Course 1 Laboratory

Second Semester

Experiment:
The shear modulus of steel

1
 Elasticity: The Shear Modulus of Steel

1 Aim of the experiment

The aim of this experiment is to measure the elastic modulus with as high precision as
possible. You will also find this experiment a valuable practice in error analysis and
planning an experiment.

2 Skills Checklist
At the end of the experiment, check that you have mastered and understood the
following main features:

• Basic error analysis,

• Use of error analysis to help plan an experiment,
• Selection of appropriate measuring instruments,
• Use and reading of standard mechanical instruments,
• Measurements of quantities to the
appropriate degree of accuracy.

3 Introduction
The shear or rigidity modulus of a material may
be measured by forming it into a wire and l
studying the torsional (twisting) oscillations of a
suspended load. Concentrate on the experimental
set-up, measurements and error estimates, rather
than on the detailed theory of the experiment. In
this experiment, precise length and timing R1 R2
measurements are needed. tube

4 Apparatus d d
A steel wire is suspended from a tripod, and Figure 1. Diagram of the apparatus.
supports a horizontal tube bearing two sliding
rings R1 and R2 each of mass M (fig. 1). You are also supplied with a stop-watch for
timing measurements. The shear modulus n is given by
16πlM
n= (1)
ma 4
where a is the radius of the wire, l is the length of the wire (Figure 1), M is the mass if
a sliding ring, and m is the slope of the graph of T2 against d2, with T being the period
of oscillation, and d the distance of the centre of the rings from the axis of rotation as
in Figure (1).

4.1 Obtaining your graph

The object of this part of the exercise is to obtain a good value for the gradient of the
graph of T2 against d2 and to use the theory of errors to give cross checks on the
experiment.

2
Take measurements of the period T for several values of d by finding the time taken
for about 10 oscillations. You may find it convenient to use clamped vernier callipers,
as shown in Figure 2, to set the position of each ring. Once the thickness of the
moveable rings and the centre suspension ring are known, the value of d can be
deduced.

R1 R2
tube

Figure 2. Measurement of d using the Vernier callipers.

A short vertical rod is provided to serve as a reference mark in timing oscillations;

where should you see it in place? We suggest that you find T for at least 6 values of d,
but no more than 10. There is no need to take more than one reading for each value of
T. Plot the graph of T2 against d2 as you take results, not afterward. This will enable
you to spot “rogue” results immediately they arise, and you can check them at once.
Use a least squares fitting programme (Mathcad: lsq3_3) to find the best value
of the gradient, m.
You will find that the experimental points will not all lie exactly in the straight line.
To investigate these deviations choose some arbitrary value of d, and time a few
batches of, say, 10 swings each. Calculate (you can use Excel) the mean value of T
and the standard deviation of the observations about the mean, i.e.
n

∑ (x − x )
i =1
i
2

σ=
n −1
NOT the standard error of the mean, which is
 i=n _ 2
σ
∑  xi − x 
i =1 
=  .
n n(n − 1)
This figure should be the error of each point on your graph.
• Put error bars equal to ± the standard deviation on each point of your graph.
About 2/3 of the error bars should intersect the best fit line; is this true for
your results?
• Compare the standard deviation with the residuals (y-ycalc) given in the
computer fit (you will have to extract them); are they comparable?
• What are the main sources of error in T? Will timing for longer reduce the
percentage of error?

The computer will give a figure for the error in m. To see if this is consistent with
the accuracy of the points on your graph, use the following (very rough) means of
estimating the error in m. Assuming that the error in d is negligible, and that there

3
 are p points on the graph, the fractional error of m will be roughly 1 / p times
the fractional error of T2, which in turn is twice the fractional error in T. Using
your value for the error in one point, estimate the percent error in m, and compare
with the computor prediction. The two results should agree to within a factor of 2
or so.

5 Planning the rest of your work

The object of this part of the exercise (which should take no more than about 30
minutes) is to use the theory of proprgation of errors as an aid in planning, the rest of
th e experiment, i.e. how accurately do you need to measure quantities such as a or l,
and what apparatus is needed to measure them.

The fractional error in n in terms of the fractional errors in a, l, M and m is given by

2 2 2 2 2
 σ n   4σ a   σ l   σ M   σ m 
  =  +  +  +  ( 2)
 n   a   l  M   m

For maximum efficiency the percent errors in m, a4, l, and M should be equal, i.e.
each term in this equation should be equal. Firstly make estimates of the errors in a, l
and M. Knowing the error in m, use equation (2) to estimate the errors σa, σl, σM that
you can tolerate in a, l, and M. What error in n would you expect to get at the end of
the experiment? Also investigate how the error in n depends on the errors in m, etc.
numerically using Excel.

Given that the resolution of a wood rule is 1 mm, that of a good steel rule is 0.1 mm,
that of a Vernier gauge is 0.02mm and that of a micrometer screw-gauge is 0.001mm
(1µm), select the appropriate instrument to make accurate measurements of a and l.
(See Squires sections 6.1 to 6.5.).

DISCUSS YOUR FINDINGS WITH A DEMONSTRATOR BEFORE GOING ANY

FURTHER.

6 Completing the Experiment

Make measurements of a, l, and M, and also find the errors in measurement. The
following points should be noted.

1. Make repeated measurements. The errors you want are the standard errors of the
means, NOT the standard deviation (see equations in section 4.1). The errors you
have estimated are target errors, and you may not be able to match the target with
the instruments available. Do not try to better the target unless to do so involves
no more effort!

2. When measuring the radius of the wire check it for uniformity (Squires, section
8.6). Take at least twelve measurements at different positions and for different
orientations. Do your measurements suggest the wire has a uniform cross-section?
Is there any evidence that it is not circular? Is the accuracy of your final value of
a4 limited by the non-uniformity of the wire or by the precision of your measuring
instrument?

4
3. When you use a micrometer, use the ratchet mechanism to tighten the jaws; do not
twist the main barrel. Unless you are a skilled operator, using the barrel will give
erratic results, and you may damage the object being measured if it is squashed
too fiercely. Remember it is possible to read a micrometer to 0.1 of the smallest
graduation on its barrel.

4. Before taking any more measurements, calculate the percentage standard

deviation of the mean value of a and hence find the percent error in a4. Is this
comparable with the target error you calculated in the planning? If not, do you
need to do anything about it, and, if it is necessary, can you do anything about it?

5. To find l, tie a short length of thread to the wire if there isn’t one there already.
Measure the distance between the top and bottom jaws holding the wire and the
marker thread. Why is this preferable to trying to measure the distance between
the jaws directly? Obtain several values, (should you move the marker thread
between each reading?), and consider whether the error in is comparable with that
in a4. Would more observations of l produce a useful reduction in this error?

6. A balance is available for determining the masses of the two rings. In the
formulae, M is the average mass of the two.

Use equation (1) to calculate the shear modulus of the steel. Use equation (2) to find
the error in your value of n, using your best values for σa, etc.

Questions
1. The wire provided is made of steel. Compare your results with values of n for
steel given in books of tables. Take the experimental uncertainty in your value of
n into account.
2. Consider the advantages and disadvantages of this method compared with the
direct twisting of the wire by applying a known couple. Can you think of any
other way of finding the rigidity modulus?
3. When finding the error bars for the graph, why are you told to use the standard
deviations of the readings of T, and not the standard error of the mean?
4. From your results, what are the main sources of error in n? Is it feasible, bearing
in mind the time and equipment, to reduce these errors? Is it really essential to
reduce the errors?

References
Spracking M.T., Liquids and Solids, Routledge and Kegan Paul, London, 1985.
Newman F.H., Searle V.H.L., The General Principles of Matter, 5th Edition, Edward
Arnold, London, 1957.

7 Appendix 1: Elastic Moduli

These notes summarise the general forms of the response of a solid body to applied
forces. Further details can be found in Sprackling, chapter 2, or Newman and Searle,
chapter 5.
The moduli of elasticity of a material are measures of its resistance to a change of size
or shape under the influence of a set of applied forces. The applied forces constitute

5
stresses, expressed as a force per unit area, and the resulting deformation is described
as a strain which is the ratio of the change in some dimension to an original
dimension. If the strain returns to zero when the stress is removed, the deformation is
said to be elastic. In many materials for small elastic strains, the deformation obeys
Hooke’s law which states that the stress is proportional to the strain. The constant of
proportionality is the elastic modulus, so that

stress
Sheer Modulus =
strain

Materials can be deformed in several different ways, corresponding to different

moduli of elasticity. For isotropic materials (i.e. those whose properties are the same
in all directions) there are three moduli of particular importance.

1. The bulk modulus, K, corresponds to a change of volume without change of

shape. This applies to deformation under a uniform hydrostatic pressure. The
stress is the pressure p, and the strain is the change in volume -δV (negative
because the change is a decrease if the pressure increases) divided by the
original volume V. The bulk modulus is
p
K=
 δV 
− 
 V 

2. The shear modulus, n, also known as the modulus of rigidity, corresponds to a

change of shape at constant volume. The stress is the tangential force F
applied over an area A divided by A, and the strain is represented by the angle
of shear, φ. The modulus is given by
F
 
n= 
A
φ

3. Young modulus, E, is used to characterise a change in length when a tensile or

compressive stress is applied, with no external forces applied to the side
surfaces of the specimen. The stress is the tensile force F divided by the cross-
sectional area A and the strain is the change in length δl divided by the original
length l. The modulus is given by E
F
 
E= 
A
 l
δ
 
l 

At the same time, there is also a contraction or expansion at right angles to the tensile
or compressive stress. The ratio of the magnitude of this transverse strain to the
principal strain is called the Poisson ratio, σ.

These various moduli are related by the equations

6
 E E
K= n=
3(1 − σ ) 2(1 + σ )

All types of elastic deformation of isotropic media can be described in terms of any
two of these moduli. Note however that an elastic modulus only has meaning if
Hooke’s law is obeyed; the ratio stress/strains not constant for a non-linear material,
even if it is perfectly elastic.

8 Appendix 2: Theory of Torsional oscillations

The theory of the torsion of a cylindrical wire is given in Sprackling, section 2.12, and
in Newman and Searle, section 5.7. These show that if a wire of length l and radius a
is twisted through an angle φ by application of a couple, the couple per unit angle of
twist τ (the torsional rigidity) is
πna 4
τ= ( A1)
2l
where n is the shear modulus of the material of the wire. When a body of moment of
inertia ℑ is suspended from the wire, torsional oscillations may be set up, with period
ℑ
T = 2π ( A2)
τ

In the apparatus used here (known as Maxwell’s needle), the body consists of the
horizontal tube carrying the rings R1 and R2 of equal mass M. When the rings are a
distance d from the mid point, then
ℑ = ℑ0 + 2 Md 2 ( A3)

where ℑ0 is the moment of inertia of the tube and central suspension plus the moment
of inertia of the rings about axes through their centres parallel to the wires.
Combining equations (A2) and (A3),
2 8π 2 Md 2 4π 2 ℑ0
T = + ( A4 )
τ τ

8π 2 M
Thus a graph of T2 against d2 will be a straight line, with gradient m= and
τ
4π 2 ℑ0
intercept . Since m may be measured, τ and hence n may be calculated using
τ
the formula
16πlM
n= ( A5)
ma 4

This is equation (1) of the main text.