8.

03 Lecture 4
Coupled oscillators

In general, the motion of coupled systems can be extremely complicated. Editors note: watch the
video lectures to see examples of complicated coupled oscillators.
Let’s consider an example:

There are many kinds of motion in this system! If you stare at it long enough you an identify a
special kind of motion! The “normal mode:” every part of the system is oscillating at the same
phase and the same frequency. We will later realize the most general motion is a superposition of
the normal modes. We can understand the system systematically step-by-step.
In general, coupled oscillators are complicated but there are easier cases we can solve, being guided
by our physical intuition.
Can you guess the normal modes of this example?
 2 = k 2 = 4k 2k
Where in Mode A we have !A m and in Mode B we have !B 2m = m
Is there a Mode C? Yes!

Where there is no force, and !C = 0 because

there is no oscillation. The whole system is simply
translating.

To summarize:

Mode A:
x1 = 0
x2 = A cos (!A t + °A )
x3 = −A cos (!A t + °A )
Mode B:
x1 = B cos (!B t + °B )
x2 = −B cos (!B t + °B )
x3 = −B cos (!B t + °B )
Mode C:
x1 = x2 = x3 = c + vt

Therefore the general solution is:

x1 = 0 + B cos (!B t + °B ) + c + vt
x2 = A cos (!A t + °A ) − B cos (!B t + °B ) + c + vt
x3 = −A cos (!A t + °A ) − B cos (!B t + °B ) + c + vt

Where A, B, C, °A , °B and v are constants to be determined by the initial conditions. Here we have
3 second order di erential equations with 6 unknown constants.

From the analysis of the force diagram analysis

we get:

2mẍ1 = k(x2 − x1 ) + k(x3 − x1 )

mẍ2 = k(x1 − x2 )
mẍ3 = k(x1 − x3 )

2
We can reorganize:

2mẍ1 = −2kx1 + kx2 + kx3

mẍ2 = kx1 − kx2 + 0x3
mẍ3 = kx1 + 0x2 − kx3

Now our job is to solve the equations. It is possible to solve this coupled set of di erential equations
directly, but we can use a matrix as a tool to help us. We convert everything to matrices. Our
equation of motion is now
MX ¨ = −KX

where: 2 3 2 3 2 3
2m 0 0 x1 2k −k −k
M = 4 0 m 05 X = 4x2 5 and K = 4−k k 0 5
6 7 6 7 6 7
0 0 m x3 −k 0 k

We go to complex notation where Xj = Re[Zj ] and Z  ei(!t+°) A and A is a column vector

(A1 , A2 , A3 )
Solving the equation of motion:

M Z¨ = −KZ
M ! 2 Z = KZ
M ! 2 A = KA
! 2 A = M −1 KA
) (M −1 K−! 2 I)A = 0

Where I is the identity matrix. To have a soluation we need to solve

det[M −1 K − ! 2 I] = 0
−k −k
2 3
k
m − !2 2m 2m
(M −1 K − ! 2 I) = 4 −k k
− !2 0 5
6 7
m m
−k k
m 0 m − !2
Define !02  k/m
−! 2 −! 2
0 1
!2 − !2 0 0
B 0 2 2 2
det @ −!0 !02 − ! 2 0 A=0
C

−!02 0 !02 − ! 2

1 1
(!02 − ! 2 )3 − !04 (!02 − ! 2 ) − !04 (!02 − ! 2 ) = 0
2 2
(!0 − ! )(!0 − 2!0 ! + ! 4 − !04 ) = 0
2 2 4 2 2

(!02 − ! 2 )! 2 (! 2 − 2!02 ) = 0

3
 p
) ! = !0 , 2!0 , 0
s s
k 2k
!= , , 0
m m
We get the same result! To get the relative amplitude of a normal made: Plut in the normal
mode frequency you get in the equation (M −1 K − ! 2 I)A = 0. For example: take Mode B where
! = !B = 2k/m

0 = 2kA1 + kA2 + kA3

0 = kA1 + kA2 + 0 ) A1 = −A2 = −A3
0 = kA1 + 0 + kA3
In Mode B we had
0 1 0 1
x1 = B cos (!B t + °B ) x1 1
x2 = −B cos (!B t + °B ) or X = @x2 A = B @−1A cos (!B t + °B )
B C B C
x3 = −B cos (!B t + °B ) x3 −1
It turns out this is just the simple harmonic oscillator!
p
Take Mode A as an example where ! = !A = k/m. Plug in this frequency (into the equa-
tion (M −1 K − ! 2 I)A = 0) to solve: A1 = 0 and A2 = −A3
x1 = 0
x2 = A cos (!A t + °A )
x3 = −A cos (!A t + °A )

There is an alternative way we can solve for the normal modes. We can define the length of the
spring as l0 and define a new origin:

4
From the analysis of the force diagram we get:
2mẍ1 = k(x2 − x1 − l0 ) + k(x3 − x1 − l0 )
mẍ2 = k(x1 − x2 + l0 )
mẍ3 = k(x1 − x3 + l0 )
Redefine the x2 and x3 coordinates:
x02 = x2 − l0 x03 = x3 − l0
Now we have
2mẍ1 = k(x02 − x1 ) + k(x03 − x1 )
mẍ02 = k(x1 − x02 )
mẍ03 = k(x1 − x03 )
Reorganizing:
2mẍ1 = −2kx1 + kx02 + kx03
mẍ02 = kx1 − kx02 + 0x03
0
mẍ3 = kx1 + 0x02 − kx03
Now use the definition of normal mode:
0 1 20 1 3
x1 A1
B 0C 6B C i(!t+°) 7
@x2 A = Re 4@A2 A e 5
0
x3 A3
−2m! 2 A1 = −2kA1 + kA2 + kA3 0 = (−2k + 2m! 2 )A1 + kA2 + kA3
−m! 2 A2 = kA1 − kA2 + 0A2 ) 0 = kA1 + (m! 2 − k)A2 + 0A3
−m! 2 A3 = kA1 + 0A2 − kA3 0 = KA1 + 0A2 + (m! 2 − k)A3
Rewrite in matrix notation:
0 10 1
(2m! 2 − 2k ) k k A1
k 2
(m! − k) 0 A @A2 A = 0
B CB C
@
k 0 (m! 2 − k) A3
To get a solution we need to solve the equation where the determinant of the left matrix is zero.
0 1
(2m! 2 − 2k) k k
det @ k (m! 2 − k) 0 A=0
B C
k 0 2
(m! − k)

(2m! 2 − 2k)(m! 2 − k)2 − 2k 2 (m! 2 − k) = 0

h i
(m! 2 − k) (2m! 2 − 2k)(m! 2 − k) − 2k 2 = 0
(m! 2 − k)! 2 (2m2 ! 2 − 4km) = 0

s s
2k k
!= , , 0
m m
Get the same result!

5
MIT OpenCourseWare
https://ocw.mit.edu

8.03SC Physics III: Vibrations and Waves

Fall 2016

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