Escolar Documentos
Profissional Documentos
Cultura Documentos
Dr Francesco Fornetti
Published by Explore RF Ltd, Bristol, United Kingdom
ALL RIGHTS RESERVED. This book contains material protected under International and Federal
Copyright Laws and Treaties. Any unauthorized reprint or use of this material is prohibited. No part
of this book may be reproduced or transmitted in any form or by any means, electronic or
mechanical, including photocopying, recording, or by any information storage and retrieval system
without express written permission from the author / publisher.
While the author and the publishers believe that the information and guidance given in this work are
correct, all parties must rely upon their own skill and judgement when making use of them. Neither
the authors nor the publishers assume any liability to anyone for any loss or damage caused by any
error or omission in the work, whether such error or omission is the result of negligence or any other
cause. Any and all such liability is disclaimed. For permission requests, write to the publisher at the
address below:
Explore RF Ltd
Trym Lodge
Henbury Road
Westbury-On-Trym
Bristol BS9 3HQ
United Kingdom
Any unauthorised broadcasting, public performance, copying or re-recording of the enclosed digital
media will constitute an infringement of copyright.
Printed in the UK
ISBN: 978-0-9576635-1-0
Table of Contents
Contents
I
Table of Contents
II
Table of Contents
III
1.1 Making the transition from DC and AC to Radio Frequency
It has now been ten years since I was first introduced to the basic concepts of RF and
Microwave theory and I recall how excited I was as an undergraduate student to start learning and
understanding this ‘mysterious’ and fascinating subject. Since then I have been to numerous
courses and have tackled a number of engineering tasks in the field. I have also had the opportunity
to teach the subject myself and invariably I can see that the students seem to struggle with exactly
the same concepts which I myself was a bit baffled by back in the days. This myth of RF Engineering
being a bit of a black-magic branch of electronics is not totally unfounded because it often takes
people years to master it and an inexperienced engineer may often be faced with seemingly
inexplicable phenomena.. I feel that the problem lies with the approach which is taken to teaching
RF & Microwave Engineering which, despite being rigorous and formally correct, focuses too soon
and too heavily on the maths when the physical concepts and phenomena should be given priority
and wide breath. Also, and more importantly, it fails to bridge the gap between what happens in AC
and DC circuits, which most students are familiar with, and what happens at Radio and Microwave
Frequencies. This lack of clarity, created by this lack of connection, is further exacerbated by the use
of similar terminology which has a very different meaning at DC/AC and RF.
With this book I will attempt to bridge this gap and make the sharp seam between low and
high frequency theory and techniques, which the classical teaching creates, a great deal smoother
and much easier to overcome.
Most textbooks and courses start with the concept of wavelength and how this quantity
allows you to determine whether you may use a lumped circuit model or whether you need to apply
RF techniques. I, on the other hand, will begin with revising the basic concepts of voltage, current
and impedance in DC and AC circuits and how they relate to Electric and Magnetic fields within such
circuits.
Subsequently I will clarify some mathematical concepts which are often used in the
treatment of the subject and connect them to the physical phenomena that they represent.
I will then introduce the wavelength not just as a simple formula but in terms of its physical
significance and explain why it is a good indication of what techniques need to be used to analyse a
circuit. From there we will move on to the concept of impedance at high frequency and how this
relates to voltages, currents and, more importantly at radio frequency, Electric and Magnetic fields.
A good understanding of these concepts forms the perfect platform onto which RF &
Microwave knowledge may be built and is key to understanding transmission lines, ubiquitous
building blocks of high frequency circuits.
Transmission lines, as most of you will know, are all about delivering power from a generator
to a load. The way we achieve maximum power transfer and maximum efficiency is however very
different for high and low frequency circuits. This is because, while at low frequencies (e.g. 50Hz),
power is conveyed predominantly as voltage and current and is ‘quasi’ electrostatic in field form, at
high frequencies the power is predominantly conveyed as ‘electromagnetic fields’.
This text presents a detailed analysis of transmission lines terminated with different loads. The
concepts of incident and reflected voltages are illustrated both in the case of continuous wave and
pulsed signals. The applications of “special” lines are also illustrated together with the practical
implementation and simulation of the most common transmission line types.
1
CHAPTER 1 - Fundamentals of Electrical Circuits
After the conceptual and practical treatment of transmission lines, the maths which is used to
design, model and analyse such elements is illustrated. This leads to the mathematical definition of
reflection coefficient and impedance seen for a single-port network and then extends to the
definition of network parameters for two-port networks. Graphical tools for impedance
representation based on such parameters are then introduced (Smith Chart).
The basic concepts of impedance at high frequency are then revisited and enhanced to
include series and parallel passive networks comprising of both resistive and reactive elements.
The basics of impedance matching are then explained on a conceptual basis and some basic
examples are illustrated which describe L-section matching and series to parallel conversions.
The simplification of impedance matching through the use of Impedance and Admittance
Smith charts is then illustrated and a number of video tutorials based on simulations are made
available to the reader.
Finally the basic techniques for high frequency amplifier design are illustrated and a number
of video tutorials are provided to corroborate the students’ learning.
I firmly believe that Radio Frequency and Microwave Engineering is not a game of Sudoku
where you just apply rules to get the result that you want. To really be a successful designer you
must understand the concepts not just the procedures behind high frequency circuits.
In particular, an understanding of what happens in the time domain and what an impedance
represent is truly crucial when working with advanced amplifier topologies such as Doherty and
when trying to model devices and systems.
It is highly recommended that the reader who is not familiar with the AWR Microwave Office
simulation software watches the introductory video (video 1.1) before any of the others.
2
1.2 Voltage (V) and Electric Field (E)
Most Electronic Engineers are familiar with voltages and with the way they are used to
represent the behaviour of AC and DC electrical circuits. One should not forget however that Voltage
is also called Electric Potential and it represents the potential energy (per unit charge) associated
with electric forces. The Electric Potential Energy is analogous to the Gravitational Potential Energy.
This is illustrated by the example below
Just as the pumping of water to a higher level results in energy being stored, "pumping"
electrons to create an electric charge imbalance results in a certain amount of energy being stored in
that imbalance. And, just as providing a way for water to flow back down from the heights of the
reservoir results in a release of that stored energy, providing a way for electrons to flow back to their
original "levels" results in a release of stored energy.
When the electrons are poised in that static condition (up in the reservoir), the energy
stored there is called potential energy, because it has the possibility (potential) to be released and
do some work. This potential energy, stored in the form of an electric charge imbalance and capable
of provoking electrons to flow through a conductor, is often expressed in terms of voltage, which is a
measure of potential energy per unit charge of electrons.
We cannot assess the amount of stored energy in a water reservoir simply by measuring the
volume of water, we must also consider how far it will drop from its initial height. Likewise, the
potential energy available for moving electrons from one point to another is relative to those two
points. Therefore, voltage is always expressed as a quantity between two points. The analogy of a
mass potentially "dropping" from one height to another is why the voltage between two points is
sometimes called a voltage drop. Such a voltage drop ΔV may be expressed mathematically as
shown by equation (1.2-1)
3
CHAPTER 1 - Fundamentals of Electrical Circuits
(1.2-1)
Equation (1.2-1) demonstrates how, under certain assumptions, electric circuit analysis may
be greatly simplified by using a scalar quantity, voltage, in place of the electric field, which is vector
entity and hence more complex to handle. In particular in DC and AC circuits the interest is in what
happens at the terminals of each component, for example the voltage at the terminals of a resistor
or capacitor or the voltage drop across it. Since a uniform and conservative field exists in resistors,
capacitors and inductors1, voltage may be used to characterise the effect of the electric field on such
components.
This simplification, which uses a scalar quantity (voltage) to represent the effects of an electric
field (vector) and is appropriate for AC and DC circuits, may not apply when higher frequencies are
used as we will see in due course. What should also be pointed out is that voltage is not a physical
quantity, it simply represents an effect of the real entity which causes it i.e. the electric field. The
presence of water in the reservoir is an effect of the pump pushing the water up to it, so the pump is
what physically changes the level of water in the reservoir, the reservoir filling up is just an effect
which may be quantified and used to represent the work done by the pump.
1
This is true only for ideal components however it is a good enough approximation in most practical cases
when only passive circuit elements are involved and the frequency of excitation remains low.
4
1.3 Current (I) and Magnetic Field (H)
As shown in Figure 1.3-1, the lines of the magnetic field are concentric circles for a long, straight,
current-carrying wire and their direction is given by the right hand rule. Most of you will be familiar
with the Biot-Savart law which allows the vector field B to be calculated. Its modulus is shown in
equation (1.3-1)
| | (1.3-1)
Where represents the distance from the wire and the magnetic permeability of space. In as
much as the B field is considered appropriate to represent magnetic fields in free space, when the
generated fields pass through magnetic materials which themselves contribute internal magnetic
fields, ambiguities can arise about what part of the field comes from the external currents and what
comes from the material itself. It has been common practice to define another magnetic field
quantity, usually called the "magnetic field strength" designated by H. In free space a very simple
relationship exists between the H and B fields
(1.3-2)
Equations (1.3-1) and (1.3-2) show that B and H have the same direction and orientation but
the magnitude of H is obtained by dividing the magnitude of the B field by the scalar quantity .
However when magnetic materials are present, this relationship no longer holds true and B and H
may be different not only in magnitude but also in direction and orientation. Because of this, the H
field is a more general representation of the magnetic field and hence, from here forth, by magnetic
field we will intend the H field, unless otherwise specified. This is also the field which is used at radio
frequency.
Equation (1.3-1) shows that a scalar quantity, current, may be used to describe the flow of
charge in an electric circuit. This, as in the case of voltage and electric field, greatly simplifies the
analysis of the circuit but it is appropriate only under specific assumptions which are generally
satisfied at low frequency. At higher frequencies however, a current alone may not be sufficient to
describe the effect of a magnetic field and hence the vector H may need to be used instead.
5
CHAPTER 1 - Fundamentals of Electrical Circuits
In this section we will explain how the values of resistance R, capacitance C, and inductance L, not
only provide a way to relate voltages and currents for resistors, capacitors and inductors but also a
link between the moduli of Electric and Magnetic fields. For the reader’s convenience the classical
V-I equations for resistors (1.4-1)(a), capacitors (1.4-1)(b) and inductors (1.4-1)(c) are shown below.
( ) ( )
( ) ( ) ( ) ( ) ( ) (1.4-1)
In the following three sections we will look at how we can approach the abovementioned circuit
elements from a field perspective.
1.4.1 Resistors
Let us now consider a simple resistor in the form a high resistance wire. This could be the filament of
a light bulb for instance (Figure 1.4-1) but also a heating element used in household appliances.
For continuity, the current inside the resistor will be the same as the current on the wires connected
to its terminals, let us call it .
Leads
I
I I
a b
Through equation (1.3-1) we can relate the current thought the resistor to the magnetic field inside
it.
(1.4-2)
Also, since the electric field inside the resistor may be considered uniform, it is easy to relate the
voltage across it to the respective electric field through equation (1.2-1)
∫ ∫ (1.4-3)
6
1.4 V & I or E & H ?
If we now calculate the ratio of voltage and current from (1.4-2), (1.4-3), and (1.4-1)(a) we obtain:
(1.4-4)
(1.4-5)
Equation (1.4-5) shows that relates and at a specific point in space around the resistor
(identified by radial distance from its axis) in a linear fashion. This derivation is not entirely
accurate as one would need to take into account material properties to be fully rigorous.
Nonetheless it demonstrates how, in principle, the value of the resistance gives an indication of
the relative magnitudes of vector quantities and , not just of the scalar quantities and which
may be used to represent and in low frequency circuits.
1.4.2 Capacitors
(1.4-6)
(1.4-7)
where is the sheet charge density on each plate, is the absolute value of the charge on either
plate and is the surface area of each plate.
Now let us take a look at the magnetic field between the plates of the capacitor. It would
appear that, since no current, in its classical form2, can flow between the plates, we are unable to
use equation (1.3-1) to obtain the magnetic field from the current. This in fact is not the case
since a displacement current exists between the plates which is equal in magnitude to the current
flowing in and out of the terminals of the capacitor. This comes from Maxwell’s generalisation of
Ampere’s law and its proof is beyond the scope of this treatment3.
2
conduction current
3
For the reader who wishes to research this further, the forth Maxwell equation is used to come to this result
∮ ⁄
7
CHAPTER 1 - Fundamentals of Electrical Circuits
A complex number effectively allows you to represent two values ( and ) with one entity
and comprises of a real part and an imaginary part , as shown below
The letter represents the imaginary constant6 which is equal to √ . Because is multiplied by the
imaginary constant , there is no way to mix it with , which is the beauty of complex numbers.
The number may be plotted on a Cartesian graph where the abscissa represents real values and
the ordinate represents imaginary values as shown in Figure 1.5-5.
Equivalently, may be represented in polar form by means of its modulus and angle as shown in
Figure 1.5-6.
6
is sometimes called
16
1.5 The significance of reactive components equations
(1.5-11)(a)
Now, you will recall, from the physics of rotational motion, that if we are moving around a circle at
an angular speed , the angle which we travel in time is equal to . Equation (1.5-11)(a) may
therefore be rewritten as
(1.5-13)
The term may be represented in the complex plane, as a vector rotating anticlockwise as time
increases, with angular speed . Equally the term represents a vector in the complex plane,
rotating clockwise at radians per second. This is illustrated in Figure 1.5-9. So the numerator of
equation (1.5-13) is just the sum of these two vectors of equal magnitudes which are rotating at the
same speed but in opposite directions. From Figure 1.5-9 (a)-(f), it is apparent that the sum of these
two vectors always lies on the real axis. If we plot the modulus of this sum on the y-axis of an
ordinary Cartesian graph versus time, as shown on the right-hand pane of Figure 1.5-9 (a)-(f), then
we can see that we get a cosine-like function. This is also illustrated in an animated fashion in video
1.5
You may notice however that this function varies between -2 and 2, i.e. is twice the amplitude of a
cosine function. This is why equation (1.5-13) prescribes that we should halve it to get the cosine.
Understanding complex exponentials and what they represent is key to understanding the
mathematical representation of waves and their propagation though transmission lines and media.
As we will see in section 2.14, just as light shining through a glass is partly reflected and partly shone
through, electromagnetic waves and signals undergo the same process at radio frequency. The
reflected and transmitted signals, which travel in opposite directions, may be represented by
complex exponentials rotating around in opposite directions on the complex plane, in a similar
fashion to the ones shown in Figure 1.5-9.
For sine functions we may use a similar equation
(1.5-11)(b)
21
1.6 Animations and Video Tutorials
The reader may now wish to watch some illustrative videos and animations which give further
insight into the topics explained so far in this chapter. A brief summary of the content of this
material is presented below.
VIDEO
CONTENT
REFERENCE
This animation shows Euler’s formula for cosine, eq. (1.5-13), from a graphical
1.5 view point. It shows how two vectors rotating around in the complex plane may
be used to represent a cosine function.
This animation shows Euler’s formula for sine, eq.(1.5-11)(b), from a graphical
1.6 view point. It shows how two vectors rotating around in the complex plane may
be used to represent a sine function.
27
2.2 The true sense of Wavelength
When the frequency of excitation of our circuits is relatively low, the effects of the connecting
wires are limited to minor losses and can often be ignored. When working at higher frequencies
however, the effect of such “wires” can no longer be ignored. In fact they are so significant that we
do not use wires at Radio Frequency, but appropriately designed interconnecting structures called
“transmission lines”.
Now where’s the boundary between low and high frequency and when do we need to use
transmission lines in place of wires?
Well, this comes down to two factors: the maximum excitation frequency of our circuit and its
dimensions.
Let us consider a 100 km power line which carries a 50Hz signal, as shown in Figure 2.2-1.
Power Line
100 km
VS RL
Assuming that the signal will travel down the line at the speed of light10, it will get from the
transmitting to the receiving end in a time equal to
Now let us look at how quickly our signal changes. Well if the frequency is 50Hz then its period is
equal to
This means that every 20 ms our signal11 starts from zero, goes up to a positive peak, then a negative
peak and after 20 ms goes back to zero again and the same pattern repeats.
Now in the that it has taken the signal to travel to the end of the power line, the voltage at
the generator end has not changed by much. In fact, if we turn on our signal generator at time
, by time the signal reaches the end of the line at , the signal generator has
12
only gone through 6% of the signal period .
10
The speed of light in vacuum is approximately equal to ⁄
11
For simplicity assume a sinusoidal signal with an initial phase of zero degrees, ( )
12
This figure is calculated as
31
CHAPTER 2 - Conveying Power at Radio Frequency
This means that, at any instant in time, the signal that we observe at the generator end VS, and the
signal at the end of the power line VL , have a very little phase difference. This is shown in Figure
2.2-2.
Figure 2.2-2 Phase Difference between Generator and Load Signals for a 100km power line
This small phase difference over such a large distance is due to the fact that the speed at which our
signal changes is much lower than the speed at which is propagates.
However, if we made our transmission line much longer, then we would increase the phase
difference between VS and VL considerably. If, for instance, our line was of such a length that, by the
time the signal appeared at the load end, the generator’s signal had gone through a quarter of its
period, we would see a 90⁰ phase shift between VS and VL at any instant in time.
This is shown in Figure 2.2-3.
Figure 2.2-3 Phase Difference between Generator and Load Signals for a 1500km power line
To find the length over which our signal suffers a 90⁰ delay, we simply need to multiply the speed of
light by a quarter of the period as shown by equation (2.2-1).
32
2.2 The true sense of Wavelength
(2.2-1)
From this calculation it is apparent is that, at 50Hz, we need to travel a very long way in order for the
finite propagation speed of the signal to cause appreciable delays. However if the period of our
signal was much shorter, i.e. if our signal varied much faster with time, then the distance over which
such delays become appreciable would be considerably reduced, as we will see shortly.
Now one may ask, “If we have a generator with a load at the end of it, what does it matter if
the signal gets to it with a bit of delay? After all, that is inevitable because of the finite speed of
propagation so what’s the big deal?”.
Well, that is a very fair point and indeed if you have a pure sinusoid travelling down a power line,
into some sort of load which is going to rectify your AC signal and turn it into DC then there is no
problem. However, suppose that you have additional elements connected along your line and that
the distance between such elements is of the order of 100’s of kilometres as shown in Figure 2.2-4.
VS L C RL
In this case, at any instant in time, the phase offsets with respect to the signal generator are
45⁰ across the inductor, 90⁰ across the capacitor and 135⁰ across the resistor due to the finite speed
of propagation of the signal alone. The inductor and the capacitor will also introduce phase changes
themselves which in turn will take time to propagate to other components. This complicates circuit
analysis considerably since the voltages across the various circuit elements are not cophasal. In a
small electrical circuit instead, you would have distances of the order of meters between the
elements and one may therefore assume that the signal propagates at infinite speed. This means
that voltages appear ‘instantly’ across shunt elements and are cophasal which simplifies matters
considerably.
Let us stop for a moment and reflect on how we could avoid these problems. There are three
things we could do:
Option 1 is clearly impossible, but the other two are possible to some extent.
However, in a power line, you are unlikely to get this sort of scenario with components at such large
distances between one another so there’s no reason to worry.
33
2.3 Transmission Lines – an Introduction
We will start our treatment of transmission lines with one the most common types, the
coaxial cable. A coaxial cable is a two-conductor cable made of a single conductor surrounded by a
braided wire jacket, with a plastic insulating material separating the two. The outer conductor
completely surrounds the inner conductor and the two conductors are insulated from each other for
the entire length of the cable.
If an ohmmeter was employed to check the cable’s resistance, it would show the two
conductors to be completely insulated from one other, with nearly infinite resistance between the
two. This is due to the fact that an ohmmeter would use a continuous direct current (DC) to perform
such a measurement. However, the cable’s response to short voltage pulses and high-frequency
signals would be quite different because of the effects of capacitance and inductance distributed
along the length of the cable. When the applied voltage changes rapidly, the cable presents a finite
impedance to the signal source (typically 50 or 75 Ω), and draws a current proportional to the
applied voltage. When such stimuli are used, this pair of wires becomes an important circuit element
with its own characteristic properties which we refer to as a transmission line.
Let us now consider a set of parallel wires of infinite length (Figure 2.3-2), with nothing
connected at the end. What would happen when we close the switch? Would there be no current at
all?
37
CHAPTER 2 - Conveying Power at Radio Frequency
Even if we assume that we are using wires with zero resistance, there would still exists some
capacitance along the cable due to the fact that any pair of conductors separated by an insulating
medium creates capacitance between those conductors (Figure 2.3-3).
When en electric filed is applied to our “pair of wires”, the capacitance which exists between them
means that a current proportional to the rate of change of voltage over time will be drawn. This is
described by the equation (1.4-1)(b), ⁄ . This capacitance stores the energy provided by the
electric field created by the voltage source. According to the equation, an instant rise in applied
voltage (produced by a perfect switch closure) gives rise to an infinite charging current. However,
the current drawn by a pair of parallel wires will not be infinite, because there exists series
impedance along the wires due to inductance (Figure 2.3-4).
Infinite Length
Switch
DC
Source
Such an inductance comes from the fact that current through any conductor develops a magnetic
field of proportional magnitude and energy is stored in this magnetic field (Figure 2.3-5). This stored
energy comes from the electric field created by the source and this transformation of energy from
electric into magnetic manifests itself as a voltage drop governed by the inductance equation
(1.4-1)(c), ⁄ . This voltage drop limits the rate of change of voltage across the distributed
capacitance, preventing the current from ever reaching an infinite magnitude.
38
2.3 Transmission Lines – an Introduction
Figure 2.3-5 Electric and Magnetic fields as signal propagates down a transmission line
Assuming that there is air between our two conductors, voltages and currents will propagate down
our transmission line at nearly the speed of light, progressively charging the distributed capacitance
and drawing a constant current of limited magnitude from the battery. This is shown in Figure 2.3-6.
Since the wires are infinitely long, their distributed capacitance will never fully charge to the source
voltage and this pair of wires will continue to draw a constant current from the source so long as the
switch is closed, behaving as a constant load. When we consider our “wires” under these conditions
we cannot just see them as a pair of conductors, we must treat them as a transmission line.
So our infinitely long transmission line behaves as a constant load and its response to the applied
voltage is resistive rather than reactive, despite the fact that the line comprises solely of inductance
and capacitance (assuming that the wires have zero resistance). This is because, from the battery’s
perspective, there is no difference between a resistor eternally dissipating energy and an infinite
transmission line eternally absorbing energy. The ratio of the battery voltage and the constant
current that an infinite line would draw is called the characteristic impedance ( ) of the line, and it
is determined by the values of our distributed capacitance and inductance, which in turn are fixed by
the geometry of the two conductors.
39
2.8 Reflection and Transmission in unterminated, shorted, matched and mismatched line
We will now look at what happens when we terminate an actual transmission line, with
characteristic impedance Z0 =50Ω, in different impedances. The circuit which we will be analysing is
shown in Figure 2.8-1 and we will be using sinusoidal stimuli from now on.
Z0
Transmission Line
VS ZL
Figure 2.8-2 (a) and Figure 2.8-2 (b) show incident, reflected, and transmitted voltages for an open-
circuit and a short-circuit termination respectively.
──── Incident
──── Reflected
(a)
──── Incident
──── Reflected
(b)
Figure 2.8-2 Incident, Reflected and Transmitted voltage for a line terminated with (a) an open circuit (b) a short circuit
As is to be expected, if there is a short or open circuit at the end of the line, the entire signal is
reflected.
73
CHAPTER 2 - Conveying Power at Radio Frequency
──── Incident
──── Reflected
(a)
──── Incident
──── Reflected
(b)
Figure 2.8-3 Incident, Reflected and Transmitted voltage for a line terminated with (a) 25 Ω (b) 100 Ω
However, when a finite load, whose impedance does not match the characteristic impedance of the
line, is placed at its end the reflection is not quite as severe. A fraction of the voltage, proportional to
the mismatch between ZL and Z0 gets reflected but the remainder does get delivered to the load.
This is shown in Figure 2.8-3.
In the following subsections the graphs will only show the incident and reflected voltages
along the line to avoid crowding the graphs. However the transmitted voltage is easily derived since
it will have an identical phase to the incident voltage and an amplitude equal to the amplitude
difference between incident and reflected voltage. This is demonstrated in Figure 2.8-3.
Now let us examine each termination in more details by taking a snapshot of incident and
reflected voltages along the line at different instants in time between zero and seconds.
represents the period of the signal travelling down the line.
A useful indicator of the proportion of the signal which gets reflected is the reflection
coefficient . This quantity is extremely useful at Radio and Microwave frequencies and we will
define it and describe it appropriately in section 2.14. is a complex number, however, for the time
being, we will just look at its modulus | | and interpret it as an indicator of the relative magnitudes
of reflected and incident signals. is defined in such a way that when the entire incident signal gets
reflected its modulus is 1 whereas when no reflection occur its modulus is zero. is clearly stated on
every graph in sections 2.8.1 to 2.8.5.
These graphs are snapshots taken from a transmission line animation, designed by the author, which
may be downloaded for free at http://docfrankie.com/index_files/matlab_anim.htm
74
2.8 Reflection and Transmission in unterminated, shorted, matched and mismatched line
In this section we will show snapshots of what is happening in our transmission line at
different instants in time when the line terminated with an open circuit.
In each figure, the top graph shows the incident wave (in red), which travels to the right-
hand side of the page, and the reflected wave (in green) which travels in the opposite direction.
The bottom graph shows total voltage (in blue), which is the sum of incident and reflected waves,
along with its envelope (dashed magenta line).
Note that if we travel an integer number wavelengths, the signals at the two ends of the line have
the same phase!
We have picked two points on the transmitted and reflected waves (α and β) and we will
now see how they move along a line terminated with an open circuit.
α β
t=0s
At t=0s, the incident and reflected voltages are 180° out of phase hence their sum is zero!
It is apparent that when the signal finds an open circuit at the end of the line it gets reflected in its
entirety.
75
CHAPTER 2 - Conveying Power at Radio Frequency
α β
t=0.1 T
Incident and reflected voltages are no longer 180° out of phase as the waves move towards one
another hence their sum is finite and shown in blue on the standing wave plot. represents the
period of the signal travelling down the line.
α β
t=0.19T
When Incident and reflected voltages overlap the standing wave amplitude is at a maximum.
76
2.8 Reflection and Transmission in unterminated, shorted, matched and mismatched line
β α
t=0.31T
As the waves move away from each other, their phase difference increases hence the amplitude of
the standing wave decreases..
β α
t=0.45T
77
2.9 Transmission Lines Applied to High Frequency Circuits
Z0
RS Transmission Line
ZIN
VS ZL
Figure 2.9-1 Input Impedance of a Transmission line of characteristic impedance Z0 terminated with load impedance ZL
( )
(2.9-1)
( )
Where is the length of the line and is the propagation constant. , which is equal to ⁄ , is
used to account for the specific speed of propagation at which the signal travels in the transmission
line of interest. Note that, although the speed of propagation does not appear explicitely in the
expression for , at a specific frequency, and may interchangebly be used (eq (2.2-4)).
We will be looking at how equation (2.9-1) is derived in section 3.5 but for now let us just
use it to understand the behaviour of our transmission line.
First of all let us take a look at the ( ) function shown in Figure 2.9-2. Notice that the
periodicity of the trigonometric tangent is just and not 2 as in sine and cosine functions. This
periodicity is one of the features of this trigonometric function which makes it suitable to represent
the impedance of transmission lines, as we will in section 2.10.
93
CHAPTER 2 - Conveying Power at Radio Frequency
Let us now consider some significant values of (since is fixed for a specific transmission
line and frequency) and see how they affect the impedance of our line.
For , equation (2.9-1) becomes
( ) ( )
( )
( )
So we have found that, if a transmission line which is half a wavelength long is placed between
source and load, our transmission line will become transparent. This will be true for every value of
which makes equal to (or an integer multiple of ) and hence ( ) equal to zero. Let us find
the values of for which this is true.
94
2.9 Transmission Lines Applied to High Frequency Circuits
Secondly, since the wavelength changes with frequency, our circuit would only behave as expected
at a specific frequency and perhaps over a narrow band around it which is often undesirable. Thirdly,
when non-deal lines are employed, losses are incurred which are proportional to length. Lastly,
often enough, the load impedance will be some complex quantity and you may want to use equation
(2.9-1) to your advantage to turn the impedance seen by the generator into a more suitable value,
usually the same as the generator’s internal impedance. As you may recall, the maximum power
transfer theorem, which largely applies at Radio Frequency, states that the power delivered to the
load is maximized when RL=RS. Figure 2.9-3 and Figure 2.9-4 demonstrate this result at DC and AC
respectively.
DC Power Transfer
RES
ID=RS
R=50 Ohm 20
Power (mW) 15
RES 10
DCVS ID=R
ID=V1 R=RL Ohm
V=2 V
5
0
0 20 40 60 80 100
Load Resistance ()
RES
AC Power transfer (50Hz)
ID=RS1
R=50 Ohm 10
8
Power (mW)
6
ACVS
RES
ID=V1
Mag=2 V
ID=R1 4
R=RL Ohm
Ang=0 Deg
Offset=0 V
DCVal=0 V 2
0
0 20 40 60 80 100
Load Resistance ()
Additional details about how to setup the simulations shown in Figure 2.9-3 and Figure 2.9-4 are
given in the videos below
95
CHAPTER 2 - Conveying Power at Radio Frequency
A coaxial line usually comes in the form of a coaxial cable with connectors at either side. A
typical cable structure is shown in Figure 2.11-1.
The metallic shield is used as a ground reference while the signal is carried by the centre core.
Figure 2.11-2 Cross-section of a coaxial cable showing inner radius and outer radius
Figure 2.11-2 shows a cross-section of the coaxial cable and clearly indicates the dimensions
which are used to calculate the characteristic impedance of the line, . The grey area between the
conductors is usually some type of dielectric but air would also work. The dielectric constant of
the material interposed between the two conductors does influence the impedance of the line
therefore this quantity is also included in the calculation. This is shown by equation (2.11-1).
( ) (2.11-1)
√
Note that in many textbooks, and indeed in Microwave Office, diameters are used in place of radii
for this calculation. However it is apparent that either may be used without the value of being
affected.
104
2.11 Transmission lines – Design and Practical Realisation
We must also remember that, when a dielectric is present, the velocity of propagation is
also affected and it usually becomes a fraction of the speed of light, . Such a fraction may be
determined by means of the dielectric constant , as shown by equation (2.11-2).
(2.11-2)
√
A reduced speed of propagation of the signal also affects the wavelength as you may recall from
section 2.2 and equation (2.2-4). The effective wavelength for the signal travelling along our
coaxial line therefore becomes
(2.11-3)
√
We will not show the derivation of equation (2.11-1) however this equation is very useful to
understand how the characteristic impedance of the coaxial line is affected by the various
parameters. The first thing to notice is that it is not just or that affects the impedance value, but
their ratio!
Now let us try to reason a little. If we increase the radius of the inner conductor (thereby
decreasing ⁄ ) we will facilitate the flow of electrons by providing a larger cross sectional area. The
resistance to current flow will therefore be lower which translates into a lower .
Also, increasing reduces the distance between the surface of the inner conductor and the inner
surface of the outer conductor. A shorter distance between the facing surfaces of inner and outer
conductors reduces the voltage between them and hence decreases the impedance. Remember
that such a voltage may be expressed as shown in equation (2.11-4) since the electric field19 is radial
as illustrated in Figure 2.11-3. is the charge density per unit length.
( ) ( ) ∫ ( ) (2.11-4)
Figure 2.11-3 Electric field inside a coaxial line. The line is effectively a cylindrical capacitor.
19
The electric field may be easily calculated in this case by using Gauss’s Law:
105
CHAPTER 2 - Conveying Power at Radio Frequency
When the radius of the outer conductor is increased, the opposite happens: the voltage
between inner and outer conductors increases and hence so does the impedance of the line.
Remember that ( ) is a positive and monotonically increasing function when . This condition
is always verified in this case since is always greater then .
There is also another way to look at this. We can think of our coaxial line as a cylindrical
capacitor. Much in same way as in the case of parallel plate capacitors20, as the distance between
the two conductors increases, the capacitance decreases. In the case of a capacitor, a lower
means higher impedance as shown by equation (1.5-6).
Now, as you may recall from Figure 2.3-4, a transmission line may be represented as a
network of capacitors and inductors so, if we reduce the capacitance value, thereby increasing the
impedance of the capacitors, we would expect the impedance of the line to increase also. This ties in
well with equation (2.11-1) shown in graphical form in Figure 2.11-4 where is plotted versus ⁄
at a fixed .
80
60
Z0, ()
40
20
0
1 2 3 4 5 6 7 8 9 10
b/a
Figure 2.11-4 Characteristic impedance of a coaxial line vs the b/a ratio,
Now let us re-examine Figure 2.11-3 and pick up again the analogy with a parallel plate
capacitor. Just as in that geometry, as the dielectric constant of the dielectric between the coaxial
line conductors increases, so does the capacitance . This in turn means that the impedance of the
capacitors which model the line (Figure 2.3-4) decreases and hence so does the characteristic
impedance of the line, . This is yet again confirmed by equation (2.11-1) plotted vs at a
constant ⁄ in Figure 2.11-5.
From a field viewpoint, the dielectric reduces the electric field by the factor i.e.
This means that, as the dielectric constant increases, the Electric field decreases and consequently,
voltage (equation (2.11-4)) and impedance decrease also.
20
For a parallel plate capacitor ⁄ , where is the distance between the plates
106
2.11 Transmission lines – Design and Practical Realisation
70
60
Z0, ()
50
40
30
20
0 2 4 6 8 10 12
r
Core Dielectric
Dielectric OD
Type Impedance diameter diameter Application
Type mm
(mm) (mm)
PF cable television,
RG-6/U 75 1.0 mm (Polyethylene 1.64 4.7 6.86 satellite television and
Foam) cable modems
PE
Radio communication
RG-58/U 50 0.81 mm (Solid 2.3 2.9 5
and amateur radio
Polyethylene)
Radio communication
and amateur radio,
RG-213/U 50 7×0.0296 in Cu PE 2.3 7.2 10.3 EMC test antenna
cables. Typically lower
loss than RG58
Video 2.5 shows how to simulate a coaxial cable with specific dimensions and dielectric
material in Microwave Office. It also shows the effect that changing such dimensions has on the
characteristic impedance of the coaxial transmission line.
107
CHAPTER 2 - Conveying Power at Radio Frequency
Coaxial lines are extensively used at RF and Microwave frequencies since they are ideal
structures to contain the electric and magnetic fields generated at high frequency.
This is because they support the TEM21 mode of propagation which means that they keep Electric
and Magnetic fields perpendicular to one another and to the direction of propagation. The speed of
propagation is of course affected by the dielectric medium used inside the line but, since this
medium is homogeneous, such a speed may be easily calculated by means of equation (2.11-2).
Figure 2.11-6 Electric and Magnetic Field inside a coaxial line, the direction of propagation is perpendicular to the page
The problem with coaxial lines is that they are not very compact, cannot be easily integrated
with printed circuits and are hard to prototype. For example, you could not easily construct a
quarter-wave transformer (section 2.9) by means of a coaxial cable and you would need to use
connectors to join the lines.
An easier way to implement transmission lines with a specific characteristic impedance is to
use microstrip lines. Such lines are very simple to construct and their basic structure is shown in
Figure 2.11-7.
They consist of a large sheet of metal, usually copper, called ground plane, with a slab of dielectric
material on top. On top of the dielectric, a strip of metal is used to create the equivalent of the inner
conductor in a coaxial line which will carry the actual signal.
The obvious problem with this type of line is that both magnetic and electric fields will be partially
contained in the dielectric material and partially in the air surrounding the line. This is shown in
Figure 2.11-8.
21
Transverse ElectroMagnetic
108
2.11 Transmission lines – Design and Practical Realisation
This means that the signal propagates at different speeds above ( ) and below
( ⁄√ ) the signal line and this influences the direction of the fields also. Nevertheless,
if the cross-section geometrical size of the microstrip line is much smaller than the wavelength of the
signal22, we may assume that the signal propagates approximately in a TEM mode. This type of
propagation is called quasi-TEM.
(2.11-5)
√ ⁄
The characteristic impedance of the line may then be calculated by means of equation (2.11-6).
( ) ⁄
√ (2.11-6)
⁄
{ √ [ ⁄ ( ⁄ )]
Also, the velocity of propagation along the line and effective wavelength become
( ) ( ) (2.11-7)
√ √
With microstrip you usually start with a specific characteristic impedance that you aim to
achieve, then pick a substrate with dielectric constant and height and calculate the width
that the line needs to be to achieve the desired impedance. A more useful formula than (2.11-6)
would therefore be one which allows us to calculate given , and . This formula does exist
but it is quite complex!
An easier way to design a microstrip line is to use the TXLine tool included in Microwave Office
shown in figure 2.11-9.
22
As a rule of thumb, in order to prevent higher-order transmission modes affecting the wave propagation,
you should limit the thickness of your microstrip substrate to 10% of a wavelength.
23
In some textbooks two different expressions are given for , when ⁄ and ⁄ however the
difference is usually very small and, due to the empirical nature of the equations, often negligible. We follow
the Pozar approach which only uses one formula for
109
CHAPTER 2 - Conveying Power at Radio Frequency
The reader may now wish to watch some illustrative videos which give further insight into the topics
explained so far in this chapter. A brief of the content of this material is shown below.
VIDEO
CONTENT
REFERENCE
Time domain reflectometry – part 1. This video tutorial shows how to use the
respective simulation file (TDR_video_2_1_and_3_4.emp) to replicate a time
2.1
domain reflectometer. Pulse propagation and reflection in the case of matched,
open-circuited and short-circuited lines are also illustrated.
The video tutorial uses MWO to show the effect that the various parameters of a
2.5
coaxial transmission line have on its characteristic impedance.
The video tutorial uses MWO to show the effect that the various parameters of a
2.6
microstrip transmission line have on its characteristic impedance.
114
3.2 Reflection Coefficient
Notice how the ratio of the total voltage and total current for the impedance is fixed by itself!
The reflection coefficient measured at the load terminals, , which is just the ratio of reflected and
incident voltages at z=0, may be expressed as
( )
( )
( ) (3.2-2)
You may say, “This is all well and good but how can I perform a measurement right at the terminals
of ?”. You would be right, most of the time you can’t do that and there will be a length of line,
albeit short, between your measurement point and your load impedance. So the question is, “If I
don’t measure my and right at the terminals of the load how is the reflection coefficient
affected?”
Well the answer is quite simple. Your reflection coefficient is a complex quantity which has
magnitude and phase. The magnitude will be fixed once you have picked but the phase will
change depending on where you measure. Let us remember that, at any point along the line, is the
ratio of and and hence ( ) may be expressed as
(3.2-3)
( )
You can see that the modulus of the complex number which represents depends solely on A and B
which are fixed for a specific termination. Its phase however depends on the point along your
transmission line where you are measuring your !
Now using (3.1-3) we can rewrite (3.2-3) as
(3.2-4)
( )
The part represents an exponential decay which accounts for resistive losses whereas the
accounts for phase shifts as mentioned in section 2.14.2.2 and in section 1.5.3.
125
CHAPTER 3 - Foundations of RF & Microwave Circuit Characterisation
Remember how in section 2.9 we expressed the physical length of the line in terms of
fractions of the wavelength ? Well, since the reflection coefficient is a complex quantity which
may be plotted on a polar plot, expressing the length of a transmission line in a form which helps
visualising its location on this type of plot may be very useful. We therefore define a measure, ,
termed electrical length, which is expressed in degrees or radians. Its definition is shown in equation
(3.2-5).
(3.2-5)
With this measure, we are effectively scaling the ratio between and by a or 360° angle. We
will see shortly how this can be very useful when using polar plots to represent the reflection
coefficient.
Let’s see what correspondence there is between the physical length of a line (in fractions of
a wavelength) and its electrical length. Let us start with a line long (Figure 3.2-1). From (3.2-5)
we get .
θ = 180°
Figure 3.2-1 Physical and electrical lengths for a line, =50Ω =100Ω
Now remember that the reflection coefficient measures reflections so, if the transmission line
between measurement point and load has an electrical length of or 180°, by the time the incident
signal reaches the load, gets reflected and comes back to the measurement point, it will have
travelled twice the length of the line i.e. or 360°. This is why the electrical length is multiplied
by 2 in the exponent of the exponential in equation (3.2-4)!
This means that an electrical length of 180°, gets the reflection coefficient to rotate by 360° and
hence takes us all the way around the polar plot thereby making travel back to the starting point. It
is just as if the line was invisible! This was also shown in section 2.9 and is further illustrated in Figure
3.2-2 and Figure 3.2-3.
126
3.6 The Smith Chart
+j1
+j0.5 +j2
+j0.2 +j5
0.2
0.5
2
1
5
0.0
-j0.2 -j5
-j0.5 -j2
(a) (b)
-j1
+j0.5 +j2
+j0.2 +j5
0.2
0.5
2
1
0.0
-j0.2 -j5
-j0.5 -j2
27
This is always true for passive networks but not always true for active circuits!
141
CHAPTER 3 - Foundations of RF & Microwave Circuit Characterisation
The impedance chart shown in Figure 3.6-1 (b), is called Smith chart. We will look at how it is
constructed in more detail in section 4.4 but for now let us just look at how it works!
Remember how the complex number which represents any arbitrary impedance may be expressed
in Cartesian form (1.5.3.1) as
The values of and are the values which we can plot on (or read from) the Smith chart.
The circles which intersect the central axis of the chart (Figure 3.6-3) are used to represent the real
part of our normalised impedance, .
+j1
+j0.5 +j2
+j0.2 +j5
0.5
0.2
2
1
0.0
-j0.2 -j5
-j0.5 -j2
-j1
+j1
+j0.5 +j2
+j0.2 +j5
0.5
0.2
2
1
0.0
-j0.2 -j5
-j0.5 -j2
-j1
Figure 3.6-4 Smith Chart peripheral values represent reactance (the imaginary constant j is displayed for clarity)
142
3.6 The Smith Chart
The reader may now wish to watch some illustrative videos which give further insight into the topics
explained so far in this chapter. A brief description of the content of these video tutorials is
provided below.
VIDEO
CONTENT
REFERENCE
This video demonstrates how to set up simulations to measure and display the
reflection coefficient for a simple circuit of the type shown below.
l ,θ
ZIN (l)
3.1
This tutorial illustrates the basic principles of operation of a slotted line and how
to set up a simulation that may emulate it. It also shows the voltage and
3.3
impedance profiles along the length of the line in both a matched and
mismatched case.
Time domain reflectometry – part 2. This video tutorial shows how to use the
respective simulation file (TDR_video_2_1_and_3_4.emp) to replicate a time
3.4 domain reflectometer. Pulse propagation and reflection in the case of a line
mismatched with finite load impedances which are lower ( ) or greater
( ) than the characteristic impedance of the line, are shown.
It is recommended that the readers should try to set up similar simulations to those illustrated in the
video tutorials and explore this type of circuits. This will help them gain further insight and perhaps
find the answers to any doubts or questions that they may have.
145
4.1 Introduction
4 Impedance Matching
4.1 Introduction
In this section we will take a step back first and revisit the maximum power transfer theorem
when passive components characterised by complex (capacitors, inductors) and real (resistors)
impedances (section 1.5.3) are present in the circuit.
As you may recall from Figure 2.9-4, in the case of resistive circuits, maximum power transfer
is achieved when load and source resistances are equal. What happens however when source and
load impedances are complex?
This case is similar to the resistive one, in that the real part of the source and load impedances must
be the same however the sign of their imaginary parts must be opposite. In mathematical terms the
load impedance must be the complex conjugate of the source impedance.
(4.1-1)
Let’s see why this is. Our aim is to convey as much power to our load as we can. Any reactive
elements which the generator sees, will act to store some of the useful energy which could instead
be delivered to the load. If the load and source impedances are equal to the conjugate of one
another, then the reactive elements will be virtually invisible to the generator and all the useful
power will be delivered to the resistive load. This is shown in Figure 4.1-1, where eq. (4.1-1) is
satisfied when .
RS jXS RS
ZS -jXL
ZL
RL
RL
29
Figure 4.1-1 Source impedance driving its complex conjugate and resulting equivalent circuit
Often enough the signal generator has a purely resistive impedance which, in many RF and
Microwave systems is 50Ω The load to which it is connected however is often characterised by a
complex impedance. This could be the input port of a transistor for instance. This is shown in Figure
4.1-2. We may therefore need to add some additional elements between our source and load to
ensure that maximum power transfer is achieved. This is called a matching network and its purpose
is to make the load impedance “look like” the complex conjugate of the source impedance (Figure
4.1-3).
29
The impedance of reactive components is frequency dependent. Here we are assuming a sinusoidal steady
state.
151
CHAPTER 4 - Impedance Matching
CAP
ID=C1
RES
RS = 50 Ω CL = -1j Ω C=1e-6 uF
ID=R1
R=1 Ohm
ACVS
RES
ID=V1
ID=R2
Mag=1 V RL = 50 Ω R=1 Ohm
Ang=0 Deg
Offset=0 V
VS ZL =50-j Ω
DCVal=0 V
Figure 4.1-2 Load impedance (50 – j Ω) does not match source impedance (50 Ω)
IND
ID=L1
L=1 nH
RES RS = 50 Ω LL = 1j Ω CL = -1j Ω
ID=R1
R=50 Ohm
ACVS
RES
ID=V1
ID=R2
Mag=1 V
Ang=0 Deg VS ZL = 50Ω RL = 50 Ω
R=50 O
Offset=0 V Matching
DCVal=0 V
Network
Figure 4.1-3 Matching network transforms load impedance, as seen by source terminals, into source impedance.
When it comes to impedance matching there are three main combinations of load and source
impedances which we may encounter:
Although it is quite easy to just regurgitate established procedures and the respective equations,
throughout this section we will try to explain things from a conceptual point of view, as we have
done throughout this treatment. To this end we will first start with a review of voltages and currents
across parallel and series combinations of resistive and reactive components (RLC networks).
Subsequently we will illustrate the techniques employed to design matching networks for purely
resistive impedances (section 4.3) and then for any arbitrary complex impedances (section 4.4).
152
4.6 S-Parameters, Impedance and Smith Charts in MWO
Pincident
θElectrical length length, θ
Electrical
Let us first consider Figure 4.6-1 which is essentially the same as Figure 4.4-1 where we have
added directional power meters and have specified the length of the line by its electrical length θ
instead of its physical length .
How can we implement this circuit in MWO?
195
CHAPTER 4 - Impedance Matching
The simulation equivalent of the circuit of Figure 4.6-1 is shown in Figure 4.6-2. Corresponding
elements in these two figures are shown by frames of the same type and colour.
First and foremost let us talk about the “PORT” element in Figure 4.6-2. This element
represents a signal generator, with internal impedance specified by the parameter ‘Z’, which is also
able to measure the power that flows out of its terminal and the power that is reflected back into it.
Its circuital equivalent is shown in the dashed frame in Figure 4.6-1. When we have more than one
port, some of the ports may simply behave as an impedance equal to ‘Z’ ohm furnished with power
meters as shown in Figure 4.6-3. The latter are defined as ‘passive’ ports whereas and the former as
‘active’ ports.
Pincident
θElectrical length length, θ
Electrical
The second element if Figure 4.6-2 represents a transmission of line whose length is
specified in terms of its electrical length ‘TLIN’. For this element we need to specify the characteristic
impedance (Z0), which we have chosen to be the same as that of PORT 1, the electrical length (EL) in
degrees (20° in this case) and the frequency F0 which is used to calculate the electrical EL33.
Finally we have the load impedance ZL, which we have implemented with an element called
34
‘IMPED’ . This element allows us to specify explicitly the real and imaginary parts of a complex
impedance.
(EL)
196
4.6 S-Parameters, Impedance and Smith Charts in MWO
Now, if we want to measure the values of the impedance we must first measure ( ) as
seen by PORT 1 (Figure 4.6-4). To measure our reflection coefficient in MWO, we will use an S11
measurement which, as we have seen in section 3.6 is the same as the reflection coefficient for a
one-port network.
We would then normally use equation (4.1-1) to convert this value into an impedance
( ), however MWO allows us to carry out this conversion in several, simpler ways. The easiest
one would entail plotting the reflection coefficient on a Smith Chart and reading the value of the
impedance directly from the chart markings or by using a marker.
Be careful though! This impedance is not ZL! It is the combination of a line, with electrical length EL
and characteristic impedance Z0, and ZL!!
Remember that the impedance seen at PORT 1 may be derived from the phase and
magnitude of the reflection coefficient S11 ( eq.(3.5-3)). Now the magnitude of S11 is determined by
ZL alone and hence is not affected by the length of the line EL (section 3.1), but its phase is. In a
laboratory environment we would have apply a technique called de-embedding (section 3.8) in
order to work out the phase of S11 right at the load terminals and then calculate ZL from the
‘adjusted’ S11. This may of course be done in MWO also, however the simulator also allows us to
probe and measure impedances in many places in our circuit which in the real world would not be
directly accessible. This is very useful!
Now that we have illustrated some basic simulation concepts, we will delegate the teaching
of the remainder of the material on passive circuits to our video tutorials. These will explain in great
details how to analyse and design passive circuits at RF and Microwave frequencies. In particular the
reader will learn how to design various types of matching networks which are essential building
blocks of most high frequency circuits.
197
CHAPTER 4 - Impedance Matching
A brief summary of the content of the training videos which are relevant to this chapter is shown
below.
VIDEO
CONTENT
REFERENCE
Resistors, capacitors and inductors at RF. Time domain Voltage and current
waveforms at high frequency and their relationship to components impedance.
4.1 Comparison of mathematical impedance calculation and impedance values
obtained from voltage and current phase and amplitude relationships.
RC circuits and RL circuits.
198
4.7 Video Tutorials
Transmission line matching. Open and short circuited transmission line stubs.
Comparison between capacitors and open-circuited transmission line stubs.
Comparison between inductors and short-circuited transmission line stubs.
Demonstration of the fact that stubs of any type can behave as both inductors
and capacitors depending on electrical length. Clear demarcation of electrical
4.14
length boundaries.
Matching networks comprising exclusively of distributed transmission line
elements. Using I-Match for distributed matching. Comparison of various types of
matching networks (for identical terminations), in terms of insertion loss, return
loss and frequency profiles, using I-Match.
199
CHAPTER 5 - Amplifier Design
5 Amplifier Design
5.1 Introduction – The transistor at Radio Frequency
In this section we will be looking at the most common methods employed to perform high-
frequency amplifier design. Modern amplifier design is almost invariably carried out with the aid of
simulation tools which employ specific models for the active devices at the heart of the amplifier.
The accuracy of the simulation results is chiefly governed by the accuracy of such models. These may
be obtained in a number of ways but they can all be grouped into two main categories: linear and
non-linear. Linear models may be employed under small signal conditions, when the amplitude of
the signals involved is such that the inherent non-linearities of the active device may be neglected.
These include linear electrical models such as the hybrid-π (section 5.1.1) and S-parameter models
(section 3.7). Non-linear models may also be of circuital nature but they would include non-linear
elements such as diodes for instance. Other non-linear models are based on the actual physics of the
devices and are considerably more complex but often more accurate than electrical models. Other
characterisation methods exist but are beyond the scope of this chapter.
Although this chapter provides some concise reference material in written form, the
majority of the teaching is delegated to the video tutorials listed in section 5.5.
Let us first take a look at the small-signal35 model of a BJT in common emitter configuration.
This type of model is usually referred to as hybrid-π Notice how this electrical model comprises
exclusively of linear elements. From this model it is apparent that amplifier design at high frequency
is a task that would be very difficult to accomplish without a thorough understanding of impedance
matching concepts and of the elements which are utilised to perform it.
r '
b c
r '
bb Cc
B C
I '
B C
B
r ' I
b e rce B
'
Ce E
E E
Figure 5.1-1 Small-signal, hybrid-π model of a BJT
First and foremost this model reminds us of the fact that the bipolar junction transistor is a
current amplifier. We have a small current flowing from the base through to the emitter terminal
( ) and this gets amplified by a factor of thereby allowing a current equal to to flow
35
Small-signal modelling is a common analysis technique in electrical engineering which is used to
approximate the behaviour of nonlinear devices with linear equations. This linearization is performed about
the DC bias point of the device and can be accurate for small excursions about this point.
200
5.1 Introduction – The transistor at Radio Frequency
between collector and emitter terminals (red ovals Figure 5.1-1). The resistance represents the
input resistance of the transistor and it is the resistance which occurs at the base-emitter junction of
the forward biased transistor. Typical values range around 1000 Ω.
The output resistance between the collector and emitter terminals is usually rather high
in value. Since this resistance appears in parallel with the load and with the current generator
supplying the load, we want it to be as high as possible. This way the current through it would be
minimised and we would maximise the load current and hence the power delivered to our load. This
resistance is also responsible for the Early effect i.e. the IV curves sloping up in the active region of
the transistor as shown in Figure 5.1-2.
IC
IB5
IB4
IB3
IB2
IB1
VCE
Figure 5.1-2 I-V Characteristic for a BJT
The other important elements are and which allow feedback from the output to the
input to take place. The most significant of these two is the feedback capacitance , which is
formed at the reverse-biased collector-to-base junction of the transistor. As the frequency of
operation of the transistor increases, the impedance of decreases and thus has a much greater
effect on the transistor operation. A typical value for this component might be 3 pF. The feedback
resistance , which is the resistance appearing from the base to the collector of the transistor, is
a very large (> 5 MΩ). Since the impedance of a parallel can at most be equal to the lowest of the
two, as the frequency increases the impedance of the parallel will be dominated by . Because
thruough this capacitance current can flow back to the input from our current generator instead of
going to the load, may act as to decease the gain.
The other elements in this model are:
- Base spreading resistance ( ) which is an inevitable resistance that occurs at the junction
between the base contact and the semiconductor material that composes the base. Its value
is usually of the order of tens of ohms.
- Emitter diffusion capacitance ( ) which is the sum of the emitter diffusion capacitance and
the emitter junction capacitance, both of which are associated with the physics of the semi-
conductor junction itself.
201
CHAPTER 5 - Amplifier Design
A comprehensive set of videos tutorials on amplifier design is available to the reader. These are
essential to the understanding of amplifier design and form the core of the amplifier section. The
textbook should just be used for a reference and a brief introduction.
VIDEO
CONTENT
REFERENCE
220
5.5 Video Tutorials
Operating Gain Circles Part 1. Writing 2-port S-parameter files (.s2p) and using
them for amplifier design. Amplifier design for a specific gain with a stable
5.8
device. Operating gain circles (GPC_MAX) and maximum available gain (GMAX).
Effects of this technique on input and output VSWR. Using sub-circuits in MWO.
Operating Gain Circles Part 2. Amplifier design for a specific gain with an
5.9
unstable device. Operating gain circles (GPCIR) and maximum stable gain (MSG).
Available Gain Circles. Amplifier design for a specific gain with an unstable device
5.10
using available gain circles (GACIR). MDIF transistor models.
221
References
[1] Gonzalez, G., Microwave transistor amplifiers : analysis and design. 2nd ed. ed. 1997, Upper
Saddle River, N.J. London: Prentice Hall ; Prentice Hall International.
[2] Bowick, C. and P. Elsevier Science, RF circuit design. 1997, Boston: Newnes.
[3] Besser, L. and R. Gilmore, Practical RF circuit design for modern wireless systems. 2003,
Boston, Mass., London: Artech House.
[4] Gilmore, R. and L. Besser, Practical RF circuit design for modern wireless systems. 2003,
Boston, Mass. ; London: Artech House.
[5] Halliday, D., R. Resnick, and K.S. Krane, Physics. 5th ed. ed. 2002, New York ; Chichester:
Wiley.
[6] Pozar, D.M., Microwave engineering. 2nd ed. ed. 1998, New York ; Chichester: Wiley.
[7] Maas, S.A., The RF and microwave circuit design cookbook. 1998, Boston ; London: Artech
House.
222