 A.
914
EX A M PLE 17.1 4 . In the circuit o f Figure 17.22, /j(7) = /j= 6 A. Find the minimum value o f
the stored energy and the corresponding value o f /2(7) =
A.
M = 3H
9H
4H
FIGURE 17.22 Coupled inductors for calculating the stored energy in Example 17.14.
S o lu tio n
d l2
clh
= 18 - 36 + 72 = 54 J.
7. IDEAL TRANSFORMERS
Developing the Equations for an Ideal Transformer
Two coupled dijfereutial equations containing three parameters Z.j, Z<2, and M characterize the cou
pled coils of Figure 17.9. By imposing two idealized conditions on these parameters, the pair o f
differential equations can be approximated by a pair o f algebraic equations.
Idealization 1: The coupled inductors have unity coupling; i.e., M~ =
pling coefficient ^ = 1.
Effect o f Idealization 1: With ^ = 1, the coupled coils have the voltage transform ation
property
915
where a\ s 2i constant and both v-^ and V2 are the voltage dropsfrom dotted {position A) to
dt
dt
dt
dt
d t}
Although the idealization o f a unity coupling coefficient is not achievable in practice, coupling
coefficients near unity are achievable by winding the turns o f two inductors very closely together
so that n^ l y all the flux that Unks one coil also links the other coil. The constant a = L^/M =
M/L^ = V V Z ^ in equation 17.33 reduces to the ratio of the numbers of turns, denoted by N-^
and N 2 , of the two coils, i.e., a = //j/A^2* Thus a is called the turns ratio. When ^ = 1, it is possi
ble to show that
W ith unity coupled coils, equation 17.33 indicates that the voltages v^{t) and V2 {t), both from dot
ted {position A o f Figure 17.9) to undotted terminals, always have the same polarity. W ith coupling
less than unity, it is possible for v-^ and V2 to have opposite polarities at some time instants, as
shown in Example 17-10.
Idealization 2: In addition to unity coupling, the coupled coils have infinite mutual and
self-inductances.
Effect o f Idealizations 1 and 2: W ith k= \ and idealization 2, the pair o f coils has the cur
rent transformation property
i2 {t) = -ai^{t)
where both / j and
(17.34)
are the currents entering the dotted terminals of the coils, as per Figure
d(Ly.
dt\M ^
-------- ^/l + /9
(17.35a)
916
Reasonably assuming that the voltage v^it) is bounded, i.e., there is some finite constant
that |i'|(f)|s
for all t, then
lim
M-*oo
such
\M
V2
where if |y2(/)|^
for all t.
lim
L^-*o
In other words, as Zj <, Z2
^ (a ii + / 2 ) - 0
(17.36a)
ai^{t) + i2 {t) = C
(17.36b)
whose solution is
for some constant C. Now assume that the coils are unenergized prior to the application o f exci
tations; i.e., at some time f = Zq in the past, /j(Zq) =
cuit. It follows that the constant C in equation 17.36b is zero, and consequently /2(^) = -^^|(^)We remark that this derivation is valid for non-dc voltages and currents. The negative sign in equa
tion 17.33 implies that at any time, if a current enters one coil at the dotted terminal, then the
current in the other coil must leave the dotted terminal.
The condition o f infinite inductances (Z,p Lj, M) is another idealization that is not realizable in
practice. However, this condition can be approximatedhy using a magnetic material with very high
permeability as the common core for the two coils.
r\
917
V2 (0
N2
/tCO
Ni
FIG U RE 17.23 Symbol and defining equations for an ideal transformer, (a) Both currents enter
the dotted terminals, (b)
In Figure 17.23 two vertical bars serve as a reminder o f the presence o f a ferromagnetic core in the
physical device. The word ideal may or may not appear in the schematic diagram. Again, the
mathematical model o f an ideal transformer depends only on the turns ratio a : 1 and the relative
dot positions. To avoid the negative sign in the current relationship, an alternative labeling o f volt
ages and currents as shown in Figure 17.23b may be used. The subscript p stands for the primary
coil, which is connected to a power source, and s for the secondary coil, which is connected to a
load. Note that
tion o f Figure 17.23b is more commonly used in the study o f electric power flow.
One important simplification resulting from the idealizations is that an ideal transformer is char
acterized by two algebraic cc[\.\7iX.\ons in terms o f its terminal voltages and currents through a sin
gle parameter a, the turns ratio. This is to be contrasted with a pair o f coupled coils, characterized
by two differential equations containing three parameters, Z,p L-,, and M.
018
EXA M PLE 17 .1 5 . For the circuit o f Figure 17.24, suppose /?^ = 10 0.,a = 0.1,/ ?^ = 1000 Q, and
/j(r),
and /-,(/).
i,(t)
ijlt)
a: 1
V ,(0
a V 2 (/ )
= =
'l(')
2 V2 ( 0
2n
n n .
in n n
,n
-'2 (')
V](/) = ------ -
/?_, + /?,
It follows that
vAr)
10cos(300;rr)
10
viU) =
= -0 .1 cos(300;rr) A
919
V,(s)
V,(s)
ViCs)
Vj(.s)
a:1
<
+
Z,(s)
:
ideal
^ ^ ^ V^(s)
aVjis)
2
.
Z:,,{s) = =
= a ---= a ZAs)
/|Cv) J i i s )
-hAs)
^
c rz ^ is)
Vi(s)
Z,(>v) + Zi{s)
z,{s) + a -Z ^s )
V|(5)
in which case
Vsis)
Finally,
, , , ( , ) . Z2<) =
Zi_U)
Z,(.s) + a-Z^(s)
920
Z-^p) in Example 17.16 For the case where the dot on the
which the primary is resonant, i.e., the source sees a pure resistance.
ANSW ER.
a
J lc
Find/jW and/j(r).
=
^
The calculation o f
fer calculations) that we call it the im pedance transform ation property o f an ideal transformer.
Z Js ) = a^~Z^{s)
(17.37)
where a is the turns ratio taken in the direction from source to load. (See Figure 17.26; the
dots are not marked on the figure because their positions are immaterial for this application.)
921
10.V
Find
ANSW ER.
.s- + 10
E X A M PLE 17.17. For the circuit o f Figure 17.25, find the instantaneous power delivered by the
ideal transformer and the energ)' stored by the ideal transformer.
SO L U T IO N
With conversion o f the voltages and currents to the time domain, the power delivered to ideal
transformer is by definition
pit) = V j/|
i't
+ vo/^ = (c /i'-)) =- +
~ \
= 0
a}
Hence, considered as a single unit, an ideal transformer neither generates nor consumes instanta
neous power; whatever instantaneous power is received at one side must transfer to the other side.
Furthermore, since p{t) is identically zero, its integral with respect to t is also zero. Thus, an ideal
transformer cannot store any energy.
In summary, the instantaneous power dehvered to one side o f an ideal transformer is transferred
to whatever loads the other side, and as a result no energy is stored in the ideal transformer.
Exercises. 1. In Figure 17.26 suppose the primar)' o f the transformer has <7 = 0.1 and is connect
ed to a source having
= 10 Q. Find Zj{s) for maximum power transfer to the load.
ANSWHR. Z^U) = 1 kLl
2. In Figure 17.26 suppose the primary o f the transformer again has a = 0.1 and is connected to
a source having Z^^(/100) = 10 + J5 0 Cl. Find Z^iJlOO) for maximum power transfer to the load.
AN SW ER. Z^(/100) = 1000 -y '5000 LI
922
Divide rhe first equation by Ms and the second equation by L 2 S. To observe the frequenc)'-dependent behavior, set s =j(x) to obtain
Mjw
L 2 J (0
= ^ I \ i j ( o ) + I- , ( j co ) = a I i { j c o ) + 12 (j co )
Lj
Practical transformers have rsvo customary properties: (i) the frequency content o f the voltages and
currents are band limited, meaning that the transformer is guaranteed to operate only over a
restricted frequency range, 0 <
^ O) :S
and |K,(/a))| rs
^ co :s
I M,
so that
Viij(o)
Mco
0 and
V2 O )
Li(o
sO
E X A M PLE 17 .1 8 . In Figure 17.27 an ideal transformer steps down the voltage o f a 2400
source to pow'er 10 incandescent lamps in parallel, each drawing 0.5 A.
(a)
(b)
transformer
source
2400 V
rms
load
10
1760
turns
88
turns
10 incandescent lamps
represented by resistors
FIGURE 17.27 A transformer is used to step down a source voltage to meet the lamp specs.
S olution
Since only magnitudes are involved in this problem, dot positions on the transformer and refer
ence directions for voltages and currents arc immaterial. The turns ratio is a =1760/88, Since 2400
=
923
88
^load
1,760
2 .4 0 0 = 120 V,
Since 10 lamps use 0.5 A each, the total current used is 5 A. Using the current transformation
property o f the ideal transformer,
/.V
1,760
X 50 = 0.25 A
Exercise. In the circuit o f Figure 17.26, represent each lamp by a 200 H resistance. Suppose the
source is limited to delivering a maximum o f 3 kVA. How many lamps in parallel can be con
nected to the secondary o f the transformer?
ANSWER: 41
E X A M PLE 17.19. Figure 17.28 shows a simplified model o f an audio amplifier containing an
ideal transformer. The input voltage is 1
z 4 Q resistance.
(a)
fier (i.e., with the transformer removed and the resistor connected across A-B).
(b)
With the transformer and load connected as per Figure 17.28, with turns ratio a= 5, find
the average power delivered to the load.
(c)
If the turns ratio a is adjustable, what value allows maximum power transfer to the load?
W hat is the value o f the maximum power?
a:1
speaker
FIGURE 17.28 An ideal transformer used for maximum power transfer.
So
lu t io n
(a) No transformer and resistor connected across A-B. First observe that the magnitude o f the current
through the 4 Q resistor is
200
4Q
9 0 0 -^ 4
= 0 .2 2 1 2 A
) 2 i
100
pri
(200|V ,|)=20 V
900+100
(c) For the maximum power transfer, the turns ratio a should match the secondary impedance to
that o f the primary; i.e., the 4
to match
^ = 15
|= 100 V, meaning that
f A ^
= 6.667 V. Thus
--------= 1 1 .1 1 W.
Exercise. Repeat all parts o f Example 17.19 for a loudspeaker with a resistance o f 16 ohms.
ANSW ERS: i).-(,28 W.
W, 11.11 W
92 S
Because o f the ferromagnetic core used in its construction, a practical transformer is inherently a
nonlinear device. Nevertheless, a first level o f approximation represents a practical transformer, by
the circuit o f either Figure 17.29b or Figure 17.29c. Figure 17.29a is ideal, whereas Figure 17.29b
contains extra inductances to account for leakage fluxes and other magnetic phenomena. These
inductances give rise to a finite usable bandwidth for a practical transformer. Figure 17.29c shows
additional resistances to account for internal power losses. Thus the ideal transform in conjunc
tion with other circuit elements can be used to approximate a real transformer. Specifically, trans
formers do not work for dc or for very high-frequency signals. The circuit model o f Figure 17.29b
makes this behavior clear: the inductance
and
produce open circuits at high frequencies. A second level o f approximation would begin to
N ,:N ,
+
(a)
-I-
N.
V.
N,
N.
ideal
N ,:N ,
L /Y Y \
-h
(b)
V,
ideal
L.
+
V,
Ideal
F IG IJR F 17.29 Linear transformer models, (a) Ideal case, (b) Extra inductances to account for leak
age fluxes and other magnetic phenomena, (c) Extra resistances to account for internal power losses.
and
age inductances, respectively. They are due to that part o f the magnetic flux that links w'ith one
winding but not the other.
and
926
and
primary and secondary windings, respectively. R^^ accounts for the power loss in the iron core due
L^, and
can be
calculated from knowledge o f the physical layout and materials properties used in the construc
tion o f the transformer. This task usually requires a large set o f design formulas and (empirical)
charts. After a transformer is constructed, these parameters can be experimentally determined.
Books on ac machinery set forth such experimental techniques.
From a circuits perspective, once the model parameters in Figure 17.30 are known, the steadystate analysis o f a circuit containing practical transformers requires no more than the phasor
method or the Laplace transform method for its analysis. l\vo examples will now be given.
Example 17.20 shows the effect these parameters have on frequency response.
E XA M PLE 17.2 0 . Example 17.19 considered an audio amplifier with an ideal coupling trans
former. Figure 17.30 shows the same amplifier with a practical transformer in place o f the ideal
one. O ur objective is to illustrate the effect a practical transformer has on the power delivered to
the 4 Q. speaker as a function o f frequency. We assume our turns ratio ^ = 15, which produced
maximum power transfer in Example 17.19. We use the model o f Figure 17.29c with
(equivalent winding resistance), ^ ^ = 8 mH (effective leakage inductance),
and
= 40 Q
1 M ii (core loss),
= 1 H (magnetizing inductance). For simplicit)^ we set Z,^= /?^= 0 in the model o f Figure
17.29c.
Lp= 8mH
1 5 :1
40
Our analysis begins with the writing o f two loop equations in terms o f /,(;) and l 2 {s):
-20()K = [ , + /? . + L,,^ +
" M
"
] /, - - V V L / ,
R , ^ L , s }
R, + L,s -
= 1 H,
j
1 -I-
R,
92-7
- 2 0 W , .( R , .
For the second loop, again neglecting
we have
In matrix form
- 2 0 0 V/,,-
(/?! + Rp + {Lp +
LjS
-L,s
/l1
{R^q + L,s)
h =
Vin
+ (Ri + f<p)Req
Our purpose is to show how average power to the load varies as a function o f frequency in a prac
tical transformer. To this end, for v,(/) = yj2cos(cot) V, the power delivered to the load is the
power delivered to the primary o f the ideal transformer, which is
/> ,> ) = ,,|/2(/ )P
Recall that the effective value o f the input voltage is 1 V. To obtain the desired plot o f P iif) (dis
played in Figure 17.31), consider the following MATLAB code.
R1 = 900; Rp = 40; Lm = 1; Lp = 0.016;
Req = 900;
num = [-200 0];
den = [Lp*Lm ((Rl+Rp)*Lm +Req*(Lp+Lm )) (Rl+Rp)*R eq];
w = linspace(0,2*pi*le4,500);
I2 = freqs(num,den,w);
I2mag = abs(I2);
PL = Req*I2mag.^2;
plot(w/(2*pi),PL)
grid
xlabel(Frequency in Hz)
ylabel(Average Power to Load)
928
TD
TO
O
4-'
$
0
Q.
q;
cn
(Z
I
01
>
<
Frequency in Hz
(a)
"O
fD
OJ
5
Cl.
0)
cn
2
oj
<
Frequency in Hz
(b)
FIG URE 17.31 Plot of the power delivered to the speaker load as a function of frequency in Hz. (a)
Overall response, showing sharp decrease with increasing frequency, (b) Low-frequency response,
showing significant reduction in power transfer.
The plot shows that the power is down from the peak by about 0.5 W at 4 5 0 0 Hz, and, at 10
kHz, the power delivered to the load is down about 2.5 W. Also note that a practical transformer
does not operate at dc and at very low frequencies, as demonstrated in Figure 17.31b.
In the next example we consider voltage drops in commercial power lines using practical trans
formers.
)2 9
EX A M PLE 17.21. Fhe pniccical transformer in Figure 17.32a is designed for operation at 60 Hz
and 1100/220 V, i.e., ideally a 5:1 step-down transformer. Using the model o f Figure 17.29c, we
represent the circuit in Figure 17.32b with the following parameters:
winding resistance),
age reactance),
= 250 Q (mag
netization reactance). R^^^is very large and its effect is neglected. If the load draws 100 A at a power
factor o f 0.6 lagging, i.e., I 2 = 100 Z. -53.13^^ A, and the load voltage is V , = 220 Z. 0^ V, what
is the magnitude o f the needed source voltage to achieve the desired load voltage-current values in
the presence o f non-idealities o f the transformer?
ac
High Voltage
Source
Low
Voltage
Load
(a)
Load
5:1
ac
High Voltage
Source
ideal
(b)
FIGURE 17.32 Analysis of a circuit containing a practical transformer, (a) A practical transformer
for stepping down ac voltage, (b) A circuit model utilizing an ideal transformer.
So
lu t io n
R^ + jX^ = 0 .0 0 2
+ y'0.016 Q
930
1 107Z 0
= 2 3 .7 ^ -5 9 .5 4 x (0.05 + ;0 .4 ) = 9 .5 5 ^ 2 3 .3 4 V
Therefore, the source voltage is
9.55Z -23.34 + 1 107Z .0.207 = 1 1 15 .8 z l0 .4 V
Thus, the magnitude o f the source voltage is 1115.8 V, which is 15.8 V higher (to overcome the
non-ideal effects) than what would be needed for an ideal transformer.
931
as
shown in Figure 17.33a. Figures 17.33b and 17.33c show two equivalent circuits, each consisting
o f one inductor and one ideal transformer whose turns ratio depends on the values o f Z,j and Z,,.
k=1
1:
o+
+
V,
ideal
(b)
d
Zo .
Vj = L i\ + , h L\------ JL\L-t = L i -------- h M
'
' dt
cit
^
dt
di
dt
V
Next, using the voltage transformation propert)',
^+U = M + U
dt
dt
dt
These two equations are exacdy the same as equations 17.6. Hence, the circuits o f Figure 17.33a
and Figure 17.33b are equivalent. A similar derivation proves the equivalence o f the circuits o f
Figures 17.33a and 17.33c.
Exercises. 1. Verify the equivalence o f the circuits o f Figures 17.33a and 17.33c.
2. In Figures 17.34a and b, let Z, = 2.4 H, Z-2 = 6 mH, L j = 3 mH, and
= 0.2 H. Find
k=1
k=1
z in
^ --
>
L
(a)
(a)
FIGURE 17.34
E X A M PLE 17.2 2 . The circuit o f Figure 17.35a has a unity coupling coefficient. Find the band
width, the center frequency, and the maximum voltage gain.
300 Q
50 nH
M = 350nH
2450 nH
104.5 pF
(a)
soon
1 :7
14.7 kn
300 0
5 1 2 0 .5 pF
+
V
5 0 nH
V.
300 0
(0
FIGURE 17.35 Analysis of a unit)' coupled circuit, (a) A circuit containing unit)' coupled inductors,
(b) An equivalent circuit utilizing ideal transformers, (c) Equivalent circuit after reflecting load
impedances to primary.
So
lu t io n
The coefFicient o f coupling is = M l -y/^1^2 = 3501 yjSO x 2 ,4 5 0 = 1. Replacing the coupled coils
with the equivalent circuit o f Figure 17.33b yields the circuit o f Figure 17.35b. From the imped
ance transformation property o f an ideal transformer, looking into the primary we see an imped
ance 1/49 times the load impedance. Therefore, looking into the primary, we see a resistance o f
14,700/49 = 300 Q. in parallel with a capacitance o f 49 x 104.5 pF = 5120.5 pF.
Figure 17.35c captures the new equivalent circuit. Th e band-pass characteristics o f H {s) =
Vl(s)
Vsis)
follow the analysis done in Example 16.2. Consistent with the notation o f Example 16.2, we have
/?= 300//300 = 150 0 , 1 = 50 nH, C = 5,120.5 pF
CO,=Wp =
*
= 62.5 X 10^ rad/sec
V 5 0 x 10"^ X 5 ,1 2 0 .5 X 1 0 "*
Q _ ___ _ ___________
RC
150 X 5 ,1 2 0 .5 x l 0 '
\H(/
I
'io)\
'huav =
ID = 0 .5
From the voltage transformation propert)' o f an ideal transformer,
only quantities affected for the transfer function
)34
The above example illustrates how a simple substitution o f an equivalent model for coupled
inductors can reduce a circuit to a simpler form for which the analysis is straightforward.
Exercise. Solve Example 17.22 again using the equivalent circuit o f Figure 17.33c.
ANSW'HRS: I he same as given in Example 17.22, of course.
M :L
M :L
< o
+
k'L.
ideal
ideal
(b)
(c)
o
+ i.
O
+
L.:M
/YY\
< o
+
V,
k'L
ideal
ideal
(d)
(e)
FIG U RF 17.36 Four different models for coupled coils with /'< 1, cach consisting of one ideal
transformer and rwo self-inductances, (a) Coupled inductor with k < 1. (b) One model using an
ideal transformer and two inductances, (c), (d), (e) Three alternative models.
= 6 mH, L^ = 5 mFI,
= 1.6 H, and k =
j = .
93
(a)
(b)
FIGURE 17.37
it is possible to subtract
a small inductance from Z,j such that the remaining inductance, Z,'|, satisfies the condition L
= M".
In other words, the new inductance, L' ^ together with L-, and M, forms a unity coupling system. Since
Z,'j = M^IL) = k~L^, the induaance to be subtracted is equal to Zj - Z .'j = (1 -
This inductance
must be added back in series with Z,', to obtain a model for the original coupled inductors. The mod
els o f Figures 17.36b and c result. Repe-ating the process on
e. A total o f four equivalent circuits is possible. Each consists o f two uncoupled inductors and one ideal
transformer. Clearly, when k= 1, the models in Figure 17.36 reduce to those in Figure 17.33.
As an application o f the models o f Figure 17.36, let us reconsider Example 17.10.
and Vjit) at the frequency 1 rad/sec
ent, the circuit responses will contain both a transient and a steady-state component.
1Q
----------O ----------+
M = 2H
(a)
i
2 :5
+
V,
2F
(b)
F I G U R E 1 7 .3 8 Stead y -state analysis o f co u p led in d u cto rs u sing th e ideal tra n sfo rm er m od el.
S olu tion
,.2
For the circuit o f Figure 17.38a, k" = ------- = 0 .8 . Substituting the model o f Figure 17.36b
L| Li
for the coupled inductors in Figure 17.38a, we obtain the equivalent circuit o f Figure 17.38b.
From the impedance transformation property o f the ideal transformer,
It follows that
z . , , a ) = 0 .8 y , i z , ,, ,( ;) = 5
; j ^
= - 0 .0 8 8 9 ,Q
^ 0 .0 8 8 3 4 5 Z - 96.34'^ V
Further,
^
0.2y 7 0 .0 8 8 9
1 + 0.2./ - 7X).0889
^ ^ ^ I0 4 2 Z 8 3 .6 6 ^ V
SUMMARY
This chapter has examined the phenomenon o f induced voltage in one inductor caused by a
change o f current in another inductor. A new parameter called the tnutttal inductance (A/) between
the coils was introduced. M was defined as a constant and is present in the coupled differential
and 5-domain equation models o f the coupled inductors. As illustrated in the chapter, M can be
determined experimentally. This treatment has avoided digressing into the study o f magnetic cir
cuits, which in fact underlies a rigorous development o f the concepts in this chapter. From the cir
cuit analysis perspective, this mathematical treatment is adequate. However, for a deeper under
standing o f the physical phenomena, one must study the principles o f magnetic circuits.
O f foremost importance in analyzing a circuit containing mutual inductance is the formulation o f
correct time domain or frequency domain equations in particular, the correct signs for the
mutual terms. For this reason, considerable time was spent on the dot convention. Once the equa
tions are correctly written, we may use any o f the techniques (loop or node analysis) studied ear
lier to anal)'ze the circuit.
A proof o f M p =
= M was given that made use o f the passivit}' principle and some trigono
metric identities. With the establishment o f this equalit)', computing the energ)' stored in the cou
pled inductors follows from simple integration. The expression for the stored energ)' and the passiyitv principle then led to an upper bound for the value ot the mutual inductance, namely M <
l .
An ideal transformer was defined as a device satisfying both the voltage transformation and cur
rent transformation properties. For practical transformers, these two properties hold only approx
imately for example, over specified voltage and current ranges as well as over specified frequen
cy ranges.
Transformers have broad applicability, h'or example, in power engineering, transformers are used
to step up or step down the voltages. In communication engineering, transformers are used to
change a load impedance for the purpose o f maximum power transfer.
Although ideal transformers can only be approximated in the real world, they nevertheless remain
an important basic circuit element because a practical transformer or a pair o f coupled inductors
can be modeled by an ideal transformer and some additional R, L, and/or C elements. The use o f
such models simplifies many analysis problems and gives physical insight into the operation o f a
circuit.
sign for the case where both currents enter (or leave) dotted terminals and the sign for
the case where one current enters a dotted terminal and the other leaves the dotted ter
minal.
Ideal transformer: rwo network branches satisfying both the voltage transformation and current
transformation properties exactly.
Impedance transformation property: when the secondary o f an ideal transformer is terminated
in an impedance Z(;), the input impedance across the primary is equal to {N^lN-,)^Z{s)
= a^Z{s).
Models for a pair of coupled inductors: a pair o f coupled inductors can be represented by one
ideal transformer and one inductance for the case ^ = 1. For coupling coefficients k less
than 1, the representation requires one ideal transformer and two self-inductances.
Mutual inductance: a real number, usually denoted by M, that determines the induced voltage
in one coil due to the change o f current in another coil.
r>
Chapter 17 Magnetically Coupled Circiiits and Transformers
938
71-equivalent for coupled inductors: if two inductors have one terminal in common, then the
three-terminal coupled inductors are equivalent to three uncoupled induaors (one of
which may have a negative inductance) connected in the 7t-form.
Primary: The winding (coil) in a transformer that is connected to a power source.
Secondary: The winding (coil) in a transformer that is connected to a load.
T-equivalent for coupled inductors: if two inductors have one terminal in common, then the
three-terminal coupled inductors are equivalent to three uncoupled inductors (one of
which may have a negative inductance) connected in the T-form.
Transformer: a practical device that satisfies approximately the voltage transformation and cur
rent transformation properties.
Unity coupling: coefficient of coupling k= I.
Voltage transformation property: for unity coupled inductors, the ratio |i/j(/) : V2 {t)\ is a con
stant equal to the turns ratio N^INj =
r>
n
o
r>
939
Problems
(a)
in terms o f /^(r)
(b)
(It
V9 (r) = M
M = 2H
20
(It
L9 ^
dt
~ dt
4H
3H
Figure P I7.3
M
4. For the circuit shown in Figure
P17.4;
(a) Express vj^t) and v^{t) in terms o f
/^(^) and /^(r) for the dot in position
C.
(b) Obtain the ^-domain equations
containing the initial /^(0~) and
by taking the Laplace trans
Figure P i7.1
2. For the circuit o f Figure P I7.2, A/ = 2 H;
compute v^{t) when
= I t cos(20/)//(f) mA
Figure P i7.4
5. In the circuit o f Figure PI 7.5, if/j(^) = Atu{t)
A, a voltage
= 40//(r) mV is observ'ed.
Determine the placement o f the dots and the
value o f M. Repeat if
observed.
Figure P I 7.2
= -80u{t) mV is
940
(b)
i,(t)
>
Figure P i7.5
6. Consider the laboratory setup o f Figure
PI 7.6a.
(a)
(b)
Figure P i7.7
ANSW1-:RS; (a) 4 H: (b)
A. Find
D
(a)
i,(A)
1>
t (msec)
1\ / ^ 2
Kc
loop a, ( I , , - Zp -
loop b,
-1
(b)
v,(V)
t (msec)
----
-2
(c)
COMPUTING EQUIVALENT
INDUCTANCES, Z,.^(s), AND
RESPONSES FOR SIMPLE
CIRCUITS
Figure P I7.6
9. (a)
7. Consider the circuit shown in Figure P i 7.7.
es.
(b)
M l
M = 2H
T Y V -^ Y Y V SW
3H ^
v.(t)
M = 2.5H
30 n
2H
r m
IF
''.w
Figure P I7.11
(a)
AN SW ERS: (a)
.(t)
/Y Y V TYYV
3H ^
(D
= 1 H, (c) i^ J i) = U-'uU)
= 2 rad/sec, Q = 0.5
5H
M = 2.5 H
30 0
/Y Y V
P17.12.
(a)
(b)
Figure P I7.9
10. (a)
H{s) =
(c)
= 30(1 -
If
V, com
pute
mand in MATLAB.
tion A.
(c)
Vin(s)
0.5H
(d)
v ,Jt )
(n
0.1 H
state value o f
/Y Y \
2H
Figure P I7.12
'y
.w o
IY Y \
Figure P i7.10
(a)
(b)
Figure P I 7.13
942
A N SW ERS; (a)
(b)
A/-)/(A, + / . . + 2/\/)
- 2A/);
the dots?
Z(s)
(b)
(a)
Figure P I7.14
AN SW ERS: (a) Z,(.v) =
RL.s
-cir
(a)
(b)
Figure P I7.17
i f
ANSW1{RS: (a) L\ -
ZJs)
Z Js )
M-
(a)
Figure P I7.15
18. Find the input admittance o f each o f the
AN SW ERS: (a) 25 rad/sec, 0 12; (b) I rad/sec,
Z j j i u ) ^ 0.5 Q
OZ,(s)
6H
8H
(a)
943
A N SW ER:
C.v-(. v -!-4) + .V+2
C = 1 / 6 F ; Z;(.v) =
Cv(.v + 2)
M = 2H
OZ,(s)
4H
Zin(s) =
(b)
when
Figure P I7.18
= /?^ =
o f the cir
non.
AN SW ER: Z,(.v) = - ^
9.V- +6s+\
3.S- + 10.V + 3
GENERAL ANALYSIS OF
CIRCUITS WITH COUPLED
INDUCTORS
23. For the circuit shown in Figure P i 7.23,
= 1
AN SW ER;
(a)
M-
.v
V.
If a dot is placed at position A, com
Li + Ljj
(b)
to the
P17.20.
2H
2H
= 10 V, find
Figure P i 7.20
21. For the circuit shown in Figure P I 7.21,
find
es resonance to occur at
oj
= 2 rad/sec.
v.(t)
Figure P i7.24
25. If a 900 Q resistor is connected across the
F igu re P i 7 .2 1
V2(.v)
G ,{s) =
C H EC K :
Cramers rule.
(d)
9000.V
C^.=
= 30e-^',Ki) mA
/^^(/) = 60(1 - e~^^u{t) niA. Hint;
Apply linearit)' to the answers to parts
= 30u{t) mA
(a)
(b)
(c)
( a + 500)(5 + 9000)
zero-input response.
(b)
Suppose
(c)
Suppose
0.625 H
-L
for / > 0.
<^10
0.1 F
for r > 0.
= - 6 u{-t) + Guit) V;
Figure P I7.28
AN SW ER rd): v^ Jt) = 1.6<---'//(r) - GAe-^'u{t)
\
200 0
M = 0.5 H
o
+
vJt)
0.2 H'
v,(t)
5H
o
If
response,
o f Problem 19.
Figure P i7.26
27. For the circuit shown in Figure P i 7.27,
(b)
Find /,(r).
(c)
Find /,(r)
i j t ) = 100(r) niA.
(a)
(b)
o+
v Jt)
0.1 H
B
50 mF
Figure P i7.27
A N S W E R (b ):
')Ma-1 ) A
\0.2 H ''
= 0 .1
sin ( l()/)//(/) \
(a)
(b)
Find
F in d Z J ^ ) = /? + Z .,W .
(c)
Find
945
(d)
Y = ..........
H = det(Y)/det(A)
F in d G 2 = ^
simple(H)
(e)
Find Gy =
(f)
to
input
v j t ) = \6 .8 le^^^u{t) V.
(b)
(^ + 1X^ + 3)
R.
+
V,(t)
'.w Q
-3 s
C H EC K : H (s) =
syms t h
200
h = ilaplace(H)
0.2 H
(c)
ZJs)
Using
the
residue
command
or
Figure P17.30
31. Suppose that
= 3yf20cos{t)u(t) A
(d)
1.5 0
LAB as follows:
(i)
1H
Kj,
Kj, / p
l2-
Figure P I7.31
Z(s)
h (s y
? ? h is y
h (s )
? ?
h is )
Vinisy
0
(b)
syms s lin A b Y H
A = Coefficient Matrix in part 4
% Y is the coefficient matrix A with the column
corresponding to V2 replaced by the % vector b
with /y^ (s) set to 1.
(c)
Find V2 (f).
C H EC K : V2 (f) =
946
+
V,(t)
P 17.33.
(a)
Find
(b)
andZ.W .
(c)
T?2 = 80 Q. Suppose
V.
(a)
(b)
inductors at ^ = 0 .
(e)
= 120(-r) - I 2 0 u{t)
and
(c)
equations computed
cuit.
(e)
Now suppose
V, Find
j
/2 5 0
in part (b),
-----
L (t)(
S
0 .2 m F
0.2 H '
L
v jt )i
Figure P I7.33
34. Consider the circuit o f Figure P I 7 .3 4 , in
which
= 0.9 H , 2 = 0-6 H,
0.2 H , and
/?2 = 100 Q. Let
= 120 V. The switch S
has been closed for a long time (i.e, the circuit
has reached equilibrium) and is then opened at
f= 0.
(a)
(c)
(d)
(b)
Figure P17.35
computed
in
part
(b)
94'
Figure P I7.38
.ANSWI-R: k = 0.2S
30
fY Y Y jy Y V
39. Consider the circuit o f Figure PI 7.39. Find
20
CO
(b)
Hh
k=0.5
Figure P i7.36
20 0
= 1H
COUPLING COEFFICIENT
PROBLEMS AND ENERGY
CALCULATIONS
40 0
0.27 H
0 .1 2 H .
200 0
Figure P I7.39
A N SW l'R: 3 3 .33 ul-
40.
at
CO
lent circuit.
is zero.
k=0.5
20 0
40 0
0.27 H
0.12
0.6 mF
T l
Figure P I7.37
AN SW ER: 0.1 ml-
Figure P I7.40
AN SW ER: k = 0.946
at
CO
(a)
(b)
is
9i.S
(c)
Q, and H
40 n >V
(s)
Figure P i7.42
43. Let /(r) = 2 cos(10r + 3 5 ) A in the circuit
Figure P i7.41
Find k.
(b)
Figure P i7.43
K ..,(5 )
(c)
(a)
A dot is in position A.
obtain
(b)
A dot is in position B.
magnitude
frequency
fv c t
Q>
where
H^{s)=
*
Figure P i7.44
A+ 0.05
characteristic
o f //2W
Figure P I7.45.
(a)
If /, = 2 A and
= - 3 A, find the
stored energy.
(b)
= (^-05 rad/scc,
(c)
(d)
equals ( 0 .5 7 ,7j2
9H
0 .5 ^ 2 ^ +
46.
equals (0.57,,7,^ +
-Lii\ ( 0 +
(a)
(b)
(c)
for / > 0.
Hints:
1.
+V'2/2)^//
Recall that
1 J
= f (vi/, + v^h)dt\
JL
,{t)
L,
L,
R, >
v,(t)
L,
v,(,) 0 1 , ( 0
Figure P I7.48
Figure P I7.46
950
ANALYSIS OF CIRCUITS
CONTAINING IDEAL
TRANSFORMERS
vJt)
Figure P I7.52
(c)
0.5
and /2(r).
Find the power delivered to the load.
(d)
CH EC K :
0.2
= 300 kH, R =
= -0 .0 1 3 3 3 3
(a)
with 16
(b)
speaker?
Com pute
(d)
_ V3/3
po w e r
Vi/
l'//i
phono
input
cartridge
transformer
Amplifier
output
loud
transformer
speaker
Figure P I7.55
A N S W E R : (a) 2 0 , 5, 100 W ; (b) 6 4 W
O '
95 1
o
56. The circuit shown in Figure P I 7.5 6 crude
20
K = .
Sw/
approximate
(a)
w
P I7.57a, where
and R = 200 Q.
(c)
(d)
(e)
(f)
The
voltage
across
and
(a)
is shown
= 0.9 H,
in
Figure
= 9.6 H, = 20,
current
tweeter.
source.
ent source.
O '
model
Source
If the
adapter
is
'O
O '
inductors as shown in
Figure P I 7 .57b, deter
L2> and M.
C H EC K : i = 4 . . ^ ^ = 8 V ^ ,
Pl.m = 27.778
57. Today almost every electronic
O
ac voltage o f 110
to, say, 3
rms or 6 V__
rmsbefore conversion to
a low dc voltage by the use of a rectifier. An
adapter is basically a (practical) transformer. As
O
O
AC adapter as
a pair of coupled
inductors
(b)
AC Adapter as a
practical transformer
(a)
Figure P i7.57
952
(c)
24 mi l, 0.48 H
=10/^^) V.
Find the steady-state response to
= 20 cos(20r) V using the phasor
method.
(d)
(c)
(d)
(a)
1 :10
lo n
Figure P i7.60
(e)
position A-C.
i l = ^ = = O.I
M
ideal
L2
Figure P I7.58
(a)
P 17.59.
(a)
(b)
(c)
= 0.1.
1:10
20
v(s)
1F
0.01 F
Figure P I7.59
(c)
60. Consider the circuit o f Figure P I 7.60.
Suppose the transformer dots are in positions
A-B.
(a)
H {s) =
(b)
C o m p u te
the
V,U)
zero-state
response,
953
64.
Rest of Circuit
PI 7 . 6 4 .
(a)
(b)
(c)
Figure P I7.61
and ^ = 10.
(d)
the
impulse
and
step
responses.
(c)
and Q^. Is
Find
toolbox o f MATLAB.
(e)
Compute
(b)
Let /? = 13
high? If
1. 8 8 8 3 c o s ( 1 0 V i 0 r ) -h 9 . 0 9 4 8 sin(10>/i0r)]//(O
Q =
Plot the magnitude response o f the
transfer
function
to
verif}'
your
answers.
lo ,
'
2sin(^^)
A 6a = arcsm
and
to PJ
9S4
Figure P I7.67
Figure 1 17.65
AN SW ER: Vv(.)
(H =
^th (
it
144
/?^ =
^(/)= 10>/2cos{l0-^/) V.
(a)
73.5
Suppose
= 25 H,
and
= 1 (equivalent to a
^ "I-------
isolation).
Find
^2, find
(b)
maximum
^ 2
power transfer to
= 5 cos(2/) V. Find
256
and |V^ |.
Suppose
= 1 (equivalent to a
straight-through connection with dc
isolation).
Find
for maximum
Suppose
and
Pi
ideal
transformer
value o f PL.ave'
What is the maximum value o f P Lave
if//, and a-^ are both adjustable?
Figure P I7.66
67. In the circuit o f Figure P 17.67, suppose
= 10 n ,
(a)
(b)
= 40 U, and R^ = 2 0..
Figure P I7.68
ANSWL'.R: (a) 0.1479 watt, (b) 0.64 watt, (c)
- 2,
for
transfer.
Figure
PI 7 .6 9 .
The
source
voltage
v^(t) = y / l C 0 S ((0 / ) V, where to = 10^ rad/sec.
(a)
955
O '
each o f the cases cited below.
Figure P I7.70
L^, and M.
fimction
Vsis)
why?
500 a
1 kO
^
O '
''.W
Ak. Ak.
(p
Lossless
N
10 nF
(a)
(b)
100
Figure P I7.69
70. Consider the resonant circuit shown in
Figure P I7 .7 0 containing a pair of coupled
inductors. Suppose
O
= 20
Zj = 1.5 mH, M
O
o
o
< J
O
(d)
krad/sec.
Find the voltage gain at resonance;
i.e., find
J| ,
and Q of
the circuit.
O '
W
Compute (O^,
approximate
|//(/(0^^^)|,
values
of
gain H{s) =
(b)
(d)
Figure P I7.71
and the
X I t P nitl/scc,
= SO, ; i i k1
= 10
APPLICATIONS
72. Recently, coupled coils have been manufac
tured on printed circuit boards by printing two
spiral coils on opposite sides o f a board. Such
coils are called coreless PCB transformers and
can be used to effect electrical isolation as well as
energ}' and signal transfer. This problem illus
trates the use o f the models developed in this
chapter for an approximate analysis o f the fre
quency response o f coreless PCB transformers.
18.26.
= 9.13
2000 a .
(b)
text.
For a first analysis, assume that the
circuit.
(c)
V;
= 0 .1
Figure P I7.73
If the coupling coefficient ^ = 0 (or Af = 0),
then the steady-state output magnitude is
= 0. Further, as shown in Chapter 16, [V qI =
Ql
|VJ.
Q = 20. M
= 0 . 05
mi l .
r nttijnux = 1
ANSWTRS:
Cj = C2 =
then
kQ^
'out
1+ (A Q )and
*c\
i+{kQf
(b)
max
'out
when
k=~,
C H A P T E R
Two-Ports
960
Chapter 18 Two-Ports
Input
Signal
Output
Signal
Often the circuit berween the ports is highly complex. This chapter looks at shorthand methods
for analyzing the input-output properties o f two-ports without having to deal directly with a pos
sibly highly complex circuit internal to the two-port. The chapter will provide a variet)- o f meth
ods for analyzing amplifier circuits.
CHAPTER OBJECTIVES
1.
2.
3.
4.
Develop specific formulas for the input impedance, output impedance, and voltage gain
o f two-ports driven by a practical source voltage and terminated by a load impedance.
5.
Introduce and interpret the notion of reciprocity in terms o f the various two-port parameters.
SECTION HEADINGS
1.
Introduction
2.
O ne-Port Networks
3.
4.
5.
6.
%1
Chapter 18 'Iwo-Ports
7.
Hybrid Parameters
8.
Transmission Parameters
9.
10.
Reciprocity
11.
12.
Summary
13.
14.
Problems
1. INTRODUCTION
Figure 18.1 shows a general one-port, whose two terminals satisfy the propert)' that for any volt
age Kj across them, the current entering one terminal, say, /,, equals the current leaving the sec
ond terminal. A resistor is a one-port: the currcnt entering one terminal equals the current leav
ing the other terminal. A capacitor and an inductor are also one-ports. K general one-port contains
any number o f interconnected resistors, capacitors, inductors, and other devices. In a one-port,
only the relationship between the port voltage and current is o f interest. For example, the port
Cdv^Jdt
voltage and current in a resistor, capacitor, and inductor satisfy the relationships
- Iq and Ldi^ldt=
emphasizing input-output properties rather than detailed internal models. Thevenin and Norton
equivalent circuits determine one-port models when only a pair o f terminals o f a network is o f
interest.
O+
V.
One
Port
O+
V,
Two
Port
-O
+
V,
-o
(a)
(b)
and V-,, to the (output) port currents, /j and 7^. Impedance parameters
(termed 2-parameters) relate the two-port (input) currents to the two-port (output) voltages.
Other t}pes o f parameters investigated in this chapter are hybrid or A-parametcrs, and trans
mission or /-parameters.
962
Chapter 18 Two-Pons
Source One-Port
FIG U RE 18.2 A general rwo-port connccced to a source one-port (represented by a Thevenin equiv
alent) and terminated in a load one-port (represented by a Thevenin equivalent impedance).
In practice, two-ports often represent coupling devices in which a source delivers energy to a load
through the two-port network as suggested in Figure 18.2. For example, stereo amplifiers take a
small low-powcr audio signal and increase its power so that it will drive a speaker system.
Determining and knowing ratios such as the voltage gain, current gain, and power gain o f a twoport is ver\' important in dealing with a source that delivers power through a t\vo-port to a load.
This chapter develops various formulas for computing these gains for each type o f two-port
parameter.
2. ONE-PORT NETWORKS
Basic Im pedance Calculations
As mentioned in the introduction, the current entering one terminal o f a one-port, illustrated in
Figure 18.1, must equal the current leaving the second terminal for any voltage
minals. We begin study o f one-ports by exploring two impedance calculations o f a transistor ampli
fier circuit that are pertinent to ba.sic electronic analysis. The impedance or admittance seen at a port
is fundamental to the behavior o f a network to which a one-port or two-port is connected.
EXA M PLE 18.1. The circuit o f Figure 18.3 is pertinent to a simplified model o f a common-collector stage o f a transistor amplifier circuit. Specifically, the one-port o f Figure 18.3 models the input
impedance. In a common-collector amplifier stage, the impedance Z j is very large and is often neg
lected, i.e., we assume |Z,| = =c over the useful bandwidth o f the circuit. The goal o f this example is
to compute Z
and interpret Zy^^ in terms o f a transistor current gain parameter denoted by (3.
963
Chapter 18 Two-Ports
So
lu t io n
Since the input impedance is the ratio o f the input voltage to the input current,
Zi=-^=(P+l)Zi,
Thus, for a large P, say, 150, the input impedance can reach very high levels for reasonably sized
impedances Z^. When amplifying voltage signals, we desire to have the ratio o f the internal source
impedance to the amplifier input impedance be small. Conversely, the amplifier input impedancc
should be large relative to the internal source impedance.
Exercises. 1. In the circuit o f Figure 18.3, suppose P = 100, Z j = oo, and the load resistance Z^ =
16 Q. If a sinusoidal voltage
the average powers
and
'o i l ! _
Io,
(P + l ) '
and |/j|.
= 200 niA
Chapter 18 Two-Ports
A second example that is common to basic electronic analysis depicts a circuit used in the analy
sis o f field-cfFect transistors (FETs).
E X A M P L E 18.2. Figure 18.5a represents a simplified model o f the input impedance o f a fieldeffect transistor circuit. As in Example 18.1, we neglect the large impedance in parallel with^^^K,.
Compute the input impedance
in terms o f Z j, Z 2, Z 3, and
FIC'ilRE 18.5. Simplified input impedance model of a field-effect transistor (FET) circuit.
S o lu tio n
left) has been attached across the port terminals. At the bottom node, /, + g,V-[ + ^ = 0. Since
K, = Z,/, and /2 =
Now, since Z, =
VAvi
Z ; = Z , + (1 + g Z | ) Z2 = Z j + (1 +
Z j) z ,.
ance o f a 0.1 l-iF capacitor. Find the frequency (in Hz) at which the magnitude ofZy^^ is 707 Q..
ANSW i-R; 1(I03.J 11/
96 S
Chapter 18 Two-Ports
EXAMPLE 18.3. Consider the circuit ofFfigiire 18.6. Find the Tlievenin and Norton equivalent circuits.
I.
I.
1 :a
----->--------+
9
>
+ C
A +
Kw = - ^ I a + V 2 = ^ I a * V i
Cs
Cs
Therefore,
yAB =
+ R cr
Cs
^)
Thus Z,|^{s) = + R cr, V^J,s) = aRIi^j, and /^(jr) = V ( ------------------- , as set frorthL m cFigures
18.7a and b.
Cs
Cs
1 + Ra^
aRCs
RCa^s + 1
(a)
Cs
+ Ra2
(b)
FIGURE 18.7 (a) Thevenin and (b) Norton equivalent circuits of Figure 18.6.
= 5, C = 10 mF, and
= U.04.
966
Chapter 18 Two-Ports
A technique known as matrix partitioning often simplifies the calculation o f Thevenin and
Norton equivalent circuits. Example 18.4 illustrates the technique.
EXA M PLE 18.4. Find the Thevenin equivalent circuit for the network o f Figure 18.8.
So
lu t io n
Step 1. Construct the loop equations. By inspection, the loop equations for the circuit satisfy
v,
+ 18
2.V
-8
-10
-8
18
-4
-10
-4
16
/3 .
13 6
s
0
(18.1)
r\v,i
\V,2
vvv>
where
= 136/j,
(18.2)
Step 2. Solve the partitioned matrix equations for K, in terms o f /j arid the vector [ K,,
The
and
+'^12
h
h
(1 8 .3 )
967
Chapter 18 Two-Ports
V2 I
= W21/1 + VV22
(18.4)
^3
Since
is invertible (for all passive networks), solving equation 18.4 for the vector [/j
yields
V2 I
(18.5)
Substituting equation 18.5 into equation 18.3 produces
[V2
(18.6)
^3
MATLAB or its equivalent, or possibly a symbolic manipulation software package, allows the
matrices in this equation to be conveniently computed. Comparing equation 18.6 with the struc
ture o f a Thevenin equivalent circuit,
(18.7)
Z , ( .) =
(18.8)
Is
-1 -8
-1 0
18
-4
-4
18
18
-4
-4
16
-1
- 8
-1 0
.
_ .V+ 0 .0 9 5
= 5 .2 6 5 --------------
- 10 ]
-1 1361
s
0
84
^
Q
968
Chapter 18 Two-Ports
5.3
s + 0.095
+ \- M
Exercise. Suppose the capacitor in Example 18.4 is changed to a 2 H resistor. Compute the new
Z,,[s).
AN SW l-ll: Z^,,{s) = 7.2647 i l via the MA'I'LAB code
AVI 1 = 20; W 12 = [-8 -10]:
\V21 = [-8: -101: \V22 = [18 -4:-4 16]:
vZth = \V11 - \Vl2'inv(\X/22)*\X'21
V.
o
Two
Port
-O
+
V,
-o
FIG URE 18.10 Standard two-port configuration having four external variables: /j, I^, Kj, and V-,.
969
Chapter 18 Two-Ports
ADMITTANCE PARAMETERS
'rhroughout this and later discussions, assume that the two-port ot Figure 18.10 has no inter
nal independent sources and that all dynamic elements are initially relaxed, i.e., have no ini
tial conditions. Under these assumptions, the admittance parameters or ^-parameters o f a
two-port are expressions for the terminal currents, /j and I 2 , in terms o f the port voltages,
and ^2
/r
-Vli
V|2' -\/,-
^2
.^21
>22
(18.9a)
^2
(I8 .9 h )
-II
V,
as fol
( 18. 10)
vs=o
Vl=0
V2 I =
V, =0
= 0, the^^y are
often called short-circuit adm ittance parameters. I'he unit o f an admittance parameter is S.
Some examples will illustrate convenient methods for computing the ^-parameters.
EXA M PLE 18.5. Compute the short-circuit admittance parameters o f the circuit in Figure 18.'
and
970
Chapter 18 Two-Ports
* y 2 V ._ * K, ( K , - K,) = ( g ^ - r ,) . (K j * K,)
In matrix form, these equations are
h
K, + K3
I2
A ' -^3
Yl
V/,-
A -Vii
V21
.^2.
V,2
V/,-
V22
vs.
in which case
y\ \ - ^1 + ^3>^12 -
^V>'21 -
~ ^3 '*^^>'22 " ^2
^3'
Exercise. Suppose the controlled current source in Figure 18.11 is reconnected across
with the
= g ,~
and y ,, = )\ +
In the next example, we combine the method o f matrix partitioning with the use o f nodal equa
tions to compute the y-parameters.
EXA M PLE 18.6. Compute the ^-parameters for the circuit o f Figure 18.12.
^TT
Computation o f the j'-parameters will again proceed by the method o f nodal analysis. Consider,
for example, node 1, in which the current /, must be determined, /j has the form
where each variable is understood to be a function o f s. The coefficient K j, is simply the sum o f
the admittances incident at node 1. The coefficient
admittances between nodes 1 and 2, and
nodes 1 and 3. Similarly, I 2 = V-y]
incident at node 2, etc. Hence, in matrix form, the node equations o f the circuit o f Figure 18.12
are
97
Chapter 18 Two-Ports
.v+ 3
w
=
-s
: -2
- .V
5 + 21 -1
r\y,i
iM
11 : M/pi
VV2V V w ji
V2
This nodal equation matrix is symmetric, because the /^LC network contains no dependent
sources. Using the method o f matrix partitioning introduced by equation 18.6 yields
/r
/2.
-\/,-
.^ 2 .
-li
-
5 + 3
- 5
i - + 2.2
-(5 + 0 .4 )
-(5 + 0 .4 )
5 + 1.8
'
1 T
+ 2 5
[2
1]]
V,-
+ 2 .7
- ( 5
- ( .s + 0 .9 )
+ 0 .9 )!
.s + 2.3
Find the y-parameters by using nodal analysis in conjunction with the ideal transformer equations.
First, at port 1,
/, = K, + (V , - V,) = 2Vi -
= 2V , - - ^ 2
(1 8 .1 2 )
Now considering that al 2 = -/, a node equation at the primary o f the transformer yields
Chapter 18 Two-Ports
(V i-V i)
h = h = -a
a
= - - V ,+ 4 - V - ,.
a~
(18.13)
where the last equality uses the relationship aV^ = V2. Putting equations 18.12 and 18.13 in matrix
form yields the ^'-parameter relationship
r\/,i
a
Exercise. In the circuit o f Figure 18.13, the top resistor is changed from 1 Q. to 0.25
turns ratio a =2 , find the new ^/-parameters.
A N SW I- R ;
-2
1.25
With the
.S.
+J'22 ^2
These interpretations lead to the tw o-dependent source equivalent circuit o f a rwo-port repre
sented by the short-circuit admittance parameters. (See Figure 18.14b.) This equivalent circuit
aids the computation o f input and output impedances and voltage gain formulas. Note: in this
chapter and later chapters the standard resistance symbol often is used to designate a general
impedance or admittance.
Chapter 18 Two-Ports
-o
O+
V.
o-
+
V,
y y.2
L
V2,
V22 J
(a)
FIGURE 18.14 (a) Short-circuit admittance parameters, (b) Their two-dependent source equivalent
circuit. The dotted line at the bottom of (b) indicates that the two halves are not necessarily connected.
O
+
-o
+
V,
Vn y ,2
L V21 V22 J
Vll
>'12
\V
.'21
V22
V2
in Figure 1 8 .1 5 , we obtain
Chapter 18 Two-Ports
97-1
Incorporating this terminal condition into the two-port ^-parameter equation yields
>11
>12 '
V/,-
/l'
>11
Vl2
>21
>22
.^^2.
>21
y ii +
det
V. =
Because
^2
in terms o f /j results in
/l
0
>12
>22 + y'L
>12
det >11
>21 >22 +
is the ratio o f /, to
rv,
(>22 + ^l ) ^1
>l 1(>22 +
" >12>21
(18.14)
>22 +
E X A M P L E 18.8. Derive the input admittance o f equation 18.14 using the two-dependent
source equivalent o f Figure 18.16. Here we avoid the solution o f simultaneous equations while
increasing insight into the operation o f the two-port.
I.
FIG URll 18.16 Input admittance calculation using twodependent source equivalent circuit.
S o l u t io n
W ith regard to the right side o f Figure 18.16, the current ^^21
.V2 I
(18.15)
Equation 18.15 says that the voltage V2 equals the current, -^21
Now consider the left side o f Figure 18.16. Here
>l I -
impedance,
>'12>'2I
>22 +
Again, y,>, =
Vi
/
1
+ K^)
9 5
Chapter 18 Two-Ports
Exercise. Let
denote a source admittance. Show that the output admittance o f Figure 18.17 is
J 12J 2 I
(18.16)
V-
-o
-O+
Vu Vl2
V.
V,
V2, V22
-o
2
G ain C alculations
Our objective now is to derive a formula for the voltage gain of a doubly terminated r\vo-port, as
illustrated in Figure 18.18. Again, t h e s y m b o l s denote general admittances rather than the tra
ditional conductances.
yi2^2
V.
V,
y
'22
V2/,
admittances.
The specific goal is to derive the voltage gain formula
Ys
->21
(18.17)
The overall gain calculation breaks down into two cascaded gain calculations as follows:
^
V,
V/.
V/,
Gi/ = = X
v;
V,
V/,
(18.18)
Chapter 18 Two-Forts
VS
To compute the gain G ^.2 = from directly from equation 18.15,
V /.=------>22 + Yl
which implies
^
V21
^ v2= = M
>22 +
(I8 .1 9 b )
Equation 18.17 follovv.s by substituting equations 18.19a and 18.19b into equation 18.18.
Exercises. 1. In the circuit o f Figure 18 . 18, 7 ,, = 5 ,^ p = -0 .2 , y 21= 50, j/2-) = 1 (all *n mS),
1 k n and /?^ = 2 kQ. Find Z- and G y = VJV^.
A N SW ER:
= 8 5 .' U. 6\ -
2.(>32
o f the circuit o f Figure 18.18.
ANSW^HR;
IMPEDANCE PARAMETERS
For the nuo-port o f Figure 18.10, the z-parameters zij relate the port currents to the port voltages
according to the mati ix equation
Vj-
^11
-12'
V2
.^21
^22_
(18.20)
under the assumption o f zero initial conditions and no internal independent sources. Therefore,
from equation 18.20, each individual z-parameter is defined according to the formulas
h h=0
-1 2
/,=0
(18.21)
Z2\ =
^22
h=0
/,=0
Since each zij is defined with one o f the ports open-circuited, i.e., /j = 0 or /-, = 0, the z ip are
called open-circiiit im pedance param eters. Their unit is the ohm.
Chapter 18 Two-Ports
Rather than apply the z-parameter definitions o f equation 18.21, we will use mesh analysis to
obtain equation 18.20 directly.
Step 1. A loop equation at port 1 yields
.y
.V
s/
(18.22)
+ (/] + h ) = /i + f 3 + ] h
s /
(18.23)
^11
- 12'
^21
~22
s
]0
.y
ay + 10
In the next example, we utilize the technique o f matrix partitioning to compute the z-parameters
o f a n:-network.
E X A M PLE 18.10. Compute the z-parameters o f the Jt-network o f Figure 18.20.
9~S
Chapter 18 Two-Ports
S o l u t io n
^2
0
z,
0 1
-z ,
Z, j
Z2
^2
Z|
Z2 j Z| + Z2 + Z3
Hence, using the matrix partitioning formula derived in Example 18.4, we obtain
^11
^12 '
^22
1
0
z,1
0-
^ 2.
Zi + Z 2 + Z 3
- Z iZ i'
Z] + Z 2 + Z 3
^2
^2
-Z ,Z 2
^2
Z | (Z 2 + Z 3 )
^ 1^2
Z|Z2
Z 2 (Z ]+ Z 3 )
z, + z , + z .
Exercises. 1. In Example 18.10, suppose Z j = Z-, = 1 H and Z 3 is a 1 H inductor. Find the zparameters.
AN SW ER:
-11
" i:
-21
-22
.V+ 1
.v+2
1
.v + 1
2. Now suppose that Z , and Z 2 are changed to 1 F capacitors and Z 3 is a 1 H inductor. Find the
new 2-parameters.
A.N'SWER:
-II
- 12'
"21
-22
.v(.v-+2]
.V -
+1
1
I
.V -
+ 1
979
Chapter 18 Two-Ports
Z21
^12'
^22
h'
-11
^12'
>21
Z22
^2
and
-1 Vi'
^12'
'y\\
312 -Vi-
.>21
.''22 .^2.
^ ^ and^j
-1
^22_
V|1
>12'
>21
>22
^11
^12'
>11
>12'
>21
-22.
>21
>22.
which case
(18.24a)
-1
(18.24b)
212'
5 + 1
^22
.9-^2
1
5+ 1
Find the ^-parameters. Can you construct a three-element passive circuit that has these ^/-parameters?
ANSWER: ;
H inductor.
>11
>12 _ 1 '.v + l
>'21
>22
.V -1
-1
.v + l
Despite this inverse relationship, some circuits have z-parameters but not ^-parameters, and vice
versa, as illustrated by the following example.
EXAM PLE 18.11. Compute the z-parameters o f the circuit of Figure 18.21. Do the/-parameters
exist?
O+
-O
V.
V,
o-
-o
980
Chapter 18 Two-Ports
S o l u t io n
R-
vs.
A
^2
I2
= 0. Because the [z-^ matrix does
not have an inverse, the circuit has no y-parameters. One can check j/, j directly to verify this claim.
Consider Figure 18.22. Because Vj = 0, there is also a short circuit across Kj, making the ratio
vi 1 =
Vi
^-,=0
undefined.
O-
V,
v ,= o
= 0.
Here,
equals the sum o f two voltages: z^ j/j plus the voltage due to a current-controlled volt
age source given by z^-^Ij- This is precisely the left-hand portion o f Figure 18.23b. A similar inter
pretation follows for Vj = -221A
O
+
V,
^2, ^22
(a)
Mr-
-0
^12
-
V,2
.
-0
o
^22
V,
+
V,
"211
(b)
F IG U R E 1 8 .2 3 (a) z-paramcters for modeled nervvork. (b) Tw o-dependent source equivalent circuit.
98
Chapter 18 Two-Ports
This equivalent circuit proves useful for computing voltage gains and input and output imped
ances o f terminated two-ports. It should also be noted that there are other, equivalent circuits that
interpret the 2-parameters. A similar remark can be made for j/-parameters.
^oiit ~ ^22
-12-21
! 1 + ^v
(18.25)
Vj = Z22^2
In calculating
source,
^21^1
(18.26)
which is the Thevenin impedance seen by the load, the independent voltage
Solving for /, and substituting into equation 18.26 yields the output impedance formula o f equa
tion 18.25.
Chapter 18 Two-Ports
Exercises. 1. Repeat the preceding derivation using a matrix method and Cramers rule.
2. Derive the following formula for the input impedance o f a terminated rwo-port:
^12^21
A n = ^ \ ---------(18.27)
<22 +
Gain Calculations
The next phase o f our rwo-port analysis is to repeat the ^'-parameter derivation o f the voltage gain
o f the rwo-port in the context o f 2-parameters. Specifically, our aim is to compute
The ratio
^2
^2
C 8-28)
Z/,
(18.29)
-----^21^1
+ ^22
^2
Hence,
+ <-22 ^in
Substituting equations 18.29 and 18.30 into equation 18.28 yields the voltage gain:
-21
Gy = Gy2Gy\ =
\Z22 + ^ L / \ ^in +
(18.31)
/
An application o f this formula and its derivation to a cascaded network of two-ports (two transis
tor amplifier circuits) appears in the next example.
E X A M PLE 18.12. Consider the nerwork o f Figure 18.25, which represents a two-stage (transis
tor) amplifier configuration. Each stage utilizes the same transistor in a difi^erent circuit configu
ration. The first stage is an amplification stage that will amplify a small source voltage to a much
larger one. The second stage is an impedance-matching stage used to match the load to the out
put impedance o f the amplifier circuitry to achieve maximum power transfer, at least approxi
mately. The 2-parameters, in ohms, for each stage in Figure 18.25 are given by
983
Chapter 18 Two-Ports
350
2.667-
-1 0 ^
6667
21 =
; Z2 =
1.0262 X 10^
1.0258
K/
6 ,7 9 1 '
(18.32)
6 ,7 9 4
(a)
and
(b)
(c)
Check the matching o f the load and output impedance o f the amplifier circuit.
160
Step 1. Determine
^ straightforward application o f equation 18.27 using the z-parameters
o f stage 2 produces the impedance
6 7 9 1 x 1 .0 2 5 8 x10^
+ Z,
~ ^0 1 1 / ^ 2
Z,
z,|
(18.33)
6794+16
(18.34)
The gain here is small, but remember that this stages real purpose is impedance matching, not
amplification. By proper choice o f the z-parameters, the output impedance will approximately
match that o f the load. This allows us to dispense with an expensive and bulky impedance-match
ing transformer.
Step 3. Compute
The input impedance seen at the front end o f stage 1 follows from equation 18.27:
4 , =
- ^ 1 2 ^ 1 - = 350 +
Z22+^
li
^
6 . 6 6 7 + 1 ,2 3 9 .8
= 6 8 7.3 n
(1 8 .3 5 )
Chapter 18 Two-Iorts
^vl - T 7 -
j" ^21
^
\ Z^. + Zj,j) \^/J + ^22/
-10^
^(
1,239.8
[ 7 5 + 687 .3 ] 1 1 ,2 3 9 .8 + 6.6 6 7 ;
(18.36)
= -2 0 5 .7 .
Here the large gain of stage 1 leads to significant amplification o f the input signal. For example,
a 40 mV sine wave would be amplified to a little over 8 V, which can drive a small speaker.
Step 5 . Compute the overall voltage gain, Gy=
is then easy to compute the output impedance o f stage 2, again using equation 18.25. The answer
is
Exercises. 1. For the circuit o f Figure 18.25, verify that the output impedance equals 17 Q.
2. In Figure 18.25, call the 2 kQ resistor R-,. Find a new value of Rj so that
exactly equals 16 Q.
ANSW ER: 1.783 k il
3 . Suppose the 75 O. source resistance in Example 18.12 is changed to 300
twin line connection bet\veen an ideal voltage source and the first amplifier stage. Redo the example.
ANSW ER: Numerical values are obtained using the h)llowing M AFI.AB code;
1^ = 300;
z z ll = 1.0262e6: zzl2 = 6791;
7.7.21 = 1.0258e6; /.z22 = 6794;
x.ll = 350; vM = 2 .6 6 "; z21 = -le 6 ;
z22 = 6667; R2 = 2e3; ZL = 16;
Zin2 = zzl 1 - z/.I 2'/.z21/(/z22 + ZI.)
(;2 = (zz21/Zin2)-ZL/(Zl.+zz22)
ZLl = R 2 Zin2/(R2+Zin2)
Zin = z ll - zl2'z21/(z22 + Z L l )
G1 = (z21/(Rs+Zin))-ZLl/(ZI.Uz22)
Gv = G 1 *G2
Zoutl = z22-zl2*z21/ (7ll+ R s)
Zs2 = R 2*Zoutl/(R 2+Zouil)
Zout = zz22-zz 12' zzl I /(zz 1 1+Zs2)
985
Chapter 18 Two-Ports
7. HYBRID PARAMETERS
Basic D efinitions and Equivalences
As we have seen, some circuits have ^-parameters but not z-parameters, and vice versa. A circuit
element that has neither is the ideal transformer.
EXA M PLE 18.13. This example shows that the ideal transformer o f Figure 18.26 has neither znor/-parameters. From the definition o f an ideal transformer (Chapter 17), V2 = a V and /j =
be
expressed as functions o f K, and Vj- Hence, an ideal transformer has neither j/- nor z-parameters.
1 :a
o+
V,
Two-port circuits having neither z- nor/- parameters require an alternative modeling technique.
The hybrid parameters offer one o f several alternatives.
HYBRID PARAMETERS
Hybrid parameters, h-jy are a cross between y- and 2-parameters: a voltage
and a current
are outputs, with /, and Vj as inputs. Specifically, if the rwo-porr o f Figure 18.10 contains no
internal independent sources and has no initial stored energy, then the hybrid parameters are
defined by the matrix equation
Vi-
/'II
im -
/'21
hi
V2
(18.37)
As with both/- and 2-parameters, we interpret equation 18.37 as a two-dependent source equiv
alent circuit, as illustrated in Figure 18.27.
)86
Chapter 18 'Two-Ports
'2
+
V.
-o
+
h,2~
_ h21 h22 _
V.
o-
-o
(a)
FIGURE 18.27 (a) Hybrid parameters, (b) Two-dependent source equivalent circuit.
Exercise. Justify the two-dependent source equivalent circuit interpretation o f Figure 18.27; i.e.,
apply KVL and KCL to the circuit o f Figure 18.27b to derive equation 18.37.
The definition o f each /^-parameter follows directly from either the preceding equivalent circuit or
from equation 18.37. For example,
(18.38a)
V'-,=0
Because Aj| is the ratio o f an input voltage to an input current, it is an input impedance. Since
^2 = 0, /?], is termed the short-circuit input impedance. Notice, however, that A, j is simply
related to both the
and the z-parameters as follows:
I
/ , = ^
'I I
-12^21
(18.38b)
The second //-parameter, />2 p is called the short-circuit forward current gain, since it is the ratio
o f I j to /] under the condition K, = 0, i.e.,
(18.39a)
V^=0
From the 2-parameter equation Vj = ^21A
pretation.
^'>2-^2
^2
(18.39b)
^-,=0
^22
/,=0
(18.40a)
Since it is the ratio o f Kj to Vj under the condition that port 1 is open-circuited, i.e., /j = 0, it is called
the reverse open-circuit voltage gain. Interpreting //p in terms o f the y-parameters, we obtain
987
Chapter 18 Two-Ports
>'12
I,=0
(18.40b)
'"'I
h
l' n = -
(18.41a)
/,=0
and has units o f S. The word open-circuit suggests a z-parameter interpretation. Considering
that V-y = 221-^1
h ~
1 _
V12J 21
- >'22-----------:22
311
/,=0
(18.41b)
This relationship is similar (notice the subscripts) to equation 18.38b, which determines the shortcircuit input impedance.
We will return to these equivalences later, after we gain some computational experience.
E XA M PLE 18.14. Consider the two-port o f Figure 18.28, whose front end is a short circuit and
whose secondary is an open circuit. Thus it has neither z- nor ^-parameters. Our objective is to
compute the /^-parameters.
I
L
-O
+
O+
V,
-o
FIG URE 18.28 Simple two-port with /^-parameters but neither z- nor ^-parameters.
SO L U T IO N .
By inspection, the A-parameters are
V,-
O' /r
/2.
V2
T he second example illustrates the computation o f /^-parameters for an ideal transformer circuit.
Chapter 18 Two-Porcs
EXA M PLE 18.15. Find the /^-parameters o f the c%vo-port in Figure 18.29.
1 :a
O--------+
------o
*+
A
V.
V,
V,
Step 1. Construct an equation fo r K,. From the primar}' and secondary voltage relationship o f an
ideal transformer,
V, = - v - ,
By KVL at port 2, =
a ~
a ~
(18.42)
/ 2 = -/ ,
Cl
(18.43)
Step 3. Write equations 18.42 an d 18.43 in matrix form, and solve fo r Kj and I-, in terms ofl^ and
rv/,1
/o
- a
/ll
(18.44)
1
-V,-
h
0
r 0n
- -
-M
h'
VS
\R
r
a~
\'
Cl
-1
^2
Chapter 18 Two-Ports
= MR
R.
O----+
V,
'2
o-
V,
-o
= R^
v-,=o
Step 2. Find
V-,=0
Step 3. Find A,2 using equation 18.40a. W ith /j = 0, i.e., port 1 open-circuited, K, = V2 since
there is no current through Ry Hence,
/,=0
Step 4. Find A22 using equation 18.4 la. Again with /j = 0 , /2 = (C^ +
= Cs + Gt
/,=0
In summar\-,
\Rx
-1
Cs + G^
in which case
Chapter 18 Two-Ports
990
Exercises. 1. If all parameters in Example 18.16 are 1 with proper units, find the /;-parametcrs.
ANSW ER:
2. If the capacitor in Example 18.16 becomes an Z. H inductor, find the new A-parameters.
ANSW ER:
-1
+ G-.
Lv
h
It follows immediately that
/hi
(18.45)
= A|j/| +
1,22+ Yl
/,
(18.46)
~ ^*22
//|o/h 1
h\ 1 +
(18.47)
Chapter 18 Two-Ports
991
Knowledge o f the input and output impedances/admittances permits us to derive various gain for
mulas in terms o f the //-parameters. For example, consider again the left half o f Figure 18.31. Since
the input impedance is known from equation 18.46,
h=^V \
(18.48a)
^2 = ------ ^ / | = ------ V,
1122+Yl
^in
(18.48b)
^v2 = 7 7 =
+ W
^1
(18.49)
Voltage division at the front end o f Figure 18.31 yields the other voltage gain formula,
G, _
Zi + Z^
(18.50)
The overall voltage gain is the product o f equations 18.49 and 18.50, i.e.,
^ ^
7 7 = C j.|Gi,2 = -
/?21
^2 ^ ^lM\
/1 h->'^ + )' /
8. TRANSMISSION PARAMETERS
Transmission or f-parameters were first used by power system engineers for transmission line
analysis and are still so used today. They are sometimes called ABCD parameters.
V92
Chapter 18 Two-Iorts
TRANSMISSION PARAMETERS
T he /-parameter representation has the matrix relationship
Vj-
^11
hi
V2
/ l.
fl\
t il
-I2
( 18.52)
^ " ^ 2 /, = 0
^-,=0
(18.53)
v^=o
/,=0
The matrix equation 18.52 leads directly to the relationships o f equations 18.53 by setting
the appropriate quantity, I 2 or Vj, to zero.
In computing a single t-j vs'ith equations 18.53, some care must be exercised in exciting the circuit.
By definition,
h= ()
The ordinary interpretation o f this equation is: apply an input V-, and find an output
the condition that
= 0. 'I'hen
under
to that
o f the input K,, i.e., a reverse voltage gain when port 2 is open-circuited. This situation causes a
predicament: an independent voltage source for Vj causes a current
^11
/7 = 0
The quantit)'
we excite port 1 by Kj with port 2 open-circuited, which forces /-, = 0. It is then straightforward
to calculate /j j as the inverse o f
EXA M PLE 18 .1 7 . Consider again an ideal transformer circuit, shown in Figure 18.32. Here V-,
and /j =
0 \ V2
a
0
-/o
Chapter 18 Two-Ports
1 :a
O
+
-o
+
V.
Input and output impcdance calculations for terminated two-ports modeled by f-parameters do
not follow the usual pattern. Nevertheless, the t\s'o-port of Figure 18.33 has input impedance
7
_ h\'^L +^12
(18.54)
^out
= ^22^v *^^12
^21^5 +^11
(18.55)
+
V,
Exercise. The rwo-port o f Figure 18.33 has r-parameters r, j = 0 .0025, ^p= 500
^ ,^ 21= 3 .1 2 5 x 1 0 -8 S, and
= 0.00625. IfZ^ = 200 kQ. and
= 20 kH, find
AN SW ER: 80 kL X 200 k L l
One o f the most important characteristics o f /-parameters is the ease with which one can use them
to determine the overall /-parameters o f cascaded rwo-ports, as illustrated in the next example.
994
Chapter 18 Two-Ports
EXA M PLE 18 .1 8 . Compute the f-parameters o f the cascaded two-port o f Figure 18.34 in terms
of
and Tj, the t^vo-port parameter matrices o f the first and second sections, respectively.
/ l.
But since V2 =
r ^21
and
= 7-,
[ K;*T 1
.-^ 2 .
.-^ 4 .
r^ ii
=
V3-
= T\1
7 i1
-h
= T\Tl
1 L
h
.-^ 4 .
implying that the f-parameter matrix o f the cascade o f Figure 18.35 is simply
Tj.
Exercise. The f-parameters o f the two-ports in Figure 18.34 are (in standard units)
0.1
T\1 =
0
10
and Ti =
2
0.1
10'
1
Find the f-parameters o f the cascade (T j comes first, followed on the right by Tj ) Then find Vj
when a voltage source o f 4 V is applied to port 1 o f the cascade and port 2 o f the cascade is opencircuited.
A N S W I- R : T =
0.2
10
995
Chapter 18 Two-Ports
Vl-
^11 hi2
h'
.^2.
.^1
.'^2.
and
Vl'
Zii
Z\2
.'^2.
Z21
^22. h .
The trick is to rewrite the jzf-parameter equations so that Vj and I2 are on the left with /j and Vj
on the right:
~ ^ \ lh - ^\\h
Z2 2 I 2 - ^ \ h ~ ^2
-Zi2
y ,-
Zii
Z22
h.
-Z 21
(18.56)
Solving equation 18.56 for the vector [V^ /j]^under the proviso that ^^2
T ,-
1 -Zl2
0 Z22
-1
Zii
O'
-Z2\
ZIIZ22 212^21
Z22
0 yields
h'
V2
(18.57)
Z12 r/i
-Z 2 1
V2
Thus, we have used matrix methods to direcdy compute the A-parameters in terms of the 2^-parameters under the condition that 222 ^
hu
h i'
211^22 Zl2^21
Z\2
h\
^22,
Z22
-Z 2I
(18.58)
All other relationships are derived in a similar manner. For example, to express j^-parameters in
terms of /^-parameters, one must rewrite the A-parameter equations so that and /j appear on the
left-hand side with
and
on the right. Then using matrix form and inverting the appropri
ate matrix under the condition of a nonzero determinant produces the desired result.
Vw>
Table 18.1 specifies the interrelationships among all the parameters studied thus far.
Chapter 18 * Two-Ports
2-Paranietcrs
z-Parameters
y-Parameter
^11
^12
Z21
Z2 2
^22
A-^21
/;-Parameter
/-Parameter
j'-Paramcters
>'22
Aj
-> 21
Av
~>'12
Ay
>11
Av
A/;
>11
>21
/!21
hn.
/222
>12
>22
!h i
I1 2 2
J_
12
Ac
/-Parameters
//2i
_- h 12
^21
^21
ti2
/21
h\
[^22
-A t
^12
^12
-1
^11
/12
^12
^12
A/
^11
Az
A'
^12
^22
Z2 2
>I1
>11
^21
>21
A>-
Z22
Z2 2
>l 1
>^l 1
^11
A zl
-> 22
-1
-A/2
- h 11
^21
^21
>21
V21
/?2i
^22
-^ >
-> 11
-/?22
_^21
^21
>21
J 21
/2.21
/Ml
Ar
/Ml
A::
fh 1
[^11
/M2
^22
/22
-1
^21
^22
^22
Il2\
^11
^12
-1
^21
^22
h')')
Chapter 18 * Two-Ports
Exercises. 1. Use the code below to create an m-file in MATLAB for conversion o f 2-parameters
to /^-parameters. Verify that Z = [1 2;3 4] produces H = [-0 .5 0 .5 ;-0 .7 5 0.25].
% convert z parameters to h parameters
function [h, hi I,h l2 ,h 2 1 ,h 2 2 ] = ztoh(z)
zl 1 = z ( l,l) ; zl2= z(l,2);z21= z(2,l);z22= z(2,2);
deltaz = z ir z 2 2 -z l2 " z 2 1 ;
h 11 = deltaz/z22;
hi 2 =zl2/z22;
h21 = -z21/z22;
h22 = l/z22;
h = [ h l l h l2 ; h21 h22];
2. Use the code below to create an m-file in MATLAB for conversion o f z-parameters to ^-parameters. Verify that Z = [1 2;3 4] produces T = [1/3 -2/ 3;l/ 3 4/3].
%converting z to t paramters
function [t,tl I,tl2 ,t2 1 ,t2 2 ] = ztot(z)
zl 1 = z (l,l); z l2 = z (l,2 ); z21=z(2,l); z22=z(2,2);
deltaz = z ll * z 2 2 - z l2 * z 2 1 :
tl 1 = zll/z21;
t l2 = deltaz/z21;
t21 = l/z21;
t22 = z22/z21;
t = [ t i l tl2 ; t21 t22];
3. Write m-files for the remaining items in the conversion table for your own future use.
10. RECIPROCITY
Writing node equations for an ordinar)' linear circuit leads to a matrix equation having the form
A'
>11
y\n
Vi'
v^i
=Y
y'n1 y'nn
Often the node admittance matrix Y = [y^^] is symmetric, i-e-.7,y = yjj for ; ^ J. Such networks are
termed reciprocal.
998
Chapter 18 Two-lorts
RECIPROCAL NETWORKS
Any circuit that has a symmetric coefficient matrix either in a nodal equation or loop equa
tion representation is said to be reciprocal. Further, a two-port represented by either ^-param
eters or ^-parameters is said to be reciprocal if Z j 2 = Z21
~^2r
From Chapter 3 we know that circuits without dependent sources have symmetric coefficient
matrices in both the nodal and loop equation representations. On the other hand, the symmetr)'
o f the z-parameters and j'-parameters is typically lost when dependent sources are present. In gen
eral, we can prove that a circuit containing R s, L's, C s, and transformers, but no dependent sources or
independent sources, is a reciprocal network. From the definition o f a reciprocal two-port, we can
conclude further that if the hybrid and/or f-parameters exist, then from Table 18.1
= ^21
and/or
Conversely, any two-port that has parameters satisfying these conditions is said to be reciprocal.
Exercise. Recall that the relationship between the ^-parameters and z-parameters is
-11
-12
'II
A/
^21
^21
ho
^21
^21
^21
Proving that any two-port created from a reciprocal network has symmetric z-parameters is
straightforward. We write the loop equations in matrix form with
/p
ages and currents o f ports 1 and 2. Since the underlying netw'ork is reciprocal by assumption, its
loop equation has a symmetric coefficient matrix partitioned as shown:
Vi-
^11
-12
-1 3
vs
-12
-22
^23
-In
^13
^ 23
^33
-3/j
'1/j
'-2/1
-3/?
VV
999
Chapter 18 Two-Ports
^^2
-li
\V,| - VV,2VV2VVV2i
^11
^12
/l1
C21
Z2 2
symmetric matrix Wj-y is symmetric, and since the sum o f two symmetric matrices is symmetric,
the resulting z-parameters are also symmetric. The symmetry o f the ^-parameters follows by the
symmetry o f the inverse o f a symmetric matrix.
We now set forth physical interpretations o f reciprocit)^
are
the same as in Figures 18.35a and b, then by reciprocit)' the zero-state short-circuit responses, I 2 a and l-^y, coincide.
FIG URE 18.35 Equivalence o f short-circuit zero-state responses induced by voltage sources for
reciprocal netw'orks.
What reciprocity interpretation 1 says is that if one applies a voltage at port 1 and measures the
short-circuit current at port 2 with an ideal ammeter (zero meter resistance), then applying the
same voltage at port 2 would result in measurement o f the same short-circuit current at port 1.
Conversely, we can show that if reciprocit}' interpretation 1 is true, then the j'-parameters are sym
metric. To see this, observe that the configuration o f Figure 18.35a implies
1()()()
Chapter 18 Two-Ports
_
^<>1
Vo I ----
ha
_ ^Ollt
2 -
^in r,=o
^2h
symmetric j^-parameters.
coincide.
Reciprocit)' interpretation 2 says that if one injects a current at port 1 and measures the voltage at
port 2 with an ideal voltmeter (infinite input resistance), injecting the same current at port 2
would result in measurement o f the same voltage at port 1.
^2V
1001
Chapter 18 Two-Ports
cuit current ratio in Figure 18.37a is equal to the open-circuit voltage ratio in Figure 18.37b.
Exercise. Show that if reciprocity interpretation 3 is assumed true, it follows that A|, =
the forw ard short-circuit current gain,
^in
\i.e.,
.
V;
The three reciprocity interpretations have rigorous proofs. They however, are beyond the scope ol
the text.
Finally, a two-port that is reciprocal has an equivalent circuit representation with no dependent
sources. For example, suppose a reciprocal two-port has the z-parameter representation
V\ - 2|I /j + 2|2 A
and
~22
^2
1002
Chapter 18 Two-Ports
in which case
(18.59a)
and
(18.59b)
liquations 18.59 have the so-called T-equivalent circuit interpretation given by Figure 18.38.
I.
^22 ^12
-o
+
V.
V,
Exercise. Suppose the r\vo-port o f Figure 18.39 is reciprocal and modeled by ^param eter equa
tions. Compute K, ^ Yj, and
in terms o f the ^/-parameters, |, j/p,
AN SW ER:
1 .J'l2 -^22
O-
-'12
-o
-I-
-I-
V.
V,
1003
Chapter 18 Two-Ports
V.
General
two-port
V,
V.
'^Common'^
ground
two-port
V,
W
1.
L
(b)
(a)
and
18.41 shows three typical interconnection structures: parallel, series, and cascade. To avoid over
crowding o f symbols, Figure 18.41 omits all voltage and current reference labels. These labels are
understood to conform to those in Figure 18.40a. In particular, note that at each port the current
entering one terminal must equal the current leaving the other terminal for the two-port param
eters to be valid or meaningful.
An interconnection o f rwo-ports has a new set o f z-, y-, /;-, or ^-parameters obtained very simply
from the individual t^vo-port parameters. The interconnected rwo-ports o f Figure 18.41 have new
parameters computed from those o f
1.
and
as follows:
2.
3.
(a)
(b)
1004
Chapter 18 Two-Porrs
-
Na
-#
#
N
(c)
FIGURE 18.41 Three typical interconnections of rvvo-ports. (a) Parallel connection,
(b) Series connection, (c) Cascade connection.
where Y, Z, and T denote the admittance, impedance, and transmission parameter matrices,
respectively, and the subscripts a and b refer to the networks
and N^,
where
Vi.
r/i.
ha
and similarly for the voltage and current vectors o f N^. From Figure 1 8 .4 la,
and
and
= Z^/^. But V=
2.
k. h-
(18.63)
r = r ,r ,r ,...
(18.64)
Equations 18.62 and 18.64 for the cascade connection hold whether or not the component rwoports are o f the common-ground t)'pe. However, equations 18.60 and 18.63 for parallel connec
tions in general \\o\(\ only for common-ground two-port connections, as shown in Figure 18.4 la.
Similarly, the series connection equation 18.61 holds only for the case illustrated in Figure 1 8 .4 lb.
Series connection o f two general rwo-ports (Figure 18.40a) or series connection o f more than t\vo
common-ground two-ports (Figure 18.40b) requires an ideal transformer for coupling, as demon
strated in the homework problems. Examples 18.19 and 18.20 explain why equations 18.60 and
18.61 fail when two non-common-ground two-ports are connected together.
Chapter 18 Two-Ports
100=)
E X A M PLE 18 .1 9 . This example illustrates the difficult)^ with a non-common-ground series con
nection. Consider the two-port shown in Figure 18.42, which is a series connection o f two com
ponent two-ports. The z-parameters o f the individual two-ports are given by
2
and
Zf,u -
+ Zy when
(b)
Show that Z =
+ Zy if
(c)
(a)
=6
Rx+2
/?3 + 2
= Rj = 0.
(a) Observe that the parallel connection o f the 6 H and 3 ^ resistors is 2 Q. Thus, by direct cal
culation, the z-parameter matrix o f the interconnected two-port is
Z=
6
4
4'
6
2
^ Zn
i + Z;,
I) =
r
2
-1-
10
2-
(b) With R ,= R . = 0,
Z,, =
On the other hand, by direct calculation, the z-parameter matrix o f the interconnected two-port is
4
Z=
24
+ Ay =
nor
acts as a two-
port, as defined in Figure 18.40. This can be understood by inspecting Figure 18.43, with the
indicated the loop currents.
Chapter 18 Two-Ports
1006
------- O-
-oV
V,
-------o-
leads to /j ^ /*.
is in general nonzero. Observe that 1*^ = 1^ the current entering the top terminal does not equal
the current leaving the bottom terminal. With unequal terminal currents,
no longer has a z-
parameter characterization, because the 2-parameter definition requires equal currents entering
and leaving the terminal pair, as per Figure 18.40. On the other hand, if
= /?, = 0 in Figure
18.43, then the third mesh equation is satisfied for arbitrary values o f the mesh currents /j, Ij, and
7^, as there is no resistance at all in the third mesh. In particular, let
and the 2-parameter characterization o f
= R^ =
0, equation 18.61 holds because Figure 18.43 now has the same interconnection as depicted in
Figure 18.4 lb.
E X A M PLE 18.20. This example illustrates the problem o f a non-common-ground parallel con
nection. Figure 18.44 shows rwo two-ports connected in parallel. Before the connection, each
two-port has ^-parameters
0.7
- 0 .2
- 0 .2
0.7
S.
After the connection, by direct calculation, the new two-port has ^-parameter matrix
Y=
Clearly,
r 1.625
-0 .6 2 5
-0 .6 2 5 1.625
S.
Chapter 18 Two-Ports
lOO'
FIG URE 18.44 New rvvo-port N subject to invalid application of equation 18.60 for non-commonground rwo-ports in parallel.
The reason for the failure o f equation 18.60 under these circumstances is the same as for the cir
cuit o f Example 18.19. If a voltage source is applied to port 1 o f N, we find that the currents
and /j'^^ are not equal. Thus,
eters in forming N.
There are, however, some special cases o f non-common-ground two-ports for which equation
18.60 holds for a parallel connection. The following is one example.
E X A M PLE 18.21. fhis example illustrates how to achieve a parallel interconnection o f two gen
eral rwo-ports o f Figure 18.40a so that equation 18.63 remains valid. Reconsider the tw'o-ports
and
(a)
(b)
and
is Y = Y^ + Y^,
Chapter 18 Two-Ports
1 008
FIG URE 18.45 Equation 18.63 holds for these non-common-ground rwo-ports. (a) Parallel connec
tion of rwo general rvvo-ports. (b) Justification of equal currents at rw'o terminals of each port.
S o l u t io n
were enclosed in a box with only the leads observable, the two-port properties would
be identical.
(b) To show that the ^/-parameters of N (the interconnection o f N * and
must first show that
(ii)
and 12 1 , - / ,j
are Y =
+ Y^, we
Chapter 18 Two-Ports
l()())
To confirm these equalities, consider Figure 18.45b, which represents Figure 18.45a with all the
nonessential components removed to avoid overcrowding figure. Four Gaussian surfaces are
drawn,
5,^,
and 5,^^. Note that all Gaussian surfaces go through the core o f the trans
former. Recall that KCL holds for a Gaussian surface: the algebraic sum o f the currents entering
(leaving) the surface is zero. This immediately asserts the validit)' o f statements (i) and (ii) above.
Hence, within the interconnection, the two-ports N * and
continue to act as individual twoports. Therefore, equation 18.60 remains valid, i.e.,
n .4
Y = Y a -fY b =
- 0 .4
- 0 .4
1.4
This example extends directly to multiple parallel interconnections. Although the precise condi
tions for the applicability o f equations 18.60 and 18.61 to the parallel and series connections o f
non-common-ground two-ports are known, they are not practical enough to be included here.
Our emphasis is on interconnections o f common-ground two-ports, which occur most often in
practice.
12. SUMMARY
This chapter presented a unified setting for one-port analysis while providing a comprehensive
extension to two-ports. Two-ports are common to numerous real-world systems such as the utili
ty power grid that delivers power to a home from a generating facility through a transmission net
work. Another representative two-port is a telephone system that delivers a speakers voice to a lis
tener by sending a converted electrical signal through a transmission network. The characteriza
tion o f a two-port for such systems is done through their input-output properties. Four sets o f
characterizing parameters were developed: impedance or z-parameters, adm ittance or ^-parame
ters, hybrid or //-parameters, and transmission or f-parameters. In order to analyze various
aspects o f a system characterized by a two-port, formulas for computing the input
impedance/admittance, the output impedance/admittance, the voltage gain, etc. were derived.
Quantities such as voltage and power gain are very important aspects o f amplifier analysis and
design, as illustrated in Example 18.12, which depicts a two-stage transistor amplifier configura
tion. Although Example 18.12 utilized the medium o f z-parameters, the more customar)' medi
um for transistor amplifier design is //-parameters.
Conditions and formulas for parallel connection o f two-ports were presented in terms of^-parameters while series connections were studied using 2-parameters. Formulas for determining the
transmission parameters o f cascades o f rwo-ports were also developed. In addition, the chapter
introduced and interpreted the notion o f reciprocit)' in terms o f the different two-port parame
ters. Reciprocal circuits generally contain only R's, Z.s, Cs and transformers. Under certain restrict
ed conditions a reciprocal network may contain a dependent source, as the homework problems
will investigate.
101 U
Chapter 18 Two-Ports
are
Impedance or z-parameters: descriptive two-port parameters in which the port voltages are func
tions o f the port currents.
Norton equivalent of one-port: a current source in parallel with the Thevenin impedance.
Open-circuit impedance parameters; the impedance or z-parameters.
Open-circuit output admittance: the hybrid parameter hj-,Output admittance: the admittance seen at port 2 o f a rvvo-port possibly terminated by a source
impedancc.
Output impedance: the impedance seen at port 2 o f a two-port possibly terminated by a source
impedance.
Partitioned matrix; a matrix that is partitioned into submatrices for easier solution o f sets o f
equations.
TU-equivalent circuit: equivalent circuit o f a reciprocal two-port containing three general imped
ances in the form o f k, as in Figure 18.39.
Reciprocal network: a network whose node equations or loop equations have a symmetric coef
ficient matrix.
Terminated two-port: a two-port attached to a load impedance and a source with, in general, a
nonzero impedance.
Thevenin equivalent of a one-port: Voltage source in series with the Thevenin impedance.
Transmission or ^-parameters: parameters where
and
are expressed as functions o f Vj and
-h -
Two-dependent source equivalent circuit: equivalent circuit tor a two-port containing imped
ances/admittances and two dependent sources.
Chapter 18 Two-Ports
P r o b le m s
+ V -
ONE-PORTS
1. In Figure PI 8.1, suppose
n , /?^ = 1 k n , and (3 = 49. Find
port. Then find
100 V,Z^ = 20
o f rhe one-
Figure P I8.3
C H EC K : (a) 1 1 k n
4. Consider the circuit o f Figure P I 8.4.
= \ 0 j2 cos(lOOOf) A, a = \0, R
Suppose
= i o n , and C = 1 ^iF.
Figure P I8.1
(a)
(b)
the load.
C H E C K : 50 V
2. In Figure P I 8.2, suppose
2500
= 100 V, Z j =
(c)
1 :a
'(t)
Figure P i8.4
C H EC K ; (c) 1 kQ resistor in series with 1 H
inductor
Figure P I8.2
C H EC K : 50 W
given by
Z 3 = 1 kl^, and
vr
= 4 mS,
(a)
(b)
0
(a)
10
-a
0.5
-1
0.5
(b)
Cramers rule.
If = 2 and a voltage source is con
nected so that V|(/)= 12V 2 cos( 2 /)
V, find the average power consumed
Chapter 18 Two-Ports
1012
by the one-port.
(c)
CO
=2 rad/sec.
(a)
i-Port
(b)
If
Figure P I8.5
AN SW ER: (a) Ga U, (b) 12 W. (c) 8.64 \V
(c)
If /? = 25
rad/sec.
(d)
-------syx/'-----b :1
Figure P i8.6
ANSWKR; (a) Z,;, = .s- + 0.8. I
Figure P I8.8
= -2 /
hR
2h~ - 2 h + \
40
^tUII
2h~ - 2 h + \
C= 0.02 F and
/. = 2 H
/-P A R A M ET ER S
9. Find the v-parameters o f the two-port in
Figure P i 8.9.
40
in
R
^lll
I "\/V^
I
20
'2 0
J_ '
16
Figure P I 8.7
CH ECK:
= 0.5
Figure P i 8.9
Chapter 18 Two-Ports
10.
1013
where
(a)
(b)
Figure P I8.12.
+
1: a
V.
O------------(b)
Figure P I8.10
>11
>12
>'21
>22
= 2, /?, = 2 Q,
= 16 Q,
= 320 Q,
(c)
Suppose K| = K2 = K3 = (^ + 1),
= 80 Q.
Compute the short-circuit admittance
is+\r
parameters.
(b)
V/, is) =
(c)
64
------- , find /, is) and I,{s).
5 + 16
Figure P I 8.13
14. In the circuit o f Figure P 18.14, /?, = 2 0 .,
= 2 Q, /?3 = 2 Q, C, = 0.5 F, Cj = 0.25 F,
and C3 = 0.5 F.
(a)
Figure P I8.11
(b)
yu^=
".V li
.Vi 2
>21
}'22
\Ou{t)V.
1014
Chapter 18 Two-Ports
1n
which
He
He
2F
1F
10
2F
in
.'^11
.^12'
>21
.'22.
60
= 1 k n , and
- o .r
1
m S,
= 2 kn .
(a)
(b)
Figure P I8.14
G,. = ^ .
K,
Suppose
V/ -
(a)
(b)
= 0 .6 .
(a)
Findj/j,.
(b)
and /U = 4.
(a)
(a)
(b)
boxed rwo-port.
If port 2 is terminated in a 2 S resistor,
compute
and the voltage gain
(b)
(c)
Compute
(d)
_ V jis)
G .,=
Vds)
boxed two-port.
o f the complete circuit.
.(s)
Figure P i 8.16
Chapter 18 Two-Ports
1015
(d)
Figure PI8.22
CHECK: (b) 0.25 / (s + 1), (d) transient
response e~^u{t) V, steady-state response
(-cos(2f) + 0.5 sin(2/))(f) V
Yz
23. Figure P I8.23 represents a two-stage ampli
fier. Suppose
CHECK: (a)
240 rad/sec
y.. y.j
Figure PI 8.21
= 50 mH, (b) 400 rad/sec, (c)
-0.64'
r, =
25
mS, Y2 =
0.4
- 0 .0 0 r
7.5
0.025
= 150 Q,
= 2 kn, and /?! = 2 kQ. Find
the volt^e gain
z-' _ VOM/
(b)
^in
Chapter 18 Two-Ports
101(>
z-PARAMETERS
Stage 1
25. (a)
Y.
Figure P I8.23
(b)
AN SW ERS: (a)
(b )6 V = S 0 0
= 500 LlS:
= 10 inS,
K(.v) =
(a)
Figure P I8.25
V2(-v)
V'|(5)
(b)
(b)
If
(c)
0.25 sec.
Is rhe circuit stable? Explain your rea
= -4u{t) V, find
0 < r<
(c)
soning.
id)
Find i/-)(0.25~).
capacitor, and
(e)
2K
S-+4
(0
y^(s).
Compute Zj(s) if ^ = 2.
(g)
Compute
I, A
-^ s/ V -
z.
b;1
Y=
Q v ,
Vm y,2
Figure P I8.26
C H EC K : (c) i^2 sss(^) = 0.707A>/lsin(2/-45) V
F'igure P i8.24
C H EC K : (d) i^M 25) = 3.463 V, (g)
3.436 f-('-'^ -5 ),; ( / - 0 . 2 5 ) V
Chapter 18 Two-Ports
1017
(a)
(b)
(c)
------ ^
r
11
It
0.5 F
-------
\f
1
< 1
_ 0.25 F
>
*
r\
0.5 F
----- 0
D-----
Figure P i8.29
1:4
-> / W -
R.
60
??1
-100
??
10
+ Z^is) = , Z^ = 10 Q, and
V,
tS^
(a)
(b)
(c)
Figure PI8.27
CHECK: (a) Z =
m '
^21
^22
find Z j, Z2 , Z3 , and
the ^-parameters.
(c)
3 1 . The
in terms of
V,
V.
-o
Figure PI8.28
CHECK:
(a)
(b)
(c)
5^ + 100 1
205
79
Z] H- Z3 Z3
99
O '
Figure PI8.30
If Z iW = Z 2 W = ^,Z3W = l . r ^ = - 5 ,
-------
= 10 .
Figure P I8.31
Chapter 18 Two-Ports
1018
(b)
s+
Z=
(a)
(b)
0.5
0.5
0.5
,v +
0.5
b :l
aR
2-Port
absorbed by Z^.
For CO = 0 and cu = 1 rad/sec, find the
(c)
Figure P I8.33
79
5a
'V R,
; (b) 2 watts
''O
2-Port
ANSW ERS:
2.v' + 2.V- + 2.V + I
1000
-1000
(a)
0.25
1 + uy
a surprise to you?
(b)
V + 2.V" + 2.V +
and
(?).v- + (?).v-(-(?)
0.5
If
If
o f the circuit.
(c) 100 W an d SOW
(d)
Z=
aRn
a~Rv
2a/?o
2 6 a /e
versa).
(e)
(a)
Chapter 18 Two-Ports
1019
The two-port
hasj^parameters and the twoport Ny has ^-parameters as given below (with
standard port labeling and units):
2-Port
'O
0.1
2 (s)
Figure PI8.34
- 0.2
2
-1 0 ^
01
20
62.582
1.20751
63.75
1.25
respectively.
(a) Compute the input impedances, Z -^2
andZ;,.
(b) Compute the voltage gain, VgJVj(c) Compute the power gain,
p
^ o u t^ o u t
(a)
(b)
(c)
kQ
(d)
0 . 1'
36
0.1
40 4
FindZ^,
FindZ,..
Find the gains
V,
G .= ^ ,G 2 = ^ ,G 3 = ^ ,a n d
Ha
Hfe
^ o u t^ o u t
..
vi/,
37.
^ equivalent circuits for a pair of coupled
put impedance of the amplifier circuit.
inductors using z-parameters. Each equivalent
circuit consists of one ideal transformer and
I.
I,
L
two inductances. The analysis of a coupled
circuit with the use of such equivalent circuits
'len is very often more illuminating than writing
and solving simultaneous equations.
(a) Find the z-parameters of the two-port
Nj of Figure PI 8,37a.
(b) Show that two-port N j of Figure
Figure P18.35
P I 8 .3 7 b has the same z-parameters as
N j.
CHECK: m = 2 kii and ^y = -9.8147
(c) Show that two-port N3 of Figure
''s
P18.37C has the same z-parameters as
36. Consider the cascaded two-port in Figure
N,.
P I8.36, in which = 10 Q, /?| = 20 Q, and Rj^
(d) Use the equivalent circuit N2 of pan (b)
=4Q .
and the properties of an ideal trans
former to find 0)^, the bandwidth, and
' ^
Vi max
Chapter 18 Two-Ports
10 2 0
1 :n
(d)
(a)Two-port N,
Figure P I8.38
m v_o
(1-k^) L,
1 :n
(b) Two-port Nj
k y ~ L ,: >/l .
o -fY Y \
(1-k^) L,
NT
N2
Ideal Transformer
Figure P I8.39
ideal transformer
(c) Two-port Nj
k= 1
Figure P I8.37
(a)
/i-PARAMETERS
38. Find the /^-parameters for each two-port
shown in Figure P I 8.38 assuming the standard
labeling and units.
----- o
o----- r^ s/ V '------ o
Figure P I8.40
41. (a)
-o
o-
(b)
(a)
1: n
Z,
A
111I
(0
Chapter 18 Two-Ports
(b)
1021
rv / W -o z
+
2-Port
Figure PI8.42
43. Consider the amplifier network of Figure
P I8.43. Stage 1 is a common-emitter stage that
drives stage 2 , the common-collector stage.
Such an amplifier combination might be used
to drive a low-impedance load. Suppose Z^ = 2
Idi, Z^ = 3 kn, Z^ = 64 2, and the A-parameters in standard units of the two stages are
O '
(a)
H ,=
Ho =
Figure P18.41
42. The /f-parameters of the two-port of Figure
P18.32 are
= 250 SI, h -^2 - 2-5 x 10^, />2 i
= 125 , h22 = 2.25 mS, -2^ = 1 kH, and
=
500 a
(a) FindZ. andZ^^,.
(c)
50
0.05x10-3
nooo
0.966
-51
0.8x10-3
''out
C .-
(d)
V,/,
(d)
(b)
(c)
(b)
(a)
2000
the
the
the
the
Chapter 18 Two-Ports
1022
' 6
Common
Common
emitter
collector
stage-1
stage-2
Figure PI8.43
44. Reconsider Problem 43. If-2^ is adjustable,
find the value of
so that
= Z^ = 64 Q.
Then find Z^- and
G. = ^
0.001
2.5
0.2
H;V2 =
10
0.051
-125
0.5
^ = 0.8
2-port
N1
CHECK:
02-port
N2
Figure PI8.45
= 0.8 and G^2 ~ ~ 1 0 0
Figure PI8.46
CHECK: />2i = 200,
= 1800
L Q ,Z } = 2 k Q
4 7 . In this problem you are to design an ampli
fier circuit represented by the doubly terminated
equivalent circuit shown in Figure PI8.47. This
means you will be given certain amplifier speci
fications that will allow you to determine the
parameters of the amplifier circuit.
(i)
= 40 Q.
G, = -^ = 100;
and (iii) the source resistance is Z^ = 8 Idi.
(ii)
Chapter 18 Two-Ports
1023
(b)
800 Q..
.....
(ill)
(iv)
V.
25
V,
26
_L =
O'
>2
T =
Vo
= -1 0 0 .
(c)
-o
o-
-O
O-
a.
(c)
Compute//jj.
Compute the input impedance
(d)
(e)
(a)
Compue *2 ,
Compute the ratio o f the power deliv
(b)
-o
o-
-o
o-
(C )
Figure P i8.49
ANSWHRS:
1
2i
: (d)
1
>2
5 0 .(a)
Figure P I8.47
Z.
ification
(b)
= 0.01
//
\ h i
when /j = 0.
2 xlO -
0.01
n: 1
AN SW ER: l/?,/l =
4W .3S
n: 1
t-PARAMETERS
4 9 .(a)
(b)
Chapter 18 Two-Ports
1024
n: 1
c^
JY Y Y -6
4H
(0
Figure P i8.50
Figure P I8.53
51. Find the /-parameters o f the ner\vork in
Figure P I 8.51.
4; 1
(a)
2H
ti\^L
Figure P I8.51
-4
ANSWlUl:
0
1
1
0
4_
(b)
2.v^
-4
-S.v
^out ~
t~>2^s
(c)
V,
Derive the voltage gain Gj.] = .
Kv
(d)
Figure P I8.52
Figure P I8.54
7
12
Ml
= Mo
= Ir.
^1
compute Vjit)-
hi
-0 .2 5
^22
-h
Vi
<21
-21
o.
'21
-21
102S
Chapter 18 Two-Ports
57.
- ZpZ-,j.
where Az =
y0.022, r ,2 = 40
0.1
-0 .0 0 1 5
-0 .0 0 5
(b)
j = 0.895 +
= - 2 .6 1 7 x 105
+ y i.l0 2 X 10-3
=o.895 + y0.022.
(a)
If the receiving end (port 2) draws a
0.1
y l8 0 iX
conditions that
= 10 cos(lOOr) V
II V j = 1 3 4 .0 0 0 Z 0 V, and a resistive
C = 1 mF.
1 :n
2-port
V.I
NT
(a)
PARAMETER CONVERSION
AND INTERCONNECTION OF
TWO-PORTS
2-port
N1
is formed by
(b)
I.
'
port parameters.
+
(a)
2-port
V,
N1
^^\\,new =
= 2Ml
,, + Z and
+ Z. Other z- and h-
(b)
Figure P I8.56
- 0 . 0 0 l.v-0 .1
\i\ =
- 0 .0 0 2
0.02
0.02
0.0075
0.025
-0.1
-0.1
-0 .0 0 2 -0 .0 0 5 5
; (b )
= 2 . 3 0 5 c o s ( l 0 0 ^ - 1 2 . 5 3 " ) A.
1026
(c)
(d)
Chapter 18 Two-Ports
Figure PI8.61
62. Compute the j/-parameters of the two-port
of Figure PI 8.62.
(b)
12
r )
(a)
(b)
10
port 1 port 2
1O
------- -------
10
On / \ / V
(0
Figure P I8.60
<
^
10
(d)
Figure PI 8.63
r ^
Chapter 18 Two-Ports
102'
64. Compute any one set o f the z-, y-, and iparameters for each of the circuits o f Figure
P I 8.64. Then obtain the remaining two sets by
the use of the conversion table (Table 18.1).
yYY\
160
1 H Ij
0.5 F
'
1#
m \
Figure P I 8.64
Y=
"4
-1
.-2
65. For the circuit o f Figure P 18.65, find the zparameters o f each rvvo-port,
2-parameters
of the intercon
Figure P I 8.67
nected two-port.
AN SW ER:
\-ij
v-l-0.5
5 4-0.5
.v+l.O
.v-h2.()
Q.
Z =
Figure P I 8.65
66. For the circuit o f Figure P I 8.66, find the^parameters o f each two-port,
2-parameters
o f the intercon
= 2 iX
-2
= 1 iX and
a.
102.S
Chapter 18 Two-Ports
Figure P I 8.69
A N SW ER:
[3.2
l.S
Q.
- 0 .8
72. Two rwo-ports
/?2 = 2 Q , /?3 = 3
parameters of N .
Figure P I 8.72
73. Consider the two-port in Figure P i 8.73,
which depicts a transistor amplifier stage for a
microphone. Suppose that the transistor has h71. Two two-ports
and
are connected in
^7 = 1 Q, and
= 8 n . /?5 = 2 n ,
^h\
=2h.
eters.
(c)
1029
Chapter 18 Two-Ports
(cl)
A N SW ER:
30.15) V
(h)
r r
..
,
.
r tn,o-rc
Consider the circuit or Figure P I S . 75, m
jn
reciprocal?
(b)
Figure P i 8.73
(c)
RC
combination,
and
RECIPROCITY
find
If
response,
(b)
Figure P I 8.75
A N S W FR; (a) r
i,(t)
vjt)
(a)
v,W
7.5
5.0
5.0
4.0
2-Port
v/t)
2-Port
(b)
Figure P i 8.74
(b)
= 1 Lh
1.25
-1 .2 5
1.25
0.25
1030
Chapter 18 Two-Ports
H (s) =
njs) ^ ____
d(s)
+ ypis + l
yyy\
R.
Figure P I8.76
77. Again consider the circuit o f Figure PI 8.76,
in which 7?^ = 1 Q,
= 0.5 2, Z = 1.673 H
^in (^)
5^ + >/2 J + 1
= 1 Q,
r\
r s
resistance ratio
r\
R.
can be derived from another with
Rc
a
n
C H EC K : L = 3 .3 4 6 6 H or I = 0.8 9 6 6 H in
series R^ with AT= 2
l^O ltoa;
U
II
OI
CT
fO 'O
:=
e
E
k.
O C
OJ
1032
A crossover netiuork (low-pass or high-pass filter) separates the frequencies so that the woofer
receives the low-frequency content of the music and the tweeter the higher-frequency content. In
the magnitude response plot illustrated in the figure, both curves have the same 3 dB frequency
at 2000 Hz. This frequenq is called the crossover frequency. This chapter explores the basic design
principles and realizations o f low-pass and high-pass filters. Some band-pass filtering is also dis
cussed. Some simple crossover circuits are considered in the problems section.
CHAPTER OBJECTIVES
1.
2.
Set forth the maximally flat Butterworth magnitude response and associated Butterworth
transfer function.
3.
Present a step-by-step design algorithm for finding the filter order and associated
4.
Present basic passive and active circuits that realize a Butterworth transfer function.
6.
7.
Present basic passive and active circuits that realize a Butterworth high-pass filter.
8.
Introduce an algorithm for the design of a band-pass filter using frequency transformations.
SECTION HEADINGS
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
Summary
13.
14.
1033
Switched capacitor networks/filters, an idea introduced in Chapter 13, are something of a hybrid
between analog and digital filters. Both are beyond the scope o f this chapter, which takes a cir
cuits viewpoint on some basic analog filtering concepts and techniques.
Analog filters process the actual input waveform with circuits composed of discrete components
such as resistors, capacitors, inductors, and op amps. Analog filters are of two types, passive and
active. Passive analog filters are composed only of resistors, capacitors, and inductors. Active ana
log filters consist o f resistors, capacitors, and op amps or other types of active elements.
B a s ic T erm in olog y
A filter is a device (often an electrical circuit) that shapes or modifies the fi^quency content (spearum)
of a signal or waveform. We represent a filter by a transfer ftmction H{s) whose frequency response is
The gain, gain magnitude, or frequency response magnitude is |//(^)|. If
is normal
ized so that its maximum gain is 1, then the gain in dB (decibels) is Gjg{vi) = 20 log]Q|//(/ci))|. In this
chapter we will assume that the maximum value of |//(/o))| has been normalized to 1.
An important frequency called the cutofF frequency (or frequencies), also called the half-power
point (or points), is that frequency, denoted O)^ for which
2
H (j(Oc)
'
= 2
Since power is proportional to voltage squared or current squared (gain squared) for a fixed load
resistance, we arrive at the terminology of half-power. This is also called the 3d B down point
O '
because
Often, design specifications for a filter are expressed in terms of attenuation or loss rather than
gain. This results in several definitions that are dual to the gain-related definitions above. The loss
fiinction, denoted H (/w), is the reciprocal o f the transfer ftmction. Thus, the attenuation, or loss
magnitude, is
H (jw )
'
HiJco)
(19.1)
With these definitions established, the chapter will examine low-pass (LP), high-pass (HP), andband pass (BP) filters, leaving the study o f band-reject filters to higher-level texts. A low-pass fil
ter is a device (typically a circuit) offering very little attenuation to the low-frequency content
(low-frequency spectrum) of signals while significandy attenuating (blocking) the high-frequency
content o f those signals. High-pass filters do the opposite: they block the low-frequency content
and allow the high-frequency content of a signal to pass through. Finally, band-pass filters, as
described in Chapter 16, allow a band o f frequencies to pass while significantly attenuating those
outside the band. Interestingly, the general practical design of all such filters, especially for the
high-order case, is generally based on a low-pass prototype design; the low-pass prototype is trans
formed into a high-pass or band-pass type using a frequency transformation.
H )34
As a firsr step in exploring the design o f such filters, vve describe Butterworth (low-pass) transfer
functions. Butterworth was the name o f the English engineer who first developed this special class
o f transfer functions in his paper O n the Theory of Filter Amplifiers. ' The next step is presen
tation of an algorithm that adapts these basic Butterworth transfer functions into ones that meet
a given set of L P filter design specifications. Once the transfer function is known, an engineer
must implement this transfer function as a passive or active circuit. This chapter will also outline
Lowpass Filter
and (to^,
3.
4.
5.
0)^
<
6.
7.
CD d e f i n e s t h e s t o p b a n d ,
(0 <
103s
F IG U R E
The shaded region in Figure 19.2 represents a brickwall. The attenuation of the filter in dB, i.e.,
/l(co) =-20 log,Q|//(/w)| = 20 log,o|/^(/w)|, must reside outside the shaded region. Finding a (nor
malized) filter transfer function that meets the brickwall specs is callcd the approximation prob
lem. Once the approximation problem is solved by identification o f the proper normalized trans
fer function, the next step is to construct a circuit realization of the normalized transfer function
and then to frequency-scale to obtain the proper pass-band edge frequency, and finally to magnitude-scale to obtain the proper impedance levels.
The simplest technique for solving the approximation problem is with 3dB
normalized
H ( jc o )
( (ti )
(19.2)
1+
U c/
where
(tip
specified constant. As mentioned earlier, the term 3 dB down point arises here because, for all w,
10
lO g lQ
1+
( coY
= 10 1og,ol2j = 3 dB
i.e., there is 3 dB of loss at (O = co^ W hen e = 1 in equation 19.2, the pass-band edge fre
quency
CO,.
^3dBNLP ( J ^ ) ~ =
(0
1 + Q^
( 1 9 .3 )
1036
Equation 19.3 denotes the wtli-order 3 dB normalized Butterworth magnitude response. The words
3 dB normalized refer to the hict that at Q = 1, the loss is 3 dB (the gain is -3 dB). Remember
that the actual filter transfer function depends on a proper choice of
ify shortly.
Before proceeding further, we ask a critical question: Does this kind o f representation make sensed To
answer this question, note that the dB loss of equation 19.3 is
A{Q) = lossidB) =
2 0
logiQ H^dBNLpU^) =
1 0
logSlO 1 + Q^
( 1 9 .4 )
Plotting this function for various ;/s as a function of normali/xd frequency, Q, for a normalized
stop band edge frequency,
= 3.5,
= 3 dB, and
to lie outside the brickwalls o f Figure 19.2 and hence can validly be used to meet a L P filter spec
ification.
-XD
Normalized Frequency, Q
FIGURE 19.3 Plot of the normalized Butterworth magnitude responses for n = 2, 3,
1()3
and (o)^,
The idea
is to use equation 19.2 and these specs to determine a proper value for the filter order , a prop
er cutoff frequency (o^., and the normalized transfer function.
1 -I-
(^ s\
2: 1 0 ^ - 1
/'" p V
(19.5b)
Dividing the left and right sides of equation 19.5a by the left and right sides o f equation 19.5b
(this maintains the inequalit)' o f equation 19.5a) yields
(w X
(19.6)
gio
(19.7)
na
( COs^
lo g lO
(Or, J
\ ^ P
Thus n can be any integer satisfying inequalit}' 19.7. Usually one takes the smallest such n.
103S
(M,
LOcjnm
CO.
2 ,^ A A min
= COc,max
(19.8)
__ j
To derive the range given in equation 19.8, we reconsider equations 19.5. Taking the 2th root of
both sides of equation 19.5a and solving for
hand, taking the 2th root of equation 19.5b and solving for to^ yields the left side of equation 19.8.
E xercise. Verify the mathematical details in the above paragraph on the derivation of equation
19.8.
- (2000 Hz,3 dB). Find (a) the minimum filter order , (b)
and co^
(c) and
the normalized L P squared magnitude function, and then (d) plot the magnitude responses for
the cases where O)^ =
and a)^ =
I n x 2500.
Step 1. Find the minimum filter order n from equation 19.7. Here
/
1 0 ^ ^ - _1 '
iogio
_ I
(
log 10
l O - '- l ^
10 - 1
- = 2.68
ogio(4)
1
ogio
\
P/
implying that the minimum filter order is 3. This can also be accomplished in M A T L A B as follows:
n=buttord(wp,ws,Amax,Amin,V)
n =3
CD,
COc \ m
2 ^ 1 A "max _ I
and
2jt X 2000
CO,
COc , m
ax
min _
^ 1 0 ^ -1
1
1+
(O
co^
1039
Step 4. Plotting this function over the frequenc)' range 0 ^ 2jvf^ 2 ttx 2500 can be achieved with
the following M A T L A B code. The resulting plot is displayed in Figure 19.4.
f= 0:4:2500;
hl = sqrt(l. ./(I + (2*pi*f ./wcmin).^6));
h2 = sqrt(l. ./(I + (2*pi*f ./wcmiix),^6));
plot(f,-20*logl 0(h 1),f,-20*logl 0(h2))
grid
Frequency in Hz
F IG U R E 19.4. Plot of third-order LP filter response using
Observe that with the choice o f
= (jo^
and
. In practice, to allow for element tolerances, one would choose O)^ somewhere in
between.
the loss
and A { lo^) s
1 04
.......................
;
0.8
I /'
i/
0.6
0.4
0.2
I**
-0.2
-0.4
-0.6
-0.8
i \
:
i
/
X
-1
-1
-0.5
1
0.5
1
1
Real part
F IG U R E 19.5 Plot of the zeros of H (s) H {-s) = 1 +
As per our earlier statement, from stability considerations we pick the left half plane zeros to form
i-e.,
Hhat = poly(zeros(l :2))
Hhat =
l.OOOOe+00 1.4l42e+00 l.OOOOe+00
W e conclude that
r + y[2s+\
For higher-order cases, the procedure of Example 9.3 can again be used. The results are given in
Table 19.1, which presents the 3 dB normalized Buttenvorth loss functions. In practice, one never
computes the actual transfer function o f a filter that meets a set of non-normalized specs. Rather,
one realizes the normalized loss or transfer function and then magnitude- and frequency-scales to
obtain the proper circuit. This process is illustrated in the next section.
1042
5.
^3 d B N L p (^)
S+ 1
S~ + y / l s + 1
+ 2 r + 2s + 1
{s + l ) ( j ^ + s + \ ) =
{ r + 0 . 7 6 5 3 7 s + 1)(^2 + 1 . 8 4 7 7 6 i + 1 )
= / + 2 .6 1 3 1 P + 3 .4 1 4 2 j2 ^ 2 . 6 1 3 b + 1
{s +
5
= P
l)(j^
+ 0 . 6 1 8 0 3 ^ + l ) ( j ^ + 1 . 6 1 8 0 3 j + I)
+ 3 .2 3 6 1 /
+ 5 .2 3 6 1 P
+ 5.2 3 6 1 ^ 2 + 3 . 2 3 6 b
+ 1
R.
LC
out
Network
-o
40 db
2db
> f(Hz)
100 Hz
1000 Hz
So
1043
lu t io n
= 2 d B) and (o)^ =
= 40 dB).
loglO
1 0 - _ 1
2 0 0 0 .T
o g lO
- =
2.1164
200 jt
COn
CO.
0j^=(0 c.min
in which case
(O,
2 0 0 JZ
Step 3. Looking up the third-order normalized Butterworth loss fiinction and inverting to obtain
the transfer function yields
H MBNLP
(19.12)
Step 4. Choose a candidate passive circuit and obtain its transfer function,
W e wind up with the circuit o f Figure 19.8. Note that we will adjust the
is)
. Vouti^)
Vin(s)
/VY\
in
V.
Vc - - ! - K u, =
Ls
A t the o u tp u t node.
v,
(19.13a)
1044
--V /C +
Lv
C-iS H-----
(1 9 .1 3 b )
Ls
Solving equation 19.13b for Kq and eliminating K q from equation 19.13a yields
" - ( 'J - i r M
'//il-V
+-
C|
C 1 + C 2------------ i ~
------LC jC t
LC^C2
s ~ + -^ -------^ s +
Equating the
an d
.
1
,
. Cl +
.
= 2 , ------ = 1, and ---- = 2
C|
LC^C')
LC\C~)
2
Clearly, Cj = 0.5 F and
LCi
C 1+ C 2
0 .5 + C 9
LC ,C 2
1c
1r
r-
In summary, C, = 0.5 F,
^11
3
Step 5. Frequency-scale the circuit to obtain the desired, a n d magnitude-scale to obtain the specified
= 100. Thus,
Q
..........
^ ^
= 7.2773 HF.
C.
______
^
.
K ,L
= 2 1.832 ^F, and L ,,,, = ^ = 0.19406 H.
is to be 50 Q and O)^ =
ment values.
.\ N S W F R S : 7.38~2 iiK 22.162 uF. and 0.0492 H
In the above example, there is no load attached to the filter. A situation in which there is no source
resistance but a load resistance is given in the problems at the end of the chapter. In addition, there
is a problem containing both a source and a load resistance.
Here we have used a coefficient matching technique that is manageable for orders 1, 2, and 3. For
higher orders the method is unwieldy and we must use the methods o f nerwork synthesis studied
in other courses or resort to normalized filter tabulations given in filter handbooks.
1045
K.
H?iclBNLp{^) = --
S~ + .V + 1
where
= 1. The gains
= H{s)Hi,{s)
(19.15)
i + 1
and /iQ, are present in equation 19.15 to allow the individual stages
to have dc gains different from 1, as is sometimes necessary. The overall transfer function is real
ized as a cascade of
= -------- a n d H ij { s ) =
s~ + .s -I" 1
with the constraint
.v+1
= 1. All transfer functions beyond first order require one or more sec-
ond-order stages in cascade for their realization. The use of a basic second-order active circuit is
key to the realization of any of these second-order sections.
note that the properties of an ideal op amp force the voltage across 7?^ to be
Vo, u t
o r equivalently,
1046
This observation permits us to simplify the diagram of Figure 19.9 to the equivalent circuit of
Figure 19.10.
identifies node a.
Construction o f H { s ) now proceeds by nodal analysis. At node a the node between C , and
Ri,
1
( 1
(19.16)
Similarly at node b,
(19.17)
Eliminating
( 5)
^
K
R\R2C\C2
K ,A ^ )
ViU)
S + 1 ------------ h
]-K \
H-------
-I
R 2 C 2 )
U ^ lC ,
(19.18)
H-----------------------R\Ri C iC 2
w h e r e / r = l / , , / l / = 1 + /? ^ //? ^
H [s) =
K
s^ + s+ \
(19.19)
The design template derives from equating the denominator coefficients of equations 19.18 and
19. 19. The dc gain can be adjusted using other methods such as input attenuation (to be described
later). In particular, equating the denominator coefficients o f equations 19.18 and 19.19 requires
that
104'
1- K
/?,C,
/?2Ci
R 2C 2
(19.20a)
and
(19.20b)
R\R2C\C2
The solutions to equations 19.20a and 19.20b are not unique because there are t%vo equations in
five unknowns. This means that we can impose up to three additional constraints to produce dif
ferent solutions. Different solutions produce the proper filtering action, but have different behav
iors in terms o f the sensitivit)' of the frequenq' response to variations nominal resistance and
capacitor values. Also, different designs have different ratios of clement values that may have prac
tical significance.
One popular and robust design is the Saraga design. For the (normali/ed) Saraga design, the three addi-
R.
R]
v3
tional constraints imposed on the solution of equations 19.20 are C, = 1, Cj = >/3Q, and =
which were chosen to minimize certain sensitivities in the circuit performance. Solving equations
19.20a
and
19.20b
for
R ,,
C 2 = IC \ = S q . R] = Q~\ R2 =
/?2,
and
I/V ,
using
these
additional
K = 4 / 3 , Rs = R^/3.
constraints
yields
equation 1 9 . 1 9 is given in Figure 1 9 . 1 1 . Again we note that the dc gain K = 413 will need to be
adjusted by other means, to be described. Frequency and magnitude scaling are necessary to
achieve proper cutoff frequencies and impedance levels.
F IG U R F. 19.11 Normalized Saraga design of Sallen and Key circuit for realizing the
transfer function of equation 19.19.
W ith this design in hand, we now redo Example 19.4 using an active realization.
1048
where
5 + 1
S~ + .S + I
S~ +
.V +
we have Q = 1. The Saraga design element values are C-, = 1 F, Cj = >/3 F, /?, = I Q, 7?2
.V + .V+ 1
To obtain the correct
K,, _ 3/4
s -t-1
(19.21)
.V+
Observe that the choice o f /Q, makes A'/Q, = U the correct overall dc gain. The transfer function
of equation 19.21 is realized with the leaky integrator circuit o f Figure 19.12, whose transfer func
tion
Hl,^ar(s) =
3/4
R ,C
.y-l-
5+1
R'yC
R,
1049
Combining the second-order and first-order sections o f the normalized design produces the cir
cuit of Figure 19.13.
K,n =
c o ld
K jC ,,,,,
= 14.555 x 1 0 ^
2 I 8 . 7 ; r l O ^
Hence the final parameter values for the second-order section are
/ ?2
= 14.55
14 .5 5 kO
1050
This design assumes the fiirer is driven by a volrage source with a very small resistance; otherwise
the source impedance must be considered part o f the front end series resistance. Also, we could
have magnitude-scaled each section separately (but chose not to).
1 and
= 0, /T = 1, /^, = /?2 = 1
= /?2 = 1
K
s~ +
.V
C\C2
+1
+ --
C,
.V +
(19.22)
--------
C ,C 2
Matching the denominator coefficients produces Cj = 2Q and C j = 1/2Q. Since the dc gain with
K = 1 is 1, there is no need to modify the overall circuit gain. Figure 19.15 illustrates design A.
C, = 2Q
Q- Thus,
/?jCj =
= ^2^2'
these choices,--------- =
= 1 Q, and
1 F,
1
= C j/ G ) = Q, w h ic h is better than the 4Q^ ratio of D esign A . Figure 19.16 shows the result
in g circu it.
lO r
C =1
F IG U R E 19.16 Design B for realizing the normalized transfer function of equation 19.19,
assuming standard units.
(b)
F IG U R E 19.17 Illustration of input attenuation concept, (a) Original circuit,
(b) Circuit with input attenuation.
10 ^ 2
Figure 19.17a represents the original active network, while Figure 19.17b represents the modified
network. For the filtering characteristics to remain invariant, the impedance looking into the
dashed boxes must remain at /?, for both circuits. Thus, the parallel combination of
and
(19.23)
a =
^4
/?3 + /?4
(19.24)
=^
a
and
(19.25)
\ -a
Thus, one can reduce the gain o f the Sallen and Key low-pass circuit via the simple technique of
input attenuation.
H ( j Q ) = 0,
for A: = 1,2,...,
cIQ'^
2/2-1
(19.26)
at Q = 0 and
cf
dQ^
H (jQ )\ = 0 .5 { 2 n ) \ * 0
( , 9.27)
at Q = 0. This is consistent with the notion of being maximally flat. To verify equations 19.26 and
19.27, observe that if
1, then
Equations 19.26 and 19.27 follow after differentiation of equation 19.28 2u times and evaluation
of each derivative at Q = 0.
Several closing remarks are now in order. First, the cutoff frequency is the half-power point, or the
3 dB down point. The terminology follows because |MyO)| = 1 (the gain at dc is unity) and
10^ 3
H (j\ ) = - ^ = 0.707,
V2
yielding half power for normalized Butterworth gain functions H{s). Second, for large Q,
-n
[20 logjQ(Q)]
indicating that the gain rolls off with a slope proportional to the number of poles, n, o f the gain
function. Specifically, the slope equals -20 dB per decade. These statements are also valid if to
replaces Q.
the inverse function o f a low-pass filter. In fact, low-pass and high-pass specifications are related
by a simple inversion of O). In particular, we define a H P (high-pass) to L P (low-pass) frequen
cy transformation as
^ =
(19.29)
This frequency transformation applied to the briclc\vall specifications o f Figure 19.18 yields the
set of normalized low-pass specifications given by Figure 19.19.
10 V i
A(co)
min
3dB
A_..> (0
0
CO.
A(co)
1''VniH _ I
log
clO
10
o g lO
_
lo2 10
(19.30)
( Q ,\
By convention we take the filter order to be the smallest integer n satisfying equation 19.30. For
a passive realization, once n is known, one would realize the resulting N L P Butterworth transfer
I05S
J_
Cs
Ls + /?< +
_
1
Cs
LC
S
+ S+ -
LC
given by
Q.
Qc ju i n
where
(19.31)
Once
/YY\
u o .C
V,
C/Q,
10
H H p{s) = H N Lpi \ = H
This is equivalent to doing a special frequenq' transformation on the circuit elements o f Figure
19.21. Specifically
1056
(O.
(19.32)
is called the L P to H P frequency transformation and has a natural interpretation in terms of
the inductors and capacitors of a passive network. In particular, a capacitor C with impedance
Cs
becomes an inductor o f value
, I.e.,
C(0:
Cs
(0),
C oj,
A similar substitution shows that an Inductor L with impedance Ls becomes a capacitor of value
Figure 19.22 illustrates the transformation.
/YY\
-O
1
L=
Cco
o- fY Y \
-o
-o
1
c = Leo.
F IG U R E 19.22 Circuit element change under the LP to H P transformation.
Hence the L P circuit of Figure 19.21 becomes the H P circuit o f Figure 19.23.
Cnew=
L CO
F IG U R H 19.23. H P circuit derived from the N L P reali/aiion of L'igure 19.21. Observe that
to.
It remains only to magnitude-scale this circuit to obtain an acceptable passive H P filter. W e illus
trate the above ideas in the following example.
lOS"
>0)
2000n
200n
Solution
9000 t
Q =
the H P to N L P trans
Step 2. Detennine the Butterworth transfer function. As these specifications correspond to those of
Example 19.3, one computes the filter order n = 3. From Tiible 19.1, the third-order Butterworth
transfer function is
v-' + 2 .v + 2.V + 1
Step 3. Realize the third-order 3 dB N LP transfer fin ctio n . According to Example 19.4, the 3 dB
N L P Butterworth transfer function can be reali7.ed by the passive RLC circuit o f Figure 19.8, with
values assigned as in Figure 19.25.
/YY\
1n
V,
C^=1.5F
-O
+
V
-y
C =0.5 F
-o
---- c
F I G U R E 19.25 Third-order 3 d B N L P filter circuit o f Example 19.4 having 3 d B down point at 1 rad/sec.
1058
3 dB
frequency
of the actual N L P
equivalent filter is calculated from equation 19.31, using the left inequality:
1/6
Step 5. Frequency-scale by
= 1.0935
quency-scaled by the factor Kj-= 1.0935 to produce the network o f Figure 19.26.
/YY\
V.
-O
+
1.219H
1o
- J
'' N
0.4573 F
1.3717 F
t
6. Apply the LP to H P circuit element transformation detailed in Figure 19.22. Converting the
circuit of Figure 19.26 to the required H P circuit yields the network of Figure 19.27.
0.1305 mF
-O
+
1Q
V.
V
0.3481 mH
0.116mH
13.1 mF
-O
+
lOO
V.
1.16 mH
3.481 mH
-o
F IG U R I: 19.28 Passive circuit realizing the high-pass specifications of Figure 19.24 with 10
source impedance.
E x e r c is e . Use SPIC E to confirm the high-pass characteristic o f the circuit o f Figure 19.28.
10^)
The preceding example completes our basic discussion o f passive high-pass filtering.
H 3dBNLP {s) =
( i - +l ) |. v
ij
+ 2 a + 2.S + 1
One approach to H P design first realizes this 3 dB N L P transfer function as a cascade o f a secondorder active circuit with a first-order circuitin Example 19.5, this resulted in the circuit o f Figure
19.13, redrawn as Figure 19.29.
1O
R
(19.33)
C
1060
Applying the transformation o f equation 19.33 to the circuit of Figure 19.29 produces the circuit
o f Figure 19.30, with one exception to the rule: the feedback resistors on the second-order section
that connect to the inverting input node are not changed to capacitors; these resistors only set up
a voltage division, so it is unnecessar)' to replace them.
1.
In terms of transfer functions, the operation o f equation 19.33 changes s in the 3 dB N L P trans
fer function to \/s. Specifically,
H 7.CIBNHP
HM BNLP W =
s
M BNLP
(^ + 0
^ -f-5-f-l
Notice that each section of the transfer Rinction has a (normalized) H P characteristic. For example,
s+ 1
= 0.
In
-1 5
- 1
(19.34)
1061
= 13.05
t<fCne.
Hence the final circuit design is given by Figure 19.31.
13.3 nF
7.54 kO
Practical
Voltage
Follower
Circuit
1062
A Loss(dB)
> CO
0)
0).
0)
P2
10^2
and 0)^2
band edge frequencies. The center of the B P filter is defined as the geometric mean of pass-band
edge frequencies, i.e.,
= yjojp\(0p2
Given the B P brickwall specs, the design procedure begins with a frequency transformation to
convert these specs to equivalent N L P specs, i.e., (Q^ = 1,
and (Q^.,
To illustrate how
as Z? = 0)^2
as
^.v/ =
Iko,i
(19.35)
for / = 1,2. Given these two numbers, the N L P equivalent specs are
= '
(19.36)
To illustrate the need for the magnitude sign in equation 19.35 and the need for the minimum
function in equation 19.36, let us do a simple example by breaking apart the formula o f equation
19.35 and
rad/sec and
1063
= ^0 rad/sec. Then
- yj(0p^(0^y2 = ^
of equation 19.35 in which B is not present and the absolute value signs are gone. Thus
_ [-6
i=\
I 6
i=l
a ; ,,
[19.2
i= 2
One can view this partial transformation as the generation of two distinct L P filters: one defined
on negative frequency (Q^
6,
= - 6,
= 19.2). O f course, we cannot really have two distinct L P filters, one for negative fre
quency and another for positive frequency. O u r mathematics requires that the magnitude respons
es of each filter be symmetric because magnitude is an even function of (o. Thus the only way to
properly interpret the above transformation is that it implicitly generates two distinct low-pass fil
ters, one of which is more stringent than the other. Specifically, the negative filter is really a LP
filter with edge frequencies (Q^, = 6, Q^, = 15), and the positive filter is really a L P filter with
edge frequencies
^,v/
Q^,
|3.2
/= 2
differs for the rwo filters. The first filter is more stringent than the second, prompting the need
for a minimum fiinction in equation 19.36. Since our edge frequencies are always specified on the
positive axis and since the magnitude response is symmetric, w'e insert the absolute value signs in
equation 19.35. Finally, in order to have Q^- =
1 we divide by
= 26 krad/sec,
= 3 dB, and
= ^7 krad/sec,
= 30.6
= 14 d k Design a minimum-order
Butterworth passive B P filter meeting these specs, assuming the filter is driven by a voltage source
in series with a
10 kQ
source resistance.
S o l u t io n
K=
997560000
B = w'p2 - w pl
B =
2000
ws = [wsl ws2];
Wsi = abs((%vs M - K) ./(B*ws))
1064
W si =
6.1838e+00 5.0195e+00
W s = m in(W si)
W s =
5.0195e+00
Step 2. Find the filter order.
n = buttord(l,W s,Am ax,Am in,V
n= 1
has transfer f^unction Hf^j^p{s) = ----,whose passive circuit realization is given in Figure 19.34.
.V+ 1
-O
+
1Q
1F
.v +
(19.37)
Bs
This frequency transformation causes a change in the capacitive and inductive circuit elements o f
the N L P circuit. From equation 19.37, the impedance o f an inductor changes as follows:
Ls
( s - + K\
Bs
L
= s +
B
1
(19.38)
\LK I
rhe expression on the right of equation 19.38 is a sum of impedances. This sum, then, is a cir
cuit composed o f an inductor in series with a capacitor as shown in Figure 19.35.
/YY\
L
L
B
3.
LK
106S
= .V+
Bs
(19.39)
XK)
The expression on rhe right is a sum of admittances. Hence under the frequency transformation,
a capacitor becomes a parallel L C combination as illustrated in Figure 19.36.
C
CK
K
------- O
+
1O
1
'^out
------- O
F IG U R E 19.37. BP circuit without magnitude scaling.
Step 5. Miignitudc-scale to obtain the correct source resistance. W ith
= 50 nF and
-o
I
_ BK
bp
k"
-o
F IG U R E 19.38. Final BP design.
The final transfer function o f the circuit is
/Y gp (.v)-
Bs
This transfer function and circuit show a clear similarity to the material developed in Chapter 16.
In contrast to the above procedure, the development o f Chapter 16 works only for w =
1 in
the
equivalent N L P circuit, which results in a B P circuit o f order 2. I f the order o f the N L P equiva-
'0 6 6
lent were 2 or higher, rhe B P order would be at least 4 and the background from Chapter 16
would prove inadequate. The next example illustrates how design is done for higher-order circuits.
= 2 dB, and
= ^^00 rad/sec,
= 3000
100 Q.
lu t io n
Part
1. F in d equivaletn N L P specs.
Define K =
K = 12000000
W si = 6.5000e+00 4.0000e+00
Ws =4
Therefore, the equivalent N L P specs arc
= 2) and (Q^ = 4,
= 20)
n=2
From Table 19.1, the second-order Butterworth filter is given by
" T T T ir r T
(19.40)
1067
yvYV
-o
1n
'- 0
F IG U R E 19.39. Second-order circuit.
The circuit o f Figure 19.39 has transfer function
LC
1
(19.41)
H----- A + ------
LC
H and C = V2 F.
V2
spec at Q = 1. For this we use
1
c,w/
in which case
- W - \A^mat _ I
= 1.1435 rad/sec
= 0.61838 H and
= 1.2368 F.
at Q =
1 or, equivaiendy,
IY Y \
IQ
-O
0 .6 1 8 3 H
+
V out
1 .2 3 6 8 F
o
, - H},dBNLP
u
n^LPkS)
1.3076
U .1 4 3 5 J
-1-
(19.42)
\ \ A 435}
+1
+ 1.617 Ly-f-1.3076
Part 3. Realize the passive band-pass filter. To compute the desired B P transfer function, one would
replace s in
as follows:
1068
+Wp\COp2
'
Bs
One could then attempt to realize the B P transfer function directly. This is numerically unwise
and, fortunately, unnecessary. A simple, numerically sound procedure is to replace each inductor
and capacitor in the N L P realization by an equivalent circuit representing the above frequency
transformation. These equivalent circuits were developed in equations 19.38 and 19.39, and illus
trated in Figures 19.35 and 19.36, respectively. The substitution scheme is repeated in Figure
19.41.
/YTV
L
L
B
_B
LK
(a)
(b)
F IG U R E 19.41. Illustration of the LP to B P transformation in terms of inductors and capacitors.
Given the transformations of Figure 19.41, the following M A F L A B code produces the desired B P
circuit parameter values, assuming the final source resistance is
% Insert L, C, B, K, and Km
L = 0.61838; C = 1.2368;
B = 1000; K = 12000000; Km = 100;
% Compute K1
K f = sqrt(K)
K f= 3.464 le+03
% Compute B P circuit parameters associated with L
Llb p = Km *L/B
L I bp = 6.1838e-02
C lb p = B/(L*K )/Km
C lb p = 1.3476e-06
% Compute B P circuit parameters associated with C
>)C2bp = C /(B*K m )
C2bp= 1.2368e-05
L2bp = K m *B /(C K)
L2bp = 6.7378e-03
100 O..
1069
The final circuit is given in Figure 19.42, where R^= 100 ^2,
= 61.8 m H ,
- 1.35 |.iF,
^Ibp ~
LI bp
Cl bp
R.
-O
+
C2bp
V.
^out
L2bp ^
Active B P design is beyond the scope of this text, but coverage can be found in more advanced
texts.
1H
(19.43a)
\I n )
/3 > t ^
s in
S in
a-) =
\ 2nj
I In ]
co s*
sm
(19.43b)
sm
2/1
( 1 9 .4 3 c )
In
1070
a = ;jsin
JT
(19.43d)
\In }
Hence for the fourth-order filter above there are two possible circuit realizations, given in Figure
19.43.
0.3827 H
1.577 H
/YYY.
/YY\
1Q
V.
1.5307 F
1.0824 F
(a)
-O
+
1071
0.3827 H
1.577 H
/YYY
/YY\
1Q
1.0824 F
-O
+
1.5307 F
-O
(b)
F IG U R E 19.43. (a) R ^ = \ Q . (b) 7?/ = 1.
All that remains in the design procedure is frequency and magnitude scaHng, which are covered in
earher sections o f the chapter.
12. SUMMARY
This chapter has covered the basics of Butterworth LP, HP, and B P fiher design. Such design tech
niques build on a set of filter specifications requiring that the desired filter magnitude response lie
outside certain briclcvvall regions. Finding transfer functions that meet a set o f briclcwall con
straints is called the approxim ation problem . This chapter developed algorithms for finding
Butterworth transfer functions for the LP, HP, and B P cases. In addition, basic passive realizations
were presented as well as active circuit realizations, except in the B P case. In the active case, the
focus was the Sallen and Key low-pass and high-pass topologies. Both the passive and active t)^pes
of realization are built around the coefficient-matching technique, which associates the coeffi
cients in the circuit transfer function with the coefficients of the desired transfer function; one
then solves the resulting equations for appropriate circuit parameter values.
There are, of course, many other types of filter transfer functions; Chebyshev, inverse Chebyshev,
and elliptic are other well-known t)'^pes. Also, in addition to low- and high-pass filters, there are
band-pass, band-reject, and magnitude and phase equalizers. To add to the richness of the area of
filtering, there are analog passive, analog active, recursive digital, non-recursive digital, and
switched capacitor implementations o f all o f these filter types. The preceding exposition is mere
ly a drop in a very large and fascinating bucket of filter design challenges.
1072
Attenuation (dB); the loss magnitude expressed in dB, i.e., y4((o) = - 2 0 logjQ|//(/(D)|.
BP to LP frequency transfonnation: the generation of equivalent normalized LP specs from a
given set o f BP specs.
Band-pass filter: a filter that passes any frequency within the band 0)^j s O) s
while signifi-
Cutoff frequency: the frequency at which the magnitude response o f the filter is 3 dB down from
'
Filten a circuit or device that significandy attenuates the frequency content o f signals in certain
frequency bands and passes the frequency content within certain other, user-specified fre
quency bands in the sinusoidal steady state.
Frequency response magnitude: magnitude of the transfer function as a function of j(a, i.e.,
^
'
Low-pass filter: a filter that passes the low-fi-equency content o f signals and significantly attenu
ates the high-frequency content.
'
'
K r3
'Problems
LOW-PASS BASICS
1. Filters can have magnitude
responses quite
BP
Radian Frequency
(a)
Radian Frequency
(b)
Figure P i 9.2
3. A N L P filter transfer function
(0
Figure P I 9.1
0.65378
. r -h 0.803816435 + 0.82306043
10 7 4
= 2 dB)
1 rad/sec.
(a)
1.
= > / i # lA.
= e.
-1
(c o A
(OPJ
7. Suppose //jW and
third-order N L P
(b)
(c)
//(0) =
1.
(a)
Findand//.W .
(b)
< 5 rad/sec on
0 < O)
(c)
Determine H{s).
(b)
(d)
1 kHz.
LOW-PASS APPROXIMATION
8. Find a minimum-order Butterworth transfer
Problem 14 as follows:
P i 9.8 as follows:
(i)
First realize
to
(a)
(b)
(c)
range o f allowable^-
and the
unit circle.
to,
0),
AAn
sections.
(e)
6. Recall equation
19.2, i.e.,
H iJ c o )
2n
+ '
.2n
1+
(0
' CO '
<f
the coefficient
e , (O^, and to^.. Show that for any
coeff
L P brickwall
specification the range o f e
brick
given by
is
1075
and
Loss (db)
(a)
(b)
8.
(c)
0.3 dB
Figure P I 9.8
(e)
Using (JL)^
unit circle;
(0
n = buttord(lp,fs,Ama.\,Amin. V)
fcmin = fp/((10A(0.1*Amax)-l)^(l/(2*n)))
wcmin = 2 pi Fcmin
Ic = fcmin;
using
[z,p,k] = buttap(n)
instead o f
M odify the
zplane(p)
grid
pause
znew = z'w c
pnew = p \vc
knew = k wc'^n
f=
0:fc/50:l. 2 * f s ;
Loss (db)
h=
freqs(knew* poiy(znew),poIy(pne\v),2' p i' t ):
plot(f,abs(h))
grid
pause
plot{l',20 lo g l 0(abs(h)))
srid
2db
^cnuix instead o f
L A B code of Problem
8 using
2000n
M odify the M A T
8 as necessary.
> 0)
lOOOn
= (1 dB, 75 Hz)
Figure P I 9.12
(a)
1076
(b)
range o f allowable
(c)
(d)
' " O
Figure P I 9.13
tions.
(e)
Using
(a)
= 500 Hz.
(.v) =
LC
<
briclcwall specs?
LC
(b)
PASSIVE LOW-PASS
REALIZATION
(c)
3 dB frequency
(a)
= 1000 Hz.
L and C.
= s~ +
(b)
LC
,
,
-s+R^C
LC
(d)
1 |.iF. Find
ues of L and R jj
(D
Figure P i 9.14
L and C
(d)
1 uF.
(a)
^in
LC
1
R,
1
1 + ^sl^L
s~ -f- ---- - H----- 5 +
CR,
L
LC
1077
(b)
R ,L C
H{s) =
T
.V +
rad/sec.
s + + ! / ,
LC
(c)
(b)
If Rj= ^
quency is 5 kHz.
tion.
(c)
R,> 1
Figure P I 9.15
16. W ith
1Q
LC
H .A s) =
s~ +
2
---- f- -- y -I-----\C
U
LC
4 [d l
H i s ) = T.V
K
-I- J 2
:s
fer function
H(s) =
V'^p
(a)
this case?
(b)
(c)
response.
10 nF.
( s
(b)
^
^in
becomes
1078
and
1 and
then fre
magnitude-scale
to
A N S W E K . C, = C , = 0 .106 u F and
= 0.2122 H.
R^= 100
(a)
lY T v
L^LjC
p .
I \
n
----- 1
^ ^1
^2/
( L\ + Lo + C\
2
+ ------ =------ j + ---------L 1L2C
}
L^LjC
(a)
yvYY
(b)
characteristic.
(c)
(b)
filter having
kn.
(d)
10
/Y Y \
+
Vo,
/Y Y \
L,
K
p
10
I
>
>
Figure P i 9.20
Figure P I 9.19
A N S W F R S : (b) A, = L , = ] H, C = 2 1-: (c)
(a)
circuit is
L C .C .
c > ( .0 = -
c ,l c'-2
(b)
L C .C t
5 4-
2I.
denominator o f
with the
denominator o f the third-order 3 dB
(a)
1F
^oiit _
L C .C
l '-2
^21
and Z. = 2 H.
(b)
1079
N LP
-V21
V22 +
Vin
(c)
part (a).
(c)
^'Kntin
rad/sec,
=1
(d)
'
7i-network).
integrator circuit.
(a)
u s i n g R^^^ = R^ + R ^.
(b)
Realize
the
first-order
normalized
^ 3dBNLP^^'^) - ~rr~r7 - r
and R^= \ Q.
(e)
V in is )
Hint: Let C= I E
(c)
is 1 k n .
S+\
22. (a)
Find the 3 dB
N LP
Butterworth
Figure P i 9.23
1080
m s ) =_
*out _
( s
0).
(a)
(b)
(c)
+1
(c)
(b)
and
ter values.
(d)
terms of (O^.
C H EC K : 0)^ = 0.5098co^
r>
R.
R.
= 1 kn
the
down
= 1
C H EC K :
dB
frequency,
0
+
0)^ in terms of (O .
C H EC K :
(0^ = 0 .6 4 3 6 co^
r^
n
Figure P I9.25
26. A second-order normalized Butterworth fil
ter can be realized by the Sallen and Key circuit
described in section 5. In the final design, the
filter is to have a dc gain o f 1, a 3 dB down fre
quency
of 10 nF.
(a)
Figure P I9.24
n
n
(b)
transfer function
K
(c)
of
n
o
and
n
1081
8 as a cascade
(b)
(c)
^'^cmhr
(d)
C H E C K :^ ,= 2 d B
28. Repeat Problem 27, except use design A
instead ol rhe Saraga design for the Sallen and
Key portion.
point should be at
Key portion.
LC
1
1
s~ -I- -------.V-f"
R jC
LC
~>
(a)
(b)
C = l/ V 2
a N H P circuit with
(a)
/?,,
3 dB
normalized
second-order
(c)
(d)
------------
0.803825+ 0.82306
Hint: Frequenc)--scale, s
(DqJ where
filter (3 dB
N H P ).
normalized H P
\ Q. and
100
1082
(c)
Figure P I9.34
(d)
r -W --/Y Y \
10
-
Hr
3
S
\ 2
Cj
LC jC j
+ Co
------
-o
+
1o
vJt)(
1
LC\C2
+ ----------
LC\C2
Figure P I9.36
Compute Z in H and Cj and C2 in F
(a)
a
CH ECK : q = 1.5 F and L = - H
order prototype.
(b)
(a)
(d)
max
and
for the
(b)
(c)
Determine
NLP filter.
design.
tion
VquA^)
Vin(s)
36.
second-order
high-pass
LC1C2
c, +c
c,c
1^2
LC 1C 2
LC 1C2
(b)
and^.
Realize the second-order 3 dB NLP
Butterworth filter using the circuit of
Figure P I9.36; i.e., compute values for L
and C.
(d)
(e)
1083
IQ
c.
load resistance of 75
(a)
Figure P I 9.37
(b)
= 2 dB, and
= 500 Hz, ^ =
(c)
= 25 dB. The
jy y
There is no load
6\
L,
1o
Figure P I 9.41
HIGH-PASS ACTIVE
REALIZATION
42. Consider the set of H P specs for which
- 0)
Figure P I 9.38
43.
(600
34.
2000 Hz,
= 2 dB, and
= 30 dB. The
H {s) =
K s-
where A" = 1 +
A second-order H P
kHz.
1o s^
/Y Y l
L
8Q woofer
design using S P IC E .
80 tweeter
Figure P i 9.47
AN.SWl-.R: L = 63(> u H , C = 9.95 uF'
48. The crossover nersvork of the previous
problem provides first-order Butterworth low-
Figure P 19.44
pass and high-pass response curves. Better45. Again consider the circuit o f Figure P i 9.44.
(a)
= 3 d B) and
resistor values
MISCELLANEOUS
47. A certain audio amplifier has a very low
internal resistance. It is therefore approximate
ly represented by an ideal voltage sourcc. The
8 ohms.
m H. ( = ".0337 itl-.
Brief Introduction to
Fourier Series
CHAPTER OBJECTIVES
1.
Introduce the concept o f and calculation procedure for the Fourier series of a periodic
signal.
2.
Describe the relationship between the complex and the real Fourier series representations.
3.
4.
Show how the basic Fourier series properties can be used to compute the Fourier series
of a wide range of signals from a few basic ones.
SECTION HEADINGS
1.
2.
Introduction
The Fourier Series: Trigonometric and Exponential Forms
3.
4.
Additional Properties and Com putational Shortcuts for the Fourier Series
Representation
Summary
5.
6.
Problems
1. INTRODUCTION
Non-sinusoidal periodic waveforms are an important class of signals in electrical systems. Some
prominent examples are the square waveform used to clock a digital computer and the sawtooth
waveform used to control the horizontal motion of the electron beam of a cathode ray T V pic
ture tube. Non-sinusoidal periodic functions also have importance for non-electrical systems. In
fact, the study of heat flow in a metal rod led the French mathematician J. B. J. Fourier to invent
the trigonometric series representation o f a periodic function. Today the series bears his name.
The Fourier series o f a periodic signal exciting a linear circuit or system leads to a simplified
understanding o f the effect o f the system on the original periodic signal. This idea is briefly
explained in the next few paragraphs and illustrated in Example 20.1.
1086
W hen a periodic input excites a linear circuit, there are many ways to determine the steady-state
output. Using a Fourier series method of analysis, the input is first resolved into the sum o f a dc
component and infinitely many ac components at harmonically related frequencies. For example,
a
1 kHz
square wave voltage with zero mean and a 0.5n V peak-to-peak value, w ithy{f) = {u{t)
/(/ - n T )
( 20. 1a)
H=0
has the Fourier series representation
2/2 + 1
/i=0
( 20. 1b)
w h e re ^Q =
2jtx
1000 rad/sec. In the Fourier series representation (which we will later develop), we
observe that the signal has a zero average dc value and harmonically related frequency components
= H {j{2n + 1)Wq)V^ , assuming that the transfer function H{s) is stable. Again by
linearity, the actual time domain output is then computed for each
tions are then summed to obtain the steady-state part o f the complete response. Example 20.1
illustrates the particulars.
input voltage (Figure 20.1a) with E = 3 0 jt V and T = 4 sec. Find the first four components o f the
output voltage v^{t) in the steady state.
>
r
b
(a)
v Jt)
v jt)
(b)
1087
S o l u t io n
Step 1. Determine the Fourier series representation o f v-^^it). The fundamental frequency o f v-^^{t) is
Jq = 1/7 = 0.25 Hz or C0q = I jifo = 0.5jt rad/sec. As shown in die next section, u-^^{t) has the
infinite (Fourier) series representation
s in ((2 + ^
=0n
For this signal the average dc value is
15 ti. The
nil)
2// + 1
^ s) =
Vi,
. v+1
Step 3. D eterm ine the magnitude an d phase o f H{s) at s = j( 0 . For sinusoidal steady-state analysis,
set s =jo ) to obtain
Jd
H(jco) = -
JOJ + 1
=H/
in which case
, 6 = - tan ^(w )
yjco^ + 1
Step 4. Find steady-state responses to a ll components o f Vj^^{t). Table
20.1
0)
Wo
Input magnitude
157:
60
5o
20
12
0.5370
0.2075
0.1263
0 (degrees)
-57.52
-78.02
-82.74
Output magnitude
1571
32.22
4.150
1.516
Output angle
-57.52
-78.02
-82.74
Input angle'
Step 5. Apply superposition to obtain the steady-state portion o f the complete response. Neglecting har
monics o f scventii order and higher, the approximate steady-state solution is
Vg{t) = 15j*r + 3 2 .22sin((/jyr- 57.52") + 4.15 sin(3w Q /-7 8 .0 2 ) + 1.516 sin(5wQ - 8 2 ,7 4 ) + ...
This response shows the effect of the system on each component of the input signal and how it in
turn affects the overall steady-state output response. Since the time constant o f the circuit is 1 sec,
this steady-state response more or less constitutes the actual response for r > 5 sec.
Exercise. Use MATLAB to plot the approximate waveform for vj^t) for 0 < r < 8 sec, based on
Tiible 20.1. From the plot identif)'
In Example 20.1, the first step o f the solution was to represent a periodic waveform as a sum o f
sinusoidal components, called the Fourier series. Section 2 covers the definition and basic proper
ties o f Fourier series. Section 3 describes several shortcuts for computing the Fourier coefficients
and identifies other important properties. Since, in practice, only a finite number o f terms can be
considered, the Fourier series method yields only an approximate solution.
Because many mathematical and engineering handbooks have extensive tables o f Fourier series o f
different waveforms, it is convenient to use these tables in much the same way as one uses a table
o f integrals or a table o f Liplace transforms. The Fourier series o f some basic signals are provid
ed later, in Table 20.3. The use o f this table, together with some properties and shortcuts dis
cussed in section 3, make the study o f the Fourier series method much more palatable to begin
ning students o f circuit analysis.
(20.2)
r i s the period o f the signal. The fundamental period is the smallest positive real number Tq for
which equation 20.2 holds; ^ = l/7j, is called the fundamental frequency (in hertz) o f the sig
nal; cOq = I k/ q = 271/7*0 is the fundamental angular frequenc)^ (in rad/sec). T he sinusoidal wave
form o f an ac power source and the square wave form used to clock a digital computer are com
mon periodic signals. Figure 20.2 shows a portion o f a hypothetical periodic signal.
1089
(20.3)
Equivalently,
/ (O =
coa{mor + d)
(20.4)
where
(20.5a)
and
(20.5b)
Both infinite series, equations 20.3 and 20.4, are called the trigonometric Fourier series repre
sentations
We note that equations 20.5a and 20.5b follow fi-om the trigonometric identit)'
A cos(.v) + fisin(.v) =
0 j. The term d-, cos(2o;(jt + 6-,) is called the second harmonic o^J{t), with amplitude d-y and phase
angle
and similarly for the other terms d^^cos(;/WQt + 0^^), which are the ;/th harmonics as indi
As illustrated in Example 20.1, given any periodic function/r), it is important to determine the
coefficients in equation 20.3 or equation 20.4. For the purpose o f easier calculation, it is advan
tageous to introduce the equivalent complex Fourier series representation o f the periodic func
tion y(f).
/ (0 =
(20.6a)
n = -ao
where it can be shown that
In + T,
c =
(20.6b)
Since we have two (allegedly) equivalent forms, let us now develop the relationship between the
real and complex forms o f the Fourier series. Recall the Euler identities:
cos( a-)
/ (')=
/I = - oc
/I = I
= -y + ^
n = \\
= ^
+ I
^----------- + h
(0.5(.; - jh
2j
+ 0.5( + jb
/! = 1
Equating coefficients yields
=
(20.7a)
and
c_ = 0.5
+ jh ) = c*
(20.7b)
(20.8a)
(20.8b)
1091
(20.8c)
^0 = ^0 = ^
(20.8d)
= ^ c ,n = 1, 2, 3 ,...
(20.8e)
In the real trigonometric Fourier series equations 20.3 and 20.4, the summation is over positive
int^er values o f, whereas in the complex exponential Fourier series the summation extends over
int^ers n such that -oo < < oo. While each term in equation 20.3 or equation 20.4 has a wave
form displayable on an oscilloscope, each individual term in the complex exponential Fourier
series lacks such a clear physical picture. However, two conjugate terms in the complex exponen
tial Fourier series always combine to yield a real-time signal d^cos{no)Qt + 6^.
To develop equation 20.6b for the coefficient
r > " 0 'to obtain
* = -00
Jto
k = -oo
(20.9)
* = -00
it = -00
'v ^
'
'
0
cTq
for k ^ n
for k = n
(20.10)
tn + T
Tn
(20.11)
The lower limit of integration, ?q, can be any real number, but is usually chosen to be 0 or - T qH,
whichever is more convenient. In addition, Tq will sometimes be written simply as T.
A hand computation of coefficients would proceed by first computing for = 0, 1,..., using equa
tion 20.11. One would then obtain and using equations 20.8a and b, and and 0^ by equa
tions 20.8c, d, and e. Other formulas are available for obtaining the Fourier coefficients by integrals
involving sine and cosuie functions. However, equation 20.11 is preferred because an integration
involving exponential fimctions is often simpler than an int^ration involving sinusoidal flmctions.
y
w
EXA M PLE 2 0 .2 . Find the trigonometric Fourier series for the square wave signal o f Figure 20.3.
S olution
The fundamental period o f/ f) is Tq = T. By inspection, the average (dc) value
is
(20.12)
A .(
JT^
sin n , / ; = l , 2 , . . .
2}
Jtn
Thus our signal has the complex Fourier series of equation 20.6 with
jTn
Substituting these coefficients into equation 20.3 yields the following trigonometric Fourier series
for the square wave o f Figure 20.3:
A 2A /
I
I
/ ( 0 = T *' c o s ( o ; o / ) - - c o s ( 3 a ;o O + 7 C o s ( 5 a ;o / ) - .. .
2
7T \
3
5
(20.13)
1093
a horizontal and/or vertical movement o f the plot without any rotation. A translation o f the waveform
in the vertical direction causes a change in the dc level and affects only the coefficients
d^y and Cq.
A translation o f the waveform in the horizontal direction causes a time shift that changes only the
angles 0,^ and has no effect on the amplitude
Fourier series o f a periodic functiony(f), and c be those for another periodic function g{t).
If^ f) is a translation o(J{t) consisting o f a dc-level increase
right by
then
g it)= p -r J)^ K
(20.14a)
^ 0= do = CQ + K = dQ + K = + K
/; = 1 , 2 , . . .
=1
= Z.C,,
, /? = l 2 , . . .
(2 0 .l4 b )
(20.14c)
(2 0 .l4 d )
(2 0 .l4 e )
The proof o f this property is straightforward and is left as an exercise. Note that equation 2 0 .1 4c
indicates that a time shift o f the signal affects only the phase angles o f the harmonics; the ampli
tudes o f all harmonics remain unchanged.
Exercise. Supposey(f) in Example 20.2 has A = 30 and suppose ^ f) =J{t) - 10 . Find the coeffi
c i e n t s a n d
o f the Fourier series o f ^ r).
AN SW I'R:
= 0.5^/,, = 3; all other coefficients arc unLh.mgcd, i.e.,
= sin n
because h.. = 0.
We will now use the translation property to obtain the Fourier series o f a square wave that is a
translation o f Figure 20.3.
EXA M PLE 2 0 .3
The square wave g{t) shown in Figure 20.4 is anti-symmetrical with respect to the origin. Find
the complex and real Fourier series for^^) given the Fourier series (equation 20.12) ofJ{t) depict
ed in Figure 20.3.
u rn
g(t)
>
'
A/2
w
-0.5T
-I
0.5T
-A/2
FIG URE 20.4 A square wave anti-symmetrical with respcct to the origin.
S olution
The curve
S ( 0 = / ((-/ ,/ ) + a: = /
By equations 20.14 and 20.13,
COS ncoQt
(
= cos ncoQf - n
\
2}
jr 1
2A
7Z
-----CO.S 3W(,/ 3
2}
?>7Z
+ - COS
5
5 jt
(20.15)
For the case o f a square wave signal, by choosing the time origin and dc level properly, the result
ant plot displays a special kind of symmetry that results in the disappearance o f all sine terms or
all cosine terms. The square waves o f Figures 20.3 and 20.4 are special cases o f the periodic func
tions amenable to such simplifications. The general case is given by the following statement.
Symmetry properties o f the Fourier series
(1) If a periodic functiony(/) is an even function, i.e.,
has only cosine terms and possibly a constant term.
cos(cl)/)
and the square wave o f Figure 20.3. The plot o f an odd function is anti-sym-
metrical about the vertical axis. Examples o f odd functions are sin(o)/) and the square wave o f
Figure 20.4. The proofs o f the symmetry properties are left as homework problems.
1095
Exercise. Suppose
the coefficients
and
AN SW ER:
lA
(h) = (i{) = 0. and hir n ^ 0. d = a = -------sin
.111
-T\
n .h = 0 .
~)
To simplify the calculation o f the Fourier coefficients, we should attempt to relocate the time ori
gin or change the dc level so that the new function ^t) displays even or odd function symmetry.
This may not be possible for an arbitrary periodic signal. When it is possible, we will calculate
the Fourier coefficients o f the new function ^t), which has only cosine terms or sine terms, and
then use the translation propert)' to obtain the Fourier coefficients for the original function//) .
A waveform o f particular importance in signal analysis is the periodic rectangular signal shown in
Figure 20.5. The fundamental period is T, and the pulse width is fiT. The constant (5 is called
the duty cycle, usually expressed as a percentage o f T. The square wave o f Figure 20.3 is a spe
cial ca.se o f the rectangular wave o f Figure 20.5 with a 50% duty cycle.
EXA M PLE 2 0 .4
Find the trigonometric Fourier series for the rectangular waveform o f Figure 20.5.
S olution
T he procedure is almost identical to that used in Example 20,2 for a square wave. By inspection,
the average value
is
(20.16a)
To calculate the other Fourier coefficients, choose fg = -7 7 2 . Equation 2 0 .10 then yields
pT
fW
~E.
= sin(/7/3.T), /? = 1 ,2 ,...
Tin
( 2 0 . 16b)
10 9 6
(20.16c)
jzn
b.. = 0
(2 0 .l6 d )
Specifying the coefficients completes tine determination ot the Fourier series, i.e.,
2A ^ (sin(nl^jr)\
) = p/\ 1------- y
(1,1
n= \
COS(nwt)
/I = l^
A very important conclusion about the rectangular wave can be drawn by examining equations
20.16: as the ratio o f the pulse width to the period becomes very small, the magnitudes o f the
fundamental and all harmonic components converge to twice the average (dc) values. To see this,
recall that sin(.v)/x approaches 1 as .v approaches 0. From equation 20.16c, we may rewrite
a
'n = sin(///i.;r)= 2ftA
jTn
It follows that
2fM as ft
as
nftji
0.
To give some concrete feel to this property, lable 20.2 gives the ratios o f
^'(ly
= and
^av
for w = 1, ..., 9 for the case o f ft = 0 .0 1 . Answers are rounded o ff to three digits after the decimal
point. Note that when the periodic rectangular signal is shifted vertically, the ratio
is affected, but the ratio
remains unchanged.
TABLE 20.2
A m p lit u d e s o f t h e F i r s t N in e H a r m o n i c s f o r t h e C a s e P = 0 .0 1
F
av
F
av
The constant
2.000
1.999
1.998
1.995
1.992
1.989
1.984
1.979
1.974
1.000
1.000
0.999
0.998
0 .996
0.994
0.992
0 .990
0 .987
property holds approximately true when the pulse width is a very small fraction
o f the period 7q. Even i f a waveform is not rectangular, if the pulse width is very small compared
109:
to its period, then the nearly constant propcrrv' o f ~ iind ^ continues to hold, as long as the
I
pulse is o f a single polarit)'. For example, consider the periodic short pulse shown in Figure
In calculating the Fourier coefficients
gral, originally (/q,
+ 7 ),
f;;
are changed to
(-(^ 7 7 2 ,
^ 7 7 2 ).
As
2 0 .1 1 ,
2 0 .6 .
in the integrand has a value very close to 1 in the new time interval, as long as n, the
harmonic order being considered, is not very high. Therefore, for pulses o f narrow width we have
(xT
aT
(20.17a)
rhus in terms of the
have
(20.17b)
This result is pertinent to the approximate analysis o f a rectifier circuit covered in other texts or
in the second edition o f this text.
To this point, we have calculated the Fourier coefficients only for some very simple periodic sig
nals. The evaluation o f the integral in equation 20.11 becomes much more involved when the
signal j{t) is not rectangular. Fortunately, many mathematical and engineering handbooks now
include comprehensive tables o f Fourier scries. From a utility point o f view, one may use these
tables much the same as one uses a table o f integrals or a table o f Laplace transforms. In effect,
the need to carrj' out the integration in equation 20.11 is not compelling in practice.
In many applications, it is important to know the average power o f a (periodic) signal and the
magnitude o f its various harmonics. From equation 20.5, the effective value o f the dc compo
nent is |<^q|, and those for the fundamental and various harmonics are d^j! yfl , /; = 1 , 2 , . . . . It is
easy to show that the effective value, or the rms value, o f/ r) is
(20.18a)
where the d- coefficients are from the Fourier series o f equation 20.4. If fit) (current or voltage) drives
a 1 ohm resistor, then the average power absorbed by the resistor is
^eff Hence we
say that the average power o f a periodic signal y(/) represented by a Fourier series is given by
1098
r~\
1
k|
(20.18b)
n=-oo
The relationships indicated in equations 20.18 are often termed Parsevals theorem.
The information on the phase angle is important when one wishes to construct the time domain
response in steady state. For the time domain problem, the Fourier series method yields only an
approximate solution, because one can only sum a finite number of terms in the series.
C on v erg en ce o f th e F o u r ie r S eries
Convergence of the Fourier series is an intricate mathematical problem, the details of which are
beyond the scope of this text. On the other hand, it is important to be aware of the ways in which
the Fourier series may or may not converge to a giveny(/). Our discussion is not comprehensive,
but is adequate for our present purposes.
To begin, we define a partial sum of terms of the complex Fourier series of a function ^(/) as
sn )=
k= -N
From our experience thus far, 5yy(f) must in some way approximatey(r). The difference between
j{t) and its approximation, 5yy(r), is defined as the error
^at( 0 = / ( 0 - 5 ^ ( 0
r\
e ^ (l) dt
This is often called the energy in the error signal, as energy is proportional to the integral of the
squared magnitude of a function. It turns out that for ftmaions having a Fourier series, the choice
of the Fourier coefficients minimizes E[^ for each N. Further, for such fiinctions,
0 as
00, i.e., the energy in the error goes to zero as JV becomes large. This does not mean that at each r,
fij) and its Fourier series are equal; it merely means that the energy in the error goes to zero.
Continuous and piecewise continuous periodic functions have Fourier series representations. A
piecewise continuous fiinction, such as a square wave, is a function that (1) has a finite number
of discontinuities over each period but is otherwise continuous, and (2) has well-defined rightand left-hand limits as the function approaches a point of discontinuity. For piecewise continu
ous functions, it turns out that the Fourier series converges to a value halfway between the values
of the left- and right-hand limits of the function around the point of discontinuity. Even so,
-* 0 as TV- 00 for piecewise continuous functions.
There are many functions that are not piecewise continuous and yet have a Fourier series. A set
of conditions that is sufficient, but not necessary, for a function to have a Fourier series repre
sentation is called the Dirichlet conditions.
r\
1099
D irichlet conditions
Condition 1. Over any period, [fg, tQ+T\,J{t) must have the property that
In the language o f matliematics, this means that J{t) is absolutely integrable. The conse
quence o f this property is that each o f the Fourier coefficients
exist.
Condition 2. Over any period o f the signal, there must be only a finite number o f min
ima and maxima. In other vv^ords, functions like sin(l/^) are excluded. In the language o f
mathematics, a function that has only a finite number o f maxima and minima over any
finite interval is said to be o f hounded variation.
Condition 3. Over any period, y(r) can have only a finite number o f discontinuities.
As mentioned, at points o f discontinuity, the Fourier series will converge to a value midway between
the left- and right-hand values o f the function next to the discontinuity. There may be other differ
ences as well. Despite these differences, the energy between the function
representation is zero; i.e.,
with N approaching oo, goes to zero. Thus, for all practical purpos
es, the functions are identical. This practical equivalence allows us to analyze how a circuit responds
to a s i g n a l b y analyzing how the circuit responds to each o f its Fourier series components.
j{t) has an analytic expression, then its Fourier coefficients can often be calculated from equation
20.11, The properties discussed below are o f great value in simplifying the calculation o f Fourier
coefficients. Their proofs are fairly straightforward and arc left as homework problems.
The linearity property: Letyj(/) and f^{t) be periodic w'ith fundamental period T. \^j{t)
= KJ'^{t) + K^fyU), then the Fourier coefficients o^j{t) may be expressed in terms o f those
of/j(f) and ^ (r) according to the following formulas:
(20.19a)
(20.19b)
and
(20.19c)
In general,
1 100
n s
This property allows us to easily obtain the Fourier series of a sum of periodic signals when the
Fourier series of the individual signals is already known.
DEFINITION
A periodic function y(/) is said to be half-wave symmetric if
-0.57) = -j{t) for all t
In words, y(r) is half-wave symmetric if a half-period shift of the plot combined with a flip about
the horizontal axis results in the identical function y(/). Some simple examples of half-wave sym
metric functions are sin(ft)r), cos{(Ot), and the square wave of Figure 20.4,
The half-wave symmetry property: A half-wave symmetric periodic function f j ) con
tains only odd harmonics.
The waveform of Figure 20.4 illustrates this property.
The dertvadve/integration proper^.
denote the ^
Ik)
c =
derivative of a periodic
(k)
(t) of satisfy
20a)
c<*)
/ _ y'
cos(jc)-cos(y) = - 2 s i n ( ^ ^ - j sin (^ -Y -j
(20.21a)
(20.21b)
!01
E X A M PLE 2 0 . 5 . Find rhe Fourier series for rhe periodic impulse trainy(^) shown in Figure 20.7.
jk
A
(A)
-21
A
(A)
k.
(A)
A
(A)
(A)
-T
2T
(20.22a)
T
for all n. Hence
Equation 20.22b states that, for a periodic impulse train, all harmonic components have magni
tude equal to twice the average value. This is the limiting case o f the short pulse propert)- stated
in section 2. The next example uses the derivative and integral properties o f the Fourier series to
develop an alternative derivation o f the Fourier series for a periodic rectangular pulse train, derived
earlier by the use o f equation 20.11.
E X A M PLE 2 0 .6 . Find the Fourier series for the periodic rectangular pulses^(f) shown in Figure
20.8a. (This corresponds to item 2 in Table 20.3.)
S olution
Figures 20.8a and b show ^(r) and its derivative,^ (t). The latter may be written as the sum o f
two shifted impulse trains:
{
f n( t ) = f6 t +
\
^ r
-fa
t-
/(5T
T \\
\ ~ r;
(2 0 .2 3 )
102
(A)
(A)
-T
T
2
i(-A )
(A)
-f
^
2
T
,,(-A )
>
(-A)
(b)
FIG U RE 20.8 (a) Periodic rectangular pulse crain and (b) its derivative. The parenthetical values,
{A) and {-A), next to each impulse in part (b) denote the weight of that impulse, i.e., its area.
Using the time shift property (translation in the horizontal direction), together with equations
20.22b and 20.21a,
f'p ( 0 = Y,
"
= I
2A
- cos{na)Qt - n^jv))
(20.24)
4/\
(sin(/7o;oOsin(/3;r))
^
(2 0 .2 5 )
1103
E XA M PLE 2 0 .7 . Find the Fourier series for the half-wave rectified sine wave fhsi^) shown in
Figure 20.9a (item 10 in Table 20.3).
A f.(t)
(0
FIGURE 20.9 (a) Half-wave rectified sine wave as the product of two functions: (b) a cosine wave
and (c) a square wave.
S olution
The periodic fiinction j{t) o f Figure 20.9a may be viewed as the product o f the sinusoidal wave
y4cos((0Qr) and the square wave^(r), shown in Figures 20.9b and c, respectively. Using the Fourier
series for^(^) given by equation 20.13, we have
//i.v(0= [ ^ c o s ( w o O lA ( 0
M
= y4cos(o;o0
9 /
1
1
\1
~ cos(ojo^) cos(3w()/) + cos(5o;o/) - ...
;r ^ 3
5
(20.26)
104
7z I
A
lA
= - + - cos(w()/) +
IT
- c o s ( 2 ft ;,)0 - cos(4coqi)
Id
JT \ J
T T -co.s( 2 mu/
4 ,r _ 1
(20.27)
harmonics.
E X A M PLE 2 0 ,8 . Find the Fourier series for the full-wave rectified sine wavej^(r) shown in Figure
20.10a (item 9 in 'I'able 20.3).
lOS
> t
FIGURK 20.10 (a) Full-wavc rcctificd sine wave as the sum of the two signals in (b) and (c).
S olution
One approach is to apply the same technique as in Example 20.7. Specifically,
is the sum o f the two waveforms shown in Figures 20.10b and c; i.e.,
(20.28)
lA
4/4
(20.29)
4 ;r -I
The derivative propert}' is particularly useful for tackling periodic piecewise linear waveforms.
Piecewise linear waveforms consist o f straight-line segments. From differentiating once, or at most
twice, impulses appear. The integration given by equation 20.11 is trivial if the integrand contains
a shifted impulse function. This fact, together with the derivative and integral properties, reduce
the task o f calculating the Fourier coefficients for piecewise linear waveforms to some complex
number arithmetic. These examples and several other commonly encountered periodic waveforms
have Fourier series as given in Table 20.3. Engineering and mathematical handbooks contain
much more comprehensive tables. O f course, when a waveform does not appear in a table, the
Fourier coefficients must be computed manually or numerically.
106
TABLE 20.3
107
= 0-5a/\
sin(2//o;r))
0.5
n (a + P)ji
b= 0
Tq = 0.5T,
f(t) =
(Oq =
2(0
^ ^ cos(2(o?)
3
4A
K
cos(4w0 ^
15
cos(6(o/)
35
(-l)"^'cos(2wO
7
4 fr-i
"*
Jl
/ (/ ) = - + - c o s (w o / ) +
n+1
v H L _
^ k
ave=A/n
rms=A/2
4n^ - 1
COS (inojQt)
108
4. SUMMARY
Given that many mathematical and engineering handbooks have extensive lists o f the Fourier
series o f common signals, this chapter has taken a practical approach to the calculation o f the
Fourier series and its application to circuit analysis. The idea is to use tables such as Table 20.3 in
the same way engineers have come to use integral tables. The keys to using such tables for the com
putation o f the Fourier series o f a waveform are the various properties that allow one to convert a
known series into one that fits a new waveform. The idea here is to express the new waveform as
a translation of, a linear combination of, a ^-fold derivative of, or a ^-fold integral o f signals with
known Fourier series as in Table 20.3, or any mixture o f these operations. The Fourier coefficients
o f the new signal can then be expressed in terms o f the Fourier coefficients o f signals with known
Fourier series.
Knowledge o f the Fourier coefficients o f a signal such as the output o f an audio amplifier allows
one to investigate phenomena including the distortion introduced by the amplifier. In the case o f
a dc power supply, such knowledge allows us to characterize the degree o f unwanted ripple in the
output o f a rectifier circuit. In addition, the application o f Fourier series plays an important role
in the computation o f steady-state circuit responses to periodic input signals.
nent ofy(r).
Derivative/integration property: Let/^^H^) denote the ^ h derivative o f a periodic function y(^).
Then the Fourier coefficients o f the ^ h derivative,
and conversely,
satisf}'
0.
D uty cycle: for a rectangular signal having fundamental period T as illustrated in Figure 20.5, the
duty cycle is the constant (3 that determines the pulse width ftT.
Effective value (rms value);
|/l=-cc
coefficients are from the Fourier series o f equation 20.4 and the
20.6b.
Even function: y(/) =
Fundamental com ponent (first harm onic): the first term under the summation sign in equa
tion 20.4,
Fundamental frequency (in hertz): /o = . where T^^ is the fundamental period. Note:
_
2:7r .
^0
Wq = 2;r/() = is the fundamental angular frequency in rad/sec.
Fundamental period: the smallest positive real number
1109
harmonics.
Lineari^ property: let^(z) and^(/) be periodic with fundamental period T.
= ATj^(/) +
A!^(r), then the Fourier coefficients o^j{t) may be expressed in terms of those
and
fj{t) according to the formulas c^ =
+ K2C2,
= ^\^\n
In general,
the angle 0^ is the same for all n.
Odd function:/r) =
Periodic signal, y(f): A signal whose waveform repeats every T seconds. Mathematically, for some
+ 7)
Symmetry properties of Fourier series: (1) If a periodic function ^(/) is an even function, then
its Fourier series has only cosine terms and possibly a constant term. (2) If a periodic
function y(/) is an odd function, then its Fourier series has only sine terms.
Translation of a plot: horizontal and/or vertical movement of the plot of a function without any
rotation.
Translation proper^ of Fourier series: if fit) is a translation of J{t) consisting of a dc level
increase K and a time shift (delay) to the right by then ^/) = j { t +K. See equa
tion 20.14.
Trigonometric Fourier series: representation of a periodic signal f j ) in terms of sines and
cosines, as given in equations 20.3 or 20.4.
10
ANSWF^R: (a)
^ ^ r o b le m s
= - 0 .7 2 1 3
0.0795
FOURIER SERIES
COEFFICIENTS BY
INTEGRATION
cos(4.t /- 8 6 .8 4 ')
-0.5 T
f(t)
0.5 T
(1)
-2
(b)
(c)
/ f) = [2 + 1 .5 sin (5 0 0 r)
(a)
A^ f(t)
(-1)
>r
and
tions
(a)
(-1)
>r
Figure P20.3
(1)
+ ...
>^ f(t)
-2 -1
1
+ / 2 ; / . t ) ; (b) fit)
P20.3.
-3
(1)
= 0 .5 /(Io g ^ .
+ 0 . 1 5 8 c o s ( 2 . t / - 8 3 . 7 ' ) +
1 2
(-1)
>r
>
(b)
Figure P20.1
- 2 cos(2000r)]cos(10^V)
20.22b for/^(r).
Note that f{ t ) = A -
Use equation
by carrying out
and 6^ o f the
Figure P20.2
1111
-1
Figure P20.6
-t
10.
before.
result o f Problem 9.
>
Figure P20.10
8. Consider the isosceles triangular wave shown
in Figure P20.8, which is item 4 ofTable 20.3.
Fourier coefficients
impulse trains.
amount.
f(t)
> t
Figure P20.8
12. Find the Fourier series coefficients Tq, cp
9. Consider the clipped sawtooth wave shown
in Figure P20.9, which is item 6 ofTable 20.3.
1 112
\H{joj)\ = 10
and
-2
Figure P 20.12
/LH ( jco) = -
5000
Express
(a )
SINUSOIDAL STEADY-STATE
ANALYSIS WITH MULITFREQUENCY INPUTS
X 45^'
frequency components.
(b)
Express
(c )
frequency components.
If
is expressed in the form
+ 8 0 V 2 co s(5 x 377/ + 5 0 ^ )V
= 78.54 a
is an ampli
Find
(b)
at steady state.
lope o f
and the
the amplification is 10
jy r \
v(t)
10H
4 pF
Fig P20.14
Figure P20.13
ANS\X'ER; (a)
0.1 cos((<o
+
r.... (M = 200 + >/2
42.6cos((o/ - 1 7 5 .4 ") +
+ 0.1 cos(((/^ +
2 cos((oj. -
= 0.2 cos(((/;. -
10
Jt) +
+ 2 cos(a;^./) + 0 .2 c o s ((o ;.
V; (b)
+ 4.5) + cos((;. -
y 9) + 20 cos(oj^j') + 2 cos((o>. +
=
+
-
\Un{l) n , and
and ;; = 0 , 1 , 2 ,
(a)
Show that the response due to the first
J{t)
is a sequence o f impulse
13
(b)
'w '
(c)
(d)
(e)
^Ollt ( 0 =
(b)
(c)
----- u
(a)
,I
ijt )
(Q)
(Q)
-T
1r
hnis)
2T
Figure P20.15
Consider again the circuit of Figure
(a)
C5 H----
(Q)
(b)
16.
H {s) =
(d)
VJt)
RC
_Z_
= ! = I.4 4 2 Q
M 2)
\ - e RC
O
+
C
\ -e
RC
14
2 h /lC
Tq - = 2vtVZc .
Jo
value.
(a)
o f poles at
CJand
interval o f length
Wq =
2ji
Tj = ---- ) differs
(X)j
2m + 1
then /^(/) is periodic. Sketch t\vo
then the
CJ
in terms o f R, L, and C.
(d)
^= -
sin(w^r).
(b)
(b)
cycles of
for (;/ - 1)
(c)
(a)
>I
>
(Q)
(Q)
-T
< r < ;/
(Q)
2T
(b)
Figure P 20.17
18. The LC resonant circuit in Figure P 2 0 .l8 a
is initially relaxed and the impulse train
iin(0 = Q ^ H r - n T )
/i=(i
If ^ = 0 .7213 a
C = 1 F, and L =
1115
(a)
12.1)
'.,W = 2<'-3I<[cos(2 w ) -
1/
I V
''max - V*mm
^oul,mm ~ *min
o .5 7
L,
\+ e
I f /? = 9.747 a , C= 1 F, I = 0.02533
H, Q = 1 coulomb and T = \ sec, in
which case a = 0.95, find the steadystate response for 0 < ^< 1 sec.
V
.
= V^max ^out,max
(b)
(c)
u (t)
V
f
' nt/TT - V' tff
0.5T
\+ g RC
(a)
U t)
ik
(C8
(Q)
-T
(Q)
2T
(a)
(b)
.v jt)
V
Figure P20.18
CHECK: (d) //(5) =
's-^
(e)/ / (5 ) = _ ----------
fflin
r + 1.3863^ + 39.96
s
r +0.102586j + 34.481049
and
= 20tf*5^'[cos(2;) 0.0082 sin(2;r^)] V for 0 < ^< 1
19. Consider the first-order RC circuit of Figure
P20.19a, where the input v-JJ) is a square wave
with period T, as shown in Figure P20.19b.
From the analysis methods discussed in Chapter
8, the output
consists of exponential rises
and decays, as illustrated in Figure P20.19c.
...
-0.5T
0.5T
0.5T
(b)
(c)
Figure P20.19
1.5T
16
= 0.25, v^Q-) =
V, and
(a)
= 1
H(s) =
/ [Output(/)]
H{0)
/ [InputC/)]
1+T5
(b)
Output(/),, = H{0)
V
* max - V^min
(c)
l+ e ^
and
0
0.57
''max - V* mm
utput(0 ,av =
is a 1 kHz
H{s) =
RC
s+
RC
(a)
(b)
values o f
Problem 20.
(c)
item 2
(e)
50 kO
Figure P20.21
A.\S\X ER,S: (a)
H (s) =
-S
i()()0.v +
: ( b ) 0 .2 4 5 V
R.
'. W
R.
v jt )
' ' 6
O '
Figure P20.23
(a)
Avjt)
pT
> t
T+pT
(b)
Figure P20.22
23. Consider the second-order circuit of Figure
P20.23, in which = Ifl, /?2 = 8 Q, /?3 = 6 2,
Cj = 0.25 F, and
= 0.125 E
(a) Derive the transfer function
16
3
r + 7J + 10
If the input voltage is a square wave
shown in Figure 20.23 with peak-topeak voltage 18 V and T = \ sec, find
the Fourier series for
and list the
first four components.
(c) Find the approximate average power
delivered to R2 considering only the
first four components.
(d) Assuming the input has been applied
for a long time so that the circuit is in
steady state,
find the expression for
during
the time interval (0, 7). Use MATLAB to plot
the waveform. Hint: Expand H{s) into par
tial fractions. Represent the system as the paral
lel connection of two first-order subsystems.
Then use the result of Problem 17 to solve each
subsystem. Finally, combine the solutions in
time domain to obtain the expression for
(b)
Vw/
------ o
120
Index
121
Index
applications, 6 4 5 - 6 5 2
principle of, 3 0 6 , 6 4 5
Conservation o f complex power in AC
Current, 2 , 4 0
alternating, 5, 8 - 1 0
branch, 107, 110, 128
circuits, 515
Conservation o f power (energy), 19, 2 0 , 3 6 ,
4 0 ,5 1 5 . 5 3 1 ,5 3 7
Constant inputs, second-order linear
networks with, 4 0 0 411
Constant-resistance network, 6 4 0
Continuity property o f the capacitor, 2 8 7 ,
306
direct, 8 - 9 , 3 0 , 33
loop, 110, 128, 141
mesh, 128, 141
Current division, 8 9 , 4 8 2 , 6 0 4 , 6 5 8
formula, 6 8 - 7 3 , 4 5 8
Current source, 4 0 , 5 6
nonideal, 8 5 - 8 8
v-i characteristics, 56
Current transformation property, 9 1 5
306
's ^
Controlled source, 2 8 , 2 9 , 7 8 , 80
D C (d irea current)
defined, 41
Convolution, 7 6 3 - 8 0 9
averaging by finite time integrator
circuit, 7 6 3
7 3 6 -7 3 9
257
O '
defined, 7 6 6 , 7 9 7
graphical convolution, 7 7 4 7 8 1 ,
by integral, 7 9 9 - 8 0 2
Dielectric, 2 8 4 , 3 0 6
Difference amplifier, 1 6 3 -1 6 4 , 167, 177
Differential equations, 3 6 4
time-invariant circuits, 7 9 3 - 7 9 7
Convolution algebra, 7 8 1 , 7 9 7
ordinary, 591
series-parallel R L C circuit, 3 8 6 - 3 8 7
C o m er (break) frequency, 7 4 0
Coulom b, 3 , 3 0 6
Coupled circuits.
circuits
Coupled inductors, models for pair of, 9 3 0 ,
937
Coupled inductors (coils), 8 8 6 - 8 8 7 ,
9 3 0 -9 3 6 , 937
Coupling network, SSS calculation and, 52 9 ,
530
Cram ers rule, 140
8 0 7 -8 0 9
'Sw/
257
Decade, 7 3 3 , 7 4 0
convolution theorem, 7 7 3 , 7 9 8
8 0 2 -8 0 4
Decibel (dB ), 7 3 3 ,7 4 0
convolution algebra, 7 8 1 - 7 8 6 , 8 0 4 - 8 0 6
Diodes, 3 0 3
Dirac delta function, 591
Dirichlet conditions, 1 0 9 8 -1 0 9 9
Distributivity, 7 6 6 , 7 9 8
D ot notation, 5 4 6 , 5 9 1 , 9 0 5 , 9 3 7
mutual inductance and, 8 8 6 - 8 9 3
Driving point admittance, 6 2 6
Driving point impedance, 6 2 6
D uty cycle, 1095, 1108
122
Index
circuit, 2 , 1 4 -1 5
low-pass approximation, 1 0 7 4 - 1 0 7 6
1 0 7 3 -1 0 7 4
passive low-pass realization, 1 0 7 6 -1 0 7 9
singly terminated Bunerw orth low-pass
networks, 1 0 6 9 -1 0 7 1
types of, 1 0 3 2 -1 0 3 4
net e n e i^ , 2 7 9
Final-value theorem, 7 2 6 - 7 2 9 , 7 4 0
n o n -D C power and, 2 2 - 2 5
Finite pole, 6 8 5
potential, 16
stored, 9 1 0 - 9 1 4
Finite zero, 6 8 5
Equilibrium state, 3 3 9 , 3 6 4
Equilibrium value, 3 5 4
Equivalent resistance, 6 5 , 6 9 , 70
r>
networks and, 4 1 2
Euler identity, 4 3 6 , 4 8 2
response classifications, 3 5 2
Eulers formula, 3 8 9
sawtooth, 3 2 2 - 3 2 3 , 3 6 5
r>
Equivalent circuits, 6 1 8 - 6 2 5
Equivalent networks, 2 1 2 - 2 1 3
r\
circuits
Equivalent, defined, 2 5 7
Equivalent conductance, 6 9
Exponential decay, 3 9 2
Fluorescence, 5 1 8 , 6 0 3 - 6 0 4 , 6 2 3 - 6 2 5
Farad (F), 2 8 4 , 3 0 6
Forced response, 3 5 2 , 3 6 4 , 7 0 0 , 7 4 0
Femto, 2 9
Forcing function, 3 3 8
Fields, 11, 12
Fourier series, 1 0 8 5 -1 1 1 7
5 7 6 -5 7 9
Filtering, 1 0 3 1 -1 0 8 4
active low-pass design, 1 0 7 9 -1 0 8 1
active realization o f high-pass filters,
1 0 5 9 -1 0 6 2
a a iv e realization o f low-pass Butterworth
filters, 1 0 4 5 -1 0 5 1
band-pass filter design, 1 0 6 2 -1 0 6 9
Butterworth loss function properties,
1 0 5 2 -1 0 5 3
Butterworth passive realization,
1 0 4 2 -1 0 4 4
1110-1112
1123
Index
Four-terminal device, 80
Frequenqr, 4 3 3 ,4 8 2
T seconds, 7 8 9
Ideal insulator, 41
magnitude, 1 0 3 3 , 1072
Ideal integrator, 3 6 0
Frequency s ^ i n g , 5 8 2 - 5 8 3 , 7 4 0
Laplace transform m ethod and, 5 8 2 - 5 8 3
m agnitude scaling and, 7 1 4 - 7 2 6
Frequency shift property, 5 7 5 - 5 7 6 , 591
Fundamental com ponent (first harm onic),
10 8 9 , 1108
Fundamental frequency, 1088, 1108
O '
o f the capacitor, 4 5 2 , 6 0 8 , 6 5 8
Fundamental period o f p e r i o d i c 7 4 0
defined, 6 0 6 , 6 5 8
Fuse, 3 2
Gain
in d B , 10 3 3 . 1072
power, 8 0
voltage, 8 0
G ain magnitude, 10 3 3 , 1072
phasor impedance, 4 5 4 , 4 5 5 - 4 6 1
Generalized conductance, 6 0 6
o f resistor, 6 0 7 - 6 0 8 , 6 5 8
w
w
o
H enry (H ), 2 7 3 , 2 7 4 , 3 0 7
H enry (H ), 2 7 3 , 2 7 4 , 3 0 7
H i^ -p a s s circuit, 4 6 9 , 4 8 2
Inductors, 15, 2 6 9 - 3 1 9 .
also Filtering
H P (high-pass) to LP (low-pass) f l u e n c y
transform ation, 1 0 5 3 , 1072
Hybrid (A-) parameters, 9 6 1 , 9 8 5 - 9 9 1 ,
1010
Impulse response, 6 2 5 , 6 3 1 - 6 3 4 , 7 1 9 , 7 4 0
Half-power frequencies, 8 1 5
two-ports and, 9 7 6 - 9 8 1 , 9 8 1 - 9 8 4
Impedance (z-) parameters, 9 6 1 , 9 7 6 - 9 8 1 ,
Half-power points, 1 0 3 3 ,1 0 7 2
applications, manipulation, 6 0 9 - 6 1 7
Laplace transform m ethod and, 5 4 4
O '
applications, 6 0 5 - 6 0 9 , 6 5 9 - 6 6 2
Laplace transform analysis circuit
1 0 1 0 , 1 0 2 0 -1 0 2 3
Ideal banery, 84
in parallel, 2 9 5 - 2 9 7
Ideal co nd u aor, 3 , 4 , 41
24
Index
n
phasor relationships for resistors,
capacitors, and, 4 4 9 - 4 5 5
series and parallel inductors and
statements of, 6 1 , 6 3
capacitors, 2 9 3 - 3 0 2
Initial-value theorem, 7 2 6 - 7 2 9 , 7 4 0
Ladder network, 2 0 6 - 2 0 8
Input admittance, 9 7 4 , 10 1 0
Lagging, 4 5 1 , 5 1 8 , 5 2 6
Input attenuation, 1 0 5 1 - 1 0 5 2
Instability, 3 6 2
basic signals, 5 4 8 - 5 5 4
5 0 0 - 5 0 5 , 5 3 2 , 533
Insulators, 3 , 41
In t^ ra tin g factor m ethod, 3 2 7 , 3 6 4
Integration, Laplace transform m ethod and,
5 7 9 -5 8 0 , 5 7 9 -5 8 2
Integrator, 3 5 9 , 3 6 4 , 6 2 9 , 6 5 8
basics, 5 4 3 - 6 0 2
Instantaneous power, 2 2 - 2 5 , 4 1 , 2 7 9 , 3 0 7 ,
Instantaneous stored e n e ^ , 2 8 0 , 291
/n
representation, 5 9 2 - 5 9 5
circuit responses, applied to
r>
o
differential equations, 6 0 0 - 6 0 2
frequency shift property, 5 7 5 - 5 7 6
integration, 5 7 9 - 5 8 0
inverse Laplace transform, 5 6 5 - 5 7 5
ideal, 3 6 0
leaky, 3 5 9 , 3 6 4
one-sided, 5 5 4 - 5 6 3
Int^ro-difFerential equations, 5 8 5 - 5 9 0
overview, 5 4 7 - 5 4 8
Internal resistance, 6 4
properties, 5 8 4
properties, finding, 5 9 5 - 5 9 8
Joule, 12
5 4 4 -5 4 7
solution o f integro-difierential
KUo, 2 9
time-/frequency-scaling property,
equations, 5 8 5 - 5 9 0
5 8 2 -5 8 3
Vi relationship o f capacitor, 5 8 0 - 5 8 2
circuit applications, 6 0 3 - 6 8 1
design o f general summ ing
in t^ rators,
652-^57
capacitors and, 2 8 7
manipulation, 6 0 9 - 6 1 7
nodal and loop analysis in x-domain,
6 3 4 -6 4 0
op amp integrator design and, 6 8 0
explained, 6 0 - 6 4
defm itions, 5 3 - 5 4
first order R L and R C circuits, 3 4 8
r\
5 3 - 5 5 , 6 0 - 6 4 . 90
fluorescence, 6 0 3 - 6 0 4 , 6 2 3 - 6 2 5
r\
conditions, 6 6 8 - 6 7 6
sawtooth waveform generadon and,
6 8 0 -6 8 1
112^
Index
switching problems, 6 7 6 - 6 7 8
transfer functions, 6 2 5 -6 3 0 , 6 6 2 - 6 6 8
convolution and, 7 7 2 - 7 7 4
Loading effect, 7 6
Loop analysis, 1 0 7 -1 5 4
concepts of, 1 1 0 -1 1 1
explained, 1 2 8 -1 3 9
constant inputs, 4 0 0
transfer function applications, 6 8 3 -7 6 1
applications and bode techniques.
7 6 0 -7 6 1
classification o f responses, 6 9 3 -7 0 1
D C m otors and, 7 3 6 - 7 3 9
electric m otor analysis and, 6 8 3
poles, zeros, 6 8 5 - 6 9 3 , 7 4 2 - 7 4 4
Lossy device, 3 0 7
7 0 1 -7 0 7
stability problems, 7 4 5 - 7 4 6
steady-state calculation, 7 4 9 - 7 5 3
L C circuit, 3 7 9 , 3 8 0
See also
See also
Filtering
V J
frequency response, 7 0 7 - 7 1 4 ,7 5 3 - 7 5 9
scaling, 7 1 4 - 7 2 6
1 4 6 -1 4 8
general SSS analysis, 4 9 2 - 4 9 4
history of, 1 0 7 -1 0 8
bode plots, 7 3 0 - 7 3 6
switching in R L C circuits, 6 4 0 - 6 4 5
3 8 4 ,4 1 7
basics of, 1 0 3 4 -1 0 3 6
low-pass filter (brickwall) specification.
1034
parallel, 385
undriven, 4 1 8
1072
Leading, 451
Leaky integrator circuit, 3 5 9 , 3 6 4 , 5 8 8 - 5 9 0
A C adaptors and, 8 8 3
Linear circuit, 7 4 0 .
transformation; Superposition
transformers, 9 5 0 - 9 5 6
9 0 1 -9 0 9
analysis o f circuits containing ideal
applications, 9 5 6 - 9 5 7
secondary, 8 9 5 - 9 0 0
analysis of, with terminated secondary.
calculations, 9 0 9 - 9 1 4
com puting equivalent inductances.
1126
Index
r\
diflferential equation, Laplace transform,
phasor models of, 8 9 3 - 8 9 5
dot placement,
9 3 9 -9 4 0
Modulus, 4 3 5 , 4 8 2
M ultiplication-by-r property, 5 5 9 - 5 6 0
Mutual inductance
r^
ideal transformers, 9 1 4 - 9 2 4
magnetically coupled, defined, 883
Nano, 2 9
417, 687
Natural modes o f vibration, 3 2 6
8 8 6 -8 9 3
M agnitude (modulus), 4 3 3 , 4 8 2
M agnitude response, 7 0 8 - 7 0 9 , 7 4 0
Network equivalency, 3 7 3
7 4 0 -7 4 1
Networks
r s
equivalent networks, 2 1 2 - 2 1 3
M A TLA B
two-terminal networks, 2 0 9 , 2 1 2 , 2 1 5 ,
2 2 8 ,2 5 7
defined, 110, l 4 l
r>
M atrix notation, 6 3 6
applications, 6 3 4 - 6 4 0
and, 461
r>
terminology, 1 0 9 -1 1 0
Node, 54, 90
Node voltage, 6 0 , 9 0 , 109, I 4 l
Maxwells equations, 2 7 3 , 3 0 7
Mega, 2 9
M em ory
capacitors and, 2 8 6
inductors and, 2 7 5
Memoryless device, 41
Normalized circuits, 7 1 4
Metastable, 6 9 0
Normalized fi^quency, 1 0 3 5 - 1 0 3 6
m ho, 3 0 , 4 0 , 41
N orton, E . L , 2 2 7 , 1069
M icro, 29
N orton theorem, 2 2 7 - 2 6 8 .
MilU, 2 9
Modified loop equations, 136
Modified n o ^ analysis (M D A ), 110, 123,
Nonlinear, 8 0
3 7 9 -3 8 0
r\
Mesh, I 4 l
140, 141
r\
2 4 9 -2 5 6 , 257, 5 3 9 -5 4 2 .
r>
See also
1127
Index
Overdamped circuit, 3 8 9 , 3 9 2 , 4 1 7
Parallel, 3 5 , 37
Parallel circuit, 90
Parallel connection, 53
two-ports and, 1 0 0 2 -1 0 0 9
Parallel resistance, 6 8 - 7 3
Parallel resonant circuit, 8 4 2 - 8 4 6
Parallel R L C circuit, 8 3 3 - 8 3 8
op amp circuits, 2 4 6 - 2 4 9
SSS analysis, 4 9 1 - 4 9 2
two-ports and, 9 6 4 - 9 6 8
Octave, 7 3 3 , 741
Passive circuit, 3 3 9
O hm , 2 9 , 41
O hm s law, 3 0 - 3 7 , 41
linearity, superposition, source
transformation and, 191
phasor relationships for resistors,
inductors, capacitors, 4 4 9 - 4 5 5
O ne-port networks, 9 6 1 , 9 6 2 - 9 6 8 ,
1011-1012
O n the T heory o f Filter Amplifiers
(Butterw orth), 1034
theorems for, 2 3 6
N ortons theorem for, 231
Thevenins theorem for, 231
Passive R L C circuit, 3 6 5
Passive sign convention, 1 5 - 1 7 , 2 2 , 2 3 , 27,
3 0 ,4 1 , 54
Passivity principle for inductors, 9 0 9
Period o f waveform, 4 3 3
Phasors.
defined, 4 8 2
phasor analysis, 4 3 5
idealized, 1 5 7 -1 6 6
capacitors, 4 4 9 455
representations o f sinusoidal signals,
4 4 7 -4 4 9
steady-state circuit analysis using,
4 6 1 -4 6 6
Photo timer, 3 7 7
Pico, 29
Oscillation
oscillator circuit, 3 8 5 , 4 1 7
sinusoidal, 378
Polar coordinates, 4 3 6 , 4 8 2
W ien bridge, 4 1 2 , 4 1 4 , 4 7 5 - 4 7 6
Poles, 5 5 7
O pen circuit, 3 0 , 3 7 , 41
coupled circuits and, 8 9 5 - 9 0 0
'< J
1128
Index
R C circuit.
function applications, 6 8 5 - 6 9 3 ,
Reactance, 4 5 6 , 4 8 2 , 8 4 3
7 4 2 -7 4 4
Reactive power, 5 1 0 - 5 1 5 , 5 3 2
pole frequency, 8 1 9
Real part, 4 3 5 , 4 8 2
pole Q. 8 5 2 , 861
Real power, 5 3 2
pole-zero, 8 1 6
Reciprocity
o f rational fiinction
fmite, 5 9 1 ,6 8 5 ,7 4 1
simple, 5 9 1 , 741
repeated, 5 6 9 - 5 7 2 , 6 8 5
Potential difference, 12, 16
Power, 1 6 - 2 5 , 4 1 .
state (SSS)
Resistance, 3, 3 0 , 4 1 , 4 8 2
converter, 2 4 7
complex, 5 1 0 - 5 1 5 , 5 3 1 , 5 3 6 - 5 3 7
equivalent, 6 5 , 6 9 , 70
internal, 64
negative, 6 8 , 2 4 6
output, 2 5 0
parallel, 6 8 - 7 3
instantaneous, 2 2 - 2 5 , 4 1 , 2 7 9 , 3 0 7 ,
phasor admittance/impedance, 4 5 6
Resistivity, 3 3 , 41
power gain, 80
impedance of, 6 0 7 - 6 0 8 , 6 5 8
reactive, 5 1 0 - 5 1 5 , 5 3 2
nonlinear, 41
real, 5 3 2
Resonance.
phenomenon, 8 3 9 - 8 4 2
Practical inductors, 8 3 0
resonance frequency, 8 6 2
Practical source, 80
Primary coils, 8 9 8 , 9 1 7 , 9 3 8
Product over sum rule, 71
Proportionality, 2 0 1 - 2 0 8
8 4 2 -8 4 6
r\
r\
r\
n
classifications, 3 5 2
complete, 3 5 2 , 3 6 4
forced, 3 5 2 , 3 6 4
r\
property, 215
proportionality property, 2 0 1 - 2 0 8
r\
capacitors, and, 4 4 9 - 4 5 5
Power gain, 80
Practical transformers, 9 2 4 - 9 3 0
r>
transfer, 79
7 9 4 -7 9 5
Rectangular coordinates, 4 3 6 , 4 8 2
available, 528
5 0 0 - 5 0 5 , 53 2 , 533
r>
5 1 5 . 5 3 1 ,5 3 7
r\
apparent, 5 1 0 - 5 1 5 , 531
calculation, 5 0 0 - 5 0 5 , 5 33, 5 3 4 - 5 3 6
natural, 3 5 2 , 3 6 5
source-free, 3 2 8 - 3 3 6 , 3 6 5
step, 3 3 6 - 3 4 7 , 365
Quality factor, 8 1 6
o f a band-pass circuit, 861
o f a capacitor, 861
o f a coil, 861
o f L and C components, 8 3 0 - 8 3 3
o f a reactive com ponent, 861
unstable, 3 3 5 , 3 6 5
zero-input, 3 2 8 - 3 3 6 , 3 5 2 , 3 6 5
zero-state, 3 5 2 , 365
Reverse open-circuit voltage gain, 9 8 6 - 9 8 7 ,
1010
Reverse voltage gain, 9 9 2
R L C circuit.
r\
circuits
computing SSS response and, 4 4 2 444
1129
Index
damping and, 3 9 2
Series connections, 1 0 0 2 -1 0 0 9
Series inductors, 2 9 3 - 2 9 5
Series-parallel circuit, 73
applications, 6 4 0 - 6 4 5
parallel, 8 3 3 - 8 3 8
passive R C or R L circuits
vs., 3 9 7 - 4 0 0
source-free, 3 8 3 - 4 0 0
Series-parallel impedance
circuits
O '
capacitors, 301
inductors, 2 9 7 - 2 9 8
resonance and, 8 3 9
definitions, 5 1 - 1 0 6 , 5 3 - 5 4
dependent sources and, 7 8 - 8 4 , 1 0 3 -1 0 4
explained, 6 4 - 6 8
M A TLA B and, 1 0 4 - 1 0 6
passive R L C circuit, 3 6 5
sawtooth, 3 2 2 - 3 2 3 , 3 6 5
source-free or zero-input response, 3 2 8 -3 3 6
steady-state circuit analysis using phasors,
4 6 1 -4 6 6
6 8 -7 3 , 9 6 -9 8
RgO and related calculations of, 9 9 - 1 0 3
series-parallel interconnections, 7 3 - 7 8
Series-parallel resonant circuit, 8 4 6 -8 5 1
Series-parallel R LC circuit, 3 8 6 - 3 8 7
Sifting property, 55 3 , 5 9 1 , 7 6 7 , 7 9 8
Simultaneous equations, 9 3 0
Butterworth filter, 6 3 4 - 6 3 5 , 8 5 6 - 8 5 8 ,
1 0 4 5 -1 0 5 1 .
Saraga design, 1 0 4 7 -1 0 5 0
Saturation, 164, 171
Saturation regions, 178
Sawtoodi waveform, 3 2 2 - 3 2 3 , 3 65, 6 8 0 -6 8 1
Single-frequency analysis, 5 0 9 - 5 1 0
Sinusoidal oscillation, 3 7 8 , 4 1 2 - 4 1 5
Sinusoidal steady state (SSS)
analysis, 4 3 1 497
complex exponential forcing
functions in, 4 4 4 - 4 4 7
5-domain
defined. 4 3 3 , 4 8 2
Secondary coils, 8 9 8 , 9 1 7 , 9 3 8
Second order linear circuits, 3 7 9 430
with constant inputs, 4 0 0 -4 1 1
defined, 4 1 7
discharging a capacitor through an
inductor, 3 8 2 - 3 8 5
oscillator application, 4 1 2 - 4 1 5
source-free, 3 8 5 - 4 0 0
Second order time domain methods, 5 4 4 -5 4 7
Seleaivity, 8 1 6 , 8 6 2
Semiconductors, 4 , 5
Series circuit, 90
application, 4 3 1 4 3 2
Laplace transform analysis and,
7 0 1 -7 0 7
naive technique for computing,
442^44
nodal analysis o f pressure-sensing
device, 4 7 7 -4 8 1
phasor impedance and admittance,
4 5 5 -4 6 1
phasor representations o f sinusoidal
signals, 4 4 7 449
o f R L C circuit, 6 0 4
1130
Index
r>
using phasors, 4 6 1 4 6 6
Square wave, 3 4 5 , 3 4 6
power calculation, 4 9 9 - 5 4 2
Stable circuit, 4 3 3
defined, 4 8 2
stability and boimdedness, 6 9 0
Stable transfer function, 741
Steady state, 4 3 3 .
power, 5 0 6 - 5 0 9
instantaneous and average powers,
Steady-state analysis, 6 0 4 , 6 5 8
Steady-state calculation, 7 4 9 - 7 5 3
5 0 0 -5 0 5
m aximum power transfer in SSS,
5 2 6 -5 3 0
Step function, 4 1 7
Step response, 3 6 5 , 4 1 7 , 6 3 1 - 6 3 4 , 741
Stored energy, 3 7 6 , 9 1 0 - 9 1 4
Stray capacitance, 3 6 1 , 3 6 2 , 3 6 5
Sm oothing, 2 6 9 - 2 7 0 , 3 0 3 - 3 0 5
Supermesh, 136
Solar cell, 3 7 6
Supernode, 123
Source, 16
Superposition, 1 9 1 -2 2 5 , 3 6 5 .
n
See also
controlled, 2 8 , 2 9 , 7 8 , 80
current, 56
o f average power, 5 0 3 - 5 0 5
linearity and, 1 9 3 -2 0 0 , 3 4 7 - 3 5 2
proportionality and, 2 0 1 - 2 0 8
floating, 1 2 3 -1 2 8 , 140, 1 8 8 -1 8 9
ideal current, 2 7 , 38
n
n
networks, 2 1 2 - 2 1 3
independent current, 2 7 , 38
independent (ideal) current, 2 7 , 38
r\
Susceptance, 4 8 2 , 8 6 2
nonideal, 8 5 - 8 8
nonlinear, 80
Switching
first order R L and R C circuits, 3 2 3 , 3 4 5
nonlinear controlled, 80
in R L C circuit, 6 4 0 - 6 4 5
practical, 2 6 , 7 9 , 8 0 , 81
v-i characteristics, 8 0 , 86
voltage, 6 4
1109
Source free, 4 1 7
Source-free response, 3 2 8 - 3 3 6 , 365
System identification, 7 6 5
r>
Tank circuit, 8 4 3 , 8 6 2
r\
See also
Linearity; Superposition
defmed, 2 1 5
Tank frequency, 8 4 3 , 8 6 2
Temperature measurement, nodal and loop
r\
analysis, 143
explained, 2 0 8 - 2 1 2
Tera, 29
applications, 6 5 9 - 6 6 2
linearity and, 1 9 3 -2 0 0
Thevenin, M . L ., 2 2 7
equivalent networks, 2 1 2 - 2 1 3
source transformation theorem for
independent sources, 2 0 9
Source transformation property, 6 1 5 , 6 5 8
S P IC E (Simulation Program with Integrated
Circuit Emphasis), 178
first order R L and R C circuits, 3 4 6
for frequency response plots, 4 7 2 - 4 7 6
W ien bridge oscillator and, 4 1 4
r\
Thevenin theorem, 2 2 7 - 2 6 8 .
See also
1131
Index
O
Laplace transform analysis circuit
applications, 6 5 9 - 6 6 2
M axim um Power Transfer theorem and,
2 4 9 -2 5 6
N orton theorem and
Transfer resistance, 7 9
Transformers.
Transient analysis, 5 4 4 , 6 5 8
Transient response, 4 3 3 , 6 9 6 , 741
Translation o f a plot, 1 0 9 2 - 1 0 9 3 , 1109
3 dB bandwidth, 8 1 5 - 8 1 6 , 861
1 0 9 3 -1 0 9 4 , 1109
Transmission
{t-) parameters, 9 6 1 , 9 9 1 - 9 9 4 ,
1010, 1 0 2 3 -1 0 2 5
Triangular waveform, 3 5 5 - 3 5 7
Trigonom etric Fourier series, 1089, 1109
o f B P filter, 1071
Tuned circuit, 8 4 4 , 8 6 2
o f LP filter, 1071
9 7 2 - 9 7 3 , 1010
Two-ports, 9 5 9 - 1 0 3 0
admittance parameters, 9 6 8 - 9 7 3
T im e constant, 3 2 9 , 3 6 5
amplifiers, 9 5 9 - 9 6 0
T im e differentiadon formula, 5 7 6 - 5 7 8
See Convolution
T im e invariance, 6 5 8 , 7 9 3 , 7 9 8
parameters, 9 9 4 - 9 9 7
A-parameters, 9 6 1 , 9 8 5 - 9 9 1 , 1 0 2 0 -1 0 2 3
impedance and gain calculations o f
T im e shift property, 5 5 8 - 5 5 9 , 5 9 1 , 7 6 7
z-parameters, 9 8 1 - 9 8 4
Touch-tone phones, 8 1 1 - 8 1 2
impedance parameters, 9 7 6 -9 8 1
Transconductance, 7 9
one-port networks, 9 6 1 , 9 6 2 - 9 6 8 ,
Transfer admittance, 6 2 6
1 0 1 1 -1 0 1 2
Transfer conductance, 7 9
Transfer function, 6 2 5 - 6 3 0 , 6 5 8 , 6 8 3 - 7 6 1
1 0 0 2 -1 0 0 9
parameter conversion and inter
connection o f two-ports, 1 0 2 5 -1 0 2 9
bode plots, 7 3 0 - 7 3 6
reciprocity, 9 9 7 - 1 0 0 2 , 1 0 2 9 -1 0 3 0
classification o f responses, 6 9 3 -7 0 1
^-parameters, 9 6 1 , 9 9 1 - 9 9 4 , 1 0 2 3 -1 0 2 5
D C m otors and, 7 3 6 - 7 3 9
transmission parameters, 9 9 1 - 9 9 4
frequency response, 7 0 7 - 7 1 4 , 7 5 3 - 7 5 9
frequency scaling and magnitude scaling,
7 1 4 -7 2 6
initial- and finai-value theorems,
7 2 6 - 7 2 9 ,7 5 9 - 7 6 0
Laplace transform analysis circuit
applications, 6 2 5 - 6 3 0
-V
ideal, 9 1 4 - 9 2 4 , 9 3 0 - 9 3 6 , 9 5 0 - 9 5 6
practical, 9 2 4 - 9 3 0
231
two-ports and, 9 6 4 - 9 6 8
defined, 885
circuits
superposition and, 3 4 9
Vw^
steady-state calculation, 7 4 9 - 7 5 3
Transfer impedance, 6 2 6
'O
7 0 1 -7 0 7
stability problems, 7 4 5 - 7 4 6
active networks, 2 4 1 - 2 4 5
op amp circuits, 2 4 6 - 2 4 9
L >
two-ports, 9 7 3 - 9 7 6 , 1 0 1 2 -1 0 1 6
z-parameters, 1 0 1 6 - 1 0 2 0
Two-terminal circuit clem ent, 58
Two-terminal device, 15, 3 0
Tw o-terminal networks, 2 0 9 , 2 1 2 , 2 1 5
Thevenin and Nortons theorems, 228, 257
V 257
Undamped circuit, 3 8 5 , 4 1 7 , 8 6 2
1132
Index
n
Undamped natural frequency, 8 4 3 , 8 6 2
Wattage, 19, 41
Underdamped circuit, 3 9 0 , 3 9 2 , 4 1 7
Waveform
periodic current, 2 7 9 - 2 8 0
r\
n
U nity coupling, 9 3 1 , 9 3 8
period of, 4 3 3
sawtooth, 3 2 2 - 3 2 3 , 3 6 5
Unstable response, 3 3 5 , 3 6 5
sinusoidal voltage, 4 1 2
v-i characteristic, 2 6 , 3 7 - 3 8 , 41
v-i relationship o f capacitor
Laplace transform m ethod, 5 8 0 - 5 8 2
Virtual ground, 160, 178
triangular, 3 5 5 - 3 5 7 , 6 8 0 -6 8 1
W ien bridge oscillator, 4 1 2 , 4 1 4 , 4 7 5 476
W ire, resistance o f copper wire, 33.
See also
Resistivity
V 257
j^-parameters
9 7 3 - 9 7 6 ,1 0 1 2 - 1 0 1 6
Voltage, 2, 16
branch, 107
gain, 117
node, 6 0
regulators, 3 0 3
Voltagp division, 9 0 ,2 9 5 ,2 9 9 ,3 0 0 ,4 8 2 ,6 5 8
Zero
finite zero, 6 8 5
Laplace transform analysis transfer
formula, 6 6 , 4 5 6
function applicadons, 6 8 5 - 6 9 3 ,
7 4 2 -7 4 4
applications, 6 5 9 - 6 6 2
Voltage drop, 10
Voltage follower, 1 64, 178
Zero-input response, 3 2 8 - 3 3 6 , 3 5 2 , 3 6 5 ,
Voltage gain, 80
Zero-state response, 3 5 2 , 3 6 5 , 6 9 3 , 7 4 1 , 7 9 8
Z Js). 9 4 0 - 9 4 3
2-parameters, 9 7 6 - 9 8 1 , 1 0 1 6 -1 0 2 0
Voltage regulator, 3 0 7
4 8 2 , 6 9 3 , 741
Zeros o f rational function, 5 6 6 , 5 91, 741
Voltage source, 4 1 , 6 4
nonideal, 8 6
v-i charaaeristics, 8 0 , 8 6
Voltmeter, 14, 7 6
sensitivity, 7 7
n
o
n
r s