Taylor

Classical Mechanics [Taylor, J.R.
]
Solution Manual
Written by
JGSK
Last Updated
April 4, 2018
Contents
1 Newton’s Laws of Motion 2
2 Projectiles and Charged Particles 3
3 Momentum and Angular Momentum 4
4 Energy 5
5 Oscillations 6
6 Calculus of Variations 7
7 Lagrange’s Equations 10
8 Two-Body Central-Force Problems 16
9 Mechanics in Non-inertial Frames 19
10 Rotational Motion of Rigid Bodies 20
11 Coupled Oscillators and Normal Modes 36
12 Nonlinear Mechanics and Chaos 46
13 Hamiltonian Mechanics 47
14 Collision Theory 60
15 Special Relativity 61
16 Continuum Mechanics 62
1
Chapter 1
Newton’s Laws of Motion
1.1. No plans to do this chapter yet.
2
Chapter 2
Projectiles and Charged Particles
3
Chapter 3
Momentum and Angular Momentum
4
Chapter 4
Energy
5
Chapter 5
Oscillations
6
Chapter 6
Calculus of Variations
6.1. The shortest path between two points on a curved surface, such as the surface of a sphere, is called a
geodesic. To find a geodesic, one has first to set up an integral that gives the length of a path on the surface
in question. This will always be similar to the integral (6.2) but may be more complicated (depending on the
nature of the surface) and may involve different coordinates than x and y. To illustrate this, use spherical
polar coordinates (r , θ, ϕ) to show that the length of a path joining two points on a sphere of radius R is
∫ θ2 q
L=R 1 + sin2 θϕ 0(θ )2 dθ
θ1
if (θ 1 , ϕ 1 ) and (θ 2 , ϕ 2 ) specify the two points and we assume that the path is expressed as ϕ = ϕ(θ ).
Solution. Credit: http://www.physics.ohio-state.edu/~mathur/5300/5300ass2sol.pdf
6.2. Do the same as in Problem 6.1 but find the length L of a path on a cylinder of radius R, using cylindrical
polar coordinates (ρ, ϕ, z). Assume that the path is specified in the form ϕ = ϕ(z).
6.3. Consider a ray of light travelling in a vacuum from point P 1 to P2 by way of the point Q on a plane mirror.
Show that Fermat’s principle implies that, on the actual path followed, Q lies in the same vertical plane as
P1 and P2 and obeys the law of reflection, that θ 1 = θ 2 .
Solution. We consider the light ray travelling the path P1QP2 , and employ the notations given in the hints:
P1 = (0, y1 , 0), P2 = (x 2 , y2 , 0) and Q = (x, 0, z) (refer to Fig. 6.8 in the textbook). Then by Fermat’s principle,
light travels between two points along the path that requires the least time. Since the speed of light is
7
Classical Mechanics [Taylor, J.R.]
Solution Manual
constant throughout the path, the minimisation of the travel time is equivalent to the minimisation of the
path length L of P1QP 2 :
q q
L(z) = x 2 + y12 + z 2 + (x 2 − x)2 + y22 + z 2 (6.1)
dL z z !
⇒ =q +q =0 (6.2)
dz
x 2 + y12 + z 2 (x 2 − x)2 + y22 + z 2
Clearly, the derivative is 0 only when z = 0, i.e., Q is in the xy plane (shown).

Following this conclusion, we now consider the light ray constrained in the xy plane. The minimisation of
L(z = 0) with respect to x (because of Fermat’s principle) then gives:
! dL x (x 2 − x)
0= =q −q (6.3)
dx
x 2 + y12 (x 2 − x)2 + y22
x (x 2 − x)
q =q (6.4)
x 2 + y12 (x 2 − x)2 + y22
sin θ 1 = sin θ 2 (6.5)
and hence θ 1 = θ 2 (shown).
6.4. A ray of light travels from point P1 in a medium of refractive index n 1 to P2 in a medium of index n 2 , by
way of the point Q on the plane interface between the two media. Show that Fermat’s principle implies
that, on the actual path followed, Q lies in the same vertical plane as P1 and P2 and obeys Snell’s law, that
n 1 sin θ 1 = n 2 sin θ 2 .
Solution. Credit: http://chriskranenberg.wix.com/taylor-mechanics-solutions#!
6.5. Fermat’s principle is often stated as “the travel time of a ray of light, moving from point A to B, is minimum
along the actual path.” Strictly speaking, it should say that the time is stationary, not minimum. In fact
one can construct situations for which the time is maximum along the actual path. Here is one: Consider
the concave, hemispherical mirror shown in Figure 6.10 (in textbook), with A and B at opposite ends of a
diameter. Consider a ray of light travelling in a vacuum from A to B with one reflection at P, in the same
vertical plane as A and B. According to the law of reflection, the actual path goes via point Po at the bottom
of the hemisphere (θ = 0). Find the time of travel along the path APB as a function of θ and show that it is
maximum at P = Po .
Solution. Credit: http://chriskranenberg.wix.com/taylor-mechanics-solutions#!
8
Solution Manual
6.7. Consider a right circular cylinder of radius R centered on the z axis. Find the equation giving ϕ as a function
of z for the geodesic (shortest path) on the cylinder between two points with cylindrical polar coordinates
(R, ϕ 1 , z 1 ) and (R, ϕ 2 , z 2 ). Describe the geodesic. Is it unique? By imagining the surface of the cylinder
unwrapped and laid out flat, explain why the geodesic has the form it does.
√
6.8. Verify that the speed of the roller coaster car in Example 6.2 (in textbook) is 2дy. (Assume the wheels have
negligible mass and neglect friction.)
Solution. Applying conservation of energy when the car is released at height y (only potential energy) and
when it is at ground level (y = 0; only kinetic energy) gives:
1 2
mv = mдy
2
⇒ v = 2дy
p
which is what we needed to show.
[Remaining solutions to Chapter 6 can be found in http://chriskranenberg.wix.com/taylor-mechanics-solutions#

!]
9
Chapter 7
Lagrange’s Equations
7.1. Write down the Lagrangian for a projectile (subject to no air resistance) in terms of its Cartesian coordinates
(x, y, z), with z measured vertically upward. Find the three Lagrange equations and show that they are
exactly what you would expect for the equations of motion.
Solution. The Lagrangian can be easily identified and written as such:
L =T −U
1 (7.1)
= m(xÛ 2 + yÛ2 + zÛ2 ) − mдz
2
A direct application of the Euler-Lagrange equations yields:
∂L d ∂L
= ⇒ mxÜ = 0 (7.2)
∂x dt ∂xÛ
∂L d ∂L
= ⇒ myÜ = 0 (7.3)
∂y dt ∂yÛ
∂L d ∂L
= ⇒ mzÜ = −mд (7.4)
∂z dt ∂zÛ
All three of which are expected for a projectile in free fall. Note the negative sign on the z direction is due
to the chosen sign convention.
7.8. (a) Write down the Lagrangian L(x 1 , x 2 , xÛ1 , xÛ2 ) for two particles of equal masses, m 1 = m 2 = m, confined
to the x axis and connected by a spring with potential energy U = 12 kx 2 . [Here x is the extension of the
spring, x = (x 1 − x 2 − l), where l is the spring’s unstretched length, and I assume that mass 1 remains to the
right of mass 2 at all times.] (b) Rewrite L in terms of the new variables X = 12 (x 1 + x 2 ) (the CM position)
and x (the extension), and write down the two Lagrange equations for X and x. (c) Solve for X (t) and x(t)
and describe the motion.
10
Solution Manual
Solution. (a) The Lagrangian is simply:
1 1
L = m(xÛ1 2 + xÛ2 2 ) − k(x 1 − x 2 − l)2 (7.5)
2 2
(b) We note that we can manipulate the new variables X = 12 (x 1 + x 2 ) and x = (x 1 − x 2 − l), as such:
1
XÛ = (xÛ1 + xÛ2 ) and xÛ = xÛ1 − xÛ2 (7.6)
2
1
xÛ1 =XÛ + xÛ
2 (7.7)
1
xÛ2 =XÛ − xÛ
2
We can then rewrite the Lagrangian from Eq. (7.5) as:
1 1 2 1 2 1

L = m (X + x)
Û Û + (X − x)
Û Û − kx 2
2 2 2 2
(7.8)
1 1 1

= m 2XÛ 2 + xÛ 2 − kx 2
2 2 2
The Lagrange equations can be written fairly easily, one for each coordinate.
∂L d ∂L
(1) : = (7.9)
∂x dt ∂xÛ
1 2k
−kx = mxÜ ⇒ xÜ = − x (7.10)
2 m
∂L d ∂L
(2) : = (7.11)
∂X dt ∂XÛ
0 = 2mXÜ ⇒ XÜ = 0 (7.12)
(c) Now we are required to solve for X (t) and x(t). We tackle the easier of the two first:
XÜ = 0 ⇒ X (t) = v 0t + X 0 (7.13)
where we have taken the liberty to introduce two integration constants that depend upon the initial condi-
tions. As for the other coordinate x, we invoke the familiar SHM solution:
r !
2k 2k
xÜ = − x ⇒ x(t) = A cos +δ (7.14)
m m
where we have introduced two other integration constants A and δ . The center of mass moves with constant
velocity since no external forces are acting on it. The extension of the spring undergoes simple harmonic
motion, which just means that the two masses are oscillating with respect to each other.
11
Solution Manual
7.33. A bar of soap (mass m) is at rest on a frictionless rectangular plate that rests on a horizontal table. At time
t = 0, I start raising one edge of the plate so that the plate pivots about the opposite edge with constant
angular velocity ω, and the soap starts to slide toward the downhill edge. Show that the equation of motion
for the soap has the form xÜ − ω 2x = −д sin ωt, where x is the soap’s distance from the downhill edge. Solve
this for x(t), given that x(0) = x 0 . [You can easily solve the homogeneous equation; for a particular solution
try x = B sin ωt and solve for B.]
Solution. To obtain the equation of motion, we shall utilise the Lagrange equations. First, we write the
Lagrangian, noting that the motion of the soap consists of both translational and rotational motion:
1 1 2
L = mxÛ 2 + Iω − mдx sin ωt
2 2
1 1 2 2
= mxÛ 2 + mx ω − mдx sin ωt (7.15)
2 2
Applying the Lagrange equations, we get
d ∂L ∂L
= ⇒ mxÜ = mω 2x − mд sin ωt
dt ∂xÛ ∂x
xÜ − ω 2x = −д sin ωt (shown) (7.16)
Note that Eq. (7.16) is a second order, inhomogeneous differential equation. We first consider the homoge-
neous equation, which has a simple harmonic motion solution.
x H (t) = A1e ωt + A2e −ωt (7.17)
Now we consider the particular solution. We make a trigonometric ansatz x P = B sin ωt. Substituting into
Eq. (7.16),
−ω 2 B sin ωt − ω 2 B sin ωt = −д sin ωt (7.18)

д
⇒ B= (7.19)
2ω 2
We can then compose our general solution for x(t):
x(t) = x H + x P
д
= A1e ωt + A2e −ωt + sin ωt (7.20)
2ω 2
We have the initial conditions: x(0) = x 0 and x(0)

Û = 0, which yield us:
д
A1 + A2 = x 0 and ω(A1 − A2 ) + =0 (7.21)
2ω
12
Solution Manual
Using these conditions, we can solve for the integration constants A1 and A2 :
x0 д
A1 = − 2 (7.22)
2 4ω
x0 д
A2 = + 2 (7.23)
2 4ω
Finally, we are able to construct our general equation as:

x
0 д ωt x 0 д −ωt д
x(t) = − e + + e + 2 sin ωt (7.24)
2 4ω 2 2 4ω 2 2ω
д д
= x 0 cosh ωt − 2 sinh ωt + 2 sin ωt (7.25)
2ω 2ω
7.40. The "spherical pendulum" is just a simple pendulum that is free to move in any sideways direction. (By
contrast a "simple pendulum" is confined to a single vertical plane.) The bob of a spherical pendulum moves
on a sphere, centered on the point of support with radius r = R, the length of the pendulum. A convenient
choice of coordinates is spherical polars, r , θ, ϕ, with the origin at the point of support and the polar axis
(z−axis) pointing straight down. The two variables θ and ϕ make a good choice of generalized coordinates.
(a) Find the Lagrangian and the two Lagrange equations. (b) Explain what the ϕ equation tells us about the
z component of angular momentum lz . (c) For the special case that ϕ =const, describe what the θ equation
tells us. (d) Use the ϕ equation to replace ϕÛ by lz in the θ equation and discuss the existence of an angle θ 0
at which θ can remain constant. Why is this motion called a conical pendulum? (e) Show that if θ = θ 0 + ϵ,
with ϵ small, then θ oscillates about θ 0 in harmonic motion. Describe the motion of the pendulum’s bob.
Figure 7.1: Sketch of the problem
13
Solution Manual
Solution. (a) With reference to Figure 7.1, we can write the Lagrangian in spherical coordinates as:
1
L = m(R 2θÛ2 + R 2ϕÛ sin2 θ ) + mдR cos θ (7.26)
2
Then, applying Lagrange equation to the two generalized coordinates:
d ∂L ∂L
=
dt ∂θ
Û ∂θ
⇒ mR 2θÜ = mR 2ϕÛ2 sin θ cos θ − mдR sin θ (7.27)
d ∂L ∂L
=
dt ∂ϕÛ ∂ϕ
⇒ mR 2ϕÛ sin2 θ = lz = const (7.28)
We leave it to the reader to simplify Eq. (7.27). As for the latter equation, we have noted that the Lagrangian
∂L
is independent of ϕ, such that ∂ϕ = 0.
(b) Equation (7.28) tells us that the z component of the angular momentum is constant.
(c) If ϕ = const, then ϕÛ = 0, and Eq. (7.27) reduces to:
д
θÜ = − sin θ (7.29)
R
That is, we can describe the motion of the pendulum like a simple pendulum, at a fixed plane characterized
by ϕ = ϕ 0 , for some fixed ϕ 0 .
(d) We want to first express ϕÛ in terms of lz (using Eq. (7.28), then substitute it into Eq. (7.27).
lz
ϕÛ = (7.30)
mR 2 sin2 θ
2
lz д

θÜ = 2
sin θ cos θ − sin θ
2
mR sin θ R
(7.31)
2
l cos θ д
= 2z 4 3 − sin θ
m R sin θ R
To explore a possible stationary solution of θ , we consider θÜ = 0. Following from Equation (7.31), we obtain:
lz2 cos θ д
= sin θ (7.32)
m2 R 4 sin3 θ R
cos θ = k sin4 θ = k(1 − cos2 θ )2 (7.33)
where we have introduced the constant k = дm2 R 3 /lz2 . To analyze the solution, we substitute x = cos θ , into
Equation (7.33) where x ∈ [0, 1]. WolframAlpha gives us the following plot: We note that there is only one
solution for which Equation (7.33) is satisfied. This would correspond to the solution θ 0 , as specified in the
question.
(e) Consider a perturbation, θ = θ 0 +ϵ. We can Taylor expand the following quantities about the equilibrium
14
Solution Manual
Figure 7.2: Plot of Equation (7.33)
position x 0 , and also write down the derivatives of θ :
cos θ ≈ cos θ 0 − ϵ sin θ 0 (7.34)

sin θ ≈ sin θ 0 + ϵ cos θ 0 (7.35)
θÜ = ϵÜ (7.36)
Before we substitute these equations into Equation (7.31), we recall that at equilibrium, Equation (7.32) is
satisfied. That is, we have:
lz2 д sin4 θ 0
2 4
= (7.37)
mR R cos θ 0
which we can use to simplify our expression slightly later on. With the last 4 equations from above, Equa-
tion (7.31) gives us:
д sin4 θ 0 д
ϵÜ ≈ (cos θ 0 − ϵ sin θ 0 )(sin θ 0 + ϵ cos θ 0 )−3 − (sin θ 0 + ϵ cos θ 0 ) (7.38)
R cos θ 0 R
д sin4 θ 0 д
= (cos θ 0 − ϵ sin θ 0 )(sin−3 θ 0 − 3ϵ sin−4 θ 0 cos θ 0 ) − (sin θ 0 + ϵ cos θ 0 ) (7.39)
R cos θ 0 R
After some simplification and ignoring terms in ϵ 2 , we obtain:
д 1 + 3 cos2 θ 0

ϵÜ = − ϵ (7.40)
R cos θ 0
This shows thatthe pendulum will oscillate in a simple harmonic motion, with angular frequency approxi-
д 1 + 3 cos2 θ 0
mately ω 2 = , about the equilibrium position θ 0 .
R cos θ 0
15
Chapter 8
Two-Body Central-Force Problems
8.13. Two particles whose reduced mass is µ interact via a potential energy U = 12 kr 2 , where r is the distance
between them. (a) Make a sketch showing U (r ), the centrifugal potential energy Ucf (r ), and the effective
potential energy Ueff (r ). (Treat the angular momentum l as a known, fixed constant.) (b) Find the “equilib-
rium” separation r 0 , the distance at which the two particles can circle each other with constant r . [Hint:
This requires that dUeff /dr be zero.] (c) By making a Taylor expansion of Ueff (r ) about the equilibrium point
r 0 and neglecting all terms in (r − r 0 )3 and higher, find the frequency of small oscillations about the circular
orbit if the particles are disturbed a little from the separation r 0 .
Solution. (a) A sketch of all three graphs are shown in the figure below.
Figure 8.1: Graphs of the potentials U (r ), Ucf (r ) and Ueff (r )
16
Solution Manual
We can write the potentials as such:
1
U (r ) = kr 2 (8.1)
2
l2
Ucf (r ) = (8.2)
2µr 2
1 l2
Ueff (r ) = U (r ) + Ucf (r ) = kr 2 + (8.3)
2 2µr 2
dUeff
(b) The equilibrium separation by setting dr = 0. We obtain:
dUeff l2
= kr 0 − 3 = 0
dr µr 0
2 1/4
l
⇒ r0 = (8.4)
µk
(c) Let us define r = r 0 + ϵ. The Taylor expansion can be performed about the equilibrium r 0 , noting that
the first derivative was calculated before, and is equal to 0 at r 0 :
d2Ueff (r )

2
Ueff (r ) = Ueff (r 0 ) + ϵ +...
dr 2 r =r 0 +ϵ
1 l2 3l 2

≈ kr 02 + + ϵ 2
k +
2 2µr 02 µ(r 0 + ϵ)4
1 2 l2 3l 2 −4

2
≈ kr 0 + +ϵ k + (r + ϵ r 0 )−5
2 2µr 02 µ 0
1 l2 3l 2 3l 2

≈ kr 02 + 2
+ ϵ 2
k + 4
+ ϵ3 5 (8.5)
2 2µr 0 µr 0 µr 0
For small oscillations ϵ, we can ignore terms of ϵ 3 and above, and note that the effective potential takes
on the form of a harmonic oscillator potential (e.g. one of an elastic spring). That is, it takes the form of
U (x) = U (0) + 12 k 0x 2 , and the frequency of oscillation for such a system is ω = k 0/m. Making such an
p
observation, we can deduce the ‘spring constant’ k 0 and oscillation frequency ω:
3l 2

k 0 = 2 k + 4 = 2k + 6k = 8k
µr 0
s s
k0 8k
ω= = (8.6)
µ µ
where we have used the result from Eq. (8.4) to rewrite r 0 in terms of k, the central force constant. Do not
confuse this with the spring constant k 0 used in the analogy earlier.
8.34. Suppose that we decide to send a spacecraft to Neptune, using the simple transfer described in Example
8.6 (c.f. textbook pg 318). The craft starts in a circular orbit close to the Earth (radius 1 AU) and is to end up
17
Solution Manual
Figure 8.2: Transfer orbit for Problem 8.8.34
in a circular orbit near Neptune (radius about 30 AU). Use Kepler’s third law to show that the transfer will
take about 31 years.
Solution. The transfer orbit in question is depicted in Figure 8.2, with the radius of the Earth orbit as r E and
the radius of the Neptune orbit as r N .
a 3 GM
Kepler’s third law is stated as: = .
τ3 4π 2
We know (from the figure) that the sum of the radii equals twice the semi-major axis a of the elliptical orbit.
Explicitly, we have
r E + r N = 2a ⇒ a = 15.5 AU (8.7)
And so to get from P to P 0:

r
1 1 4π 2a 3
Time of transfer = τ =
2 2 GM
= 1.926 × 109 s ' 31 years (shown) (8.8)
Note that I used the following constants: 1 AU = 1.496 × 1011 m, G = 6.67 × 10−11 Nm2 kg−2 , and M Sun =
1.989 × 1030 kg.
18
Chapter 9
Mechanics in Non-inertial Frames
19
Chapter 10
Rotational Motion of Rigid Bodies
10.1. The result that m α rα0 = 0, can be paraphrased to say that the position vector of the CM relative to the
Í
CM is zero, and, in this form, is nearly obvious. Nevertheless, to be sure that you understand the result,
prove it by solving rα = R + rα0 for rα0 and substituting into the sum concerned.
Solution. We start with the equation given in the question, and substituting into the sum as suggested:
rα = R + rα0 ⇒ rα0 = rα − R (10.1)

Õ Õ Õ
m α rα0 = (m α rα ) − R mα (10.2)
where the sum is taken implicitly over α. We also have the following expression for the position of the
center of mass:
1 Õ
R= (m α rα ) (10.3)
M
Using this relation, and noting that the term m α is simply the total mass M, Equation (10.2) becomes:
Í
Õ
m α rα0 = RM − RM = 0 (10.4)
We have thus shown that the position of the CM relative to itself is zero.
10.3. Five equal point masses are placed at the five corners of a square pyramid whose square base is centered
on the origin in the xy plane, with side L, and whose apex is on the z axis at a height H above the origin.
Find the CM of the five mass system.
Solution. The (x, y, z) coordinates of the 4 corners are (L/2, L/2, 0), (L/2, −L/2, 0), (−L/2, L/2, 0), (−L/2, −L/2, 0),
while the coordinates of the apex is given by (0, 0, H ). Let m be the mass of one point mass. To calculate the
20
Solution Manual
CM of the system, we perform the following:
m xi
1 L L L L
Í
i
x CM = =
+ − − =0 (10.5)
5m 5 2 2 2 2
m i yi 1 L L L L
Í
yCM = = + − − =0 (10.6)
5m 5 2 2 2 2
m i zi H
Í
z CM = = (10.7)
5m 5
The coordinates of the CM is given by (0, 0, H /5).
10.4. The calculation of centers of mass or moments of inertia usually involves doing an integral, most often a
volume integral, and such integrals are often best done in spherical polar coordinates. Prove that:
∫ ∫ ∫ ∫
2
dV f (r) = dr r dθ sin θ dϕ f (r , θ, ϕ).
[Think about the small volume dV enclosed between r and r + dr , θ and θ + dθ , and ϕ and ϕ + dϕ.] If the
volume integral on the left runs over all space, what are the limits of the three integrals on the right?
Figure 10.1: Taken from http://astro-learned.blogspot.sg/
Solution. Referring to the figure, a small volume element dV can be approximated to a cuboid with volume:
dV = (dr )(rdθ )(r sin θ dϕ) = r 2 sin θ drdθdϕ (10.8)
Substituting this into the integral, taking note of the independence (or dependence) of the integrand with
respect to the 3 variables, we obtain, trivially, the result:
∫ ∫ ∫ ∫
dV f (r) = dr r 2 dθ sin θ dϕ f (r , θ, ϕ). (10.9)
21
Solution Manual
Another approach to this question is by considering the Jacobian. We express the Cartesian coordinates in
terms of the spherical polar coordinates:
x = r sin θ cos ϕ, y = r sin θ sin ϕ, z = r cos θ . (10.10)
We then have the volume element as dV = dxdydz = J (r , θ, ϕ) drdθdϕ, with J (r , θ, ϕ) being:

∂x ∂x ∂x

∂r ∂ϕ ∂θ

sin θ cos ϕ −r sin θ sin ϕ r cos θ cos ϕ
∂y ∂y ∂y

J (r , θ, ϕ) = = sin θ sin ϕ r sin θ cos ϕ r cos θ sin ϕ

∂r ∂ϕ ∂θ
cos θ

∂z ∂z ∂z 0 −r sin θ

∂r ∂ϕ ∂θ

= r 2 sin θ (10.11)
This yields us the same result as above. The integration limits for a volume integral over all space are:
0 ≤ r ≤ ∞, 0 ≤ ϕ ≤ 2π , 0 ≤ θ ≤ π .
10.5. A uniform solid hemisphere of radius R has its flat base in the xy plane, with its center at the origin. Use
the result of Problem 10.4 to find the center of mass.
Solution. Due to the inherent spherical symmetry of the object, one might expect the x and y coordinates
of the center of mass to be zero. Nevertheless, they can be computed by:
x dm y dm
∫ ∫
x CM = ∫ =0 yCM = ∫ =0 (10.12)
dm dm
Let ρ be the density of the hemisphere. The z coordinate can be computed in a similar fashion:
z dm ρ r cos θ dV
∫ ∫
z CM = ∫ =
dm ρ dV
∫
2 dr π /2 sin θ cos θ dθ 2π dϕ
∫R
r r
∫ ∫
·
= 0 ∫R 0 0
r 2 dr π /2 sin θ dθ 2π dϕ
∫ ∫
0 0 0
(R 4 /4) (1/2) (2π ) 3R
= = (10.13)
(R 3 /3) (1) (2π ) 8
10.6. (a) Find the CM of a uniform hemispherical shell of inner and outer radii a and b and mass M positioned
as in Problem 10.5. (b) What becomes of your answer when a = 0? (c) What if b → a?
22
Solution Manual
Solution. (a) Once again, by symmetry we have x CM = yCM = 0. We set up a similar integral to compute
z CM , but with modified limits:
z dm ρ r cos θdV
∫ ∫
z CM = ∫ =
dm ρ dV
∫
∫b ∫ π /2 ∫ 2π
a
r · r 2 dr 0 sin θ cos θ dθ 0 dϕ
= ∫b ∫ π /2 ∫ 2π
a
r 2 dr 0 sin θ dθ 0 dϕ
1 4 4
4 (b − a ) (1/2) (2π )
= 1 3 3
3 (b − a ) (1) (2π )
3 b − a4
4
= (10.14)
8 b 3 − a3
(b) Note that when a = 0, we are simply computing the CM of a uniform hemisphere (as in Problem 10.4),
with b = R. It is then no surprise that our answer here is 38 b, which corresponds with our answer from the
previous problem.
(c) When b → a, the hemispherical shell becomes infinitely thin, with radius a. We would expect that z CM
approaches a/2. Now let us compute it. Note that here, a is simply a number, and so we are only treating b
as a variable as it tends towards a. Since we have the limit of [0]/[0], we can use L’Hopital’s rule:
3 b 4 − a4

z CM = lim
b→a 8 b 3 − a 3
3
3 4b
= lim
8 b→a 3b 2
3 4a 3 a

= = (10.15)
8 3a 2 2
10.8. A uniform thin wire lies along the y axis between y = ±L/2. It is now bent toward the left into an arc of a
circle with radius R, leaving the midpoint at the origin and tangent to the y axis. Find the CM. Comment on
your answer for the cases that R → ∞ and that 2πR = L.
Solution. Credit: http://www.physics.umd.edu/courses/Phys410/gates/Phys410_Solution_07.pdf

10.9. The moment of inertia of a continuous mass distribution with density ϱ is obtained by converting the sum
Iz = m α ρ α2 into the volume integral ρ 2 dm = ρ 2 ϱ dV . Find the moment of inertia of a uniform circular
Í ∫ ∫
cylinder of radius R and mass M for rotation about its axis. Explain why the products of inertia are zero.
23
Solution Manual
Solution. For continuous masses, we have:

∫ ∫ h ∫ 2π ∫ R
Iz = ϱ ρ 2 dV = ϱ ρ 2 · ρ dρdθ dz
0 0 0
4
R
= ϱ (h)(2π ) (10.16)
4
M
Since the density ϱ = , we finally obtain the moment of inertia Iz :
πR 2h
MR 2
Iz = (10.17)
2
The products of inertia are zero because of cylindrical symmetry. More explicitly, we have, for example:
∫
I xy = ϱ −xy dV
∫ h ∫ R ∫ 2π
= −ϱ dz ρ 2 · ρdρ sin θ cos θ dθ
0 0 0
| {z }
=0
=0
10.10. (a) A thin uniform rod of mass M and length L lies on the x axis with one end at the origin. Find its
moment of inertia for rotation about the z axis. (b) What if the rod’s center is at the origin?
Solution. (a) The linear mass density of the (uniform) rod is µ = M/L. When the rotation is about the z axis,
each infinitesimal mass element dm of the rod is a distance x away from the rotation axis. So the moment
of inertia (about the end of the rod) can then be found easily:
L L
M
∫ ∫
I= x 2 dm = x2 dx (10.18)
0 0 L
1
= ML2 (10.19)
3
(b) If the rotation axis passes through the rod’s center, then the integration limits change appropriately to
yield:
L/2 L/2
M
∫ ∫
I0 = x 2 dm = x2 dx (10.20)
−L/2 −L/2 L
1
= ML2 (10.21)
12
which is precisely that predicted by the parallel-axis theorem as well.
24
Solution Manual
10.15. (a) Write down the integral (as in Problem 10.9) for the moment of inertia of a uniform cube of side a
and mass M, rotating about an edge, and show that it is equal to 23 Ma 2 . (b) If I balance the cube on an
edge in unstable equilibrium on a rough table, it will eventually topple and rotate until it hits the table. By
considering the energy of the cube, find its angular velocity just before it hits the table. (Assume the edge
does not slide on the table.)
Solution. (a) The details of this integral can be found in the textbook’s Example 10.2. I shall present the final
answer here without proof:
∫ a,a,a
y2 + z2 −xy −xz
I =ρ x 2 + z2 −yz ®® dV
© ª
−yx
0,0,0

« −zx −zy x 2 + y2¬
2 5
a − 14 a 5 − 14 a 5
© 3
M 1 5 2 5
ª
= 3 − 4 a 3a − 14 a 5 ®
®
a ®
1 5
«− 4 a − 14 a 5 2 5
3 a ¬
8 −3 −3
Ma 2 ©
= (10.22)
ª
−3 8 −3®®
12
«−3 −3 8¬
Hence we can directly read off the moment of inertia about an edge: I = 23 Ma 2 .
(b) When the cube is resting stably on the table, the height of the CM above the table is a/2. When it is held
at an unstable equilibrium (assuming balanced on an edge at a 45◦ angle), the height of the CM above the
a a
table is = √ . Hence we can calculate the change in the potential energy of the cube as it falls:
2 tan 45◦ 2
a a 1 1

∆P .E. = Mд √ − = Mдa √ − (10.23)
2 2 2 2
1 2
The corresponding gain in kinetic energy is ∆K .E. = Iω , where ω is the angular velocity just before it
2
hits the table and I is the moment of inertia about an edge (calculated in (a)). Equating the two energies
together, and solving for ω:
3д √
ω2 = ( 2 − 1) (10.24)
2a
10.23. Consider a rigid plane body or "lamina", such as a flat piece of sheet metal, rotating about a point O in
the body. If we choose axes so that the lamina lies in the xy plane, which elements of the inertia tensor I
are automatically zero? Prove that Izz = I x x + Iyy .
25
Solution Manual
Solution. If the lamina is on the xy plane, then z = 0 automatically. Hence, products of inertia containing
z is thus automatically zero. Explicitly, I x z = Izx = Izy = Iyz = 0. We will not be able to deduce anything
about Iyx and I xy .
Consider:
∫
Izz = ρ x 2 + y 2dV
∫ ∫
=ρ x 2 + z 2dV + ρ y 2 + z 2dV
= Iyy + I x x (10.25)
We can perform the second step because z 2 = 0.
10.25. (a) Find all nine elements of the moment of inertia tensor with respect to the CM of a uniform cuboid (a
rectangular brick shape) whose sides are 2a, 2b and 2c in the x, y, z directions and whose mass is M. Explain
clearly why you could write down the off-diagonal elements without doing any integration. (b) Combine
the results of part (a) and Problem 10.24 to find the moment of inertia tensor of the same cuboid with respect
to the corner A at (a, b, c). (c) What is the angular momentum about A if the cuboid is spinning with angular
velocity ω around the edge through A and parallel to the x axis?
Solution. (a)
∭c, b, a y 2 + z 2 −xy −xz

I =ρ x2 + z2 −yz ®® dV (10.26)
© ª
−yx

−c,−b,−a « −zx −zy x 2 + y2¬
Identify the density to be ρ = 8abc .

M
We compute the diagonal elements first.
∫ c ∫ b ∫ a
Ix x = ρ y 2 + z 2 dxdydz
−c −b −a
M
∫ c∫ b
= 2a(y 2 + z 2 ) dydz
8abc −c −b
∫ c 3
M 2b
= + 2bz 2 dz
4bc −c 3
M 2b 2c 2c 3 M

= + = (b 2 + c 2 )
2c 3 3 3
M 2
Iyy = · · · = (a + c 2 )
3
M 2
Izz = · · · = (a + b 2 )
3
where the other two moments of inertia can be directly calculated, or inferred from I x x (due to the ’symmetry’
present).
26
Solution Manual
The off-diagonal elements are zero because: −xy, −yz, −xz etc. are all odd functions in x, y, z. Integrating
from −a to a of x, for example, will yield 0.
(b) We shift the ‘origin’ by d = (a, b, c). Using the results from Problem 10.24, the modified moment of
inertia is given by:
b2 + c2 0 0 b2 + c2 −ab −ac
M
I0 = a2 + c2 0 ®® + M −ba a2 + c2
© ª © ª
0 −bc ®®
3
« 0 0 a2 + b 2 ¬ « −ca −cb a2 + b 2 ¬
4b 2 + 4c 2 −3ab −3ac
M ©
= 4a 2 4c 2 (10.27)
ª
−3ba + −3bc ®®
3
« −3ca −3cb 4a 2 + 4b 2 ¬
ω
(c) If the angular velocity ω is parallel to the x axis, we can write ω = 0 ®®. The angular momentum is then
© ª
«0¬
given as the matrix multiplication:
4(b 2 + c 2 )
Mω ©
L = Iω = (10.28)
ª
−3ba ®®
3
« −3ca ¬
10.26. (a) Prove that in cylindrical polar coordinates, a volume integral takes the form
∫ ∫ ∫ ∫
dV f (r) = ρdρ dϕ dz f (ρ, ϕ, z).
(b) Show that the moment of inertia of a uniform solid cone pivoted at its tip and rotating about its axis is
given by the integral:
∫ ∫ h ∫ 2π ∫ r
2
Izz = ϱ dV ρ = ϱ dz dϕ ρdρ ρ 2 ,
V 0 0 0
3 2
explaining clearly the limits of integration. Show that the integral evaluates to 10 MR . (c) Prove also that
3 2
Ix x = 20 M(R + 4h 2 ).
Solution. (a) Using cylindrical polar coordinates, we have the following coordinate transformations: x =
ρ cos ϕ, y = ρ sin ϕ. Hence consider the Jacobian for a small volume element:
∂x ∂x ∂x

∂ρ ∂ϕ ∂z

cos ϕ −ρ sin ϕ 0
∂y ∂y ∂y

J (ρ, ϕ, z) = = sin ϕ ρ cos ϕ 0 = ρ (10.29)

∂ρ ∂ϕ ∂z
0

∂z ∂z ∂z 0 1

∂ρ ∂ϕ ∂z

And so we have dV = ρ dρdϕ dz, which then yields us the integral as required.
27
Solution Manual
(b) If we were to use cylindrical coordinates, Izz will be calculated as:

∫ ∫
Izz = ϱ x 2 + y 2 dV = ϱ ρ 2 dV
V V
Utilising the result from part (a), we then have:

∫ h ∫ 2π ∫ r
Izz = ϱ dz dϕ ρ 3 dρ (10.30)
0 0 0
M
where we have the density: ϱ = . The volume could have been obtained by considering the triple
1/3 πR 2h
integral dV , but here we quote a direct result.
∭
The limits of integration are as such:
• h is the height of the cone.

• 2π is taken due to the cylindrical symmetry.
z
• r = R · is derived from considering a set of similar triangles in Figure 10.2, with R being the radius
h
of the base of the cone.
Let us now evaluate this integral.
2π
h
R 4z 4
∫ ∫
Izz = ϱ dz dϕ
0 0 4h 4
πR 4z 4
h
∫
=ϱ dz
0 2h 4
3M πR h 4
= ·
πR 2h 10
3 2
= MR (shown) (10.31)
10
(c) Using the same notations as before, I x x is calculated as:

∫ ∫
Ix x = ϱ y 2 + z 2 dV = ϱ ρ 2 sin2 ϕ + z 2 dV
V V
∫ h ∫ 2π ∫ r
=ϱ dz dϕ ρ 3 sin2 ϕ + ρz 2 dρ
0 0 0
2π 4 4
h
Rz R 2z 4
∫ ∫
=ϱ dz sin2 ϕ + dϕ
0 0 4h 4 2h 2
h
πR 4z 4 πR 2z 4
∫
=ϱ + dz
0 4h 4 h2
3
= M(R 2 + 4h 2 ) (shown)
20
28
Solution Manual
Figure 10.2: Similar triangles in a cone for Problems 10.26 and 10.27.
10.27. Find the inertia tensor for a uniform, thin hollow cone, such as an ice-cream cone, of mass M, height h,
and base radius R, spinning about its pointed end.
Solution. We approach the problem similar to Problem 10.26, and first calculate the density ϱ = M/A, where
A is the curved surface area of the cone.
The area element on the surface of the cone dA is given by:
dA = (r dθ )(ds) (10.32)
where ds is the infinitesimal slanted length, denoted in Figure 10.2. Then, integrating over azimuthal angles
θ and the entire slanted length S:
∫ ∫ S ∫ 2π
A= dA = r dθ ds (10.33)
0 0
= πRS (10.34)
M
⇒ ϱ= (10.35)
πRS
where the relation r = RS s was substituted (obtained by considering a pair of similar triangles in Figure 10.2).
The inertia tensor is then given by:
∬ ©y 2 + z 2 −xy −xz
I =ϱ x2 + z2 −yz ®® dA
ª
−yx

« −zx −zy x 2 + y2¬
∫ S∫ 2π
r 2 sin2 θ + z 2 −r 2 sin θ cos θ −rz cos θ
=ϱ −r 2 sin θ cos θ r 2 cos2 θ + z 2 −rz sin θ ®® r dθ ds (10.36)
© ª
0 0

« −rz cos θ −rz sin θ r2 ¬
29
Solution Manual
Note that off-diagonal elements equal to zero because

∫ 2π ∫ 2π ∫ 2π
sin θ dθ = cos θ dθ = sin θ cos θ dθ = 0
0 0 0
∫ 2π ∫ 2π
2
Using the fact that sin θ dθ = cos2 θ dθ = π , and the relation z = S s,
h
we have the diagonal
0 0
components of the inertia tensor to be:
S
R3 3 Rh 2 3
∫
I x x = Iyy = ϱ π 3 s + 2π 3 s ds
0 S S
M
= (R 2 + 2h 2 ) (10.37a)
4
S
R3 3
∫
Izz = ϱ 2π s ds
0 S3
MR 2
= (10.37b)
2
Finally, we obtain the inertia tensor for a hollow cone to be:
R 2 + 2h 2 0 0
M ©
I= R2 2h 2 (10.38)
ª
0 + 0 ®®
4
« 0 0 2R 2 ¬
10.35. A rigid body consists of three masses fastened as follows: m at (a, 0, 0), 2m at (0, a, a) and 3m at (0, a, −a).
(a) Find the inertia tensor I. (b) Find the principal moments and a set of orthogonal principle axes.
Solution. (a) The inertia tensor of a rigid body with discrete masses is given by:
y 2 + zi2 −x i yi −x i zi
Õ © i
I= mi −yi x i x i2 + zi2
ª
−yi zi ®®
i
« −zi x i −zi yi x i2 + yi2 ¬
m(0) + 2m(2a 2 ) + 3m(2a 2 ) 0 0

= 2 2 2 −2m(a ) − 3m(−a 2 ) ®®
2
© ª
0 m(a ) + 2m(a ) + 3m(a )
« 0 −2m(a 2 ) − 3m(−a 2 ) m(a 2 ) + 2m(a 2 ) + 3m(a 2 )¬
10 0 0
= ma 2 0 6 1®®
© ª
« 0 1 6¬
(b) We have to find directions of ω such that Iω = λω, where λ is the moment of inertia about the principal
axes. This eigenvalue problem has a non-trivial solution when det(I − λ1) = 0.
30
Solution Manual
1
Firstly, by inspection, we know that ω1 = 0 is an eigenvector, or equivalently, one of the principal axes
0
directions, with λ 1 = 10ma 2 . To check that this is true, one only needs to substitute ω1 into Iω, and verify
that one obtains λ 1ω.
For the other two solutions, we just need to solve:

6ma 2 − λ ma 2

=0

ma 2 2
6ma − λ

⇒ (6ma 2 − λ2 )2 − (ma 2 )2 = 0
And so the eigenvalues are:
λ 2 = 5ma 2 and λ 3 = 7ma 2
When λ 2 = 5ma 2 , we have:

! ! !
1 1 ωy 0
ma 2 =
1 1 ωz 0
⇒ ωy = 1, ωz = −1
Normalizing, we obtain
0
1 © ª
ω2 = √ 1 ®® (10.39)
2
«−1¬
Then when λ 3 = 7ma 2 , we have, instead:

! ! !
2 −1 1 ωy 0
ma =
1 −1 ωz 0
⇒ ωy = 1, ωz = 1
0
1 © ª®
ω3 = √ 1® (10.40)
2
«1¬
Summarizing, with the set of orthogonal principle axes {ω1 , ω2 , ω3 }, the inertia tensor consisting of the
principal moments is given by:
10ma 2 0 0
I 0 = 0 5ma 2 (10.41)
© ª
0 ®®
« 0 0 7ma 2 ¬
31
Solution Manual
10.36. A rigid body consists of three equal masses (m) fastened at the positions (a, 0, 0), (0, a, 2a), and (0, 2a, a).
(a) Find the inertia tensor I . (b) Find the principal moments and a set of orthogonal principal axes.
Solution. (a) We take reference from the previous problem (Problem 10.35). A similar, trivial calculation can
be performed to obtain the inertia tensor as:
y 2 + zi2 −x i yi −x i zi
Õ © i
I= mi −yi x i x i2 + zi2
ª
−yi zi ®®
i
« −zi x i −zi yi x i2 + yi2 ¬
10 0 0
2
= ma 0
© ª
6 −4®®
«0 −4 6 ¬
(b) Similar to the previous problem, we have to find directions of ω (principal axes), such that we have
Iω = λω, with λ being the moment of inertia about each principal axis. This eigenvalue equation has
non-trivial solutions if and only if det(I − λ1) = 0.
1
From the form of the inertia tensor above, we observe by inspection that the first eigenvector is ω 1 = 0 ,
0
with an eigenvalue of λ 1 = 10ma 2 .
For the other 2 axes, we consider the determinant:

6ma 2 − λ −4ma 2
=0

−4ma 2 6ma 2 − λ

⇒ (6ma 2 − λ2 )2 − (4ma 2 )2 = 0
λ 2 = 10ma 2 and λ 3 = 2ma 2
When λ 2 = 10ma 2 , we have:

! ! !
2 −4 −4 ωy 0
ma =
−4 −4 ωz 0
⇒ ωy = 1, ωz = −1
0
1 © ª
ω 2 = √ 1 ®® (10.42)
2
«−1¬
32
Solution Manual
(0, 1, 0)
dy
(0, 0, 0) (1, 0, 0)
Figure 10.3: Metal Triangle for Problem 10.37
Then when λ 3 = 2ma 2 , we have, instead:

! ! !
4 −4 ωy 0
ma 2 =
−4 4 ωz 0
⇒ ωy = 1, ωz = 1
0
1 © ª
ω 3 = √ 1®® (10.43)
2
«1¬
10ma 2 0 0
I = 0 10ma 2 (10.44)
0
© ª
0 ®®
« 0 0 2ma 2 ¬
10.37. A thin, flat, uniform metal triangle lies in the xy plane with its corners at (1, 0, 0), (0, 1, 0) and the origin.
Its surface density (mass/area) is σ = 24. (a) Find the triangle’s inertia tensor I . (b) What are its principal
moments and the corresponding axes?
Solution. (a) The inertia tensor is given by:
∫ ©y 2 + z 2 −xy −xz
I =σ x2 + z2 −yz ®® dA
ª
−yx

A
« −zx −zy x 2 + y2¬
∫ © y2 −xy 0
(z=0)
= σ x2 ® dA (10.45)
ª
−yx
0 ®
A
« 0 0 x 2 + y2¬
We shall compute the elements of the inertia tensor component wise. Take reference from Figure 10.3,
33
Solution Manual
noting that the hypotenuse is characterized as: y = 1 − x:

∫ 1 ∫ 1−y
Ix x = σ y 2 dx dy
0 0
∫ 1
2
=σ y (1 − y) dy = 2 (10.46)
0
∫ 1 ∫ 1−x
Iyy =σ x 2 dy dx
0 0
∫ 1
2
=σ x (1 − x) dx = 2 (10.47)
0
∫ 1 ∫ 1−y
Izz = σ x 2 + y 2 dx dy
0 0
= I x x + Iyy = 4 (c.f. Problem 10.23) (10.48)
∫ 1 ∫ 1−y
I xy = Iyx =σ −xy dx dy
0 0
∫ 1
(1 − y)2
=σ −y dy = −1 (10.49)
0 2
Putting all the numbers together, we get the inertia tensor to be:
2 −1 0
I = −1 2 (10.50)
© ª
0®®
«0 0 4¬
(b) For the set of principal axes {ω1 , ω2 , ω3 }, we necessarily have Iω i = λi ω i for i = 1, 2, 3. Note that λi is
the moment of inertia about the principal axis in the direction of ωi . The eigenvalue equation has non-trivial
solutions if and only if det(I − λ1)= 0.
0
By inspection, we have that ω3 = 0 is an eigenvector, with corresponding eigenvalue/ moment of inertia
1
λ 3 = 4. Verify this by ensuring the eigenvalue equation stated above is satisfied for this choice of ω3 .
As for the remaining two principal axes, we shall consider:

2 − λ −1
=0

2 − λ

−1
⇒ (2 − λ)2 − 1 = 0
λ1 = 1 and λ2 = 3
34
Solution Manual
When λ 1 = 1, we have:
! ! !
1 −1 ωx 0
=
−1 1 ωy 0
⇒ ωx = 1, ωy = 1
1
1 © ª®
ω1 = √ 1® (10.51)
2
«0¬
Then when λ 2 = 3, we have:

! ! !
−1 −1 ωx 0
=
−1 −1 ωy 0
⇒ ωx = 1, ωy = −1
1
1 © ª
ω2 = √ −1®® (10.52)
2
«0¬
1 0 0
I 0 = 0 (10.53)
© ª
3 0®®
«0 0 4¬
35
Chapter 11
Coupled Oscillators and Normal Modes
11.5. (a) Find the normal frequencies, ω1 and ω2 , for the two carts shown in Figure 11.1, assuming that m 1 = m 2
and k 1 = k 2 . (b) Find and describe the motion for each of the normal modes in turn.
Solution. (a) Let m 1 = m 2 = m, k 1 = k 2 = k, and rightwards be positive. We can write down the forces
acting on each cart:
F 1 = k 2x 2 + (−k 1 − k 2 )x 1 = kx 2 − 2kx 1 (11.1)

F 2 = −k 2x 2 + k 2x 1 = −kx 2 + kx 1 (11.2)
Using Newton’s second law, we express the above two equations as a single matrix equation:
! ! !
mxÜ1 −2k k x1
= (11.3)
mxÜ2 k −k x2
We look for the normal mode solutions that take the form x = Ae iωt , where x and A are column vectors.
Substituting into Eq. (11.3),
! ! !
2 A1 −2k k A1
−mω = (11.4)
A2 k −k A2
Figure 11.1: Two carts for Problems 11.5 and 11.6
36
Solution Manual
For non-trivial solutions, we have:

−2k + mω 2 k

=0

k 2
−k + mω

⇒ k 2 − 3kmω 2 + (mω 2 )2 = 0
We obtain the eigenfrequencies to be:

√
(3 ± 5)
ω2 = ω0 , (11.5)
2
where ω0 = k/m. This works out to be ω 1 = 0.62ω0 and ω 2 = 1.62ω0 approximately.

(b) When ω = ω 1 , we have:
√
−1− 5
! ! !
2 k √
k A1
=
0
−1− 5
k 2 k
A2
0
√
1+ 5
⇒ A1 = 1 and A2 = (11.6)
2
And so we have the amplitudes (the reader can normalize the expression if he so wishes) to be:
! !
A1 1
= √
1+ 5
(11.7)
A2 2
This means that both masses oscillate in phase, with the amplitude of mass 2 to be approximately 1.62 times
that of mass 1.
When ω = ω2 , we then have:
√
−1+ 5
! ! !
2 k √
k A1
=
0
−1+ 5
k 2 k
A2
0
√
1− 5
⇒ A1 = 1 and A2 = (11.8)
2
The amplitudes, when written in vector form is:

! !
A1 1
= √
1− 5
(11.9)
A2 2
In this situation, both masses oscillate out of phase, with the amplitude of mass 2 being approximately 0.62
times that of mass 1.
11.6. Answer the same questions as in Problem 11.5 but for the case that m 1 = m 2 and k 1 = 3k 2 /2. (Write k 1 = 3k
and k 2 = 2k.) Explain the motion in the two normal modes.
37
Solution Manual
Solution. We use the same sign conventions as Problem 11.5, with m 1 = m 2 = m and k 1 = 3k, k 2 = 2k. The
forces on each mass is then:
F 1 = k 2x 2 + (−k 1 − k 2 )x 1 = 2kx 2 − 5kx 1 (11.10)

F 2 = −k 2x 2 + k 2x 1 = −2kx 2 + 2kx 1 (11.11)
We then use Newton’s second law to express the two equations above in a single matrix equation:
! ! !
mxÜ1 −5k 2k x1
= (11.12)
mxÜ2 2k −2k x2
Normal mode solutions: x = Ae iωt . Substituting into the above equation, we obtain:
! ! !
A1 −5k 2k A1
−mω 2 = (11.13)
A2 2k −2k A2
For non-trivial solutions, we have the determinant vanishing:

−5k + mω 2 2k
=0

2k −2k + mω 2

⇒ 6k 2 − 7kmω 2 + (mω 2 )2 = 0
We obtain the eigenfrequencies to be:
ω 2 = 6ω 0 or ω0 , (11.14)
√
where ω 0 = k/m. And so ω1 = 6ω0 and ω 2 = ω0 .
√
When ω1 = 6ω0 , we have:
! ! !
k 2k A1 0
=
2k 4k A2 0
⇒ A1 = 2 and A2 = −1 (11.15)
In vector form,
! !
A1 2
= (11.16)
A2 −1
With this normal mode, mass 1 and mass 2 are oscillating out of phase, with the amplitude of mass 1 being
twice that of mass 2.
38
Solution Manual
We do the same for ω2 = ω0 . We get:

! ! !
−4k 2k A1 0
=
2k −k A2 0
⇒ A1 = 1 and A2 = 2 (11.17)
In vector form,
! !
A1 1
= (11.18)
A2 2
With this normal mode, mass 1 and mass 2 are oscillating in phase, with the amplitude of mass 1 being half
that of mass 2.
11.7. The most general motion of the two carts of Section 11.2 is given by (11.21) (c.f. textbook pg 440), with the
constants A1 , A2 , δ 1 and δ 2 determined by the initial conditions. (a) Show that (11.21) can be rewritten as
" # " #
1 1
x(t) = (B 1 cos ω1t + C 1 sin ω1t) + (B 2 cos ω2t + C 2 sin ω 2t) .
1 −1
This form is usually a little more convenient for matching to given initial conditions. (b) If the carts are
released from rest at positions x 1 (0) = x 2 (0) = A, find the coefficients B 1 , B 2 , C 1 and C 2 and plot x 1 (t) and
x 2 (t). Take A = ω 1 = 1 and 0 ≤ t ≤ 30 for your plots. (c) Same as part (b) except that x 1 (0) = A but x 2 (0) = 0.
Solution. (a) Rewriting (11.21) from the textbook, making use of trigonometric identities:
" # " #
1 1
(11.21): x(t) = A1 cos(ω1t − δ 1 ) + A2 cos(ω2t − δ 2 )
1 −1
" # " #
1 1
= A1 (cos(ω1t) cos δ 1 + sin(ω 1t) sin δ 1 ) + A2 (cos(ω 2t) cos δ 2 + sin(ω 2t) sin δ 2 )
1 −1
Absorbing the time-independent terms (e.g. sin δ 1 , cos δ 2 etc.) into the arbitrary constants, we then obtain
the required expression:
" # " #
1 1
x(t) = (B 1 cos ω1t + C 1 sin ω 1t) + (B 2 cos ω2t + C 2 sin ω 2t) (shown)
1 −1
(b) Given the initial conditions x 1 (0) = x 2 (0) = A,

" # " # " #
A 1 1
= B1 + B2
A 1 −1
39
Solution Manual
Figure 11.2: Plots for x 1 (t ) and x 2 (t ) for Problem 11.7
Solving, we obtain
B 1 = A and B 2 = 0 (11.19)
We are also given that they are released from rest, so xÛ1 (0) = xÛ2 (0) = 0. So we have:
" # " # " #
0 1 1
= ω 1C 1 + ω2C 2
0 1 −1
Solving, we obtain
C1 = C2 = 0 (11.20)
Taking A = ω1 = 1, we have x 1 (t) = x 2 (t) = cos t. The plots can be found in Fig. 11.2.
(c) With the initial conditions x 1 (0) = A, x 2 (0) = 0, we can obtain the coefficients:
" # " # " #
A 1 1
= B1 + B2
0 1 −1
Solving, we obtain
B 1 = B 2 = A/2 (11.21)
The initial conditions for velocities are unchanged, so C 1 = C 2 = 0 from above. And so, taking A = ω1 = 1,
1
we have x 1 (t) = 2 cos t + 12 cos ω2t and x 2 (t) = 1
2 cos t − 21 cos ω2t. No information is given about ω 2 , so we
assume it is different from ω1 , and set it as an arbitrary number ω 2 = 1.7. The relevant plots can be found
in Figure 11.2.
11.8. Same as Problem 11.7 but in part (b) the carts are at their equilibrium positions at t = 0 and are kicked
away from each other, each with speed v 0 . In part (c), the carts start out at their equilibrium positions and
cart 2 has speed v 0 to the right but cart 1 has initial speed 0. Take v 0 = ω1 = 1 and 0 ≤ t ≤ 30 for your plots.
40
Solution Manual
Solution. We have (from previous problem) the expression for x(t):

" # " #
1 1
x(t) = (B 1 cos ω 1t + C 1 sin ω1t) + (B 2 cos ω2t + C 2 sin ω 2t)
1 −1
We still have the carts starting off at positions x 1 (0) = x 2 (0) = A = 1, but now xÛ1 (0) = v 0 = 1 and
xÛ2 (0) = −v 0 = −1. As from part (b) previously, we apply the initial conditions for position and yield B 1 = 1
and B 2 = 0. Now we consider the initial velocities:
" # " # " #
1 1 1
= ω1C 1 + ω2C 2
−1 1 −1
Taking ω1 = 1 as suggested, and solving simultaneously, we obtain
C1 = 0 and C 2 = 1/ω2 (11.22)
Combining all the information, we have the equations of motion:
1
x 1 (t) = cos t + sin ω 2t
ω2
(11.23)
1
x 2 (t) = cos t − sin ω 2t
ω2
For the plots, we use an arbitrary value of ω2 = 1.7, as before. See Fig. 11.3 for the plots.
For part (c), we have the initial conditions: positions x 1 (0) = A but x 2 (0) = 0, and velocities xÛ1 (0) = 0 and
xÛ2 (0) = v 0 = 1. Since the initial positions are unchanged, we have B 1 = B 2 = A/2 (See previous problem
part (c)). The velocity conditions give us:
" # " # " #
0 1 1
= ω1C 1 + ω2C 2
1 1 −1
Taking ω1 = 1 as suggested, and solving simultaneously, we obtain
C 1 = 1/2 and C 2 = −1/2ω2 (11.24)
Using the above coefficients, we have the equations of motion to be:
1 1

x 1 (t) = cos t + sin t + cos ω2t − sin ω2t
2 ω2
(11.25)
1 1

x 2 (t) = cos t + sin t − cos ω2t + sin ω2t
2 ω2
The plots can be found in Fig. 11.3.
41
Solution Manual
Figure 11.3: Plots for x 1 (t ) and x 2 (t ) for Problem 11.8
11.9. (a) Write down the equations of motion (11.2) for the equal-mass carts of Section 11.2 (c.f. textbook pg 418,
421) with three identical springs. Show that the change of variables to the normal coordinates ξ 1 = 12 (x 1 +x 2 )
and ξ 2 = 12 (x 1 − x 2 ) leads to uncoupled equations for ξ 1 and ξ 2 . (b) Solve for ξ 1 and ξ 2 and hence write down
the general solution for x 1 and x 2 .
Solution. (a) Given the carts have equal masses, and the springs are identical, we have: m 1 = m 2 = m and
k 1 = k 2 = k 3 = k. We can then write the equation of motion (from (11.2) of textbook) as:
mxÜ1 = −2kx 1 + kx 2
(11.26)
mxÜ2 = kx 1 − 2kx 2
Adding the two equations from Eq. (11.26), and making a change of variables with ξ 1 :
m(xÜ1 + xÜ2 ) = −k(x 1 + x 2 )

k
⇒ ξÜ1 = − ξ 1 (11.27)
m
Similarly, by subtracting one equation from the other in Eq. (11.26), and making a change of variables with
ξ2:
m(xÜ1 − xÜ2 ) = −3k(x 1 − x 2 )

3k
⇒ ξÜ2 = − ξ 2 (11.28)
m
The above two equations demonstrate that appropriate change of variables lead to uncoupled equations for
ξ 1 and ξ 2 .
42
Solution Manual
Figure 11.4: Two carts, Three springs for Problem 11.9
(b) The solutions for ξ 1 and ξ 2 are straightforward.
ξ 1 = A1 cos(ω0t + δ 1 )
√ (11.29)
ξ 2 = A2 cos 3 ω0t + δ 2
where ω0 = k/m and A1 , A2 , δ 1 , δ 2 are arbitrary constants.

p
We note that x 1 = ξ 1 + ξ 2 and x 2 = ξ 1 − ξ 2 , so the general solutions are:

√
x 1 = A1 cos(ω0t + δ 1 ) + A2 cos 3 ω 0t + δ 2
√ (11.30)
x 2 = A1 cos(ω0t + δ 1 ) − A2 cos 3 ω 0t + δ 2
11.10. (a) Write down the equations of motion corresponding to (11.2) for the equal mass carts of Section 11.2
with three identical springs (c.f. textbook pg 418,421), but with each cart subject to a linear resistive force
−bv (same coefficient b for both carts). (b) Show that if you change variables to the normal coordinates
1 1
ξ1 = 2 (x 1 + x 2 ) and ξ 2 = 2 (x 1 − x 2 ), the equations of motion for ξ 1 and ξ 2 are uncoupled. (c) Write
down the general solutions for the normal coordinates and hence for x 1 and x 2 . (Assume b is small, so
that the oscillations are underdamped.) (d) Find x 1 (t) and x 2 (t) for the initial conditions x 1 (0) = A and
x 2 (0) = v 1 (0) = v 2 (0) = 0, and plot them for 0 ≤ t ≤ 10π using the values A = k = m = 1, and b = 0.1.
Solution. (a) Let us have all the masses equal to m, and all the spring constants equal to k. We can simply
modify the equations in (11.2) from the textbook to include the effects of the resistive force.
mxÜ1 = −bxÛ1 − 2kx 1 + kx 2

(11.31)
mxÜ2 = −bxÛ2 + kx 1 − 2kx 2
(b) Let us consider the sum of the two equations from Eq. (11.31), and make the appropriate substitution
43
Solution Manual
with ξ 1 = 21 (x 1 + x 2 ) and its derivatives:
m(xÜ1 + xÜ2 ) = −b(xÛ1 + xÛ2 ) − k(x 1 + x 2 )

b k
ξÜ1 = − ξÛ1 − ξ 1 (11.32)
m m
Consider taking the difference instead, and using the change of variables with ξ 2 = 21 (x 1 − x 2 ):
m(xÜ1 − xÜ2 ) = −b(xÛ1 − xÛ2 ) − 3k(x 1 − x 2 )

b 3k
ξÜ2 = − ξÛ2 − ξ 2 (11.33)
m m
It is evident that the above two equations of motion are uncoupled.

(c) Before we solve the equations, it is helpful to relabel the constants with β = b/2m and ω02 = k/m. The
two equations then become:
ξÜ1 = −2β ξÛ1 − ω02 ξ 1

(11.34)
ξÜ2 = −2β ξÛ2 − 3ω 02 ξ 2
The equations in Eq. (11.34) take the form of that describing a damped oscillator. In general, for q
the differ-
ential equation yÜ + 2βyÛ + ω02y = 0, we have the solution y = A1e ω+ t + A2e ω− t , with ω± = −β ± i ω02 − β 2 .
One can verify this by substituting this into the differential equation and checking that it equals zero. We
apply this general solution to Eq. (11.34):
√ √
ω 02 −β 2 t ω02 −β 2 t
ξ 1 = A1e −β t +i + B 1e −β t −i
= e −β t (A1e iω1 t + B 1e −iω1 t ) (11.35)
√ 2 2 √ 2 2
ξ 2 = A2e −β t +i 3ω0 −β t + B 2e −β t −i 3ω0 −β t
= e −β t (A2e iω2 t + B 2e −iω2 t ) (11.36)
q q
where we have defined the real constants ω1 = ω02 − β 2 and ω 2 = 3ω02 − β 2 for conciseness; where A1 ,
A2 , B 1 and B 2 are all arbitrary constants.
The expressions for x 1 (t) and x 2 (t) are easily obtained by taking x 1 = ξ 1 + ξ 2 and x 2 = ξ 1 − ξ 2 .
(d) The given initial conditions imply that ξ 1 (0) = ξ 2 (0) = A/2, and ξÛ1 (0) = ξÛ2 (0) = 0. We can solve for the
arbitrary constants based on the initial conditions. I will demonstrate the calculation for A1 and B 1 :
A A
ξ 1 (0) = ⇒ A1 + B 1 = (11.37)
2 2
Aβ
ξÛ1 (0) = 0 ⇒ A1 − B 1 = − i (11.38)
2ω1
And so, we have:
A β A β
A1 = (1 − i) and B1 = (1 + i) (11.39)
4 ω1 4 ω1
44
Solution Manual
Figure 11.5: Plots of x 1 (t ) and x 2 (t ) for Problem 11.10.
Substitute this back into the expression for ξ 1 in Eq. (11.35) (and we find that the trigonometric expression
is more compact here):
A β A β

ξ 1 (t) = e−β t
(1 − i)e iω1 t + (1 + i)e −iω1 t
4 ω1 4 ω1
A −β t β

= e cos ω1t + sin ω1t (11.40)
2 ω1
We can perform similar calculations for A2 , B 2 and ξ 2 . The results are presented below:
A β
A2 = (1 − i)
4 ω2
A β
B 2 = (1 + i) (11.41)
4 ω2
A −β t β

ξ 2 (t) = e cos ω 2t + sin ω 2t
2 ω2
Once again, we have x 1 (t) and x 2 (t) by taking the sum and difference of ξ 1 and ξ 2 respectively. The plots
are shown in Fig. 11.5.
45
Chapter 12
Nonlinear Mechanics and Chaos
46
Chapter 13
Hamiltonian Mechanics
13.3. Consider the Atwood Machine of Fig. 13.1, but suppose that the pulley is a uniform disc of mass M and
radius R. Using x as your generalized coordinate, write down the Lagrangian, the generalized momentum
p, and the Hamiltonian H = pxÛ − L. Find Hamilton’s equations and use them to find the acceleration x.
Ü
x
y
m1
m2
Figure 13.1: The Atwood Machine
MR 2 xÛ
Solution. The moment of inertia of the disc is , and angular velocity of the disc is given by ω = . Since
2 R
the length of the string is fixed (say l), we can write the following equation to incorporate the constraint:
y + x + πR = l ⇒ y = −x + const (13.1)
47
Solution Manual
Now we can express the kinetic and potential energies in terms of a single coordinate x.
2
1 2 1 MR 2 xÛ
T = (m 1 + m 2 )xÛ +
2 2 2 R
1 1

= m 1 + m 2 + M xÛ 2 (13.2)
2 2
U = −m 2дy − m 1дx
= (m 2 − m 1 )дx + const (13.3)
The Lagrangian is therefore:
L =T −U
1 1

= m 1 + m 2 + M xÛ 2 + (m 1 − m 2 )дx + const (13.4)
2 2
The generalized momentum p can be written as such:
∂L 1

p= = m 1 + m 2 + M xÛ (13.5)
∂xÛ 2
The Hamiltonian is then given by:
H = pxÛ − L
!
p2 1 p2
= − − (m 1 − m 2 )дx − const
m 1 + m 2 + 12 M 2 m 1 + m 2 + 12 M
!
1 p2
= − (m 1 − m 2 )дx − const (13.6)
2 m 1 + m 2 + 12 M
We note that H is just the total energy of the system, that is, H = T + U . The Hamilton’s equations are
given by:
∂H
xÛ =
∂p
p
= (13.7)
m 1 + m 2 + 12 M
∂H
pÛ =
∂x
= (m 2 − m 1 )д (13.8)
48
Solution Manual
We can obtain the acceleration rather trivially from Hamilton’s equations (Eqs. (13.7) and (13.8)).
pÛ
xÜ =
m 1 + m 2 + 12 M
(m 2 − m 1 )д
= (13.9)
m 1 + m 2 + 12 M
13.5. A bead of mass m is threaded on a frictionless wire that is bent into a helix with cylindrical polar coordinates
(ρ, ϕ, z) satisfying z = cϕ and ρ = R, with c and R constants. The z axis points vertically up and gravity
vertically down. Using ϕ as your generalized coordinate, write down the kinetic and potential energies,
and hence the Hamiltonian H as a function of ϕ and its conjugate momentum p. Write down Hamilton’s
equations and solve for ϕÜ and hence z.
Ü Explain your result in terms of Newtonian mechanics and discuss
the special case that R = 0.
Solution. In cylindrical coordinates, we have, in general, the kinetic energy T to be:
1
T = m ρÛ2 + ρ 2ϕÛ2 + zÛ2 (13.10)
2
Substituting in the constraints z = cϕ and ρ = R, we obtain

1 2
2 Û2
T ϕ = m R ϕ + c ϕÛ
Û (13.11)
2
The potential energy is simply given by the gravitational potential energy:
U (ϕ) = mдz = mдcϕ (13.12)
We can write the Lagrangian L as:
1
L = T − U = m R 2 + c 2 ϕÛ2 − mдcϕ (13.13)

2
The next step is to find the canonical momentum p and substitute it into the general expression for the
49
Solution Manual
Hamiltonian.
∂L
p= = m R 2 + c 2 ϕÛ (13.14)

∂ϕÛ
H = pϕÛ − L (13.15)
p2 p2

= − − mдcϕ (13.16)
mR 2 (R 2 + c 2 ) 2m (R 2 + c 2 )
p2
= + mдcϕ (13.17)
2m (R 2 + c 2 )
Now, we turn to Hamilton’s equations by taking appropriate derivatives of the Hamiltonian.
∂H p
ϕÛ = = (13.18)
∂p m (R 2 + c 2 )
∂H
pÛ = − = −mдc (13.19)
∂ϕ
We are able to solve for ϕÜ and zÜ as such:
pÛ дc
ϕÜ = =− 2 (13.20)
m (R 2 + c 2 ) R + c2
дc 2
zÜ = c ϕÜ = − 2 (13.21)
R + c2
Let us unwrap the helix into a 2-D plane. When we go round the helix an angle of ∆ϕ = 2π , the vertical
height we have gained is z = 2πc, while the ’horizontal’ distance travelled is 2πR. That is, we have the
following:
−д −д sin α
2πc
α
2πR
Figure 13.2: Unwrapped helix into a 2-D plane
where the angle tan α is given by c/R. From this, we note that we can write Eq. (13.21) as:
zÜ = −д sin2 α (13.22)
An object sliding down a slope of angle α has a component of acceleration down the slope given by at anд =
д sin α. So the bead has a tangential acceleration at anд along the wire. The vertical component of this
acceleration (along the z axis) is zÜ = at anд sin α = д sin2 α. This is precisely the acceleration calculated in
Eq. (13.22).
If R = 0, then by Eq. (13.21), zÜ = −д, as expected, since α = π /2.
50
Solution Manual
13.6. In discussing the oscillation of a cart on the end of a spring, we almost always ignore the mass of the spring.
Set up the Hamiltonian H for a cart of mass m on a spring (force constant k) whose mass M is not negligible,
using the extension x of the spring as the generalized coordinate. Solve Hamilton’s equations and show
that the mass oscillates with angular frequency ω = k/(m + M/3). That is, the effect of the spring’s mass
p
is to add M/3 to m. (Assume that the spring’s mass is distributed uniformly and that it stretches uniformly.)
m
y(s)
s y(L) = x
ds
L
Solution. With reference to the figure, let L be the length of the spring, while ρ be the mass per unit length
of spring.
Consider an element A of spring of length ds and mass ρds, at a distance s from the fixed end. Let y(s) be
the displacement of the element A, and y(L) = x be the displacement of mass m at s = L.
Assume that the spring stretches uniformly, i.e. y(s) = xs/L, with y(0) = 0, y(L) = x. We can then express
the velocity and kinetic energy of element A as:
xs
Û
yÛ =
L
1 ρ xÛ 2s 2
TA = (ρ ds)yÛ2 = ds (13.23)
2 2L2
Thus, the kinetic energy of the spring is given by:
ρ xÛ 2 L 2
∫
Tspring = s ds
2L2 0
1 1
= ρ LxÛ 2 = M xÛ 2 (13.24)
6 6
where in the last line we have identified the total mass of the spring as M = ρL.
With the kinetic energy of the cart as Tcart = 12 mxÛ 2 , we have the total KE:
1 1
T = Tspring + Tcart = M xÛ 2 + mxÛ 2
6 2
1
= m effxÛ 2 (13.25)
2
where m eff = m + 31 M is the effective mass.
The potential energy can be identified to be U = 21 kx 2 , and thus the Lagrangian is:
1 1
L = m effxÛ 2 − kx 2 (13.26)
2 2
We use the Lagrangian to derive the momentum in terms of the velocities, and then use the inverse relation
51
Solution Manual
to obtain the Hamiltonian H (px , x), as such:
∂L
px = = m effxÛ (13.27)
∂xÛ
px2 1
H = px xÛ − L = + kx 2 (13.28)
2m eff 2
The Hamilton’s equations of motions give us:
∂H px ∂H
xÛ = = pÛx = − = −kx (13.29)
∂px m eff ∂x
Differentiating the first of the two equations with respect to time, and using the second equation, we obtain:
pÛx k
xÜ = =− x = −ω 2x (13.30)
m eff m eff
where ω 2 = k/m eff . More explicitly, this is the equation of simple harmonic motion with frequency
s s
k k
ω= = (13.31)
m eff m + M/3
13.17. Consider the mass confined to the surface of a cone described in Example 13.4 (page 533). We saw that
there are solutions for which the mass remains at the fixed height z = z 0 , with fixed angular velocity ϕÛ0 ,
say. (a) For any chosen value of pϕ , use (13.34) to get an equation that gives the corresponding value of the
height z 0 . (b) Use the equations of motion to show that this motion is stable. That is, show that if the orbit
has z = z 0 + ϵ, with ϵ small, then ϵ will oscillate about zero. (c) Show that the angular frequency of these
√
oscillations is ω = 3ϕÛ0 sin α, where α is the half angle of the cone (tan α = c where c is the constant in
ρ = cz). (d) Find the angle α for which the frequency of oscillation ω is equal to the orbital angular velocity
ϕÛ0 , and describe the motion in this case.
Solution. The two equations in (13.34) from the textbook (pg 534) are:
pz
zÛ = (13.32)
m(c 2 + 1)
pϕ2
pÛz = − mд (13.33)
mc 2z 3
The above equations were obtained from constructing the Lagrangian, and writing the Lagrange equations.
For completeness, the Hamiltonian of the system can be written as:
pz2 pϕ2
" #
1
H= + + mдz (13.34)
2m c 2 + 1 c 2z 2
52
Solution Manual
Figure 13.3: A mass m constrained to move on the surface of a cone. Taken from textbook pg 533.
(a) To maintain at a fixed height, we require zÛ = 0, which from Eq. (13.32), we infer that it is necessary that
pz = 0. And so pÛz = 0. From Eq. (13.33), we see that the height z 0 which corresponds to a particular value
of pϕ is:
! 1/3
pϕ2
z0 = (13.35)
m 2 c 2д
(b) From Eq. (13.32), we differentiate once with respect to time to get an equation of motion for z as:
pÛz
zÜ =
m(c 2 + 1)
pϕ2
!
1
= − mд (13.36)
m(c 2 + 1) mc 2z 3
We consider a perturbation of z about the equilibrium z 0 , that is, z = z 0 + ϵ, for small ϵ. We have the
following approximations:
zÜ = ϵÜ (13.37)
ϵ

z −3 = z 0−3 1 − 3 (13.38)
z0
Substituting these relations into Eq. (13.36), we obtain:
3 pϕ2
ϵÜ = − ϵ (13.39)
m(c 2 + 1) mc 2z 04
pϕ2
where we have used the fact that = mд (from Eq. (13.35) in an intermediate step. To simplify Eq. (13.39)
mc 2z 03
53
Solution Manual
further, we recall a relation between ϕÛ and pϕ derived from the Hamilton’s equations.
∂H pϕ
ϕÛ = = (13.40)
∂pϕ mc 2z 2
We use this to replace pϕ in Eq. (13.39):
3ϕÛ2c 2
ϵÜ = − ϵ (13.41)
c2 + 1
This takes on the form of a SHM equation. (shown)
(c) The angular frequency of the oscillation can be simply obtained from the SHM equation:
s
3ϕÛ2c 2 √ Û c
ω= = 3ϕ √ (13.42)
c2 + 1 2
c +1
Finally, we can note that the fraction above is none other than sin α, where α is defined in the following
√
way (c.f. Fig. 13.3): tan α = cz/z. Hence, we have shown that ω = 3ϕÛ0 sin α.
√
(d) In order for ω = ϕÛ0 , we require sin α = 1/ 3, or:
α = 35.26◦ (13.43)
In this case, the mass’ rotational frequency equals its oscillation frequency. As a result, its orbit is closed.
13.23. Consider the modified Atwood machine shown in Fig. 13.4. The two weights on the left have equal masses
m and are connected by a massless spring of force constant k. The weight on the right has mass M = 2m, and
the pulley is massless and frictionless. The coordinate x is the extension of the spring from its equilibrium
length; that is, the length of the spring is le + x where le is the equilibrium length (with all the weights
in position and M held stationary). (a) Show that the total potential energy (spring + gravitational) is just
U = 21 kx 2 (plus a constant that we can take to be zero). (b) Find the two momenta conjugate to x and y.
Solve for xÛ and y,
Û and write down the Hamiltonian. Show that the coordinate y is ignorable. (c) Write
down the four Hamilton equations and solve them for the following initial conditions: You hold the mass
M fixed with the whole system in equilibrium and y = y0 . Still holding M fixed, you pull the lower mass m
down a distance x 0 and at time t = 0 you let go of both masses. Describe the motion. In particular, find the
frequency with which x oscillates.
Solution. (a) Let string length be L. We take downwards to be positive, and the origin to be the centre of the
pulley wheel. We also assume the radius of the pulley to be 0, without any loss in generality. In addition,
let us define the original, unstretched length of the spring (without load) to be l 0 . That is, le = l 0 + mд/k,
where k is the spring constant.
54
Solution Manual
Figure 13.4: Modified Atwood machine for Problem 13.23
The two components of the potential energy can be written as:
Ugrav = −mдy − mд(y + le + x) − 2m(L − y)д

= −mдle − 2mLд − mдx (13.44)
1
Uspring = k(x + mд/k)2 (13.45)
2
Summing the two potential energies give us:
1 2 (mд)2

U = (−mдle − 2mLд − mдx) + kx + mдx +
2 2k
1
= kx 2 + const (shown) (13.46)
2
Note that we can ignore the constant, and take it to be equal to zero. Henceforth, the potential energy shall
be computed only with U = 21 kx 2 .
(b) We first complete the expression for the Lagrangian (L = T − U ) by considering the kinetic energy T of
the system:
1 1 1
T = myÛ2 + (2m)yÛ2 + m(yÛ + x)
Û2 (13.47)
2 2 2
Then, we can write the momentum conjugates as:
∂L ∂T
px = = = m(yÛ + x)
Û
∂xÛ ∂xÛ
(13.48)
∂L ∂T
py = = = 3myÛ + m(yÛ + x)
Û = m(4yÛ + x)
Û
∂yÛ ∂yÛ
55
Solution Manual
Solving the above equation for xÛ and y,

Û we find their expressions:
4px − py
xÛ =
3m
(13.49)
py − px
yÛ =
3m
The Hamiltonian can thus be written as:
H = xp
Û x + yp
Û y −L
1 1 1

= (py − px )2 + px2 + kx 2 (13.50)
2m 3 2
after some simplification. Note that this is simply the sum of the kinetic energies and potential energy. Thus
H represents the total energy of the system.
∂H
Since pÛy = − = 0, py is constant and thus y must be an ignorable coordinate. (shown)
∂y
(c) The four Hamilton equations are:
∂H ∂H
pÛx = − = −kx pÛy = − =0 (13.51)
∂x ∂y
∂H 4px − py ∂H py − px
xÛ = = yÛ = = (13.52)
∂px 3m ∂py 3m
noting that the expressions in Eq. (13.52) matches with those from Eq. (13.49).
To solve these equations, we shall use the following initial conditions: x(0) = x 0 , y(0) = y0 , x(0)
Û = y(0)
Û = 0.
Furthermore, with Eq. (13.48), we also have px (0) = py (0) = 0.
For the coordinate in x, we differentiate the expression for xÛ to get:
4pÛx − pÛy 4k
xÜ = = − x, (13.53)
3m 3m
where we have substituted the expressions for pÛx and pÛy from Eq. (13.51). One can easily recognise this
taking the form of a simple harmonic motion, where the position and velocity is given by:
r !
4k
x(t) = A cos t +δ (13.54)
3m
r r !
4k 4k
x(t)
Û =− A sin t +δ (13.55)
3m 3m
where A and δ can be found by considering the initial conditions. It turns out that δ = 0 and A = x 0 , after a
quick substitution of the initial conditions. As such, the mass attached to the spring is undergoing simple
56
Solution Manual
harmonic motion described by:
r !
4k
x(t) = x 0 cos t (13.56)
3m
r
4k
with angular frequency ω = t . As for the motion characterized by the coordinate y, since we know
3m
that (i) py is constant in time, and (ii) py (0) = 0, we have, necessarily, that py = 0 for all time t. From the
second equation in Eq. (13.48), we have:
r r !
xÛ 1 4k 4k
yÛ = − ⇒ yÛ = x 0 sin t (13.57)
4 4 3m 3m
A quick integration (and applying initial conditions) yields us:
r !
1 1 4k
y(t) = (y0 + x 0 ) − x 0 cos t (13.58)
4 4 3m
It is not surprising to see that y is executing simple harmonic motion as well.
13.25. Here is another example of a canonical transformation, which is still too simple to be of any real use,
but does nevertheless illustrate the power of these changes of coordinates. (a) Consider a system with one
degree of freedom and Hamiltonian H = H (q, p) and a new pair of coordinates Q and P defined so that
√ √
q = 2P sin Q and p = 2P cos Q. Prove that if ∂H ∂H
∂q = −p and ∂p = q,
Û Û it automatically follows that ∂H
∂Q = −P
Û
∂H
and ∂P = Q.Û In other words, the Hamiltonian formalism applies just as well to the new coordinates as to
the old. (b) Show that the Hamiltonian of a one-dimensional harmonic oscillator with mass m = 1 and force
constant k = 1 is H = 12 (q 2 + p 2 ). (c) Show that if you rewrite this Hamiltonian in terms of the coordinates
Q and P defined above, then Q is ignorable. What is P? (d) Solve the Hamiltonian equation for Q(t) and
verify that, when rewritten for q, your solution gives the expected behaviour.
Solution. (a) We have the following relation given in the question:

√
q = 2P sin Q
√ (13.59)
p = 2P cos Q
We compute the relevant derivatives below:
∂q sin Q ∂q √
= √ = 2P cos Q (13.60)
∂P 2P ∂Q
∂p cos Q ∂p √
= √ = − 2P sin Q (13.61)
∂P 2P ∂Q
57
Solution Manual
∂H ∂H
Consider the Hamiltonian equation for Q, recalling that = qÛ and = −pÛ :
∂p ∂q
∂H ∂H ∂p ∂H ∂q
= +
∂Q ∂p ∂Q ∂q ∂Q
∂p ∂q
= qÛ − pÛ
∂Q ∂Q
∂q Û ∂q Û ∂p ∂p Û ∂p Û ∂q

= Q+ P − Q+ P
∂Q ∂P ∂Q ∂Q ∂P ∂Q
∂q ∂p ∂p ∂q ∂q ∂p ∂p ∂q

= QÛ − + PÛ −
∂Q ∂Q ∂Q ∂Q ∂P ∂Q ∂P ∂Q
∂q ∂p ∂p ∂q

= 0 + PÛ −
∂P ∂Q ∂P ∂Q
= PÛ − sin2 Q − cos2 Q = −PÛ (shown)

Similarly, we can consider the Hamiltonian equation for P:
∂H ∂H ∂p ∂H ∂q
= +
∂P ∂p ∂P ∂q ∂P
∂p ∂q
= qÛ − pÛ
∂P ∂P
∂q Û ∂q Û ∂p ∂p Û ∂p Û ∂q

= Q+ P − Q+ P
∂Q ∂P ∂P ∂Q ∂P ∂P
∂q ∂p ∂p ∂q ∂q ∂p ∂p ∂q

=Q Û − +P Û −
∂Q ∂P ∂Q ∂P ∂P ∂P ∂P ∂P
∂q ∂p ∂p ∂q

= QÛ − +0
∂Q ∂P ∂Q ∂P
= QÛ cos2 Q + sin2 Q = QÛ (shown)

(b) The Lagrangian for the 1-D harmonic oscillator is simply given by the quantity T − U :
1 1
L = T − U = mqÛ2 − kq 2 (13.62)
2 2
We can express the velocity as a function of the momentum:
∂L p
p= = mqÛ ⇒ qÛ = (13.63)
∂qÛ m
Thus the Hamiltonian can be written as such:
1 p 2 1 2
H =T +U = m + kq
2 m 2
1 2 2
= (q + p ) (shown) (13.64)
2
where the last line was obtained by setting k = m = 1, as given in the question.
58
Solution Manual
(c) Using the coordinate transformations defined in Eq. (13.59), we get:
1
H= 2P sin2 Q + 2P cos2 Q = P (13.65)

2
And so P is the Hamiltonian of the system. It can also be observed that Q is no longer present in the
Hamiltonian, and so, is an ignorable coordinate.
(d) We have from Hamilton’s equations:
∂H
QÛ = =1 (13.66)
∂P
As such, we can easily find Q(t) by integrating, then solving for q:
Q =t +δ (13.67)
√
q = 2 sin t + δ (13.68)
where δ is a constant of integration. This gives the expected behaviour (a simple harmonic oscillation).
59
Chapter 14
Collision Theory
60
Chapter 15
Special Relativity
61
Chapter 16
Continuum Mechanics
62

Taylor

Diunggah oleh

Informasi Dokumen

Judul Asli

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

Taylor

Diunggah oleh

Hak Cipta:

Format Tersedia

Classical Mechanics [Taylor, J.R.

1 Newton’s Laws of Motion 2

2 Projectiles and Charged Particles 3

3 Momentum and Angular Momentum 4

8 Two-Body Central-Force Problems 16

9 Mechanics in Non-inertial Frames 19

10 Rotational Motion of Rigid Bodies 20

11 Coupled Oscillators and Normal Modes 36

12 Nonlinear Mechanics and Chaos 46

Newton’s Laws of Motion

1.1. No plans to do this chapter yet.

Projectiles and Charged Particles

2.1. No plans to do this chapter yet.

Momentum and Angular Momentum

3.1. No plans to do this chapter yet.

4.1. No plans to do this chapter yet.

5.1. No plans to do this chapter yet.

Solution. Credit: http://www.physics.ohio-state.edu/~mathur/5300/5300ass2sol.pdf

Solution. Credit: http://www.physics.ohio-state.edu/~mathur/5300/5300ass2sol.pdf

Clearly, the derivative is 0 only when z = 0, i.e., Q is in the xy plane (shown).

sin θ 1 = sin θ 2 (6.5)

and hence θ 1 = θ 2 (shown).

Solution. Credit: http://chriskranenberg.wix.com/taylor-mechanics-solutions#!

Solution. Credit: http://chriskranenberg.wix.com/taylor-mechanics-solutions#!

Solution. Credit: http://www.physics.ohio-state.edu/~mathur/5300/5300ass2sol.pdf

which is what we needed to show.

[Remaining solutions to Chapter 6 can be found in http://chriskranenberg.wix.com/taylor-mechanics-solutions#

Solution. The Lagrangian can be easily identified and written as such:

A direct application of the Euler-Lagrange equations yields:

Solution. (a) The Lagrangian is simply:

We can then rewrite the Lagrangian from Eq. (7.5) as:

Applying the Lagrange equations, we get

x H (t) = A1e ωt + A2e −ωt (7.17)

−ω 2 B sin ωt − ω 2 B sin ωt = −д sin ωt (7.18)

We can then compose our general solution for x(t):

We have the initial conditions: x(0) = x 0 and x(0)

Finally, we are able to construct our general equation as:

Figure 7.1: Sketch of the problem

Then, applying Lagrange equation to the two generalized coordinates:

⇒ mR 2ϕÛ sin2 θ = lz = const (7.28)

(c) If ϕ = const, then ϕÛ = 0, and Eq. (7.27) reduces to:

Figure 7.2: Plot of Equation (7.33)

position x 0 , and also write down the derivatives of θ :

cos θ ≈ cos θ 0 − ϵ sin θ 0 (7.34)

After some simplification and ignoring terms in ϵ 2 , we obtain:

Two-Body Central-Force Problems

Figure 8.1: Graphs of the potentials U (r ), Ucf (r ) and Ueff (r )

We can write the potentials as such:

observation, we can deduce the ‘spring constant’ k 0 and oscillation frequency ω:

Figure 8.2: Transfer orbit for Problem 8.8.34

And so to get from P to P 0:

Mechanics in Non-inertial Frames

9.1. No plans to do this chapter yet.

Rotational Motion of Rigid Bodies

rα = R + rα0 ⇒ rα0 = rα − R (10.1)

CM of the system, we perform the following:

The coordinates of the CM is given by (0, 0, H /5).

Figure 10.1: Taken from http://astro-learned.blogspot.sg/

dV = (dr )(rdθ )(r sin θ dϕ) = r 2 sin θ drdθdϕ (10.8)

x = r sin θ cos ϕ, y = r sin θ sin ϕ, z = r cos θ . (10.10)

We then have the volume element as dV = dxdydz = J (r , θ, ϕ) drdθdϕ, with J (r , θ, ϕ) being:

Solution. Credit: http://www.physics.umd.edu/courses/Phys410/gates/Phys410_Solution_07.pdf

Solution. For continuous masses, we have: