University of Alabama

Department of Physics and Astronomy

PH 301 / LeClair Fall 2018

Problem Set 6
Solutions

1. (a) Write down the Lagrangian L(x1 , x2 , ẋ1 , ẋ2 ) for two particles of equal mass m1 = m2 = m,
confined to the x axis and connected by a spring with potential energy U = 21 x2 , where x is the
extension of the spring, x = (x1 − x2 − l) and l is the spring’s unstretched length. You can assume
mass 1 remains to the right of mass 2 at all times. (b) Rewrite L in terms of the new variables
X = 21 (x1 +x2 ) (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.

Solution: With the given coordinates,

1 1
T = mẋ21 + mẋ22 (1)
2 2
1 1
U = k (x1 − x2 − l)2 = kx2 (2)
2 2
1
1
L = T − U = m ẋ1 + ẋ22 − kx2
2
(3)
2 2

With the new variables, we note x1 = X + 21 x+ 12 l and x2 = X − 12 x− 12 l. Taking the time derivatives
and substituting,
" 2  #
1 2

1 1 1 1 1
L= m Ẋ + ẋ + Ẋ − ẋ − kx2 = mẊ 2 + mẋ2 − kx2 (4)
2 2 2 2 4 2

The two Lagrange equations are:

∂L d ∂L
= or 0 = 2mẌ (5)
∂X dt ∂ Ẋ
∂L d ∂L 1
= or − kx = mẍ (6)
∂x dt ∂ ẋ 2

The X equation implies motion at constant velocity, Ẋ = const = Vo , so X(t) = Vo t + Xo . In other

words, the CM moves like a free particle, which makes sense since there are no external forces. The
x equation is just SHM:

2k
ẍ = − x (7)
m

where we know the general solution is x(t) = A cos (ωt − δ). The two masses oscillate in and out
 p
relative to each other with frequency ω = 2k/m. We can view this result in several ways. On one
hand, the spring compresses twice as much with both masses moving compared to only one mass
moving. On the other hand, we could say it is as though the spring has an effective force constant
2k with both masses moving.

2. A mass m is suspended from a massless string, the other end of which is wrapped several times
around a horizontal cylinder of radius R and moment of inertia I, which is free to rotate about a
fixed horizontal axle. Using a suitable coordinate, set up the Lagrangian and the Lagrange equations
of motion, and find the acceleration of the mass m. (The kinetic energy of a rotating cylinder is
1 2
2 Iω .)

Solution: W need to account for the translational kinetic energy of the mass m moving and the
rotational kinetic energy of the cylinder. If the rope is not to slip on the cylinder, then the angular
velocity ω for the cylinder must match the linear velocity ẋ of the mass m, ẋ = Rω. Put another
way, the arclength a point on the surface of the cylinder moves through in a time t must match the
distance the mass has fallen, so x = θR. Taking a time derivative and noting θ̇ = ω, we have the
same condition. Thus,
 
1 1 1 1 I 2 1 I
T = mẋ2 + Iω 2 = mẋ2 + ẋ = m + 2 ẋ2 (8)
2 2 2 2 R2 2 R

Taking the zero of potential energy to be the starting position of the mass m, with +x downward,
we have U = −mgx, and thus
 
1 I
L= m + 2 ẋ2 + mgx (9)
2 R
∂L
= mg (10)
∂x  
d ∂L I
= m + 2 ẍ (11)
dt ∂ ẋ R
g
ẍ = I
(12)
1 + mR 2

3. Using the usual angle ϕ as generalized coordinate, write down the Lagrangian for a simple
pendulum of length l suspended from the ceiling of an elevator that is accelerating upward with
constant acceleration a. (Be careful when writing T ; it is probably safest to write the bob’s velocity
in component form.) Find the Lagrange equation of motion and show that it is the same as that
for a normal, non-accelerating pendulum, except that g has been replaced by g + a. In particular,
p
the angular frequency of small oscillations is (g + a)/l.

Solution: The elevator is a non-inertial frame (primed coordinates), so we should analyze the
situation from a reference frame on the ground (unprimed coordinates). The easiest way forward is
to write down the coordinates of the mass m from a cartesian reference frame based on the ground,
with y upward and x to the right. With respect to the elevator, with the origin at the support
point, the mass’ position is (x0 , y 0 ) = (l sin ϕ, −l cos ϕ). We add to this the position of the elevator
relative to the ground, (0, 12 at2 ), and find the position of the mass as:

x = l sin ϕ (13)
1
y = at2 − l cos ϕ (14)
2

Taking the time derivative, and remembering the chain rule,

ẋ = vx = lϕ̇ cos ϕ (15)

ẏ = vy = lϕ̇ sin ϕ + at (16)

The height of the bob above the ground is y = 21 at2 − l cos ϕ, making the potential energy mgy.
We can then construct the Lagrangian, with v 2 = vx2 + vy2 :
 
1 1 1 2
L = mv 2 − mgy = m a2 t2 + 2atlϕ̇ sin ϕ + l2 ϕ̇2 − mg


at − l cos ϕ (17)
2 2 2

The Lagrange equation is then readily obtained. First the relevant derivatives.

∂L
= matlϕ̇ cos ϕ − mgl sin ϕ (18)
∂ϕ
∂L
= ml2 ϕ̇ + matl sin ϕ (19)
∂ ϕ̇
d ∂L
= ml2 ϕ̈ + matlϕ̇ cos ϕ + mal sin ϕ (20)
dt ∂ ϕ̇

The first and third lines must be equal according to the Lagrange equation.

matlϕ̇ cos ϕ − mgl sin ϕ = ml2 ϕ̈ + matlϕ̇ cos ϕ + mal sin ϕ (21)
−mgl sin ϕ = ml2 ϕ̈ + mal sin ϕ (22)
−g sin ϕ = lϕ̈ + a sin ϕ (23)
lϕ̈ = − (g + a) sin ϕ (24)

This is the normal equation for a pendulum except that the acceleration is g + a rather
q than g. For
g+a
small angles, where sin ϕ ≈ ϕ, this leads to a frequency of small oscillations ω = l .

4. We saw in example 7.3 in the text that the acceleration of an Atwood machine is ẍ = (m1 −
m2 )g/(m1 +m2 ). It is sometimes claimed that this result is “obvious” because, it is said, the effective
force on the system is (m1 − m2 )g and the effective mass is (m1 + m2 ). This is not, perhaps, all that
obvious, but it does emerge very naturally in the Lagrangian approach. Recall that the Lagrange
equation can be thought of as

(generalized force) = (rate of change of generalized momentum) (25)

Show that for the Atwood machine the generalized force is (m1 − m2 )g and the generalized momen-
tum is (m1 + m2 )ẋ.

Solution: We’ve already considered the Atwood machine in Chapter 7 (example 7.3), so we already
know the Lagrangian:

1
L=T −U = (m1 + m2 ) ẋ2 + (m1 − m2 ) gx (26)
2

Using Lagrange’s equation,

∂L
= (m1 − m2 ) g = (generalized force) (27)
∂x
d ∂L d
= (m1 + m2 ) ẋ = (m1 + m2 ) ẍ = (generalized momentum) (28)
dt ∂ ẋ dt

5. A pendulum consists of a mass m and a massless stick of length l. The pendulum support
oscillates horizontally with a position given by x(t) = A cos ωt. What is the general solution for the
angle of the pendulum as a function of time?

Solution: As with problem 3, we need to choose a reference frame on the ground since the pendulum
support is accelerating. Again, we write down the position with respect to the ground, with x
horizontal and y vertical. With x(t) = A cos ωt for the support, the position of the mass is

(x, y)m = (x + l sin θ, −l cos θ) (29)

Now we take the time derivative to get the velocity components:

ẋm = ẋ + lθ̇ cos θ (30)

ẏm = lθ̇ sin θ (31)
2
vm = ẋ2m + 2
ẏm 2 2 2 2 2 2 2 2 2 2
= ẋ + 2ẋlθ̇ cos θ + l θ̇ cos θ + l θ̇ sin θ = ẋ + 2ẋlθ̇ cos θ + l θ̇ (32)

The potential energy with y measured from the vertical position of the support is U = −mgy =
−mgl cos θ, so the Lagrangian is
 1  
L = m ẋ2 + 2ẋlθ̇ cos θ + l2 θ̇2 + mgl cos θ (33)
2

Clearly our generalized coordinates are x and θ. Since the Lagrangian is independent of x, the
coordinate x is ignorable, and Lagrange’s equation for x will only indicate that momentum is
conserved along x. Applying Lagrange’s equation for θ,

d ∂L ∂L
= (34)
dt ∂ θ̇ ∂θ
d  2 
ml θ̇ + mlẋ cos θ = −mlẋθ̇ sin θ − mgl sin θ (35)
dt
lθ̈ + ẍ cos θ = −g sin θ (36)

Plugging in ẍ = −ω 2 A cos ωt,

lθ̈ − Aω 2 cos (ωt) cos θ + g sin θ = 0 (37)

It sort of looks like a normal pendulum equation, except that we have both horizontal and vertical
components of acceleration. We can apply a small angle approximation for θ:

A 2 g
θ̈ − ω cos (ωt) + = 0 (38)
l l
A
θ̈ + ωo θ = ω 2 cos (ωt)
2
(39)
l
p
Here ωo = g/l. This is just the equation of a driven oscillator without damping, whose solution
we investigated in Ch. 5:

(A/l)ω 2
θ(t) = cos ωt + C cos (ωo t + ϕ) (40)
ωo2 − ω 2

where C and ϕ are constants, determined by the initial conditions.

6. Consider a mass m connected to a spring of force constant k, confined to the x axis. We know
p
the system executes simple harmonic motion with frequency ω = k/m if the spring is massless.
Using the Lagrangian approach, you can find the effect of the spring’s mass M as follows. (a)
Assuming the spring is uniform and stretches uniformly, show that its kinetic energy is 16 M ẋ2 . (As
usual x is the extension of the spring from its equilibrium length.) Write down the Lagrangian for
the system of cart plus spring (note U = 12 kx2 still holds.) (b) Write down the Lagrange equation
p
and show that the mass still executes SHM, but with angular frequency ω = k/(m + M/3); that
is, the effect of the spring’s mass M is just to add M/3.

Solution: With one end of the spring fixed, the other end moves with the velocity of the mass ẋ,
and we presume the velocity varies linearly from one end to the other. If we compress or extend
the spring by some amount x, its current length is then l + x. If the velocity is to vary linearly over
the whole length, at some position x0 along the spring the velocity is

x0
v(x0 ) = v (41)
l+x

At position x0 , the mass dm of some infinitesimal segment dx0 would be the fractional mass of the
segment (the length of the segment divided by the total length) multiplied by the mass of the spring
M:

dx0
dm = M (42)
l+x

Putting the two together we can find the kinetic energy of our infinitesimal segment of the spring:

1
dT = v(x0 )2 dm (43)
2

Now we just integrate over the length of the spring, from 0 to l + x. Note v = ẋ.

Zl+x Zl+x Zl+x

1 x02 v 2 M 0 1 ẋ2 M
T = dT = dx = x02 dx0 (44)
2 (l + x)2 l + x 2 (l + x)3
0 0 0
1 M x03 l+x 1
T = ẋ2 = M ẋ2 (45)
2 (l + x)3 3 0 6

So the spring behaves as though 1/3 of it moves with the full velocity ẋ. The Lagrangian is now
readily constructed

1 1 1
L = M ẋ2 + mẋ2 − kx2 (46)
6 2 2

The Lagrange equations show that we do indeed have SHM at the specified frequency.

∂L
= −kx (47)
∂x
d ∂L 1 ∂L
= M ẍ + mẍ = − (48)
dt ∂ ẋ 3 ∂x
k
=⇒ ẍ = x (49)
m + 31 M
s
k
=⇒ ω= (50)
m + 13 M