Vous êtes sur la page 1sur 746

Lecture notes Radio Engineering by Jorma Kekalainen

RADIO ENGINEERING

Jorma Kekalainen

Contents
Electromagnetic radiation
Radio communication
Electrical Noise
Communication link calculations
Microwaves
Maxwells equations
Application of Maxwells equations
Some passive microwave devices
Signal flow graphs and scattering parameters
Introduction to transmission line theory
Smith chart and impedance matching
Waveguide theory
Radio wave propagation in more or less free space
Microwave link design
2

1
Lecture notes Radio Engineering by Jorma Kekalainen

Books
Following references can be useful:
Bennett, Electrical Noise
Freeman, Telecommunication System Engineering
Freeman, Radio System Design for Telecommunications
Forouzan, Data Communications and Networking
Gardiol, Introduction to Microwaves
Goldsmith, Wireless Communications
Krauss, Bostian, Raab, Solid State Radio Engineering
Lee, Statistical Theory of Communication
Ramo, Whinnery, Van Duzer, Fields and Waves in Communication
Electronics
Roddy, Satellite Communications
Skolnik, Introduction to Radar Systems
3

Electromagnetic radiation

Nature of
electromagnetic radiation

2
Lecture notes Radio Engineering by Jorma Kekalainen

Radiation models
Electromagnetic radiation is usually described in one of three
ways:
Ray model
In the classical physics light consisted of rays that could be
reflected and refracted through mirrors and prisms etc.
Although this simple description cannot explain many of the
phenomena we meet in radio and optical communications, ray
model is much-used in telecommunications.
The problem is that when we try to study rays very closely
they start behaving like waves.
In general, the ray model is good when the distances involved
in the device are much larger than the wavelength.
5

Radiation models
Wave model
In the context of telecommunications, the best
way of regarding electromagnetic radiation is
to think of it as an electromagnetic wave.
It is a fair generalization to say that
electromagnetic radiation may be looked on as
a wave in situations where we are studying
transmission or propagation.

3
Lecture notes Radio Engineering by Jorma Kekalainen

Radiation models
Particle model
In many contexts, electromagnetic radiation behaves as
though it consists of tiny particles called photons
(quanta).
Electromagnetic radiation may often be regarded as a particle
when we are studying its interactions with matter.
Note. We are used to thinking of electrons as classical
particles but quantum mechanics assigns to them wave
properties, matter waves, as it assigns to the electromagnetic
waves particle like properties. This is wave particle duality.
The fact is that rays, photons and waves are all useful
analogies that help us to understand what electromagnetic
radiation really is from different viewpoints. 7

Electromagnetic waves
In the electromagnetic wave theory, radiation beam is
represented by electromagnetic waves propagating in space.
An electromagnetic wave consists of two fields
an electric field and
a magnetic field.
Both of these fields have
a direction and
a strength (or amplitude).
These fields oscillate in time and space as the beam
propagates.
Within the electromagnetic wave the two fields are oriented at
90 to one another.
8

4
Lecture notes Radio Engineering by Jorma Kekalainen

Electromagnetic waves
The fields move (by
definition at the speed of
light) in a direction at 90 to
both of them.
In three dimensions, we
could consider the electric
field to be oriented on the x-
axis, and the magnetic field
on the y-axis.
Direction of travelling
wave would then be along
the z-direction.

Travelling wave
In three dimensions, we could consider the electric field to be oriented on
the x-axis, and the magnetic field on the y-axis.
Direction of travelling wave would then be along the z-direction.

The field strength changes periodically.


The rate of oscillation is the frequency of the wave.
The distance travelled during one period of oscillation is the wavelength.
10

5
Lecture notes Radio Engineering by Jorma Kekalainen

Polarization of a vector field


The polarization of any vector field is defined by the
geometrical locus of the tip of the vector, as it varies during
one period.
Hence, the polarization describes the behavior of the field (the
direction and amplitude) as a function of time.
In case of electromagnetic wave the polarization of wave is
defined by means of the electric vector, as it varies during one
period.
In an isotropic medium, the direction of oscillation is always
perpendicular to the direction of propagation.
If the direction of oscillation (in the transverse plane) is
random the wave is said to be unpolarized.
11
Note: For all thermal sources in nature, the direction of vibration is random.

Polarization types
A wave is said to be elliptically polarized if the curve traced by the
end point of the electric field vector is an ellipse (in the xy-plane).
The special cases of the elliptically polarized wave are:
circularly polarized wave
Electromagnetic wave is said to be circularly polarized if the electric
field vector uniformly rotates in the xy-plane.
So, wave is circularly polarized when the tip of the electric field vector
describes a circle.
linearly polarized wave
Electromagnetic wave is said to be linearly polarized if the electric
field vector vibrates in a constant direction in the xy-plane (plane
polarized wave).
So, wave is linearly polarized when the tip of the electric field vector
moves along a straight line.
12

6
Lecture notes Radio Engineering by Jorma Kekalainen

Electromagnetic spectrum

13

Electromagnetic spectrum

14

7
Lecture notes Radio Engineering by Jorma Kekalainen

Wavelength and frequency ranges of


electromagnetic radiation

15

Fractions and multiples

Note: Angstrom is a unit of length used in the study of spectra (spectroscopy). 16

8
Lecture notes Radio Engineering by Jorma Kekalainen

Maxwells equations
Maxwells equations govern the propagation of
electromagnetic waves in various media, including
free space,
waveguides,
optical fibers, and
optical crystals.
All
electric,
magnetic,
electromagnetic, and
optical phenomena
are governed by the same fundamental laws of
electromagnetism.
17

Maxwells equations
These laws are written mathematically in terms
of the Maxwells equations:
D
B 0
B
E 0
t
D
H J
t
18

9
Lecture notes Radio Engineering by Jorma Kekalainen

Maxwells equations
In these equations, the quantities D and B are called the electric
displacement or the electric flux density vector (in coulombs per square
meter) and magnetic induction or the magnetic flux density vector (in
webers per square meter), respectively.
These two vectors (D ,B) include the effect of the electromagnetic field on
matter.
The quantities E and H are the electric field vector (in volts per meter) and
magnetic field vector (in amperes per meter), respectively.
These two field vectors (E , H) are employed to describe an
electromagnetic field or wave.
The quantities and J are the electric charge density (in coulombs per
cubic meter) and electric current density vector (in amperes per square
meter), respectively.
The electric charge and current may be considered the source of the
electromagnetic radiation, represented by the E and H vector fields.
19

Maxwells equations
These four equations completely determine the electromagnetic field and
are the fundamental equations of the theory of such fields, that is, of
electrodynamics.
Maxwells equations cannot be solved uniquely without the known
relationships between
B and H
E and D .
To obtain a unique determination of the field vectors, Maxwells equations
must be supplemented by the so-called material equations,

D E For anisotropic
material e.g.
B H
where the parameters and are are known as the dielectric tensor (or
permittivity tensor) and the permeability tensor, respectively.
20

10
Lecture notes Radio Engineering by Jorma Kekalainen

EM field and matter


When an electromagnetic field is present in matter, the electric field
can perturb the motion of electrons and produce a distribution of
charge separation.
This leads to a dipole polarization per unit volume.

D E 0 E P
where P is the electric polarization and 0 is the dielectric constant of
a vacuum.
Analogously, the magnetic field can also produce a magnetization in
materials.
B H 0 H M
where M is the magnetic polarization and 0 is the magnetic
constant of a vacuum.
21

EM field and matter

Dipole polarization in material

Vacuum

22

11
Lecture notes Radio Engineering by Jorma Kekalainen

Notes
Note. The permittivity of a vacuum is constant
0=8.854 x 10-12 F/m.
The constant 0 is the permeability of a vacuum and
has a value of 0=410-7 H/m.

Note. If the material medium is isotropic, both


material parameters and reduce to scalars.
If the medium is isotropic and homogenous, material
parameters reduce to constants.

23

Material parameters
For most applications, material parameters can be assumed to
be independent of the field strengths.
However, if the fields are sufficiently strong, the dependence
of these quantities on E and H must be considered.
These strong fields produce nonlinear electromagnetic effects,
which may be harmful or useful depending on the application.
The fields can be sufficiently strong, for example, if we focus
an intense laser beam on a thin fiber or apply a strong static
electric field to an electro-optic crystal, or transmit very high
power RF beam through the atmosphere.

24

12
Lecture notes Radio Engineering by Jorma Kekalainen

Sourceless medium and EM waves


In communications, we often deal with the
propagation of electromagnetic radiation in such
regions of space where both charge density and
current density are zero.
In fact, if we set = 0 and J = 0 in Maxwells
equations, we find that nonzero solutions exist.
This means that an electromagnetic field can exist
even in the absence of any charges or currents.
Electromagnetic fields occurring in media in the
absence of charges are called electromagnetic
waves.
25

Electromagnetic radiation

Remote sensing

13
Lecture notes Radio Engineering by Jorma Kekalainen

Remote sensing

27

Remote sensing
Note various paths
Source to sensor direct?
Source to surface to sensor
Sensor can also be source
radar, lidar, sonar (i.e. active remote sensing)
Reflected and emitted components
Several components of final signal captured at sensor

Note: Radar (RAdio Detection And Ranging) is an object detection system which uses radio
waves to determine the range, altitude, direction, or speed of objects.
Lidar (Light Detection And Ranging) is an optical remote sensing technology that can measure
the distance to, or other properties of, targets by illuminating the target with laser light and
analyzing the backscattered light.
Sonar (originally an acronym for SOund Navigation And Ranging) is a technique that uses sound
propagation (usually underwater, as in submarine navigation) to navigate, communicate with or
detect objects on or under the surface of the water. 28

14
Lecture notes Radio Engineering by Jorma Kekalainen

Gathering information about the


environment using EM waves
Electromagnetic radiation is convenient for
information transfer because:
EM radiation travels at speed of light
the fastest possible speed
EM radiation travels along a straight line
EM radiation does not need medium to travel
EM radiation interacts with matter
EM radiation wavelength allows to interact with targets
of various size
EM radiation is everywhere, because every body in
the universe radiates EM energy
29

Electromagnetic environment on
Earth
Mainly determined by solar radiation and properties
of atmosphere
Solar radiation spectrum

30

15
Lecture notes Radio Engineering by Jorma Kekalainen

Atmospheric attenuation

31

Atmospheric windows

Radiowave window
Visible window
(clouds are transparent)

32

16
Lecture notes Radio Engineering by Jorma Kekalainen

Energy transport
Conduction
transfer of molecular kinetic (motion) energy due to contact
heat energy moves from T1 to T2 where T1 > T2
Convection
movement of hot material from one place to another
e.g. Hot air rises
Radiation
results whenever an electrical charge is accelerated
propagates via EM waves, through vacuum & over long distances
hence of interest for remote sensing

33

Electromagnetic radiation: Wave model

Wave model of EM energy James Clerk Maxwell (1831-1879)

Unified theories of electricity and magnetism (via


Newton, Faraday, Kelvin, Ampre)
Oscillating electric charge produces magnetic field (and
vice versa)
Can be described by four differential equations
Calculated speed of EM wave in a vacuum

34

17
Lecture notes Radio Engineering by Jorma Kekalainen

Electromagnetic wave

Electric field (E) perpendicular to magnetic field (M)


Travels at velocity c (3x108 ms-1, in a vacuum)

35

Wave motion
All waves characterized
by:
Wavelength, (m)
Amplitude, a (m)
v

Velocity, v (m/s)
Frequency, f (s-1 or
Hz)
Sometimes period, T
(time for one
oscillation i.e. 1/f)

36

18
Lecture notes Radio Engineering by Jorma Kekalainen

Wave motion
Velocity, frequency and wavelength related by

1

f
proportional to 1/ (constant of proportionality is wave
velocity v

v f

37

Sinusoidals

Note angles in radians (rad)


360 = 2 rad, so 1 rad = 360/2 57.3
Rad to deg. (*180/) and deg. to rad (* /180)
38

19
Lecture notes Radio Engineering by Jorma Kekalainen

Maxwells equations
Four equations relating vector
E
electric (E) and vector magnetic 0
fields (B)
B 0
0 is permittivity of free space
B
E
0 is permeability of free space t
E
B 0 J 0 0
t

1
0
c 0
2

James Clerk Maxwell (1831-1879) 39

Maxwells equations
1. Gauss law for electricity: The electric flux out of any
closed surface is proportional to the total charge E
0
enclosed within the surface
2. Gauss law for magnetism: The net magnetic flux out B 0
of any closed surface is zero (i.e. magnetic monopoles
do not exist)

3. Faradays Law of Induction: Line integral of electric B


field around a closed loop is equal to negative of rate of E
change of magnetic flux through area enclosed by the loop. t

4. Amperes Law: For a static electric field, the line integral


of the magnetic field around a closed loop is proportional to E
B 0 J 00
t
the electric current flowing through the loop.

Note: is divergence operator and x is curl operator. 40

20
Lecture notes Radio Engineering by Jorma Kekalainen

EM spectrum
EM Spectrum
Continuous range of EM radiation
From very short wavelengths (<10-12m)
o high energy
To very long wavelengths (cm, m, km)
o low energy
Energy is related to wavelength (and hence frequency)

41

Electromagnetic spectrum

42

21
Lecture notes Radio Engineering by Jorma Kekalainen

Units
EM wavelength is m, but various prefixes
o cm (10-2m)
o mm (10-3m)
o micron or micrometer, m (10-6m)
o Angstrom, (10-8m)
o nanometer, nm (10-9)
f is cycles/second or Hertz (Hz)

Note: Sometimes it is used wavenumber, k = 1/ [m-1]. 43

Electromagnetic spectrum

Energy radiated from sun or active sensor


Energy 1/wavelength (1/)
shorter (higher f) higher energy
longer (lower f) lower energy
44

22
Lecture notes Radio Engineering by Jorma Kekalainen

EM spectrum

We will see how energy is related to frequency, f (and hence


inversely proportional to wavelength, ).
When radiation passes from one medium to another, speed of
light (c) and change, hence f stays the same.

45

Electromagnetic spectrum: visible


Visible part
from visible blue (shorter )
to visible red (longer )
~0.4 to ~0.7m
Violet: 0.4 - 0.446 m
Blue: 0.446 - 0.500 m
Green: 0.500 - 0.578 m
Yellow: 0.578 - 0.592 m
Orange: 0.592 - 0.620 m
Red: 0.620 - 0.7 m

46

23
Lecture notes Radio Engineering by Jorma Kekalainen

Electromagnetic spectrum: IR
Longer wavelengths (sub-mm)
Lower energy than visible
Arbitrary cutoff
IR regions covers
reflective (shortwave IR SWIR)
and emissive (longwave or thermal
IR TIR)
region just longer than visible
known as near-IR NIR.

47

Electromagnetic spectrum: microwave


Longer wavelengths
RADAR
1 mm to 1 m (300GHz-
300MHz)
various bands used by
RADAR instruments
long so low energy,
hence need to use own
energy source (active
wave)

48

24
Lecture notes Radio Engineering by Jorma Kekalainen

Blackbody
All objects above absolute zero (0 K or -273 C)
radiate EM energy (due to vibration of atoms)
We can use concept of a perfect blackbody
o Absorbs and re-radiates all radiation incident upon it at
maximum possible rate per unit area (Wm-2), at each
wavelength, , for a given temperature T (in K)

Energy from a blackbody?

49

Stefan-Boltzmann Law
Total emitted radiation from a blackbody, M, in Wm-2,
described by Stefan-Boltzmann Law

M T 4
T is temperature of the object in K; and = is Stefan-
Boltzmann constant 5.6697x10-8 Wm-2K-4
So energy T4
o Tsun 6000K M,sun 73.5 MWm-2
o TEarth 300K M , Earth 460 Wm-2

50

25
Lecture notes Radio Engineering by Jorma Kekalainen

Stefan-Boltzmann Law

51

Stefan-Boltzmann Law

Note that peak of suns energy around 0.5 m


o negligible after 4-6m
Peak of Earths radiant energy around 10 m
o negligible before ~ 4m
Total energy in each case is area under curve

52

26
Lecture notes Radio Engineering by Jorma Kekalainen

Generalization of Stefan-Boltzmann Law

Radiation emitted from unit area of any plane surface


with emissivity of (<1) can be written
= Tn
where n is a numerical index
For grey surface where is nearly independent of, n
=4.
When radiation emitted predominantly at < m , n > 4.
When radiation emitted predominantly at > m , n < 4.

53

Peak of emitted radiation: Wiens Law


Wien deduced from thermodynamic principles that
energy per unit wavelength E() is function of T and

f (T )
E (
5
At what m is maximum radiant energy emitted?
Comparing blackbodies at different T, note mT is
constant, k 2897mK i.e. m = k/T
o m, sun 0.48m
o m, Earth 9.66m
54

27
Lecture notes Radio Engineering by Jorma Kekalainen

Wiens (Displacement) Law

Increase
(displacement) in m
as T reduces
Increasing

Straight line in log-


log space

55

Particle model of radiation


Newton proposed wave theory of light (EMR)
in 1666
o Observation of light separating into spectrum

Einstein explained photoelectric effect by


proposing photon theory of light
o Photons individual packets (quanta) of energy
o Photons possess energy and momentum

Light has both wave- and particle-like


properties
o Wave-particle duality

56

28
Lecture notes Radio Engineering by Jorma Kekalainen

Particle model of radiation


EMR intimately related to atomic structure and energy.
Atom: positive (+) charged nucleus (protons +neutrons) &
negative () charged electrons bound in orbits.
o Electron orbits are fixed at certain levels, each level corresponding to
a particular electron energy.
o Change of orbit either requires energy (work done), or releases
energy
o Minimum energy required to move electron up a full energy level.
o Once shifted to higher energy state, atom is excited, and possesses
potential energy.
o Released as electron falls back to lower energy level.

57

Particle model of radiation


As electron falls back, quantum of EMR (photons) emitted.
o Electron energy levels are unevenly spaced and characteristic of a
particular element (basis of spectroscopy)

Bohr and Planck recognized discrete nature of transitions.


Relationship between frequency of radiation (wave theory) of
emitted photon (particle theory)

E hf
E is energy of a quantum in Joules (J); h is Planck constant
(6.626x10-34Js) and f is frequency of radiation

58

29
Lecture notes Radio Engineering by Jorma Kekalainen

Particle model of radiation


If we remember that velocity v = f and in this case v is
actually c, speed of light then
hc
E

Energy of emitted radiation is inversely proportional to
o longer (larger) lower energy
o shorter (smaller) higher energy
Implication for remote sensing: harder to detect longer
radiation (e.g. thermal) as it has lower energy.

59

Particle model of radiation

60

30
Lecture notes Radio Engineering by Jorma Kekalainen

Plancks Law of blackbody radiation


Planck was able to explain energy spectrum of blackbody
Based on quantum theory rather than classical mechanics

2c 2 h 1
E
5 hc
e kT
1
dE()/d gives constant of Wiens Law
E() over all results in Stefan-Boltzmann relation
Blackbody energy function of , and T

61

Plancks Law
Explains/predicts shape of blackbody curve
Use to predict how much energy lies between given
o Crucial for remote sensing

62

31
Lecture notes Radio Engineering by Jorma Kekalainen

Consequences of Plancks Law: Plants


Chlorophyll a, b absorption spectra

Photosynthetic pigments
o Basic driver of nearly all life on
Earth!
o Source of all fossil fuel
63

Applications of Plancks Law


Fractional energy from 0 to i.e. F0 ? Integrate Planck
function.
Note Eb (,T), emissive power of blackbody at , is function of
product T only, so....
Radiant energy from 0 to

E0 , T
F0 , T
T 4

Total radiant energy


for = 0 to =

64

32
Lecture notes Radio Engineering by Jorma Kekalainen

Applications of Plancks Law: example


Q: What fraction of the total power radiated by a black body at 5770 K fall, in
the UV (0 < 0.38m)?
o Need table of integral values of F0
o So, T = 0.38m * 5770K = 2193mK l T (mmK x10 3) F 0 l (l T)
(dimensionless)
o Or 2.193x103 mK i.e. between 2 and 3 2 .067
3 .273
o Interpolate between F0 (2x103) and F0 (3x103) 4 .481
5 .634


F00.38 , T F00.38 2 x103

2.193 2
0.193
6
8
.738
.856


F00.38 3x10 F00.38 2 x10
3

3
32
10
12
.914
.945
14 .963

F00.38 , T 0.067
16 .974
18 .981
0.193 20 .986
0.273 0.067

Finally, F00.38 =0.193*(0.273-0.067)+0.067=0.11


i.e. ~11% of total solar energy lies in UV between 0 and 0.38 m
65

Problem
Show that ~38% of total energy radiated
by the sun lies in the visible region
(0.38m < 0.7m) assuming that
T (mK x103) F0(T)
solar T = 5770K (dimensionless)
2 .067
3 .273
o Hint: We already know F(0.38m), so 4 .481
calculate F(0.7m) and interpolate 5
6
.634
.738
8 .856
10 .914
12 .945
14 .963
16 .974
18 .981
20 .986

66

33
Lecture notes Radio Engineering by Jorma Kekalainen

Deviation from blackbody assumption

67

Recap
Objects can be approximated as blackbodies
o Radiant energy T4
EM spectrum from sun a continuum peaking at ~0.48m
o ~39% energy between 0.38 and 0.7 in visible region

Plancks Law - shape of power spectrum for given T (Wm-2 m-1)


o Integrate over all to get total radiant power emitted by blackbody per
unit area
Stefan-Boltzmann Law M = T4 (Wm-2)
o Differentiate to get Wiens law
Location of max = k/T where k 2898mK

68

34
Lecture notes Radio Engineering by Jorma Kekalainen

Radio communication

Radio communication
system

Invention of radio waves


Hertz proved the existence of the
electromagnetic waves in Karlsruhe Institute
of Technology

Heinrich Hertz: I do not think that the wireless


waves...will have any practical application 70

35
Lecture notes Radio Engineering by Jorma Kekalainen

Electromagnetic spectrum

71

Radio engineering
Radio engineering is a technical science that deals with
the electromagnetic waves at the frequencies below the
infrared frequencies (e.g. from 30 Hz up to 3 THz)
Word radio comes from Latin: radius and/or radiare
(= to radiate)
An oscillator generates high-frequency alternating
current which is radiated as radio waves by the
antennas
Millimeter wave frequencies become more important
all the time.
lack of frequencies, higher data rate, improved technology,
new applications

Note: Radio spectrum is only a tiny fraction of electromagnetic spectrum


near dc or zero frequency. 72

36
Lecture notes Radio Engineering by Jorma Kekalainen

Electromagnetic spectrum

73

Radio transmission

(a) In the VLF, LF, and MF bands, radio waves follow the
curvature of the earth.
(b) In the HF band, they bounce off the ionosphere.

Microwave transmission uses frequencies above


VHF
74

37
Lecture notes Radio Engineering by Jorma Kekalainen

Radio communication system


Transmitters
Frequency stability, power
Antennas
Gain, beam width, sidelobe
level, bandwidth, noise
temperature
Radio channel
Loss, noise, fading, etc.
Receiver
Selectivity, noise
Modulation
Analog, digital

75

Transmitting system

76

38
Lecture notes Radio Engineering by Jorma Kekalainen

Transmitter
Stable transmit frequency: stability 10-9 - 10-10 / day
Phase-locked loop (PLL) used to lock the frequency
Reference by using a stable quartz crystal resonator
oscillator
Modern fast way: direct digital synthesis (DDS) => waveform
created with digital-to-analog converter (DAC)
Simple direct conversion transmitter with PLL
oscillator

77

Transmitter
Low or medium power transmitters use semiconductors
Solid-state transmitters can now provide tens of kW for TV
bands
The highest power transmitters use electron tubes
Such as magnetrons in radars because high peak power is
needed
International broadcasting stations at low frequencies could
transmit even 1 MW but modern radio communication
mostly uses much less transmit power
Portable devices work typically with powers up to 1 - 2 W; short
range devices even with much less power
Transmit power is controlled as needed and best for service,
without excessive power (e.g. cell phone systems)

78

39
Lecture notes Radio Engineering by Jorma Kekalainen

Receiver
Good filters, an accurate LO frequency, low-noise components
LO can be realized similarly as in transmitters
Superheterodyne receiver uses an intermediate frequency
stage - or two at different frequencies
Received signal level may vary a lot => automatic gain
control
Demodulator extracts information from the IF signal
Typical simple radio communication receiver

79

Receiver specifications
Several parameters determine the ability of a receiver to
successfully demodulate a radio signal.
Receiver sensitivity is the minimum input voltage that
produces a specified signal-to-noise ratio at the output of
the IF section.
Dynamic range is the difference between the largest input
signal that will not become distorted and the smallest signal
that can be discerned, and is measured in dB.
Selectivity specifies a receivers ability to discriminate
against adjacent channel signals.
It is a function of the IF strips BPF.
Noise figure indicates how much the receiver degrades the
input signals signal-to-noise ratio.

80

40
Lecture notes Radio Engineering by Jorma Kekalainen

Superheterodyne receiver

81

RF and IF input spectrum

82

41
Lecture notes Radio Engineering by Jorma Kekalainen

Double-conversion receiver

83

Direct-conversion receiver
Frequency conversion direct to baseband for data
demodulation
Attractive for realization in integrated circuits (IC)
Characteristics more limited than in complex receiver
designs

84

42
Lecture notes Radio Engineering by Jorma Kekalainen

Software radio
Progressively faster digital technology allows also
the use of software defined radio
Typical radio device components used only as antenna
and in receiver front-end and then ADC, followed by
digital processing with software
Digital technology and software-defined
operation is used also for transmitters
After software controlled parts and DAC follow radio
devices: up-converter, power amplifier, filter and
antenna

85

Software defined radio


A software defined radio (SDR) is one that can be configured to any
radio or frequency standard through the use of software.
A software defined radio would be able to work on different broadband
networks and would be able to transfer to another network seamlessly
while traveling outside of the users home network.
A software defined radios best advantage is its great flexibility to be
programmed for emerging wireless standards.
It can be dynamically updated with new software without any changes
in hardware and infrastructure.
Software could just download interface automatically.
This is the job of the packet layer, which will split the data into small
packets.

86

43
Lecture notes Radio Engineering by Jorma Kekalainen

Software Defined Radio (SDR)


A conceptual definition of the software defined radio (SDR) is
presented in the following figure

87

Transceiver
Much of radio communication needs a two-way device to
both transmit and receive => transceiver combines these
Basic transceiver device problem: high power TX & low
power RX
How to get necessary isolation when connecting to
common antenna?
If a frequency division duplexing (FDD) system is used => a
duplexer filter separates TX and RX bands (which are different)
If a time division duplexing (TDD) system is used => separate
TX and RX times allow simply a switch to be used only
transmitter or receiver is connected to antenna at any time, not
both simultaneously
For example, cell phone network devices may use both systems

88

44
Lecture notes Radio Engineering by Jorma Kekalainen

Signals and waveforms


Communication systems transmit information in the form of
electrical signals that represent
speech,
music,
television pictures,
data, and so forth.
The waveforms of these signals are complex and continually
changing, but the frequency spectrum of the signals is usually
limited to a specified bandwidth either by the nature of the
signal source or by filters in the transmitting equipment.
Since many of these signals occupy a frequency band that
extends downward to a few hertz (bandbase signals), they
cannot be transmitted in their original form over a common
transmission path because it would not be possible to separate
them at the receiving end.
89

Multiplexing
A separate transmission line or separate radio
path for each signal would not be feasible from
either an economic or a practical standpoint.
Hence the overall communication system must
provide a means for simultaneous transmission
of a number of signals either by shifting them
into different parts of the frequency spectrum
or by sending samples of the signals on a time-
shared basis.

90

45
Lecture notes Radio Engineering by Jorma Kekalainen

Need for higher frequencies


The wavelength () in meters of a radio wave is given
by c/f, in which c is the velocity of light (3 x 108
meters per second in atmosphere), and f is in hertz.
A radio antenna should have a physical size of one-
half wavelength or more for reasonable efficiency.
Hence, as the transmission frequency is increased, the
physical size and cost of the antenna are reduced and
its efficiency increases.

91

Elements of a radio system

A simplified block
diagram of a radio
transmitter and
receiver illustrates
the signal
processing that
takes place in a
radio system.

92

46
Lecture notes Radio Engineering by Jorma Kekalainen

1. Message signal source

The source of the message signal may be


a microphone,
phono pickup,
television camera, or
other device
that transforms the desire information into an
electrical signal.

93

2. Amplifier and 3. Oscillator


The signal is amplified and often passed
through a low-pass filter to limit the
bandwidth.
The RF oscillator establishes the carrier
frequency or some submultiple of it.
Since good frequency stability is required to
keep the transmitter on its assigned frequency,
the oscillator is often controlled by a quartz
crystal.

94

47
Lecture notes Radio Engineering by Jorma Kekalainen

4. RF amplifier and frequency


multiplier
One or more amplifier stages increase the
power level of the signal from the oscillator to
that needed for input to the modulator.
Using frequency multiplication the final carrier
frequency can be a multiple of the oscillator
frequency.

95

5. Modulator
The modulator combines the signal and carrier
frequency components to produce one of the varieties
of modulated waves (AM, ...)
In our simplified system the output signal spectrum
lies in the vicinity of the desired RF carrier frequency.
In many transmitters a second oscillator and mixer
(similar to blocks 10 and 11) are inserted between
blocks 5 and 6 in order to shift the modulated wave to
a higher frequency range.

96
Note: The carrier may be a sinucoidal wave or a train of pulses.

48
Lecture notes Radio Engineering by Jorma Kekalainen

6. RF amplifier and 7. Transmitting


antenna
Additional amplification may be required after
modulation to bring the power level of the signal to
the desired value for input to the antenna.
The transmitting antenna converts the RF energy into
an electromagnetic wave of the desired polarization.
If a single (fixed) receiver is to be reached, the
antenna is designed to direct as much of the radiated
energy as possible toward the receiving antenna.
For broadcasting service suitable transmitting antenna
is omnidirectional.

97

8. Receiving antenna
The receiving antenna may be omnidirectional for
general service or highly directional for point-to point
communication.
The wave propagated from the transmitter induces a
small voltage in the receiving antenna.
The range of amplitudes of the induced antenna
voltage may be from tens of millivolts to less than 1
microvolt, depending upon a wide variety of
conditions.

98

49
Lecture notes Radio Engineering by Jorma Kekalainen

9. RF amplifier
The RF amplifier stage increases the signal power to
a level suitable for input to the mixer and it helps to
isolate the local oscillator from the antenna.
This stage does not have a high degree of frequency
selectivity but does serve to reject signals at
frequencies far from the desired channel.
The increase in signal power level prior to mixing is
desirable because of the noise that is inevitably
introduced in the mixer stage.

99

10. Local oscillator and 11. Mixer


The local oscillator in the receiver is tuned to produce
a frequency fLO that differs from the incoming signal
frequency fRF by the intermediate frequency fIF; that
is, fLO can be equal to fRF + fIF or fRF - fIF .
The mixer is a nonlinear device that shifts the
received signal at fRF to the intermediate frequency
fIF.
Modulation on the received carrier is also
transformed to the intermediate frequency.

100

50
Lecture notes Radio Engineering by Jorma Kekalainen

12. IF amplifier
The IF amplifier increases the signal to a level
suitable for detection and provides most of the
frequency selectivity necessary to pass the desired
signal and filter out the undesired signals that are
found in the mixer output.
Because the tuned circuits in blocks 11 and 12 always
operate at a fixed frequency (fIF), they can be
designed to provide good selectivity.
Ceramic or crystal filters are often used.

101

13.-15. Detector, Audio/video


amplifier and Output device
The detector recovers the original message signal
from the modulated IF input.
The audio or video amplifier increases the power
level of the detector output to a value suitable for
driving a loudspeaker or other output device.
The output device converts the signal information
back to its original form (sound waves, picture, etc.).

102

51
Lecture notes Radio Engineering by Jorma Kekalainen

Radio transmitter and receiver block


diagram
Note: In practice, there
are so many variations
in transmitter and
receiver systems that no
single block diagram
could even be
considered typical.

103

Note: Electrical noise


In addition to the desired signal that is
processed by the receiver, electrical noise is
added in the transmission path, and is
generated within
the RF amplifier,
local oscillator,
mixer, and so forth.

104

52
Lecture notes Radio Engineering by Jorma Kekalainen

Radio communication

Radio transmission

Power gain
Let it be an linear time-invariant (LTI) system whose
input signal has average power Pin.

If the system is distortionless, the average signal


power at the output will be proportional to Pin.
Thus, the systems power gain is

106

53
Lecture notes Radio Engineering by Jorma Kekalainen

Power gain in decibels

Systems that include amplification may have


very large values of g, so well find it
convenient to express power gain in decibels
(dB) defined as

107

Power gain in decibels


Since the decibel is a logarithmic unit, it converts
powers of 10 to products of 10.
For instance, g = 10e becomes gdB = e x 10 dB.
Power gain is always positive, of course, but
negative dB values occur when g < 1.0 = 100 and
hence gdB < 0 dB.
Note that 0 dB corresponds to unity gain (g = 1).
Given a value in dB, the antilogarithmic value is

108

54
Lecture notes Radio Engineering by Jorma Kekalainen

Absolute units: dBW and dBm

While decibels represent power ratios, signal


power itself may be expressed in dB if you
divide P by one watt or one milliwatt, as
follows:

109

Power gain in dB-units


Rewriting the definition equation of power
gain as (Pout/1 mW) = g(Pin/1 mW) and taking
the logarithm of both sides then yields the dB
equation

110

55
Lecture notes Radio Engineering by Jorma Kekalainen

Transfer function
Now consider a system described by its transfer
function H(f).
A sinusoidal input with amplitude Ax produces the
output amplitude Py
Ay = |H(f)|Ax,
and the normalized signal powers are
Px= Ax 2/2 and Py = Ay2/2= |H[f)|2Px.
If the system has the same impedance level at input
and output
Py/ Px = Pout/ Pin = |H[f)|2=g
111

Square of amplitude function


|H[f)|2
In any case |H[f)|2 tells us how the power gain
varies as a function of frequency.
For a useful measure of frequency dependence
in terms of signal power we take

This represents the relative gain in dB.

112

56
Lecture notes Radio Engineering by Jorma Kekalainen

Problems

(a) Verify that dBm = dBW + 30 dB.


(b) Show that if |H(f)|dB = -3 dB then
|H(f)|1/2.

113

Solution
(a) Verify that dBm = dBW + 30 dB.

114

57
Lecture notes Radio Engineering by Jorma Kekalainen

Solution

(b) Show that if |H(f)|dB = -3 dB then


|H(f)|1/2.

115

Transmission loss
In addition to any signal distortion, a
transmission system also reduces the power
level or strength of the output signal.
This signal-strength reduction is expressed in
terms of transmission power loss.
Although transmission loss can be
compensated by power amplification, the ever-
present electrical noise may prevent successful
signal recovery in the case of large
transmission loss.
116

58
Lecture notes Radio Engineering by Jorma Kekalainen

Transmission loss

Any passive transmission medium has power


loss rather than gain, since Pout < Pin
We therefore prefer to work with the
transmission loss or attenuation

Hence
117

Attenuation coefficient
In the case of transmission lines, coaxial cables,
and waveguides, the output power decreases
exponentially with distance.
Well write this relation in the form

where l is the path length between source and


destination and is the attenuation coefficient in
dB per unit length.
Then

118

59
Lecture notes Radio Engineering by Jorma Kekalainen

Example

A 40-km cable system has Pin 2 W and a


repeater with 64-dB gain is inserted 24 km
from the input.
The cable sections have = 2.5dB/km.
Use dB equations to find the signal power at:
(a) the repeaters input;
(b) the final output.

119

Solution
Use dB equations to find the signal power at:
(a) the repeaters input;
(b) the final output.

120

60
Lecture notes Radio Engineering by Jorma Kekalainen

Radio transmission
Signal transmission by radio wave propagation can
reduce the required number of cable repeaters, and has
the additional advantage of eliminating long cables.
Here we examine the transmission loss for line-of-
sight (LOS) propagation (illustrated in the following
Fig.) where the radio wave travels a direct path from
transmitting to receiving antenna.
This propagation mode is commonly employed for
long-distance communication at frequencies above
100 MHz.

121

Line-of-sight (LOS) radio transmission

122

61
Lecture notes Radio Engineering by Jorma Kekalainen

Free-space loss
The free-space loss on a line-of-sight path is due to
spherical dispersion of the radio wave.
This loss is given by

in which is the wavelength, the signal frequency, l


the distance (path length), and c the speed of light.
If we express l in kilometers and f in gigahertz (109 Hz)
then

123

Free-space loss(dB) log(path


length)
We see that LdB increases as the logarithm of the path
length.
Thus, for instance, doubling the path length increases
the loss by only 6 dB. (Prove this as homework!)
We also see that LdB increases as the logarithm of the
frequency.
Thus, for instance, doubling the frequency increases
the loss by 6 dB.

Note: Here LdB is the free-space line-of-sight attenuation and doubling the path length
can violate this free-space line-of-sight assumption. 124

62
Lecture notes Radio Engineering by Jorma Kekalainen

Antenna gain

Furthermore, directional antennas have a


focusing effect that acts like amplification in
the sense that

where gT and gR represent the antenna gains at


the transmitter and receiver.

125

Effective aperture area


The maximum transmitting or receiving gain of an antenna
with effective aperture area Ae is

where c 3 x 105 km/s.


The value of Ae for a horn or dish antenna approximately
equals its physical area.
More accurately, the value of Ae for a circular dish antenna
equals its physical area (DA/2)2 multiplied by the aperture
efficiency 0.5 0.70.
Large parabolic dishes may provide gains in excess of 60 dB.

126

63
Lecture notes Radio Engineering by Jorma Kekalainen

Example
Satellite relay system shows a transoceanic radio system with a satellite
relay serving as a repeater.
The satellite is about 40000 km from either ground station, and the signal
frequency is 6 GHz.
The satellite has a repeater amplifier with gamp = 80 dB, and the input
power at the ground station is Pin = 100 W .

127

Solution
Free space loss equation gives the uplink and
downlink path loss

Since we have all the data in dB, we can


compute Pout, by adding gains and subtracting
losses in the following tabulation:

128

64
Lecture notes Radio Engineering by Jorma Kekalainen

Solution

Note: Very small power levels are typical for satellite systems. 129

Ground reflection
Often, there are multiple
signals arriving to the
receiver due to the
reflections from the
ground, water, building
etc.
Multi-path propagation
For the horizontal polarization the
Very simple case: direct reflection coefficient is -1.
Expression r1+r2-r0 2h1h2/d
wave and one reflected
wave. =2/

130

65
Lecture notes Radio Engineering by Jorma Kekalainen

Destructive reflection case


In certain situation, the phase difference between the
direct and reflected wave is 180 degrees and thus the
waves cancel each other out in the receiver.
The total electric field strength is 0.
Condition for that is

Obviously, there is also possible that both waves are


summed together constructively as is the case in the
next example.
131

Example
There is a lake between a transmitter and a receiver (distance
d).
The height of the transmitter antenna is H (H<<d).
Calculate the optimum height of the receiver antenna when the
frequency is f.
The polarization is horizontal and the lake is an ideal reflector.

132

66
Lecture notes Radio Engineering by Jorma Kekalainen

Solution
There is a lake between a transmitter and a receiver (distance d).
The height of the transmitter antenna is H (H<<d).
Calculate the optimum height of the receiver antenna when the frequency is
f.
The polarization is horizontal and the lake is an ideal reflector.
The total electric field at the antenna site is

=2/

The maximum field strength is obtained when

Obviously, the smallest height (n=0) is the optimum height.


So
133

Numerical version
There is a lake between a television transmitter
and a receiver (distance 5 km).
The height of the transmitter antenna is 50 m.
Calculate the optimum height of the receiver
antenna when the frequency is 500 MHz.
The polarization is horizontal and the lake is
an ideal reflector.
The optimum height is
h=15 m
134

67
Lecture notes Radio Engineering by Jorma Kekalainen

Multi-path propagation
The waves reflect and refract from obstacles and
they arrive to the receiver along different paths.
In the antenna the waves interfere and thus in the
receiver only one signal appear.
The phase difference between the arriving signals affect the
total received signal
When the receiver moves, (e.g. in mobile
communication), the received signal changes
significantly as a function of the distance.
Fast signal fading takes place within the distance of one
wavelength.
135

Antenna parameters
Radiation pattern (Directional pattern )
The directional dependence
of the power density radiated
or received by the antenna

Directivity D

136

68
Lecture notes Radio Engineering by Jorma Kekalainen

Antenna parameters
Radiation efficiency rad

Gain G

Effective area Aeff

Area efficiency coefficient for aperture antennas

137

Antenna parameters
Input impedance Zant
Zant(f) = Rrad(f)+ Rloss(f) + j X(f)
When the input impedance Zant of the antenna
differs from the characteristic impedance of the
feed line Z0, part of the power is reflected

Polarization
Is defined as the direction of radiated/received
electric field vector of the antenna.

138

69
Lecture notes Radio Engineering by Jorma Kekalainen

Low-frequency antennas

f< SHF
Wire antennas, microstrip antennas

Solid bow tie


139

High-frequency antennas

0.1 mm (3 THz)
Reflectors, horns, dielectric antennas

140

70
Lecture notes Radio Engineering by Jorma Kekalainen

Friis formula

Received power

Note:
141

Example: Link budget


In a 58-GHz link radio, the gain of antennas is
34 dBi, the transmitter power is 5 dBm and the
sensitivity of the receiver is -73 dBm. What is
the largest distance that can be implemented
with this link radio?

142

71
Lecture notes Radio Engineering by Jorma Kekalainen

Solution

The gain of the antennas

The transmitted and


received powers

The wavelength

The largest distance r can be


solved from Friis formula

143

Non-idealities
It is impossible to implement this theoretical distance in
practice, because there are many non-idealities, so the distance
is significantly shorter due to the following factors:
The attenuation in the air (moisture, rain)
For instance, at 58 GHz the attenuation of the air is significant due to
the resonance mode of oxygen molecules at 60 GHz
Obstacles (buildings, forest, hilly terrain etc.) between the
antennas
Reflections from the ground and other obstacles
Polarization mismatching
Impedance mismatching
Pointing error of the antennas
144

72
Lecture notes Radio Engineering by Jorma Kekalainen

Propagation mechanisms of radio


waves
Line-of-sight propagation
Ionospheric propagation
Ground propagation (surface wave)
Reflection (refraction, diffraction)
Atmospheric scattering

145

Propagation mechanisms of radio


waves
Line-of-sight propagation
The most common and
important mechanism
Scattering from the in-
homogeneity of the
atmosphere
communication far beyond
the radio horizon
Reflection from the
ionosphere
Ground wave
146

73
Lecture notes Radio Engineering by Jorma Kekalainen

1st Fresnel zone

147

Problem
There is a 45-km radio link from a mainland to
an island. The operation frequency is 6 GHz.
What is the required heights of the link
antennas when the first Fresnel ellipsoid is free
of obstacles?

148

74
Lecture notes Radio Engineering by Jorma Kekalainen

Problem
r0=45000m, f=6GHz, K = 4/3, R=6375km

149

Solution
Radius of the 1st Fesnel zone

Radius in middle ( r 1=r2r0/2,


because hF<<r0 ) is

Applying Pythagoras to Fig.


and solving the equation

we get the antenna heights

150

75
Lecture notes Radio Engineering by Jorma Kekalainen

Atmospheric scattering
Inhomogeneity in the atmosphere cause scattering
We can think that scatterer operates as a isotropic antenna
which radiates power uniformly in all directions, taking its
power from passing by EM wave.

Scattering is always present in the atmosphere but its


attenuation effect for line-of-sight communication
is infinitesimally small in clear atmosphere and minor
in heavy rain up to 10 GHz.
151

Atmospheric scattering
Sometimes this weak
scattering effect can be
exploited.
In that case the radio wave
propagates behind an
obstacle or far beyond the
horizon where it would not
propagate otherwise.
In weather radars, the
scattering signals from
hydrometeors (raindrops,
fog, hails or snow flakes) are
measured. 152

76
Lecture notes Radio Engineering by Jorma Kekalainen

Summary
A radio communication system is an economic compromise
between information transfer capacity needs, available
frequency bands, propagation and equipment characteristics
(propagation attenuation, size of the antenna ).

Low frequencies (long wavelength)


Small propagation attenuation
Large-sized antennas
Low data rate (a small number of different frequency bands)
High frequencies (short wavelength)
Higher propagation attenuation (and shadow areas caused by terrain)
Small-sized antennas
High data rate
153

Examples
FM radio waves propagate very well because the frequency is
small (about 100 MHz), but the required antenna is large-
sized.
GSM 900 (900 MHz) operate well in rural areas where the
waves can propagate far
GSM 1800 (1800 MHz) operate better in urban areas where
the base stations are close (cell size is small) and a number of
frequency bands can be larger than at 900 MHz.

154

77
Lecture notes Radio Engineering by Jorma Kekalainen

Radio communication

Fundamental limitations

Fundamental limitations
There are two general kinds of constraints when designing a
communication system.
On the one hand are the technological problems, including
hardware availability,
economic factors,
federal regulations, and so on.
These are problems can be solved in theory, even though perfect
solutions may not be practical.
On the other hand are the fundamental physical limitations dictating
by the laws of nature.
These limitations ultimately dictate what can or cannot be
accomplished, irrespective of the technological problems.
The fundamental limitations of information transmission by
electrical means are bandwidth and noise.

156

78
Lecture notes Radio Engineering by Jorma Kekalainen

Bandwidth (B)
The concept of bandwidth applies to both signals and
systems as a measure of speed.
When a signal changes rapidly with time, its frequency
content, or spectrum, extends over a wide range and
we say that the signal has a large bandwidth.
Similarly, the ability of a system to follow signal
variations is reflected in its usable frequency response
or transmission bandwidth.
All electrical systems contain energy-storage elements,
and stored energy cannot be changed instantaneously.
Consequently, every communication system has a finite
bandwidth B that limits the rate of signal variations.

157

Transmission bandwidth (BT)


Communication under real-time conditions requires
sufficient transmission bandwidth to accommodate the
signal spectrum.
Otherwise, severe distortion will result.
Thus, for example, a bandwidth of several megahertz is
needed for a video signal, while the much slower variations
of a voice signal fit into BT 3 kHz.
For a digital signal with r symbols per second, the
bandwidth must be BT r/2.
In the case of information transmission without a real-time
constraint, the available transmission bandwidth determines
the maximum signal speed.
The time required to transmit a given amount of information
is therefore proportional to 1/BT.
158

79
Lecture notes Radio Engineering by Jorma Kekalainen

Noise
Noise imposes a second limitation on information
transmission.
Why is noise unavoidable?
According to kinetic theory at any temperature above
absolute zero, thermal energy causes microscopic
particles to exhibit random motion.
In electrical communication systems the random
motion of charged particles such as electrons generates
random currents or voltages called thermal noise.
There are also other types of noise, but thermal noise
appears in every communication system.

159

SNR
We measure noise relative to an information signal in
terms of the signal-to-noise power ratio (SNR or S/N).
Thermal noise power is ordinarily quite small, and S/N
can be so large that the noise goes unnoticed.
At lower values of S/N, however, noise degrades
fidelity in analog communication and produces bit
errors in digital communication.
These problems become most severe on long-distance
links when the transmission loss reduces the received
signal power down to the noise level.
Amplification at the receiver is then to no use, because
the noise will be amplified along with the signal.

160

80
Lecture notes Radio Engineering by Jorma Kekalainen

Hartley-Shannon law
Taking both limitations into account, Shannon
(1948) stated that the rate of information
transmission cannot exceed the channel capacity
C=Blog2(1 +S/N)
with the zero-error probability (pe0).
This relationship, known as the Hartley-Shannon
law, sets an upper limit on the (error-free)
performance of a communication system with a
given bandwidth and signal-to-noise ratio.
161

Ideal communication systems


Every realistic communication system is subject to some
power and bandwidth limitations.
If we exclude radio transmission with fading, careful
system design can largely eliminate all other
contaminations except unavoidable thermal noise.
The AWGN (Additive White Gaussian Noise) channel
serves as a reasonable model under these conditions, and
the Hartley-Shannon law gives the maximum rate for
reliable communication.
Therefore, we here define
An ideal communication system is one that achieve
nearly error-free information transmission at a rate
approaching R=Blog(1+S/N)
162

81
Lecture notes Radio Engineering by Jorma Kekalainen

Electrical Noise

Introduction

Noise and interference


Unwanted electric signals come from a variety of sources,
generally classified as human interference or naturally
occurring noise.
Human interference comes from other communication
systems, ignition and commutator sparking, 50 or 60 cycle
hum, and so forth; natural noise-producing phenomena
include atmospheric disturbances, extraterrestrial radiation,
and circuit noise.
By careful engineering, the effects of many unwanted
signals can be reduced or eliminated completely.
But there always remain certain inescapable random
signals, that present a fundamental limit to systems
performance.

164

82
Lecture notes Radio Engineering by Jorma Kekalainen

Noise is always here!


Noise is always present in communications systems, but under
normal operating conditions it goes unnoticed because the
signal levels are much higher than the noise levels.
A weak signal accompanied by noise can be amplified if the
associated noise level is low relative to the signal level.
However, if the noise level is close to the signal level,
amplification will be useless because any amplifier will
amplify both the incoming signal and the incoming noise as
well as adding more noise of its own.
This process is evident in any receiver when an incoming
signal fades into the noise or the external noise level rises to
the point that it drowns out the signal.

165

Noise limits performance


Transmission loss usually results in a very weak
signal at the input of a communication receiver.
Consequently, to obtain an adequate signal level
for further processing, the front end of a typical
receiver includes several stages of amplification.
High-gain amplifiers amplify any noise that
accompanies the received signal, and they also
add their own internally generated noise.
An accurate assessment of system performance
must therefore take account of the amplifier noise.
166

83
Lecture notes Radio Engineering by Jorma Kekalainen

Circuit and device noise

Noise generated within an electrical circuit


may come from numerous sources involving
several different physical phenomena.
Here well describe major types of device
noise found in communication systems, and
well develop appropriate circuit models and
analysis methods.

167

Signal and noise


To simplify the mathematical details, the
desired signal is assumed to be a sinusoid or a
group of sinusoids that comprise the
transmitted information.
This may be called a deterministic signal.
Noise is defined as any extraneous electrical
disturbance tending to interfere with the
normal reception of the transmitted signal.

168

84
Lecture notes Radio Engineering by Jorma Kekalainen

Random noise
Noise can consist of signals from unwanted
sources plus random fluctuations of voltages
and currents caused by physical phenomena.
There are various types of random noise
thermal noise,
shot noise, and
flicker noise

169

Signal power/Noise power


One of the goals of communications system design is to keep the ratio of
average (or peak) signal power to average noise power so large that the
noise has no harmful effects on system performance.
Techniques for doing this include
(1) using powerful transmitters and high-gain antennas to develop
strong signals at the receiver,
(2) designing amplifier and mixer circuits so that they introduce a
minimum amount of additional noise when processing signals, and
(3) using modulation or coding schemes that facilitate the separation of
signaIs from noise.
In the case of man made noise sources (e.g., automobile ignitions) there
exists the fourth option of suppressing the noise at its source by filtering,
bypassing, or redesign.
Attention is usually given to each of these four options, and the mix
selected is determined by such factors as cost, weight, and efficiency.

170

85
Lecture notes Radio Engineering by Jorma Kekalainen

Stationarity

The assumption of stationarity is reasonable


for most electrical noise processes, thereby
allowing us to represent noise sources in the
frequency domain.
For convenience, well adopt the common
practice of working entirely with positive
frequency and one-sided frequency functions
as distinguished from the two-sided functions.

171

Example: 2-sided 1-sided spectrum


function
To clarify this distinction, let G(f) be the two-sided available
power spectrum of some noise source.
Since G(f) has even symmetry, the corresponding one-sided
available power density will be defined by

p f pa f 2G f , f 0
and the total available noise power is


N pa f df
0

Likewise, the one-sided mean square voltage and current densities are equal twice the
172
two-sided frequency functions for f0.

86
Lecture notes Radio Engineering by Jorma Kekalainen

Stochastic nature of noise


It is desirable to characterize electrical noise as accurately as possible.
A common characteristic of most types of noise, however, is its nondeterministic
nature: that is,
the exact waveform of the noise cannot be predicted.
A measure of the amount of electrical noise can be obtained by connecting a meter
across a noise source to measure the
average,
peak,
rectified-average, or
rms voltage (or current).
Relationships between these quantities are different for different types of noise, that
is, the average value may be zero whereas the others are not.
The rms voltage (or current) can be used to calculate the average noise power
delivered to a resistive load.
As will be shown, the measured value depends upon the spectrum of the noise
source and the frequency response of the measuring instrument.

173

Power spectral density


The frequency-domain characterization of noise can
be given by means of a power spectral density curve
[the units are watts per hertz (W/Hz)].
The simplest kind of noise to work with would be one
whose spectral density is flat over the frequency
range of interest.
A common example of noise with a flat frequency
spectrum is thermal noise associated with Brownian
motion of elections in a conductor.

174

87
Lecture notes Radio Engineering by Jorma Kekalainen

Thermal noise

One unavoidable cause of electrical noise is


the thermal motion of electrons in conducting
mediawires, resistors, etc.
As long as communication systems are
constructed from such material, this thermal
noise will be with us.

175

Thermal noise
Available noise power of thermal noise in a 1-Hz bandwidth is
given by
p(f) = kT,
where k is Boltzmanns constant and T is the temperature of
the noise source in kelvin units.
This expression applies from direct current (dc) to high
microwave frequencies.
If the bandwidth were allowed to extend to infinity, the
available power of a thermal noise would appear to be
unlimited also.
However, results from quantum mechanics indicate that
physical thermal noise sources, although extremely wide band,
have a power spectrum that drops to zero at arbitrarily high
frequencies.
176

88
Lecture notes Radio Engineering by Jorma Kekalainen

Equivalent thermal noise source


Thermal noise is both the most common type of noise
encountered in radio circuits and the easiest to
analyze.
Other types of noise will be represented by equivalent
thermal noise sources.
Although the common emphasis will be on those
aspects of noise necessary to the design of
communications system that operate under average
conditions, much more can be done with the theory of
noise in terms of developing probability distributions
for system outage time, and etc.

177

Thermal noise in resistors and


networks
As the name implies, thermal noise is due to the random
motion of charge carriers in any conducting medium whose
temperature is above absolute zero.
The velocity of this motion increases with temperature in such
a way that the electrical noise power density produced is
proportional to the resistance of the conductor and to its
absolute temperature, hence the name thermal noise.
It is also called white noise because it has been shown both
theoretically and experimentally to have a uniform spectrum
up to frequencies on the order of 1013 (just as white light is
composed of all colors of the visible spectrum).

178

89
Lecture notes Radio Engineering by Jorma Kekalainen

Noise equivalent circuit


A metallic resistor may be considered a thermal noise
source that can be presented by either of the noise
equivalent circuits shown below.

179

The mean-square values


The mean-square noise
voltage (Vn2) and current
(In2) are given by the
following expressions in
which R is the resistance, G
= 1/ R the conductance, T
the temperature of the
resistor in kelvin units, k
Boltzmanns constant
(1.38*10-23J/K), and B the
bandwidth in hertz in which
the noise is observed.

180

90
Lecture notes Radio Engineering by Jorma Kekalainen

PN B
The noise power that is transmitted through a
circuit is proportional to the circuit bandwidth.
Consequently, the circuit bandwidth should
never be greater than that necessary to transmit
the desired signal if the maximum output
signal-to-noise ratio (SNR) is to be achieved.

181

Example
Calculate the mean-square noise voltage
produced in a 100-k resistor in a bandwidth
of 106 Hz at room temperature (T=20C = 293
K).
Vn2=4kTRB

Note: The rms voltage is (Vn2)40.3 V 182

91
Lecture notes Radio Engineering by Jorma Kekalainen

Comment

If this 100-k resistor were in the input circuit


of an electronic voltmeter that had a bandwidth
of 1 MHz, no amount of gain built into the
voltmeter would enable it to measure signals
below 1 millivolt (mV) with accuracy.

183

Frequency domain models of a noise


source

Thvenin circuit Norton circuit

184

92
Lecture notes Radio Engineering by Jorma Kekalainen

Thvenin equivalent circuit


The previous Figure shows the frequency-domain Thvenin
circuit model of a stationary but otherwise arbitrary noise
source with noiseless internal impedance Z(f).
The function vn2(f) represents the open-circuit mean square
voltage density, defined such that

Here p(f) [=pa(f)] is one-sided (f0) available power density

185

Norton equivalent circuit


Converting Thevenin source to a Norton
equivalent circuit gives the source model with

which represents the short-circuit mean square


current density.
This equation also expresses Ohms law in the
form needed for circuit noise analysis.

186

93
Lecture notes Radio Engineering by Jorma Kekalainen

Example
Suppose the source in question happens to be a
thermal resistance R at temperature T.
Then the one-sided available power density
p f kT
and setting Z(f) = R yields
4kT
vn2 f 4 RkT and in2 f
R
These constant densities correspond to white
noise, at least up to infrared frequencies.
187

Thevenin equivalent circuit for noise


calculation
A resistive network driven by a voltage source
Circuits containing more
than one resistor may be
analyzed by reducing
them to one (Thvenin)
equivalent resistance
and applying
Thevenin equivalent circuit for noise computation
Vn2=4kTRTB , RTRTh

to obtain the mean-


square noise voltage..

188

94
Lecture notes Radio Engineering by Jorma Kekalainen

Note
The noise Thvenin equivalent is different from the signal Thevenin equivalent
of the same circuit, where a signal source is driving a hypothetical noiseless
load resistor Ri (representing the input of an amplifier) through three noisy
resistors R1, R2, and R3.
Thevenin equivalent circuit for signal computation

The signal voltage at the load is found by the conventional voltage division, but
the noise voltage is found from a circuit in which the noise source is related to
the Thevenin resistance of the resistive network rather than the voltage divider
ratio.

189

Note
The process of combining the resistors in a network to obtain
an effective noise resistance is equivalent to combining the
mean-square noise voltages of the resistors themselves to
obtain the net mean-square noise voltage.
This is consistent with the statistical principle that if two or
more independent random processes are combined, the mean-
square value of the resultant is obtained by adding the mean-
square value of each process.
Thus the mean-square noise voltage of a group of resistors
connected in series is the sum of the mean-square noise
voltage across each resistor.

190

95
Lecture notes Radio Engineering by Jorma Kekalainen

Effect of reactive elements


Resistance noise occurs in almost all circuits, but
reactive elements may alter the frequency density.
In particular, let Figure be a one-port (two-terminal)
network containing only resistance, capacitance, and
inductance, and having the equivalent impedance Z(f)
= R(f) + jX(f).

191

Nyquists formula
When the resistances are in thermal equilibrium at
temperature T, Nyquists formula states that

vn2 f 4 R f kT
Hence, the mean square voltage density takes the
shape of the equivalent resistance R(f).
Nyquists equation includes the special case of an
all-resistive network whose equivalent resistance
will be independent of f.

192

96
Lecture notes Radio Engineering by Jorma Kekalainen

Nyquists formula proof

We prove Nyquists formula by considering


the complete circuit in Fig. below, where the
source has been connected to a load resistance
RL at the same temperature T.

193

Nyquists formula proof


This load resistance, of course, generates
thermal noise with vL2(f) = 4RLkT.
Since the two noise sources are physically
independent, superposition applies and we can
calculate the mean square current densities in
each direction, namely,

194

97
Lecture notes Radio Engineering by Jorma Kekalainen

Proof of Nyquists formula


The average power delivered from the source to the load in
any frequency band df is RL iL2(f) df, while the load delivers
R(f)iZ2(f) df back to the source.
Now the net power transfer must be zero at every frequency
for the circuit to be in thermal equilibrium.
Therefore,

and Nyquists formula follows after cancellation.

Note: Nyquists formula does not hold when the resistances are at different
temperatures or the network contains non-thermal sources. However, such cases are
easily analyzed provided that the sources are independent. Using superposition and
summing mean square values the resulting frequency density can be found. 195

Thermal noise by Nyquist


Nyquist has determined the thermal noise output of a
network containing both resistive and reactive
elements.
At a port in such a circuit, the mean-square thermal
noise voltage is given by

in which R(f) is the resistive (real) part of the input


impedance at frequency f and the integration is performed
over the bandwidth (B) of interest.

196

98
Lecture notes Radio Engineering by Jorma Kekalainen

Example

The thermal noise generated in the


conductance of a parallel GC circuit is
represented by a current source with
In2(f )=4kTGf
over any small frequency increment f.

197

Example
The admittance of the circuit is

Y G j C
The mean-square voltage at the port is

I n2( f )
V 2
n ( f ) 2
Y

198

99
Lecture notes Radio Engineering by Jorma Kekalainen

Example

By integration over infinite bandwidth the total


mean-square noise voltage is obtained by
4kTG
Vn2 df
0 G (2fC ) 2
2

The resistance at the port is

G
R f
G (2fC ) 2
2

199

Example
To complete the calculation of Vn2, it is
convenient to change the variable of
integration to and after some manipulations

kT
Vn2
C

Note: Suprisingly, Vn2depends on C but not on R (or G), even though the noise source is
actually thermal resistance! The explanation of this paradoxical independence from R is
included to the equivalent noise bandwidth (B1/R) of GC-filter. 200

100
Lecture notes Radio Engineering by Jorma Kekalainen

Example

Substitution of the half-power frequency of RC


circuit,
G
3dB 2f 3dB
C

into the previous integral yields after


integration the result

Vn2 4kTR f 3dB
2
201

Example
By the comparison of
Vn2 4kTRB
with the previous result,

Vn2 4kTR f 3dB
2
the equivalent noise bandwidth is seen to be
equal to /2 times the half-power bandwidth.

202

101
Lecture notes Radio Engineering by Jorma Kekalainen

Problem

Verify that the change in variable from f to


in
4kTG
V
n
2
df
0 G 2 (2fC ) 2
yields

4kT 1
2G 0
V
n
2
2
d
C
1
G 203

Problem
Then introduce the substitution given by
G
3dB 2f 3dB
C
and change the variable of integration to
x = /3dB.
Show that the value of the definite integral that
results yields
kT
Vn2 4kTR f 3 dB
2 C 204

102
Lecture notes Radio Engineering by Jorma Kekalainen

Noise in receiving antennas


The resistance measured at the terminals of an antenna is on
the order of 70 ohms (half-wave dipole) to 300 ohms (folded
dipole).
This resistance value is primarily the radiation resistance,
which accounts for power that is radiated from the antenna.
The ohmic resistance contributed by the resistance of the
antenna conductors is usually negligible in comparison with
the radiation resistance.

Note: An exception occurs in low-frequency antennas in which the dlmensions are a


small fraction of a wavelength. Here the ohmic resistance may predominate,
205

Antenna noise sources


A receiving antenna exhibits noise at its
terminals from two sources:
(1) the thermal noise generated in its ohmic
resistance (usually negligible), and
(2) the noise received from external sources.
(Any body with temperature greater than 0 K
radiates noise energy.)

206

103
Lecture notes Radio Engineering by Jorma Kekalainen

Antenna noise temperature


The received noise is represented as though it were
thermal noise generated in a fictitious resistance equal
to the radiation resistance, at a temperature TA that
would account for the noise actually measured.
This is called the noise temperature of the antenna.

207

Example
Suppose that a 200-ohm antenna exhibits an
rms noise voltage of 0.1 V at its terminals,
when measured in a bandwidth B = 10 kHz.
By the use of the equation

Vn2 4kTA RB
Vn2 1014
TA 90.6 K
4kRB 4 1.38 10 23 200 104
208

104
Lecture notes Radio Engineering by Jorma Kekalainen

Equivalent noise temperature

Thus the noise at the antenna terminals is


equivalent to that from a 200-ohm resistor at a
temperature of 90.6 K.
Note that other parts of a receiving system can
also be characterized by equivalent noise
temperatures in order to simplify the
computation of signal-to-noise ratio at the
output of a receiver.

209

Shot noise
The most common type of non-thermal noise in
electrical circuits is shot noise.
This phenomenon occurs whenever charged
particles cross a potential barrier as in
semiconductor junctions or vacuum tubes.
Small variations of kinetic energy among the
individual particles cause random fluctuations of
the total current.
Schottky first studied this effect in a vacuum-tube
diode operated under temperature-limited
conditions.

210

105
Lecture notes Radio Engineering by Jorma Kekalainen

Shot noise
Schottky found that shot noise could be
represented by a current source with

in which q is the electronic charge (1.6 x 10-19 C)


and I is the DC current.
Since this simple equation is independent of both
frequency and temperature, shot noise seems to be
non-thermal white noise.
Later investigations have shown that in2(f)
actually decreases at frequencies above f1/ ,
where denotes the average particle transit time.
211

Shot noise in junction diode


Schottkys result also holds for the
semiconductor junction diode

The net DC current consists of two


components, given by the diode equation

212

106
Lecture notes Radio Engineering by Jorma Kekalainen

Shot noise in junction diode


The net DC current consists of two components,
given by the diode equation

where V is the junction voltage and Is is the


reverse saturation current.
The two current components produce statistically
independent shot noise, so the total mean square
noise current density becomes

213

Semiconductor junction diodes noise


model
Figure below shows the complete noise source model,
including the diodes dynamic resistance

Unlike ohmic resistance, dynamic resistance is noiseless since


it does not correspond to any power dissipation.

Semiconductor junction diodes noise model

214

107
Lecture notes Radio Engineering by Jorma Kekalainen

Non-white noise
Transistors, vacuum tubes, resistors, and other devices exhibit a low-
frequency phenomenon known as flicker noiseoften called one-
over-f noise because the mean square density is proportional to 1/fn
with n1.
Some semiconductor devices also produce burst or popcorn noise,
whose waveform resembles the random telegraph wave.
Flicker and burst noise pose serious problems for low-frequency
applications, but they usually can be ignored at frequencies above a
few kilohertz.
At much higher frequencies, capacitive coupling and various other
effects tend to increase noise in electronic devices.
Non-white noise may or may not be significant, depending on the
device and the application.

215

Semiconductor noise
Figure illustrates the frequency variation of
semiconductor noise caused by flicker, burst,
and white noise.

216

108
Lecture notes Radio Engineering by Jorma Kekalainen

Noise in multi-terminal devices


Because resistors and antennas are two-terminal devices, it is easy to
describe their noise characteristics in terms of a noise temperature or an
equivalent noise resistance.
The situation is more complicated for transistors and other multiterminal
circuit elements because their internally generated noise depends upon
temperature,
operating point, and
input and output terminations.
For noise calculations in transistor circuits, the transistors are conveniently
represented as black boxes with specified noise figures and the physical
causes of the transistor noise are represented by equivalent noise sources.

Semiconductor manufacturers give data-sheet information on the measured noise


figures for a variety of operating conditions. 217

Electrical Noise

Noise terms

109
Lecture notes Radio Engineering by Jorma Kekalainen

Definitions of noise terms

Various terms are used to define and compare


the relative amounts of noise produced in
electrical systems.
The following definitions and discussion will
provide the basis for understanding the
nomenclature and for computing the overall
effect of noise in a system.

219

Signal-to-Noise Ratio (SNR)


In a specified bandwidth, the signal-to-noise
ratio is defined as the ratio of signal power to
noise power at a port.
PS Vs2
SNR
PN Vn2
where Vs and Vn are the rms signal and noise
voltages, respectively.
In decibels, P
SNRdB 10 log10 S
PN 220

110
Lecture notes Radio Engineering by Jorma Kekalainen

Minimum SNR

The larger the SNR, the less the signal is


corrupted by the noise.
The lowest permissible value of SNR depends
upon the application.
Approximate minimum value is 10 dB at the
detector input of a receiver.

221

Noise sources
Note that as a signal passes through a cascade of
amplifier stages, the SNR continually decreases
because each stage adds additional noise.
In most systems, however, the amplified output noise
is due primarily to
(1) the noise present along with the input signal, and
(2) the noise contributed by the first two stages
(such as the RF amplifier and mixer stages in a
receiver).

222

111
Lecture notes Radio Engineering by Jorma Kekalainen

Noise bandwidth

The most common sources of noise (thermal


and shot noise) have an essentially uniform
spectral distribution so that the noise
transmitted through an amplifier is determined
by the bandwidth of the amplifier.

Shot noise is generated in semiconductors. It arises because the current crossing


junctions is made up of charge carriers that are emitted randomly; the number of these
carriers fluctuates statistically from instant to instant. Shot noise has essentially a flat
223
spectral distribution and is treated in the same manner as thermal noise.

Noise bandwidth
If the amplifier had a Constant gain characteristic, with
cutoff at fc
constant gain Av up
to some frequency fc
and zero gain
thereafter, the noise
bandwidth B would Filter with voltage gain Av (f), power
clearly be equal to fc. gain proportional to Av(f)2.

224

112
Lecture notes Radio Engineering by Jorma Kekalainen

Limited frequency response

Generally, however, the frequency response is


limited (by the shunt capacitance or by a tuned
circuit) so that an abrupt cutoff of the
frequency response is not achieved.
Thus a more sophisticated determination of
noise bandwidth is required.

225

Filter
Consider a filter, as shown in figure, that has
voltage gain Av(f)=V2/V1.
Filter with voltage gain Av (f), power
gain proportional to |Av(f)|2.

226

113
Lecture notes Radio Engineering by Jorma Kekalainen

Frequency response

Since the noise power is related to the mean-


square voltage, it is convenient to plot the
frequency response in terms of lAv(f)I2, as
shown by the solid line in Fig., where lAml2 is
the maximum value of this curve.

Illustration of
noise-equivalent
bandwidth B,
defined by equal
areas under the
dashed and solid
curves. 227

Filter output
If the input to this filter is white noise with mean-
square voltage v1n2/Hz, the corresponding mean-
square output voltage in a 1-hertz interval at
frequency f is

Addition of all such increments over the frequency


band yields

228

114
Lecture notes Radio Engineering by Jorma Kekalainen

Noise-equivalent bandwidth
The value of the integral on the right side is the area
under the solid-line curve of |Av(f)l2 in Fig..
The dashed line shows a rectangular spectrum of the
same maximum height lAm|2 and with bandwidth B.

Illustration of
noise-equivalent
bandwidth B,
defined by equal
areas under the
dashed and solid
curves. 229

Noise-equivalent bandwidth (B or BN)

The noise-equivalent bandwidth B is the value


that gives equal areas under the solid- and
dashed-line curves such that

or

230

115
Lecture notes Radio Engineering by Jorma Kekalainen

Comment
This is the value of B that should be used in equations
earlier, and obviously it must be evaluated for the
particular system being analyzed.
The integral is not always easy to evaluate.
However, in many RF amplifiers, the bandwidth is
established by tuned RLC circuits for which the noise
bandwidth is /2 times the 3-dB bandwidth of the
circuit.

231

Available noise power


The available power Pa of a source is the
maximum power that can be drawn from the
source.
If the source has internal impedance Zs = R +
jX, the maximum power will be delivered to a
conjugate-matched load (ZL = R - jX).

232

116
Lecture notes Radio Engineering by Jorma Kekalainen

Available noise power

If the open-circuit voltage of the source is V,


the maximum power transfer theorem yields

If R is a source of thermal noise, V2=4kTRB,


and

233

Available noise power


The available noise power in a bandwidth of 1
Hz is

Since Pa is independent of R, the available


noise power from all nonzero finite resistances
is the same.

234

117
Lecture notes Radio Engineering by Jorma Kekalainen

Amplifier noise
A detailed circuit model showing all the
individual noise sources within an amplifier
would be very complicated and of little practical
value.
Consequently, alternative methods have been
devised for the analysis of noise in amplifiers.
Two particularly useful measures of amplifier
noise are the effective noise temperature and the
noise figure.
Both of these measures involve the concept of
available power gain.

235

Model of a general two-port network.

A general two-port network that could serve as


a model for an amplifier stage, filter, or other
network.

236

118
Lecture notes Radio Engineering by Jorma Kekalainen

Model of a general two-port network.

The source has open-circuit voltage Vs and internal


impedance Zs.
The two-port has input impedance Z1, output
impedance Z2, and open-circuit output voltage Vo.
The load is an impedance ZL.

The open-circuit voltage gain of the two-port will be


called H(f)

237

Available signal power


The available signal power from the source, as
defined previously, is

Similarly, the available signal power at the


output port is

238
Note: Rs and R2 are the real parts of corresponding impedances.

119
Lecture notes Radio Engineering by Jorma Kekalainen

Available power gain

The available power gain Ga of the network is


defined as the ratio Pao/Pas, that is,

Note that this expression does not account for


impedance mismatches at the input and output
ports.

239

Voltage relations
Examination of the network yields the
following relations:

Z1 Z s 1 Z1 1
Vs V1
Z1 Vs Z 1 Z s V1
Construction of Vo/Vs:
Vo 1 Z1 V Z1
Vo o H f
Vs Vs Z1 Z s V1 Z1 Z s
240

120
Lecture notes Radio Engineering by Jorma Kekalainen

Available power gain


Substitution of the expressions

into

gives the following relation for available power gain


at frequency f

241

Remark
The two-port model applies only if the network is
unilateral (no reverse transmission).
For this case it is worth emphasizing that Ga given by
the previous expression is independent of the value of
ZL.
The available gain of several cascaded unilateral
networks is equal to the product of the Ga values of
the individual networks.

Note that Ga is not the actual gain of the network the value of Ga would be obtained in
242
practice only with ideal matched conditions (Z1 = Zs*and ZL = Z2*) at both ports.

121
Lecture notes Radio Engineering by Jorma Kekalainen

Definition: Equivalent noise temperature

The noise temperature at any port of a network


is defined as follows:
A noise source that has an available power Pa in a
small frequency interval f has an equivalent noise
temperature equal to
Te = Pa/kf.
If the noise power spectrum of the source is not
flat, Pa and Te are frequency-dependent.

Note:
243

Excess noise temperature (Tx)


Noise generators used for amplifier testing are
often calibrated in terms of excess noise
temperature,

in which T is the noise temperature of the


source and T0=290 K is the standard reference
temperature.

244

122
Lecture notes Radio Engineering by Jorma Kekalainen

Noiseless network

If a thermal noise source of temperature T is


connected to a noiseless network with small
bandwidth f and available gain Ga(f), the
available noise power from the source is

and the available noise power at the output of


the network is

245

Model of a noisy network


If the network is noisy, it will produce additional
noise power, Pne, at the output.
With the same input noise as before, the output noise
power will be

246

123
Lecture notes Radio Engineering by Jorma Kekalainen

Noiseless network arrangement

Now replace the noisy network with a


noiseless network having the same available
power gain Ga(f), and account for the output
noise Pne by means of an extra noise source on
the input side.

247

Effective input noise temperature

The temperature Te of this extra source is


adjusted to produce Pne at the output,

and

The value Te is called the effective input noise


temperature of the network.
Note: This way of representing network noise is very useful in the determination of
248
overall signal-to-noise ratios of cascaded amplifiers.

124
Lecture notes Radio Engineering by Jorma Kekalainen

Noise Figure (NF or F)


It seems to be the variety of ways in which the
relative noisiness of a device or system can be
expressed.
Fortunately, a single index called the noise figure can
be used to compare noise performance.
The noise figure (NF) of a two-port network gives a
measure of the degradation of the SNR between the
input and output ports.

Note: Watch out for several variations of the term with subtle differences that can be
249
confusing.

I/O-powers
Figure shows a noisy network with input
signal and noise powers and Psi and Pni
respectively, and corresponding output signal
and noise powers Pso and Pno.

Signal and noise powers at the input


and output of a two-port network.
250

125
Lecture notes Radio Engineering by Jorma Kekalainen

Definion: Noise Figure (Noise Factor)

The noise figure is defined over a specified


bandwidth as

251

NFdB (or FdB)


The value of NF is often expressed in decibels
through the relation

For a noise-free network, the input and output


SNRs will be equal and NF = 1 or NFdB=0.
Practical circuits always have larger noise
figures than this.

252

126
Lecture notes Radio Engineering by Jorma Kekalainen

Additional specifications for N F


Equation

provides a conceptual definition of the noise figure, but it


requires further qualification to make it more precise.
The ratio Psi/Pso is equal to 1/Ga(f), where Ga(f) is the frequency-
dependent available power gain of the network.
The variation of Ga(f) with frequency must be taken into account.
Furthermore, the input power from the signal source is a function of
temperature.
In order to obtain a standard value for NF, the source temperature must be assumed
to be 290 K.
These considerations have led to the following definitions.

253

Spot noise figure


At a selected input frequency, the spot noise
figure is the ratio of
(1) the total available noise power per unit
bandwidth at the output port to
(2) the portion produced at the input
frequency by the input termination, whose
noise temperature is 290 K.

254

127
Lecture notes Radio Engineering by Jorma Kekalainen

Spot noise figure

If the network provides a conjugate match at


the input port, the available power from the
standard-temperature source in a 1-Hz
bandwidth is equal to Pa(1Hz)=kT0.
Hence, the spot noise figure is given by

255

Note
In practice, the value of Pno is measured over a
small bandwidth f that is more than 1 Hz due
to the practical limitation on filter bandwidth.
The equation for NF then becomes

256

128
Lecture notes Radio Engineering by Jorma Kekalainen

Average noise figure


Over wide bandwidths in which Ga(f) varies
appreciably, the average noise figure NF is
given by

in which Pno is the total noise power delivered


to the output termination in noise bandwidth B,
and Gmax is the maximum value of |Ga(f)|.
257

Overall NF of cascaded networks


Of primary interest in the evaluation of noise
performance of multistage amplifiers (such as
radio receivers) is the overall noise figure of
the system.
In general, the overall noise figure is evaluated
for a bandwidth B that is the bandwidth of the
overall system.
In a radio receiver the IF stages are narrow-band
and they determine the amount of noise that
reaches the detector.

258

129
Lecture notes Radio Engineering by Jorma Kekalainen

Output noise power

From

if the source temperature T is assumed to be T0


(required for the standard definition of NF),
the output noise power of a single stage in a
small frequency band is

259

Relation between NF and Te


Substitution of the previous equation into

gives

Thus the noise figure referred to a standard-temperature


source can be expressed in terms of the equivalent noise
temperature of the network, and vice versa.
260

130
Lecture notes Radio Engineering by Jorma Kekalainen

Two networks in cascade

Figure shows two cascaded networks with


their respective available gains, effective noise
temperatures, and noise figures.
A noise source at standard temperature is input
to the system.

261

Available output noise power of two


stages
Small bandwidth f is assumed for all parts of the system.
The available output noise power of two cascaded
network is found by the use of the output noise power of a
single stage

and some regrouping of terms

262

131
Lecture notes Radio Engineering by Jorma Kekalainen

Overall effective (input) noise


temperature
A comparison of this relationship with the equations

leads to an expression for the effective input


temperature of the two networks in cascade

Note that the input source temperature does not appear in the expression.
Pne is a symbol for additional noise power at the output caused by noisy network. 263

Overall effective (input) noise


temperature
Here Te1,2 is the effective input temperature
that accounts for all of the output noise
introduced by the noisy networks.
For n networks in cascade, the corresponding
expression is

264

132
Lecture notes Radio Engineering by Jorma Kekalainen

Overall Noise Figure


Introduction of the network noise figures by the use
of

in

leads to the following expression for the overall noise


figure

265

Conclusion

It is apparent from this relation that the NF of the first


stage in a system has the predominant effect on the
overall NF, unless Ga1 is small or NF2 is large.
Therefore, the system designer should always try to
minimize the noise produced in the first stage by the
choice of low-noise transistors and the selection of
operating conditions that minimize noise.

266

133
Lecture notes Radio Engineering by Jorma Kekalainen

If Ts T0=290K
By definition, the standard noise figure is the
ratio of the input SNR to the output SNR with
the input source at the standard 290 K noise
temperature.
In many practical cases, however, the input
noise temperature is not 290 K and the
standard noise temperature does not describe
accurately the SNR degradation from input to
output of a system.

267

Actual noise figure


A true measure of the SNR degradation is the
actual noise figure defined by

in which TsT0.
This value is related to the standard noise
figure evaluated for Ts=T0 by the relation

268

134
Lecture notes Radio Engineering by Jorma Kekalainen

Example
Calculate effective temperatures and noise figures for the
receiver front end shown in below. The antenna has a
(radiation) resistance of 70 ohms and an effective temperature
of 20 K due primarily to external radiation.

The noise contributed by the receiver local oscillator is assumed to be negligible (which
269
is not always the case).

Solution
Noise figures and gains for the RF amplifier
and mixer are given in decibels, and must be
converted to actual values for use in the
computation.
For the RF amplifier, NF1=2, and Ga1 = 10.
For the mixer, NF2 = 4.47 and Ga2 = 7.94.

270

135
Lecture notes Radio Engineering by Jorma Kekalainen

Solution

By the use of

the effective noise temperatures of the two


units are found to be
Te1=290 K and
Te2=1006 K.

271

Solution
Then from

the effective input temperature of the receiver


(excluding the antenna) is found as

Note: Here Tr Te1,2 272

136
Lecture notes Radio Engineering by Jorma Kekalainen

Solution
The overall noise figure can be calculated from
either

or

273

Comment
In the preceding example, the value of NF was found
to be 2.35 or 3.7 dB.
However, since the antenna temperature was 20 K,
the use of

to calculate the actual noise figure yields


NFact = 1 + (2.35-1)(290/20) = 20.43 or 13.1 dB.
By definition, the standard noise figure is the ratio of the input SNR to the output SNR
with the input source at the standard 290 K noise temperature. Here, however, the input
noise temperature is not 290 K and the standard noise temperature does not describe
274
accurately the SNR degradation from input to output of a system.

137
Lecture notes Radio Engineering by Jorma Kekalainen

Amplifier noise considerations


For calculations of NF or Te of a system
comprising one or more stages, knowledge is
required of
(1) the noise delivered from the signal source,
(2) the noise-equivalent bandwidth B,
(3) the thermal noise generated in various resistances
in the circuit, and
(4) the noise generated within the solid-state devices.

Noise produced within semiconductors can be predicted by the use of


equivalent circuits. However, because there are so many variables to be taken
into account that such predictions are best made by the use of analysis
275
programs.

Electrical Noise

Noise calculations

138
Lecture notes Radio Engineering by Jorma Kekalainen

Example

Verify that the change in variable from f to


in
4kTG
V
n
2
df
0 G 2 (2fC ) 2
yields

4kT 1
2G 0
V
n
2
2
d
C
1
G 277

Solution
d
f df
2 2
1 4kTG
Vn2
2 G (C ) 2
0 2
d

4kTG d
2 0

2G C 2
1 ( )
G
4kT d

2G 0 1 ( C ) 2
G

278

139
Lecture notes Radio Engineering by Jorma Kekalainen

Evaluation of the previous integral


4kT d G
Vn2
2G 0 1 ( C ) 2
, 3dB , (G || C - circuit)
C
G
4kT d G

2G 1 ( ) 2
0
,x
3dB
d 3dB dx dx
C
3dB
4kT d 4kTG dx
2G 0 1 ( ) 2 2GC 0 1 x 2

3dB
4kT kT
arctan x0


2C C

0
2
279

Example
Show that
4kTG
V
n
2
df
0 G 2 (2fC ) 2
yields
1 G
Vn2 4kTR f 3dB , when f 3dB
2 2 C

280

140
Lecture notes Radio Engineering by Jorma Kekalainen

Solution
4kTG 4kTG df
Vn2 df
0 G 2 ( 2fC ) 2 G 2 0 2fC 2
1 ( )
G
4kT df 1 2C

G 0 f
,
f 3dB

G
1 ( )2
f 3dB
4kTf 3dB dx f

G 0 1 x2
,x
f 3dB
df f 3dB dx
1
G
4kTf 3dB
arctan x 0 4kTR f 3dB
R


G 2
0
2

281

Example
According to quantum mechanics the spectral density of
thermal noise can be expressed in form

hf V2
Gnoise f hf
Hz
e kT
1
where k=1.38*10-23 J/K
h=6.62*10-34 Js
is the proportionality symbol

a) Show that Gnoise(f) is essentially constant in the region of the radio


spectrum.
b) Determine an expression for Gnoise(0).
Note: 0/0 is not the correct answer.
282

141
Lecture notes Radio Engineering by Jorma Kekalainen

Spectral density function

Thermal noise level [dBW] T=290K

Frequency [GHz]

283

Solution a)
Of course, we can approximate with series expansion of e

2
hf
hf
hf kT
e kT
1
kT 2
hf hf
, 1
kT kT
Now we get
hf
G f kT , 0 1
kT
284

142
Lecture notes Radio Engineering by Jorma Kekalainen

Solution a)
Lets select the standard temperature T=T0=290K and a
rough but conservative estimate

hf upper kT0
0.1 1 f upper 0.1
kT0 h
Substituting numerical values kT0 4 10 21 J
and h 6.62 10 -34 Js
f upper 1012 Hz
This upper limit is in the infrared portion of the
electromagnetic spectrum.

285

LHospitals rule

If lim M f lim N f 0
f 0 f 0

then we can calculate Mf


lim
f 0 N f
using LHospitals rule
Mf M f
lim N f lim N f
f 0 f 0

whenever the limit on the right can be found.

286

143
Lecture notes Radio Engineering by Jorma Kekalainen

Gnoise(0)
Lets denote Mf hf V2
Gnoise f
N f hf
Hz
e kT
1
d
hf
Mf M f df
lim
f 0 N f
lim
f 0 N f
lim
f 0 d
hf

e kT 1
df

h kT
lim hf
lim hf
kT
f 0 h kT f 0
e e kT
kT

Gnoise(0) kT 287

Example
Reduce the circuit

to the Thevenin equivalent

288

144
Lecture notes Radio Engineering by Jorma Kekalainen

Solution

289

Solution

Note: Removing the voltage source means that we replace the source voltage with
short-circuit.
Note: In circuit noise analysis we simply use the thermal noise source model
vn2(f)=4kTRTh , not the signal source model based on the voltage division as in this
problem. Circuits containing more than one resistor may be analyzed by reducing them
to one (Thvenin) equivalent resistance RTh and applying vn2 (f)=4kTRTh to obtain the
mean-square noise voltage density. So the noise Thevenin equivalent of such a circuit
is then a voltage source with this mean-square voltage density in series with an ideal
(noiseless) resistor RTh. 290

145
Lecture notes Radio Engineering by Jorma Kekalainen

Example
In the network below R1 and R2 are thermal resistances
at different temperatures (T1T2). Let R2 = R1 = R and L
= R/(2). Obtain expressions for vn2(f) and p(f).

291

Solution
Now resistances are at different temperatures
(T1T2) Nyquists formula

vn2 f 4 R f kT
is not valid.
However, the sources are independent.
Thus, we can use superposition and sum mean square
values to find resulting frequency densities.

292

146
Lecture notes Radio Engineering by Jorma Kekalainen

Solution
Applying superposition the total short-circuit mean square
current density is
vn21 f vn22 f
in21 f in22 f 2
2
Z1 Z2
where vn21 f 4 RkT1 and vn22 f 4 RkT2
2
R
R jfR R fR
2 2
R j 2f 2 2
Z1
2
2
Z2 R2
4k T1
in2 f T2
R 1 f 2

293

Solution
Z Th f ( R jfR) || R
R( R jfR) R1 jf

R ( R jfR) 2 jf

Z Th f
2 R1 jf
2



R2 1 f 2
2 jf
2
4 f 2
According to Ohm' s law

v f Z Th f
2
i f
2 2
4kR T1 1 f 2 T2
n n
4 f 2
294

147
Lecture notes Radio Engineering by Jorma Kekalainen

Solution
R1 jf
Z Th f
2 jf

ReZ Th f Re
R1 jf 2 jf R 2 f 2
2 jf 2 jf 4 f 2

p f
vn2 f
k

T1 1 f 2 T2
4 ReZ Th f 2 f 2

T1 T2 T
Note : p f kT
295

Note
When the resistances are in thermal equilibrium at
temperature T, Nyquists formula states that
vn2 f 4 RTh f kT , RTh f ReZTh f
Z Th f RTh f jX Th f
Hence, the mean square voltage density takes the
shape of the equivalent resistance RTh(f).
Nyquists equation includes the special case of an all-
resistive network whose equivalent resistance will be
independent of f.

296

148
Lecture notes Radio Engineering by Jorma Kekalainen

Note

If in the previous problem (T1=T2=T) Nyquists


formula is valid and then

vn2 f 4 RTh f kT , RTh f ReZ Th f


2f2
4kTR and
4f2
vn2 f
p f kT
4 ReZ Th f

297

Example
Lets study the circuit model of a noiseless amplifier inserted
between a source and a load.
For simplicity, we have omitted any reactances that might be
associated with the source, amplifier, or load impedances.
In this analysis we use spectral density functions

298

149
Lecture notes Radio Engineering by Jorma Kekalainen

The available power density


The amplifier is characterized by an input resistance ri,
output resistance ro, and voltage transfer function H(f).
The source generates a mean-square voltage density
vs2(f) representing noise or an information signal or
bothand the available power density from the source is
pas(f) = vs2(f)/4Rs.
The available power density at the output of the amplifier
is

299

Amplifiers available power gain


We define the amplifiers available power gain
ga(f) as the ratio of these available power
densities, i.e.,

The actual power gain of an amplifier equals


the available power gain when impedances are
matched to obtain maximum power transfer at
input and output.
300

150
Lecture notes Radio Engineering by Jorma Kekalainen

Noisy amplifier
Well assume from now on that the source generates
white noise, thermal or non-thermal, with noise
temperature Ts.
Then pas(f) = kTs and the available noise power
density at the output of a noiseless amplifier will be

pao f g a f pas f g a f kTs

But a noisy amplifier contributes additional internally


generated noise pint(f).

301

Available output power density


Since the internal noise is independent of the source noise, we
write

where pint(f) stands for the available power density of the


internal noise seen at the output.
Figure depicts the previous equation in the form of a block
diagram.

302

151
Lecture notes Radio Engineering by Jorma Kekalainen

Total available output noise power

Integration then yields the total available output noise power



pao f df kTs g a f df pint f df
0 0 0
an expression that calls for some simplifications.
Most amplifiers in a communication system have a frequency-
selective response, with maximum power gain g and noise
equivalent bandwidth BN.
These parameters are related to ga(f) by

303

Effective noise temperature


So the first term of No reduces to kTs gBN.
Next, to simplify the second term, we define the
effective noise temperature of the amplifier to be
1
Te pint f df
gkBN 0

Hence, the total output noise power becomes


N o kTs gBN gkBN Te gk Ts Te BN
It is diagrammatically portrayed in the following
figure.
304

152
Lecture notes Radio Engineering by Jorma Kekalainen

Total output noise power

This diagram brings out the fact that Te


represents the internal noise referred to the
input and thereby expedites calculations of
signal-to-noise ratios.

305

Block-diagram representation of a
noisy amplifier

306

153
Lecture notes Radio Engineering by Jorma Kekalainen

Calculations of signal-to-noise ratios


The figure represents a noisy amplifier with signal plus white noise
at the input.

The available signal power from the source is Ss and the signal
spectrum falls within the passband of the amplifier so the available
signal power at the output will be
So = gSs.

307

Output signal-to-noise ratio


Thus, using the two previous equations, the
output signal-to-noise ratio is
S gS Ss
s
N o N o k Ts Te BN
Note that the gain g has canceled out in
numerator and denominator.

308

154
Lecture notes Radio Engineering by Jorma Kekalainen

Note
Although the source noise does not necessarily have a defined
bandwidth, the source signal-to-noise ratio is taken by convention to
be
S def . S s

N s kTs BN
which just corresponds to the signal-to-noise ratio produced by an
ideal noiseless filter with unit gain and bandwidth BN.

309

(S/N)o expressed with (S/N)s


However, reforming (S/N)o expression of the noisy
amplifier produces
S gS Ss Ss
s
N o N o k Ts Te BN 1 Te kT B
Ts s N
1 S

1 Te N s
Ts

S def . S s
where
N s kTs BN 310

155
Lecture notes Radio Engineering by Jorma Kekalainen

SNR comments
Note that in case of a noisy amplifier
(S/N)o< (S/N)s.
We also see that the degradation of signal-to-
noise ratio due to a noisy amplifier depends upon
the value of effective noise temperature relative to
the source noise temperature.
In particular, if Te<<Ts then (S/N)o (S/N)s
Meaning that under this condition the internal
noise has little effect and the amplifier appears to
be noiseless.

311

Mismatching comments
When impedances are not matched at the input or
output all signal and noise powers will be less
than the available powers.
Nonetheless, the previous SNR calculations are
still valid because they express power ratios
measured at specific points, so the impedance
mismatch factor cancels out along with the gain.
The effective noise temperature is therefore a
significant parameter, irrespective of impedance
matching.

312

156
Lecture notes Radio Engineering by Jorma Kekalainen

Noise figure (F)

Another measure of amplifier noise is the


noise figure F, defined such that

when Ts T0

Since So= gSs and (S/N)o = Ss/(kT0 BN) when


Ts=T0
No T
F 1 e
gkT0 BN T0
313

F vs. Te
Conversely
Te F 1T0

A very noisy amplifier has Te>>T0 , and F>>


1 in which case we generally express the
value of F in decibels.
A low-noise amplifier has Te < T0 and 1 < F <
2 in which case we usually work with the
effective noise temperature.
314

157
Lecture notes Radio Engineering by Jorma Kekalainen

Noise parameters of some


microwave amplifiers

315

Spot noise
Equation
No T
F 1 e
gkT0 BN T0
defines the average or integrated noise figure in the sense that
No involves the integral of p0(f) over all frequency.
But sometimes we need to know how the internal noise varies
with frequency.
The spot noise figure F(f) contains this information in the form

po f
F f , when Ts T0
kTs g a f 316

158
Lecture notes Radio Engineering by Jorma Kekalainen

Amplifier noise measurement

Measuring absolute noise power is a difficult ,


so clever techniques have been developed for
amplifier noise measurement with a relative
power meter connected at the output.
One technique utilizes a calibrated source of
white noise, such as diode noise generator,
impedance-matched to the input of the
amplifier.

317

Noise measurement procedure


1 Set the noise source temperature at Ts=T0 and record the power
meter reading N1. This value corresponds to

N1 CN o Cgk T0 Te BN
where the proportionality constant C includes any impedance
mismatch factor at the output.
2 Increase the source temperature to Ts=T0+Tx such that the
power meter reading has doubled

N 2 Cgk T0 Tx Te BN 2N1

318

159
Lecture notes Radio Engineering by Jorma Kekalainen

Noise measurement procedure

Then form
N 2 Tx T0 Te
2 Te Tx T0
N1 T0 Te
The final result follows substituting Te
to noise figure expression
No T T T T
F 1 e 1 x 0 x
gkT0 BN T0 T0 T0

Note: We dont need to know g, BN, or the constant C. 319

Example
An amplifier with g = 60 dB and BN =2 MHz has No = 40 nW
when the source noise is at room temperature.
(a) Find the effective noise temperature Te and noise figure F.
(b) Calculate the increased source temperature Ts needed for
the second step of the previous measurement procedure.

Note: For computational purposes, the quantity kT can be


rewritten as
T T
kT kT0 4 10 21 J
T0 T0
T0 290K is the standard room temperature 320

160
Lecture notes Radio Engineering by Jorma Kekalainen

Solution
An amplifier with g = 60 dB and BN =2 MHz has No = 40 nW
when the source noise is at room temperature.
(a) Find the effective noise temperature Te and noise figure F.
The total output noise power
N o gk T0 Te BN 106 k T0 Te 2 10 6
T0 Te
2 1012 4 10 21 40 10 9
T0
T0 Te
So 5 Te 5T0 T0 4T0
T0
Te 4T
F 1 1 0 5
T0 T0
321

Solution
An amplifier with g = 60 dB and BN =2 MHz has No = 40 nW
when the source noise is at room temperature.
(b) Calculate the increased source temperature Ts needed
for the second step of the previous measurement procedure
From a)
T
F x 5 Tx 5T0
T0

and the increased source temperature

Ts =T0+Tx=6T0=6*290=1740K

322

161
Lecture notes Radio Engineering by Jorma Kekalainen

Lossy two-port networks


Lossy two-port networks are such as transmission lines and
connecting cables.
Power loss implies dissipation by internal resistance.
Consequently, the internal noise is thermal noise at the
ambient temperature Tamb and pint(f)=kTamb.
However, we cannot use the previous noisy amplifier model
presented in the following slice because lossy two-ports are
bilateral, meaning that a portion of the internal noise flows
back to the input.

323

Block-diagram representation of a
noisy amplifier

Total output noise power

N o kTs gBN gkBN Te gk Ts Te BN


324

162
Lecture notes Radio Engineering by Jorma Kekalainen

Bilateral lossy two-port


When impedances are matched, a bilateral lossy two-port has
constant gain g < 1 in both directions, so gpint(f) flows back to
the input while (1-g) pint(f) goes to the output.
The total available noise power in bandwidth BN at the output
thus becomes
N o gkTs BN 1 g kTamb BN
1
gk Ts L 1Tamb BN , L
g
where L is the transmission loss or attenuation.

325

Bilateral lossy two-port


Comparing our expression for No with Eq.
N o kTs gBN gkBN Te gk Ts Te BN

N o gkTs BN 1 g kTamb BN
1
gk Ts L 1Tamb BN , L
g
we obtain the effective noise temperature
Te L 1Tamb

326

163
Lecture notes Radio Engineering by Jorma Kekalainen

Bilateral lossy two-port


Substituting the effective noise temperature
expression
Te L 1Tamb
to the noise figure equation

No T
F 1 e
gkT0 BN T0

we obtain
Tamb
F 1 L 1
T0
Note: If a lossy two-port is at room temperature, then Tamb = T0 and F = L. 327

System noise calculations


Here we take up the analysis of cascade-
connected systems that include amplifiers and
other noisy or lossy two-port networks.
Our objective is to develop expressions for the
overall performance of the system in terms of
the parameters of the individual stages.

328

164
Lecture notes Radio Engineering by Jorma Kekalainen

Example: Two noisy two-ports


Examine the cascade of two noisy two-ports, where
subscripts identify the maximum power gain, noise
bandwidth, and effective noise temperature of each
stage.

We assume that both stages are linear and time


invariant (LTI).

329

BN B2B1 g = g1g2
We further assume that the passband of the second stage falls
within the passband of the first stage, so B2 B1 and the
overall noise bandwidth is
BN B2.
This condition reflects the sensible strategy of designing the
last stage to mop up any remaining noise that falls outside the
signal band.
The overall power gain then equals the product
g = g1g2
since the first stage amplifies everything passed by the second
stage.

330

165
Lecture notes Radio Engineering by Jorma Kekalainen

Total output noise power

The total output noise power consists of three


terms:
source noise amplified by both stages;
internal noise from the first stage, amplified by the
second stage; and
internal noise from the second stage.
Thus,
N o gkTs BN g 2 g1kT1 BN g 2 kT2 BN
T
gk Ts T1 2 BN
g1 331

Overall effective noise temperature


and noise figure
The overall effective noise temperature is
T2
Te T1
g1
The overall noise figure is
T1 T F 1
F 1 2 F1 2
T0 g1T0 g1
which follows from the general relationship

Te
F 1 .
T0
332

166
Lecture notes Radio Engineering by Jorma Kekalainen

General case
The foregoing analysis readily generalizes to the case of three or
more cascaded LTI two-ports.
The overall effective noise temperature and noise figure are given
by
T2 T
Te T1 3
g1 g1 g 2
F2 1 F3 1
F F1
g1 g1 g 2
Both expressions bring out the fact that the first stage plays a critical
role and must be given careful attention in system design.
333

Example: Lossy 1. stage


Suppose the first stage happens to be a connecting cable or any other lossy
two-port.
From the noise viewpoint, the attenuation has a double harmful effect since
g1 = 1/L1 < 1 and

T1 L1 1Tamb Te L 1Tamb
The overall noise temperature thus becomes
T2 T
Te T1 3
g1 g1 g 2
L1T3
L1 1Tamb L1T2
g2
Note: L1 (>1) multiplies the noise temperatures of all subsequent stages. 334

167
Lecture notes Radio Engineering by Jorma Kekalainen

Example: Preamplification principle


Suppose the first stage is a preamplifier with sufficiently large gain
g1 that the overall noise temperature reduces

T2 T
Te T1 3
g1 g1 g 2
T1 , g1
The system noise is then determined primarily by the preamplifier.
The remaining stages provide additional amplification and filtering,
amplifying the signal and noise without appreciably changing the
signal-to-noise ratio.
The design of low-noise receivers is usually based on this
preamplification principle.

335

Complete receiver
Here a complete communications receiver has been
divided into two major parts, a predetection unit
followed by a detector.
The detector processes the amplified signal plus noise
and carries out a nonlinear operation, i.e., analog
demodulation or digital regeneration.
It is the reasonable assumption that the detector
introduces negligible noise compared to the amplified
noise coming from the predetection unit.
We are concerned here with the predetection signal-to-
noise ratio denoted by (S/N)R.
Note: When the predetection unit includes a frequency converter, as in a
superheterodyne receiver, its conversion gain takes the place of available power gain. 336

168
Lecture notes Radio Engineering by Jorma Kekalainen

Communications receiver
(a) Communications receiver; (b) noise model of predetection unit.

BT is the transmission
bandwidth required
for the signal

337

Predection unit
The predetection portion of a receiver is a cascade of
noisy amplifiers and other functional blocks that act
as LTI two-ports under the usual small-signal
conditions.
Hence, as indicated in the previous block diagram a),
the entire predetection unit can be characterized by its
overall effective noise temperature calculated from
the general equation

T2 T
Te T1 3
g1 g1 g 2
338

169
Lecture notes Radio Engineering by Jorma Kekalainen

Predetection signal-to-noise ratio


(S/N)R
For a well-designed receiver, the predetection noise bandwidth
essentially equals the transmission bandwidth BT required for
the signal.
If the available signal power at the receiver input is SR and the
accompanying noise has temperature TR then the previously
presented SNR equation becomes
S S SR S
R gS s Ss

N R pBT k TR Te BT
N N o k Ts Te BN
o

where p k TR Te kTN

339

System noise temperature


The sum
TN TR Te
is called the system noise temperature, and p
represents the total noise power density referred to
the input of an equivalent noiseless receiver
corresponding to the diagram below.

340

170
Lecture notes Radio Engineering by Jorma Kekalainen

Example
The signal received at a satellite ground station is
extremely weak.
Fortunately, the accompanying noise comes
primarily from cold atmospheric phenomena
and has a very low temperature.
Minimizing the receiver noise is therefore
essential.
In contrast, a receiving antenna pointed at or
below the horizon picks up blackbody radiation
from the hot earth; then TRT0 and receiver
noise will be relatively less important.

341

Example
Calculate a) the overall effective noise temperature Te
b) the system noise temperature
c) the predetection signal-to-noise ratio(S/N)R

Figure depicts an illustrative low-noise microwave receiver for a satellite signal


with frequency modulation. The waveguide is part of the antenna feed structure
and introduces a small loss. Two preamplifiers are employed to mitigate the
noise of the high-gain FM receiver. 342

171
Lecture notes Radio Engineering by Jorma Kekalainen

Solution a)
Calculate a) the overall effective noise temperature Te
b) the system noise temperature
c) the predetection signal-to-noise ratio (S/N)R
T1 L1 1Tamb Te L 1Tamb
T2 T
Te T1 3
g1 g1 g 2
L1T3
L1 1Tamb L1T2
g2
1.05 170 1.05 1860
Te (1.05 1)290 1.05 9
100 100 10
14.5 9.5 1.8 2.0 27.8 K

Note: The waveguide loss accounts for half of Te, while the noise from the FM
receiver has been nearly washed out by the preamplification gain. 343

Solution b)
b) Calculate the system noise temperature

TN TR Te 30 K 27.8 K 57.8K
0.2T0

344

172
Lecture notes Radio Engineering by Jorma Kekalainen

c) Predetection (S/N)R
(a) Communications receiver; (b) noise model of predetection unit.

BT is the transmission
bandwidth required
for the signal

S S SR 10 12
R 50 17 dB
N R pBT k TR Te BT 0.2 4 10 25 10
- 21 6

where p k TR Te kTN k 57.8 k 0.2T0 0.2 kT0 0.2 4 10 21

Note: This SNR would be too low for analog communication without the further
345
improvement caused by FM demodulation.

Example
Suppose the parametric amplifier in the figure below could be
mounted directly on the antenna, ahead of the waveguide.
Find the system noise temperature TN with and without the
FET preamplifier.

346

173
Lecture notes Radio Engineering by Jorma Kekalainen

Solution
Find the system noise temperature TN with and without the
FET preamplifier.

With FET: 14.5


Te 9 1.8 2.0 12.9 K
100
TN TR Te 30 12.9 42.9 K

Without FET: 14.5 1.05 1860


Te 9 28.7 K
100 100
TN TR Te 30 28.7 58.7 K

347
Note: FET increases (S/N)R by 58.7/42.9=1.371.4 dB

Electrical Noise

Recap

174
Lecture notes Radio Engineering by Jorma Kekalainen

Noise
Thermal noise: electron collisions with the lattice structure in
resistive materials
For a resistor, we have a noise voltage

Shot noise: charge is not a continuous quantity


1/f-noise at low frequencies
Quantum noise: power in quanta (sub-mm )
Noise is characterized by noise temperature or noise factor
Noise factor describes the decrease of signal-to-noise ratio
caused by the system noise
At radio frequencies, the noise mainly considered is so called
white noise noise which has flat frequency spectrum 349

Receiver noise
Characterization of receiver noise (as well as noise of other devices)
relies on some main points and parameters to be remembered
For use as a noise reference:
Available noise power of a matched load
For use as parameters of the receiver or device:
Noise factor (noise figure when in dB)
Noise temperature
Noise temperature is applicable also e.g. to antenna then it is
called antenna noise temperature
Also noise bandwidth is a useful and effective concept
Idea: All of these simplify handling noise very complex subject
otherwise

350

175
Lecture notes Radio Engineering by Jorma Kekalainen

Reference noise
Available noise power of a matched load

Resistive loads are used as reference noise sources in


calculations but also in practice (matched resistive
terminations in device inputs)
Available noise power suitably depends on considered
bandwidth and the physical temperature of the load
Physical temperature can be measured quite accurately => a
good resistive termination is a good absolute reference for noise
when we are characterizing receivers and their sensitivity
Physical temperature is useful alone for defining noise power
density

351

Reference noise temperature T0


For definitions and related calculations, a commonly
defined true reference is always needed
In radio engineering we use standard temperature T0 = 290
K (= exact constant!) as the assumed physical temperature
of an absolute noise reference termination
Thus we get e.g. noise factor results without any definition
dependent errors or guessing
But as a result of the choice 290 K, it is often
convenient that T0 anyway happens to be near typical
room temperature
Conclusion: T0 = 290 K is actually constant - not
variable such as any room temperature or physical
temperature can be

352

176
Lecture notes Radio Engineering by Jorma Kekalainen

Noise factor
Noise factor (F) indicates how many times larger the output
noise power of a device is compared to a noiseless device
when both have a 290 K resistive termination in input

Nout includes the noise added by the device as well as the


amplified input noise NinG => F > 1
F has to be unambiguous for a device => F is specified with
using known defined Nin = kT0B

Note: Here noise temperature has to be reference value T0 = 290 K, not some
freely chosen temperature 353

F =SNRin/SNRout

Noise factor also provides a way to connect


signal-to-noise ratios (S/N) in input and output
of a two-port
We get the S/N relation in the noise factor
equation introducing input signal power Sin

354

177
Lecture notes Radio Engineering by Jorma Kekalainen

Noise figure

Noise figure is noise factor in dB:


F [dB] = 10 log10(F)
Typically amplifier datasheets give noise
figure specifications

Note: Some authors use these two concepts just the other way round. 355

Effective noise temperature


For receiver devices (and others) we can define noise also by
using effective noise temperature
This is always calculated as in the input of the device so that
we can use a well-defined equivalent source to help in noise
calculations
Otherwise noise in a device would be a very complex matter to
handle
Definition: Effective noise temperature Te of a device =
physical temperature of a matched load connected to a similar
noiseless device to obtain output power equal to the noise
power of the device

356

178
Lecture notes Radio Engineering by Jorma Kekalainen

Relation between Te and F

Relating Te to the noise factor F:

357

Derivation of the previous relation

Note: Te is not any physical temperature when linked to the device


as a parameter. 358

179
Lecture notes Radio Engineering by Jorma Kekalainen

Noise bandwidth

Gain of a receiver is usually not constant as a


function of frequency
Thats why for calculations of total noise we
need to define an effective noise bandwidth
to use with maximum gain:

Equal
areas

Note: Scale of G is in absolute values, not dB. 359

Note
Through this way we can use Gt,max and Bn as
simple parameters in noise power calculations
of device.

360

180
Lecture notes Radio Engineering by Jorma Kekalainen

Example
Picture shows a G-curve of a receiver low-noise amplifier
(LNA) with 13 dB of maximum gain.
Define the noise bandwidth Bn of the LNA.
G(f)

361

Noise of an attenuator
Noise temperature of a resistive attenuator

Note that waveguides and all kinds of wires


and cables have loss and thus can be
considered as attenuators adding noise =>
Attenuator noise has practical importance
especially in sensitive receivers

362

181
Lecture notes Radio Engineering by Jorma Kekalainen

Noise factor of attenuator


Noise factor of a resistive attenuator is

If we assume that Tphys T0 as might be in practice =>


we get for attenuator noise factor in this simple case:

Thus in practice about at room temperature, the noise


factor and loss of an attenuator are more or less equal
in value!
363

Receiver noise
Receivers are chains of stages - some noise is added
and also amplified in them - how to calculate it all?
Noise factor and noise temperature of a chain of
stages

Friis formulas for


matched stages

364

182
Lecture notes Radio Engineering by Jorma Kekalainen

Comment on receiver noise


Note: If G1 is large, the first stage dominates

Basically this is good because we can focus on


getting a low-noise first stage with gain
First stage noise dominance is very understandable:
this noise besides input noise - is amplified most to
the final output in a chain of amplifiers, i.e., by the
product of all gains: G1 G2 G3
In fact, it follows that the noise added by the final
stage is only a negligible part of the total noise in the
output!
365

Comparing amplifiers
When amplifiers are compared, the noise
measure counts:

Smaller M is better =>a low-noise amplifier with a


small noise factor but with also low gain may not
be a desirable choice
Friis formula can always be used for final check

366

183
Lecture notes Radio Engineering by Jorma Kekalainen

Example
Lets consider a simple example of two amplifiers in
chain
How about combined noise?
First amplifier (LNA stage): T1 = 100 K, G1 = 13 dB (=>
G1 = 20)
Second amplifier (gain stage): T2 = 300 K, G2 = 26 dB
(=> G2 = 400)

367

Comment on example
We notice that of the noise added by amplifiers,
100 K/115 K 87 % is from the 1st and only 15
K/115 K 13 % from the 2nd amplifier => we
have a typical dominating first stage
If we use the amplifier chain with a 290 K input
termination, then 290 K /(115 K + 290 K) 72 %
of noise in the chain output is amplified noise
from the input termination while amplifiers
contribute 25 % and 4 %, respectively
In practice, of course, input noise depends on
application

368

184
Lecture notes Radio Engineering by Jorma Kekalainen

Noise temperatures of receivers


Background noise sources as a reference

Molecular
resonances of water
vapour and oxygen:
loss as in an
attenuator => also
noise
Background
noise of the
Big Bang

Note: The HEMT or High Electron Mobility Transistor is a form of field effect transistor,
FET, that is used to provide very high levels of performance at microwave frequencies. 369

Antenna noise temperature


Brightness of black surface

Noise power received by an antenna

Note: Noise power received by an antenna is constant although brightness B


increases as 1/ 2 or f 2 . This is because antenna beam solid angle changes as
2 or 1/f 2 and compensates. 370

185
Lecture notes Radio Engineering by Jorma Kekalainen

Noise temperature of the sky

371

Radio link
Signal-to-noise ratio is a critical parameter for a radio link
For example, frequency modulation requires SNR 10 dB
Usually also some margin is needed due to fading etc.
SNR depends on received signal power and noise power of the
receiving system
By calculating SNR we get a link budget where the effects of
different factors are easily seen

372

186
Lecture notes Radio Engineering by Jorma Kekalainen

Radio link
Received signal power Input noise power of
receiver system

Figure of merit G/T


So SNR is

373

Example: Signal bandwidths

374

187
Lecture notes Radio Engineering by Jorma Kekalainen

Example: Satellite-TV connection at 12 GHz

Link budget factors are often considered just like in a budget


of dB values

Budget calculation result: Necessary receiving system


parameters


375

Example

If for a circular dish antenna DA = 0.4 m Gr


= 31.8 dB (using aperture efficiency 0.6)
max. TS = (31.8-2.4) dBK
Thus TS 870 K TR 720 K (when TA =
150 K)

376

188
Lecture notes Radio Engineering by Jorma Kekalainen

Note: Effective aperture area


The maximum transmitting or receiving gain
of an antenna with effective aperture area Ae is

where c 3 x 105 km/s.


The value of Ae for a circular dish antenna
equals its physical area (DA/2)2 multiplied by
the aperture efficiency 0.5 0.70.

377

Note: Figure of merit


Gr/Ts is the figure of merit for a receiving
system.
Gr is the receiving antenna gain.
Ts is the system noise temperature.
System noise temperature = antenna noise
temperature + receiver noise temperature
(LNA)

378

189
Lecture notes Radio Engineering by Jorma Kekalainen

Communication link
calculations

Communication link
All communications links comprise
a transmitter,
a channel and
a receiver.
A channel might be
a radio or optical link through the atmosphere,
an optical fiber,
a coaxial cable,
a twisted pair cable, or
a telephone wire on a pole.
Links have performance attributes throughput/capacity,
error rates, bandwidth etc.
The type of channel determines the design of the
transmitter and the receiver.

380

190
Lecture notes Radio Engineering by Jorma Kekalainen

Shannons model

381

Elements of a communication link

Send messages or information from a


source to a destination

382

191
Lecture notes Radio Engineering by Jorma Kekalainen

Classes of transmission media

383

Transmission media
There are basically two types of media, wired or
wireless.
Here, we are dealing with the wireless medium
only.
Electromagnetic waves travel through the earths
atmosphere in the following four ways:
Ground wave propagation
Ionospheric propagation
Tropospheric scattering
Line of sight

384

192
Lecture notes Radio Engineering by Jorma Kekalainen

Ground wave propagation


Frequencies below 30 MHz propagate along the earths
curvature guided by the surface and are called ground waves.
This guided wave has two main components:
The direct wave, and
The reflected component
This mode is also called ducting.

385

Ionospheric propagation
Frequencies between 3 and 50 MHz are
reflected by the ionosphere and can travel
much further than ground waves.

386

193
Lecture notes Radio Engineering by Jorma Kekalainen

Tropospheric scattering
Above VHF, we see a phenomena where the signals are
scattered by the troposphere.
The scattered waves are very weak, but they can be received
and demodulated.
This mode of media behavior is called tropospheric scattering.

387

Line of sight
This is the mode for satellite communications and also for
terrestrial microwave links.
For frequencies above 3 GHz to about 10 GHz, the earths
atmosphere offers practically no degradation.
Frequencies above 10 GHz suffer from oxygen and water
vapor absorption and cause degradation.

388

194
Lecture notes Radio Engineering by Jorma Kekalainen

Link budgets
The designer hopes that the signal will complete its trip
through link chain with just enough power to be decoded at
the receiver with the desired signal quality.
With digital signals, the quality is measured by the Bit Error
Rate (BER).
If we want our signal to have a low BER, we would start it out
with higher power and then make sure that along the way it
has enough power available at every stage to maintain this
BER.
The signal can get extra power infusion along the way from
intermediate amplifiers such as microwave repeaters for
telephone links or from satellite transponders for satellite links.

389

How we specify the quality of signal


transmission
The BER, as a measure of the signal quality, is the
most important figure of merits in all link budgets.
The BER is a function of a quantity called Eb/N0, the
bit energy per noise-density of the signal.
For a QPSK signal in an Additive White-Gaussian-
Noise (AWGN) channel, the BER is given by

390

195
Lecture notes Radio Engineering by Jorma Kekalainen

Erfc
According to the previous formula the BER of
any signal is related to its Eb/N0 by the
complementary error function

Erfc-function is found tabulated in most


telecommunications textbooks and is available
as a built-in function in most math programs.
391

The above equation when plotted has a classic


waterfall shape when plotted on a log-log
scale.
The BER is inversely related to Eb/N0.
Higher Eb/N0 means better quality or smaller
BER.

392

196
Lecture notes Radio Engineering by Jorma Kekalainen

BER of a signal

393

Signal power vs. noise power


Instead of monitoring just the signal power we
will actually monitor a ratio of the signal power to
the noise power injected along the way.
This means that in doing link budgets, we will
keep track of not just the signal power and how it
is getting attenuated, but also where the noise is
entering into the link and how much.
This dual role makes things complicated because
there are numerous sources of noise.

394

197
Lecture notes Radio Engineering by Jorma Kekalainen

Eb/N0 - a measure for digital links

Eb/N0 is the most common parameter used to


compare communication systems even when
they have differing bit rates, modulations, and
media.
Lets take a closer look at the Eb/N0.
The quantity Eb is a measure of the bit energy.
The energy is the capacity to do work and
energy expended per time is called power.

395

Bit energy
To compute Eb, we divide the average signal power by
its bit rate

This makes sense because the average power is the


energy per unit time, and the bit rate is the number of
bits per unit time.
So the division removes the units of time leaving
energy per bit.
We can also write the above equation in an alternate
amplitude-squared form with the amplitude-squared
representing the Pavg

396

198
Lecture notes Radio Engineering by Jorma Kekalainen

Noise density
In the denominator of Eb/N0, the quantity N0 is called
the noise density.
It is the total noise power in the frequency band of the
signal divided by the bandwidth of the signal.
Noise density is measured as Watts/Hz and is the
noise power in one Hz of bandwidth.

where PN = noise power and BN = noise bandwidth.


The units of N0 are Joules.

397

C/N and C/N0 - a measure of analog


links
For analog signals, we use a quantity called C/N0 in the
same way as Eb/N0, where C is the signal power.
C and Eb are related by the bit rate.
So you will typically see C/N0 specified for the analog
portions (or the passband signals) of the link and Eb/N0
for the digital (or the baseband) portions.
C in the expression (C/N) is simply the carrier power in
the whole useable bandwidth, where C in the
expression (C/N0) is carrier power per unit bandwidth.

398

199
Lecture notes Radio Engineering by Jorma Kekalainen

Relations between Eb/N0 and C/N0


and C/N
Previously, we know that
C = Energy per bit x bit rate
= Eb x Rb, from which we get

Note both C/N0 and Eb/N0 are densities so we do not need to


specify the bandwidth of the signal.
But to convert (C/N0) to (C/N), need to divide by the signal
bandwidth

N=N0B
399

Link and link equation


A link consists of three parts.
1. Transmitter
2. Receiver
3. Transmission media
The very simplest form of a link equation is
written as
Preceived = Power of the transmitter + Gain of
the transmitting antenna + Gain
of the receiving antenna - Sum of
all losses

400

200
Lecture notes Radio Engineering by Jorma Kekalainen

Transmitter

A transmitter receives baseband data,


modulates onto a higher frequency carrier,
amplifies it and broadcasts it via an antenna.
The two main items that are associated with
transmitters are
1. Flux density
2. EIRP

401

Flux density
The flux density is a measure of energy that is available
for gathering from a particular source.
It is sometimes called the Radio Power of a Source.
The flux density is defined by

where G = gain of the transmitting antenna and P =


transmitter power in watts.
The flux density is a measure of the amount of energy that
is received at a distance r from a transmitter of gain G and
transmit power P watts.

402

201
Lecture notes Radio Engineering by Jorma Kekalainen

Amplifier + Antenna = Transmitter


A very basic transmitter consists of an amplifier and an antenna

The amplifier puts out a certain amount of power and the antenna is
said to have a particular gain that further amplifies this power.
This combination is here called the transmitter.
Usually, lossy elements such as wires connect these two components
in the preferred direction of radiation of the antenna.
These losses are included in the quoted EIRP figure for the
transmitter.

403

EIRP Effective Isotropic Radiated


Power
EIRP is closely related to the flux density.
Where flux density is energy as measured a distance
away from the source, EIRP is a measure only of the
transmitted power.
For a transmitter EIRP is defined as the combination
of
EIRP = Power of transmitter x Gain of the
antenna
= Pamp x Gantenna

404
Note: EIRP is the numerator of the flux density equation.

202
Lecture notes Radio Engineering by Jorma Kekalainen

Note
There are two hidden assumption in EIRP.
First is that the transmitter is putting out the
maximum power that it can, and second, that
the EIRP figure is delivered at the antennas
boresight.
So if you happen to have your antenna pointed
not quite straight into the boresight of the
transmitting antenna then you will not get the
quoted EIRP.

405

Received power
EIRP and the flux density both tells us something
about a transmitter but nothing about what is
actually received.
To compute power received by a receiver at a
distance r from the source, we need to multiply
the flux density with the receiving antennas area.
The flux density is energy per unit area per unit
time.
The only useable part of this energy is what is
accepted by the receiving antenna.
406

203
Lecture notes Radio Engineering by Jorma Kekalainen

Effective area
So the power received is equal to the flux density
times the receiving area

The effective receiving area (not actually a physical


area but strongly related to it) of any antenna is
defined by

where GR is the gain of the receiving antenna and is


the wavelength.

407

Power equation
Now we can write the expression for computing the received
power as

This equation says that if we know the gain of the receiving


antenna, the EIRP (= GTPT) of the transmitter, the operating
frequency, and the distance between the two, then we can
calculate the received power.
We can rewrite the above in dB as

408

204
Lecture notes Radio Engineering by Jorma Kekalainen

Free Space Loss (FSL)

The last portion of the expression

containing the ratio of the distance r to the


wavelength , i.e. the number of wavelengths
in the distance, is called the Free Space Loss
(FSL).

409

Noise
This topic causes a lot of confusion, particularly when dealing
with link budgets.
As we can see, our important parameters Eb/N0, C/N, C/N0 all
have this noise term on the denominator.
All objects not at absolute zero emit electromagnetic radiation.
The band of frequencies emitted are a function of the
temperature of the object.
E.g. a light bulb emits many different frequencies.
Most of its radiation is in the range of infra-red light and
ultraviolet frequencies which we can see and feel.

410

205
Lecture notes Radio Engineering by Jorma Kekalainen

Noise
The sun puts out visible noise in the light wave
frequencies among many others that we can
not see such as X-rays and infra-red.
The noise coming to us from the galaxies is
typically in microwave frequencies.
The statistics of this noise is well described by
quantum physics.
The black body radiation problem was first
solved by Max Planck in 1901.

411

Spectral radiance
According Planck's law
hf
G f , T hf
e kT
1
where G(f,T) is the spectral radiance or the energy per unit
time (or the power) radiated per unit area of emitting surface in
the normal direction per unit solid angle per unit frequency by
a black body at temperature T; h is the Planck constant; k is the
Boltzmann constant; f is the frequency of the electromagnetic
radiation; and T is the absolute temperature of the body.

412

206
Lecture notes Radio Engineering by Jorma Kekalainen

hf vs. kT
For radio and microwave frequencies, the factor hf is quite
small relative to the factor kT in the nominal range of room
temperatures, say 290 degrees in Kelvin.
Thus the exponential function in the expression can be
approximated by the first two terms.

413

Thermal noise power


Now we get

We can interpret that this is the power at frequency f, provided


f is small enough such that hf << kT.
Thus it seems appropriate to refer to the low-frequency noise
as thermal noise (it is proportional to the temperature) and the
high-frequency noise, from somewhere sub-optical on up, as
the quantum noise, since it is proportional to Plancks
constant h.

414

207
Lecture notes Radio Engineering by Jorma Kekalainen

Note

415

Thermal noise

416

208
Lecture notes Radio Engineering by Jorma Kekalainen

Note: Quantum noise


When f becomes large enough that this kT
approximation no longer holds, the frequencies
are in the generalized optical (i.e., infrared,
visual optical, ultraviolet, and X-ray, )
range.
In this frequency range, the higher order hf
terms eventually dominate the frequency-
dependent noise called quantum noise.

417

Noise definition 1
The thermal noise power in a bandwidth B is
kTB.
We can define noise density simply as the
power of the noise signal divided by its
bandwidth

418

209
Lecture notes Radio Engineering by Jorma Kekalainen

Noise definition 2
We can define noise density

This definition is consistent with widespread usage,


where T is considered to be the system temperature,
not the ambient or room temperature.
In the radio and microwave bands, the spectral
density is taken as N0 for a one-sided spectrum, and
as N0/2 for a two-sided spectrum.
The noise power will be N0B in all cases where this
convention is applied.
419

Noise power
Lets now set these two definitions for noise
density N0 equal

So, we can write

where k = Boltzmanns constant 1.38 x 10-23


Joules/Kelvin, T in Kelvins and BN in Hertz, and
PN in Watts.

420

210
Lecture notes Radio Engineering by Jorma Kekalainen

Noise bandwidth
There is no confusion when talking about bit rates but
unfortunately there are many different ways of defining
bandwidth.
The most common definition of bandwidth is the distance
from one pass-band edge to another where the edge is
defined as the 3 dB below the maximum.
What is the noise bandwidth in the equation for noise?
In simple terms, it is the noise that is allowed to enter into
the system by the receive filter.
The idea of noise bandwidth is to cover all power passed-in
by the filter.
Generally, the noise bandwidth is approximately 1.12 times
the 3 dB bandwidth of the signal.

421

G/T
G/T is a very important parameter of receivers.
The T is the thermal noise temperature of the receiver and it
impacts the ability of the receiver to see a signal in the noise.
Just as we characterize a transmitter by its EIRP, we use G/T in
a similar way to specify receivers.
In dB, G/T is the difference in the gain of the receiving
antenna gain and its noise temperature

where GR = Gain of the receiving antenna and


T = Thermal noise temperature of the receiver

422

211
Lecture notes Radio Engineering by Jorma Kekalainen

Note
This handy G/T variable allows us to compare
receiving systems of all kinds.
A G/T of 15 dB is better than 10 dB.
This parameter is usually given for earth stations
as well as for satellite receivers, and does not
need to be calculated.
In doing link budgets, we will assume that it is
given.
Actually calculating the G/T or the Noise Figure
of a receiver is a topic in itself and worthy of the
own story.
423

Losses experienced in LOS links


The losses experienced by the signal fall into
these categories
Free Space Loss (FSL)
Rain
Antenna misalignment
Gaseous absorption

424

212
Lecture notes Radio Engineering by Jorma Kekalainen

Computing free space loss


For line of sight links, this loss is a function of the square of
the distance.
For a signal going from ground to the satellite, the free space
loss is largest of all other types of losses.
It can be simplified and written as

where r is the distance and f the frequency.

425

Rain
Signal attenuation due to rain is the second most
significant after free space loss.
It is particularly significant for higher SHF frequencies.
We have to deal with rain losses for both uplinks and
downlinks.
It also varies a great deal from location to location since
it is a function of the rain rate.
The attenuation can vary from 0.1 dB in California to
over 10 dB in Seattle.
Providing for this large attenuation in satellite links
results in over design of the system for areas which
have little rain.

426

213
Lecture notes Radio Engineering by Jorma Kekalainen

Countermeasures for rain attenuation

Accommodations are made by providing ground or


site diversity, which just means that there are two or
more receivers instead of one which may or may not be
geographically separated.
Ground diversity such as having another ground station
located a few kilometers away in a rainy region can
reduce the rain attenuation by more than half.
Other ways to accommodate for location-specific rain
attenuation is to allow higher power for the
transmitters and variable error correcting codes and
variable data rates.

427

Rain models
There many popular rain models that help us
compute the rain loss.
E.g., these are
NASA Rain Attenuation Model
CCIR Rain Attenuation Model
The result from each can vary by 1-2 dB
depending on how the regions are defined and
estimate of rain rate is made.

428

214
Lecture notes Radio Engineering by Jorma Kekalainen

Antenna misalignment
The antenna gain plays an important role in the link
calculations and we have assumed that the receiving
antenna and the transmitting antenna are oriented perfectly
so that the maximum gain of the receiving antenna is
aligned with the uplink.
The gain varies a great deal off the bore sight and as shown
in the figure below, unless we have perfect alignment, we
are going to have losses associated with this.
There are two parts to this loss.
One is at the transmitter, if its antenna is not pointed to deliver
maximum gain and
the second is at the receiving antenna is not pointed to receive
the maximum gain.
Antenna pointing is an important operation and calibrations
are performed when the system is set up.
429

Antenna misalignment
To receive maximum EIRP, both receiving and
transmitting antenna have to be aligned
perfectly

430

215
Lecture notes Radio Engineering by Jorma Kekalainen

Gaseous absorption

The attenuation caused by clouds and fog due


to the phenomena of gaseous absorption has to
be considered particularly for frequencies
above 10 GHz.
These effects are primarily due to amplitude
reduction which reduces signal power, but
water can also affect the polarization of
signals.

431

Combining link power and noise


Now we write the link equation in terms of
C/N.
Carrier power (C) is the sum of the EIRP of
the transmitter, the gain of the receiver and any
associated implementation losses all expressed
in dB
C = EIRP + G Losses
So that is all the power available to the signal.

432

216
Lecture notes Radio Engineering by Jorma Kekalainen

(C/N)dB and (C/N0)dB


Remember that
PN N = k T BN =N0BN
In dB form we can write
(C/N)dB = EIRP + G/T - Losses - k - BN
If we convert C/N to C/N0 form
C/N=C/(N0BN) C/N0=(C/N)BN
(C/N0)dB = EIRP + G/T - Losses - k
Note: The bandwidth has dropped out. C/N0 is then a figure that is independent
of bandwidth. However, you will hear much more often about C/N (also referred
to as SNR) that C/N0. 433

Eb/N0

C/N is used for analog signals, and Eb/N0 for


digital data signals.
Conversion to Eb/N0 requires knowledge of the
symbol/bit rate and C/N the knowledge of the
bandwidth.
Summarizing these conversions in dB form
Eb/N0 = C/N0 - Rb
C/N = C/N0 - BN

434

217
Lecture notes Radio Engineering by Jorma Kekalainen

Symbol rate and bit rate


Nowadays nearly all the BER curves are given as a function
of the bit energy or Eb/N0.
However, the modem part of the communications systems
operate on the basis of symbol rates and not bit rates.
In a binary case, a symbol consists of just one bit, so the bit
rate of a binary system is equal to the symbol rate.
In a QPSK system, one symbol represents two bits, so here
the bit rate is twice symbol rate.
In 8-PSK system, one symbol stands for three bits, so now
the bit rate is three times the symbol rate.
The advantage is of these higher order modulations is that
higher bit rates can be contained in the same bandwidth as
the binary signal.

435

Higher order modulations


We can make a symbol stand for as many bits
as we want but since we find that the BER
increases faster than the bit rate, which
requires more power to overcome.
If the system is bandwidth-limited and but has
plenty of power, then these higher order
modulations make sense but e.g. for satellites
which are power limited (due to their being
weight-limited), they are not used.

436

218
Lecture notes Radio Engineering by Jorma Kekalainen

Coding gain
E.g. in bank transactions over the line to an another
bank, an error-free transmission is wanted.
Typically a data transmission at about 10-12 BER is
considered nearly error free or quasi error-free using
the official term.
To get a picture of 10-12 BER, this is equivalent to about
one error per hundred million pages of text transmitted.
A typical satellite link can provide, for a reasonable
power level, only about 10-3 BER.
This is equal to about ten errors per page.
This may be OK if this a voice signal or music but data
at this error rate is quite unacceptable.

437

BER vs. Eb/N0

438

219
Lecture notes Radio Engineering by Jorma Kekalainen

Error correction codes


Nowadays error correction codes (ERC) is nearly
always used on all types of links.
There are many different varieties of ERC
suitable for all different types of link.
In all cases, the use of coding is a trade-off
between power and payload.
The bits used to convey coding cannot be used for
information, effectively reducing the useable bit
rate for a given bandwidth and power level.
439

Simple coding example


Consider e.g. a group of 7 bits.
We assign the last bit as a parity bit.
This means that when the blocks of bits is checked at the
receiver, if the first six bits are even (equal number of 0 and 1)
then, the parity bit is set to 0, and if odd then it is set to 1.
Now we can tell if a received block is erroneous.
So by using a parity bit, we now have a way to tell if the block
has a bit error.
If the first six bit of the received block have an even number of
1s, as in the following figure, we know there has been an error
somewhere.

440

220
Lecture notes Radio Engineering by Jorma Kekalainen

Simple coding example


This type of code is called a block code because
it deals with a group of bits at one time.
It is said to have a rate of 6/7, as the code uses 1
out of 7 bits for coding purposes and only 6 out 7
bits can be used for information.
In return for an Eb/N0 of 10 dB, it gives us an
approximate BER of 10-9 instead of the 10-5 we
would have gotten otherwise.
So in return for a loss of about 15% of the bit rate
to coding overhead, we get coarsely 5 dB
improvement in Eb/N0.

441

BER curves
Lets take a look at the
following BER curves.
The right most curve is
for an uncoded QPSK
signal, the other two for
coded links. BER
First thing we see is that
the use of valid coding
moves the BER curve to
the left.
Otherwise there would be
no point to use error Eb/N0(dB)
correcting codes.

442

221
Lecture notes Radio Engineering by Jorma Kekalainen

BER curves
These curves say that
we can get the same
BER but for a lesser
Eb/N0.
Of course, our effective BER
data rate is reduced.
But we wont worry
about the reduced data
rate, if we can live with
it.
Eb/N0(dB)

443

Example
We see that for a desired
BER of 10-5, an
uncoded QPSK signal
requires an Eb/N0 of 9.6
dB. BER

Now if we add coding,


for the first code we
need only 4.4 dB and
the second code requires
Eb/N0(dB)
only 3.4 dB.

444

222
Lecture notes Radio Engineering by Jorma Kekalainen

Coding gain
Coding gain is the difference between Eb/N0 needed for
an uncoded system and Eb/N0 needed for a coded
system.
We define the coding gain as
Coding Gain = Uncoded Eb/N0 - Coded Eb/N0
for any particular BER.
The coding gain is a function of the BER level chosen.
So before the coding gain calculation we need to know
what BER level is desired for the link.
E.g., a BER of 10-5 can be chosen for voice and 10-10
10-12 for data links.
445

Coding gain
The coded BER curves
shown beside are usually
produced by testing or
simulation.
In scientific literature there
can be found BER curves for
BER
various types of codes.
In most cases this loss in
useful bit rate is accepted for
reduction in transmit power.
This is true now for all types
of links including wireless Eb/N0 (dB)
phones as well.
446

223
Lecture notes Radio Engineering by Jorma Kekalainen

Link margin

The goal of the link budget exercise is to have


a reasonable margin for the chosen data rate,
bandwidth and EIRP and the G/T figures.
Often some adjusting and readjusting is needed
to get the desired link margin.

Margin

447

Note
Other factors such as interference from
adjacent channels and adjacent links will affect
the link budgets.

448

224
Lecture notes Radio Engineering by Jorma Kekalainen

Microwaves

General description and


definitions

Applications
Microwaves are conventionally encountered in
three kinds of applications:
1. radar, used for detection and measurements;
2. radiocommunications, for point-to-point links,
most particularly for satellite and space
communications;
3. heating, drying, cooking of many different types
of materials.

450

225
Lecture notes Radio Engineering by Jorma Kekalainen

Applications
The majority of applications of microwaves are related
to radar and communication systems.
Microwave communication systems handle a large
fraction of the worlds international and other long-haul
telephone, data, and television transmissions.
Most of the current wireless telecommunications
systems, such as terrestrial radio links, satellite
communications , wireless local and personal or body
area networks (WLAN, WPAN, WBAN), cellular
systems, and global positioning satellite (GPS) systems,
operate in the UHF EHF range, and thus rely on
microwave technology.

451

Applications
Information applications are in the field of information
acquisition and transfer.
In radar, microwaves permit acquisition of certain
information, which is contained within the echo signal from a
target.
The location and characteristics of the latter are determined by
comparing the signal received after reflection to the emitted
signal.
In communications, microwaves provide the support to the
information which is to be transmitted from very small to very
large distances.

452

226
Lecture notes Radio Engineering by Jorma Kekalainen

Applications

Energy applications utilize a most remarkable


property of microwaves, that is in-depth
heating.
This is an energy transfer application, which
has found a number of uses in everyday life,
from the kitchen to the medical treatment of
various ailments.

453

Microwave field
The microwave field covers quite a wide range of
applications of very diverse nature.
Many different problems are found in the field.
On the fundamental level, basic theories quickly lead
to very complex mathematics.
In rather sharp contrast, many practical applications
are still based on the traditional cut and try
experimental approach or/and simulation.

454

227
Lecture notes Radio Engineering by Jorma Kekalainen

Theoretical models
The same variety is encountered in the use of models at different levels of
complexity.
The study of waveguides and cavities can be done by considering the
electrical and magnetic fields, described in terms of Maxwells
equations.
The same macroscopic model is also adequate to consider the interactions
between electromagnetic fields and an electron beam, as they occur in most
generating and amplifying vacuum tubes.
For other active components, such as the solid state sources, the use of
quantum physics is required.
The characterization and measurement of components is done in terms of
circuit theory, using the scattering matrix formalism and Kirchhoffs
laws.

455

What are microwaves?


If you consult Webster you find the word
microwave defined as a very short
electromagnetic wave.
The key word in this definition is short, which
implies length (a wavelength) is involved.
The wavelength () of an electromagnetic
wave is equal to the velocity (v) of that wave
divided by its frequency (f) or =v/f.

456

228
Lecture notes Radio Engineering by Jorma Kekalainen

What are microwaves?

Therefore, the wavelength of an electro-


magnetic wave is inversely proportional to its
frequency; that is, the higher the frequency, the
shorter the wavelength.
So what are microwaves?
In simple terms, they are radio waves of a very
high frequency.

457

What are microwaves?


Exactly where the microwave spectrum begins
is a matter of some discussion.
There are those who say it begins at 300 MHz
( = 1 m in air) and those who maintain it
starts at 1.0 GHz ( = 30 cm in air).

458

229
Lecture notes Radio Engineering by Jorma Kekalainen

Compromise definition

The microwave spectrum should begin where


conventional components (lumped capacitors,
resistors, and coils) no longer exhibit their
proper parameters and where it becomes
practical and advantageous to build
(distributed) components in stripline and
microstrip, as well as in coaxial, form.

459

Skin effect
The origin of the microwave spectrum should be where
conventional components cease to display normal parameters.
But why would such a change occur?
One key reason is a phenomenon known as the skin effect.
At higher frequencies the energy of a wave tends to use only
the outside skin of a conductor for the transferring of energy
because larger and larger flux are set up around a conductor,
increasing its inductance.
By once the inductive reactance (XL) increases when the
frequency does, thereby offering a larger resistance to the
energy.
Thus the high frequency component searches for the path of
lowest resistance on the outer skin of the conductor.
460

230
Lecture notes Radio Engineering by Jorma Kekalainen

Skin effect
Because of this phenomenon the energy travels only on the outside
edge of the conductors; lumped components thus become very
unpredictable and impractical at microwave frequencies.
It has been observed that some capacitors actually look inductive at
high frequencies.
The capacitance set up in coils and wire-wound resistors when used
at microwave frequencies completely cancel out the intended
functions of each of these components.
Special materials, usually of a ceramic base, are used in the
production of capacitors and resistors for use at the higher
frequencies.
Coils are usually made of high impedance lines cut to a specific
length.
461

Official definition: Microwave region

The frequency range extending from 300 MHz up


to 300 GHz is generally known as microwaves,
which thus characterize signals having between
300 million and 300 billion periods per second
(from 3 108 to 3 1011 Hz).
These limits are to some extent arbitrary: they
permit to position the microwave domain between
the waves used for conventional radio and TV
diffusion (lower ranges) and infrared rays (higher
ranges).

462

231
Lecture notes Radio Engineering by Jorma Kekalainen

UHF, SHF, and EHF decades


A finer designation of the frequency bands divides
them into decades.
The microwave range more or less covers the
three decades called
ultra-high frequencies (UHF),
super-high frequencies (SFH), and
extra-high frequencies (EHF).
The upper ranges of electromagnetic radiation are
known as infrared, visible rays, and ultraviolets.
Microwaves are divided into still narrower bands,
corresponding to specific wave-guide sizes.
463

US MICROWAVE BAND

464

232
Lecture notes Radio Engineering by Jorma Kekalainen

Energy of a microwave photon


In quantum physics, electromagnetic radiation
is considered to be a flow of photons of energy
hf, where f is the frequency and h is Plancks
constant ( 6.6310-24 Js).
A microwave photon thus has an energy
located in the range between roughly 1.2 10-6
and 1.2 10-3 eV.

465
1 eV=1e*1V=1.6*10-19 As*1V=1.6*10-19J

Magnitude of the periods and


wavelengths
The period T of a microwave signal, defined as
the inverse of the frequency f is located
between 3 ns (nanoseconds) and 3 ps
(picoseconds).
The wavelength of a microwave signal for an
electromagnetic wave propagating across
space at the velocity of light c0 = 2.997925 ...
108 m/s is defined by = c0/f= c0T,
somewhere between 1 m and 1 mm.

466

233
Lecture notes Radio Engineering by Jorma Kekalainen

Remark
The term microwaves, used to define a range
of frequencies, denotes the smallness of the
wavelength encountered, as compared to those
utilized for conventional radio and television.
The use of the comparative prefix micro may
lead to confusions here too: one would expect
microwaves to denote wavelengths in the
micrometer range and not, as is actually the
case, in the millimeter to meter ranges.

467

Subdivisions of the EM spectrum.

468

234
Lecture notes Radio Engineering by Jorma Kekalainen

Alternate names: Decimeter,


centimeter, and millimeter waves

The three decade ranges can also be called


decimeter waves (UHF),
centimeter waves (SHF), and
millimeter waves (EHF).

469

Division, nomenclature, and allocation of microwaves

470

235
Lecture notes Radio Engineering by Jorma Kekalainen

Frequency allocation
Use of radio frequency bands is regulated by governments.
ITU works on standards for frequency allocation.

471

Dimensional comment
The wavelength of a microwave signal is of the same
order of magnitude as the devices used to produce it
and to transmit it.
It is not possible to assume that devices are merely
dimensionless points in space, as is done in circuit
theory approximations.
Also, the term voltage is not defined in a unique way,
since the electric field does not derive from a scalar
potential.
On the other hand, it is neither possible to assume that
devices become large with respect to wavelength, as is
the case for geometrical optics.

472

236
Lecture notes Radio Engineering by Jorma Kekalainen

Wave theory

Microwave problems must be considered in


terms of electric and magnetic fields, as
defined in Maxwells model.
Microwaves actually helped to demonstrate the
validity of this model.
The study and utilization of the microwave
domain required the development of specific
new techniques and devices.

473

Comparable field: Acoustics


Since the velocity of sound in air is of the order of 300 m/s,
roughly one million times smaller than the velocity of light,
wavelengths between 1 mm and 1 m correspond, in
acoustics, to frequencies between 300 Hz and 300 kHz.
This range contains most known acoustic sources, in
particular the human voice and music.
A similarity does therefore exist between microwaves and
acoustics, two fields in which devices have the same size as
wavelengths.
Certain methods developed for acoustical applications can
be directly transposed to microwaves and vice versa.

474

237
Lecture notes Radio Engineering by Jorma Kekalainen

Microwaves

Properties of microwaves

Bandwidth
The very wide frequency bands available at microwaves are most
favorable for radio communications.
The rate of transmission of a channel being directly proportional to
its bandwidth, a simple calculation shows that over the 300 MHz to
300 GHz frequency range, 999 times more information can be
transmitted over a specified time period than in all the lower
frequency bands taken together.
As a result, the use of microwaves permits meeting the increasing
need for communications channels.
This particular property is directly related to the signal frequency.
Following the same line of reasoning, even much larger amounts of
information could, in principle, be transmitted over the infrared and
visible spectra, using lasers and optical fiber systems.

476

238
Lecture notes Radio Engineering by Jorma Kekalainen

Transparency of the ionosphere

The ionosphere forms a group of ionized


layers of electron plasma, which surround the
earth at altitudes ranging from 50 to 10,000
km.
These ionized layers are produced by the
impact of solar radiation, so that their
parameters (density, height) fluctuate widely
between day and night, exhibiting seasonal
variations and changes related to the solar
activity. 477

Transparency of the ionosphere


The electromagnetic propagation within the ionosphere is similar to that in
a waveguide.
Signals at frequencies lower than 10 - 40 MHz (cut-off frequency) are
partially or totally reflected.
This property is used to realize multiple reflection links at short waves.
Higher frequency signals travel across the ionosphere, but experience
distortion, which decreases with frequency.
Microwave signals, well above the ionospheric cut-off, are hardly affected
at all, at sufficiently small power levels.
For these reasons, microwaves are utilized for satellite communications and
space transmissions.
Radioastronomy deals with microwave radiation emanating from distant
stars, galaxies, and quasars.
A quasi-stellar radio source ("quasar") is a very energetic, distant, and compact
region in the center of a massive galaxy surrounding its central supermassive black
478
hole.

239
Lecture notes Radio Engineering by Jorma Kekalainen

Partial transparency of the


atmosphere
The various atmospheric components (oxygen,
nitrogen, water vapor, carbon dioxide) and the
many elements in suspension (water droplets,
ice crystals, dust, smoke) do not significantly
affect signals having frequencies smaller than
about 10 GHz.
Higher frequency signals experience several
unwanted effects: absorption, depolarization,
and scintillation.
479

Electromagnetic noise
The noise picked up by an antenna directed skyward, in the absence of
signal, goes through a relatively flat minimum over the 1 to 10 GHz
frequency range.
The received noise power is the product of the equivalent noise temperature
(in Kelvin) by Boltzmanns constant (kB =1.3804 10-23 J/K) and by the
receivers bandwidth.
Over the 1-10 GHz band, the noise temperature decreases below 10 Kelvin.
In practical terms, this means that it is within this frequency range that
signals of the lowest amplitude can be detected, and thus the most sensitive
receivers can be designed.
For instance, signals of extremely low levels received after transmission
across planetary space are often in the neighborhood of 3 GHz.
A further requirement for the receiver is not to produce too much noise,
which would degrade the input signal-to-noise ratio.

480

240
Lecture notes Radio Engineering by Jorma Kekalainen

Directivity of antennas
The width of the beam radiated by an antenna is directly
proportional to the ratio of the wavelength to the antennas
largest dimension.
When transmitting a signal from one point to another
(microwave link), or when determining the origin of a
reflection (radar), a narrow beamwidth is required.
It is then either necessary to have a large size antenna, which is
often not convenient for mechanical reasons, or to utilize a
signal of high frequency.
Microwaves are well suited for such applications.

481

Note
Even narrower beamwidths are obtained using
signals in the visible range (laser).
As the beam is then extremely narrow, it must
be directed quite accurately towards the
detector, and pointing problems may become
significant.

482

241
Lecture notes Radio Engineering by Jorma Kekalainen

Reflection on targets
The effective reflection area of an object depends in a
very sensitive manner on the ratio of the objects size
to the wavelength.
When the reflecting element is much smaller than the
wavelength, the reflection becomes vanishingly
small.
On the other hand, when the wavelength becomes
much smaller than the object, the effective reflection
area for a metallic element is approximately its cross
section transverse to the beam.
483

Remark
Centimeter waves thus detect objects of meter
size, but are not affected by raindrops.
The latter can, on the other hand, perturb the
detection at millimeter waves.

484

242
Lecture notes Radio Engineering by Jorma Kekalainen

Interaction with matter

When an electromagnetic wave impinges on a


material sample, it is preferentially absorbed at
particular frequencies, which are the resonant
frequencies of the material.
The resonances observed at microwaves
depend on the molecular composition of
matter.
This effect is put to good use in chemical and
physical analyses.
485

Interaction with matter


In particular, water strongly absorbs all
microwaves, a specific property which leads to
two types of applications:
1. microwave heating, utilized for the cooking of
food, the drying and thermal processing of
numerous materials, and the medical treatment of a
number of ailments by hyperthermia;
2. the detection and measurement of moisture
contained within materials.

486

243
Lecture notes Radio Engineering by Jorma Kekalainen

Non-ionizing radiation

The energy of a molecular bond is larger, by


several orders of magnitude, than the energy
belonging to a microwave photon.
This means that a photon, at microwave
frequencies, does not possess sufficient energy
to break a chemical link, for instance by
photoelectric effect.
Microwaves are thus a non-ionizing form of
radiation.
487

Note
At frequencies far above microwaves, a photon
can have enough energy to extract an electron
and produce ionization.
This happens in the visible spectrum,
ultraviolet (suntan), X and gamma rays.

488

244
Lecture notes Radio Engineering by Jorma Kekalainen

Stable oscillation frequencies

The most stable known atomic oscillators,


hydrogen, cesium, and rubidium exhibit
extremely stable oscillations within the
microwave range.
As a result, all atomic clocks and frequency
standards make use of microwaves.

489

Note
The set of properties listed previously makes
microwaves a privileged field for a large
number of applications, such as satellite
communications and radar, but also heating
and measurement.
Some of these applications simply could not
exist without microwaves.
For others, microwaves provide the best
compromise among various requirements to be
satisfied simultaneously.
490

245
Lecture notes Radio Engineering by Jorma Kekalainen

Microwaves

Historical landmarks

First experiments in radio


communications
The spark generators used by Hertz and
Marconi for their first radio transmissions in
1888 and 1890 were non-coherent emitters,
generating a very broad spectrum of noise
extending well into the microwave range.

492

246
Lecture notes Radio Engineering by Jorma Kekalainen

First waveguide

In 1894, Oliver Lodge surrounded a spark


generator with a metal tube and noted that the
resulting radiated signal exhibited directional
properties.
This effect was at the time considered a
laboratory curiosity, and it did not lead to
practical applications.
With the advent of vacuum tubes, the further
development of radio was hence oriented
towards lower frequencies. 493

Communications experiments
The first practical experiments carried out with
the purpose of using microwaves for the
transmission of information are credited to
George Southworth, at the Bell Telephone
Laboratories, in the US during the 1920s and
1930s.
Wave propagation along copper water pipes
was investigated.
Since then, users of microwaves have often been
nicknamed plumbers.
494

247
Lecture notes Radio Engineering by Jorma Kekalainen

Development of radar

The advent of microwave techniques is quite


closely linked to the development of radar
during World War II.
The basic principle, already outlined by Robert
Watson-Watt around 1930, makes use of the
electromagnetic echo produced by a target.

495

Development of radar
A short signal pulse is launched, the time
between its departure and the return of a
reflected signal is measured.
To ensure an adequate detection of targets,
signals of increasingly higher frequencies were
found to be needed.
The magnetron was developed for this
purpose, being the first tube having a high
power generating capability at microwaves.
496

248
Lecture notes Radio Engineering by Jorma Kekalainen

Development of radar
The design and the industrial fabrication of radar
systems started around 1940.
Radar played a significant role during World War II,
among others in the well known Battle of Britain.
Some of the research and development work was de-
classified at the end of the war.
Particularly worth mentioning is the 25-volume set
published by the Massachusetts Institute of Technology
(Radiation Laboratory Series), covering all aspects of radar
design, including the foundations of microwaves.

497

Development of microwave tubes


Microwave generators were at first vacuum
tubes, specifically designed for use within
radar systems.
The magnetron is based on the interactions of
an electron flow within crossed electric and
magnetic fields, an effect studied since the
1920s.
It became operative at the beginning of World
War II.
498

249
Lecture notes Radio Engineering by Jorma Kekalainen

Development of microwave tubes

The klystron was invented in 1935 by Russell


and Sigurd Varian.
Many different kinds of microwave generators
were developed later on, the most significant
ones being
the backward-wave oscillator (BWO) and
the traveling wave tube (TWT).

499

Ferrite devices
The first non-reciprocal linear passive device
appeared in 1956, a gyrator.
The many isolators and circulators developed since
then are mostly devices for protection, decoupling,
and control.
They are nowadays found in most microwave
systems.
An important property of a gyrator is that it inverts the current-voltage characteristic of
an electrical component or network.
An isolator is a two-port device that transmits microwave or radio frequency power in
one direction only.
A circulator is a passive non-reciprocal three - or four-port device, in which a microwave
or radio frequency signal entering any port is transmitted to the next port in rotation only.
500

250
Lecture notes Radio Engineering by Jorma Kekalainen

Satellite communications
The first satellite, Sputnik I, was placed in orbit on
October 4 1957.
In 1962 Telstar was launched, the first
communications satellite placed in a low earth orbit.
Three years later, in 1965, the satellite Early Bird was
placed in a geostationary orbit (remaining over a
fixed location on the equator).
Since then, successive generations of satellites are
playing an important role in communications, mostly
for international links, but also within domestic
networks.
501

Microwave solid state two-ports


Active semiconductor devices appeared on the market
in the 1960s, replacing vacuum tubes as microwave
sources for low and medium power levels.
The first device of this kind developed was the Gunn
diode, based on a physical phenomenon discovered
by J.B Gunn in 1962.
Diode generators and amplifiers were increasingly
being replaced, since the end of the 1970s, by
transistors, either bipolar or MESFETs
(MetalsemiconductorFET).
502

251
Lecture notes Radio Engineering by Jorma Kekalainen

Microwave printed circuits

Over the same time span, metallic waveguides


are often superseded by more compact circuits
in microstrip, slot line, or coplanar line,
realized by means of printed circuit
techniques.

503

Maxwells equations

Local and global form


versions

252
Lecture notes Radio Engineering by Jorma Kekalainen

Maxwells equations
Maxwells equations govern the propagation of
electromagnetic waves in various media, including
free space,
waveguides,
optical fibers, and
optical crystals.
All
electric,
magnetic,
electromagnetic, and
optical phenomena
are governed by the same fundamental laws of
electromagnetism.
505

Gauss' Law (M I)
Gauss' Law says how the electric field behaves around electric charges.
Gauss' Law can be written in terms of the electric flux density (D) and the
electric charge density (v) as:

The symbol is the divergence operator.


This local form equation is valid at any point in space.
Thus, if there exists electric charge somewhere, then the divergence of D at
that point is nonzero, otherwise it is equal to zero.
The divergence of the D field over any region (volume) of space is exactly
equal to the net amount of charge in that region.

506

253
Lecture notes Radio Engineering by Jorma Kekalainen

Note: Divergence ()
Divergence at a point (x,y,z) is the measure of the vector flow out of a
surface surrounding that point.
Divergence is a specific measure of how fast the vector field is changing in
the x, y, and z directions.
If a vector function A is given by:

Then the divergence of A is the sum of how fast the vector function is
changing:

507

Example
Consider the vector function A

Thus divergence of A at any point (x,y,z) in the space is

If we want to know the divergence at (x,y,z)=(1,2,3) then we can calculate


the divergence of A at this point is 2+6*3 = 20.
We can find the divergence at any point in space because we know the
functions defining the vector A, and then calculated the derivates (the rate
of changes).

508

254
Lecture notes Radio Engineering by Jorma Kekalainen

M I (Gauss Law)

Gauss' law says that electric field lines diverge


away from electric charges.
This means that positive charge acts as a source of
electric field.
Gauss' law means that negative charge acts as a
sink for electric field.
This all means electric field lines start and stop on
electric charges.

509

M II (Gauss Law for magnetism)


Gauss' magnetism law states that the divergence of
the magnetic flux density (B) is zero.

Since B and the magnetic field H are related by the


permeability , we note that the divergence of the
magnetic field is also zero.

The divergence of the B or H field is always zero


through any volume.
510

255
Lecture notes Radio Engineering by Jorma Kekalainen

M II (Gauss Law for magnetism)

The M II says that magnetic charge (magnetic


monopoles) do not exist.
This equation states that the magnetic field
tends to wrap around things - since the
divergence is zero the fields tend to form
closed loops.

511

M III (Faradays Law)

Faraday's law says that a time-varying


magnetic field will generate a circulating
electric field.
This means we have two ways of generating
electric fields
from electric charges (or flowing electric charge,
current) or
from a time-varying (changing) magnetic field.

512

256
Lecture notes Radio Engineering by Jorma Kekalainen

Note: Curl ()
The curl is a measure of the rotation of a vector field.
Let us take a vector field, A(x,y,z), we can write A as:

The curl of A is defined to be:

513

M IV (generalized Amperes Law)


Ampere's law says that a flowing electric current gives rise to
a magnetic field that circles the wire.
In addition to this, the generalized Amperes law also says that
a time-varying electric field gives rise to a magnetic field that
encircles the electric field - this is the displacement current
term (also called Maxwells addition to Amperes law).
This means there are two ways to generate a circulating
(solenoidal) magnetic field
a flowing electric current or
a changing electric field.

514

257
Lecture notes Radio Engineering by Jorma Kekalainen

Local (point) form versions


Maxwell's equations can be rewritten with only E and H
present.
This means we express D, B and J using E, H and E.
We can do this by using the material equations:

515

Global form version


If the local form of Maxwell's equations are true at
every point, then we can integrate them over any volume
(V) or through any surface and they will still be true.
Total charge inside the
volume V with the (closed)
boundary surface S

Elementary vector dS is ortogonal to the surface S and |dS| is the area element.
516
Elementary vector dL is parallel to the closed loop L and |dL| the length element.

258
Lecture notes Radio Engineering by Jorma Kekalainen

Maxwells equations

517

Note
The global form version can be derived from
the local form Maxwells equations using the
following theorems:
Divergence
theorem

Stokes
theorem

Note: These theorems are valid with an arbitrary vector field A. 518

259
Lecture notes Radio Engineering by Jorma Kekalainen

Example

Derive the global form version from the local


form Gauss laws (M I and M II) using the
Divergence theorem:
A surface integral


A dS

S

519

Solution
Derive the global form version (surface integrals) from the
local form Gauss laws (M I and M II) using the Divergence
theorem (Gauss theorem):

Thus

520

260
Lecture notes Radio Engineering by Jorma Kekalainen

Solution
Divergence theorem (Gauss theorem):

Thus

521

Example
Derive the global form version (line integrals) from
the local form Faradays law and Amperes law (M III
and M IV) using the Stokes theorem:

522

261
Lecture notes Radio Engineering by Jorma Kekalainen

Solution
Stokes theorem:

523

Solution
Stokes theorem:

524

262
Lecture notes Radio Engineering by Jorma Kekalainen

Electromagnetic waves
Let's look at the M III and M IV equations that govern all of
electric and magnetic field propagation:

Faraday's law says that a time-varying magnetic field will


give rise to a circulating electric field.
In practice, fields do not continually grow or continually
shrink - they oscillate about an average.
So the magnetic field oscillates - which means that also the
electric field oscillates.
525

Electromagnetic waves
Ampere's law says that a time-varying electric field generates a magnetic
field that encircles the electric field.
In the same way, the electric field will be oscillating in time, the encircling
magnetic field will be oscillating in time as well.
So, a time-varying electric field generates a time-varying magnetic field
and vice versa.
This continuous coupling between these two fields produces the
propagation of electromagnetic waves.
All electromagnetic waves (radiation) radio waves, visible light, x-rays
, consist solely of propagating electromagnetic fields.
From these two basic Maxwells equations (M III, M IV) we can determine
how electromagnetic waves propagate through medium and
how they interact with materials.

526

263
Lecture notes Radio Engineering by Jorma Kekalainen

Origin of an electromagnetic wave

527

Maxwells equations

Complex formalism and


phasors

264
Lecture notes Radio Engineering by Jorma Kekalainen

Linear systems and


eigenfunctions
In electromagnetics, we often deals with monochromatic
waves, which serves as a carrier of information in
electromagnetic communication.
The field vectors of monochromatic waves are sinusoids of
time.
More complex waveforms can always be developed over a
finite or infinite sum of sine waves.
The selection of sine waves as the basis for decomposition
results from the fact that, in linear systems the eigenfunctions
are sine waves, which move along within the system without
deformation, even though their amplitude and phase do vary
with position.
529

Phasor
Nonsidusoidal signals change shape while propagating in
dispersive systems, either through amplitude distortion or
through phase distortion.
For simplifying the algebraic manipulations, sine waves are
often represented in complex notation by means of phasors
and phasor-vectors.
To illustrate this complex formalism, we consider simple
functions

where is the angular frequency (in units of radian per second), |X|
and X are the amplitudes, and is the phase.
530

265
Lecture notes Radio Engineering by Jorma Kekalainen

Phasor
If we define the complex amplitude of x(t) by

then we can write

In complex-function formalism, we will use the complex


function

instead of the real-function

531

Phasor transformation
When this happens it is always understood that we mean the
real part of x(t), which is x(t).
In linear operations the replacement of the real form x(t) by the
complex form x(t) poses no problems, when it is understood
that the final result is the real part of the complex form result.
Actually, we can reduce algebraic manipulations with phasor
transformation:

532

266
Lecture notes Radio Engineering by Jorma Kekalainen

Example

Some transformation pairs:

533

Note
x(t) is called an instantaneous value and it is
time-function

Its complex counterpart x(t) is also time-


function.
However, phasor X is not time-function.

534

267
Lecture notes Radio Engineering by Jorma Kekalainen

Phasor-vector

Correspondingly to the previous phasor


representation, the actual physical vector field
X(r,t), function of space and time, is related to
the corresponding phasor-vector X(r) by the
relationship

where vector r represents the position in space, t


the time and the angular frequency.
The modulus of the phasor-vector X(r) is the peak
value of the corresponding physical field.
535

Non-linear operation
In operations that involve product of
sinusoidal time-functions or monochromatic
field vectors such as powers, energies, energy
densities, we must use the real form of
functions and vectors, because generally

536

268
Lecture notes Radio Engineering by Jorma Kekalainen

Example

The product of two sinusoidal functions


x(t)y(t) can be calculated in real form or using
complex time-functions or using the special
expression of phasors

where the asterisk superscript indicates the


complex conjugate.
537

Time-average
Occasionally in electromagnetics, we have to deal with waves
with extremely short periods (T=/c<10-12s).
Most detectors are unable to respond instantaneously at such a
rapid change, but they can find the time-averaged values.
It is often necessary to find the time average of the product of
two sinusoidal functions of the same frequency:

538

269
Lecture notes Radio Engineering by Jorma Kekalainen

Example

Convert Telegraphers equations

to phasor form.

539

Solution
Let us replace u(x,t) with a corresponding
rotating phasor u(x,t) in both equations

Lets do the same operation for i(x,t)

540

270
Lecture notes Radio Engineering by Jorma Kekalainen

Solution

Lets take partial derivates from rotating


phasors u(x,t) and i(x,t)

541

Solution
Lets substitute phasor partial derivates to the
original Telegraphers equations

542

271
Lecture notes Radio Engineering by Jorma Kekalainen

Note

If we choose a sine function as a base


function

543

Sinusoidal waves, phasor-vectors


The theoretical developments presented here
consider exclusively time-dependent
sinusoidal wave signals, represented in
complex notation by means of phasor-vectors.
More complex signal patterns can, in principle,
always be developed over a finite or infinite
sum of sinusoidal waves (Fourier series,
Fourier transform).

544

272
Lecture notes Radio Engineering by Jorma Kekalainen

Why sinusoidals?
The selection of sinusoidal waves as the basis for
decomposition results from the fact that, in all
systems of linear equations the eigenfunctions are
sine waves, which move along within the system
without deformation, even though their amplitude and
phase do vary with position.
In every dispersive system, nonsinusoidal signals, on
the contrary, change shape while propagating, either
through amplitude distortion or through phase
distortion.

545

X(r,t) and X(r)


The actual physical field X(r,t), function of space and
time, is related to the corresponding phasor-vector
X(r) by the relationship

where vector r represents the position in space, t


the time and =2 the angular frequency
corresponding to frequency .

Note: Underlining means that a quantity is complex (phasor or phasor-vector). Bold face
italics denote vectors. The modulus of the phasor-vector X(r) is the effective value of the
corresponding physical field. 546

273
Lecture notes Radio Engineering by Jorma Kekalainen

Maxwells equations in complex


notation
Maxwells equations, in term of phasor-vectors in a
linear, homogeneous, possibly lossy, isotropic
medium are given by

where

547

Phasor-quantities

548

274
Lecture notes Radio Engineering by Jorma Kekalainen

Note

The phasor form or the time-harmonic form


of Maxwell's equations is perfectly
legitimate, because
this form tells us how the waves behave if they are
oscillating at frequency f, and
all waves can be decomposed into the sum of
simple oscillating waves.

549

Maxwells equations

Material parameters and


boundary conditions

275
Lecture notes Radio Engineering by Jorma Kekalainen

Parameters of the medium

The terms r and r are respectively the


relative permittivity and permeability of the
medium.
The loss tangent tan = / characterizes the
attenuation of the electric field in a lossy
material.

551

Perfect electric conductor


In many metals and semiconductors, the current
density J is many orders of magnitude larger than the
displacement current D/t.
It is then often assumed that, in the first
approximation, , the conductor being then a
perfect electric conductor (pec).
However, for physical reasons the current density J
cannot become infinite: the above condition means
that the electric field E must vanish within a pec.

552

276
Lecture notes Radio Engineering by Jorma Kekalainen

Perfect magnetic conductor

Similarly, in ferromagnetic materials the


permeability becomes quite large, and is
often assumed to be infinite, as a first-order
approximation, the conductor being then a
perfect magnetic conductor (pmc).
In this instance, the induction field B cannot
become infinite, so that the magnetic field H
must vanish within a pmc.

553

Boundary conditions
On the interface separating two different materials,
none of which is a perfect electric conductor, the
tangential components of the electric and of the
magnetic field are continuous, as expressed by

where n is the unit vector normal to the interface,


directed from medium 2 towards medium 1, and
where indices 1 and 2 specify, respectively, the
medium within which the field is defined.
554

277
Lecture notes Radio Engineering by Jorma Kekalainen

Boundary conditions

When 0, the boundary conditions for the


normal components of the phasor-vectors are
automatically satisfied when equations

are met.

555

Perfect electric conductor


On the surface of a perfect electric conductor
(pec, =), the electric phasor-vector must
meet the condition

This condition is approximately satisfied on a


metallic surface (short-circuit).

556

278
Lecture notes Radio Engineering by Jorma Kekalainen

Perfect magnetic conductor

On the surface of a perfect magnetic conductor


(pmc, =) without surface current, the
magnetic phasor-vector must meet the
condition

Magnetic materials no longer have, at


microwaves, a sufficiently large permeability
to satisfy the above condition.
557

Boundary conditions on the


normal components
Boundary conditions on the normal components of the electric
field yield:

Surface charges s can only be found on the interface between


two materials when at least one of the two has a non-zero
conductivity .
Similarly, the magnetic field must satisfy

558

279
Lecture notes Radio Engineering by Jorma Kekalainen

Note: Magnetic field in a perfect


electric conductor
Since the electric field E must vanish inside of
a perfect electric conductor, it follows from

that the magnetic field must also vanish there


when 0.

559

Constants for vacuum


In vacuum, and approximately in air, the electric
displacement D is related to the electric field E by the
electric constant 0
D 1 As
0 109
E 36 Vm

560

280
Lecture notes Radio Engineering by Jorma Kekalainen

Constants for vacuum


Similarly, the magnetic constant 0 is the ratio
of the magnetic induction B by the magnetic
field H in all nonmagnetic materials

561

The velocity of light


The velocity of light in vacuum c0 is related to
the product of those two constants by

562

281
Lecture notes Radio Engineering by Jorma Kekalainen

The characteristic impedance

The characteristic impedance of vacuum Z0 is


defined as:

563

Continuity for fields at boundary


At first let us consider the Faradays law (M III) in the integral
form

applied to a path formed by moving distance l along one side


of the boundary between any two materials, and returning on
the other side, an infinitesimal distance into the second
medium.

564

282
Lecture notes Radio Engineering by Jorma Kekalainen

Tangential components of electric


field
The line integral of electric field is

Since the path (infinitesimal loop) is an infinitesimal distance


on either side of the boundary, its area is zero and therefore the
contribution from changing magnetic flux is zero so long as
rate of change of magnetic flux density is finite.
When

565

Tangential components of electric


field
Consequently,

Thus tangential (parallel with the boundary)


components of electric field are continuous across the
boundary surface or

where n is unit normal (vector) to boundary surface.

566

283
Lecture notes Radio Engineering by Jorma Kekalainen

Tangential components of magnetic


field
Similarly, the generalized Amperes law in global form may be
applied to a like path with its two sides on the two sides of the
boundary as in the previous case.
The line integral of magnetic field (M IV) is

When

Since the path (infinitesimal loop) is an infinitesimal distance


on either side of the boundary, its (loop) area is zero and
therefore the contribution from changing magnetic flux is zero
so long as rate of change of magnetic flux density is finite.
567

Tangential components of magnetic


field
Thus

((Providing that J
vector is finite.))

Thus tangential components of electric and magnetic


field must be equal on the two sides of any boundary
between physically real media.
568

284
Lecture notes Radio Engineering by Jorma Kekalainen

Normal components
If two very small elements of area S are considered, one on either side of
the boundary between any two materials, with a surface charge density p
existing on the boundary, the application of Gauss law (M I)

to this elemental volume

569

Normal components (cont.)


gives

That is, for a charge-free boundary, normal components of electric flux


density are continuous; for a boundary with charges, they are discontinuous
by the amount of the surface charge density p.
With no magnetic charge term on the RHS of M II

a development corresponding to the above shows that always the magnetic


flux density is continuous:
Bn1-Bn2=0 570

285
Lecture notes Radio Engineering by Jorma Kekalainen

Boundary between different


materials

571

Isotropy
Inside the great majority of materials utilized
in practical applications, phasor vectors D and
E are parallel to one another, or colinear.
The same occurs for B and H.
Such a material is then called isotropic.
The permittivity and the permeability are
both scalar quantities.

572

286
Lecture notes Radio Engineering by Jorma Kekalainen

Anisotropy
For a family of magnetic dielectrics called ferrites,
the phasor-vectors B and H of a microwave signal are
not parallel.
These are anisotropic materials, for which the
permeability becomes a tensor, represented by a 3
3 matrix.
11 12 13 H x
B H 21 22 23 H y
31 32 33 H z

Within certain anisotropic materials, the reciprocity


is no longer applicable.
573

Note: Reciprocal networks


If a network does not contain non-reciprocal devices or materials* (i.e.
ferrites, or active devices), then the network is reciprocal.

Zij Z ji Y ij Y ji Note: The inverse


of a symmetric
matrix is symmetric.
Z and Y are symmetric

A reciprocal material is one that has reciprocal permittivity and permeability


tensors.
A reciprocal device is one that is made from reciprocal materials.
Example of a nonreciprocal material: ferrite
This material is very useful for making isolators and circulators.
574

287
Lecture notes Radio Engineering by Jorma Kekalainen

Example: Reciprocal networks


We can show that the equivalent circuits for reciprocal two-port
networks are:
Z11 Z 21 Z 22 Z 21

T-equivalent Z 21

Y21

-equivalent Y11 Y21 Y22 Y21

575

Note: Reciprocal materials


D E
B

Dx xx xy xz Ex Bx xx xy xz x
D
yy yz E y B
y yx yy yz y
y yx
Dz zx zy zz Ez Bz zx zy zz z

Reciprocal: ij ji , ij ji

j 0
0
Ferrite: 0 j is not symmetric!
0 0 1
576

288
Lecture notes Radio Engineering by Jorma Kekalainen

577

Example: Sourceless, homogeneous,


isotropic and linear dielectric region

578

289
Lecture notes Radio Engineering by Jorma Kekalainen

Example: Eliminating magnetic field


vector

579

Example: Wave equation

580

290
Lecture notes Radio Engineering by Jorma Kekalainen

Example: Cartesian form

581

Example: Cartesian component form

582

291
Lecture notes Radio Engineering by Jorma Kekalainen

Example: One-dimensional wave


equation

583

Note

584

292
Lecture notes Radio Engineering by Jorma Kekalainen

Maxwells equations

Dynamics Quasistatics

Maxwells equations in free space

586

293
Lecture notes Radio Engineering by Jorma Kekalainen

Physical quantities and Units


Quantity Unit/Value

587

Electrostatics and magnetostatics


Suppose we restrict ourselves to time-independent situations (i.e.
nothing is varying with time everything is static).
We get two sets of equations for electric and magnetic fields that are
completely independent and uncoupled:

Equations of Electrostatics Equations of Magnetostatics

Electric fields are produced by Magnetic fields are produced by


only electric charges. only electric currents.
In electrostatics problems one In magnetostatics problems one
needs to determine electric field needs to determine magnetic field
given some charge distribution. given some current distribution.

588

294
Lecture notes Radio Engineering by Jorma Kekalainen

Slow changes
The restriction to completely time-independent situations is too
limiting and often unnecessary.
What if things are changing slowly in time.
Allowing for slow time variations, one often uses the equations of
electroquasistatics and magnetoquasistatics.
Equations of Electroquasistatics Equations of Magnetoquasistatics

Electric fields are produced by Magnetic fields are produced by


only electric charges. only electric currents.
Once the electric field is Once the magnetic field is
determined, the magnetic field can determined, the electric field can be
be found by the last equation. found by the last equation.
589

How slow is slow enough ?


Electromagnetic waves and signals move at the
speed c (speed of light)
Time variations are considered slow if the time
scales over which things are changing are
much longer compared to the time taken by
light to cover distances equal to the length
scales of the problem.

590

295
Lecture notes Radio Engineering by Jorma Kekalainen

Example
Lets take an amplifier chip (max. dimension=3 cm) operating
from 100 MHz to 10 GHz. Is quasistatics analysis valid ?
100 MHz operation
Time scale of the problem = 1/(100 MHz) = 10 ns
Length scale of the problem = 3 cm
Time taken by light to travel 3 cm = 0.1 ns
Since 10 ns >> 0.1 ns, quasistatics is a valid means of analysis
at 100 MHz
10 GHz operation
Time scale of the problem = 1/(10 GHz) = 0.1 ns
Length scale of the problem = 3 cm
Time taken by light to travel 3 cm = 0.1 ns
Quasistatics is not a valid means of analysis at 10 GHz 591

Validity scale of quasistatic


analysis
Electromagnetic wave frequency f and wavelength are related to the
speed of the wave c by the relation:
f=c
Let: L = length scale of the problem
T = time scale of the problem 1/f
Condition for quasitatic analysis to be valid:
T>>L/c
cT>>L
c/f >>L
>>L
Quasistatic analysis is valid if the wavelength of electromagnetic wave at
the frequency of interest is much longer than the length scales involved in
the problem.

592

296
Lecture notes Radio Engineering by Jorma Kekalainen

Polarization states of plane


waves
Linear, circular, and
elliptical polarization

Polarization of a field
The polarization of any vector field is defined by the
geometrical locus of the tip of the vector, as it varies during
one period T.
Equation

may be written in the following form:

This is actually the equation of an ellipse having two conjugate


half-axes X(r,0) and X(r,T/4), which defines the plane of the
ellipse as well as its dimension.
594

297
Lecture notes Radio Engineering by Jorma Kekalainen

Polarization of a field

In the most general case, therefore, a field is


elliptically polarized.
There are two limiting cases:
1. when the two vectors are directed parallel to
each other, or when one of the two vanishes, the
ellipse shrinks to a straight line.
The field is then linearly polarized, and the condition

is satisfied.

595
The asterisk * denotes the complex conjugate.

Polarization of a field
2. when the two vectors have the same length and
are perpendicular to each other, the ellipse
becomes a circle, and the field is circularly
polarized.
This occurs when

and the field may be described by

where e0 and eT/4 are unit vectors along the two


directions X(r,0), and X(r,T/4), respectively.

596

298
Lecture notes Radio Engineering by Jorma Kekalainen

Polarization of a field

An elliptical A linear
polarization polarization

A circular
polarization

597

Comment
Any field may be expressed by the sum of two
linearly polarized fields: this is actually done in

Any field may also be decomposed into two


circularly polarized fields, rotating in opposite
directions.
The long half-axis of the ellipse is then the sum of the two
radii, and the short axis their difference (which vanishes for
a linear polarization).

598

299
Lecture notes Radio Engineering by Jorma Kekalainen

Example: Circularly polarized plane waves

Note: The x- and y-components of the E-field have the same amplitude but
599
are 90-degrees out of phase.

Example: Circularly polarized plane waves

Note: The E-field at z = 0 never goes to zero but keeps rotating in a


600
circular trajectory.

300
Lecture notes Radio Engineering by Jorma Kekalainen

Example: Circularly polarized plane waves

So what is happening in space when wave proceeds?


If one takes a snapshot of a circularly polarized wave at any instant, then he
will see the picture shown below

The E-field vector does not change in magnitude but its direction twists
in space.
An observer sitting in the path of the wave will see the E-field vector rotate
in a circular trajectory at his location as the wave passes by.

Note: The H-field at each point is orthogonal to the E-field. 601

Example: Linearly vs. circularly polarized


plane waves

602

301
Lecture notes Radio Engineering by Jorma Kekalainen

Handedness of circularly polarized plane


waves
Right-hand circular polarization Left-hand circular polarization
The wave is right-hand circularly The wave is left-hand circularly
polarized if at any location in space polarized if at any location in space
when the thumb of the right hand when the thumb of the left hand
points in the direction of points in the direction of
propagation, the fingers curl in the propagation, the fingers curl in the
direction of rotation of the E-field direction of rotation of the E-field
vector in a plane perpendicular to vector in a plane perpendicular to
the direction of propagation. the direction of propagation.

603

Example: Elliptically polarized plane waves

604

302
Lecture notes Radio Engineering by Jorma Kekalainen

Example: Special cases

605

Example: Special cases

606

303
Lecture notes Radio Engineering by Jorma Kekalainen

Example: Special cases

607

Example: Linear and/or circular?


A circularly polarized wave is a linear superposition of two linearly
polarized waves

Similarly, a linearly polarized wave is a linear superposition of two


circularly polarized waves

Any arbitrary polarization state can be expressed as a linear superposition


of appropriate basis states chosen from either linear or circular basis.

608

304
Lecture notes Radio Engineering by Jorma Kekalainen

Comment

609

Application of Maxwells
equations
Wave propagation

305
Lecture notes Radio Engineering by Jorma Kekalainen

Maxwells equations
Let us consider a region in space containing no sources so
there will be no charge and current terms in Maxwells
equations.
Arbitrary time variations of the fields will be permitted and the
medium is to be considered homogeneous, isotropic, linear,
and with zero conductivity.
So, we have equations in form

611

Separation of fields
To attempt a solution of a group of simultaneous equations, it is usually a
good plan to separate the various functions of position, such as D and B
here, to arrive at equations that give the distributions of each.
Thus, lets try to separate E and H to the own expressions.
First, taking the curl of M III, we see that

Time and space partial


derivatives may be taken
in any order.

612

306
Lecture notes Radio Engineering by Jorma Kekalainen

Wave equation

The left side may be expanded, using a vector identity, to give

A source-free region has been specified.


Thus

which is the general form of the wave equation for the


medium considered.

Note: This form applies as well to the magnetic field, as may seen by a similar derivation started
613
by taking the curl of M IV.

Wave equation
From this simple special case many of the characteristics of
electromagnetic waves can be found.
Recalling that in rectangular coordinates the Laplacian of a
vector can be separated into the vector sum of the Laplacian of
the components, we have in Cartesian form

614

307
Lecture notes Radio Engineering by Jorma Kekalainen

Solution of the one-dimensional


wave equation
For the case of variation in the z direction only, Laplacian
reduces to the form of the one-dimensional wave equation.

It was shown there that the equation has a general solution in


the form

Where F1 and F2 are arbitrary functions, which have


corresponding derivates (F, F) and

615

Uniform plane waves


The first term of the solution represents the
wave or function F1 traveling with velocity v
and unchanging form in the positive z
direction.
The second term represents the wave or
function F2 traveling with velocity v and
unchanging form in the negative z direction.
These functions F1 and F2 are called uniform
plane waves.
Note: Radio waves at some distance from the antenna and the ground are approximately of this
616
simple form with space variations in one direction only.

308
Lecture notes Radio Engineering by Jorma Kekalainen

Application of Maxwells
Equations
Poyntings theorem and
energy relations in an
electromagnetic field

Energy transfer through


electromagnetic fields
The simple transmission line waves can transfer energy
from one point to another.
Energy transfer may also be accomplished through more
general types of electromagnetic waves, the amount of the
energy depending on the magnitudes, distribution, and
phases of the electric and magnetic fields of the wave.
Let us take a region in which permittivity and
permeability may be functions of position but not of time.
Maxwells equations, written in terms of the total fields,
currents, and charges of a region, describe the
electromagnetic behavior of the region.

618

309
Lecture notes Radio Engineering by Jorma Kekalainen

Derivation of Poyntings theorem


Let us take a

Using a general equivalence of vector operations, we get

Next we substitute M III and M IV

to the previous equivalence

619

Derivation of Poyntings theorem

Now we integrate the previous equation over the volume V

From the Gauss theorem, the volume integral of div(E x H)


equals the surface integral of (E x H) over the boundary.

620
Note: All points of the closed surface A are on the boundary surface of the volume V.

310
Lecture notes Radio Engineering by Jorma Kekalainen

Note: Gauss theorem (Divergence


theorem)

A surface integral


A dS

S

621

Poyntings theorem

This form is valid for general media, but in the common case
of linear, isotropic, time-invariant materials, material
parameters are scalars independent of time.
Then
B H

D E

622

311
Lecture notes Radio Engineering by Jorma Kekalainen

Interpretation of Poyntings theorem

The term H2/2 represents the energy storage per unit


volume for a static magnetic field.
Energy density
of magnetic field

If this interpretation is extended by definition to any


magnetic field, the first term of volume integral
represents the time rate of increase of the stored
energy in the magnetic fields of the region.

Vs
Henry Vs
A
m m Am 623

Interpretation of Poyntings theorem

The term E2/2 represents the energy storage


per unit volume for a static electric field.
Energy density
of electric field

Similarly thinking as above, the second term of


volume integral represents the time rate of
increase of the stored energy in the electric
fields of the region.
As
Faraday V As

m m Vm 624

312
Lecture notes Radio Engineering by Jorma Kekalainen

Interpretation of Poyntings theorem

The third term is the usual ohmic term and so represents


energy dissipated in heat per unit time in medium.
All the net energy term must have been supplied externally.
Thus the term on the right (surface integral) represents the
energy flow into the volume per unit time.
Changing sign, the rate of energy flow out through the
enclosing surface is
V A W J
S E H m m m 2
sm 2
where

P S A W J
m2 W
m2 s
is called Poyntings vector.
625

Poyntings vector
It is convenient to think of the Poyntings
vector

as the vector giving direction and magnitude of


energy flow at any point in space.
V A W J
S E H m m m 2
sm 2

Note:
P S A W m 2 W J

m2 s
626

313
Lecture notes Radio Engineering by Jorma Kekalainen

Poyntings vector

If E and H
are phasors (complex vectors), then

An average power flow is expressed by

627

Example
Let it be

where A unit vector parallel with the


direction of propagation

628

314
Lecture notes Radio Engineering by Jorma Kekalainen

Application of Maxwells
equations
Plane waves

Plane wave approximation


Plane waves are good approximations to real waves in many
practical situations.
Radio waves at large distances from the transmitter, or from
diffracting objects, have negligible curvature and are well
represented by plane waves.
Much of optics utilizes the plane wave approximation.
More complicated electromagnetic wave patterns can be
considered as superposition of plane waves, so in this sense
the plane waves are basic building blocks for all wave
problems.
Even when that approach is not followed, the basic ideas of
propagation, reflection, and refraction, which are met simply
here, help the understanding of other wave problems.
630

315
Lecture notes Radio Engineering by Jorma Kekalainen

Uniform plane waves in a perfect


dielectric (free-space)
We restrict our attention to media for which and are scalar
constants.
For a uniform plane wave, variations in two directions, say x
and y, are assumed to be zero.
Maxwells curl equation in rectangular coordinates

631

Maxwells curl equation in rectangular


coordinates
Maxwells curl equation in rectangular coordinates reduce to

We see first from (3) and (6) that Ez and Hz must both be zero, except
possibly for constant (static) parts which are not of interest to us in the
wave solution.

That is, the electric and magnetic fields of this simple wave are both
transverse to the direction of propagation.

Note:
632

316
Lecture notes Radio Engineering by Jorma Kekalainen

Wave equation and its solution


Next, if (2) is differentiated with respect to z, (4) with respect
to t, and the two results combined, we obtain the one-
dimensional form of the wave equation in Ex

The solution represents wave traveling with velocity v in the


positive z direction and another traveling with the same
velocity in the negative z direction.

633

Wave traveling positive z-


direction
For free space

Lets concentrate on the wave traveling with velocity v in the


positive z direction

From (2) we may find a relation for H in the positive wave:

634

317
Lecture notes Radio Engineering by Jorma Kekalainen

Wave impedance

Integrating, and ignoring (static) constants of integration, we


have

The wave impedance of the medium is

635

Wave impedance

The quantity may also be considered a


constant of the medium, and it will be a useful
parameter in the analysis of electromagnetic
waves.
It is also known as the intrinsic impedance of
the medium.
Thus, its dimension is ohm ().
For vacuum (free space),

636

318
Lecture notes Radio Engineering by Jorma Kekalainen

Application of Maxwells
equations
A plane wave solution of
the wave equation in a
conducting medium

Conductivity
In the previous section we considered a uniform plane wave
propagating in free space.
Here we consider a uniform plane wave propagation in a
conductor.
The term conductor will be applied to those materials in which
collisions determine the movement of the charges when an
electric field is applied.
Let it be (>0) the constant conductivity of the homogenous
medium.
Thus the current density resulting from movement of the
charges in conducting medium is given by Ohms law:

638

319
Lecture notes Radio Engineering by Jorma Kekalainen

Maxwells curl equations in time-


harmonic case
Now Maxwells curl equations can expressed in the form

Applying curl operation to Faradays law (M III) we can


eliminate magnetic field

639

Note: Sinusoidal fields


A real physical field can be expressed in complex phasor
form

So

Now the field quantities


in this formula collection
are not instantaneous
values but complex
phasors.

640

320
Lecture notes Radio Engineering by Jorma Kekalainen

Wave equation in conductive


medium
Taking divergence from (M IV) and using assumed
homogeneity of the medium (, ) and generally valid vector
identities

641
Note: The divergence of the curl of any vector is zero.

Solution of the wave equation


If we mark

The wave equation reduces to the form

in the case of plane wave where an electric field has only x-


component.
A solution of this wave equation is

where is a complex propagation coefficient and C1 and C2


are some constants.
642

321
Lecture notes Radio Engineering by Jorma Kekalainen

Propagation coefficient
The complex propagation coefficient can be written

Using boundary conditions arbitrary constants can


be determined.
643

Attenuating waveform
Here we assume

And

Finally, the inverse transform gives

644

322
Lecture notes Radio Engineering by Jorma Kekalainen

Problem

Show that the expressions of


attenuation coefficient () and
phase coefficient () for the
electro-magnetic wave in
lossy medium can be
presented in the following
forms a) and b).

645

Solution

646

323
Lecture notes Radio Engineering by Jorma Kekalainen

Solution

647

Solution

648

324
Lecture notes Radio Engineering by Jorma Kekalainen

Problem
Calculate the attenuation (dB/m) in the
(saltless) lake water (r=80, =10 mS/m)
and in the sea water (r=80, =5 S/m) for an
electro-magnetic wave with frequency
a) f=10 kHz
b) f=10 MHz
c) f=10 GHz
and assuming that a field strength varies
according the distance in the form
E exp(-z), where z is a distance

649

Solution
Attenuation
expressed in dB

650

325
Lecture notes Radio Engineering by Jorma Kekalainen

Solution (r=80, =10 mS/m)

651

Solution (r=80, =10 mS/m)

652

326
Lecture notes Radio Engineering by Jorma Kekalainen

Solution (r=80, =5 S/m)

653

Solution (r=80, =5 S/m)

654

327
Lecture notes Radio Engineering by Jorma Kekalainen

Problem
Show that the propagation velocity (v) of the electromagnetic
wave in lossy medium can be presented in the following forms
a) and b).

655

Solution

656

328
Lecture notes Radio Engineering by Jorma Kekalainen

Solution

657

Problem

Calculate the propagation velocity of the


electromagnetic wave
a) in the air (r = 1, 0)
b) in the lake water (r = 80, = 10 mS/m)
c) in the sea water (r = 80, = 5 S/m)
when corresponding vacuum values are
08.85410-12 As/(Vm), c02.998 108 m/s and
f=100 MHz.

658

329
Lecture notes Radio Engineering by Jorma Kekalainen

Solution

659

Solution

660

330
Lecture notes Radio Engineering by Jorma Kekalainen

Solution

661

Application of Maxwells
equations
Penetration of
electromagnetic fields
into a good conductor

331
Lecture notes Radio Engineering by Jorma Kekalainen

Plane conductor

The simplest case to solve is a plane conductor


of infinite depth, and with no field variations
along the width or length dimension.
This case is frequently taken as that a
conductor filling the half space z > 0 in a
rectangular coordinate system with the x-y
plane coinciding with the conductor surface.

663

Good conductor
Lets look at penetration of uniform field into good
conductor filling half-space 0<z<.
If we assume that at z=0 an electric field has value
Ex(z=0)=E0 and propagation direction is z, then we
can use previously derived plane wave equation and
its solution.
So

For any good conductor the following relation is


valid

664

332
Lecture notes Radio Engineering by Jorma Kekalainen

Amperes law for good conductors

Now, we can write (M IV) in phasor notation

Note: For all but the poorest conductors (such as earth) the displacement current term is
665
completely negligible compared with conduction current term for radio frequencies at least.

Approximation of propagation
coefficient
Lets return to the previous solution of the wave equation

where

Now

Proofing as homework!

666

333
Lecture notes Radio Engineering by Jorma Kekalainen

Skin depth or depth of


penetration
Substituting the previous expression to Ex we get

where
Skin depth or depth of penetration

The magnetic field and the current density are governed by


the same form differential equations as the electric field.
Thus

667

Skin depth or depth of


penetration
It is evident from that the magnitudes of the
fields and current density decrease
exponentially with penetration into the
conductor, and z=s has the significance of the
depth at which the magnitudes have decreased
to 1/e (about 36.8 per cent) of their values at
the surface (E0, H0, J0).
The quantity s is called depth of penetration
or skin depth.

Note: symbol is also used as a symbol of skin depth or depth of penetration.668

334
Lecture notes Radio Engineering by Jorma Kekalainen

Skin depth or depth of


penetration
According to


We see that s decreases, when increase
f

E.g. for aluminium s1 mm, f=10kHz


s1 m, f=10GHz

669

Note
Depth of penetration is smaller, the higher the
conductivity, the higher the permeability, and the higher
the frequency, since it is inversely proportional to the
square root of each of these.
The fields and current penetrate deeper than the depth
s; this s is merely the point at which they have
decreased to 1/e of their values at the surface.
The concept as stated here applies strictly to plane
conductors but it may be extended to conductors of
other shapes so long as the value of s calculated is
much smaller than any curvatures of the surfaces or
curvature of surface
s
thickness of conductor 670

335
Lecture notes Radio Engineering by Jorma Kekalainen

Comment
In spite of the infinite depth requirement, the analysis
of this case is practically important to many conductors
of finite extent, and with curved surfaces, because at
high frequencies significant fields are concentrated very
near the surface, so that radius of curvature and
conductor depth may be taken as infinite in this
comparison.
Moreover, any field variations along the length or width
dimension due to curvature, edge effects, or variations
along a wavelength are ordinarily so small compared
with the variations into the conductor that they may be
neglected.

671

Problem
The surface value of the field is E0. Determine
the magnitude of field penetrated s meters
deep into the good conductor.

672

336
Lecture notes Radio Engineering by Jorma Kekalainen

Problem

The surface value of the field is E0. Determine


the magnitude of field penetrated s meters
deep into the good conductor.

Let
The magnitude of the field is
0.368 E0

673

Problem
Show that the unit of s defined previously is
meter.
1
s
f

674

337
Lecture notes Radio Engineering by Jorma Kekalainen

Problem

Show that the unit of s defined previously is


meter. 1
s
f

Vs A
1 B m 2 Vs J m 2 A
s , ,
f H A Am E V Vm
m m
1
s m
1 Vs A
s Am Vm
675

Problem

Defining wavelength as = v/f =1/(f) show


that, for a good conductor, depth of penetration
is always a very small quantity compared with
waveIength or s<<.

676

338
Lecture notes Radio Engineering by Jorma Kekalainen

Solution

1
s , for a good conductor
f
1 1 1
s2
f v 1

1 v 2
v v
2f 2 f 2 2
s 1 1
0.23 s
2 4.4

677

Problem
Show that

when

678

339
Lecture notes Radio Engineering by Jorma Kekalainen

Solution

679

Application of Maxwells
equations
Field theoretic problems

340
Lecture notes Radio Engineering by Jorma Kekalainen

Problem 1
Write Maxwells equations

in Cartesian (xyz) coordinates, when

681

Solution
Write Maxwells equations in Cartesian (xyz)
coordinates.

682

341
Lecture notes Radio Engineering by Jorma Kekalainen

Solution

683

Solution

684

342
Lecture notes Radio Engineering by Jorma Kekalainen

Solution

685

Solution

686

343
Lecture notes Radio Engineering by Jorma Kekalainen

Problem 2

Derive the global form version from the local


form Gauss laws (M I and M II) using the
Divergence theorem:
A surface integral


A dS

S

687

Solution
Derive the global form version (surface integrals) from the
local form Gauss laws (M I and M II) using the Divergence
theorem (Gauss theorem):

Thus

688

344
Lecture notes Radio Engineering by Jorma Kekalainen

Solution
Divergence theorem (Gauss theorem):

Thus

689

Problem 3
Derive the global form version (line integrals) from the local
form Faradays law and Amperes law (M III and M IV) using
the Stokes theorem:

690

345
Lecture notes Radio Engineering by Jorma Kekalainen

Solution
Stokes theorem:

691

Solution
Stokes theorem:

692

346
Lecture notes Radio Engineering by Jorma Kekalainen

Problem 4

693

Solution
Now Maxwells equations are

Operations:
M IV and D E

M III E B H
B H

t t

694

347
Lecture notes Radio Engineering by Jorma Kekalainen

Solution

On the other side we have well-known vector


identity

695

Problem 5

696

348
Lecture notes Radio Engineering by Jorma Kekalainen

Solution 5

697

Solution 5

698

349
Lecture notes Radio Engineering by Jorma Kekalainen

Problem 6

699

Solution 6
Let us derivate the solution candidate Ex-

Let us substitute the derivates to the


wave equation Ex- is a solution of
the wave equation

700

350
Lecture notes Radio Engineering by Jorma Kekalainen

Problem 7

701

Solution 7

702

351
Lecture notes Radio Engineering by Jorma Kekalainen

Solution 7

703

Solution 7

704

352
Lecture notes Radio Engineering by Jorma Kekalainen

Solution 7

705

Solution 7

706

353
Lecture notes Radio Engineering by Jorma Kekalainen

Problem 8

707

Solution 8

708

354
Lecture notes Radio Engineering by Jorma Kekalainen

Solution 8

709

Solution 8

710

355
Lecture notes Radio Engineering by Jorma Kekalainen

Problem 9

711

Solution 9
Now

712

356
Lecture notes Radio Engineering by Jorma Kekalainen

Solution 9

713

Solution 9

714

357
Lecture notes Radio Engineering by Jorma Kekalainen

Solution 9

715

Problem 10
Write the local form Maxwells equation in the
following special cases:
1 for static fields
2 for sourceless fields
3 for sinusoidal fields in phasor form

716

358
Lecture notes Radio Engineering by Jorma Kekalainen

Solution 10

717

Solution 10

718

359
Lecture notes Radio Engineering by Jorma Kekalainen

Solution 10

719

Solution 10

720

360
Lecture notes Radio Engineering by Jorma Kekalainen

Solution 10

721

Note

722

361
Lecture notes Radio Engineering by Jorma Kekalainen

Problem 11

723

Solution 11

724

362
Lecture notes Radio Engineering by Jorma Kekalainen

Solution 11

725

Note

726

363
Lecture notes Radio Engineering by Jorma Kekalainen

Problem 12

727

Solution 12

728

364
Lecture notes Radio Engineering by Jorma Kekalainen

Problem 13

729

Solution 13

730

365
Lecture notes Radio Engineering by Jorma Kekalainen

Solution 13

731

Problem 14

732

366
Lecture notes Radio Engineering by Jorma Kekalainen

Solution 14

733

Solution 14

734

367
Lecture notes Radio Engineering by Jorma Kekalainen

Solution 14

735

Solution 14

736

368
Lecture notes Radio Engineering by Jorma Kekalainen

Solution 14

737

Notes

738

369
Lecture notes Radio Engineering by Jorma Kekalainen

Problem 15

739

Solution 15

740

370
Lecture notes Radio Engineering by Jorma Kekalainen

Solution 15

741

Problem 16

742

371
Lecture notes Radio Engineering by Jorma Kekalainen

Solution 16

743

Problem 17

744

372
Lecture notes Radio Engineering by Jorma Kekalainen

Solution 17

745

Problem 18

746

373
Lecture notes Radio Engineering by Jorma Kekalainen

Solution 18

747

Note

748

374
Lecture notes Radio Engineering by Jorma Kekalainen

Problem 19

749

Solution 19

750

375
Lecture notes Radio Engineering by Jorma Kekalainen

Problem 20

751

Solution 20

752

376
Lecture notes Radio Engineering by Jorma Kekalainen

Passive microwave devices

Definitions and classification


of transmission lines and
waveguides

Uniformity along the direction of


propagation
The propagation characteristics of a
transmission line or a waveguide can be
determined from the study of the
electromagnetic fields within the general
structure depicted in the figure below.

Waveguide of arbitrary
cross section, uniform
along the direction of
propagation.

754

377
Lecture notes Radio Engineering by Jorma Kekalainen

Uniform transmission line or


waveguide
The cross section of this structure is arbitrary.
It can contain several different propagation media (air,
dielectrics, ferrites), in which case it is called inhomogeneous.
It can also contain, or be surrounded by, metallic conductors.
The direction of propagation is straightforward: the
longitudinal coordinate axis z is located along this direction.
A displacement along this axis does not modify either the
geometry, nor the material properties: these quantities are thus
independent of the longitudinal coordinate z (translation-
invariance).
The transmission line or waveguide is then called uniform.

755

Selection of a coordinate system


The system of coordinates used to study the propagation must
be well adapted to the transverse geometry of the line.
Boundaries and interfaces should be described by simple
geometrical expressions, so that boundary conditions for the
fields can be satisfied.
In all uniform structures, the coordinate system contains the
coordinate z, defined in the previous section, and two
coordinates located in the transverse plane.
For problems exhibiting a symmetry of revolution, the
(circular) cylindrical coordinate system is the obvious choice.
A rectangular waveguide is best studied in a rectangular
reference base.
756

378
Lecture notes Radio Engineering by Jorma Kekalainen

Example: Cartesian coordinate


system

All mean exactly the same


thing but just a different
notation for the unit vectors.

757

Example: Cylindrical coordinate


system

Both mean exactly the same thing


but just a different notation for the
unit vectors.

758

379
Lecture notes Radio Engineering by Jorma Kekalainen

Example: Spherical coordinate


system

Both mean exactly the same thing


but just a different notation for the
unit vectors.

759

Example: Vector field

In engineering terms, a vector field implies a


vector associated with every point is space.

Electric field: E(x,y,z,t ) or E(r ,t )

Magnetic field: H(x,y,z,t ) or H(r ,t )

760

380
Lecture notes Radio Engineering by Jorma Kekalainen

Separation

Transverse planes are always perpendicular to


the direction of propagation.
Transverse and longitudinal dependencies of
the fields are thus independent of each other,
and the differential operator del or nabla can
be expressed by the sum of a transverse part
(index t) and a longitudinal one:

761

Rectangular coordinates
In the rectangular coordinate system, the transverse part of the
operator can be further separated, in terms of the coordinates x
and y:

In all other coordinate systems, the transverse operator cannot


be separated in this manner!
Similarly, the phasor-vectors have transverse and longitudinal
components:

The same is true for the vector r, which indicates position:

762

381
Lecture notes Radio Engineering by Jorma Kekalainen

Consequence of separation
Within every system of cylindrical coordinates
(which correspond to uniform lines), Maxwells
equations are partially separable.
The general solution for the electromagnetic fields is
then the sum of several terms, each term being the
product of a function of the transverse coordinate rt
by a function of the longitudinal variable z.
The transverse and longitudinal dependencies of the
fields can then be studied independently of each
other.
763

Longitudinal dependence of the


fields: Wave equation
Applying the method of separation of variables, and
then grouping the resulting terms, a wave equation
involving the direction of propagation is obtained for
every field component in an isotropic structure

The complex constant 2 results from the separation


process.
Its value depends on frequency, material properties
and the particular solution considered for the fields.
764

382
Lecture notes Radio Engineering by Jorma Kekalainen

General solution

Wave equation is satisfied by the general form


of solution:

The two terms correspond, respectively, to the


forward and to the backward waves, functions
Xi+ (rt) and Xi- (rt) defining their transverse
dependencies in the plane z= 0.

765

Propagation factor and related


quantities
The meaning of the propagation factor
= +j :
The real part , expressed in Neper per meter
(Np/m), is called attenuation per unit length.
It represents the damping of the wave traveling along
the transmission line.
Its inverse = 1/ is the skin depth, the distance over
which the signal amplitude decreases to 1/e of its initial
value.

766

383
Lecture notes Radio Engineering by Jorma Kekalainen

Propagation factor and related


quantities
The meaning of the propagation factor
= +j :
The imaginary part , measured in radians per
meter (rad/m), is the phaseshift per unit length.
It indicates the phase variation of the wave along the
direction of propagation.
The wavelength along the line or waveguide, denoted
by g, is inversely proportional to :

767

Propagation velocities
The two propagation velocities, the phase and group
velocities, are respectively defined by the relations

An observer following a zero of the field moves at the


phase velocity.
An observer following the envelope of a modulated
signal at the group velocity.

768

384
Lecture notes Radio Engineering by Jorma Kekalainen

Guided waves
Theoretically the simplest case is to deal with waves
that propagated in infinite size media.
For many applications it is desirable to have
electromagnetic energy be guided in much the same
way as water flow is guided by having it flow in
pipes.
Transmission lines are the simplest structures that
guide electromagnetic waves.
Transmission line would be any two arbitrary shaped
metal conductors that are long and uniform in at least
one dimension.
769

Definition: Transmission line


Transmission line is the preferred term to
describe transmission systems with two or
more metallic conductors electrically insulated
from one another
for instance the two-wire line and the coaxial line.

770

385
Lecture notes Radio Engineering by Jorma Kekalainen

Some transmission line structures


Capacitance and inductance per unit length for each structure

771

E-fields and H-fields for the


previous transmission lines

772

386
Lecture notes Radio Engineering by Jorma Kekalainen

Some transmission line structures


Capacitance and inductance per unit length for each structure

773

E-fields and H-fields for the


previous transmission lines

774

387
Lecture notes Radio Engineering by Jorma Kekalainen

Definition: Waveguide
The propagation in a waveguide is generally
ensured by successive reflections on the guide
boundaries.
These are conducting walls in the case of metallic
waveguides.
Dielectric waveguides and optical fibers
utilize the total reflection on the interface
between two dielectric materials.

775

Definition: Modes of propagation


The solution of Maxwells equations, subject
to the transverse boundary conditions in the
structure of transmission line or waveguide is
an eigenvalue problem.
A set of solutions is obtained, which are called
the modes of propagation.
They are the eigenvectors of the problem, each
one associated with an eigenvalue (propagation
factor, transverse wavenumber).

776

388
Lecture notes Radio Engineering by Jorma Kekalainen

Modes of propagation
The modes of propagation form an infinite set,
which is discrete for a closed structure
(surrounded by a metal boundary),
[to which must be added a continuous spectrum of
radiating modes when the structure is open (fields
extending to infinity)].
Each mode possesses specific propagation
properties:
attenuation and phase-shift per unit length,
propagation velocities,
cut-off frequency. 777

Multimode propagation
When propagation of a signal takes place at the same time over several
modes, the difference in the propagation velocities may produce
distortions of the signal.
To avoid this unwanted effect in radar and in communications, the shape
and the dimensions of the line are adjusted so that only one mode can
propagate at the signal frequency.
The possible existence of several propagating modes limits the available
frequency band in all transmission lines.
The length of the line or the bit rate of the signal are reduced to maintain
distortion within acceptable limits.
The presence (or absence) of a particular mode of propagation on a
transmission line depends on the excitation, i.e., on the boundary conditions
at both ends of the line (generator and load).

Note: In fiber optics, also multimode operation (thick fiber) is used in short
distance communication. 778

389
Lecture notes Radio Engineering by Jorma Kekalainen

Classification of the modes of


propagation
The presence or the absence of longitudinal
field components affects the propagation
behavior of the modes.
Four mode categories can exist, as shown
below.

779

Classification of lines and


waveguides
A large number of different structures can be used to
transmit electromagnetic signals:
1. either open (radiation can take place) or closed (fields
entirely enclosed within a conducting envelope);
2. either homogeneous (one single propagation medium
without transverse dependence of the material properties),
or inhomogeneous (several different propagating media, or
a single medium having a continuous variation of the
material properties in the transverse plane);
3. either conductorless, or possessing one or more
conductors.

780

390
Lecture notes Radio Engineering by Jorma Kekalainen

Classification of lines and


waveguides

Note: When a waveguide or a transmission line (for instance a coaxial line) is partially
781
filled with dielectric, they are respectively called loaded waveguide or loaded line.

Cross sections of the transmission


lines

782

391
Lecture notes Radio Engineering by Jorma Kekalainen

Cross sections of the waveguides

783

Comment
Early microwave systems relied on waveguide and coaxial lines for
transmission line media.
Waveguide has the advantage of high power-handling capability and low
loss but is bulky and expensive.
Coaxial line has very high bandwidth and is convenient for test
applications, but is a difficult medium to fabricate complex microwave
components.
Planar transmission lines provide an alternative, in the form of stripline,
microstrip, slotline, coplanar waveguide, and many other types of related
geometries.
Such transmission lines are compact, low in cost, and are capable of being
easily integrated with active devices such as diodes and transistors to form
microwave integrated circuits.

784

392
Lecture notes Radio Engineering by Jorma Kekalainen

Characteristics of some
transmission media

785

Notes
Wire pair is distance limited. The shorter the pair length, the higher the
bit rate.
LOS microwave is limited by statute (legal limitations), meaning by the
ITU Radio Regulations and the national regulatory authority.
Satellite communication faces the same legal limitations. Geostationary
orbit (GEO) satellites have long delays, which could affect interactive data
systems. Only one GEO satellite relay allowed for a voice connectivity.
The limits of fiber optics are still being explored. All terrestrial buried and
aerial cable systems are vulnerable to severing by natural disaster or by
man.
Coaxial cable is limited by amplitude-frequency response characteristics.
In nearly every instance, fiber optic cable connectivity is preferred.

786

393
Lecture notes Radio Engineering by Jorma Kekalainen

Twisted pair

(a) Category 3 UTP (Unshielded Twisted Pair), 16 Mbps.


(b) Category 5 UTP, 100 Mbps.
Category 6 and 7, 250 600 Mbps.

787

Coaxial cable

50 ohm and 75 ohm 1 GHz

788

394
Lecture notes Radio Engineering by Jorma Kekalainen

Optical fiber

(a) Three examples of a light ray from inside a silica fiber


impinging on the air/silica boundary at different angles.
(b) Light trapped by total internal reflection.

Multimode uses total reflection.


Single mode 50 Gbps over 100 km.

789

Transmission of light through fiber

Attenuation = 10*log10(transmitted power/received power)


f = c/ f = 300,000,000/1.3*10^-6 = 2.31*10^14 = 231 THz.
Bandwidth = 300,000,000*(1/1.22 1/1.37)*10^6 = 0.27*10^14 = 27000 GHz 790

395
Lecture notes Radio Engineering by Jorma Kekalainen

Fiber cables

(a) Side view of a single fiber.


(b) End view of a sheath with three
fibers.

791

Fiber vs. copper

Repeaters: copper <5 km vs. fiber >50 km.


1000 twisted pairs 1 km long weights 8 tons
vs. 2 fibers with more capacity weight 100
kg.
Fiber doesnt leak the light excellent
security.

792

396
Lecture notes Radio Engineering by Jorma Kekalainen

Example: Bit rate


Coax 750 MHz bandwidth
If we assume that it is possible to code 8 bps/Hz of
bandwidth.
6 Gbps.

Fiber:
f = c/ f = |c

For = 1.3 m band = 0.17 m that gives the


bandwidth f = 40 THz
To get 40 Tbps coding 1 bps/Hz of bandwidth is
needed. 793

Transmission lines and


waveguides
Recap

397
Lecture notes Radio Engineering by Jorma Kekalainen

Guiding structures

A guiding structure is one that carries a signal


(or power) from one point to another.

There are three common types:


Transmission lines
Fiber-optic guides
Waveguides

Note: An alternative to guiding structures is wireless transmission in


more or less free-space using antennas. 795
795

Transmission line properties

Has two conductors running parallel.


Can propagate a signal at any frequency (in theory).
Becomes lossy at high frequency.
Can handle low or moderate amounts of power.
Does not have signal distortion, unless there is loss.
May or may not be immune to interference.
Does not have Ez or Hz components of the fields (TEMz).

796

398
Lecture notes Radio Engineering by Jorma Kekalainen

Example: Transmission line


CAT 5 cable (twisted pair)

Coaxial cable

The two wires of the transmission line are twisted to reduce


interference and radiation from discontinuities. 797

Symmetric transmission line

The twin lead is an open type of transmission


line the fields extend out to infinity.

+ -
The extended fields may cause
interference with nearby
objects. (This may be improved
by using twisted pair.)

Note: Having fields that extend to infinity is not the same thing as
having radiation, however. 798

399
Lecture notes Radio Engineering by Jorma Kekalainen

Example: Transmission lines


Transmission lines on printed-circuit boards

w
r
h
r h w

Microstrip Stripline

w w w

r h r h

Coplanar strips Coplanar waveguide (CPW)


799

RF waveguide properties
Is a single hollow metal pipe.
Can propagate a signal only at high frequency: > c
The width must be at least one-half of a wavelength.
Has signal distortion, even in the lossless case.
Immune to interference.
Can handle large amounts of power.
Has low loss (compared with a transmission line).
Has either Ez or Hz component of the fields (TMz or
TEz).

800

400
Lecture notes Radio Engineering by Jorma Kekalainen

Fiber-optic guide

Properties
Uses a dielectric rod
Can be made very low loss
Has minimal signal distortion
Very immune to interference
Not suitable for high power
Has both Ez and Hz
components of the fields

801

Two types of fiber-optic guides

1) Single-mode fiber
Carries a single mode, as with the mode on a
transmission line or waveguide.
Requires the fiber diameter to be small relative to a
wavelength.

2) Multi-mode fiber
Has a fiber diameter that is large relative to a
wavelength.
Its operation can be described using simple ray
theory and the principle of total internal
reflection (critical angle effect).

802

401
Lecture notes Radio Engineering by Jorma Kekalainen

Optical fiber

Higher index core region

803

Transmission line theory


Lumped circuit elements: resistors, capacitors,
inductors
neglect time delays (phase)

Distributed circuit elements: transmission lines

account for propagation and


time delays (phase change)

We need transmission line theory whenever the


length of a line is significant compared with a
wavelength.
804

402
Lecture notes Radio Engineering by Jorma Kekalainen

Passive microwave
components
Directional couplers

Components
Each individual component has a specific
purpose.
Its function may be to couple, mix, filter, isolate,
divide, add, act on a systems phase, convert RF
to DC, attenuate, or any combination of the
above.
Basic questions for component:
How does it work?
What does a data sheet or catalog say about that
component?

806

403
Lecture notes Radio Engineering by Jorma Kekalainen

Directional coupler
One of the most widely used components in microwaves is the
directional coupler.
It may be employed by itself to monitor, sample, or attenuate; or it
may be used as part of another component, such as a mixer.
The basic definition of couple, from which the word coupler is
derived, is to bring two electrical circuits into such close proximity
as to permit mutual influence.
In other words, two circuits are side-by-side so that they have an
effect on one another.
The circuits could be two pieces of waveguide, two coaxial lines,
or two copper lines in a stripline coupler.
One of the circuits is the main line of the coupler; the other circuit is
the coupled line.

807

Basic directional coupler

808

404
Lecture notes Radio Engineering by Jorma Kekalainen

Directional couplers

a coaxial coupler

a stripline coupler

a waveguide coupler

809

Directional coupler
Figure helps in our explanation of the term directional.
If energy is applied to port 1 in (a), most of the energy appears
at port 2, while a fraction appears at port 3.
The actual amount at port 3 depends on the coupling value;
standard couplers are of 6, 10, 20, and 30 dB.

Note: A 3 dB coupler, called a quadrature hybrid, is also available. 810

405
Lecture notes Radio Engineering by Jorma Kekalainen

Dual-directional coupler
One method to expand the basic directional coupler is
through the construction of a dual-directional coupler.
Just as the name implies, this procedure involves two
couplers put back-to-back.
The one great advantage of this type of coupler is that it
allows you to monitor both forward and reflected power
at the same time.
It aids in the obtaining of data from a device which
indicates the input match and, thus, its VSWR.

811

Specifications
Below are sample specifications for a directional coupler as
they could appear on a data sheet or in a catalog.

812

406
Lecture notes Radio Engineering by Jorma Kekalainen

Specifications

Connectors This portion of coupler


specifications is very important since it tells
you if you either need adapters to fit the
coupler into your system.
Note that additional adapters mean higher
overall losses in a system.
Connectors may be type N, BNC, SMA, APC-
7, or any special variety required.

813

3- and 4 port coupler

814

407
Lecture notes Radio Engineering by Jorma Kekalainen

Specifications
Coupled Line VSWR This specification
indicates what sort of match the coupled port
(port 3) offers to any external load.
A low VSWR would mean a good match and
would cause little or no problems to any
external circuit.

815

Specifications
Coupling The ratio of the power available
at the coupled port to the power at the input
port. (Port 3 to port 1).
It is the amount of attenuation of the input
power as a result of the coupling structure.
This term is expressed in dB and has standard
values of 6, 10, 20, and 30.

816

408
Lecture notes Radio Engineering by Jorma Kekalainen

Specifications

Coupling Deviation This measurement


indicates how flat the coupling is over a
specified band of operation; it is expressed as a
(+) or () dB value.
If we rely heavily on a specified value of
coupling over a wide range of frequencies,
coupling deviation should be low (e.g. 1 dB).

817

Specifications
Directivity This figure indicates how
accurately you are able to measure parameters in
your system.
It is the difference between the desired and
undesired couplings.
If a coupler has low directivity, the forward and
reverse powers in the coupler interfere with one
another and cause great inaccuracies to occur.
Ports 1 and 4 in are the undesired coupling; 1 and 3 are
the desired combination.

818

409
Lecture notes Radio Engineering by Jorma Kekalainen

Specifications

Frequency This term indicates the range of


frequencies for which all the other
specifications are true.
Above and below this range, none of the
specifications apply.
Most couplers are designed to operate over an
octave band (1-2 GHz, 2-4 GHz, etc.).

819

Specifications
Impedance Impedance is an RF
resistance reading which ensures that any
component with a similar impedance value
connected to the coupler will operate properly.
The most common impedance is 50 ohms,
which is considered to be a standard in most
systems.

820

410
Lecture notes Radio Engineering by Jorma Kekalainen

Specifications
Input VSWR This
specification is identical to
the coupled line VSWR in
that it is an indication of
what sort of match is offered
by the input port of the
coupler;
it is usually of low-value,
1.15:1 for example.
In Figure (A) it is the match
at port 1; in (B) it is either
port 1 or 2. 821

Specifications
Insertion Loss The insertion loss indicates the power lost
in the main line of the coupler primarily through dissipation.
In coaxial coupler it is simply a loss which occurs in a conductor as
energy is passed through it;
in a stripline coupler, the copper losses of the conductors; and
in a waveguide coupler, a loss within the guide with the only
contributing factor usually the length of the guide.
In Figures (A) and (B), the insertion loss is that from either
ports 1 and 2 or 2 and 1.
In lower value couplers (3 dB or 6 dB, for example) the
coupled energy due to the coupled port also is included in the
total through-line insertion loss, thereby increasing the overall
value.
822

411
Lecture notes Radio Engineering by Jorma Kekalainen

Specifications

Model Number The model number is


simply a way of referring to and identifying
the coupler.
In most cases the last two numerals indicate
the coupling value.

823

Specifications
Power Rating This specification refers to
the amount of both CW and peak power that
the coupler is capable of handling.
Any power level greater than this value may
cause arcing or a deterioration of performance.
This term is one of the most important of
which to take care if you will be operating at a
high power level.

824

412
Lecture notes Radio Engineering by Jorma Kekalainen

Specifications
Reverse Power Rating This term applies only to a
3-port coupler; it is usually a lower value than that of
the forward power rating because the fourth port is
internally terminated and so, generally, can not handle
any amount of power.
This specification is also quoted as a CW and peak
power number; it is the power handling capability of
the coupler if power is applied in a reverse direction.
At port 2 instead of port 1 in Figure (A).

825

Specifications
Size the actual physical size of the coupler.
This term is of a secondary nature except when
you are required to put the coupler in a restricted
space.
Weight Weight is self-explanatory term
which only comes into play when an overall
weight restriction is placed on a system, such
as with airborne equipment.

826

413
Lecture notes Radio Engineering by Jorma Kekalainen

Directional coupler symbols

827

Applications of directional couplers

The most obvious use of a directional coupler is to


monitor an output power.
Its directional properties ensure that only the forward
power is coupled; its property of circuit isolation
leaves the main line power undisturbed.
Figure shows the orientation of a directional coupler
when used as a power monitor.

828

414
Lecture notes Radio Engineering by Jorma Kekalainen

Applications of directional couplers

A scheme similar to the monitoring technique described earlier


is shown in figure below.
The big difference in such a setup is that a diode detector is placed at
the coupled port to provide a dc voltage proportional to any main line
power variations.
That voltage operates leveling circuitry which corrects the RF level and
eliminates differences in the output power over a certain specified
frequency band.
This technique is used widely where swept frequency measurements
are taken.

829

Applications of directional couplers

Figure below shows an application which takes advantage of


the directional couplers ability to handle higher CW power.
It is an actual example of a system used to check a 100 W power
amplifier.
The only components available were two-watt attenuators, five-watt
terminations, and the directional coupler.
It can be seen that a more adequate setup was created by utilizing the
couplers 200 watt capability and directivity of 25-30 dB.
None of the components used had its power rating exceeded; the
directional coupler was doing the majority of the work.

830

415
Lecture notes Radio Engineering by Jorma Kekalainen

Applications of directional couplers

Figure shows an example where the coupler is


an essential part of another component.
In this case it is a very important part of a single
diode mixer.
The circuit is of a very basic mixer, but it shows
that by using a coupler the RF and LO signals have
a high degree of mutual isolation.

831

Applications of directional couplers

The dual-directional coupler is very useful in VSWR


measurements.
In Figure it is employed to measure the return loss which then
is converted to VSWR.
Power meter # 1 reads the forward power to the device under
test; meter # 2 measures any reflected power.
By comparing these two figures a return loss is determined
and converted to the input VSWR of the particular device
under test.

832

416
Lecture notes Radio Engineering by Jorma Kekalainen

Applications of directional couplers

833

Summary
Directional coupler is a component which allows
two microwave circuits to be combined into one
integrated system in one direction, while being
completely isolated from each other in the
opposite direction.
Its property of being directional, its isolation, its
low input and coupled line VSWR, and its
adaptability to being 3- or 4-port or dual-
directional make the directional coupler one of
the most widely used components in microwaves.

834

417
Lecture notes Radio Engineering by Jorma Kekalainen

Passive microwave
components
Detectors

Detector
The key to a reliable, trouble-free system is
simplicity.
One component whose operation is based on
simplicity is the microwave detector.
This concept of a simple and reliable
component is the very reason why the detector
is used so extensively in microwave systems.

836

418
Lecture notes Radio Engineering by Jorma Kekalainen

Basic microwave detector

The basic construction of the microwave


detector is illustrated in figure below.
The heart of the detector is, of course, the
diode itself.
e.g. Schottky diodes are often used because of their
good RF characteristics, which include high
sensitivity.

837

Basic microwave detector


The diode is matched to the driving circuit (usually 50
ohms) so that maximum power transfer is obtained.
In this manner maximum efficiency of the diode is
ensured since any reflections due to mismatches are
eliminated and, thus, all of the input power reaches the
diode.
The dc return, besides acting as a ground for the diode,
also has the second function of acting as an RF choke
so that no RF is shunted to ground.
A low-pass filter is placed after the diode to eliminate
all of the high frequency ripple caused by the detection
process and allow only dc to be present at the output.

838

419
Lecture notes Radio Engineering by Jorma Kekalainen

Specifications
A typical set of specifications for a microwave detector

839

Specifications
Frequency Response Frequency response
is a term used to indicate the RF performance
of the detector.
It is sometimes given in decibels per octave or
in one figure for overall performance.
Basically it is a measure of the variation in
sensitivity of the detector, expressed in dB.

840

420
Lecture notes Radio Engineering by Jorma Kekalainen

Specifications
Maximum Input Power This term refers to
one of the most important considerations when
using a detector the highest level of RF power
that can be applied to its input.
It is always good practice to stay below this figure
by a reasonable margin.
The reason that there is a maximum level is that
there is a diode directly in line in the circuit;
wherever there is a semiconductor there is a
certain power level that could destroy the
detector.

841

Specifications
Polarity Polarity is either positive or
negative.
It is important to know the polarity of the
output of your detector since it will have to be
joined with some external circuit that requires
a certain polarity to operate properly.

842

421
Lecture notes Radio Engineering by Jorma Kekalainen

Specifications
A typical sensitivity curve for a microwave detector

843

Specifications
Sensitivity From the previous sensitivity curve
for a microwave detector it can be seen that one
factor affecting sensitivity is the RF power
applied to the input of the detector.
Therefore, the detector sensitivity is the dc volt-
age produced at the output for a specific power
input, usually expressed in millivolts per
microwatt CW.
Sensitivity is simply how much power you need
to produce a certain voltage.

844

422
Lecture notes Radio Engineering by Jorma Kekalainen

Detector applications

The microwave detector finds appications


wherever there is a need to control or convert
an RF signal to a DC voltage.
E.g. for display purposes.

845

Monitoring
One of the most common applications is
monitoring an RF line.
Power is coupled off the main line;
the detector provides a DC voltage proportional
to this power.
The monitoring device gives an instant reading of
the main line power.

846

423
Lecture notes Radio Engineering by Jorma Kekalainen

Automatic Leveling Control (ALC)


A leveled power is needed, e.g., in swept measurements.
A coupler and detector work together to sample the main
line RF;
the detector converts the RF to DC to activate the leveling
circuits.
This DC voltage causes either the RF power out of the source to
be attenuated or the attenuation to be removed in order to raise
the power level.
The result of this instantaneous pattern of action and reaction of
coupler and detector is that the RF power remains level at the
system output.
Automatic Leveling Control (ALC) in microwave systems often
employs just such a process.

847

Detecting for display on a scope


It is impractical to attempt to view signals at
microwave frequencies directly on a conventional
oscilloscope.
Without a Network Analyzer figure below is the
answer.
A simple combination of an attenuator and a detector do the
job.
The attenuator is of great importance in the setup, it keeps
power both below a maximum power level and in the
linear range of the sensitivity curve of the detector, giving
much better results.

848

424
Lecture notes Radio Engineering by Jorma Kekalainen

Passive microwave
components
Mixers

Mixer

A schematic representation of a mixer

850

425
Lecture notes Radio Engineering by Jorma Kekalainen

Microwave mixer circuits

A basic block diagram of the three


fundamental parts of a mixer an input
coupling network, the diode circuitry, and an
output filtering network.

851

Input coupling network


The input coupling network is the circuitry
which allows the RF and LO signals to be
combined properly before being applied to the
diode portion.
This network could be a directional coupler, a
transformer, or a quadrature hybrid;
its prime feature should be that, under operating
conditions, it combines the two signals while
providing isolation between them.

852

426
Lecture notes Radio Engineering by Jorma Kekalainen

Mixing process
The diode circuitry consists of the diode(s).
This part is where the actual mixing process occurs.
When the RF and LO signals are applied to the diode
circuitry, many combinations of the two are generated.
The signals which are most prominent are each of the
originals and their sum and difference.
In other words, if the RF signal is called f1, and the LO
signal called f2, the output would have f1, f2, f1 + f2, f1 -
f2, and every other combination of fundamentals and
harmonics, (2f1, 2f2, 2(f1 + f2), etc.)

853

Output filtering
These combinations of frequencies are the reason why the final
block the output filtering network is needed in a mixer.
Usually, the upper or lower sideband (that is, f1 + f2 or f1-f2 ) is the
desired output;
everything else must be eliminated to produce the desired and
necessary clean signal.
The filtering usually is accomplished by a low-pass network which
might be as simple as a shunt capacitor or as complex as a cascade
of low-pass filters; it could be a series of trap circuits which would
eliminate only specific unwanted signals.
For certain applications, the filtering could be performed by a band-
pass or high-pass filter.
The type of filtering and device used depends, of course, upon the
individual situation.

854

427
Lecture notes Radio Engineering by Jorma Kekalainen

Basic types of mixers

Three basic configurations


single diode,
balanced, and
image rejection.

855

Single diode mixer


The advantage of the single diode mixer is its
simplicity of construction.
The disadvantages,
very narrow band device
No noise eliminating ability, which is a primary
concern of the sophisticated low-noise systems
however, far outweigh the advantages.

856

428
Lecture notes Radio Engineering by Jorma Kekalainen

Balanced mixer
Here the input network is termed a balanced coupling
network, either a 90 or 180 hybrid.
Whichever type is used, the efficiency of the coupling
network is vastly greater than that of any single diode
mixer.
The increased efficiency of the hybrids means that virtually
all of the power reaches the diodes, and there is very little
loss due to mismatches.
So a conversion loss through the mixer is reasonably low.
Also, this type of coupling network usually operates over an
octave band, another great improvement.

857

Balanced mixer
A mixing apparatus or configuration which
contains four matched diodes called a quad.
A quad makes up a balanced configuration
with two matched sections in each side, having
the effect of lowering conversion loss and
noise.

858

429
Lecture notes Radio Engineering by Jorma Kekalainen

The advantages of balanced


mixers
Balanced mixers have very good noise
characteristics and low conversion losses;
They operate over a wide band of frequencies
(both LO and RF);
They have an acceptable input VSWR on the
RF and LO ports.

859

Specifications
A typical data sheet for a balanced mixer

860

430
Lecture notes Radio Engineering by Jorma Kekalainen

Specifications

Conversion Loss Conversion loss is the


level of the IF output as compared to the level
of the RF input.
Usually expressed in dB, it can be considered
as a measure of the efficiency of the mixer;
it indicates how efficiently the two input
frequencies are combined to form the IF output.

861

Specifications
Isolation It may be any one of three
combinations of measurements referred to by this
term RF/LO, RF/IF, or LO/IF.
In each case, isolation is a measure of the
undesired frequency at an auxiliary port.
An example would be as follows the RF/LO
isolation figure, in dB, is the amount of RF power
measured at the LO port compared to that at the
RF port.
The other two isolation figures are found the same
way.

862

431
Lecture notes Radio Engineering by Jorma Kekalainen

Specifications

LO Power LO power is the amount,


usually expressed in dBm, necessary to
provide proper operation of the mixer.
This power level is used to bias the diodes in a
non-linear region so that mixing action
occurs.

863

Specifications
Noise Figure Every network, whether a mixer or other
component, has a certain amount of internal noise
associated with its operation.
A comparison of this level of noise to that of the desired
signal in the network is called the signal-to-noise ratio.
There are input and output signal-to-noise ratios for each
network under consideration.
The noise figure measures the reduction of a signal-to-noise
ratio by the network under consideration (for example, the
balanced mixer).
It is a term expressed in dB which compares the input and
output signal-to-noise ratios.

864

432
Lecture notes Radio Engineering by Jorma Kekalainen

Noise Figure

Expressed mathematically it is:

865

Specifications
VSWR VSWR is an important term which
indicates the match, or return loss, of each port
under actual operating conditions.
If measurements are made under abnormal
circumstances, the figures obtained are useless.
Only when the diodes in the mixer are being
driven properly are true readings obtained.

866

433
Lecture notes Radio Engineering by Jorma Kekalainen

Mixer applications
The mixer finds many applications in microwave systems.
In a receiver front end the mixer is used to convert the high
frequency input signal to a lower frequency in order to obtain the
transmitted information.
It is much easier to operate on this lower frequency than on the input
signal as it appears at the antenna.
The same concept is used in microwave network analyzers.
The input GHz signals are converted down even to the kHz range for
easy operation and greater accuracy.
In applications of this type, it is very important to have low noise
mixers with a high degree of isolation and minimum conversion loss.

867

Mixer applications
Other utilizations are as an up-converter in a
transmitter (to increase the frequency for
transmission), as a balanced modulator (which
incorporates a 180 hybrid and reversed diodes
in a balanced mixer) and in many instances
where signal processing must be
accomplished.

868

434
Lecture notes Radio Engineering by Jorma Kekalainen

Image-rejection mixer
The image-rejection mixer gives the desired sideband (f1 + f2 or f1 - f2) with
a minimum of output filtering.
In a basic mixer the local oscillator may be above or below the RF signal in
frequency.
Here, let us suppose an LO below the RF.
When the mixer is operating there is an lF signal equal to the difference
between the two frequencies.
This operation is a proper one.
However, there is also a signal separated from the RF by the same IF
amount but it is above the RF signal and is called the image frequency
generated by the mixer.
Obviously there are occasions where this signal would create problems.
Thus, rejecting is needed.
The image- rejection mixer provides the solution.

869

Image-rejection mixer
Figure shows a block diagram of an image-
rejection mixer.

870

435
Lecture notes Radio Engineering by Jorma Kekalainen

Image-rejection mixer
The input coupling network consists of a combination of a 90
hybrid for the RF, a 0 power divider for the LO, and the input
coupling networks of the two very important balanced mixers.
The diode circuitry is that contained in the mixers; it is
identical to that previously discussed two diodes in each
mixer.
The output filtering consists of the 90 IF hybrid, a quadrature
hybrid designed to operate at the IF frequency.
One output port (No. 1) produces the upper image frequency;
the other (No. 2), creates the lower.

871

Pros and cons


The quadrature characteristic (90 phase
relationship) of the hybrid results in only one
frequency being present at each port when the
opposite port is terminated, thus providing
rejection of the opposite image frequency.
When a clean IF signal is wanted, this ability is
of great importance.
The disadvantage of the image-rejection mixer
is its complexity.

872

436
Lecture notes Radio Engineering by Jorma Kekalainen

Passive microwave
components
Quadrature hybrid

Quadrature hybrid
Definition 1: A quadrature hybrid is a 4-port, 3
dB coupler capable of dividing an input signal
into two mutually isolated quadrature phased
outputs while maintaining isolation of the
fourth port from the input.
Definition 2: A quadrature hybrid is a
directional coupler whose two outputs are
equal in amplitude and separated from one
another by a constant 90 phase.

874

437
Lecture notes Radio Engineering by Jorma Kekalainen

Quadrature hybrid symbols

That of having output signals in quadrature,


separated 90 from one another, is one of the
most important property of the quadrature
hybrid component. In Figure ports 2 and 3
would be in quadrature with an input at either
port 1 or 4.

875

Typical data sheet for a quadrature


hybrid

876

438
Lecture notes Radio Engineering by Jorma Kekalainen

Specifications
Amplitude Balance This figure is a
comparison of the power levels of the two
output ports on the hybrid ideally there
should be no difference in amplitude between
the two outputs, but in the real world there
naturally is some variation.
It is necessary for this figure to be as low as
possible so that both circuits driven have the
same level applied to them.

877

Specifications
Insertion Loss Figure illustrates the insertion loss of
a quadrature hybrid basically the difference between
the maximum height of the coupled curve (1-2) and the
straight-through curve (1-3).
This internal loss of the device causes its output
response to deviate from an ideal 3 dB figure across the
desired band.

878

439
Lecture notes Radio Engineering by Jorma Kekalainen

Specifications
Isolation This term tells you how good a
conjugate pair is formed by the terminals.
The isolation between terminals is very
important to consider if the hybrid is going to
be used to drive two active devices, such as
transistors or diodes.
If the figure is not high enough, any imbalance
in one device affects both of them and may
eventually cause one or both to be destroyed.
A network has a conjugate pair if an input applied to one of the branches has no effect on the
879
other; that is, two branches form a conjugate pair if they are isolated from one another.

Specifications
Phase Balance The quadrature hybrid is
designed to have a 90 separation between the
output terminals.
But, because of variations within the hybrid,
the phase of the output varies with frequency.
The phase balance figure on a data sheet is a
measure of how well the two output terminals
track over the band of operation.

880

440
Lecture notes Radio Engineering by Jorma Kekalainen

Applications of quadrature
hybrids
Since they provide isolation between the RF
and LO while applying equal signals to each of
the diodes, quadrature hybrids are an integral
part of a balanced mixer.

881

Applications of quadrature
hybrids
They are ideal for a matched detector, also
because of the equal level applied to the
detector diodes.

882

441
Lecture notes Radio Engineering by Jorma Kekalainen

RF and DC combination

Quadrature hybrids are also appropriate as a


means of DC biasing an RF circuit they
supply a continuous DC path while providing
isolation for RF.

883

Applications of quadrature
hybrids
A common application is combining of amplifiers.
The ability of the hybrid to provide isolated signals of
equal amplitude to a device is of great value.
The quadrature effect (90 separation) allows the split
of signals at the input and their recombination at the
output.
Combining amplifiers

884

442
Lecture notes Radio Engineering by Jorma Kekalainen

Applications of quadrature hybrids


Other areas in which quadrature hybrids are used as 2-way
power dividers, time delay circuits, and modulators, and as
variations on the simple PIN switch.
Obviously, the quadrature hybrid is a component with many
applications.
When a device is needed to provide two output ports of equal
amplitude and which are isolated from one another, the hybrid
satisfies that requirement.
Its quadrature characteristics are also of great value when
division and recombination are to be accomplished with active
devices.

The PIN diode is a semiconductor device that consists of three regions:


an intrinsic region is layered between an n-doped and a p-doped region. The PIN diode
885
can control large RF and microwave signals with comparably small DC signals.

Signal flow graphs and


scattering parameters

443
Lecture notes Radio Engineering by Jorma Kekalainen

Multiport networks
A general circuit can be represented by a multi-port network, where the
ports are defined as access terminals at which we can define voltages and
currents.
Note: Equal and opposite currents are
assumed on the two wires of a port.
Examples:

One-port network I1 I1

V1 V1

Two-port network
I2
I1 I2 I1

V1 V2 V1
V2

887
887

Signal flow graphs


The signal flow graph is a very useful
technique for the analysis of microwave
networks in terms of transmitted and reflected
waves.
Main phases of the signal flow graph
technique use:
the construction of the flow graph itself,
the reduction, or solution, of the flow graph.

888

444
Lecture notes Radio Engineering by Jorma Kekalainen

Features and the construction of


the flow graph
The primary components of a signal flow
graph are nodes and branches:
Nodes: Each port, i, of a microwave network has
two nodes, ai and bi.
Node ai is identified with a wave entering port i, while
node bi, is identified with a wave reflected from port i.
Branches: A branch is a directed path between an
a-node and a b-node, representing signal flow
from node a to node b.
Every branch has an associated S parameter or
reflection or transmission coefficient.
889

Signal flow graph of a two-port


network
It is useful to consider the flow graph of an arbitrary two-port
network with incident and reflected waves at each port, and the
corresponding signal flow graph representation.
Definition of incident and reflected waves

Signal flow graph

890

445
Lecture notes Radio Engineering by Jorma Kekalainen

Graphical illustration of the


network behavior
The flow graph gives an intuitive graphical illustration of the
network behavior.
For example, a wave of amplitude a1 incident at port 1 is split, with part
going through S11 and out port 1 as a reflected wave and part transmitted
through S21 to node b2.
At node b2, the wave goes out port 2; if a load with nonzero reflection
coefficient is connected at port 2 this wave will be at 1east partly reflected
and reenter the two-port network at node a2. Part of the wave can be
reflected back out port 2 via S22, and part can be transmitted out port 1
through S12.

Source Load
end end

891

Examples
The signal flow graph representations of a
one-port network and a voltage source

892

446
Lecture notes Radio Engineering by Jorma Kekalainen

Flow graph algebra

Once a microwave network has been


represented in signal flow graph form, it is not
very difficult to solve the ratio of any
combination of wave amplitudes.
This can be done using basic decomposition
rules or applying Masons rule.

893

What are S-parameters?


S-parameters or scattering parameters are
reflection and transmission coefficients.
Transmission coefficients are commonly called
gains or attenuations; reflection coefficients
are directly related to VSWRs and
impedances.
S-parameters are vector quantities they
give magnitude and phase information.

894

447
Lecture notes Radio Engineering by Jorma Kekalainen

Parameter models

Conceptually, s parameters are like h or


z parameters because they describe the
inputs and outputs of a black box.
The inputs and outputs are in terms of power
for s parameters;
for h, y and z parameters, they are voltages
and currents.

895

I/O-signals
Using the convention that a is a signal into a
port and b is a signal out, the figure below
helps to explain s parameters.

Test device

896

448
Lecture notes Radio Engineering by Jorma Kekalainen

I/O-signals

Here, a and b are the square roots of


power;
(a1)2 is the power incident at port 1 and
(b2)2 is the power leaving port 2.
The diagram shows the relationship between
the s parameters and the as and bs.
Test device

897

Reflection and transmission


coefficients
For example, a signal, a1, is partially reflected at port
1; the rest of the signal is transmitted through the
device and out of port 2.
The fraction of a1 that is reflected at port 1 is s11; the
fraction of a1 that is transmitted is s21.
Similarly, the fraction of a2 that is reflected at port 2
is s22, and the fraction s12 is transmitted.
Test device

898

449
Lecture notes Radio Engineering by Jorma Kekalainen

I/O-relations

The signal, b1, leaving port 1 is the sum of the


fraction of a1 that is reflected at port 1 and the
fraction of a2 that is transmitted from port 2.
Thus, the outputs can be related to the inputs
by the equations:
b1 = s11a1 + s12a2
b2 = s21a1 + s22a2

899

R-T coefficients
From the previous equations
b1 = s11a1 + s12a2
b2 = s21a1 + s22a2
when a2=0 s11=b1/a1 and s21=b2/a1

when a1=0 s12=b1/a2 and s22=b2/a2

900

450
Lecture notes Radio Engineering by Jorma Kekalainen

Measuring R-T coefficients


The setup below shows how s11 and s21 are measured.

Port 1 is driven and a2 is made zero by terminating


the 50 transmission line coming out of port 2 in its
characteristic 50 impedance.
This termination ensures that none of the transmitted
signal, b2, is reflected toward the test device.
901

Measuring transmission
coefficients
Similarly, the setup for measuring s12 and s22 is:

Compare this to the previous setup for measuring s11


and s21 .

902

451
Lecture notes Radio Engineering by Jorma Kekalainen

Device under test


The S-parameters can be measured by
embedding the two-port network (the device-
under-test) in a transmission line whose ends
are connected to a network analyzer.

An experimental setup:

903

Two-port network under test


Figure shows more details of the connection.
The generator and load impedances are configured by
the network analyzer.
The connections can be reversed, with the generator
connected to port 2 and the load to port 1.

904

452
Lecture notes Radio Engineering by Jorma Kekalainen

Two-port network under test


The two line segments of lengths 11,12 are assumed to have characteristic
impedance equal to the reference impedance Z0.
Then, the wave variables a1, b1 and a2, b2 are recognized as normalized
versions of forward and backward traveling waves.
The network analyzer measures the waves a1, b1 and a2, b2 at the
generator and load ends of the line segments, as shown in the figure.
From these, the waves at the inputs of the two-port can be determined.

905

Comment
S-parameters are vector quantities they
give magnitude and phase information.
Most measurements of microwave components
have been measured historically only in terms
of magnitude.
Obviously because it was too difficult to
obtain both phase and magnitude information.

906

453
Lecture notes Radio Engineering by Jorma Kekalainen

Parameters
To represent multi-port networks we use:

- Z (impedance) parameters Not easily


- Y (admittance) parameters measurable at
- h (hybrid) parameters high
- ABCD parameters frequency

Measurable at high
- S (scattering) parameters
frequency

907

Summary
2-port and corresponding S-formalism

908

454
Lecture notes Radio Engineering by Jorma Kekalainen

Summary
N-port and corresponding S-matrix

909

Summary
At high frequencies, Z, Y, h & ABCD parameters are
difficult (if not impossible) to measure.
V and I are not uniquely defined.
Even if defined, V and I are very difficult
to measure.
Required open and short-circuit conditions are
often difficult to achieve.
Scattering (S) parameters are often the best
representation for multi-port networks at high
frequencies.
Instead of open and short-circuit conditions only matched
terminal is needed.
910

455
Lecture notes Radio Engineering by Jorma Kekalainen

Summary
b1 0 S11a1 0 S12a2 0
b2 0 S21a1 0 S22a2 0
Scattering parameters
b1 0 Output is
S11 matched input reflection coefficient
a1 0 a 0
2 with output matched
b1 0 Input is reverse transmission coefficient
S12 matched
a2 0 a 0 with input matched
1

b2 0 Output is forward transmission coefficient


S 21 with output matched
a1 0 a 0 matched
2

output reflection coefficient


b 0 Input is
with input matched
S 22 2 matched
a2 0 a 0
1 911

Summary

912

456
Lecture notes Radio Engineering by Jorma Kekalainen

Introduction to transmission
line theory

Wave-guiding structures
A wave-guiding structure is one that carries a signal (or
power) from one point to another.

There are three common types:


Transmission lines
Fiber-optic guides
RF waveguides

Note: An alternative to waveguiding structures is wireless transmission in more or less free-


space using antennas. 914

457
Lecture notes Radio Engineering by Jorma Kekalainen

Transmission line properties

Transmission line
Has two conductors running parallel
Can propagate a signal at any frequency (in theory)
Becomes lossy at high frequency
Can handle low or moderate amounts of power
Does not have signal distortion, unless there is loss
More or less immune to interference
Does not have Ez or Hz components of the fields (TEMz)

915

Note: Compare properties


RF waveguide Is a single hollow metal pipe
Can propagate a signal only at high frequency: >
c
The width must be at least one-half of a wavelength
Has signal distortion, even in the lossless case
Immune to interference
Can handle large amounts of power
Has low loss (compared with a transmission line)
Has either Ez or Hz component of the fields (TMz or
TEz)
Optical fiber wave guide Uses a dielectric rod
Can propagate a signal at any frequency (in theory)
Can be made very low loss
Has minimal signal distortion
Very immune to interference
Not suitable for high power
Has both Ez and Hz components of the fields 916

458
Lecture notes Radio Engineering by Jorma Kekalainen

Example: Transmission line


Twisted pair

Coaxial cable

The two wires of the transmission line are twisted to reduce interference and radiation from
discontinuities. 917

Example: Transmission lines


Transmission lines on printed-circuit boards

w
r
h
r h w

Microstrip Stripline
w w w

r h r h

Coplanar strips Coplanar waveguide (CPW)

Using transmission lines to synthesize loads or filters is very useful in microwave


engineering. E.g., A microwave filter constructed from microstrip. 918

459
Lecture notes Radio Engineering by Jorma Kekalainen

Transmission line theory


Lumped circuit elements: resistors, capacitors,
inductors
neglect time delays (phase)

Distributed circuit elements: transmission lines

account for propagation and


time delays (phase change)

We need transmission-line theory whenever the


length of a line is significant compared with a
wavelength.
919

Lumped-element model of a
transmission line
Transmission line = the length of the line > /10 (or /30,
definition issue)
A very short piece (dz << ) of a transmission line can be
modeled with a lumped-element equivalent circuit.

920

460
Lecture notes Radio Engineering by Jorma Kekalainen

Transmission line
Two conductors

Four per-unit-length parameters:

C = capacitance/length [F/m]
L = inductance/length [H/m]
R = resistance/length [/m] z

G = conductance/length [ S/m]
921

Transmission line model


i z, t
+++++++
----------
v z, t

i(z,t) Rz Lz i(z+z,t)

+ +

v(z,t) Gz Cz v(z+z,t)

- -
z
922

461
Lecture notes Radio Engineering by Jorma Kekalainen

Transmission line model


i(z,t) Rz Lz i(z+z,t)

+ +

v(z,t) Gz Cz v(z+z,t)

- -
z

i ( z , t )
v( z, t ) v( z z , t ) i ( z, t ) Rz Lz
t
v( z z, t )
i ( z, t ) i ( z z , t ) v( z z, t ) G z C z
t
923

Transmission line equations


Thus,

v( z z, t ) v( z, t ) i ( z , t )
Ri ( z, t ) L
z t
i ( z z , t ) i ( z , t ) v( z z , t )
Gv( z z , t ) C
z t
Now let z 0:

v i
Ri L
z t Telegraphers
Equations
i v
Gv C
z t 924

462
Lecture notes Radio Engineering by Jorma Kekalainen

Derivation of the wave equation


To combine these, take the derivative of the first
one with respect to z:

v
2
i i
R L
z 2
z z t Switch the
order of the
i i derivatives.
R L
z t z
v
R Gv C
t
v v
2

L G C
t t 2

925

Wave equation

v
2
v v v 2

R Gv C L G C
z 2
t t t 2

Thus, we have:

v
2
v v
2

RG v ( RC LG ) LC 0
z 2
t t
2

The same equation also holds for i v.


926

463
Lecture notes Radio Engineering by Jorma Kekalainen

Time-harmonic version
Time-harmonic waves:

v 2
v v
2

RG v ( RC LG ) LC 0
z 2
t t
2

dV
2

RG V ( RC LG ) jV LC ( )V 0 2

dz 2

927

Time-harmonic wave equation


dV
2

RG V j ( RC LG )V LC V 2

dz 2

Note that the coefficient of V can express as a product

RG j ( RC LG ) LC ( R j L)(G j C )
2

Z R j L = series impedance/length

Y G jC = parallel admittance/length

dV2

Then we can write: ( ZY )V


dz 2
928

464
Lecture notes Radio Engineering by Jorma Kekalainen

Solution
ZY dV 2

( )V
2
Let Then 2

dz 2

Solution: V ( z ) Ae Be z z

is called the "propagation constant.

( R j L)(G jC )
1/2
Convention:
z z e j / 2
principal square root

j attenuatio n constant

0, 0 phase constant 929

Phasor time function


Forward travelling wave (a wave traveling in the positive z direction):

V ( z ) V0 e z V0 e z e j z


v ( z, t ) Re V0 e z e j z e jt
Re V 0

e j e z e j z e
jt
The wave repeats when:

V0 e z cos t z 2
g
t 0
g
Thus:
V e
2
z

930

465
Lecture notes Radio Engineering by Jorma Kekalainen

Attenuating wave

2 2
The local maximum values when:

g

t 0
g
V 0

e z

931

Phase velocity
Track the velocity of a fixed point on the wave (a point of constant
phase), e.g., the crest.

vp (phase velocity)

v ( z, t ) V0 e z cos(t z )
932

466
Lecture notes Radio Engineering by Jorma Kekalainen

Phase velocity

Set: t z constant
dz
Take d/dt: 0
dt
dz

dt
In expanded form:

Hence
v v
Im ( R j L)(G jC )
p 1/ 2
p

933

Characteristic impedance Z0
I+ (z)
+
V+(z)
- z

A wave is traveling in the positive z direction.

V ( z)
Z0
I (z)
V ( z ) V0 e z V0
so Z0
I ( z ) I 0 e z I0

Note: Z0 is a number, not a function of z. 934

467
Lecture notes Radio Engineering by Jorma Kekalainen

Characteristic impedance Z0
Use Telegraphers Equation:

v i
Ri L
z t

dV
so RI j LI Note:

dz V ( z ) V0 e z
ZI I ( z ) I 0 e z

Thus V0 e z ZI 0 e z
935

Characteristic impedance Z0


1/2
V Z Z
From this we have: Z 0 0
2
ZY
I0 Y

Using
Z R j L
Y G jC

1/2
We have R j L
Z0
G j C

936

468
Lecture notes Radio Engineering by Jorma Kekalainen

Forward and reverse wave


V z V0 e z V0 e z

V0 e j e z e jz V0 e j e z e jz

wave in +z
Note: direction wave in -z
direction

v z , t Re V z e jt
V0 e z cos t z
V0 e z cos t z
937

Reverse wave

I - (z)
+
V -(z)
- z

A wave is traveling in the negative z direction.

V ( z) V (z)
Z0 so Z0
I ( z) I (z)

Note: The reference directions for voltage and current are the same as for the
forward wave. 938

469
Lecture notes Radio Engineering by Jorma Kekalainen

Superposition of forward and backward


waves
I (z)
+
V (z)
- z

A superposition of forward and


General case: backward traveling waves:

V ( z ) V0 e z V0 e z
1
I ( z) V0 e z V0 e z
Z0

Note: The reference directions for voltage and current are the same for forward
and backward waves. 939

Recap

940

470
Lecture notes Radio Engineering by Jorma Kekalainen

Recap

V z V0 e z V0 e z Guided wavelength g

I z
V z V z 2
e
m
0 0
e g
Z0 Z0
1

j R j L G jC 2
Phase velocity vp
1
R j L
2
Z0 vp [m/s]
G jC
941

Losless transmission line


R 0, G 0

j ( R j L)(G j C )
1/ 2

j LC

so 0
vp
LC

1/ 2
R j L L 1
Z0 Z0 vp
G jC C LC
Note: Now Z0 and vp are independent of frequency. 942

471
Lecture notes Radio Engineering by Jorma Kekalainen

Lossless transmission line


1
vp
LC
If the medium between the two conductors is homogeneous
(uniform) and is characterized by (, ), then it can be shown
that
LC

The speed of an electromagnetic 1


cd
wave in a dielectric medium is

Thus v p cd
The phase velocity does not depend on the frequency, and it is
always the speed of light in this kind of material. 943

Terminated transmission line


Terminating impedance (load)
V z V e 0
z
V e0
z

Amplitude of voltage wave


propagating in positive z
direction at z = 0.
Amplitude of voltage wave
propagating in negative z
direction at z = 0.

Where do we assign z = 0?
z
The usual choice is at the load.

Note: The length l measures distance from the load l=-z 944

472
Lecture notes Radio Engineering by Jorma Kekalainen

Change of reference plane


Terminating impedance (load)
V z V0 e z V0 e z

What if we know amplitudes


V+ and V- at z=-l


Can we use z = - l as
a reference plane?

V0 V 0 V e V V 0 e

V0 V 0 V e
Hence

V z V e V e
z z

945

Change of reference plane


Terminating impedance (load)


Compare:

V z V 0 e z V 0 e z

V z V e V e
z ( ) z ( )

Note: This is simply a change of reference plane, from z = 0 to z = -l. 946

473
Lecture notes Radio Engineering by Jorma Kekalainen

What is V(-l )?
Terminating impedance (load)
V z V0 e z V0 e z

What is V(-l )?

Wave propagating Wave propagating


forwards (load) backwards

The current at z = - l is then

l = distance away from load


947

Load reflection coefficient (L)

Total voltage at distance


l from the load

Amplitude of voltage wave


propagating towards load at Amplitude of voltage
L = Load reflection coefficient
the load position (z = 0). wave propagating away
from load at the load
position (z = 0).

Similarly,

Note: l = Reflection coefficient at z = - 948


l

474
Lecture notes Radio Engineering by Jorma Kekalainen

Input impedance Z(-l )

Z0 ,

Input impedance seen looking towards


load at z = -l .
949

Relation between ZL and L


At the load (l = 0):

Z L Z0
L
Z L Z0

Recall:

Thus, Z L Z0 2
1 e
Z L Z0
Z Z 0
Z Z 2
1 L 0
e
L Z Z 0
950

475
Lecture notes Radio Engineering by Jorma Kekalainen

Local impedance along terminated


transmission line
Simplifying, we have
Z L Z0 2
1 e
Z Z0 Z Z L Z0 Z L Z0 e
2
e l
Z Z0 L
0
2 l
Z L Z0 Z L Z0 e
Z Z 2
1 L 0
e e
Z L Z0
Z Z 0 e Z L Z 0 e e x e x e x e x
Z0 L
cosh x and sinh x

Z L Z0 e Z L Z 0 e
2 2

Z cosh Z0 sinh cosh l


Z0 L
Z0 cosh Z L sinh cosh l

Thus, we have

Z Z0 tanh
Z Z0 L
Z0 Z L tanh
951

Terminated lossless transmission line


j j

V V0 e j 1 L e2 j Impedance is periodic with


period g/2 because tan( )
V0 j repeats when
I e 1 Le2 j
Z0
1 Le 2 j
2
Z Z 0 2 j
1 Le
g
Z jZ0 tan g / 2
Z Z0 L
Z jZ tan
0 L
e x e x ex ex
cosh x and sinh x
Note: 2
e jx e jx
2
e jx e jx
tanh tanh j j tan
cosh jx cosx and sinh jx j sin x 952
2 2

476
Lecture notes Radio Engineering by Jorma Kekalainen

Lossless transmission line


From now on in our transmission line discussion we will assume that
the transmission line is lossless.

Z0 ,

953

Matched load

Z0 ,

1 Matched load: (ZL=Z0)

No reflection from the load

For any l 954

477
Lecture notes Radio Engineering by Jorma Kekalainen

Short-Circuit (S-C) load


2 Short-circuit load: (ZL = 0)

Z0 ,


Note: 2
g Always imaginary!

/ g

S-C can become an O-C with a


g/4 transmission line
955

Example
Find the voltage at any point on the line.

Z0

956

478
Lecture notes Radio Engineering by Jorma Kekalainen

Example

Note:

At l = d :

Zin j d 1
V0 VTH e j 2 d
Zin ZTH 1 Le
Thus

Zin j d 1 Le j 2
V VTH e j 2 d
Z m ZTH 1 Le 957

Example
1 L e j 2 d
Zin Z d Z0 j 2 d
1 Le

958

479
Lecture notes Radio Engineering by Jorma Kekalainen

Example
After algebraic manipulation we have

where

Therefore, we have the following alternative form for the result:

Z0 j d 1 L e j 2
V VTH e j 2d
Z0 ZTH 1 S Le
959

Example

Z0

Z 0 j d 1 L e j 2
V VTH e j 2 d
Z0 ZTH 1 S Le

Note: Voltage wave that would exist if there were no reflections from the load (a
960
matched load or a semi-infinite transmission line ).

480
Lecture notes Radio Engineering by Jorma Kekalainen

Time-averaged power flow

Z0 ,
At a distance l from the load:

Note: V V0 e 1 Le 2
V0
I e 1 L e2
Z0
j

*
If Z0 real (low-loss transmission line) L e2 *L e2
Le2 L e2
*

pure imaginary
961

Time-averaged power flow

Z0 ,
Low-loss line

Lossless line ( = 0): P


962

481
Lecture notes Radio Engineering by Jorma Kekalainen

Quarter-wave (g/4) transformer

Z jZ0T tan Z0 Z0T ZL


Zin Z0T L
Z0T jZ L tan Zin

g 2 g Matching:
in 0 Z in Z 0
4 g 4 2
Z 02T
Z0
ZL
This requires ZL to be real.
Thus
Hence
2
Z
Z in Z 0T Z 0 Z L
0T 1/2
ZL
963

Voltage Standing Wave Ratio (VSWR)

Z0 ,

1+ L
V (z)
1 V0
Vmax V0 1 L
1- L
Vmin V0 1 L z
z / 2
z0

Vmax 1 L
Voltage Standing Wave Ratio VSWR VSWR
Vmin 1 L 964

482
Lecture notes Radio Engineering by Jorma Kekalainen

Example: Coaxial cable


Unit length
a

Coaxial cable b r ,
Parameters per unit length (C, L, G, R)

For a TEMz mode, the shape of the fields is independent of frequency, and hence
parameter calculation can be performed using electro- and magnetostatics.

2 0 r 2
C [F/m] G [S/m]
b b b
ln L 0 r ln [H/m] ln
a 2 a a
965

Example: Coaxial cable

2 0 r
C [F/m]
b
ln
a LC 0 0 r r
0 r b
L ln [H/m]
2 a

This result actually holds for any transmission line.

966

483
Lecture notes Radio Engineering by Jorma Kekalainen

Example: Coaxial cable


L
For a lossless cable: Z0
C
2 0 r 0 r b
C [ F/m] L ln [H/m]
b 2 a
ln
a

r 1 b
Z 0 0 ln [ ]
r 2 a

0
0 376.7303 []
0
967

Example: Coaxial cable

2
C [F/m] 0 r
b
ln
a
G C
2
G [S/m]
b
ln
a

This result actually holds for any transmission line.

968

484
Lecture notes Radio Engineering by Jorma Kekalainen

Example: Coaxial cable


G C

More generally:

G
tan The loss tangent would arise
C from conductivity effects.

G The loss tangent actually arises from both


tan conductivity loss and polarization loss
C (molecular friction loss).

Note: It is the loss tangent that is usually (approximately) constant for a


material, over a wide range of frequencies. 969

Note
General expression for loss tangent:

e j Effective permittivity that accounts for conductivity


' j ' ' j

e ' j e ' '

Loss due to molecular friction Loss due to conductivity

The loss tangent tan


' '
'' characterizes the attenuation of
tan e the electric field in a lossy
e' '
material.
970

485
Lecture notes Radio Engineering by Jorma Kekalainen

Example: Coaxial cable


R (resistance / unit length): R Ra Rb
1
Ra Rsa
Rs = surface resistance of metal 2 a
1
b , rb Rb Rsb
2 b
a , ra 1 1
Rsa Rsb
a a a b b

2 2
b a b
0 ra a 0 rb b
Is the skin depth, the distance over which the signal amplitude decreases to 1/e of
its surface value. 971

Example: Coaxial cable


L
(1) Z 0lossless characteristic impedance of line (neglecting loss)
C

(2) LC 0 0 r r

G 1 1
(3) tan (4) R Rsa Rsb
C 2 a 2 b

Equations (1) and (2) can be used to find L and C if we know the material
properties and the characteristic impedance of the lossless line.

Equation (3) can be used to find G if we know the material loss tangent.

Equation (4) can be used to find R.


972

486
Lecture notes Radio Engineering by Jorma Kekalainen

Example: Coaxial cable


All four per-unit-length parameters can be found
from material parameters, Z0lossless and R.

L Z 0lossless

C / Z 0lossless

G C tan

RR

973

Limitations of transmission line theory


At high frequency, discontinuity effects can become important.
transmitted

incident

Bend

reflected

The simple transmission line model does not account for the bend.
ZTH

+- Z0 ZL

974

487
Lecture notes Radio Engineering by Jorma Kekalainen

Idealizations of transmission line theory

At high frequency, radiation effects can become important.


We want energy to travel from the generator to the load without radiation.

ZTH

+- Z0 ZL

When will radiation occur?

975

Idealizations of transmission line


theory
The coaxial cable is a perfectly shielded
system there is never any radiation at any
frequency, or under any circumstances.

r a
z b

The fields are confined to the region


between the two conductors.

976

488
Lecture notes Radio Engineering by Jorma Kekalainen

Idealizations of transmission line


theory
The twin lead is an open type of transmission
line the fields extend out to infinity.

+ -
The extended fields may cause
interference with nearby objects.
(This may be improved by using
twisted pair.)

977
Note: Having fields that extend to infinity is not the same thing as having radiation.

Idealizations of transmission line


theory
The infinite twin lead will not radiate by itself,
regardless of how far apart the lines are.

backward
forward

No attenuation on an infinite lossless line.

The forward and backward waves represent an exact solution


to Maxwells equations on the infinite line, at any frequency.
978

489
Lecture notes Radio Engineering by Jorma Kekalainen

Limitations of transmission line theory


A discontinuity on the twin lead will cause radiation to occur.

Incident wave
pipe

Reflected wave

Note: Radiation
effects increase Incident wave
as the frequency
increases.

h bend

Reflected wave 979

Limitations of transmission line theory


To reduce radiation effects of the twin lead at discontinuities:

1) Reduce the separation distance h (keep h << ).


2) Twist the lines (twisted pair).

CAT 5 cable
(twisted pair)
980

490
Lecture notes Radio Engineering by Jorma Kekalainen

Introduction to transmission
line theory
Recap and supplement

Transmission lines

982

491
Lecture notes Radio Engineering by Jorma Kekalainen

Transmission line circuit

983

Boundary condition at load

This gives us the backward going


wave amplitude in terms of the
forward going wave amplitude.

984

492
Lecture notes Radio Engineering by Jorma Kekalainen

Load reflections

985

Transmission line as an impedance


transformator

986

493
Lecture notes Radio Engineering by Jorma Kekalainen

Transmission line as an impedance


transformator

987

Equivalent circuit

988

494
Lecture notes Radio Engineering by Jorma Kekalainen

Special case: Short-circuit load

Note: Impedance seen at the source end is inductive. The transmission line appears
989
like an inductor.

Special case: Short-circuit load

990

495
Lecture notes Radio Engineering by Jorma Kekalainen

Special case: Open-circuit load

Note: Impedance seen at the source end is capacitive. The transmission line appears
991
like a capacitor.

Special case: Open-circuit load

992

496
Lecture notes Radio Engineering by Jorma Kekalainen

Matched load

For matched loads:


The load reflection coefficient L is zero.
There is no reflected wave generated at the load end ( i.e. V_ = 0 ).
The impedance seen at the source end is Z0 irrespective of the length of
the transmission line.

993

Periodic impedance transformation

Note: Impedance is periodic with period equal to half-wavelength. 994

497
Lecture notes Radio Engineering by Jorma Kekalainen

Voltage Standing Wave Ratio (VSWR)

The interference of forward and backward going waves leads


standing-wave behavior.

Voltage Standing Wave Ratio = VSWR =


995

Equivalent circuit

Thus, we have produced the equivalent impedance


Z(z=-l) of the transmission line + load as seen by the
source.

996

498
Lecture notes Radio Engineering by Jorma Kekalainen

Example
Reduce the transmission line circuit

to the Thevenin equivalent at the load terminals:

Note: Now we want to produce the Thevenin equivalent of the source +


transmission line as seen by the load. 997

Example
1 To find Zth short the voltage source and find the impedance
looking in from the load terminals

2 To find Vth remove the load and find the voltage at the load
terminals

998

499
Lecture notes Radio Engineering by Jorma Kekalainen

Cascaded infinite transmission lines


Let us study the situation where we have two unmatched transmission
lines connected together.

Now

Boundary Conditions:
(1) At z=0 the voltage on both the transmission lines must be the same.
(2) At z=0 the current on both the transmission lines must be the same.

999

Example: Power splitting in microwave circuits

Input wave: Transmitted wave:


Reflected wave:

Goal: Express V+2 and V-1 in terms of V+1.

1 Cast the circuit in the following equivalent form and find V-1

2 Voltage V+2 is the same as the voltage across the impedance Zo2
in the equivalent circuit

1000

500
Lecture notes Radio Engineering by Jorma Kekalainen

Transmission line junctions and


discontinuities

1001

Transmission line junctions and


discontinuities

1002

501
Lecture notes Radio Engineering by Jorma Kekalainen

Multi-sectional transmission line


Question: How does one solve a problem like this?

1003

Example
Question: How does one solve a problem like this?

1 Replace the last line with a lumped equivalent impedance


(corresponding to an infinite line).

2 Replace the middle line with the impedance Z(z=- )

1004

502
Lecture notes Radio Engineering by Jorma Kekalainen

Matching transmission lines


Is it possible to use a transmission line to perfectly match two dissimilar
transmission lines so that there is no reflection?

What is the appropriate impedance Zo2? What is the appropriate


length ?
Use a Quarter-wave transformer:

A quarter-wavelength long transmission


line inverts the normalized impedance1005

Matching transmission lines

1006

503
Lecture notes Radio Engineering by Jorma Kekalainen

Smith chart

Smith chart
The Smith chart is a graphical aid that is very useful when
solving transmission line problems.
The key to its understanding is to realize that it is essentially a
polar plot of the voltage reflection coefficient ().
Let the reflection coefficient be expressed in polar form as

e j

Then the magnitude | | is plotted as a radius (| | 1) from the


center of the chart and the angle (-180 180 ) is
measured from the right-hand side of the horizontal diameter.
Any passively realizable (| | 1) reflection coefficient can
then be plotted as a unique point on the Smith chart.
1008

504
Lecture notes Radio Engineering by Jorma Kekalainen

1009

Normalized impedance
The real utility of the Smith chart, however, lies in the fact that
it can be used to convert from reflection coefficients to
normalized impedances (or admittances), and vice versa using
the impedance (or admittance) circles printed on the chart.
When dealing with impedances on a Smith chart, normalized
quantities are generally used, which will be denoted by
lowercase letters (z).
The normalization constant is usually the characteristic
impedance of the line (Z0).
Thus, z= Z/Z0 represents the normalized version of the
impedance Z.

1010

505
Lecture notes Radio Engineering by Jorma Kekalainen

Terminated transmission line

Total voltage at distance


l from the load

Amplitude of voltage wave


propagating towards load, Amplitude of voltage L Load reflection coefficient
at the load position (z = 0). wave propagating away
from load, at the load
position (z = 0).

Similarly,

Note: l reflection coefficient at z = - 1011


l

Generalized reflection coefficient

Impedance
calculation:

Generalized reflection coefficient: 1012

506
Lecture notes Radio Engineering by Jorma Kekalainen

Generalized reflection coefficient

For
Re Z L 0
L 1

Proof:

Lossless transmission line ( = 0) L


RL jX L Z 0
RL jX L Z0
R Z 0 jX L
L
RL Z0 jX L

R Z 0 X L2
2
2
L L
RL Z0 X L2
2
1013

Complex plane
Im Decreasing l (toward load)
R j I
j 2
Le l
L
j L 2
L e L Increasing l (toward
L generator)

Re
L 2

Lossless line

1014

507
Lecture notes Radio Engineering by Jorma Kekalainen

Impedance (Z) chart


1
Z Z0 ,
1
Z 1
Zn
Z0 1
Define:
Zn Rn jX n ; R j I

Substitute into above expression for Zn(-l ):

1 R j I
Rn jX n
1 j
R I

Next, multiply both sides by the RHS denominator term and equate real and
imaginary parts. Then solve the resulting equations for R and I in terms of Rn
and Xn. This gives two equations. 1015

Impedance (Z) chart


1) Equation 1:
2
Rn 2 1 Equation for a circle
R I in the plane:
1 Rn 1 Rn
I

1016

508
Lecture notes Radio Engineering by Jorma Kekalainen

Impedance (Z) chart


2 2
1 1
R 1 I
2
2) Equation 2:
Xn Xn
I
Equation for a circle in the plane:
0 Xn 1 Xn 1

1 Xn

Xn 0
R

X n 1
0 X n 1
X n 1
1017

Impedance (Z) chart


Smith Chart (Z-Chart) Clear-cut version

Xn = 1

Rn = 1

Xn = -1

1018

509
Lecture notes Radio Engineering by Jorma Kekalainen

Impedance (Z) chart

Xn = 1 plane
Rn = 1

Xn = -1

1019

Impedance transformation along


transmission line

Define a position dependent reflection


coefficient as: (z) is a complex number
of magnitude never
greater than unity.
Normalized resistance

Normalized reactance
Thus

1020

510
Lecture notes Radio Engineering by Jorma Kekalainen

Complex plane (-plane)

1021

-plane with normalized resistance curves

1022

511
Lecture notes Radio Engineering by Jorma Kekalainen

-plane with normalized reactance curves

1023

Smith chart

The Smith chart enables


one to read values of both
the normalized resistance
Rn(z) and the normalized
reactance Xn(z) values on
the -plane.

1024

512
Lecture notes Radio Engineering by Jorma Kekalainen

Which chart to use?


When adding elements in series use Z-chart.

When adding elements in parallel use Y-chart.

1025

Example: How to use the Smith chart


for impedances
We wish to determine the input impedance of a transmission line into which have
been inserted series components, and which is terminated into a mismatched load
with complex impedance.
A computation would require several successive applications:

Zk ZL= load impedance


s = distance from load ZL
One for every section of line.

513
Lecture notes Radio Engineering by Jorma Kekalainen

Example
In the Smith chart, the sequence of operations is
carried out graphically: every section of line produces
a rotation around the center of the chart, the rotation
angle being related to the length of the section,
divided by the wavelength.
To insert a series impedance, one simply adds its
value (normalized to the line impedance) to the one at
the line's input, obtained on the Smith chart.
When adding a reactance, one moves on a circle R =
constant; when adding a resistance, on a circle X =
constant.
1027

Example
Considering the circuit below the
procedure followed is:
1. The load terminating the line (at
right) has an impedance ZL = (2.15 -
j3)Zc.
The corresponding point A in the Smith chart
chart is located at the coordinates 2.15
- j3.
2. The line section, long of 0.087,
produces a rotation from A to B by an
angle 0.087 x 720 = 62.5 around the
chart's center, moving towards the
generator (decreasing values of z),
i.e., clockwise.
The impedance at point B read on the
chart is (0.3 - j1)Zc.
One wavelength, , corresponds to
two full turns, i.e., a 720 angle.

Zc is the characteristic impedance of the line. 1028

514
Lecture notes Radio Engineering by Jorma Kekalainen

Example
3. One then adds an impedance of j0.8
Zc, bringing to the point C at (0.3 -
j0.2)Zc.
The move from B to C follows a circle Smith chart
R = constant.
4. The section of line between C and D
produces a second rotation around the
chart's center, by an angle 0.174 x 720
= 125.3.
The input impedance of this section of
line at point D is (0.65 + j1)Zc.
5. To this last value is added the
impedance of the series capacitor -j
0.6Zc, yielding point E at (0.65 +
j0.4)Zc.
The displacement is done along a circle
R = constant (as between B and C).
Zc is the characteristic impedance of the line. 1029
1029

Example
6. The addition of a series resistance
between F and G leads to a move
along a circle X = constant, between
coordinates R = 1.6 Zc and R = 3 Zc.
7. Finally, the line section between G
and H produces a last rotation around
the center of the chart, by an angle 0.1
x 720 = 72 , yielding the input
impedance to the whole assembly as
(0.9 j1.15)Zc

Zc is the characteristic impedance of the line. 1030


1030

515
Lecture notes Radio Engineering by Jorma Kekalainen

Example

1031

Example: Calculation impedances


using Maple

1032

516
Lecture notes Radio Engineering by Jorma Kekalainen

Example

1033

Example

1034

517
Lecture notes Radio Engineering by Jorma Kekalainen

Example

1035

Example

ZL = (0.9 j1.13) Zc

1036

518
Lecture notes Radio Engineering by Jorma Kekalainen

Smith chart for admittances


When connecting components in shunt on the line (which is most often preferred at
microwaves), one wishes to use local admittances, rather than impedances.
The relationship of the local admittance, Yn, normalized to the characteristic
admittance Yc of the line, with the reflection factor, is given by:

It is the same relationship as the one for the impedance, in which is simply
replaced by .
In the Smith chart, the value of the normalized admittance Yn/Yc is obtained by
taking the image of the point Zn/Zc across the center of the chart.
This actually provides a simple graphical method to determine the inverse of a
complex number.

1037

Example: How to use the Smith


chart for admittance
The input admittance of a transmission line,
into which a capacitor, an inductor and a
resistor have been shunt-connected between
sections of line, is to be determined in the
following circuit.

1038

519
Lecture notes Radio Engineering by Jorma Kekalainen

Example
The procedure followed is exactly the same as the one outlined previously for
series-connected elements in the Smith chart for impedances.
The very same sequence A-H is followed, this time for admittances in the Smith
chart.
The input admittance obtained is YH = (0.9 j1.13) Yc.

YH = (0.9 j1.13) Yc

1039

Example: Simultaneous use of the Smith


chart for shunt and series components
We may, of course, utilize the Smith chart to connect both series impedances and
shunt admittances, taking advantage of the central imaging property indicated in

A T-circuit, having one shunt susceptance between two series reactances, is


connected to a load ZL

ZL = (2.5 j0.5)Zc

1040

520
Lecture notes Radio Engineering by Jorma Kekalainen

Example
1. The load impedance is here ZL = (2.5 j0.5)Zc so that the starting point A is
located at the coordinates 2.5 - j0.5 in the Smith chart for impedances.
2. A series capacitance is then added.
Remaining on the impedance chart, one moves along an R = constant circle, by a
distance corresponding to the reactance j2Zc; i.e., from - j0.5 to - j2.5 , reaching
point B.

ZL = (2.5 j0.)Zc

1041
1041

Example
3. The next component to be connected is a shunt susceptance.
This means that the input admittance must be determined.
This is obtained by jumping over the center of the chart, from point B to its image
B' , this time in the admittance chart.
4. The addition of a shunt susceptance -j0.6Yc is obtained by moving along a circle
of constant conductance G = constant, reaching point C' (still in the admittance
chart).

ZL = (2.5 j0.)Zc

1042
1042

521
Lecture notes Radio Engineering by Jorma Kekalainen

Example
5. The next component is connected in series, so that one has to return to the
impedance chart.
This is done by going to point C, symmetrically located across the center of the
chart.
6. The second series reactance -j2Zc is then added, in the same way as the first one
(moving along a circle R = constant).

ZL = (2.5 j0.)Zc

1043
1043

Example
The point D reached is, for the particular set of values selected, the center of the
chart.
The T-circuit, containing only lossless reactive components, has actually matched
the load ZL to the transmission line of characteristic impedance Zc (reflectionless
match).

ZL = (2.5 j0.)Zc

1044

522
Lecture notes Radio Engineering by Jorma Kekalainen

Example: Calculation using Maple

ZL = (2.5 j0.5)Zc

1045

Example: Calculation using Maple


This T-circuit has matched the load ZL to
the transmission line of characteristic
impedance Zc (reflectionless match).

ZD = (1.0 + 0.0j)

ZL = (2.5 j0.5)Zc

1046

523
Lecture notes Radio Engineering by Jorma Kekalainen

Impedance matching

Impedance matching
Impedance matching is an important practical problem in
microwave engineering.
In practice, impedance matching is a part of the larger design
process for a microwave component or system.
The basic idea of impedance matching is illustrated in figure,

which shows that an impedance matching network is placed


between a load impedance and a transmission line.

1048

524
Lecture notes Radio Engineering by Jorma Kekalainen

Matching network
The matching network is ideally lossless, to avoid
unnecessary loss of power, and is usually designed so
that the impedance seen looking into the matching
network is Z0.
Then reflections are eliminated on the transmission
line to the left of the matching network, although
there will be multiple reflections between the
matching network and the load.

1049

Benefits of impedance matching


Impedance matching or tuning is important for the
following reasons:
Maximum power is delivered when the load is matched to
the line (assuming the generator is matched), and power
loss in the feed line is minimized.
Impedance matching of receiver components (antenna,
low-noise amplifier, etc.) improves the signal-to-noise ratio
of the system.
Impedance matching in a power distribution network (such
as an antenna array feed network) will reduce amplitude
and phase errors.

1050

525
Lecture notes Radio Engineering by Jorma Kekalainen

Matching load for maximum power


In many RF circuits it is extremely desirable to be able to transfer
maximum possible time-average power to a load impedance:

What is the time-average power delivered to the load?

Maximum time-average power delivered to the load:

1051

Matching load in transmission lines

1052

526
Lecture notes Radio Engineering by Jorma Kekalainen

Example: Matching load in


transmission line

1 Find L

2 Find such that the impedance Z(z=-) has a real part of 50

1053

Example: Matching load in


transmission line
Find the desired using Smith chart

1054

527
Lecture notes Radio Engineering by Jorma Kekalainen

Example: Matching load in


transmission line

But there is a problem !

Well get rid of this unwanted reactance using stub tuners.


Stub tuners are short stubs of transmission lines that are
used to cancel out unwanted reactances in microwave circuits. 1055

Choice of matching network


As long as the load impedance, ZL, has some nonzero
real part, a matching network can always be found.
Many choices are available, however, so factors that
may be important in the selection of a particular
matching network include the following:
ComplexityNearly always in case of engineering
solutions, the simplest design that satisfies the required
specifications is the most preferable.
A simpler matching network is usually cheaper, more reliable, and
less lossy than a more complex design.

1056

528
Lecture notes Radio Engineering by Jorma Kekalainen

Choice of matching network


BandwidthAny type of matching network can ideally
give a perfect match (zero reflection) at a single frequency.
In many applications, however, it is desirable to match a load over a
band of frequencies.
There are several ways of doing this with, of course, a corresponding
increase in complexity.
ImplementationDepending on the type of transmission
line or waveguide being used, one type of matching network
may be preferable compared to another.
For example, tuning stubs are much easier to implement in
waveguide than multi-section quarter-wave transformers.
AdjustabilityIn some applications the matching network
may require adjustment to match a variable load impedance.
Some types of matching networks are better than others in this
regard.
1057

Basic impedance matching


methods
Reactive lumped elements
(L-section networks)
Inductors and capacitors
Distributed elements
Tuning stub
Quarter-wave transformer
Resistive matching
(attenuator)
Suits only for the applications
which do not suffer from the
limited power and additional
heat => generally not
acceptable

1058

529
Lecture notes Radio Engineering by Jorma Kekalainen

Matching with lumped elements


Probably the simplest type of matching network is the L
section, which uses two reactive elements to match an
arbitrary load impedance to a transmission line.
There are two possible configurations for this network, as
shown in figure below.

In either of the configurations the reactive elements may be


either inductors or capacitors, depending on the load
impedance.
Thus, there are eight distinct possibilities for the matching
circuit for various load impedances.
1059

L-matching possibilities

Eight possible configurations of the discrete two-


component matching networks.

1060

530
Lecture notes Radio Engineering by Jorma Kekalainen

Matching with lumped elements


If the frequency is low enough and/or the circuit size
is small enough, actual lumped-element capacitors
and inductors can be used.
This may be feasible for frequencies up to about 1
GHz or so, although modern microwave integrated
circuits may be small enough so that lumped
elements can be used at higher frequencies as well.
There is a large range of frequencies and circuit sizes
where lumped elements may not be realizable.
This is a limitation of the L section matching
technique.
1061

Matching with distributed


elements
Usually, distributed elements are suitable
above 1 GHz.
Basic methods are tuning stub and quarter-
wavelength transformer.
Distributed-component matching circuits can
be integrated on the surface of a printed circuit
board.

1062

531
Lecture notes Radio Engineering by Jorma Kekalainen

Single-stub tuning
A matching technique that uses a single open-circuited or
short-circuited length of transmission line (a stub),
connected either in parallel or in series with the transmission
feed line at a certain distance from the load, as shown in figure
below.

1063

Stub tuning parameters


A stub tuning circuit is convenient from a microwave
fabrication aspect, since lumped elements are not
required.
The shunt tuning stub is especially easy to fabricate in
microstrip or stripline form.
In single-stub tuning, the two adjustable parameters
are the distance (d) from the load to the stub position,
and the value of susceptance or reactance provided
by the shunt or series stub.

1064

532
Lecture notes Radio Engineering by Jorma Kekalainen

Matching conditions
For the shunt-stub case, the basic idea is to select d so
that the admittance Y, seen looking into the line at
distance d from the load is of the form Y0 + jB.
Then the stub susceptance is chosen as -jB, resulting
in a matched condition.
For the series stub case, the distance d is selected so
that the impedance Z, seen looking into the line at a
distance d from the load is of the form Z0 + jX.
Then the stub reactance is chosen as -jX, resulting in
a matched condition.
1065

Open-circuited stub vs. short-


circuited stub
The proper length of open or shorted transmission line can
provide any desired value of reactance or susceptance.
For a given susceptance or reactance, the difference in lengths
of an open- or short-circuited stub is /4.
For transmission line media such as microstrip or stripline,
open-circuited stubs are easier to fabricate since a via hole
through the substrate to the ground plane is not needed.
For lines like coax or waveguide short-circuited stubs are
usually preferred, because the cross- sectional area of such an
open-circuited line may be large enough (electrically) to
radiate, in which case the stub is no longer purely reactive.

1066

533
Lecture notes Radio Engineering by Jorma Kekalainen

Stub tuners
Stub tuners are short stubs of transmission lines that are used to cancel
out unwanted reactances in RF circuits.

1067

Matching load using series stub tuner


From the previous example

k=2/
1068

534
Lecture notes Radio Engineering by Jorma Kekalainen

Smith chart: Open-circuit stub tuner

1069

Normalized admittances on Smith chart

1070

535
Lecture notes Radio Engineering by Jorma Kekalainen

Matching load using parallel stub


tuner
It is easier to work with admittances than impedances when parallel
stubs are used since admittances in parallel can simply add.

1071

Matching load using parallel stub


tuner

1072

536
Lecture notes Radio Engineering by Jorma Kekalainen

Matching load using parallel stub


tuner
We have up to now:

k=2/ 1073

Matching load using parallel stub


tuner
Open-Circuit stub tuner on Smith chart

1074

537
Lecture notes Radio Engineering by Jorma Kekalainen

Double-stub tuning
The single-stub tuners are able to match any load impedance
(as long as it has a nonzero real part) to a transmission line, but
suffer from the disadvantage of requiring a variable length of
line between the load and the stub.
This may not be a problem for a fixed matching circuit, but
would probably pose some difficulty if an adjustable tuner
was desired.
In this case, the double-stub tuner, which uses two tuning
stubs in fixed positions, can be used.
Such tuners are often fabricated in coaxial line, with adjustable
stubs connected in parallel to the main coaxial line.
The double-stub tuner cannot match all load impedances.
1075

Double-stub tuner
The double-stub tuner circuit the load may be an arbitrary distance from the
first stub.

The stubs shown in figure are shunt stubs which are usually easier to
implement in practice than are series stubs: the latter could be used just as
well, in principle.
In either case, the stubs can be open-circuited or short-circuited.

1076

538
Lecture notes Radio Engineering by Jorma Kekalainen

Quarterwave transformer
The quarter-wave transformer is a simple circuit for
matching a real load impedance to a transmission
line.
An additional feature of the quarter-wave transformer
is that it can be extended to multi-section designs for
broader bandwidth.
If only a narrowband impedance match is required, a
single-section transformer may suffice.

1077

Quarterwave transformer
One drawback of the quarter-wave transformer is that
it can only match a real load impedance.
However, a complex load impedance can always be
transformed to a real impedance by using an
appropriate length of transmission line between the
load and the transformer, or by using an appropriate
series or shunt reactive stub.
These techniques will usually alter the frequency
dependence of the equivalent load, which often has
the effect of reducing the bandwidth of the match.
1078

539
Lecture notes Radio Engineering by Jorma Kekalainen

0/4-transformer
The single-section quarter wave matching transformer circuit
is shown in figure .

The characteristic impedance of the matching section is

Z1 Z 0 Z L

At the design frequency f0, the electrical length of the


matching section is 0/4, but at other frequencies the length is
different, so a perfect match is no longer obtained.
1079

Example
Matching with a quarter-wave transformer:
An 8 load is matched to a 50 line through a
20 transformer.

1080

540
Lecture notes Radio Engineering by Jorma Kekalainen

Quarter-wave transformer

k=2/

1081

/4-transformer on Smith chart


For example:

1082

541
Lecture notes Radio Engineering by Jorma Kekalainen

Waveguide theory

Reflection and
transmission of EM waves
at interfaces

Waves at an interface
Let us study a plane wave given by:

incident upon an interface between medium i and medium t

When the incident wave strikes the interface, it generates a


transmitted wave and a reflected wave.
We need to find the amplitudes of these reflected and transmitted
waves.
1084

542
Lecture notes Radio Engineering by Jorma Kekalainen

Incident, transmitted, and reflected


waves E-fields

Note: In the above equations, Er and Et are the unknowns that we need to find
in terms of the incident field amplitude Ei. 1085

Boundary conditions at interface E-


fields

Boundary conditions:
(1) At z = 0 the E-field parallel to the interface must be continuous
This gives:

1086

543
Lecture notes Radio Engineering by Jorma Kekalainen

Boundary conditions at interface H-


fields

(2) At z = 0 the H-field parallel to the interface must be continuous (no


surface currents)-

The wave impedance


of the medium =(/)

This gives: The other equation was:

1087

Reflection and transmission


coefficients

These are three different ways


of writing the same results using:

1) Impedances
2) Wavevectors
3) Refractive indices

1088

544
Lecture notes Radio Engineering by Jorma Kekalainen

Reflection at the perfect metal surface

For perfect metals:

Reflection and transmission coefficients for free-space and perfect metal


interface:

1089

Reflection at the perfect metal surface


standing waves

Standing waves

The interference between


incident and reflected waves
gives rise to standing waves

1090

545
Lecture notes Radio Engineering by Jorma Kekalainen

Reflection at the perfect metal surface


currents

The magnetic field is parallel to


the surface and peaks at the metal
surface.
There must be surface currents
(recall the boundary conditions
for the magnetic field). 1091

Reflection and transmission


coefficients for non-perfect metal

For non-perfect metals:

Reflection and
transmission coefficients
for non-perfect metal or
good conductor are
complex numbers. 1092

546
Lecture notes Radio Engineering by Jorma Kekalainen

Transmitted wave in non-perfect metal

The wave penetrates a few skin depths into the non-perfect metal.

Current flows within a layer a few skin


depths thick inside the non-perfect metal
or good conductor.

1093

Wave reflection and VSWR

E max Ei 1 E min Ei 1

VSWR is found by taking


Voltage Standing
VSWR
E max

1 the ratio of the magnitude
of the maximum field value
Wave Ration:
E min 1 (wherever that occurs) to
the minimum field value
(wherever that occurs). 1094

547
Lecture notes Radio Engineering by Jorma Kekalainen

Wave reflection and VSWR


No reflection (|| = 0)

VSWR
E max

1 1
E min 1
Small reflection (|| = 0.25) VSWR=1.66

Large reflection (|| = 0.75) VSWR=7

Complete reflection (|| = 1) VSWR=

1095

Waves in free space and in


transmission lines
Free space Transmission line

1096

548
Lecture notes Radio Engineering by Jorma Kekalainen

Comparison: Waves at transmission lines


and interfaces

1097

Three-layer structure
How do we calculate the reflection coefficient for the structure?

Note: We can use the same method solving the equivalent transmission line
problem!

1098

549
Lecture notes Radio Engineering by Jorma Kekalainen

Three-layer structure

1099

Three-layer structure

1100

550
Lecture notes Radio Engineering by Jorma Kekalainen

Dielectric Anti-Reflection (AR)


coatings

1101

Dielectric Anti-Reflection (AR) coatings

1102

551
Lecture notes Radio Engineering by Jorma Kekalainen

Dielectric Anti-Reflection (AR) coatings


Question: How do /4 long matching layers work?

1103

Dielectric High-Reflection (HR) coatings


What if we want to increase the reflectivity?
A periodic stack of high index-low index layers can be used as an HR
coating!

1104

552
Lecture notes Radio Engineering by Jorma Kekalainen

Dielectric High-Reflection (HR) coatings

1105

Dielectric High-Reflection (HR)


coatings

1106

553
Lecture notes Radio Engineering by Jorma Kekalainen

Dielectric High-Reflection (HR)


coatings

Magnitude of can be made arbitrarily close to unity by:

1) Choosing a large number N of pairs of high index-low index layers.


2) Choosing the difference between 1 and 2 to be large.
1107

Waveguide theory

Guided waves in parallel


plate metal waveguides

554
Lecture notes Radio Engineering by Jorma Kekalainen

Parallel plate metal waveguide


Let us study a parallel plate waveguide (shown below).
We have studied such structures in the context of transmission lines.
Thus, we know that they can guide TEM waves (Transverse Electric
and Magnetic). in which both the electric and magnetic fields point in
direction perpendicular to the propagation direction.
But these structures can guide more than just the TEM waves that
we have considered so far.

1109

Complex wave equations


Consider a parallel plate waveguide:

1110

555
Lecture notes Radio Engineering by Jorma Kekalainen

TE (guided) modes

1111

TE modes

1112

556
Lecture notes Radio Engineering by Jorma Kekalainen

Dispersion relation for TE mode

1113

Cut-off frequency for TE mode

1114

557
Lecture notes Radio Engineering by Jorma Kekalainen

TE guided modes Magnetic field

1115

TE guided modes
A parallel plate waveguide:

1116

558
Lecture notes Radio Engineering by Jorma Kekalainen

TM guided modes

1117

TM guided modes


1118

559
Lecture notes Radio Engineering by Jorma Kekalainen

TM guided modes Electric field

1119

TM guided modes Dispersion


relation

1120

560
Lecture notes Radio Engineering by Jorma Kekalainen

TM guided modes Cut-off frequency

1121

TM guided modes - Summary

Note: The TM 0 mode is just the TEM mode that we were dealing with
1122
transmission lines.

561
Lecture notes Radio Engineering by Jorma Kekalainen

Waveguide theory

Rectangular metal
waveguides

Rectangular metal waveguides


Rectangular metal waveguides are usually made of
copper and the best are gold plated.
They are commonly used to guide electromagnetic
power when dealing with high power levels (radars,
satellite and space communications, wireless/mobile
base stations, etc.)
Integrated versions are used in sub-millimeter
wavelength ultrahigh speed electronics (e.g. Gunn
oscillators, superconducting THz electronics, nonlinear
Schottky diode mixers) operating at frequencies between
300 GHz to 1000 GHz.

1124

562
Lecture notes Radio Engineering by Jorma Kekalainen

Rectangular metal waveguide


Like a parallel plate metal waveguide that is closed by metal walls on
the remaining two sides.

1125

TE guided modes

1126

563
Lecture notes Radio Engineering by Jorma Kekalainen

TE guided modes

1127

TE guided modes

1128

564
Lecture notes Radio Engineering by Jorma Kekalainen

TEmn modes

Note: The TE00 mode does not exist (i.e. it corresponds to field being
trivially zero everywhere) 1129

Cut-off frequencies of TEmn modes

Note: If the frequency is less than the cut-off frequency then kz


becomes entirely imaginary and the mode does not propagate but decays
1130
exponentially with distance.

565
Lecture notes Radio Engineering by Jorma Kekalainen

Dispersion relations of TEmn modes

1131

TM guided modes

1132

566
Lecture notes Radio Engineering by Jorma Kekalainen

TM guided modes

1133

TM guided modes

1134

567
Lecture notes Radio Engineering by Jorma Kekalainen

TM guided modes

Note: The TM00 , TMm0 TM0n modes do not exist (i.e. they correspond to
field being trivially zero everywhere). 1135

Cut-off frequency of TMmn modes

Note: If the frequency is less than the cut-off frequency then kz


becomes entirely imaginary and the mode does not propagate but decays
exponentially with distance. 1136

568
Lecture notes Radio Engineering by Jorma Kekalainen

Dispersion relation of TM modes

1137

Waveguide to coaxial cable


adapters
Sometimes it is necessary to transfer power between a waveguide and
a coaxial cable.

Loop solution Pin solution

1138

569
Lecture notes Radio Engineering by Jorma Kekalainen

Waveguide theory

Dielectric waveguides

Dielectric waveguide
TE-wave undergoing total internal reflection:

1140

570
Lecture notes Radio Engineering by Jorma Kekalainen

Dielectric waveguide Total internal


reflection
One can have a guided wave that is bouncing between two
dielectric interfaces due to total internal reflection and moving
in the z-direction.

1141

Dielectric slab waveguide

1142

571
Lecture notes Radio Engineering by Jorma Kekalainen

Comparison between dielectric and metal


waveguides

1143

Optical communications: Optical


fibers

1144

572
Lecture notes Radio Engineering by Jorma Kekalainen

What is cut-off in optical fiber case ?

So what does cut-off mean? It means that the


wave is no longer being guided through total
internal reflection since i < c.
1145

Cut-off frequency

1146

573
Lecture notes Radio Engineering by Jorma Kekalainen

Radio wave propagation in


more or less free space
Propagation mechanisms

Introduction
The majority of applications of microwaves are related
to radar and communication systems.
Microwave communication systems handle a large
fraction of the worlds international and other long-haul
telephone, data, and television transmissions.
Most of the current wireless telecommunications
systems, such as terrestrial radio links, satellite
communications , wireless local area computer
networks (WLAN), cellular systems, and global
positioning satellite (GPS) systems, operate in the UHF
EHF range, and thus rely on microwave technology.

1148

574
Lecture notes Radio Engineering by Jorma Kekalainen

International organizations

The use of radio frequencies is controlled by


the international organizations:
ITU (International Telecommunications Union)
and ETSI (European Telecommunication
Standards Institute), for example
Allocation the limited resources of frequency spectrum
and geostationary orbits.
Formulation of rules and procedures to avoid mutual
disturbance of services.

1149

Frequency allocation

1150

575
Lecture notes Radio Engineering by Jorma Kekalainen

Maxwells equations
The initial understanding of radio wave propagation goes back to the
work of James Clerk Maxwell, who in 1864 formulated the theory
of electromagnetic propagation which predicted the existence of
radio waves.
Electromagnetic waves propagate through environments where they
are reflected, scattered, and diffracted by walls, terrain, buildings,
and other objects.
The ultimate details of this propagation can be obtained by solving
Maxwells equations with boundary conditions that express the
physical characteristics of these obstructing objects.
This requires the calculation of the Radar Cross Section (RCS) of
large and complex structures.
Since these calculations are difficult, and many times the necessary
parameters are not available, approximations have been developed
to characterize signal propagation without resorting to Maxwells
equations.
1151

Propagation model categories


The simplest model for signal propagation: free space path
loss.
A signal propagating between two points with no attenuation
(absorption) or reflection follows the free space propagation
law.
Ray tracing propagation models are used to approximate wave
propagation according to Maxwells equations, and are
accurate models when the number of multipath components is
small and the physical environment is known.
Ray tracing models depend heavily on the geometry and
dielectric properties of the region through which the signal
propagates.
Empirical models use parameters based on measurements for
both indoor and outdoor channels.
When the number of multipath components is large, or the
geometry and dielectric properties of the propagation
environment are unknown, statistical models must be used. 1152

576
Lecture notes Radio Engineering by Jorma Kekalainen

Propagation models
The most common approximations use ray-tracing techniques.
These techniques approximate the propagation of electromagnetic waves
by representing the wavefronts as rays: the model determines the reflection
and refraction effects on the wavefront but ignores the more complex
scattering phenomenon predicted by Maxwells coupled differential
equations.
The simplest ray-tracing model is the two-ray model, which quite
accurately describes signal propagation when there is one direct path
between the transmitter and receiver and one reflected path.
The reflected path typically bounces off the ground, and the two-ray model
is a good approximation for propagation along highways, rural roads, and
over water.
More complex General Ray Tracing (GRT) models are needed to describe
additional reflected, scattered, or diffracted components
Many propagation environments are not accurately enough described by
ray tracing models.
In these cases it is common to develop analytical models based on
empirical measurements.
These are called empirical models.
1153

Statistical models
Often because of complexity and variability of the radio channel it
doesn't pay to develop a deterministic channel model.
For these cases statistical models are used.
The attenuation caused by signal path obstructions such as buildings
or other objects is typically characterized statistically.
Statistical models are also used to characterize the constructive and
destructive interference for a large number of multipath components.
Statistical models are most accurate in environments with fairly
regular geometries and uniform dielectric properties.
Indoor environments tend to be less regular than outdoor
environments, since the geometric and dielectric characteristics
change dramatically depending on whether the indoor environment
is an open factory, office, flat or metal machine factory.
For these environments computer-aided modeling tools are available
to predict signal propagation characteristics.

1154

577
Lecture notes Radio Engineering by Jorma Kekalainen

Microwave propagation
In free-space, electromagnetic waves propagate in straight
lines without absorption or other adverse effects.
Free-space, however, is an idealization that is only
approximated when microwave energy propagates through
the atmosphere or in the presence of the earth.
In practice the performance of a communication, radar, or
radiometry system may be seriously affected by propagation
effects such as reflection, refraction, attenuation,
diffraction, or scattering.
It is important to realize that more or less free-space
propagation effects cannot be quantified in any exact or
rigorous sense, but can only be described in terms of their
statistics.

1155

Radio propagation mechanisms


Propagation in free space is straightforward.
Received power is proportional to 1/d in free space.
(d = distance
Receiving power additionally influenced by between sender
fading (frequency dependent) and receiver)

shadowing
reflection at large obstacles
refraction depending on the density of a medium
scattering at small obstacles
diffraction at edges

fading shadowing reflection refraction scattering diffraction 1156

578
Lecture notes Radio Engineering by Jorma Kekalainen

Electromagnetic spectrum and


atmospheric opacity

1157

Low frequency

1158

579
Lecture notes Radio Engineering by Jorma Kekalainen

Microwave frequency bands

FSW = Free Space Wave, SW = Scattered Wave

1159

ELF and ULF


Extremely Low Frequency (ELF), f =30-300
Hz, =10000-1000km.
Ultra Low Frequency (ULF), f =0.3-3 kHz,
=1000-100km.
Waves do not penetrate ionosphere.
Antennas would be extremely large (yet << ).
Very low data rate.
Submarine communications, radio
communications to mines.

1160

580
Lecture notes Radio Engineering by Jorma Kekalainen

VLF

Very Low Frequency: f =3-30kHz, =100-10


km
Waves do not penetrate ionosphere guided
waves between the earth and the ionosphere.
Antenna height << .
Telegraph to sea vessels, time standards.

1161

LF
Low Frequency (LF), f =30-300 kHz, = 10-1
km.
Waves do not penetrate ionosphere guided
wave between the earth and the ionosphere.
Surface wave propagates beyond the horizon.
Antenna height <<.
Broadcasting service, radio communications to
boats, radio navigation (Decca, Loran).
Interference between the ionosphere wave and
the ground wave fading.

1162

581
Lecture notes Radio Engineering by Jorma Kekalainen

Propagation mechanism: Surface


(ground) wave

Surface wave or ground wave for LF and MF broadcasting (f < 0.5MHz)

1163

Ground wave propagation over a


curved surface
Sommerfelds theory or its approximations for the ground wave
propagation are valid for a flat surface and thus, can be used if the
following conditions for the distance and the antenna height are
valid

a is the radius of the earth


Example: Let f = 300 MHz dmax = 34 km and hmax =180 m.

If the flat ground assumptions are violated, the curvature of the


Earth should be taken into account in the modeling.
The higher the frequency the higher the losses.
Thus, the best frequency regime for the ground wave propagation is
VLF-MF bands (3 kHz-3 MHz).

1164

582
Lecture notes Radio Engineering by Jorma Kekalainen

Note: Conductor vs. insulator


Radio wave see conducting, non-magnetic
medium (r=1)

as a good conductor if

as a good insulator if

1165

Example
Dry (bad) ground: r 4, 0.001 S/m
good insulator if f >> 4.5 MHz (VHF and higher
frequencies)
good conductor if f << 4.5 MHz (LF and lower
frequencies)
Sea water: r 80, 5 S/m
good insulator if f >> 1.1 GHz (EHF and higher
frequencies)
good conductor if f << 1.1 GHz (HF and lower
frequencies)

Note: 0.001 S/m=1 mmho/m 1166

583
Lecture notes Radio Engineering by Jorma Kekalainen

Note: Skin depth vs. attenuation


coefficient
Skin depth () is a distance s0= which the wave [y(s)] decreases
to 1/e of its initial (surface) value [y(0)]

y(s=)=(1/e)y(0)0.37y(0)

where y(s)=y(0)exp(-s), the attenuation coefficient (attenuation


per unit length)
If set s= then

y(s=)= y(0)exp(-)= y(0)exp(-1) =1

The inverse of the skin depth is =1/ is the attenuation


coefficient.

1167

MF
Medium Frequency (MF), f = 0.3-3 MHz, = 1-
0.1 km.
Ionosphere waves attenuate in daytime (D-layer
absorption).
Ionospheric refraction at night from E- and F-
layers (the range of ionospheric wave at night
1000 km)
Surface wave propagates beyond the horizon but
attenuates (range 100km).
Antenna heights < .

1168

584
Lecture notes Radio Engineering by Jorma Kekalainen

HF

High Frequency (HF), f =3-30MHz, =100-


10m
Waves can penetrate the ionosphere, E- and F-
layers can cause refraction.
Antennas: monopoles, dipoles and antenna
arrays.
Very long communications can be achieved.

1169

Propagation mechanism: Ionosphere


(sky) wave
Medium (MF) and short wave (HF) communication over long distances

1170

585
Lecture notes Radio Engineering by Jorma Kekalainen

VHF
Very High Frequency (VHF), f = 30-300 MHz = 10-1
m.
Usually no refraction of waves from the ionosphere.
Direct wave is the main propagation mechanism but
scattering from troposphere and ionosphere, and
troposphere ducting (waveguide type propagation) is
possible.
Hills attenuates waves, reflections from surfaces,
multipath propagation.
Helix and yagi antennas usable.
100 km range for FM broadcasting service.
Radio beacons, mobile communication in sea, air and land
applications.

1171

Ultra High Frequency (UHF)


f =0.3 -3 GHz =100-10 cm.
Scattering from troposphere can provide even 300-600 km
communication links.
Direct wave is the main propagation mechanism and tropospheric
ducting (waveguide type propagation) is possible.
Hills shadow propagating waves.
Antennas are usually Yagis and parabolic reflectors (dishes).
TV broadcasts, radio navigation (GPS), air surveillance radar,
mobile communications etc.
Line-of sight communications (LOS)
Rain attenuation is negligible.

1172

586
Lecture notes Radio Engineering by Jorma Kekalainen

Super High Frequency (SHF)


f = 3-30 GHz = 10-1 cm.
Direct wave
Hydrometeors (rain and fog) attenuate waves
(>10GHz)
Horn antennas and parabolic reflectors.
Radio links, satellite communications, radars,
remote sensing, mobile communications,
WLAN, WPAN, etc.
Line - of sight communications

1173

Extremely High Frequency (EHF)

f =30-300GHz =10-1mm.
Direct wave limited by strong attenuation to
gases, rain and mist.
Atmosphere gases absorb waves (windows).
Small parabolic reflectors, lenses.
Short line-of sight communication links,
remote sensing from satellite
Scientific and experimental use.

1174

587
Lecture notes Radio Engineering by Jorma Kekalainen

Atmospheric attenuation

1175

Propagation mechanisms
Free space and reflected wave

TV and radio broadcast, navigation, wireless communication


etc. f > 30 MHz

1176

588
Lecture notes Radio Engineering by Jorma Kekalainen

Troposphere scattering
Communication over horizon.
Using UHF and higher frequencies.
Military applications.

1177

Radio propagation mechanisms

1178

589
Lecture notes Radio Engineering by Jorma Kekalainen

Line of sight (LOS) transmission link

1179

Friis formula for a radio link


Assume two antennas whose mutual distance is r and gains
along the line of sight G1 and G2
Antenna 1 transmits power P1. What is the received power, P2?
The power density, S1, at the antenna 2:

Antenna 2 receives the power:

1180

590
Lecture notes Radio Engineering by Jorma Kekalainen

Friis formula for a radio link


Since power density in terms of the electric field
is
The wave impedance (a
vacuum value is 120377)

Friis formula can be given in terms of radiated


field by the transmitter

Here it is assumed that the antennas are in free


space and in the far field.

1181

Note: Far-field concept


The near and far field are defined in terms of the Fraunhofer distance.
The Fraunhofer distance is the value of:
d f D

D 2

where D is the largest physical dimension of the radiator (or the diameter of
the antenna) and is the wavelength of the radio wave.
This distance provides the limit between the Fresnel and Fraunhofer region.
Additionally, a far-field region distance must satisfy the condition

where df is the far-field distance.


The far-field distance is the distance
from the transmitting antenna to the
beginning of the Fraunhofer region, or far field.

1182

591
Lecture notes Radio Engineering by Jorma Kekalainen

Far-field Radiation zone Free


space
In the far-field, the shape of the antenna pattern is
independent of distance from the source.
The far-field is also frequently referred to as the
"radiation zone", or "free-space".
A more precise definition is given by the propagation
properties.
The radiation zone is important because far-field falls off in
amplitude by 1/r (E1/r, r>>).
This means that in the far-field the total energy per unit
area at a distance r is proportional to 1/r2.
The area of the sphere is proportional to r2, so the total
energy passing through the sphere is constant.

1183

Environment factor
The effect of real conditions for the propagated
field can be modeled by environment factor,
F, which is the ratio of free-space and true
fields, E1 and E1true, respectively:
E1true =FE1
Hence, free-space Friis formula can be
generalized to real conditions with the
environment factor

1184

592
Lecture notes Radio Engineering by Jorma Kekalainen

Design inequality
Friis formula leads to a design inequality for a radio link

where P2,min is the minimum value for the received power.


Minimize the right hand side of the inequality by choosing first the
frequency f=c/ F and P2,min which is restricted by noise.
Antenna gains, G1 and G2, and transmitted power P1 state antenna
technology and costs.
The environment factor, F, depends heavily on frequency and
propagation mechanism.
The exact value is indeterminable but its average and statistics are
interesting average field and its variation.

1185

Propagation in more or less free-


space
Propagation in free space:
An antenna radiation power P1 gives rise to electric field

Propagation near ground:

1186

593
Lecture notes Radio Engineering by Jorma Kekalainen

Ray theory
In the far field the Poynting vector has only a radial
component then propagation can be modeled as flux
tubes in which power is conserved
Thin flux tubes are rays that are
straight lines in a homogeneous medium
curved lines in a nonhomogeneous medium

Note: The Poynting vector S=ExH represents the energy flux (in W/m2) of an electromagnetic
1187
field.

Ray theory
The fundamental
assumption of the ray
theory is that obstacles
shadow only those rays
that are impinging on it
without affecting the
others.
The validity of the ray
theory is the better the
greater is the frequency.

1188

594
Lecture notes Radio Engineering by Jorma Kekalainen

Atmospheric propagation
While the electromagnetic properties of air are
very close to those of a vacuum (r=1, r =1), they
are not rigorously identical.
When considering a long trajectory (path) within
the atmosphere, the variations due to the pressure
p, the temperature T, and the moisture level
have to be considered.
Electromagnetic rays do not travel in a straight
line within the atmosphere, but follow slightly
curved trajectories.

1189

Inhomogeneous atmosphere
The troposphere is very slightly inhomogeneous
caused by the gravitation which makes the gas
mixture called air thinner as height increases.
Free electrons in the ionospheric layers change the
effective permittivity as the function of height.
Inhomogeneous media bend radio waves or radio
rays.

1190

595
Lecture notes Radio Engineering by Jorma Kekalainen

Atmosphere

1191

Standard atmosphere
The relative permittivity of air at microwaves is given by an
experimental expression:

where p is the barometric pressure in millibar, T the


temperature in Kelvin, the partial pressure of water vapor in
millibar ( or hPa) and n the refractive index.
A large number of measurements led to the establishment of an
average profile for r as a function of height h, called the
standard atmosphere.
Note: Waves in the visible spectrum are not affected by moisture: their refractive index is given
by the equation, letting =0. 1192

596
Lecture notes Radio Engineering by Jorma Kekalainen

Atmospheric effects
The relative permittivity of the atmosphere is
close to unity.
It generally decreases (approaches unity as
altitude increases, since pressure and humidity
decrease with height faster than does temperature.
This change in permittivity with altitude causes
radio waves to bend toward the earth.
Such refraction of radio waves can sometimes be
useful, since it may extend the range of radar and
communication systems beyond the limit imposed
by the presence of the earths horizon.
Note: In a radar system, refraction effects can lead to errors when determining the elevation of
1193
a target close to the horizon.

Refractivity of the standard


troposphere
The refraction index of the troposphere
depends exponentially of the height:

N is the modified refraction index, refractivity


Ns = 315 in the refractivity at ground level under
normal atmospheric conditions (temperature,
atmospheric pressure, the partial pressure of water
vapor)
hs=7.35km

1194

597
Lecture notes Radio Engineering by Jorma Kekalainen

Refractivity
At low altitudes up to h = 1 km the linearized model is valid
for the refractivity

Since the velocity of a radio wave is inversely proportional to


the refractive index (n) the wavefronts tend to bend towards
increasing direction of n.

Note: The refractivity function is spherical, but a planar model can be used as an approximation
1195
for low altitudes.

Standard atmosphere

1196

598
Lecture notes Radio Engineering by Jorma Kekalainen

Radio wave propagation through


inhomogeneous medium
Consider a radio ray traveling in a planar medium whose refraction index
depends only height coordinate n = n(h) (a model for the troposphere)
Let us study the system of two homogeneous layers, having respectively
the refraction indices n, n+dn

dn

dh
d

Note: The equation for a bending radio ray can be derived assuming that the refraction index
1197
changes stepwise over layer boundaries.

Curvature radius of the radio ray


Snells law of refraction:
n (h) sin [ (h)] = constant

From this the radius of curvature of the ray R (h) can


be solved:
1198

599
Lecture notes Radio Engineering by Jorma Kekalainen

Bending direction

Radio ray bends towards electrically denser


medium, i.e., towards the direction of higher
refraction index n, , i.e., where n is higher.

1199

Linear approximation for n(h)


Consider a linearly changing n(h) for which

constant, if n1 and /2

This means circular ray path in the troposphere for


small elevation angles!
Note: Here is assumed that radio ray propagates nearly as a direction of the Earths surface. So
we can assume n 1 and approximately zero elevation angle ( 0), which means /2. 1200

600
Lecture notes Radio Engineering by Jorma Kekalainen

Alternative trajectories

1201

Effective radius of the earth


If a radio wave propagates in the troposphere nearly
parallel to the surface of the ground i.e. /2,
the radius of curvature of the path is then approximately

the bending path is almost circular in standard


tropospheric conditions which means


where the real Earth radius a = 6370 km.

1202

601
Lecture notes Radio Engineering by Jorma Kekalainen

Affine transformation
The analysis of radio wave propagation is difficult
since both the wave and the surface are bending.
Using an affine transformation one of the paths
can be made straight.
After an affine transformation, only one curved line is
left.
This transformation must be done so that the
distance between the lines do not change.
So, the height of radio ray from the Earths
surface is the same as before the transformation.

1203

Affine transformation
Assume the path height h(d)<< d << a or R

Cosine law:

Omitting the small terms:

1204

602
Lecture notes Radio Engineering by Jorma Kekalainen

Effective radius
The height of the radio path h(d),

remains the same if values a ae and R Re such that

1205

Effective radius
Now we can straighten the radio path by setting Re=


In standard tropospheric conditions R 4a corresponding the
effective radius of the Earth ae = Ka = (4/3)a = 8500km
This transformation replaces the refractive index of the
troposphere and its derivate to ne and dne/dh=0
How does the refractive index change in the affine
transformation?

1206

603
Lecture notes Radio Engineering by Jorma Kekalainen

Example
If an antenna is at a height h, above the earth, simple
geometry gives the line-of- sight distance to the horizon
as
d 2ah
where a is the radius of the earth.
From the previous study we see that the effect of
refraction on range can be accounted for by using an
effective earth radius ae=Ka where k> 1.
A value commonly used is K = 4/3, but this is only an
average value.
Actually, bending depends on weather conditions.

1207

Geometric and radio horizon


If an antenna is at a height h, above the earth, simple geometry
gives the line-of- sight distance to the horizon.
If we assume idealized free-space type straight line propagation, we
get the distance d to the geometric horizon

d a h2 a 2 a 2 2 ah h 2 a 2
2ah h 2 2ah , when h a

In standard troposphere, of course, the refraction bends the


radio wave.
So, the distance dst to horizon in the standard troposphere is
4
d st 2 ae h 2 Kah d K d 1.15d
3
1208

604
Lecture notes Radio Engineering by Jorma Kekalainen

Radio and optical horizon (smooth


earth) in the standard atmosphere

1209

Note
Prerequisites to exploit effective Earth radius
dn/dh = constant ( n(h)=constant*h+Cint)
propagation almost parallel to the surface (radio
wave path is circular)
For microwave links, the elevation angle is very small.
Operation area is small enough (radius < 100 km
from transmitter)

1210

605
Lecture notes Radio Engineering by Jorma Kekalainen

Profile chart
Antenna heights and terrain heights representations are defined
with respect to the ground surface (sea level)
the design of radio mast height is difficult

The height function is a parabola!

1211

Profile chart
Knife edge diffraction causes attenuation (shadowing
effect)! Increase antenna height (h2) to avoid
shadowing.

1212

606
Lecture notes Radio Engineering by Jorma Kekalainen

Profile chart usage


The middle point of the radio link
is set to the center.
The highest points from the
terrain profile between Rx and Tx
antennas should be marked to the
chart.
The radius of the Fresnel zone
should be calculated at the highest
points.
The line-of-sight path is drawn to
the chart such that the highest
points are below the Fresnel
ellipsoid.
The needed radio mast heights are
read from the chart.

1213

Note
Radio ray propagation in the troposphere is
relatively easy to analyze because its path is
practically circular besides in very exceptional
temperature conditions and in the case of too
steep elevation angle.
The spherical inhomogeneity of troposphere can be
approximated by a planar linear model for the
refractive index.
In the ionosphere the ray path must be analyzed
with more rigorous method using nonlinear
spherical permittivity model.

1214

607
Lecture notes Radio Engineering by Jorma Kekalainen

Real atmosphere
The standard atmospheric profile is only a
statistical value, representing an average taken over a
large sample of experimental data.
In fact, the term K depends upon the latitude, varying
between 1.2 and 1.5.
When the atmosphere fluctuates, n may even locally
increase with altitude: waves are then trapped within a
channel (anomalous refraction, mirages).
Shadow regions are located next to such channels,
where it is apparently impossible to emit or to receive.
Such anomalies often take place over sea shore.

1215

Standard atmosphere

1216

608
Lecture notes Radio Engineering by Jorma Kekalainen

Ducting in the troposphere


Weather conditions can produce a temperature
inversion, where the temperature increases with
altitude.
This condition can sometimes lead to ducting (also
called trapping, or anomalous propagation), where a
radio wave can propagate long distances parallel to the
earths surface, via the duct created by the layer of air
along the temperature inversion.
The situation is very similar to propagation in a
dielectric waveguide.
Such ducts can range in height from 15 150 m, and
may be near the earths surface, or higher in altitude.

1217

Tropospheric channel
Temperature inversion causes channels, or ducts, of cool air to
form between layers of warm air, which can cause radio waves
to travel far beyond the normal line-of-