....................
`
`jjjjjjjj;;;;
`
`.::.
`
`;:
`
`;
`
`.........
`
`;;;:.;
`
`............. :; ..............
`
`; ........
`
`:
`
`.........................
`
`--
`
`__:
`
`........... --:~
`
`-::-.............--::-:-.¢:
`
`:-::-::.
`
`:-:-::-::-.
`
`===================================
`
`~vVIRELES S COMMUNICATIONS
`DESIGN HANDBOOK
`Aspects of Noise, Nterference,
`and Environmental Concerns
`
`:;:;
`
`:;;; ::::::::::::
`
`;;:
`
`::::::;:.
`
`............. ;:.:
`
`.............. ;:;;
`
`.................. :;::~ ................. ;::::
`
`;:.;
`
`, .......
`
`:::;::; ............
`
`VOLUME !' SPACE INTERFERENCE
`
`REINALDO PEREZ
`Spacecraft Design
`Jet Propulsion Laboratory
`California Institute of Technology
`
`ACADEMIC PRESS
`San Diego London
`New York Sydney
`
`Boston
`Tokyo Toronto
`
`Rembrandt Exhibit 2004
`Qualcomm v. Rembrandt
`IPR2020-00510
`
`Page 1 of 22
`
`

`

`~ i s book is printed on acid-free paper.
`
`Copyright © 1998 by Academic Press
`
`All rights reserved.
`No. part of this publication may be reproduced or transmitted in any form or by any
`means, elec~onic or mechanical, including photocopy, recording, or any information
`storage and retrieval system, without permission in writing from the publisher.
`
`ACADEMIC PRESS
`525 B Street, Suite 1900, San Diego, CA 92101~4495, USA
`http://www.apnet.com
`
`Academic Press
`.2.4-28 Oval Road, London ~ I
`http://www~hbuk.co.ukJap/
`
`7DX, UK
`
`Library of Congress Catalo~g-in-Publication Data
`
`Perez, Reinaldo.
`"Wireless communications design handbook : aspects of noise, interference, and environ-
`mental concerns / Reinaldo Perez.
`p. cm.
`Contents: v. 1. Space interference- v. 2. Terrestrial and mobile interference- v.. 3.
`Interference into circuits.
`ISBN 0-12-550721-6 (volume I); 0-12-550723~2 (volume 2); 0~ 12-550722-4 (volume 3)
`1~ Electromagnetic interference~ 2. Wireless communication systems--Equipment and
`supplies.
`I. Title.
`TK7867.2.P47
`1998
`621.382'24-<tc21
`
`98-16901
`CIP
`
`Transferred to Digital Printing 2006
`98 99 00 01 02 IP 9 8 7 6 5 4 3 2 I
`
`Page 2 of 22
`
`

`

`Chapter 6 Noise Representations in
`Transponders and Multiple Access
`
`6.0 Imroducfion
`
`At ~ e h e ~ of a satellite communications system is the transponder. The transpon-
`der consists of input and output filters, up and down converters, phase-locked
`loops, and traveling wave tube amplifiers ( T ~ A s . ) More modem transponders
`systems are using solid state power amplifiers (SSPAs)~ We now consider the
`nonlinear behavior of the transponder. A block diagram representation of a typical
`transponder was shown in Figure 5.36.
`Let the input of the transponder be represented by
`
`Si(t) = A cos(wet + f l ~ ) ,
`
`(6.1)
`
`where oJ c = 2~rfc is the angular career frequency and phase of the input signal.
`The transponder output can ~
`represented as
`
`Sou.t = g(A)cos(wct + A ~ + f(A)).
`
`(6.2)
`
`If g(A) and f(A) are independent of oJ c, let Ak(O, Wc + ~oj,, and flOk(t) ) denote
`the envelope, the angular carrier frequency, and the phase of the kth carrier, w c
`is the midband frequency or center frequency, which can take any value within
`a transponder bandwidth. For m number of modulated c ~ e r s , access to a Nan-
`sponder input can be represented by
`
`Si(t) = ~ Ak(t) cos{(co c + oJ~)t +fk(O(t)}-
`
`Ak(t)COS(Wkt +fk(O(t)) COS coot
`
`-
`
`Ak(t) sin(o4t +fk(~O)
`
`sin wd
`
`.
`
`.
`
`= X(O cos o)ot - Y(t) sin wet = ~ . ~ + y2 cos Wct+ tan- 1
`
`(6.3)
`
`202
`
`Page 3 of 22
`
`

`

`The co~esponding transponder output is
`
`6.0. Introduction
`
`203
`
`{
`
`= Re g(~,~2 + y2).ex p
`
`(,v/X2 + y2)
`
`]
`
`•
`
`(X + jg) exp(jo4t )
`
`}
`
`(6.4)
`
`Define the double Fourier transfo~
`
`. .
`
`)
`
`.
`
`\ ~
`
`exp
`
`... (
`
`r 2)
`
`(X + jY) exp[~ juX - jvY] dx dy,
`
`which means
`
`Therefore,
`
`2 7 r
`
`After further mathematical manipulations,
`
`&re(t) = R e exp(jo4t )
`
`{
`
`kkk
`
`. . .
`
`~Kl(~,t + f(Sl) + jKe(w2t + f(02)]
`exPL + " " +JKn( wnt + f( On)
`
`)
`• N(k) ,
`
`where K~, K2 . . . .
`and
`
`, K~ can be zero or any integers either positive or negative,
`
`N(k)
`
`(2 7r) 2 )
`_
`
`exp[jf ( ~ ) ] ( X
`
`+ jD H J K ~ ( A e ~ )
`e=1
`
`(6.5)
`
`× exp
`
`Ke tan-i
`
`exp[ - j ~ - jvy]&dy du dv,
`
`£ = 1
`
`Page 4 of 22
`
`

`

`204
`
`6. Noise Representations in Transponders and• Multiple Access
`
`where .Ix is the Kth-order Bessel function. Using the polar coordinate• transforma-
`tion
`
`X = pcos (,
`Y = p s i n ( ,
`
`u = ysin: r/
`v = y c o s r /
`
`and performing the integration, on .£ and r/ simplifies the: expression to the
`following two cases. For K l + K 2 + . . .
`+ K , , - 1,
`
`= f f
`
`0 o
`
`C =1
`
`p g<p exp(jf(P)) " Jl(Y P) dT dp
`
`N(k) = O.
`+ K,, ~ 1,
`and for K l + K 2 + . . .
`Finally, the output of the transponder can be expressed as
`
`Sou t (it) = Re{N(k)exp[j(~ t + f{"(#)]},
`
`(6.6)
`
`where
`
`m
`
`f~(O) = ~ Ktfe(O)
`
`m
`
`~ 1
`
`In the case of numerical computation, ~e. factor g(p)exp[ff(p)] can be approxi-
`mated by
`
`L
`g(p)exp[jf (p)] = ~ beJ~(oeCp),
`
`where L is the number of coefficients needed, J! is the first-order Bessel function,
`and a = 2~(period of the Fourier series). For a given transponder, the characteris-
`tics g(p) and tip) are known. Since g(p) and f(~) are given, the coefficients be
`can be obtained by an approximation, and N(k) reduces to
`
`L
`
`tn
`
`~=1
`
`#=1
`
`(6,7)
`
`which outlines the amplitude for each input signal. The be are determined by
`best fit from the input data of g(p) andfip) in terms of least-square error. Computer
`programs can calculate Sout(t ) when Si(t ), g(~, and tip) are given.
`
`Page 5 of 22
`
`

`

`6.1. Traveling Wave Tube Amplifiers in Satellite Transponders
`
`205
`
`6.1 Traveling Wave ~ b e Amplifiers in
`Satellite Transponders
`
`together with the klystron are known
`The traveling wave tube amplifier ( ~ T A )
`as linear-beam types of microwave device.s. Two type of TWTAs, the helix (for
`broadband applications) and the coupled cavity (for high-power applications) are
`predominant. Such devices are used for frequencies ranging from below 1 GHz
`and with power ranging from watts to megawatts, In satellite systems the TWTA
`is used as the final amplifier before the signal is transmitted.
`It was Lindbland [4] who first described the helix traveling wave amplifiers
`and was the first to explain that a synchronous interaction between an electron
`beam and the RF wave on a helix could produce amplification of a signal on.
`the helix. As previously stated, there are two basic types of traveling wave tubes.
`The. helix TWTA shown, in Figure 6.1 is used mostly in satellite communications
`with RF power ranging from tens to hundreds of watts and also for broadband
`applications. The coupled-cavity TWTA is used mostly in radar applications,
`since it is capable of providing several megawatts of power, but at the expense
`of a very limited bandwidth.
`Let us assume that a transmission line is bent into a helix, as shown in Fig-
`ure 6.2. An RF signal that is applied to the left end of the helix will travel at
`the speed of light in a helical path along the length of the conductor. The velocity
`in the axial direction, the x direction in Figure 6.2, will be the velocity through
`the helix re.ducexl by the helix pitch. The polarity of the signal wilt alternate
`every half, wavelength along the helix conductor. As shown in Figure 6.2, the
`electric field lines extend from regions of positive charge to regions of negative
`
`Magnetic F~using Fie{d
`
`RF Output
`
`.
`
`i L j
`
`y
`
`~
`
`............................................. ~
`
`~
`Etectron Gun
`
`L
`\
`
`~ E~ectron Beam
`
`Hetix SJow-Wa.ve
`Circuit
`
`Collector
`
`Figure 6.1 Description of a helix ~ T amplifier,
`
`Page 6 of 22
`
`

`

`206
`
`6. Noise Representations in Transponders and Multiple Access
`
`Z
`
`Figure 6.2 Electron velocity and direction through, the TWTA helix.
`
`charge. Fuahermore, there, is also an electric field inside the helix with large
`axial components. When an. electron beam (generated by the electron gun. in
`Figure 6.1) is injected along the axis of the helix, the axial electric field compo-
`nents accelerate some electrons and will also slow down other electrons, in Figure
`6.2, the forces on the electrons would be toward the regions denoted by 1 and
`away from the regions denoted by 2. The field pattern will vary sinusoidally in
`the axial direction. If the axial velocity of the electric field and the electron beam
`is the same, the electrons will experience a continuous force toward region 1 as
`the electron, beam goes through the helix. The electrons will start bunching in
`region 1. The field produced by the bunching electrons in the beam will make
`the electrons, on the helix to move away from the region near 1 and toward the
`2 regions. This causes the field on the helix to change in two ways:
`
`1.. The electron current flowing to the !eft on the helix from region 1 is
`that of cu~ent flowing to the right. This current, causes a positive voltage
`region on the helix m the left of region 1. In the same way the electron
`current flowing to the fight on the helix from mgi.on 1 produces a negative
`voltage to the right side of 1.
`2. When the beam-wave interaction continues, the induced voltage wave-
`form becomes much larger than the input wavefo~.
`
`When the voltage wa.vefo~ on the helix shifts to the left, decelerating field
`regions start moving into the electron bunching regions; energy is extracted from.
`the decelerating bunches, and then transferred to the circuit field, which produces
`
`Page 7 of 22
`
`

`

`6.1. Traveling Wave Tube Amplifiers in Satellite Transponders
`
`207
`
`amplification of that field. Figure 6.3 shows the relationship between the electron
`beam and the growth of circuit voltage.
`There are basically three major sources of noise in ~ T A : shot noise, velocity
`noise, and thermal noise. Shot noise results from the discrete nature of the electron
`as a result of the fact that etec.tron emission from the cathode is a random process..
`"velocity noise results from the wide Maxwellian distribution in velocities of
`electrons emitted from the cathode. Thermal agitation noise is present because
`of the RF circuits, which: have loss.
`In addition to these ever-present noise sources, there are other sources of noise
`that depend mainly on the design and construction of a TWTA. Most of these noise
`sources can be diminished greatly by good design practices and manufacturing
`procedures. Some of these noise sources come as a result of partition effects,
`flicker effects, collision ionization, secondary and reflected electrons, lens effects,
`noise growth, poor insulation, yawing insulator charges, microphonics resulting:
`from vibration, multipactors, and power-supply-induced noise.
`The most important noise in a TWTA is produced by the cathode shot and
`velocity. This noise is modified as the electron beam travels through the electron
`gun and into the RF region. An important factor to consider is flaat the noise on
`the electron beam propagates like any other RF signal on the electron k a m .
`Noise travels in fast and slow space-charge waves. The two sources of this noise
`
`Circuit voltage
`,\,
`,
`x,,,
`
`,
`
`,/
`
`Electron density
`/
`,
`,
`,
`
`
`..-x..,
`
`,
`
`t.,),
`
`\,
`
`),
`
`~,
`
`!:
`i
`
`1
`i
`
`I
`.4,
`
`1
`J~k
`
`I
`
`/
`I
`/4"{
`
`-"',
`
`t
`'
`
`f
`
`/
`
`i
`i
`
`'k
`\
`
`I
`t Jt~
`1 f
`/1
`
`"V : V ?2
`
`decelerating field
`
`distance along circuit in half wavelength
`
`Figure 6.3 Voltage and charge density buildup in a TWTA.
`
`Page 8 of 22
`
`

`

`208
`
`6. Noise Representations in ~ansponders and Multiple Access
`
`..................................................................................
`
`I Cathode
`I Surface
`I
`[
`
`.....................................................................................................................
`
`Electron Gun
`f Space Chargel Low Velocity
`i Minimum
`I Correlation
`I
`i
`!
`I
`
`= ..................................
`
`-.
`
`..................
`
`l High Voltage
`I Acceleration
`t
`t
`
`~
`
`RF SECTION
`I Amptification
`1
`i
`I
`
`V
`
`t
`
`["
`'~~
`t
`i Shot Current;
`
`,
`IN°iseli,
`[
`~-#.du£1=!'!#~
`
`I
`
`,
`t\
`t
`
`Drift
`"
`
`. i~ i,.~ ~ Impedance
`~-- .
`~Transform
`"
`l,
`I,
`I
`'~'~ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
`J l
`l
`I
`
`.
`
`.
`
`.
`
`
`
`Figure 6.4 Noise regions in a TWTA electron gun.
`
`we shot noise and velocity noise, and since they are uncorrelated, each source
`produces fast and slow waves. As these noise waves travel along the beam, noise
`current and velocity standing waves axe independently produced.
`When there is no correlation between current and velocity noise, the minimum
`noise figure for a TWTA is around 6.5 dB, with the correlation noise figure "below
`3 dB. It has also been shown that a large axial magnetic field in the acceleration
`region at the input section of the RF structure is capable of inhibiting noise.
`In assessing the noise sources in a TWTA, Figure 6.4 shows the four noise
`regions into which the electron gun can be divided. The figure also shows the
`RF section noise regions. In the cathode region and cathode surface, shot and
`velocity noise are the main noise sources. In the space-charge minimum region,
`low-velocity electrons are returned to the cathode by the potential of the space-
`charge region. The amount by which the potential is depressed fluctuates as the
`cathode current fluctuates because of shot noise. In a low-noise TWTA, electrons
`drift at low velocities for a large distance in the low-velocity region. ~ e n
`the
`electron beam leaves the low-velocity correlation region, it undergoes acceleration
`in the high-voltage acceleration region. Space-charge waves are sent and a stand-
`ing wave pattern is produced. ~ i s sudden change in impedance produces an
`increase in standing wave ratio, and this causes an increase in the noise figure
`for a low-noise TWTA.
`
`6.2 Distortions in TWTAs
`
`The power output vs power input of a ~÷¢TA is mostly linear, and signal distor-
`tions are small. However, as the TWTA is driven into saturation, the transform
`
`Page 9 of 22
`
`

`

`6.2. Distortions in TWTAs
`
`209
`
`function becomes nonlinear, and as a result distortions do occur in amptitude~
`modulated (AM) or phase-modulated (PM) signals. These distortions occur for
`AM/AM conversion, AM/PM conversion, harmonic generation, and intermodula~
`tion distortions.
`The relationship between RF power output and RF power input for a TWTA
`is shown in Figure 6.5. Gain is defined as the rate of output RF power to input
`RF power. There are basically two types of signal gains measured: the saturated
`gain and small-signal gain. AM/AM conversion is a measure of output RF power
`that results fi'om a change in RF input power. It can also be calculated as the
`slope of the curve of 1~ power output, vs RF input power of Figure 6.5. It is
`therefore often the case that power levels must be reduced because of saturation.
`AM/PM conversion is a measure of the change in TWTA phase length resulting
`from a change in RF drive level. As the drive RF power is increased, more power
`is extracted from the electron beam and the velocity of the beam is reduced. As
`the beam vel~ity decreases, the velocity of the RF wave is reduced, and this
`increases the phase length of the TWTA. ~.e typical Figure 6.6 shows the
`variation of phase length with RF input power. As the drive power is increased,
`approaching saturation, the rate of phase change increases rapidly and then
`decreases as the TWTA saturates. As before, the slope of the phase length curve
`in degrees/dB is the AM/PM conversion.
`
`Power t
`
`RF
`Output
`
`Saturated Region
`
`Linear
`Region
`
`Slope = AM/PM Conversion
`
`RF Input Power
`Figure 6.5 Linear and saturated regions in a TWTA,
`
`Page 10 of 22
`
`

`

`210
`
`6. Noise Representations in Transponders and Multiple Access
`
`k Phase Length
`
`Slope = AM/PM conversion
`
`RF Power Input
`Figure 6.6 Power output and phase length variation in TWTA for a given input
`power.
`
`As before, a maximum allowed value of AM/PM conversion is usually speci-
`fied (3-10 dB below the input power that causes saturation).
`
`6.2.1 1NTERMODULATION DISTORTION
`
`Intemaodulafion disto~ion is defined as the production of new output signals
`which are created from the nonlinear combination of two or more input signals.
`This inte~odulation occurs because of the nonlinearity in the amplification
`process~ The order of the intermodulation product depends on how many input
`signals are mixed and which h ~ o n i c s of each of those input signals have mixed.
`Second-order and third-order intermodulation products are defined as follows.
`Second-order intermodulafion woducts are
`
`-f~ - A =A
`U, + A = A
`f2 -f~ =/4
`f~ - A =A,
`
`where frequencies f3, f4, and f5 axe the undesirable distortion products that show
`up at the output. ~ird-order intermodulation products axe
`
`~ + A=f6
`~ + ¾ = ~
`~ - A = A
`
`Page 11 of 22
`
`

`

`6.2. Distortions in TWTAs
`
`211
`
`If fl and f2 are very closely spaced, then the third-order products 2fl - fl are
`the most difficult to deal with. These spurious signals fail. in the vicinity of f~
`and may show up in the receiving passband with sufficient amplitude m cause
`interference problems.
`R e fourth-order intermodulation products are
`
`- A =flo
`
`3A
`
`The fifth-order inte~odulation products are
`
`2A
`3ji
`3f2 -- 2Ji =ft3"
`
`These possible intermodulation products were. produced by just two input
`signals. Figure 6.7 shows a graphical representation of some of the intermodula-
`tion products.
`Finally, Figure 6.8 shows a plot of power output vs power input, not. only for
`the transmitted signal, but also for the intermodulation products. Notice that as
`shown in the figure, the slope of the second, order inte.~odu!ation product is two
`times the slope of the desired signal output. R e slope of the third inte~odula.tion
`product is three times that of the output signals. Notice also that both of the.
`inte~.odulation plots intersect the. output level because they are offset from and
`at. different slopes than the output level. These points where the interseztion
`occurs are called intercept points.
`
`Signal 1
`
`Signal 2
`
`:. ..............
`
`==: ...................................................
`
`...,,,:.,
`
`,:,.,,:,.
`
`..................................................
`
`.
`
`...............................
`
`fl
`2fl "f2
`3fl " 2f2
`Figure 6.7 Some intermodulation products from two signals.
`
`2f2 "fl 3f2- 2fl
`
`f2
`
`Page 12 of 22
`
`

`

`212
`
`6. Noise Representations in Transponders and Multiple Access
`
`/•'•t'--
`Second-order intercept
`Power
`~
`.
`/~.
`7~
`
`..................................................... ~ird-order intercept
`JL Output:
`
`dB
`z S ~ i g n a l
`
`Output
`..//
`.
`...
`(
`)
`/
`~"
`'
`~
`
`Second-order Intermod
`Third-order Intermod
`
`ut
`
`Figure 6.8
`
`Input/output power relationship for transmitted and intermodulation prod-
`ucts~
`
`6.3 Multiple Access in Satellites
`
`Multiple access in satellite communications means sharing all the satellite's
`resources. The theoretical and technological procedures encountered in accessing
`a communication network by multiple stations are referred to as multiple access
`processes. The methods of transmitting information in a simultaneous manner,
`including that of point-to-point in earth stations, to share a common resource
`network are referred to as multiple access techniques.
`In satellite communications, access techniques allow the satellite to be shared
`for mulfidestination signal transmission. There are basically three different (often
`called oahogonal) access techniques. For the technique dealing basically with
`divisions in time slots, we have time-division multiple access (TDMA). For the
`technique dealing with divisions in a frequency band, we have frequency-division
`multiple access (FDMA). When ~ M A
`is combined with TDMA, we have a
`time frequency (TF) access scheme, which includes frequency hopping (FH) and
`time hopping (TH). The third access technique is code-division multiple access
`(CDMA). CDMA is a TF scheme which makes use of the elements or cells of
`the time and frequency two-dimensional plane as a result of both the time and
`frequency divisions. In CDMA the access signal, can be represented by a collection.
`
`Page 13 of 22
`
`

`

`6.3. Multiple Access in Satellites
`
`213
`
`of occupied cells in. a time-frequency matrix: of the TF scheme. All three satellite
`access schemes are shown in Figure 6.9.
`
`6.3.1 FREQUENCY-DIVISION MULTIPLE ACCESS
`
`FDMA uses frequency allocation among a set of each stations to share a series of
`frequencies. The transponder bandwidth is divided into a series of nonoverlapping
`frequency slots, each of which constitutes an access channel. Because the nonline-
`arity of the satellite transponder causes intermodulation products, nonlinear signal
`power transfer, and intelligible crosstalk, the TWTA cannot be operated to full
`capacity. Therefore, the power efficiency of an FDMA system is decreased
`because of the n e ~ to decrease the power of the TWTA. Other potential problems
`in ~ M A are as follows:
`
`1. During • uplink, RF power coordination is needed to prevent the suppres-
`sion of weaker signal by stronger signals
`
`2. Low flexibility; need to slowly change preassigned, traffic patterns
`
`3. Carrying capacity of traffic is relatively tow
`
`fn+Afn *
`fn
`
`...................
`f
`............................................................
`
`fl+Afl
`i i= i ............
`.........
`fl
`(a) FDMA Scheme
`
`"
`
`iii~time
`
`f
`
`C4 4 _.
`
`f
`C23
`
`_ _
`
`I
`
`~ ...............
`
`.........
`
`,,,:
`
`._
`
`t!
`
`tn
`tl+Atl
`(b) TDMA Scheme
`
`tn+Atn
`
`(c) CDMA Scheme
`
`Fig~Jre 6.9 Different satellite access schemes~
`
`Page 14 of 22
`
`

`

`214
`
`6. Noise Representations in ~ansponders and Multiple Access
`
`6.3.2 TIME-DIVISION MULTIPLE ACCESS
`
`The time-division multiple access scheme is an access technique in which, signals
`from each earth station are. sent into burst (time segment) and in a sequential
`manner, gaining access to a common satellite transponder within a periodic
`TDMA frame. Each earth station burst (see Figure. 6.10) occupies a particul~
`slot in a basic time frame: without overlapping with bursts from other earth
`stations. Burst overlapping can be avoided by burst scheduling and network
`synchronization. The TDMA scheme has the highest degree of satellite efficiency,
`but requires excellent timing and network control with other eaah stations. TDMA
`is less susceptible to interference, since it can operate near saturation without
`producing significant interference. Therefore, in order to saturate TDMA, the
`transponder must be capable of producing very high power transmissions. In
`TDMA, at each instant of time there is only one carrier using the transponder,
`and there is no intermodulation problem. However, one main disadvantage of
`TDMA is that it requires network timing and synchronization. It is possible to
`supply vari~ie time slots, but these slots must remain unchanged during op-
`e
`'
`eration, limiting the syst m s ability to accommodate rapidly changing traffic
`patterns, because accurate synchronization is needed. TDMA needs long synchro-
`nization sequences, which decrease the efficiency of the system.. Furthe~ore,
`though there are no intermodulation problems, TDMA is still susceptible to
`nonlinearity problems in the form of intersymbol interference, which needs to
`be minimized by selective filtering.. One of the main advantages of TDMA over
`FDMA is that in ~ M A ,
`the earth station must transmit and receive on multiple
`frequencies to achieve a desired traffic plan. Therefore, there is a large number
`
`N N N N i II
`
`N
`l
`
`Figure 6.10 The TDMA scheme at a system level.
`
`Page 15 of 22
`
`

`

`6.3. Multiple Access in Satellites
`
`215
`
`of frequency-selective upconversion and downconversion chain.s. In a TDMA
`system, the needed selectivity is obtained in. time rather than frequency.
`In a multiple-beam satellite system, the stations of each beam communicate
`with other stations in all other beams. If FDMA methods were to be used, the
`satellite's up-beam must be routed to the down-beams through transponder filter
`banks and additive combiners which sum the noise from all. up!ink beams, creating
`an overall low S/N ratio. The use of TDMA allows the use of a satellite switch
`which selectively connects individual up-beams to individual down-beams, elimi-
`nating uplink noise problems. Furthermore, by synchronizing the locations and
`the duration of individual station traffic bursts and the location and dwell, times
`of satellite switch beam-m-beam connections, the total traffic among all beams
`can be accommodated in a simple and optimum manner.
`
`6.3.3 BRIEF OVERVIEW OF TDMA ARCHITECTURE
`
`Each of the traffic bursts shown in Figure 6.10 looks like the frame, shown in
`Figure. 6.11. Each frame begins with a reference burst. The locations of traffic
`bursts are referenced to the time of occurrence of the reference burst. Traffic
`bursts originate from a given emeth station and caxry the traffic from that station
`to all destinations in a digital transmission format. In Figure 6.11, the start of
`a. traffic burst from station 1 occurs at time T~ after the reference burst. The
`traffic burst from station 2 occurs at time ~ after the reference burst, and so on.
`The position and duration of each traffic burst relative m the reference bm'st are
`
`'I~~, Reference Burst
`
`,
`
`m
`
`traffic burst
`
`r | .........
`
`t
`
`" ' T1 ...... " 7
`i
`
`T2
`1
`F, q ........................................................
`
`Tn
`
`!
`I
`
`I
`
`Reference
`
`Burst
`
`•
`
`•
`
`I
`I
`I
`I
`I
`'
`
`r
`
`i
`I
`i
`I
`I
`I
`
`t
`Frame m
`
`TDMA
`
`frame
`
`I
`Frame m+t
`
`Figure 6.11 Example of TDMA frame.
`
`Page 16 of 22
`
`

`

`216
`
`6. Noise Representations in Transponders and Multiple Access
`
`~ rvice Channel
`
`.
`
`.
`
`.
`
`.
`
`.
`
`
`
`..........
`
` mir, gRooove,4Wo
`
`,,
`
`.....
`11
`
`....
`
`.......... 1
`P:lV°cel
`
`.......
`
`i,,0,oo
`" . . . .
`
`Delay Channel
`
`"
`
`"
`
`Fibre 6.12 Architecture of a reference burst.
`
`scheduled to a protocol established by the network. These protocols, however,
`may change depending on the traffic demand.
`As previously stated, reference bursts are emitted by a reference station and
`are the basis for synchronizing all other stations in a network. The architecture of
`a reference burst is shown in Figure 6.12. h i s reference burst has the information
`needed by other stations to obtain the exact location of their bursts in the frame.
`The reference burst is made up of three parts. The first is the "carrier and bit
`timing recovery," which serves the objective of look5ffg at a received station to
`the carrier frequency and the bit timing clock of the burst. This is followed by
`the "unique word." The unique word is used to reference the time of occurrence
`of a burst, and it also marks the symbol time reference for d~oding information
`in the traffic p~t of a burst. Following the unique word, the reference burst
`contains a
`serv ce channel' and order features which can be used to support
`operating system protocols and for utility teletype and voice communications
`among stations. Following all these and last is the control and delay channel,
`which serves to corr'anunicate information for the control of burst positions to
`station.s in the network.
`The traffic burst architecture is shown in Figure 6.13. Traffic bursts are syn-
`chronized relative to the reference burst to occupy assigned l~ations in the
`TDMA frame~ Notice in the figure that the "preamble" of a traffic burst has
`the same architecture as the reference burst; actually, the c ~ e r and bit t i ~ n g
`recove~ and unique word are the same as in the reference burst.
`
`Figure 6.13 Tra~c burst arc~tecture.
`
`Page 17 of 22
`
`

`

`6.3. Multiple Access in Satellites
`
`217
`
`6.3,4 CODE-DIVISION MULTIPLE ACCESS
`
`Code-division multiple access systems originate fiom a more general class of
`system called nonorthogonal systems (which also includes spread spectrum and
`time/frequency hopping). Actually, most nonorthogonal systems have the com-
`mon property of spread spectrum. In a CDMA system, a message spectrum can
`be spread by pseudorandom (PR) code sequences to enable users to obtain access
`through a corp~on signaling channel.
`There are three basic methods which can produce a spread-spectrum effect:
`direct frequency, frequency hopping, and time hopping. CDMA is an extension
`of direct-frequency spread spectrum and frequency-hopping spread spectrum.
`CDMA systems provide multiple access communications capabilities. In CDMA
`each user has an individual, distinctive pseudonoise (PN) code. If these codes
`are uncorrelated with each other, then within the same mobile cell, K independent
`users can transmit at the same time and in the same radio bandwidth. The receivers
`decorrelate (or despread) the information and regenerate the desired data stream
`DM( 0, where m = 1 . . . . .
`k. In Figure 6.14, a CDMA spectral concept of
`several direct-sequence spread-spectrum carriers is illustrated.
`If the received power Ps of all these signals is the same, then any one of the
`desired signals will be interfered by m - 1 equal-power CDMA signals. Therefore,
`the RF received carfier-to-interference ratio (C/1 (dBm)) = 10 log (l/m), which
`is a negative number known as "self-interference" caused by the other m - 1
`carriers that simultaneously occupy the same bandwidth as the mth desired c ~ e r .
`The probability of error (Pc) caused by the self-interference of the m simultane-
`ous equal-power received signals is given by
`
`Pc = ~ erfc
`
`~
`
`,
`
`(6.8)
`
`where jQ is the carrier frequency, fb is the bit rate, and m is the number of users.
`It is assumed in this equation that thermal noise is negligible and that the powers
`received are the same and uncorre!ated codes.
`~ e more general error probability for a general CDMA was derived through
`a theorem of higher moments, and the expression is given by Equation (6.9),
`where h is the magnitude of desired signal, n is the: zero mean additive white
`Gaussian noise, z is the interference from m - 1 users, D is the maximum values
`of the interference variable z, m l = E(z2), m2 = E(z4), and ~
`is the variance of n.
`The bounds can be evaluated once we know the values of h, m m~, m2, and
`D. The terms m~ and m2 depend on the probability density function of the
`
`Page 18 of 22
`
`

`

`218
`
`6. Noise Representations in Transponders and Multiple Access
`
`Dt(t)
`Gl(t)
`
`]~
`D2(t )
`~(t)
`D3(t)
`c~(t)
`Din(t) = ruth data source (m users)
`~(t) = ruth PN code (m users)
`l Space
`
`[~
`D4(t)
`~(t)
`
`,.,
`
`e
`
`Dm(t )
`Gin(t)
`
`. . . . . . . . . . . . . . . . . . . . . . . .
`
`DS(0
`
`tuency
`
`Time
`Figure 6.14 CDMA system with m users sharing ~e same receiver satellite.
`
`interference. For PN code sequ~.nces, the density function can be approximated
`as Gaussian. Notice from the following equation that the error probability is
`highly sensitive to the number of' users in the CDMA system.
`{h-(rnl)"4\]
`+ erfc\ ................ ; : ~ , .................. ]
`
`{h + (ml)l/4~
`Pe (up~r bound)= effc \
`~
`j
`
`
`+ erfc(~)+
`
`[erfc/h + D)
`
`ert~(h.....--.D't, effc( h__~)]
`
`mt
`
`• k, ' Y ~ ] + ~
`
`~ - m~
`',, m2
`.
`
`h ~ tm2j
`P~ (lower bound) = effc h + (m2)t!4']
`~ r ' ~
`) + effc ~r-~
`•
`
`h
`|
`j - 2 erfc ~ ~
`
`m~
`
`+ 2 erfc (
`
`) h
`
`(6.9)
`
`Page 19 of 22
`
`

`

`6.3. Multiple Access in Satellites
`
`219
`
`A general power-bandwidth relationship of spread spectrum systems is given in
`terms of signal-to-noise ratio at the receiving detector output after going through
`a nonlinear transponder of usable bandwidth Bt (Hz):
`
`(57N) =-~ kBm~(~z) /
`
`1 ~ (~N)rJ"
`
`(6,1,0)
`
`Here, (C/N) is the received carrier-to-noise ratio in system bandwidth B t (Hz),
`Bmc (Hz) is the message bandwidth, and m is the number of simultaneous signals
`present at the transponder input, all assumed to be of equal power. A comparison
`of FDMA, TDMA, and CDMA for small satellite network earth stations is given
`in Table 6.1, where K is the number of accesses.
`A block diagram of a CDMA transmitter and receiver is shown in Fig-
`ures 6.t5 and 6,16, We are assuming a modulated binary PSK signal S(O given
`by
`
`m
`
`S~i(t) = ~ ~
`i = t
`
`Di(0 cos 2 ~fif t,
`
`(6.11)
`
`Table 6.1 A Comparison of FDMA, TDM~, and CDMA
`
`Power
`Type of
`Access Eff:w&ncy BandwMth
`
`Jamming
`
`Oper~on
`Feature
`
`Complexity
`
`FDMA
`
`0.8
`
`10/3K.I ~3
`
`TDMA
`
`0.8--0.95 10/3KI.3
`
`CDM_A
`
`0.8
`
`3//I'~ K
`~-~,.C) K - 1
`
`Most
`vulnerable
`
`Moderate
`
`Uplink power
`coordination.
`Frequency
`assignment
`changes with
`changing the
`number of
`access.
`Synchronization is Most
`re