DEPARTMENT OF PHYSICS

INDIAN INSTITUTE OF TECHNOLOGY, MADRAS

PH1010 Physics I Problem Set 2 - Solutions Nov 2022

1. (a) ~r(t) = ~a cos ωt + ~b sin ωt. Since ~r(t) is a linear combinations of the fixed vectors
~a and ~b at all times, the motion is in the plane (passing through origin) spanned
by ~a and ~b. Let us reorient our axes such that this plane is the xy-plane, so that
~a and ~b has no z-component. Then

x(t) = ax cos ωt + bx sin ωt, y(t) = ay cos ωt + by sin ωt. (1)

We readily see that

by x(t) − bx y(t) ay x(t) − ax y(t)
cos ωt = , sin ωt = − (2)
ax by − ay bx ax by − ay bx
Therefore, we have
(by x(t) − bx y(t))2 (ay x(t) − ax y(t))2
+ =1 (3)
(ax by − ay bx )2 (ax by − ay bx )2
Any equation of such a form can be reduced generically to either an ellipse or a
hyperbola by reducing to the standard form:
(x(t) cos α + y(t) sin α)2 (−x(t) sin α + y(t) cos α)2
± ± = 1. (4)
A2 B2
By equating coefficients of x2 , y 2 and xy in the two expressions we can determine
A, B and α. The signs are chosen such that A and B are real.
(b) Angular momentum about the origin :
~ = ~r × p~
L
= ~r × m~r˙
   
= m ~a cos ωt + ~b sin ωt × −~aω sin ωt + ~bω cos ωt
 
~
= mω ~a × b .

The angular momentum is constant.

2. (a)

dE(θ, θ̇) ∂E dθ ∂E dθ̇

= +
dt ∂θ dt ∂ θ̇ dt
= mgl(sin θ)θ̇ + ml2 θ̇θ̈
 
= mlθ̇ g sin θ + lθ̈

1
 dE
But the potential does not have an explicit time dependence, hence dt
= 0. This
should follow from the equation of motion
g
θ̈ = − sin θ (5)
l

(b) For small enough θ, sin θ ≈ θ and the EOM becomes

g
θ̈ = − θ
l
This is the EOM of a simple harmonic oscillator (i.e.q ẍ = −ω 2 x). The angular
frequency ω = gl and the timeperiod T = 2π = 2π gl .
p
ω

(c)
V (θ) = mgl (1 − cos θ)
dV
V 0 (θ) = = mgl sin θ
dθ
V 00 (θ) = mgl cos θ
The critical points for V (θ) are θ = nπ, n ∈ Z.
V 00 (2nπ) > 0 implies the minima of the potential, hence (0, 0) in the (θ, θ̇) plane
are the stable equilibrium points (the black dots in the top contour plot of Figure
1).
V 00 ((2n + 1)π) < 0 implies the maxima of the potential, hence (±π, 0) in the (θ, θ̇)
plane are the unstable equilibrium points (the red dots in the top contour plot of
Figure 1).
(d) For the plots below we have taken m = 1, l = 1, g = 10. Therefore, one can easily
find out the corresponding phase space region from the contour labels.
From phase space contour it is clear that for E < 2mgl (i.e. the region E < 20
in the attached plot (we have taken m = 1, l = 1, g = 10)) the motion is periodic
and for E > 2mgl the motion is not periodic.

2
 20 20
6

÷
35 15 35 15 35 35 15

10 30 30 20 30 20 10 30
4

25 5 > 20
25 >5 u 25 25 5
^^ 35
2 × ✓
n

30
✓
n

✓ 25
p 0 ••
•••
••
a
•••
••
•••
••

20

^
v n
✓ ✓ 15
-2
10
L
5 25 25 r
25 5 25
5
r v
-4 ,

30 20 10 20 30 20 10 30 20 20 30 10
✓
r

15 35 35 35 15 35
-6
20 20
-10 -5 0 5 10

θ
20

15
V(θ)

10

-10 -5 0 5 10

Figure 1: Figure on the bottom is simply the plot of the potential. Figure on the top is the
corresponding contour plot of phase space. The above plots are for m = 1, l = 1, g = 10.

(e) For given energy E = 2mgl, the particle will just able to reach θ = π and we have
1
2mgl = ml2 θ̇2 + mgl (1 − cos θ)
r2
2g
⇒ θ̇ = (1 + cos θ)
l
Z T /2 s Z
π
l dθ
⇒ dt = p
0 2g 0 (1 + cos θ)

But the integrand diverges as θ → π. Hence, it will take infinite time just to
reach half oscillation. In fact, there is no oscillations for E = 2mgl !

Something more: Let’s consider E < 2mgl. In this case the energy equation is
as follows

3
 1
mgl (1 − cos θ0 ) = ml2 θ̇2 + mgl (1 − cos θ)
2
where θ0 is the amplitude. The time period of the oscillation is

s
Z T Z θ0
l dθ
dt = p
0 2g −θ0 (cos θ − cos θ0 )
s Z
θ0
l dθ
=2 p
2g 0 (cos θ − cos θ0 )
s Z
θ0
1 l dθ
cos θ = 1 − 2 sin2 (θ/2)


=2 2 r
sin (θ0 /2) 2g 0 sin2 (θ/2)

1 − sin 2 (θ /2)
0
s Z
l π/2 sin2 (θ/2)
 
dβ 2
=4 sin β = , x = sin (θ0 /2)
g 0 1 − x2 sin2 β sin2 (θ0 /2)

For small oscillations, x → 0 and we recover the results for part (b) of the question.
For finite amplitude the above integral can exactly be evaluated as an elliptical
integral of the first kind.
For E > 2mgl there is no periodic motion as we have already seen it form the
phase space contour.

4
 (f) Substituting E = 2mgl in the energy equation, we get,

r
g
θ̇ = 2 cos (θ/2)
l
Z Z r
dθ g
⇒ = 2 dt
cos (θ/2) l
  r
θ π g
⇒ ln tan + =t
4 4 l
 √ 
⇒ θ(t) = 4 arctan et g/l − π

subject to the initial condition θ(0) = 0.1 If we consider θ̇ = −2 gl cos (θ/2) then
p

the corresponding solution is

 √ 
θ(t) = 4 arctan e−t g/l − π

Let’s analyze: The above solution clearly shows that θ → π only when t → ∞.
We have already seen this in the previous part. For all other time, θ < π.
Therefore, the trajectory asymptotically approaches the equilibrium point. This
is indeed a consequence of the fact that phase space trajectories don’t intersect.
The equilibrium point (π, 0) is itself a phase space trajectory and can not be
intersected by another.

Summary of the simple pendulum:

• 0 < E  2mgl: This is the case of small oscillations and behave as a SHO.
The phase space trajectory is closed.
• 0  E < 2mgl: The motion is oscillatory but not simple harmonic.The phase
space trajectory is also closed. The time period can be evaluated using ellip-
tical integral of the first kind.
• E = 2mgl: The motion is Not oscillatory. There are three phase space tra-
jectories possible for this case. The two are which approaches θ = ±π asymp-
totically, and third one is the fixed point (π, 0). It takes infinite time to reach
the fixed point for the dynamic trajectory. This singular trajectory in phase
space separates two types of motion: those with E < 2mgl, which oscillate
as usual (referred to sometimes as “librations”), and those with E > 2mgl,
which rotate fully in either clockwise or anticlockwise directions (referred to
as “rotations”). Thus the dynamic trajectories with E = 2mgl are called
separatrices.
• E > 2mgl: The motion is not periodic, and also eternally clockwise or anti-
clockwise.
1
While we have explicitly integrated the large amplitude trajectory that corresponds to E = 2mgl we
cannot do so for any other energy exactly using elementary functions. For E 6= 2mgl the exact solutions are
in terms of the Jacobi elliptic functions.

5
3. Consider the potential V (x) = Kx2n . From energy conservation, we can write the
formula for the time period. The turning points are
1
±(E/K) 2n .

Using symmetry of the potential (it is an even function), we get

1
m (E/K) 2n
rZ
dx
T = 4 √
2 0 E − Kx2n
r Z 1
m − 1 1−n du
= 4 K 2n E 2n √ , (6)
2 0 1 − u2n
1−n
where u = (K/E)1/(2n) x. Therefore, T ∝ E 2n .
We can also get this from dimensional analysis. The time period T should be a function
of the mass m, the constant K and the conserved energy E of the motion. So

T ∝ E a K b mc (7)

Equating mass, length and time dimensions on both sides, we get a = (1 − n)/(2n),
b = −1/(2n) and c = 1/2.

6
4. • E < −1: As tanh x is bounded below by −1, there is no phase space region that
would satisfy this energy.
• E = −1: The energy equation is

p2
+ tanh x = −1
2
As x → −∞ momentum p and time derivative of the momentum (which is −∇V )
both vanishes. Hence, (−∞, 0) is an equilibrium point (fixed point) in the phase
space. Therefore, for E = −1 the phase space trajectory is a single point (−∞, 0).
• |E| < 1: The energy equation is

p2
+ tanh x = E
2
p
As x → −∞ momentum p → ± 2(1 + E). The particle will continue to move
upto the point where E = tanh x in other words it will continue to move upto
x = tanh−1 E and then goes back to −∞.
• E = 1: The energy equation is

p2
+ tanh x = 1
2
√
As x → −∞ momentum p → ±2. As x → 0 momentum p → ± 2. As x → ∞
momentum p → 0. Here, (∞, 0) is an equilibrium point in the phase space (as
momentum and time derivative of momentum both vanishes).
• E > 1: The energy equation is

p2
+ tanh x = E
2
p √
As x → −∞ momentum p →p± 2(1 + E). As x → 0 momentum p → ± 2E.
As x → ∞ momentum p → ± 2(−1 + E).
Now compiling all the above analysis the phase space contour looks as follows:

5. Consider the potential given by v(x) = −A|x|n . The initial conditions given in the
problem states that the particle barely reach the point x = 0. This would imply that
E = 0, (If E < 0, the particle does not reach x = 0, and if E > 0, then ẋ2 > 0, at
x = 0). Thus, we have
2A n
ẋ2 = |x| .
m
Thus, the time taken t to go from an initial point x0 to 0 is given by:
Z 0 Z −x0 r
dx m dx
t= q = .
x0 2A
|x|n
/2 0 2A xn/2
m

7
Performing the integration, we get the time t as
r
m x(−n/2+1) −x0
t= | .
2A (−n/2 + 1) 0

For this to be finite −n/2 + 1 > 0, which in turn implies that n < 2.

8
 0.30.25 0.1
0.250.3 0.30
0.4 > ✓
^
) 0.2
0.2
0.25
0.2
m=1
a

0.15
p 0.0 0.15
^
-000
✓ 0.20
l =1

÷
 ,
-0.2 0.2 0.15 g =10
0.25 0.05
-0.4 0.10
0.3 0.250.3
0.2
-0.2 -0.1 0.0 0.1 0.2
0.05
θ

6 80 90 90 80 90
35 35 35
7 10 70
30 10 30 60 > 80
4 ✓ 30 50
25 20 20 25 30 >
> 20 40 20 20
5 10 > 10 ✓ > 70
2 an

[
j
a
^ 25
60
p 0 Go
n n
BB
✓
✓
BO

20
p 0 BO n Br
✓
BB
n

50
-2 , her ihr
10

-5

-4
25 20 5 y 20 25 15 e
50
20
30<
40
sr
40
30 30 60
-10 70 30
35 15 35 10
-6 80 90 90 80
20
-3 -2 -1 0 1 2 3 -6 -4 -2 0 2 4 6
5
10
θ θ

Figure 2: The above plots are in the unit m = 1, l = 1, g = 10. The upper left figure is the
contour plot of phase space for E  2mgl. The bottom left is for E < 2mgl. The bottom
right is for E > 0. The red dots are the unstable fixed points and black dots are the stable
fixed points. the blue arrows shows the flow direction of phase space trajectory.

9
 1.0

0.5

V(x)
0.0

-0.5

-1.0

-10 -5 0 5 10

Figure 3: Plot of V (x) = tanh x

4
7
6 8

> 5
4

2 ? >
' 3
1 8
> 2 >
7
>
I
6

> 5
p 0 0
<
≥
4

L 1 3

< < 2
L
L 1
-2 2 < 3
< < 0

4
: 5

7
6 8
-4
-4 -2 0 2 4

x
Figure 4: Phase space contour for V (x) = tanh x. Both the fixed points are (−∞, 0) and
(∞, 0).

10