Chapter 3

Oscillations in systems with a few

particles: normal modes
 Chapter 3

Introduction Our Chapter 2 discussion of the oscillations of a single mass

attached to a spring was a review of material you learned in
previous physics courses. In this chapter, we would like to
start with a specific PHY 241 topic: the oscillations of
several - perhaps many - masses connected by springs.
Ultimately, this will lead to the concept of waves.

Several complications arise when we consider the motion of

many interacting masses. To describe this motion, we need a
position function for each mass. Hence there will be an
equation of motion associated with each of these functions.
However, these equations will not be independent, because
the motion of one of the masses affects the motion of the
others. Hence we cannot solve the equations one at a time:
we must solve all the equations simultaneously! This appears
to be an impossible task, particularly for cases like a chain of
hundreds of beads connected by springs. Fortunately, the
solution of this problem is not as complicated as it sounds
at this time. Very soon, youÕll solve the problem of N beads
connected by springs, with N as large as you want.

The two-mass Equations of motion and their solutions

Let us start with a simple problem, illustrated in Fig. 1. Two
problem
000
111
identical masses are connected by identical springs.
111
000 000
111
111
000
111
000 x1 x2
111
000
111
000 111
000
111
000 111
000
111
000
111
000 K
m
K
m
K 111
000
111
000
Figure 1 Two masses connected by springs.

Let us assume that the displacements are restricted to the

horizontal direction, so that we need only two functions to
characterize the motion completely: the position x 1 (t) for
the mass on the left and the position x 2 (t) for the mass on
the right. We also assume that the springs are not stretched
when the masses are at rest. It will become clear as we

phy 241 MenŽndez 46

 Chapter 3

proceed that it is not convenient to refer the two position

functions to the same horizontal axis. Instead, as indicated
in the figure, we measure x1(t) from the equilibrium position
of the first mass and x2(t) from the equilibrium position of
the second mass. Before we even write down the equations
of motion for the system, we know that the solutions must
have four independent constants. This is because we can
specify the initial position and velocity of each particle
arbitrarily .

The first step to solve our problem is to write down the

equation of motion for each particle. This is, of course,
NewtonÕs second law. Consider first particle 1 (the particle
on the left). There are two forces acting on this particle: a
force exerted by the spring on the left, which is identical to
the force exerted by a spring on an isolated mass, and a force
exerted by the middle spring. The elongation of this spring
depends on the difference x1-x2. This is easy to see: if the two
masses are displaced by the same amount in any direction,
the middle spring is not stretched and exerts no force.
Hence the force exerted by this spring must be given either
by K(x 1 -x 2 ) or by K(x 2 -x 1 ). The trickiest point is to deter-
mine which of these two expressions is the correct one. This
is best done by inspection of the figure and considering any
specific displacements. For example, if mass 2 moves to the
right and mass 1 doesnÕt move, the force exerted by the
middle spring is to the right, which is positive for our axis
convention. But K(x1-x2) would be negative in such case, so
that the right expression for the force of the middle spring
on mass 1 is K(x 2 -x 1 ). The equation of motion of mass 1
becomes

m ẍ 1 = − K x 1 + K (x 2 − x 1 ) (1)

Note that this equation involves the motion of mass 2, as we

discussed in the introduction. So we cannot solve for the
motion of mass 1 alone: we must solve simultaneously the
equation for mass 2. This equation can be deduced along
the same lines as Eq. (1). It is given by

phy 241 MenŽndez 47

 Chapter 3

m ẍ 2 = − K x 2 − K (x 2 − x 1 ) (2)

The first term in the right-hand side is the force exerted by

the spring to the right of mass 2. The second term is the
force exerted on mass 2 by the middle spring. Note that this
force is equal to minus the force exerted by this spring on
mass 1. This is consistent with NewtonÕs third law.

The procedure to solve a set of simultaneous differential

equations is similar to the approach used to solve a set of
simultaneous algebraic equations: you multiply the equations
by carefully chosen constants and add them up, so as to get
an equation for a single function. In the case under consider-
ation, the simplicity of our problem dictates a very simple
linear combination. Suppose we add Eqs. (1) and (2). We
obtain

m (ẍ 1 + ẍ 2 ) = − K (x 1 + x 2 ) (3)

If we now that we define a function XA(t) = x1(t) + x2(t), the

second derivative of this function is d2 X A /dt 2 = x¨ 1 + x¨ 2 , so
that the function X A (t) satisfies the same equation as the
simple harmonic oscillator: d 2 X A /dt2 = -(K/m)XA . But we
already know the solution to the simple harmonic oscillator
equation, so that we can immediately write

X A (t) = x 1 (t) + x 2 (t) = 2A A cos (ω At + α A) (4)

with ωA = (K/m)1/2 . We have inserted here a factor of 2 for

later convenience, but this obviously does not affect the
generality of our solution. If we now subtract Eqs. (1) and
(2), we obtain a differential equation for the difference
between x1(t) and x2(t):

m (ẍ 1 − ẍ 2 ) = − 3K (x 1 − x 2 ) (5)

This equation is also a harmonic oscillator equation for the

function XB(t) = x1(t) - x2(t). Thus its solution is

X B (t) = x 1 (t) − x 2 (t) = 2A B cos (ω Bt + α B ) (6)

with ω B = (3K/m) 1/2.. Again, we have inserted here a factor

phy 241 MenŽndez 48

 Chapter 3

of 2 for later convenience. It is now easy to combine Eqs.

(4) and (6) to obtain the final expressions for x1(t) and x2(t):

x 1 (t) = A A cos (ω At + α A) + A B cos (ω Bt + α B ) (7)

and

x 2 (t) = A A cos (ω At + α A) − A B cos (ω Bt + α B ) (8)

Notice that there are four free constants in these expressions:

AA, AB, αΑ, and αB. These constants are determined from the
four initial conditions (initial position and velocities of the
two masses) corresponding to this problem. Since we can
satisfy any set of four initial conditions with four constants,
our solution is complete.

Normal modes
The motion of the two masses in the above problem,
represented by Eqs. (7) and (8), is not simple harmonic. If
you plot the functions x1(t) or x2(t), you obtain a complicat-
ed curve. This is shown in Fig. 2.
0.3
Displacement (m)

0.2
0.1
0.0
−0.1
−0.2
−0.3
0.0 0.5 1.0 1.5 2.0

0.3
D isplacement (m)

0.2
0.1
0.0
−0.1
−0.2
−0.3
0.0 0.5 1.0 1.5 2.0

Time(sec)
Figure 2 Displacements of mass 1 (top figure) and
mass 2 (bottom figure) for arbitrary choices
of the initial conditions.

phy 241 MenŽndez 49

 Chapter 3

But with an appropriate choice of the initial conditions, it is

possible to make the system execute simple harmonic
oscillator vibrations. For example, if AB = 0, the two masses
vibrate like a simple oscillator of frequency ωA. This is called
a normal mode. Normal mode A is shown here in Fig. 3
0.2

Displacement (m)
0.1

0.0

−0.1

−0.2
0.0 0.5 1.0 1.5 2.0

1.5
Displacement (m)

1.0
0.5
0.0
−0.5
−1.0
−1.5
0.0 0.5 1.0 1.5 2.0

Time(sec)

Figure 3 Displacements of mass 1 (top panel) and

mass 2 (bottom panel) for choices of the
initial conditions that produce normal
mode A.

On the other hand, if A A = 0, we excite a second normal

mode, for which the two masses oscillate with frequency ω B.
This mode is shown in Fig. 4

phy 241 MenŽndez 50

 Chapter 3

1.5

Displacement (m)
1.0
0.5
0.0
−0.5
−1.0
−1.5
0.0 0.5 1.0 1.5 2.0

1.5

Displacement (m)
1.0
0.5
0.0
−0.5
−1.0
−1.5
0.0 0.5 1.0 1.5 2.0

Time(sec)

Figure 4 Displacements of mass 1 (top panel) and

mass 2 (bottom figure) for a choice of the
initial conditions that leads to normal
mode B.

The initial conditions needed to excite these normal modes

are particularly simple. To excite normal mode A, our initial
conditions must be such that AB = 0. This means that (see
Eq. 6) x 1 (t) - x 2 (t) = 0 at all times. Remember that the
difference x 1 (t) - x 2 (t) satisfies a harmonic oscillator equa-
tion. The only way to keep a harmonic oscillator at rest is to
place it at the equilibrium position and set its initial
velocity equal to zero. Thus in our case we need x1(0)-x2(0)
=0 and v1(0)-v2(0) = 0. In other words, to excite mode A we

phy 241 MenŽndez 51

 Chapter 3

need to select the initial displacement and velocity of the

two masses in such a way that they are equal.

In order to excite mode B, we must make A A = 0. This is

achieved if x1(0) = -x2(0) and v1(0) = -v2(0).

More about normal modes

For the two normal modes of our example

x 1 (t) = x 2 (t) (mode A) (9)

and

x 1 (t) = − x 2 (t) (mode B) (10)

This means that a normal mode has a characteristic displace-

000
111 000
111
ment pattern. For example, the displacement pattern for

111
000 111
000
mode A is indicated in Fig. 5.

111
000
111 111
000
000
111
111
000
000
111
000
111
000
111
000
Figure 5 Displacement pattern for the excitation of mode A

Of course, after a certain time the direction of the arrows

will be reversed, but the point is that they remain equal at

000
111 000
111
all times. Similarly, the displacement pattern for mode B is

111
000 111
000
given in Fig. 6.

111
000
111
000 111
000
111
000
111
000 111
000
111
000 111
000
Figure 6 Displacement pattern for the excitation of mode B

At a later time, the two arrows will point away from each
other, so that they remain equal in magnitude but opposite
in direction at all times. From these displacement patterns

phy 241 MenŽndez 52

 Chapter 3

it is easy to understand the frequencies we found for the two

normal modes. For normal mode A, the middle spring is
never stretched. Hence the forces acting on the masses
originate only in the left and right springs, as if the masses
were attached to a single spring. This explains the simple
expression ω A = (K/m) 1/2 for the frequency of this mode.
For normal mode B, the middle spring is stretched by twice
the amount the other two springs are stretched. Hence the
force acting on the masses is three times the force acting on
them in mode A. This explains the frequency ωB = (3K/m)1/2.

If the system is not in a normal mode, there is no character-

istic displacement pattern. This is because the ratio
x1 (t)/x2 (t) is not constant, as can be seen from Eqs. 7 and 8
and Fig. 2. Since the ratio is not constant, the relative
magnitude and direction of the displacement arrows will
change with time. In Fig. 2, for example, x 1 (t) < x2 (t)
initially, but at later times x1(t) > x 2(t). This is because the
ratio x1(t) /x 2 (t) depends on time unless the system is in a
normal mode.

There are normal modes for any system of masses and

springs. In a more complicated system (different masses,
springs with different spring constants, etc.) the displace-
ments of the masses for a normal mode X are given by

x 1 (t) = A X (1) cos (ω Xt + α X)

,
x 2 (t) = A X (2) cos (ω Xt + α X) (11)

where the ratio x1 (t)/x2 (t) = A X (1)/A X (2) is a constant, but

not necessarily +1 or -1as in our simple example.

A way to find the ratio A X (1)/A X (2) (in other words, a way
to find the displacement pattern for the normal mode) is to
realize that in a normal mode X all masses oscillate with the
same frequency ω X. From our solution to the simple
harmonic oscillator problem, we now that the square of this
frequency is equal to minus the force per unit mass per unit
displacement. By equating such an expression for all masses,

phy 241 MenŽndez 53

 Chapter 3

we find an equation for the amplitudes. Let us apply this

method to our simple example. From Eqs. (1) and (2),
minus the force per unit mass per unit displacement is given
by
− K x 1 + K (x 2 − x 1 ) − K x 2 − K (x 2 − x 1 )
ω X2 = − =− ,
mx 1 mx 2 (12)

where the equal signs are only valid if the system is in a

normal mode, for it is only in this case that the oscillations
occur at a single frequency. But if the system is in a normal
mode, we can use Eq. (11) for the displacements, so that
Eq.(12) becomes

ω X2 =
− K A X (1) + K A X (1) − A X (2) − K A X (2) − K A X (1) − A X (2)
− = −
mA X (1) mA X (2)
(13)

Notice that the time-dependent cosine functions cancel out

in this expression because they are identical for x 1 (t) and
x 2 (t), since the system is assumed to be in normal mode X.
Eq. (13) can be simplified to

A X (2) A X (1) − A X (2) = A X (1) A X (2) − A X (1) (14)

This equation has two solutions: AX(1) = AX(2) and AX(1) =

-AX(2). These two solutions correspond to normal modes A
and B. This can be seen by comparing with Eqs. (7) and (8)
and by replacing the solutions into Eq. (13), which yields the
frequencies ω A = m and ω B = m , respectively.
2 K 2 3K

Systematic search for normal modes

In the previous sections, we discussed two methods of
finding the normal modes of a system. First, we solved the
two coupled equations (1) and (2) by finding an appropriate
combination of these two equations that leads to simple
harmonic oscillator solutions for the two normal modes. In
the previous paragraph, we found the normal modes by using

phy 241 MenŽndez 54

 Chapter 3

the condition that the two masses oscillate with the same
frequency. These methods become increasingly more
complicated if the masses are not all equal, if the spring
constants happen to be different from spring to spring, and,
most importantly, if the number of masses connected by
springs is larger than 2. Fortunately, there is a systematic
way of solving the equations of motion for coupled springs
that overcomes these difficulties. Suppose that we have two
simultaneous equations of the form
d2 x 1
= − a1 1 x 1 − a1 2 x 2
dt 2
,
d2 x 2
= − a2 1 x 1 − a2 2 x 2
dt 2 (15)

where the coefficients a 11 , a12, a 21 , and a22 are constants.

(Note that Eqs. (1) and (2) are of this form, with a11 = a22
= 2K/m and a 12 = a 21 = -K/m).Now we assume that there
exist solutions in the form of normal modes, given by Eq.
(11). This implies
2 2
d x1 d x2
= − ω x1, = − ω X2 x 2
2
2 X 2
dt dt (16)

Substituting into Eqs. (15), and rearranging, we obtain

(a 1 1 − ω )x 1 + a 1 2 x 2 = 0,
2

a2 1 x 1 + (a 2 2 − ω 2 )x 2 = 0. (17)

These are two standard simultaneous algebraic equations.

You remember from your algebra courses that for such a
system to have a solution other than the trivial solution (x1
= x2 = 0), the determinant of the coefficients must vanish:

a1 1 − ω 2 a1 2
/ (a 1 1 − ω 2 )(a 2 2 − ω 2 ) − a1 2 a2 1 = 0.
a2 1 a2 2 − ω 2 (18)

This is a quadratic equation in ω 2, from which two solutions

ωΑ and ωB are obtained. These are precisely the frequencies
of the two normal modes. Once these frequencies are

phy 241 MenŽndez 55

 Chapter 3

determined, the relationship between the coefficients A X(1)

and AX(2) follows from Eqs. 17:

ω A − a1 1
2
x2 A A (2)
= =
x1 mode A
A A (1) a1 2

ω B − a1 1
2
x2 A B (2)
= =
x1 mode B
A B (1) a1 2 (19)

where we have used the first equation in (17). We could

have used the second equation to obtain exactly the same
result; this is because the two equations are made propor-
tional to each other by the requirement that the determi-
nant of the coefficients be zero. You should verify that by
applying Eq. (18) and Eq. (19) to our initial problem you
obtain ω A2 = m , ω B2 = m , and AA(1)/AA(2) = 1, AB(1)/AB(2) =
K 3K

-1. This is of course the same solution we found before. The

advantage of the systematic method, however, is that the
solution is not more difficult if the masses or the springs are
different, whereas the other methods of solving the equa-
tions of motion become more complicated.

Finally, one can write the most general solution as a linear

combination of the two normal modes:

x 1 (t) = A A (1) cos (ω At + α A) + A B (1) cos (ω Bt + α B )

x 2 (t) = A A (2) cos (ω At + α A) + A B (2) cos (ω Bt + α B ) (20)

This has six constants, but due to Eqs. (19) only four are
independent. This is exactly what we need to accommodate
the initial position and velocity of the two masses.

The systematic method can be generalized to an arbitrary

number of masses connected by springs. There are as many
normal modes as degrees of freedom in the system,
whereby the number of degrees of freedom is the number of
functions of time we are trying to find. For example, in the
problem of the two masses connected to springs we are
looking for the functions x 1 (t) and x 2 (t). Thus this system
has two degrees of freedom and two normal modes. If we

phy 241 MenŽndez 56

 Chapter 3

had three masses moving along a line, we would be looking

at a system with three degrees of freedom and three normal
modes. The systematic method can still be applied, but we
would have one extra equation of motion in Eq.(15). This
leads to a 3 × 3 determinant instead of the 2 × 2 system we
discussed above. As more masses are added to the system,
the determinant becomes larger and larger. By the time we
have 5 or more masses, it must be solved numerically.
Unfortunately, the numerical algorithms used to solve
determinants have a serious shortcoming: the time needed
by a computer to solve the problem is proportional to the
cube of the number of masses. If 10 seconds are needed to
solve the problem of 3 masses, the computer requires 80
seconds to solve a problem that involves six masses. This
makes it impossible to compute the vibrations of atoms in
many types of crystals. For example, in a fast state-of-the-art
workstation the time needed to solve the problem of 600
interconnected masses is of the order of 1 hour. 6000
masses would require 42 days! However, even more masses
are needed to study certain phenomena, such as the motion
of dislocations in crystals. Worse, in a real crystal the atoms
can move in three directions, so that for each mass we have
3 degrees of freedom. Hence the number of normal modes
in a three-dimensional system with N masses is 3N and the
determinant to solve has a 3N × 3N size. One of the Òhot
topicsÓ in condensed matter physics today is the develop-
ment of Òorder NÓ methods, for which the computer time is
proportional to N rather than N3.

Guessing the normal modes

In cases where the system under investigation has some
ÒniceÓ symmetries, such as the mass-spring system in Fig. 1,
it is possible to guess the displacement pattern and the
frequency of the normal modes. In a normal mode, all parts
of the system move with the same frequency, so that the
magnitude of the restoring force per unit mass per unit
displacement must be the same. If all masses are equal, as in
Fig. 1, this is equivalent to requiring that the restoring force

phy 241 MenŽndez 57

 Chapter 3

per unit length be the same for all masses. Notice that this is
precisely the case for the displacement patterns of Figs. 5
and 6. The restoring force per unit displacement is K for the
two masses in Fig. 5 and 3K for the two masses in Fig. 6
Hence the frequency of the two modes are (K/m)1/2 and
(3K/m)1/2, respectively.

Notice that it would have been impossible to guess the

displacement patterns and the frequencies of the normal
modes if the masses had not been equal or if all springs had
had different constants. It is the symmetry of the problem
that allows us to guess its normal modes. The relationship
between normal modes and symmetry is so important that it
has been investigated systematically using group theory, the
branch of mathematics that deals with symmetry. Using
group theory techniques (which are beyond the scope of this
course) one can use symmetry properties to determine the
form of the displacement patterns. This is very important in
connection with the above-mentioned problem of solving
large determinants: using group theory, one can often reduce
a problem that consists of a large determinant into several
smaller problems that consist of smaller determinants. A
classical example is that of a periodic system: a system that
repeats itself in space, such as a perfect crystal. When group
theory is applied to this problem, it can be shown that the
very large determinant for the normal modes of the system
can be reduced to a set of small determinants whose size is
equal to the number of atoms in the unit that repeats itself.
This number is usually small. For example, silicon, the most
important material for modern technology, has only 2 atoms
per Òunit cellÓ. Hence the vibrations of silicon can be
calculated solving small 6 × 6 determinants. (It is 6 × 6 and
not 2 × 2 because each atom in a real crystal has three
degrees of freedom x, y, and z).

Damped normal modes

So far we have neglected the effect of damping in our
discussion of normal modes. In a real system there is always
damping, for the normal mode oscillations eventually die

phy 241 MenŽndez 58

 Chapter 3

out, in much the same way the vibration of a single mass is

damped due to friction. Friction adds a term of the form
-λv to the equations of motion of the masses in the system.
For example, Eqs. (1) and (2) become

mx¨ 1 = − K x 1 + K ‡ x 2 − x 1 Ž − λ 1 x˙ 1 (21)

and

mx¨ 2 = − K x 2 − K ‡ x 2 − x 1 Ž − λ 2 x˙ 2 (22)

If we assume the same friction coefficient for the two

masses, i.e., λ 1 =λ 2=λ, we obtain, by adding and subtracting
these equations,

m ‡ x¨ 1 + x¨ 2 Ž = − K ‡ x 1 + x 2 Ž − λ ‡ x˙ 1 + x˙ 2 Ž (23)

and

m ‡ x¨ 1 − x¨ 2 Ž = − 3K ‡ x 1 − x 2 Ž − λ ‡ x˙ 1 − x˙ 2 Ž (24)

These are damped harmonic oscillator equations for the

variables x 1 (t)+ x 2 (t) and x 1 (t)- x 2 (t). From Chapter 2, we
already know the solution to this type of equation. We
obtain

x 1 (t) + x 2 (t) = 2A A e − γt cos ‡ ω At + α A Ž

,
x 1 (t) − x 2 (t) = 2A B e − γt cos ‡ ω Bt + α B Ž (25)

with γ = λ/2m, ω A2 = m − γ 2, and ω B2 = m − γ 2. The motion of the

K 3K

individual masses is finally described by

x 1 (t) = A A e − γt cos ‡ ω At + α A Ž + A B e − γt cos ‡ ω Bt + α B Ž

x 2 (t) = A A e − γt cos ‡ ω At + α A Ž − A B e − γt cos ‡ ω Bt + α B Ž (26)

Hence the solution to the damped system of connected

masses is particularly simple: each normal mode behaves like
an independent damped harmonic oscillator. This simplicity,
however, is somewhat fortuitous. If λ 1 ≠λ 2 , a not very
unreasonable assumption, we can no longer separate the
equations of motion into two independent equations for the

phy 241 MenŽndez 59

 Chapter 3

functions x1(t)+x2(t) and x1(t)-x2(t). Physically, this means

that damping can ÒmixÓ normal modes: if you select your
initial conditions so that only one mode is excited, other
normal modes may be indirectly excited due to the effect of
damping. In addition to this complication, our mathemati-
cal treatment of friction misses an important feature of
damping in systems with several moving parts. We model
friction as a force acting on the vibrating masses. In real
systems, however, friction forces are distributed over the
entire system. In the case of masses and springs, an impor-
tant component of the friction force acts on the springs
themselves, which become warmer and dissipate energy as
they contract and expand repeatedly. In the case of a single
mass, this effect can be included as an effective friction force
assumed to act on the mass. For systems with more than
one mass, however, the distinction between friction from
the springs and friction from the masses is important.
Consider for example the two normal modes A and B
depicted in Figs. 5 and 6. If the middle spring makes an
important contribution to damping, its contribution will not
show up in mode A, because the middle spring is never
stretched in this mode. By contrast, the spring is com-
pressed and expanded in mode B. Therefore, we would
expect a larger damping for mode B. This conclusion is very
general:

In a system with several normal modes, the

normal mode with the lowest frequency is
usually the one with the smallest damping.
This is because there are less moving parts
for this mode.

For example, if you shake an aquarium tank, you will excite

many normal modes of the water in that aquarium. After a
while, however, you will see very clearly the lowest mode, for
which the water goes up at one end and down at the other
end of the aquarium.

Forced oscillations of normal modes

Suppose that in our system of two masses we apply a force

phy 241 MenŽndez 60

 Chapter 3

F 0 cos ω f t to the first mass. The equations of motion

become

mx¨ 1 = − K x 1 + K ‡ x 2 − x 1 Ž − λ x˙ 1 + F 0 cos ω f t (27)

and

mx¨ 2 = − K x 2 − K ‡ x 2 − x 1 Ž − λ x˙ 2 (28)

By adding and subtracting these equations we obtain two

forced oscillator equations for the variables x1 (t)+x2 (t) and
x1(t)-x2(t):

m ‡ x¨ 1 + x¨ 2 Ž = − K ‡ x 1 + x 2 Ž − λ ‡ x˙ 1 + x˙ 2 Ž + F 0 cos ω f t
m ‡ x¨ 1 − x¨ 2 Ž = − 3K ‡ x 1 − x 2 Ž − λ ‡ x˙ 1 − x˙ 2 Ž + F 0 cos ω f t (29)

Notice that the two normal modes are driven by the same
force. This is due to our particular choice for the external
force. It is possible (see end of chapter problem) to apply
forces on the masses so that only one of the normal modes is
excited. On the other hand, even though the two normal
modes are driven by the same force, their response will be
different depending on the value of the frequency ω f. If ωf is
K
very close to m , mode A will be in resonance, so that its
amplitude will usually be much larger. Conversely, if ω f is
3K
close to m , mode B will be in resonance. If the separation
between the frequencies of the modes is much larger than
the broadening parameter γ, it is possible to resonantly
excite only one normal mode. But if the separation is of the
order of γ, one may not be able to isolate the modes by
tuning the exciting frequency ω f. Sometimes, this may lead
to catastrophe. The famous Tacoma Narrows Bridge had
two normal modes: an up-and-down motion and a torsional
motion. Their frequency difference was of the order of γ.
When the wind excited the up-and -down normal mode, the
torsional motion picked up a significant amount of the
external energy. Eventually, the bridge collapsed due to the
torsional motion.

phy 241 MenŽndez 61

 Chapter 3

b) Show that the displacement as a function of time

Problems for the mass on the left can be written as x(t) = A
cos ω1 t cos ω2 t, where ω 1 = (ω A +ω B )/2 and ω 2 =
(ωA-ωB)/2.
1. Two 1 kg masses are connected by springs as in
Fig. 1. The spring constant of the three springs is K c) Guess the values of the frequencies ω A and ω B if
= 100 N/m. The mass on the left is initially the middle spring is replaced by a different spring of
displaced to the right by 0.2 m. The mass on the constant KÕ.
right is initially displaced to the right by 0.1 m.
d) Study the motion of the mass on the left in the
Their initial velocities are zero. Calculate the
limit KÕ→0. Graph this displacement as a function
position as a function of time for the two masses
both analytically and numerically using the step-by- of time. Invent your own parameters.
step procedure discussed in previous chapters. 5. Discuss the normal modes of vibration of the
Graph together the analytical and the numerical C O 2 molecule. Go to the library, find the
solutions.
frequencies of the normal modes, and deduce the
2. Suppose that in the previous problem the mass value of the spring constant between the carbon and

00
11
on the right is m = 2 kg. Find the normal mode oxygen atoms.

11
00
frequencies. Because the different masses destroy
O C O
11
00
the symmetry of the problem, it may be easier to
use the systematic approach discussed in the text.

3. Determine the mode displacement patterns and

6. Find the frequencies of the two normal modes of
the mode frequencies for the system below. Rather
oscillation of the coupled LC circuits shown in the
than performing an actual calculation, which
figure. Compare your solutions with the case of two
involves a 3× 3 determinant, try to guess a
masses connected by springs.
displacement pattern and require that the force per
unit mass per unit displacement be the same for all Ia Ib
masses. To simplify the notation, you might want

00
11 00
11
to use a , b, and c rather than AX (1), A X (2), and
AX(3)

11
00 11
00 L L
11
00 K
m
K
m
K
m
K
11
00 Q1
C C
Q2
C
Q3
-Q 1 -Q 2 -Q 3

4. Suppose that in the mass-spring system of Fig. 1

the mass on the left is initially displaced by an
amount A. The other mass is, initially, at the 7. Write down the steady -state solutions to Eq.
equilibrium position. The initial velocities are zero. (29). On the same graph, plot the amplitude of the
two normal mode components of the motion of
a) Find the expression that describes the position as mass 1 as a function of ω f . Invent your own
a function of time for the two masses. parameters. Discuss the conditions under which

phy 241 MenŽndez 62

 Chapter 3

one of the amplitudes is much larger than the other.

For what values of the parameters is this possible?

8. For the two-mass system of Fig. 1, describe how

to apply forces so that only one of the normal
modes is excited.

phy 241 MenŽndez 63