You are on page 1of 21

EECS 117

Lecture 16: Magnetic Flux and Magnetization

Prof. Niknejad University of California, Berkeley

EECS 117 Lecture 16 p. 1/2

Magnetic Flux
Magnetic ux plays an important role in many EM problems (in analogy with electric charge)
=
S
C

B dS

=
S

B dS 0

but net ux can certainly cross an open surface.

University of California, Berkeley

Magnetic Flux and Vector Potential

S2

S1

Magnetic ux is independent of the surface but only depends on the curve bounding the surface. This is easy to show since
=
S

B dS =
S

A dS =
C

A d

=
S1

B dS =
S2

B dS

I1 B field

S2

Consider the ux crossing surface S2 due to a current owing in loop I1

21 =
S2

B1 dS

Likewise, the self-ux of a loop is dened by the ux crossing the surface of a path when a current is owing in the path
11 =
S1
University of California, Berkeley

B1 dS

The Geometry of Flux Calculations

The ux is linearly proportional to the current and otherwise only a function of the geometry of the path To see this, lets calculate 21 for lamental loops
21 =
C2

A1 d2

but
1 A1 = 41 0
C1

I1 d1 R R1

substituting, we have a double integral

21 I1 = 41 0 d1 R R1 d2

C2

C1

Geometry of Flux (cont)

R2 - R 1 d1 R1 R2 d2

We thus have a simple formula that only involves the magnitude of the current and the average distance between every two points on the loops
21 I1 = 41 0 d1 d2 R2 R1

C2

C1

Mutual and Self Inductance

Since the ux is proportional to the current by a geometric factor, we may write
21 = M21 I1

We call the factor M21 the mutual inductance

M21 21 1 = = I1 41 0 d1 d2 R2 R1

C2

C1

The units of M are H since [] = H/m. Its clear that mutual inductance is reciprocal, M21 = M12 The self-ux mutual inductance is simply called the self-inductance and donated by L1 = M11
University of California, Berkeley

System of Mutual Inductance Equations

If we generalize to a system of current loops we have a system of equations
1 = L1 I1 + M12 I2 + . . . M1N IN . . . N = MN 1 I1 + MN 2 I2 + . . . LN IN

Or in matrix form = M i, where M is the inductance matrix. This equation resembles q = Cv, where C is the capacitance matrix.

Solenoid Magnetic Field

We have seen that a tightly wound long long solenoid has B = 0 outside and Bx = 0 inside, so that by Ampres law
By = N I0

where N is the number of current loops crossing the surface of the path. The vertical magnetic eld is therefore constant
N I0 By = = 0 In
University of California, Berkeley

EECS 117 Lecture 16 p. 9/2

Solenoid Inductance
The ux per turn is therefore simply given by
turn = a2 By

The total ux through N turns is thus

= N turn = N a2 By N 2 a2 I = 0 The solenoid inductance is thus 0 N 2 a2 L= = I

EECS 117 Lecture 16 p. 10/2

Coaxial Conductor
In transmission line problems, we need to compute inductance/unit length. Consider the shaded area from r = a to r = b
S

The external ux (excluding the volume of the ideal conductors) is given by

b

=
a

0 I B dr = 2
University of California, Berkeley

b a

dr 0 I b = ln r 2 a

Coaxial Transmission Line (cont)

The inductance per unit length is therefore
0 L = ln 2

b a

[H/m]

Recall that the product of inductance and capacitance per unit length is a constant
1 LC = 2 c

where c is the speed of light in the medium. Thus we can also calculate the capacitance per unit length without any extra work.

EECS 117 Lecture 16 p. 12/2

Magnetization Vector
Wed like to study magnetic elds in magnetic materials. Lets dene the magnetization vector as
M = lim mk V
k

V 0

where mk is the magnetic dipole of an atom or molecule The vector potential due to these magnetic dipoles is given by in a differential volume dv is given by
M r dA = 0 dv 2 4R

so
0 A= 4
University of California, Berkeley

M r dv 2 R

Vector Potential
Using
0 A= 4

1 R

r = 2 R

M
V

1 R

dv

M R

1 = M + R

1 R

We can thus break the vector potential into two terms

0 A= 4 M 0 dv R 4
V

M R

dv

Another Divergence Theorem

Consider the vector u = a v, where a is an arbitrary constant. Then
u = (a v) = ( a)v( v)a = ( v)a

( v) adV =
V S

((a v) u) dS

(a v n) dS

Another Divergence Thm (cont)

Since the vector a is constant, we can pull it out of the integrals
a
V

( v) dV = a
S

r n dS

( v) dV =
V S

r n dS

V

M R

dv =
S

Mn dS R

Vector Potential due to Magnetization

The vector potential due to magnetization has a volume component and a surface component
0 A= 4 M 0 dv + R 4 Mn dSdv R

We can thus dene an equivalent magnetic volume current density Jm = M and an equivalent magnetic surface current density
Js = M n

Volume and Surface Currents

Js

Js

Js

In the gure above, we can see that for uniform magnetization, all the internal currents cancel and only the magnetization vector on the boundary (surface) contributes to the integral

EECS 117 Lecture 16 p. 18/2

Relative Permeability
We can include the effects of materials on the macroscopic magnetic eld by including a volume current M in Ampres eq
B = 0 J = 0 (J + M)

or
B M=J 0

B M H= 0

The units of H, the magnetic eld, are A/m

University of California, Berkeley

Amperes Equation for Media

We can thus state that for any medium under static conditions H=J equivalently
H d = I
C

Linear materials respond to the external eld in a linear fashion, so M = m H so

B = (1 + m )H = H

or

1 H= B
University of California, Berkeley

EECS 117 Lecture 16 p. 20/2

Magnetic Materials
Magnetic materials are classied as follows Diamagnetic: r 1, usually m is a small negative number Paramagnetic: r 1, usually m is a small positive number Ferromagnetic: r 1, thus m is a large positive number Most materials in nature are diamagnetic. To fully understand the magnetic behavior of materials requires a detailed study (and quantum mechanics) In this class we mostly assume 0