Introduction to CMOS Logic Circuits

Digital circuit Design : An Overview

Digital IC technologies and logic circuit families

Logic families
Bipolar Uses bipolar junction transistors npn or pnp silicon structure Small current into very thin base layer controls large currents between emitter and collector Base currents limit integration density MOS - Uses nMOS and pMOS Metal Oxide Semiconductor Field Effect Transistors nMOS and pMOS MOSFETS Voltage applied to insulated gate controls current between source and drain Low power allows very high integration

Transistor Evolution
Transistor Bipolar

Bardeen et.al. (Bell Labs) in 1947

transistor Shockley in 1949

First
First First TTL ECL

bipolar digital logic gate Harris in 1956

monolithic IC Jack Kilby in 1958 commercial IC logic gates Fairchild 1960 1962 into the 1990s 1974 into the 1980s

MOS Technology Evolution

MOSFET transistor - Lilienfeld (Canada) in 1925 and Heil (England) in 1935 CMOS 1960s, but plagued with manufacturing problems (used in watches due to their power limitations) PMOS in 1960s (calculators) NMOS in 1970s (4004, 8080) for speed CMOS in 1980s preferred MOSFET technology because of power benefits BiCMOS, Gallium-Arsenide, Silicon-Germanium SOI, Copper-Low K, strained silicon, High-k gate oxide...

Bipolar
TTL (Transistor-transistor logic) had been used for many years . ECL (Emitter Coupled Logic) : basic element is the differential BJT pair. BiCMOS : combines the high speed of BJTs with low power dissipation of CMOS . GaAs : for very high speed due to the high carrier mobility . Has not demonstrated its potential commercially .

CMOS
Replaced NMOS (much lower power dissipation) Small size, ease of fabrication Channel length has decreased significantly (as short as 0.06 m or shorter) Low power dissipation than bipolar logic circuits ( can pack more) . High input impedance of MOS transistors can be used to storage charge temporarily (not in bipolar) High levels of integration for both logic (chapter 10) and memory circuits. Dynamic logic to further reduce power dissipation and to increase speed performance .

Features to be Considered
Interface circuits for different families Logic flexibility Speed Complex functions Noise immunity Temperature Power dissipation Co$t

Design Abstraction Levels

SYSTEM

MODULE
+ GATE

CIRCUIT

DEVICE G S n+ D n+

N-type and P-type

MOS Structure
NMOS

L Distance between the source and the drain (Channel length)

Channel

W Width of the channel

n+ - Silicon doped with higher concentration p - Silicon substrate doped with p+ - Silicon doped with higher concentration

Introduction to CMOS Logic Circuits

CMOS stands for Complementary Metal Oxide Semiconductor
Complementary: there are N-type and P-type transistors. N-type transistors use electrons as the current carriers. P-type transistors use holes as the current carriers.
Electrons are free carriers in the conduction band with energy of Ec or just above the conduction band edge. Free electrons are generated by doping the silicon with an N-type impurity such as phosphorous or arsenic. A hole is a current carrier due to the absence of an electron in a covalent bond state, i.e. a missing electron which would otherwise be part of a silicon-to-silicon bond. Holes are free carriers in the valence band with energy of Ev or just below the valence band edge. Holes are generated by doping the silicon with a P-type impurity such as boron.

Metal: the gate of the transistor was made of aluminum metal in the early days, but is made of polysilicon today (for the past 25 years or more). Oxide: silicon dioxide is the material between the gate and the channel Semiconductor: the semiconductor material is silicon, a type IV element in the periodic chart. Each silicon atom bonds to four other silicon atoms in a tetrahedral crystal structure.

CMOS NMOS and PMOS Transistors

oxide

gate
N+ N N+ P+

oxide
gate
P+

source

P substrate

drain

source

N well

drain

N channel device

P channel device

N channel device: built directly in the P substrate with N-doped source and drain junctions and normally N-doped gate conductor
Requires positive voltage applied to gate and drain (with respect to source) for electrons to flow from source to drain (thought of as positive drain current)

P channel device: built in an N-well (a deep N-type junction diffused into the P substrate) with P-doped source and drain junctions and N or P-doped gate
Requires negative voltage applied to gate and drain (with respect to source) for electrons to flow from drain to source (thought of as negative drain current)

N-FET and P-FET Devices as Switches

NMOS Device:
positive voltage (1 or high) on gate relative to source turns device ON and allows positive current to flow from drain to source (switch closed) zero volts on gate (0 or low) turns device OFF (open circuit) Source (vs drain) is the most negative terminal
drain

oxide

N+ N
source

gate

N+
drain

P substrate

N channel device source

gate

PMOS Device:
Negative voltage (0 or low) on gate relative to source turns device ON and allows (negative) current to flow from drain to source (closes switch) Zero volts on gate relative to source (1 or high) turns device OFF (closes switch) source (vs drain) is the most positive terminal

substrate N-FET device schematic

oxide gate

source

P+
N well

P+
drain

gate

P channel device source substrate P-FET device schematic drain

3-D view of an NMOS

polysilicon gate W tox n+ L p-type body n+ SiO2 gate oxide (good insulator, ox = 3.9)

Symbols
NMOS
Vg + Vgs Vs + Vgd Vds + Vd

Vgs = Vg Vs Vgd = Vg Vd Vds = Vd Vs = Vgs - Vgd

Operation
Three regions of operation Cutoff Linear Saturation

Cut-Off Mode
Vgs = 0

+ s n+

+ d n+

Vgd

=0

p-type body b

No channel formed

Operation
Linear Mode
Vgs > Vt

+ s n+

+ d n+

Vgd = Vgs

Vds = 0

p-type body b

Vgs > Vt

+ s n+

+ d n+

Vgs > Vgd > Vt Ids 0 < Vds < Vgs-Vt

p-type body b

Operation
Saturation Mode
Vgs > Vt

+ -

+ -

Vgd < Vt

d Ids Vds > Vgs-Vt

n+ p-type body b

n+

The voltage drop at the induced channel (from the pinch-off point to the source remains fixed at Vgs Vt

pMOS Transistor
Similar, but doping and voltages reversed
Body tied to high voltage (VDD) Gate low: transistor ON Gate high: transistor OFF Bubble indicates inverted behavior
Source Polysilicon SiO2 Gate Drain

p+ n

p+ bulk Si

Power Supply Voltage

GND = 0 V In 1980s, VDD = 5V VDD has decreased in modern processes
High VDD would damage modern tiny transistors Lower VDD saves power

VDD = 3.3, 2.5, 1.8, 1.5, 1.2, 1.0,

Transistors as Switches
We can view MOS transistors as electrically controlled switches Voltage at gate controls path from source to drain
g=0 d nMOS g s s d ON s s s d OFF s d OFF g=1 d ON

d pMOS g

MOS Transistors
Voltage-controlled resistance

PMOS

NMOS

MOSFET as Switches

MOSFET: Metal-Oxide-Semiconductor Field-Effect Transistor nFET: an n-channel MOSFET that uses negatively charged electrons for electrical current flow pFET: a p-channel MOSFET that uses positive charges for current flow In many ways, MOSFETs behave like the idealized switches introduced in the previous section
(a) nFET symbol

The voltage applied to the gate determines the current flow between the source and drain terminals
(b) pFET symbol Figure 1 Symbols used for nFETs and pFETs

MOSFET as Switches

Early generations of silicon MOS logic circuits used both positive and negative supply voltages as Figure 2 showing In modern designs require only a single positive voltage VDD and the ground connection, e.g. VDD = 5 V and 3.3 V or lower The relationship between logic variables x and its voltages Vx
0 Vx VDD 0V VDD
(1)

Figure 2 Dual power supply voltages

x 0 means that Vx x 1 means that Vx

(2)

(a) Power supply connection (b) Logic definitions Figure 3 Single voltage power supply

Switching Characteristics of MOSFET

In general,

Low voltages correspond to logic 0 values High voltages correspond to logic 1 values

The transition region between the highest

logic 0 voltage and the lowest logic 1 voltage is undefined

(a) Open

(b) Closed

Figure 4 nFET switching characteristics

nFET
y x A which is valid iff A 1
(3)

pFET
y x A which is valid iff A 0
(4)
(a) Open (b) Closed

Figure 5 pFET switching characteristics

nMOS FET Threshold Voltages

An nFET is characterized by a threshold voltage VTn that is positive, typical is around VTn = 0.5 V to 0.7 V If VGSn VTn , then the transistor acts like an open (off) circuit and there is no current flow between the drain and source If VGSn VTn , then the nFET drain and source are connected and the equivalent switch is closed (on)

(a) Gate-source voltage

Thus, to define the voltage VA that is associated with the binary variable A
VA VGSn
(5)
(b) Logic translation Figure 6 Threshold voltage of an nFET

pMOS FET Threshold Voltages

An pFET is characterized by a threshold voltage VTp that is negative, typical is around VTp = 0.5 V to 0.8 V
If VSGp VTp , then the transistor acts like an open (off)

switch and there is no current flow between the drain and source
(a) Source-gate voltage

If VSGp VTp , then the pFET drain and source are

connected and the equivalent switch is closed (on)

Thus, to the applied voltage VA we first sum voltage to write

VA VSGp VDD
(6) (7)

VA

0V

VA VDD VSGp
VDD VTp

VA VDD

(9)
(b) Logic translation Figure 7 pFET threshold voltage

(8) Note that the transition between a logic 0 and a logic 1 is at (8) !

nFET Pass Characteristics

An ideal electrical switch can pass any voltage applied to it As Figure 8 (b), the output voltage Vy is reduced to a value
V1 VDD VTn
(10)

since

VGSn VTn

(a) Logic 0 transfer

Which is less than the input voltage VDD, called

threshold voltage loss

Thus, we say that the nFET can only pass a weak logic 1; in other word, the nFET is said to pass a strong logic 0 can pass a voltage in the range [0, V1]

(b) Logic 1 transfer Figure 8 nFET pass characteristics

Static CMOS Circuit

At every point in time (except during the switching transients) each gate output is connected to either VDD or V ss via a low-resistive path. The outputs of the gates assume at all times the value of the Boolean function, implemented by the circuit (ignoring, once again, the transient effects during switching periods).

This is in contrast to the dynamic circuit class, which relies on temporary storage of signal values on the capacitance of high impedance circuit nodes.

Static CMOS
VDD In1 In2 In3

PUN

PMOS Only
F=G

In1 In2 In3

PDN

NMOS Only

VSS

PUN and PDN are Dual Networks

Simple CMOS Circuits: The Inverter Gate

Vdd

The simplest complementary MOS (CMOS) circuit is the inverter:

NMOS & PMOS gates are connected together as the input NMOS & PMOS drains are connected together as the output NMOS & PMOS sources are connected to Gnd and Vdd, respectively. NMOS substrate is normally connected to Gnd for all NMOS devices in the circuit PMOS well is normally connected to Vdd (most positive voltage in circuit) for all PMOS devices

Inverter Schematic

PMOS source

P-MOS
PMOS drain Vin Vout

NMOS drain

N-MOS
NMOS source

Operation:
If Vin is down (0 volts), NMOS is OFF and PMOS is ON pulling Vout to Vdd (high = 1) If Vin is up (at Vdd), NMOS is ON hard and PMOS is OFF pulling Vout low to Gnd (0) With Vin at 0 or Vdd, no dc current flows in inverter

Gnd

Inverter Symbol

CMOS Inverter
A 0 1 Y

VDD A Y

GND

CMOS Inverter
A 0 1 Y
0
A=1

VDD OFF
Y=0

ON
A Y

GND

CMOS Inverter
A 0 1 Y 1 0
A=0

VDD ON
Y=1

OFF
A Y

GND

Series connected MOSFETS

Parallel Connected MOSFETS

Series and Parallel

nMOS: 1 = ON pMOS: 0 = ON Series: both must be ON Parallel: either can be ON
a g1 g2 b (a) 0 0 b OFF a 0 0 b (b) b ON 0 1 b OFF a 0 1 b OFF a 1 0 b OFF a 1 0 b OFF a 1 1 b OFF a 1 1 b ON a a

a g1 g2

a g1 b (c) g2 0

a 0 b OFF 0

a 1 b ON 1

a 0 b ON 1

a 1 b ON

a g1 b (d) g2 0

a 0 b ON 0

a 1 b ON 1

a 0 b ON 1

a 1 b OFF

Pull-up / Pull-down Model

Typical CMOS gate can be viewed as consisting of two parts
pull-up network and pull-down network
VDD A B C Pull-up output

A B C

Pulldown GND

Pull-up / Pull-down Model

High level inputs to the PDN cause switches to close If there is a closed switch path thru PDN, then output is low Low level inputs to the PUN cause switches to close If there is a closed switch path thru PUN, then output is high

CMOS NAND Gate

A 0 0 1 1 B 0 1 0 1 Y

Y A B

CMOS NAND Gate

A 0 0 1 1 B 0 1 0 1 Y 1
A=0 B=0 ON ON Y=1 OFF OFF

CMOS NAND Gate

A 0 0 1 1 B 0 1 0 1 Y 1 1
OFF A=0 B=1 ON Y=1 OFF ON

CMOS NAND Gate

A 0 0 1 1 B 0 1 0 1 Y 1 1 1
ON A=1 B=0 OFF Y=1 ON OFF

CMOS NAND Gate

A 0 0 1 1 B 0 1 0 1 Y 1 1 1 0
OFF A=1 B=1 OFF Y=0 ON ON

CMOS NOR Gate

A 0 0 1 1 B 0 1 0 1 Y 1 0 0 0

A B Y

3-input NAND Gate

Y pulls low if ALL inputs are 1 Y pulls high if ANY input is 0

3-input NAND Gate

Y pulls low if ALL inputs are 1 Y pulls high if ANY input is 0
Y A B C

Logic Identities

Logic Analysis -Inverter

Logic Analysis NAND Gate

Logic Analysis NOR Gate

Pull-up / Pull-down Model

A
A and ( B or C) B C Since hign level signals on the inputs cause the PDN to close switches, we get a Boolean expression for the input which creates a closed path thru PDN

If a closed path exists in PDN, then the output is pulled low. Thus the logic function realized is the complement (inverted) version of the Boolean expression. output

A B C

Pulldown GND

not (A and ( B or C))

Pull-up / Pull-down Model

What happens when the Boolean expression is false? Since there is no path thru PDN, the output could float. In order to make the output high, the PUN must have a path which connects VDD to the output.
Vdd

B A

Observe: take the expression for PDN and use DeMorgans Law to write it in terms of complemented input variables. Complemented variables are true when the input level is low. Thus, this gives exactly the form of C the PUN
In this case: not A or ( not B and not C)

Example Gate: COMPLEX CMOS GATE

B A C D

VDD

OUT = D + A (B+C)
A D B C

Example Gate: COMPLEX CMOS GATE

VDD B A C D

OUT = D + A (B+C)
A D B C

VDD B A C D

OUT = D + A (B+C)
A D B C

Examples of pull-down networks

Examples of pull-up networks

VDD PMOSs A B C NMOSs GND Out

Vcc

Gnd

Vcc C E D

D Vcc

F A B F Gnd A

C
D E

Gnd

Stick Diagram, Interconnected

83

Two Versions of C (A + B)
A
VDD

C
VDD

GND

GND

Line of diffusion layout abutting source-drain connections

Note crossover eliminated by A B C ordering

 Example: x y + x z + x v = x (y + z + v) 5 operations: 1 AND, 2 OR # txs = ___? 3 operations: 3 AND, 2 OR # txs = ___?

Construct a CMOS logic gate to implement the function:

F = a (b + c)

14 transistors (cascaded gates)

Construct a CMOS logic gate to implement the function:

F = a (b + c)

Complex gates in CMOS logic

A complex logic gate is one that implements a function that can provide the basic NOT, AND and OR operation but integrates them into a single circuit. CMOS is ideally suited for creating gates that have logic equations by exhibiting the following, 1) AND-OR-INVERT - AOI form 2) OR-AND-INVERT - OAI form An AOI logic equation is equivalent to a complemented SOP from, while an AOI equation is equivalent to a complemented POS structure. In CMOS, output always produces NOT operation acting on input variable

AOI Logic Function (OR) Design of XOR gate using CMOS logic.
AND-OR-INVERT logic function(AOI) implements operation in the order AND,OR,NOT. For example , Let us consider the function Y = AB+CD i.e., Y = NOT((A AND B)OR (C AND D))

The AOI logic gate implementation for Y

CMOS implementation for Y

Step 1: Draw A.B (AND) function first by connecting 2 nMOS transistors in series.

Step 2: Draw C.D implementation, by using 2 nMOS transistors in series.

Step 3:

Y = A.B+C.D , In this function A.B and C.D are added, for addition , we have to draw parallel connection. So, A.B series connected in parallel with C.D as shown in figure.

Step 4: Draw pMOS connection,

I. In nMOS A,B connected in series. So, in pMOS side, A.B should be connected in parallel. II. In nMOS C,D connected in series. So, in pMOS side, C.D should be connected inparallel. III. A.B and C.D networks are connected in parallel in nMOS side. So, in pMOS side, A.B and C.D networks should be connected in series. IV. In pMOS multiplication should be drawn in parallel, then addition should be drawn in series as shown in figure.

Step 5: Take output at the point in between nMOS and pMOS networks.

CMOS implementation for Y

Example: x = ab+cd

OAI Logic Function (OR) Design of XNOR gate using CMOS logic. OR-AND-INVERT logic function(AOI) implements operation in the order OR,AND,NOT. For example , Let us consider the function Y = (A+B).(C+D) i.e., Y = NOT((A OR B)AND (C OR D))

CMOS implementation for Y

OAI22 Logic Graph

A B C D X = (A+B)(C+D) C A D B

A B C D

Structured Logic Design (1/4)

CMOS logic gates are intrinsically inverting

Output always produces a NOT operation acting on the input variables

Figure 2.42 Origin of the inverting characteristic of CMOS gates

Structured Logic Design (2/4)

(a) Series-connected nFETs

Figure 2.44 nFET AOI circuit

(b) Parallel-connected nFETs Figure 2.43 nFET logic formation

Figure 2.45 nFET OAI circuit

Structured Logic Design (3/4)

(a) Parallel-connected pFETs

(a) pFET AOI circuit

(b) pFET OAI circuit (b) Series-connected pFETs Figure 2.46 pFET logic formation Figure 2.47 pFET arrays for AOI and OAI gates

Structured Logic Design (4/4)

(a) AOI circuit

(b) OAI circuit

Figure 2.48 Complete CMOS AOI and OAI circuits

Bubble Pushing

(a) NAND - OR (a) Parallel-connected pFETs

(b) NOR - AND Figure 2.52 Bubble pushing using DeMorgan rules (b) Series-connected pFETs Figure 2.51 Assert-low models for pFETs

XOR and XNOR gates

XOR and XNOR Gates

An important example of using an AOI circuit is constructing Exclusive-OR (XOR) and Exclusive-NOR circuits
a b a b a b a b a b a b
a
a

(2.71) (2.72) (2.73) (2.74)

(a) Exclusive-OR (b) Exclusive-NOR

b (a

b) a b a b

b a b a b

Figure 2.57 AOI XOR and XNOR gates

(a) AOI22

(b) AOI321

(c) AOI221

Figure 2.58 General naming convention Figure 2.56 XOR

E Vcc

Gnd

logic cascade in cmos

Half-Adder
Logic Equations
Truth Table

x 0 0 1 1

y 0 1 0 1

C 0 0 0 1

S 0 1 1 0

C = x y S = x y
Schematic

Full Adder
Adding two single-bit binary values, X, Y with a carry input bit C-in produces a sum bit S and a carry out C-out bit.
Inputs Outputs

Sum S
XY C-in 0

00 0
1 3

01
2

11
6

10
4

Full Adder Truth Table

X 0 0 0 0 1 1 1 1 Y 0 0 1 1 0 0 1 1 C-in 0 1 0 1 0 1 0 1 S 0 1 1 0 1 0 0 1 C-out 0 0 0 1 0 1 1 1

1
7 5

1 1
Y C-in

S = XY(C-in) + XY(C-in) + XY(C-in) + XY(C-in) S= X Y (C-in)

Carry C-out
XY C-in 0

00 0
1 3

01
2

11
6

10
4

1
7
5

1
Y

C-in

S(X,Y, C-in) = (1,2,4,7) C-out(x, y, C-in) = (3,5,6,7)

C-out = XY + X(C-in) + Y(C-in)

Addition of Binary Numbers

Full Adder. The full adder is the fundamental building block of most arithmetic circuits:

ai

bi

Cout

Full Adder

Cin

si
The sum and carry outputs are described as:

si

aibi ci aibi ci aibi ci aibi ci

ci 1 ai bi ci ai bi ci ai bi ci ai bi ci ai bi ai ci bi ci

Full Adder Circuit Using AND-OR

X X X Y Y C-in C-in
X Y C-in X Y C-in X Y C-in X Y C-in XYC-in XYC-in

XYC-in

Sum S

C-in

XYC-in

X Y

XY

C-out

Full Adder
S

C-in

X C-in Y C-in

XC-in

C-out

YC-in

Cont..

Addition of Binary Numbers

Inputs
ci 0 0 0 0 1 1 1 1 ai 0 0 1 1 0 0 1 1 bi 0 1 0 1 0 1 0 1 si 0 1 1 0 1 0 0 1

Outputs
ci+1 0 0 0 1 0 1 1 1

Propagate Generate

Propagate
Generate

PGK

For a full adder, define what happens to carries

Generate: Cout = 1 independent of C

G=AB P=A B

Propagate: Cout = C

Kill: Cout = 0 independent of C

K = ~A ~B

Full-Adder Implementation
Full Adder operations is defined by equations:

si

aibici aibi ci aibi ci aibi ci

ai

bi

ci

pi

ci

ci 1 aibici aibi ci aibi

Carry-Propagate: pi and Carry-Generate

gi
bi

pici

ai bi

ai

gi

ai bi
cout
One-bit adder could be implemented as shown

cin

si

High-Speed Addition
ci
ai
1

gi

pi ci
pi ai bi

bi

gi

ai bi
0

cout

si

pi

ci

s 1

cin

One-bit adder could be implemented more efficiently because MUX is faster

si

Multiplexers

2:1 multiplexer chooses between two inputs

S
S 0 0 1 D1 X X 0 D0 0 1 X Y

D0 D1

0 Y 1

Multiplexers

2:1 multiplexer chooses between two inputs

S
S 0 0 1 D1 X X 0 D0 0 1 X Y 0 1 0

D0 D1

0 Y 1

Gate-Level Mux Design

SD1 SD0 (too many transistors)

How many transistors are needed?

Gate-Level Mux Design

SD1 SD0 (too many transistors)

D1 S D0

How many transistors are needed? 20

D1 S D0

4 2 4

2 4 2 2

Inverting Mux

Inverting multiplexer
Use compound AOI22 Or pair of tristate inverters Essentially the same thing

Noninverting multiplexer adds an inverter

D0 S S

S D1 Y S D1 1 D0

S 0 Y

Tristate Inverter

Tristate inverter produces restored output

Violates conduction complement rule Because we want a Z output

A EN Y EN

Tristate Inverter

Tristate inverter produces restored output

Violates conduction complement rule Because we want a Z output

A A EN Y EN Y

EN = 0 Y = 'Z'

EN = 1 Y=A

D0 S S

D1 S Y S D1 1

S D0 0 Y

Inverting Mux

Inverting multiplexer
Use compound AOI22 Or pair of tristate inverters Essentially the same thing

Noninverting multiplexer adds an inverter

D0 S S S D1 Y S S S D1 D0 S D1 S Y D0 0 Y 1 S

Complimentary Static CMOS Full Adder

VDD VDD A B A B Ci A B VDD Ci A B

X
A

Ci

Ci A B

S
Ci

VDD
A Co B Ci A

28 Transistors

A Better Structure: The Mirror Adder

VDD VDD A "0"-Propagate Ci "1"-Propagate A B A Generate B A B Ci A B B B Kill A Co Ci S Ci A B VDD Ci A B

24 transistors
and 2 inverter Total 24+4 =28 Transistors

Mirror Adder
Stick Diagram

VDD

Ci Co

A Ci

Co

Ci

S GND
139

OAI/AOI duality
A B C OUT

Vdd B C

A B C OUT

A
Out B C Gnd A

Review: Construction of PDN

NMOS devices in series implement a NAND function
AB

A
B

NMOS devices in parallel implement a NOR function

A+B

 Scale both W and L no effective change in W/L increases gate capacitance

 Series Transistors
increases effective L

 Parallel Transistors
increases effective W

Analysis of CMOS gates

Represent on transistors as resistors

1
W

1
R

W W

R R

Transistors in series resistances in series

Effective resistance = 2R Effective width = W

Analysis of CMOS gates, cont

Represent on transistors as resistors
W
W

0
0

Transistors in parallel resistances in parallel Effective resistance = R Effective width = 2W

CMOS gate design

Designing a CMOS gate: Find pulldown NMOS network from logic function or by inspection Find pullup PMOS network - By inspection - Using logic function - Using dual network approach Size transistors using equivalent inverter - Find worst-case pullup and pulldown paths - Size to meet rise/fall or threshold requirements

Equivalent inverter: effective width to length ratios (model I)

Parallel combination Series combination

For the NOR gate the effective width of the drivers transistors doubles. That means the effective aspect ratio is increased.

For the NAND gate the effective length of the driver transistors doubles. That means the effective aspect ratio is decreased.
.

Exercise: Design and compare two input NAND and NOR gates

Find W/L relationship between NMOS and PMOS transistors

Assume K'n=2K'p

NAND gate NOR gate

Cmos inverter : summary

1.

2. 3.

4.
F or an inverter the (W/L) of the NMOS is (W/L)n and that for the PMOS is (W/L)p for Fall time = Rise Time. For the worst case design of the NAND gate, you should set the W/L ratio of the NMOS transistors to 2*(W/L)n (i.e. twice that of the inverter) and that of the PMOS to (W/L)p (i.e. equal to that of the inverter)

Transistor sizing CMOS NOR

CMOS NAND Gates

A B AB

A
B

Adding transistors in series (without sizing) slows down the circuit

Optimized NAND Gates W/L ratio of PMos and NMos transistors

Size Devices for the Worst Case

Series transistors: Increase W to reduce Reff.

Parallel transistors: assume the worst case, i.e. only one of the parallel transistor is ON.

Transistor Sizing Without Velocity Saturation

Assumption: Equal rise delay and fall delay Consideration: Effective Resistance

Trade-Off

Increase W to reduce the effective Resistance for the pull down network. The area is increased.

i input CMOS NAND i WP /LP =2.5 WN /LN Two input CMOS NAND 2 WP /LP =2.5 WN /LN

i input CMOS NOR WP /LP =2.5 i WN /LN

TWO input CMOS NOR

WP /LP =2.5 (2) WN /LN = 5 WN /LN

Transistor Sizing Without Velocity Saturation

Total area of the NMOS and PMOS channels is seen as L=2u w=4u ratio 2.5 times CMOS NAND 2 input gate 2( 8u)(2u) + 2 (10u)(2u) = 72um2 CMOS INVERTER gate ( 4u)(2u) + (10u)(2u) = 28um2

WN LN + WP LP

CMOS NOR 2 input gate 2( 4u)(2u) + 2 (20u)(2u) = 96um2

Example

Total area of the NMOS and PMOS channels is seen as

WN LN + WP LP
L=2u w=4u ratio 2.5 time CMOS NAND gate -2 input gate 2( 8u)(2u) + 2 (10u)(2u) = 72um2

CMOS NOR 2 input gate 2( 4u)(2u) + 2 (20u)(2u) = 96um2

Cmos inverter: Charging

Cmos inverter : Discharging

A fast gate is built either by keeping the output capacitance small or by decreasing the on-resistance of the transistor The latter is achieved by increasing the W/L ratio of the device

Symmetrical propagation delay

Input Pattern Effects on Delay-I

Delay is dependent on the pattern of inputs

Rp A Rn A Rn B B

Rp

Low to high transition

both inputs go low

- delay is 0.69 Rp/2 CL since two p-resistors are on in parallel

CL

one input goes low

- delay is 0.69 Rp CL

Cint

High to low transition

both inputs go high

- delay is 0.69 2Rn CL

Adding transistors in series (without sizing) slows down the circuit

Delay Dependence on Input Patterns-II

The reason for difference in low to high transition is due to internal node capacitance and high to low is Due to initial state of the internal nodes.
3 2.5 2

2-input NAND with NMOS = 0.5 m/0.25 m PMOS = 0.75 m/0.25 m CL = 10 fF

A=B=1 0
Input Data Delay (psec) 69 62 50 35 76 57

Voltage, V

1.5 1 0.5 0 -0.5 0 100 200

A=1

0, B=1
Rp A Rn A Cint B CL Rp

Pattern A=B=0 1 A=1, B=0 1 A= 0 1, B=1 A=B=1 0 A=1, B=1 0 A= 1 0, B=1

A=1, B=1 0

300

400

time, psec

Conclusions: Estimates of delay can be fairly Rn complex have to consider internal node B capacitances and the data patterns

Transistor Sizing a Complex CMOS Gate

B A 4 3

8 6

C
D 4 6

8 6

OUT = D + A (B + C)

A
D 1 B 2C

2
2

Red sizing assuming Rp = Rn Follow short path first; note PMOS for C and B 4 rather than 3 average in pull-up chain of three (4+4+2)/3 = 3 Also note structure of pull-up and pull-down to minimize diffusion cap at output (e.g., single PMOS drain connected to output) Green for symmetric response and for performance (where Rn = 3 Rp) Sizing rules of thumb PMOS = 3 * NMOS 1 in series = 1 2 in series = 2 3 in series = 3 etc.

Tutorial on Transistor Sizing Problem #1 (Static CMOS logic): Design a 3-input CMOS NAND gate (PUN/PDN) with fan-out of 3. Total output load of the NAND gate is equal to 15fF and n/p = 2.5. For 0.35m process technology tox = 7.6*10-9m, ox = 35*10-12F/m. Compare the above design with that of a 3-input NOR (PUN/PDN) gate. State any benefits of one implementation over the other. For the sake of simplicity assume all capacitance is lumped and gate capacitance neglecting diffusion and wiring capacitance. Solution: Cox = oxWL/ tox = 15fF. So W = 9.2m. Leff = 0.35m. Assuming output load is all gate capacitance. This is a simplifying assumption made for this problem. A more realistic approach would be to calculate the diffusion and wiring capacitances as well. Equivalent inverter for fan-out of 3 and n/p = 2.5 would result in: Wp = 2.5*Wn for equal rise and fall times. Also, Wp + Wn = 9.2/3 = 3.16m for fan-out of 3. Solving the above equations we have, Wp = 2.23m and Wn = 0.89m.
A 2.23 B 2.23 C 2.23

A B C

2.67 2.67 2.67

NAND implementation NAND implementation: Therefore, for the 3-ip NAND gate implementation, each PDN n-MOS transistor will be: 3*0.89m = 2.67m

and each p-MOS transistor in the PUN network will be: 2.23m NOR implementation: For a 3-ip NOR gate implementation, each PDN n-MOS transistor will be: 0.89m Each PUN p-MOS transistor will be 3*2.23m = 6.69m Area comparison: Total gate area of NAND gate is: 3*(2.67+2.23)m = 14.7m Total gate area of NOR gate is: 3*(6.69+0.89) m = 22.7m
A 0.89

6.69

6.69

6.69 C

0.89

0.89

NOR implementation

Thus we infer that the NAND gate has less area and power compared to the NOR gate for identical loading and fan-out conditions and is the preferred implementation.