Y

Lectures

NO C P
on

CO Y
O
Satellite Communications

OP
PY
Systems
D O NO T C
Assoc. Professor PhD Stefan Zhelev
Shumen University
Faculty of Technical Sciences
DO NO
T
Com
man
d satellitte
Uplin
satellitte k
T
Tele
metr
Dow y
transport nlink
Tracking,
Telemetry,
DO

Command &
Monitoring

wire wire
fiber fiber
cable cable
users
video receiving
transmitting voice data graund
graund station
station

PR = EIRP + GR − LFS , dB
 Y
NO C P
CO Y
O
OP
PY
D O NO T C
DO NO
T
T
DO

ISBN 978-954-577-637-3
електронно издание
© Стефан Желев – автор
© Шуменски университет “Епископ Константин Преславски”
Шумен, 2012
 Contents

Lecture 1. Introduction to Satellite Communications

Y
1. A little History of Satellite Communications
2. Satellite Communications Segments
3. Satellite Orbits

NO C P
4. Frequency Band

CO Y
4.1. Regulatory Process
4.2. The electromagnetic frequency spectrum

O
5. Signal processing elements in satellite communications

OP
PY
Lecture 2. A Link Budget Calculation and Analysis
1. The quality of signal transmission
D O NO T C
2. Elements of the link
2.1. Effective Isotropic Radiated Power
2.2. Power flux density
2.3. Antenna Gain
2.4 Free-Space Path Loss
2.5. System Noise
2.6. Link Performance Parameters
DO NO

3. Link Budget
T
3.1. Frequency Translation Satellite
3.2. The on-board processing satellite

Lecture 3. Forward Error Correcting in Satellite Communications

T
1. The necessity of channel coding
2. Types of error control
3. Concept of Hamming Weight and Hamming Distance
DO

4. Reed Solomon code

4.1. Reed Solomon coding
4.2. Reed Solomon decoding
5. Interleaving
6. Convolution code
6.1. Coder and coding.
6.2. State representation and state diagram
6.3. The Viterbi Decoding Algorithm
6.4. Implementation of the Viterbi Decoder
6.5. Recursive Systematic Convolutional Encoder

Lecture 4. Multiple Access in Satellite Communications

1. Introduction of Multiple Access in Satellite Communications
2. Frequency Division Multiple Access
2.1. Frequency Division Multiple Access with PCM/TDM/PSK application
 2.2. Frequency Division Multiple Access with PCM/SCPC/PSK application
3. Time Division Multiple Access
3.1. Time Division Multiple Access with PCM/TDM/PSK application
3.2. TDMA Frame Efficiency

Y
3.3. TDMA capacity
3.4. Switching in satellite TDMA

NO C P
4. Code Division Multiple Access

CO Y
4.1. Direct Sequence Spread Spectrum
4.2. Frequency Hopping Spread Spectrum.

O
OP
Problems

PY
D O NO T C
DO NO
T
T
DO
 Satellite Communications Systems
Lecture 1.
Introduction to Satellite Communications

1. A little History of Satellite Communications

Y
2. Satellite Communications Segments
3. Satellite Orbits
4. Frequency Band

NO C P
CO Y
4.1. Regulatory Process
4.2. The electromagnetic frequency spectrum
5. Signal processing elements in satellite communications

O
OP
PY
1. A little History of Satellite Communications
D O NO T C
The unique feature of communications by satellites is there ability to provide
distance insensitive point to multipoint communications. These links can be between
fixed terminals, mobile terminals and both fixed and mobile terminals on lend, on the
air, and at sea as it is shown in Figure 1.

C om
mand
DO NO

Uplin satellitte
satellitte k
T
Tele
metr
D ow y
transport nlink
Tracking,
Telemetry,
Command &
T
Monitoring

wire wire
fiber fiber
cable cable
DO

users
video receiving
transmitting voice data graund
graund station
station
Figure 1. The idea of a communications satellite network

Communications by satellites offers a number of features that are not readily

available with alternative modes of transmission, such as terrestrial microwave, cable,
or fiber networks.
Some of the advantages of satellite communications are:
Diverse User Networks. The typical communications satellite can see large
areas of the earth, and link together many users simultaneously. Satellites are particu-
larly useful for accessing remote areas or communities not otherwise accessible by
terrestrial means.
High Capacity. Satellite communications links operate with high carrier fre-
quencies, with large information bandwidths. Capacities of typical communications

1
 satellites are very high (above 100Mbps ), and can provide services for several hundred
video channels or several tens of thousands of voice or data links.
Low Error Rates. Bit Error Rate (BER) is a measure of quality. Bit errors on
a digital satellite link tend to be random, allowing statistical detection and error cor-

Y
rection techniques to be used. Error rates can be routinely achieved efficiently and
reliably with standard equipment.
Distance Independent Costs. The cost of satellite transmission is independent

NO C P
of the distance between the transmitting and receiving earth stations.

CO Y
Fixed Broadcast Costs. The cost of satellite broadcast transmission, that is,
transmission from one transmit ground terminal to a number of receiving ground ter-

O
minals, is independent of the number of ground terminals receiving the transmission.

OP
PY
A little History of Satellite Communications
The first idea of communications by satellite was given by Arthur C. Clarke in
D O NO T C
his classic paper “Extraterrestrial Relays”, 1945 1 .
Satellite communications began in October 1957 with the launch by the former
USSR a small satellite called Sputnik 1 (4.10.1957), then 3.11.1957 was launched
Sputnik 2 with Laika.
The first communications by artificial satellite was accomplished by
SCORE (Signal Communicating by Orbiting Relay Equipment), launched by the Air
Force, USA into a low (160 by 1280 km) orbit in December 1958. SCORE relayed a
DO NO

recorded voice message, from one earth station to another. SCORE broadcast a mes-
T
sage from President Eisenhower to stations around the world. The maximum message
length was four minutes, and the relay operated on a 150MHz uplink and 108MHz
downlink. SCORE, powered by battery only, operated for 12 days before its battery
failed, and decayed out of orbit 22 days later 2 .
T
1960’s – First satellite communications:
• 1960 First passive communication satellite (Large balloons, ECHO satel-
DO

lites 1 and 2, launched by the National Aeronautics and Space Administration

(NASA) in August 1960 and January 1964, respectively).
• 1962: First active communication satellite (The TELSTAR Satellites 1 and
2, launched into low orbits by NASA for AT&T/Bell Telephone Laboratories in July
1962 and May 1963, were the first active wideband communications satellites).
• 1963,1964: First satellite into geostationary (GEO) orbit (Syncom 1 failed,
SYNCOM 2 and 3 were placed in orbit in July 1963 and July 1964, Syncom, with
7,4GHz uplink and 1,8GHz downlink frequencies, employed two 500 KHz channels for
two-way narrowband communications, and one 5MHz channel for one-way wideband
transmission).
• 1964: International Telecomm. Satellite Organization (INTELSAT) created.
• 1965 First successful communications GEO (Early Bird / INTELSAT 1).

1
A.C. Clarke, ‘Extraterrestrial Relays,’Wireless World, Vol. 51, pp. 305 308, October 1945.
2
M.I. Davis and G.N. Krassner, ‘SCORE First Communications Satellite,’ Journal of American Rocket Society,
Vol. 4, May 1959.

2
 1970’s – GEO Applications Development, DBS:
• 1972 First domestic satellite system operational (ANIKA launched in No-
vember 1972 by NASA for Telsat Canada, was the first domestic commercial com-
munications satellite. Two later ANIKAs were launched in April 1973 and May 1975.
The satellites, built by HughesAircraft Company, operated at C-band and had 12

Y
transponders, each 36MHz wide).
• 1975 First successful direct broadcast experiment (USA-India).

NO C P
• 1977 A plan for direct broadcast sattellites (DBS) assigned by the ITU

CO Y
• 1979 International Mobile Satellite Organization (Inmarsat) established.
1980’s – GEO Applications Expanded, Mobile:

O
• 1981 First reusable launch vehicle flight.

OP
• 1982 International maritime communications made operational.

PY
• 1984 First direct-to-home broadcast system operational (Japan).
• 1987 Successful trials of land-mobile communications (Inmarsat).
D O NO T C
• 1987 TVSAT: First DBS-satellite (Direct Broadcast Satellite, Television-
broadcasts directly to home)
• 1989-90 Global mobile communication service extended to land mobile and
aeronautical use (Inmarsat)
1990+’s NGSO applications development and GEO expansion
• 1990-95:
- Proposals of non-geostationary (NGSO) systems for mobile communications.
DO NO

- Continuing growth of VSATs (Very Small Aperture Terminal) around the

T
world.
- Spectrum allocation for non-GEO systems.
- Continuing growth of DBS. DirectTV created.
T
• 1997:
- Launch of first batch of LEO for hand-held terminals (Iridium).
- Voice-service portables and paging-service pocket size mobile terminals
DO

launched (Inmarsat).
• 1998-2000: Mobile LEO systems initiate service and fail afterwards (Irid-
ium, Globalstar).

2. Satellite Communications Segments

The satellite communications have two areas or segments: the space segment
and the ground (or earth) segment.
The space segment includes the satellite (or satellites) in orbit in the system,
and the ground station that provides the operational control of the satellite(s) in orbit.
The ground station is variously referred to as the Tracking, Telemetry, Command
(TT&C) or the Tracking, Telemetry, Command and Monitoring (TTC&M) station.
Two elements of the space segment of a communications satellite system are shown
on Figure 2 - the satellite and TTC&M station.
A communications satellite is an orbiting artificial earth satellite. It receives a
communications signal from a transmitting ground station, amplifies or processes it,
and transmits it back to the ground stations for reception.

3
 The TTC&M station provides essential spacecraft management and control
functions to keep the satellite operating safely in orbit. The TTC&M links between
the spacecraft and the ground are usually separate from the user communications
links and links may operate in the same frequency bands or in other bands. The space
segment equipment carried aboard the satellite has two subsystems: the bus and the

Y
payload.

NO C P
Co
mm

CO Y
Up and
link
Te
le
Do metr
wn y

O
link

OP
PY
Graund Station
Tracking,
D O NO T C
Telemetry,
Command &
Monitoring

Figure 2. The space segment for a communications satellite network

The bus has the subsystems that support the satellite: the physical structure,
power subsystem, attitude and orbital control subsystem, thermal control subsystem,
and command and telemetry subsystem.
The payload on a satellite is the equipment that provides the services, and con-
DO NO

sists of the communications equipment that provides the relay link between the up-
T
and downlinks from the ground. That equipment is called the transponder. The an-
tennas on the satellite and the transponder receives the uplink signal, amplifies (or
processes the signal), and then reformats and transmits the signal back to the ground
T
received station.
That portion of the link from the earth station to the satellite is called the up-
link, and another - from the satellite to the ground is called the downlink.
The ground segment terminals consist of three basic types:
DO

- fixed (in-place) terminals;

- transportable terminals;
- mobile terminals.
The TTC&M ground station are not included in the ground segment.
Fixed terminals may be providing different types of services, but they are de-
signed to access the satellite while fixed in-place on the ground. Small terminals used
in private networks (VSATs) are examples of fixed terminals, or terminals mounted
on residence buildings used to receive broadcast satellite signals.
Transportable terminals are designed to be movable. Examples of the trans-
portable terminal are satellite news gathering (SGN) trucks, which move to locations,
stop in place, and then deploy an antenna to establish links to the satellite.
Mobile terminals are designed to communicate with the satellite while in mo-
tion. They are further defined as land mobile, aeronautical mobile, or maritime mo-
bile, depending on their locations.

4
 3. Satellite Orbits
There are four most commonly used orbits in satellite communications.
Geostationary (Geosynchronous) Earth Orbit (GEO)
The satellite remains fixed (or approximately fixed) over one point on the
equator. The parameters for the evaluation of the GEO link are:

Y
Range (distance) from the earth station (ES) to the satellite, in km
Geostationary Radius: rS = 42164,17km ;

NO C P
Geostationary Height (Altitude): hGSO = rs − re = 35786,43km ;

CO Y
Equatorial Radius: re = 6378,14km ;
Advantages of Geostationary Earth Orbit:

O
- The period of revolution for the geostationary orbit is 23 h 56 min 4,091 s, so

OP
the visibility of the satellite is 24 h;

PY
- A satellite in GSO sees about one-third of the earth’s surface, so three GEO
D O NO T C
satellites, placed 1200 apart in the equatorial plane, could provide global coverage;
- An antennas on the ground, once aimed at the satellite, need not continue to
rotate.
Disadvantages of the GEO are:
- the gravity of the sun and moon disturb the orbit;
- the geostationary orbit’s finite capacity and satellites using the same frequen-
cies must be separated to prevent mutual interference;
- providing coverage of high latitudes (above 800 ) is generally not possible, so
DO NO

the polar aeries cannot be achieved with a geostationary constellation.

T
Low Earth Orbit (LEO)
LEO satellites move faster than the rotation of earth. The typically altitudes are
T
from 160 to 2500 km. The satellite is not at a fixed location in the sky and the visibil-
ity of the satellite is 5 ÷ 15 min , so it needed many satellites to see the earth’s surface.
Some current LEO satellite networks operate with 12, 24, or 66 satellites to achieve
the desired coverage.
DO

Applications of LEO satellites are: Communications; Military surveillance;

Weather; Atmospheric studies; Earth observation.

Medium Earth Orbit (MEO)

Satellites at Medium Earth Orbit (MEO) operate in the range between LEO and
GSO, typically at altitudes of 10 000 to 15 000 km.
Applications: used for meteorological, remote sensing and position location
applications (GPS).

High Elliptic Earth Orbit (HEO)

Satellites operating in high elliptical (high eccentricity) orbits (HEO) are used
to provide coverage to high latitude areas not reachable by GSO, and those that re-
quire longer contact periods than available with LEO satellites. Russian use of Mol-
nya and Tundra elliptical orbits for satellite television to the high northern latitude

5
 Russian states is well known. These orbits have a perigee altitude of about 1000 km,
and an apogee altitude of nearly 40 000 km.
Satellite orbits that are not synchronous, such as the LEO, MEO, or HEO, are
often referred to as non-geosynchronous orbit (NGSO) satellites.
Orbital parameters are:

Y
- Apogee – the point farthest from earth;
- Perigee – the point of closest approach to earth;
- Line of Apsides – the line joining the perigee and apogee through the center

NO C P
CO Y
of the earth;
- Ascending Node – the point where the orbit crosses the equatorial plane, go-
ing from south to north;

O
OP
- Descending Node – the point where the orbit crosses the equatorial plane, go-

PY
ing from north to south;
- Line of Nodes – the line joining the ascending and descending nodes through
D O NO T C
the center of the earth;
- Argument of Perigee – the angle from ascending node to perigee, measured in
the orbital plane;
- Right Ascension of the Ascending Node, – the angle measured eastward, in
the equatorial plane, from the line to the first point of Aries (Y) to the ascending node.
Table 1. The characteristics of satellite orbits
LEO МЕО GEO НЕО
DO NO

Altitude, km 500÷3 000 10 000÷14 000 35 786 500÷50 000

T
23 h 56 min. 4,091s
Period, h 1÷3 6÷8 3÷24
one sidereal day
Delay, ms 6÷30 70÷120 240÷280 50÷320
Visibility 5÷15 min. some hours 24 h 2÷12 h
Tracking, Telemetry,
T
complexity less complexity simply complexity
Command & Monitoring
Outlay for launch a satel-
low high high high
lite into orbit
Numbers of satellites to
DO

66 <24 3
cover the earth
- telephony;
- Communications;
- Broadcasting;
- Military surveillance;
- Point to multi-point
Applications - Weather; GPS
communications;
- Atmospheric studies;
- Mobile services;
- Earth observation.
- Weather observation.

The parameters for the evaluation of the GEO link are:

- Range (d-distance) from the earth station (ES) to the satellite, in km;
- Azimuth angle ( Θ ) from the ES to the satellite, in degrees. The azimuth angle
is the angle measured in the horizontal plane of the location between the direction of
geographic north and the intersection of the plane containing the point considered, the
satellite and the centre of the earth.
- Elevation angle ( ϕ ) from the ES to the satellite, in degrees. The elevation an-
gle is the angle between the horizon at the point considered and the satellite, meas-

6
 ured in the plane containing the point considered, the satellite and the centre of the
earth.
Additional parameters are:
Equatorial Radius: re = 6378,14km ; Eccentricity of the earth: ee = 0,08182 ; Differ-

Y
ential longitude, B = le − ls - the difference between the earth station and satellite longi-
tudes; le earth station longitude, in degrees; ls satellite longitude, in degrees; LE earth
station latitude, in degrees; LS satellite latitude, in degrees; H earth station altitude

NO C P
above sea level, in km.

CO Y
Longitudes east of the Greenwich Meridian and latitudes north of the Equator
are positive; longitudes west of the Greenwich Meridian and latitudes south of the

O
Equator are negative.

OP
PY
The distance between earth station and satellite can be calculated as 3 :
d = R 2 + rs − 2 R.rs . cosψ E . cos B , km
2
(1)
D O NO T C
⎛ ⎞
l =⎜ + H ⎟. cos( LE ) ;
re
where R = l 2 + z 2 ;
⎜ 1 − e 2 . sin 2 ( L ) ⎟
⎝ e E ⎠
⎛ re (1 − ee2 ) ⎞ ⎛z⎞
z =⎜ + H ⎟. sin( LE ) ; ψ E = tg −1 ⎜ ⎟ .
⎜ 1 − e 2 sin 2 ( L ) ⎟ ⎝l⎠
⎝ e E ⎠
Elevation angle to the satellite:
⎛ re + h
DO NO

⎞
ϕ = cos −1 ⎜ 1 − cos 2 ( B). cos 2 ( LE ) ⎟ . (2)
T
⎝ d ⎠
Azimuth angle to the satellite:
The azimuth angle is determined from the intermediate angle A from one of
four possible conditions, based on the relative location of the earth station and the
T
subsatellite point on the earth’s surface. The condition is determined by standing at
the earth station (ES) and looking in the direction of the subsatellite point (SS) 4 (Fig-
ure 3).
DO

⎛ sin( B ) ⎞
Α = sin −1 ⎜⎜ ⎟,
⎟ (3)
⎝ sin β ⎠
where β = cos −1[cos( B). cos( LE )] , B = le − ls .
N N N N

Θ Θ
ES ES
SS SS SS SS

Θ Θ
ES ES
a) b) c) d)

Figure 3. Determination of azimuth angle to the satellite

3
Ippolito, Louis J., Satellite communications systems engineering: atmospheric effects, satellite link design, and system
Performance, 2008 JohnWiley & Sons Ltd.
4
Ippolito, Louis J., Satellite communications systems engineering: atmospheric effects, satellite link design, and system
Performance, 2008 JohnWiley & Sons Ltd.

7
 The resulting equation to determine the azimuth angle Θ for each of the four
conditions is given in Table 2.

Table 2. Determination of azimuth angle from intermediate angle A

Condition SS point of ES Θ= Figure

Y
northeast Α 1а
northwest 3600 − Α 1b
southeast 1800 − Α 1c

NO C P
southwest 1800 + Α 1d

CO Y
More then four possible conditions, there are two additional cases:

O
- If the earth station is located at the same longitude as the subsatellite point,

OP
PY
the azimuth angle will be 1800 if the earth station is in the northern hemisphere and 0◦
if the earth station is in the southern hemisphere.
D O NO T C
- If the earth station is located on the equator, the azimuth angle will be 90◦ if
the earth station is west of the subsatellite point and 2700 if the earth station is east of
the subsatellite point.

Example 1. The earth station is located at Coventry, England.

Latitude: LE =52°25'=52.420, N=+52.42; Longitude: 010 28'W = 1.47W , le = −1.47 ;
Altitude: H=0km.
Satellite: Longitude: 19.20 E, lS = +19,2 .
DO NO

Find the rang, azimuth angle ( Θ ) and elevation angle ( ϕ ).

T
1) Determine the differential longitude
B = le − ls = −1.47 − 19.2 = −20,67 ;
2) Determine the earth radius at the earth station, R, for the calculation of the range
T
⎛ ⎞ ⎛ 6378,14 ⎞
l= ⎜ re ⎟
+ H . cos(LE ) = ⎜ + 0 ⎟. cos(52,42) = 3897,8km
⎜ 1 − e .sin ( L ) ⎟ ⎜ ⎟
⎝ 1 − 0,08182 .sin (52,42)
2 2 2 2
⎝ e E ⎠ ⎠
DO

⎛ re (1 − ee2 ) ⎞ ⎛ 6378,14.0,9933 ⎞
z =⎜ + H ⎟. sin( LE ) = ⎜ + 0 ⎟.0,7925 = 5030,87 km
⎜ 1 − e 2 . sin 2 ( L ) ⎟ ⎝ 0,998 ⎠
⎝ e E ⎠
⎛z⎞
ψ E = tg −1 ⎜ ⎟ = 52,230 ; R = l 2 + z 2 = 6364,15 km
⎝l⎠
3) Determine the range
d = R 2 + rs − 2 R.rs . cosψ E . cos B =
2

= 6367,84 2 + 42164,17 2 − 2.6367,84. 42164,17. 0,6125. 0,9356 = 38866,9km

4) Determine the elevation angle
⎛r +h ⎞
ϕ = cos −1 ⎜ e 1 − cos 2 ( B). cos 2 ( LE ) ⎟ = 27,010
⎝ d ⎠
5) Determine the azimuth angle
β = cos −1 [cos( B). cos( LE )] = cos −1 (0,9356.0,6099) = 55,210
⎛ sin( B ) ⎞ ⎛ 0,353 ⎞
Α = sin −1 ⎜⎜ ⎟⎟ = sin −1 ⎜ ⎟ = 25,47 0
⎝ sin β ⎠ ⎝ 0,821 ⎠

8
 Since the subsatellite point SS is southeast of the earth station ES according Table 2:
Θ = 1800 − A = 154,530 .

4. Frequency Band
4.1. Regulatory Process

Y
The satellite communications system parameters that are under the regulatory
include:
- choice of radiating frequency;

NO C P
CO Y
- maximum allowable radiated power;
- orbit locations (slots) for GSO.
The allocation and regulation of the frequency spectrum is colled spectrum (or

O
OP
frequency) management.

PY
The Europe countries have active organizations involved with spectrum man-
agement. They are responsible for the development of satellite systems or the provi-
D O NO T C
sion of satellite based services. Besides national organizations there is international
management by the International Telecommunications Union (ITU), with headquar-
ters in Geneva, Switzerland. The ITU was formed in 1932 from the International
Telegraph Union, created in 1865. It is a United Nations Specialized Agency, cur-
rently with over 190 members.
The ITU has three primary functions:
- allocations and use of the radio-frequency spectrum;
DO NO

- telecommunications standardization;
T
- development and expansion of worldwide telecommunications.
According to these three functions the ITU have three sectors:
- the Radiocommunications Sector (ITU-R), responsible for frequency alloca-
tions and use of the radio-frequency spectrum;
T
- the Telecommunications Standards Sector (ITU-T), responsible for telecom-
munications standards; and
- the Telecommunications Development Sector (ITU-D), responsible for the
DO

development and expansion of worldwide telecommunications.

The specific frequency bands and other regulatory factors for a particular satel-
lite system have two performances:
- service(s) to be provided by the satellite system/network; and
- location(s) of the satellite system/network ground terminals.
They both together determine the frequency band, or bands, where the satellite
system may operate.

4.2. The electromagnetic frequency spectrum

The wavelength of the free space path signal is the principal parameter that de-
termines the interaction effects of the atmosphere, and the resulting link path degra-
dations.
Communications systems employ the electromagnetic frequency spectrum
shown in Table 3 and Figure 3 (according to International Telecommunication Union
ITU).

9
 Table 3. Electromagnetic frequency spectrum
Range Frequency Length of the wave
Extremely low frequency ELF 3 ÷ 30 Hz 100000 ÷10000 km
Super low frequency SLF 30 ÷ 300 Hz 10000 ÷1000 km
Ultra low frequency ULF 300 ÷ 3000 Hz 1000 ÷100 km

Y
Very Low frequency VLF 3 ÷ 30 kHz 100 ÷10 km
Low frequency LF 30 kHz ÷ 300 kHz 10 km ÷1 km
Medium frequency MF 300 kHz ÷ 3 MHz 1 km ÷100 m
High frequency HF 3 MHz ÷ 30 MHz 100 m ÷10 m

NO C P
Very high frequency VHF 30 MHz ÷ 300 MHz 10 m ÷ 1 m

CO Y
Ultra high frequency UHF 300 MHz ÷ 3 GHz 1 m ÷ 10 cm
Super high frequency SHF 3 GHz ÷ 30 GHz 10 cm ÷ 1 cm

O
Extremely high frequency EHF 30 GHz ÷ 300 GHz 1 cm ÷ 1 mm

OP
Terahertz THz 300GHz ÷3000GHz 1 mm ÷ 0.1 mm

PY
Infrared IR 3000GHz ÷ 400THz 0,1mm÷0,75μm
Visible 400 THz ÷ 750 THz 0,75μm÷0,4μm
Ultraviolet UV 750 THz ÷ 3 PHz 0,4μm÷0,1μm
D O NO T C
DO NO
T
T
DO

8
Figure 3. Electromagnetic frequency spectrum ( λ = c ≈ 3.10 m / s )
ν 1/ s

The frequencies used for satellite communications are accommodated between

Super high frequency and Extremely high frequency bands (radio waves) as shown in
Table 4.
Table 4. Range of frequencies used for satellite communications
Bandwidth,
Range Carrier, GHz Direction Service
MHz
1,5-1,6 Downlink Mobile 100
L
1,6-1,7 Uplink Mobile 100
S 2,5-2,6 Downlink Broadcast 100
3,4-4,2 Downlink Fixed 800
C 4,5-4,8 Downlink Fixed 300
5,9-7 Uplink Fixed 1100
7,2-7,7 Downlink Army 500
X
7,9-8,4 Uplink Army 500

10
 10,7-11,7 Downlink Fixed 1000
11,7-12,5 Downlink Broadcast 800
12,5-12,75 Downlink Fixed (trade) 250
Ku
12,75-13,25 Uplink Fixed (trade) 250
14-14,8 Uplink Fixed 800

Y
17,3-18,3 Uplink Fixed 1000
17,7-20,2 Downlink Fixed 2500
20,2-21,2 Downlink Mobile 1000
Ka 22,5-23 Downlink Broadcast 500

NO C P
27-30 Uplink Fixed 3000

CO Y
30 - 31 Uplink Mobile 1000

Most commercial communications satellites now operate with a 500 MHz band-

O
OP
width both on the uplink and on the downlink. They use a frequency spectrum

PY
6 / 4GHz and 14 / 12GHz (uplink/downlink). The typical 500 MHz satellite bandwidth can
be segmented into many satellite transponders bandwidths. For example, eight trans-
D O NO T C
ponders can be provided, each with a 54 MHz nominal bandwidth and center-to-center
frequency spacing 61MHz . The C-band Fixed Satellite Services (FSS) operate with a
500MHz bandwidth. A typical design would accommodate 12 transponders, each
with a bandwidth of 36 MHz, with guard bands of 4MHz between each (Figure 4). A
typical commercial communications satellite today can have 24 to 48 transponders,
operating in the C-band, Ku-band, or Ka-bands.
DO NO

500MHz
T
vertical plarization
4 36
6105 6145
T
МHz

6085 6125 6165

horizontal polarization

Figure 4. A 500MHz bandwidth accommodate 12 transponders, each with a bandwidth of 36

DO

MHz, with guard bands of 4MHz

Modern communications satellites also employ polarization frequency reuse to

increase (doubled) the number of transponders: two carriers are used at the same fre-
quency, but with orthogonal polarization. Both linear polarization (horizontal and
vertical sense) and circular polarization (right-hand and left-hand sense) have been
used.

5. Signal processing elements in satellite communications

Figure 5 shows a the basic signal processing elements in satellite communica-
tions and the techniques available for the provision of baseband formatting, source
combining, and carrier modulation in satellite communications systems.
Acronyms:
SSB/SC single sideband suppressed carrier
DSB/SC double sideband suppressed carrier
FDM frequency division multiplexing

11
 TDM time division multiplexing
FM/FDM frequency division multiplex
SCPC single channel per carrier
MCPC multiple channel per carrier
FSK frequency shift keying

Y
BPSK binary phase shift keying
QPSK quadrature phase shift keying
QAM quadrature amplitude modulation
FDMA frequency division multiple access

NO C P
TDMA time division multiple access

CO Y
CDMA code division multiple access
NIC nearly instantaneous compounding

O
CVSD continuously variable slope delta modulation

OP
ADPCM adaptive differential pulse code modulation

PY
Baseband Source Carrier Multiple Transmission
Formating Combining Modulation Access Channel
D O NO T C
Analog: FM/FDM
Voice SSB/SC SCPC, FDMA
Video DSB/SC FDM MCPC TDMA

PCM FSK
Digital: BPSK
NIC FDMA
Data QPSK
CVSD TDMA
Voice/Video
ADPCM TDM QAM CDMA
Source Multiple
Source
Multi-Plexer modulator Transmiter
Coding Access

uplink
DO NO
T
Satellite

downlink

Source Demulti- De- Multiple

Receiver
T
Destination Decoder Plexer modulator Access

The same signal processing

The same signal processing

DO

Figure 5. The basic signal processing elements in satellite communications

The Source information may be analog or digital. The first three elements
(baseband formatting, source combining, and carrier modulation) prepare the signal
for introduction to the transmission channel (the ground-to-satellite-to-ground RF
channel). The received signal at the Destination location is subjected to a reverse se-
quence of processing.

5.1. Analog Systems

In early generation satellites analog transmission has dominated satellite com-
munications since its inception. Even today many satellite systems still transmit te-
lephony and television signals using frequency modulation.
Baseband Formatting - Analog Signals
Analog baseband voice spectrum is typically reduced to the band 300–3400 Hz
for electronic transmission. The voice signals are placed on sub-carriers to allow for
propagation through the network (figure 6). The sub-carrier format is either Single

12
 Sideband Suppressed Carrier (SSB/SC) or Double Sideband Suppressed Carrier
(DSB/SC).

Amp
SSB/SC

Y
Amp

Мix filter
16,6 19,7 20 KHz

300 3400 Hz

Amp
NO C P
CO Y
20KHz DSB/SC

16,6 19,7 20 20,3 23,4 KHz

O
Figure 6. Analog voice baseband formats

OP
PY
D O NO T C
video

audio

color
DO NO

PAL chanal spectrum

T
7,5МHz

-10

-20
T
-30

-40

-50
DO

-60

5,5МHz МHz
Video
info 4,436618МHz Color Audio
info info

Figure 7. Analog video NTSC and PAL composite baseband signal spectrum

Analog video baseband consists of a composite signal that includes video,

color, and audio information (figure 7). The specific format of the components de-
pends on the standard employed for color sub-carrier modulation. Three standards are
in use: NTSC (National Television System Comm.), PAL (Phase Alternation Line),
SECAM. A modulation format for the satellite channel is a frequency modulation
(FM). The composite NTSC signal and all audio channel sub-carriers are combined
and the resulting signal frequency modulates the RF carrier, resulting in a total RF
spectrum of approximately 36 MHz. It is shown in figure 8.

13
 COMPOSITE NTSC 4,2MHz

audio 1

C
O

Y
M
6,2MHz fC =±18MHz
B
audio 2 I
N
E fC

NO C P
R

CO Y
6,8MHz
audio 3

O
OP
PY
7,2MHz

Figure 8. Signal processing format and spectrum for analog video satellite transmission
D O NO T C
Source Combining - Analog Signals
Source combining involves the combining of multiple sources into a single
signal, which then modulates an RF carrier for transmission through the communica-
tions channel. The preferred combining method for analog data is Frequency Division
Multiplex (FDM).
The ITU-T FDM standard for FDM is shown in figure 9.
Group 12 Voice Channels,
DO NO

... 4 kHz separation,

60-108KHz
T
+ + +
SuperGroup 60 Voice Channels,
... 312-552KHz
х5

Basic MasterGroup
T
... 300 Voice Channels,
х5 812-2044KHz

Super MasterGroup
900 Voice Channels,
DO

х3 8516-12388KHz

Figure 9. ITU-T FDM standard for analog voice frequency division multiplexing

Carrier Modulation - Analog Signals

Amplitude modulation (AM) was the first method to carry communications on
an RF carrier. AM is still used in satellite communications for voice and data com-
munications. Amplitude modulation is produced by mixing the information signal
with the RF carrier in a product modulator (mixer) providing a modulated RF carrier
where the amplitude envelope is proportional to the information signal.
Suppressed carrier amplitude modulation is the preferred AM modulation im-
plementation for satellite communications. Both single sideband suppressed carrier
AM (AM-SSB/SC) and double sideband suppressed carrier AM (AM-DSB/SC) are
used for subcarrier components in satellite communications.
Another technique to modulate carrier of analog signal is Frequency modula-
tion (FM). FM was used extensively in satellite communications for telephony and
14
 video transmissions on early generation analog based systems, many of which are
still in use.

5.2. Digital Signals

Y
Baseband Formatting
Digital signals dominate satellite communications systems, for data, voice, im-
aging, and video applications. Digital formatted signals allow for more comprehen-
sive processing capabilities regarding coding, error correction, and data reformatting.

NO C P
CO Y
The basis for digital communications is the binary digital (2-level) format. There are
many of binary waveforms used for the encoding of baseband data - unipolar NRZ
and polar NRZ, polar RZ, split phase (Manchester) coding, alternate mark inversion

O
OP
(AMI) and etc.

PY
A measure the quality of digital signals is Bit Error Rate (BER). The bit error
rate or bit error ratio is the number of bit errors divided by the total number of trans-
D O NO T C
ferred bits during a studied time interval. BER is an unitless performance measure,
often expressed as a percentage number. The BER is often expressed as a function of
the Eb / N 0 , (energy per bit to noise power spectral density ratio).
A second step in digital baseband formatting is multi-level coding, where the
binary bit stream is combined into groups, called symbols. It reduces the required
bandwidth. When two consecutive bits are combined, they are forming a group of
two bits, there are four possible combinations of the two bits, resulting in quaternary
DO NO

encoding. If three consecutive bits are combined, forming a group of three bits, there
T
are eight possible combinations, resulting in 8-level encoding.
The number of possible levels for “m-ary signal” is
m = 2N b
(1)
The number of bits per symbol is
T
N b = log 2 m (2)
The symbol duration is N b times the bit duration Tb , i.e.,
TS = Tb .N b (3)
DO

The symbol rate is

Rb
RS = (4)
Nb
where Rb is the bit rate.
The symbol rate is often expressed in units of baud, i.e., the baud rate. A
transmission rate of RS simbols / s is the same as a rate of RS baud . The baud rate is
equal to the bit rate only for N b = 1 , that is, for binary signals.
The second possibility is when the original source data is analog (voice or
video). Conversion to digital form is required before digital formatting is performed.
The most popular baseband formatting technique for analog source data is Pulse
Code Modulation (PCM).
In addition to conventional PCM there are some other digital voice source cod-
ing techniques used for the communications systems, as:
Adaptive differential PCM (ADPCM)

15
 ADPCM also uses differential encoding, but takes the mean-square value in the
sampling process. It requires fewer coding bits than PCM.
Adaptive delta modulation (ADM) or continuously variable slope delta modu-
lation (CVSD)
ADM uses differential encoding – only changes are transmitted. ADM pro-

Y
vides acceptable voice at 24–32 kbps, providing a more spectral efficient option.
Nearly instantaneous compounding (NIC)
NIC achieves bit-rate reduction by taking advantage of short-term redundancy

NO C P
CO Y
in human speech. It can achieve data rates approaching 1/2 that required for PCM,
because of a more efficient use of the frequency spectrum.

O
OP
Source Combining

PY
Digital satellite communications systems use Time division multiplex (TDM).
This technique combines multiple digitally encoded signals into a composite signal at
D O NO T C
a bit rate equal to or greater than the sum of the input rates. Multiple PCM bit streams
are combined in a TDM multiplexer, which generates a TDM composite bit sequence
that drives the RF modulator (Figure 10). Because PCM samples the incoming sig-
nals 8000 times per second, each sample occupies 1 / 8000s ( 125μs ). According rec-
ommendation G.702 ITU-T TDM is organized into a well structured hierarchy.
PCM
Binary
Sequence
DO NO

PCM
T
Modulator TDM
MULTIPLEXER

Composite
Analog Signals

PCM Sequence
Modulator
To RF
…...

Modulator
T
PCM
Modulator

Figure 10. Time division multiplexing source combining process

DO

Two TDM standards for voice circuits are in global use: DS or T-carrier TDM
signalling and CEPT TDM signalling. Each hierarchy starts with 64 kbps analog
voice, but the subsequent TDM levels consist of different combinations, as shown in
Figure 12. The systems used in Europe and North America are different (Figure 12).
The North American standard is based on a 24-channel PCM system, whereas the
European system is based on 30/32 channels. This system contains 30 speech chan-
nels, a synchronisation channel and a signalling channel, and the gross line bit rate of
the system is 2.048 Mbps (32 x 64 Kbps). The system can be adapted for common
channel signalling, providing 31 data channels and employing a single synchronisa-
tion channel.
The following details refer to the European system (Figure 11). The 30/32
channel system uses a frame and multiframe structure, with each frame consisting of
32 pulse channel time slots numbered 0-31. Slot 0 contains the Frame Alignment
Word (FAW) and Frame Service Word (FSW). Slots 1-15 and 17-31 are used for
digitised speech (channels 1-15 and 16-30 respectively). In each digitised speech
16
 channel, the first bit is used to signify the polarity of the sample, and the remaining
bits represent the amplitude of the sample. The duration of each bit on a PCM system
is 0,488μs . Each time slot is therefore 3,904μs ( 8bits * 0,488μs ). Each frame therefore
occupies 125μs ( 32 *3,904μs ).

Y
NO C P
CO Y
O
OP
PY
D O NO T C
DO NO
T
Figure 11. The frame and multiframe structures for a 30/32 channel PCM system
T
Japan USA Europe
DO

397200 kbit/s 564992 kbit/s

*4 *4
139264 kbit/s
97728 kbit/s 274176 kbit/s
1920 channels
*6 *3
*3 *4
34368 kbit/s
32064 kbit/s 44736 kbit/s
480 channels
*5 *7 *4
8448 kbit/s
6312 kbit/s 120 channels
*3 *4
*4
2048 kbit/s
1544 kbit/s 30 channels

*24 *30
64 kbit/s
ISDN
Figure 12. Standardized TDM structures

17
 In order for signalling information (dial pulses) for all 30 channels to be trans-
mitted, the multiframe consists of 16 frames numbered 0-15. In frame 0, slot 16 con-
tains the Multiframe Alignment Word (MFAW) and Multiframe Service Word
(MFSW). In frames 1-15, slot 16 contains signalling information for two channels.
The frame and multiframe structure are shown below. The duration of each multi-

Y
frame is 2ms ( 125μs *16 ). A TDM multiplexer can provide Bit multiplexing, Byte (8
bits symbol) multiplexing and block multiplexing.

NO C P
CO Y
Forward Error Correcting (FEC)
The purpose of Forward Error Correcting (FEC) is to improve the capacity of a

O
transmission channel by adding to the source data redundant information. It is a proc-

OP
ess known as channel coding. Channel coding consists:

PY
Reed Solomon coding;
Interleaving;
D O NO T C
Convolution code.
The place of channel coding is between multiplexing and modulating process,
as shown in Figure 13.

video
audio TV coder

M
DO NO

video U f
audio TV coder L
T
T PA
I
..................

P RS Convoluti QPSK
L interleaving
coder on coding modulater
E
X
T
DVB-S
E
R
video
audio TV coder

MPEG
DO

coder

Фиг. 13. Transmitter digital sat TV.

Channel coding is described in more detail in lecture 3.

Carrier Modulation - Digital Signals

Digital modulation is accomplished by amplitude, frequency, or phase modula-
tion of the carrier by the binary (or m − ary ) bit stream.
The major modulation formats used in satellite communications are:
- Binary Phase Shift Keying (BPSK) - polarity changes in binary signal used to
produce 180° carrier phase change;
- Differential Phase Shift Keying (DPSK) – phase shift keying where carrier
phase is changed only if current bit differs from preceding bit. A reference bit must
be sent at start of message for synchronization;
- Quadrature Phase Shift Keying (QPSK) – phase shift keying for a 4-symbol
waveform. Data bit streams are converted to two bit streams, I and Q, and then binary

18
 phase shifted as in BPSK. The adjacent phase shifts are equi-spaced by 900 . BPSK
only requires one-half the bandwidth of BPSK;
- M-ary Phase Shift Keying (MPSK) – phase shift keying for m − ary symbol
waveform;
- Minimum Shift Keying (MSK) – phase shift keying with additional process-

Y
ing to smooth data transitions, resulting in reduced bandwidth requirements;
- Quadrature Amplitude Modulation (QAM) – multilevel (higher than binary)
modulation; it is a combination of amplitude and phase modulations.

NO C P
CO Y
BPSK and QPSK are the most widely used in satellite systems.
In the simplest case PSK is based on phase switching signal to 1800
when changing from a logic “0” to “1” and “1” to “0”, but the amplitude

O
OP
and frequency remain constant. The change of phase state is finding relatively simple.

PY
It is shown in figure 14.
D O NO T C
carrier

Digital
0 1 1 0 signal
t
PSK
modulated
t
DO NO

signal
T
Figure 14. PSK modulation

Quadrature Phase Shift Keying (QPSK)

T
In QPSK the phase of the carrier is shifted between four positions that are 90
degrees apart. The binary data stream is converted into 2-bit symbols, which are used
to phase modulate the carrier. The serial-to-parallel process used to generate the two
parallel encoded bit streams is shown in figure 15. Odd numbered bits in the original
DO

data sequence p(t ) are sent to the i (in-phase) channel, producing the sequence pi (t ) .
Even numbered bits are sent to the q (quadrature) channel, producing the sequence
pq (t ) . The bit duration is doubled in the i and q channels, reducing the bit rate to ½
the original data bit rate (Figure 16).
pi(t) pi (t ) cos ω 0 (t )
+

pi(t)
a i cos ω0 (t )
c e g t
p(t) Serial to
QPSK
a b d h i
Parallel
Converter 900
+ modulated
c e f g t signal

− sin ω0 (t )
b d h
pq(t)
+

f pq(t) t − p q (t ) sin ω 0 (t )

Figure 15. QPSK modulator Figure 16. Generation of the QPSK waveform

19
 The in-phase signal is mixed directly with the carrier frequency cos(ω0t ) , while
the quadrature signal is mixed with a 900 phase shifted carrier
( cos(ω0t + 900 ) = − sin(ω0t ) ). The output of the two mixers is summed to produce the
modulator output signal:

Y
s (t ) = pi (t ) cos(ω0t ) − pq (t ) sin(ω0t ) (5)
The phase state of s(t ) depend on the bit values that compose the in-phase and
quadrature component signals. Figure 17 and Table 5 show the four possible combi-

NO C P
nations of bits and the resulting s (t ) and the phase state diagram for the four bit se-

CO Y
quences, plotted pi (t ) and pq (t ) . The symbol phases are orthogonal, 900 apart from

O
each other, with one in each quadrant.

OP
The symbol rate is two bit per symbol, Es = 2 Eb , and QPSK requires half the

PY
transmission bandwidth of BPSK.
D O NO T C
s (t ) I(t) Q(t) QPSK изход
1/ 2 1/ 2 cos(ω0t ) − sin(ω0t ) = 2 cos(ω0t + 45 )
0
11

1/ 2 −1/ 2 cos(ω0t ) + sin(ω0t ) = 2 cos(ω0t − 45 )

0
10
−1/ 2 1/ 2 − cos(ω0t ) − sin(ω0 t ) = 2 cos(ω0t + 135 )
0
01
Q(t)
00 −1/ 2 −1/ 2 − cos(ω0 t ) + sin(ω0 t ) = 2 cos(ω0 t − 135 )
0
1350 450
01 11
si(t) 00 11 10 01 1/ 2
DO NO

+1
T
0
1/ 2 I(t)
t

-1 00 10
0
? i(t) -135 -450
T
1350

450

-450 t
DO

0
-135

Figure 17. The four possible combinations of bits and the resulting s (t ) and the phase state
diagram

pi(t)
+

cos ω0 (t ) LPF

Serial to p(t)
QPSK
Parallel
modulated
900 Converter
signal

− sin ω0 (t )
+

pq(t)
LPF

Figure 18. QPSK demodulator

20
 The QPSK demodulator reverses the process as shown in Figure 18. The input
signal, which has passed through the communications channel, is split into two chan-
nels. Each channel is essentially a BPSK demodulator and the demodulation process
produces the in-phase and quadrature components pi (t ) and pq (t ) , which are then

Y
parallel-to-serial converted to produce the original data stream p(t ) .
Higher Order Phase Modulation
Higher Order Phase Modulations can achieve further reduction in symbol rate,

NO C P
and the required transmission bandwidth. For example, 8-phase shift keying (8PSK),

CO Y
which combines groups of three bits per symbol, requires a transmission channel
bandwidth of 1/3 BPSK, with the phase state diagram as shown in Figure 19.

O
OP
010

PY
001

100 000
D O NO T C
111 011

110 101

Figure 19. 8PSK phase state diagram

Disadvantage of 8PSK is requirement of twice the power over BPSK or QPSK

DO NO

to achieve the same overall performance. 8PSK Modulation is important in satellite

T
communications systems because the additional bit in the symbol can be used for er-
ror correction coding, allowing an additional approximately 3 dB of coding gain.
BER values exceeding 1.10−10 .
T
Multiple Access
A multi-user system achieves a specific point in the capacity region depending
on how the multi-user channel is shared by the users, which depends on the multiple
DO

access technique or Multiple Access (MA). It is the last technique before signal trans-
mitting.
Satellite links are designed to provide desired link availability for average con-
ditions. Satellite MA techniques interconnect ground stations through multiple satel-
lite transponders with the goal of optimizing several system attributes such as: spec-
tral efficiency; power efficiency; reduced latency; and increased throughput.
The MA methods available to the satellite system designer can be categorized
into three fundamental techniques:
• Frequency Division Multiple Access (FDMA);
• Time Division Multiple Access (TDMA);
• Code Division Multiple Access (CDMA).
FDMA systems consist of multiple carriers that are separated by frequency in
the transponder. The transmissions can be analog or digital, or combinations of both.
In TDMA the multiple carriers are separated by time in the transponder, presenting
only one carrier at any time to the transponder. TDMA is most practical for digital

21
 data only, because the transmissions are in a burst mode to provide the time division
capability. CDMA is a combination of both frequency and time separation. It is the
most complex technique, requiring several levels of synchronization at both the
transmission and reception levels. CDMA is implemented for digital data only, and
offers the highest power and spectral efficiency operation of the three fundamental

Y
techniques.
Multiple Access is described in more detail in lecture 4.

NO C P
CO Y
References
1. A.C. Clarke, “Extraterrestrial Relays”,Wireless World, Vol. 51, pp. 305 308,

O
OP
October 1945.

PY
2. Bernard Sklar, Digital Communications: Fundamentals and Applications,
Second Edition, Prentice-Hall, 2001
D O NO T C
3. Dennis Roddy, Satellite communications, McGraw-Hill Professional, 2001 -
569 p.
4. Deffebach H. L. and Frost W. O., A survey of digital baseband signaling
techniques. NASA Technical Memorandum NASATM X-64615, June, 1971.
5. Gérard Maral, Michel Bousquet, Satellite communications systems: systems,
techniques, and technology, John Wiley and Sons, 2002 – 757p.
6. Ippolito, Louis J., Satellite communications systems engineering:
DO NO

atmospheric effects, satellite link design, and system Performance, 2008 JohnWiley
T
& Sons Ltd, 396 p.
7. International Telecommunications Union, www.itu.int.
8. Korn I., Digital Communications. Van Nostrand Reinhold Company, New
York, 1985.
T
9. Lathi B. P., Modern digital and analog communications systems, Third Edi-
tion, Oxford University Press, 1998, 781 стр.
10. Michael O. Kolawole, Satellite communication engineering, Marcel
DO

Dekker, 2002 - 263 p.

11. Rappaport T. S., Wireless Communications, Chapter 3 and 4, Prentice Hall,
Upper Saddle River, New Jersey, 1996.
12. Zhili Sun, Satellite networking principles and protocols, John Wiley &
Sons, 2005. 342 p.

22
 Satellite Communications Systems
Lecture 2.
Link Budget Calculation and Analysis

Y
1. The quality of signal transmission
2. Elements of the link
2.1. Effective Isotropic Radiated Power

NO C P
2.2. Power flux density

CO Y
2.3. Antenna Gain
2.4 Free-Space Path Loss

O
2.5. System Noise

OP
2.6. Link Performance Parameters

PY
3. Link Budget
3.1. Frequency Translation Satellite
D O NO T C
3.2. The on-board processing satellite

The goal of satellite communications is to provide links between fixed termi-

nals, mobile terminals and both fixed and mobile terminals on lend, on the air, and at
sea.
A link budget is the sum of all gains and losses in the radio connection be-
tween two terminals from end to end, including antenna's, feed lines and the path be-
DO NO

tween the antenna's, but also the relevant portions of the transmitter and the receiver.
T
G G

Tx Rt
T
PT
Losses
dBm in feeder
DO

PR

Free space loss

distance d, km

Figure 1. Gains and losses in the radio connection between two terminals

1. The quality of signal transmission

The quality of digital signals (digital links) is measured by the Bit Error Rate
(BER). The bit error rate or bit error ratio is the number of bit errors divided by the
total number of transferred bits during a studied time interval. BER is a unitless per-
formance measure.
The BER is often expressed as a function of the Eb / N 0 , (energy per bit to noise
power spectral density ratio).
1
BER = erfc ( Eb N 0 ) , (1)
2
where the function erfc is called a complimentary error function, it is tabulated.
1
 The relation between BER and Eb N 0 is shown in Figure 2.
BER
1.10 -1

1.10 -2

Y
1.10 -3

1.10 -4

NO C P
CO Y
1.10 -5

1.10 -6

O
1 2 3 4 5 6 7 8 9 10 E b /N 0

OP
Figure 2. The relation between BER and Eb N 0

PY
Higher Eb N 0 means better quality.
D O NO T C
The parameter Eb N 0 is the most common to compare the digital communica-
tion systems, even they have differing bit rates or modulation. The quantity Eb is a
measure of bit energy, a ratio of the average signal power and bit rate:
Pavg W
Eb = , . (2)
Rb bit / s
DO NO

Example 1: A signal has power 15W, his bit rate is 200bps. What is the bit energy in deci-
T
bels?
Eb = 10 log15 − 10 log 200 = −11,25dB

The quantity N 0 is called the noise power density (or noise power spectral den-
T
sity). It is the total noise power in the frequency band of the signal divided by the
bandwidth of the signal:
PN W
N0 = , (3)
DO

BN Hz

Example 2: A signal has a noise power 3W, the signal bandwidth is 500 Hz,
Eb = −11,25dB . What is N 0 and Eb N 0 ?
Eb
N 0 = 10 log 3 − 10 log 500 = −22,22dB , = (−11,25) − (−22,22) = 10,97 dB
N0

The ratio (3) shows, that if bandwidth of the signal decreases noise power den-
sity increases (the same amount of noise occupies a smaller signal space) and the pa-
rameter Eb N 0 decreases. It also decreases when the bit rate increases.
The quality of analog signals is measured by the carrier to noise ratio C N or
by the carrier to noise power spectral density ratio C N 0 . The ratio C N is the carrier
power in the whole useable bandwidth, where C N 0 is the carrier power per unit
bandwidth.
The relation between C N , C N 0 and Eb N 0 is the power of carrier C = Eb .Rb .
2
 C E ⎛C ⎞ E
= b . Rb , in dB: ⎜⎜ ⎟ = b + Rb ,
⎟ (4)
N0 N0 ⎝ N 0 ⎠ dB N 0
C Eb Rb ⎛C⎞ E
= . , in dB: ⎜ ⎟ = b + Rb − B , (5)
N N0 B ⎝ N ⎠ dB N 0

Y
2. Elements of the link
A link consists of three parts with specific parameters: transmitter, receiver and

NO C P
medium, as shown in figure 3.

CO Y
d

O
pt pr

OP
Transmitter Receiver

PY
gt gr
pfdr
Figure 3. The three parts of link.
D O NO T C
The parameters of the link are:
pT - transmitted power (in watts);
pR - received power (in watts);
gT - transmit antenna gain;
g R - receive antenna gain, and
DO NO

d - path distance (km).

T
2.1. Effective Isotropic Radiated Power
An important parameter in the evaluation of the RF link is the effective iso-
tropic radiated power (EIRP). The EIRP, using the parameters introduced in Figure 3,
T
is defined as
EIRP = pT .g T , W , (6)
EIRP = 10 log PT + 10 log GT , dBW , (7)
DO

where pT - transmitted power (in watts); gT - transmit antenna gain.

EIRP evaluate the energy potential of the transmitter.
Requirements for the values of EIRP are usually
- EIRP = 50 ÷ 100 dBW of the ground transmitter;
- EIRP = 45 ÷ 65dBW of the satellite transponder.
Equations (6) and (7) do not account Transmitter Feeder loss LC and Antenna
Pointing Loss LPointing due to incorrect orientation of the antenna to the satellite.
For precise calculations equation (7) is:
EIRP = 10 log PT + 10 log GT − 10 log LC − 10 log LPointing dBW , (7a)

2.2. Power flux density

A specific parameter is the power flux density (power density pPFD ). It is a
measure of energy that can be gathering from a particular source (figure 3) and it is
expressed in watts / m 2 .

3
 pT gT EIRP
pPFD = = , W / m2 , (8)
4πd 2
4πd 2

where d - path distance (km).

PPFD = EIRP − 10 log(4πd 2 ) = EIRP − 20 log(d ) − 10 log(4π ) =
(9)
= EIRP − 20 log(d ) − 10,99 dB

Y
The power flux density gives the function between transmitted and received
power. The power flux density limited service area of the satellite:

NO C P
- for Direct Broadcast System (DBS), like TV-Sat, TDF, Olimpus

CO Y
PPFD ≥ −103dBW / m 2 ;
- for multifunctional systems (Astra, Eutelsat, Intelsat)

O
PPFD ≥ −(113 ÷ 116)dBW / m .
2

OP
PY
Example 3: The transmitted power of Earth station is 10W , a distance to satellite is
d = 37500 km . The antenna gain is 50dB . What is the power flux density?
D O NO T C
EIRP = 10 log PT + 50 = 10 + 50 = 60, dBW
PPFD = EIRP − 10 log(37500) − 10,99 = 60 − 20 log(37500) − 10,99 =
2

= 60 − 91,48 − 10,99 = −42,47dB

2.3. Antenna Gain

Diagram of the broadcast antenna means that compared with an isotropic radia-
tor, it concentrates more electromagnetic energy in a definitely direction then to an-
DO NO

other. In this sense antenna gives “gain” the signal in that direction. The gain of the
T
antenna shows how much more is the gain compared to isotropic antenna gain.
In physical antennas some energy is reflected away by the structure, and some
energy is absorbed by lossy components (feeds, struts, subreflectors). In this reason to
T
account for this, an effective aperture Aeff is defined in terms of an aperture efficiency
η such that Aeff = η . A , (A is antenna area).

⎛ π .D ⎞
2
4.π . Aeff
DO

g = η .⎜ ⎟ = , (10)
⎝ λ ⎠ λ2
η = 0,55 ÷ 0,65 ; λ = c f (wave length is the distance from one wave point to the next

wave point); c - speed of the light, c = 3.10 m / s ; f is the frequency.

8

⎡ ⎛ π .D ⎞ 2 ⎤
G = 10 log ⎢η .⎜ ⎟ ⎥ = 10 logη .(10,47. f .D) , dBi
2
(11)
⎣⎢ ⎝ λ ⎠ ⎦⎥
Table 1 gives example for the values of gain depending on frequency and an-
tenna’s diameter D .

Table 1. The values of antenna gain

f, GHz 17,5 11,5 3,9 2,6 0,72
D, m \ λ, m 0,017 0,026 0,0,77 0,115 0,417
0,6 38,7 35,0 25,5 22,1 10,9
2,0 49,1 45,4 36,0 32,5 21,3
5,0 57,1 43,9 43,9 30,5 29,3

4
 Figure 4 shows a typical directional antenna pattern for a circular parabolic re-
flector antenna. The antenna pattern shows the gain as a function of the distance from
the boresight direction. The boresight direction is a direction of maximum gain, for
which the value g is determined from (11) equation. The 1/2 power beamwidth is the

Y
contained conical angle α for which the gain has dropped to 1/2 the value at bore-
sight, the power is 3 dB down from the boresight gain value.

NO C P
The antenna beamwidth for a parabolic reflector antenna can be approximately

CO Y
determined in degree from
75.λ 22,5
α≅ = , in degree . (12)

O
D D. f

OP
0

PY
-450 450
0,7Emax
D O NO T C
α

900

-1350 1800 1350

F (ϕ )
DO NO
T
F 2 (ϕ )
T
-1800 -1350 -900 -450 0 450 900 1350 1800

Figure 4. The typical antenna bandwidth

DO

Most antennas have sidelobes. Its are regions where the gain may increase due
to physical structure elements or the characteristics of the antenna design. It is also
possible that some energy may be present behind the physical antenna reflector.
Sidelobes are a possible source for noise and interference, for example, when the sat-
ellite ground antenna is located near to other antennas or sources of power in the
same frequency band as the satellite link.
Because of antenna beamwidth for satellite links is very small (much less than
α < 1 ), it is requiring careful antenna pointing and control to maintain the link.
0

2.4 Free-Space Path Loss

When signals are transmitting between the transmitting and receiving antennas
signal fade in space. This is due to loss of energy dissipation in the propagation. For
GEO orbits satellites the attenuation range is from 195 to 213 dB for 4 / 30 GHz.
According to figure 3 the power pR got by the receiving antenna will be
pR = pPFD . Aeff , (13)

5
 GR .λ2
where Aeff = .
4π
Replacing Aeff and pPFD (8) in equation (13),
2
⎛ λ ⎞ p .g .g
pR = pT .gT .g R ⎜ ⎟ = T T R, (14)

Y
⎝ 4πd ⎠ lFS
16π d 2 2
(4π . f .d ) 2
where lFS = = is a free space path loss.
λ2 c2

NO C P
In dB:

CO Y
⎛ 4.π .d ⎞ ⎛ 4π . f .d ⎞
LFS = 20 log⎜ ⎟ = 20 log⎜ ⎟ (15)
⎝ λ ⎠ ⎝ c ⎠

O
OP
Example 4: The frequency uplink is 6,175GHz, a distance to satellite is d = 37500km .

PY
What is the free space path loss?
(4π . f .d ) 2 4π . f .d 4.3,14.6,175.109.37,5.106
D O NO T C
LFS = = 20 log = 20 l og = 199,73 dB
c2 c 3.108

Equation for Received Power (14) in dB

PR = EIRP + GR − LFS , dB (16)
gives the basic link equation, sometimes referred to as the Link Power Budget Equa-
tion, for a satellite communications link, and is the design equation from which satel-
DO NO

lite design and performance evaluations proceed.

T
Example 5: What is the power flux density pPFD and power of received signal pR if both
the transmit and receive parabolic antennas have a diameter of 3 m, D = 3m , the transmit power is
12 watts, pT = 12W , the antenna efficiency is 55% for both antennas, η = 0,55 . The satellite is in a
T
GSO location, with a range of d = 37500km . The frequency of operation is 12 GHz.
First have to estimate antenna gain (equation 11)
GdB = 10 logη .(10,47. f .D) 2 = 10 log 0,55.109,66.144.9 = 48,93 , dBi
DO

Then EIRP is: EIRP = 10 log PT + GT = 10,79 + 48,93 = 59,72 dB

The free space path loss, in dB is (Equation 15)
(4π . f .d ) 2 4π . f .d
LFS = 10 log 2
= 20 log = 215,05 dB
c c
The received power, in db, is then found from the Equation (16):
PR = EIRP + GR − L = 59,72 + 48,93 − 215,05 = −106,4 dB
The received power in watts:
−106.4
PR = 10 10 = 2,29.10 −11 W
The power flux density, in db, is then found from the Equation (9):
PPFD = EIRP − 20 log(d ) − 10,99 = 59,72 − 20 log 37,5.106 − 10,99 = 102,75 dB .
The calculation shows, that the received power is very, very low.

6
 2.5. System Noise
Noise is a parasitic undesired signal, superimposed on the useful signal, caus-
ing adverse effects in the processing of signals. According to the physical character,
the noises are: Thermal noise, Shot noise and Flicker noise.
Sources of noise and there frequency domains are different, but in the micro-

Y
wave range (above 300 MHz) matter have two main sources. One of them is related
to random changes in the currents of the media in transitions of semiconductor struc-
tures (shot noise), and the other - with occasional changes voltages, primarily related

NO C P
CO Y
to the thermal motion of carriers in the volume of in the wires (thermal noise). Ther-
mal noise is called “white noise” and is a basic in RF band.
The noise introduced by each device in the system is quantified by the intro-

O
OP
duction of an equivalent noise temperature t E . It is defined as the temperature of a

PY
passive resistor producing a noise power per unit bandwidth.
Solid state theory gives the following value of noise voltage
D O NO T C
u N2 = 4k .t E .r.bN , V 2 (17)
where u N is a noise voltage;
r is a resistance, which occurs on the noise voltage;
⎧k = 1,37.10 −23 Ws / deg,
⎪
k is Boltzmann’s constant: ⎨k = −198dBm / K / Hz, ;
⎪k = −228,6dBW / K / Hz
⎩
DO NO

t E equivalent noise temperature in degree Kelvin;

T
bN - noise bandwidth.
The noise power is
u N2
nN = = k .t E .bN , W (18)
T
2r
The noise bandwidth is the RF bandwidth of the information-bearing signal –
usually it is the filtered bandwidth of the final detector/demodulator of the link.
A noise power density (noise power spectral density) is
DO

nN
n0 = = k .t E , W / Hz (19)
bN

2.5.1. Noise figure

The noise figure NF is defined by considering the ratio of the desired signal
power to noise power ratio at the input of the device, to the signal power to noise
power ratio at the output of the device
( p / n) IN
nf = (20)
( p / n)OUT
In other wise
pIN
( p / n) IN k .t0 .b t +t t
nf = = = 0 E = 1+ E (21)
( p / n)OUT g. pIN t0 t0
g .k .(t0 + t E ).b

7
 where t0 is the input reference temperature, usually set at 290 K , and b is the noise
bandwidth.
⎛ t ⎞
NF = 10 log⎜⎜1 + E ⎟⎟, dB (22)
⎝ t0 ⎠

Y
⎛ t ⎞
The term in brackets, ⎜⎜1 + E ⎟⎟ , is sometimes referred to as the noise factor,
t ⎝ 0 ⎠
when expressed as a numerical value.

NO C P
CO Y
Corollary:
nout = nf .g.nIN (23)

O
t E = t0 .(nf − 1), K (24)

OP
NF

PY
t E = t0 .(10 10
− 1), dB (25)
nN = k .t0 .(nf − 1).bN , W (26)
D O NO T C
2.5.2. Noise Temperature
Noise temperature concern three types of devices: active devices, passive de-
vices, and the receiver antenna system.
Active devices in the communications system are amplifiers and other compo-
nents (upconverters, downconverters, mixers, active filters, modulators, demodulators,
and etc.). They provide an output power that is greater than the input power (increase
DO NO

the signal level and input noise, g>1).

T
For active devices noise temperature can be expressed by equations (24) and
(25).
Passive devices are waveguides, diplexers, filters, and switches. They reduce
the level of the power passing through the device (decrease the signal level, g<1).
T
A passive device is defined by the loss factor:
pIN
l= (27)
pOUT
DO

where pIN and pout are the powers into and out of the device, respectively.
The input to the ideal amplifier noise contribution of the passive device will be
t E = t0 .(l − 1) = 290.(l − 1), K (28)

2.5.3. Receiver Antenna Noise

There are two possible ways to introduce noise into the system at the receiver
antenna: from the physical antenna structure (main reflector, subreflector, struts, etc.),
it is called antenna losses, and from the radio path - radio noise or sky noise.
The antenna loss is usually included as part of the antenna aperture efficiency,
η = 0,55 ÷ 0,65 , and does not need to be included in link power budget calculations di-
rectly.
Radio noise can be introduced into the transmission path from both natural and
human induced sources. This noise power will add to the system noise through an
increase in the antenna temperature of the receiver.
The natural components present in radio noise on a satellite link are:

8
 - Galactic noise: ≈ 2,4 K for frequencies above about 1GHz ;
- Atmospheric constituents that absorbs the radiowave and will emit energy in
the form of noise, like oxygen, water vapor, clouds, and rain (most severe for fre-
quencies above about 10 GHz).

Y
- Extraterrestrial sources - the moon, sun, and planets.
Human sources of radio noise consist of interference noise in the same infor-
mation bandwidth induced from:

NO C P
- communications links, both satellite and terrestrial;

CO Y
- machinery;
- other electronic devices that may be in the vicinity of the ground terminal.

O
For the antenna pointing to the sky (ground station antenna) the output noise

OP
power from the antenna has two components which are represented by the sky tem-

PY
perature, Tsky , and the earth temperature Tearth .
Sky temperature, Tsky is due to noise originating in the atmosphere. It varies
D O NO T C
with frequency and the elevation angle ϕ of the antenna. The sky temperature is
higher for ϕ = 00 (antenna pointing to the horizon) because of the longer path of the
radiation through the atmosphere. Elevation angles of less than 10° are usually
avoided.
Figure 5 show Tsky for different frequency ranges 1 .
DO NO
T
Tsky φ=00
T
φ=650

φ=800
DO

φ=870

φ=900

Figure 5. Sky noise for clear air and 7.5 g/m3 of water vapour concentration (φ is the
elevation angle)

The earth temperature Tearth interference noise enter the system through the
sidelobes or backlobes of the ground receiver antenna, it is often difficult to quantify
interference noise directly.
For the antenna pointing to the earth (satellite antenna) the noise temperatutre
of the antenna is about 290K, the physical temperature of the earth.

1
J P Silver. Satellite Communications Tutorial, E-mail: [email protected]

9
 2.5.4 System Noise Temperature
The noise contributions of each device in the communications transmission
path, including sky noise, will combine to produce a total system noise temperature.
Consider a typical satellite receiver system with the components shown in Fig-
ure 6: an antenna with a noise temperature of t A ; a low noise amplifier (LNA) with a

Y
gain of g RF and noise temperature of t RF ; a cable with a line loss of 1 / l and noise
temperature of tC ; a downconverter (mixer) with a gain of g M and noise temperature

NO C P
of t M ; and finally an intermediate frequency (I.F.) filter and amplifier with a gain of

CO Y
g IF and t IF .

O
OP
Receiver
Down

PY
Converter
line (Mixer)
D O NO T C
1/l gM IF Filter IF Amplifier
tA tC LNA
gRF tM
tRF gIF
tIF

Figure 6. Receiver System Components

The receiver system noise temperature will be

DO NO

tM t IF
t R = t RF + + , (28)
T
g RF g RF .g M

Example 6: A receiver with a low noise amplifier (LNA) with a gain of GRF = 23dB
( g RF = 200 ) and noise temperature of t RF = 50 K ; a downconverter (mixer) with a gain of
T
GM = −10dB ( g M = 0,1 ) and noise temperature of t M = 500 K , an intermediate frequency (I.F.)
unit: t IF = 1000 K , GIF = 30dB ( g IF = 1000 ).What is the receiver system noise temperature.
t t IF 500 1000
t R = t RF + M + = 50 + + = 102,5 K .
DO

g RF g RF .g M 200 200.0,1

The total system noise temperature will be

1 t ⎛ 1⎞
t S = (t A + tC ) + t R = A + t0 ⎜1 − ⎟ + t R , (29)
l l ⎝ l⎠
The system noise temperature t S represents the noise present at the antenna
terminals from all the front-end devices.

2.5.5. The quality of the receiver

The ratio of receiver antenna gain to the receiver system noise temperature
GR / t S indicates figure of merit.
G
= GR − TS , dB / K (30)
T

10
 Example 7: Calculate Receiver G/T (dB/K) of a satellite having antenna gain 42 , over all
receiver noise temperature t S = 75K .
G
= GR − 10 log t S = 42 − 18,75 = 23,25 dB / K
T
The minimum values for sat TV are:

Y
- 6dB / K for an individual receiver;
- 14dB / K for a joint receiver.

NO C P
CO Y
2.6. Link Performance Parameters
2.6.1. Carrier-to-Noise Ratio
The average RF carrier power to noise power ratio C N is carrier power in the

O
OP
whole useable bandwidth.

PY
The C N can be expressed in terms of the eirp, G/T, and other link parameters
developed earlier
D O NO T C
pR ⎛ C ⎞ pT .gT .g R
=⎜ ⎟= ; (31)
nR ⎝ N ⎠ k .t S .bN .(lFS .lad )
where lad is sum of all other losses. The other losses could be from the free
space path itself, such as rain attenuation, atmospheric attenuation, etc., or from
hardware elements such as antenna feeds, line losses, etc.
⎛C⎞
⎜ ⎟ = EIRP + R − (LFS + Lad ) − 228,6 − 10 log bN , dB ,
G
(32)
⎝N⎠
DO NO

TS
where the EIRP is in dBW , the bandwidth BN is in dBHz , and
T
k = −228,6 dBW / K / Hz .
The larger the ratio C N is the better. Typical communications links require
minimum C N values of 6 to 10 dB for acceptable performance. The performance of
T
the link will be degraded in two ways: if the carrier power c, is reduced, and/or if the
noise power nB , increases.
DO

2.6.2. Carrier-to-Noise Density

The carrier-to-noise density is defined in terms of noise power density, defined
by Equation (19):
nN k .t S .bN
n0 = = = k .t S (33)
bN bN
Equation (4) and (5) give the relation between C N and C N 0 by the band-
width BN .
c c 1 ⎛C⎞ ⎛ C ⎞
= . , ⎜ ⎟ = ⎜⎜ ⎟ − BN , dB . (34)
n n0 bN ⎝ N ⎠ dB ⎝ N 0 ⎠⎟
⎛C ⎞
⎟⎟ = ⎛⎜ ⎞⎟ + BN , dBHz ,
c c C
= . bN , ⎜⎜ (35)
n0 n ⎝ N 0 ⎠ dB ⎝ N ⎠
The carrier-to-noise density behaves similarly to the carrier-to-noise ratio in
terms of system performance. The larger value C N 0 - better performance.

11
 2.6.3. Energy-Per-Bit to Noise Density
For digital communications systems, the bit energy, eb , is more useful than car-
rier power in describing the performance of the link. The bit energy is related to the
carrier power c = eb .rb (or eb = c.Tb , Tb is the bit duration).

Y
The ratio eb n0 is related to c n0 :
⎛ eb ⎞ ⎛ c ⎞ 1 ⎛ c ⎞
⎜⎜ ⎟⎟ = ⎜⎜ ⎟⎟ . = ⎜⎜ ⎟⎟.Tb , (36)
⎝ n0 ⎠ ⎝ n0 ⎠ rb ⎝ n0 ⎠

NO C P
CO Y
⎛ eb ⎞ ⎛ c ⎞ bN
⎜⎜ ⎟⎟ = ⎜ ⎟. , (37)
⎝ n0 ⎠ ⎝ n ⎠ rb

O
The eb n0 will be numerically equal to the c n when the bit rate ( rb , bps ) is

OP
equal to the noise bandwidth ( bN , Hz ).

PY
D O NO T C
Example 8. 2 :
There is a satellite with a range 40,000 km (range); transmitted power pT = 2W ( 3 dB ); an-
tenna gain gT = 17 dB ( 50 ); channel system noise temperature t S = 152 K and bandwidth
bN = 500MHz of the satellite channel; Frequency = 11GHz.
What are:
a) Power Flux Density pPFD to the received terminal;
b) Received power pR to the received terminal with antenna effective aperture Aeff = 10m 2 .
DO NO

c) Received antenna gain g R ;

T
d) Received ratio c n
Results:
a) Power Flux Density by (8) and (9):
g p 2.50
pPFD = T T2 = = 4,97.10 −15 ,W / m 2 ;
T
4πd 4π .(4.107 ) 2
PPFD = 3 + 17 − 11 − 152 = −143dBW / m 2
DO

b) Received power pR by (13):

pR = pPFD . Aeff = 4.97.10 −15.10 = 4.97.10 −14 W
PR = 4.97.10−14 W = −133dBW

c) Gain of receiving antenna g R by (10):

λ = (c / f ) = 3.108 / 11.109 = 0.0273 ,
4π . Ae 4.3,14.10 125.6
GR = = = = 167467; GR = 52,24dB
λ 2
0,0273 2
7,5.10− 4
d) Carrier-to-Noise Ratio
It can determinate noise power first
p N = n = k .t0 .bN = 1,37.10 −23.152.500.106 = 104,12.10 −20 ,
PN = −119,8 dBW .
Received power is PR = C = 4.97.10 −14 W = −133dBW ;

2
Mohamed Khedr, Лекция: Satellite Communication Systems, http://webmail.aast.edu/~khedr

12
 C
Carrier-to-Noise Ratio: = −133 − (−119.8) = −13.2dBW .
N

3. Link Budget
The link budget determines transmitter power needed to obtain suffi-

Y
cient strong signal in the receiver. The quality of the satellite link is estimated by the
Carrier-to-Noise Ratio in the receiving earth station or terms of energy per bit to
noise density.

NO C P
CO Y
Path noise is added to the signal at the uplink and downlink. Path noise is the
sum of additive noise effects such as noise caused by atmospheric gases, clouds, rain,
depolarization, surface emissions, or extra-terrestrial sources. Path losses are intro-

O
OP
duced in the uplink and the downlink signal paths. Path loss is the sum of signal

PY
power losses caused by effects such as gaseous attenuation, rain or cloud attenuation,
scintillation loss, angle of arrival loss, or antenna gain degradation.
D O NO T C
The total system carrier-to-noise ratio, (c n) S , is determined by developing the
system equations for the total link, including the path degradation parameters.
The parameters used in the link calculations are shown in Figure 7.

nfNSR, pSR, gSR pST, gST

2 3
stelitte
dD
U

, fD
L
U,

,L
DO NO
,f

D
dU

1
T
4
pGT, gGT nfNGR, pGR, gGR
Transmit Receive
terminal terminal

Figure 7. The parameters used in the link calculations

T
The communications satellite transponder is implemented in one of two gen-
eral types:
DO

1) the conventional frequency translation (FT) satellite, which receives the up-
link signal, amplifies, and then reformats and transmits the signal back to the ground
received station, and
2) the on-board processing (OBP) satellite, which utilizes on-board detection
and remodulation to provide two independent communications links (uplink and
downlink). Because of that satellite link is assessed in two ways depending on the
type of transponder.

3.1. Frequency Translation (FT) Satellite

Figure 8 shows an example of the conventional frequency translation trans-
ponder. In the direct transponder, the uplink frequency is converted to the downlink
frequency, and after one or more stages of amplification, re-transmitted to the ground.
Signal degradations and noise introduced on the uplink are translated to the downlink,
and the total performance of the system will be dependent on both links.

13
 1 1
2 IFA
2
up 3 3 down
converter 4 4 converter

multiplexer
demultiplexer
Мix 5 IFA 5 Мix
6 6

……...
reseiver
fif 7 7

Y
fup 8 8 fdwn PA
9 9
10 10
11 11
LO LO
12 12

NO C P
IFA

CO Y
Figure 8. An example of conventional frequency translation transponder

O
3.1.1. Uplink

OP
To calculate uplink it is starting at the transmit terminal (point 1, Figure 7).

PY
The eirp of the ground transmit terminal is:
eirp = pGT .g GT , W , (38)
D O NO T C
The carrier power received at the satellite antenna (point 2) is
pGT .g GT .g SR
cSR = , (39)
lUfs .lUad
where lUfs is the uplink free space path loss, a lUad is the uplink path loss, and gGT , g SR
are the transmit and receive antenna gains, respectively.
The noise power at the satellite antenna, point (2), is
1
DO NO

nSR = k .tU (1 − ).bU + k .t SA .bU + k .290.(nf NSR − 1).bU , (40)

T
lUad
where k is Boltzmann’s constant, bU is the uplink information bandwidth, t SA is the
satellite receiver antenna temperature, nf NSR is the satellite receiver noise figure, and
tU is the mean temperature of the uplink atmospheric path.
T
The uplink carrier-to-noise ratio, at point (2), is given:
⎛c⎞ c pGT .g GT .g SR
⎜ ⎟ = SR = . (41)
⎝ n ⎠U nSR ⎡ 1 ⎤
DO

lUfs .lUad .k .⎢tU (1 − ) + t SA + 290.(nf NSR − 1)⎥.bU

⎣ lUad ⎦
This result (41) gives the uplink carrier-to-noise ratio.

3.1.2. Downlink
The procedure is the same that was used for the uplink.
The carrier power received at the received terminal (point 4) is
pST .g ST .g GR
cGR = , (42)
lDfs .lDad
The noise power at the satellite antenna, (point 4), is
1
nGR = k .t D (1 − ).bD + k .tGA .bD + k .290.(nf NGR − 1).bD . (43)
lDad
The downlink carrier-to-noise ratio, at point (4), is:

14
 ⎛c⎞ c pST .g ST .g GR
⎜ ⎟ = GR = . (44)
⎝ n ⎠ D nGR ⎡ 1 ⎤
lDfs .lDad .k .⎢t D (1 − ) + tGA + 290.(nf NGR − 1)⎥.bD
⎣ lDad ⎦
This result (44) gives the downlink carrier-to-noise ratio.

Y
3.1.3. Determination the carrier-to-noise ratio of total link
Consideration should be given to two important conditions specific to the FT

NO C P
satellite:

CO Y
- the downlink transmits power, pST , for a frequency translation satellite will
contain both the desired carrier component, cST , and noise introduced by the uplink

O
and by the satellite system itself, nST :

OP
PY
pST = cST + nST ; (45)
- since there is no on-board processing of the information signal, the satellite
D O NO T C
input carrier-to-noise ratio must equal the satellite output carrier-to-noise ratio:
⎛c⎞ ⎛c⎞
⎜ ⎟ =⎜ ⎟ , (46)
⎝ n ⎠ IN ⎝ n ⎠OUT
that is to say that all noise introduced by the satellite system is accounted for by nf SR .
So, in terms of the link parameters
⎛c⎞ c c
⎜ ⎟ = ST = SR (47)
⎝ n ⎠U nST nSR
DO NO

Replacing nST in Equation (47) with the Equation (45) condition,

T
⎛c⎞ c cST 1
⎜ ⎟ = ST = = ,
⎝ n ⎠U nST pST − cST pST
−1
cST
T
consequently
1 p
= ST − 1 . (48)
⎛c⎞ cST
⎜ ⎟
DO

⎝ n ⎠U
From (48) solving for cST , and next replacing cST in Equation (47) with the
Equation (45) condition:
pST ⎛c⎞ c p − nST
cST = , ⎜ ⎟ = ST = ST (49)
1 ⎝ n ⎠U nST nST
1+
⎛c⎞
⎜ ⎟
⎝ n ⎠U
The (49) is transforming and solving for nST ,
⎛c⎞ p pST
⎜ ⎟ = ST − 1 , nST = . (50)
⎝ n ⎠U nST ⎛c⎞
1+ ⎜ ⎟
⎝ n ⎠U
Next should calculate the necessary carrier power, which must be acceptably
the ground station (point 4) for overall link.
The desired carrier power received at the satellite antenna (point 4) is

15
 cST .g ST .g GR
c'GR = , (51)
lDfs .lDad
Replacing cST from (49) in (51) the result will be

Y
pST .g ST .g GR
c'GR = , (52)
1
1+ lDfs .lDad
⎛c⎞
⎜ ⎟

NO C P
⎝ n ⎠U

CO Y
consequently
cGR
c'GR = , (53)

O
1
1+

OP
⎛c⎞

PY
⎜ ⎟
⎝ n ⎠U
The total noise power received on the ground, n'GR , will be the sum of the noise
D O NO T C
introduced on the downlink, Equation (43), and the noise transferred from the uplink,
Equation (50), and after some transformation
cGR
n'GR = nGR + . (54)
⎛c⎞
1+ ⎜ ⎟
⎝ n ⎠U
The total carrier-to-noise ratio for the frequency translation transponder is then
found as the ratio of
DO NO
T
⎛c⎞ c'
⎜ ⎟ = GR . (55)
⎝ n ⎠ Σ n'GR
Replacing n'GR from (54) and c'GR from (53) in (55) can get the total carrier-to-
noise ratio
T
cGR
1
1+
⎛c⎞
⎜ ⎟
DO

⎛c⎞ ⎝ n ⎠U
⎜ ⎟ = (56)
⎝ n ⎠Σ n + cGR
⎛c⎞
GR
1+ ⎜ ⎟
⎝ n ⎠U
cGR ⎛ c ⎞
After transforming (56) and allude that = ⎜ ⎟ result for the FT satellite is:
nGR ⎝ n ⎠ D
⎛c⎞ ⎛c⎞
FD ⎜ ⎟ .⎜ ⎟
⎛c⎞ ⎝ n ⎠U ⎝ n ⎠ D
⎜ ⎟ = (57)
⎝ n ⎠Σ ⎛c⎞ ⎛c⎞
1+ ⎜ ⎟ + ⎜ ⎟
⎝ n ⎠U ⎝ n ⎠ D

If both ⎛⎜ ⎞⎟ >> 1 and ⎛⎜ ⎞⎟ >> 1 , equation (57) reduces to

c c
⎝ n ⎠D ⎝ n ⎠U
−1 −1 −1
⎡⎛ c ⎞ FD ⎤ ⎡⎛ c ⎞ ⎤ ⎡⎛ c ⎞ ⎤
⎢⎜ ⎟ ⎥ ≈ ⎢⎜ ⎟ ⎥ + ⎢⎜ ⎟ ⎥ (58)
⎣⎢⎝ n ⎠Σ ⎦⎥ ⎣⎝ n ⎠U ⎦ ⎣⎝ n ⎠ D ⎦

16
 This result is usually acceptable for satellite link analysis, because ⎛⎜ ⎞⎟ and
c
⎝ n ⎠D
⎛c⎞
⎜ ⎟ are generally much greater than 1.
⎝ n ⎠U

Y
3.1.4. Determination the carrier-to-noise density ratio of total link
The total carrier-to-noise density, (c n0 ) , can easily be shown to be of the same

NO C P
form as Equation (57)

CO Y
⎛c ⎞ ⎛c ⎞
FD ⎜⎜ ⎟⎟ .⎜⎜ ⎟⎟
⎛c ⎞
= ⎝ ⎠U ⎝ ⎠ D ,
n0 n0

O
⎜⎜ ⎟⎟ (59)
⎛c ⎞ ⎛c ⎞

OP
⎝ n0 ⎠ Σ 1 + ⎜⎜ ⎟⎟ + ⎜⎜ ⎟⎟

PY
⎝ n0 ⎠U ⎝ n0 ⎠ D
−1 −1 −1
⎡⎛ c ⎞ FD ⎤ ⎡⎛ c ⎞ ⎤ ⎡⎛ c ⎞ ⎤
D O NO T C
⎢⎜⎜ ⎟⎟ ⎥ ≈ ⎢⎜⎜ ⎟⎟ ⎥ + ⎢⎜⎜ ⎟⎟ ⎥ , (60)
⎢⎣⎝ n0 ⎠ Σ ⎥⎦ ⎣⎢⎝ n0 ⎠U ⎦⎥ ⎣⎢⎝ n0 ⎠ D ⎥⎦
where
⎛c ⎞ pGT .g GT .g SR
⎜⎜ ⎟⎟ = ,
⎝ n0 ⎠U l .l .k .⎡t (1 − 1 ) + t + 290.(nf − 1)⎤
Ufs Uad ⎢U SA NSR ⎥
⎣ lUad ⎦
⎛c ⎞
DO NO

pST .g ST .g GR
⎜⎜ ⎟⎟ = .
⎝ n0 ⎠ D l .l .k .⎡t (1 − 1 ) + t + 290.(nf ⎤
T
Dfs Dad ⎢D GA NGR − 1) ⎥
⎣ lDad ⎦

3.1.5. Determination the energy-per-bit to noise density

T
The ratio eb n0 is related to c n0 :
⎛ eb ⎞ ⎛ c ⎞ 1 ⎛ c ⎞
⎜⎜ ⎟⎟ = ⎜⎜ ⎟⎟ . = ⎜⎜ ⎟⎟.Tb , (61)
⎝ n0 ⎠ ⎝ n0 ⎠ rb ⎝ n0 ⎠
DO

Inserting Equation (61) into Equation (59) the energy-per-bit to noise density
can be found:
⎛ eb ⎞ ⎛ eb ⎞
FD ⎜⎜ ⎟⎟ .⎜⎜ ⎟⎟
⎛ eb ⎞
= ⎝ ⎠U ⎝ ⎠ D ,
n0 n0
⎜⎜ ⎟⎟ (62)
⎝ n0 ⎠Σ ⎛e ⎞ ⎛e ⎞
1 + ⎜⎜ b ⎟⎟ + ⎜⎜ b ⎟⎟
⎝ n0 ⎠U ⎝ n0 ⎠ D
−1 −1 −1
⎡⎛ e ⎞ FD ⎤ ⎡⎛ e ⎞ ⎤ ⎡⎛ e ⎞ ⎤
⎢⎜⎜ b ⎟⎟ ⎥ ≈ ⎢⎜⎜ b ⎟⎟ ⎥ + ⎢⎜⎜ b ⎟⎟ ⎥ . (63)
⎢⎣⎝ n0 ⎠Σ ⎥⎦ ⎢⎣⎝ n0 ⎠U ⎥⎦ ⎢⎣⎝ n0 ⎠ D ⎥⎦
Equations (62) and (63) give the probability of error for the overall end-to-end
digital link.

3.2. The on-board processing (OBP) satellite

The on-board processing satellite utilizes on-board detection and remodulation
to provide two independent communications links (uplink and downlink). The infor-
17
 mation signal on the uplink at a carrier frequency, fU , after passing through a low
noise receiver, is demodulated, and the baseband signal, at f BB , is amplified and en-
hanced by one or more signal processing techniques. The processed baseband signal
is then re-modulated on the downlink, at the carrier frequency, f D , for transmission to

Y
the downlink ground terminal. Figure 9 shows an example of the on-board processing
satellite.

NO C P
BASEBAND

CO Y
De-
receiver PROCESSOR Modulator
modulator
fup Amplify
fdwnPA
fbaseband Error correction fbaseband

O
Coding

OP
LO Switching
LO

PY
...

Фиг. 9. The on-board processing transponder.

D O NO T C
Degradations on the uplink can be compensated for by the on-board processing,
and are not transferred to the downlink, consequently, satellites employing on-board
processing essentially separate the uplink and downlink. Noise induced on the uplink
does not degrade the downlink because the waveform is reduced to baseband and re-
generated for downlink transmission.
DO NO

3.2.1 Uplink and Downlink for OBP satellite

The downlink carrier-to-noise ratio, (c n )D , or energy-per-bit to noise density,
T
(eb n0 )D , for an onboard processing satellite system is essentially independent of the
uplink carrier-to-noise ratio. Therefore the link equations (41) and (44) for (c n )U and
(c n )D previously developed for the frequency translation transponder, are applicable
T
to the OBP satellite uplink and downlink.
OBP
⎛c⎞ pGT .g GT .g SR
⎜ ⎟ = . (64)
⎡ ⎤
DO

⎝ n ⎠U 1
lUfs .lUad .k .⎢tU (1 − ) + t SA + 290.(nf NSR − 1)⎥.bU
⎣ lUad ⎦
OBP
⎛c⎞ pST .g ST .g GR
⎜ ⎟ = . (65)
⎝ n ⎠D ⎡ 1 ⎤
lDfs .lDad .k .⎢t D (1 − ) + tGA + 290.(nf NGR − 1)⎥.bD
⎣ lDad ⎦
Similarly, it can estimate satellite link by the ratio energy-per-bit to noise den-
sity.
OBP
⎛ eb ⎞ 1 pGT .g GT .g SR
⎜⎜ ⎟⎟ = . (66)
⎝ n0 ⎠U rU ⎡ 1 ⎤
lUfs .lUad .k .⎢tU (1 − ) + t SA + 290.(nf NSR − 1)⎥
⎣ lUad ⎦
OBP
⎛e ⎞ 1 pST .g ST .g GR
⎜⎜ b ⎟⎟ = . (67)
⎝ n0 ⎠ D rD ⎡ 1 ⎤
lDfs .lDad .k .⎢t D (1 − ) + tGA + 290.(nf NGR − 1)⎥
⎣ lDad ⎦
where rU and rD are the uplink and downlink data rates, respectively.

18
 3.2.2 Composite OBP Performance
The overall composite link performance for the OBP satellite is described by
its bit error performance, or the probability of error, PE . The overall error perform-

Y
ance of the on-board processing transponder will depend on both the uplink and
downlink error probabilities. A bit will be correct in the total link if either the bit is
correct on both the uplink and downlink, or if it is in error on both links. The overall
probability that a bit is correct, PCOR , is therefore 3

NO C P
CO Y
PCOR = (1 − PU ).(1 − PD ) + PU .PD = 1 − ( PU + PD ) + 2 PU .PD , (68)
where PU is the probability of a bit error on the uplink, PD is the probability of a

O
bit error on the downlink; (1 − PU ) is the probability of correct bit on uplink, (1 − PD ) is

OP
PY
the probability of correct bit on downlink.
The probability of a bit error on the total link is
D O NO T C
PEOBP = 1 − PCOR = PU + PD − 2 PU .PD . (69)
The composite link probability of error will be dependent on the uplink and
downlink parameters and their impact on the (eb n0 ) for each link. A specific modula-
tion must be specified to determine the relationship between the bit error probability
and the (eb n0 ) for each link. For example, the procedure for the determination of the
composite error performance for an on-board processing transponder system will be
demonstrated for a binary frequency-shift keying (BFSK) system with noncoherent
DO NO

detection. The bit error probability is given by Equation (70) 4

T
1 ⎛ Eb ⎞
1 − 2 ⎜⎜ N ⎟⎟
PE = e ⎝ 0 ⎠ (70)
2
Equations (69) and (70) give the bit error probability for the total link
T
1 ⎛ Eb ⎞ 1 ⎛ Eb ⎞ 1 ⎡ ⎛ Eb ⎞ ⎛ Eb ⎞ ⎤
1 − 2 ⎜⎜ N ⎟⎟ 1 − 2 ⎜⎜ N ⎟⎟ 1 − 2 ⎢⎢⎜⎜ N ⎟⎟ +⎜⎜ N ⎟⎟ ⎥⎥
PE = e ⎝ 0 ⎠U + e ⎝ 0 ⎠D − e ⎣⎝ 0 ⎠U ⎝ 0 ⎠D ⎦ (71)
2 2 2
⎛e ⎞ ⎛ c ⎞ 1 ⎛c⎞ ⎛ c⎞
DO

The ratio eb n0 is related to c n0 by (61): ⎜⎜ b ⎟⎟ = ⎜⎜ ⎟⎟ . = ⎜⎜ ⎟⎟.Tb , and ⎜⎜ ⎟⎟ can

⎝ n0 ⎠ ⎝ n0 ⎠ rb ⎝ n0 ⎠ ⎝ n0 ⎠
c eirp.g R
be expressed as = , therefore,
n0 l fs .lad .k .t
⎧ 1 ⎛⎜ EIRPU .⎛⎜⎝ g t ⎞⎟ ⎞
⎟
⎠U ⎟ 1 ⎜
⎛ EIRP .⎛ g ⎞ ⎞
⎜ U ⎜ t ⎟ ⎟
⎝ ⎠D ⎟ 1 ⎢
⎡ EIRP .⎛ g ⎞
U ⎜ t ⎟
⎝ ⎠U
EIRPU .⎛⎜ g ⎞⎟
⎝ t ⎠D
⎤
⎥ ⎫
⎪ − 2.k .rU ⎜⎜ l fs .lad ⎟
−
2.k .rD ⎜ ⎟
− ⎢ + ⎥ ⎪
⎛ eb ⎞ ⎪ ⎜ ⎟ ⎜
l fs .l ad
⎟
2 k ⎢ rU .l fs . Lad
⎢⎣
rD .l fs . Lad ⎥
⎥⎦ ⎪
⎜⎜ ⎟⎟ = −2. ln ⎨e ⎝ ⎠U
+e ⎝ ⎠D
−e ⎬ (72)
⎝ n0 ⎠Σ ⎪ ⎪
⎪ ⎪
⎩ ⎭

The result (72) is available for the OBP satellite with the binary frequency-shift
keying system.

3
Ippolito, Louis J., Satellite communications systems engineering: atmospheric effects, satellite link design, and system
Performance, 2008 JohnWiley & Sons Ltd, 396 p.
4
B. Sklar, Digital Communications – Fundamentals and Applications, Prentice Hall, Englewood Cliffs, NJ, 1988.

19
 References
1. Bernard Sklar, Digital Communications: Fundamentals and Applications, Second Edition,
Prentice-Hall, 2001.
2. Broadcom Corporation and Cisco Systems, Digital Transmission: Carrier-to-Noise Ratio,

Y
Signal-to-Noise Ratio, and Modulation Error Ratio.
http://www.broadcom.com/docs/general/Broadcom-Cisco_CNR-SNR-MER.pdf
3. Dennis Roddy, Satellite communications, McGraw-Hill Professional, 2001 - 569 p.
4. Intuitive Guide to Principles of Communications, Link Budgets, www.complextoreal.com

NO C P
CO Y
5. Ippolito, Louis J., Satellite communications systems engineering: atmospheric effects,
satellite link design, and system Performance, 2008 JohnWiley & Sons Ltd, 396 p.
6. Mohamed Khedr, Lecture: Satellite Communication Systems,

O
http://webmail.aast.edu/~khedr

OP
7. RPC Telecommunications Ltd., http://www.satcom.co.uk/

PY
8. Silver J. P., Satellite Communications Tutorial, E-mail: [email protected]
D O NO T C
DO NO
T
T
DO

20
 Satellite Communications Systems

Lecture 3.
Forward Error Correcting in Satellite Communications

Y
1. The necessity of channel coding

NO C P
2. Types of error control

CO Y
3. Concept of Hamming Weight and Hamming Distance
4. Reed Solomon code
4.1. Reed Solomon coding

O
OP
4.2. Reed Solomon decoding

PY
5. Interleaving
6. Convolution code
D O NO T C
6.1. Coder and coding.
6.2. State representation and state diagram
6.3. The Viterbi Decoding Algorithm
6.4. Implementation of the Viterbi Decoder
6.5. Recursive Systematic Convolutional Encoder
DO NO
T
T
DO

1
 1. The necessity of channel coding.
The designation of Source Coding is to get a multiplexing and compressing in-
formation. The purpose of Channel Coding is decreasing of errors when signals are
transmitted by noisy channel. The capabilities of Channel Coding were investigated
by Claude Elwood Shannon (30.04.1916–24.02.2001) and written in “A mathemati-

Y
cal theory of communication” 1 .
A measure the quality of digital signals is Bit Error Rate (BER). The bit error
rate or bit error ratio is the number of bit errors divided by the total number of trans-

NO C P
CO Y
ferred bits during a studied time interval. BER is a unitless performance measure, of-
ten expressed as a percentage number. The BER is often expressed as a function of
the Eb / N 0 , (energy per bit to noise power spectral density ratio). Figure 1 is shows a

O
OP
comparison of typical coded versus uncoded error performance. If it is needed to pro-

PY
Eb
vide BER = 1.10−5 , the uncoded signal needs a ratio = 9.4 dB , but coded 1 / 7 - only
N0
D O NO T C
Eb
= 4.6 dB . The deference between values of ratio Eb N 0 coded and uncoded signal
N0
is called Coding Gain (G).
⎛E ⎞ ⎛E ⎞
G = ⎜⎜ b ⎟⎟ − ⎜⎜ b ⎟⎟ , dB (1)
⎝ N 0 ⎠U ⎝ N 0 ⎠C
DO NO

BER
T
1.10-1

1.10-2
un
co
co

de
co

1.10-3
de

d
de

T
d1
d1

/7
/5

1.10-4

gain
DO

1.10-5

1.10-6
1 2 3 4 5 6 7 8 9 10 Eb/N0

Figure 1. Comparison of typical ceded versus uncoded error performance.

2. Types of error control.

There are two basic methods for error control:
- Error detection and retransmission utilizes parity bits to detect an error (parity
bits are redundant bits added to the data);
- Forward Error Correction (FEC) - by parity bits it can detect and correct the
errors.
By first method the receiver detect that an error has been made, but does not at-
tempt to correct the error, it requests the transmitter retransmits the information. The
first method required two-way link for dialog between the transmitter and receiver.

1
C. E. Shannon, A mathematical theory of communication, Bell Syst. Tech., 1948.

2
 That technique is known as Automatic Repeat Request (ARQ). It can be realize by
three ways:
- stop and wait ARQ - the transmitter waits for acknowledgement of each
transmission before the next transmission; if some of the message is received with an
error, the receiver responds with negative acknowledgement and transmitter repeats

Y
this message;
- continuous ARQ with pullback - both the transmitter and receiver work si-
multaneously, the transmitter is sending messages and receiver is sending acknowl-

NO C P
CO Y
edgement data; if some of the message is received with an error, the receiver re-
sponds with negative acknowledgement and transmitter retransmits all messages
started with the corrupted message to the moment of the negative acknowledgement;

O
OP
- continuous ARQ with selective repeat - this method is the same as the first,

PY
but the transmitter retransmits only the corrupted message.
Which ARQ procedure to choice is a trade-off between the requirements for
D O NO T C
the efficient of the communication resources.
Therefore, there are developed methods for error correction by encoding the
digital signal (ERC error correction coding). ERC has already been used in all types
of satellite links. General of the various methods of error correction is the addition of
information surplus, called a control feature (additional information) to the basic flow
of data through which some errors can be eliminated during signal reception. The re-
ceiving system detect and possibly correct errors caused by corruption from the
DO NO

channel. This is called Forward Error Correcting (FEC). The purpose of FEC is to
T
improve the capacity of a transmission channel by adding to the source data redun-
dant information. It is a process known as channel coding. Channel coding consists:
Reed Solomon coding;
Interleaving;
T
Convolution code.
Reed Solomon codes operate on multi-bit symbols rather than individual bits
like binary codes. They are a “block” coding technique requiring the addition of re-
DO

dundant parity symbols to the data to enable error correction. The block code oper-
ates on a block of bits. Each block is processed as a single unit by both the encoder
and decoder. Using a preset algorithm, to a group of bits and it add a coded part to
make a larger block. This block is checked at the receiver. The receiver then makes a
decision about the validity of the received sequence.
Convolutional codes are referred to as continuous codes as they operate on a
certain number of bits continuously.
Interleaving has mitigating properties for fading channels and works well in
conjunction with these two types of coding. Standard interleavers scramble code bits
among multiple blocks so that they are not contiguous when transmitted; as a result,
any bursty errors caused by channel corruption are spread out, into more-random er-
rors after deinterleaving.
For years, convolutional coding with Viterbi decoding has been the predomi-
nant FEC technique used in space communications, particularly in geostationary sat-
ellite communication networks, such as VSAT (very small aperture terminal) net-
works. Then, convolutional coding with Viterbi decoding has begun to be supple-
3
 mented in the satellite communications with Reed-Solomon coding as serially con-
catenated block and convolutional coding. Typically, the information to be transmit-
ted is first encoded with the Reed-Solomon code, then with the convolutional code.
On the receiving end, Viterbi decoding is performed first, followed by Reed-Solomon
decoding. This is the technique that is used in most of the direct-broadcast satellite

Y
systems and in several of the newer VSAT products.
In 1993, Claude Berrou developed the turbo code, and now it is the most pow-
erful forward error-correction code. Communication systems with a turbo code can

NO C P
CO Y
approach the theoretical limit of channel capacity, called Shannon Limit. The turbo
coder-decoder configuration must include some arrangement of at least two compo-
nent encoders that are separated by an interleaver. The interleaver in a turbo encoder

O
OP
serves a different purpose than interleavers used by other parts of a communication

PY
system. The interleaver in a turbo encoder is designed that the second encoder gets an
interleaved version of the same data block that went into the first encoder; thus, the
D O NO T C
second encoder generates an independent set of code bits.
The place of channel coding is between multiplexing and modulating process,
as shown in Figure 2.

video
audio TV coder

M
DO NO

video U f
audio TV coder
T
L
T PA
I
..................

P
RS Convolution QPSK
L interleaving
coder coding modulater
E
X Channel coding
T
E
R
video
audio TV coder
DO

MPEG
coder

Figure 2. Transmitter digital sat TV.

3. Concept of Hamming Weight and Hamming Distance

The Hamming weight is the largest number of ones in a valid codeword.
Ordinarily, distance is measured by Euclidean concepts such as lengths, angles
and vectors. In the digital world, distances are measured between two binary words
by the Hamming distance. The Hamming distance is the number of disagreements
between two binary sequences of the same size. It is a measure of how apart binary
objects are.
Example 1: The Hamming distance between sequences A = 001 and B = 101 is d ( A, B) = 1 .
The Hamming distance between sequences A = 00101101 and B = 10110101 is d ( A, B) = 3 . The
Hamming weight of a sequence A = 00101101 is 4; Hamming weight of a sequence B = 10110101 is
5.
The Hamming distance can be search not only between two binary sequences,
but also between all binary words in the same code space. The comparison will give
4
 different Hamming distances. The minimum number of symbols that differs the
words in the same code space is called the minimum Hamming distance. The number
of detected errors depending on the minimum Hamming distance.
t = d min − 1 , d min ≥ t + 1 , (2)

Y
i.e. to detect one error the code has to contain the minimum Hamming distance d = 2 ,
to detect two errors - d = 3 , …
To correct the errors codes have to get more parity bits or symbols. For exam-

NO C P
ple, for a sequence of 6 bits, there are 6 different ways to get a 1 bit error and there

CO Y
6!
are 15 different ways we can get 2-bit errors ( = 15 ). The number of errors that
2!.4!

O
can be correct is:

OP
d min − 1

PY
t= , d min ≥ 2t + 1 , (3)
2
D O NO T C
i.e. to correct one error the code has to contain the minimum Hamming distance d = 3 .

4. Reed Solomon code

In 1960, Irving Reed and Gus Solomon published a paper in the Journal of the
Society for Industrial and Applied Mathematics. This paper described a new class of
error-correcting codes that are now called Reed-Solomon (R-S) codes. Reed Solomon
codes are linear block nonbinary cyclic codes with symbols made up of m-bit se-
quences, where m is any positive integer having a value greater than 2. R-S (n, k)
DO NO

codes on m-bit symbols exist for all n and k for which

T
k < n < 2m + 2 , (4)
where k is the number of data symbols being encoded, and n is the total number of
code symbols in the encoded block, n − k = 2t is the number of parity symbols.
T
For Reed-Solomon codes, the code minimum distance is given by 2
dmin = n - k + 1. (5)
If combine Equation (3) and (5) can be express the number of errors that can
DO

be correct
d min − 1 n − k
t= = , (6)
2 2
n−k n−k
where is the largest integer not to exceed ; a Reed–Solomon code can cor-
2 2
rect up to half as many errors as there are redundant symbols added to the block.

2t = 16 parity
k=188 information symbols
symbols
Block of n=204 symbols

Figure 3. Block of RS (204,188) code - 204 symbols

Reed Solomon codes have information surplus about 8% and BER = 10−4 ÷ 10−11 .
For example, RS (204,188) code for DVB systems uses 8 bits symbols, and blocks

2
Gallager, R. G., Information Theory and Reliable Communication, New York: John Wiley and Sons, 1968.

5
 have a length 188 information symbols and 16 parity symbols, as it is shown in Fig-
ure 3.
Reed Solomon codes are based on finite fields, called Galois fields (GF)
with 2 m elements. Every finite field can be representing by a fixed length binary
word. For any number p there exists a finite field GF ( p) that contains p elements. It

Y
can extend to a field of p m elements, where m is a nonzero positive integer. The new
field is called an extension field of GF ( p) , and denoted by GF ( p m ) . GF ( p m ) contains

NO C P
as a subset the elements of GF ( p) . Symbols from the extension field GF ( p m ) are used

CO Y
in the construction of Reed-Solomon (R-S) codes.
The binary field GF ( p) is a subfield of the extension field GF ( p m ) . Besides the

O
numbers 0 and 1, there are additional unique elements in the extension field that will

OP
PY
be represented with a new symbol α . Each nonzero element in GF ( p m ) can be repre-
sented by a power of α . An infinite set of elements, F, is formed by starting with the
elements {0 ,1, α } , and generating additional elements by progressively multiplying
D O NO T C
the last entry by α :
F = {0,1, α , α 2 , α 3 ,.., α j ,...} = {0, α 0 , α 1 , α 2 , α 3 ,.., α j ,...}. (7)
The condition that closes the set of field elements under multiplication is char-
acterized by the irreducible polynomial shown below:
m
α (2 −1)
+1 = 0
DO NO

or
T
m
α (2 −1)
=1= α0. (8)
Using this polynomial constraint, any field element that has a power equal to or
greater than 2m − 1 can be reduced to an element with a power less than 2m − 1 :
T
m m
α (2 +n )
= α (2 −1)
.α ( n+1) = α ( n+1) . (9)
Thus, Equation (8) can be used to form the finite sequence F * from the infinite
sequence F
{ } { }
DO

m m m m
F * = 0,1, α , α 2 , α 3 ,.., α 2 −2
, α 2 −1 , α 2 ,... = 0, α 0 , α 1 , α 2 , α 3 ,.., α 2 −2
, α 0 , α 1 , α 2 ,... (10)
From Equation (10) the elements of the finite field GF(2m) are
{
GF (2 m ) = 0, α 0 , α 1 , α 2 , α 3 ,.., α 2
m
−2
}. (11)
Each of the 2 m elements of the finite field, GF ( p m ) , can be represented as a
distinct polynomial of degree (m - 1) or less where at least one of the m coefficients
of ai ( X ) is nonzero. The degree of a polynomial is the value of its highest-order ex-
ponent.
α i = ai ( X ) = ai ,0 + ai ,1 X + ai , 2 X 2 + ... + ai ,m−1 X m−1 , (12)
where i = 0,1, 2, ..., 2m − 2 .

Example 2: Table 1 shows the possible nine elements of field GF (23 ) , m = 3 , where every
row contain values of coefficients ai ,0 , ai ,1 и ai , 2 of Equation (12). Seven elements {ai } are differ
m 3
each other and zero α 0 , α 1 , α 2 , α 3 , α 4 , α 5 , α 6 , because from (8) α ( 2 −1)
= α ( 2 −1) = α 0 .

6
 Table 1. Mapping field elements in terms of basis elements for GF(8) with 1 + X + X 3
Basic elements
0
X X1 X2
0

Y
0 0 0
α0 1 0 0

Elements of the field

α1 0 1 0
α2 0 0 1

NO C P
CO Y
α3 1 1 0
α4 0 1 1

O
α5 1 1 1

OP
α6 1 0 1

PY
α 7 =α 0 1 0 0
D O NO T C
One of the benefits of using extension field elements {ai } in place of binary
elements is the compact notation that facilitates the mathematical representation of
nonbinary encoding and decoding processes. Addition of two elements of the finite
field is then defined as the modulo-2 sum of each of the polynomial coefficients of
equal powers
α i + α j = (ai ,0 + a j , 0 ) + (ai ,1 + a j ,1 ) X + ... + (ai ,m−1 + a j ,m−1 ) X m−1 (13)
DO NO

Table 2 shows some of primitive polynomials. A class of polynomials called

T
primitive polynomials define the finite fields GF ( p m ) .

Table 2. Some of primitive polynomials

m m
T
3 1+ X + X 3 14 1 + X + X 6 + X 10 + X 14
4 1+ X + X 4
15 1 + X + X 15
5 1+ X 2 + X 5 16 1 + X + X 3 + X 12 + X 16
DO

6 1+ X + X 6 17 1 + X 3 + X 17
7 1+ X 3 + X 7 18 1 + X 7 + X 18
8 1+ X 2 + X 3 + X 4 + X 8 19 1 + X + X 2 + X 5 + X 19
9 1+ X 4 + X 9 20 1 + X 3 + X 20
10 1 + X 3 + X 10 21 1 + X 2 + X 21
11 1 + X 2 + X 11 22 1 + X + X 22
12 1 + X + X 4 + X 6 + X 12 23 1 + X 5 + X 23
13 1 + X + X 3 + X 4 + X 13 24 1 + X + X 2 + X 7 + X 24

First of them f ( X ) = 1 + X + X 3 defines finite field GF (23 ) . There are 23 = 8 ele-

ments in the field defined by f ( X ) . Now the roots of f ( X ) = 0 must be found. The
familiar binary elements, 1 and 0, do not satisfy (are not roots of) the polynomial
f ( X ) = 1 + X + X 3 , since f (1) = 1 and f (0) = 0 (using modulo-2). A fundamental theo-
rem of algebra states that a polynomial of degree m must have precisely m roots,

7
 consequently f ( X ) = 0 has three roots, and if α is a root of the polynomial f ( X ) it is
possible to write the following:
1 + α + α 3 = 0 , α 3 = −1 − α = 1 + α , (14)
because of − 1 = +1 in the binary field.

Y
Thus, α 3 is expressed as a weighted sum of α − terms having lower orders. In
fact all powers of α can be so expressed.
α 4 = α .α 3 = α .(1 + α ) = α + α 2 ,

NO C P
CO Y
α 5 = α .α 4 = α 2 + α 3 = 1 + α + α 2
α 6 = α .α 5 = α (1 + α + α 2 ) = α + α 2 + α 3 = 1 + α 2

O
α 7 = α .α 6 = α (1 + α 2 ) = α + α 3 = 1 = α 0 . (15)

OP
PY
Therefore the eight finite field elements of GF (23 ) are
{0,α 0
, α 1 , α 2 , α 3 , α 4 , α 5 , α 6 }. (16)
D O NO T C
4.1. Reed Solomon coding
The generating polynomial for an R-S code in terms of the parameters n, k, t,
and any positive integer m > 2 is
g ( X ) = g 0 + g1 X + g 2 X 2 + ... + a2t −1 X 2t −1 + X 2t . (17)
where n − k = 2t is the number of parity symbols, and t is the symbol-error correcting
DO NO

capability of the code.

T
Since the generator polynomial is of degree 2t, there must be precisely 2t suc-
cessive powers of α that are roots of the polynomial. We designate the roots of g ( X )
as α 1 , α 2 , α 3 ,...,α 2t . It is not necessary to start with the root α ; starting with any power
of α is possible.
T
For example, there is the (7, 3) R-S code, t = 2 . The generator polynomial in
terms of its 2t = n - k = 4 roots can be described, as follows:
g ( X ) = ( X − α ).( X − α 2 ).( X − α 3 ).( X − α 4 ) = X 4 − α 3 X 3 + α 0 X 2 − α 1 X + α 3
DO

g( X ) = α 3 + α1X + α 0 X 2 + α 3 X 3 + X 4 . (18)
R-S codes are cyclic codes and encoding in systematic form is analogous to the
binary encoding procedure. The information message (a polynomial, m( X ) ) can be
multiplied by X n−k so that it is right-shifted n − k positions and then appending a par-
ity polynomial, p( X ) , by placing it in the leftmost n − k stages. Then it can divide
X n − k .m( X ) by the generator polynomial g ( X ) , which is written in the following form:
m( X ).x n−k = q( X ).g ( X ) + p ( X ) , (19)
where q( X ) and p( X ) are quotient and remainder polynomials, respectively.
As in the binary case, the remainder p( X ) is the parity. Equation (19) can also
be expressed as follows:
p ( X ) = m( X ).x n − k ⊕ g ( X ) . (20)
The resulting codeword polynomial, U ( X ) can be written as
U ( X ) = p( X ) + m( X ). X n−k . (21)

8
 Example 3: There is a message that consist three symbols 010 110 111, and it has to be
coded by RS(7,3) code with the generator polynomial (18): g ( X ) = α 3 + α 1 X + α 0 X 2 + α 3 X 3 + X 4 .
First, according to table 1 these three symbols are α 1 , α 3 , α 5 . The information message

Y
polynomial is m( X ) = α 1 + α 3 X + α 5 X 2 , it has to multiply by X k = X 4 . The result is
X 4 .m( X ) = α 1 X 4 + α 3 X 5 + α 5 X 6 . Next, it has to divide this upshifted message polynomial by the
generator polynomial (18) to get a parity polynomial, p ( X ) = α 0 + α 2 X + α 4 X 2 + α 6 X 3 . Then,

NO C P
from Equation (21), the codeword polynomial is:

CO Y
U ( X ) = α 0 + α 2 X + α 4 X 2 + α 6 X 3 + α1X 4 + α 3 X 5 + α 5 X 6 .

O
The circuit to encode a symbol sequence in systematic form with the (n, k) R-S

OP
code with the generator polynomial requires the implementation of a linear feedback

PY
shift register (LFSR) circuit. The linear feedback shift register is a dynamic memory
of binary 0 and 1. The memory consist a number of cages that determines the mem-
D O NO T C
ory. When in the most left stage enters a new symbol all next shift to the right, and
this one that is the rightmost go out of the register.
Figure 4 shows a circuit of LFSR for RS (7,3) code from example 3.
switch 1

α3 α1 α0 α3
DO NO

Register Register Register Register

T
A Q1 A Q1 A Q1 A Q1
To output
symbols
D Q4
+ D Q4
+ D Q4
+ D Q4
+ register
ENB ENB ENB ENB

2
010 110 111(α , α , α )
1 3 5

input message symbol 1

T
sequence switch 2

Figure 4. A linear feedback shift register of RS (7,3) code

DO

The encoder forms codewords in the steps, as follows:

1) Switch 1 is closed during the first k clock cycles to allow shifting the mes-
sage symbols into the (n − k ) − stage shift register.
2) Switch 2 is in the position 1 during the first k clock cycles in order to allow
simultaneous transfer of the message symbols directly to an output register (not
shown in Figure 4).
3) After 3rd clock cycle all message symbols are transferred to the output reg-
ister, switch 1 is opened and switch 2 is moved to the position 2.
4) The remaining (n − k ) clock cycles the parity symbols contained in the shift
register are moved to the output register.
5) The total number of clock cycles is equal to n , and the contents of the out-
put register is the codeword polynomial U ( X ) = p( X ) + m( X ). X n−k , where p( X ) is the
parity polynomial and m( X ) the information message in polynomial form.

9
 During the transfer time the rightmost symbol of the information sequence is
the earliest symbol, and the rightmost bit is the earliest bit. Table 3 shows the opera-
tional steps during the first k = 3 shifts of the encoding circuit of Figure 4.

Table 3. The operational steps to shift symbols in the register

Y
входна после- такт съдържание на регистъра обратна връзка
дователност
α1 α 3 α 5 0 0 0 0 0 α5

NO C P
α1 α 3 1 α1 α6 α5 α1 α 0 (α 1 ⊕ α 3 )

CO Y
α1 2 α3 0 α2 α2 α 4 (α 2 ⊕ α 1 )
- α0 α2 α4 α6 -

O
3

OP
PY
Explanations:
α 5 .α 3 = α 8 = α 1.α 7 = α 1.1 = α 1 - first clock, first register;
D O NO T C
α 5 .α 1 = α 6 - first clock, second register;
α 5 .α 0 = α 5 .1 = α 5 - first clock, third register;
α 5 .α 3 = α 8 = α 1.α 7 = α 1.1 = α 1 - first clock, forth register;
α 0 .α 3 ⊕ α 5 = 1.α 3 ⊕ α 5 = α 2 (001) - second clock, first register;
α 0 .α 0 ⊕ α 6 = 1.1 ⊕ α 6 = 1 ⊕ 1 ⊕ α 2 = α 2 (001) - second clock, second register;
α 0 .α 1 ⊕ α 1 = 1.α 1 ⊕ α 1 = 0 (000) - second clock, third register;
α 0 .α 3 = 1.α 3 = α 3 (110) - second clock, forth register;
DO NO

.................................................................................................
T
Thus, after the third clock cycle, the register contents are the four parity sym-
bols, α 0 , α 2 , α 4 ,α 6 . Then, switch 1 of the circuit is opened, switch 2 is moved to the
position 2, and the parity symbols from the register are shifted to the output. The out-
T
put codeword, U ( X ) , is:
U ( X ) = α 0 + α 2 X + α 4 X 2 + α 6 X 3 + α1X 4 + α 3 X 5 + α 5 X 6 =
(22)
DO

= (100) + (001) X + (011) X 2 + (101) X 3 + (010) X 4 + (110) X 5 + (111) X 6

4.2. Reed Solomon decoding 3

The presence of noise in the communication channel causes errors in the in-
formation and the received message can be different from original message. For in-
stant, assume that during transmission the codeword (22) becomes corrupted so that
two symbols are received in error. Remember that the number of errors corresponds
to the maximum error-correcting capability of the code. The receiver will get a mes-
sage:
r ( X ) = U ( X ) + e( X ) , (23)
where e( X ) is a polynomial of error.
For this example of seven-symbol codeword, the error pattern, e( X ) , can be
described in polynomial form as:

3
Bernard Sklar, Digital Communications: Fundamentals and Applications, Second Edition, Prentice-Hall, 2001.

10
 6
e( X ) = ∑ en . X n . (24)
n =0

Assume that corrupted symbols are forth and fifth, polynomial of error is:
e( X ) = 0 + 0 X + 0 X 2 + α 2 X 3 + α 5 X 4 + 0 X 5 + 0 X 6
(25)

Y
= 000 + 000 X + 000 X 2 + 001X 3 + 111X 4 + 000 X 5 + 000 X 6
According to Equation (23) the received message will be (using modulo-2 on
Equations (22) and (25)):

NO C P
CO Y
r ( X ) = 100 + 001X + 011X 2 + 100 X 3 + 101X 4 + 110 X 5 + 111X 6 =
, (26)
= α 0 + α 2 X + α 4 X 2 + α 0 X 3 + α 6 X 4 + α 3 X 5+α 5 X 6

O
because of 101 ⊕ 001 = 100 = α 0 (forth symbol) and 010 ⊕ 111 = 101 = α 6 ( fifth symbol).

OP
In this example, there are four unknowns—two error locations and two error

PY
values. Therefore, four equations are required for their solution.
The decoding order is as follows:
D O NO T C
- Syndrome Computation;
- Error Location;
- Error Values Computation;
- Correcting the Received Polynomial with Estimates of the Error Polynomial.

1) Syndrome Computation
The syndrome is the result of a verification on r ( X ) to determine whether r ( X )
DO NO

is a valid member of the codeword set. The syndrome S has value 0 when r ( X ) is a
T
member. Any nonzero value of S indicates the presence of errors. The syndrome S is
made up of n − k symbols, {Si } (i = 1,..., n − k ) . Thus, for (7,3)RS code, there are four
symbols in every syndrome vector; their values can be computed from the received
T
polynomial, r ( X ) . It can be made by the structure of the code
U ( X ) = m( X ).g ( X ) . (27)
This structure shows that every valid codeword polynomial U ( X ) is a multiple
DO

of the generator polynomial g ( X ) . Consequently, the roots of g ( X ) must also be the

roots of U ( X ) . Thus, each of the roots of g ( X ) should yield zero r ( X ) (received mes-
sage) only when it is a valid codeword. Any errors will give a nonzero result in one
or more of the computations. The computation of a syndrome symbol can be de-
scribed as follows:
Si = r ( X ) X =α = r (α i ), i = 1, 2, ..., n − k ( i = 1,2,3,4 for example )
i (28)
In the example, r ( X ) contain two corrupted by error symbols (Equation 25).
The computation of the four syndrome symbols is:
S1 = r (α ) = α 0 + α 3 + α 6 + α 3 + α 10 + α 8 + α 11 =
(29)
= α 0 + α 3 + α 6 + α 3 +α 2 + α1 + α 4 = α 3
S 2 = r (α 2 ) = α 0 + α 4 + α 8 + α 6 + α 14 + α 13 + α 17 =
(30)
= α 0 + α 4 + α1 + α 6 + α 0 + α 6 + α 3 = α 5

11
 S3 = r (α 3 ) = α 0 + α 5 + α 10 + α 9 + α 18 + α 18 + α 23 =
(31)
=α0 +α5 +α3 +α2 +α4 +α4 +α2 =α6
S 4 = r (α 4 ) = α 0 + α 6 + α 12 + α 12 + α 22 + α 23 + α 29 =
(32)
= α 0 + α 6 + α 5 + α 5 + α1 + α 2 + α 2 = 0

Y
The result shows, that received codeword contains errors.

2) Error Location

NO C P
Suppose there are v errors in the codeword at location X J , X J ,..., X J . Then,

CO Y
1 2 v

the error polynomial e( X ) shown in Equations (24) and (25) can be written as fol-

O
lows:

OP
e( X ) = e j X J + e j X J + ... + e j X J .
1 2 v
(33)

PY
1 2 v

The indices 1,2,..., v refer to the first, second, …, v th errors, and the index j re-
D O NO T C
fers to the error location. To correct the corrupted codeword, each error value e jl and
its location X jl , ( l = 1,2,..., v ) must be determined. An error locator number can be
mark as β l = α l . Next, by substituting α i into the received polynomial for i = 1,2,...,2t
j

can be obtaining the n − k = 2t syndrome symbols:

S1 = r (α ) = e j1 β1 + e j2 β 2 + ... + e jv β v
S 2 = r (α 2 ) = e j1 β 12 + e j2 β 22 + ... + e jv β v2
DO NO
T
...
S 2t = r (α 2t ) = e j1 β 12t + e j2 β 22t + ... + e jv β v2t (34)
There are 2t unknowns ( t error values and t locations), and 2t simultaneous
equations. However, these 2t simultaneous equations cannot be solved in the usual
T
way because they are nonlinear (as some of the unknowns have exponents). Any
technique that solves this system of equations is known as a Reed-Solomon decoding
algorithm.
DO

If a nonzero syndrome vector (one or more of its symbols are nonzero) has
been computed, that signifies that an error has been received. Next, it is necessary to
learn the location of the error or errors. An error-locator polynomial, σ ( X ) , can be
defined as
σ ( X ) = (1 + β1 X )(1 + β 2 X )...(1 + β v X ) = 1 + σ 1 X + σ 2 X 2 + ... + σ v X v . (35)
1 1 1
The roots of σ ( X ) are , ,..., . The reciprocal values of the roots of σ ( X )
β1 β 2 βv
show the error-location numbers of the error polynomial e(X ) . Then, using autore-
gressive modelling techniques 4 [7], it can form a matrix from the syndromes, where
the first t syndromes are used to predict the next syndrome.

4
Blahut, R. E., Theory and Practice of Error Control Codes (Reading, MA: Addison-Wesley, 1983)

12
 ⎡ S1 S 2 S3 ... St −1 St ⎤ ⎡σ t ⎤ ⎡− St +1 ⎤
⎢S S
⎢ 2 3 S 4 ... St St +1 ⎥⎥ ⎢⎢σ t −1 ⎥⎥ ⎢⎢− St +2 ⎥⎥
⎢... ⎥ ⎢... ⎥ = ⎢... ⎥. (36)
⎢ ⎥⎢ ⎥ ⎢ ⎥
⎢ St −1 St St +1 ... S 2t −3 S 2t −2 ⎥ ⎢σ 2 ⎥ ⎢− S 2t −1 ⎥

Y
⎢ St St +1
⎣ St +2 ... S 2t −2 S 2t −1 ⎥⎦ ⎢⎣σ 1 ⎥⎦ ⎢⎣− S 2t ⎥⎦
For the (7, 3) RS code, the matrix size is 2 × 2, and the model is written as:

NO C P
⎡ S1 S 2 ⎤ ⎡σ 2 ⎤ ⎡ S3 ⎤
⎢ S S ⎥ ⎢σ ⎥ = ⎢ S ⎥ (37)

CO Y
⎣ 2 3 ⎦ ⎣ 1 ⎦ ⎣ 4⎦
⎡α 3 α 5 ⎤ ⎡σ 2 ⎤ ⎡α 6 ⎤

O
⎢ 5 6 ⎥⎢ ⎥ = ⎢ ⎥ (38)
⎥⎦ ⎣σ 1 ⎦ ⎣0 ⎦

OP
⎢⎣α α

PY
To solve for the coefficients σ 1 and σ 2 and of the error-locator polynomial,
σ ( X ) (35), first can take the inverse of the matrix in Equation (38). The inverse rule
D O NO T C
of a matrix [A] is:
cofactor [ A]
Inv [ A] = .
det [ A]
Consequently,
⎡α 3 α 5 ⎤
⎥ = α α −α α = α +α = α +α = α
3 6 5 5 9 10 2 3 5
det ⎢ 5 6
⎢⎣α α ⎥⎦
DO NO

⎡α 3 α 5 ⎤ ⎡α 6 α 5 ⎤
T
cofactor ⎢ 5 6 ⎥= ⎢ 5 3 ⎥
⎢⎣α α ⎥⎦ ⎢⎣α α ⎥⎦
⎡α 6 α 5 ⎤
⎢ ⎥
⎡α 3 α 5 ⎤ ⎣⎢α 5 α 3 ⎥⎦ −5 α
⎡ 6 α5 ⎤ 2 α
⎡ 6 α 5 ⎤ ⎡α 8 α 7 ⎤ ⎡α 1 α 0 ⎤
T
Inv ⎢ 5 6 ⎥= =α ⎢ 5 3 ⎥ =α ⎢ 5 3 ⎥ = ⎢ 7 5 ⎥ = ⎢ 0 5 ⎥ (39)
⎣⎢α α ⎦⎥ α5 ⎣⎢α α ⎦⎥ ⎣⎢α α ⎦⎥ ⎣⎢α α ⎥⎦ ⎣⎢α α ⎥⎦
Verification: If the inversion was performed correctly, the multiplication of the
DO

original matrix by the inverted matrix should yield an identity matrix.

⎡α 3 α 5 ⎤ ⎡α 1 α 0 ⎤ ⎡α 4 + α 5 α 3 + α 10 ⎤ ⎡1 0 ⎤
⎢ 5 6 ⎥⎢ 0 5 ⎥=⎢ 6 11 ⎥
=⎢ ⎥. (40)
⎢⎣α α ⎥⎦ ⎢⎣α α ⎥⎦ ⎢⎣α + α α + α ⎥⎦ ⎣0 1 ⎦
6 5

The coefficients σ 1 and σ 2 of the error-locator polynomial, σ ( X ) , are solving

from Equation (38)
⎡σ 2 ⎤ ⎡α α ⎤ ⎡α 6 ⎤ ⎡α 7 ⎤ ⎡α 0 ⎤
1 0

=
⎢σ ⎥ ⎢ 0 5 ⎥ ⎢ ⎥ = ⎢ 6⎥ = ⎢ 6⎥ (41)
⎣ 1 ⎦ ⎢⎣α α ⎥⎦ ⎣0 ⎦ ⎢⎣α ⎥⎦ ⎢⎣α ⎥⎦
From Equations (41) and (35) it can represent σ ( X ) :
σ ( X ) = 1 + σ1 X + σ 2 X 2 = 1 + α 6 X + α 0 X 2 (42)
The roots of σ ( X ) are the reciprocals of the error locations. Once these roots
are located, the error locations will be known. Because of each of the field elements
can be root of σ ( X ) (if it yields σ ( X ) = 0 ), it has to solve the σ ( X ) polynomial with
each of the field elements:

13
 σ (α 0 ) = α 0 + α 6 + α 0 = α 6 ≠ 0
σ (α 1 ) = α 0 + α 7 + α 2 = α 2 ≠ 0
σ (α 2 ) = α 0 + α 8 + α 4 = α 6 ≠ 0
σ (α 3 ) = α 0 + α 9 + α 6 = 0 ⇒ error

Y
σ (α 4 ) = α 0 + α 10 + α 8 = 0 ⇒ error
σ (α 5 ) = α 0 + α 11 + α 10 = α 2 ≠ 0

NO C P
σ (α 6 ) = α 0 + α 12 + α 12 = α 0 ≠ 0

CO Y
According to Equation (35), the error locations are at the inverse of the roots of

O
the polynomial. Therefore σ (α 3 ) = 0 indicates that one root exits at 1 βl = α 3 . Thus,

OP
β l = 1 / α 3 = α 4 . Similarly, σ (α 4 ) = 0 indicates that another root exits at 1 β l ' = α 4 . Thus,

PY
β l ' = 1 / α 4 = α 3 , where l , l ' ,..., l v refer to the first, second, …, v th error. Therefore, in this
D O NO T C
example, there are two-symbol errors, so that the error polynomial is of the following
form:
e( X ) = e j X J + e j X J
1
1
2
2
(43)
Thus, the two errors were found at locations α 3 and α 4 , i.e. β1 = α j = α 3 and 1

β2 = α j = α 4 .
2

2) Error Values Computation

DO NO

To calculate the error value in the location β1 = α 3 and β 2 = α 4 can be used any
T
of the four syndrome equations (34). For instance, assume that are S1 and S2
S1 = r (α ) = e1β1 + e2 β 2
S 2 = r (α 2 ) = e1β 12 + e2 β 22 , (44)
T
where indexes of the errors are 1 and 2, because they are associated with locations
β1 = α 3 and β 2 = α 4 .
These equations in matrix form are:
DO

⎡ β1 β 2 ⎤ ⎡e1 ⎤ ⎡ S1 ⎤
⎢ 2 2 ⎥⎢ ⎥ = ⎢ ⎥ (45)
⎣ β1 β 2 ⎦ ⎣e2 ⎦ ⎣ S 2 ⎦
⎡α 3 α 4 ⎤ ⎡e1 ⎤ ⎡α 3 ⎤
⎢ 6 8 ⎥ ⎢ ⎥ = ⎢ 5⎥ (46)
⎣⎢α α ⎥⎦ ⎣e2 ⎦ ⎣⎢α ⎦⎥
To solve for the error values e1 and e2, the matrix in Equation (46) is inverted
⎡α 1 α 4 ⎤ ⎡α 1 α 4 ⎤
⎢ 6 3 ⎥ ⎢ 6 3 ⎥
⎡α 3 α 4 ⎤ ⎢⎣α α ⎥⎦ ⎢⎣α α ⎥⎦ −6 α
⎡ 1 α4 ⎤ 1 α
⎡ 1 α4 ⎤
inv ⎢ 6 8 ⎥ = 3 1 = =α ⎢ 6 3 ⎥ =α ⎢ 6 3 ⎥ =
⎢⎣α α ⎥⎦ α α − α α α 4 +α3
6 4
⎢⎣α α ⎥⎦ ⎢⎣α α ⎥⎦ (47)
⎡α 2 α 5 ⎤ ⎡α 2 α 5 ⎤
=⎢ 7 4 ⎥=⎢ 0 4 ⎥
⎢⎣α α ⎥⎦ ⎢⎣α α ⎥⎦
It can calculate the error value from Equations (47) and (46)

14
 ⎡e1 ⎤ ⎡α α ⎤ ⎡α 3 ⎤ ⎡α 5 + α 10 ⎤ ⎡α 5 + α 3 ⎤ ⎡α 2 ⎤
2 5

=
⎢e ⎥ ⎢ 0 4 ⎥ ⎢ 5⎥ = ⎢ 3 9 ⎥
=⎢ 3 2⎥
= ⎢ 5⎥ (48)
⎣ 2 ⎦ ⎢⎣α α ⎥⎦ ⎢⎣α ⎥⎦ ⎢⎣α + α ⎥⎦ ⎢⎣α + α ⎥⎦ ⎢⎣α ⎥⎦

4) Correcting the Received Polynomial with Estimates of the Error Polynomial

Y
Using Equations (43) and (48) the estimated error polynomial is
)
e ( X ) = e1 X J + e2 X J = α 2 X 3 + α 5 X 4
1 2
(49)

NO C P
The algorithm that can repair the received polynomial is

CO Y
) ) )
U ( X ) = r ( X ) + e ( X ) = U ( X ) + e( X ) + e ( X ) (50)
r ( X ) = 100 + 001X + 011X 2 + 100 X 3 + 101X 4 + 110 X 5 + 111X 6

O
)

OP
e ( X ) = 000 + 000 X + 000 X 2 + 001X 3 + 111X 4 + 000 X 5 + 000 X 6

PY
)
U ( X ) = 100 + 001X + 011X 2 + 101X 3 + 010X 4 + 110 X 5 + 111X 6 =
(51)
= α 0 + α 2 X + α 4 X 2 + α 6 X 3 + α 1 X 4 + α 3 X 5+α 5 X 6
D O NO T C
Thus, received information message that is located in the rightmost k = 3 sym-
bols is the same as transmitted 010 110 111 → α 1 α 3 α 5 .
Advantages of RS codes:
- The RS code achieves the largest possible code minimum distance for any
linear code with the same encoder input and output block lengths.
- The RS code can be configured with long block lengths (in bits) with less de-
DO NO

coding time than other codes of similar lengths; This is because the decoder logic
T
works with symbol-based rather than bit-based arithmetic.
- R-S Codes Perform Well Against Burst Noise.

5. Interleaving
T
Interleaving is a powerful technique that can be used in digital communications
systems to enhance the random error correcting capabilities of block codes such as
Reed-Solomon codes to the point that they can be effective in a burst noise environ-
DO

ment. Interleaving is helpful when burst errors are present.

The transmitted message can be received with errors because of system’s noise
environment. Errors which can be classified into three broad categories 5 :
1) Random errors - the bit errors are independent of each other. Additive noise
typically causes random errors.
2) Burst errors - the bit errors occur sequentially in time and as groups.
3) Impulse errors - large blocks of the data are full of errors.
Random errors occur in a channel when individual bits in the transmitted mes-
sage are corrupted by noise. They can be described as some single erroneous bits in
the message. Thermal noise in communications channels is the main cause for ran-
dom errors. Reed-Solomon codes are designed specifically for correcting random er-
rors. If the number of errors per data block is small, (up to 4% byte errors) they can
be totally corrected by Reed-Solomon codes.

5
Interleaving for Burst Error Correction, AHA products groop, Moscow, Idaho, USA, www.aha.com.

15
 Burst errors are characterized as a series of adjacent erroneous bits in a signal.
They can be caused by fading in a communication channel and its can be difficult to
correct for some codes. However, block codes (in particular Reed-Solomon codes)
can handle burst noises effectively. The ability of a block code to correct burst errors
depends upon the number of errors in the signal. Interleaving can be used to extend

Y
the error correcting capability of the Reed-Solomon code.
Impulse errors can cause catastrophic failures in the communications system.
In general, all coding systems fail to reconstruct the message in the presence of catas-

NO C P
CO Y
trophic errors. However, certain codes like the R-S codes can detect the presence of a
catastrophic error by examining the received message.
The interleaver subsystem rearranges the encoded symbols over multiple code

O
OP
blocks. This effectively spreads out long burst noise sequences so they appear to the

PY
decoder as independent random symbol errors or shorter burst errors.
As it was shown in charter 4, (n, k) Reed Solomon codes are linear block non-
D O NO T C
binary cyclic codes with symbols, where k is the number of data symbols being en-
coded, and n is the total number of code symbols in the encoded block, n − k = 2t is
the number of parity symbols. This encoded block of length n symbols is called a
d min − 1 n − k
codeword. The number of errors that can be correct is t = = .
2 2
Suppose there are number of p corrupted by errors symbols in the block. If
p ≤ t errors will be located and corrected. If bursts of length p > t the error correcting
DO NO

code will fail. Let assume that t < p ≤ t * i , where i is called a interleaving depth (it is
T
an integer):
i≥ p , (52)
t
The Reed-Solomon (n,k) code can be used if we can spread the burst error se-
T
quence over several code blocks so that each block has no more than t errors which
can then be corrected. This can be accomplished using block interleaving.
DO

Example 4: There is RS (255,235) code with t = 10 , i = 5 . Without interleaver the data

bytes output from the Reed-Solomon encoder would appear as shown below in Figure 5, where
bytes numbered 0 to 234 are the data bytes and bytes 235 to 254 are the parity check bytes.

a0...a254 b0...b254 c0...c254 d0...d254 e0...e254

Figure 5. Data bytes output from the encoder for RS (255,235) code

The sequence of bytes enters in interleaver RAM row by row to build an i * k

data table as shown in Figure 6.

16
 Data bytes Parity bytes

codeword A a0 a1 a2 ... a234 a235 a236 ... a254

codeword B b0 b1 b2 ... b234 b235 b236 ... b254

Y
codeword C c0 c1 c2 ... c234 c235 c236 ... c254

codeword D d0 d1 d2 ... d234 d235 d236 ... d254

codeword E e0 e1 e2 ... e234 e235 e236 ... e254

NO C P
CO Y
Figure 6. Interleaver Table – Data Bytes are Input Row by Row and Output Column by
Column

O
OP
To interleave the data bytes its is sent to the output devise column by column

PY
as it is shown in Figure 7.
D O NO T C
a0 b0 c0 d0 e0 a1 b1 ... a254 b254 c254 d254 e254

Figure 7. Data Bytes Output from the Block Interleaver

6. Convolution code
Possibilities of convolutional coding have been introduced by Peter Elias
(1923-2001) in 1955. Convolutional coding offers an alternative to block codes for
DO NO

transmission over a noisy channel (Convolutional encoding with Viterbi decoding is a

T
Forward Error Correcting technique that is particularly suited to a channel in which
the transmitted signal is corrupted mainly by additive white gaussian noise). It has
been used in space communications and wireless communications. An advantage of
convolutional coding is that it can be applied to a continuous data stream as well as to
T
blocks of data.
Convolutional codes are commonly characterized by parameters: n, k, K, and r,
where
DO

n is a number of outputs bits;

k is a number of inputs bits;
K is a constraint length which equals the shift register stages of the encoder,
i.e. the number of information bits in the register memory, which take a part in cod-
ing process;
r is a ratio of information bits to transmitted bits k / n , called a code rate; it is a
measure of efficiency of the code. Smaller coding rate gives more powerful code due
to extra redundancy and less bandwidth efficiency.
The principle of “overlap” is extensively used in convolutional codes. For ex-
ample, in a rate k / n convolutional code, each input is “overlapped” with several pre-
vious inputs to produce each pair of encoded symbols. It is not able to divide coded
sequence into blocks as each coded symbol pair is interlocked with its neighbours.
Longer constraint lengths produce more powerful codes, but the complexity of de-
coding operations increases exponentially with constraint length.

17
 6.1. Coder and coding.
A convolutional encoder is a Mealy 6 machine, where the output is a function of
the current state and the current input. Figure 8 shows a ( n, k , m ) convolutional en-
coder.

Y
A A xor B
1 3 n B

input output
A B A xor B

NO C P
Output 0 0 0
k input

CO Y
codeword 1
information data 0 1
U=U1, U2,.., Un
1 2 3 ... m 1 0 1
Sfift register 1 1 0

O
...

OP
2

PY
Figure 8. ( n, k , m ) convolutional encoder
D O NO T C
It consists of one or more shift registers, multiple XOR gates and an output
switch. The stream of information bits flows in to the shift register from one end and
is shifted out at the other end every clock time. XOR gates are connected to some
stages of the shift registers to generate the output. There is no theoretical basis for the
optimal location of the shift register stages to be connected to XOR gates. It is based
on an empirical approach. The location of stages as well as the number of memory
DO NO

elements determines the minimum Hamming distance. The minimum Hamming dis-
T
tance determines the maximal number of correctable bits.
The output switch consecutively reads the outputs of XOR gates and generates
the output codeword U i = G (mi ) . Frequency clock switching depends on code rate
T
r = k / n . Thus, the frequency clock switching has to be n times higher by comparison
with input clock frequency. In practical use there are systematic and non-systematic
convolutional codes depending on the generator polynomial. For example, if the gen-
erator polynomials g1 ( x) = 1 or g 2 ( x) = 1 for a rate 1/2 code, the information sequence
DO

would appear directly in the output and the code becomes systematic. Figure 9 shows
idea of systematic convolutional code, figure 10 shows encoder of non-systematic
convolutional code.
m(t) output u1

Multiplexer u
S1 D
m(t)
m(n) m(n-1) m(n-2) S2

Shift register
C ENB
synchronization

output u2

Figure 9. Coder of systematic convolutional code

6
George H. Mealy „A Method for Synthesizing Sequential Circuits". 1955, Bell Systems Technical Journal 34: pp.
1045–1079.

18
 XOR
output u1

Double
speed
m input Multiplexer
output
information bit S1 D
u

Y
stream
m(n) m(n-1) m(n-2) S2
Shift register
C ENB
synchronization

output u2

NO C P
CO Y
XOR

Figure 10. Coder of non-systematic convolutional code

O
OP
One of the advantages of a systematic code is that it is simple to extract the in-

PY
formation sequence for a decoder.
The encoder shown in figure 10 has a code rate r = 1 / 2 . This means there are
D O NO T C
two output bits for each input bit. Here the output bits are transmitted one after an-
other, two per clock cycle. The two output can be represent as binary vectors [111]
and [101] are known as the generating vectors or generating polynomials for the code:
g1 ( x) = 1 + x + x 2
g 2 ( x) = 1 + x 2 . (53)
where m(n) is a last entered bit, m(n − 1) is a bit one clock time older than m(n) , and
DO NO

m(n − 2) is a bit two clock times older than m(n) .

T
After every clock time two bits appear in output: the first from output u1 and
the second from output u2
u1 = m(n) ⊕ m(n − 1) ⊕ m(n − 2)
T
u2 = m(n) ⊕ m(n − 2) . (54)
The correlation encoder can be described as a Mealy machine. The state is the
two bits in the shift register. Let the first input bit to the shift register be m(n) = 1 , and
DO

let the flip-flops be reset to zero so m(n − 1) = 0 and m(n − 2) = 0 . This state is
S0 = 00 = [m(n − 1), m(n − 2)] .
The output is
u1 = m(n) ⊕ m(n − 1) ⊕ m(n − 2) = 1 ⊕ 0 ⊕ 0 = 1 ,
u 2 = m ( n ) ⊕ m ( n − 2) = 1 ⊕ 0 = 1 ,
u = [u1 , u2 ] = 11
After the clock, state bit m(n − 1) = 0 will shift right into m(n − 2) , the input
m(n) = 1 will shift right into m(n − 1) , and the next state will be S1 = 10 = [m(n), m(n − 1)] .
In the first stage can enter m(n + 1) = 0 or m(n + 1) = 1 with equal probability.
If there is m(n + 1) = 0 the output is
u1 = m(n + 1) ⊕ m(n) ⊕ m(n − 1) = 0 ⊕ 1 ⊕ 0 = 1 ,
u2 = m(n + 1) ⊕ m(n − 1) = 0 ⊕ 0 = 0 ,

19
 u = [u1 , u2 ] = 10 ,
if there is m(n + 1) = 1 the output is
u1 = m(n + 1) ⊕ m(n) ⊕ m(n − 1) = 1 ⊕ 1 ⊕ 0 = 0 ,
u2 = m(n + 1) ⊕ m(n − 1) = 1 ⊕ 0 = 1 ,

Y
u = [u1 , u2 ] = 01 .
In a clock time 3 the bit m(n − 1) = 0 out of register, the bit m(n) = 1 will be

NO C P
moved to right in the third register, the bit m(n + 1) = 0 or 1 will move in the second

CO Y
stage. In the first enters new bit m(n + 2) , which can also be 0 or 1. The states in a
clock 3 are S 2 = 01 = [m(n + 1), m(n)] or S3 = 11 = [m(n + 1), m(n)] . The output can be

O
OP
If m(n + 2) = 0 , m(n + 1) = 0 :

PY
u1 = m(n + 2) ⊕ m(n + 1) ⊕ m(n) = 0 ⊕ 0 ⊕ 1 = 1 ,
u 2 = m ( n + 2) ⊕ m ( n ) = 0 ⊕ 1 = 1 ,
D O NO T C
u = [u1 , u2 ] = 11 ,
If m(n + 2) = 1 , m(n + 1) = 0 :
u1 = m(n + 2) ⊕ m(n + 1) ⊕ m(n) = 1 ⊕ 0 ⊕ 1 = 0 ,
u 2 = m ( n + 2) ⊕ m ( n ) = 1 ⊕ 1 = 0 ,
u = [u1 , u2 ] = 00 .
DO NO

If m(n + 2) = 0 , m(n + 1) = 1 :
T
u1 = m(n + 2) ⊕ m(n + 1) ⊕ m(n) = 0 ⊕ 1 ⊕ 1 = 0 ,
u 2 = m ( n + 2) ⊕ m ( n ) = 0 ⊕ 1 = 1 ,
u = [u1 , u2 ] = 01 .
T
If m(n + 2) = 1 , m(n + 1) = 1 :
u1 = m(n + 2) ⊕ m(n + 1) ⊕ m(n) = 1 ⊕ 1 ⊕ 1 = 1 ,
DO

u 2 = m ( n + 2) ⊕ m ( n ) = 1 ⊕ 1 = 0 ,
u = [u1 , u2 ] = 10 .

Example 5: Let the information data consist three bits m = [101] . The encoder structure is
shown in Fig. 10. The stages of the shift register are reset and the first bit received (clock time 1) is
m(n) = 1 . So in the first cycle in the output will
u1 = 1 ⊕ 0 ⊕ 0 = 1 , u2 = 1 ⊕ 0 = 1 , u = [u1 , u2 ] = 11 .
After the clock a bit m(n + 1) = 0 enters. The result is
u1 = 0 ⊕ 1 ⊕ 0 = 1 , u2 = 0 ⊕ 0 = 0 , u = [u1 , u2 ] = 10 .
Then (at time 3) a bit m(n + 2) = 1 enters. The result is
u1 = 1 ⊕ 0 ⊕ 1 = 0 , u2 = 1 ⊕ 1 = 0 , u = [u1 , u2 ] = 00 .
After that a tail of two zero-bits is appended to data bits to clear out the memory after encod-
ing the last bit (all-zero); the result is two more pairs of output symbols:
t = 4 : u = [u1 , u2 ] = 10 ,

20
 t = 5 : u = [u1 , u2 ] = 11 .
The output coded message is 11 10 00 10 11 .

L
The ratio Reff = is called effective code rate, where L is the
n[( L / k ) + ( K − 1)]

Y
number of data bits. For example 5: L = 3 , n = 2 , k = 1 , K = 3 and
L 3 3
Reff = = = which is less than r = 1 / 2 .

NO C P
n[( L / k ) + ( K − 1)] 2[(3 / 1) + (3 − 1)] 10

CO Y
The effective code rate falls below code rate r because the added zero-bits do
not carry information.

O
OP
Example 6: Let the information data consist three bits m = [101] . The encoder structure is

PY
shown in Fig. 11. This is a coder with a constraint length K = 3 and code rate r = 1 / 3 .
The memory of the shift register is clean and the first bit received (clock time 1) is m(n) = 1 .
D O NO T C
So in the first cycle in the output will
u1 = 1 ⊕ 0 ⊕ 0 = 1 , u2 = 1 ⊕ 0 = 1 , u3 = 0 ⊕ 0 = 0 u = [u1 , u2 , u3 ] = 110 .
After the clock a bit m(n + 1) = 0 enters. The result is
u1 = 0 ⊕ 1 ⊕ 0 = 1 , u2 = 0 ⊕ 0 = 0 , u3 = 1 ⊕ 0 = 1 u = [u1 , u2 , u3 ] = 101 .
When t = 3 m(n + 2) = 1 enters. The output result is
u1 = 1 ⊕ 0 ⊕ 1 = 0 , u2 = 1 ⊕ 1 = 0 , u3 = 0 ⊕ 1 = 1 u = [u1 , u2 , u3 ] = 001 .
DO NO
T
m input output u1
information bit
stream output u2 u
T
m(n) m(n-1) m(n-2)
output u3

synchronization
DO

Figure 11. Coder with a constraint length K = 3 and code rate r = 1 / 3

After that a tail of two zero-bits is appended to data bits to clear out the memory after encod-
ing the last bit (all-zero).
t = 4 : u1 = 0 ⊕ 1 ⊕ 0 = 1 , u2 = 0 ⊕ 0 = 0 , u3 = 1 ⊕ 0 = 1 u = [u1 , u2 , u3 ] = 101 .
t = 5 : u1 = 0 ⊕ 0 ⊕ 1 = 1 , u2 = 0 ⊕ 1 = 1 , u3 = 0 ⊕ 1 = 1 u = [u1 , u2 , u3 ] = 111 .
The output coded message is 110 101 001 101 111 .

The coder is represented by set of generator polynomials, one for each of the n
modulo-2 adders. Each polynomial is of degree K − 1 or less and describes the con-
nection of the shift register to that modulo-2 adder. The coefficients can be either 1 or
0. The value 1 means that connection exists between the stage of shift register and the
modulo-2 adder. The value 0 means that connection does not exist. For instant, the

21
 encoder shown in figure 10 can be representing by generator polynomials (53), the
encoder shown in figure 11 can be representing by generator polynomials
g1 ( x) = 1 + x + x 2
g 2 ( x) = 1 + x 2

Y
g 3 ( x) = x + x 2 . (55)
The information message from Example 5 can be representing by generator
polynomial m( x) = 1 + x 2 . The output u1 is formed by multiplication of g1 ( x) with m( x) :

NO C P
CO Y
g1 ( x).m( x) = (1 + x + x 2 ).(1 + x 2 ) = 1 + x + x 3 + x 4 ,
and

O
g 2 ( x).m( x) = (1 + x 2 ).(1 + x 2 ) = 1 + x 4 .

OP
These Equations can represent with coefficients:

PY
g1 ( x).m( x) = 1 + 1.x + 0.x 2 + 1.x 3 + 1.x 4 ;
g 2 ( x).m( x) = 1 + 0.x + 0.x 2 + 0.x 3 + 1.x 4 ;
D O NO T C
U ( x) = (1,1) + (1,0).x + (0,0).x 2 + (1,0).x 3 + (1,1).x 4 ;
U = G (m) = 11 10 00 10 11 .
The result of U ( x) is the same as that in Example 5.
For Example 6:
g1 ( x).m( x) = (1 + x + x 2 ).(1 + x 2 ) = 1 + 1.x + 0.x 2 + 1.x 3 + 1.x 4
g 2 ( x).m( x) = (1 + x 2 ).(1 + x 2 ) = 1 + 0.x + 0.x 2 + 0.x 3 + 1.x 4
DO NO

g 3 ( x).m( x) = ( x + x 2 ).(1 + x 2 ) = 0 + 1.x + 1.x 2 + 1.x 3 + 1.x 4

T
U = 110 101 001 101 111 .

Example 7: Let the information data consist three bits m = [1011] . The encoder structure is
shown in Fig. 10. This is a coder with a constraint length K = 3 and code rate r = 1 / 2 .
T
The work of the encoder to create a codeword can be presented in tabular form, as it is
shown below.
DO

№ of Input bit, Contents of the State, cycle i State, cycle u1 u2 u

cicle cycle i register i +1
000 S 0 = 00 S 0 = 00 - - -
1 1 100 S 0 = 00 S1 = 10 1 1 11
2 0 010 S1 = 10 S 2 = 01 1 0 10
3 1 101 S 2 = 01 S1 = 10 0 0 00
4 1 110 S1 = 10 S3 = 11 0 1 01
5 0 011 S3 = 11 S 2 = 01 0 1 01
6 0 001 S 2 = 01 S 0 = 00 1 1 11

The output codeword is 11 10 00 01 01 11 .

6.2. State representation and state diagram

A convolutional encoder is a Mealy machine (finite-state machine), where the
output is a function of the current state and the current input. There are a finite num-
ber of states that the machine can encounter. The state consists of the smallest quan-
22
 tity of information that with a current input bit can predict the output of the machine.
For a rate 1 / n convolutional encoder the state is represented by contents of ( K − 1)
rightmost stages.
There are some alternative ways of describing a convolutional code. It can be

Y
expressed as a state diagram, a tree diagram, or a trellis diagram.
A State Diagram shows all possible present states of the encoder as well all the
possible state transitions that may occur. In order to create the state diagram, a state

NO C P
transition table may first be made, showing the next state for each possible combina-

CO Y
tion of the present state and input to the decoder. For example it can draw the table
for the encoder shown in figure 10.

O
OP
Table 4: State Transition Table

PY
Next State if
Current State
input m = 0 input m = 1
D O NO T C
00 00 10
10 01 11
01 00 10
11 01 11

Table 5 shows the change in output for each combination of input and previous
output.
Table 5: Output Table
DO NO

Current out- Output simbols if

T
put input m = 0 input m = 1
00 00 11
10 10 01
01 11 00
T
11 01 10

The state diagram is created using the information from Table 4 and Table 5,
DO

as shown in Figure 12.

m = 0;u=00

00
1

S0
=1

m
;u

=0
=1

;u
m

=1
1

m = 1 ;u = 00
10 01
S1 S2
m = 0 ;u= 10
m

01
=1

=
;u

;u
=0

=0

11
1

S3

m = 1;u= 10

Figure 12. State Diagram (coder К=3, R = 1/ 2 , Figure 10)

23
 For each state, there can be two outgoing transitions; one corresponding to a
‘0’ input bit and the other corresponding to a ‘1’ input bit. The values inside the cir-
cles indicate the state (contents of ( K − 1) rightmost stages). The values on the arrows
indicate the output of the encoder. A solid line corresponds to input 0, a dotted line –

Y
to input 1. The State Diagram shows an entire picture the all possible states and their
output code words. The disadvantage of the state diagram is the lack of measurement
of time, i.e. it is not clear what state of the encoder in which time is obtained.

NO C P
A Tree Diagram shows the passage of time. In each cycle coding procedure is

CO Y
described in a passage in the branches from left to right as each branch shows the
codeword. If the input bit is 1, the codeword is found by moving to the next right and

O
below branch, if the input bit is zero - the right and above. Any input sequence can be

OP
traced through a path in this diagram which forms the corresponding code word. This

PY
path can also be called the code word path. For instance, Figure 13 shows a tree dia-
gram of coder, shown in Figure 10.
D O NO T C
00 S 0 00
00 S0 11
11 S 1 10
00 S0 01
10 S 2 11
11 S 1 00
01 S 3 01
00 S0 10
11 S 0 00
DO NO

10 S 2 11
T
00 S 1 10
11 S 1 01
01 S 2 11
01 S 3 00
10 S 3 01
10
0
T
1 00
00 S 0
11 S 0 11
10
11 S 1
DO

10 S 2 01
10 S 2 11
00 S 1 00
01
01 S 3
11 S 1 10
00
11 S 0
01 S 2 11
10
00 S 1
01 S 3 01
01 S 2 11
10 S 3 00
10 S 3 01
10

Figure 13. Tree Diagram of coder, shown in Figure 10

The disadvantage is the progressive increase in the branches if information

message bits (input) are too much. After the first K branches (constraint length) tree
is repeated.
Trellis Encoding Diagram present the sequence of events in linear time in a
compact form. In constructing the chart along the horizontal axis has time clocks and
24
 the vertical - all possible states. Figure 14 shows first two clocks of a trellis diagram
for the encoder given in Figure 10.
0;00
S0 S0
1;1
1

Y
вход m ; изход u
m=1

11
S1 S1

0;
m=0

00
1;
NO C P 0;
10

CO Y
1;
S2 S2

01
1
0 ;0

O
1;10

OP
S3 S3

PY
t=n t=n+1

Figure 14. Trellis structure for R=1/2, k=3 convolutional code (first two clocks)
D O NO T C
In the trellis diagram, nodes correspond to the states of the encoder. From an
initial state ( S0 ) the trellis records the possible transitions to the next states for each
possible input pattern. Figure 15 shows a trellis diagram for the encoder given in Fig-
ure 10 and input message m = 101100 .
0;00 0;00 0;00 0;00 0;00 0;00
S0 S0 S0 S0 S0 S0 S0
DO NO
1;

1;
1;

1;
1;
1;
11

11

11
11

11
11

1
1

0;1
1

0;1
0;1

0;1

S1 S1 S1 S1 S1 S1 S1
00

00

00

00
1;

1;

1;

1;
T 0;

0;

0;
0;

0;

10

10

10
10
10

S2 S2 S2 S2 S2 S2 S2
1;0
1; 0
DO

1;0
1;0

01

1;0
01

01

01
0;
1
1

1
1
0;

1
0;

0;

1;10 1;10 1;10 1;10

S3 S3 S3 S3 S3 S3 S3

t=1 t=2 t=3 t=4 t=5 t=6 t=7

Figure 15. Trellis Diagram for the encoder given in Figure 10 and input message
m = 101100

For this code, there are four possible encoder states. Each row of nodes is rep-
resenting the same state of the encoder at different time steps. The first row nodes
shows the state S0 = 00 , the second - S1 = 10 , the third - S 2 = 01 , and fifth - S3 = 11 .
Thus, the trellis diagram requires 2 K −1 nodes to represent 2 K −1 possible encoder states.
Each of the states can be entered from either of two preceding states. At every time
step there are two branches output from each node where an input “0” to the encoder
corresponds to the upper branch and a “1” input to the lower branch. Since the initial
condition of the encoder is S0 = 00 , and the two memory flushing bits are zeroes, the
lines start out at S0 = 00 and end up at the same state. For example, the input sequence
25
 1 0 1 1 0 0 corresponds to the particular path, shown with thick red lines, through the
trellis. The output codeword is 11 10 00 01 01 11 , and states are S 0 , S1 , S 2 , S1 , S3 , S 2 , S0 .

6.3. The Viterbi Decoding Algorithm

Y
There are two types of decoding algorithms used with convolutional encoding:
sequential decoding and Viterbi decoding algorithm. Sequential decoding has the ad-
vantage that it can perform very well with long-constraint-length convolutional codes,

NO C P
but it has a variable decoding time. The Viterbi decoding algorithm was developed by

CO Y
Andrew J. Viterbi 7 in 1967. This algorithm performs a maximum likelihood decoding
(maximum likelihood detector). Viterbi decoding has the advantage that it has a fixed

O
decoding time and it has good performance in hardware. Disadvantage is exponential

OP
growth of its computational requirements as a function of the constraint length. In

PY
this reason, it is usually limited to constraint lengths of K = 9 or less.
The algorithm includes calculating a distance between received signal at time
D O NO T C
ti , i = 1,2,3,... and all trellis paths entering each state at time ti . The metric of calculat-
ing is the Hamming distance between the received channel symbol pair and the all
possible channel symbol pairs. The Hamming distance is computed by counting how
many bits are different between the received channel symbol pair and the possible
channel symbol pairs (see part 3). The results can be zero, one, ... The Hamming dis-
tance values is computing at each time instant for the paths between the states at the
previous time instant and the states at the current time instant are called branch met-
DO NO

rics. When the encoding process is presented with trellis diagram it shows that the
T
coded output word is formed as a path through the states of the encoder at any clock
time linear in time. The decoder receives the sequence of bits of the codeword and
must determine the path through the states that will determine the decoded sequence
of bits. Since it is unknown whether the received sequence without error, the path in
T
trellis diagram is chosen according to the smallest Hamming distance.
Figure 16 shows how to choice the path in trellis diagram according to the
smallest Hamming distance measured in every clock time. Suppose the received bits
DO

have an error so instead of 11 10 , one receives 11 11 . The lines start out at S0 = 00 and
according to two possible outputs (00 and 11) the path can only go through state
S 0 = 00 or state S1 = 10 . Suppose at t = 1 , it is received 11. The only possible channel
symbol pairs that could have received are 00 and 11. The Hamming distance between
11 and 11 is zero. The Hamming distance between 11 and 00 is two. The Hamming
distance between the received input and the bits for the transition is shown in box.
The choice is S0 → S1 because of less Hamming distance. The next input has an error.
Possible paths, 10 or 01, are at Hamming distance 1. At t = 2 it can going from S3 to
state S3 (Hamming distance is 1) or from S3 to state S 2 (Hamming distance is 1, too), or
from S 2 to state S1 (Hamming distance is 0) and from S 2 to state S0 (Hamming distance
is 2). Thus at t = 1 the proper path was not obvious, at t = 2 , the choice is clearer. The
choice is a path through the trellis based on the path Hamming distance or path metric,

7
Viterbi A. J., Error Bounds for Convolutional Codes and an Asymptotically Optimum Decoding Algorithm", IEEE
Transactions on Information Theory, Volume IT-13, pages 260-269, April, 1967.

26
 which is the sum of the Hamming distances as one steps along a path through the trel-
lis. The path S0 → S1 → S 2 → S1 has a Hamming distances sum equal to 1. The Ham-
ming distances sums of other paths are higher then this one. At t = 3 it can going from
S1 to state S3 (Hamming distance is 0) or to S 2 (Hamming distance is 2). Thus, the

Y
choice is going according to minimum Hamming distance in every clock time.
Transmitted
11 10 00 01 01 11 codeword
error
Received
11 11 00 01 01 11

NO C P
sequence

CO Y
0;00 2
S0 S0 S0 S0 S0 S0 S0
1;
11

Hamming

O
0

distance

1 2

OP
0
0;1
S1 S1 S1 S1 S1 S1 S1

PY
0;1

2
0

00
00

1;
1;
0;

0;
10

10
D O NO T C 1

2
1;0
S2 S2 S2 S2 S2 S2 S2

1
0

0
1
1;0

01
01

0;
1

0;
1

1;10 1 1;10 2
S3 S3 S3 S3 S3 S3 S3

t=0 t=1 t=2 t=3 t=4 t=5 t=6

0 Decoded
1 0 1 1 0
word

Figure 16. To choice the path in trellis diagram according to the smallest Hamming distance
DO NO
T
Figure 17 illustrates how to choice the path in trellis diagram according to the
smallest sum of the Hamming distances that is called path metric.
Transmitted
11 10 00 01 01 11 codeword
error
T
Received
11 11 00 01 01 11
sequence
5
0;00 2 0;00 1 0;00 1 0;00 2
S0 S0 S0 S0 S0 S0 S0
1
1;
11
1;

1;
1;
DO

11
11

11
1

1
0

1 1
2

0
1 1
1
0;1

0;1

S1 S1 S1
1

S1 S1 S1 S1 3
0;1

0;1
1

1
0

2
0

00
1;
00

00
1;
1;
0;

1;
0;
0;

0;
10

10
10

10
1;0

2
1

S2 S2 S2 S2 S2 S2 S2
0

4
0
1

0
1;0
1;0

1;0
10
01

01

01
1

0;
1

1
0;

0;

0;
0
1

1;10 1 1;10 2 1;10 2 1;10 1 4

S3 S3 S3 S3 S3 S3 S3

t=0 t=1 t=2 t=3 t=4 t=5 t=6

Decoded
1 0 1 1 0 0
word

Figure 17. The choice the path according to the smallest sum of the Hamming distances
(path metric)

At t = 6 , one path has a total distance of 1 from the input data. The others have
a distance of 3, 4 or 5. Thus the most likely path is S 0 , S1 , S 2 , S1 , S3 , S 2 , S0 with a path

27
 distance of 1, and the corresponding output decoded word is 101100 . In decoders
where the input is an analog signal, the distance between the actual and expected
voltage may be measured, and the sum of the squares of the errors might be used for
the branch metric.
The minimal path metric is the rule that the Viterbi decoder exploits to recover

Y
the original message. It is the maximum-likelihood path.

6.4. Implementation of the Viterbi Decoder

NO C P
CO Y
A Viterbi decoding algorithm consists of the three major parts, shown in Fig-
ure 18:
1. Branch metric calculation – calculation of a distance between the input pair

O
OP
of bits and the other possible pairs (“00”, “01”, “10”, “11”).

PY
2. Path metric calculation – for every encoder state, calculate a metric for the
survivor path (a path with the minimum metric) ending in this state.
D O NO T C
3. Traceback (survivor path decoding) – it is necessary for hardware imple-
mentations that don't store full information about the survivor paths, but store only
one bit decision every time when one survivor path is selected from the two.
Received Branch metric Path metric Traceback Decoded
sequence calculation calculation word

Figure 18. The major parts of the Viterbi decoding algorithm

DO NO

The Branch metric unit (BMU) receives input data from the channel and com-
T
putes a metric for each state and input combination. Branch metric calculation is dif-
ferent for hard decision and soft decision decoders. For a hard decision decoder, a
branch metric is a Hamming distance between the received pair of bits and the
T
“ideal” pair. Therefore, for every input pair a branch metric can take values of 0, 1
and 2. For a soft decision decoder, the received code words are quantised into differ-
ent levels according to the signal strength then the BMU maps the levels of code
words into BMs according to their likelihood. For a soft decision decoder, a branch
DO

metric is measured using the Euclidean distance:

Bm = ( x − x0 ) 2 + ( y − y0 ) 2 , (56)
where x is the first received bit in the pair, y – the second, x0 and y0 – the “ideal”
values.
Path metric is calculated by a number of units called “Add-Compare-Select”
(ACS). The ACS unit adds the current metric to the accumulated metric for each path
and determines the least metric for each state of the trellis. The ACS retrieves the ac-
cumulated metric from the register files, and then adds the current metric. The result
is stored back in the register files. This procedure is repeated for every encoder state.
- Add – a new value of the state metrics has to be computed at each time in-
stant; for a given state are known two states on the previous step which can move to
this state, and the output bit pairs that correspond to these transitions. To calculate
new path metrics, it has to add the previous path metrics with the corresponding
branch metrics, i.e. the state metrics have to be updated every clock cycle.

28
 - Compare and select – There are two paths, ending in a given state. One of
them (with greater metric) is rejected.
As there are 2 K −1 encoder states, we have 2 K −1 survivor paths at any given time.
The traceback unit (TBU) traces back the trellis after a block of data (deter-
mined by the trace back length) has been processed by the ACS. First, the TBU estab-

Y
lishes an optimal path by starting from the node of minimum metric and traces back
the path in the trellis all the way to the beginning of the trellis diagram. Then, the
original data is determined.

NO C P
CO Y
6.5. Recursive Systematic Convolutional Encoder
The recursive systematic convolutional (RSC) encoder is produced from the

O
OP
nonrecursive nonsystematic (conventional) convolutional encoder by feeding back

PY
one of its encoded outputs to its input. Figure 19a,b shows an example of conven-
tional convolutional encoder and recursive systematic convolutional encoder.
D O NO T C
MF(X)
F(X)

M(X) e e e
M(X) M0(X) e e e
e(n) e(n)
(n-1) (n-2) (n-...) (n-1) (n-2) (n-...)

U(X) U(X)
DO NO

G(X) G(X)
T
a) b)
Figure 19. Example of conventional convolutional encoder and recursive systematic convo-
lutional encoder.
T
These encoders are represented by: M ( X ) is a polynomial of information mes-
sage; U ( X ) is a coded message (word); G ( X ) is a nonrecursive generator polynomial;
DO

F ( X ) is a feedback polynomial. The output transfer function for the encoder, shown
in Figure 19a, is
U ( X ) = M ( X ).G( X ) , (57)
for the recursive encoder
GR ( X ) = U ( X ) / M ( X ) , (58)
M 0 ( X ) = M ( X ) + M F ( X ) , M F ( X ) = M 0 ( X ).F ( X ) ,
M 0 ( X ) = M ( X ) + M 0 ( X ).F ( X ) → M ( X ) = M 0 ( X )[1 − F ( X )] , U ( X ) = M 0 ( X ).G ( X )
From Equations (57) and (58) the recursive polynomial is
GR ( X ) = U ( X ) / M ( X ) = G ( X ) /[1 − F ( X )] , (59)
i.e. the recursive polynomial depends on the nonrecursive generator polynomial G ( X )
and the feedback polynomial F ( X ) .
It has been proven that the feedback polynomial F ( X ) has to be a primitive
polynomial as a condition for stable operation of the recursive encoder.

29
 Advantages of the recursive systematic convolutional encoder (the positive ef-
fect of the feedback) are:
- Reducing the number of states of the encoder, which increases the efficiency
of decoding algorithm, for example, if the degree of the nonrecursive generator poly-
nomial G ( X ) is p , the degree of the feedback polynomial F ( X ) is q and q < p , af-

Y
ter division the degree of the recursive polynomial decreases, which reduces the con-
straint length of recursive code; this allows for synthesis of recursive code with a

NO C P
small number of states and simplified decoding algorithm;

CO Y
- Increasing the average distance between output code sequences (increasing
the code Hemming weight), which achieves better results for error correction.

O
OP
PY
References
D O NO T C
1. Bernard Sklar, Digital Communications: Fundamentals and Applications, Second Edi-
tion, Prentice-Hall, 2001.
2. Berrou C., Glavieux A., Thitimajshima P., Near Shannon Limit Error-Correcting Coding
and Decoding: Turbo Codes // Proc. of the Intern. Conf. on Commun (Geneva, Switzerland). May
1993, pp.1064–1070.
3. Berrou C., Glavieux A., Near Optimum Error Correcting Coding and Decoding: Turbo-
Codes, IEEE Trans. On Comm., Vol. 44, No. 10, October 1996.
4. Blahut, R. E., Theory and Practice of Error Control Codes, MA: Addison-Wesley, 1983.
DO NO

5. C. E. Shannon, A mathematical theory of communication, Bell Syst. Tech., 1948.

6. Dennis Roddy, Satellite communications, McGraw-Hill Professional, 2001 - 569 p.
T
7. Fred Daneshgaran, Massimiliano Laddomada, and Marina Mondin, High-Rate Recursive
Convolutional Codes for Concatenated Channel Codes, IEEE Transactions on Communications,
Vol. 52, No. 11, november 2004.
8. Fred Ma, John Knight, Convolution Codes, http://read.pudn.com/, 2001.
T
9. Forney G., D., The Viterby algorithm. IEEE, vol. 61, n. 3, Marsh, 1978, pp. 268-278.
10. Gallager, R. G., Information Theory and Reliable Communication, New York: John
Wiley and Sons, 1968.
11. Gérard Maral, Michel Bousquet, Satellite communications systems: systems, techniques,
DO

and technology, John Wiley and Sons, 2002 - 757 pp.

12. George H. Mealy „A Method for Synthesizing Sequential Circuits". 1955, Bell Systems
Technical Journal 34: pp. 1045–1079.
13. Imai H., Hirakawa S., A New Multilevel Coding Method Using Error Correcting Codes.
IEEE Trans. Info Theory, vol. IT-23, n. 3, pp. 371–377, May 1977.
14. Intuitive Guide to Principles of Communications, Coding Concepts and Block Coding,
www.complextoreal.com
15. Interleaving for Burst Error Correction, AHA products groop, Moscow, Idaho, USA,
www.aha.com.
16. International Telecommunications Union, www.itu.int.
17. Ippolito, Louis J., Satellite communications systems engineering: atmospheric effects,
satellite link design, and system Performance, JohnWiley & Sons Ltd, 2008, 396 p.
18. Mealy, George H., A Method for Synthesizing Sequential Circuits, Bell Systems Tech-
nical Journal 34,1955: pp. 1045–1079.
19. Michael Kolawole, Satellite communication engineering, Marcel Dekker, 2002, 263 p.
20. RPC Telecommunications Ltd., http://www.satcom.co.uk/.
21. Soleymani M. R., Yingzi Gao, Vilaipornsawai U., Turbo coding for satellite and wire-
less communications. Kluwer Academic Publishers, Dordrecht. 2002. 214 pp.

30
 22. Stewart Robert W., Daniel Garsia-Alis, Concise DSP Tutorial from Digital Communica-
tions: Fundamentals and Applications, Blue Box Multimedia, 2001, http://www.DSPedia.com
23. Viterbi A. J., Convolutional codes and their performance in communication systems,
IEEE Trans. Commun., vol. 10, pp. 751–772, Oct. 1971.
24. Viterbi A. J., Error Bounds for Convolutional Codes and an Asymptotically Optimum

Y
Decoding Algorithm", IEEE Transactions on Information Theory, Volume IT-13, pages 260-269,
April, 1967.

NO C P
CO Y
O
OP
PY
D O NO T C
DO NO
T
T
DO

31
 Satellite Communications Systems

Lecture 4.

Y
Multiple Access in Satellite Communications
1. Introduction of Multiple Access in Satellite Communications
2. Frequency Division Multiple Access

NO C P
2.1. Frequency Division Multiple Access with PCM/TDM/PSK application

CO Y
2.2. Frequency Division Multiple Access with PCM/SCPC/PSK application
3. Time Division Multiple Access

O
3.1. Time Division Multiple Access with PCM/TDM/PSK application

OP
3.2. TDMA Frame Efficiency

PY
3.3. TDMA capacity
D O NO T C
3.4. Switching in satellite TDMA
4. Code Division Multiple Access
4.1. Direct Sequence Spread Spectrum
4.2. Frequency Hopping Spread Spectrum.

1. Introduction of Multiple Access in Satellite Communications

A multi-user system achieves a specific point in the capacity region depending
DO NO

on how the multi-user channel is shared by the users, which depends on the multiple
T
access technique or Multiple Access (MA). It is the last technique before signal
transmitting.
Satellite links are designed to provide desired link availability for average
conditions. Satellite Multiple Access techniques interconnect ground stations through
T
multiple satellite transponders. The goal is optimizing some of system parameters
such as power efficiency; spectral efficiency and etc. These techniques are applicable
to both fixed and mobile users. The satellite transponder may have different
DO

configurations, depending on the application and the satellite payload design.

Figure 1 illustrates four basic multiple access configurations 1 .
The first configuration is single channel per carrier and single carrier per
transponder (a). Single baseband channel modulate an RF carrier. The baseband
channel could be analog (voice or video) or a digital (data, voice, or video).The
modulation could be analog (amplitude or frequency modulation), or digital
(frequency shift keying or various forms of phase shift keying). The second
possibility is multiple channels per carrier and multiple carriers per transponder (b).
To avoid intermodulation noise, the final amplifier in the transponder is usually
operated in a power backoff mode. The next case (c) shows that multiple baseband
channels are multiplexed onto a single data stream before carrier modulation
(multiple channels per carrier, single carrier per transponder). The last case consists

1
Ippolito, Louis J., Satellite communications systems engineering : atmospheric effects, satellite link design, and
system Performance, 2008 JohnWiley & Sons Ltd, 396 pp.

1
 of multiple multiplexed baseband channels modulating multiple RF carriers, with the
multiple carriers per single transponder (d). This case also requires operating in a
power backoff mode to avoid intermodulation noise. Cases (a) and (c) is called single
carrier per transponder. The final amplifier can operate in full power, and provide

Y
high power efficiency. When configuration is multiple carriers per transponder (b and
d) power backoff can be several dB, resulting in lower power efficiency in
comparison with cases (a) and (c).

NO C P
CO Y
BBC carrier
BBC carrier BBC carrier
BBC carrier
а) b)

O
OP
BBC
BBC carrier

PY
BBC
BBC
BBC BBC carrier
BBC carrier BBC
D O NO T C
BBC
BBC
BBC carrier
BBC
c) d)

BB – Baseband channel;
Carrier – Modulated Radio Frequency Carrier;
– Satellite Transponder.

Figure 1. Illustration of the four basic multiple access configurations

DO NO

The satellite multiple access methods can be categorized into three mainly
T
three conventional techniques:
• Frequency Division Multiple Access (FDMA);
• Time Division Multiple Access (TDMA);
• Code Division Multiple Access (CDMA).
T
FDMA system separates multiple carriers in the transponder by frequency. The
transmissions can be analog or digital, or combinations of both. TDMA system
separates the multiple carriers by time in the transponder, presenting only one carrier
DO

at any time to the transponder. TDMA is most practical for digital data only, because
the transmissions are in a burst mode to provide the time division capability. CDMA
is a combination of both frequency and time separation. As the most complex
technique, CDMA requires several levels of synchronization at both the transmission
and reception levels and it is implemented for digital data only. CDMA offers the
highest power and spectral efficiency operation of all three techniques.
On the base of these three fundamental techniques are developed some
secondary access techniques which are putting into practice:
9 Demand Assigned Multiple Access (DAMA) - changes signal configuration
to respond to changes in user demand. FDMA or TDMA networks can be operated as
an assigned-on-demand DAMA network. CDMA is a DAMA network by design,
because it is a random access system by its implementation.
9 Space Division Multiple Access (SDMA) - is a promising multiple access
technology for improving capacity by the spatial filtering capability of adaptive

2
 antennas. It spatially separated physical links by different antenna beams, cells,
sectored antennas, signal polarization, etc. It is employable with any of the three
basic multiple access techniques. SDMA is an essential element of mobile satellite
networks, which employ multibeam satellites and it can be applied to all other

Y
multiple access schemes.
9 Satellite Switched TDMA (SS/TDMA) - employs sequenced beam
switching to add an additional level of multiple access in a frequency translation

NO C P
CO Y
satellite. The switching is accomplished at radio frequency (RF) or at an intermediate
frequency (IF).
9 Orthogonal frequency division division multiple access (OFDMA) - a

O
OP
special form of multicarrier modulation, can be used for multiplexing for multiple

PY
users. An OFDMA system is defined as one in which each user occupies a subset of
subcarriers, and each carrier is assigned exclusively to only one user at any time.
D O NO T C
Advantages of OFDMA over OFDM-TDMA and OFDM-CDMA include elimination
of intracell interference and exploitation of network/multiuser diversity.

2. Frequency Division Multiple Access

Frequency division multiple access (FDMA) was the first technique
implemented on satellite systems. FDMA divides all frequency band into many
frequency channels and assigns a separate frequency channel on demand to each
DO NO

earth station. Figure 2 shows functions of FDMA process.

T
f
Friquency
band 3 { ES 3 f1
Friquency
{ ES 2
T
band 2 f2
f3 ES 1

band 1 {
Friquency f1 f2 f3
ES 1
ES 2
t ES 3
DO

Figure 2. Principle of the Frequency Division Multiple Access

Figure 2 shows an example for three earth (ground) stations (ES) accessing a
single frequency translation satellite transponder. Each station is assigned a specific
frequency band for its uplink ( f1 , f 2 and f 3 ). Frequency guard bands are used to avoid
interference between the user slots. All stations receive total spectrum on downlink.
The receiving station must be able to receive the full spectrum and can select the
desired carrier for demodulation or detection.
The main advantages of this technique are:
- independent of the channels;
- no framing or synchronization bits needed;
- simplicity;
- transmissions can be analog or digital, or combinations of both;
- FDMA is most useful for applications where a full time channel is desired –
for example, video distribution;

3
 - the least expensive to implement technique.
The main disadvantages are:
- lower power efficiency because of the multiple carriers in the transponder
final power amplifier requires power backoff operation to avoid intermodulation

Y
noise;
- lower flexibility if it is necessary another system configuration (change
frequency plan);

NO C P
CO Y
- when number of users increase system efficiency decrease.
Multiple access system performance must be analyzed by considered the
specific signal processing used in the satellite communications for both analog and

O
OP
digital signals.

PY
2.1. Frequency Division Multiple Access with PCM/TDM/PSK application.
D O NO T C
Some of the most common techniques in satellite communications are a
PCM/TDM (pulse code modulation/time division multiplexed). Its application used
for voice communications. By PSM analog signals are converted into digital and
combined using TDM hierarchy according ITU-R Recommendation. The first level
consists of 30 channels which of them with 64kbit / s , multiplexed to a 2,048Mbit / s
TDM bit rate. The carrier modulation is phase shift keying (BPSK or QPSK).
One of the most important parameter for evaluation is the capacity of the
DO NO

multiple access system. It determines the maximum number of users that can access
T
the satellite and serves as the basis for decisions on demand access options on the link.
The capacity for the PCM/TDM/PSK/FDMA digital multiple channel per carrier
(MCPC) system is determined by the following steps 2 .
1) Determine the composite carrier-to-noise density available on the RF link
T
(C N 0 )T .
2) Determine the required carrier-to-noise ratio (C N )t required to support each
individual MCPC carrier at the desired BER
DO

⎛E ⎞
, ⎛⎜ ⎞⎟ = ⎜⎜ b ⎟⎟ + Rb − WN + M i + M A , dB ,
C Eb Rb C
= . (1)
N N 0 WN ⎝ N ⎠t ⎝ N 0 ⎠t
where Rb is a data rate of digital signal, in dB; WN - noise bandwidth of carrier, in dB;
⎛ ⎞
C - carrier power; ⎜⎜ Eb ⎟⎟ - the ratio, required for the threshold BER; M i - MODEM
⎝ N 0 ⎠t
implementation margin, in dB ( ≈ (1 ÷ 3)dB ); M A - Adjacent Channel Interference
margin, in dB ≈ (1 ÷ 2)dB .
The bit rate required to support each channel depends on the specific PCM
baseband formatting: PCM: 64 kbps/voice channel, ADPCM: 32 kbps/voice channel.
The noise bandwidth depends on the carrier modulation

2
Ippolito, Louis J., Satellite communications systems engineering : atmospheric effects, satellite link design, and
system Performance, 2008 JohnWiley & Sons Ltd, 396 pp.

4
 for BPSK WN = (1,2 * TDM bit rate) + 20% ,
TDM bit rate
for QPSK WN = (1,2 * ) + 20% . (2)
2

Y
The coefficient 1,2 account for differences between real and ideal
characteristics of bandpass filter and the 20% factor is included to account for guard
bands.

NO C P
For example, the first level of hierarchy structure for 64kbit / s PCM voice,

CO Y
TDM bit rate is 2,048Mbit / s , then
BPSK WN = (1,2 * 2048) + 0,2(1,2 * 2048) = 2,949 MHz ,

O
2048 2048

OP
QPSK WN = (1,2 * ) + 0,2(1,2 * ) = 1,474MHz .

PY
2 2
D O NO T C
3) Determine the required carrier-to-noise density for each carrier:
⎛ C ⎞ ⎛C⎞
⎜⎜ ⎟⎟ = ⎜ ⎟ + WN , dB (3)
⎝ 0 ⎠t ⎝ N ⎠t
N
4) Compare (C N 0 )t with the total available RF link carrier-to-noise
ratio (C N 0 )T from 1) to determine the number of carriers, n p , that can be supported:
⎛C ⎞ ⎛C ⎞
⎜⎜ ⎟⎟ = ⎜⎜ ⎟⎟ + 10 log n p (4)
DO NO

⎝ N 0 ⎠T ⎝ N 0 ⎠t
T
From equation (4) it can be determinate the number of carriers, which can be
supported:
⎛ C ⎞ ⎛ C ⎞
⎜⎜ ⎟⎟ −⎜⎜ ⎟⎟
T
⎝ N 0 ⎠T ⎝ N0 ⎠t
n p = 10 10
(5)
This result (5) has to round to the next lowest integer. It gives the power
limited capacity of the system.
DO

5) Determine the bandwidth limited capacity of the system, nW , from

wTR
nW = , (6)
wN
where wTR is a satellite transponder bandwidth, wN - noise bandwidth (including guard
bands).
The result (6) (rounded to the next lowest integer), gives the bandwidth limited
capacity of the system.
6) Determine the system capacity, C, from the lower of the values of n p or nW :
C = n p or nB (7)
The result C from (7) is the maximum number of carriers that can be used in
the PCM/TDM/PSK/FDMA link, within the power and bandwidth limitations of the
system.

5
 2.2. Frequency Division Multiple Access with PCM/SCPC/PSK application.
PCM/SCPC/PSK is a digital baseband single carrier per channel (SCPC)
system used for data and voice applications. It is not necessity of multiplexing. Each
incoming signal is analog to digital converted, and modulated RF carrier by BPSK or

Y
QPSK for transmission over the satellite channel. A pair of channel frequencies is
used for voice communications, one for each direction of transmission.
One advantage of this SCPC FDMA approach is that it can operate as a

NO C P
CO Y
demand assignment access, where the carrier is turned off when not in use. The
system can also use voice activation, which makes use of the statistics of voice
conversations to share the SCPC carrier with multiple users.

O
OP
It is known, that a typical voice channel conversation is active only about 40%

PY
of the time in any one direction. A voice activation factor (VA) is used to quantify
the improvement possible in the network. For example, a 36MHz transponder has a
D O NO T C
bandwidth limited capacity of 800 SCPC channels, using 45 KHz channel spacing:
36.103 / 45 = 800 . The 800 channels correspond to 400 simultaneous conversations.
The capacity calculations for PCM/SCPC/PSK/FDMA are similar to the
procedure, which is discussed in the previous section 1.1. Single channel carrier noise
bandwidths and data rates are used to determine performance. The VA factor, if used,
is applied to increase the power limited capacity, n p , (4)). The bandwidth limited
capacity, nW , (5) is determined from
DO NO
T
wTR
nB = , (8)
wC
where wTR is a satellite transponder bandwidth, wC - individual channel bandwidth
(including guard bands).
T
The lower value of n p or nW determines the capacity C of the system.
The representative FDMA wireless cellular standards include Advanced
DO

Mobile Phone System (AMPS) in the United States, Nordic Mobile Telephones
(NMT) in Europe, and Total Access Communications System (TACS) in the United
Kingdom 3 . Long Term Evolution 4 (LTE) implements Orthogonal Frequency
Division Multiplexing (OFDM) for its downlink and Single-Carrier Frequency-
Division Multiple Access (SC-FDMA) for its uplink.

3. Time Division Multiple Access

Time Division Multiple Access (TDMA) is another multiple access technique
employed in the digital wireless communication systems. This method has been
applied in the late 1980s. The frequency band for the each station is the same, and the
separation of the carriers in the transponder is done in time (time intervals), and only

3
Kiseon Kim, Insoo Koo, CDMA Systems Capacity Engineering. Artech House INC., 2005, 201 pp.
4
3rd Generation Partnership Project Long Term Evolution – standard improving Universal Mobile Telecommunications
System (UMTS)

6
 one user is allowed to either transmit or receive the information data in each interval
(slot).
Figure 3 shows a principle of the TDMA process as three ground stations
accessing a single frequency translation satellite transponder. Defined time frames are

Y
formatting all the transmission time. Each frame is divided into time intervals (time
slots), which are at least as much as possible access. Equal time intervals in each
frame are designed for individual stations. Each station can transmit in a specific time

NO C P
CO Y
interval (slot) for its uplink transmission of a burst of data in sequence and each
station has use of the full transponder bandwidth during its time slot. The time slot
can be changed on demand. Guard bands are used between the time slots to avoid

O
OP
interference.

PY
f frame Time
D O NO T C
slot
fra
m e
ES 3

ES 3
ES 1

ES 2

ES 1

ES 2

ES 1
{
{
{

slot 1 slot 2 slot 3 t ES 2

ES 3

Figure 3. Principle of the Time Division Multiple Access

DO NO
T
The information that has been sent in a time slots is called sequence. During
an active connection the information is transmitted as sequence, not continuously. At
each time frame can be transmitted synchronization, control and user information.
TDMA is a convenient method for transmitting digital data because by using digital
T
buffer memory it is converted into time slots and then the receiver restores them. The
required capacity of the buffer is determined by:
cBuf = rb .t F (9)
DO

where rb is a bite rate, t F is a period of the time slots.

Downlink transmission consists of interleaved set of packets from all the
ground stations. A receiving station detects and demultiplexes bursts and delivers
information to the users. A reference station is used to establish the synchronization
reference clock and provide burst time operational data to the network.
Figure 4 shows a typical time frame. The range of the frame is from 1 to 20 ms.
Each station burst contains a preamble and traffic data. The preamble contains
synchronization and station identification data. The reference burst, from the
reference station, is usually at the start of each frame, and provides the network
synchronization and operational information.
Station bursts do not need to be identical in duration. Its can be longer for
heavier traffic stations or during higher use periods. The specific allocation of burst
times for each of the stations within the frame is called the burst time plan. The burst
time plan can be changed to adapt for changing traffic patterns.

7
 Gard bands

TDMA frame

Reference station station station

Y
burst 1 2 n

Service

NO C P
User information
information

CO Y
Frequency
Gard Service
and time Identification

O
band channel
control

OP
Figure 4. Example of the typical time frame.

PY
D O NO T C
Advantages of TDMA:
- TDMA can easily adapt to transmission of data as well as voice
communication; it has an ability to carry 64 kbps to 120 Mbps of data rates and this
allows the operator to do services like multimedia and videoconferencing, fax, voice;
- TDMA provides no interference from simultaneous transmissions because it
separates users according to time;
- TDMA provides users with an extended battery life, since it transmits only
DO NO

portion of the time during conversations;

T
- less stringent power control due to reduced interference between users;
- TDMA is the most cost effective technology to convert an analog system to
digital.
Disadvantages of TDMA:
T
- higher synchronization overhead;
- the user might be disconnected - because of the users has a predefined time
slot when they moving from one cell site to other, if all the time slots in this cell are
DO

full;
- the user could not receive a dial tone, if all the time slots in the cell in which
the user is currently in are already occupied;
- multipath distortion - to overcome distortion, a time limit can be used on the
system, once the time limit is expired the signal can be ignored.

3.1. Time Division Multiple Access with PCM/TDM/PSK application.

One of the most common structures found applications in VSAT networks,
consists of a PCM or ADPCM baseband formatting technique (period of the frame
has to be multiple of 125μs ), TDM multiplexing and modulation by binary PSK or
quadrature QPSK. The information is uploaded to one carrier, covering the entire
transponder bandwidth. The data rates are usually 60Mbps when the bandwidth is
36 MHz or 130Mbps when the bandwidth is 72 MHz . The synchronization of the frame

8
 is maintained by using the codeword includes a sequence of 24 to 48 bits. The
codeword is repeated in each frame.
The preamble consists of some components, each with a specific purpose in the
TDMA process. Typical components of the TDMA preamble and reference burst are

Y
summarized in Table 1 for a representative operational system, the INTELSAT
TDMA system, deployed on many early INTELSAT satellites 5 . The INTELSAT
TDMA system consists of two reference bursts per frame and operates with a 2 ms

NO C P
CO Y
frame period.
Table 1. Intelsat TDMA preamble and reference burst structure.

O
Components Description Number

OP
of bits

PY
Preamble
CBR Carrier and synchronizing signal for detector 352
D O NO T C
bit-timing recovery
UW Unique word burst code word 48
TTY Teletype operational data communications between 16
stations
SC Service channel carries network protocol and alarm messages 16
VOW Voice order wire voice communications between stations 2*64
(*2)
Reference Burst
DO NO

CBR Carrier and synchronizing signal for detector 352

T
bit-timing recovery
UW Unique word burst code word 48
TTY Teletype operational data communications between 16
stations
T
SC Service channel carries network protocol and alarm messages 16
VOW Voice order wire voice communications between stations 2*64
(*2)
CDC Coordination and used to transfer acquisition, synchronization, 16
DO

delay channel control, and monitoring info to stations

Total number of bits in the preamble is 560, the reference burst consists 576
bits.

3.2. TDMA Frame Efficiency

The efficiency of the TDMA frame is defined as the ratio of number of bits
available for traffic and total number of bits in frame 6

5
Dennis Roddy, Satellite communications, McGraw-Hill Professional, 2001, 569 pp.
6
Ippolito, Louis J., Satellite communications systems engineering: atmospheric effects, satellite link design, and system
Performance, JohnWiley & Sons Ltd, 2008, 396 pp.

9
 number of bits available for the traffic
ηF = =
total number of bits in the frame
(10)
number of overhead bits
= 1−

Y
total number of bits in the frame

nr .br + nt .bp + (nr + nt )bg

ηF = 1 − , (11)
rT .t F

NO C P
CO Y
where t F is the TDMA frame time, in s; rT is the total TDMA bit rate, in bps; nr -
number of reference stations; nt - number of traffic bursts; br - number of bits in

O
reference burst; bp - number of bits in the preamble; bg - number of bits in guard band.

OP
PY
The frame efficiency shows how mach of the frame is in used. It will change
when increase the total number of bits (a longer frame time) or lowering the overhead
D O NO T C
in the frame (no traffic bits). The optimum operating structure occurs by providing
the longest possible frame time with the lowest total number bits allocated to
overhead functions, but the longest frame time needs of largest memory buffers,
which increases the overall delay of the signal.

Example 1. How calculate the frame efficiency, if TDMA frame has a frame shows in Table
1 with length 2ms , a total number of bits in the preamble is bp = 560bits , in the reference burst -
br = 576bits , number of bits in guard band - bg = 206bits , there are two reference stations, each
DO NO
T
transmitting a reference burst in the frame nr = 2 .
Evaluate the TDMA network in terms of the maximum number of traffic terminals and the
operating TDMA date rate for a desired minimum frame efficiency of η F = 0,95 .
From Equation 11 the frame efficiency is
T
n .b + n .b + (nr + nt )bg 2.576 + nt .560 + (2 + nt )206 1564 + 766nt
ηF = 1 − r r t p = 1− = 1− .
rT .t F rT .0,002 rT .0,002
1) If the TDMA data rate is set at 120 Mbps, the number of terminals that could be
DO

supported at a 95% frame efficiency is

1564 + 766nt
0,95 = 1 − ⇒ nt = 13,6; nt ≈ 13 traffic terminals.
120.106.0,002
2) for a network with a 13 traffic terminals (13 time slots) and η F = 0,95 it can be achieved
data rate of
1564 + 766.13
0,95 = 1 − ⇒ rT = 115,22Mbps .
rT .0,002
The result chows, if the number of traffic terminals is fixed, network performance can be
optimized by setting the data rate rT at the minimum value to achieve the desired efficiency η F .
Therefore the network ground terminals can operate with variable TDMA data rates.

3.3. TDMA capacity

To evaluate a capacity for any type of data source bit stream (voice, video, data
or combination of these) the network channel capacity for a TDMA network can be
evaluated in terms of an equivalent voice-channel capacity, nC . The equivalent voice

10
 channel capacity is defined as ratio of available information bit rate, ri and equivalent
voice channel bit rate, rC
ri
nC = (12)

Y
rC
The available information bit rate, ri , represents that part of the total bit rate
available for information (the total bit rate minus the bit rate allocated to overhead

NO C P
CO Y
functions). The equivalent voice channel bit rate is usually defined as the standard
PCM bit rate rC = 64kbps .

O
Order for determination of TDMA capacity:

OP
1) Determine the composite carrier-to-noise ratio available on the RF link

PY
(C N 0 )T .
2) Determine the carrier-to-noise ratio required to achieve the threshold BER
D O NO T C
desired for the TDMA network
⎛ C ⎞ ⎛⎜ Eb ⎞⎟
⎜ ⎟ =⎜ + RT − WN + M i + M A , dB , (13)
⎝ N ⎠t ⎝ N 0 ⎟⎠t
where RT is the TDMA data rate at the desired frame efficiency, η F , in dB; N 0 -
power flux density for a unit bandwidth, in dB; WN - noise bandwidth of carrier, in
⎛ Eb ⎞
DO NO

dB; C - power of carrier, in dB; ⎜⎜ ⎟⎟ - the ratio, required for the threshold BER; M i
⎝ N 0 ⎠t
T
- modem implementation margin (the deviation of modem performance from the
ideal), in dB, ≈ (1 ÷ 3)dB ; M A - adjacent channel interference margin ≈ (1 ÷ 2)dB .
3) Determine the TDMA data rate from
T
⎛C⎞ ⎛C⎞
⎜ ⎟ ≥⎜ ⎟ . (14)
⎝ N ⎠ t ⎝ N ⎠T
4) Determine the TDMA capacity from (12)
DO

The frame parameters for the reference bursts, traffic bursts, and guard bands
defined are: t F - the TDMA frame time, in s; rT - the total TDMA bit rate, in bps; nr -
number of reference stations; nt - number of traffic bursts; br - number of bits in
reference burst; bT - number of bits in total TDMA frame; bp - number of bits in the
preamble; bg - number of bits in guard band.
Therefore, it can determinate the follow bit rates:
bT
Total TDMA Bit Rate : rT =
tF
bp
Reference Burst Bit Rate : rp =
tF
b
Reference Burst Bit Rate : rr = r
tF

11
 bg
Guard Time Bit Rate : rg =
tF
The available bit traffic rate is

Y
ri = rT − nr ( rr + rg ) − nt ( rp + rg ) (15)
Therefore
ri rT − nr (rr + rg ) − nt (rp + rg )

NO C P
nC = = (16)

CO Y
rC rC
This result provides the number of equivalent voice channels that can be

O
supported by the TDMA network for the specified TDMA bit rate, TDMA frame

OP
efficiency, and frame parameters.

PY
Example 2. To calculate the TDMA capacity in number of equivalent voice channel nC
D O NO T C
with data from Example 1. There is standard PCM bit rate rC = 64kbps and first level of digital
plesiochronous hierarchy structure rC = 2048kbps .
First define the following data rates in bps:
b
Total TDMA Bit Rate : rT = T = 115,22Mbps ;
tF
b 560
Reference Burst Bit Rate : rp = p = = 280kbps ;
t F 0,002
DO NO

576
T
Reference Burst Bit Rate : rr = br = = 288kbps ;
t F 0,002
b 206
Guard Time Bit Rate : rg = g = = 103kbps .
t F 0,002
T
The available bit traffic rate
ri = rT − nr (rr + rg ) − nt (rp + rg ) = 115,22.106 − 2.(288+103).103 −13.(280+103).103 = 112,818Mbps
For standard PCM bit rate rC = 64kbps the number of equivalent voice channel is:
DO

ri 112,818.106
nC = = = 1762,8; ⇒ nC = 1762 channels .
rC 64.103
For bit rate rC = 2048kbps the number of equivalent voice channel is:
ri 112,818.106
nC = = = 55,1; ⇒ nC = 55 channels .
rC 2048.103
This number of equivalent voice channels can be supported to maintain a frame efficiency of
η F = 0,95 .

3.4. Switching in satellite TDMA.

Switching in satellite TDMA (Satellite Switched TDMA) lies in the rapid
reconfiguration of the diagram of satellite antenna to improve the possibilities for
providing additional access to traditional TDMA system. Satellite Switched TDMA
adds antenna beam switching to provide additional multiple access capability to adapt
to changing demand requirements. The on-board switching is accomplished at

12
 intermediate frequency with an n× n switch matrix. Switching is done in
synchronization with the TDMA bursts from the ground stations.
Figure 5 shows example for configuration of a 3× 3 satellite switched TDMA

Y
architecture. The network consists of three regional beams, designated as A, B, and C
beams. The switch matrix mode is shown on the right of the figure, labeled Mode 1,
Mode 2, or Mode 3. The options for switching in this case are nS = 6 (showing only

NO C P
three of them). Table 2 shows the number of full versions of switching time for this

CO Y
example.

O
OP 3

PY
D O NO T C

2
1
A1 A2 B1 B2 C1 C2

Figure 5. Example of a 3× 3 satellite switched TDMA architecture

DO NO
T
Table 2. The number of full versions of switching time for 3× 3 satellite switched TDMA
architecture.
Downlink beam
Uplink
Switch Switch Switch Switch Switch Switch
T
beam
position 1 position 2 position 3 position 4 position 5 position 6
А В С А А В С
В А В С В С А
DO

С С А В С А В

For N number of Regions (beams) nS = N! . As N increases, the number of

switch positions becomes quite large, for example:
N = 3 : nS = 6 ;
N = 4 : nS = 24 ;
N = 5 : nS = 120 …
The switch consists of coaxial cables or waveguides. Switching elements are
ferrites, diodes or FET transistors and dual-gate FET. The architecture is usually
limited to three or four beams. For larger matrices grow much weight and dimensions
of the switching matrix.

13
 The major TDMA standards contain Global System Mobile (GSM) in Europe
and Interim Standard 54/136 (IS-54/136) in North America 7 . GSM was developed in
1990 for second generation (2G) digital cellular mobile communications in Europe.
Systems based on this standard were first deployed in 18 European countries in 1991.

Y
By the end of 1993, it was adopted in nine more European countries, as well as
Australia, Hong Kong, much of Asia, South America, and now the United States.

NO C P
CO Y
4. Code Division Multiple Access
The third technique Code Division Multiple Access (CDMA) uses basic
principles of the two previous methods. It is a combination of both frequency and

O
OP
time separation. The station has all the bandwidth as in TDMA all in the time, as in

PY
FDMA.
CDMA, and. The main characteristics of the CDMA are:
D O NO T C
- Works with digital formatted data only;
- Offers the highest power and spectral efficiency operation of the three
fundamental techniques;
- All users use same frequency and may transmit simultaneously;
- Narrowband message signal multiplied by wideband spreading signal, or
codeword;
- Each user has its own pseudo-codeword (orthogonal to others) and receivers
DO NO

detect only the desired codeword, all others appear as noise.

T
In this method, each signal is transmitted to the satellite using the same carrier
frequency and occupies the entire bandwidth of the transponder, but carriers differ in
the way of coding, i.e. they are modulated with a separate code. The receiver on the
other side must recognize this particular code; otherwise it can not choose the desired
T
signal. Code modulation used in CDMA is different from channel coding and
modulation.
DO

each uplink station is assigned

f time slot and frequency band

{
in coded sequence
band 3 ES 3 ES 1 ES 2
slot

{
fra
m e
band 2 ES 2 ES 3 ES 1

band 1 { ES 1 ES 2 ES 3

ES 3
ES 1

t
slot 1 slot 2 slot 3 ES 2
downlink receiver must know
code to detect the signal

Figure 6. Principle of the Code Division Multiple Access

Figure 6 shows the principle of the Code Division Multiple Access. At each
station is given time interval (time slot) and bandwidth. The information is sent to the

7
Rappaport, T. S., Wireless Communications, Englewood Cliffs, NJ: Prentice-Hall, 2002.

14
 satellite (“uplink”) as a coded sequence as a series. From satellite to terminals
(“downlink”) information packages are interleaved, how it is shown in Figure 6. The
ground station must know the number and frequency of its time interval to adopt its
number of information packets. The ground station separates useful signal from the

Y
noise.
The used code sequence has to meet two conditions: to prevent unauthorized
access and to be short enough because there can be problems with synchronization or

NO C P
CO Y
delay of the signal. The most appropriate code sequence that satisfies these conditions
is a pseudorandom sequence of binary symbols. It is a binary sequence with a final
length in which the position of each bit is random. The autocorrelation of the pseudo-

O
OP
random sequence is similar to the autocorrelation of white noise. Therefore, this

PY
sequence is called pseudonoise, PN. What is the difference between random and
pseudorandom signal? A random signal cannot be predicted, its future value can be
D O NO T C
described statistical. A pseudorandom signal is not random at all; it is defined as
deterministic periodic signal that is known to both transmitter and receiver. It must
comply with three conditions:
- Balance property – in each period of the sequence the number of binary ones
differs from the number of binary zeroes by at most one digit;
- Run property - the pseudorandom sequence consist sequences of a same type
of binary digits (for example, only ones or only zeroes); the appearance of the
DO NO

opposite digit in a sequence starts a new run; the length of the run is equal of the
T
number of digits in the run;
- correlation property – if the period of the sequence is compared term by term
with any cyclic shift of itself, the number of agreements has to differ from the
number of disagreements by not more than one count.
T
The PN sequence used in CDMA systems is generated using sequential logic
circuits and a n-stage feedback shift register. Figure 7 shows an example of a 4-stage
feedback shift register and table of contents in order of a clock, used to generate the
DO

PN sequence. At each clock pulse the binary sequences are shifted through the shift
registers one stage to the right. The feedback logic consists of exclusive-OR gates
generated by a unique algorithm. The output of the stages are logically combined and
fed back as input, generating a PN sequence at the final output. Since the last state
1000 corresponds to the initial one, it is obviously that the register repeats foregoing
sequence after 15 clock times. The output sequence of the example on Figure 7 is
000111101011001, where the left digits are the earliest.
The output sequences can be classified as either maximal length (ML) or non
maximal length (NML). The number of non-zero states that are possible for this
linear PN sequence generator, called its maximal length will be
p = ML = 2 n − 1 . (17)
The sequence has a maximal length if for an n-stage linear feedback shift
register the sequence repetition period p is (17).

15
 Register 1 Register 2 Register 3 Register 4 Output
A Q1 A Q1 A Q1 A Q1 pseudorandom
D Q4 D Q4 D Q4
+ D Q4
sequence

Y
000111101011001
ENB ENB ENB ENB

1 1 0 0 0
2 0 1 0 0
3 0 0 1 0

NO C P
4 0 0 0 1

CO Y
5 1 0 0 1
6 1 1 0 1
7 1 1 1 1
8 1 1 1 0

O
9 0 1 1 1

OP
10 1 0 1 0
11 0 1 0 1

PY
12 1 0 1 1
13 1 1 0 0
14 0 1 1 0
D O NO T C
15 0 0 1 1
16 1 0 0 0
clock
Contents of registers

Figure 7. An example of a 4-stage feedback shift register and table of contents in order of a
clock

It can test the sequence for the randomness properties. First is balance property.
DO NO

The sequence has seven zeroes and eighth ones, consequently the sequence meets the
T
balance condition. Second condition, the sequence consists of regularly repeated
same number (in Fig. 7 they are separated by vertical lines). The number of digits is
eight. Four of them (50%) have one digit, two (25%) - two digits, and the other with
more than two digits. So the run property is met. Third condition: It was found that
T
normalized autocorrelation function of periodic pulse signal with a period ML = 2n − 1
is as follows:
1
R (τ ) = .Δk , (18)
DO

ML
where Δk is a difference between the number of agreements and the number of
disagreements when comparing part of the sequence and its shifted copy.
It is obviously, when τ = 0 (the signal and its copy match) autocorrelation
function is one R(τ ) = 1 . After shifting the copy off less than ML autocorrelation
1
function is R(τ ) = − . A comparison of the signal and its copy shifted by one
ML
register to the right gives the following result:

0 0 0 1 1 1 1 0 1 0 1 1 0 0 1
1 0 0 0 1 1 1 1 0 1 0 1 1 0 0
no yes yes no yes yes yes no no no no yes no yes no

There are seven agreements and eight disagreements. From Equation (18)

16
 1 1
R(τ ) = .(7 − 8) = −
15 15
So, the third condition is done. If the copy of the signal is shifted by two or
more register (but not more than 15) is obtained the same result. The synthesis of an

Y
appropriate structure of pseudorandom sequence generator depends on the meeting
the conditions for randomness. Not every structure can maximize the length and
satisfy these randomness properties.

NO C P
The generated PN sequence of binary symbols is combined with digital's data

CO Y
to obtain the number of pseudorandom sequence of binary data. Figure 8 shows the
process used to generate the PN data stream. Two digital signals are entering in the

O
modulator: the input digital data stream m(t ) at a rate rb = 1 tb and pseudorandom

OP
PY
sequence of binary symbols pPN (t ) at a chip rate rCH = 1 tCH . A PN cycle is called chip
clock.
D O NO T C
PN chip clock
A A xor B
t B
1 bit
period
input output
0 1 0 0 A B A xor B
m(t) 0 0 0
0 1 1
DO NO

1 0 1
1 0 1 1 0 0 1 00 1 1
chip 1 1 0
T
period pPN(t)

1 01 1 0 0 11
s(t)
T
Figure 8. An example of the generation a PN sequence of data

The PN sequence is modulo-2 added to the data sequence m(t) to produce the
data stream
DO

s (t ) = m(t ) ⊕ pPN (t ) . (19)

Figure 8 shows that the PN data stream is at the chip rate rCH = 1 tCH , which is
higher than the original data rate rb = 1 tb , that is rCH >> rb . In the example shown on
the figure rCH = 6rb .
The condition rCH >> rb is very important for the successful implementation of
CDMA. It is the reason that CDMA is often referred to as spread spectrum or spread
spectrum multiple access, because the original data sequence is “spread” out over a
much greater frequency band in the transmission channel.
The chip rate is selected so as to extend the signal over the total available
channel bandwidth. Large spreading ratios are typical, for example, in mobile
satellite voice networks, the original 16 kbps voice data stream may be spread at a
chip rate to produce a PN data stream that operates over an 8MHz RF channel

17
 bandwidth. This is a spreading factor of 500, assuming 1 bit/Hz modulation such as
BPSK.
What is Spreading Factor, SF? The ratio of both chip and bit rates indicates
how many times spectrum of producing data stream has extended compared with the

Y
input digital data stream and it is called Spreading Factor or a processing gain when
is calculated in dB:
⎛r ⎞

NO C P
rCH
SF = ; gain BPSK = 10 log ⎜⎜ CH ⎟⎟ , (20)

CO Y
rb ⎝ rb ⎠
because for a 1bps / Hz modulation system, such as BPSK, WRF ≈ rch ; chip rate

O
rCH = 1 tCH , data rate rb = 1 tb , tb = n.tch .

OP
PY
For example shown on the figure 8 rCH = 6rb ,
rCH 6rb
SF = = = 6,
D O NO T C
rb rb
⎛r ⎞ ⎛ 6r ⎞
gain BPSK = 10 log⎜⎜ CH ⎟⎟ = 10 log⎜⎜ b ⎟⎟ = 10 log 6 = 7,78 dB .
⎝ rb ⎠ ⎝ rb ⎠
For example, the standard IS-95, which was introduced as a second generation
(2G) wireless cellular communications in 1990 in North America, and later in Japan
and Korea, has a chip rate rch = 1,2288 and spreading factor SF = 64 .
From equation (19) follow spreading the bandwidth needed for transmission of
DO NO

information. Therefore, this access method is often called spread spectrum multiple
T
access. The spread spectrum is defined as a communication method in which the
signal occupies a bandwidth, greater than the minimum required to transmit
information and that spreading code is independent of the transmitted information.
T
There are three basic types of CDMA:
- Direct Sequence Spread Spectrum, DS-SS - digital data directly recode with
pseudo-random sequence of binary symbols (Fig. 8);
- Frequency Hopping Spread Spectrum, FH-SS - change of carrier frequency,
DO

i.e. useful signals are transmitted on different carrier frequencies generated pseudo
manner;
- Time Hopping, TH - intermittent switching time interval defined by pseudo-
random sequence.
In commercial communications are primarily used the first two methods. The
third method, TH-CDMA is used in cases of intentional spread of disturbance as
possible to hide the details of the signal from the enemy.

4.1. Direct Sequence Spread Spectrum

The structure of the system with direct sequence spread spectrum (DS-SS) is
shown in Figure 9. It includes a phase modulator, PN code modulator, balanced
demodulator with bandpass filter in the receiver and demodulator phase.

18
 sm(t) smp(t)
Data in Phase PN Cod Balanced Phase Data
m(t) Mod Mod Demod Demod out

Y
pPN(t) pPN(t)

2 E cos(ω 0 t ) PN PN 2 E cos(ω 0 t )

NO C P
CO Y
Figure 9. A structure of the system with direct sequence spread spectrum (DS-SS)

The information bitstream is phase modulated onto a carrier (phase modulator),

O
OP
then enter in the PN Code Modulator which phase modulates the RF carrier to

PY
produce the spread signal. After passing through the satellite channel, the signal is
‘despread’ with a balanced modulator, then phase demodulated to produce the
D O NO T C
original data bitstream.
After phase modulation with a carrier in Phase Modulator input data stream
m(t ) with the rate rb = 1 tb and bandwidth WRb is as follows:
s m (t ) = 2 E cos[ω 0 t + ψ m (t )], 0 ≤ t ≤ T ,
where ψ (t ) is the information bearing phase modulation.
The PN Code Modulator phase modulates the data modulated signal sm (t ) with
the PN sequence pPN (t ) with chip rate rCH . The output of the PN modulator is
DO NO
T
smp (t ) = 2 E cos[ω 0 t + ψ m (t ) + ψ p (t )] , (21)
Thus, the signal to be transmitted to the receiver over the satellite is spread in
frequency by the PN sequence to a bandwidth of WR . CH
T
It is known that ideal suppressed carrier binary phase shift keying (BPSK)
phase is taking M = 2 discrete values, i.e. it switches to 1800 according to a logic “0”
and “1”. Then the expression (21) can be written as a multiplication of the carrier and
m(t ) information antipodes impulses with a values + 1 or − 1 :
DO

sm (t ) = 2 E .m (t ) cos(ω 0 t ) . (22)
The PN sequence pPN (t ) is also antipodal pulse stream with a values + 1 or − 1 ,
Equation (22) can be written as
smp (t ) = 2 E .m (t ). p PN (t ) cos(ω 0 t ) . (23)
The phase of the carrier equals ψ m (t ) +ψ p (t ) = 1800 , if the modulo 2 sum of data
and code is binary 1 ( m (t ) + p PN (t ) = 1 ); and the phase of the carrier is zero
ψ m (t ) +ψ p (t ) = 0 0 when the modulo 2 sum of data and code is binary 0 ( m (t ) + p PN (t ) = 0 ).
Figure 10 shows the structure according to Equation (23)

19
 s*mp(t) s*m(t) m*(t)
sm(t) smp(t)
Data in Phase
m(t) + Mod + Phase
Mod
BPF

Y
pPN(t) pPN(t)

2 E cos(ω 0 t ) 2 E cos(ω 0 t )

NO C P
а) b)

CO Y
Figure 10. An elements of the modulator (a) and demodulator (b) of the direct sequence
spread spectrum system

O
OP
The received spread signal is multiplied by a stored replica of pPN (t ) (Fig.

PY
10,b). The output of the balanced demodulator is then
s *mp (t ) = 2 E .m (t ). p PN
2
(t ) cos(ω 0 t ) . (24)
D O NO T C
Since the pPN (t ) is a binary signal, therefore, pPN
2
(t ) = 1 and
s *m (t ) = 2 E .m (t ). cos(ω 0 t ) (25)
or
sm* (t ) = 2 E cos[ω 0 t + ψ *m (t )] , (26)
and the information ψ m* (t ) can be recovered through the final phase demodulator.
DO NO

If the receiver PN code differs from that of the transmitter, random phase
T
modulation occurs and spread spectrum signal looks as noise. The bandpass filter
(with bandwidth WRb ) removes high frequency sidebands.
Figure 11 shows an example of the direct sequence spread spectrum
modulation and demodulation, shown in Figure 10.
T
In Fig. 11a, b are shown input data sequence m(t ) with the bit rate rb = 1 tb and
bandwidth WRb , and pseudorandom sequence of binary symbols pPN (t ) at a chip rate
DO

rCH = 1 tCH . In Fig. 11c is shown the result of modulo-2 addition of the data sequence
m(t ) and code sequence pPN (t ) . Fig. 11d shows the value of the phase of spread
spectrum signal that is π if the modulo 2 sum of data and code is binary 1; and it is a
zero when the modulo 2 sum of data and code is binary 0.
The demodulation of a signal begins with dispreading of the spectrum by
multiplying the received signal with a synchronized replica of the code ψ *p (t ) as the
antipodal phase shift.
The carrier phase is determined as a result of modulo-2 addition of d) and e)
(Fig. 11f). Fig. 11g shows the result of recovering the data waveform by the use of
BPSK demodulator.

20
 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0
1
a) m(t)
t

Y
-1
1
1 0 1 1 0 1 0 0 0 1 1 0 0 1 1 1 0 1 0 0 1 1 0 0
b) pPN(t)
t
-1

NO C P
CO Y
0 1 0 0 1 0 1 1 1 0 0 1 0 1 1 1 0 1 0 0 1 1 0 0
c) m(t).pPN(t)
t

O
d) ψ m (t ) +ψ p (t ) 0 π 0 0 π 0 π π π 0 0 π 0 π π π 0 π 0 0 π π 0 0

OP
t
ψ *p (t ) π π π 0 π 0 0 0 π π 0 0 π π π 0 π 0 0 π π 0 0

PY
e) 0
t
ψ m* (t ) π π π π π π π π π π π π
f) 0 0 0 0 0 0 0 0 0 0 0 0
D O NO T C
t
1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0
1
g)
m*(t)
t
-1

Figure 11. An example of the direct sequence spread spectrum modulation and
demodulation
DO NO
T
4.2. Frequency Hopping Spread Spectrum.
The main point in Frequency Hopping Spread Spectrum, FH-SS, is that the
carrier frequency is changed in accordance of a PN sequence, producing a sequence
of modulated data bursts with time varying pseudorandom carrier frequencies.
T
The all possible carriers frequencies available for frequency hopping in FH-SS
is called the hop set. Each of the hopped channels contains adequate RF bandwidth
for the modulated information, usually a form of frequency shift keying (FSK).
DO

There are two bandwidths defined in FH-SS operation:

- Instantaneous Bandwidth, wbb – the baseband bandwidth of the channel used
in the hopset.
- Total Hopping Bandwidth, wRF – the total RF bandwidth over which hopping
occurs.
As larger the ratio of wRF to wbb , so better the spread spectrum performance of
the FH-SS system. Figure 12 shows the elements of the FH-SS satellite system.

21
 Data Carrier Trans Carrier Data
MFSK Receiver MFSK

Y
in Mod mitter Demod out
fC
sequence
Frequency Frequency Synchronization
Syntesizer System

NO C P
Syntesizer

CO Y
pPN(t) pPN(t)

PN PN

O
OP
Figure 12. The elements of the FH-SS system.

PY
This technique commonly used M-ary frequency shift keying MFSK where
D O NO T C
k = log 2 M information bits are used to determine which frequency of all M is to be
transmitted. There are two steps of modulation - data modulation and frequency
hopping modulation. The data modulated signal is PN modulated with a PN sequence
of carrier frequencies, f C , generated from the PN sequence pPN (t ) . The frequency-
hopped signal is transmitted through the satellite channel. Receiver reverses the
signal processing steps. The received signal is frequency hopping demodulated in the
demodulator using a stored replica of the PN sequence. Then the dehopped signal is
DO NO

demodulated by the data demodulator to develop the input data stream. Commonly in
T
used is noncoherent detector because it is difficultly to keep the phase coherence in
the frequency hopping.

Example 3. A hopping bandwidth is 840 MHz and frequency step size is Δf = 100 Hz .
T
What is the minimum number of PN chips required for each frequency word?
W 840.106
The number of frequency are M = RF = = 84.105 ;
Δf 100
DO

Minimum number of chips = log 2 84.105 = 23 .

22
 References
1. Bernard Sklar, Digital Communications: Fundamentals and Applications,
Second Edition, Prentice-Hall, 2001.

Y
2. Broadcom Corporation and Cisco Systems, Digital Transmission: Carrier-
to-Noise Ratio, Signal-to-Noise Ratio, and Modulation Error Ratio.
http://www.broadcom.com/docs/general/Broadcom-Cisco_CNR-SNR-MER.pdf

NO C P
CO Y
3. Gilhousen, K. S., et al., On the Capacity of a Cellular CDMA System, IEEE
Trans. On Vehicular Technology, 1991, pp. 303–312.
4. Dennis Roddy, Satellite communications, McGraw-Hill Professional, 2001,

O
OP
pp.569.

PY
5. Intuitive Guide to Principles of Communications, CDMA: The Concept of
signal spreading and its uses in communications, www.complextoreal.com
D O NO T C
6. Ippolito, Louis J., Satellite communications systems engineering:
atmospheric effects, satellite link design, and system Performance, JohnWiley &
Sons Ltd, 2008, pp.396.
7. Kiseon Kim, Insoo Koo, CDMA Systems Capacity Engineering. Artech
House INC., 2005, pp.201.
8. Kim, K. I., Handbook of CDMA System Design, Engineering and
Optimization, Englewood Cliffs, NJ: Prentice Hall, 2000.
DO NO

9. Koo, I., et al, Sensitivity Analysis for Capacity Increase on the DS-CDMA
T
System, Proc.of JCCI, 1997, pp. 447–451.
10. Pickholtz R. L., Schilling D. L., Milstein L. B., Theory of Spread Spectrum
Communications: A Tutorial. IEEE Transaction on Communications, Vol. com-30,
No. 5, May 1982.
T
11. Pickholtz R. L., Schilling D. L., Milstein L. B., Revisions to Theory of
Spread Spectrum Communications: A Tutorial. IEEE Transaction on
Communications, Vol. com-32, No. 2, February 1984.
DO

12. Zhili Sun, Satellite networking principles and protocols, John Wiley & Sons,
2005. pp. 342.

Problems and questions

1. What are the similarities and differences between multiple access and multiplexing?
2. What are the benefits of CDMA versus FDMA and TDMA?
3. Determine the maximum number of users that can access the satellite for the
PCM/TDM/PSK/FDMA digital multiple channels per carrier system.
4. Calculate the capacity for PCM/SCPC/PSK/FDMA system.
5. A TDMA system operates on 100 Mbps with a 2ms frame time. Assume that all slots are
equal length and guard lines between slots are 1μs . Calculate the efficiency for the case of 1, 5 and
50 slots per frame. Calculate the efficiency again if it is required 500bits preamble for each slot.
6. The TDMA frame consists of two reference bursts per frame, with a variable number of
traffic bursts, depending on load demand and service area coverage. QPSK modulation (2

23
 bits/symbol) is used, with a total frame length of 120832 symbols. The preamble in each traffic
burst is 280 symbols long, the control and delay channel is 8 symbols, and the guard band interval
is 103 symbols. Calculate the frame efficiency for a frame consisting of 14 traffic bursts per frame.
7. Determine the voice-channel capacity for the TDMA frame of problem 6.The voice

Y
channel is the standard PCM format (64 kbps) with QPSK modulation. The frame period is 2 ms.
Assume a speech activity factor of 1.
8. Describe randomness properties that make pseudorandom sequence to be random.
⎛ ⎞ ⎛ ⎞

NO C P
9. What is the average BER for the network with an ⎜⎜ Eb ⎟⎟ = 3dB and ⎜⎜ Eb ⎟⎟ = 30dB if the PN

CO Y
⎝ N0 ⎠ ⎝ N0 ⎠
sequence for a BPSK direct sequence spread spectrum network operates at 512 chips per symbol
period? The network consists of 150 users with equal received power at the demodulator.

O
OP
PY
D O NO T C
DO NO
T
T
DO

24
 Problems

1. Satellite Communications Segments are:…?

Y
2. Why Geostationary Earth Orbit (GEO) is called Geosynchronous?
3. Find the rang, azimuth angle and elevation angle if the earth station is
located at: (a) Coventry, England; (b) Riga, Latvija; (c) Istambul, Turkey.

NO C P
CO Y
Satellite Longitude is 19.20 E.
4. Which type of satellite orbit provides the best performance for each of

O
the following characteristics: (a) Minimum latency (time delay) for voice and

OP
data networks; (b) Best coverage of high latitude locations; (c) Minimum free

PY
space path loss; (d) Ground terminals with minimum antenna tracking required.
5. The frequencies used for satellite communications are accommodated
D O NO T C
between…?
6. Why communications satellites also employ polarization frequency
reuse?
7. What is a satellite communications link budget?
8. A signal has power 20W, his bit rate is 600bps. What is the bit energy
in decibels?
9. Describe the relation between C N , C N 0 and Eb N 0 .
DO NO

10. The transmitted power of Earth station is 20W , a distance to satellite

T
is d = 37500km . The antenna gain is 45dB . What is the power flux density?
11. How the values of antenna gain is depending on frequency and
antenna’s diameter?
12. The frequency uplink is 6,175GHz, a distance to satellite is
T
d = 37500km . What is the free space path loss?
13. What is the power flux density and power of received signal if both
the transmit and receive parabolic antennas have a diameter of 3 m, the transmit
DO

power is 15 watts, the antenna efficiency is 55% for both antennas. The satellite
is in a GSO location, with a range of d = 37500km . The frequency of operation is
6 GHz.
14. A receiver with a low noise amplifier with a gain of GRF = 23dB and
noise temperature of t RF = 50 K ; a downconverter with a gain of GM = −10dB and
noise temperature of t M = 500 K , an intermediate frequency unit: t IF = 1000 K ,
GIF = 30dB .What is the receiver system noise temperature.
15. Calculate Receiver G/T (dB/K) of a satellite having antenna gain 45 ,
over all receiver noise temperature t S = 70 K .
16. A fixed service satellite uplink operating at 14 GHz consists of a 3m
antenna diameter ground terminal with a 1kW transmitter. The receiver on the
GSO satellite has a 2m antenna and a receive system noise temperature of 500 K.
The free space path loss for the link is 202,5 dB. Assume an efficiency of 55%
 for both antennas. What is the carrier-to-noise density for the uplink under clear
sky conditions with a gaseous attenuation loss of 2,5 dB on the link?
17. There is a satellite with a range d = 37500km ; transmitted power
pT = 2W ; antenna gain gT = 22 dB ; channel system noise temperature t S = 160 K

Y
and bandwidth bN = 500MHz of the satellite channel; Frequency = 12GHz. What
are: (a) Power Flux Density to the received terminal; (b) Received power to the

NO C P
received terminal with antenna effective aperture Aeff = 10m 2 ; (c) Received

CO Y
antenna gain g R ; (d) Received ratio c n .
18. A satellite network operates with a frequency translation transponder

O
and provides a 64 kbps BFSK data link. The requirement for the link BER is

OP
PY
5.10−5 . (a) What is the required composite Eb N 0 for the link? (b) The downlink
C N 0 for the link is 60 dBHz. What uplink C N 0 would be required to maintain
D O NO T C
the BER requirement?
19. What are the main differences, advantages and disadvantages
between: (a) FDMA and TDMA; (b) TDMA and CDMA?
20. A communications satellite transponder with a 42MHz usable
bandwidth operates with multiple FDMA carriers. Each FDMA carrier requires
a bandwidth of 7,5MHz and an EIRP of 16 dBw. The total available EIRP for
the link is 24 dBw. Determine the maximum number of carriers that can access
DO NO

the wireless link if assume 10% guard bands and neglect implementation
T
margins.
21. A direct-sequence spread spectrum CDMA satellite network operates
with 8 kbps voice channels. The interfering noise density on the network is
measured as 6 dB above the thermal noise level. Determine the processing gain
T
for a 6MHz spread spectrum bandwidth.
22. A hopping bandwidth is 820 MHz and frequency step size is
Δf = 50 Hz . What is the minimum number of PN chips required for each
DO

frequency word?
23. The Hamming distance between sequences A = 00101101 and
B = 10110101 is …?. The Hamming weight of a sequence A = 00101101 is…?
24. What is the symbol error correcting capabilities of a RS(7,3) code.
Now many bits are there per symbol?
25. What is the codeword polynomial of a message that consist three
symbols 010 110 111 and has to be coded by RS(7,3) code with the generator
polynomial g ( X ) = α 3 + α 1 X + α 0 X 2 + α 3 X 3 + X 4 ?
26. Use the generator polynomial for the RS(7,3) code to encode the
massage 10110101 .
27. Convolutional codes are commonly characterized by parameters…?
28. Draw the trellis diagram for the K=3, code rate r = 1 / 3 code generated
by g1 ( x) = 1 + x + x 2 , g 2 ( x) = 1 + x 2 and g3 ( x) = x + x 2 .
 29. Describe the generator polynomials of the coder in Figure P1(a) and
(b).
30. What is the output coded message generated by encoders, shown in
Figure P1 (a) and (b)?

Y
m(t) output u1

Multiplexer u
S1 D m input output u1
m(t)

NO C P
information bit
m(n) m(n-1) m(n-2) S2 output u2 u

CO Y
stream
Shift register m(n) m(n-1) m(n-2)
C ENB
synchronization output u3

synchronization
output u2

O
OP
a) b)

PY
Figure P1. Coders of convolutional code
D O NO T C
31. What is a trellis diagram for the encoder given in Figure P1(b) and
input message m = 101100 .
32. Choice the path in trellis diagram for the output coded message
generated by encoder, shown in Figure P1(b).
33. What are the main differences, advantages and disadvantages between
recursive and nonrecursive systematic convolutional encoder.
DO NO
T
T
DO