PHYSICS

Only study guide for

PHY2606

Waves

Compiled by
Prof SA Rakitianski

Revised by
Ms MM Tibane
Mr SJ Moloi

Department of Physics
University of South Africa
Pretoria
© 2010 University of South Africa

All rights reserved

Printed and published by the

University of South Africa
Muckleneuk, Pretoria

Page layout by the Department

PHY2606/1/2011
1 Introduction 1

1.1 Welcome . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.2 The study material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.3 Assignments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2 The course 6

2.1 Simple Harmonic Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2.2 Damped Simple Harmonic Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

2.3 The Forced Oscillator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

2.4 Coupled Oscillations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

2.5 Chapter 5: Transverse Wave Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

2.6 Fourier Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

2.7 Waves in Optical Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

2.8 Interference and Diffraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

3 Examination Preparation 35

3.1 What is the examination all about? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

3.2 Trial examination Paper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

3.3 Discussion of the Trial examination Paper . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

3.4 Preparation for the examination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

3.5 The examination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

3.6 Model Answers to Trial examination Paper . . . . . . . . . . . . . . . . . . . . . . . . . . 42

Chapter 1

Introduction

1.1 Welcome
We are glad that you got through all the first level modules and are about to embark on something new.
Congratulations and welcome!
What is new and special about this module? Is it something new, indeed? Yes, it is. So far you have
studied a course of the so-called General Physics where the central role is given to a physical phenomenon
while the mathematical equation appears as a result of a quantitative understanding of it. In other words,
in General Physics we go from the phenomena to the equation and very often even do not derive the
equation but prove it by certain experiments. Of course, there is nothing wrong with this, and, as a
matter of fact, much (if not most) of the physical knowledge was generated by scientists in just this way.
This, however, makes general physics consisting of different subjects such as mechanics, optics, molecular
physics, electricity, etc., seem to have nothing in common with each other. But actually they have much
in common. Physics is a whole science not fragmented into a mosaic of different subjects. This will
become clear when you start studying Theoretical Physics.

Of course, theoretical physics also has its deep roots in experiments, but is constructed differently. It is
built up like a mathematical theory: Certain axioms (such as conservation of energy, momentum, etc.)
and assumptions about the forces acting in a physical system, serve as the foundation from which various
conclusions are derived via pure logical reasoning. There is no need to refer to experiments in order to
prove equations which are derived in this way. Various physical phenomena are rather used to illustrate
how these equations work. Furthermore, the equations are very often used to predict new, unknown
phenomena.

The beauty of theoretical physics is that it reveals general features of seemingly different physical pro-
cesses, such as vibrations of atoms in a molecule, oscillations of an electric current in a circuit, and
swings of a pendulum. The PHY2606 is the first module where you will become acquainted with the
methods of theoretical physics. Vibrations and waves which you will be studying in this module, are the
most fundamental phenomena in physics. Just look around, and you will find something oscillating or
producing waves. So, the time and effort invested in studying this module, will not be wasted. It will
help you in your future studies and work, whatever career you will pursue.

1
 2

We welcome you in taking your first step into the beautiful and challenging world of theoretical physics.

1.2 The study material

The present module is based on the sixth edition of the book ”The Physics of Vibrations and Waves” by
H. J. Pain. It was published in 2005 by John Wiley & Sons. This book is quite thick because it covers all
aspects of the theory and even describes several most important areas of its applications. In principle, it
would be very useful to go through the whole book, but we understand, of course, that this is unrealistic.
We have chosen certain sections of the book (see the list below) which we prescribe as the compulsory
material. The other sections, may be read or consulted in the future when studying Electrodynamics,
Quantum Mechanics, and other theoretical subjects.

The book by Pain is a kind of encyclopedia on vibrations and waves. Please do not despair and do not
feel inferior just because you cannot read the whole book all at once. As a matter of fact, nobody does
it in this way. Usually, when studying a theory, we learn the main principles first and then come back to
the details when we start implementing them.

Before you start reading the prescribed textbook, it would be very helpful if you could spend some time
in refreshing your memory on oscillations and waves. Just take one of the books on General Physics,
such as the courses by Cutnell and Johnson, or Halliday, Resnick and Walker, and browse through the
relevant pages. This will put you in the appropriate state of mind.

The only drawback of Pain’s text is that the sections and equations are not numbered (this is, however,
compensated by it’s clarity). As a result, each section and equation as to be referred to explicitly whenever
it is necessary. The following chapters and sections of Pain’s book are to be studied:

• Chapter 1: Simple Harmonic Motion

– Introductory pages
– Displacement in Simple Harmonic Motion
– Velocity and Acceleration in Simple Harmonic Motion
– Energy of Simple Harmonic Oscillator
– Simple Harmonic Oscillations in an Electrical System
– Superposition of Two Simple Harmonic Vibrations in One Dimension
– Superposition of a Large Number n of Simple Harmonic Vibrations of Equal Amplitude a
and Equal Successive Phase Difference δ
– Some Useful Mathematics
– Chapter 2: Damped Simple Harmonic Motion
– The entire chapter
– Chapter 3: The Forced Oscillator
– The operation of i upon a Vector
 3 PHY2606/1

– Vector form of Ohm’s Law

– The Impedance of a Mechanical Circuit
– Behaviour of a Forced Oscillator
– Behaviour of Velocity v in Magnitude and Phase versus Driving Force Frequency ω
– Behaviour of Displacement versus Driving Force Frequency ω
– Power Supplied to Oscillator by the Driving Force
– Variation of Pav with ω. Absorption Resonance Curve
– The Q-Value in Terms of the Resonance Absorption Bandwidth
– The Q-Value as an Amplification Factor
– The Effect of the Transient Term

• Chapter 4: Coupled Oscillations

– Stiffness (or Capacitance) Coupled Oscillators

– Normal Co-ordinates, Degrees of Freedom and Normal Modes of Vibration
– The General Method for Finding Normal Mode Frequencies, Matricies, Eigenvectors and
Eigenvalues
– Coupled Oscillations of a Loaded String
– The Wave Equation

• Chapter 5: Transverse Wave Motion

– Partial Differentiation
– Waves
– Velocities in Wave Motion
– The Wave Equation
– Solution of the Wave Equation
– Characteristic Impedance of a String (the string as a forced oscillator)
– Reflexion and Transmission of Waves on a String at a Boundary
– Reflexion and Transmission of Energy
– The Reflected and Transmitted Intensity Coefficients
– The Matching of Impedances
– Standing Waves on a String of Fixed Length
– Energy of a Vibrating String
– Energy in each Normal Mode of a Vibrating String
– Standing Wave Ratio
– Wave Groups and Group Velocity
– Doppler Effect

• Chapter 10: Fourier Methods

– Fourier Series
 4

– Application of Fourier Sine Series to a Triangular Function

– Application to the Energy in the Normal Modes of a Vibrating String
– Fourier Series Analysis of a Rectangular Velocity Pulse on a String
– The Spectrum of a Fourier Series

• Chapter 11: Waves in Optical Systems

– Light. Waves or Rays?

– Fermat’s Principle
– The Law of Reflection
– The Law of Refraction

• Chapter 12: Interference and Diffraction

– Interference
– Division of Amplitude
– Newton’s Rings
– Michelson’s Spectral Interferometer*1
– Fabry–Perot Interferometer*
– Division of Wavefront
– Interference from Two Equal Sources of Separation f
– Interference from Linear Array of N Equal Sources
– Diffraction*
– Intensity Distribution for Interference with Diffraction from N Identical Slits
– Fraunhofer Diffraction for Two Equal Slits (N = 2)
– Transmission Diffraction Grating (N Large)
– Fresnel Diffraction

We estimated that these sections, comprising about 170 pages, can be studied (including writing the
assignments) in 120 study hours. In doing this estimation, we assumed that during one hour you can
go through 3 pages, working out all the mathematical derivations. So, approximately half of the time
should be devoted to reading while the other half to solving the assignment problems and preparing for
the examination..

All the prescribed material is contained in the textbook. In Part 2 of this study guide we have tried to
clarify some points in Pain’s text. However, the study guide is no substitute for the book. You can read
our comments only where you feel they will be of help to you, otherwise you may just ignore them. This,
however, does not apply to Part 3 which is entirely devoted to examination preparation. We strongly
recommend you to read it at various stages throughout the course, perhaps after completing each assign-
ment. The information contained therein could save you wasting time when the pressure is on, i.e. the
few days before the examination.

1
These sections are recommended but are not compulsory for studying
 5 PHY2606/1

1.3 Assignments
All the assignment problems are taken from the prescribed textbook. Their numbers (fortunately, Pain
at least numbered the problems) are given in the Tutorial Letter 101. Submit your assignment scripts
in good time. Rather submit an incomplete assignment on or before the due date, than a complete but
late assignment. The reason being that complete solutions of the assignment problems are sent out to all
students soon after the due date. It is clear that we cannot accept any late assignments in such a situation.

On receiving our solutions, take time to work through them, paying particular attention to any problems
with which you may have struggled. Firstly, this exercise will often assist you with the completion of
subsequent assignments, and secondly, it will be a part of your preparation for the examination. Keep
in mind that assignment problems are not only prescribed to monitor your progress, but also to accel-
erate and entrench your learning. We, therefore, advise you to read all the problems at the end of each
prescribed chapter and to think about possible ways of solving them. Of course, you should submit only
the prescribed problems, but for yourself make an attempt to solve all of them. Definitely, these efforts
will be rewarded in the examination.

Write your scripts neatly! Try to make your handwriting as clear as possible. Give some logical expla-
nations of how you came from one formula to another. Believe me, these are not trifles. These points
psychologically influence a lecturer when he is marking your assignment. You can gain or loose 10% of
the mark on these trivialities. Sometimes it is crucial.
Chapter 2

The course

2.1 Simple Harmonic Motion

What is important to understand from the outset is that pure harmonic motion is only possible when a
force F , acting on a moving body, is proportional to its displacement x from the equilibrium position,
i.e.
F = −sx . (2.1)
Always keep in mind, however, that a simple pendulum or a ball attached to a spring, is not necessarily
a change of the linear co-ordinate x (using a bit of abstract thinking).

For an electrical LC circuit, for example, x is a charge of the capacitor equilibrium position for which
the charge is zero. In this case you can consider the disturbance x of the circuit as displacement of a
certain number of electrons from one plate of the capacitor to the other. The electric force (the so-called
electromotive force) is proportional to the charge and tries to return the electrons back, moving them
through the coil. The coil, however, cannot pass all the charge immediately, because it has an inductance
(kind of inertia). The current flowing through the coil, increases and reaches its maximum when the
capacitor is totally discharged (equilibrium point). But the current in the coil cannot stop at this point,
again due to the coil inductance. By the time the current decreases to zero, too many electrons have
been transferred and the capacitor is charged again but with different signs of the plates. The process
repeats itself in the opposite direction.

The oscillations are possible when a system has the two essential properties: stiffness (which generates
the restoring force) and inertia (which keeps the system moving when it reaches the equilibrium). The
frequency of free oscillations is, therefore, determined by these two properties.

What makes similar very different physical processes, like vibrations of a ball on a spring and oscillations
of the capacitor charge in the LC circuit, is the that in these processes the displacement x satisfies the
equation
ẍ + ω 2 x = 0 , (2.2)
from which we can find how the displacement changes with time. Perhaps you are not satisfied with the

6
 7 PHY2606/1

way this equation is solved in the textbook where the solution is simply stated. There are techniques to
solve simple differential equations, and you should soon encounter them in your mathematical courses,
if you have not already done so. However, the method used in the book is quite acceptable (especially
for simple equations like this). A solution is guessed and substituted into the original equation. If the
equation is satisfied then our guess was correct.

You will soon learn that any linear second–order differential equation has two linearly independent solu-
tions. Any linear combination of them is also a solution. In the case of Eq. (2.2) we have two independent
solutions, x = cos ωt and x = sin ωt (they are independent because no coefficient can transform sinωt
into cosωt). Their linear combination
x = A cos ωt + B sin ωt (2.3)
is also a solution, and a choice of the two parameters A and B determine a particular solution obeying
certain initial conditions.

Generally speaking, the constants A and B can be complex, but in this case they are not totally inde-
pendent and should be subjected to an additional constrained condition because each complex constant,
in fact, represents two real constants (real and imaginary parts) while, according to the general theory of
differential equations, only two parameters determine a particular solution. In the section ’Some Useful
Mathematics’ you will find that a complex exponential function
X = aeiωt+iφ (2.4)
also satisfies Eq. (2.2). This is because, according to the Moivre’s theorem, which states that
eiϕ = cos ϕ + i sin ϕ , (2.5)
the function (2.4) is just a linear combination of cos and sin, i.e.
X = a cos(ωt + φ) + ia sin(ωt + φ) . (2.6)
If we choose the amplitude a and the phase–shift φ to be real then these two parameters can be fixed
by the initial conditions. Sometimes it is very useful to write the general solution of Eq. (2.2) in the
form (2.4). Perhaps you are confused why a real displacement can be described by a complex function.
There is nothing mystical about it. Any point z on a complex plane can be indicated either by giving
its co-ordinates, Re z and Im z, or alternatively, by giving its distance r from the origin and the angle ϕ
between the real axis and the beam directed to this point as is shown in Fig.2.1. Taking into account
Moivre’s formula (2.5), we can write this complex number as
z = reiϕ . (2.7)
We see that the function (2.4) is of the same form with r = a and ϕ = ωt + φ. Since ϕ uniformly
increases with time, we can say that the function (2.4) describes a point in the complex x–plane moving
along a circular trajectory with a constant angular velocity ω in the anti-clockwise direction. The ampli-
tude a gives the radius of the circle and the phase–shift φ the initial angular position of the point at t = 0.

The beauty of the form (2.4) is that one complex number X contains information about both the ampli-
tude and the current phase ϕ. Furthermore, as you have seen, this form relates oscillations to uniform
circular motion. When z moves along a circle, its real and imaginary parts harmonically oscillate, viz.
Re z = a cos(ωt + φ) , (2.8)

Im z = a sin(ωt + φ) . (2.9)
 8
Im

Im z rz

~r

< 
ϕ
Re
Re z

Figure 2.1:

Therefore, to describe the oscillations of the displacement x we can use either Re z or Im z. Actually,
there is no principal difference between them as the cos and sin are just the same functions, simply shifted
by π/2 from each other. The choice between (2.8) and (2.9) is a matter of choosing the phase–shift φ
(which would differ by π/2). In different textbooks you can find formulae with cos as often as with sin.
However, since Pain opted for sin, we assume that
 
x = Im aeiωt+iφ . (2.10)

Complex representation of oscillations is very convenient, (I would rather say indispensable) when we deal
with superposition of two or more oscillatory motions having the same frequency but different phases. To
obtain the total displacement and the resulting phase, we need just to add all the complex displacements
which contribute to it. We need not worry about the phase differences among them. The correct phase
of the total displacement will appear automatically as the phase of the resulting complex number, if the
summation is correctly made.

The easiest way to do such a summation is to represent the complex numbers by vectors having two
components: real and imaginary parts, and apply the rules of the vector addition. As you know, to make
a sum of several vectors, we draw them one after another (putting the beginning of the next vector at
the end of the previous one) and connect the beginning of the first vector with the end of the last one.
This connecting vector is their sum. A very useful example of vector addition is given in Fig. 1.11 (page
21 of the textbook). The formula for the length of the resulting vector,
sin nδ/2
R=a ,
sin δ/2
derived there, is used in the last of the prescribed chapters to describe the diffraction in the Fraunhofer’s
approximation.
 9 PHY2606/1

2.2 Damped Simple Harmonic Motion

As we have already mentioned in the Introduction, the specific characteristic of Theoretical Physics is that
it goes from mathematical equations to physical phenomena. In the previous chapter we assumed that a
moving object feels only the restoring force which is proportional to the displacement. This assumption
led us to the following equation of motion (recall that F = mẍ)

mẍ = −sx , (2.11)

the solution of which lead to the discovery of the harmonic oscillations. In this chapter, we make one
step forward by adding another term to the right hand side of Eq. (2.11), namely, the resistance term.
We assume that the resistance force is proportional to velocity, hence instead of (2.11), we have

mẍ = −sx − rẋ . (2.12)

The rest of this chapter is devoted to searching for various solutions of Eq. (2.12) and their physical
interpretation.

At the right hand side of Eq. (2.12) we have two different forces. One of them is proportional to the
displacement and the other to velocity. Since the displacement is maximal when the velocity is zero and
vice versa, these forces are acting out of phase: when one of them is growing the other one is reducing.
Moreover, they always act in the opposite direction. During the oscillations, we have an interplay between
them and the ratio of their strengths is very important. The values of the coefficients s and r determine
which one is stronger.

As discussed in the book, there are three possible relations between these coefficients, which lead to three
different types of solutions of Eq. (2.12):

1. HEAVY DAMPING
The condition for heavy damping is, r2 /4m2 > s/m ,
which leads to aperiodic motion. The system, when disturbed, does not oscillate but returns slowly
to the equilibrium point. In its final stages the displacement is diminished exponentially.
2. CRITICAL DAMPING
Hence we have the condition r2 /4m2 = s/m ,
which also leads to aperiodic motion similar to the case of heavy damping. What is important for
applications is that under this condition of critical damping a system requires the minimal time
to come to rest (less than for heavy or light damping).
3. LIGHT DAMPING
Light damping occurs when r2 /4m2 < s/m ,
leads to exponentially decaying oscillations.

Of course, as far as the course of ’Vibrations and Waves’ is concerned, the first two of these three types
of solutions are irrelevant, but they describe some realistic physical processes. We found them while
analysing all possible solutions. This is an example of how theoretical physics can lead us to something
which we did not mean to study, and reveal hidden relations among seemingly different phenomena. A
prominent theorist, L. D. Landau, used to say that equations are more clever than we are.
 10

For all three cases, Eq. (2.12) is solved by guessing a solution and then verifying that it satisfies the equa-
tion. Note that in constructing a general solution, we always introduce two arbitrary constants (because
the equation is of the second order) which are to be fixed by some initial conditions.

You should spend some time to ponder over the physical meaning of the parameters characterising the
speed of damping. These parameters are:

• Logarithmic decrement
which shows the logarithmic ratio of the amplitudes for two successive swings;
• Relaxation time (lifetime)
which shows how long it takes before the amplitude decreases by a factor of 2.71828;
• Quality factor (Q-value)
which is the number of cycles after which the system practically comes to rest.
Indeed, in the course of Q cycles (i.e. after a time
2π
QT = Q = 2πm/r)
ω′
the amplitude is reduced by a factor
   
r r 2πm
exp − QT = exp −
2m 2m r

= e− π

≈ 0.043 .

Note that here we used the scientific notation exp(z) ≡ ez for the exponential function. Pain
gives a different interpretation to the Q-value, namely, as the number of radians through which the
damped system oscillates as its energy decreases in 2.71828.... times, which, of course, is also true.
But I feel that a number of radians is something more abstract and less tangible than a number
of swings. Thus, if somebody says to you that a particular pendulum has the Q-value of 50, you
may think in terms of radians, but you should also immediately realize that the pendulum will
make about 50 swings before the oscillations become imperceptible.
 11 PHY2606/1

2.3 The Forced Oscillator

In my opinion, the way Pain started this chapter is a bit confusing. To be consistent, we should begin
from the equation of motion as we did before. In this chapter we add a new term, F0 cos ωt, to its right
hand side, viz.
mẍ = −sx − rẋ + F0 cos ωt , (2.13)
which represents an external harmonic force acting on the oscillating system. The rest of the chapter is
a study of the consequences of the inclusion of this force.

In the first section, Pain explains some mathematical tricks involving the complex unit i. When he speaks
about the operation of i upon a vector, you must bear in mind that he means vectors in the complex
plane, like the one which I showed in Fig. 2.1 of this study guide. This operation is applicable not to any
vector but only to vectors indicating points in the complex plane. As we have seen, such a point
can be parametrized by its distance r from the origin and the angle ϕ between the real axis and the line
directed to it, as follows
z = reiϕ . (2.14)
Since the complex unit is  
θ theta
i = eiπ/2 = cos + isin =i (2.15)
2 2
the point
iz = rei(ϕ+π/2) (2.16)
is at the same distance from the origin but in the direction which is turned by 90◦ anti-clockwise. In
other words, when a vector ~r (in the complex plane) is multiplied by i, it turns by π/2. Actually, we do
not need to use the term vector when dealing with a complex plane. A complex number itself has two
components and exactly indicates its position in the plane. Unfortunately, in many textbooks (like the
one you are studying) the term vector in the complex plane is used.
REMEMBER: A VECTOR IN THE COMPLEX PLANE IS SIMPLY A COMPLEX NUM-
BER.

If you forget about this, you may be confused with such formulae as, for example, the so-called vector
form of Ohm’s law,
V = IZ . (2.17)
Indeed, what kind of product do we have here? When two (genuine) vectors are multiplied, we can make
either a scalar product (the result is a scalar) or a vector product (the result is a vector normal to the
plane). Apparently, none of these is applicable to Eq. (2.17). This is because the “vectors” used in this
equation, in fact, are not ordinary vectors. They are just complex numbers. Thus,
their multiplication produces another complex number. You should remember this to avoid confusion.

As can be seen in the section devoted to Ohm’s law, complex numbers are very useful for describing
oscillations in electrical circuits. The (complex) impedance is a natural generalization of the ordinary
(real) resistance. The logic in this section is quite clear. However, the formula for the mechanical
impedance  s
Zm = r + i ωm − , (2.18)
ω
it seems to appear out of the blue in the subsequent section. You should not worry too much about it.
This is just a definition of the mechanical impedance. Intuitively, you can understand why it is defined
 12

in such a “strange” way. The three terms at the right hand side of Eq. (2.18), namely,
s
r + iωm − i , (2.19)
ω
stem from the corresponding term of
rẋ + mẍ + sx , (2.20)
The terms in (2.20) represent the acting forces. To obtain the impedance the force should be divided by
the velocity ẋ. Divide these terms by ẋ, using the maximum values for x, ẋ, and ẍ, namely,

x = a

ẋ = aω

ẍ = aω 2 ,

and taking into account that the velocity leads the displacement by a phase angle of π/2 (this means
that in the ratio x/ẋ, velocity must be multiplied by i) and the acceleration leads the velocity also by
π/2 (in the ratio ẍ/ẋ, the acceleration must be multiplied by i), we obtain exactly the terms (2.19).

Anyway, if you are still not satisfied with these arguments, disregard it. In the next section, the right
hand side of Eq. (2.18) appears in a natural way in the denominator of the amplitude of steady state
oscillations,
−iF0
A= , (2.21)
ω[r + i(ωm − s/ω)]
where we can use the notation (2.18) in order to simplify this and many subsequent formulae. This is
another way of defining the mechanical impedance.

This was a digression from our main stream of developing the theory, which flows from the equation of
motion. It is convenient to rewrite Eq. (2.13) in the form

mẍ + rẋ + sx = F0 cos ωt , (2.22)

where the driving force is an inhomogeneous term added to the corresponding equation of the previous
chapter. According to the general theory of differential equations of the type (2.22) (which is linear and
with constant coefficients), the general solution of an inhomogeneous equation can be found as a sum
of the general solution of the corresponding homogeneous equation (with zero right hand side) and any
particular solution of the inhomogeneous equation. Thus, we can write the general solution of (2.22) as
r !
 r  s r2
x(t) = C exp − t exp it − + A exp(iωt) . (2.23)
2m m 4m2 | {z }
| {z }
general solution of particular solution of
inhomogeneous eq. homogeneous eq.

The first term of this solution is the so-called transient term which plays a role only at the beginning
of the oscillations and exponentially dies away. The second term describes the steady state oscillations.
From the physical point of view, the transiet term describes certain instabilities of the oscillations, which
take place immediately after the driving force is switched on, and lasts until the system “tunes” to this
force.
 13 PHY2606/1

The amplitudes of the displacement, the velocity, and the acceleration for the steady state oscillations,
F0
max{|x|} = , (2.24)
ωZm
F0
max{|ẋ|} = , (2.25)
Zm
ωF0
max{|ẍ|} = , (2.26)
Zm
are functions of the driving force frequency ω. All three of them in their denominators have Zm which is
minimal at the so-called resonance frequency
r
s
ω0 = . (2.27)
m

However, since the displacement and acceleration amplitudes have also a factor ω in the denominator
and numerator respectively, their maxima are slightly shifted towards lower and higher values of ω. The
velocity amplitude attains its maximum exactly at ω = ω0 . So does the power absorbed by the system.
The sharpness of the resonance curves depends on the rate of the energy dissipation, or in other words,
on the resistance r which determines the width,
r
Γ= , (2.28)
m
of the power absorption curve at the half–height.

The Q–value is a very important parameter characterizing resonant properties of an oscillator:

• it tells us how many cycles are needed to dissipate all the energy after the driving force is switched
off;
• it gives the ratio of the resonance frequency to its width, Q = ω0 /Γ (note that, the greater ω0 the
greater Q with the same Γ);
• it shows us how many times the oscillation amplitude is amplified when we “tune” the frequency
to ω0 .
 14

2.4 Coupled Oscillations

In the first section, we just make the coordinate transformation

X =x+y , (2.29)
Y =x−y , (2.30)

which enables us to decouple the equations of motion. As soon as they are decoupled, we immediately
obtain the so-called mode frequencies (the coefficients in the decoupled equations). Actually, we man-
aged to decouple the equations so easily because the chosen system was very simple. Indeed, if you try
to do the same when the masses or the lengths of the pendulums are different, you will immediately find
that the simple transformations (??) and (2.29) do not help and a more general method is needed. So,
how can we proceed in such cases?

The first two sections, gave us an idea of normal modes. Note that the term normal is used to indicate
that the modes are independent or, in a sense, perpendicular to each other, the same as perpendicular
vectors which have no projections onto each other. In reality, you can come across very complicated os-
cillators having many degrees of freedom and, therefore, many normal modes. In principle, for each these
oscillators there are specific transformations which decouple the equations of motion, but nobody knows
them. If you are lucky, in certain simple cases, you will find them, but how to find them in general cases?
There are special mathematical methods for finding the transformations, but they are too complicated
to be included in this course. So, what to do?

The third section gives an answer to this question. It does not prescribe a recipe for finding the normal
coordinates, but describes how to find the frequencies of the normal modes. The idea is very simple. Since
in each mode all parts of the system move with the same frequency, we can substitute into equations of
motion the exponential solutions, ∼ exp(iωt), with the same ω for all physical (not normal!) coordinates
and find out which values of ω satisfy the system of resulting equations. Actually, the problem is reduced
to solving a system of linear algebraic equations. In this way you will find all the normal frequencies
(eigenvalues) without knowing anything about the appropriate coordinate transformations. In fact, you
do not need the normal coordinates themselves. What is essential is the normal frequencies. Indeed, if
you apply an external harmonic force to a multi-dimensional oscillator, it will show a resonant behaviour
when the frequency of the force is close to any of the mode frequencies.

This algebraic method of finding normal frequencies is then applied to ‘Coupled oscillations of a loaded
string’. Reading this section, you will probably be confused by the subscripts r, n + 1, and J. Don’t
panic! Actually, it is easy. To make the formulae understandable, choose a specific n and try to rewrite
all the formulae for this particular number of masses.

Let us consider a system of three oscillators as is shown in Fig. 2.2. In this case n = 3, but since the
subscript r runs from 0 to n + 1, we have to add two extra masses at both ends of the string, which are
numbered by 0 and 4. Actually, these extra masses are “fictitious” because they do not move. They are
needed only to make the equation of motion,
T
ÿr = (yr−1 − 2yr + yr+1 ) , r = 1, 2, 3 (2.31)
ma
universally applicable for all oscillating masses including those which are next to the boundaries. Indeed,
if we take, for examinationple, r = 1, Eq. (2.31) requires yr−1 that is y0 . Similarly, we need y4 at the
 15 PHY2606/1

0 1 2 3 4
@
@ z z
@
 @ H
HH
@
@  @ H
HH
@  @
@z Hz
  HH
 @
@ @
@ @
@ @
@
@z
@ @
@
@

Figure 2.2:

right hand side boundary.

Actually, Eq. (2.31) is equivalent to the following system of coupled equations


T

 ÿ1 = (y0 − 2y1 + y2 ) ,



 ma



T
ÿ2 = (y1 − 2y2 + y3 ) , (2.32)


 ma





 ÿ T
3 = (y2 − 2y3 + y4 ) ,
ma
where y0 = y4 = 0. Substituting into these equations y1 = A1 eiωt , y2 = A2 eiωt , and y3 = A3 eiωt , we
obtain the linear algebraic system of equations,
  
 maω 2

 2 − A1 − A2 = 0,


 T


  
 maω 2
−A1 + 2− A2 − A3 = 0,

 T



  


 maω 2
 − A2 + 2− A3 = 0 ,
T

for the three amplitudes. In the matrix form it reads

   
maω 2
2 − −1 0    
 T
  A1   0
 
   

 maω 2 
   
 −1 2− −1    
 T  A  
= 0

(2.33)
  2   


      

 maω 2 
   
 0 −1 2− 
 T  A3 0
 16

Since this is a homogeneous system (the right hand side is zero) it has a nontrivial (non-zero) solution if
and only if its determinant is zero (see any book on linear algebra), i.e.
   
maω 2
 2− T −1 0 
 
   

 maω 2 

 −1 2− −1 
det  T =0. (2.34)
 
   

 maω 2  
 0 −1 2− 
 T 

Eq. (2.34) is of the third order in ω 2 and, therefore, can give us three different solutions, three normal
frequencies ω1 , ω2 , and ω3 . Pain uses the notation ωJ for them. In our case the subscript J takes only
three values, 1, 2, 3.

This method of finding ωJ works practically in all cases. If for examinationple, the oscillating bodies
shown in Fig. 2.2 had different masses we would obtain a system of equations similar to (2.33), which
would involve some mass–dependent coefficients. We would still be able to solve such a system in the
same manner.

However, the way Pain solves the problem for an arbitrary number of oscillators on a string is not
universal. It is essentially based on the symmetry of the system. Firstly, it uses the fact that all the
masses are equal, and secondly, that they are equally spaced along the string. As a result, he obtains the
ratio
Ar−1 + Ar+1
Ar
does not depend on r, i.e. is the same for any three neighbouring masses.

You may be puzzled: Why does he substitute Ceirθ instead of Ar . I remind you that in mathematics
‘the result justifies (or excuses) the means’. In other words, no matter how you derived the solution, it is
acceptable if and only if it is correct. We can always verify a solution by substituting it into the original
equation.

We can guess that Ar changes along the string (different r corresponds to different positions of the
oscillators on the string) harmonically, i.e. the dependence of Ar on r should be described either by
sin or cos functions. We have learnt from experience (practical and theoretical) that standing waves,
have nodes and anti-nodes with harmonically increasing and decreasing amplitude between them. At the
beginning of this study guide I explained that the complex exponential function

exp(irθ) = cos rθ + i sin rθ ,

represents harmonic oscillations in two perpendicular directions, and we chose to use its imaginary part
to describe real oscillations. Thus, the attempt to find an amplitude dependence on r in the form

Ar = C exp(irθ)

looks reasonable. You can easily verify that the solution

rJπ
Ar = C sin
n+1
 17 PHY2606/1

satisfies the equation. Moreover, this formula coincides with our experience. Indeed, when J is large
(large frequency ωJ ) the sin generates more nodes.

The derivation of the wave equation in the last section of this chapter is straightforward. Imagine what
happens if we put many oscillators on the string, reducing their separation and masses. As a result
we will obtain a real massive string (not an infinitely light one which connects the balls depicted on
Fig. 2.2) every part of which oscillates harmonically. Instead of r which represented the ordering number
of the balls, we indicate the position of an element of the string by the continuous variable x. The finite
differences become infinitely small and the derivative appears. Perhaps, it is worthwhile to go through
the first section (‘Partial Differentiation’) of the next chapter before you start reading the last section of
this one.
 18

2.5 Chapter 5: Transverse Wave Motion

In this chapter the wave equation obtained at the end of the previous one, is once more derived but
in a different way. You might get the feeling that this derivation is a bit clumsy, but it is correct. All
the “approximations” used, are justified by the choice of a very short piece of the string. Actually, they
are not approximations but exact relations when we take an infinitely short piece. Remember how the
derivative is introduced in the calculus. This is something similar. The formula
"  2 #1/2
∂y
ds = 1 + dx
∂x

which seems to appear out of the blue, is in fact the Pythagoras’ theorem for the triangle shown in
Fig. 2.3. The formula    
∂y ∂y
−
∂2y ∂x x+dx ∂x x
=
∂x2 dx
is just a definition of the second derivative, where we should consider dx as a finite interval which
eventually becomes infinitely small.


 !
ds  ∂y

 dx
  ∂x

dx

Figure 2.3:

Solution of the Wave Equation

The wave equation describes all types of waves. This is why any function f (ct−x) or f (ct+x) satisfies this
equation. In this chapter we study only harmonic waves and hence use the sine function for f . In general,
the form of the wave is determined by the time dependence of an external force which sends the wave
through the medium. This dependence can be periodic but not necessarily harmonic. For examinationple,
we can produce a wave with the profile shown in Fig. 2.4, which is also a possible solution f (ct − x) of
the wave equation but, obviously, is not sine or cosine function (soon, in Chapter 10, you will learn that
such fancy waves can be described as superpositions of many harmonic waves moving through the same
medium at the same time).

You may be confused when on page 119 you encounter the formula

y = a sin(ωt − φ)

with the minus sign in front of the φ. Although the choice of the phase–shift is a matter of convention,
to be consistent, it would be better to use the same formula as in Chapter 1, namely,

y = a sin(ωt + φ) .
 19 PHY2606/1

p pppppp pp pppp p pppppp pp pppp

y 6
p
ppp pp JJ

ppp ppp J ppp pp

ppp p p ppp

p J
J

J
J

J
J -
J

pppppp
x
pp p ppp
J

J pp ppp

Jp p p p p pp pp

Figure 2.4:

We would then obtain φ = −2πx/λ for a wave moving to the right, and φ = +2πx/λ for a wave moving
to the left.

On page 120, Pain introduces the phase velocity as the partial derivative ∂x/∂t. You may be puzzled:
how can we differentiate one independent variable over another independent variable? The explanation
is simple: among three variables x, y, t, we can always choose any two to be independent while the third
will be a function of them. If for examinationple, we have

y = a sin(ωt − kx) ,

then
y
arcsin = ωt − kx ,
a
and therefore
1 y
x=
ωt − arcsin . (2.35)
k a
In the last equation x is a function of t and y, and we can find the phase velocity as

∂x ω
=
∂t k
(2.36)
= c.

Characteristic Impedance and Transmission of Waves

The characteristic impedance is a very important property characterizing the transmission line. When
sending signals through a communication line, we want to minimize the losses of their energy. One obvi-
ous reason for attenuation of the signals is the energy dissipation which can only be reduced by using high
quality lines. However, if your line consists of two or more pieces (which is usual), you may encounter
significant losses even if all the pieces are of the best quality but have different impedances. This is
because waves are reflected at the joints unless the impedances Zl and Zr at the left √ and right sides are
equal or matched via a transition piece of the length λ/4 and the impedance Z = Zl Zr .
 20

Although in this Chapter we study only mechanical lines for which the impedance Z = ρc is the product
of the linear density and the wave velocity, this simple rule of the impedance matching is universal and
applicable to all kinds of the transmitting lines, including optical and electromagnetic. More about elec-
tromagnetic transmission lines can be found in Chapter 7 which, however, is not prescribed due to the
limitation of the study time. To those who are planning their career in industry related to radio, TV,
computer hardware, networks, etc., I would advice to read Chapter 7 anyway.

Standing Waves

When reading about standing waves, you should take heed of the stationary wave equation which can be
written as  2 
d 2
+ k y(x) = 0 (2.37)
dx2
for a wave function y(x) of only one variable. Then the total wave function is the product y(x) exp(iωt)
where the harmonic time dependence exp(iωt) is factorized.

If the string is clamped at both ends, this equation has solutions not for any k but for certain discrete
values k1 , k2 , k3 , . . . , which are called eigenvalues. In this case the solution y(x) represents the station-
ary distribution of the amplitude along the x–axis.

Energy of a Vibrating String

Perhaps, the following formulae will help you to understand the derivations given on page 133,
√ 1
1+ε≈1+ ε for small ε ,
2
s  2  2
∂y 1 ∂y
1+ −1≈1+ −1 ,
∂x 2 ∂x
 2
1 ∂y
T (x) = T (0) + s , T (0) = const ,
2 ∂x
s 
 2  2  4
∂y 1 ∂y s ∂y
T (x)  1+ − 1 ≈ T (0)
 + ,
∂x 2 ∂x 4 ∂x
 4  2
∂y ∂y
≪ .
∂x ∂x
When calculating the energy in each normal mode, you need the following integrals

Zℓ
ωn x
sin2 dx (2.38)
c
0

and
Zℓ
ωn x
cos2 dx . (2.39)
c
0
 21 PHY2606/1

They can be found using well-known formulae.

Zπ
π
sin2 nxdx = for a non-zero integer n (2.40)
2
0

and
Zπ
π
cos2 nxdx = for a non-zero integer n . (2.41)
2
0

Indeed, replacing the variable x in the integral (2.38) by a new variable z which is related to x as
ℓ
x= z,
π
we find that
ℓ
dx = dz
π
and the point x = ℓ corresponds to z = π. Hence,
Zℓ Zπ
2 ωn x ℓ ωn ℓz
sin dx = sin2 dz
c π πc
0 0

Zπ
ℓ
= sin2 nzdz
π
0

ℓ
= .
2
The same trick can be used to calculate the integral (2.39).

Doppler Effect

This very important physical effect is described by Pain in a rather formal way. Perhaps, it wouldn’t be
a bad idea to refresh your memory about the physics of the Doppler effect by reading the corresponding
section in any physics textbook which you used at the first level.

There are three essentially different situations which you may encounter when solving problems related
to the Doppler effect:

1. The source is moving while the detector is stationary.

2. The source is stationary while the detector is moving.
3. Both, source and detector, are moving.

In all these cases, if separation between the source and the detector decreases, the effective
frequency registered by the detector is increased, and vice versa. But the formula for calculating the
effective frequency is different for these three cases. Let u and v be the source and detector velocities
which are considered as positive if they are directed towards decreasing the separation and otherwise as
negative. Then it can be shown that
 22

c
moving source: ν′ = ν
c−u

c+v
moving detector: ν′ = ν
c

c+v
moving source and detector: ν′ = ν
c−u

The last of these three formulae is obtained by multipying two “amplification” factors c/(c − u) and
(c + v)/c, which work simultaneously when both the source and the detector are moving.

When using these formulae, REMEMBER the above convention about the signs of the velocities u
and v.
 23 PHY2606/1

2.6 Fourier Methods

The basic conclusion of this chapter is that any (reasonable) periodic function can be written as a series of
harmonic functions. Although this chapter is rather mathematical, the possibility of describing any wave
as a superposition of different harmonic waves is of the utmost importance to physics. Indeed, in practice
we never have pure harmonic oscillations and waves. For example, when you are listening to a symphonic
orchestra (or a rock group, if you prefer) the sound waves reaching your ears, are absolutely different
from a sine or cosine function. However, using the Fourier technique, you can decompose these waves
into the elementary components, i.e. into many harmonic waves of different frequencies and amplitudes.
Each of these harmonic waves move independently, and only their joint action on your ears gives you the
impression of hearing the music in its full beauty. Therefore, everything that we studied before about
(ideal) harmonic waves can be used for practical purposes despite the fact that real waves are not ideal.

In mathematics, very often we decompose something into its elementary components. For examinationple,
any vector ~a in a three-dimensional space can be written as a sum of three (elementary) unit vectors ~i,
~j, ~k directed along the Cartesian coordinate axes, with certain amplitudes ax , ay , and az , viz.

~a = ax~i + ay~j + az~k . (2.42)

The important thing in such a decomposition is that the elementary vectors are orthogonal to each other,
which implies that the scalar product of any pair of them is zero,

(~i · ~j) = 0 , (2.43)

(~i · ~k) = 0 , (2.44)

(~j · ~k) = 0 , (2.45)

but the scalar product of a unit vector with itself is one unit,

(~i · ~i) = 1 , (2.46)

(~j · ~j) = 1 , (2.47)

(~k · ~k) = 1 . (2.48)

The orthogonality of the elementary vectors enables us to easily find the expansion coefficients ax , ay ,
and az . Indeed, they are scalar products of ~a with ~i, ~j, and ~k respectively. For examinationple,

ax = (~a · ~i) . (2.49)

Mathematicians have moved along this way a little bit further. They found that the above idea of or-
thogonal vectors can be generalized in two aspects. Firstly, they considered abstract spaces with an
infinite number of orthogonal vectors. And secondly, they introduced the so-called functional vectors,
which actually are functions (such as sin or cos). You may ask me: “What would a scalar product for
such vectors be?” A very reasonable question, indeed!

A scalar product (ϕ1 , ϕ2 ) of two functions ϕ1 (x) and ϕ2 (x) is usually defined as an integral

Zb
(ϕ1 , ϕ2 ) = ϕ1 (x)ϕ2 (x)dx (2.50)
a
 24

of their product over a certain interval [a, b]. In fact, many infinite sets of functions are known, which
form orthonormal bases similar to the set of the three vectors ~i, ~j, and ~k, in the sense that the functions
constituting such a set are orthogonal to each other and normalized (a scalar product to itself is a unit),
i.e. 
Zb  0, if m 6= n ,
ϕm (x)ϕn (x)dx = (2.51)
 1, if m = n ,
a

Therefore, similarly (2.42), an arbitrary function f (x) can be expanded into its “elementary” components
as
f (x) = a1 ϕ1 (x) + a2 ϕ2 (x) + a3 ϕ3 (x) + . . .

∞
P (2.52)
= an ϕn (x) ,
n=1

where the expansion coefficients an are just the integrals

Zb
an = f (x)ϕn (x)dx . (2.53)
a

These coefficients are the “amplitudes” representing each elementary vector in the expansion. Since
in (2.52) we have an infinite number of terms, this sum has a meaning only if each subsequent term
contributes less than the previous one when n → ∞. Thus

an −→ 0 (2.54)
n→∞

and therefore, if we omit all the terms after a certain large n = N , the error might be small and the finite
sum approximately equal to f (x).

f (x) ≈ a1 ϕ1 (x) + a2 ϕ2 (x) + a3 ϕ3 (x) + . . . + aN ϕN (x) , (2.55)

This sum is used to approximate a complicated function f (x) by more simple functions ϕn (x). The more
terms included in this sum, the better such an approximation will be.

There are many orthonormal sets of functions {ϕn (x)}, such as the Bessel functions, Legendre polyno-
mials, Hermite polynomials, etc., which are used for this type of expansion and approximation. In the
theory of oscillations and waves, there is one such set which is of special interest and importance. This
special set comprises the sin and cos functions which are orthogonal on the interval [0, 2π],

Z2π  0, if m 6= n ,
sin(mx) sin(nx)dx = (2.56)
 π, if m = n ,
0

Z2π  0, if m 6= n ,
cos(mx) cos(nx)dx = (2.57)
 π, if m = n ,
0

Z2π
sin(mx) cos(nx)dx = 0, for all m and n . (2.58)
0
 25 PHY2606/1

They are normalized not to unity but to π. This is why, in the formulae defining the expansion coeffi-
cients, we have to include a factor 1/π.

This orthogonal basis is most suitable when it comes to expansion of periodic functions. Furthermore,
when applying such an expansion to the wave motion, each term has a physical meaning of an elementary
harmonic component which is really present and, in principle, can be extracted from the wave by certain
physical means.

We can always change the variable to make the sin and cos functions orthogonal on an interval different
from [0, 2π]. First of all, we note that the integration interval in (??), (?? and (2.56) is not necessarily
[0, 2π]. It can be any interval of length 2π, say from a to a + 2π, because the integrand repeats itself after
this length. Therefore, instead of (??,??,2.56), we can write, for example,

Zπ  0, if m 6= n ,
sin(mx) sin(nx)dx = (2.59)
 π, if m = n ,
−π

Zπ  0, if m 6= n ,
cos(mx) cos(nx)dx = (2.60)
 π, if m = n ,
−π

Zπ
sin(mx) cos(nx)dx = 0, for all m and n , (2.61)
−π

which is also correct.

Now, let a periodic function f (x) repeat itself after length ℓ and is given on the interval [−ℓ/2, +ℓ/2]. To
expand it, we need a basis orthogonal on this interval. How should we modify the trigonometric basis
{sin mx}, {cos nx} for this purpose? We can change the integration variable, which leaves the value of the
integral unchanged. The variable x should be replaced by another variable, say, z such that if z = −ℓ/2
the corresponding x = −π, and if z = +ℓ/2 then x = +π. This is achieved if we let

2π
x= z, (2.62)
ℓ

which gives us
2π
dx = dz (2.63)
ℓ
and

Zℓ/2


 0, if m 6= n ,

sin(2mπz/ℓ) sin(2nπz/ℓ)dz = ℓ/2, if m = n , (2.64)


−ℓ/2 


Zℓ/2


 0, if m 6= n ,

cos(2mπz/ℓ) cos(2nπz/ℓ)dz = ℓ/2, if m = n , (2.65)


−ℓ/2 

 26

Zℓ/2
sin(2mπz/ℓ) cos(2nπz/ℓ)dz = 0, for all m and n . (2.66)
−ℓ/2

To find the expansion coefficients (amplitudes) an and bn we should integrate the products of f (x) with
sin(2nπx/ℓ) and cos(2nπx/ℓ) over the same interval, [−ℓ/2, ℓ/2] and divide the integrals by ℓ/2 (because
of the normalization).

Analogously, we can derive formulae for the expansion of a function given on an interval [0, ℓ]. This is
used in the section ’Application to the Energy in the Normal Modes of a Vibrating String’, where the
mode energies are obtained for a string of length ℓ which at the initial moment is plucked from the middle
as is shown in Fig. 2.5.
y(x)

d HH
 H
 H
HH
 H x
0 ℓ/2 ℓ

Figure 2.5:

Note that the function given in this figure, is not periodic. It simply is not defined outside the interval
[0, ℓ] because the string does not exist there. But we can apply the Fourier expansion anyway! The
resulting formula (see page 283 of the textbook),
X 8d  nπ  ω x
n
y0 (x) = sin sin ,
n2 π 2 2 c
n odd

gives the correct triangle function on [0, ℓ] and periodically repeats itself outside this interval as is shown
in Fig. 2.6, but we simply ignore everything which is to the left of the point x = 0 and to the right
of x = ℓ. Therefore, we can expand not only an infinite wave but also any isolated pulse (signal) into
y(x)

H H
 H H   HH
 H
HH  H
H x

 

HH  0 ℓ HH  
H HH

Figure 2.6:

harmonic constituents. This is very important and is always used in the signal processing.

As a result of the Fourier analysis, a continuous function f (x) is represented by a discrete set of numbers,
{an } and {bn }, which are the amplitudes of different spectral components. This is extensively used in
 27 PHY2606/1

modern digital technique. Signals (for examinationple, music) can be electronically decomposed into
elementary harmonic components and their amplitudes thus obtained (the digits) can be recorded or
transmitted. The numbers are usually recorded in the so-called binary form, when all numbers are
represented by certain sequences of digits 0 and 1. To reproduce the initial signal afterwards, an electronic
device generates harmonic signals of different frequencies and mixes them using different amplitudes
according to the received ’digits’ {an } and {bn }. This digital way of recording or transmitting signals is
much more reliable than the traditional one (when you record the function f (x) itself), because the digits
are much more difficult to distort. For examinationple, if you recorded the digit 1 but it was somehow
distorted to 0.7, the electronic device would still be able to distinguish it from 0.
 28

2.7 Waves in Optical Systems

Despite its title, the main part of this chapter is about the geometrical optics which deals with linear
light rays experiencing abrupt changes of direction at various surfaces. Of course, we know that light is
an electromagnetic wave, but its wave nature does not exhibit itself in simple optical systems involving
mirrors and thin lenses. However, the basic laws (I would rather say phenomenological rules) of geometri-
cal optics, such as the laws of reflection and refraction, can only be explained and formally derived using
the notions of a wave and wavefront.

We prescribe only the first four sections of this chapter, because the laws derived here are used in the next
chapter where the wave nature of the light is fully exploited to demonstrate such pure wave phenomena
as interference and diffraction. Before proceeding to the next chapter, you should know that:

• The incident and reflected rays are in the same plane as the normal to the surface, and their angles
with this normal are equal to each other.
• When light travels from a material with refractive index n1 into a material with refractive index
n2 , the incident and refracted rays belong to the same plane as the normal to the surface. The
angle of refraction θ2 is related to the angle of incidence θ1 by

n1 sin θ1 = n2 sin θ2 .

• Refractive index n of a medium is a ratio of the speed of light c in the vacuum to the speed of light
cm in this medium, i. e.
c
n= .
cm
• The refractive index of the air is approximately 1.
• The optical path length in the medium is the product nx, where x is the ordinary physical length of
the path. In other words, the optical path length is the corresponding length in the vacuum, which
can be passed by the light during the same time (because in the vacuum the light travels n times
faster).
• When light travels through a material with a smaller refractive index towards a material with
a larger refractive index, reflection at the boundary occurs together with a phase change that is
equivalent to one half of a wavelength.
• When light travels from a larger towards a smaller refractive index, there is no phase change upon
reflection at the boundary.
 29 PHY2606/1

2.8 Interference and Diffraction

The most spectacular and unmistakably identifiable phenomena of interference and diffraction come from
the optics. This is why Pain presents these general wave phenomena as the wave properties of light.
However, you should bear in mind that exactly the same can happen with radio and sound waves or with
water waves on the surface of a pond. The interference and diffraction are universal properties of all
kinds of wave motion.

To ease the reading of this chapter, I would advise you first to refresh your memory of the optical phenom-
ena by reading (or just browsing through) the corresponding pages of any textbook on general physics
which you used at the first level. In principle, the basic concepts, which are given in the first four sections
of the previous chapter, are enough for this purpose. However, some less formal text would give you the
“feeling” of what happens with light in this or that situation.

Division of Amplitude
To follow the derivation of the optical path difference for the two rays shown in Fig. 12.1 of the textbook,
you need to accurately write trigonometric formulae for the triangles involved. Note that 6 CAD = π/2−i
and AC = AD cos(π/2 − i) = AD sin i.

As it follows from the subsequent formulae, Pain denotes the total phase difference between the two
beams by δ, and thus it should be written as
2π
δ= 2nt cos θ + π , (2.67)
λ
where the additional term π comes from the reflection because n > 1. Unfortunately, Pain uses the
same symbol t for time and for thickness of the plane parallel slab. Be careful not to mix
them up .

Note, there is a misprint on page 335: the formula for y1 should be written as follows

y1 = a sin(ωt + δ/2) .

In a wave of light, the electric and magnetic fields E and B oscillate harmoncally. The intensity of light
perceived by the human eye, is proportional to the energy density and therefore to the square of the
electric or magnetic field, i. e.

energy density = ǫ0 E 2
1 2
= B .
µ0
In the formulae of this section, y1 , y2 , and their superposition R = y1 +y2 mean either electric or magnetic
field (they both have the same dependence on t and x). This is why the light intensity is just R2 .

The intensity is zero when

δ π
= + mπ , (m is an integer) ,
2 2
and attains its maximum when
δ
= mπ , (m is an integer) .
2
 30

Using δ given by Eq. 2.67, we find

π π π
2nt cos θ + = + mπ ,
λ 2 2

2nt cos θ = mλ zero intensity ,

and

'$
π π
2nt cos θ + = mπ ,
λ 2

 
1

p&%
2nt cos θ = m− λ maximum intensity .
2
In the last formula, we can write (m + 1/2) as well because m is an arbitrary integer number (positive
or negative) and we can always shift it by 1.

Newton’s Rings
To derive the approximate formula t ≈ r2 /2R for the thickness of the air gap between a spherical and a
flat surface consider a circle shown in Fig. 2.7. The equation of this curve is
y

Figure 2.7:

x2 + (R − y)2 = R2 ,
from which we find
p
y = R− R 2 − x2

r  x 2
= R−R 1− .
R
Since we consider the area very close to the point where √the circle touches the x-axis (x/R ≪ 1), we can
replace the square root using the approximate formula 1 − ε ≈ 1 − ε/2. As a result we obtained the
formula we wanted to prove, namely,
x2
y≈ .
2R
 31 PHY2606/1

Michelson’s Spectral Interferometer

An interferometer is an apparatus that can be used to measure the wavelength of light by utilizing
interference between two light waves. In the Michelson’s interferometer, the light (from the same source) is
divided into two parts which, after passing through different paths, are again combined. If the optical path
lengths are equal or differ by an integer number of wavelengths, one observes a constructive interference
(bright field). If they differ by half–integer number of wavelengths, a viewer sees a dark field. Moving one
of the mirrors to or from the centre, a viewer will observe a sequence of bright and dark fields. When he
shifted the mirror to a distance d such that a bright field changes to a dark field, this distance is equal to
one quarter of the wavelength, because the light passes the distance d twice. In principle, by measuring
the shift d, one can find the wavelength as
λ = 4d .
This however, is not practical because a typical wavelength of a visible light is very small (∼ 10−4 cm)
and it is impossible to measure such small shifts of the mirror with sufficient accuracy.

Fortunately, there is a way around this difficulty. One can slowly move the mirror to a rather long
distance D and count the corresponding number N of changes between the bright to dark fields. Then
the wavelength is
D
λ=4 ,
N
and hence a very high accuracy can be achieved.

Fabry–Perot Interferometer
A light ray passing through the system at an angle of incidence θ, may pass directly through or it may be
reflected one or more times before passing through. The total intensity of the passing light periodically
depends on both the angle θ and the wavelength,
Imax
It =   2 ,
4R 2πd cos θ
1+ sin
(1 − R)2 λ

which enables us to distinguish two different (but close to each other) wavelengths when they are both
produced by the same source.

Division of Wavefront
The formulae derived here (as well as the way they are derived), are very similar to those given in the
section Division of Amplitude. The idea is the same: we just calculate the path difference and then
the corresponding phase difference. At the points for which this difference is equivalent to an integer
number of wavelength, we have fully constructive interference, and at the points for which it is equal to
a half–integer number of wavelength, the interference is fully destructive. This simple idea underpins all
the theories (however complicated they look) concerning interference and diffraction.

Interference from Two Equal Sources of Separation f

In this section Pain uses the formula for light intensity,
δ
I = 4a2 sin2 (ωt − kX) cos2 , (2.68)
2
 32

which was derived in the previous section. You may be confused: why does he drop the time dependent
factor sin2 (ωt − kX)? This is because he considers the space distribution of the bright and dark areas.
Actually, the light always oscillates at every point (because it is a wave), but these oscillations are so rapid
that we never see them. If the phase differences do not change with time, we can see only a stationary
picture which has nothing to do with the time dependent factor. In other words, the “stationary” formula
δ
I = 4a2 cos2 (2.69)
2
describes the space distribution of the intensity amplitude.

Spatial Coherence
The important formula here is
AB/d ≪ λ/f ,
which can be rewritten as
λd
AB ≪ .
f
This means that to observe the fringes, we need either very small source (AB) or very small separation
f between the slits, or else to put the source very far from the slits (large d). Of course, the best way is
to have small source situated very far from the slits which in turn are very close to each other.

Interference from Linear Array of N Equal Sources

Before starting this section, you should go back to Chapter 1 and read once more about the superposition
of a large number of simple harmonic vibrations of equal amplitudes and equal successive phase differences.
The mathematics used here is quite simple. The only difficulty may be caused by the approximation
sin2 N β N 2β2
2 −→ , β = 0, ±π, ±2π, . . . , (2.70)
sin β β2
which follows from the formula
sin ε ≈ ε (ε → 0) . (2.71)
Indeed, for the value β = 0 Eq. (2.70) follows immediately from Eq. (2.71). Near the point β = π, we
write β = π + ε and using the reduction formula
sin(π + ε) = − sin ε ,
arrive at the same result. The other values of β can be treated analogously.

Diffraction
Note the difference between Fraunhofer and Fresnel diffractions. As a matter of fact, this is the same
phenomenon but viewed at different distances from the diffracting system.
When discussing Fraunhofer’s diffraction from a single narrow slit, Pain writes the formula (page 369)
sin2 α sin2 α
N 2 Is = I 0 ,
α2 α2
i.e., he replaces the product N 2 Is by the total intensity I0 . This is so because Is = a2 (where a is the
amplitude of an individual wave) and hence
N 2 Is = (N a)2

= A2 (square of the total amplitude)

= I0 .
 33 PHY2606/1

If you understood the diffraction from one slit, the other three sections describing diffraction from N slits,
two slits, and grating are easy reading.

Fresnel Diffraction
I must admit that the Fresnel theory is rather tough. In the textbook, however, Pain avoids complicated
formulae. Instead, he tries to give you an intuitive insight into this phenomenon. First of all you should
understand the difference between Fraunhofer’s and Fresnel’s description of the diffraction. Fraunhofer
assumes that we observe the light pattern on a screen which is far enough from the diffracting object
(slit, grating, etc.) that all the waves (rays) arriving at the screen may be considered as parallel to each
other. This is an approximation which makes Fraunhofer’s theory much easier than Fresnel’s theory.

To find the character of the fringes, we need to know how the resulting light intensity on the screen
depends on the position of the observation point P . The total amplitude of the light wave at a point P
is the sum of contributions from all the elements of the wavefront at the slit. These contributions arrive
at the point P with different phases. In other words, they can be represented as vectors in the complex
plane, having different angles. Therefore, to find the total amplitude, we should make a vector sum of all
these vectors. As you know, to make a sum of several vectors, we draw them one after another (putting
the beginning of the next vector at the end of the previous one) and connect the beginning of the first
vector to the end of the last one. This connecting vector is their sum.

As a result of the Fraunhofer’s approximation, the phase differences between the waves coming from any
two neighbouring elements of the wave front are the same (as is shown in Fig. 12.20 of the textbook).
Therefore, when calculating the total effect of all the elements, we perform the vector summation of all
the individual contributions as is shown in Fig. 1.11 where each vector (contribution from an individual
element of the wavefront) is turned by the same angle δ relative to the previous vector and the resulting
curve is circular.

If, however, the screen is close to the diffracting object (Fresnel’s regime), the diffracting rays cannot be
considered as parallel and Fig. 12.20 is not correct anymore. Each element of the wavefront sends its
waves to the point of observation at its own angle θ (see Fig. 2.8). Therefore, the phase difference

-
- geometric shadow
p p pppppppp pp pppppp p pp pp pp pp pppppp p ppp
-
pp pppppp p p pppppp p ppppp
-
p
p pppp ppppppppp p pp pppppppppppppp pp
- 6
pp pppppppppppppppp
- P
pppppppp
-
-
-
-
light - slit - I
pppppppp
-
pp ppppppppppppppp p p
-

p pppp ppppppppp p pp ppppppppp pppp ppp

-
-
p p pppppppp pp pppppp p pp pp pp pp pppppp p ppp
p p pppppp p p pppppp p ppppp
- screen
-
- ?
-
- geometric shadow
-

Figure 2.8:
 34

2π
δ= f sin θ
λ
between a pair of neighbouring elements is not the same for all the elements. The variation of δ trans-
forms the circular curve of Fig. 1.11 into Cornu’s spiral shown in Fig. 12.34. Since the angle θ can
take positive as well as negative values, this spiral has two symmetric branches. We always assume
that the wavefront element situated just opposite (at the same level as) the observation point P , has
a zero phase shift, and calculate the phase differences for the other elements relative to it. Therefore
the centre of Cornu’s curve (where the two branches of the spiral are connected) corresponds to this
element. The elements situated above it, are represented by the corresponding number of the elements
of the Cornu’s curve in the first quadrant, and those situated below it, by the corresponding number in
the third quadrant. Thus, the number of the elements of Cornu’s curve which are involved in a particular
situation, is exactly equal to the number of elements forming the wavefront at the slit, and the centre
of this curve relates to the point P . If the point P is opposite the middle of the slit, the part of the
Cornu’s curve which is involved, is symmetric. If P is above the middle line (as is shown in Fig. 2.8), the
number of elements of this curve is the same, but its lower branch contributes more than its upper branch.

The resulting light amplitude at point P is the distance between the end points of the involved part of
Cornu’s curve. It is clear that this distance is different for different positions of P . For examinationple,
when we go along the spiral, the distance to any fixed point on the other branch periodically lengthens
and shortens until we reach the eye of the spiral. This is why the diffraction fringes appear as is shown
at the right hand side of Fig. 2.8.

If we make the slit wider (by moving the edges away), the fringes move together with the edges, and the
central area remains uniformely illuminated. In the limiting case, when one of the edges is removed (to
infinity), only one fringe pattern remains.
 35

Chapter 3

Examination Preparation

3.1 What is the examination all about?

The only means to evaluate the knowledge you acquired by studying the textbook is a written exami-
nationination, where you will be given several questions of various levels of difficulty. By means of the
examinationination, we determine two things: Have you understood the physics and mathematics that
we covered in the course, and can you apply these methods to real problems. Consequently each of the
examination questions set, will fall into one of the following categories:

1. Theory
A theory question will ask you to reproduce a result that was obtained in one of the prescribed
sections of the main body of the textbook. Of course, you need not reproduce it word for word,
but your answer must be complete in order to be awarded a full mark. You should write, not only
mathematical derivations, but accompany them with explanations showing the logical connections
between all successive steps of the derivation.

2. Assignments
If you recall how long it took you to solve some of the prescribed assignment problems, you
will appreciate that it would be unfair to set unknown problems of the same standard in the
examination. However, we do need to test your problem–solving ability. Consequently, we will
include problems which are of the same type as those in the assignments. However, the problems are
slightly altered to avoid getting an exact reproduction of the assignment as set out in the solutions
you received during the year. the problem is slightly altered so that a simple reproduction of the
solutions that are sent out during the year will not be exactly correct.

3. Unknown questions.
If you want to obtain a really high score for this course, you will have to do some thinking in
the examination hall (thinking always pays off!). These questions will not necessarily be based
directly on any of the prescribed study material. However, they will be quite tractable to a person
who has a good working knowledge of the contents of the course. Usually, the answers to these
questions can be found using physical insight rather than complicated mathematics.

35
 36

Typically, twenty five percent of the mark will be awarded for unknown questions. The other two types
of the questions are of equal importance, each being worth thirty to forty per cent of the examination
mark.

Many of the formulae required for solving the theory and unknown problems will be provided. However,
you will be expected to memorise the important ones, and most of what was needed for the assignment
problems. Of course, the more you remember, the better.

Apart from the sample paper contained in this study guide, you will have received the previous year’s
paper (Tutorial Letter 401). It will be useful to consult these papers, but remember that the question
paper will be unique. New papers are set each year according to the above guidelines.
 37 PHY2606/1

3.2 Trial examination Paper

This is what an examination paper may look like:

Duration: 2 hours 125 marks

The use of a pocket calculator is permissible.

It is not necessary to answer all the questions because the total mark would be more than 100% /

Information which you may find useful is provided at the end of the paper.

1. For a vibrator with mass m = 0.010 kg and stiffness s = 36 N/m,

(a) what value of the resistance r would make the amplitude decrease from A0 to A0 /e in 1.0 s?
(3)
(b) What is the Q–value of such a system? (3)
(c) What value of r would produce critical damping? (4)
[13]
2. A vibrator with mass m, stiffness s, and resistance r is set into steady oscillations by a harmonic
driving force. Show that the displacement amplitude is a maximum at the driving frequency
r
s r2
ω= − .
m 2m2
[18]
3. Two masses m1 and m2 are joined by a spring of stiffnes s. They can vibrate along the line of
their centres, moving alternately towards and away from each other. For this system, show

(i) that the frequency of oscillations ω is given by

p
s/µ
 38

where
m1 m2
µ =
m1 + m2
is the reduced mass. (12)
(ii) Discuss the case
m 1 ≪ m2
(6)

[18]
4. (a) Derive the wave equation for a perfectly flexible taut string. (10)
(b) Derive the travelling harmonic wave equation. (9)
[19]
5. The relation between the impedance Z and the refractive index n of a dielectric is given by Z = 1/n.
Light travelling in free space, enters a glass lens which has a refractive index of 1.5 for a free space
wavelength of 5.5 × 10−7 m. Show that reflections at this wavelength are avoided by coating which
has a refractive index 1.22 and thickness 1.12 × 10−7 m. [13]
6. Light from a star of wavelength 6 × 10−7 m is found to be shifted (its wavelength is lengthened)
10−11 m towards the red when compared with the wavelength from an identical laboratory source.
If the velocity of light is 3×108 m/s, find the velocity at which the earth and the star are separating.
[13]
7. Find the first three terms of the Fourier series for a signal given by



 0 , x<0





 x2 , 0 ≤ x ≤ 2π
y(x) =




 0 , 2π < x





[13]
8. A mixture of light of wavelengths λ and λ + ∆λ (where ∆λ ≪ λ) falls on an N –slit diffraction
grating with spacing f . Show that the angular separation of the two principal maxima of order n
is
n∆λ
∆θ = .
f cos θ
For a grating with f = 2.5 µm, evaluate ∆θ for the sodium D lines (589.0 nm and 589.6 nm) with
n = 1.
[18]

USEFUL INFORMATION:
Z
x2 sin xdx = −x2 cos x + 2x sin x + 2 cos x

Z
x2 cos xdx = x2 sin x + 2x cos x + 2 sin x
 39 PHY2606/1

3.3 Discussion of the Trial examination Paper

To illustrate what was said before, let us examinationine the trial paper. As it could be guessed from the
outset, the questions cover all the prescribed chapters. In the context of the symbolic division into the
three categories (theory, assignment, unknown), they can be summarized as follows:

question number type mark

1 assignment 13
2 theory 18
3 unknown 18
4 theory 19
5 assignment 13
6 assignment 13
7 assignment 13
8 unknown 18

The relative importance of the three question types is seen by the marks allocated to each. It is not
possible to examine every part of the course, but we will always try to distribute the questions uniformly
across the prescribed material. The fact that this paper gives the theory questions on the forced vibrations
and diffraction, is just a nuance of this particular examination. It is not a characteristic of a typical paper.

3.4 Preparation for the examination

Now, when you have some insight into how the paper will be structured, you can prepare accordingly.

The best way to prepare for the theory questions is to go through the book one page at a time. Every
time a result is derived, identify the start of the argument. Then note how this leads on to the final result.
Make sure that you understand any assumptions made, physical constraints imposed, and mathematics
used.

Construct a schematic logical sequence of the derivation. For examinationple, the result given to the
second question of the above Trial examination Paper, can be obtained via the following steps:

START from the equation of

motion:
mẍ + rẋ + sx = F0 eiωt
 40

⇓
Try the complex solution:
x = Aeiωt

⇓
find the complex amplitude:
F0
A=
iωr + s − ω 2 m

⇓
Find the real amplitude:
F0
A = |A| = r s 2
ω r2 + − ωm
ω

⇓
Obtain the derivative:
" r #
d s 2
ω r2 + − ωm
dω ω

⇓
Solve the equation for ω:
" r #
d s 2
ω r + 2 − ωm =0
dω ω

You should memorize the sequence of these steps. It is enough to remember the formulae given only in
the top two boxes. The other formulae can be easily derived if you remember which step is next.

Now we come to the important part: Close the book and try to reproduce the above
reasoning to the paper. If you are successful, go on to the following pages. If you get stuck,
consult the textbook. Then repeat the work on paper until you can do it without referring to the book.
Remember that in the examination hall you won’t have the option of referring to the textbook!
 41 PHY2606/1

The assignment questions should be the easiest to learn. You would have thought through each one
before submitting your assignments. Also by the time it comes to preparing for the examination, you
would have received solutions from us. Again you need to practice solving each of the problems without
notes to assist you. It is also very useful to construct a schematic logical sequence for each solution.
Remind yourself of what the problem is actually about, so that you will be prepared to deal with any
modifications which may be made.

Your ability to answer the unknown questions should be cultivated throughout the year. However, the
following exercise may help to sharpen your skills in this direction during the days before the examination.
Read through some of the questions in the textbook which were not prescribed. Try to figure out how
you would approach the solution. To actually do all the calculations may require more time than you
have available, but you can try to draw a schematic logical sequence again. As soon as you understand
which steps are needed to find the solution, you are already close to a victory.

There are also many practical things to be considered when preparing for an examination. Make sure of
the time and venue, get enough sleep, take extra pens, ensure your calculator is working, etc. Above all:
Be Calm and confident. Don’t underestimate the importance of these factors.

3.5 The examination

There are many practical tips for writing examinations which, if taken, can add several per cent onto
your result. You have heard them many times, but it is a good idea to remind yourself of them just
before writing the examination. I list a few here:

• Write something! If you offer no answer to a question, the maximum that you can score is zero.
As soon as you write something, the minimum score that you can be awarded is zero (feel the
difference!). There is no negative marking.
• If a question is taking a disproportionate amount of time, move on. You want to be earning the
maximum marks per minute possible, particularly during the early stages of the paper.
• Don’t cross any of your work out until you have replaced it. If you offer two solutions to the same
question, it is likely that only the first will be considered for marks. However, if you write one and
a quarter solutions, and the one is crossed out, only the quarter solution will be eligible for marks.
• Write your answers neatly! Try to make your handwriting as clear as possible. Put some logical
explanations of how you come from one formula to another (see how I did it in the next section).
Believe me, these are not trifles. These points psychologically influence a person who is marking
your paper. You can gain or loose 10% of the mark on these trifles. Sometimes it can be crucial.
• Befor attacking the questions, look to see what is given at the end of the paper (if anything).
Often examinationinees struggle with a question when much of what they require for the solution
is given.
• And finally the ubiquitous “Read the question”! This particularly concerns the assignment ques-
tions which may differ in a small way from what you are expecting.
 42

3.6 Model Answers to Trial examination Paper

1.
2. (a) We assume that the motion is oscillatory (because the question concernes the amplitude) and
therefore the damping is below the critical value. The time dependence of the displacement for a
lightly damped system is given by

x(t) = A(t) sin(ωt + φ) ,

where the amplitude A(t) is exponentially diminishing, namely,

 
rt
A(t) = A0 exp − .
2m

At the initial moment (t = 0) we have

A(0) = A0 .
At the moment t1 = 1.0 s the amplitude is

A(t1 ) = A0 /e .

This can be achieved if

rt1
=1,
2m
hence
2m
r =
t1

2 · 0.010 kg
=
1.0 s

= 0.02 kg/s

= 0.02 N/s/m .

(b) The quality factor is given by

ω′ m
Q= ,
r
where the friction modified frequency ω ′ is
r
′ s r2
ω = − ,
m 4m2
therefore
r
m s r2
Q = −
r m 4m2
r
ms 1
= −
r2 4
 43 PHY2606/1

s
0.010 kg 36 N/m 1
= −
0.0004 (N s/m)2 4

≈ 30 .

(c) Critical damping takes place if

r2 s
2
= ,
4m m
that is when
√
r = 2 ms
p
= 2 0.01 kg · 36 N/m

= 1.2 N s/m .

3. The equation of motion, describing forced oscillations reads

mẍ + rẋ + sx = F0 eiωt ,

where F0 is the amplitude of the driving force. One way to find a solution of a differential equation
is to guess it and to check if the guess is correct by substituting it into the equation. Since we
know that the motion is oscillatory, we can try the solution in the form

x = A exp (iωt) .

With such x the equation of motion becomes

−Aω 2 m + iωAr + As eiωt = F0 eiωt ,

from which we find that our guess is a true solution if and only if the amplitude A is given by
F0
A= .
iωr + s − ω 2 m
This is complex amplitude which includes the phase–shift factor. To obtain the real amplitude
of the√displacement, we need to find its absolute value. As we know, for any complex number
|z| = zz ∗ , hence

A = kAk
s
F02
= 2
(iωr + s − ω m) (−iωr + s − ω 2 m)

F0
= r s 2
ω r2 + − ωm
ω
For a steady state, the displacement amplitude does not depend on time. In this respect it is a
constant. But it is different for different choices of the driving force frequency ω, and, therefore, is
a function of ω, which we can denote as A(ω). We have to find the frequency at which this function
 44

attains its maximum value. This can be done by locating the point at which its first derivative,
dA/dω, is zero. Performing the differentiation, we find


d r2 − 2sm + 2ω 2 m2 F0
A(ω) = 3/2
.
dω ω 2 [r2 + (s/ω − ωm)2 ]

This derivative can be zero if

r2 − 2sm + 2ω 2 m2 = 0 ,
from which we immediately find r
s r2
ω= − .
m 2m2

4. Let x1 and x2 be the displacements of the masses m1 and m2 from their equilibrium positions
which are at the left and right hand sides of the spring as is shown in the Figure. For both masses,

equilibrium equilibrium
'$

m1 m2

WWWWWWWWWWWWWWWWWWWWWWW

&%
x1 x2

we consider the displacement as positive when a mass is to the right of its equilibrium point, and
as negative otherwise. Therefore, the total increase x of the spring length is

x = x2 − x1 ,

which is positive when the spring lengthens and negative otherwise. The restoring force Fs which
drives the masses towards the equilibrium positions is proportional to the total length change, i.
e.
|Fs | = |sx| .
According to Newton’s law, this force acts on both masses but in opposite directions. For the mass
at the right hand side, a positive x gives a negative restoring force. For the mass at the left hand
side the sign of the force must be opposite. Therefore, the equations of motion for the masses have
the following form

m1 ẍ1 = +s(x2 − x1 ) ,

m2 ẍ2 = −s(x2 − x1 ) .

Let us multiply the first of these equations by m2 and the second by m1 , and subtract the former
from the latter. The resulting equation,

m1 m2 (ẍ2 − ẍ1 ) = −s(m1 + m2 )(x2 − x1 ) ,

can be rewritten as
m1 + m2
ẍ + s x=0.
m1 m2
 45 PHY2606/1

Now, comparing it with the standard form of the equation describing harmonic oscillations, we see
that the frequency is r
m1 + m2
ω= s ,
m1 m2
which is exactly what we had to prove. Instead of the mass entering the standard formula for the
frequency, we obtained the so-called reduced mass
m1 m2
µ= .
m1 + m2

If one of the bodies is much heavier than the other (say m1 ≪ m2 ), then it practically remains at
rest while the light one makes all the movements. In such a case the frequency must be determined
by the mass of the light body. Indeed, rearranging the above equation for µ as
m1
µ=
m1 /m2 + 1

and assuming that m1 /m2 ≈ 0, we obtain that

µ ≈ m1 if m1 ≪ m2 .

5. If the string has mass per unit length ρ, then the mass of a segment ds long is simply ρds. It
follows from the Pythagorean theorem that
s  2
∂y
ds = 1 + dx ,
∂x

which, due to the fact that ∂y/∂x ≪ 1, can be approximated as

"  2 #
1 ∂y
ds ≈ 1+ dx
2 ∂x

≈ dx .

Therefore the mass of this segment is approximately equal to ρdx. The perpendicular force F

displacement
y 6 T~
θ+ dθ
1

 
ds
∂y
θ
∂x

- x
x x + dx
T~
 46

pushing the string element up and down, is the difference between vertical components of the
tension forces at its right and left ends, i. e.

F = T sin(θ + dθ) − T sin(θ) ,

which can be approximated as follows

"    #
∂y ∂y
F ≈ −
∂x x+dx ∂x x

because sin θ ≈ tan θ = ∂y/∂x. Since dx is an infinitely small interval

   
∂y ∂y ∂2y
− = dx .
∂x x+dx ∂x x ∂x2

Therefore Newton’s second law for vertical motion of the string element reads
∂2y ∂2y
ρdx 2
= T 2 dx ,
∂t ∂x
which gives us the wave equation in the form
∂2y 1 ∂2y
− =0,
∂x2 c2 ∂t2
p
where c = T /ρ. Any (differentiable) function of the type f (ct − x) is a solution of this equation.
Indeed
∂y
= −f ′ (ct − x) ,
∂x

∂2y
= f ′′ (ct − x) ,
∂x2

∂y
= cf ′ (ct − x) ,
∂t

∂2y
= c2 f ′′ (ct − x) ,
∂t2

1 2 ′′
f ′′ − c f = 0.
c2
The equation for a travelling harmonic wave can therefore be written as
2π(ct − x)
y(x, t) = a sin
λ
or
y(x, t) = a sin(ωt − kx) ,
where ω = 2πc/λ and k = 2π/λ.

6. The light wave passes from the medium having the impedance Z1 = 1 (free space) to the medium
having the impedance Z2 (glass) via an intermediate medium (coating) with the impedance Zi .
We know that if p
Zi = Z1 Z2 ,
 47 PHY2606/1

and the thickness of the coupling medium is λi /4 (where λi is the wavelength in this medium)
then all the energy carried by the wave, is transmitted through the connection without losses.
Since Z1 = 1, Z2 = 1/n2 , and Zi = 1/ni we have
√
1/ni = 1/ n2 ,
√
ni = n2
√
= 1.5

≈ 1.22 ,
which coincides with the value given in the problem. The necessary thickness of the coupling
medium is
d = λ2 /4

= λ1 /(4ni )
√
= λ1 /(4 n2 )

5.5 × 10−7 m
= √
4 1.5

≈ 1.12 × 10−7 m ,
which also coincides with the value given in the problem.

7. The general formula expressing the observed wave frequency ν in terms of the source frequency νs
and the detector and the source velocities (vd and vs ) reads
c ± vd
ν = νs ,
c ± vs
where the choice of the signs is determined by the direction of the motion. In the denominator, the
sign ’plus’ should be used if the source moves away from the detector, and ’minus’ otherwise. In
the numerator, the signs ’plus’ or ’minus’ are used when the detector moves to or from the source
respectively. Since in this problem the star and the earth are separating, we have
c − vd
ν = νs .
c + vs
The emitted and the observed wavelengths are
λs = c/νs ,

λ = c/ν .
Hence
ν λs
=
νs λ

λ + ∆λ
=
λ

∆λ
= 1+ ,
λ
 48

where ∆λ = 10−11 m. In the frame of reference connected to the earth, vd = 0 and vs = v where
v is the velocity of the separation. Therefore
∆λ c
1+ = ,
λ c+v

cλ
v = c− ,
λ + ∆λ
 
1
v = c 1−
1 + ∆λ/λ
 
m 8 1
= 3 × 10 1−
s 1 + 10−11 /(6 × 10−7 )

m
≈ 5000 .
s

8. The coefficients of the first three terms of the expansion

a0
y(x) = + a1 cos x + b1 sin x + . . .
2
are given by
Z 2π
1
a0 = y(x)dx ,
π 0

Z 2π
1
a1 = y(x) cos xdx ,
π 0

Z 2π
1
b1 = y(x) sin xdx .
π 0

Substituting y(x), we find

Z 2π
1
a0 = x2 dx
π 0

2π
x3 
=
3π 0

8π 2
= ,
3
Z 2π
1
a1 = x2 cos xdx
π 0

1 2
2π
= x sin x + 2x cos x + 2 sin x 0
π

= 4,
 49 PHY2606/1

Z 2π
1
b1 = x2 sin xdx
π 0

1
2π
= −x2 cos x + 2x sin x + 2 cos x 0
π

= −4π .

Therefore
y(x) = 4π 2 + 4 cos x − 4π sin x + . . . .

9. The angular position of the principle maxima are fixed by

nλ
sin θ = .
f
If the wavelength is λ + ∆λ, then
n(λ + ∆λ)
sin(θ + ∆θ) =
f
n(λ + ∆λ)
=⇒ sin θ cos ∆θ + cos θ sin ∆θ = .
f
If ∆λ is small, ∆θ will be small, so

sin ∆θ ≈ ∆θ ,

cos ∆θ ≈ 1.

Therefore
n(λ + ∆λ)
sin θ + cos θ∆θ = .
f
Substituting sin θ given by the first equation, we find
nλ n(λ + ∆λ)
+ cos θ∆θ =
f f
n∆λ
∆θ = .
f cos θ
Sodium D lines:
p
cos θ = 1 − sin2 θ
p
= 1 − λ2 /f 2 ,
∆λ
∆θ = p
f 2 − λ2
0.6 × 10−9 m
= √
6.25 × 10−12 m2 − 5892 × 10−18 m2

≈ 0.00025 rad .