Massachusetts Institute of Technology OpenCourseWare

8.03SC Fall 2012

Problem Set #3 Solutions

Problem 3.1: (French 5-10)1 Coupled Oscillators using two springs

Let the displacement from the equilibrium posi­
tions for masses m1 and m2 be x1 and x2 respec­
tively. Then the tensions in the two strings are T1 = |kx1 |
T2 = |k(x2 − x1 )|
T1 = kx1 and T2 = k(x2 − x1 ), respectively. Now Equilibrium

m1 x¨1 = +k(x2 − x1 ) − kx1 T1 x1

m2 x¨2 = −k(x2 − x1 ) T2
+
Equilibrium
Substituting m1 = m2 = m and ωs2 = k/m:
x¨1 = ωs2 (x2 − 2x1 ) T2
x2

x¨2 = ωs2 (x1 − x2 ) (1)

Let x1 = C1 cos(ωt), and x2 = C2 cos(ωt). Now using these in Eq. 1
− ω 2 C1 + 2ωs2 C1 = ωs2 C2
−ω 2 C2 + ωs2 C2 = ωs2 C1 (2)

Method I: Without using Cramer’s Rule

From Eq. 2 we get
C1 ωs2 ωs2 − ω 2
= = ωs4 = 2ωs4 − 3ωs2 ω 2 + ω 4 ω 4 − 3ωs2 ω 2 + ωs4 = 0
C2 2ωs2 − ω 2 ωs
2
2 3ωs2 ± 9ωs4 − 4ωs4 √ ωs2 √
k
ω = = 3± 5 = 3± 5
2 2 2m
√ √
√ k √ k ω+ 3+ 5 5+1
ω+ = (3 + 5) ω− = (3 − 5) = √ =√
2m 2m ω− 3− 5 5−1
√ C1 ωs2 2ωs2 2
For ω+ = (3 + 5)k/2m = 2
= √ = √
C2 2ωs − ω+
2
4ωs − (3 + 5)ωs
2 2 1− 5
√ C1 ωs2 2ωs2 2
For ω− = (3 − 5)k/2m = 2
= √ = √
C2 2ωs2 − ω− 4ωs2 − (3 − 5)ωs2 1+ 5
Method II: Using Cramer’s Rule
On collecting coefficients of C1 and C2 in Eq. 2 we get (2ωs2 − ω 2 )C1 − ωs2 C2 = 0 and
−ωs2 C1 + (ωs2 − ω 2 )C2 = 0.
1
The notation “French” indicates where this problem is located in one of the textbooks used for 8.03 in 2004:
French, A. P. Vibrations and Waves. The M.I.T. Introductory Physics Series. Cambridge, MA: Massachusetts
Institute of Technology, 1971. ISBN-10: 0393099369; ISBN-13: 9780393099362.
 0
−ωs2 2ωs2 − ω 2 0

0
ω 2 − ω 2 −ωs2 0

s
C1 =
C2 =

2ωs2 − ω 2 −ωs2 2ωs2 − ω 2 −ωs2

−ω 2 ωs2 − ω 2

−ω 2 ωs2 − ω 2

s s

Non-zero solutions for C1 and C2 only possible if

2ωs2 − ω 2 −ωs2
=
0 ⇒
ω 4 − 3ωs2 ω 2 + ωs4 = 0

−ωs2 ωs2 − ω 2
This is the same as the last equation in line 1 of Method I. From here on, the solution is identical.
Problem 3.2: (French 5-11) Coupled spring and pendulum
a) The tension in the string is T ≈ M2 g. l

The equation of motion for mass M2 in the x- +

θ
direction is as follows N = M1 g + T cos θ

M2 x¨2 = −M2 g sin(θ)

T
T cos θ ≈ M g
2

g
M1 T sin θ ≈ M2 g sin θ
M2 x¨2 = −M2 (x2 − x1 ) M 2

l T sin θ ≈ M g sin θ
2 kx1
T
and for mass M1 is M g2 T cos θ ≈ M2 g
g
M1 x¨1 = −kx1 + M2 (x2 − x1 )
l g M1 g
2 2
b) & c) Substituting ωs = k/M2 , ωp = g/l and M1 = M2 = M we get
x¨2 + ωp2 x2 − ωp2 x1 = 0

x¨1 + (ωs2 + ωp2 )x1 − ωp2 x2 = 0

Let x1 = C1 cos(ωt), x2 = C2 cos(ωt). − ω 2 C2 + ωp2 C2 = ωp2 C1

Substituting gives:
−ω 2 C1 + (ωs2 + ωs2 )C1 = ωp2 C2 (3)
Method I: Without using Cramer’s Rule

C1 −ω 2 + ωp2 ωp2
= = ωp4 = ω 4 − ω 2 ωs2 − 2ω 2 ωp2 + ωp2 ωs2 + ωp4
C2 ωp2 −ω 2 + ωp2 + ωs2

2ωp2 + ωs2 1 q
ω 4 − (2ωp2 + ωs2 )ω 2 + ωp2 ωs2 = 0 ω2 = ± (2ωp2 + ωs2 )2 − 4ωp2 ωs2
2 2
�1/2
2ωp2 + ωs2 1
q 4
�
2
2
2ωp + ωs2 1

4 4 1/2
ω =
±
4ωp + ωs4 ω± = ± 4ωp + ωs
2 2
2 2

C1 2
−ω+ + ωp2 −ωs2 − 4ωp4 + ωs4
For ω+ = =
C2 ωp2 2ωp2
q
2
C1 2
−ω− + ωp2 −ω s + 4ωp4 + ωs4 )
For ω− = =
C2 ωp2 2ωp2

MIT OCW 8.03SC 2 Problem Set #3 Solutions

Method II: Using Cramer’s Rule

Collecting coefficients in Eq. 3: ωp2 C1 + (ω 2 − ωp 2 )C2 = 0 and (−ω 2 + ωp2 + ωs2 )C1 − ωp2 C2 = 0.

 
 0 ω2 − ω2 
 p 
0 −ωp2

C1 = 

 ωp 2 ω 2 − ωp2 

 −ω 2 + ω 2 + ω 2 −ωp2 
 p s

2


 ω p 0 

 −ω 2 + ω 2 + ω 2 0 
p s
C2 =  2
ω 2 − ωp2

 ωp 

 −ω 2 + ωp2 + ωs2 −ωp2
 

Non-zero values of C1 and C2 only possible if

ωp 2 ω 2 − ωp2 

⇒ −ωp4 − (ω 2 − ωp2 )(−ω 2 + ωp2 + ωs2 ) = 0


 −ω 2 + ω 2 + ω 2  = 0
p s −ωp2

ω 4 − (ωs2 + 2ωp2 )ω 2 + ωp2 ωs2 = 0

This is the same as found for Method I (see second line). From here on, the solution is identical.
Problem 3.3: (Bekefi & Barrett 1.16)2 Coupled oscillators using three springs
Side (a) of the figure shows the sys­
x1
tem at rest and side (b) shows it at
some random time t. Displacements
y1
from Equilibrium are x1 and x2 . Now d1
y1 = d1 + d2 + x1 and y2 = d2 + x2
x2
a) The equations of motion are: +
mẍ1 = −2kx1 − k(x1 − x2 )
d2 y2
⇒ ẍ1 + 3ω02 x1 − ω02 x2 = 0
mẍ2 = +k(x1 − x2 )
(a) (b)
⇒ ẍ2 + ω02 x2 − ω02 x1 = 0 x2

where ω02 = k/m

b) Substituting x1 = A cos(ωt) and x2 = B cos(ωt) in the equations of motion gives
A(3ω02 − ω 2 ) = Bω02 and Aω02 = B(ω02 − ω 2 )
2
The notation “Bekefi & Barrett” indicates where this problem is located in one of the textbooks used in 8.03
in 2004: Bekefi, George, and Alan H. Barrett Electromagnetic Vibrations, Waves, and Radiation. Cambridge, MA:
MIT Press, 1977. ISBN: 9780262520478.

MIT OCW 8.03SC 3 Problem Set #3 Solutions

 A ω02 ω02 − ω 2
= = ω04 = 3ω04 − 4ω 2 ω02 + ω 4
B 3ω02 − ω 2 ω02
√
ω 4 − 4ω02 ω 2 + 2ω04 = 0 2
ω± = ω02 (2 ± 2)

√ B √ √ B √
For ω1 = ω0 (2 − 2)1/2 =1+ 2 For ω2 = ω0 (2 + 2)1/2 =1− 2
A A
Hence the general solutions are:
y1 (t) = d1 + d2 + x1 (t) = d1 + d2 + A cos(ω1 t + α) + B cos(ω2 t + β)
√ √
y2 (t) = d2 + x2 (t) = d1 + d2 + (1 + 2)A cos(ω1 t + α) + (1 − 2)B cos(ω2 t + β) (4)

x1
x1
c) Side (a) of the figure shows
y1
the normal mode with higher d1

frequency ω2 such that x2 (t) =

√
(1 − 2)x1 (t). Side (b) shows x2
x2 +
the normal mode with lower
frequency ω1 such that x2 (t) = d2 y2
√
(1 + 2)x1 (t).

(a) (b)
x2

Problem 3.4: Driven coupled oscillator

a) The equation of motion for mass M2 is unchanged M2 x¨2 =

−M2 g sin(θ) and for mass M1 is +
g X0 cos(ωt)
M1 x¨1 = −k[x1 − X(t)] + M2 (x2 − x1 )
l M1

g k

M1 x¨1 + kx1 + M2 (x1 − x2 ) = kX0 cos(ωt)

l x1 θ

b) Substituting ωs2 = k/M2 , ωp2 = g/l and M1 = M2 = M ,

x2
x¨2 + ωp2 x2 − ωp2 x1 = 0 M2

Equilibrium for both masses

x¨1 + (ωs2 + ωp2 )x1 − ωp2 x2 = ωs2 X0 (5)
Let x1 = C1 cos(ωt) x2 = C2 cos(ωt). Now using these in Eq. 5
ωp2 C1 + (ω 2 − ωp2 )C2 = 0 (−ω 2 + ωs2 + ωs2 )C1 − ωp2 C2 = ωs2 X0 (6)

MIT OCW 8.03SC 4 Problem Set #3 Solutions

  
 0
 ω 2 − ωp2 

 ω2X −ωp2 kX0 (g − lω 2 )

s 0
C1 =  =

 ωp2 ω 2 − ωp2  M lω 4 − (2M g + kl)ω 2 + kg

 −ω 2 + ωp2 + ωs2 −ωp2
 


2


 ω p 0 

 −ω 2 + ω 2 + ω 2 ω 2 X0  kgX0
p s s
C2 = =

ω2 ω 2 − ωp2 
 M lω 4 − (2M g + kl)ω 2 + kg
 2 p2

−ω + ωp + ωs2 −ωp2 


These are the steady state solutions. The general problem is a linear combination between the
transient problem and the steady state solutions. Notice that the transient problem has four
adjustable parameters which follow from the initial conditions.
c) The figure shows the plot of am­
Plot of amplitudes as function of t
plitudes C1 (blue line) and C2 (red 5

line) as a function of the frequency.

4
Note: At ω = 0, the amplitudes are
3
C1 = C2 = X0 . This figure is unrealis­ tp
tic. It was derived (i) under the small 2

angle approximation and (ii) for zero 1

C/X0 A

damping. Thus, the very large ampli­

0
tude for C1 and C2 as shown are mean­ tï t+
ï1
ingless. If you add sufficient damping,
and if you cannot make the small an­ ï2

gle approximations, because the angles ï3

are large, the problem becomes sub­

ï4
stantially more complicated. But it
ï5
can be solved numerically. You will 0 1 2 3 4 5 6 7 8 9 10
tA
then find meaningful values for the
amplitudes. A more insightful way to express (and plot) the amplitude of the pendulum would be
to do this in terms of the angle θ, rather than C2 .
d) We can note from the functional form of C2 that it cannot haverthe value zero (except for
g
ω → ∞). However, C1 will be zero when g − lω 2 = 0 ω = ωp = C1 → 0. This is the
l
resonance frequency of the pendulum. Thus, at this frequency, the two horizontal forces on the
upper mass, kXo cos(ωt) and T sin(θ), cancel. Since sin(θ) < 1, kXo must always be smaller than
T . At first sight, this inequality seems a bit bizarre, as, according to our derivation, the frequency
at which the upper mass stands still is independent of the spring constant k. Also, keep in mind

MIT OCW 8.03SC 5 Problem Set #3 Solutions

that we never had to make any assumption regarding k in our derivation (the inequality must have
been met automatically without our realizing it).
You SHOULD also ask yourself the question: How on Earth can the pendulum swing if the
mass attached to the spring does not move at all; what is driving the pendulum? The
answer is simple: it is not possible! It is only possible in our dream-world of zero damping. In
the presence of damping, no matter how little, the peculiar state is unstable. This can easily be
seen as follows.
Assume that the system is in that state. That means that at any moment in time the net horizontal
force on the pendulum mass is zero. Thus, the vectorial sum of the spring force and T sin(θ) must
be ZERO. However, if the mass on the spring is not moving, the pendulum is no longer driven,
and thus its amplitude will decay, and the net force on the mass on the spring is no longer zero,
and thus that mass will start to move. Thus, the peculiar state is unstable. You will be able to go
through that “special” state by varying ω, but you cannot ”stop” there. However, I demonstrated
in lectures (9/28) using 3 different driven systems, that you can get very close to those “special”
states, and that is already amazing (and very non-intuitive).

MIT OCW 8.03SC 6 Problem Set #3 Solutions

MIT OpenCourseWare
http://ocw.mit.edu

8.03SC Physics III: Vibrations and Waves

Fall 2012

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.