Phys 7221 Homework # 8

Gabriela Gonzalez
November 15, 2006

Derivation 5-6: Torque free symmetric top

In a torque free, symmetric top, with Ix = Iy = I , the angular velocity vector in body
coordinates with axes along the principal axes, is given by
x = 0 cos t
y = 0 sin t
z = 3
with = 3 (I3 ?I)/I. Also in body coordinates, the angular momentum vector L has
components
Lx = Ix = I0 cos t
Ly = Iy = I0 sin t
Lz = I3 z = I3 3
We see that the component of the angular momentum along the top symmetry axis, Lz ,
is constant, and the the components perpendicular to the symmetry axis rotate about
the z axis with angular velocity . Since the top is free of external torques, the angular
momentum vector L is constant in an inertial system. In an inertial system, it is the tops
symmetry axis the one that rotates about the L direction. We can use Euler angles to
describe the rotation between body axes and inertial axes. Comparing the expressions for
the components of in body axes we obtained and the expressions for the same components
in terms of Eulers angles in (4.87):
x = 0 cos t = sin sin + cos
y = 0 sin t = sin cos + sin
z = 3 = cos +
we recognize that the Euler angle is constant, the bodys rotation about the symmetry
axis is = /2 , and
cos = 3 = 3 + = I3 3 /I
sin = 0
1

The angle is the angle between the symmetry axis and the angular momentum vector,
and is determined by initial conditions. The bodys symmetry axis rotates about the angular
momentum with constant angular velocity = I3 3 /I cos . We now use the Euler angles
we obtained to calculate the components of in the inertial system, using the expression
from Derivation 15 in Chapter 4:
+ 0 )
x = cos + sin sin = sin sin(t
+ 0 )
y = sin sin cos = sincos(t
z = cos + = cos + I3 3 /I
We see that the component of the angular momentum along the z-axis, or the direction of
the angular momentum vector L, is constant, and the component of the angular velocity

perpendicular
~ and L is given
q to L is rotating with angular velocity . The angle between
0
2
2
by sin = x + omegay / = sin /. The angle between and the symmetry axis is
q
given by sin 00 = x20 + y20 / = 0 /. Then we have
sin 0 = sin / = sin sin 00 /0 = sin 00 /
where we have used 0 = sin , which we had obtained when solving for the Euler angles.
We can also use moments of inertia for an expression of sin 0 :
sin 0 = sin 00 / = I cos sin 00 /I3 3 = ((I3 I)/I3 ) cos sin 00
For the Earth considered as a free symmetric top, we have (I3 I)/I 3 103 , so the
angle 0 is very small, independent of values of , 00 : the angular velocity vector
~ is very
close to the angular momentum vector L. The measured distance 2R sin 00 is about 10m,
so the distance R sin 0 = (I3 I)/IR sin 00 cos 15mm cos < 1.5cm.
As seen in the body axes, the angular velocity vector describes a cone of aperture angle
about the symmetry axis: this is called the body cone. As seen in the inertial frame, the
angular velocity vector describes a cone of aperture angle , about the angular momentum
vector: this is called the space cone. Both cones share the angular momentum vector along
their sides at any given instant. The angular velocity vector is the instantaneous axis of
rotation, so the cones are rolling without slipping on each other.
A very nice page with animations showing this example, by Prof. Eugene Butikov, can
be found in http://faculty.ifmo.ru/butikov/Applets/Precession.html , from which
Fig.1 is a snapshot.

Exercise 5-15
Consider a flat rigid body in the shape of a right triangle with uniform mass density
= M/A, and area A = a2 /2, where a is the length of the equal sides of the triangle.
2

Figure 1: A rotating torque free symmetric top (left), and the associated space and body
cones. The conserved angular momentum vector (blue) is along the z axis; the (red)
instantaneous angular velocity vector is at the intersection of the cones, precessing about
the z axis. From http://faculty.ifmo.ru/butikov/Applets/Precession.html

Figure 2: Exercise 5-15

Let us choose the right angle corner of the triangle as the origin of a coordinate system
with the x, y axis along the sides of the triangle. The boundary of the triangle is given by
x + y = a; the mass elements on the surface will have coordinates (x,y,0) with (x,y) within
the triangle. The elements of the inertia tensor in such a system are
Z
Z a Z ax
Z a
(a x)3
a4
M a2
Ixx =
(y 2 + z 2 )dA =
dx
y 2 dy =
dx =
=
3
12
6
A
0
0
0
Z a Z ay
Z
M a2
2
2
x2 dx =
dy
(x + z )dA =
Iyy =
6
0
0
A
Z
2
Ma
Ixx =
(x2 + y 2 )dA = Ixx + Iyy =
3
A
Z a
Z ax
Z a
Z
(a x)2
a4
M a2
xdx
ydy =
xdx
xydA =
Izz = Iyx =
=
=
2
24
12
0
0
0
A
Z
Ixz =
xzdA = 0 = Iyz
A

2 1 0
M a2
1 2 0
I=
12
0
0 4
We look for eigenvalues of the inertia tensor, which will be solutions to the equation
det(I I1) = 0, or


 2k I

k
0



 = (4k I)((2k I)2 k 2 )
2k I
0
det(I I1) =  k


0
0
4k I 
with k = M a2 /12. The three real, positive solutions are the principal moments of inertia:
(I1 , I2 , I3 ) = (k, 3k, 4k) = (M a2 /12, M a2 /4, M a2 /3).
The principal axes are the eigenvectors corresponding to each eigenvalue. If the eigenvectors have components ~ni = (nix , niy , niz ), the equations are I ~ni = i k~ni with i =
1, 3, 4. The equations for n1 are:
I n1 = kn1

2 1 0
n1x
n1x
k 1 2 0 n1y = k n1y
0
0 4
n1z
n1z

2n1x n1y
n1x
n1x + 2n1y = n1y n1x = n1y , n1z = 0.
4n1z
n1z
4

The equations for n2 are:

I n2 = 3kn2

2 1 0
n2x
n2x
k 1 2 0 n2y = 3k n2y
0
0 4
n2z
n2z

2n2x n2y
n2x
n2x + 2n2y = 3 n2y n1x = n1y , n1z = 0
4n2z
n2z
The equations for n3 are:
I n3 = 4kn3

2 1 0
n3x
n3x
k 1 2 0 n3y = 4k n3y
0
0 4
n3z
n3z

2n2x n2y
n3x
n2x + 2n2y = 4 n3y n3x = n3y = 0, n3z 6= 0
4n2z
n3z
The eigenvectors with unit magnitude are then

n1 = (1/ 2)(1, 1, 0)

n2 = (1/ 2)(1, 1, 0)
n3 = (0, 0, 1)

Exercise 5-17: A rolling cone

The instantaneous axis of rotation, and thus the direction of the angular velocity
~ is along
the line of contact of the cone with the surface. The center of mass is at a distance 3h/4
along the axis of the cone, which then will be at a vertical height a = 3h sin /4. The total
mass of the cone is h3 tan2 /3.
Each cross section of the cone is like a disk perpendicular to the cones axis (and thus
tilted with respect to the vertical), moving in a circle. The velocity of the center of mass
has magnitude v, and is related to the angular velocity is = v/a sin . We also know that,
if there is no slipping, the velocity f the center of mass is related to the angular velocity as
v = a cos , where is the angle along the circle, and = 2/ . Then = cot .
One principal axis of the cone is along its axis, say x3 , with moment of inertia I3 ;
the other two axes are in the plane perpendicular to its axis, with moments of inertia
5

I1 = I2 = I. The angular velocity will have a component cos along the cone axis, and
a component sin on the plane perpendicular to the axis.
The moments of inertia with respect to the cones axis are I30 = 3M R2 /10, I 0 =
3M (h2 +R2 /4)/5, where R = h tan is the radius of the base. The moments of inertia with
respect to the center of mass at a distance a = 3h/4 along axis, are I3 = I30 = 3M R2 /10,
I = I 0 M a2 = 3M (R2 + h2 /4)/20.
If we choose the center of mass as the origin of the body axes, the kinetic energy is
T = (1/2)M v 2 + (1/2)I3 2 cos2 + (1/2)I1 2 sin2 = 3M h2 2 (1 + 5 cos2 )/40.

Exercise 5-18
A weightless bar of length l has two masses of mass m at the two ends, and is rotating
uniformly about an axis passing through the bars center, making an angle with the bar.

Figure 3: Exercise 5-18

The principal body axes for a bar are one along the bar itself, say the z 0 axis, and the
other two axes, x0 and y 0 , in the plane perpendicular to the bar. The principal moments of
inertia are then I3 = 0, and I1 = I2 = 2ml2 . Since the axis of rotation is not perpendicular
to the bar, the angular velocity (along the axis of rotation by definition) will have a
component along the bar, // = 3 = cos and a component in the plane perpendicular
to the the bar, = sin . We can choose the x0 axis along the component of the angular
velocity in the plane perpendicular to the bar, so 1 = sin and 2 = 0. For uniform
6

velocity, Euler equations are then

N1 = I1 1 2 3 (I2 I3 ) = 0
N2 = I2 2 3 1 (I3 I1 ) = 2ml2 2 sin cos
N3 = I3 3 1 2 (I1 I2 ) = 0
The torque is then along the y 0 axis: a direction perpendicular to the bar, and perpendicular to the angular velocity (which is in the x0 z 0 plane).
We can also calculate the torque from the time derivative of the angular momentum
vector in an inertial frame. If we set up an inertial frame with the origin in the center of
the bar, and the z axis along the rotation axis, then the masses have position vectors
r1 = l(sin cos(t), sin sin(t), cos )
r2 = l( sin cos(t), sin sin(t), cos ) = r1 .
The velocities are
v1 = l( sin sin(t), sin cos(t), 0)
v2 = v1 .
The angular momentum of each mass is
L1 = r1 mv1
= ml2 (sin cos(t), sin sin(t), cos ) ( sin sin(t), sin cos(t), 0)
= ml2 sin ( cos cos(t), cos sin(t), sin )
L2 = r2 mv2 = L1
The total angular momentum is
L = L1 + L2 = 2ml2 sin (cos cos(t), cos sin(t), sin )
and the total external torque
N=

dL
= 2ml2 2 sin cos (sin(t), cos(t), 0)
dt

The torque is a rotating vector, perpendicular to the angular velocity (which is along
the z axis), and perpendicular to the bar (which is along r1 ), just as we had obtained from
Eulers equations.

5-20: A plane physical pendulum

Consider a uniform rod of length l and mass m, suspended in a vertical plane by one end.
At the other end, there is a uniform disk of radius a and mass M attached, which can
rotate freely in the vertical plane. The systems configuration can be described at any time
with two angles: the angle the rod makes with the vertical direction, and the angle a
reference direction on the the disk makes with the vertical. If the disk is rigidly attached
to the rod, then = . Let us take an inertial system with the origin at the suspension
point, a horizontal x axis, and the y axis pointing down along the vertical direction. The
The kinetic energy of
velocity of a mass element dm a distance s along the rod will be s.
the rod is then
Z l
s2 2 ds = (1/2)(m/l)(l3 /3)2 = ml2 2 /6.
T rod = (1/2)(m/l)
0

The velocity of points in the disk are equal equal to r a + a r0 , where r0 is the position
vector of the mass element in the reference frame fixed to the rotating disk, ra is the position
of the center of mass of of the disk, and the the angular velocity vector a describes the
rotation of the disk in an inertial system. The magnitude of the angular velocity vector a
and its direction is perpendicular to the motion plane. The squared speed of mass
is =,
elements will then be v 2 = va2 + 2ra ( a r0 ) + a2 r02 , and the kinetic energy of the disk
will be
Z
Z
1
1
2
T disk =
v dm =
(va2 + 2ra ( a r0 ) + a2 r02 )dm.
2
2
We recognize that the third term will lead to a term (1/2)I0 a2 in the kinetic energy,
with I0 = M a2 /2 the moment of inertia of the disk with
R respect to the center of mass.
The integral of the second term will vanish, since r0 dm is the position of the center
of mass in a coordinate system where the center of mass is at the origin.
The velocity of the center of mass of the disk r a is equal to r 0 + a a, where r0 is the
position vector of the attachment point, and a is the position of the center of mass of the
disk with respect to the attachment point. Then
va2 = (r0 + a a) (r0 + a a)
= v02 + 2v0 ( a a) + a2 a2
= l2 2 + 2 a (a v0 ) + a2 2
The velocity of the attachment point v0 is tangent to the disk, so the direction of the
cross product a v0 is perpendicular to the plane of motion, and a (a v0 ) = a |a v0 |.
The position vector of the attachment point is r0 = l(cos , sin , 0), and its velocity vector
sin , cos , 0). The position of the center of mass of the disk with respect to
is v0 = l(
the attachment point is a = a(cos , sin , 0). Thus,
sin , cos , 0)
a v0 = a(cos , sin , 0) l(
8

 0, cos cos + sin sin )

= al(0,

= al cos( )k
The kinetic energy of the disk is then
Tdisk =
=
=

1
M va2 + I0 a2
2
1
1 M a2 2
M (l2 2 + 2al cos( ) + a2 2 ) +

2
2 2
1
3
M l2 2 + M al cos( ) + M a2 2
2
4

and the total kinetic energy is

T

= Trod + Tdisk
3
1 2 2 1

) + M a2 2
ml + M l2 2 + M aldotcos(
=
6
2
4
1
3
2 2
2

=
(3M + m)l + M alcos(
) + M a 2
6
4

The potential energy of the rod is Vrod = mg(l/2) cos , and the potential energy of the
disk is Vdisk = M g(l cos + a cos ), so the total potential energy is
1
1
V = mgl cos M g(l cos + a cos ) = (m + 2M )gl cos M ga cos
2
2
The Lagrangian is
1
3
1
L = T V = M l2 2 + M al cos( ) + M a2 2 + (m + 2M )gl cos + M ga cos
2
4
2
Lagranges equation for is:

 
1
d  2

M l + M al cos( ) M al sin( ) (m + 2M )glsin = 0

dt
2
1
M l2 + M al cos( ) M al 2 sin( ) + (m + 2M )gl sin = 0
2
Lagranges equation for is

 

d 3
2

M a + M al cos( ) M al sin( ) M gasin = 0

dt 2
3
M a2 + M al cos( ) + M al2 cos( ) + mga sin = 0
2
9

Exercise 25: A rolling sphere

The sphere is a rigid body described by the position of its center of mass r = (x, y, z), and
its orientation, defined by Euler angles , , . The constraint of the sphere remaining on
the surface (the effect of the normal force responding to gravity) is
z=R
(if the origin of the coordinate system is on the surface). At any instant, the sphere is
rotating about an axis on its point of contact with the surface, which is the instantaneous
axis of rotation and the direction of the angular velocity vector , and thus = (x , y , z ).
If the sphere is not slipping, the angular velocity vector is perpendicular to the velocity
of the center of mass, v = 0, which we can express as a condition for an angle to
exist (defining the direction of the horizontal component of the angular velocity) such that
= ( cos , sin , z ) and v = v(sin , cos ). Moreover, not slipping also means
that the instantaneous velocity of the contact point is zero:
0 = vc = v + r0
= v(sin , cos , 0) + ( cos , sin , z ) (0, 0, R)
= (v R)(sin , cos , 0)
which then means
v = R.
The angular velocity vector is related to Euler angles, and thus our constraints are
Rx = v cos = y = R( cos + sin sin )
Ry = v sin = x = R( sin sin cos )
It may be useful to compare these conditions with the problem of a rolling disk considered
in Chapter 1: our equations are the same as (1.39), since the disk would have = 0.
but that is not a constraint, of course. The angular
We also know that z = cos + ,
velocity z is the instantaneous spinning velocity of the sphere about a vertical axis.
The sphere then has 6 coordinates, and three constraints, so the system only has three
degrees of freedom. Wed like to choose these as the three Euler angles, for example, using
the two costraints to solve for the x, y. However, the constraints are non-holonomic and
do not allow us to integrate them for x, y. To prove the constraints are non-holonomic, we
consider the differentials
dfx (x, , , ) = R(d sin d sin cos ) dx
dfy (y, , , ) = R(d cos + d sin sin ) + dy

10

and we prove they are not exact differentials (they are not derivatives of a function):
d dfx
d
d dfx
=
R sin = R cos 6=
=0
d d
d
d d
d dfy
d
d dfy
=
R cos = R sin 6=
=0
d d
d
d d
We now want to write Lagranges equations of motion, using Lagrange multipliers.
The potential energy is constant, so we only have kinetic energy. The kinetic energy has a
translational part, and a rotational part. The rotational energy is especially simple since
the sphere has identical moments of inertia about the principal axes:
L=T

=
=
=

1
1
mv v + I
2
2
1
1
2
2
2)
m(x + y ) + I(( cos + sin sin )2 + ( sin sin cos )2 + ( cos + )
2
2
1
1
m(x 2 + y 2 ) + I(2 + 2 + 2 + 2 cos )
2
2

The non-holonomic constraints are

Fx = R( sin sin cos ) x
Fy = R( cos + sin sin ) + y

Lagranges equations are

Fx
= x
x
Fy
= y
y
y
Fy
Fx
x
+ y


R(x sin + y cos )
Fy
Fx
x
+ y

m
x = x
m
y =
I + I sin =
=
I + I( cos sin ) =

= R sin (x cos + y sin )

dz
d
I + I( cos sin ) = 0 = I ( + cos ) = I
dt
dt
We have then 7(!) differential equations (five Lagrange equations and two constraints)
for the eight unknowns x, y, , , , x , y .

11

The last equation (Lagranges equation for ) says that the angular velocity z is
constant. This is because is a cyclical variable in the Lagrangian and because the
Notice that is also a cyclical variable in the Lagrangian,
constraints do not depend on .

but it is not associated with a conserved quantity because the constraints depend on .
Since there is no dissipation or forces doing any work, the energy (equal to the kinetic
energy) is conserved:
1
1
1
1
E = T + V = Ttr + Trot + 0 = mv v + I = mv 2 + I(2 + z2 )
2
2
2
2
Due to the constraint v = R, the energy is equal to E = 12 (mR2 +I)2 + 12 Iz2 . Since z is
constant, then we know that (and thus also v) are constant. Therefore, the translational
kinetic energy 12 mv 2 and the rotational energy 21 I(2 + z2 ) are separately conserved.

Exercise 30: Closing tilted door

The doors axis of rotation is along the hinges, which make an angle with the vertical
(normally, on firm ground and for a well aligned door, of course, = 0). The position
of the door as a rigid body is determined by the angle about its axis of rotation. The
principal axes of the door are an axis perpendicular to the door, and two axes in the door
plane, parallel to the doors sides (assumed to be straight) .
We choose a body coordinate system with an origin in the bottom doors corner along
the hinged side. We choose an inertial coordinate system with the same origin, a vertical
axis z pointing up, and two perpendicular axis x, y in the horizontal plane. We choose
body axes x0 along the short side of the door, z 0 along its hinged side (the axis of rotation),
and y 0 perpendicular to the door.
Following the convention for Euler angles, and placing the door in the x0 z 0 plane in
Figure 4.7, we see that the angle is the angle between the door and the vertical; the angle
= 0, and the angle is the one that defines the position (angle) of the door.
The initial position of the door (lifted by 90 degrees) has an angle = 0. The final
equilibrium position of the door (when the door is shut) has an angle = /2.
The gravitational potential energys is V = mg r0 = mgz, where r0 is the position of
the center of mass in the inertial system, and z its vertical component.
The position of the center of mass in the body axes system is r00 = (x0 , y 0 , z 0 ) =
(w, 0, h)/2, where w is the doors width and h is the doors height. In the inertial system, the position of the center of mass is given by the transformation (4.47), r0 = A1 r00 .
The vertical coordinate will be
z = x0 sin sin + y 0 sin cos + z 0 cos = (w/2) sin sin + (h/2) cos
The gravitational potential energy is then
V = mgz = mg(w sin sin + h cos )/2
12

Since the motion is a pure rotation, the kientic energy has the form T = (1/2)I0 2 ,
where I0 is the moment of inertia about the axis of rotation (the doors hinged side, the
z 0 axis), and = is the angular velocity: Trot = 12 Iz 0 2 . The moment of inertia of the
door with respect to the z 0 axis, is calcualted integrating over the points on the door (all
with y 0 = 0 coordinates:
Z
Z
Z
m w3
m w 02 0 h 0
1
02
02
0 0
dz =
Iz 0 = (x + y )dx dz =
x dx
h = mw2 .
wh 0
wh 3
3
0
The kinetic energy is then
1
1
T = Iz 0 2 = mw2 2 .
2
6
The total energy is
mg
1
(w sin sin + h cos ).
E = T + V = mw2 2 +
6
2
In the initial position, = 0 and = 0. Since the energy is conserved, we obtain
mg
1
mgh
E = mw2 2 +
(w sin sin + h cos ) =
cos ,
6
2
2
an expression we can use as a differential equation for :
3g sin
2 =
sin
w
Notice that since /2 < < 0, the expression on the left hand side is a positive
expression. The angular velocity when the door reaches the equilibrium position at =
/2 (where it will not stop, but oscillate about, if it can go through the shut position) is
then
r
3g sin
f =
.
w
The time it will take to reach that position can be obtained the equation for :
r
d
3g sin
=

sin
dt
w
r
w
d

dt =
3g sin sin
r Z /2
w
d 0

t =
3g 0
sin
r
Z 0
w
d 0

=
3g sin /2 sin 0
13

w
2F (/4, 2)
3g sin
r
w
= 1.51
g sin




w 1.51 2
0.9m
1.51 2
sin =
=
= 1.3
g
t
9.8m/s2
3s
=

Three seconds is a looong time for a door to close, so we obtain a small hinge angle.
The smallest
p time for the door to close is when the hinges are horizontal, = /2 and
t = 1.51 w/g = 0.45sec. The wider the door, the longer it takes to close.

14