You are on page 1of 9

Balance and its Relationship to Rotational Inertia

Suppose we have some sort of distributed mass pivoted about an axis, and we want to
know where to put the axis to minimise the rotational inertia. This would be important if, for
example, we wanted maximise the acceleration of the distributed mass.

Balance Position for Two Point Masses


Lets start off gently by considering two small masses joined by a rigid rod of negligible
mass.
L

m1

kL

m2

The masses, shown in blue, have a definite diameter just to show their position, but we
want to consider them as effectively having a negligibly small diameter, making them point
masses. The triangular shape is intended to represent a point about which the rigid rod is
free to rotate in the plane of the page, a pivot point.
We know from experiment that this arrangement will balance if the product of force and
distance from the pivot is equal on both sides. This is expressed mathematically as
m1 g k L m2 g 1 k L

where g is the gravitational constant, having units of force/mass.


The g and L terms immediately cancel leaving
m1 k m2 1 k

which we re-arrange
m1
k 1
m2
1
k
m
1 1
m2

m2
m1 m2

Check this to see if it makes sense. If m 1 = m2 then k = which makes sense by


symmetry. If m1 is very large compared to m2 then k is very small so the pivot is close to
m1. This again makes sense.

Leslie Green CEng MIEE

1 of 9

June 2015

Minimum Inertia for Two Point Masses


L
kL

m1

m2

Rotational Inertia, I, is found by summing the products of mass and the squared distance
from the pivot point. This is easier to see as an equation rather than written in a long
sentence!
2
2
I m1 kL m2 1 k L
We want to know what the correct value of k is in order to minimise I. If we know m1 and
m2 then we can plot a graph of I versus k and see how it varies. But for the general case
we would have to draw lots of graphs with different ratios of m 1 to m2 and try to discover
the relationship. Fortunately mathematicians figured out how to solve this class of
problems more than 300 years ago. The method involves differential calculus. We
consider everything on the right hand side of the equation as constant apart from k and
ask the question, what is the slope of the curve of I against k at any point?
slope of I against k

I
m1 L2 2k m2 L2 21 k
k

You may be used to seeing derivates written like this,

dI
. My using the curly d may have
dk

confused you. Dont worry! I am using the symbol for a partial derivative. The only
difference is that there are lots of variables in the equation. By using a partial derivative I
am saying, Lets just ignore that fact that there are lots of variables by pretending that
everything else is constant.
We can look at the equation for the slope and see that it is negative when k=0 but changes
to positive when k=1. The slope must therefore go through zero. So it goes down, reaches
a minimum, and then goes back up again. When the slope is zero, that must be the
minimum value. This means that
m1 L2 2k m2 L2 21 k
Cancel the 2 and the L2 to get

m1 k m2 1 k

If that doesnt look familiar than turn back one page and look at the value of k required for
balance. So we get minimum inertia when we rotate two masses about their point of
balance. To find out what the value of the inertia actually is we plug
m2
2
2
k
into
I m1 kL m2 1 k L
m1 m2

Leslie Green CEng MIEE

2 of 9

June 2015

giving

m2
I m1 L
m1 m2
2

m2

m2 L 1
m1 m2

Then re-arrange in a series of simple steps

m1m22
m1

IL
m2 L2
2
m1 m2
m1 m2

m1m22
m2 m12
2
I L
L
m1 m2 2
m1 m2 2
2

I L2

m1m2
m1 m2 L2 m1m2
2
m1 m2
m1 m2
I

m1m2 2
L
m1 m2

This Mathcad plot shows how the inertia varies with pivot point position between the two
masses. The plot has been scaled so that changing the relative masses of m 1 and m2 does
not change the inertia at the optimum pivot point.
Pivot Point Changes Rotational Inertia
4

3.5

Rotational Inertia

I( k 1) 2.5
I( k 2)
I( k 3)

1.5

0.5

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Per-Unit Pivot Position Between Two Masses

Leslie Green CEng MIEE

3 of 9

June 2015

Balance Point for Numerous Point Masses

m1L1

m2 L 2

m3L3

mNLN

L
This time all distances are measured from an arbitrary datum on the left. Each mass is
labelled with its mass-moment (the product of mass times distance from the datum).
Gravity is ignored because it cancels out, as we saw previously. The balance point is at L.
We require L times the sum of masses acting at L to be the same as the sum of the
individual mass-moments. The drawing shows 4 masses, but it is intended to illustrate N
masses, where N could vary from 1 to infinity. Without loss of generality we index (label)
the masses and distances, starting from the left.
L m1 m2 m3 mN m1 L1 m2 L2 m3 L3 mN LN

The three dots, an ellipsis, means there are missing terms in between, depending on the
value of N. We can write the same equation using the sigma notation, which just means
sum the terms from the lower index value (the r ) to the higher index value.

m m L
N

r 1

r 1

This is a convenient notation in terms of entering the equation into a computer because it
represents a for-loop. In this case
for r =1 to N step 1

or

for (r=1; r<=N, r++)

if you prefer C

Re-arranging the equation

mr Lr

r 1
N

mr

r 1

which probably looks pretty scary if you are not used to the sigma notation!
Just to see what this means in practice we consider the original case of two masses. We
need to put the measurement datum somewhere, and it happens to be convenient to put it
at the centre of m1.

Then

mr m1 m2

r 1

Leslie Green CEng MIEE

4 of 9

June 2015

And

mr Lr m1 L1 m2 L2 m1 0 m2 L2 m2 L2

r 1

m2
L2
m1 m2

The distance of the balance point from m 1 (L) is the distance between the two masses (L2)
multiplied by the k value we calculated on the first page. So the formula agrees with our
first attempt, provided we are carefully about handling the changing definition of the
variables.

Leslie Green CEng MIEE

5 of 9

June 2015

Minimum Inertia Pivot for Numerous Point Masses


m1

m2

m3

mN

L
This time we again measure distances from an arbitrary datum, but we calculate moments
of inertia relative to the pivot point (distance L from the arbitrary datum).

m L L
N

r 1

Calculate the partial derivative of I with respect to L


I
2
L

m L L
N

r 1

Find the point where the partial derivative is zero

mr Lr L 0

r 1

mr Lr

r 1

mr L

r 1

L is a multiplicative constant which we can take outside of the summation symbol since it
is common to all terms. Then

mr Lr

r 1
N

mr

r 1

But this is exactly the same equation we had for the balance point. Thus for any
configuration of masses, the balance point is the correct place to achieve minimum
inertia for rotation.
Before we proceed to derive an equation for the minimum inertia of the set of masses, we
need to simplify our notation. We have been explicitly writing out the summation index and
the start and end values. But since we know what they are, lets just not bother writing
them anymore, and read the summation as all of them.
Now we can write things like

mr Lr

mr

The required minimum inertia is


I

m L
r

Leslie Green CEng MIEE

m L
r

2
r

L2 2 LLr

6 of 9

m L
r

2
r

L2

m
r

2 L

m L
r

June 2015

I have written brackets around each of the summation terms to help you to see where the
summation ends. People often omit these brackets and you are left to guess where any
particular summation ends! Typically a plus or minus sign signifies a break in the
summation, and using that rule the brackets above would not be necessary.
This inertia equation is getting difficult to work with, so I am going to break the right hand
side into three parts.
A

so that

B L2

mr L2r

C L

mr

mr Lr

I A B 2C

First substitute for L in the equation for B


BL

mr

mr Lr

mr

Then substitute for L in the equation for C


C L

m L
r

m L

I A B 2C A B 2 B A B

Writing the total mass as

mr

1
mr L2r

mr Lr

which we should now check for the two mass situation solved previously. Again we set m 1
as the reference datum, making L1=0.
Then

I m2 L22

1
m2 L2 2 m2 L22 1 m2 m1m2 L22
m1 m2
m1 m2 m1 m2

which is the same as we found previously.

Leslie Green CEng MIEE

7 of 9

June 2015

If the Pivot is in the Wrong Place, Can Adding Mass Reduce the
Rotational Inertia?
We know the correct place for the pivot is the balance position, so maybe it is possible to
rebalance the mechanism and achieve something useful without moving the pivot point.
L

m1

kL

m2

Lets accidentally put the pivot point too far to the left so that
m1 g k L m2 g 1 k L

and therefore
We have

m1 k m2 1 k
I m1k 2 L2 m2 1 k L2
2

Notice that all terms in this equation have definite fixed values by the way the problem has
been set up. If we now add an extra mass, m 3, at an offset of x to the left of the pivot point
to improve the balance, what happens to the inertia?
I m1k 2 L2 m2 1 k L2 m3 x 2 L2
2

When presented in this way it is self-evident that the inertia increases.


There is another point to make here about the balancing mass, m 3. Suppose we pick
particular values of m3 and x to achieve balance, so that
m1 k m3 x m2 1 k

This balance condition will still be met with a new position


However, the added inertia changes from

x p

m3 x 2 pm3
p

and a new mass p m3 .


m3 x 2
p

So, for example, doubling the mass at half the offset halves the additional inertia.
We are now in a position to answer a similar question: if the pivot has been placed
correctly, can the inertia be reduced by reducing some portion of the mass, thereby
creating an imbalance. Well we have
2
I m1k 2 L2 m2 1 k L2
and we are proposing to reduce either m1 or m2. Obviously the inertia will go down!

Leslie Green CEng MIEE

8 of 9

June 2015

We can illustrate this idea with another plot. This time the inertia is directly calculated with
no re-normalisation. For the red curve we have one mass of 1.0 units and the other is 1.5
units. When the pivot is at the 1.5 unit mass we have 1.0 units of inertia whereas when the
pivot is the other side we have 1.5 units. For the blue curve the 1.5 unit mass is reduced to
1.4 units. The whole curve is lower everywhere and whilst k=0.6 is not the exact minimum
of the blue curve it is still lower than the optimum (k=0.6) point on the red curve.
Reducing Mass Always Reduces Rotational Inertia
1.5

1.4

1.3

Rotational Inertia

1.2

1.1
J( k 1.5)
J( k 1.4)

0.9

0.8

0.7

0.6

0.5

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Per-Unit Pivot Position Between Two Masses

Whilst unbalancing a shaft may seem like a way of getting a small reduction in rotational
inertia this is in all probability a really bad idea for a practical application. It is probable that
an unbalanced shaft would cause unnecessary stress on the bearings, causing them to fail
prematurely.

Summary
Rotate a load about its balance point to minimise its (rotational) inertia.
Never add mass to create a balance point if you are trying to minimise inertia.
Removing mass and creating an imbalance will reduce inertia, but the imbalance
may cause bigger problems.
Further reading: The Parallel Axis Theorem

Leslie Green CEng MIEE

9 of 9

June 2015

You might also like