Electromagnetism 70006 Answers to Problem Set 10 Spring 2006

1. Jackson Prob. 5.15: Shielded Bilar Circuit: Two wires carrying oppo-
sitely directed currents are surrounded by a cylindrical shell of inner radius
a, outer radius b, and relative permeability
r
.
(a) Determine the magnetic potential for two wires; the rst is located
at x = d/2 and carries current I in the -z direction and the second
is located at x = d/2 and carries current I in the z direction.

m
=
I
2
ln
_
_

2
+d
2
/4 d cos
_

2
+d
2
/4 +d cos
_

I
2
cos

,
where is measured counter-clockwise from the x axis.
(b) Find the potential in the three regions. We may assume that only
terms in the expansion of the potential in cylindrical coordinates
proportional to cos contribute:

m
=
_
A +

_
cos < a
=
_
B +
C

_
cos a < < b
=
D

cos b < ,
where = I/(2) and where (A, B, C, D) are unknown expansion
coecients to be determined by boundary coecients on the two
surfaces = a and = b. These conditions;
m
continuous and
normal component of B continuous, lead to the equations:
A+

a
2
= B +
C
a
2
A+

a
2
=
r
_
B +
C
a
2
_
B +
C
b
2
=
D
b
2

r
_
B +
C
b
2
_
=
D
b
2
Solving, we nd

a
=
_
(a
2
b
2
)
_

2
r
1
_
a
2
(b
2
(
r
+ 1)
2
a
2
(
r
1)
2
)
+

_
cos

b
=
_
2(
r
1)
b
2
(
r
+ 1)
2
a
2
(
r
1)
2
+
2b
2
(
r
+ 1)
b
2
(
r
+ 1)
2
a
2
(
r
1)
2
1

_
cos

c
=
_
4b
2

r
b
2
(
r
+ 1)
2
a
2
(
r
1)
2
1

_
cos
1
Substituting into the earlier expression gives explicit results for
m
in each region. In particular, outside the shield we nd a dipole
potential with coecient proportional to that of the two wires ();
the coecient of proportionality is
F =
4b
2

r
b
2
(
r
+ 1)
2
a
2
(
r
1)
2
.
In problem 5.14, a uniform external eld maintained its form but
was reduced in strength inside a cylindrical shield. Here an internal
dipole eld maintains its form but is reduced in strength outside a
cylindrical shield.
(c) For
r
1 and b = a+t with t b, we nd (
r
= 200, b = 1.25 cm,
t = 0.3 mm)
F
2b
t
r
= 0.417
2. Jackson: Prob. 5.24: For a conducting plane with a circular hole and a
tangential eld H
0
on one side:
(a) Determine H
(1)
on the side with H
0
for > a. We have for z = 0

(1)
(, ) =
2aH
0

_

0
j
1
(ka)J
1
(k)dk sin
=
H
0

_
sin
1
_
a

_
a

1
a
2

2
_
sin. > a.
For > a, we nd
H

=

(1)

=
H
0

_
_
a
3
_
1
a
2

2

3
+
a
_
1
a
2

2

sin
1
_
a

_
_
_
sin
H

=

(1)

=
H
0

_
a

1
a
2

2
sin
1
_
a

_
_
cos
H
x
= H

cos H

sin =
H
0

a
3

2
sin2
_

2
a
2
=
2H
0

a
3

4
xy
_

2
a
2
2
H
y
= H

sin +H

cos =
H
0

_
a
_

2
a
2
sin
1
_
a

a
3

2
cos 2
_

2
a
2
_
=
2H
0

a
3

4
y
2
_

2
a
2
+
H
0

_
a

2
_

2
a
2
sin
1
_
a

__
(b) Sketch the surface currents above and below the plane. Above the
plane both H
0
and H
(1)
contribute to the current:
(K
x
, K
y
) = (H
0
H
(1)
y
, H
(1)
x
)
while below, only H
(1)
contributes:
(K
x
, K
y
) = (H
(1)
y
, H
(1)
x
)
3. Jackson Prob. 5.25: A rectangular loop carrying current I
1
interacts with
a wire carrying current I
2
. The center of the loop is a distance d from
the wire and two sides of the loop of length a are parallel to the wire;
the sides of length b make angle with the plane of the wire and the line
from the wire to the center of the loop The direction of the current in the
side nearest the wire is in the same direction as I
2
. Set up a coordinate
system with the loop in the xy plane and center of the loop at the origin;
3
the sides a run parallel to y and are located at x = b/2; the sides b are
parallel to the x axis. The wire located at z = d sin, x = d cos and
I
2
ows along +y. In this coordinate system, the vector potential of the
wire has only a y component and
A
2
=
I
2
4
ln[(x d cos )
2
+ (z d sin)
2
] y
(a) The interaction energy is
W
12
= I
1
_
dl
1
A
2
=

0
I
1
I
2
4
a ln
_
(b/2 d cos )
2
+ (d sin)
2
(b/2 d cos )
2
+ (d sin)
2
_
=

0
I
1
I
2
4
a ln
_
4d
2
+b
2
+ 4db cos
4d
2
+b
2
4db cos
_
where only the two sides parallel to y contribute.
(b) Calculate the force on the loop We have in the xy plane
B
x
(x, 0) =
A
y
z

z=0
=
d sin
(x d cos )
2
+d
2
sin
2

B
z
(x, 0) =
A
y
x

z=0
=
(x d cos )
(x d cos )
2
+d
2
sin
2

The force on the two sides of the rectangle of length b precisely cancel.
The x and z components of the force on the two sides of length a are
F
x
= I
1
[B
z
(b/2, 0) B
z
(b/2, 0)] =
2
0
I
1
I
2
ab
_
4d
2
cos(2) b
2
_
(b
4
8d
2
cos(2)b
2
+ 16d
4
)
F
z
= I
2
[B
x
(b/2) B
z
(b/2, 0) =
8
0
I
1
I
2
ab d
2
sin(2)
(b
4
8d
2
cos(2)b
2
+ 16d
4
)
(c) Repeat for the case where the rectangle of sides a, b is replaced by a
circle of radius a. In this case, we write
W
12
= I
1
_
2
0
a cos A
y
(a cos , 0) d,
where we have used the fact that x = a cos and dl
y
= a cos d
along the circle. We expand the A
y
in a series in powers of 1/d and
carry out the integral term by term to nd
W
12
= I
1
I
2

0
a
_
cos()a
2d
+
cos(3)a
3
8d
3
+
cos(5)a
5
16d
5
+
5 cos(7)a
7
128d
7
+
7 cos(9)a
9
256d
9
_
4
The same series results if we evaluate W
12
= I
1

2
, where
2
is the
magnetic ux through the circle. Note that the term in parentheses
above can be written
_

_
=
_
z
2
+
z
3
8
+
z
5
16
+
5z
7
128
+
7z
9
256
_
,
Where
z =
a
d
e
i
Moreover,
1

1 z
2
z
=
z
2
+
z
3
8
+
z
5
16
+
5z
7
128
+
7z
9
256
Thus, we may write
W
12
= I
1
I
2

0
a
_
1

1 z
2
z
_
with z =
a
d
e
i
This correct answer is close (but not identical) to the answer given
in the text. Indeed, if we assumed

_
1

1 z
2
z
_
=
1
_
1 (z)
2
z
,
then we would recover the result in the text.
Find the force.
F =

i I
1
_
2
0
a cos B
z
(a cos , 0) d

k I
1
_
2
0
a cos B
x
(a cos , 0) d
Again, expanding the potential and carrying out the integrations
leads to We nd
F
x
=
0
I
1
I
2
_
cos(2)a
2
2d
2
+
3 cos(4)a
4
8d
4
+
5 cos(6)a
6
16d
6
+
35 cos(8)a
8
128d
8
+
63 cos(10)a
10
256d
10
_
F
z
=
0
I
1
I
2
_
sin(2)a
2
2d
2
+
3 sin(4)a
4
8d
4
+
5 sin(6)a
6
16d
6
+
35 sin(8)a
8
128d
8
+
63 sin(10)a
10
256d
10
_
Again, we can identify the two series: Consider the function
G(z) =
1

1 z
2
1 =
z
2
2
+
3z
4
8
+
5z
6
16
+
35z
8
128
+
63z
10
256
5
Comparing, we nd
F
x
=
0
I
1
I
2
G
_
a
d
e
i
_
F
z
=
0
I
1
I
2
G
_
a
d
e
i
_
(d) Express the energies for large d in terms of moments of loops. For
the rectangular loop:
W
12
=

0
I
1
I
2
4
a ln
_
4d
2
+b
2
+ 4db cos
4d
2
+b
2
4db cos
_

0
I
1
I
2
4
a
2b cos
d
= (I
1
ab)

0
I
2
2d
cos = m
1
B
2z
For the circular loop:
W
12
= I
1
I
2

0
a
_
cos()a
2d
+
cos(3)a
3
8d
3
+
cos(5)a
5
16d
5
+
I
1
I
2

0
a
cos()a
2d
= (I
1
a
2
)

0
I
2
2d
cos = m
1
B
2z
In both cases, the + sign is a result of the fact that the moment and
the normal component of the eld are in opposite directions.
4. Jackson Prob. 5.34: Two identical circular loops are located a distance R
apart on a common axis,
(a) Find M
12
using A

from Prob. 5.10b:

A

(, z) =

0
I
1
a
2
_

0
J
1
(ka)J
1
(k)e
k|z|
dk
W
12
= I
2
_
2
0
a dA

(a, R) =
0
I
1
I
2
a
2
_

0
J
1
(ka)J
1
(ka)e
kR
dk
Leading to the result
M
12
=
0
a
2
_

0
[J
1
(ka)]
2
e
kR
dk
(b) Assuming a << R, we obtain an asymptotic series in R by expanding
J
(
ka) in a power series
[J
1
(ka)]
2

a
2
k
2
4

a
4
k
4
16
+
5a
6
k
6
768

7a
8
k
8
18432
Integrating, we nd
M
12
=

0
a
2
_
a
3
R
3

3a
5
R
5
+
75a
7
8R
7

245a
9
8R
9
+
_
6
(c) Find the mutual inductance for co-planer loops with centers sepa-
rated by R. The axial B
z
eld from the loop centered at the origin
is
B
z
(z) =

0
I
1
2
a
2
[a
2
+z
2
]
3/2
This eld can be derived from a scalar potential

m
(z) =

0
I
1
2
_
1
z

z
2
+a
2
_
=

0
I
1
2
_
a
2
2z
2

3a
4
8z
4
+
5a
6
16z
6

35a
8
128z
8
+
63a
10
256z
10
+
_
.
Analytically continuing the potential leads to

m
(r, ) =

0
I
1
2
_
a
2
2r
2
P
1
(cos )
3a
4
8r
4
P
3
(cos ) +
5a
6
16r
6
P
5
(cos )

35a
8
128r
8
P
7
(cos ) +
63a
10
256r
10
P
9
(cos ) +
_
We need B
z
at large values of r and = /2. We nd
B
z
(r, /2) =
1
r

=/2
=

0
I
1
2
_
a
2
2r
3
+
9a
4
16r
5
+
75a
6
128r
7
+
1225a
8
2048r
9
+
19845a
10
32768r
11
_
Introduce the vector centered on the second loop. Then we may
write bmr = R+, where R is the vecror from the center of the rst
loop to the center of the second. We may replace
r
_
R
2
+
2
+ 2R cos
where is the polar angle with respect to the center of the second
loop and carry out a second expansion of B
z
with respect to R. With
this in hand, we calculate the ux
2
the through the second loop.
First, integrating B
z
over the polar angle , we obtain
_
2
0
B
z
d =

0
I
1

2
_
a
2
R
3
+
_
9a
4
8
+
9
2
a
2
4
_
1
R
5
+
_
75a
6
64
+
225
2
a
4
32
+
225
4
a
2
64
_
1
R
7
+
_
1225a
8
1024
+
3675
2
a
6
256
+
11025
4
na
4
512
+
1225
6
a
2
256
_
1
R
9
_
7
To evaluate the ux through the second loop, we integrate the pre-
vious result over

2
=
_
a
0
d
_
2
0
B
z
d =

0
I
1
a
4
_
a
3
R
3
+
9a
5
4R
5
+
375a
7
64R
7
+
8575a
9
512R
9
_
Given that
2
= M
12
I
1
, we may write
M
12
=

0
a
4
_
a
3
R
3
+
9a
5
4R
5
+
375a
7
64R
7
+
8575a
9
512R
9
_
(d) calculate the force in each case. For the co-planar loops, the only
non-vanishing component of the force on the second loop is
F
x
= I
2
_

0
a cos B
z
( = a, )d
=

0
I
1
I
2

2
_
3a
4
2R
4
+
45a
6
8R
6
+
2625a
8
128R
8
+
77175a
10
1024R
10
_
The force is repulsive and along the line joining the centers of the
loops.
In the case of the co-axial loops, only the component F
z
of the force
on the second loop contributes:
F
z
= cos F
r
sinF

.
Now, at the location of the second loop, components of the force are
F
r
= 2a I
2
B

(r, )
F

= 2a I
2
B
r
(r, ),
where r =

a
2
+z
2
and = arccos(z/

a
2
+z
2
). Therefore,
F
z
= 2a I
2
(cos B

(r, ) sinB
r
(r, ))
Substituting and expanding the elds in z, one obtains
F
z
= I
1
I
2

3a
4
2R
4
+
15a
6
2R
6

525a
8
16R
8
+
2205a
10
16R
10
_
The force is attractive and along z.
8