Sei sulla pagina 1di 209

ECE 303 - Sum10 Notes Set 1: Transmission Lines and the Wave Equation 1

INTRODUCTION

One purpose of a transmission line is to transport electromagnetic energy from one location
to another. For example, we might wish to “send” a digital bit from CPU to a memory
chip in a computer or “send” an analog television signal down a coaxial cable. We can
understand many applications by using the concepts of current and voltage from basic
circuits. This approach is the main topic of this set of Notes, which leads to a very important
mathematical result called the Wave Equation.

This set of Notes shows how circuit theory, which you studied in ECE 200 and ECE 211,
can be used to derive a mathematical characterization of transission lines. For future refer-
ence, here are some areas where transmission line analysis is sometimes used:

Coaxial Cables
Coaxial cables provide connection media in the Cable TV and Cable Modem net-
works. Coax can also connect RF (Radio Frequency) devices like antennas to display
devices, like televisions.

Twisted-Pair Wire Pairs


Twisted pairs of insulated wires, usually copper, is used to connect wireline tele-
phones to an access point in the Public Switched Telephone Network. Twisted-pair
copper is also used in Digital Subscriber Line (DSL) internet access. Twisted-pair
can also be used in short-length high-speed computer networks.

Computer Bus
As computer systems become “faster”, the pulse widths become shorter and the con-
ducting paths in a computer system become transmission lines. You can think of a
16-bit bus as being 16 parallel transmission lines, located very close together.

We will first derive the general transmission lines equations. After that we will study
some of the mathematical properties of simple lossless transmission lines. Properties and
characteristics of other lines will be the topics of later sets of Notes.
ECE 303 - Sum10 Notes Set 1: Transmission Lines and the Wave Equation 2

TRANSMISSION LINES

Most metallic transmission lines have two conductors in proximity, which are separated by
an insulator. Here are physical forms for some common transmission lines (for clarity the
insulator material is not shown):

COAXIAL CABLE

TWISTED−PAIR
COPPER
x=0 x x +∆ x

There is insulation between the conductors in both examples shown above. The twisted-
pair of lines do not touch in the diagram; the view is like we are looking down on the
wire-pair. The length from x to x + ∆x is called an infinitestimal section. You can think of
the line as being composed of a large number of these infinitesimal sections.

QUESTION:
In order to anlayze and design many devices, a property we need to understand is how
voltage v(x; t ) and current i(x; t ) are related for the transmission line. For example, for the
resistor we have v(t ) = i(t ) R. We need a similar result for the transmission line.

APPROACH:
) Use Kirchoff’s laws to relate voltage and current over the infinitesimal section ∆x.
) Extrapolate results to find the differential equation relating voltage and current for any
value of x.
ECE 303 - Sum10 Notes Set 1: Transmission Lines and the Wave Equation 3

CIRCUIT MODEL

In this course, we will consider transmission lines having two metallic conductors in prox-
imity. One useful circuit model of a ∆x section for this type of transmission line is shown
below. We will discuss in class the physical origin of the parameters in the figure:

i(x,t) L ∆x R∆x i(x + ∆ x, t)

+ +
v(x, t) v(x + ∆ x , t)
G∆ x C ∆x
− −
x x + ∆x
The above circuit models the transmission line as having resistance. inductance, capaci-
tance, and conductance for any infinitesimal segment of the line. In this circuit model the
parameters are:
R = RESISTANCE per unit length ( like Ω/m )
L = INDUCTANCE per unit length ( like H/m )
C = CAPACITANCE per unit length ( like F/m )
G = CONDUCTANCE per unit length ( like S/m )

These parameters quantify the series and parallel properties of the line:

 You can think of R and L as series, or in-line, parameters of the conductors. For
example, a wire has a series resistance.
 You can think of C and G as parallel parameters between the two conductors. For
example, two conductors close together form a capacitor.

We next use Kirchoff’s Laws to obtain two equations in the two unknowns v(x; t ) and i(x; t ).
ECE 303 - Sum10 Notes Set 1: Transmission Lines and the Wave Equation 4

KIRCHOFF’S LAWS

LOOP EQUATION:
To derive one of the needed equations, use Kirchoff’s Voltage Law around the outer loop
of the ∆x section. This procedure obtains the following equation:


v(x; t ) + R∆x  i(x; t ) + L∆x  i(x; t ) + v(x + ∆x; t ) = 0 (1 )
∂t
Note that we have to use a Partial Derivative in equation (1) to quantify the voltage across
the inductor. This is because we have two independent variables, x and t, and we are
keeping x constant while we take the derivative with respect to t. Next, divide both sides of
(1) by ∆x and rearrange the resulting terms:

v(x + ∆x; t ) v(x; t ) ∂


= R i(x; t ) L i(x; t ) (2)
∆x ∂t
Now take the limit of both sides of (2) as ∆x ! 0. The left-hand side of (2) becomes a
partial derivative:
∂ ∂
v(x; t ) = R i(x; t ) L i(x; t ) (3)
∂x ∂t
Equation (3) is a Partial Differential Equation (PDE). It tells us the voltage (and current)
are functions of both time and distance on the line.

NODE EQUATION:
To derive the other equation relating v(x; t ) and i(x; t ), use Kirchoff’s Current Law at the
NODE in the ∆x-section and follow the above limit procedure. This process will eventually
give
∂ ∂
i(x; t ) = G v(x; t ) C v(x; t ) (4)
∂x ∂t

Problem: 1-1
ECE 303 - Sum10 Notes Set 1: Transmission Lines and the Wave Equation 5

TRANSMISSION LINE EQUATIONS

The two equations (3) and (4) derived on the previous page are sometimes collectively
called the Transmission Line Equations. Although v(x; t ) and i(x; t ) are strictly functions
of both x and t, there is a simpler notation we can use for our course. To simplify the
notation, let v(x; t ) be replaced by simply v and let i(x; t ) be replaced by simply i. Using
these simplifications, the two partial differential equations in (3) and (4) then become

∂v ∂i
= Ri L (5a)
∂x ∂t
∂i ∂v
= Gv C (5b)
∂x ∂t

We will call the partial differential equations in (5) the Transmission Line Equations. They
are the first of many partial differential equations we will be using in this course. The
solutions v = v(x; t ) and i = i(x; t ) from (5) give the manner in which the voltage and
current vary as a function of time t and distance x on the transmission line.

In these partial differential equations, one position on the line may have a different voltage
and/or different current than another position on the line. You may have seen this type
of behavior before: think of a string vibrating between two fixed end points (like a guitar
string). The parts of the string near the middle, away from the fixed ends, vibrate with the
maximum amplitude, while the portions of the string near the fixed ends vibrate very little.
Thus, the amplitude A = A(x; t ) of the vibrating string becomes a function of both position
x on the string as well as time t.

This vibrating string example is mentioned for familiarity. The electrical transmission line
has a different mathematical form for its solution, which is what we will discuss next.
ECE 303 - Sum10 Notes Set 1: Transmission Lines and the Wave Equation 6

WAVE EQUATION SOLUTIONS

In general, none of R; L; G, or C in (5) are zero, and the solution process for (5) to obtain
v(x; t ) and i(x; t ) is quite involved. There is a set of Supplementary Notes accessible from
the course Home Page called “Solution of Transmission Line Equations” which investigates
actually solving the equations in (5). If you are interested in this solution method these
Supplementary Notes should be useful. However, in the work which follows we will just
use the solutions as needed to understand properties of transmission lines.

Whereas the solution of (5) for general R; L; G, and C is quite complicated, there is one
choice for R and G which leads to a very simple math solution and which describes a
situation common to many applications of transmission lines. Such a situation arises for
the line having R = 0 and G = 0 in (5). (We will discuss in class situations in which
this is a valid approximation.) This line is called the Lossless Line and the PDE resulting
from R = 0 and G = 0 is sometimes called the Wave Equation. To develop this equation,
substitute R = G = 0 in equations (5a) and (5b), giving

∂v ∂i
= L (6a)
∂x ∂t
∂i ∂v
= C (6b)
∂x ∂t
It is straightforward to show that the two first-order partial differential equations in (6) are
equivalent to one second-order partial differential equation given by

∂2 v ∂2 v
= LC (7 )
∂x2 ∂t 2
Equation (7) for voltage v = v(x; t ) is sometimes called the Wave Equation. (There is also a
second-order wave equation for current i.) On the next page we discuss the solution to the
Wave Equation.

Problems: 1-2, 1-3


ECE 303 - Sum10 Notes Set 1: Transmission Lines and the Wave Equation 7

SOLUTION PROPERTIES

The solution v(x; t ) to (7) has a very simple form. Let g(t ) be any voltage waveshape which
can be successfully introduced to the point x = 0+ on the infinite length transmission line.
Then the solution for voltage and current at any point x and at any time t is
 x
v = v(x; t ) = g t (8a)
vp
v(x; t )
i = i(x; t ) = (8b)
R0
where the parameter v p in (8a) is called the Phase Velocity (in m/sec) and the parameter R0
in (8b) is called the Characteristic Resistance. These parameters are computed by
r
vp = p1 ; R0 =
L
C
(9 )
LC
You can see the Supplementary Notes on Solution of Transmission Line Equations for the
specific solution method, as well as the derivation of v p and R0 in (9).

Useful Solutions:
It is straightforward (but somewhat lengthy) to show that the following functions are solu-
tions to the Wave Equation :
v1 (x; t ) = e j ( ω t β x )
v2 (x; t ) = cos(ω t β x + φ)
v3 (x; t ) = sin(ω t β x)
where the parameter β in these equations is called the phase constant and is defined by
ω
β = (10)
vp
We will frequently use these solutions in our work since they are easily differentiable with
respect to both time t and space x.

Problems: 1-4 through 1-8


ECE 303 - Sum10 Notes Set 1: Transmission Lines and the Wave Equation 8

TRANSFORMATION OF VARIABLES

x
We have now seen that transmission lines techniques introduce the “ t ” transforma-
vp
tion into the argument of functions. Let’s now examine in detail the implications of this
transformation.

QUESTION:
x
What does the transformation term “t ” mean in the argument of the “g”
vp
function in equation (8a) ?

ANSWER:
As t increases, this term moves the “g” shape in the direction of propagation, with
a velocity v p .

We can show this property by choosing a simple “g” function and mathematically inves-
x
tigating the effect of t . Let a simple rectangular pulse be defined by
vp

A; 0  α  T
g(α) = (11)
0; else
 x x
To get g t , substitute α = t in equation (11):
vp vp
(
 x A; 0  t T
x
g t = vp (12)
vp 0; else
ECE 303 - Sum10 Notes Set 1: Transmission Lines and the Wave Equation 9

PROPAGATION PROPERTIES

Now let’s locate the non-zero portion of the “g” function on the x-axis.
 x
QUESTION: Where is g t in (12) non-zero on the x-axis?
vp

ANSWER: Solve the inequality on the right side of equation (12) such that only x is in
between the inequality signs, as shown below:

0 t T
x
(13)
vp
. &x
0 t
x
Lower : t < T : Upper
vp vp

t
x x
< T t
vp vp

vp t  x x  vp t vpT

& Combine: .
v p (t T) x vp t (14)

Equation (14) shows that the non-zero portion of the “g” function is localized to a particular
region in x. Note also from (14) that time t is in the upper and lower bounds of this region.
This means that the region, and hence the “g” function, moves as time t changes. We can
think of this as a pulse moving through space as time increases.
ECE 303 - Sum10 Notes Set 1: Transmission Lines and the Wave Equation 10

PROPAGATION EXAMPLE I

Thus we see that the inequality on the right side of (12) is equivalent to the inequality in
(14). Substituting (14) into (12) then gives
 
x A; v p (t T) x vp t
g t = (15)
vp 0; else
x
The argument t moves the “g” waveform to the right as t increases. The movement
vp
of the waveform in space is like a “wave” moving. We next examine two examples which
demonstrate how this movement occurs.

EXAMPLE: Let v p = 2(10)8 m/sec. and let the duration of the pulse T = 1 µsec. Find the
position of the pulse in (15) above at t = 0.

APPROACH: At t = 0 we have v pt = 0 and v p (t T) = 2(10)8 ( 10 6) = -200. Thus,


 
x A; 200 x 0
g t = (16)
vp 0; else

A representation of this pulse is shown below.

x = −200 x=0 x=x


L
physical cable
region

At time t=0, the pulse is “still in the source” ... the pulse has not yet entered the PHYSICAL
LINE. The next example will show how the pulse travels down the line.

Problems: 1-9, 1-10, 1-11


ECE 303 - Sum10 Notes Set 1: Transmission Lines and the Wave Equation 11

PROPAGATION EXAMPLE II

As time t increases, the pulse travels down the transmission line. The example below will
show how this occurs.

EXAMPLE:
As in the previous example, let v p = 2(10)8 m/sec, T = 1 µsec. Find the position on the
line of the pulse at t = 2 µsec.

APPROACH:
At t = 2(10) 6, we have the values given below:
h i
v pt = 2(10)8  2(10) 6
= 400 ; v p (t T) = 2(10)
8
2(10) 6
10 6
= 200

Substitute these values into equation (29), giving


 
x A; 200  x  400
g t = (17)
vp 0; else

This pulse is shown below.

x=0 x = 200 x = 400

As time t increases, the rectangular pulse travels farther to the right (increasing x) until it
reaches the end of the line.

Problems: 1-12, 1-13, 1-14


ECE 303 - Sum10 Notes Set 2: Propagation on Transmission Lines 1

INTRODUCTION

In the previous set of Notes we developed some of the mathematics quantifying the current
and voltage on a transmission line. In some cases, the line may be approximated as the
Lossless Line, which led to the Wave Equation. We then solved the Wave Equation and
briefly examined some of the solution properties.

With the introduction of transmission lines, we now find we have two seemingly different
methods for analyzing electrical systems: (i) Circuit Theory, which you learned about in
ECE 211, and (ii) Transmission Line techniques (like wave theory), which we are currently
studying. One of the things we will learn in this set of Notes is when Circuit Theory is
admissable and when we must apply transmission line techniques.

At some level, every circuit has transmission line characteristics. However, often the layout
of the circuit, the size of the circuit, and the frequency of operation are such that Circuit
Theory techniques are sufficiently accurate for successful analysis and design. In other
cases, such as very high frequencies and/or very large circuits, we may have to use trans-
mission line techniques. We will learn in this set of Notes how to understand some of
these considerations and make the choice between circuit techniques and transmission line
techniques.

In this set of Notes we will also develop an understanding of the physical meaning of the
solution to the Wave Equation. We will see that the solution has a very satisfying physical
and geometrical interpretation which can sometimes save us subtantial mathematical effort.
ECE 303 - Sum10 Notes Set 2: Propagation on Transmission Lines 2

THE WAVE EQUATION

Understanding when to use transmission line techniques requires us to examine in detail


the solution to the wave equation. From our previous work we derived the Wave Equation
for the lossless transmission line. We found that this equation is a second-order PDE for
the voltage v = v(x; t ) given by
∂2 v ∂2 v
= LC (1 )
∂x2 ∂t 2
where L and C are the per unit Inductance and Capacitance, respectively. In Notes Set 1 we
showed that the solution to the wave equation for v in (1) is given by

x 
v = v(x; t ) = g t (2)
vp
where the parameter v p in (2) is called the Phase Velocity (in m/sec). Recall that v p and the
Characteristic Resistance, R0 , of the lossless line are computed by
r

vp = p1 ; R0 =
L
C
(3a; 3b)
LC
The function “g” in (2) is any physically realizable function, like a pulse or a sinusoid,
which can be successfully coupled to the line. It is straightforward to show that among the
solutions to the wave equation are the following:
v1 (x; t ) = e j ( ω t βx )

v2 (x; t ) = cos(ω t β x + φ)
v3 (x; t ) = sin(ω t β x)
The parameter β in these equations is called the phase constant, defined by

ω 2π
β = = (4a)
vp λ
where ω = 2π f in (4a) is the radian frequency of the sinusoid and λ is the wavelength,
defined by
vp
λ= (4b)
f
We will frequently use these solutions in our work since they are differentiable with respect
to both time t and space x.
ECE 303 - Sum10 Notes Set 2: Propagation on Transmission Lines 3

CIRCUIT THEORY OR WAVE TECHNIQUES ?

In ECE 211 you developed the ability to perform circuit analysis. However, understanding
the current material requires us to learn about waves and propagation. At first glance these
seem to be two very different approaches to finding currents and voltages. For the following
discussion let’s adopt the following informal definitions:

Circuit Techniques :
Circuit Techniques are the application of Kirchoff’s Laws, ordinary derivatives, and
the RLC voltage/current relations like v(t ) = i(t ) R, v(t ) = L di=dt, etc. to find volt-
ages and currents in a device or system.

Transmission Line Techniques :


Transmission Line Techniques are the application of the wave equation, partial deriva-
tives, and reflections (which will be developed in a later Notes Set) to find voltages
and currents in a device or system.

Under some constraints, transmission line techniques and circuit techniques give the same
answers to problems. Consequently, circuit theory is widely used in situations where these
constraints are met. But there remains the questions of (i) what these constraints are, and (ii)
when these constraints are met. Understanding the implications of the solution to the wave
equation, given by equation (2) on the previous page, will help us answer the following
fundamental question concerning the analysis and design of systems and circuits :

Given a circuit layout of a particular size and operating frequency, when can
we analyze/design the circuit using Circuit Techniques and when must we use
Transmission Line techniques ?

The answer to this question depends on the relation between the physical size of the layout
and the wavelength of the propagating voltage/current wave. We can understand this by
examining an example which we begin on the following page.
ECE 303 - Sum10 Notes Set 2: Propagation on Transmission Lines 4

SOURCE-LINE-LOAD SYSTEM

The figure below shows a lossless transmission line connecting an ideal source, with source
resistance Rs , to a load RL = R0 , where R0 is the characteristic resistance given by (3b).
(As we will learn later, setting RL = R0 will eliminate reflections, but RL = R0 will still
demonstrate the need for transmission line techniques.) The source-line connection is at
x = 0 and the line-load connection is at x = xL :

Rs

+ ( R0 )
vs (t) − RL = R
0

source line region load


region region
x=0 x = xL

The figure above looks very much like a circuit schematic. Therefore, let’s redraw the
source-line-load system and attempt to analyze it using circuit theory. In the figure below,
the “X” represent locations at x = 0+ and x = xL . Measuring the voltages at these points
provides the voltages v0 (t ) at x = 0+ and v1 (t ) at x = xL .

x=0 x = xL
Rs
X X
+ +
vs (t) v0(t) v1 (t)
+
− RL = R0
x = 0+ x = x−
L
− −
X X

The point x = xL is the last point on the transmission line before the connection with the
load. On the next page we will apply circuit theory to obtain v0 (t ) and v1 (t ) .
ECE 303 - Sum10 Notes Set 2: Propagation on Transmission Lines 5

CIRCUITS-BASED SOLUTION

Let’s analyze the bottom figure on the previous page using the circuits skills you learned in
ECE 211. Let the transmission line be lossless with v p = 2(10)8 m/sec. Since the line is a
lossless conductor, then circuit-theory suggests the points x = 0+ and x = xL should have
the same voltage, that is,

v0 (t ) = v1 (t ) (5)

Suppose vs (t ) is one-period (Tp seconds) of a sine wave at frequency f Hertz. Let this
source produce the initial voltage on the line given by

A sin(2π f t ); 0  t  Tp


v0 (t ) = (6 )
0; else

Combining (5) and (6) then gives the circuits-based result

A sin(2π f t ); 0  t  Tp


v1 (t ) = (7 )
0; else

Note that the result in (7) is independent of the actual Hz frequency f , the propagation
velocity v p , or the length of the line.

Therefore, the circuits-based approach, represented by the results in equations (5) - (7),
suggests the voltage on the two ends of the line is the same. This means the voltages are
the same regardless of whether the line is 1 m long or 0.01 m long or 10 m long. However,
let’s see what happens when we use the solutions required by the wave equation (that is,
solutions of the type given in equation (2) ). This is done on the next page.
ECE 303 - Sum10 Notes Set 2: Propagation on Transmission Lines 6

WAVE-BASED SOLUTION

We can see the limitation of the circuits-based result by relating the voltages in the previous
figures to the solutions to the wave-equation. The solution from equation (2) for the lossless
wave equation is repeated here for easy reference:

x 
v(x; t ) = g t (8)
vp

where g is the wave shape appearing on the first point of the line, that is, at x = 0+ . But
v0 (t ) has been defined to be the voltage at x = 0+ . Therefore, equation (8) thus relates the
voltages at each end of the line to solutions of the wave equation. At x = 0+ we have

x 
v0 (t ) = v(x; t ) = g t = g(t ) (9)
x=0+ v p x=0+

At the load end of the line we have x = xL and therefore equation (8) gives

x  
xL 
v1 (t ) = v(x; t ) = g t = g t (10)
x=xL v p x=xL vp

Comparing (10) to (9) we see an equivalent way of expressing v1 (t ) is



xL 
v1 (t ) = v0 t (11)
vp

Notice how the waveform at the end of the line is a time-delayed replica of the waveform
at the beginning of the line. “Time” is a very appealing dimension for humans, but the
quantity of Phase (in radians) is frequently more useful for understanding sinusoidal results.
We will explore the implications of this phase on the following page.
ECE 303 - Sum10 Notes Set 2: Propagation on Transmission Lines 7

IMPLICATION OF PHASE

To see the implication of phase, let’s examine what occurs at the load end of the line. At the
load end of the line x = xL . Therefore, replace t by t (xL =v p ) in the sinusoid waveform
in equation (6). Then using equation (11) provides an alternate expression for v1 (t ) :
8 
xL 

xL  <
A sin 2π f (t
vp
) ; 0 t xL
vp
 Tp
v1 (t ) = v0 t = (12)
vp :
0; else

Use (4b) to simplify the sinusoidal and the inequality on the right side of (12), obtaining
8 
xL 
v1 (t ) =
<
A sin 2π f t 2π
λ
;
xL
vp
t  Tp +
xL
vp (13)
:
0; else

Equation (13) can be simplified using two substitutions. Let the phase angle θL and the
time delay tL be defined by
xL xL
θL = 2π ; tL = (14; 15)
λ vp
Substituting (14) and (15) into (13) then provides the more compact form
(  

v1 (t ) =
A sin 2π f t θL ; tL  t  Tp + tL
(16)
0; else

Now recall (6) for v0 (t ), the signal at the beginning of the line, repeated here for reference:

A sin(2π f t ); 0  t  Tp


v0 (t ) = (17)
0; else

Comparing (16) and (17), we see that v1 (t ), the signal at x = xL , is not the same as v0 (t ),
the signal at x = 0+ . There is a time-delay between the two signals, as expressed in the
time range on the right side of (16). The non-zero portion of v0 (t ) begins at time t = 0, but
the non-zero portion of v1 (t ) does not begin until time tL .
ECE 303 - Sum10 Notes Set 2: Propagation on Transmission Lines 8

IMPLICATION OF PHASE (cont.)

The time displacement between the two signals v0 (t ) and v1 (t ) on the previous page is
interesting, but it does not entirely answer the question of whether we need to use circuit
theory or transmission line techniques. For answering this question, it turns out that the
phase delay θL , given by the negative phase component in the sinusoid in (16), is the more
useful quantity. This phase delay is the quantity which helps us answer the question of
when circuit theory is valid.

From equation (14), we see that θL becomes non-negligible as the length of the line xL
becomes an appreciable fraction of the wavelength λ. We will work two examples on the
following pages which will help us understand the relation between θL and λ. In each of
these examples you can refer to the line diagrams on page 4 for definition of parameters.
Also for simplicity, we will assume that there are no reflections from the load and that the
propagation velocity is independent of frequency.
ECE 303 - Sum10 Notes Set 2: Propagation on Transmission Lines 9

EXAMPLE: LOW FREQUENCY

Let’s first examine an example which connects a low-frequency source to a load. Let the
length of the lossless transmission line be xL = 1 meter and let the propagation velocity v p
be v p = 2(10)8 m/sec. Using (15), this gives the time length of the line as

1
tL = = 5 nanosec = 000005 ms
: (18)
2(10)8

Expressing tL in milliseconds (ms) will be useful in the final result. Let the frequency of
the low-frequency sinusoidal source be f = 200 Hz, in which case the time period Tp is

1 1
Tp = = = 5 ms (19)
f 200
This is a signal in the “audio” range. From (4b) the wavelength of this sinusoid is

vp 2(10)8
λ = = = 106 meters (20)
f 200
The ratio of the length of the line (1 m) to the wavelength for this case is therefore
xL 1 6
= = 10 (21)
λ 106
Equation (21) shows that the size of the entire source/line/load system is very small when
measured in units of wavelengths. Substituting (21) into (14) gives the phase term

θL = 2π 10 6
rad  3:6(10) 4
degrees (22)

Finally, substitute (18), (19) and (22) into (16) and (17), obtaining the expression for the
signals at the beginning and the end of the line:

A sin(2π 200t ); 0  t  5 ms
v0 (t ) = (23)
0; else
(  
v1 (t ) =
A sin 2π 200t 0:00036 Æ ; : 000005 ms  t  5 000005 ms
:
(24)
0; else
ECE 303 - Sum10 Notes Set 2: Propagation on Transmission Lines 10

EXAMPLE: LOW FREQUENCY (cont.)

The phase value θL = 2π 10 6 rad  3:6(10) 4 degrees in (24) is very small compared
to one sinusoidal cycle, which is 2π radians or 360 degrees. This means that the signal
v1 (t ), at the load end of the line, is essentially the same as v0 (t ), the signal at the beginning
of the line. Their plots “overlap in time”, as discussed in class. You can see this by plotting
v0 (t ) and v1 (t ) from (23) and (24) on the same graph, as shown below:

x = 1, f = 200 v0 (t)
L
v1 (t)
A

t (ms)
o t=5
t=0 0.00036 phase
difference

The plot above shows v0 (t ), and then v1 (t ) displaced by 3:6(10) 4 degrees. The sinusoids
are virtually indistinguishable. In fact, the displacement of v1 (t ) in the figure is actually
exaggerated for illustration purposes.

Plotting the two graphs shows easily that the two waveforms are essentially the same. Since
circuit theory predicted that v0 (t ) = v1 (t ), this is equivalent to saying that circuit theory is
valid for analyzing this system.
ECE 303 - Sum10 Notes Set 2: Propagation on Transmission Lines 11

EXAMPLE: HIGH FREQUENCY

Now let’s examine an example which connects a high-frequency source to a load using
the same xL = 1 meter line. This means that tL , the time length of the line, is still tL =
5 nanosec, as given by (18). However, now let the frequency of the sinusoidal source be
f = 100 MHz, in which case the period Tp in seconds is

1 1
Tp = = = 10 nanosec (25)
f 108
This is an electrical signal in the “FM-radio” range. The electrical wavelength of this
sinusoid is
vp 2(10)8
λ = = = 2 meters (26)
f 100(10)6
The ratio of the line length (1 m) to the wavelength for this high-frequency case is thus
xL 1
= = 0:5 (27)
λ 2
Equation (27) shows that now the size of the source/line/load system is non-negligible
when measured in units of wavelengths. Substituting (27) into (14) gives the phase term

θL = 2π (0:5) = π rad = 180 degrees (28)

Finally, substitute (18), (25) and (28) into (16) and (17), obtaining the expressions for the
signals at the beginning and the end of the line:

A sin(2π 108 t ); 0  t  10 nanosec
v0 (t ) = (29)
0; else

(  
v1 (t ) =
A sin 2π 108 t 180Æ ; 5 nanosec t 15 nanosec
(30)
0; else
ECE 303 - Sum10 Notes Set 2: Propagation on Transmission Lines 12

EXAMPLE: HIGH FREQUENCY (cont.)

The results in (29) and (30) mean that the signal at the end of the line, v1 (t ), is quite
different from v0 (t ), the signal at the beginning of the line. They do not “overlap in time”
like for the low frequency case. You can see this by plotting v0 (t ) and v1 (t ) from (29) and
(30) on the same graph, as shown below:

x = 1, f = 100 MHz v0 (t)


L
v1 (t)
A

t (nanosec)

t=0 t=5 t = 10 t = 15

From the figure above you can see that the waveforms are not the same. For example,
observe the waveform values at t = 7:5 nanosec. At the beginning of the line we have
v0 (t = 7:5 ns) = A. However, at the load end we have v1 (t = 7:5 ns) = A, the negative
of the voltage at the beginning of the line. This means a design based upon v0 (t ) = v1 (t )
would be in error.

Therefore, circuit theory like you learned in ECE 211 is not a completely accurate approach
for analyzing the currents and voltages in this “high-frequency” distributed system. How-
ever, if the line were substantially “shorter” in physical length xL , then circuit theory might
once again be valid.

The preceeding material does not mean circuit theory is not a useful technique. Provided
there are no reflections from the load back onto the line, circuit theory can be applied very
well to analyze and design systems whose sizes are “small” with respect to a wavelength.
The problems listed below give you some practice in deciding whether circuit theory or
transmission line methods are appropriate.

Problems: 2-1 through 2-4


ECE 303 - Sum10 Notes Set 3: Lossless Lines - Pulses 1

INTRODUCTION

In previous work we have learned that the general transmission line has losses due to the
in-line resistance of the conductors and the conductance between the conductors. The
conductance is due to a current path existing between the two conductors.

However, there are many applications in which, at least to a first approximation, we can
assume that the lines are lossless. Some applications which use this assumption will be
discussed in class. The lossless approximation means the per unit resistance, R, and the
per unit conductance, G, may be assumed to be zero. In fact, these were the assumptions
which led to the wave equation.

In this set of Notes we examine some properties of pulse propagation on lossless lines.
For example, a high speed computer might use rectangular pulses to transmit information
between CPU and main memory. Additionally, a local area network (LAN) might use rect-
angular pulses to transmit information between clients and servers. In these applications
the short transmission lines may often be assumed as lossless.

However, we will see that simply using a lossless line for transmission does not necessarily
remove all distortion from the transmission system. We will see that there may be reflec-
tions created by the impedance mismatches among the transmitter, the transmission line
itself, and the receiver. In this set of Notes we will study the effects of these reflections on
the transmission environment for pulses. In a future set of Notes we will study what occurs
if the transmitted signals are sinusoids.
ECE 303 - Sum10 Notes Set 3: Lossless Lines - Pulses 2

LOSSLESS TRANSMISSION LINES

In a common application, we will use a Transmission Line to connect a Source to a Load.


A Lossless Line is shown below:
Rs

+ ( R0 )
vs (t) − RL

source line region load


region region
x=0 x = xL

In the figure above, the parameters are given by:

vs (t ) = Source Voltage, Rs = Source Resistance


RL = Load Resistance, R0 = Characteristic Resistance of Line

We encountered the characteristic resistance before in Notes Set 1. The characteristic resis-
tance R0 is not a series resistance. As we saw in previous notes, the characteristic resistance
is the ratio of voltage to current on the lossless line when there are no reflections. To under-
stand this, let a voltage pulse enter the transmission line at x = 0 and t = 0. Then a physical
interpretation of R0 is given by
v(x = 0+ ; t = 0+ )
R0 = (1 )
i(x = 0+ ; t = 0+ )
For a lossless line, the characteristic resistance R0 and the velocity of propagation v p (phase
velocity) may be calculated by
r
R0 =
L
C
; vp = p1 (2 )
LC
These relations were developed in Notes Set 1.
ECE 303 - Sum10 Notes Set 3: Lossless Lines - Pulses 3

COMPUTING LINE PROPERTIES

A 5 meter lossless transmission line has per unit parameters L = 10 6 H/cm and C =
5(10) 9 F/m. The line is terminated at y = 5 m with RL = 0 (a short circuit ). Compute the
properties requested in (a), (b), (c), and (d) below:

(a) Compute v p , the velocity of propagation in m/sec;


ANSWER:
Use v p = p1 ) v p = 1:414(10)6 m/sec
LC

(b) Compute the characteristic resistance, R0 , of the line in Ohms.


ANSWER:
r
Use R0 =
L
C
) 141:2 Ω

(c) Compute the total series resistance of this 5 meter line in Ohms.
ANSWER:
The total series resistance of any lossless line is ZERO.

(d) Let the source resistance Rs = 100 Ω. A rectangular pulse on the line with amplitude
6 volts and duration 1 µsec. begins propagating toward RL at time t = 0. Plot the voltage
across RL over the range 0  t  10 µsec.
ANSWER:
The voltage drop across the short-circuit (RL = 0) is ZERO.
ECE 303 - Sum10 Notes Set 3: Lossless Lines - Pulses 4

PULSE PROPAGATION ON LOSSLESS LINE

Let’s first consider the time period when there are no reflections from the load. The ampli-
tude of the ideal pulse generated by the source is scaled as it enters the transmission line.
When there are no reflections from the load, an Equivalent Circuit is as shown below:

Rs

+ R0
vs (t) −

source line region


region
x=0

Define v0 (t ) = v(x = 0+ ; t ) as the initial voltage pulse v0 (t ) propagated on the line. The
source pulse vs (t ) and v0 (t ) are related as shown below:

vs (t) K v (t)
0
A

t t
t=0 t=T t=0 t=T

Note that the shape of vs (t ) has not changed as the pulse enters the line. Finding the
amplitude of v0 (t ) is similar to voltage division from circuits courses. The values K and A
are related by  
A=
R0
R0 + Rs
K (3 )

On the following page we will begin to examine the line as having dimensions in both soace
(meters) and time (seconds).
ECE 303 - Sum10 Notes Set 3: Lossless Lines - Pulses 5

PULSE AND LINE PROPERTIES

QUESTION: What is the distance “length” of a pulse on a line?


Answer: Suppose v0 (t ) is a 1 V pulse which is 4 µsec long in time:

v (t)
0
1

t
t=0 t = 4 µs
When this pulse is on the line it occupies a span of
d p = v p T = 2(10)8 m/sec  4(10) 6 sec = 800 m

QUESTION: How many seconds “long” in time is a transmission line which has physical
length xL meters?
Answer: Electromagnetic energy propagates at a velocity of v p m/sec on the line.
xL
Therefore it takes energy a time of tL = seconds to travel from one end of the
vp
line to the other. For example, if xL = 1 mile  1609 meters and v p = 2(10)8 m/sec,
1609 m
then the “time” length of the line is tL = = 8.05 µsec.
2(10)8 m/sec

In this course, it will be useful to think of pulses and lines as having dimensions in both
spatial and time units. We will next consider what happens to the energy when it reaches
the end of the transmission line where the load is located.

Problem: 3-1
ECE 303 - Sum10 Notes Set 3: Lossless Lines - Pulses 6

REFLECTION AND TRANSMISSION

Consider again the source/line/load transmission system on page 2. A source with source
resistance Rs is connected to a lossless transmission line with characteristic resistance R0 .
The transmission line is then connected at x = xL to a load having resistance RL . When the
energy in the pulse propagates to the load then some energy in the pulse will be transfered
into the load, but some energy might be reflected from the load. The reflection from the
load is quantified by the load reflection coefficient and the transmission into the load is
quantified by the load transmission coefficient.

LOAD REFLECTION COEFFICIENT:


Let A be the voltage amplitude of a pulse on a line which reaches the load. Then Vre f l , the
voltage amplitude of the pulse which is reflected back onto the line, is

Vre f l = ΓL A (4 )

where ΓL is the Load Reflection Coefficient, computed by the formula


RL R0
ΓL = (5 )
RL + R0
LOAD TRANSMISSION COEFFICIENT:
Let A be the voltage amplitude of a pulse on a line which reaches the load. Then Vload , the
voltage amplitude of the pulse which enters the load, is

Vload = τL A (6 )

where τL is the Load Transmission Coefficient, computed by the formula


2 RL
τL = (7 )
RL + R0
Using the preceding definitions, you can show that ΓL and τL at the load are related by

τL = 1 + ΓL (8 )
ECE 303 - Sum10 Notes Set 3: Lossless Lines - Pulses 7

REFLECTION AND TRANSMISSION (cont.)

If the load is mismatched to the line, then load reflections will travel back toward the
source. If the source is also mismatched to the line, then reflections will be created at the
line-source boundary, and additional reflections will be created. These source reflections
are then sent back toward the load. The amplitude of the source relections are determined
by the Source Reflection Coefficient.

SOURCE REFLECTION COEFFICIENT:


Let Vin be the voltage amplitude of a pulse on a line which reaches the source from the line
side. Then Vback , the voltage amplitude of the pulse which is reflected back onto the line, is

Vback = Γs Vin (9 )

where Γs is the Source Reflection Coefficient, computed by the formula


Rs R0
Γs = (10)
Rs + R0
Some computation examples for relflection coeffients are given on the following page.
ECE 303 - Sum10 Notes Set 3: Lossless Lines - Pulses 8

REFLECTION AND TRANSMISSION EXAMPLES

The reflection and transmission coefficients defined on the previous pages quantify how the
voltage of the pulse is changed as energy is either reflected back onto the line or transmitted
into the load. Here are some examples which demonstrate computations.

EXAMPLE 1:
Let Rs = 150; R0 = 100; RL = 75. Use equation (4) to compute the reflection and transmis-
sion coefficients.

150 100 75 100 2(75)


Γs = = 0.20 ; ΓL = = - 0.143 ; τL = = 0.857
150 + 100 75 + 100 75 + 100

EXAMPLE 2:
Let Rs = 100; R0 = 100; RL = 100. Compute the reflection and transmission coefficients.

100 100 100 100 2(100)


Γs = = 0 ; ΓL = = 0 ; τL = = 1.0
100 + 100 100 + 100 100 + 100

In the second example above there are no reflections. We say the source is match to the line
and the line is matched to the load.

Problems: 3-2, 3-3


ECE 303 - Sum10 Notes Set 3: Lossless Lines - Pulses 9

PULSE AND REFLECTION EXAMPLE

At t = 0 an ideal voltage source having Rs = 100 Ω produces a rectangular pulse with ampli-
tude 6 volts and duration 10 nsec. The amplitude of the pulse which initially travels toward
the load is 4 volts. The amplitude of the first reflected pulse which travels back toward the
source is 2 volts. Compute the load resistance RL and the characteristic resistance of the
line R0 .

APPROACH:

Solution for R0 :
The amplitude of the source pulse is K = 6 and A = 4 volts initially propagates on the
line. Use the relation between A and K given by euation (3) to solve for R0 :
   
A=
R0
Rs + R0
K ) 4=
R0
100 + R0
6
Solve this last equation for R0 and you will obtain
R0 = 200 Ω

Solution for RL :
To solve for RL , use the relation between A and the reflected voltage ΓL A:

ΓL A = 2 ) RL R0
RL + R0
A = 2 ) RL 200
RL + 200
4 = 2
Solve this last equation for RL and you will obtain
RL = 600 Ω.
ECE 303 - Sum10 Notes Set 3: Lossless Lines - Pulses 10

BOUNCE DIAGRAMS

Some important properties of pulse voltages are conveniently expressed in a Bounce Dia-
gram. The Bounce Diagram displays propagation and reflection using the Two-Dimensional
representation shown below:

SOURCE LOAD

x=0 x = xL
t=0
A

t = tL τ LA
ΓL A

t = 2t L
Γs ΓL A
τ L Γs ΓL A
t = 3t L

Γs ΓL A
2

 Some energy of the original pulse (with amplitude A) moving to the right is transmit-
ted into the load. This amount of voltage is τL A.

 Some energy of the original pulse (with amplitude A) moving to the right undergoes
a reflection given by ΓL . Therefore, the amplitude of the first reflected pulse moving
back to the left is ΓL A.

 The pulse moving to the left with amplitude ΓL A then encounters the source and
undergoes a reflection Γs . The amplitude of the pulse reflected from the source is
Γs ΓL A
ECE 303 - Sum10 Notes Set 3: Lossless Lines - Pulses 11

BOUNCE DIAGRAM EXAMPLE

Let A = 3 volts, T = 10 ns, v p = 2(10)8 m/sec and line length equal 20 m. Also let ΓL =
0.429 and Γs = 0:200. Plot the location and amplitude of the pulse at t = 20 ns, 120 ns,
and 240 ns using the bounce diagram.

APPROACH: Compute the necessary parameters:


20
tL = = 100 ns, d p = 2(10)8 10(10) 9 =2m
2(10)8
ΓL A = 1.29, Γs ΓL A = -0.26
t = 20 ns: d = 2(10)8 20(10) 9 = 4
t = 120 ns d = 2(10)8 120(10) 9 = 24 m 4 m toward source
t = 240 ns: d = 2(10)8 240(10) 9 = 48 m 8 m toward load

SOURCE LOAD

x=0 x = 20
t=0
3.00 (volts)

t = 100 ns
1.29
t = 200 ns
−.26
t = 300 ns

Given the bounce diagram, we can now plot the voltage on the transmission line for selected
times. This is shown on the next page.
ECE 303 - Sum10 Notes Set 3: Lossless Lines - Pulses 12

LINE “SNAPSHOTS”

Plot the voltage on the physical line (a “snapshot”) corresponding to the bounce diagram
in the previous example.

SOURCE LOAD

t = 20 ns 3

2 4
x=0 x = 20

t = 120 ns
1.29

16 18

t = 240 ns

6 8

−.26

The above plot shows the pulse changing polarity and reflecting off of both the load and the
source. We can also fix the position at any value of distance x and compute the time-domain
voltage at this position. This corresponds to the physical situation of using an oscilloscope
at x = x1 to acquire the time voltage trace v(x1 ; t ).
ECE 303 - Sum10 Notes Set 3: Lossless Lines - Pulses 13

WORKED LINE EXAMPLE - 1

The following pages are some worked examples for transmission lines and pulses. You can
study these to help you understand some of the concepts developed for voltages, pulses,
and reflections.

EXAMPLE: An ideal signal generator with impedance Rs = 70Ω is at x = 0 on a loss-


less transmission line (R0 = 70Ω, v p = 2(10)8 m/sec). This signal generator produces the
waveform 
5; t  0
vs (t ) =
0; else
The line is terminated at x = 30 m with a 50 Ω load. Plot the voltage on the line at t = 200
nanoseconds.

ANSWER:
This source signal vs (t ) is a step, which is infinite in time length. Therefore, it is also
infinite in spatial length. The amplitude which first appears on the line is
   
A=
R0
Rs + R0
K =
70
70 + 70
5 = 2:5 volts

The time length of the line is


xL 30
tL = = = 150 nsec
vp 2(10)8
At t = 200 nsec, the leading edge of the step has travelled
d = vp t = 2(10)8  200(10) 9
= 40 meters
The load reflection coefficient is
RL R0 50 70 20
ΓL = = = = 0:167
RL + R0 50 + 70 120
With this information you can obtain the following answer for the problem:

2:500; 0 < x  20
v(x) =
2:082; 20 < x < 30
ECE 303 - Sum10 Notes Set 3: Lossless Lines - Pulses 14

WORKED LINE EXAMPLE - 2

An ideal signal source with impedance Rs = 70Ω is at x = 0 on a lossless transmission


line (R0 = 70Ω, v p = 2(10)8 m/sec). This ideal source (at x = 0 ) produces the waveform

3; t  0
vs (t ) =
0; else
The line is terminated at x = 50 m with a 110 Ω load. Plot the voltage on the line at t = 400
nanoseconds.

ANSWER:
This source signal vs (t ) is a step, which is infinite in time length. Therefore, it is also
infinite in spatial
length. The amplitude which first appears on the line is
   
A=
R0
Rs + R0
K =
70
70 + 70
3 = 1:5 volts

xL
The time length of the line is tL = = 250 nsec.
vp
At t = 400 nsec, the leading edge of the step has travelled

d = vp t = 2(10)8  400(10) 9
= 80 meters

The load reflection coefficient is


RL R0 110 70 40
ΓL = = = = 0:222
RL + R0 110 + 70 180
With this information you can obtain the following answer for the problem:

1:500; 0 < x  20
v(x) =
1:833; 20 < x < 50

Problems: 3-4 through 3-11


ECE 303 - Sum10 Notes Set 4: Lossless Lines - Sinusoids 1

INTRODUCTION

In an earlier set of Notes we studied the behavior of pulses transmitted on lossless lines.
We learned that for short pulses we could actually “see” the propagation and reflection
processes using the bounce diagram. In the current set of Notes we will extend these
concepts to the case of transmitting sinusoids onto lossless lines. The propagation and
reflection processes are still present but we will see that they now take a different form.

Many information systems can be analyzed and designed by considering the input to be a
single frequency sinusoid. For example, an analog cable television system might transmit
a television signal by modulating a single high frequency sinusoid called a carrier. The
time-domain variation of the video information signal is slower (i.e., at a lower frequency)
than the rapid time-domain variations of the sinusoidal carrier. Therefore, the physical
cable tends to propagate the modulated signal very similarly to how it would propagate the
carrier by itself.

Even “digital” communication systems, like digital cable TV or digital cable modem, fre-
quently use a carrier modulation method to transmit the “digital” data. For example, a
logical 1 might be 100 cycles of a positive cosine at the carrier frequency and a logical 0
might be 100 cycles of a negative cosine at the carrier frequency. The transmitted signal
looks very much like a single frequency sinusoid except the these “phase transition” times.
Consequently, design and analysis of the physical system can often be accomplished by
considering the input as a single frequency sinusoid.

In this set of Notes we will see how the sinusoidal nature of the propagating signal requires
a different analysis than did pulse propagation. Many of the tools needed in this analysis
are complex-valued and therefore we will be using quite a bit of complex functions in this
set of Notes.
ECE 303 - Sum10 Notes Set 4: Lossless Lines - Sinusoids 2

SINUSOIDS ON LOSSLESS LINES

If the source connected to a transmission line is a sinusoidal source, the wave propagated
on the line has some specific properties:

WAVELENGTH:
The velocity of propagation v p (in m/sec), the Hertz frequency f (in Hz), and the wave-
length λ (in m) of a wave on a transmission line are related by
vp = f λ (1 )

EXAMPLE: A sinusoid with frequency of f = 400 MHz is found to have a velocity com-
puted to be v p = 2(10)8 m/sec on a transmission line. The wavelength on the line is thus
vp 2(10)8
λ= = = 0:5 meters
f 4(10)8
PHASE CONSTANT:
The phase constant β (in rad/m), is given by the equivalent expressions
ω 2π
β= = (2 )
vp λ
The parameter β is sometimes called the wavenumber, since it computes the number of
radians in one meter of line.

COMPLEX-EXPONENTIAL (PHASOR) NOTATION:


In this course, a tilde will sometimes represent a complex-exponential function (frequently
called a phasor). Here is one example:
ṽ(α) = Ae jω α (3 )
x
Here is another example. Replace α in (3) with α = t :
vp
x
Ae j(ω t βx)
jω ( t vp )
g̃ = g̃(x; t ) = Ae = (4)

Problems: 4-1, 4-2


ECE 303 - Sum10 Notes Set 4: Lossless Lines - Sinusoids 3

SINUSOIDAL WAVES

Energy is not localized in position on a transmission line when the source signal is a sinu-
soid. This requires a different mathematical approach for understanding the properties of
the signals on the line. One difference is that the load is placed at position x = 0, such that
the line extends back into negative x values. This signifies we reference voltage, current,
and impedance phenomena with respect to the load, like we sometimes reference oscillo-
scope time measurements with reference to a triggering waveform. This change of load
position will also simplify some of the resulting mathematical formulas.

Previously we found that some energy on a line will propagate toward the load resistance
RL and will be dissipated as productive “work” across the load resistance. We also found
that some energy will be reflected from the load if RL is not equal to the line characteristic
resistance R0 . This situation is represented for sinusoids in the figure below:

v~f v~b
RL

x=−d ( Ro ) x=0

In the figure above we use the “wave packet” type of symbol to represent the sinusoidal
energy on the line. The wave packet symbol ṽ f pointing toward the load in the above figure
represents the voltage in the Forward Wave (traveling toward the load). The wave packet ṽb
symbol pointing toward the source represents the voltage in the reflected Backward Wave
(traveling back toward the source). The total voltage on the line, the total current on the
line, and the line impedance at any line position can be computed using these concepts of
Forward and Backward waves. We begin a mathematical analysis of these properties on
the following page.
ECE 303 - Sum10 Notes Set 4: Lossless Lines - Sinusoids 4

FORWARD AND BACKWARD VOLTAGE WAVES

Let’s consider first the voltage wave on the line when the input is a sinusoid. Let ṽ f be the
“forward” voltage wave (due to a source) which is propagating to the right on the line:

ṽ f = Ae j ( ω t β x)
(5)

Let ṽb be the “backward” voltage wave (due to a reflection) which is propagating back to
the left on the line:
ṽb = A ΓL e j ( ω t +β x) (6)

Note in (6) that the reflection at the load has changed the amplitude of the reflected wave
to A ΓL . The reflection has also changed the direction of propagation and this is signified
by the changed sign of β x in the argument of the backward wave exponential. The total
phasor voltage wave is therefore the sum of the backward and forward waves. Adding (5)
and (6) then gives

ṽ = ṽ f + ṽb = Ae j ( ωt β x)
+ A ΓL e
j ( ω t +β x)
(7 )

Equation (7) can be simplified to provide


h i
jβx jβx
ṽ = A e + ΓL e e j ωt (8 )

Note that the form in (8) has been factored into a spatially varying function V (x) and the
time-varying phasor e j ω t to give the simple form

ṽ = V (x) e j ω t (9 )

where the spatial function V (x) in (9) is seen to be given by


h i
jβx jβx
V (x) = A e + ΓL e (10)

This form in (10) for the spatial voltage dependence will be very useful when we solve for
the current on the line. This is done on the next page.
ECE 303 - Sum10 Notes Set 4: Lossless Lines - Sinusoids 5

FORWARD AND BACKWARD CURRENT WAVES

To compute the current wave we need to consider again the PDEs for the lossless transmis-
sion line. From Notes Set 1 we derived the following for the lossless line, modified here
for phasor ṽ and ĩ:
∂ṽ ∂ĩ ∂ĩ ∂ṽ
= L ; = C (11a; 11b)
∂x ∂t ∂x ∂t

Equation (11a) requires a spatial partial derivative of ṽ. We can compute this partial deriva-
tive using (9):
∂ṽ j ωt ∂
= e V (x) (12)
∂x ∂x
Compute the derivative of V (x) in (12) using the form for V (x) given in (10):
∂ ∂ h jβx jβx
i
V (x) = Ae + A ΓL e (13)
∂x ∂x
Operating with the partial derivatives on the right of (13) then gives
   
∂ ∂ jβx ∂ jβx
V (x) = A e + A ΓL e (14)
∂x ∂x ∂x
Computing the required partial derivatives in (14) produces
∂ jβx
V (x) = jβAe + j β A ΓL e j β x (15)
∂x
Substituting (15) into (12) and simplifying then gives the expression for the right side of
(11a):
∂ṽ h i
= j β A e j ω t e j β x ΓL e j β x (16)
∂x
We will assume that the phasor current, like the phasor voltage, can also be factored into a
spatially varying function I (x) and a time-varying phasor. This provides the simple form

ĩ = I (x) e j ω t (17)

where the spatial function I (x) in (17) is still unknown. On the following page we will
complete the solution for ĩ by solving for I (x).
ECE 303 - Sum10 Notes Set 4: Lossless Lines - Sinusoids 6

CURRENT WAVE SOLUTION

The right side of (11a) requires a partial derivative of ĩ with respect to t. Using (17) this
derivative is given by
∂ĩ jωt
= j ω I (x) e (18)
∂t
Substitute (16) into the left side of (11a) and (18) into the right side of (11a). The PDE
from (11a) then becomes
h i
jβAe j ω t e jβx
ΓL e j β x = j ω L I (x) e j ω t (19)

Divide both sides of (19) by j ω L e j ω t . This gives a preliminary form for I (x):

βA h jβx
i
I (x) = e ΓL e j β x (20)
ωL
The constant multiplying the spatial exponentials in (20) can be simplified recalling from
previous work the definitions of phase constant β, propagation velocity v p , and character-
istic resistance R0 :
r
ω
β=
vp
; vp = p1 ; R0 =
L
C
(21a; 21b; 21c)
LC
Using (21a) first gives
β ω 1
= = (22)
ωL vp ω L vp L
Now use (21b) in the far right term of (22):
p p
β
= p
1 LC C
= = (23)
ωL vp L L L
Then use (21c) in the far right term of (23) to obtain the desired simplification:
p
β
= p = q =
C 1 1
(24)
ωL L L R 0
C
ECE 303 - Sum10 Notes Set 4: Lossless Lines - Sinusoids 7

CURRENT WAVE SOLUTION (cont.)

Substituting (24) into (20) gives the following form for the spatial function for current:
A h jβx
i
I (x) = e ΓL e j β x (25)
R0
The desired solution for the current on the lossless line with reflections is now obtained by
substituting (25) into (17):
A h jβx jβx
i
ĩ = e ΓL e e jω t (26)
R0
Comparing (26) with (8), we see the forward and bacward current and voltage wave are
related by
ṽ f ṽb
ĩ = ĩ f + ĩb = (27)
R0 R0
Note that the backward current has a minus sign when written in terms of the voltage. This
is because the backward current is flowing in a direction opposite to the forward current.
Given the phasor form of a voltage or current, the real-valued (i. e., measured) quantity is
obtained by taking the real part of (26) and (27):

v(x; t ) = Refṽg; i(x; t ) = Refĩg (28; 29)

To see the time-domain structure of the voltage, substitute (8) into (28), giving the follow-
ing for the real-valued voltage:
n o
v(x; t ) = Re A e j ( ω t βx)
+ A ΓL e
j ( ω t+ β x )
(30)

For our course, all A and ΓL are real. Therefore, (30) becomes
h i
v(x; t ) = A cos(ω t β x) + ΓL cos(ω t + β x) (31)

An expression for the real valued current may be obtained in a similar manner.

Problems: 4-3, 4-4, 4-5


ECE 303 - Sum10 Notes Set 4: Lossless Lines - Sinusoids 8

STANDING WAVE RATIO

Equation (31) for the real-valued voltage on the loaded line is quite complicated. When you
work Problem 4-5 you will find that the corresponding current is also quite complicated.
However, in many problems we may proceed with analysis and design without having to
know the detailed structure of voltage and current waves. The high level information in
these expressions is often summarized by the Voltage Standing Wave Ratio (VSWR). The
VSWR is defined as the following ratio:

1 + j ΓL j
1 j ΓL j
V SW R = (32)

The VSWR is essentially the ratio of maximum voltage magnitude to minimum voltage
magnitude on the transmission line. Recall that the load reflection coefficient for voltage is
given by
RL R0
ΓL = (33)
RL + R0

LIMITING CASES:
Matched Load:
For the Matched Load we have the condition RL = R0 and consequently ΓL = 0. Equation
(32) then gives VSWR = 1. We will later see that this is the case of total power flow into
the load.
“Open-Circuit” Load:
For an “open-circuit” load, we have RL ! ∞ and consequently ΓL = 1. Equation (32) then
gives VSWR ! ∞. We will later see that this is the case of zero power flow into the load.
This is equivalent to a standing wave on the line.
“Short-Circuit” Load:
For a “short-circuit” load, we have RL = 0 and consequently ΓL = 1. Equation (32) again
gives VSWR ! ∞.

Problem: 4-6
ECE 303 - Sum10 Notes Set 5: Line Impedance 1

INTRODUCTION

In Notes Set 3 we studied the problem of reflection and transmission of pulses on trans-
mission lines. We found that when the reflection coefficient ΓL was non-zero then voltage
pulses could reflect back onto the physical transmission line. In a similar manner, in Notes
Set 4 we extended the concept of transmission and reflection to include sinusoidal signals
on the line. The reflection/transmission environment for sinusoidal voltages and currents
on the line was seen to be very complicated.

This set of Notes provides an alternate approach to quantifying the system-level properties
of the reflected and transmitted waves on the transmission line. Remember from basic
circuits that the ratio of voltage to current is, in the general case, called Impedance. Notes
Set 4 has just introduced us to the concept of reflected/transmitted voltage and current
waves. Thus, we naturally wonder if there is some ratio of “reflected/transmitted voltage”
to “reflected/transmitted current” which has the units of impedance and which can be used
to simplify problems.

Such a ratio is called the Line Impedance, which is the topic of the present set of Notes.
This quantity might also be called the “Line-Load Impedance” since it incorporates the
effects of the load at the end of the line, as well as the length of the transmission line
itself. The line impedance will be seen to be a much simpler computation than the voltages
and currents of Notes Set 4. However, accurately using the line impedance requires more
understanding of the physical properties of the transmission line having a load. One of the
goals of the present Notes Set is to aid in this understanding.
ECE 303 - Sum10 Notes Set 5: Line Impedance 2

LINE IMPEDANCE DEFINITION

The definition of the line impedance uses the formula for the transmitted/reflected voltage
and current waves from Notes Set 4. These formulas are repereated here for easy reference.
The phasor voltage ṽ = ṽ(x; t ) was seen to be given by
h i
jβx jβx
ṽ(x; t ) = A e + ΓL e e j ωt (1 )

Similarly, the phasor current ĩ = ĩ(x; t ) is given by

A h jβx jβx
i
ĩ(x; t ) = e ΓL e e jω t (2 )
R0
We can now determine the line impedance Z (x), which is defined as the ratio

ṽ(x; t )
Z (x) = (3)
ĩ(x; t )

You can think of Z (x) as the “looking in” impedance (toward the load) at the point x.
Substituting (1) and (2) into (3) this ratio becomes:
" #
e jβx + Γ e jβx
L
Z (x) = R0 jβx Γ e jβx
(4)
e L

Now we need a useful formula for computing the looking-in line impedance. One such
form is obtained by substituting for ΓL in the numerator and denominator in (4):
2 3
jβx + RL R0 j β x
6e RL + R0
e 7
Z (x) = R0 6
4
7 (5)
jβx
RL R0 j β x 5
e e
RL + R0
ECE 303 - Sum10 Notes Set 5: Line Impedance 3

LINE IMPEDANCE COMPUTATIONS

Now multiply numerator and denominator of the right side of (5) by RL + R0 to obtain
" #
(RL + R0 )e j β x + (R R0 ) e j β x
L
Z (x) = R0 jβx
(6a)
(RL + R0 )e (RL R0 ) e j β x

Grouping the R0 and the RL terms in the right side of (6a) then gives
" #
RL ( e jβx + e jβx ) R0 ( e jβx e jβx )
Z (x) = R0 (6b)
R0 ( e j β x + e jβx ) RL ( e j β x e jβx )

Euler’s formulas for cosine and sine may be manipulated to give the following:

2 cos β x = e j β x + e jβx
; 2 j sin β x = e j β x e jβx
(6c)

Substituting (6c) into (6b) then gives


 
2 RL cos β x j 2 R0 sin β x
Z (x) = R0 (7)
2 R0 cos β x j 2 RL sin β x

sin β x
Divide numerator and denominator of (7) by cos βx, remember that tan βx = , and
cos βx
simplify the result to obtain the useful form
 
RL j R0 tan β x
Z (x) = R0 (8)
R0 j RL tan β x

Although equation (8) is mathematically valid, a change in the way of positioning the trans-
mission line with respect to the x-axis will simplify future computations. When considering
a line having a load, it will be convenient to place the load at x = 0 and let the transmission
line extend in the x region. This change in positioning is explored on the next page.
ECE 303 - Sum10 Notes Set 5: Line Impedance 4

LINES WITH LOADS

When considering a line having a load, it will be convenient to place the load at x = 0 and
let the transmission line extend in the x region. The figure below shows this geometry
and shows a forward and reflected wave.

v~f v~b
R
L

x=−d ( Ro ) x=0

In the above figure the input terminals to the transmission line at located at x = d. Define
the looking-in impedance Zin as Zin = Z (x = d ). Therefore, substitute x = d in equation
(7) for Z (x) on the previous page, and recognize that tan( βd ) = tan(βd ). Performing
these operations then provides
 
RL + j R0 tan βd
Zin = R0 (9)
R0 + j RL tan βd

Equation (9) shows that the looking-in impedance to the transmission line can have a dif-
ferent impedance depending on the value of x. One implication is that even though the
load may be fixed, a source will see a varying impedance which depends on the line length.
Another implication is that Z (x) may be complex-valued, even if the load resistance RL and
characteristic resistance of the line R0 are real-valued. We will frequently use equation (9)
to find the impedance of a length d line connected to a load.
ECE 303 - Sum10 Notes Set 5: Line Impedance 5

LINE IMPEDANCE EXAMPLES

EXAMPLE 1: A lossless transmission line operating at f = 10 MHz has v p = 2:36(10)8


m/s and R0 = 141 Ω. The line is terminated at x = 0 m with RL = 100 Ω. Compute Zin , the
“looking in” impedance, on the line at the point x = 50 m from the load.

APPROACH:
Use equation (9), repeated here, to compute this impedance:
 
RL + j R0 tan βd
Zin = Z (x = d) = R0
R0 + j RL tan βd

where d = 50 for this problem. Compute the phase constant β:

ω 2π 107
β= = = 0:266 rad/m
vp 2:36(10)8

Thus, β d = 0.266 (50) = 13.3 rad  762o . This gives tan β d = 0:90 and the looking
in impedance becomes
   o 
100 + j 141 (:90) 161:6 e j 51:8
Zin = Z ( 50) = 141 = 141 o
141 + j 100 (:90) 167:6 e j 32:6

In polar and rectangular we thus have

o
Zin = 136:2 e j 19:2 = 128:6 + j 44:8

One important point to note is that even though the load RL is a real-valued load, the line-
plus-load combination seen by the source is a complex value.
ECE 303 - Sum10 Notes Set 5: Line Impedance 6

LINE IMPEDANCE EXAMPLES (cont.)

EXAMPLE 2: A transmission line (x < 0) operating at f = 50 MHz has v p = 2(10)8 m/sec


and R0 = 100 Ω. The line is terminated at x = 0 m with a 50 Ω load. What non-zero length
must the line have such that the looking in impedance is Zin = RL = 50 Ω? (NOTE: The
trivial line length d = 0 is not allowed.)

APPROACH:
 
RL + j R0 tan βd
Zin is given by equation (9): Zin = R0 .
R0 + j RL tan βd
We want Zin = R L = 50. This occurs if tan βd = 0:
 
RL
Zin = R0 = RL
R0

But tan βd = 0 for βd = n π, where n is a positive integer. Thus,

nπ nλ λ
βd = n π ) d=
β
=

=n
2
Thus we need to compute the wavelength for the frequency f = 50 MHz:

2 (10)8
λ=
vp
f
=
5 (10)7
=4 ) d= 2n

From this last result we see an example of the periodicity of the line impedance. Specifi-
cally, the line impedance equals RL every 2 meters.

Problems: 5-1 through 5-8


ECE 303 - Sum10 Notes Set 5: Line Impedance 7

LIMITING CASES

ZERO-LENGTH LINE: d = 0
The general formula for Zin should reduce to just the load impedance when the line
length d = 0. Does it? When d = 0, we have βd = 0, and thus

tan βd = tan(0) = 0 (10)

Using (10) in equation (9) then gives


     
RL + j R0 tan βd RL + j R0  (0) RL
R0 + j RL  (0)
Zin = R0 = R0 = R0 = RL (11)
R0 + j RL tan βd R0

Yes, the general line impedance formula gives the intuitively correct answer for the
“very short” line.

LOW FREQUENCIES:
For “low” frequencies, wave theory should reduce to circuit theory; that is, two loss-
less conductors across an impedance RL should only have an impedance RL . To check
this case, note that
ω 2π f 2π
β= = =
vp fλ λ

Thus we have β d = 2π . Therefore, when d << λ, we have βd  0. This again


d
λ
gives
tan βd  tan(0) = 0
from which we again obtain
     
RL + j R0 tan βd R L + j R 0  (0 ) RL
R 0 + j R L  (0 )
Zin = R0 = R0 = R0 = RL
R0 + j RL tan βd R0

Yes, the general line impedance formula gives the intuitively correct circuit theory
answer for the “low” frequency line.
ECE 303 - Sum10 Notes Set 5: Line Impedance 8

SHORT-CIRCUITED LINES

Let a length d line be terminated in a short circuit. This is sometimes called a stub. This
line has RL = 0, as shown in the figure below:

Z in

x=−d ( Ro ) x=0

Using (9), the input impedance Zin is given by

   
RL + j R0 tan βd 0 + j R0 tan βd
Zin = Z( d) = R0 = R0 = j R0 tan βd
R0 + j RL tan βd R0 + 0
(12)

EXAMPLE:
Let λ = 1 m, f = 200 MHz, and R0 = 50 Ω. Compute Zin for the length d transmission
line terminated in a short-circuit.
APPROACH:
We have β = 2π=λ = 2π and thus Zin = j50 tan 2πd. If d = 0.1 m, then

Zin = j 50 tan(2π  0:1) = j 50 (:726) = j 36:3 ( INDUCTIVE

If d = 0.35 m, then

Zin = j 50 tan(2π  0:35) = j 50 ( 1:376) = j 68:8 ( CAPACITIVE

Problem: 5-9
ECE 303 - Sum10 Notes Set 5: Line Impedance 9

EXAMPLE FOR INDUCTOR

EXAMPLE:
It is desired to create an L = 10 6 H inductor using a short-circuit lossless transmission
line which has a length less than 1 m. The line characteristic resistance R0 = 100 Ω, the
operating frequency is f = 100 MHz, and the velocity of propagation is v p = 2(10)8 m/sec.
Compute the length of line which meets the specification.

APPROACH: We want Zin = jωL


Compute Zin :
ω 2π108
Zin = j R0 tan βd ( β=
vp
=
2(10)8

This gives Zin = j 100 tan πd


Compute jωL:
jωL = j 2π 108 (10) 6 = j200π
Equate Zin and jωL:
j100 tan πd = j200π
tan πd = 2π
πd = tan 1 (2π)
1 1
d = tan 1 (2π) = (1:413) = 0.45 m.
π π
Remember to find tan 1 (2π) in radians.

Problem: 5-10
ECE 303 - Sum10 Notes Set 5: Line Impedance 10

OPEN-CIRCUITED LINES

Let a length d line be terminated in a open circuit. This line has RL ! ∞.

Z in

x=−d ( Ro ) x=0

To compute Zin , divide numerator and denominator of Zin by RL :


2R R0 3 2 R0 3
L
+j tan βd 1+ j tan βd
6R RL 7 6 RL 7
Zin = R0 4 L 5 = R0 4 5 (13)
R 0R L R0
+ j tan βd + j tan βd
RL RL RL
Take the limit as RL ! ∞:
 
1 R0
Zin = R0 = j (14)
j tan βd tan βd

EXAMPLE:
Let λ = 1 m, f = 200 MHz, and R0 = 50 Ω. Compute Zin for the length d transmission line
terminated in an open-circuit.
APPROACH:
50
For these values we have β = 2π=λ = 2π and thus Zin = j . If d = 0.1 m, then
tan 2πd

Zin =
j50
0:726
= - j 68.9 ( CAPACITIVE

Problem: 5-11
ECE 303 - Sum10 Notes Set 5: Line Impedance 11

EXAMPLE FOR CAPACITOR

EXAMPLE: It is desired to create a C = 10 10 F capacitor using an open-circuit transmis-


sion line which has a length less than 1 m. The line characteristic resistance R0 = 50 Ω, the
operating frequency is f = 100 MHz, and the velocity of propagation is v p = 2(10)8 m/sec.
Compute the length of line which meets the specification.

ANSWER:
1 1
We want Zin = = j
jωC ωC
Compute Zin :
ω 2π108
Zin = j
R0
tan βd
( β=
vp
=
2(10)8

50
This gives Zin = j
tan πd
1
Compute j :
ωC
1 1 100
j = j = j
ωC 2π108 (10) 10 2π
1
Equate Zin and j :
ωC
tan πd
j50
tan πd
=
j100

) 50
=

100
) tan πd = π
Therefore,
πd = tan 1 (π)
1 1
d = tan 1 (π) = (1:263) = 0.402 m.
π π

Once again, remember to find tan 1 (π) in radians.


ECE 303 - Sum10 Notes Set 5: Line Impedance 12

IMPEDANCE FOR PARALLEL COMBINATIONS

Lines can be connected in parallel. For example, let Z1 be the input impedance of line 1
and let Z2 be the input impedance of line 2. The physical picture and equivalent circuit is
shown below:

RA

1
e
lin

Z1 Z2
Z in RB Z
in
line 2

Physical Lines Equivalent Circuit

The input impedances Z1 and Z2 are functions of ZA , ZB , and other line parameters as
shown on the previous page. Using circuit theory, the input impedance for the parallel
combination of lines is given by

Z1 jj Z2
Z1 Z2
Zin = = (15)
Z1 + Z2
EXAMPLE:
An open-circuit line with input impedance Z1 = j68:9 Ω and a short-circuit line with
input impedance Z2 = j36:3 Ω are connected in parallel. Compute Zin .

Z1 Z2 ( j68:9)( j36:3)
APPROACH: Zin = = = j76:7
Z1 + Z2 j68:9 + j36:3

Problems: 5-12 through 5-15


ECE 303 - Sum10 Notes Set 6: Lossy Lines 1

INTRODUCTION

In previous work we have studied lossless transmission lines. We saw that this was an
accurate model for some transmission lines which were small in length. In this set of Notes
we extend our study of transmission lines to include lossy transmission lines. We might
think of these as lines which are long enough in length such that the total line resistance
and conductance values become appreciable.

Examples of these lossy lines are common, especially in many communication systems
which use twisted-pair or coaxial cable for transmission over substantial distances. Two
examples of these lossy line systems are as follows:

Digital Subscriber Line (DSL):


Some DSL systems use the entire length of the twisted-pair copper telephone wires
which connect the subscriber with the remote Central Office. This line is the same
line which carries your voice conversations, and the line may be quite long (on the or-
der of miles for many lines). Due to the lossy nature of the twisted-pair line, DSL can
not be implemented on all of these regular telephone lines. The lossy nature creates
a maximum length of line over which high-speed DSL services can be sucessfully
implemented.

Cable TV/Cable Modem:


Many modern cable TV/cable modem systems have fiber-optic networks which carry
information from the cable service provider to a central neighborhood location. How-
ever, many of the connections from the central neighborhood location to the sub-
scriber premises is still a coaxial cable connection. The length of this last cable
connection is frequently long enough that the line must be analyzed as a lossy line.

We will see that analyzing a lossy line is more challenging than understanding the lossless
line. Mathematically, this is because the solution for the lossy line PDE is not as simple as
the Wave Equation for the lossless line. We will denote the origin of these challenges when
we encounter them in our development.
ECE 303 - Sum10 Notes Set 6: Lossy Lines 2

LOSSY LINE PDEs

A transmission line is called a lossy line when R 6= 0 and G 6= 0 in the transmission line
circuit model. To understand the lossy line we need to return to the general transmission
line equations, which were given by equations (5a) and (5b) in Notes Set 1. They are
repeated here for easy reference:
∂v ∂i
= Ri L (1a)
∂x ∂t
∂i ∂v
= Gv C (1b)
∂x ∂t
Our studies are simplified if we assume the input current and voltage for the line are pha-
sors. This is an accurate mathematical condition for many communications systems which
use what is called “carrier modulation” for sending information. Many modern analog and
digital communication systems use various types of carrier modulation methods.

The boundary conditions for the PDEs in (1) will be the conditions imposed at the transmit-
ter end (x = 0+ ) of the transmission line. Using the tilde notation for a complex “rotating”
phasor, suppose the following are the boundary conditions:

ṽ(x = 0+ ; t ) = ṽ0 (t ) = A e j ω t (2a)

ṽ0 (t ) A j ωt
ĩ(x = 0+ ; t ) = = e (2b)
Z0 Z0
where Z0 is an unknown, possibly complex, scaling factor relating phasor current and volt-
age. We have previously seen a scaling factor relating current and voltage on the lossless
line. It was the characteristic resistance R0 , which related the forward lossless current wave,
ĩ f = ĩ f (x; t ), to the forward lossless voltage wave ṽ f = ṽ f (x; t ). That relation was
ṽ f
ĩ f = (3 )
R0
Notice the similarity between equation (3) and Ohm’s Law from circuits. Therefore, we
can think of (2b) as an assumption that “Ohm’s Law holds for Transmission Lines”. This
is a reasonable assumption. If this assumption is incorrect, then our solution method will
expose an inconsistency.
ECE 303 - Sum10 Notes Set 6: Lossy Lines 3

PROPAGATION ON LOSSY LINES

The boundary conditions at x = 0, given by equation (2), represent the transmitter’s input
signal on the line. The solution we obtain for the lossy line PDE will be the voltage and
current along the entire line for any value of time t. It is straightforward to show that the
real part of the boundary condition in (2a) is equivalent to the real-valued condition

v(x = 0+ ; t ) = v0 (t ) = A cos ωt (4 )

Think of equation (4) as giving us the known, real-valued transmitter voltage. The solution
for all x will then be the propagation of the transmitter voltage and current “down the line”.

Whereas the real-valued boundary condition of (4) is understandable from a physical stand-
point, it is actually easier to mathematically solve the lossy PDE using the complex-valued
phasor representation given in equation (2). Subject to the boundary conditions of (2),
we will next show that the voltage and current phasors given in equations (5) below solve
the lossy transmission line PDEs given in equation (1). We will assume the form of the
solutions can be written as
ṽ = A e γ x e j ωt (5a)
ṽ A
ĩ = = e γ x e j ωt (5b)
Z0 Z0
where γ and Z0 are (possibly complex) constants which must be determined. For simplicity
of notation we have used the simplified forms ṽ = ṽ(x; t ) and ĩ = ĩ(x; t ).

We will next verify that the forms given in (5) are solutions of the lossy PDEs given by
(1). We will do this on the following pages by substituting (5a) and (5b) into (1) and then
testing the results for consistency. If the forms in (5) are not valid solutions for (1), then an
inconsistent or inadmissable condition will result. If no inconsistent conditions arise, then
the forms in (5) are solutions. Additionally, this approach will display the required forms
for the unknown constants γ and Z0 . We will see that they are easily calculable from the
line parameters R, L, G, and C.
ECE 303 - Sum10 Notes Set 6: Lossy Lines 4

SOLUTION PROCESS

Examining equations (1a) and (1b), we see we need to compute four partial derivatives.
Using the solution forms from (5a) and (5b) these partial derivatives are given below:

∂ṽ γ x jω t
= γAe e = γ ṽ (6a)
∂x

∂ĩ A γ x jω t γ
= γ e e = ṽ (6b)
∂x Z0 Z0

∂ṽ γ x jω t
= jω A e e = j ω ṽ (6c)
∂t

∂ĩ A γ x jω t jω
= jω e e = ṽ (6d)
∂t Z0 Z0

Now substitute the partial derivative results from (6a) - (6d) into the equations (1a) and (1b)
and simplifiy:
R j ωL R + jωL
γ ṽ = ṽ ṽ = ṽ (7a)
Z0 Z0 Z0
γ
ṽ = G ṽ j ωC ṽ = (G + jωC ) ṽ (7b)
Z0
The equations in (7a) and (7b) are consistent if the coefficients multiplying ṽ on each side
are equivalent. Therefore, equating the coefficients on both sides gives
R + j ωL
γ = (8a)
Z0
γ
= G + j ωC (8b)
Z0
One the next page we solve (8) for the constants γ and Z0 . We will see that they are
generalizations of parameters we have already seen in the context of the lossless line.
ECE 303 - Sum10 Notes Set 6: Lossy Lines 5

LINE PARAMETERS

Solving equations (8a) and (8b) for γ and Z0 gives the expressions in equations (9) and
(10) below. These two quantities are very important parameters of the lossy line. They are
called the Propagation Constant and the Characteristic Impedance:

Propagation Constant:
p
γ = (R + j ω L) (G + j ωC) (9)

Characteristic Impedance: s
R + j ωL
Z0 = (10)
G + j ωC
The propagation constant basically determines how voltage (or current) decreases along
the length of the line. The characteristic impedance determines the magnitude and phase of
the current with respect to the voltage. We will illustrate these properties on the following
pages. Note also the results for the lossless line are special cases of (9) and (10) obtained
by setting R = 0 and G = 0.

Equations (9) and (10) show that computation of the propagation constant and characteristic
impedance requires multiplications, divisions, and square roots for complex numbers. This
will require you to use some of your complex function skills you developed previously. To
help you with some of these methods a computational example appears on the next page.
ECE 303 - Sum10 Notes Set 6: Lossy Lines 6

COMPUTATION EXAMPLES

EXAMPLE: A lossy line operating at f = 10 KHz has the following per meter parameters:
R = 0:1 Ω; G = 10 µS; L = 0:2 µ H, C = 100 pF . Compute the propagation constant γ
and characteristic impedance Z0 .

APPROACH:
Refer to equations (9) and (10) for the this problem. Using the relation ω = 2π f , we have:

R + j ω L = 0:1 + j 2π 104  0:2(10) 6 = 0:1 + j 0:0125


= 0:101 e j 7:1o
h i
G + j ωC = 10 5+ j 2π 104  10 10 = 10 5 1 + j0:628
5 o
= 1:18(10) e j 32:1

To find γ :
q q
γ
o o 6 e j 39:12o
= (0:101 e j 7:1 ) (1:18(10) 5 e j 32:1 ) = 1:189(10)

o
= 0.00119 e j 19:6 = 0.0011 + j 0.0004
polar rectangular

To find Z0 :
s
0:101 e j 7:1
o p
Z0 = = 8559 e j 25o = 92.5 e j 12:5o
o
1:18(10) 5 e j 32:1

It was previously stated that the propagation constant determines how voltage decreases
along the length of the line. We will illustrate this property on the following page.

Problems: 6-1, 6-2, 6-3


ECE 303 - Sum10 Notes Set 6: Lossy Lines 7

PROPAGATION CONSTANT

Now that we know how to compute the propagation constant, let’s see how it affects the
properties of a lossy line. Recall from equation (9) the formula for the propagation constant
γ, repeated here for reference:
p
γ = (R + j ω L) (G + j ωC) (11)

In general, γ is a complex-valued parameter, with a real part α and imaginary part β:

γ = α+ jβ (12)

In equation (12), α is called the attenutaton constant and β is called the phase constant.
Once γ is computed from (11), α and β may be computed by

α = Refγg; β = Imfγg; (13)

To see the effect of these constants, substitute equation (12) for γ into the transmission line
solution for ṽ from equation (5a):
γx
ṽ = Ae e jω t = Ae ( α+ j β ) x
e jω t = Ae αx
e j(ω t βx)
(14)

Since v(x; t ), the real voltage at time t and position x, is given by v(x; t ) = Refṽg, take the
real part of both sides of (14), giving
αx
v(x; t ) = Ae cos( ωt βx) (15)

Since α > 0 for a lossy line, equation (15) shows that the amplitude of the wave decreases
as a function of distance along the line. As a practical matter, make sure you use the
appropriate units for distance x. For example, if α and β are given in per meter units, then
x also has to be in meters. If α and β are given in per mile units, then x also has to be in
miles, etc.

Problems: 6-4 through 6-8


ECE 303 - Sum10 Notes Set 6: Lossy Lines 8

PROPAGATION CONSTANT EXAMPLES

EXAMPLE 1: A lossy line with no reflections has γ = 3 + j30 (in per kilometer units). A
voltage v0 (t ) = 10 cos(ω t ) enters the transmission line at x = 0+ and flows to the right. At
position x = x1 suppose we measure v(x1 ; t ) = A1 cos(ω t 15). Assuming x1 > 0, compute
the value of A1 .

APPROACH:
Since γ = 3 + j30 we have α = 3 and β = 30. The general real-valued solution is
αx
v(x; t ) = Ae cos(ω t βx)

At x = x1 , we see that the phase is given by

βx1 = 15 ) x1 =
15
β
=
15
30
= 0:5

At x = x1 = 0:5, we see that the magnitude is given by

α x1 3 (0:5)
A1 = A e = 10 e = 2.23

EXAMPLE 2: At position x = x2 on this line e we measure v(x2 ; t ) = 5 cos(ω t 30 x2 ).


Assuming x2 > 0, compute the value of x2 .

APPROACH:
Since γ = 3 + j30 we have α = 3 and β = 30. Also, since
αx j(ω t βx) 3x j(ω t 30x)
ṽ(x; t ) = Ae e = 10e e

we have
jṽ(x t )j = 10e
;
3x

At x = x2 , this last result gives 5 = 10e 3x2 , from which e 3x2


= 0:5. Taking ln of both sides
1
of this last result gives x2 = ln(0:5) = 0.23 m
3
ECE 303 - Sum10 Notes Set 6: Lossy Lines 9

CHARACTERISTIC IMPEDANCE

We can compute the current waveform on the lossy line by using the voltage solution and
the characteristic impedance. Let’s see how the characteristic impedance affects the current
properties on the lossy line. Recall from equation (10), the characteristic impedance Z0 ,
repeated here: s
R + jωL
Z0 = (16)
G + jωC
In general, Z0 is a complex-valued parameter, having a magnitude jZ0 j and a phase φ:

Z0 = jZ0 j e jφ (17)

To see the effect of the characteristic impedance magnitude and phase on the line current,
substitute equation (17) for Z0 into the transmission line solution for ĩ from equation (5b):
A γx A A
e j ωt ( α+ jβ ) x
e j ωt αx
e j ( ωt β x φ)
ĩ =
Z0
e =
jZ0j e j φ e =
jZ0j e (18)

Since i(x; t ), the real voltage at time t and position x, is given by i(x; t ) = Refĩg, take the
real part of both sides of equation (18), giving
A αx
cos(ωt βx φ)
i(x; t ) =
jZ0 j e (19)

The result in (19) shows that jZ0 j scales the current magnitude, and φ causes the current
and voltage to be out of phase on the line. This is different from the lossless infinite length
line. In that case, voltage and current were in phase.
ECE 303 - Sum10 Notes Set 6: Lossy Lines 10

CHARACTERISTIC IMPEDANCE EXAMPLE

At ω = 106 rad/sec a lossy transmission line with no reflections has per meter parameters
R = 0.6 Ω, G = 0, C = 10 10 F, and L = UNKNOWN. The characteristic impedance is
Z0 = 80:0 j37:3. Compute the per meter inductance L.

APPROACH:
From the definition of Z0 and the problem statement we have
s
R + jωL j 25o
Z0 = = 80:0 j37:3 = 88:27e
jωC

Squaring both sides gives


R + jωL j 50o
= 7791:6e
jωC
Multiply both sides by jωC and substitute for ω, R, and C:
h i
0:6 + j106L = j106 10 10
 7791:6e j 50o
= 0:78e
j 40o

Compute the Cartesian form for the term on the right:

0:60 + j106L = 0:78 cos(40o) + j0:78 sin(40o) = 0:60 + j0:50

The real parts cancel, giving

j106 L = j0:50 ) L=
0:50
106
= 5(10) 7

Problems: 6-9 through 6-12


ECE 303 - Sum10 Notes Set 7: Power Flow on Transmission Lines 1

INTRODUCTION

One of the main purposes of transmission lines is to transport an information signal from
one place to another. For analog signals or digital information impressed on a carrier sinu-
soid, the power transported is an important measure of performance.

Computing the power transported by a transmission line will extend some of the properties
of power you learned from basic circuits. We will start this set with a review of the familiar
relations from circuits and then see how the power relations change for transmission lines.
We will study three important transmission line cases in this Notes Set:

 Lossless Lines with no reflections

 Lossy Lines with no reflections

 Lossless Lines with reflections

Lossless lines with no reflections will produce results quite similar to the circuit expressions
you learned earlier. However, we will see that the conditions of lossy lines and reflections
on lossless lines require changes in the computation of power as compared to the circuit
expressions. We will see that lossy lines produce a loss in propagated power as the power
is dissipated in the line. In lossless lines, reflections from the load cause energy to travel
backward on the line toward the source. This represents energy not efficiently transported
from source to load.
ECE 303 - Sum10 Notes Set 7: Power Flow on Transmission Lines 2

INSTANTANEOUS POWER

Let’s first review what we learned about power from basic circuits. Consider the case of
current i(t ) flowing through a resistor of R ohms, causing a voltage v(t ) across the resistor.
From basic circuits the Instantaneous Power dissipated across the resistor is

p(t ) = v(t ) i(t ) (1)

Alternate forms of p(t ) are given by

v2 (t )
p(t ) = ; p(t ) = i2 (t ) R (2a; 2b)
R
If we think of R as the load of a device, we frequently say that p(t ) is the instantaneous
power delivered to the load.

EXAMPLE:
As an example, let the voltage across a resistor R be v(t ) = 2 cos(2π500t ) and let the current
be i(t ) = 0:01 cos(2π500t ). Equation (1) provides a straightforward method to compute the
instantaneous power:
p(t ) = v(t ) i(t ) = 2 cos(2π 500t )  0 01 cos(2π 500t )
:

1 h1 i
= 0:02 cos2 (2π 500t ) = 0.02 +
cos(2π 1000t )
2 2
= 0.01 + 0.01 cos(2π 1000t ) watts

Note that p(t ) is still a function of time t. In fact, notice that p(t ) has a frequency which is
twice the frequency of the original voltage sinusoid.
ECE 303 - Sum10 Notes Set 7: Power Flow on Transmission Lines 3

TIME-AVERAGE POWER

Another useful measure of power is the Time-Average Power for a sinusoid. This is the
average power delivered to the load over one period of the sinusoid. In this course we will
use the symbol Pavg for this quantity. The definition of the time-average power is given by
Z Tp
1
Pavg = p(t ) dt (3 )
Tp t =0
This average power is similar to the “rms” measures you calculated in your circuits courses.

EXAMPLE:
As an example, let’s compute the time-average power for the example on the previous page.
For reference, the voltage v(t ) across the resistor R and the current i(t ) through the resistor
are given below:
v(t ) = 2 cos(2π 500t ) ; i(t ) = 0:01 cos(2π 500t )
The frequency f of the voltage sinusoid is seen to be 500 Hz and the period is therefore given
by Tp = 1= f = 0.002 sec. Equation (3) provides a staightforward method to compute the
time-average power:
Z 002 h i
1 :

Pavg = 0:01 + 0:01 cos(2π 1000t ) dt


0:002 t =0
Z 002
:
Z 002 :

= 500 0:01 dt + 500 0:01 cos(2π 1000t ) dt


t =0 t =0
= 0.01 + 0 = 0.01 watt

Note that there is no spatial argument (like x or y) in any of the preceding computations.
This signIfies we are computing power for a circuit which is very small compared to the
wavelength of the carrier sinusoid. However, transmission lines are frequently larger than
the wavelength, and thus the voltage and current in a transmission line DO have spatial
dependence. How is power now computed for a spatially-distributed transmission line?
This concept is studied next.

Problem: 7-1
ECE 303 - Sum10 Notes Set 7: Power Flow on Transmission Lines 4

COMPLEX-VALUED VOLTAGE AND CURRENT

Power computations for the spatially-distributed transmission line can be simplified if


we use the complex phasor notation. Purely time-domain computations are sometimes
complicated because trigonometric identities and evaluations of triginometric integrals are
frequently necessary. There is an alternative computation for time-average power using
complex-valued phasors which is often much simpler.

To develop these phasor relations, let v(t ) = A cos(ωt ) be the voltage across a resistor R.
Ohm’s law gives the current through the resistor as

v(t ) A
i(t ) = = cos(ωt ) (4 )
R R
The complex phasor voltage and current corresponding to the quantities in (4) are therefore

ṽ(t )
ṽ(t ) = A e j ω t ; ĩ(t ) = (5a; 5b)
R
The real quantities v(t ) and i(t ) are related to the phasor ṽ(t ) and ĩ(t ) by taking the real part
of the corresponding complex quantity:

v(t ) = Ref ṽ(t )g = Ref A e j ω t g = Ref A cos ωt + j A sin ωt g = A cos ωt (6a)

i(t ) = Ref ĩ(t ) g = Ref g = Ref cos ωt + j sin ωt g = cos ωt


ṽ(t ) A A A
(6b)
R R R R

With the previous definitions of complex-valued voltage and current, the time-average
power Pavg may be computed by the formula

Refṽ(t ) ĩ  (t )g
1
Pavg = (7)
2
where ĩ  (t ) is the complex conjugate of the complex current. Using equation (7) to com-
pute Pavg is frequently easier than using equation (3).
ECE 303 - Sum10 Notes Set 7: Power Flow on Transmission Lines 5

COMPLEX-VALUED POWER COMPUTATIONS

We will do an example which shows that the complex-valued computation gives the same
answer for Pavg as does the time-domain integral in equation (3).

EXAMPLE: Let the voltage across a resistor be v(t ) = 2 cos(2π 500t ) and let the current through
the resistor be i(t ) = 0:01 cos(2π 500t ). Then the complex voltage and current are:

ṽ(t ) = 2 e j 2π500t ; ĩ(t ) = 0:01 e j 2π500t

Using ṽ(t ) and ĩ(t ) in equation (7) then gives

Ref 2 e j 2π500t  0 01 e g = 12 Ref 0 02 g = 0 01 watt


1 j 2π500t
Pavg = : : :
2
Note that 0.01 watts is the same answer previously obtained using the time-domain method
for time-average power.

Computation of Pavg using the preceding complex-valued approach is frequently simpler


than the corresponding integral of trigonometric functions. In this course we will consider
Pavg for the following three cases:

 Lossless line with no reflections.


 Lossy line with no reflections.
 Lossless line with reflections.

For some transmission lines the time-average power will not be a function of position.
However, for other transmission lines the average power will be a function of position on
the transmission line. One implication is that all the power introduced to the transmission
line may not be successfully transported to the load. We will next consider each of the three
cases above individually.

Problem: 7-2
ECE 303 - Sum10 Notes Set 7: Power Flow on Transmission Lines 6

LOSSLESS LINE WITH NO REFLECTIONS

If a transmission line is lossless, then R = 0, G = 0, and thus


r
ω p L
β= =ω LC; Z0 = R0 = (8a; 8b)
vp C

With no reflections, there is only a forward voltage wave and a forward current wave:
A j(ωt
ṽ(x; t ) = ṽ f (x; t ) = Ae j(ωt β x)
; ĩ(x; t ) = ĩ f (x; t ) = e β x)
(9a; 9b)
R0
Thus using equations (9a) and (9b) in equation (7) shows that Pavg (x), the time-average
power at point x on the line, is given by
 
 1
Re ṽ(x; t ) ĩ(x; t ) = Re Ae j(ωt
1 β x) A j(ωt β x)
Pavg (x) = e
2 2 R0

A2
= (10)
2R0

Note in this case the power flow is independent of position x.

EXAMPLE: Let the voltage wave be v(x; t ) = 5 cos(ω t β x) on a lossless transmission line with
R0 = 50 Ω. Assuming no reflections, compute Pavg (x).

APPROACH: The phasors are ṽ(x; t ) = 5 e j ( ωt βx ) and ĩ(x; t ) = 0:1 e j ( ωt βx ) . Thus, using (10)
gives
Pavg (x) = Ref 5 e j( ωt βx ) 0:1 e j ( ωt βx ) g = 0:25
1
2
A2 52
Note from the given v(x; t ) and R0 that = = 0:25
2R0 2(50)

Problems: 7-3, 7-4


ECE 303 - Sum10 Notes Set 7: Power Flow on Transmission Lines 7

LOSSY LINE WITH NO REFLECTIONS

Assuming no reflections on the lossy line, the voltage and current phasors are given by the
forward waves:

γx j ωt A γx
ṽ(x; t ) = Ae e ; ĩ(x; t ) = e e jωt (11a; 11b)
Z0

where γ = α + j β is the propagation constant and Z0 = jZ0 je j φ is the characteristic impedance


of the lossy line.

Again using equation (7) shows Pavg (x), the time-average power at point x on the line, is
given by

Pavg (x) = Re ṽ(x; t ) ĩ (x; t )
1
(12)
2
Substituting the phasors from (11) into (12) then gives
 
ṽ(x; t )ĩ (x; t ) = A e αx j ( ωt β x) A ( α+ j β )x j ωt
e
jZ0 j e j φ e e (13)

Completing the computations in (13) shows that Pavg (x) becomes


 
1 A2 2αx jφ
Pavg (x) =
2
Re
jZ0j e e (14)

Since Euler’s formula gives e jφ = cos φ + j sinφ then equation (14) becomes

A2 2αx
cos φ
2jZ0 j
Pavg (x) = e (15)

Note in (15) the power is a function of position x. Since the function e 2 α x decreases as
a function of x, we see that the power transmitted by the lossy line descreases as the line
gets longer.
ECE 303 - Sum10 Notes Set 7: Power Flow on Transmission Lines 8

LOSSY LINE EXAMPLE

A lossy transmission line with no reflections has per kilometer parameters R = 300 Ω,
G = 0, C = 5(10) 8 F, and L = 6(10) 4 H. At y = 0+ , the input waveform to the line is
v(t ) = 5 cos(π 106 t ). Compute the time average power flowing on the line at y = 500 m.

APPROACH:

Compute some of the necessary terms:


R + j ω L = 300 + j π 106  6(10) 4 o
= 300 + j 1885 = 1908 e j 81
G + j ωC = 0 + j π 106  5(10) 8 o
= j 0:157 = 0:157 e j 90

To find α = Refγg first compute γ :


q p o
γ 17:3 e j 85 5
o o o :
= (1908 e j 81 ) 0:157 e j 90 ) = 300 e j 171 =

Now take the real part of γ to find α :


α = Refγg = 17:3 cos(85:5o ) = 1.38

Now find Z0 in polar form :


r
1908 e j 81o p
Z0 = = 12; 153 e j 9o = 110.2 e j 4:5o
0:157 e j 90o
Thus, Pavg becomes

A2 25
2αx
cos(φ) = 2(1:36)(0:5)
cos( 4:5o ) = 0.029 watts
2 jZ0 j
Pavg = e e
2(110:2)

Note x = 0:5 in the last computation. Since the R; L; G; and C parameters are given in per
kilometer units, we must therefore convert x into kilometers: x = 500 m = 0.5 km.

Problems: 7-5 through 7-10


ECE 303 - Sum10 Notes Set 7: Power Flow on Transmission Lines 9

LOSSLESS LINE WITH REFLECTIONS

In this case both ṽ = ṽ(x; t ) and ĩ = ĩ(x; t ) have forward and backward waves:
h i A h i
jβx jβx jω t jβx jβx
ṽ = A e + ΓL e e ; ĩ = e ΓL e e jω t (16a; 16b)
R0
The time-average power at point x on the line is again given by equation (7)

Re ṽ ĩ 
1
Pavg (x) = (17)
2
Compute the phasor product required in equation (17):
A2 h j β x i h i
ṽ ĩ  = e + ΓL e
jβx
 e j β x ΓL e jβx
R0
A2 h j 2βx j 2βx
i
= 1 ΓL e + ΓL e Γ 2
R0
A2 h  i
= 1 ΓL2 + ΓL e j 2 β x e j 2 β x
R0
Simplifying this last result using Euler’s formula gives

A2  
ṽ ĩ  = 1 ΓL2 + j 2 ΓL sin 2 β x (18)
R0
Using the result from (18) in equation (17) for the time average power formula then gives

A2 
Pavg (x) = 1 ΓL2 (19)
2R0
In equation (19) the parameter A is the amplitude of the forward voltage wave which is
propagated on the line. You can see the reflection (represented by the parameter ΓL ) has
reduced the power which successfully enters the load.
ECE 303 - Sum10 Notes Set 7: Power Flow on Transmission Lines 10

EXAMPLE WITH REFLECTIONS I

8
A forward voltage wave ṽ f (t ) = 5 e j 2π (10) t propagates from x = 0 on a lossless trans-
mission line. The line has characteristic impedance R0 = 50 Ω, v p = 2(10)8 m/sec, and is
terminated at x = 0:25 m with a 100 Ω load. Compute the time-average power delivered to
the load.

APPROACH:
Compute the reflection coefficient ΓL :

RL R0 100 50
ΓL = = = 0:333
RL + R0 100 + 50

Thus, the time average power is

A2  52 
Pavg (y) = 1 ΓL2 = 1 0:333 2
2 R0 2(50)

Completing the computations gives

Pavg (x) = 0.222 w


ECE 303 - Sum10 Notes Set 7: Power Flow on Transmission Lines 11

EXAMPLE WITH REFLECTIONS II

Suppose the voltage phasor and current phasor on a transmission line are given by

ṽ = 2 e j ( ωt β x)
+e
j ( ωt +β x )
; ĩ = 0:06 e j ( ωt βx)
0:03 e j ( ωt +β x )

In these equations ω = 108 and β = 0:5. Compute the real-valued time-average power
flowing on this line.

APPROACH:

Note that the voltage and current expressions above include a forward wave and a backward
(reflected) wave. To find the power we can use


Re ṽ ĩ 
1
Pavg (x) =
2

Compute the product ṽ ĩ 


h i h i
ṽ ĩ = 2 e j ( ωt β x ) + e j ( ωt +β x )  0:06 e j ( ωt β x) 0:03 e j ( ωt +β x )

= 0:12 : 06 e j 2 β x + :06 e j 2 β x : 03
= 0:09 + j 0:12 sin(2 β x)

Taking the real part then provides

Re f0:09 + j0:12 sin(2 β x)g


1
Pavg (x) = = 0.045 w
2

Problems: 7-11, 7-12, 7-13


ECE 303 - Sum10 Notes Set 7: Power Flow on Transmission Lines 12

LINE AS CIRCUIT ELEMENT

If we know that a line plus load has an input impedance Zin , then we may use circuit
theory to determine the time-average power flowing on the line. As an example, let a signal
source vs (t ) have source resistance Rs . Note this implies a phasor source ṽs (t ). The source
is connected to a lossless line/load having Zin , as shown below:

RS
.
vs (t) ~ Z in RL
.
x=0

Since we know the input impedance Zin we do not need specific properties of the line like
length or characteristic resistance. To determine the power flowing on the line we can use
circuit theory to determine the power across the “lumped” impedance Zin . To compute this
average power, apply equation (7) to obtain

RefṼin I˜in g
1
Pavg = (20)
2
In equation (20), Ṽin is the phasor voltage across Zin , obtained by voltage division,
 
Zin
Ṽin = ṽs (t ) (21)
Zin + Rs

Also in equation (20), I˜in is the phasor current through Zin , obtained by Ohm’s law

ṽs (t )
I˜in = (22)
Zin + Rs
ECE 303 - Sum10 Notes Set 7: Power Flow on Transmission Lines 13

LINE AS CIRCUIT ELEMENT : EXAMPLE

Here is an example of this approach. Let a signal source vs (t ) = 5 cos(1000t ) have source
resistance Rs = 70 Ω, The source is connected to a lossless line/load having “looking-in”
impedance Zin = 50 + j50 Ω, as shown in the figure on the previous page. Compute the
time-average power flowing on the transmission line.

SOLUTION:
Replace the transmission line and load in the preceding figure with the discrete impedance
Zin . Also, work the problem using phasors. This simplifies the transmission line circuit
down to a voltage supply (given by ṽs (t ) = 5 e j 1000t ) in series with Rs = 70 Ω and Zin .
Since the line is lossless, all power flowing into the line (i.e., flowing across Zin ) will be
delivered to the load. Thus, we need only to find the Pavg across Zin . Define Ṽin as the
voltage across Zin and I˜in as the current through Zin . This current and voltage are related by

Ṽin
I˜in = (23)
Zin
Using (23) the Pavg formula becomes
  
1  1
Re Ṽin I˜in = Re
Ṽin Ṽin 1 jṼin j2 
Pavg =
2 2 
Zin
=
2
Re 
Zin
(24)

Use voltage division to compute Ṽin :

Ṽin =
Zin
Zin + Rs
 ṽs(t ) = 5050 + j50
+ j50 + 70
 5 e j 1000 t = 2:715 e j (1000 t +0 391)
:
(25)

 into the right side of (24), obtaining


Substitute (25) and Z̃in
     
1 2:7152 1 7:371 1
Pavg = Re = Re = 3:686 Re (26)
2 50 j50 2 50 j50 50 j50

On the following page we will provide two methods for finding the real part of the complex
expression in (26). Being able to use both methods will be very helpful in working with
complex functions
ECE 303 - Sum10 Notes Set 7: Power Flow on Transmission Lines 14

EXAMPLE (cont.)

On this page we w provide two methods for finding the real part of the complex expression
in (26). One method uses the polar form of a complex number and the other method uses
the rectangular form.
Polar Form:
Write the denominator expression in (26) in polar form:
   
1 7:371 1 7:371 1 n o
Pavg = Re = Re = Re 0:1043 e j 0 7853
:
(27)
2 50 j50 2 70:7 e j 0 7853:
2
Rewrite (27) by factoring out the constant inside the “Re” argument and expanding the
complex exponenetial using Euler’s formula:
n o
Pavg = 0:0521 Re cos(0:7853) + j sin(0:7853) (28)

Extracting the real part in (28) then gives the desired result:
Pavg = 0:0521 cos(0:7853) = 0:0521 (0:707) = 36:9 mW (29)

Rectangular Form:
Multiply the numerator and denominator expressions in (26) by the complex conjugate of
the denominator:
   
1 7:371 50 + j50
Pavg = Re = 3:686 Re (30)
2 50 j50 (50 j50) (50 + j50)
Perform the multiplication inside the “Re” argument in equation (30):
   
50 + j50 50 50
Pavg = 3:686 Re = 3:686 Re +j (31)
2500 + 2500 5000 5000
Extracting the real part of the expression in equation (31) then gives the desired answer:
50
Pavg = 3:686 = 36:9 mW (32)
5000

Problems: 7-14 through 7-17


ECE 303 - Sum10 Notes Set 8 : Electric Charges and Fields 1

INTRODUCTION

An awareness of many important physical phenomena can be achieved by assuming the


existence of electric charges in fixed positions. This is sometimes called Electrostatics,
denoting the stationary location of the charges. A fundamental property of these stationary
electric charges is that they effect each other through a property called the Electric Field.

The effect of this electric field appears as a force exerted on one charge by another charge.
This force is not like a mechanical force, in which particles or objects must be in physical
contact to transmit the force. The charges themselves can be great distances apart and still
produce forces on each other. However, the specific type of effects depend on the motion
of the charges.

Here is a hierarchy of forces and effects produced by electric charges:

Stationary Charges
If charges are stationary, then they can only effect each other by the Electric Field.
Each charge exerts a force on the other charges. This type of force is sometimes
called the Coulomb force.

Constant Velocity Charges


Charges in constant velocity create a magnetic field. This magnetic field causes a
force which can effect other charges having non-zero velocity. A DC current in a
conductor is an example of constant velocity charges.

Accelerating Charges
Accelerating charges radiate energy away from their location. This energy is some-
times called the Electromagnetic Wave. Antennas use this principle to send television
and radio signals. An AC current is an example of accelerating charges.

In this set of Notes we will discuss the effects of stationary charges on each other. We will
also discuss the motion of charges under the influence of the electric field. In future sets of
notes we will discuss magnetic fields and electromagnetic wave propagation.
ECE 303 - Sum10 Notes Set 8 : Electric Charges and Fields 2

EXAMPLES OF CHARGE PHENOMENA

Below are some examples of how the concepts of electric charge and electric field can
describe some physical problems.

Semiconductor Devices
Some diodes use the important semiconductor structure called a PN-junction. Near
this junction positive and negative ions are held fixed in the lattice of the semiconduc-
tor crystal. These fixed ions carry a charge, which creates an electric field allowing
a preferential direction of current flow. This is sometimes quantified as the threshold
voltage, which is about 0.7 volts for silicon.
Biological Structures
Many important biological structures have electrical charges. For example, some
amino acids (which are the building blocks of proteins) are positively charged and
some amino acids are negatively charged. The attracting and repelling forces pro-
duced by these charged structures cause the resulting protein molecule to twist, turn,
and fold to assume a specific structural shape. In many cases, it is this structural
shape which determines the function of the protein.
Bio-Electricity
Positive ions such as sodium and potassium and negative ions such as chlorine are
significant for the electrical properties of biological cells. These (and other) ions are
resident in the fluids inside and outside the cell, being separated by the cell mem-
brane. However, their concentration imbalances across the cell membrane causes a
potential difference on the order of 50 - 100 millivolts to be generated across the cell
membrane. Therefore, we can think of each cell as a tiny battery, storing potential
energy and available for doing work. Cells such as nerves and neurons can generate
and propagate pulses using these ions. In this way nerves can provide electrical com-
munication pathways between sensory locations in the body and command centers
in the brain.

In the above examples, the importance of the charges is their particulate nature, rather than
the wave energy radiated by the charges as they move. We will discuss the wave nature of
propagated energy in future sets of Notes.
ECE 303 - Sum10 Notes Set 8 : Electric Charges and Fields 3

ELECTRIC FIELDS

We will first consider stationary electric charges; the study of these charges is sometimes
called Electrostatics. An electric charge produces an electric field E which affects other
charges. As a first example, let charge Q exist at the origin. At point P the electric field due
to charge Q is given by
1 Q
4πε jrj2
E = ur (1 )

where the new quantities in equation (1) are defined as follows:

 r is the vector between the location of Q and the point P,


 ur is a unit vector in the direction of r (pointing radially away from the origin),
 ε is the electrical permittivity (in Farads/m) of the medium.

The units of the electric field are volts/meter. If the charges are located in free-space, then
the permittivity (having units of farads/meter) is
ε = ε0  8:85(10) 12
(2 )

The following evaluation, obtained using equation (2), often simplifies computations:
1
4πε0
 9(10)9 (3 )

Equation (1) shows that for a charge at the origin the electric field E has magnitude j E j
and direction u r pointing directly away from the charge (which is at the origin). Let’s refer
to the direction of a general E-field vector as u E . Thus, the E-field in equation (1) can also
be written in magnitude-direction form as

E = j E j uE ) j E j = 4πε
1 Q
jrj2 (4a; 4b)

The electric field E is also a linear field. This means that the fields due to each of multiple
charges add independently. For example, let charge Q1 produce field E1 at point P and let
charge Q2 produce field E2 at point P. The total field, ET is the vector sum
ET = E1 + E2 (5 )
ECE 303 - Sum10 Notes Set 8 : Electric Charges and Fields 4

ELECTRIC FIELD EXAMPLE

Let’s work an example which shows how to compute the E-field. A charge of Q = 1 nC is
located at (x; y) = (0; 0) in free space. Find the E field at the point (x; y) = (4; 3) and write
it in the form E = E u E . This geometry is shown below:
111
000
000
111
0
1
1111111
y=3 000
111
11111
00000
0000000 0
1 E
0000000
1111111 0
1
0000000
1111111 0
1
0000000
1111111 0
1
0000000
1111111
r 0
1
0000000
1111111 0
1
0000000
1111111
Q 0
1
0000000
1111111 0
1
x=4

p p
The vector r = 4ux + 3uy . Thus jrj = 42 + 32 = 25 = 5. The unit vector is therefore

r 4 ux + 3 uy
uE =
jrj =
5
= 0.8 ux + 0.6 uy

The magnitude of the E-field is computed by


9
E=
1
4πε0
 jrQj2 = 9(10)9  1(10)
52
= 0.36

Two forms for the E-field will be useful:


h i
Magnitude-Direction: E = E uE = 0:36 0:8ux + 0:6uy
Cartesian: E = 0:288 ux + 0:216 uy

The last line is the Cartesian form. To obtain the Cartesian form, multiply the magnitude,
E, times the unit vector, u E .

Problems: 8-1, 8-2


ECE 303 - Sum10 Notes Set 8 : Electric Charges and Fields 5

LINEAR FIELD EXAMPLE

The electric field E is a linear field. This means that the fields due to each of multiple
charges add independently, as shown below:

Q2 E1
E T = E1 + E 2
y

.
E2
Q
1
x

EXAMPLE: Let Q1 = 2/9 nCoul be located at (x; y) = (1; 0) in free space and let Q2 = 1/3
nCoul be located at (x; y) = (0; 3). Compute the total E-field at the point P = (x; y) = (5; 4).

APPROACH: Draw the location of the charges and the point P:

y P = ( 5, 4)
.
Q2
r
2

r1

x
Q1

Since E1 and E2 are in the directions of r1 and r2 , you should now be able to sketch the
directions of the individual field. The specific math computations using the figure above
are carried out on the next page.
ECE 303 - Sum10 Notes Set 8 : Electric Charges and Fields 6

LINEAR FIELD EXAMPLE (cont.)

Substitute the expressions for the individual fields and factor out common terms:

Q1 Q2
4πε0jr1 j2 4πε0jr2 j2
ET = E1 + E2 = u1 + u2

 
1 Q1 Q2
=
4πε0 jr1 j 2
u1 +
jr2j2 u2
From the problem statement we can find the unit vectors and lengths:

r1 = 4 ux + 4 uy ) u1 = 0:71 ux + 0:71 uy and jr1 j2 = 32


r2 = 5 ux + uy ) u2 = 0:98 ux + 0:20 uy and jr2 j2 = 26

Substituting the above and equation (7) into the expression for ET then gives
" #
2 9 3 9
9 9 (10) 9 (10)
ET = 9(10) u1 + u2
32 26

2 h i 3 h i
= 0:71 ux + 0:71 uy + 0:98 ux + 0:20 uy
32 26

= 0:157 ux + 0:067 uy

Another form for ET is the magnitude-direction form. You should be able to show that the
boxed result above is equivalent to

ET = 0.171 u E , where u E = 0:918 ux + 0:392 uy

Problems: 8-3 through 8-6


ECE 303 - Sum10 Notes Set 8 : Electric Charges and Fields 7

ELECTRIC (COULOMB) FORCE

Once an Electric Field is established, there will be a FORCE on any other charge in that
field. This is sometimes called the Coulomb force.

Fixed and Mobile Charges


It will simplify our study to think of one charge, Q, as being in a fixed position and another
charge, q, as being free to move. In this way, Q produces the field E and q moves due to
this field. The force, in Newtons, exerted on q is

F = qE (9 )

Here is an example of how to compute the force.

EXAMPLE: h i
Let an existing electric field be E = 0:36 0:8 ux + 0:6 uy . Suppose a charge q = -5
nC is placed at the point (x; y) = (3; 4). Find the force which is exerted on charge q.
APPROACH:
Using (9), the force F on q is given by
h i
F = qE = 5(10) 9
 0 36
: 0:8 ux + 0:6 uy = 1:8(10) 9u E (10)

where u E = 0:8 ux + 0:6 uy .

The negative sign in equation (9) signifies the force is an ATTRACTION. We thus have the
following very useful result:

 Same polarity charges REPEL .


 Opposite polarity charges ATTRACT.

Problems: 8-7, 8-8


ECE 303 - Sum10 Notes Set 9 : Voltage 1

INTRODUCTION

Voltage is a type of Potential energy, which is the ability to do work. In fact, voltage is
often referred to as the electric potential. This set of Notes examines the definition and
origin of voltage using geometries of charges.

It is instructive to use some mechanical analogies to understand electric voltage. Below are
two mechanical analogies which will be very useful for us during this course:

Pendulum
One type of potential energy is the gravitational potential energy stored in a mass.
Raising the mass to a higher location increases the gravitational potential energy,
which then gives the mass the ability to do work. For example, upward displacement
of the mass at the end of a pendulum increases the gravitational potential of the mass.
Releasing the mass then causes the pendulum to move downward through its motion
trajectory.

Spring
Another type of potential energy is compressing or stretching a spring. Assume one
end of the spring is fixed and a load is attached to the other end. Then compressing
the spring “stores” potential energy in the spring. When the compressing force is
removed, the potential energy is transformed into the motion of the load.

Electrical voltage is somewhat more abstract since we don’t actually see the charges being
moved into the configuration which produces the voltage. However, we can sometimes
“see” the effect of the voltage. For example, the voltage of a flashlight battery moves
electrons through a filament, which in turn glows, producing the light.
ECE 303 - Sum10 Notes Set 9 : Voltage 2

DOT PRODUCT

The definition of voltage uses the vector concept of a dot product. It is easy to use Cartesian
coordinates to demonstrate some properties of the dot product. However, keep in mind that
the principles also apply to vectors in cylindrical and spherical coordinate systems.

Let W = Wx ux + Wy uy + Wz uz and V = Vx ux + Vy uy + Vz uz be vectors in Cartesian coor-


dinates. The Dot Product of these two vectors is given by
W  V = Wx Vx + Wy Vy + Wz Vz (2 )

An alternate formula for the dot product is


W  V = jWjjVj cos θwv (3 )

where θwv is the angle between the vectors W and V. An example will demonstrate using
the dot product.

EXAMPLE:
Let W = 3 ux + 4 uy and V = 2 ux + 5 uy . Drawing these vectors in the xy-plane shows
that the angle between them is small. To show this mathematically, equating (2) and
(3) above gives
Wx Vx + Wy Vy + Wz Vz = jWjjVj cos θwv (4 )
Now solve (4) for cos θwv , producing

cos θwv =
Wx Vx + Wy Vy + Wz Vz
jWjjVj = p p
(3)(2) + (4)(5) + (0)(0)
= 0:966
32 + 42  22 + 52
Thus, the angle θwv may thus be found as

θwv = cos 1 o
(0:966) = 15:1 = 0.263 rad.

It is frequently a good idea to draw the vectors (such as W and V above) to validate your
mathematical angle computation.

Problems: 9-1, 9-2, 9-3


ECE 303 - Sum10 Notes Set 9 : Voltage 3

CHARGES AND WORK

Now that we have reviewed the dot product, let’s examine the definition of voltage. We will
see that a configuration of charges creates the voltage “field”. For now, let’s consider the
simple charge configuration shown below. Assume both charges are positive and assume
charge Q is fixed in position. Charge qt is allowed to move.

Q q
t
A X
B
r = rA

r = rB

It is instructive to think of the interaction between Q and qt in three steps:

 The fixed charge Q produces a field E at point A.

 Thus, there is a force on qt , determined by F = qt E.

 This force F (or, equivalently, the field E) does work W in moving qt from point A to
point B.

From basic physics, the work (in Joules) done by the field (i.e., the force) is given by
Z Z
W = F  dl = qt E  dl (1 )
l l

The l subscript on the integral in (1) signifies a Line integral and dl is the line differential.
The “” in the integrand of (1) signifies the Dot Product.
ECE 303 - Sum10 Notes Set 9 : Voltage 4

DEFINITION OF VOLTAGE

Now consider the situation of using an external source of energy to move a charge qt against
an electric field E. Denote the work done by this external source as Wex . Since this work
is done against the Force (or against the Field) this work is thus the negative of the work
integral in equation (1): Z
Wex = F  dl (5 )
l
Since the force F on qt is due to the existing field E, we have F = qt E and therefore (5)
becomes Z
Wex = qt E  dl (6 )
l
The common engineering unit of Voltage, having units of Joules/Coulomb, is defined as the
work per unit positive charge done by the external source:
Z
qt E  dl
Wex l
V= = (7 )
qt qt
We will use the symbol V (or v) to denote voltage in this course. In evaluating (7) we see
that the charge qt cancels, giving the following definition of voltage in terms of the electric
field: Z
V= E  dl (8 )
l
Voltage is sometimes called a Potential Difference, which signifies voltage is a difference
in potential energy between two points. In the definition of equation (8) these two points
are the endpoints on the line integral.

You can check the sign of a voltage computation using physical arguments: If the direction
on the path l is Against the field, then work must be done in moving a positive test charge.
This should produce a Positive sign on the resulting computation of V in (8).
ECE 303 - Sum10 Notes Set 9 : Voltage 5

VOLTAGE FIELD FOR POINT CHARGE

We can use our prior results to find the scalar Voltage Field due to a point charge Q at the
origin. Due to the spherical symmetry of the E-field in this problem, it is easiest to work
in spherical coordinates. The electric field in spherical coordinates due to this charge has
been given in Notes Set 7 by
Q
E= ur (9 )
4π ε r2
The uθ and uφ components of E in (9) are zero and have been omitted for simplicity. The
line differential dl in spherical coordinates is given by :

dl = dr ur + r dθ uθ + r sin θ dφ uφ (10)

Substituting (9) and (10) into (8) then gives


Z  
V=
Q
ur  dr ur + r dθ uθ + r sin θ dφ uφ (11)
l 4π ε r
2

To show the relation between this problem for computing a line integral (in spherical co-
ordinates) and the previous example (in Cartesian), let’s divide the computation into two
steps: (a) Computation of the dot product, and (b) Evaluation of the scalar integral:

(a) Dot Product:


Note the uθ and uφ components of the E-field are zero. Therefore, the uθ and uφ com-
ponents of the line differential dl do not affect the dot product and thus equation (11)
simplifies to the following:

E  dl = ur  dr ur =
Q Q
dr (12)
4π ε r 2 4π ε r2
Substitute (12) into (11) for the voltage, giving the scalar integral
Z
Q
V= dr (13)
l 4π ε r
2

We now need to choose a path “l” and evaluate the line integral in (13). This is done on the
next page.
ECE 303 - Sum10 Notes Set 9 : Voltage 6

VOLTAGE FIELD FOR POINT CHARGE (cont.)

(b) Evaluation of Line Integral:


To evaluate the line integral in (16), we first need to choose a path “l” . Since the integrand
in (16) is only a function of r, we can therefore use a path which only varies as a function
of r, that is, a radial line intersecting the origin. This is a simple geometry which is shown
below:
circle,
radius = r0

E = Er u r at point x
Q
x ...
r
path l

Let the path start at r ! ∞, and then move inward along a radial direction such that dθ = 0
and dφ = 0 along the path. Let the stopping point on the path be r = r0 . Enforcing these
conditions on the integral in (13) thus gives
Z r = r0
Q
V (r0 ) = dr (14)
r = ∞ 4π ε r
2

Evaluating the integral in (14) then gives


Z r = r0  r0
Q dr Q 1
V (r 0 ) = = (15)
4πε ∞ r2 4πε r ∞

This example is completed on the following page.


ECE 303 - Sum10 Notes Set 9 : Voltage 7

VOLTAGE FIELD FOR POINT CHARGE (cont.)

Now we need to evaluate the limits in (15). As r ! ∞, we have ! 0 and thus the lower
1
r
limit in (15) goes to zero. Evaluation at the upper limit in (15) then gives

 
Q 1 Q
V (r0 ) = = (16)
4πε r0 4πεr0

Since r0 was completely arbitrary we may replace r0 in (16) with the spherical coordinate
radial variable r, giving the desired result:
Q
V (r) = (17)
4πεr
Equation (17) is very useful and very flexible for applications. Specifically, (17) may be
used even when the point charge is not located at the origin. For example, let Q be located
at a general point P. Then the voltage at a distance d from P is given by
Q
V (d ) = (18)
4πεd
In another application, suppose we know the value of Q but not its location. If we know the
value of voltage at point P is VP , then equation (18) allows us to determine how far point P
is from the location of Q:

VP =
Q
4π ε d
) d=
Q
4π εVp
(19)

An extension of these results shows that knowing the voltage at two points P1 and P2 allows
us to compute information about Q and/or the location of Q. This concept is illustrated in
the detailed problem solutions.

Problems: 9-4 through 9-8


ECE 303 - Sum10 Notes Set 9 : Voltage 8

VOLTAGE FOR MULTIPLE CHARGES

Voltage is also a linear field, which means the voltages due to separate charges add at a
point in space. For example, suppose we have two charges Q1 and Q2 which are fixed at
the locations x = x1 and x = x2 , and we want to compute the voltage at a point in between
the charges. This is shown in the figure below:

d1
d2
Q Q2
1

x = x1 x x = x2

To compute the voltage at the position x shown in the figure, we have

V (x) = V1 (x) + V2 (x) (20)

where V1 (x) and V2 (x) are the voltages due to Q1 and Q2 independently. Using equation
(21) for each of these voltages then gives
Q1 Q2
V (x) = + (21)
4 π ε d1 4 π ε d2
where d1 and d2 are the distances to each of the charge from the point x. Since d1 and d2
must be non-negative, equation (21) becomes
Q1 Q2
V (x) = + (22)
4 π ε (x x1 ) 4 π ε (x2 x)
This same approach can be used if the point x is outside the range x1  x  x2 . If this were
the case, draw a figure like the one above to help you compute the distances d1 and d2 in
terms of the charge locations. (Note: In either of the preceeding cases, the voltage at a
charge location is undefined, because the distance to the charge would then be zero.)

Problems: 9-9, 9-10


ECE 303 - Sum10 Notes Set 10 : Voltage Surfaces 1

INTRODUCTION

Now that we have studied voltage, we can begin to use visualization techniques for under-
standing many properties in electromagnetics. Specifically, voltage is a scalar field, which
is much simpler to understand than a vector field, such as the electric field. A scalar field
is also much simpler to draw. For example, we will see that a voltage which is a function
of two spatial variables x and y can be represented as a surface above the xy-plane.

However, a voltage “surface” is much more important than just providing us with a picture.
The voltage surface has geometrical properites, such as height, steepness, smoothness, and
curvature. We will see that there is a definite relation between the geometrical properties
of a voltage surface at a point and the corresponding electric field at that point. We can use
these geometrical properties to visualize the direction and magnitude of charge movement
without doing complicated mathematics.

In addition, there does exist a specific mathematical relation between the voltage surface
and the electric field. This means that we can rigorously compute the electric field at
any point if we know the mathematical expression for the voltage. The vector operator
linking the scalar voltage and the vector electric field is the gradient. We will discover
a very intuitive meaning for the gradient operator and use this property to enhance our
understanding of the electric field.
ECE 303 - Sum10 Notes Set 10 : Voltage Surfaces 2

PLOTTING A SURFACE

To help us eventually understand voltage surfaces, let’s first examine a simpler surface.
Plotting a simple three-dimensional surface will aid in developing visualization skills. Sup-
pose we have a surface defined by the equation
f (x; y) = x2 + y2 (1 )

For now, we don’t have to think of f (x; y) as a voltage. Just think of f (x; y) as a surface
having varying height above the xy-plane. You could use Matlab and mesh to plot this
surface very accurately. However, there are some simple ways for obtaining an approximate
sketch of the surface which will be accurate enough for our purposes.

In the left-hand figure below we have plotted the height of the surface f (x; y) above the
x-axis over the range 2  x  2 and the height of the surface above the y-axis over the
range 2  y  2. These lines trace out two parabolas above the xy-plane.

f(x,y)

f(x,y) f=4

f=4
f=4 f=4

y f = 2.5 y

f=1 f=1 f=1

1 2 x
x

Then in the right-hand figure above we have plotted the height of the surface above the xy-
plane for three values for height: f (x; y) = 1, f (x; y) = 2:5, and f (x; y) = 4. You can think
of all these lines as “stripes” on the surface f (x; y). With some imagination, you should
be able to “see” the parabolic surface f (x; y) rising “above” the xy-plane. This parabolic
surface resembles a “bowl”.
ECE 303 - Sum10 Notes Set 10 : Voltage Surfaces 3

PARTICLE MOTION

Now let’s perform a thought experiment to relate the function f (x; y) to energy and motion.
Think of the surface on the previous page as a physical surface, like a bowl, rising above
the flat xy-plane. Suppose we held a small ball on the surface directly above the point
(x; y) = (1; 1), or directly above (x; y) = ( 1; 1), as shown in the figure below. What
would happen at either place if the restraining force on the ball were removed? Obviously,
the ball would “roll down the surface”, as shown in the figure. Notice that the direction of
the roll is dependent on the initial position of the ball. Also note that the balls are rolling
toward the “bottom of the bowl”.

Path of Movement
on Surface Original Position

Original Position
f(x,y) = 2

Later Position
y
Later Position
(x,y) = (1,1)

x
(x,y) = (−1,−1)

In this example, the ball has been released in the scalar “gravitational potential energy
field”, and the ball, having positive mass, receives a force of movement due to the field.

This is exactly what happens when a charge q is released in a scalar voltage field V (x; y; z).
The charge moves under the influence of the voltage field in a manner similar to the ball
rolling down the surface. A positive charge will move from locations of higher voltage
(potential) in space to points of lower voltage (potential), much as the ball would move
from location of greater height to lower height.
ECE 303 - Sum10 Notes Set 10 : Voltage Surfaces 4

CONTOUR PLOT

Although the three-dimensional surfaces examined on the previous pages are very instruc-
tive, they are very cumbersome to reproduce every time we want to investigate charge
movement in a voltage field. Consequently, we desire some picture which is easier to draw
but which still has the same information as the three-dimensional surfaces. Such a reduced
dimension plot does exist and it is called a Contour Plot.

A contour plot is created by projecting the location of properties of the three-dimension


surface f (x; y) onto the xy-plane. To see how this is done, consider again the example of
the balls rolling down the f (x; y) surface on the previous page. The three-dimension picture
of this is shown again in the figure below:

f(x,y) = 2

y
(x,y) = (1,1)

f(x,y) = 0.5 x
(x,y) = (−1,−1)

In the above figure the balls’ path on the three-dimensional f (x; y) surface is projected
down onto the two-dimensional xy-plane. As the balls roll toward the bottom of the bowl,
their movement is directed toward the origin, which is the bottom of the bowl. In the
two-dimensional xy-plane this is represented by the two arrows pointing inward toward the
origin. We will next show how to extract this same information in a simpler manner using
the contour plot.
ECE 303 - Sum10 Notes Set 10 : Voltage Surfaces 5

CONTOUR PLOT (cont.)

Important properites in the current example are the initial locations of the balls, their initial
directions of roll, and some measure relating to the initial velocities of their roll. The
contour plot provides much of this information. Referring to the three-dimensional f (x; y)
surface on the previous page, consider the locus of all points in the xy-plane which give a
surface height f (x; y) = 2. Also consider the locus of all points in the xy-plane which give
a surface height f (x; y) = 0:5. In the xy-plane these loci are the circles x2 + y2 = 2 and
x2 + y2 = 0:5, which have been drawn in the figure below:

y
x 2+ y 2 = 2

. (x,y) = (1,1)

. Initial Position

x Later Position

(x,y) = (−1,−1)
. x 2 + y 2 = 0.5

In terms of the above contour plot, the balls were initially released at their respective points
on the circle x2 + y2 = 2 and they begin moving toward the origin. Suppose that at the
later time the balls have now rolled down to a height of f (x; y) = 0:5. Therefore, the balls’
xy-positions on the contour plot have now reached the circle x2 + y2 = 0:5. This situation is
represented by the arrows in the above figure. The arrows trace the movement of the balls
projected onto the xy-plane.

One the following pages, we will study more about contours, how to plot them, and how
they are useful in voltage problems.
ECE 303 - Sum10 Notes Set 10 : Voltage Surfaces 6

PLOTTING CONTOURS

If we know a two-dimensional function g(x; y) we can plot the contours. For example,
suppose we have the two-dimensional function

g(x; y) = x2 + 9 y2 (2 )

Suppose we want to plot the contours in the xy-plane corresponding to the surface heights

g(x; y) = K (3 )

where K is a constant. We can solve for the entire family of contours by substituting
equation (2) into equation (3):
x2 + 9 y2 = K (4 )

Equation (4) is recognized to be the equation for a family of ellipses whose size depends
on the value of K. For example, if we want the contour in the xy-plane corresponding to
the height of K = 50 on the g(x; y) surface, then we substitute K = 50 into equation (4):

x2 + 9 y2 = 50 (5 )

We can draw this ellipse by solving for the points where it crosses the x-axis and the points
where it crosses the y-axis. When it crosses the x-axis, we have y = 0 in (5), and therefore
equation (5) becomes
p
x2 = 50 ) x=  50 =  7 07
: (6 )

When the ellipse crosses the y-axis, we have x = 0 in (5), and therefore equation (5) be-
comes r
9 x = 50 ) x =  =  2:36
2 50
(7 )
9
When the ellipse crosses the line y = 0:5 x, we have y = 0:5 x in (5), and therefore equation
(5) becomes
x2 + 9 (0:5 x)2 = 50 ) x =  3:92 (8 )

These are enough points to sketch the K = 50 ellipse, which is done on the next page.
ECE 303 - Sum10 Notes Set 10 : Voltage Surfaces 7

PLOTTING CONTOURS (cont.)

From our work on the previous page, we now have enough points to sketch the K = 50
ellipse, which is shown in the figure below:

y y = 0.5x
2 2
x + 9y = 50
ellipse y = 2.36

x = 3.92
x = 7.07
x = −7.07

y = −2.36

Although plotting a single contour is sometimes useful, the real value of the contour plot
arises when when plot the contours for several heights on the surface. The figure below
plots the elliptical contours for the surface of equation (4) when K = 25; 50; 75; and 100.

x 2 + 9y 2 = K
ellipses y K = 25 50
75
100

x = 10
x = −10
x = −7.07 x = 7.07

Problems: 10-1 through 10-5


ECE 303 - Sum10 Notes Set 10 : Voltage Surfaces 8

E-FIELD USING GRADIENT

The previous material in this Notes Set discussed geometrical properties of voltage sur-
faces and contour plots. A natural question is how a voltage surface (which is a scalar
field) relates to the electric field (which is a vector field). Another question concerns the
mathematical operators which can give us quantitative properties of these fields. This next
section of Notes addresses these questions.

We found earlier that if we know the vector electric field then we can find the scalar voltage
using the line integral : Z
V = E  dl (9 )
l
Now we are interested in the reverse question: if we know a voltage function V (x; y; z),
can we find the electric field E(x; y; z) ? The answer is “Yes”, by using the mathematical
operation of the Gradient of V . The computational formula is

E= ∇V (10)

where “∇” in (10) is the gradient operator. The math term for “∇V ” is “gradient of V ”.
Note that V is a scalar, but ∇V is a vector. Once again, ∇V is coordinate-specific and there
are three forms:

∂V ∂V ∂V
Cartesian: ∇V = ux + uy + uz (11a)
∂x ∂y ∂z

∂V 1 ∂V ∂V
Cylindrical: ∇V = ur + uφ + uz (11b)
∂r r ∂φ ∂z

∂V 1 ∂V 1 ∂V
Spherical: ∇V = ur + uθ + uφ (11c)
∂r r ∂θ r sin θ ∂φ

In this course we will primarily study the Cartesian form for the gradient. We will examine
an example on the next page.
ECE 303 - Sum10 Notes Set 10 : Voltage Surfaces 9

GRADIENT EXAMPLE

Let’s work an example which shows the mechanics of computing the gradient of a scalar
function. For simplicity, the first example will use Cartesian coordinates.

EXAMPLE:
Let a scalar function of x and y in Cartesian coordinates be given by
2 3
F = F (x; y) = 4 x y + z

Compute the gradient of F, which is denoted as ∇F.

APPROACH:
From equation (11a) we have

∂F ∂F ∂F
∇F = ux + uy + uz
∂x ∂y ∂z
Compute the partial derivatives required above, giving

∇F 2 2
= 4y ux + 8xy uy + 3z uz

Note that ∇F is a vector. Notice also that ∇F has a different magnitude and direction for
different values of (x; y; z).
ECE 303 - Sum10 Notes Set 10 : Voltage Surfaces 10

E-FIELD EXAMPLE

As another example of using the gradient, let’s compute the vector E-field due to a point
charge by starting with the scalar voltage field surrounding the charge. From previous
work, the voltage field due to a charge Q at the origin is given by
Q
V (r) = (12)
4πεr
Computing the gradient in spherical coordinates will keep us from having to rewrite (12) in
terms of Cartesian x, y, and z. Now operate on both sides of (12) with the gradient operator
in spherical coordinates from equation (11c):
 ∂V 1 ∂V 1 ∂V

E = ∇V = ur + uθ + uφ (13)
∂r r ∂θ r sin θ ∂φ

But notice from (12) that V (r) is only a function of radial variable r. Therefore, the φ and
θ partial derivatives in equation (13) vanish:

∂V ∂V
= 0; =0 (14)
∂φ ∂θ
Substitute the results from (14) into the expression for E in (13):

∂V ∂
 Q
 Q ∂
1
E = ur = ur = ur (15)
∂r ∂r 4πεr 4πε ∂r r

Computing the r partial derivative in (15) then provides the desired result:
Q
E = ur (16)
4πεr2
Equation (16) is the same result as given by Coulomb’s law. However, notice that it was
computationally simpler to start with V and differentiate to obtain E (this example) that it
was to start with E and integrate to obtain V .

Problems: 10-6 through 10-9


ECE 303 - Sum10 Notes Set 10 : Voltage Surfaces 11

VOLTAGE FOR CHARGE PAIR

Some interesting physical voltage phenomena can occur when there are multiple fixed
charges. Suppose we have two charges Q1 and Q2 which are fixed at the locations x = x1
and x = x2 , as shown in the figure below:

d1
d2
Q Q2
1

x = x1 x x = x2

Let’s compute the voltage in the region x1 < x < x2 . For any x-value in this range we have
Q1 Q2
V (x) = + (17)
4 π ε d1 4 π ε d2
where d1 and d2 are the distances to each of the charge from the point x. Since d1 and d2
must be non-negative, equation (17) becomes
Q1 Q2
V (x) = + (18)
4 π ε (x x1 ) 4 π ε (x2 x)

There is some value of x in the range x1 < x < x2 which gives a minimum for voltage for
this x-range. Call this value xmin and denote the corresponding voltage as Vmin . We can find
this value xmin by solving the differential equation
d
V (x) = 0 (19)
dx
for the value of xmin . We will do an example on the next page which demonstrates this
procedure.
ECE 303 - Sum10 Notes Set 10 : Voltage Surfaces 12

VOLTAGE MINIMUM EXAMPLE

Consider again the charge configuration shown in the figure on the preceding page. There
is a voltage minimum somewhere in the region x1 < x < x2 . To show this let’s assume
free-space and work a specific example having the following parameters:
10 10
x1 = 5 ; Q1 = 10 Coul ; x2 = 10 ; Q2 = 3(10) Coul
Substituting these values into (18) and remembering that 1=4πε0 = 9(10)9, we get the
following equation:
0:9 2:7
V (x) = + (20)
(x 5) (10 x)
Now take the first derivative of (20) and set it equal to zero:
2:7 0:9
=0 (21)
(10 x)2 (x 5)2

Now add the terms in (21), and simplify, obtaining


2:7 (x 5)2 0:9 (10 x)2 = 0 (22)

Multiply out the squares in (22) and then collect terms, producing
1:8 x2 9x 22:5 = 0 (23)

You can use the quadratic formula on equation (23) to find the roots are
x = 1:83 ; x = 6:83 (24)

The positive root in (24) is in the range 5 < x < 10. Thus,
xmin = 6:83 (25)

The voltage minimum Vmin is obtained by using (20) to evaluate V (x) at x = xmin = 6:83:
0:9 2 :7
Vmin = + = 1:344 (26)
(6:83 5) (10 6:83)

Problems: 10-10, 10-11, 10-12


ECE 303 - Sum10 Notes Set 11 : Gauss’ Law 1

INTRODUCTION

A fundamental property we have seen is that charges create electric fields. Sometimes
the charges are discrete charges or point charges, as we studied previously. However, in
many physical structures it is more instructive to consider these discrete charges as being
distributed over a spatial extent. This creates the concept of a distributed region in space
which has a particular “charge density”. An example of such a charged region is the deple-
tion region near the PN-junction of a semiconductor diode.

If we begin with a knowledge of the charge density in a region, then Gauss’ Law gives us
a way to compute the electric field E over the region. This might at first seem to be more
complicated than just using Coulomb’s Law to compute the E-field given just a point charge
Q. However, the region which is modeled as having a charge density may actually comprise
an enormous number of individual discrete charges. To use Coulomb’s Law to find the E-
field component due to each of these discrete charges may not be computationally feasable.
Consequently, an alternate approach is to use Gauss’ Law, starting with the charge density
due to the individual charges.

There are two forms of Gauss’ Law: the Integral Form and the Differential Form. The in-
tegral form considers the E-field existing over a surface in space. However, the differential
form gives us E-field at every point within the surface, as well as the E-field on the surface.
For these reasons, the differential form is also often called the Point Form of Gauss’ Law.
This set of Notes studies the differential form of Gauss’ Law. The PN-junction voltage
from solid-state electronics is used as a motivating application to show us how Gauss’ Law
can be applied to a useful problem.

If you take a course in solid-state electronics you may learn about the PN-junction in much
more detail. The diffusion equation and current density equations are often used to develop
very quantitative results for hole and electron flow across the junction. In this notes Set, we
will only examine the PN-junction from the standpoint of Gauss’ Law, which is in keeping
with our study of electromagnetic phenomena.
ECE 303 - Sum10 Notes Set 11 : Gauss’ Law 2

THE P-N JUNCTION

This page describes a simple model for the hole and electron densities on the different
sides of the PN-junction, prior to their being joined to actually form the junction. On
the following page, we will place the sides in contact and observe the hole and electron
movement.

Lattice Geometry
In the figure below, the bulk material on both sides of the dotted line is silicon. Atoms from
column V in the periodic chart have been added to the silicon lattice to form the N-type
region on the right side and atoms from column III in the periodic chart have been added
to the silicon lattice to form the P-type region on the left side. These added atoms are often
called the impurities.

+4 +3 +4 +5

+ _
+4 +4 +4 +4
_
+
+3 +4 +5 +4

P−type N−type
(Acceptor) Junction (Donor)

Donors and Acceptors


At room temperature, one of the outer shell electrons from the column V impurity is held
very loosely by its nucleus and is therefore very mobile. This material is called the N-type
(or Donor) because these mobile electrons are Negative charges. Conversely, there is an
electron vacancy in the outer shell of the column III impurity, and this vacancy can accept
mobile electrons. However, it is conceptually easier to envision the electron “vacancy” as
moving through the lattice. This vacancy is called a hole, and it is a virtual Positive charge.
Thus, this material is called the P-type material (or Acceptor).
ECE 303 - Sum10 Notes Set 11 : Gauss’ Law 3

DIFFUSION OF HOLES AND ELECTRONS

We can show that the diffusion of holes and electrons across the PN-junction creates an
electric field. It is this field which eventually stops the diffusion and causes the process to
reach a steady-state.

Before Diffusion:
Consider the figure below, which shows a mobile hole and a mobile electron before dif-
fusion begins. The +3 and +5 are the column III and column V nuclei which are held
immobile in the semiconductor lattice:

BEFORE Diffusion:
Junction

+3 +5

+ _
+4 +4
P−type N−type
Diffusion

After Diffusion:
Now let the diffusion process occur and suppose the mobile electron diffuses to the left and
fills the hole. These mobile charges cancel, but the bound charge imbalances due to the +3
and +5 lattice nuclei still remain:

AFTER Diffusion: Junction

+3 +5

+4 +4
P−type N−type
ECE 303 - Sum10 Notes Set 11 : Gauss’ Law 4

ELECTRIC FIELD

Note from the bottom figure on the previous page that, after diffusion, the +3 lattice nuclei
in the P-type material is surrounded by 4 outer shell electrons. Therefore, there is a resulting
“ q” charge on the left. Additionally, we see that the diffusion of the electron from the
N-type material on the right results in a charge of +q remaining at the +5 location in the
N-type material.

Therefore, from electrostatics we see there is an Electric Field created between the +q
charge and the q charge. Thus, diffusion generates bound stationary charges and an
Electric field is created.

Since there are many mobile elections on the right, another electron will try to move to the
left by diffusion. However, note the q previously created will oppose the diffusion of the
second electron to the left. As many electrons diffuse, the opposing electric field builds up,
until no more electrons from the right can diffuse. A steady state electric field Ex (x) then
exists, as shown in the figure below:

Junction
P−type N−type
_ +
_ +
_ + E−field
_ +
_ +
x = x1 x = x2

This region between x1 and x2 in the figure is sometimes called the Depletion Region in
a course on solid-state electronics. From the figure you can see that the “bound” charges
form two regions of charge density. On the following pages, we will show how we can use
Gauss’ Law to determine the electric field if we know these original charge densities.
ECE 303 - Sum10 Notes Set 11 : Gauss’ Law 5

GAUSS’ LAW: DIFFERENTIAL FORM

This concept of a distributed charge region (such as that in the depletion region) can be
quantified by introducing the Charge Density. Charge density has the symbol ρ, and has
units of coulombs per m3 . When the charge environment is given by charge density ρ in
a material of permittivity ε, then the electric field E is given by Gauss’ Law. One form of
Gauss’ Law is called the differential form and is given by
ρ
∇ E =
ε
(1 )

The vector operation on the left side of equation (1) is called the Divergence operation. The
mathematical equation resulting from Gauss’ Law in (1) is a Partial Differential Equation
for the electric field E, given a charge density ρ and a medium with permittivity ε. The math
form of the divergence depends on the coordinate system used. For example, in Cartesian
coordinates the divergence is given by

∂Ex ∂Ey ∂Ez


∇ E =
∂x
+
∂y
+
∂z
(2 )

Substituting (2) in (1) then gives

∂Ex ∂Ey ∂Ez ρ


+ + = (3 )
∂x ∂y ∂z ε

You can think of the charges, represented by ρ = ρ(x; y; z), as creating an E-field with
components Ex , Ey , and Ez . To find these components, we would generally have to solve
equation (3) for Ex ; Ey; and Ez . However, specific problems can often be made simpler by
using geometrical considerations. To help our intuitive understanding of Gauss’ Law, let’s
first examine the geometrical properties of divergence. This will in turn help us understand
the physical interpretation of Gauss’ Law.

Problems: 11-1, 11-2


ECE 303 - Sum10 Notes Set 11 : Gauss’ Law 6

DIVERGENCE - PHYSICAL INTERPRETATION

Many problems in electromagnetics can be simplified if we understand the geometical


properties of the mathematical operators. Such an operator is the divergence operator given
in the statement of Gauss’ Law in equation (1).

To develop our understanding, let’s consider equation (3) when the partial derivaties with
respect to y and z are equal to zero and charge density ρ is only a function of x. This
does not trivialize the problem, as will be explained in class. Using these simplifications,
equation (3) becomes
dE ρ(x)
= (4 )
dx ε
where we have also set Ex = E for simplicity. Now recall the definition of the ordinary
derivative in (4):
dE E (x) E (x ∆x)
∆x ! 0
= lim (5 )
dx ∆x
If ∆x in (5) is small, but non-zero, we can approximate the derivative in (5) by the right-
hand expression without the limit operation:

∆x)
dE
dx
 E (x) E (x
∆x
(6 )

Now substitute (6) into (4) and rearrange somewhat:

ρ(x)
E (x) E (x ∆x) = ∆x (7 )
ε
Now we can investigate (7) for the two possible cases: (i) the region ∆x does not con-
tain charge, or (ii) the region ∆x does contain charge. These regions are explored on the
following page.
ECE 303 - Sum10 Notes Set 11 : Gauss’ Law 7

DIVERGENCE - PHYSICAL INTERPRETATION (cont.)

ZERO CHARGE:
If the region ∆x does not contain any charge, then we have ρ(x) = 0 and (7) becomes
E (x) = E (x ∆x) (8)

Equation (8) signifies that for no charge in the region ∆x, then the E-field is the same on
either side of the ∆x segment. This non-zero E-field has been produced by a charge not
located in the ∆x segment. A figure describing this situation is shown below:

E(x − ∆ x) E(x)

x− ∆x x
NON-ZERO CHARGE:
If the region ∆x does contain charge, then the fields on either side of the ∆x segment are
not equal. To see this, observe the situation in the figure below:

E(x − ∆ x) E(x)

E q(x − ∆ x)
q Eq (x)
x− ∆x x

The fields E (x) and E (x ∆x) due to charges not in ∆x are shown, as in the preceding figure.
However, due to the presence of charge q within ∆x, there is an additional field Eq (x ∆x)
at x ∆x and a field Eq (x) at location x. These two fields are in opposite directions, as
required by Coulomb’s Law. This example is completed on the following page.
ECE 303 - Sum10 Notes Set 11 : Gauss’ Law 8

DIVERGENCE - PHYSICAL INTERPRETATION (cont.)

To obtain the total fields at x and x ∆x we have to add the two field components existing
at each boundary, as shown on the following page. Let the subscript s, for sum, represent
the summation of the fields at the boundaries of the ∆x segment. Visually adding the E and
Eq fields shown in the previous figure then gives the resulting fields shown below:

E s(x − ∆ x) q E s (x)

x− ∆x x

You can see in the figure that the fields are no longer equal on each side on the segment.
As ∆x ! 0, there is a field discontinuity at the location of the charge, or more strictly, the
location of the charge density. With the preceding understanding, we now have a simple
interpretation of the differential form of Gauss’ Law:

If the divergence of the E-field is non-zero at a point in space, then there is


charge (or charge density) located at that point.

When we have a volume of space which has a non-zero charge density, like the PN-junction
depletion region, then the divergence of the E-field is non-zero over the entire volume.
Similarly, if we have a region where the E-field divergence has been determined to be non-
zero, then we are guaranteed that a non-zero charge density exists throughout that region.
ECE 303 - Sum10 Notes Set 11 : Gauss’ Law 9

E-FIELD DIFFERENTIAL EQUATION

We will work an example using the differential form of Gauss’ Law to demonstrate how to
find the electric fields, given the charge density.

EXAMPLE: Let a region on the x-axis have the following charge density:

K; w<x0
ρ(x) = (9 )
K; 0<x<w

Find the differential equation for E over the region w < x < w.

APPROACH: Start with Gauss’ Law in Cartesian coordinates from equation (3):

∂Ex ∂Ey ∂Ez ρ(x)


∇E = + + = (10)
∂x ∂y ∂z ε

where ρ(x) is the charge density function is given in (9). Since ρ(x) is only a function
of x, then E = Ex ux is only a function of x. Thus, only the partial derivative with
respect to x in (10) is non-zero, and equation (10) becomes

∂Ex ρ(x) K =ε; w<x0
= = (11)
∂x ε K =ε; 0<x<w

Since there is only one independent variable (x) we may replace the partial derivative
in (11) with an ordinary derivative. Equation (11) then becomes

dEx K =ε; w<x0
= (12)
dx K =ε; 0<x<w

Equation (12) gives the two differential equations which must be solved to find the Ex -field.
Other properties, such as continuity and boundary values, are necessary to actually solve
these equations for Ex .
ECE 303 - Sum10 Notes Set 11 : Gauss’ Law 10

BOUNDARY VALUE EXAMPLE

Computing a specific E-field requires that the solution of the ODE, like equation (12) on
the previous page, satisfies specific Boundary Conditions. An example of Boundary Con-
ditions is shown in the following example.

EXAMPLE: Let the charge density in a region w < x < w be given by



K; w<x0
ρ(x) = (13)
K; 0<x<w

Compute Ex (x), subject to the Boundary Conditions Ex (w) = Ex ( w) = 0.

APPROACH: Equation (13) above is actually two ODEs. Therefore, solve the ODE for
each region, and apply the boundary condition appropriate for each region.

LEFT REGION: w<x0


Since ρ(x) = K in this region, the following differential equation holds:
dEx K
= (14)
dx ε
Integrating both sides of the differential equation in (14) gives
K
Ex (x) = x + C1 (15)
ε
where C1 is a constant of integration. Apply the left-side boundary condition:

Ex (x = w) = 0 ) K
ε
( w) + C1 = 0 ) C1 =
K
ε
w (16)

Substituting the value for C1 from (16) into equation (15) for Ex (x) then gives
K K
Ex (x) = x+ w (17)
ε ε
ECE 303 - Sum10 Notes Set 11 : Gauss’ Law 11

BOUNDARY VALUE EXAMPLE (cont.)

RIGHT REGION: 0 < x w


Since ρ(x) = K in this region, the differential equation is:
dEx K
= (18)
dx ε
Integrating both sides of equation (18) gives
K
Ex (x) = x + C2 (19)
ε
where C2 is a constant of integration. Apply the right-side boundary condition to (19):

Ex (x = w) = 0 ) K
ε
(w) + C2 = 0 ) C2 =
K
ε
w

Putting this value for C2 into (19) for Ex (x) gives


K K
Ex (x) = x+ w (20)
ε ε
A plot of Ex (x) as a function of x is shown below

E x (x)
_K w
ε

x=−w x=w
x=0

Other E-field problems may be solved in a similar manner.

Problems: 11-3 through 11-9


ECE 303 Sum10 Notes Set 12 : Poisson/Laplace Equations 1

INTRODUCTION

We have previously seen how a charge density created an electric field. We have also seen
how an electric field produced a voltage difference between two points in space. Therefore,
computing a voltage difference based upon an initial charge density required the solution
of two different equations. We naturally wonder if there is perhaps a more straightforward
way of computing a voltage based upon a specified charge density.

Indeed there is such a straightforward method and it is the topic of this Notes set. Although
only one new approach is needed, there has been a historical division in the names given
to the resulting equations. This division is based upon whether the region in question has a
zero charge density or a non-zero charge density. These equations are called:

 Laplace’s Equation:
If the region under consideration has a zero charge density, the resulting equation is
called Laplace’s Equation. This equation is useful in regions which are dielectrics,
insulators, or free-space. We will use the region in the dielectric between the plates
of a capacitor to illustrate this equation.

 Poisson’s Equation:
If the region under consideration has a non-zero charge density, the resulting equation
is called Poisson’s Equation. This equation is useful in regions which have a charge
denisty, like the depletion region near a PN junction.

Laplace’s Equation is usually easier to mathematically solve. Therefore, we will concen-


trate on solving Laplace’s Equation for regions of zero charge density. However, the ex-
tension is straightforward to charged regions (Poisson’s Equation), and this is explored in
some of the problems.
ECE 303 Sum10 Notes Set 12 : Poisson/Laplace Equations 2

THE EQUATIONS

There is a way to find voltage, V , directly from charge density, ρ, without having to first
find the E-field. To derive this method, start with the point form of Gauss’ Law:
ρ
∇E = (1 )
ε
Recall that the E-field is the negative gradient of voltage

E= ∇V (2 )

Thus, substituting (2) into (1) thus gives


ρ
∇( ∇V ) = (3 )
ε
Multiplying both sides of (3) by minus one produces the form known as Poisson’s equation:
ρ
∇  (∇V ) = (4 )
ε
The form ∇  (∇V ) on the right-hand side of (4) is not a dot product ... it is the Divergence
of the Gradient of V . The simplified notation

∇2V 
= ∇ (∇V ) (5 )

is frequently used. Thus, Poisson’s equation is often written as


ρ
∇2V = (6 )
ε
It is easy to see that if the charge density ρ equals zero, then

∇2V =0 (7 )

Equation (7) is called Laplace’s Equation. It is valid wherever the medium under considera-
tion has no charges. Examples of where this occurs are in free-space away from conductors,
or in an insulator separating conductors.
ECE 303 Sum10 Notes Set 12 : Poisson/Laplace Equations 3

SECOND-ORDER DIFFERENTIAL EQUATIONS

Laplace’s Equation and Poisson’s Equation are examples of second order partial differential
equations. To solve these equations, we have to consider the specific form for ∇2V . For
example, in Cartesian coordinates we have

∂2V ∂2V ∂2V


∇V 2
= + + (8 )
∂x2 ∂y2 ∂z2
In many of our cases, there may be a spatial symmetry and consequently voltage will vary
only as a function of one variable. For example, suppose voltage varies only as a function
of x. Laplace’s equation thus becomes
 
∂2V ∂ ∂V
∇V2
= = =0 (9 )
∂x2 ∂x ∂x

Since (9) has only one independent variable, it is equivalent to the second order ordinary
differential equation (ODE) given by
 
d 2V d dV
= =0 (10)
dx2 dx dx

Equation (10) says that V = V (x) is a function whose second derivative (with respect to x)
is zero. Thus, the first derivative of V (x) is a constant, or

dV
= K1 (11)
dx
where K1 is the unknown constant. We have transformed the second-order ODE into a
simpler first-order ODE. Geometrically, this implies that the Slope of V (x) is constant. To
find V (x) we integrate the first-order ODE in (11) to obtain

V = V (x) = K1 x + K2 (12)

where K2 is another constant of integration. It is easy to see that V (x) is the equation of a
straight line. The specific constants K1 and K2 are obtained by using boundary conditions.
ECE 303 Sum10 Notes Set 12 : Poisson/Laplace Equations 4

LAPLACE’S EQUATION EXAMPLE

A parallel plate capacitor has its upper plate at z = 0 and its lower plate at z = d. The
voltage on the upper plate is V (z = 0) = V0 volts and the voltage on the lower plate is
V (z = d ) = 0 volts. Compute the voltage in the region d < z < 0.

APPROACH:
Since there is only one independent variable (z) we may replace partial derivatives with
ordinary derivatives, giving  
d dV
=0 (13)
dz dz
Integrating equation (13) once gives
dV
= K1 (14)
dz
where K1 is a constant of integration. Integrating (14) then gives
V (z) = K1 z + K2 (15)

where K2 is another constant of integration. To evaluate K1 and K2 we have to use the


boundary conditions or boundary values for the specific problem.

Boundary Conditions:
V (z = 0) = V0 : ) K1  (0) + K2 = V0 ) K2 = V0

V (z = d ) = 0: ) K1  ( d ) + K2 = 0 ) K1 =
V0
d

Using these evaluations for K1 and K2 in (14) we see that the equation for the voltage in the
dielectric is given by
V0
V (z) = z + V0 (16)
d

Problems: 12-1 through 12-5


ECE 303 Sum10 Notes Set 12 : Poisson/Laplace Equations 5

CYLINDRICAL COORDINATES

Since many electrical structures (like many wires and coaxial cables) are cylindrical it is
sometimes easier to work problems using Cylindrical Coordinates. The coordinates of a
point P in cylindrical coordinates are specified by the three values r, φ, and z, where

 r is the radial distance from the origin to the projection of P onto the xy-plane,
 φ is the rotation angle from the x-axis to the radial containing the projection of P onto
the xy-plane,
 z is the vertical distance from the xy-plane to the point P,

An example of the point P = (r; φ; z) defined in cylindrical coordinates is shown below:

z−axis

.
P = ( r, φ , z )

y−axis

r−direction
φ angle
x−axis

Here are the equations for mapping the Cartesian coordinates x; y and z into the correspond-
ing cylindrical r; φ and z
p y
r = x2 + y2 ; φ = tan 1 ; z=z
x
We will next work an example solving Laplace’s equation in cylindrical coordinates.
ECE 303 Sum10 Notes Set 12 : Poisson/Laplace Equations 6

CYLINDRICAL COORDINATE EXAMPLE

Let’s work an example using Laplace‘s Equation in cylindrical coordinates. Cylindrical


coordinates are very useful when the structure under consideration (like a wire or cable)
has a cylindrical geometry.

EXAMPLE:
Let a very long coaxial cable have inner conductor with radius r = a and outer conductor
with radius r = b. The conductors are separated by an insulating dielectric material. The
voltage on the outer conductor is V (r = b) = V0 and the voltage on the surface of the inner
conductor is V (r = a) = 0. Compute the voltage in the region a < r < b.

APPROACH:
In cylindrical coordinates the Laplacian of V is given by
 
1 ∂ ∂V 1 ∂2V ∂2V
∇V2
= r + + (17)
r ∂r ∂r r2 ∂φ2 ∂z2

By symmetry considerations, the voltage in the coaxial cable dielectric varies only as a
function of r. Therefore, the partial derivatives in (17) with respect to r and φ are zero.
That is,
∂V ∂V
= 0; =0 (18a; b)
∂z ∂φ
Using these simplifications in (17) thus produces
 
1 ∂ ∂V
∇V
2
= r (19)
r ∂r ∂r

Using (19), Laplace’s equation in cylindrical coordinates (equation (7)) thus becomes
 
1 ∂ ∂V
r =0 (20)
r ∂r ∂r

This example is continued on the next page.


ECE 303 Sum10 Notes Set 12 : Poisson/Laplace Equations 7

CYLINDRICAL COORDINATE EXAMPLE (cont.)

Multiplying both sides of (20) by r then gives


 
∂ ∂V
r =0 (21)
∂r ∂r

Finally, since there is only one independent variable (r) in (21), we can replace the par-
tial derivatives in (21) with ordinary derivatives. Therefore, (21) becomes the ordinary
differential equation  
d dV
r =0 (22)
dr dr
We next need to solve equation (22) for the unknown voltage V = V (r). To do this, integrate
both sides of (22), giving
dV
r = K1 (23)
dr
where K1 is a constant of integration. Dividing both sides of (23) by r then provides
dV K1
= (24)
dr r
Integrating (24) then gives the desired result

V (r) = K1 ln r + K2 (25)

where K2 is another constant of integration independent of r. As before, K1 and K2 are


found using the boundary conditions.

Problems: 12-6 through 12-10


ECE 303 - Sum10 Notes Set 13 : Time-Invariant Magnetic Fields 1

INTRODUCTION

Stationary electric charges produce only an electric field in space. This phenomenon is
often called Electrostatics. However, when these charges move to create a current, a mag-
netic field is also produced in space. This magnetic field can then effect other moving
charges, or, equivalently, other currents. The inclusion of magnetic field phenomena with
electric field phenomena is often called Electromagnetics. If the charges are moving with
constant velocity, then the magnetic fields produced are time-invariant. However, these
time-invariant magnetic field can vary as a function of space.

When charges accelerate, that is, change their velocity, then the magnetic field becomes
time-varying, as well as spatially-varying. Additionally, electromagnetic energy can also
be radiated in this case, which means the energy can be propagated in space. Transmit-
ting antennas are structures which support this charge acceleration and specifically allow
electromagnetic energy to be radiated in an effective manner. An AC current is also an
example of charges which undergo acceleration. Time-varying electric and magnetic fields
are examined in future sets of Notes.

This set of notes studies the magnetic phenomenon which occurs when time-invariant cur-
rent flows. An example of this time-invariant current is DC flowing in a wire or a conduc-
tor. We will find that this time-invariant current produces a time-invariant magnetic field in
space.
ECE 303 - Sum10 Notes Set 13 : Time-Invariant Magnetic Fields 2

MAGNETIC FORCE

Our primary interest in this course is the magnetic field produced by currents, that is, by
large numbers of charges in motion. However, it is instructive to first see how the magnetic
field affects one moving charge. Recall that an electric charge q placed in an electric field
E experiences a force given by F = q E. Note that this force is in the same direction as
the E field. Now let the charge be moving with vector velocity v in a region where there
is a existing magnetic flux density B, having units of Webers/m2 . Then the moving charge
experiences an additional force Fm (in newtons) given by the cross product
Fm = q ( v X B ) (1)

This magnetic force is illustrated in the sequence of figures below:

z z z
v v v

Fm
q q q
y y y
B B
x x x

The figure on the left is the charge q moving in the z-direction, with no magnetic flux
density present. The middle figure shows the appearance of the magnetic flux density B in
the x-direction. The right figure then shows the direction of the resulting force Fm along
the y-axis. This means the velocity of the charge will now change in response to the force.
Consequently, the velocity vector will no longer be aligned with the z-axis. Thus, we see
that the presence of q B-field changes the direction of q moving charge. The direction
change is orthogonal to both the initial velocity and the direction of the magnetic field.
The cross product operation quantifies this orthogonal direction. If you need a review of
the mathematics of the cross product you can see the information in the Review Notes on
Vectors.

Problems: 13-1, 13-2


ECE 303 - Sum10 Notes Set 13 : Time-Invariant Magnetic Fields 3

MAGNETIC FIELD UNITS

Now let’s examine the units of the fields related to magnetics. Even though equation (1) is a
cross product, all the terms of the cross-product will have units given by the scalar relation

F = qvB (2 )

Let’s write equation (2) in terms of the units of the quantities:


   m   
Newtons =( Coul )  sec
 Webers
m2
(3)

But from basic circuits we know that


 
Coul
= ( Amp ) (4)
sec

Using (4) in (3) then gives


   
Newtons =( Amp )  (m)  Webers
m2
(5)

Solve (5) for the units of Webers/m2 , obtaining


 
Webers Newtons
= (6)
m2 Amp - m

“Webers per meter-squared” seems to be the units of q density (due to the appearance of
meters-squared). Therefore, let’s examine what a Weber might be. Multiply both sides of
(6) by m2 , obtaining
   
Webers =
Newtons
Amp - m
 m2 =
Newton - m
Amp
(7)

But a Newton-meter is a Joule and therefore (7) becomes


 
Joules
Webers = (8)
Amp
ECE 303 - Sum10 Notes Set 13 : Time-Invariant Magnetic Fields 4

MAGNETIC FIELD UNITS (cont.)

Equation (8) has shown that the units for a Weber are “Joules per Amp”. Remember that a
Joule is one of the standard units for Work. When interpreted this way, the magnetic flux
in Webers is the amount of work a particular magnetic field can perform on a conductor
carrying one amp of current.

To gain even more familiarization with the unit of Weber, recall from our work in Notes
Set 9 that the units for the Volt are
 
Joules
Volts = (9)
Coul

When interpreted this way, the electric potential in Volts is the amount of work an electric
field can perform on a charge of one coulomb.

Comparing equations (8) and (9) shows that the Weber in magnetics is somewhat analogous
to the volt in electrostatics. The Weber is frequently called the unit for the “Magnetic Flux”.
Using this terminology, the unit of B, Webers per meter-squared, is frequently called the
unit for the “Magnetic Flux Density”. One Weber per meter-squared is also related to the
units Tesla and Gauss through the following:
Weber
1 = 1 Tesla = 104 Gauss (10)
m2
There is another field, sometimes called the Magnetic Field “Intensity”, which is denoted
by the symbol H. The H field is related to B by the equation

B = µH (11)

where µ is called the Magnetic Permeability of the medium. For free-space, µ = µ0 , which
has the value  
Henries
µ0 = 4 π(10) 7
(12)
m
The units of H are Amps/meter, and this may be a more familiar unit. Many of our problems
will be worked using the H-field.
ECE 303 - Sum10 Notes Set 13 : Time-Invariant Magnetic Fields 5

AMPERE’S CIRCUITAL LAW

Ampere’s Circuital law is a fundamental result which relates current and the magnetic field.
This law allows us to compute the magnetic field produced by a current-carrying wire. To
understand Ampere’s Law, we will use a very simple geometry. Let a uniformly distributed
current I flow inside a cylindrical conductor with cross-section surface S as shown below :

Let a closed contour C surround the conductor as shown in the figure. The magnetic field
H is related to current I by Ampere’s Circuital Law:

The Line Integral of the Magnetic Field Intensity (H) around a


Closed Path is equal to the current enclosed by the path.

A mathematical statement of qmpere’s Circuital Law is


I
H  dl = Ienc (13)
C

The left side of (13) is a line integral around the closed contour C and the right side is the
current enclosed by C. If the contour C encloses the entire wire then Ienc = I. The value of
H is found by applying (13) for the specific geometry of the current.
ECE 303 - Sum10 Notes Set 13 : Time-Invariant Magnetic Fields 6

MAGNETIC FIELD DIRECTION : QUALITATIVE

One useful result of the magnetic field definition in equation (13) occurs for a very long
straight conductor. Suppose a time-invariant (DC) current flows in a very long conductor
oriented with the z-axis as shown below. This current produces a Time-Invariant magnetic
field both inside and outside the conductor. Unlike the electric field lines, which are in a
direction radial from charges, the magnetic field lines form closed paths. The direction of
the magnetic field outside the conductor is given by the Right-Hand Rule, as shown in the
figure below:

I
CLOSED
lines of H

Ampere’s Circuital Law also holds for the magnetic field inside the conductor, but right
now we are primarily interested in the field outside the conductor. In the figure above,
notice that the H field lines extend out into the space surrounding the conductor. This
means the magnetic field shown can affect current flowing in conductors which are nearby.
To determine such an effect, we need quantitative expressions for H, which are examined
on the following pages.
ECE 303 - Sum10 Notes Set 13 : Time-Invariant Magnetic Fields 7

LONG CYLINDRICAL CONDUCTOR

Ampere’s Circuital Law in equation (13) can be used to obtain the magnetic field due to
current in a very long conductor. Let the uniformly distributed DC current I be flowing in
a cylindrical conductor (radius a) aligned in the z-direction as shown below:
z
a I

φ
x

Let a point P be defined in cylindrical coordinates as being at radial distance r from the
center of the conductor and have angle φ with respect to the x-axis. Then we can use
Ampere’s Law to show the magnetic field at P due to this current is given by
8
>
> Ir
0<ra
>
< 2π a2 uφ ;
H= (14)
>
>
>
: I uφ ; r>a
2π r
The problem below gives a derivation of the magnetic field outside the conductor, which is
the bottom relation in (14). The top relation in (14) can be derived via a similar process.

Problem: 13-3
ECE 303 - Sum10 Notes Set 13 : Time-Invariant Magnetic Fields 8

MAGNETIC FIELD DIRECTION: QUANTITATIVE

While in general the direction of the magnetic field follows the right-hand rule, in practice
we often have to compute this direction. Let’s examine a specific case to understand more
about how to compute the direction of the magnetic field. Assume the following:

 Let a current I = 0:2 amps be in the +y-direction (directed into the xz-plane) at the
point (x; z) = (2; 1).

 Find the magnetic field H at the point P = (x; z) = (5; 2).

The intercept of the current with the xz-plane is shown in the figure below:

z
q
z=2
I=0.2
p . P = (x,z)=(5, 2)

z=1 X H

x=2 x=5 x

There are three basic steps to finding the direction of the H vector in the figure above:

 Find the vector p linking the location of the current (at the point (x; z) = (2; 1)) and
the point P where we want to find the H field.

 Find the vector q which is orthogonal to the vector p.

 Use the right-hand rule to determine whether H point “up” on q or “down” q.

In the figure we have used the right hand rule to qualitatively show the direction of H. On
the following page we will quantitatively determine the H field.
ECE 303 - Sum10 Notes Set 13 : Time-Invariant Magnetic Fields 9

MAGNETIC FIELD DIRECTION: QUANTITATIVE (cont.)

The vector p is the vector which begins at the location of the current and ends at P. From
the figure on the preceeding page this vector is

p = (5 2) ux + (2 1) uz = 3 ux + uz (15)

To find q, which is a vector orthogonal to p, observe that if q is orthogonal to p, then the


dot product of p and q is zero:
pq=0 (16)

The general form for the vector q is given by

q = qx ux + qz uz (17)

where qx and qz are unknown scaling factors which must be determined. Substituting (15)
and (17) into (16) then gives the following:
   
pq = 3 ux + uz  qx ux + qz uz = 0 (18)

Computing the dot product in (18) and solving for the relation between qx and qz then gives

3 qx + qz = 0 ) qz = 3 qx (19)

Let the unit vector in the direction of q be denoted by ua . Therefore we know that

q2x + q2z = 1 (20)

Substitute (19) for qz into (20), giving

q2x + ( 3 qx )2 = 1 ) qx =  0 316: (21)


ECE 303 - Sum10 Notes Set 13 : Time-Invariant Magnetic Fields 10

MAGNETIC FIELD DIRECTION: QUANTITATIVE (cont.)

Note that (21) gives two possibilities for qx : one pointing to the right ( + 0:316) and one
pointing to the left ( 0:316). Examining the figure for this example, the right-hand rule
shows that H must point “down” and “to the right”. Therefore, we choose the positive (“to
the right”) value for qx in (21):
qx = 0:316 (22)

Substituting (22) for qx into (21) then shows that qz is given by

qz = 3 (0:316) = 0:948 (23)

Since we know that ua is in the same direction as uH , then we also have

uH = qx ux + qz uz (24)

Finally, substitute (22) and (23) into (24) to obtain desired the unit vector uH :

uH = 0:319 ux 0:948 uz (25)

Compare the quantitative result of (25) with the qualitative direction of H shown in the
earlier figure. You will see that the result in (25) does indeed “point in the correct direction”.
ECE 303 - Sum10 Notes Set 13 : Time-Invariant Magnetic Fields 11

MAGNETIC FIELD MAGNITUDE

Now let’s find the magnitude for the H-field in the previous example. First, observe that the
point P = (x; y) = (5; 2) is outside of the current carrying conductor. Let d be the distance
between the location of the current at (x; y) = (2; 1) and the point P = (5; 2). The magnitude
of H, denoted by H, is then found using the bottom relation in (14) :
I
H = (26)
2π d
p
From the problem statement and figure we know that I = 0:2 and d = 10 = 3:16. Substi-
tuting these values into (26) and simplifying then gives the result

H = 0:01 (27)

Finally, using (25) and (26) then gives the result for the vector H-field at point P :
 
H = H uH = 0:01 0:319 ux 0:948 uz (28)

Problems: 13-4, 13-5, 13-6


ECE 303 - Sum10 Notes Set 13 : Time-Invariant Magnetic Fields 12

MAGNETIC FIELD: MULTIPLE CURRENTS

There are many applications where multiple conductors in close proximity carry multiple
currents at the same time. Some common applications where this occurs are:

 Parallel data transfer in digital systems

 Simultaneous telephone voice conversations in the cable bundle outside the residence

 Simultaneous DSL/voice transmission in the cable bundle outside the residence

We can understand the effect of the magnetic field on adjacent conductors by using a simple
example. Let multiple currents IA ; IB ; and IC be positioned parallel to the y-axis, and be
flowing in the y-direction, as shown in the figure below.

HA
z=1

HB IC
HC x = −1

x=0
y IB
x=1
IA
x

Suppose we want to find the total magnetic field at the point P = (0; 0; 1) above the xy-
plane. In the figure above you can use the right-hand rule to qualitatively determine the
direction of the magnetic fields at this point due to each current. If we recognize that HA ,
HB , and HC are all in the xz-plane, then we can simplify the problem to a two-dimensional
figure as shown on the following page.
ECE 303 - Sum10 Notes Set 13 : Time-Invariant Magnetic Fields 13

MAGNETIC FIELD: MULTIPLE CURRENTS (cont.)

Since HA , HB , and HC are all in the xz-plane, then the fields and currents from the previous
page can be simplified to the two-dimensional figure as shown below :

HA
z=1 HB
HC
IC IB IA
X X X
x = −1 x=0 x=1

The currents IA , IB , and IC are all into the page, and the directions of the individual magnetic
fields are found using the right-hand rule. The total magnetic field at z = 1 is the sum of the
three individual magnetic fields.

Therefore, finding the magnetic field due to the multiple currents involves finding each of
HA , HB , and HC and then adding them together to find the total H. Finding the magnitudes
of the individual Hi is straightforward. However, finding the direction of these individual
Hi is a little harder. You might want to review Problem 13-4 and 13-5 for how to find these
directions.

On the next page we will complete the numerical example for computing the total magnetic
field at the point P = (0; 0; 1).
ECE 303 - Sum10 Notes Set 13 : Time-Invariant Magnetic Fields 14

LINEAR FIELD EXAMPLE

EXAMPLE: Find the H-field at the point (x; y; z) = (0; 0; 1) if IA = IB = IC = 1:0 in the
figure on the previous page.

APPROACH: The total H-field at point P is H = HA + HB + HC where the subscripts refer


to the current elements. All currents are “into the page”, that is, in the +y-direction.

HB : Using the right-hand rule shows the direction of HB to be ux . The magnitude is


IB 1
HB = = = 0:159
2πr 2π(1)

HA : The right-hand rule shows the direction of HA to be u A = .707ux + .707uz .


(See also Prob. 13-8 for more information.) The magnitude is

HA =
IA
2πr
= p
1
= 0:113
2π 2

HC : The right-hand rule shows the direction of HC to be u C = .707ux : 707uz .


(See also Prob. 13-8 for more information.) The magnitude is

HC =
IA
2π r
= p
1
= 0:113
2π 2

The total magnetic field is thus

H = 0:159ux + 0:113(:707ux + :707uz ) + 0:113(:707ux : 707uz )

Adding the vector terms in the above equation then gives


H = 0.319 ux

Problems: 13-7 through 13-10


ECE 303 - Sum10 Notes Set 14 : Curl and the Magnetic Field 1

INTRODUCTION

We have previously studied the integral form of Ampere’s Circuital Law. This integral form
is very descriptive and is useful for magnetic field computations in simple geometries. For
example, the integral form lets us find the magnetic field due to a cylindrical conductor.

However, there is an equivalent differential form, or point form, which is very useful in
more general environments and in our eventual use of Maxwell’s Equations. In theory, this
differential form allows us to compute the magnetic field at any point in space. Using the
differential form requires us to understand a new quantity, called Current Density, which
has a value at every point in space.

Stokes’ Theorem is the mathematical tool which we use to obtain the differential form of
Ampere’s Law from its integral form. Therefore, Stokes’ Theorem is similar in function to
the Divergence Theorem, which can be used to obtain the differential form of Gauss’ Law
from its integral form. This is similar to the manner in which we used Gauss’ Law to find
the time-invariant electric field at any point in space.

Another reason for studying curl relations concerns deriving computational results from
Maxwell’s equations. Expressing Maxwell’s equations in their differential form allows us
to theoretically compute the time-varying electric and magnetic fields at any point in space.
We will see that the spatial differential operators in Maxwell’s equations are defined by the
curl operator. Therefore, familiarization with the curl will be very valuable in later work
when we study Maxwell’s equations.
ECE 303 - Sum10 Notes Set 14 : Curl and the Magnetic Field 2

CURRENT DENSITY

To derive the differential (point) form for Ampere’s Law, we need to understand the Current
Density, which is analogous to charge density. For example, we can think of a region of
charge (Coulombs) as being distributed over a volume (with units m3 ). At any point within
the region, there is a charge density ρ having units of Coulombs/m3. We can similarly
think of the total current I (in Amps) passing through a surface (having units m2 ). At any
point on the surface, there is a current density J having units of Amps/m2. This allows
consideration of currents which vary as a function of space.

The current density J is sometimes called the Conduction Current density, because it is
due to the actual physical movement of charge carriers (like holes or ions or electrons).
These charges are usually resident in conductors. Just as there is a “macroscopic” relation
between current and voltage (Ohm’s Law), there is an equivalent “microscopic” relation
between electric field E and the vector current density J. Recall that Ohm’s Law relates the
scalar voltage V and scalar current I:
1
VI = (1 )
R
Now consider the units on both sides of equation (1),
     

Amps =
1
ohms
 ( volts ) =
volts
ohms
(2)

Divide the units of each side of (2) by m2 :


       
Amps
m2
=
volts
ohms - m2
=
1
ohms - m
 volts
m
(3)

Observing the units in (3), we see that Amps/m2 is the unit of current density. We also
see that 1/ohm-m is the unit of conductivity, and volts/m is the unit for electric field. The
symbols for these quantities are J for current density, σ for conductivity, and E for the
electric field. Therefore, equations (1) and (3) show that the relation between the vector
electric field E and the vector current density J is given by
J = σE (4 )

The result from (4) is useful for quantifying properties at points within the conductor.
ECE 303 - Sum10 Notes Set 14 : Curl and the Magnetic Field 3

CURRENT DENSITY AND CURRENT

We frequently use current when we are more interested in the macroscale properties of a
device. An example is circuit analysis, in which we study a circuit to determine its voltage
and current properties. In circuits the conductors carrying the current are not considered
relevant to the problem. However, there are many applications where the dimensions and
the geometry of the conductors are very important. One application is when the current
density is spatially varying, as in “high-frequency” applications using metallic conductors.
The general case thus requires that we consider current density, rather than current.

Obtaining the current from a current density is like obtaining the mass of an object if you
know its mass density: You would “sum up” the specific mass densities over the spatial
volume of the object. In a similar manner the total current in a conductor can be found by
“summing up” all the individual current densities passing through the cross section of the
conductor. To see how this is done, consider the cylindrical conductor in the z-direction
shown in the figure below. The current density J = J(x; y) flows only inside the conductor
and the surface S is a cross section of the conductor:

S
J = current density
J inside conductor

The total current in the conductor is the sum of all the current densities flowing through
the surface S. “Summing up” this current density is done by a Surface Integral, which is
examined on the following pages.
ECE 303 - Sum10 Notes Set 14 : Curl and the Magnetic Field 4

SURFACE INTEGRAL

Let’s examine the surface integral using a simple Cartesian example. Let Z a specific vector
field be given by G = ux + 1:5 uy + 2 uz . Evaluate the surface integral G  dS over the
h S i
rectangle in the xy-plane having corners at the points (x; y) = (0; 0); (5; 0); (5; 3); (0; 3) .

APPROACH:
Draw the geometry of the field G and the surface S:
z

G
y dS
(5,3)
(0,3)
Surface S

x
(5,0)

For simplicity in the above figure, G and dS have only been shown at one point. However,
keep in mind that that both G and dS exist over the entire surface S. This means that G  dS
has a value over the entire surface.

Direction of a Surface
Note that the surface element dS is a vector and therefore has a “direction”. One way of
denoting this direction is
dS = dS u N (5 )

where u N is the unit vector Normal to the surface element and dS is the scalar surface
differential element. Since the rectangle in the above figure is in the xy-plane, the surface
element in (5) becomes
dS = dS uz = dx dy uz (6 )
ECE 303 - Sum10 Notes Set 14 : Curl and the Magnetic Field 5

SURFACE INTEGRAL (cont.)

Note that we could also have chosen u N in the uz direction. This will only change the
sign of the surface integral. Now compute the dot product:
h i
G  dS = ux + 1:5 uy + 2 uz  dx dy uz = 2 dx dy (7 )

Substitute (7) into the expression for the surface integral and integrate over the dimensions
for x and y shown in the previous figure:
Z Z x=5 Z y=3 5 3
G  dS =

2 dx dy = 2 x y = 2(5)(3) = 30
S x=0 y=0 x=0 y=0

In summary, evaluating a surface integral has two steps:

 Compute the scalar dot product.

 Evaluate the integral over the limits defining the surface.

NOTE: If G and dS are constant with respect to space over the surface S then
Z Z Z
G  dS = G  uN dS = G  uN dS (8 )
S S S

Evaluating the last integral on the right hand side of (24) then gives
Z
G  dS = ( G  uN ) A = A j G j cos θ (9 )
S

where A on the right hand side of (9) is the area of the surface S and θ is the angle between
the surface normal and G.. These results sometimes simplify computing surface integrals.
ECE 303 - Sum10 Notes Set 14 : Curl and the Magnetic Field 6

CIRCULAR CROSS SECTION

Consider again the current density J = J(x; y) flowing through the cross section surface
S. Using the surface integral, we therefore see that Itot , the total current flowing in the
conductor, is given by Z
Itot = J  dS (10)
S
If the integration in (10) is performed in cylindrical coordinates, then dS must be expressed
in cylindrical coordinates. A circular cross section of a conductor with radius a is shown in
the figure below:

z−axis
y−axis
u
N circular
r dφ cross−section
r=a surface S

dr
x−axis

φ angle

The unit vector uN in the figure is orthogonal to the surface S. Keep in mind that a unit
vector uN exists for every point on the surface, but for simplicity only one is shown in the
figure. The computation of dS in the figure thus becomes

dS = dS uN = r dφ dr uz (11)

As shown in the figure, dr is the infinitesimal length along the r-axis and r dφ is the in-
finitesimal length along the φ angle.

Problems: 14-1, 14-2


ECE 303 - Sum10 Notes Set 14 : Curl and the Magnetic Field 7

AMPERE’S LAW AND CURRENT DENSITY

Ampere’s Circuital law may now be restated to incorporate current density. We will use
the same integral form which we used previously. Let the current density J flow inside a
cylindrical conductor with cross-section surface S as shown below :

The current density J inside the conductor creates the total current I as shown in the figure.
Let a closed contour C surround the conductor as shown in the figure. The magnetic field
H is related to current I by Ampere’s Circuital Law:

The Line Integral of the Magnetic Field Intensity (H) around a


Closed Path is equal to Ienc , the current enclosed by the path.

But now we know that the current enclosed by the path is related to the surface integral.
Therefore, the mathematical statement of Ampere’s Circuital Law becomes
I Z
H  dl = Ienc = J  dS (12)
C S

The left side of (12) is a line integral around the closed contour C and the right side is a
surface integral over the surface enclosed by C. The new information in (12) is the relation
of magnetic field H to current density J.
ECE 303 - Sum10 Notes Set 14 : Curl and the Magnetic Field 8

STOKES’ THEOREM

Having developed the concepts of current density and surface integral, we can now apply a
vector identity to derive the differential (point) form for Ampere’s Law. This vector identity
is called Stokes’ Theorem and it is very important in many areas of applied mathematics.

Let G be a general vector field which intersects a surface S. Stokes’ Theorem relates a
property of G occuring inside the surface with a property of G occuring on the boundary
of the surface. This surface S is simply a mathematical “shape” inside a boundary curve
C which encloses the surface. The mathematical shape might be flat or have curvature.
Stokes’ Theorem states that the following surface integral and line integral are equal:
Z I
( ∇ X G )  dS = G  dl (13)
S C

In (13), the left side is a surface integral and the right side is a line integral. The term
“∇ X G” on the left side integral in (13) is the Curl of G. The circle and the subscript “C”
on the right side integral in (13) means this line integral is evaluated around the Contour
(line) which encloses the surface S.

Equation (13) is a general property for any general vector field. With G replaced by the
specific magnetic field H, equation (13) becomes
Z I
(∇ X H)  dS = H  dl (14)
S C

Now recall the integral form of Ampere’s circuital law from (12), repeated here for refer-
ence: I Z
H  dl = J  dS (15)
C S
where J is the current density intersecting the surface. The derivation of the point form for
Ampere’s Law is completed on the next page.
ECE 303 - Sum10 Notes Set 14 : Curl and the Magnetic Field 9

AMPERE‘S LAW: POINT FORM

To complete the derivation of the point form for Ampere’s Law, substitute equation (15)
for the H line integral into (14). This process gives the following:
Z Z
(∇ X H)  dS = J  dS (16)
S S

Since equation (16) must hold for any arbitrary surface S, limiting arguments would show
that an equivalent requirement is that the integrands of (17) must be equal. That is, we have
the condition

∇XH = J (17)

Equation (17) is the differential point form for Ampere’s Circuital Law. Since this relation
holds at every point in space it is often also called the point form of Ampere’s Law.

Observe that equation (17) is a vector partial differential equation. If the right-hand side
(the current density J) is given, then spatial integrations may be performed to determine the
magnetic field H. Equivalently, if we know H, then performing the curl (spatial derivatives)
will provide the J which produced the magnetic field.

Note that (16) and (17) have been derived independent of coordinate system, and therefore
hold for Cartesian, Cylindrical, and Spherical coordinates. On the next few pages we will
study some specific examples for Cartesian and Cylindrical coordinates.
ECE 303 - Sum10 Notes Set 14 : Curl and the Magnetic Field 10

CURL COMPUTATIONS

Since the point-form of Ampere’s Circuital Law uses the curl, let’s develop some familiarity
with this operation. The formulas for curl in Cartesian and Cylindrical coordinates are
given below:

CARTESIAN:
Let G(x; y; z) = Gx ux + Gy uy + Gz uz . In Cartesian coordinates the curl of G is:

 

∂Gz ∂Gy
ux

ux uy uz
∂y ∂z
+
 
∂ ∂ ∂ ∂Gx ∂Gz
∇XG = = uy (18)
∂x ∂y ∂z
∂z ∂x
+
 
Gx Gy Gz ∂Gy ∂Gx
uz
∂x ∂y

CYLINDRICAL:
Let F(r; φ; z) = Fr ur + Fφ uθ + Fz uz . In cylindrical coordinates the curl of F is:

 

1 ∂Fz ∂Fφ
ur

ur r uφ uz
r ∂φ ∂z
+
 
1 ∂ ∂ ∂ ∂Fr ∂Fz
∇XF= = uφ (19)
r
∂r ∂φ ∂z
∂z ∂r
+
 
Fr rFφ Fz 1 ∂ ∂Fr
uz (rFφ )
r ∂r ∂φ
We will not use the spherical curl in this course, but spherical curl is very important in the
study of power radiated from antennas.
ECE 303 - Sum10 Notes Set 14 : Curl and the Magnetic Field 11

CURL EXAMPLE: CARTESIAN

Let’s work an example in Cartesian coordinates for computing the curl. Let the vector field
G be given by
G(x; y; z) = 3z ux + 2x uy + y2 uz (20)

Equation (20) shows that the x, y, and z components of G are given by

Gx = 3z ; Gy = 2x ; Gz = y2 (21)

Substitute the relations in (21) into equation (18) for the curl in Cartesian, giving

ux uy uz



∂ ∂ ∂
∇XG = (22)
∂x ∂y ∂z


3z 2x 2
y

Evaluating the terms in the “determinant-like” structure of (22) then gives

     
∂ 2 ∂ ∂ ∂ 2 ∂ ∂
∇XG = ux y 2x + uy 3z y + uz 2x 3z (23)
∂y ∂z ∂z ∂x ∂x ∂y

Finally, evaluating the partial derivatives in (23) gives the desired form

∇XG = 2y ux + 3 uy + 2 uz (24)

Problems: 14-3, 14-4


ECE 303 - Sum10 Notes Set 14 : Curl and the Magnetic Field 12

CURL EXAMPLE: CYLINDRICAL

Now let’s work an example in cylindrical coordinates for computing the curl. Let the vector
field F be given in cylindrical coordinates. by
2
F(r; φ; z) = 3rz ur + uφ + r φ uz (25)
r
Equation (13) shows that the r, φ, and z components of G are given by
2
Fr = 3rz ; Fφ = ; Fz = r φ (26)
r
Substitute the relations in (26) into equation (19) for the curl in cylindrical, giving

ur r uφ uz



1 ∂ ∂ ∂
∇XF=

∂r ∂φ ∂z (27)
r


2
3rz r( ) rφ
r
1
Note the term “ ” outside of the determinant in (27). This means do all of the partial
r
1
derivatives inside the determinant first, and then multiply the results by . Therefore,
r
evaluating the terms in the “determinant-like” structure of (27) then gives
   
1 ∂ ∂ 2 ∂ ∂
∇XF = ur rφ + uφ 3rz rφ
r ∂φ ∂z r ∂z ∂r

 
1 ∂ 2 ∂
+ uz r( ) 3rz (28)
r ∂r r ∂φ
ECE 303 - Sum10 Notes Set 14 : Curl and the Magnetic Field 13

CURL EXAMPLE: CYLINDRICAL (cont.)

Finally, evaluating the partial derivatives in (28) gives the desired form

∇XF = ur + (3r φ) uφ + (0) uz (29)

Since the uz term in (29) is zero, it is easier to just omit this term in the expression. There-
fore, (29) may be written in the simpler form as

∇XF = ur + (3r φ) uφ (30)

Problems: 14-5, 14-6


ECE 303 - Sum10 Notes Set 14 : Curl and the Magnetic Field 14

CURL OF H AND CURRENT DENSITY

Sometimes a simple way of computing the curl of H is to instead compute the current
density J. Recall that time-invariant current densities and magnetic fields satisfy the point
form for Ampere’s Circuital Law, repeated here for reference:
∇XH = J (31)

Equation (31) shows that knowing of J immediately gives the curl of H. For example, let a
cylindrical conductor have current in the z-direction. A cross section of the cylinder in the
z = 0 plane is shown below, in which the shaded area signifies presence of current density :
y−axis inside
outside conductor : conducting cylinder :
J= 0 J = 0

x−axis

Using (31) we now have the equivalent figure in terms of curl of H :


y−axis
inside
outside conductor : conducting cylinder :
XH= 0 XH = J

x−axis

Therefore, the curl of H is found by computing J instead. The problems listed below give
you some practice using this method.

Problems: 14-7 through 14-11


ECE 303 - Sum10 Notes Set 15 : Time-Varying Magnetic Fields 1

INTRODUCTION

As we have progressed in our study of electromagnetics, we have given increasing degrees


of freedom to the electric charge. Recall that first we required the charges to be stationary,
in which case the only field effect was an electric field quantified by Coulomb’s Law. Next,
we allowed the charges to move with constant velocity, which produced a time-invariant
magnetic field which was quantified by Ampere’s Circuital Law. The magnetic field pro-
duced by these time-invariant currents can vary as a function of space, but can not vary as
a function of time.

In this set of Notes we remove the constraint of time-invariant current. A time-varying


current does indeed produce a time-varying magnetic field in space. However, we will
show later that a time-varying current can also produce a magnetic field and electric field
which can propagate like a wave through space.

In this set of Notes we concentrate on the effects of the time-varying magnetic field with-
out investigating the wave nature of the propagating energy. We will study Faraday’s Law,
which describes how we can transform the electromagnetic field energy back into a voltage
and/or current. In a later Notes Set we will study Maxwell’s Equations, which do incorpo-
rate the propagation characteristics of the electromagnetic wave.
ECE 303 - Sum10 Notes Set 15 : Time-Varying Magnetic Fields 2

MAGNETIC FLUX

A magnetic field changing in time can induce a voltage and current in a circuit. This is
quantified by Faraday’s Law, which is the main topic of this set of notes. The mathematics
of Faraday’s Law can be simplified by using the magnetic flux density, B(t ). This magnetic
flux density is related to the magnetic field intensity, H(t ), by

B(t ) = µ H(t ) (1)

where µ is the magnetic permeability in Henries/meter of the medium. For free space the
value is
µ = µ0 = 4π(10) 7 (2 )

In equation (1) we have now allowed for the possibility of a time-varying magnetic field.
Units of these quantities are:

H(t ) ) B(t ) ) µ)
Amps Webers Henries
; ; (3 )
m m2 m
The Magnetic Flux, ΦM (t ), over a surface S is the surface integral given by
Z
ΦM (t ) = B(t )  dS (4 )
S

The surface integral in (4) is computed using the techniques of surface integrals, which we
covered in a previous set of notes. Evaluation for ΦM (t ) in equation (4) is perhaps easier if
we divide the problem into two steps:

 Compute a scalar dot product B(t )  dS.

 Evaluate an integral over the limits defining the surface.

Note that if B = B(t ) is time-varying, then ΦM = ΦM (t ) is time-varying. An example


of computing a constant ΦM is shown on the next page, after which we will compute a
time-varying ΦM (t ).
ECE 303 - Sum10 Notes Set 15 : Time-Varying Magnetic Fields 3

MAGNETIC FLUX EXAMPLE

Let’s work an example in which we compute the time-invariant magnetic flux ΦM over a
surface. Let a time-invariant current I = 0:3 be flowing in the z-direction as shown in the
figure below. A metallic conductor in free space forms a closed path over the rectangle
with corners at (x; y; z) points (3, 0, 2), (6, 0, 2), (6, 0, 7), and (3, 0, 7). Compute ΦM
passing through the rectangle if the rectangle surface normal uN = uy .

z B = Bu y
into page

z=7

z=2

x=3 x=6 x

I = 0.3
In the above figure the y-axis is into the page. Let’s first evaluate the dot product required
in equation (4) for the magmetic flux.

Dot Product :
Since the surface S is a rectangle in the xz-plane then we have the following for dS:
dS = dS uy = dx dz uy (5 )

We see that the loop is outside the current-carrying wire and we need only the formula for
B outside a current carrying wire. From our previous work, we know that the cylindrical-
coordinate expression for the H-field produced outside a long, straight wire is
I
H= uφ (6 )
2π r
Comparing (6) with the figure above we see that in the plane of the loop we have
r = x ; uφ = uy (7 )

The example is continued on the next page.


ECE 303 - Sum10 Notes Set 15 : Time-Varying Magnetic Fields 4

MAGNETIC FLUX EXAMPLE (cont.)

Substituting (7) into (6) and using the free-space result B = µ0 H then gives
µ0 I
B= uy (8 )
2π x
Now substitute (5) and (8) into the formula for the dot product required in (4) :

B  dS =  dx dz uy
µ0 I
uy (9 )
2π x
Performing the dot product in (14) then gives the desired result for the dot product:

B  dS =
µ0 I
dx dz (10)
2π x

Evaluation of Integral :
From the figure, we see that the surface is the rectangle in the xz-plane spanning 3  x  6
and 2  z  7. Substitute these limits into the limits of integration required by (4) and also
substitute B  dS given by (10) into (4). This gives
Z 7 Z 6 Z 7 Z 6
µ0 I µ0 I dx
ΦM = dx dz = dz (11)
z=2 x=3 2πx 2π z=2 x=3 x

Use I = 0:3 from the problem statement and evaluate the integrals in (11). This gives
6
0:75 µ0
ΦM = ln x (12a)
π x=3

Use µ0 = 4π(10) 7 and evaluate the expression in (12a), providing the desired result

ΦM = 2:08(10) 7 (12b)

Problems: 15-1 through 15-4


ECE 303 - Sum10 Notes Set 15 : Time-Varying Magnetic Fields 5

FARADAY’S LAW

The previous pages have constrained the magnetic field in space to be time-invariant. How-
ever, a new and extremely important phenomenon occurs when the magnetic field becomes
time-varying. Faraday’s Law states that a time-varying magnetic field causes a voltage (and
a current) to be induced in a closed path. A mathematical statement of Faraday’s Law is
d
vind (t ) = ΦM (t ) (13)
dt
where ΦM (t ) is the magnetic flux passing through the loop. The geometry of of Faraday’s
Law is shown below:

z uN
B (t) y

+
v (t)
− ind

In the example in the figure, the time-varying magnetic field intersects a circular loop in
the xy-plane. Two things to keep in mind are:

 ΦM (t ) must be time-varying; otherwise there is no induced voltage.

 The closed path (loop, rectangle, etc.) must have non-zero area; otherwise there is
no induced voltage.

On the following page we will work an example which demonstrates the computation of the
time-varying magnetic flux. After that we will give an example which shows how Faraday’s
law computes the voltage induced by the changing magnetic flux.
ECE 303 - Sum10 Notes Set 15 : Time-Varying Magnetic Fields 6

TIME-VARYING FLUX EXAMPLE

Here is an example in which the flux density B(t ) is time-varying, in which case. ΦM (t ) is
also time-varying. Let a magnetic flux density given by

B(t ) = 10 9
cos ω t uz (14)

exist in a region of free space. A metallic conductor in this region forms a closed path over
the rectangle with corners at the four (x; y; z) points (100, 4, 0), (105, 4, 0), (105, 14, 0),
and (100, 14, 0). Compute ΦM (t ) passing through the rectangle if the rectangle surface
normal uN = uz . This geometry is shown below, in which the z-axis is out of the page:

y B = B(t) u z y B = B(t) u z
into page out of page

y=14 y=14

y=4 y=4
... ... x
x=100 x=105 x
x=100 x=105

Note from (14) that the direction of the magnetic field changes depending on the polarity
of the cosine. The figure on the left above is for the case of cos ωt negative (field into the
page) and the figure on the right above is for the case of cos ωt positive (field out of the
page). However, notice the surface does not change orientation and therefore the surface
element dS is given by
dS = dS uz = dx dy uz (15)

As before, we will first compute the dot product and then evaluate the surface integral.
ECE 303 - Sum10 Notes Set 15 : Time-Varying Magnetic Fields 7

TIME-VARYING FLUX EXAMPLE (cont.)

Dot Product :
Substitute (14) and (15) into the expression B(t )  dS, producing
B(t )  dS = 10 9
cos ωt uz  dx dy uz (16)

Computing the dot product operation on the right of (16) then gives

B(t )  dS = 10 9
cos ωt dx dy (17)

Evaluation of Integral :
Substitute (17) into (4) and integrate over the dimensions of the rectangle in the xy-plane:
Z 14 Z 105
Φ(t ) = 10 9
cos ωt dx dy (18)
y=4 x=100

Since cos ω t is not a funvtion of space it can be factored out of the integral, producing
Z 14 Z 105
Φ(t ) = 10 9
cos ω t dy dx (19)
y=4 x=100

Evaluating the integrals in (19) then gives the desired answer

Φ(t ) = 10 9
cos ω t  10  5 = 5(10) 8 cos ωt (20)

Problem: 15-5
ECE 303 - Sum10 Notes Set 15 : Time-Varying Magnetic Fields 8

VOLTAGE COMPUTATION

To compute the voltage induced in a loop, we use the mathematical statement of Faraday’s
Law given by equation (13). In the following example we will use the same geometry as
in the previous example. Let’s assume that the magnetic flux is given by equation (20) and
we want to find the voltage induced in the loop shown below:

y
B = B(t) u z
y=14
.+
.−
y=4

...
x=100 x=105 x

From equation (20) we found the time-varying flux to be

ΦM (t ) = 5(10) 8
cos ωt (20)

Therefore, substituting (20) into (13) shows the induced voltage is given by

d
vind (t ) = 5(10) 8
cos ωt = 5(10) 8ω sin ωt (21)
dt

For example, if ω = 106 rad/sec, the induced voltage would be vind (t ) = 0:5 sin 106 t.

Problems: 15-6 through 15-11


ECE 303 - Sum10 Notes Set 15 : Time-Varying Magnetic Fields 9

FARADAY’S LAW : DIFFERENTIAL FORM

The integral form of Faraday’s Law given by (13) is very descriptive and is useful for
computations in simple geometries. Equation (13) is repeated here for easy reference:
Z
B(t )  dS
d d
vind (t ) = ΦM (t ) = (22)
dt dt S
where we have also expressed the magnetic flux ΦM (t ) as a surface integral. In equation
(22), S is the surface enclosed by the closed path (the circuit) which induces vind (t ). We
may move the derivative on the right hand side of (22) inside the integral by transforming
it to a partial derivative:
Z Z
∂B(t )
B(t )  dS =  dS
d
(23)
dt S S ∂t
Substituting (23) into (22) thus gives
Z
∂B(t )
vind (t ) =
∂t
 dS (24)
S

Now recall that a voltage, such as vind (t ), is the line integral of an electric field E(t ) over
a path. Faraday’s law requires that the “line” form a closed path (or closed contour C).
Therefore, since the fields are “doing the work” we have
I
vind (t ) = E(t )  dl (25)
C

where the circle on the integral in (25) signifies the closed path which encloses the surface
S. Equating (25) and (24) then gives
I Z
∂B(t )
E(t )  dl =  dS (26)
C S ∂t
We complete the derivation of the differential form of Faraday’s Law on the following page.
ECE 303 - Sum10 Notes Set 15 : Time-Varying Magnetic Fields 10

DIFFERENTIAL FORM (cont.)

Stokes’ Theorem now states that the line integral of E(t ) on the left side of (26) is equiva-
lent to a surface integral of the curl of E(t ) :
I Z  
E(t )  dl = ∇ X E(t )  dS (27)
C S

Substituting equation (27) in equation (26) then gives


Z   Z
∂B(t )
∇ X E(t )  dS =  dS (28)
S S ∂t
Note that both sides of (28) are now surface integrals over the same surface S. Now let the
size of the surface S shrink to zero. Since (28) must hold for any arbitrary surface, then
limiting arguments would show that (28) is valid if the left side terms in the dot products
are equal; that is, if
∂B(t )
∇ X E(t ) = (29)
∂t
Finally, substitue B(t ) = µ H(t ) into (29), giving

∂H(t )
∇ X E(t ) = µ (30)
∂t
Equation (30) is the point (or differential) form for Faraday’s Law. We will use the form
(30) when we develop Maxwell’s Equations.
ECE 303 - Sum10 Notes Set 16 : Maxwell’s Equations 1

INTRODUCTION

In electromagnetics, there are four vector partial differential equations which are collec-
tively known as Maxwell’s Equations. Three of them are similar to equations we have
previously studied under other names. The derivation of the fourth equation requires a
little more investigation, which is one of the topics of this set of notes.

To derive this fourth equation we will need another basic property relating charges and
current density. This property is called the continuity equation. For brevity we will not
derive the continuity equation, although it is certainly within our mathematical skill level
at this time. We will, however, outline the derivation steps in class discussions.

In this set of Notes, we will largely discuss qualitative properties of Maxwell’s equations,
and will defer specific solution of the equations until later Notes Sets. However, there is a
great deal of insight we can derive from these qualitative studies. Additionally, recalling
the relation between excitation and response from ordinary differential equations will help
us understand the implication of Maxwell’s equations.
ECE 303 - Sum10 Notes Set 16 : Maxwell’s Equations 2

DISCUSSION OF PREVIOUS RESULTS

We have previously examined several different phenomena concerning electric and mag-
netic fields. For easy reference, we will list below the differential form for some of the
important relations we have found. We will also include a brief discussion of the physical
meaning of each law.

GAUSS’ LAW:
From Notes Set 10, we learned that the differential form, or point form, of Gauss’
Law is given by
ρ
∇E = (1 )
ε
where ρ is the charge density at the point and ε is the electric permittivity. In Notes
Set 10 we also examined the physical interpretation of the divergence in (1) above,
and learned that non-zero divergence of E implies the existence of electric charges.
Depending on the sign of ρ on the right-hand side of (1), one of three possibilities
emerges:
ρ>0:
If, at a specific point, we have ρ > 0, then the right hand side of (1) means there is
a positive charge at that point. The left hand side side of (1) then means that E-field
lines are directed outward through a small closed surface surrounding the charge.
Therefore, we can think of positive charges as Sources of E-field lines.
ρ<0:
If ρ < 0, the right hand side of (1) means there is a negative charge at that point. The
left hand side side of (1) then means that E-field lines are directed inward through
a small closed surface surrounding the charge. Therefore, we can think of negative
charges as Sinks, or Terminations, of E-field lines.
ρ=0:
If ρ = 0, the right hand side of (1) means there is no charge at that point. The left
hand side side of (1) then means that all E-field lines which enter a small surface
surrounding the point will exit that surface (perhaps on the other side).
ECE 303 - Sum10 Notes Set 16 : Maxwell’s Equations 3

DISCUSSION OF PREVIOUS RESULTS (cont.)

DIVERGENCE OF MAGNETIC FIELD:


Recall the results we found for the magnetic field lines due to a constant current in a
conductor. Notes Set 12 showed that external to the conductor the lines of magnetic
field H formed closed circles around the conductor. There was no beginning and
no termination to the magnetic field lines. Therefore, for any small closed surface (a
small volume element) in the space surrounding the conductor, we have the result that
any H-field line that enters the small volume also exits the small volume. But this is
exactly the qualitative description of the mathematical condition that the divergence
of the lines must be zero. Consequently, we arrive at the second divergence equation:

∇ H = 0 (2 )

Although we have discussed the situation outside of a conductor (because it is rela-


tively easy to envision), equation (2) also holds inside the conductor. Equation (2)
states that magnetic field lines have no beginning and no end. Another implication is
that there is no magnetic counterpart to the electric charge. (Recall that electric field
lines begin on positive electric charges and terminate on negative electric charges).
Isolated magnetic “charges” in this sense do not exist.

FARADAY’S LAW:
We found in Notes Set 14 that a time-varying magnetic field could induce a voltage
and current in a closed conducting path, i. e., a circuit with a load. An easy to
understand form of Faraday’s Law for this induced voltage was given in Notes Set
14, repeated here:
d
vind (t ) = ΦM (t ) (3)
dt
where ΦM (t ) is the magnetic flux, which is related to the magnetic flux density (and
magnetic field H ) through a surface integral. Recall in Notes Set 9 that a voltage can
be expressed as a line integral involving the electric field E. Therefore, as shown in
Notes Set 14, we see that the left side of (3) is a function of the electric field E and
the right side of (3) is a function of magnetic field H.
ECE 303 - Sum10 Notes Set 16 : Maxwell’s Equations 4

DISCUSSION OF PREVIOUS RESULTS (cont.)

In Notes Set 14 a vector identity called Stokes’ Theorem was then then used to pro-
duce the differential for Faraday’s law:

∂H
∇XE= µ (4 )
∂t
The differential form in (4) is not as physically descriptive as the form in (3). How-
ever, equation (4) is the form frequently given in Maxwell’s equations, especially
when we wish to actually compute the electric and magnetic fields at a point in space.

AMPERE’S CIRCUITAL LAW :


In Notes Set 12 we studied the time-invariant magnetic field produced by a time-
invariant current. In that Notes Set we outlined the derivation of the differential form
for Ampere’s Circuital Law, repeated below:

∇XH = J (5 )

In the form given in (5), J is the current density in a conductor. For this reason, J is
sometimes called the conduction current density. Additionally, a constraint on (5) is
that J be time invariant.
ECE 303 - Sum10 Notes Set 16 : Maxwell’s Equations 5

SUMMARY OF EQUATIONS

The equations discussed in the material on the previous pages are summarized below:
ρ
∇E= (6 )
ε
∇H=0 (7 )

∂H
∇XE= µ (8 )
∂t
∇XH=J (9 )

As given above, these field equations are very general and express PDEs having time and
spatial variables. Recall that to obtain useful resuls we previously had to examine spe-
cific applications and strict geometries. Examples of this were using Gauss’ Law in one-
dimension to find the electric field for the PN-junction and using Ampere’s Circuital Law
to find the magnetic field in the vicinity of a long, straight DC-carrying conductor.

In addition, the field equations (6) - (9) above reflect the experimental nature of early elec-
tromagnetic science. It was often the case that experimental effects were observed and the
mathematical laws governing the physical principles were discovered later.

However, there is one inconsistency in the above equations, specifically equation (9). Re-
moving the inconsistency using mathematical techniques led to some very important pre-
dictions concerning physical properties of electromagnetic phenomena. This is a case
where the mathematical work preceded the experimental results. On the following pages
we will examine this inconsistency and study one approach for resolving it.
ECE 303 - Sum10 Notes Set 16 : Maxwell’s Equations 6

THE CONTINUITY EQUATION

When we deal with the differential form of equation (9), we see that “current” is represented
by current density J. Current density J (in units of amps/m2) is related to current I (in units
of amps) by the following surface integral:
Z
I = J  dS (10)
S
Therefore, you can see the formulas which involve J are actually related to current, even
though we specifically call J the current density.

A very important equation called the Continuity Equation relates the current density J to
the charge density ρ by the following:
∂ρ
∇ J =
∂t
(11)

This equation has a very intuitive physical interpretation, which we will now obtain. The
operation on the left side of (11) is the divergence of the current density. Let’s examine
what (11) implies about the charge inside a very small closed surface, that is, inside a very
small volume. If there is non-zero divergence of current density (that is, if ∇  J 6= 0 ),
then current must be flowing across the surface of the very small volume.

Since current is flowing across the surface, this means the charge density must be changing
inside the volume of the closed surface. This in turn means we must have the time derivative
of ρ inside the surface is not equal to zero. Further, the minus sign on the right side of (11)
states that if current flows outward through the closed surface (positive divergence) then
the charge density within that closed surface must decrease (negative time derivative).

Although there are many interesting implications of (11), our current interest is how the
relation in (11) affects the field equations (6) - (9). We will next see how (11) leads to
an inconsistency in Ampere’s Circuital Law listed in (9). On the following pages we will
outline an approach for removing the inconsistency. This will result in the set of equations
called Maxwell’s Equations.

Problems: 16-1 through 16-6


ECE 303 - Sum10 Notes Set 16 : Maxwell’s Equations 7

A VECTOR IDENTITY

Equation (11) incorporates the divergence of current density; the right hand side of Am-
pere’s Circuital Law in (9) incorporates the current density. Therefore, let’s take the diver-
gence of both sides of (9) and investigate the resulting equation. Taking the divergence of
both sides of (9) produces  
∇  ∇XH =∇  J (12)

By the continuity equation in (11), the left side of (12) should be the negative of the first
derivative of charge density. Therefore, substituting (11) into (12) gives
  ∂ρ
∇  ∇XH ?
=
?
∂t
(13)

Is the relation in (13) true? (We have used the question marks in (13) to denote the tentative
nature of this conclusion.) Let’s see if (13) is true.

Here is an vector operation identity from the area of mathematics called vector analysis.
For any arbitrary vector field A the following equation is true:
 
∇  ∇XA =0 (14)

Equation (14) holds for any vector field and holds regardless of the coordinate system.
Proving (14) requires more mathematics than we have had so far in this course, but showing
that (14) is true is straightforward. Lengthy, but straightforward. (You are asked to do this
in Cartesian coordinates in Problems 16-7 through 16-9.)

If we substitute the magnetic field H into the relation given by (14) we get
 
∇  ∇XH =0 (15)

But the relation (15) contradicts (13). (Equation (15) turns out to be the true relation.)
Therefore, we have an inconsistency which we must resolve.

Problems: 16-7, 16-8, 16-9


ECE 303 - Sum10 Notes Set 16 : Maxwell’s Equations 8

THE FINAL EQUATION

Here is one approach to resolving the inconsistency created by equation (9). Rewrite the
continuity equation of (11) as
∂ρ
∇  J+ =0 (16)
∂t
Since the right side of (15) is zero, substitute (16) into the right side of (15), obtaining
   ∂ρ 
∇  ∇XH =∇  J+
∂t
(17)

Recall now Gauss’ Law relating the E-field and charge density:

∇  E = ρε (18)

Solving (18) for charge density then gives


 
ρ=ε ∇ E (19)

Substituting (19) for ρ into (17) and rearranging then gives


   
∇  ∇XH =∇  J + ε ∂t∂ ∇ E (20)

The two divergences on the right side of (20) can be rearranged to give
   ∂ 
∇  ∇XH =∇  J+ ε E
∂t
(21)

Using limiting arguments, it is straightforward to show that equation (21) is valid if the
quantities inside the parentheses are equivalent; that is, if

∂E
∇ X H = J+ ε (22)
∂t
ECE 303 - Sum10 Notes Set 16 : Maxwell’s Equations 9

MAXWELL’S EQUATIONS

Equation (22) is therefore the appropriate form for the curl of H. In words, equation (22)
says a magnetic field (H) can be created by either a conduction current density J or a time-
varying electric field (the first derivative of E term). This time-varying E-field term has
substantial implications which we will explore in the next set of Notes.

For easy reference, the four field equations are listed below. Collectively, they constitute
the set of equations called Maxwell’s Equations:

ρ
∇E= (23a)
ε

∇H=0 (23b)

∂H
∇XE= µ (23c)
∂t

∂E
∇ X H = J+ ε (23d )
∂t

In the next set of Notes we will solve Maxwell’s equations for a specific set of conditions.
We will see that the set of conditions we choose leads to the phenomena of a propagating
electromagnetic wave. The problems listed below give you some practice in working with
curl differential equations.

Problems: 16-10 through 16-13


ECE 303 - Sum10 Notes Set 17: Plane Wave Propagation 1

INTRODUCTION

In the preceding set of Notes we introduced Maxwell’s equations. In the form in which
we saw them they are a very general set of equations, and apply to Cartesian, Cylindrical,
and Spherical coordinate systems. However, to gain useful results we frequently have to
formulate a problem in a particular coordinate system. It is at this point we have to evaluate
the necessary partial derivatives and then solve the remaining partial differential equation.

In this set of Notes we examine one of the simpler geometries and physical situations.
Specifically, we want to find the electric field in a non-conducting medium, at a point
which is “far” from any current sources. Although at first these constraints might seem
to be somewhat limiting, this situation accurately describes the electric field geometry in
broadcast television, broadcast radio, cell phones, wireless internet access, and many other
communications environments.

The reason for enforcing these constraints is that it greatly simplifies the mathematics of
solving Maxwell’s equations. We will be able to follow the solution process all the way
from the initial statement, using vector derivative operators, to a scalar wave equation which
we have already solved. Therefore, we will be able to use our prior knowledge of the wave
equation to infer properties of the electromagnetic field in a three-dimensional space.
ECE 303 - Sum10 Notes Set 17: Plane Wave Propagation 2

NON-CONDUCTING MEDIUM

Let’s investigate solving Maxwell’s curl equations for the E and H fields in a non-conducting
region (a medium) having zero charge density (ρ = 0). This accurately describes free-space
and links used in broadcast TV, radio, cell phones, and wireless internet access.

For easy reference, here are the two curl equations from Maxwell’s equations listed in the
preceding Notes Set:

∂H
∇XE= µ (1 )
∂t
∂E
∇ X H = J+ε (2 )
∂t

If the medium is non-conducting, then it has conductivity σ = 0. Since J = σE, this


implies J = 0. Therefore, using J = 0 in (2) gives the following two equations relating E
and H:

∂H
∇XE= µ (3 )
∂t
∂E
∇XH=ε (4 )
∂t

To find the electric and magnetic field, we have to solve equations (3) and (4) for E and H.
Even though the equations look very complicated, they are actually two equations in two
unknowns. Thus, our approach is to manipulate equations (3) and (4) via math and vector
operations to produce one equation in one unknown. This one equation might be a vector
partial differential equation, and it might be difficult to solve, but it will be one (vector)
equation in one (vector) unknown.
ECE 303 - Sum10 Notes Set 17: Plane Wave Propagation 3

SOLUTION APPROACH

To derive the desired equation, first take the curl of both sides of the equation (3):


∇X(∇XE)= µ ( ∇XH) (5)
∂t
Substitute equation (4) into the right-hand side of equation (5):

∂ ∂E ∂2 E
∇X(∇XE)= µ ( ε )= µε (6 )
∂t ∂t ∂t 2

A VECTOR IDENTITY :
There is a general vector identity which will simplify equation (6). Given a general
vector field G, the following is true:

∇X(∇XG) = ∇(∇G) ∇2 G (7 )

where ∇2 G in (7) is called the vector Laplacian of the field G.

With the general vector G in (7) replaced by the specific electric field vector E we have
ρ
∇X(∇XE) = ∇(∇E) ∇2 E = ∇( ) ∇2 E (8 )
ε
But in the non-conducting medium we have assumed that the charge density ρ = 0. Sub-
stituting ρ = 0 into (8) gives the following result:

∇X(∇XE) = ∇2 E (9 )

Substituting (9) into (6) then gives

∂2 E
∇2 E = µε (10)
∂t 2
Since ∇2 E is the vector Laplacian (instead of the scalar Laplacian), equation (10) needs
some further explanation.
ECE 303 - Sum10 Notes Set 17: Plane Wave Propagation 4

VECTOR LAPLACIAN

Suppose the electric field in equation (10) is given in Cartesian coordinates by

E = Ex ux + Ey uy + Ez uz

Then the Vector Laplacian is given by

∇2 E = ( ∇2 Ex ) ux + ( ∇2 Ey ) uy + ( ∇2 Ez ) uz (11)

Each of the three scalar Laplacians in equation (11) has three terms:
∂2 E x ∂2 E x ∂2 E x
∇2 E x = + + (12a)
∂x2 ∂y2 ∂z2

∂2 E y ∂2 E y ∂2 E y
∇2 E y = + + (12b)
∂x2 ∂y2 ∂z2
∂2 E z ∂2 E z ∂2 E z
∇2 E z = + + (12c)
∂x2 ∂y2 ∂z2

SIMPLIFICATION 1: Suppose that Ex and Ey are zero; that is

E = Ez uz = Ez (x; y; z; t ) uz (13)

Substituting the simplication (13) into equations (10) and (11) provides

∂2 Ez
∇2 Ez = µε (14)
∂t 2
Substituting (12c) into (14) then gives the following

∂2 Ez ∂2 Ez ∂2 Ez ∂2 Ez
+ + = µε (15)
∂x2 ∂y2 ∂z2 ∂t 2
Equation (15) is like the wave equation we saw when we studied transmission lines. We
will use another simplification to make equation (15) easier to visualize.
ECE 303 - Sum10 Notes Set 17: Plane Wave Propagation 5

PLANE WAVE GEOMETRY

SIMPLIFICATION 2: Suppose that Ez has no variation in the x or the z direction; that is

∂Ez ∂Ez
= = 0 (16)
∂x ∂z
The simplification in (16) leads to the field geometry below:

E E

y = y1 plane y = y2 plane

The figure above shows that Ez may only vary as a function of y. Once y is fixed, the Ez
is fixed, regardless of the coordinates of x or z. This configuation for the electric field is
called a Plane Wave.

The plane wave approximation is very widely used in communications applications. If a


receiver is located many wavelengths away from the transmitting antenna, then the received
electromagnetic field is essentially a plane wave.
ECE 303 - Sum10 Notes Set 17: Plane Wave Propagation 6

WAVE EQUATION SOLUTION

How do these simplifications affect the mathematical solution for the E-field? Substituting
equation (16) into equation (15) shows that
∂2 Ez ∂2 Ez
= µε 2 (17)
∂y2 ∂t
The partial derivative on the left hand side of (17) shows that Ez can only vary as a function
of y and t. Furthermore, the following two properties are very important:

 If time t is fixed, then Ez varies only as a function of the position y.


 If position y is fixed, then Ez varies only as a function of time t.

Equation (17) is the one-dimensional Wave Equation, just like we studied in lossless Trans-
mission Lines. Therefore, the solution to (17) is known immediately to be

Ez (y; t ) = g(t
y
vp
) ; where v p = p1µ ε (18)

In equation (18), v p is the velocity of the wave and “g” is any realizable function.

EXAMPLES: Let β = ω=v p . The following are solutions of the wave equation given by
equation (17):

 Ez (y; t ) = Ẽz (y; t ) = A e j(ω t βy) (19)


 Ez (y; t ) = A cos(ω t βy) = RefẼz (y; t )g (20)

Note that there is no attenuation of the amplitude in either of the above solutions. In actual
applications, the wave Ez -field will experience an attenuation due to Spherical Spreading.
We will not address that phenomenon in this course.

Problems: 17-1, 17-2


ECE 303 - Sum10 Notes Set 17: Plane Wave Propagation 7

MAGNETIC FIELD SOLUTION

We now have a solution which shows the E-field to be a propagating wave. How do we
determine the H-field? One method is to substitute our solution for E into Maxwell’s
equations and solve for H. Let’s choose the simple one-dimensional exponential wave
solution from (19). This represents a wave having only a uz -component and propagating in
the y-direction:
Ẽ = A e j (ω t β y ) uz = Ẽz uz (21)

Substitute Ẽz from equation (21) into equation (3), giving



ux uy uz



∂ ∂ ∂ ∂H̃
∇ X Ẽ = = µ (22)
∂x ∂y ∂z
∂t


j (ω t βy)
0 0 Ae
Evaluating the curl on the left side of (22) then provides
h i
∇ X Ẽ = ux j β A e j(ω t βy)
= j β Ẽz ux (23)

Substituting (23) into (22) thus gives


∂H̃
j β Ẽz ux = µ (24)
∂t
The left side of (24) shows that the time-varying component of H̃ has only a ux component.
Thus, we need only solve for this H̃x component. Also, note that the H̃-field is orthogonal
to both the Ẽ-field and the direction of propagation. The ODE for the H̃x -component is
obtained from (24) as
d H̃x jβ
= Ẽz (25)
dt µ
Equation (25) says that H̃x is porportional to Ẽz . We may now solve for H̃x using solution
techniques for Ordinary Differential Equations, as outlined in the problems below.

Problems: 17-3, 17-4


ECE 303 - Sum10 Notes Set 17: Plane Wave Propagation 8

INTRINSIC IMPEDANCE

There is another approach to solving for the magnetic field which emphasizes a fundamen-
tal property of the medium. Again let Ẽz have the form from equation (21). However, now
assume that the magnetic field H̃x is related to Ẽz by just an unknown scaling factor η:

Ẽz A j(ω t βy)


H̃x = = e (26)
η η
V
Note the similarity of equation (26) to Ohms Law: I = . Compute the time-derivative of
Z
(26) and substitute the result into equation (25), giving
A j(ω t
j β A e j(ω t βy)
= j ωµ e βy)
(27)
η
Solving for the unknown η then gives
j ωµ A ωµ
jβA =
η
) η=
β
(28)

p
Use the definitions of β = ω=v p and v p = 1= µ ε to simplify η in (28):
ωµ
η=
β
= vp µ = pµµ ε (29)

Performing a final simplification of (29) then gives the desired result:


r
µ
η= (30)
ε
Since the units of η are in Ohms, η is called the intrinsic impedance of the medium. In free
space, ε0 = 8:85(10) 12 and µ0 = 4π(10) 7. These values gives the following important
result for free-space:
η  377 Ohms (31)

Problem: 17-5
ECE 303 - Sum10 Notes Set 17: Plane Wave Propagation 9

FIELD DIRECTIONS

The previous example illustrates a general property: the electric field propagates with the
magnetic field. The propagating electric field can not exist without a propagating magnetic
field. Thus, the phenomenon is called an Electro-Magnetic wave. Here are some important
properties for propagating electromagnetic waves in a lossless medium:

 The E-field and H-field are orthogonal to each other.


 Both fields are orthogonal to the direction of propagation of the wave.
 The magnitude of the magnetic field is porportional to the magnitude of the electric
field: H = E =η.

The figure below illustrates the propagation of the real-valued fields E = A cos(ω t βy) uz
A
and H = cos(ω t βy) ux . These waves are propagating in the +y-direction.
η

E = A cos( ω t − β y) u z

direction of
propagation
H= _
A cos(ω t − β y) u
H
η x

x
E

Problem: 17-6
ECE 303 - Sum10 Notes Set 17: Plane Wave Propagation 10

POWER DENSITY

We have now found that the E and H fields of an electromagnetic wave are orthogonal to
the direction of propagation. There is another field which has a component aligned with
the direction of propagation. This field is the instantaneous power density field P, which is
computed using the cross-product:

P = EXH (32)

The Cross-Product operation has been studied previously. Since the result of a cross-
product is orthogonal to each vector contained in the cross-product, then (32) gives the
result that P is orthogonal to both E and H and is therefore in the direction of the wave
propagation. Since the units of P are in Watts/m2, the vector P is sometimes called the
Power Density vector. The figure below displays the relation between P, E, and H:

P
H

Let the vectors in the figure above be given by P = P uP , E = E uE , and H = H uH . Since


P, E, and H are all mutually orthogonal, this gives the following useful properties:

uE X uH = uP (33a)

uH X uP = uE (33b)

uP X uE = uH (33c)

Equations (33a) - (33c) are very important. They state that knowledge of the directions of
any two of P, E, or H allows us to compute the direction of the third.

Problems: 17-7 through 17-11


ECE 303 - Sum10 Notes Set 17: Plane Wave Propagation 11

POWER DENSITY EXAMPLE

As an example, consider an electromagnetic wave propagating in a lossless medium. Sup-


pose the real-valued electric field associated is

E = 2 cos(ω t βy) uz (34)

The argument of the cosine in equation (34) tells us the EM-wave is propagating in the
+y-direction. The directions of E and P now fix the direction of H. Using equation (33c)
we see that since uP X uE = uH , the H-field must be in the ux -direction.

To find the magnitude of H, remember that H =


E
η
. Thus, H =
2
377
 0 0053. This gives
:

the H-field as
H = 0:0053 cos(ω t β y) ux (35)

To compute the numerical value for P, use the cross-product from (32):

ux uy uz



P = EXH = 0 0 2 cos(ω t β y) (36)


0:0053 cos(ω t β y) 0 0

Evaluating this cross-product then gives

P = 0:0106 cos2 (ω t β y) uy (37)

This last result shows that propagating power flows in the uy direction. Note that the wave
itself propagates in this same direction. Also note that P in (37) is still a function of time.
Therefore, P is sometimes called the Instantaneous power density.

Problems: 17-12 through 17-18


ECE 303 - Sum10 Notes Set 17: Plane Wave Propagation 12

TIME-AVERAGE POWER DENSITY

The power density vector P in equation (37) can be written in a form which separates its
magnitude and direction:
P = P(t ) uP (38)

where uP is a unit vector in the direction of power density flow and P(t ) is the scalar
amplitude of the time-dependent power density. This direction uP is also the direction of
propagation of the electromagnetic wave. In general, P(t ) will also be a function of space,
as well as time.

The scalar P(t ) in (38) is an instantaneous field, meaning it varies over time t and space.
We are frequently interested in the time-average power density Pavg , given by

Pavg = Pavg uP (39)

Note that Pavg points in the same direction as power density. However, the scalar amplitude
Pavg now is not a function of time and is given by the average of P(t ) over one period of
the sinusoid: Z Tp
1
Pavg = P(t ) dt (40)
Tp t =0
where Tp is the time period of the sinusoidal frequency. In the complex-valued phasor
domain, an equivalent expression for the vector Pavg in (39) is given by the complex-domain
computation
1 n o
Pavg = Re Ẽ X H̃ (41)
2
Equation (41) is similar to the expression we found in previous work for the time-average
power flowing on a transmission line. However, in equation (41), Pavg is a Power Density
vector and has units of Watts/m2 , instead of watts.
ECE 303 - Sum10 Notes Set 17: Plane Wave Propagation 13

TIME-AVERAGE POWER DENSITY : EXAMPLE

As an example, let’s reconsider the real-valued example from the previous pages. The
complex-valued electric field corresponding to the real-valued E in (34) is given by

Ẽ = 2e j(ω t βy)
uz (42)

The complex-valued magnetic field corresponding to the real-valued H in (35) is given by

H̃ = 0:0053 e j(ω t β y)
ux (43)

Now compute the cross-product required in (41) :


ux uy uz

Ẽ X H̃

=

0 0 2e j (ω t βy)

= 0:0106 uy



0:0053 e j( ω t β y) 0 0

(44)

Substituting (44) into equation (41) then gives the desired result :
1 n o
Pavg = Re Ẽ X H̃ = 0:0053 uy (45)
2
ECE 303 - Sum10 Notes Set 18 : Wave Propagation in Lossy Media 1

INTRODUCTION

In previous work we studied the propagation of electromagnetic waves in free-space. We


modified the general problem to examine a wave propagating in a region where it could be
considered a plane wave. We found that this simplified situation led to the wave equation,
which showed that plane wave experienced no attenuation as it propagates.

In actual applications, we would find that there is an attenuation due to spherical spreading,
in which the transmitted power is distributed over a larger surface as the wave propagates.
However, this spherical spreading is not a function of the non-conducting region through
which the wave is propagating.

In constrast to this lossless propagation, waves propagating in conducting media do expe-


rience a loss in amplitude. For this reason, these media are sometimes called lossy media.
We will study this case in detail in this set of Notes. We will find that the wave propaga-
tion characteristics in a lossy media are substantially different from waves propagating in
non-conducting media or free-space.
ECE 303 - Sum10 Notes Set 18 : Wave Propagation in Lossy Media 2

MAXWELL’S EQUATIONS: Time-Domain

The beginning point is again Maxwell’s equations. However, now the medium in which the
wave is propagating is a lossy medium; that is, the medium has a non-zero conductivity.
Let this non-zero conductivity be σ in units of Siemens/meter. If σ > 0, then the current
density J in the conducting medium is J = σ E. Therefore, Maxwell’s two curl equations
are as previously given :

∂E
∇ X H = σE + ε (1 )
∂t
∂H
∇XE= µ (2 )
∂t

We do not need to solve (1) and (2) using the detailed approach we used previously. In fact,
we can use the form of the plane wave solution and propose a similar form for the solutions
to (1) and (2) above. Therefore, let’s assume the following phasor solutions for Ẽ and H̃:

Ẽ = F (x; y; z) e j ω t uE (3a)
1
H̃ = F (x; y; z) e j ω t uH (3b)
η

where F (x; y; z) is a scalar function which is not dependent on time t, and η is a possibly
complex constant signifying a possible phase difference between Ẽ and H̃. The unit vectors
for Ẽ and H̃ are uE and uH . Specifiying the unit vectors separately allows for the possibility
that Ẽ and H̃ are in different directions.

Here is a point to realize when we ”assume a solution form”: If the form of the assumed
solution is “wrong”, then a mathematical inconsistency will appear in the suceeding anal-
ysis. We saw this earlier when we needed to make a modification to Ampere’s Circuital
Law, such that it was consistent with the Continuity Equation.
ECE 303 - Sum10 Notes Set 18 : Wave Propagation in Lossy Media 3

MAXWELL’S EQUATIONS: Phasor-Domain

The solution forms assumed in equations (3) have separated the solution into a time-domain
part (e j ω t ) and a spatial-domain part F (x; y; z). In this approach, the time-domain part is
known once we specifiy the frequency ω. Therefore, we only have to solve Maxwell’s
equations (2) for the spatial solution F (x; y; z).

Since the forms for the solutions in equation (3) have time-domain phasors, then the
time-derivatives are easy to compute. Therefore, using (3) we have the following time-
derivatives:

∂Ẽ
= j ω F (x; y; z) e j ω t uE = j ω Ẽ (4 )
∂t
∂H̃ 1
= jω F (x; y; z) e j ω t uH = j ω H̃ (5 )
∂t η

Equations (4) and (5) show one of the benefits of working with the complex-valued phasors.
Specifically, time-differentiation translates into a j ω multiplier.

Now substitute (4) and (5) into (1) and (2), respectively. This gives the following equations
for the phasor fields Ẽ and H̃:

∇ X H̃ = (σ + j ω ε) Ẽ (6 )

∇ X Ẽ = j ω µ H̃ (7 )

These last two equations must now be solved simultaneously for Ẽ and H̃. We will do this
on the next page.
ECE 303 - Sum10 Notes Set 18 : Wave Propagation in Lossy Media 4

SOLUTIONS

We will again limit our study to plane waves propagating in a conducting medium. Let the
Ẽ-field be in the z-direction and let the wave be propagating in the y-direction. A solution
process similar to that for the non-conducting medium (like free-space) in Notes Set 16
would show that the following are the phasor solutions for the Ẽ and H̃ fields:

γ y jω t
Ẽ = Ae e uz (8a)
A γ y jω t
H̃ = e e ux (8b)
η

In equations (8a) and (8b) above, γ is called the propagation constant for the medium and
η is called the intrinsic impedance. The solution process would show that the fields have
the relatively simple forms given in (8a) and (8b), provided γ and η are computed by the
following expressions:

p
γ= jω µ (σ + jω ε) (9a)

s
jω µ
η= (9b)
σ + jω ε

The quantities in (9) are similar to the related quantities of propagation constant and charac-
teristic impedance we obtained for transmission lines. The quantities γ and η are complex
numbers and are very useful when written as

γ = α+ jβ (10a)

η = jηj e jφ (10b)
ECE 303 - Sum10 Notes Set 18 : Wave Propagation in Lossy Media 5

SOLUTIONS (cont.)

Using (10a) and (10b) for γ and η in (8a) and (8b) for Ẽ and H̃ then gives the following
phasor solutions for the fields:

αy
Ẽ = A e e j(ωt βy)
uz (11a)

A αy
e j (ωt βy φ)
H̃ =
j η je ux (11b)

We notice that the solutions in (11) are somewhat different from the solutions for free-space
propagation we obtained in Notes Set 16. Specifically, two properties are immediately
evident from equations (11a) and (11b):

 The amplitude of both Ẽ and H̃ attenuates as the wave propagates into the medium.
 The phase of the H̃-wave differs from the phase of the Ẽ-wave by the angle of the
intrinsic impedance.

These two properties directly affect the power that an incident wave can carry into a con-
ducting medium. We will explore this phenomenon in the remainder of this Notes Set.

Problems: 18-1, 18-2, 18-3


ECE 303 - Sum10 Notes Set 18 : Wave Propagation in Lossy Media 6

E-FIELD EXAMPLES

Suppose a medium existing over y  0 has the parameters (σ; µ; ε) = (4; µ0 ; 72ε0). Let a
wave with f = 5 MHz have the form
αy
Ẽ = A e e j(ωt βy)
uz (12)

as it propagates in this medium. Suppose the following boundary condition is known:

Ẽ(y = 0; t = 0) = 10 uz (13)

Do the following: (a) Compute A, (b) Find the direction of propagation, (c) Compute γ
and write it in the form γ = α + j β, (d) Write the expression for the real-valued E-field.

(a) Compute A.
To compute A, substitute y = 0 and t = 0 into (12) and equate the result to the bound-
ary condition given in (13):

Ẽ (y = 0; t = 0) = A e α ( 0 ) e j ( ω ( 0 )+β ( 0 ) ) = 10 (14)

The exponentials on the left side of (14) evaluate to unity and thus we find that

A = 10 (15)

(b) Find the direction of propagation.


Consider the exponential in (12), given by e j ( ω t β y ) . Since the spatial argument
in this exponential is given by β y, then the propagation of the wave is in the +y
direction (rightward direction).
ECE 303 - Sum10 Notes Set 18 : Wave Propagation in Lossy Media 7

E-FIELD EXAMPLES (cont.)

(c) Compute γ and write it in the form γ = α + j β.


Equation (9a) gives the expression for the propagation constant, repeated here for
reference: p
γ = j ω µ ( σ+ j ω ε ) (16)

Some of the computations required in (16) are shown below:


j ω µ = j π 107  4π 10 7
= j 39:48 (17a)

σ + j ω ε = 4 + j 107 π  72 (8 85)10
:
12
= 4+ j 0:02 (17b)

Substituting (17a) and (17b) into (16) then gives


r  
γ= j 39:48 4 + j 0:02 (18)

It is straightforward to show that (18) is equivalent to


o
γ = 12:58 e j 45 = 8:89 + j 8:89 (19)

The rectangular form for γ on the far right side of (19) then shows that

α = 8:89 and β = 8:89 (20)

(d) Write the expression for the E-field.


Take the real part of (12) to obtain the expression for the real-valued field:
αy
E=Ae cos(ω t β y) uz (21)

Substituting the computed parameters from (15) and (20) into (21), the form form
the real-valued field becomes

E= 10 e 8:89 y cos(107 π t 8:89 y) uz (22)

Problems: 18-4, 18-5


ECE 303 - Sum10 Notes Set 18 : Wave Propagation in Lossy Media 8

H-FIELD EXAMPLES

Now let’s work some examples for determining the magnetic field. For simplicity, we will
use some of the results of the previous example for the electric field. Suppose a medium
exists over y  0 and has (σ; µ; ε) = (4; µ0 ; 72ε0). Suppose an E-field propagating to the
right in this medium.has the form

E = 10 e 8:89 y
cos(107 π t 8:89y) uz (23)

Do the following: (a) Compute the direction of H, (b) Compute the value of the intrinsic
impedance η, (c) Compute the real-valued H field.

(a) Compute the direction of H.


From (23), we see the E-field of the wave is in the uz direction. Equation (23) also
shows the wave is propagating in the +y direction. Therefore power density P also
in the +y direction, and thus has a uy unit vector. In equation form we now have

uE = uz ; uP = uy (24)

The direction of H is found using the cross product relations given in Notes Set 16:

uE X uH = uP (25)

Substituting (24) into (25) then gives

uz X uH = uy (26)

Equation (26) is satisfied if


uH = ux (27)

(b) Compute η.
Equation (9b) gives the expression for the intrinsic impedance, repeated here for
reference: s
jωµ
η= (28)
σ+ j ω ε
ECE 303 - Sum10 Notes Set 18 : Wave Propagation in Lossy Media 9

H-FIELD EXAMPLES (cont.)

Some of the computations required in (28) are shown below:

j ω µ = j π 107  4 π 10 7
= j 39:48 (29a)

σ + j ω ε = 4 + j 107  π 72 (8 85)10
:
12
= 4+ j 0:02 (29b)

Substituting (29) into (28) then gives


s s
jωµ j 39:48
η = = (30)
σ+ j ω ε 4 + j 0:02

Simplifying (30) then gives the desired form for η :

η = jηje jφ = 3:14 e j 45
o
(31)

(c) Compute the real-valued H field.


With the direction of the magnetic field given from (27) as uH = ux , we can use
equation (11b) to write the phasor H̃ as follows:
A αy
e j (ωt βy φ)
H̃ =
j η je ux (32)

Substitute (15) for A, (31) for j η j and φ, and (20) for α and β into (32), giving
7π t 45o )
H̃ = 3:18 e 8:89 y
e j ( 10 8:89 y
ux (33)

Thus, the real-valued H-field is given by taking the real part of (33):

H= 3:18 e 8:89 y
cos(107 π t 8:89 y 45o ) ux (34)

Problem: 18-6
ECE 303 - Sum10 Notes Set 19 : Wave Reflection and Transmission 1

INTRODUCTION

This set of Notes considers the case when waves propagating in one medium are incident
upon the boundary containing another medium, Some of the incident wave energy may be
transmitted into the new medium and some energy may be reflected back into the original
medium. Common examples of this are the electromagnetic waves broadcast for television,
radio, cell phone, and wireless internet applications.

In many of these examples, a broadcast antenna sends electromagnetic waves out over a
region of space, but the waves do not have a great deal of directionality. Indeed, this is one
of the advantages of a broadcast medium: a user or subscriber is not limited to a specific
fixed location. Think of using a cell phone in different places in a neighborhood or campus:
you can get a “good” signal in many places. Think of using the radio in a vehicle: you can
get a particular station over a wide area.

However, this very advantage of the broadcast application leads to the following situation.
The electromagnetic waves can be reflected from nearby structures, and this can lead to
interference effects. This interference occurs because there now may be multiple paths from
the broadcast antenna to the receiving antenna. It is therefore possible for these multiple
paths to create “destructive interference” at the receiver.

This set of Notes examines wave reflection and transmission at boundaries. We will see the
intrinsic impedance of the different media determine the degrees of reflection and transmis-
sion. You are probably familiar with the wave reflection process in acoustics: it is called
an echo. We will see the electromagnetic situation is very similar in concept to the acoustic
echo.
ECE 303 - Sum10 Notes Set 19 : Wave Reflection and Transmission 2

TWO-MEDIA BOUNDARY

This page lists the properties of the two adjoining lossless media which are important for
understanding wave transmission and reflection. The figure below shows the geometry of
the boundary, and properties of medium 1 and medium 2:

Medium 1 Medium 2

( µ1, ε1) ( µ 2, ε 2 )

η1 η2

Γ τ

At the basic level, the lossless media are defined by their electric permittivity ε and mag-
netic permeability µ. This is the top level of descriptors shown in the figure above. How-
ever, we have seen that these basic descriptors define the intrinsic impedance η of the
media. Therefore, if we know the εi and µi of the ith medium, we can compute the intrinsic
impedance ηi of the ith medium using the previously obtained formula
rµ i
ηi = (1 )
εi

This computation is represented by the downward arrow to the ηi in the figure. The ar-
rows down from the ηi to the reflection coefficient, Γ, and the transmission coefficient τ
imply that Γ and τ can be computed from the ηi . Reflection coefficient and transmission
coefficient are explained on the next page.
ECE 303 - Sum10 Notes Set 19 : Wave Reflection and Transmission 3

TWO-MEDIA BOUNDARY (cont.)

The intrinsic impedance η is a very important parameter for a medium, However, if we


are interested primarily in wave transmission and reflection, then we can use the η to cal-
culate even more useful parameters. These are the values of reflection coefficient Γ and
transmission coefficient τ shown in the last line of the figure on the previous page.

The theoretical computation of Γ and τ proceeds from the basic property that the tangential
component of the E-field must be continuous across a boundary. Enforcing this constraint
and using limiting arguments gives the following computational formulas for reflection
coefficient Γ and transmission coefficent τ :

η2 η1
Γ= (2 )
η2 + η1

2 η2
τ= (3 )
η2 + η1

It is much easier to work reflection/transmission problems using reflection coefficient Γ


and transmission coefficent τ. On the following pages we will show how these parameters
can help us obtain the fields in the vicinity of a reflection/transmission boundary.

Suppose we have the case where η1 and η2 are real-valued and positive. Then from (3)
we see that the reflection coefficient Γ may be either positive or negative, depending on
the specific values of η1 and η2 . If Γ is negative this denotes a 180Æ phase change (a sign
change) in the reflected wave ER . However, note from (3) that since η2 can not be negative,
then the transmission coefficient τ can not be negative. Therefore, there is no sign change
between EI and ET as the electric field crosses the boundary between the media.

Problem: 19-1
ECE 303 - Sum10 Notes Set 19 : Wave Reflection and Transmission 4

NORMAL INCIDENCE

In this course we will only consider the case of Normal Incidence. This means the unit
normal uN of the reflecting boundary is orthogonal to the incident E-field vector. For our
example figure shown below, assume all E-fields are in the uz direction (up in the figures)
and we have two media which are both lossless. Let the phasor form for the incident wave
propagating in medium 1 be given by
j(ωt β1 y)
ẼI =Ae uz (4 )

The incidient wave is propagating in the y-direction (to the right), as shown below:

Medium 1 Medium 2
~
Incident EI
~
Transmitted E T

~
Reflected E z
R

x (out)
. y

The reflected and transmitted waves shown in the figure are easily described using the
reflection coefficient Γ and the transmission coefficient τ. These fields are given by
ẼR = A Γ e j ( ω t + β1 y) uz (5 )

j(ωt β2 y )
ẼT =Aτe uz (6 )

Note in (5) that the sign on the spatial phase term β1 y is now a positive sign. This means the
reflected wave is propagating in the y direction. Also note in (5) and (6) that the phase
constant β is a function of the particular medium. Specifically, let the ith medium have
p
velocity of propagation vi = 1= µi εi . Then we have

βi =
ω
=
p
ω µi ε i (7 )
vi
ECE 303 - Sum10 Notes Set 19 : Wave Reflection and Transmission 5

TRANSMISSION EXAMPLE

Let’s work an example for finding the Ẽ field transmitted from free-space into a medium.
Suppose an incident plane wave in free space propagates in the +y direction and has an
electric field given by
ẼI = 5 e j ( 2π 10 t β1 y ) uz
9
(8 )

where β1 is the phase constant in medium 1 (free space). This wave is incident on the
boundary of a dielectric having permeability ε = 10 ε0 and permeability µ = µ0 . Compute
the real-valued field ET transmitted into the dielectric.

APPROACH :
We will first find ẼT and then take the real part to find ET . Here are some properties of the
electric field transmitted into the dielectric:

 The Hertz (or rad/sec) frequency of the electric field does not change as the wave
enters medium 2. Thus ẼT has a Hz frequency f = 109 .

 The vector orientation of the electric field does not change as the wave enters the
medium 2. Thus ẼT has a unit vector in the uz direction.

 The direction of propagation does not change as the wave enters medium 2, but the
β constant does change. Thus ẼT has a “ β2 y” in its complex exponential.

 The amplitude of ẼT is the amplitude of ẼI scaled by the transmission coefficient τ.

Using (6) and (8) and incorporating the above properties then gives the following for ẼT :
9 t β2 y )
ẼT = 5 τ e j ( 2π 10 uz (9 )

where in (9) we will have to compute τ and β2 . We will do this on the next page.
ECE 303 - Sum10 Notes Set 19 : Wave Reflection and Transmission 6

TRANSMISSION EXAMPLE (cont.)

To compute β2 , substitute ε2 = 10 ε0 and µ2 = µ0 into equation (7):


p q
β2 = ω µ2 ε 2 = 2π 109 8:85(10) 11  4π(10) 7 (10)

Evaluating (10) then gives the following value for β2 :

β2 = 66:3 (11)

To compute τ we need η2 . To compute η2 , substitute ε2 = 10 ε0 and µ2 = µ0 into (1):

rµ s 7
2 4π(10)
η2 = = (12)
ε2 8:85(10) 11

Evaluating (12) then gives the following value for η2 :

η2 = 119:2 (13)

Since medium 1 is free-space we therefore know that η1 = 377. Substituting this value and
(13) into (3) then gives the evaluation for τ:

2 (119:2)
τ = = 0:48 (14)
119:2 + 377
Substituting (11) and (14) into (9) then gives the phasor form for the transmitted field as
9
ẼT = 2:4 e j ( 2π 10 t 66:3 y )
uz (15)

Finally, taking the real part of ẼT in (15) then provides the real-valued field transmitted
into medium 2:
ET = 2:4 cos( 2π 109 t 66:3 y ) uz (16)

Problem: 19-2
ECE 303 - Sum10 Notes Set 19 : Wave Reflection and Transmission 7

REFLECTION EXAMPLE

Now let’s examine the previous example and compute the electric field reflected back into
medium 1 (which is free-space in this example). Here are some properties of the electric
field transmitted back into free-space:

 The Hertz (or rad/sec) frequency of the reflected electric field does not change as the
wave reflects. Thus ẼR has a Hz frequency f = 109 .

 The vector orientation of the electric field does not change as the wave reflects. Thus
ẼR has a unit vector in the uz direction.

 The direction of propagation reverses as the wave reflects. Therefore, ẼR has a
“+β1 y” in its complex exponential.

 The amplitude of ẼR is the amplitude of ẼI scaled by the reflection coefficient Γ. If
Γ is negative, you can think of this a reversing the direction of ẼR . For this example,
this means ẼR has a unit vector in the uz direction.

Using (5) and (8) and incorporating the above properties then gives the following for ẼR :
9 t + β1 y )
ẼR = 5 Γ e j ( 2π 10 uz (17)

where in (17) we will have to compute Γ and β1 . To compute β1 , substitute ε1 = ε0 and


µ1 = µ0 into equation (7):
p q
β1 = ω µ2 ε 2 = 2π 109 8:85(10) 12  4π(10) 7 (18)

Evaluating (18) then gives the following value for β1 :

β1 = 20:95 (19)

This example is completed on the following page.


ECE 303 - Sum10 Notes Set 19 : Wave Reflection and Transmission 8

REFLECTION EXAMPLE (cont.)

Since medium 1 is free-space we know that η1 = 377. Substituting this value for η1 and
η2 from (13) into equation (2) then gives
119:2 377
Γ = = 0:52 (20)
119:2 + 377
Substituting (19) and (20) into (8) then gives the phasor form for the reflected field as
9 t +20:95
ẼR = 2:6 e j ( 2π 10 y)
uz (21)

Finally, taking the real part of ẼR in (21) then provides the real-valued field reflected back
into medium 1:
ER = 2:6 cos( 2π 109 t + 20:95 y ) uz (22a)

An alternative representation of ER in (22a) is to associate the minus amplitude with a


change in direction of the unit vector. Therefore, moving the minus sign to the unit vector
location in (22a) thus gives

ER = 2:6 cos( 2π 109 t + 20:95 y ) [ uz ℄ (22b)

Equation (22b) specifically denotes the “direction change” of the E-field vector as it is
reflected back into medium 1. The direction has changed from “up” (+uz ) for the incident
wave in medium 1 to “down” ( uz ) for the reflected wave. This representation will be very
helpful in understanding the total fields throughout both media.

Problem: 19-3, 19-4


ECE 303 - Sum10 Notes Set 19 : Wave Reflection and Transmission 9

INTRINSIC IMPEDANCE: CONDUCTORS

Now let’s consider the case of finding the reflection from an ideal conductor. We will model
this ideal conductor by letting conductivity σ ! ∞. (The conductivity for a good conductor
like copper is σ = 5:8(10)7 Siemens/m .) From Notes Set 17 we have seen that the intrinsic
impedance of a conductor is given by

s
jω µ
η = (23)
σ + jω ε

Divide numerator and denominator of the intrinsic impedance in equation (23) by σ:


s v
u
η =
jω µ
=
u
t jω µ
σ
(24)
σ + jω ε 1 + jωσ ε

Take the limit of this last form in (24) as σ ! ∞ :


v
u r
η = lim
u
t jω µ
σ
=
0
= 0 (25)
σ!∞ 1 + jωσ ε 1+0

Therefore, the intrinsic impedance of a ideal conductor is zero. In practice, the conductiv-
ity of actual conductors, like gold, silver, and copper, is frequently approximated as zero.
However, this is not the case for semiconductors, like silicon and germanium. Semiconduc-
tors have finite conductivities, and thus require that equation (23) be used for the intrinsic
impedance.

The relation in (25) is very similar to an approximation you used in circuits. When you an-
alyzed circuits, you may have assumed the voltage drop along the wires was zero. This was
equivalent to assuming zero resistance in the wires, which is similar to assuming infinite
conductivity of the conductor at the boundary.

Problem: 19-5
ECE 303 - Sum10 Notes Set 19 : Wave Reflection and Transmission 10

REFLECTION AND TRANSMISSION: CONDUCTORS

The fact that good conductors have approximately zero intrinsic impedance has substantial
implication for the behavior of electromagnetic waves in the vicinity of conductors. Let’s
determine how this effects the reflection and transmission coefficients for a ideal conductor.
Consider again the figure on page 4 showing the boundary between two media. Let the
medium 1 be free-space (η1 = 377) and medium 2 be an ideal conductor for which η2 = 0:

η1 = 377 ; η2 = 0 (26)

Now substitue the relations in (26) into equations (2) and (3) for the reflection coefficient
Γ and the transmission coefficient τ, obtaining

0 377
Γ= = 1 (27)
0 + 377

2(0)
τ= = 0 (28)
0 + 377

Equation (27) means all of the incident E field energy is reflected back from the ideal
conductor and equation (28) means no energy is transmitted into the conductor. We will
show that the negative sign in Γ in (27) implies a phase shift of π-radians in the wave
reflected back into the medium 1.

On the following pages we will solve for the total electric and magnetic fields at the bound-
ary between free-space and the ideal conductor. Since the boundary coincides with the
surface of the ideal conductor then this is equivalent to solving for the fields on the surface
of the conductor.

Problem: 19-6
ECE 303 - Sum10 Notes Set 19 : Wave Reflection and Transmission 11

ELECTRIC FIELD ON CONDUCTORS

You may be familiar with receiving antennas, such as those on portable televisions and
radios. These antennas are frequently constructed from a good conductor (like a metal)
which has many electrons in the conduction band. These electrons are very mobile within
the metal conductor. Consequently, they may easily be made to move in response to exter-
nal fields, like those in the incident electromagnetic wave.

We may understand the basics of how a current is generated in a receiving antenna by


computing what occurs when an EM wave is incident on the surface of an ideal conductor.
The material to follow uses the figure on page 4 for reference. You can refer to that figiure
for the field directions and propagation directions in the material below.

Let’s compute the E-field on the surface of the conductor. Refering to the figure on page 4,
let an EM wave in free-space (medium 1) be propagating such that the incidient E-field is
given by
ẼI = A e j ( ω t β1 y ) uz (29)

Let this wave be normally incident on the surface of an ideal conductor occupying the re-
gion y  0. The reflected E-field is given by equation (5), repeated here for easy reference:

ẼR = A Γ e j ( ω t + β1 y) uz (30)

Since the boundary is the ideal conductor, substitute Γ = 1 from (27) into (30). This
shows that the reflected electric field is given by

ẼR = A e j ( ω t + β1 y) uz (31)

Although (31) is mathematically correct, there is another form which will later help us find
the direction of the reflected magnetic field. If we associate the minus sign in (31) with the
unit vector, we see that (31) is equivalent to

ẼR = A e j ( ω t + β1 y) [ uz ℄ (32)

We continue this material on the following page.


ECE 303 - Sum10 Notes Set 19 : Wave Reflection and Transmission 12

ELECTRIC FIELD ON CONDUCTORS (cont.)

Now recall that any phasor electric field has the generic scalar-unit vector form given by

Ẽ = Ẽ uE (33)

Comparing the reflected electric field from (32) to the generic form in (33) then gives the
following for the reflected electric field :

Ẽ = A e j ( ω t + β1 y) ; uE = [ uz ℄ (34a; b)

Equation (34b) shows that the ideal conductor has “reversed” the vector direction of the
electric field in the reflected wave. Note also that (34a) shows the direction of propagation
has been reversed, as well.

The total E-field in medium 1 is thus the sum of the incident and reflected fields:

Ẽ1 = ẼI + ẼR (35)

Substituting (29) and (31) into (35) then gives

Ẽ1 = Ae j ( ω t β1 y )
uz A e j ( ω t +β1 y ) uz (36)

Now evaluate (36) at the boundary (y = 0) to find the phasor electric field on the surface of
the ideal conductor:

Ẽ1 (y = 0) = A e j ω t uz A e j ω t uz = 0 (37)

Equation (37) tells us that the E-field on the surface of an ideal conductor is zero. This
is similar to the circuits approximation where we assume there is no voltage loss on the
conductors between circuit elements.
ECE 303 - Sum10 Notes Set 19 : Wave Reflection and Transmission 13

MAGNETIC FIELD ON CONDUCTORS

We have now found that the electric field on the surface of the ideal conductor is zero. Now
let’s find the magnetic field on the surface. Like the E-field in medium 1, the H-field in
medium 1 is the sum of the incident and reflected magnetic fields:

H̃1 = H̃I + H̃R (38)

Refering to the field directions and propagation directions in the figure on page 4 will
help us determine the directions of H̃I and H̃R . Since we know the directions of the Ẽ
components and their propagation directions, we can use the following relation from Notes
Set 14 to find the directions of the H̃ components:

uH = uP X uE (39)

Equation (39) applies to both the incident and reflected components of the fields.

Direction of H̃I :
From equation (29) we see the the following properties for the incident wave ẼI :

uP = uy ; uE = uz (40a; b)

Substituting (40a) and (40b) into (39) then shows that the direction of the magnetic field
for the incident wave is given by

uH = uy X uz = ux (41)

Using (41) thus gives the phasor form for the incident magnetic filed as

ẼI A j (ωt β1 y )
H̃I = uH = e ux (42)
η1 η1

Direction of H̃R :
From equation (31) we see the the following properties for the reflected wave ẼI :

uP = [ uy ℄; uE = [ uz ℄ (43a; b)
ECE 303 - Sum10 Notes Set 19 : Wave Reflection and Transmission 14

MAGNETIC FIELD ON CONDUCTORS (cont.)

Substituting (43a) and (43b) into (39) then shows that the direction of the magnetic field
for the reflected wave is given by
uH = [ uy ℄ X [ uz ℄ = ux (44)

Using (44) thus gives the phasor form for the reflected magnetic field as
A j ( ω t +β1 y )
H̃R = e ux (45)
η1

Total Magnetic Field :


To find H̃1 , the total magnetic field in medium 1, substitute (42) and (45) into (38), giving
A j ( ω t β1 y ) A
H̃1 = e ux + Γ e j ( ω t +β1 y ) ux (46)
η1 η1
Evaluate (46) at the boundary (y = 0) to find the H̃-field on the surface of the ideal conduc-
tor :
A j ωt A jωt
H̃(y = 0) = e ux + e ux (47)
η1 η1
Simplifying (47) then gives the desired result for the phasor magnetic field on the surface
of the ideal conductor :
2A j ω t
H̃(y = 0) = e ux (48)
η1
Taking the real part of (48) then gives the real-valued magnetic field on the surface of the
ideal conductor :
2A
H(y = 0) = cos ωt ux (49)
η1
Equation (49) shows that the EM wave incident on the surfae of the ideal conductor has thus
created a time-varying sinusoidal magnetic field on the surface of the conductor. You can
think of this time-varying magnetic field as “moving” the mobile charges available in the
conduction band of the conductor. This constitutes a sinusoidal current on the conductor
surface.

If you think of the E and H plane waves as being “sheets” incident on the conductor, then
you can think of the current as a “sheet” moving up and down in the z-direction on the
surface of the conductor.

Potrebbero piacerti anche