Embed Size (px)
Citation preview
.
00 O
O
RADC-TR-77-347 Final Technical Report October 1977
ji
^ ADAPTIVE SAME FREqUENCY^(SFR) BREADBOARD IMPROVEMENTS
Magnavox/General Atronics Corporation
Sponsored by Defense Advanced Research Projects Agency (DoD) ARM Order No. 2750
^v
Approved for public release; distribution unlimited.
The vievs and conclusions contained in this document are those of the authors and should not be interpreted as necessarily representing the official policies, either expressed or implied, of the Defense Advanced Research Projects Agency or the U. S. Government.
♦/»«
>- ROME AIR DEVELOPMENT CENTER Q— Air Force Systems Command Sj Griffiss Air Force Base, New York 13441
D D C
CEC 2 1977
iHSEinns D
iM^i^...J,:^W.,«i:^k^^ ■■. ^■-',,ti.-li-.!i£;?j:LU>j. - -■. ■'-: ~-.i.^fcfc
^™|P^^^iwwfliK»w,*«j«w^w,!iw^^^^ v »■'• "VJwi!ij«iii5:,>»N<w»«^«wn|!.',^''v
^
This report has been reviewed by the RADC Information 0«^e (01) and ifreleasable to the National technical Infoption Service NT S). At NTIS it will be releasable to the general public, including foreign
nations.
This report has been reviewed and is approved for publication.
APPROVED:
If your address has changed or if you wish to be removed from the RADC mailing list, or if the addressee is no longer employed by your ^an"ation. pleise notify RADC (DCCR) Grifflss AFB m 13441. This will assist us in maintaining a current mailing list.
Do not return this copy. Retain or destroy.
jm^d^m^M^Mmiämäi
BEST AVAILABLE COPY
f ....
p5!m»^TO»«WfWS'^WW"''^^ ■,."«««» ■ ,/■■■•■ -.sj
Ulli' Hirtiiahrt ¥i ii i 'i
■ . ^
«.eras in,! >,■ ITjJ
.'«nfiaiiis,..
Wme iKWu^t^ BUI iectios n
o
8?
ciS7RiBiiTia;!/Av^it/,::!ii,;KiiB
_iiH: Ä7.';iL7"Kn(,;8i IPLCUL"
fl
'.
ADAPTIVE SAME FREQUENCY REPEATER (SFR) BREADBOARD IMPROVEMENTS
Burton S. Abrams Michael Caracappa Andrew E. Zeger
Contractor: Magnavox/General Atronics Corp Contract Number: F30602-74-C-02A2 Effective Date of Contract: 1A June 1976 Contract Expiration Date: 1 May 1977 Short Title of Work: Adaptive Same Frequency
Repeater (SFR; Breadboard Improvements
Program Code Number: 7G10 Period of Work Covered: May 1976 - April 1977
Principal Investigator: Andrew E. Zeger Phone: 215 2A8-3700
Project Engineer: Carmen J. Luvera Phone: 315 330-3091
Approved for public release; distribution unlimited,
This research was supported by the Defense Advanced Research Projects Agency of the Department of Defense and was monitored by Carmen J. Luvera (DCCR), Griffiss AFB NY 13441 under Contract F30602-74-C-0242,
0 D D C
AHC 2 1977
T-7T*T
UL^ELTLTE
____^___^_^__^_^^^^_
■■^'^^^w^^^^^^mm^- l'nämu HI -..
■
UNCLASSIFIED SECURITY CLASSIFICATION OF THIS PAGE (When Data Entered)
REPORT DOCUMENTATION PAGE
~ "g^PfTTTBRMiNc; ORGANIZATION NAME A, O ADDRESS
llagnavox/General Atronics Corporation' 1200 Mermaid Lane Philadelphia PA 19118
2. GOVT ACCESSION NO. 3. RECIPIENT'S CATALOG NUMBER
ADAPTIVE.JAME FREQUENCY REPEATER (SFR) BREADBOARD IMPROVEMENTS 7
'' Burton S./Abrams, Michael/Caracappa Andrew E./Zeger l/^
II. CONTROLLING OFFICE NAME AND ADDRESS Defense Advanced Research Projects Agency 1400 Wilson Boulevard Arlington VA 22209
Gd
READ INSTRUCTIONS BEFORE COMPLETING FORM
S •''YPF "f "'-r-.fr.T " ■!■■ pbcn
Final technical Xep&rt. 14 Jun 76-1 Apr 77j
.6-. REPjF-qßMiN.G-ORrv, aefoRT NUMBER GA02896-2876-36/ . 8. CONTRACT OR GRANT NUMBERfs;
F3j2l6^'2-74-C-^242>
4» i'nv*nBm!.t:tviEtrrrrnrnT< AREA ft WORK UNIT >*fM.BE
J. 27500001
T >*»MBEJ16-— ASK
'S 12. fttEjOÄT-OA**—'
Oct0*er 1977 Tr-NÜM'öWöT PAGET
14, MONITORING AGENCY NAME & ADDRESSf//d///eron( from Cnnlrollint Ollice)
Rome Air Development Center ^DCCR)^, Griffiss AFB NY 13441 v A^p J
Jki- 9
IS. SECURITY CLASK. (ol this report)
UNCLASSIFIED 15«
N/A DECLASSIFI CATION'DOWN GRADING SCHEDULE
16 DISTRIBUTION ST ATEMEN T fo/(hy s ReporO
Approved for public release; distribution unlimited,
17. DISTRIBUTION STATEMENT (oi the abstract entered In Block 20, II dllterent from Report)
Same
18. SUPPLEMENTARY NOTES
RADC Project Engineer: Carmen J. Luvera (DCCR)
19 KEY WORDS (Continue on reverse side II necessary and Irienltfy by block number)
Interference Cancellation Adaptive Antenna Array VHF--UHF Communications Same Frequency Repeater Fl - Fl Radio Relay
20. ABSTRACT (Continue on reverse side It necessary and Identify by block number)
This report documents the technical effort conducted under Amendment P0Ü003 to Contract F30602-74-C-0242. It presents the results of modifications to the real time, nonsampling, Same-Frequency Repeater breadboard equipment originally constructed under the above contract effort. The SFR is capable of relaying various types of analog and digital signals using coherent interference cancellation techniques to prevent oscillation. The main thrust of this effort was to increase the forward gain from the 67 dB obtained in
the initial effort to a performance goal of 100 - 120 dB. —
DD ,Ü FORM l j-ti AN 73 1473
......... . -■;iJ.^,M,/--. - ■>"-.-"■
EDITION OF 1 NOV 65 IS OBSOLETE
91 20'/ UNCLASSIFIED ~ Aß
SECURITY CLASSIFICATION OF THIS PAGE (Wien Data Entered)
ipppmwsiw ■ ■■!^,!w«",f¥J?wpw¥>fwwPiP'w?»K.^y^ '/
UNCLASSIFIED
M SECURITY CLASSIFICATION OF THIS PAGEfWien Dal» EnUnd)
v The actual measured forward gain achieved In this follow-on «««'«f» "Jf' This improved forward gain was attained through «odificaumn f ^ir""« J" the breadboard equipment. The basic S?R system 8tr"ctu"' fnd/^,"n^^t on which It was predicated remain Intact. The results of the Initial effort are summarized In RADC-TR-76-78 entitled. "Adaptive Same Frequency Repeater
(SFR) Study*.
The modifications that were performed and the performance results obtained in this follow-on effort are discussed In this report. Updated system and circuit diagrams are Included as well as a discussion of those technical areas where additional work would further Improve SFR equipment performance.
UNCLASSIFIED SECURITY CLASSIFICATION OF THIS PAGEfWhen Da(. Entered)
. ^»temjiMai»..^.!...,.,- .,.i,k.-:; .,-,, ',.::i.,^....^,.„.. .^t..;.;^...,.-...- .■-..;- ■■ ■-:.^..,- -,^-.- .
■ itRfii^pipiWSpsBppspwpP'W»'? ^p^Dfipippi
-
SUMMARY
This report documents the technical effort on Amendment P00003 to Contract F30602-74-C-0242. It presents the results of modification to a real time, nonsampling, Same-Frequency Repeater (SFR) breadboard. The SFR is capable of relaying vari- ous types of analog and digital signals, using coherent inter- ference cancellation techniques to prevent oscillation. The breadboard was originally constructed under the above contract.
The main thrust of this effort was to increase the forward gain from 67 dB obtained in the initial effort to 102 dB. The improvement was attained through modification of circuits in the breadboard. The basic system structure, and the concepts on which it was founded, remain intact. The results of the ini- tial effort are summarized in RADC-TR-76 - 78, entitled, "Adaptive Same Frequency Repeater (SFR) Study."
The modifications that were performed are discussed in the report. Updated system and circuit diagrams arc included. *reas where additional work would further improve performance nee discussed as well.
■' 'ir i -rfii l||<|<^M|ii.~-^»~*^"-""-- -' ■J"i'"'
■■■. ■ ■
, ■.
.
TABLE OF CONTENTS
Page
SECTION I - INTRODUCTION 1
SECTION II - REVIEW OF PREVIOUS WORK 1 . SFR Concepts 3
a. Interference Cancel! ation System 3 b . I CS AppI led to an SFR ' c. Pilot-Directed ICS 5 d. Multichannel Notch Filter ICS S c. Combination of Pilot-Direction with the Multi-
channel Notch Filter ICS 9 f. RF and IF ICS 9 g. SFP Conceptual Block Diagram 9
2 . Hardware Implementat ion 13 3. Experimental Results of Original SFR 13
SECTION III - MODIFICATIONS TO SFR BREADBOARD 16 1. Description of Modified Block Diagram ^(1
2 . System Modifications IS a. IF Rejection in RF Front End 1'^ b. Isolation of Transmitted Signal from Pilot Is
c. Redistribution of IF ICS Loop Cain 19 d . AGC Time Constant 19 e. Drivers for RF Bipolar Attenuators 19 f. High Q Notch Filters 21
SECTION IV - SFR PERFORMANCE . 28
1. Pilot Cancellation 28 2. Cancellation of Transmitted Signal and Pilot ^J
3 . Forwa rd Gain 31
SECTION V - PERFORMANCE LIMITING FACTORS 36 1. Transmitted Signal, Pilot Interaction 36 2. Noise Modulation of the RF Weight 37 3 . RF We ight Linearity 39
SECTION VI - RESULTS WITH OTHER PILOT WAVEFORMS 44 I . I;M Pilot 44 2. Straddle Pilot 4 4 3 . Separate RF and I F Pi lo.ts 46
SECTION VII - RECOMMENDATIONS FOR FURTHER WORK 48 1 . Circuit Noise in the RF Weight Control Unit ♦° 2 . i\nea r i 2at i on of the RF I CS IVe i ght 3. Straddle Pilot ' 4-1
4. Field Test 49 5 . Increased Bandwi dth 49
SECTION VIII - SUMMARY AND CONCLUSIONS ." ^3
APPENDIX A - Ci rcuit Schema t i c s S6
i i i.
KBCSBUKS PACK NOT IUüBD
HLAIK -■
-~———«rffci "
■ '^-"^■"^"^lli
1 2
3 4
6 7 8 9
10, 11 12 13, 14, 15, 16, 17, 18, 19. 20, 21 ,
22.
23. 24. 25. 26. 27. 28. 29. 30.
31 . 32 . 3 3 . 34. 35. 36. 37. 38. 39. 4 0 . 41 .
LIST OF ILLUSTRATIONS
Same-Frequency Repeater Breadboard Interference Cancellation System for Colocated Transmitter/Receiver Interference Cancellation System for SFR Pilot-Directed Single-Channel ICS to Overcome Problem of Correlation Between Transmitted and Received Wave forms
Multichannel Notch Filter ICS Concept for Broadband Cancellation Notch
Pilot-Directed Multichannel Notch Filter ICS Distribution of Cancellation between RF and IF SFR Conceptual Block Diagram SFR Functional Block Diagram Block Diagram of Improved SFR Breadboard Performance of AGC Amplifier with Slow Time Constant. Usual Single-Pole Notch Filter Another Approach to Notch Filter Synthesis Definition of Terms 30 MHz Notch Filter Schematic
Page
1
4 6
Pilot Cancellation in RF and IF ICS's Cancellation of Pilot Plus CW Transmitted Signal Test Setup for Forward Gain Measurement SFR Output for Various Input Levels Modulation Effect of ICS Circuit Noise SFR Pilot Output Showing Effect of Noise Modulation of RF Weight
Minimum Essentials for Analysis of Effect of Third- Order Weight Nonlinearity Generation of Straddle Pilot Implementation of Separate RF and IF ICS Pilots Generation of a Straddle Pilot with PN BPSK Sidebands. SFR Field Test RF ICS Complex Weight and Canceller RF Bipolar Attenuator Circuit RF Preamp and Downconverter SFR Pilot Upconverter and RF Unit On
RF and IF Correlators SFR Synchronous Detector and Lowpass Filter... RF Bipolar Attenuator Driver SFR IF Canceller and Error Signal Distribution
A
B ICS Weight Control
Notch Filter Processors, Box | E G , H SFR IF Weight Assembly
IF Bipolar Attenuator Ci rcui t.. .... .... .. IF ICS Weight Control Units [T], QT], GO- IF Bipolar Attenuator Driver Notch Filter Processors, Box SFR Signal Upconverter, Local and Reference Downconverter.
Oscil1ator Distribution
8 10 11 12 14 17 20 23 24 25 26 29, 32 33 35 38
40
41 45 47 50 51 57 58 59
60 61 6 2 63 64 65 66 6 7 6 8 69 70
71
30
LV
.-—.^_^^^__^__1__________
■!•'■■■ «Mfwt^ppmipnppKu^ : f>*,„.ijii
,:
LIST OF ILLUSTRATIONS (Continued) Page
42. SFR IF Bandpass Filter and Pilot Distribution System. 72 43. SFR 455 kHz Waveform Generator I N 1 73
--■-"--• . , • i
.:.-,-,-T^.Y,^r,^.,.,,-,T;, ^™^«^WtlW»J»,Mg»)glT51» WWFCTrWW^Jf^ft^ 'y
SECTION I
INTRODUCTION
This report documents the technical effort on /\mcndrnent P000Ü3 to Contract F30602 - 74-C-0242. It presents the- result: of modification to a real time, nonsampling Same-Frequency Repeater (SPR) breadboard. The SFR is capable of relaying v.i r i ous types of analog and digital signals, using coherent inter- ference cancellation techniques to prevent oscillation. The breadboard was originally constructed under the above con tract
The main thrust of this effort was to improve the 07 dP forward gain obtained in the initial effort to 102 dB. The improvement was attained through modifications of circuits in breadboard. The basic system structure, and the concepts on which it was founded, remain intact. A summary of the prescn performance features of the SFR breadboard is given in Table !
Table 1
Performance Features of Modified SFR Breadboard
Forward Gain Retransmitted Signal Power Center Frequency Bandwidth Output Level of Added Pilot Signal MO «liim, spn\^' o ^ , = !:
102 dB + 28 dBm, linear 291 .8 MHz 100 kHz + 16 dBm, spread over
The SFR breadboard has been packaged in a rack-mounting enclosure having a front panel height of 13-3/4 inches and a depth of 19 inches. A front view photograph is given in Figur.
The modifications that were performed are discussed in the report. Updated system and circuit diagrams arc included. Areas where additional work would further improve performance are discussed as well.
». _^^_^_^_^_^^^_—^—^^—^ immmtm —^—^^mmm I
■■p ■■■■ i#ÄWt!m'v!f"'"^"tw--■ - ■■v^|*<.^i^flii^!v*p^p!^Ä^)[«v!»fl>.'.r- ■
y
Q
< o O < w cc
er: W H <
tu W Di
p >-
u- w er w
i^iigf^ßfgi^gigg^gimmw^imiiti^m^^!^^^ ■v- -A
, *■ ^i
SECTION II
REVIEW OF PREVIOUS WORK
In this section, the initial work done by General Atronics in the development of the Same Frequency Repeater (SFR) is re- viewed. A full exposition of this earlier work is contained in the final report submitted on that program.*
1. SFR CONCEPTS
In this section the system concepts developed for the SFR are described. Each of these concepts were verified experimen- tally as part of the initial program.
a. Interference Cancellation System
The motivating concept for the SFR development was the use of an Interference Cancellation System (ICS) to protect a receiver from interference by a colocated transmitter. A block diagram illustrating the ICS in this application is given in Figure 2.
Because of imperfect isolation in the antenna network, nonzero VSWR of the antenna, and reflections from nearby objects, a small portion of the transmitted signal is coupled into the receiver output of the antenna network. The transmitted signal at this point can be further reduced by an ICS. The ICS extracts a portion of the transmitted signal through a directional coup- ler to use as a reference waveform whose amplitude and phase is adjusted in a complex weight. When the weighted reference is combined with the signal at the output of the antenna network, proper setting of the complex weight allows the transmitted signal to be cancelled while the received signal level is main- tained. The complex weight setting is adaptively controlled by a feedback system which correlates the error signal with the reference and seeks to minimize the correlation by minimizing the transmitted component of the error signal. Since the received signal and transmitted signal are generally separated in fre- quency, correlation of the received signal with the reference is rejected by lowpass filtering in the weight control unit.
*Abrams, B.S. , et al, "Adaptive Same Frequency Repeater (SFR) Study," Final Report RADC-TR-76-78, March 1976.
■aaa, M^MiMüiin _ ___
W9^ ■' ..,-....:■..-■■> ^-^ ■.::;.>-■ ■■;•■>-■,. .v-.?..-^;-. ".'..r.^.-..^^ " ~-■-■.,-■-. i^-^■.--;l-l-■—..,'.;■.■■:. ■.■■,■■ .ir-^,-.^. • ■ * y
o: <- -I
cr UJ
cc UJ > t—t
UJ o UJ cc
\
X[
.
a
_J
W z C3 »—4
1 k < » CC o cc cc UJ
i f
X _1 UJ 1- h-O _J n: ac cc i- Q. cr o 1- •- s: •-' >-• z z O UJ UJ O =3 O 3 3 O
1 t i \
wm mlmm «B ^HM MM —
UJ o z UJ cc
\ UJ u, UJ cc
—
_J
1
12
I o uj I •- I UJ to I o >- I Z (^ |UJ I cc
1^
3»
CC
cc
< a:
Q UJ t- <: o o
CM
CC cc
CD
>- 00
o t—<
I—
cc UJ Ll- CC
!_____ ■ - - üü _-_
Ml'JiltfSitlBtalift,null« ! —. .-
'
b. ICS Applied to an SFR
Once the principle of ICS isolation between a colocated receiver and transmitter is established, it can be extended to the SFR application without difficulty. This is illustrated in Figure 3. The receiver and transmitter are now tuned to the same frequency an' are processing the same waveform to make an SFR. Means must be provided now for the correlation control of the ICS to distinguish between the transmitted waveform and the received waveform.
c. Pilot-Directed ICS
The mean and not by what ICS, as shown i mitted signal i structured to k Only the pilot the complex wei receiver. Howe both have the s the pilot also
s of controllin is received is
n Figure 4. A n the same freq eep it uncorrel is applied to t ght adapts to c ver, since the ame coupling pa results in cane
g the ICS by what is transmitted provided by pilot direction of the
pilot signal is added to the trans- uency band, but with a waveform ated with the transmitted signal. he ICS correlator inputs, so that ancel the pilot coupled into the pilot and the transmitted signal ths into the ICS combiner, cancelling ellation of the transmitted signal.
d. Multichannel Notch Filter ICS
The isolation pro gain if desensitization high forward gain requir the entire band which th far is limited in broadb coupled waveform to be c whose delays may not be provided by the complex tion at only one frequen worse at frequencies far
vided to the SFR rau or oscillation is t. es a large amount 0
e relay operates. and cancellation by ancelled arrives th known or fixed. Th weight is correct f cy, and the phase e ther removed from t
st exceed the fo o be avoided. T f cancellation o The ICS describe the fact that th rough several pa us, the phase sh or perfect cain-e rror gets progre hat one.
rward bus , ver d thus e ths ift 11a- ssively
This effect, and the means for overcoming it, are illus- trated in Figure 5. If we assume that the signal to be cancelled has a flat spectrum over some range of frequencies, and that the two paths into the ICS combiner have a differential delay, the output spectrum has an M shape, as shown at the top of the figure. Cancellation is best at the center of the spectrum and gets pro- gressively worse toward its edges, generating a cancellation notch
The cancellation notch may be broadened, and hence the total rejection improved, by adding other weighted channels with frequency shaping to the ICS. The frequency shaping used is one which looks most like the output spectrum of the single channel ICS, namely, a notch filter. The notch filter channel cancels the second order terms that dominate the output residue
iv„ii.i. ...■_....: „i .; i-i^.-;,, . ..1;.„..^._.... ■ ...........^ :—,.--,■ ■■.-■.■ -- ■■ ■■■•'-■ v,.--.-,.—■■ ,.. -—... --» ...----..■.■ ,...-..^^..
EP . - . ■ ■ . . _ .. '.-■._., ,. '■.- fafc Tlflftnn'r I'I . .
ex:
QC Q
O l- —i Q. i—i et
X t^ CD ts 2: •-. 1—1 c£ CO X CÜ
s: o i—1 »—1
o; CJ et: uj
\
/
3 UJ _J o > «a: a. ►-■ z:
3 O 1—1 O UJ l/> _i a:
1 00
et 2i
o o
1- 5:
_J o LU Ll_ O LU
CC < O 3
O O
< UJ _i a: _J LU LU U- (_) LU z cc <C -
s: UJ o: x o
h- I-H O- UJ et <_> Q LU < CC
tI5 >
ijj
1—. Q s: LU 00 I— z: < < -1 QL Lü I— CC
CÜ LU O X CJ
O 1 < o; (— X _J LU
h- LU (— (— o: I— 3z 1— o: ■-. CD o o s: 1—1 ^ CJ oo LU :z
>- et CO cc
3 Q
X <
0- >- 2: —I o z: CJ o
00 2: CJ cc K-c o
U- LU LU X s» 1— et 3
h- 2: cc et LU
Q . uj et _l
1 _I X O h- CC 1—1
o = CJ —J
et 00 S
CD
O LU CD
z: o 1— 1—1 00
et s: _l _I oo LU s: CJ cc z: o et U- CJ LU >
Q et O 3 O CD O
1—< et LU z 00 o 1—1 CC I— o D. CC et OC Q LU et -
CC U- 00
cc o
oo >- 00
CO o
CC et :D _J CD —1 1—( LU U. CJ
et CJ
< «
cc o o
cc
o
et CC
o
D- i—1
o
cc D.
Mit»ffiirtiititaii»Mw»Wiiii i-1. im jum
r
I O t- oz
a. uj I- <->
OQ:
X 1
a: UJ z _l
Wo CO «a: z
-. 2: en O K^
u
rv'
1 k 1 i O oc en UJ
\ f z O 3
X _J »—< *^^« LU » H- O I— _J —1 1— 3: 0; 1- •«o D- 3: 0 1— i-> _J a: ST O l-i z z LU 1— O >-. LU O ZD a: z O UJ 3 t_> a: 0
3 0 0 0
i ^ 4 k II- 1— fc>>- 13 ^f _i _J a. ^^--^ "—i z z ^^— a. O
UJ 0 y» >- \- h- cc UJ < —■ O-
UJ 3 -1
1 a. _J z 1—1 LU
u_ C\J t^) Q. ta UJ QC
1 '
<C i^ , 1 z a: z 0 UJ 3
X X " ' \ h- 1—
Z UJ < z
Q UJ 1—
o C/1
LU
»—4
o
Q.
a: UJ
o
o
I— s: ZD co a. Z: Z CC 1—1 Q;
t— O UJ ^r h- 0 X z z (— —I <
0 0 X oo 1— —1
" a. z _i C£5 •a: CD 1—1 Z CO C3 1—(
»—i _J Q oo D- LU
or • O H- 0 a- 0 1—( Q: LU s: a- > UJ oo LU h-H 5: z
UJ ct <t 31 0 00 o: t— UJ (— 1— o; LU
^ >- x:
0 0 e? Z ^— 3- O
1— z i—1
0 >-< —J (— _J CC < l-H Q Z _l Q. UJ CD _l
—J 1—1 LU CD Q. CO CJ z rD z ►-H O Q cC (_ 0 LU CJ <C 1— _i • 1— (— LU H- >-. O D; 0 ST —I Q; —1 CO 1—1 O --i z a. 0 a. <
QC Z >- UJ (— UJ CQ X X
H- Q h- Q Z LU —1 <: « _J UJ Q:
_J O t— LU O z 0 z a: cc _l (-H (— 0 1—1 03 z a. s: 0 0 0 0 h- LU O
O oo I— z 00 CJ LU 1-H O •—< CO CO ►—1
o
LU a: _l 0 CQ u_ O UJ o; > Q. <
3 UJ
ST Q O UJ CJ > CC 1—1
UJ UJ > 0 O UJ
or. 0 (- Q
Z CO <C CJ
<* 1— O
a: Lu I— ^> Z i—i
CD z 2: »-H <; co u_ nr z
t_3 et 1 CC
CD
Q LU UJ QQ H- C_) sr LU O CC p—1
t—1 H Q <
1 _]
(— UJ O D£ _J D; ►—1 0 Q- CJ
1—1 <C '—' 2 z a. CO CD Z 1—" CO <: 00 ^ CC _J t— a.
■•=-- 1-~»-.-^... ts», ■'"■■" ^•^•^^^
____^ . i - . ■ , y.-...-*l :. " '^■..UiMmmA,* it i --. — fa*
Z3 D- I— =) o
W ik ik
(X)
rn o
J
X DC
o nr z: O o H- o o
2: t/i 0 z: KH O
1—4
o: h- UJ <=c (— —1
tn _.j _i 1—1 LU
UJ U- 0 nc 7Z ^D .-r «a: LD O CJ t—( l_ u_ 0 Q
zr z: CC
_1 CO Li.) O 2: cC 2: O < o: x: m 0 I-H Q; t- O
h- o LU O.
Q:
o D.
cr LU
Xvjj
, ■i-4'iiiit^iiifi»iWtlf fttwtWiiiitefeiifTi 1 , siu> ■■^-■-'---'■-'"
... ...i.,-,^, . "- CiA-mftteiiii « — '.
o£ the single-channel ICS, leaving higher order terms whose com- bined power level is much lower than that, of the second order term. Additional notcli filter channels may be used to further broaden the notch and improve broadband cancellation.
e. Combination of Pilot-Direction with the Multichannel Notch Filter ICS
Figure 6 illustrates the means by which the pilot direc- tion and multichannel notch filter ICS concepts are combined. The pilot supplied to the weight control units musf. undergo the same notch filter processing as the weight inputs do. Hence, there are two notch filter chains.
f. RF and IF ICS
It is important to ensure that the achievement of a large amount of cancellation is not limited by stray leakage of signals around the canceller. Just as high gain is often achieved with gain blocks at two frequencies, with frequency conversion between them, to circumvent leakage problems, the same concept can be applied to cancellation. This is illustrated in Figure 7. The first stage of cancellation is at RF, followed by downcouversion to IF, and then additional cancellation at IF. Since a single- channel ICS can provide a significant amount of cancellation before frequency dispersive effects limit it, the multichannel notch filter ICS is used at IF only.
g. SFR Conceptual Block Diagram
The concepts described in the preceding sections form the basis for the formulation of the SFR block diagram. A conceptual block diagram of the SFR is given in Figure 8. It shows a sin- gle antenna used for both transmission and reception, with a 6 dB hybrid as the antenna coupler. The received signal passes through a single-channel pilot-directed ICS at RF. The RF ICS is expected to provide 55 dB of transmit-to-receive isolation, sufficient to allow RF amplification and RF-to-IF conversion with negligible distortion.
Following RF-to-IF conversion, the signal is amplified and inserted into a three-channel pilot-directed notch filter ICS at IF. At this point the full isolation is obtained. The signal then passes through an IF filter and amplifier with AGC for out- put level control. The detected video output of the amplifier may be used to indicate signal presence to control the DC power supplied to the RF transmitter amplifier.
The pilot is generated and added to the signal at IF. It is also supplied to the IF ICS as well as to the RF ICS after
.^^i^.^.a^i^aM^^M«^'^ —•**"■-"
0.0 l-UI
xr
£*3o eo o o
oc o ou. Q£ UJ
Ul <
>-1-
to
o
tc UJ
u I- o
mi \0 UJ
z UJ z
3 X (9 U 1^ t-* U. h-
_J =>
Q UJ
u UJ QC
o t
»- o —J >—I Q.
.^-^^^-^■■..^^^.■^■.■^■......^-.^.■-^^.■■..^^^^
'' . : _. .-■ _!_.. -_ -^...^J» ■, ..- -.'. ./
<
Ui
UI CQ
a ui i- => as
oc
•—« a
C9
ui > HI
UI
uo
o
Z
z: UJ ui ;* »-- LU CTJ
LU :2t CQ r^ o
►—t
Q UJ l~ UJ a:< H- ID —1 n CX> —1 CO t-l UJ ►—* u. o Q: z h- < c/> o 1—*
Q u. o
Z o z 1—1 o 1— l-H
<c 1- _l 13 _J CQ UJ «—• o CX' z h- <c l/l o >—1
Q
CD
CJ
11
.:..:,■,..... ^- „.^.c. .^.. .^,., i .. • „,. . I
m^ir-ws ,,,,.,,..,,,.-,,.,,,, - . .
N 3: LU Lu jt a a.
OS o
CD ■o
oo US CO < a.
D> N X U-l Lu ^ o a.
>—I 00 O 3 o
o z u.
O. Si
< UJ u. z •—I
UJ Q
a 2= z o «tt-. 00 1- Q«a: «ca: o UJ a: o. eoo
z: oeoe < UJ <c Q: :»UJ t3 oz <
►—* »-H
Z _I Q o >-i a: iw; t-o o «tu. 00 o _J —1 —i _i UJ 00 UJ UJ en o > =3 _l Z UJ CD «C ■t—1 ►i =3 <J U. 1-
o: Q. n: uj UJ o 3 <-> t-iO z
«o are». o o o >-i oet- <
o= z Q: U. U. D. Z u. 0</) => UJ 00
o c^>o 1— Z I-H u z O l-l U. < •-■ b. O I— 1-H _J X o UJ —i o: UJ Q z o -^ oe z zac x ->*t*ct- t- Q 3: z
to oo z 1— O »-lO o o t-" i- s: _i _i u z: >-• U. =3 C3 o a. o: z: «t <_>
■-I cvj ro ^- in
■^...■■.A».>..,.. ~-...a-^.^iiMiill(|MM^^^ wiimm
; ■ , .•■,.■■ .^1 '. '_ '■1t.....--<.»l»i ■ nr n,- . . A
upconversion. The combined signal plu? pilot ore upconvcrted to RF and amplified to a +28 dBm level for transmission.
The concept of using an ICS to get stable forward gain has been proven. The following design features have also been experimentally verified:
a. Pilot: control of the ICS enables cancellation of the trans- mitted signal without disturbance of the ICS by the re- ceived signal .
b. High isolation can be achieved by the combination of an IP ICS with an RF ICS.
c. The multichannel notch filter ICS provides high cancella- tion over a broad band.
d. The SFR is a linear relay, with AGC control of its output- power level. Thus, it can relay multiple channels within its band.
e. The SFR operates with a single antenna for transmit and receive,
2. HARDWARE IMPLEMENTATION
A breadboard model of the SFR incorporating the features discussed above was designed and constructed according to the block diagram given in Figure 9. Each subassembly was debugged, and the subassemblies were mounted on a chassis for incorpora- tion into a rack-mounted enclosure. Tests were run on the breadboard to determine its performance, and performance limiting factors were identified.
3. EXPERIMENTAL RESULTS OF ORIGINAL SFR
T operatin They als tion whi designed the IF n mance wi the IF t through directed coupling by the t network .
he breadbo g p r i n c i p 1 o uncovere ch needed
These p otch filte th the 100 ransmitted the pilot cancellat of the pi
ransmitted
ard experiments described demonstrated the es en which the SFR design was predicated, d some problems in the breadboard implementn to be corrected for the SFR to operate as roblem areas were: bandwidth limitations in r ICS which prevented its effective perfor- kHz repeater bandwidth; stray coupling of signal into the pilot distribution network insertion coupler which degraded the pilot- ion in both the RF and IF 'S's a n d s t r a y lot into the receiver through a path not shared signal, namely, through the LO distribution
13
■aSkUÜ ■ '■•-^ .....-^^'■.«s.. .^l*'~.'-"'.' .l^.;.!^ ^.^.^..^„^..„.....„■■..■„...„..■..^.■...„..i, ^......L;»^.^^. J.,..,^,^,^,.:.,...,...i. —_
■ ■T»www9sf«w,l#|»»i*l'!-' ^«rf>,*B»»Tr--' ■ , ■ /
III CÜ o t- >- l- -1 X o 1—t z U- CO
■o LU 00 tn </1 ro «s < i- a. ÜJ oo O 00 y «t UJ < I —1 03 a. o
BS '--—id-
Q. CD
O
O
a. CD
'if. BQ siTT
xr
o a:
o
^
'o 1^
o ^ • a.; a.
oo
L__T._IZ.J
O ^) _J Q..
i
%r X
.^ UJ CO O
o •-■ —I o 3; <£
CT* O _l
LU CO
^ _1 tD «I i-i Z u. o
O I- _i a:
14
mmmmmmmmm m— ^ ■.^.^^.
mmm **•,
mwmt* ■ ^■P^P^PPI ■ wi^w^
The experiments did show, however, that the SFR operating principles were valid. Forward gain was measured to be as large as 67 dB; and with modifications to the breadboard, forward gain in excess of 100 dB appeared to be achievable.
IS
- _. —. , ;,j.i:„..-.i,.-.i.,ii.:..,...-.. ..... ,..„,. ...;; wL.m.m.ac.,.^...^a».v. mm
"^'■wfj^wwwww'w'T^^ -y •■'mm
SECTION III
MODIFICATIONS TO SFR BREADBOARD
1. DESCRIPTION OF MODIFIED BLOCK DIAGRAM
A detailed block diagram of the modified SFR breadboard is given in Figure 10. It shows how the breadboard is parti- tioned into subassemblies, designated by letters. The functions of each subassembly are shown in the diagram. Detailed schematic diagrams are given in Appendix A.
An input signal from the antenna is coupled into the Antenna Input of Box A through a 6 dB coupler. Also Present on this input are coupled components of the transmitted signal and the pUo?. The fourth port of the same 6 dB coupler serves to provide a portion of the transmitted signal and the pilot to the Re?Irence Input of Box A. The Reference Input, ^.^f ^^ dB is outputted from Box A for down-conversion to IF for further us^ in JhelF ICS. A larger portion of the Box A Reference Input "attenuated by 6 dB) is adjusted by the RF weight to effect cancellation of the coupled RF transmitted signal and pilot in a 5 dB hybrid combiner/ The RF weight is designed to impart very low intermodulation distortion upon the reference signal.
After RF cancellation, the received signal and cancella- tion residue are provided to Box B where they are amplified and d ^-c^erterto'^MHz and outputted to Box D A ^le of the RF ICS output is extracted for use as the RF ICS feedback erior signal It is chopped at a 455 kHz rate. Chopping sidebands that ajpear'in the main line of the error coupler are at multiples of 455 kSz from the RF, and will subsequently be rejected by 100 kHz wide filters at. IF.
The chopped preamp output is fed to Box C, where it is correlated with I and Q versions of the up-converted pilot to generate the I and Q control signals for the RF weight.
The output from Box A carrying the 26 dB attenuated version of the Box A Reference Input is connected to Box M where it is further attenuated by 20 dB and hen down-converted to IF The attenuation is necessary to avo i g^^^1?^1^^^^ ^^ distortion in the down-converting mi5er:1 Th^r^u ^s
S^e containing the transmitted signal and pilot at IF, forms the reference input for the IF ICS weights. It is provided to the three ?F IC^weights, located in Boxes F, G and H, from output taps on a cascade of 30 MHz notch filters located m Box E The taps are extracted by 3 dB hybrid splitters. Each stage of notch filtering is separated by an amplifier to provide isolation betweei the notches and to restore some gain to the notched signals.
16
'^—'i-^--.....,^-^..^.^..;, ^....■.,. .^.^..,i.^.^...
r. J
^ __'_ - -- ■ - - ■"■■■ ^ - "— '^■^^«—'• " *
\
^ '■■ ■
The outputs of the IF ICS weights are all combined in Box D, then amplified and combined with the IF Main Input to effect cancellation at IF. The resulting signal is further amplified and supplied to Box D', where it is split into two paths. One path .passes through a 100 kHz wide crystal bandpass filter and1 out to an IF AGC amplifier.
The other path, signal, is chopped at a three in-phase (I) and one Q output are provid weight control units (B and Q weight control vo I and Q inputs with the cascade of notch filter in Box L and is almost
serving as the IF ICS feedback error 455 kHz rate and split into six outputs,
three quadrature (Q). One I output and ed to each of the three identical IF oxes. I, J and K) . In these units the I Itajes are generated by correlating the pilot signal obtained from taps on a
s. This notch filter cascade is located identical to that in Box E.
The output of the IF AGC amplifier is further filtered by a 100 kHz wide crystal bandpass filter. It is then combined with the pilot signal in Box M' in a 21 dB coupler.
The combined signal and pilot are up-converted in Box M to the 291.8 MHz RF. They are then amplified to the point where the transmitted output signal is at +30 dBm, and the transmitted pilot is at +18 dBm.
2. SYSTEM MODIFICATIONS
The SFR was modified in several significant ways to improve system performance. These modifications will be discut^ed in this section.
a. IF Rejection in RF Front End
Two 30 MHz notch filters were incorporated into the RF section of the SFR receiver (Boxes A and B). These notch filters prevent oscillation caused by residual 30 MHz IF signal en^nating from the transmitter and proceeding through the receiver with enough gain to support an oscillation.
b. Isolation of Transmitted Signal from Pilot
Both the RF and IF ICS's are designed to be controlled by the pilot only, not by the transmitted signal. In order to keep the transmitted signal from leaking into the pilot supplied to the ICS correlators, an amplifier and attenuator were added in Box M' between the three-way splitter that divides the pilot into three paths and the 21 dB coupler in which the pilot and transmitted signal are combined for transmission.
18
-""-"-^ i I ' '■ ■■ ' — ■■ ■ „^,:^;,,...--.-J..-..„-,,i.,..;,...v.:........ -■.. .-..,^..„.la.^
—ppi^a^TBiWTfr..--.- ^!,.«,s(B^aw^iii,j,,rwM^^^^ ^ ... -^r.
Gi.
c. Redistribution of IF ICS Loop Gain
In order to improve the cancellation limitation imposed by DC offsets in IF ICS weight control units, loop gain in the IF ICS was redistributed. A 12 dB amplifier was added in the IF ICS error signal path, and a corresponding amount of DC gain was removed from the weight control units.
d. AGC Time Constant
The IF AGC amplifier (RHG #EST 3002) was supplied with a fast time constant. Thus, when an AM signal with modulating fre- quencies between 300 Hz and 3 kHz was supplied to it, the AGC would follow the AM and strip it from the carrier. Spectrally, this could be seen as greatly suppressing the modulation sidebands of an AM input signal. The AGC time constant was slowed so that no reduction in the sidebands was apparent when an AM signal with 75% modulation by 400 Hz was applied.
The slow AGC performance is shown in the spectrum analyzer photographs of Figure 11. The upper photograph shows the AM input, with the 400 Hz sidebands -10 dB below the carrier. The lower photograph shows the AGC output, with the 10 dB ratio of carrier to 400 Hz sidebands maintained. The AGC does introduce a small amount of distortion evidenced by the lower level 800 Hz and 1200 Hz sidebands at its output.
e. Drivers for RF Bipolar Attenuators
The drivers for the RF bipolar attenuators are voltage- controlled current sources incorporating nonlinear shaping to compensate for the nonlinear control characteristic of the bi- polar attenuators. The shaping is accomplished by a piecewise linear approximation to the desired nonlinear characteristic. The break points in the piecewise linear approximation were ori- ginally determined by resistive voltage dividers operating from the -15V system power supply. It was found that the small ripple on the power supply voltage caused slight variations in the con- trol currents for the bipolar attenuators, which in turn caused a low level modulation of the RF signal and pilot passing through the attenuators. This modulation introduces some low level sidebands on the output of the RF ICS, which is fed to th3 IF ICS. These sidebands are not present on the reference input to the IF ICS, and hence are not cancelled.
The break point network was redesigned using nigh value resistors to set the break points, so that a bypass capacitor provides heavy filtering before they are provided to the cur- rent drivers. Buffering of these high resistance voltages is obtained by operational amplifiers and emitter followers to provide the voltages with the necessary current drive capa- bility. The new circuit is shown in Figure 33 of Appendix A.
19
h .. ■■ .■ :^.^..A.^^,:^:^,. . ^.-u.:.^ ■..,.,.^.._.,;..:..-..^\, iUafi
|^iMi..J'!J^w.iWffli...ii;i,"""^^ ' ■- ■ vrrvrFmjrrfmvwvs'w'- .i5iwii»<»»^wi5i--^^•wmmf™. , / .r»
Horizontal Scale: 500 Hz/division
Vertical Scale: 10 dB/division +10 dBm top line
Upper Photo: Amplifier Input with 75% AM by 400 Hz sinusoid
Lower Photo: Amplifier Output
FIGURE 11
PERFORMANCE OF AGC AMPLIFIER WITH SLOW TIME CONSTANT
2 0
Li ■^■' ■■ ■■-■--■■■ .■■../.-.l.--^l.^..-. „
£. High Q Notch Filters
It was determined in the first phase of the SFR develop- ment that for the three-channel notch filter IF ICS to work properly with a pilot having 100 kHz bandwidth the notch filter bandwidth should be about 300 kHz at the 3 dB points. Inductors having high enough Q's are not available at 30 MHz to achieve this bandwidth with any reasonable notch depth (in excess of
20 dB).
An examination of the transfer function of the single- stage notch filter in Figure 12 illustrates why a narrow, deep notch is so difficult to build at RF with ordinary lumped induc- tors and capacitors. The transfer function corresponding to the notch filter in Figure 12 is
„21c2
S + Q- a.0S + Wo
H(S) - -2—r—r^~T S + Q7 "ü5 + "0
ii
(1)
The notch depth (term defined in Figure 14) at the center fre- quency is |H(ja)o) - QL/QU- Thus, for a notch depth of 0 1 (-20 dB), the unloaded Qu must be ten times larger than the loaded Qi. For a QL of 100 corresponding to a 300 kHz bandwidth at 30.0 MHz, the Qu must be 1000, an impractically high value.
Figure 13 shows the approach taken to build a narrow, deep notch at 30 MHz. The signal is divided into two channels, a straight-through channel and a channel wif.h a single-pole bandpass frequency response. When the two sign^l5»^, JubtJ^ted' the overall response is that of a single.pole notch filter but with one important feature -- the notch depth may now be con- trolled with just a gain adjustment; resonant elements with ulta-high Q's are not needed. The transfer function of a single- pole bandpass filter is
Sl * .o0S/QL ♦ ü)0
a)0 = l/ZTÜ - center frequency
Q. = R/2ü)nL - fn/Af, where Af is the 3 dB L u u bandwidth
K - gain factor
A little algebra then shows
21
■jaatiwiiaWMi» . --^ ■•■ ^^-^«^-^^■«■a--«- ; dm
- ..,, ^■.'...^ ...."_- i^fc-JLt/l^toW j l<l -J—
Kü)0S/Q
H(S) = 1 - F(S), = 1 - -j j- + %b + OJQ
CL
S2 + l-(l-K)a)0S + w2
s2 + t£ wos + (üo
(3)
Comparison o£ this expression with (1) gives the promised result for notch depth when K^l: |H(jwo)| = 1-K. Notice that for K slightly greater than one, the transfer function zeroes are to the right of the jw-axis in the S-plane, causing a phase reversal at WQ. (For K=2, H(S) would be a classic all-pass transfer function with pole-zero pairs symmetrically placed about the jw-axis.)
A circuit breadboard was constructed whose schematic diagram is shown in Figure 15. A differential amplifier, the yA733 video amplifier, performed the subtraction of the two channels. Some of the properties of the yA733 that proved de- sirable in this application are its 100 MHz bandwidth, 48 dB common mode rejection ratio at 30 MHz, and high input impedance. The 1 dB compression point in this particular circuit, is +5 dBm. The amplifier is followed by a step-up "L" section matching net- work to present a high, impedance to the amplifier output, thus eliciting the highest compression point.
A trimmer capacitor permits the center frequency of the notch to be adjust ±250 kHz about 30 MHz, although the gain factor K has to be reset for each change in the center frequency. The bandwidth of the notch can be broadened by loading the band- pass filter tank with a shunt resistance. For example, 2.7 kil- ohms shunted across the tank increased the notch bandwidth from 350 kHz to 670 kHz.
The breadboard test results for the notch filters is giVen in Table 2.
The notch filter chains in Boxes E and L were redesigned using these new notch filters. Schematics for these boxes are given in Figures 35 and 40, respectively, of Appendix A.
Setting up the notch filter must be done methodically. A random search may never get there. The tuning procedure follows:
1. Temporarily disconnect the straight-through channel to the differential amplifier.
2. Tune the input 30 MHz bandwidth filter to 30.0 MHz exactly. If the trimmer is at its minimum or maximum, use a more
22
.^■.«a^^,^., ^.»,A^.„^,.^.:..i,^.^^^ -.^ ■..■■ ,. - ,-- ■ ...- - ■ .... .■ . .
_______._.:■:_ . : i-/..«.I .;..•■ _ JiÄv ' iilmn ■ « i i . -- - ■=-*
eg(s)
S2^i^oS^c H(s) = -T—J S + Q u ns + u.
u0= i/v^rr Qu^oL/r QL= 2CJ0L/R
A j^
H x 0--OJ,
H CJ 0
QL
U 0
-^-(r
FIGURE 12 USUAL SINGLE POLE NOTCH FILTER
23
w^mi^feitt#i^MiiaB^ii^^ . . . :_ __^. . . -
_____m_________ : ••' ' ■ /,.. J i ■■^^■ri~^*m. >—
eo OH BANDPASS FILTER TRANSFER FUNCTION
u;os/0L
S2 + ^ W S + W 2
ea(s)
= H^) = S2 +^(1^) CJ0S +(J0
2
QL
s2 + ^- CJ0 s + w0 2
FIGURE 13 ANOTHER APPROACH TO NOTCH FILTER SYNTHESIS
24
..,_„ _•„ ,^_.^ __. 1iMa^.Mi&i^i^>M^.^.«a. i '■friiiiiiriiiMi
fr
dB. 20 log HÜCJ)
NOTCH DEPTH
-NOTCH BANDWIDTH
NOTCH CENTER FREQUENCY
FIGURE 14 DEFINITION OF TERMS
25
■■.-..,..-..^.^J. :-- '- - ■■ ■ ■ .-.:.:....~:L
^wgJWBSiBfi!«!^^ Bp
N
UJ
1/1
10 o:
N I
o
o CO
26
^.^...^.^^..■■.■^^.^■:^m;.^^^^^^..J>^^J,^^..
MpjpWipjM'W'WW-W!^^?^^ . . ,' • ,/■
Table 2
Test Results on the Breadboard Model
Notch filter center frequency Notch filter 3 dB bandwidth Center frequency adjustment range with trimmer capacitor
Overall -3 dB frequencies Worst-case notch depth during localized
heating or cooling with heat gun or freon
Worst-case change in center frequency under same conditions
Maximum power gain (at 31 MHz) 1 dB compression point at 31 MHz Power supply
30.0. MHz 350 kHz
^250 kHz 2 MHz S 60 MHz
25 dB
+40 kHz when cooled -25 kHz when heated 22 dB + 5 dBm -15 volts
appropriate value for the fixed 330 pf capacitor,
3. Move the inductor tap up or down until the 3 dB bandwidth is approximately 400 kHz ("up" increases bandwidth).
4. Feed in a signal at 30.0 MHz and ;:!ark the output on a spectrum analyzer as a reference point. Then disconnect the bandpass filter channel and reconnect the straight-through channel.
5. Adjust gain factor K until the same reference level is reached. Reconnect the bandpass filter channel.
6. Judiciously tweak the: trimmer capacitor and the gain factor pot to achieve the desired notch depth.
27
_____ , aate^—u—
^ _ '.--.jj -._ .a^^ififciiii'i > 11 >11—-.. *
■
SECTION IV
SFR PERFORMANCE
The ultimate figure of merit for the SFR is the amount of stable forward gain. This, in turn, is determined by the amount of isolation provided around the SFR loop. The bulk of the isolation is achieved by cancellation in the RF and IF ICS's, and a low VSWR antenna provides an additional 15 dB.
The ICS's in the SFR operate under direction of a biphase psuedonoise PSK pilot signal with 100 kHz bandwidth that is added to the transmitted signal. Ideally, whatever cancellation is achieved on the pilot signal will also be achieved on the transmitted signal. Performance results will be presented in this section showing cancellation of the pilot signal alone, cancella- tion of the pilot plus the additional transmitted signal, and forward gain.
1. PILOT CANCELLATION
In order to measure pilot signal cancellation, the SFR was opened by disconnecting the output of the IF ICS (Box D1) from the inpu^- of the 1^ AGC amplifier. All measurements were then taken at the IF ICS output to avoid the need fov accounting for gain factors between various points in the system.
A sequence of five spectrum in Figure 16 which shows the pilot photo at the upper left (#1) shows t with its input level 60 dB lower th abled. (Disabling an ICS is accomp weight reference input and terminat 60 dB reduction is done to avoid sa pilot in the SFR receiver, and to a used on the spectrum analyzer disnl All other photos in Figure 16 are t its full operating level (-8 dBm).
analyzer photo cancellation p he pilot at th an usual, and lished by disc ing the opened turation on th How the same ay as on the s aken with the
graphs is given erformance. The e IF ICS output all ICS's dis- onnecting the ports.) The
e uncancelled scale to be ubsequent photos pilot input at
Photo #2 shows the pilot output with only the RF ICS enabled. It shows cancellation of the pilot in excess of 60 dB.
Photo #3 shows the pilot output with the RF ICS and the first IF ICS loop enabled. The pilot cancellation at the band edges is not improved much over that achieved by the RF ICS alone, but the spectrum has developed a notch in it due to dispersive effects. The second and third IF ICS loops correct for these dispersive effects, as shown in photos resulting pilot cancellation is in the order of 95
#4 and dB.
#5 The
2 8
: ; , -
mmmm^mm mmmm. mm, m«m$mmm»m m
Vertical Scales: 10 dB/div, -30 dBm top
Horizontal Scales: 20 kHz/div
Photo n-. Output with 60 dB attenu- ation at input, all ICS's disabled
löa3/ -JOcfr. Photo #2: Output with full pilot input, only RF ICS enabled
JKHI MS jgW
FIGURE 16
PILOT CANCELLATION IN RF AND IF ICS'S
,•'1
m
-JWI» -; *
1 >: ;4V :^ ,;. ; '
litÜ< ■. i , 'u.ii*' .
Vertical Scales: 10 dB/div, -30 dBm top
Horizontal Scales: 20 kHz/div
Photo #3: Output with RF ICS and first IF ICS loop enabled
Photo #4: Output with RF ICS and first and second IF ICS loops enabled
Photo #5: Output with RF ICS and all three IF ICS loops enabled
FIGURE 10 [concluded]
•; (i
.-
mmm* if>m^mmmmmimm^9&t
T>J'!WTOll^«t»J|,|UW.(l|^|l.W«»»>»lIlJWW)f' 7m.^-^.^?-™^'w-'-i.'>.<!J,.™!W!m!f*'miMmm*<»;Mm nwwm!^^' p ■
2. CANCELLATION OF TRANSMITTED SIGNAL AND PILOT
Cancellation of a transmitted signal along with the pilot was tested by inserting a 30 MHz CW test signal at the IF trans- mitted signal input to Box M'. The input level was set to pro- duce a desired corresponding output level at 291.8 MHz a: the SFR antenna port. The output of the IF ICS was monitored on the spectrum analyzer.
Figure 17 shows photographs of the spectrum analyzer dis- play under two conditions of transmitted signal output. The upper photo is the result with the SFR transmitted signal output level at +23 dBm. With 15 dB isolation in the antenna, 5 dB loss in Box A, 13 dB gain in Box B, 8 dB gain in Box D, and 4 dB gain in Box D', the equivalent level at the IF ICS output with no cancellation and without saturation would be +26 dBm. The cancellation residue is -64 dBm, producing a cancellation ratio of 90 dB on the transmitted signal. This cancellation ratio holds as the frequency of the transmitted signal is moved across the 100 kHz band.
The lower photo shows the result when the SFR transmitted signal output level is +28 dBm. Although the transmitted signal is 5 dB higher than above, it may be seen from the photos that its cancellation residue has increased by 10 dB, producing a can- cellation ratio of 85 dB.
Cancellation of high level transmitted signals is not as good as pilot cancellation, and it degrades as the transmitted signal level is increased. This effect has been determined to be due to nonlinearity in the complex weight of the KF ICS, is discussed more fully in Section V-3.
It
3. FORWARD GAIN
The SFR has one RF port (the antenna port) through which both input and output signals flow. In order to separate input and output signals, a network was made with a directional coupler and attenuator, as shown in Figure 18. Loss through the network on the output signal was measured to be 17 dB. For the input signal the loss was measured to be 27 dB. Thus, the SFR forwa 1 gain may be computed using the formula
JSFR (dB) PSA(dBm) - PGEN(dBm) + 17 27
PcA(dBm) GEN (dBm) +44 (4)
where P^A is the power level displayed on the spectrum analyzer SA
'GEN is the power level from the HP608 generator
31
^_-___ _______ _ mmm _____ sm _—
/
Vertical Scale: 10 dB/division -30 dBm top line
Horizontal Scale: 20 kHz/division
Upper Photo: Transmitted signal output at +23 dBm
Lower Photo: Transmitted signal output at +28 dBm
FIGURE 17
CANCELLATION OF PILOT PLUS CW TRANSMITTED SIGNAL
3 2
r^mxr H^~«i«B^^HW"-^^P mm
00 I
•-H O CO O _l o; <—I T3 UJ D-
_j o: o I-H o S O »-H Q O
i-i 2: ac Z =3 Ul O CM DC M OS ^1 I— >- (— _l O _J i^ r^ LU «C UJ Q. Z t— 00 <:
2:
00 <: UJ 2:
<: CO o
a uj cc a: ct
O
Q-
UJ 00
1- oo
3 3
m. . mtMmimmm&mt^mmä ____^_^^_^^
J ' /
Photographs of the SFR output are shown in Figure 19 for three difference generator levels: -36 dBm in the upper photo, -41 dBm in the middle photo, and -46 dBm in the lower photo. The output display shows a constant +12 dBm output level, corres- ponding to +29 dBm out of the antenna port. The forward gain for the lowest input level is computed from Equation (4) to be 102 dB
Varying the input frequency across the 100 kHz band causes a 5-10 dB ripple in the output level when the generator level is -46 dBm. When the generator level is -36 dBm, the ripple is no larger than 1 dB. Thus, the isolation around the SFR is just sufficient to achieve the 102 dB of forward gain.
34
Vertical Scale: 10 dB/division +20 dBm top line
Horizontal Scale: 50 kHz/division
Upper Photo: Generator output at -36 dBm
Middle Photo: Generator output at -41 dBm
Lower Photo: Generator output at -46 dBm
FIGURE 19
SFR OUTPUT FOR VARIOUS INPUT LEVELS
35
— i^ü — - - ■ ■ ■ - ■
SECTION V
PERFORMANCE LIMITING FACTORS
Several factors have been identified as ones which place limits on the achievable cancellation in the ICS's, and hence on thP achievable SFR forward gain. The first factor is caused by In interaction between the pilot and the transmitted signal which woulS cISse ICS control loop instability without very narrow loop filtering. The second factor is circuit noise in the RF ICS control loop (primarily 1/f naise and popcorn noise) which the TF TCS is too sluggish to track out due to its narrow loop fil- er ng. The thirff actor is a slight amount of nonlinear distor- tion in the RF ICS complex weight which causes a small change m the ra?io of pilot-to-?ransmitted signal, but a change large enough to present cancellation of the transmitted signal in the IF ICS by the same amount hat the pilot is cancelled These three factors are discussed more explicitly in the following
subsections .
1. TRANSMITTED SIGNAL, PILOT INTERACTION
It has been found that when the ratio of transmitted signal-to-pilot is strong enough, it can cause a P^'^^cted Tcfto become unstable. This effect can be understood l^appli cation of the following line of reasoning. Assume initially thaJthe pilot-directed ICS is cancelling pilot only, and then Jhe level of the transmitted signal is gradually increased The ^11 transmitted signal is cancelled by the same amount as the Tilot However! since the transmitted signal appears in the P i ; Knni thPre are some correlation components between the Tran minted signal res due and the pilot wKich pass through the loon lowpass filter and deviate the weight from its original loop iowpass rii weißht results in an increase in cancel- r.tion^esl^e both oi fhe pilot and of the transmitted signal The increased pilot residue generates a control voltage which
high enough, cancellation abruptly ^ases. For * sP^d P^0t
in which the spacing between its spectral lines is mucn ^esb than the ICS l^op blndwidth, it has been found experimentally that the condition for stability is
Pp BCL
(5)
36
aäamlkiü __^___________________
i^qp^Enfl^^nn^^^ . • -;jr ' .
PT is the transmitted signal power P is the transmitted pilot power B^ is the pilot bandwidth BS, is the closed loop bandwidth of the ICS
Another way to view this phenomenon is that the ICS cor- relators provide a form of spread spectrum rejection of the transmitted signal in the control loop. For the control loop to remain stable, sufficient rejection must be provided so that the post-processing level of the transmitted signal is below that of the pilot.
The conclusion drawn from the result in (5) is that loss of cancellation due to transmitted signal-pilot interaction is prevented by making the loop bandwidth very narrow. If the transmitted signal is 12 dB stronger than the pilot, and the pilot bandwidth is 100 kHz, then the ICS closed loop bandwidth must satisfy
BCL < 6 kHz (6)
Since the closed loop bandwidth is the product of the open loop bandwidth and the loop voltage gain, a loop with 60 dB of gain needs on open loop bandwidth less than 6 Hz.
2. NOISE MODULATION OF THE RF WEIGHT
Circuit noise generated in the RF ICS weight control unit will cause slight variations in the setting of the complex weight These weight variations will, in turn, modulate the reference signal flowing through the weight. Since the modulation compo- nents are not present on the main input to the ICS, when that is combined with the weighted reference for cancellation, the modu- lation components will not be cancelled. This effect is illus- trated in Figure 20 both for cancellation of a CW signal and cancellation of a spread signal.
Following the RF ICS by a second ICS at IF will allow can- cellation of these noise modulation components only if the IF ICS loop bandwidth is wide enough to follow the modulation, thereby duplicating the modulation in its own weight. The ICS loop bandwidth is restricted, however, in order to prevent the transmitted signal-pilot interaction discussed previously.
In the time domain the noise modulation causes the can- cellation ssidue to fluctuate. In other words, the cancella- tion, and nence the isolation provided by the ICS's is time- varying. Once in a while the isolation is momentarily reduced to a point where it is insufficient to keep the SFR from oscil- lating. This effect may be seen by examining the SFR output
3 7
L_____________^________________^^^_^_^^^____________^_______^_
CW SIGNAL SPREAD SIGNAL
MAIN INPUT
WEIGHTED REFERENCE
CANCELLER OUTPUT
FIGURE 20 MODULATION EFFECT OF ICS CIRCUIT NOISE
38
_.__:;„.;„_ . __._._ .__ _..^iÄitf»S
mmr- . ,y
with no input signal, and full forward gain tion, the output consists only of the pilot,
Under this condi
Figure 21 shows spectrum analyzer photos of the pilot at the SFR's (antenna) output. The upper photo is the pilot output with the SFR loop open. The lower photo shows the SFR pilot out put with the SFR loop closed. Spectral lines can be seen which are not steady, but pop up at random locations and random times.
3. RF WEIGHT LINEARITY
The third factor found to be limiting SFR performance is the linearity of the complex weight with respect to RF signals flowing through it. Deviation from perfect linearity is evi- denced by the generation of third-order intcrmodulation products between the pilot and the transmitted signal. Thus, even if the pilot and transmitted signal could be cancelled perfectly, a residue would remain which is caused by third-order intermodu- lation in the weight.
A more serious problem exists, however, which arises from the same cause. That problem is that the weight nonlinearity, in addition to generating intermodulation products, also changes the ratio of the transmitted signal to the pilot. For example, if this ratio were changed by II, then perfect cancellation of the pilot would result in 40 dB cancellation of the transmitted signal.
It is possible to quantitatively relate the ratio change, and hence the achievable transmitted signal cancellation, to the level of the third-order intermod product generated. To do that we refer to the simple block diagram of Figure 22 in which the cancelled output voltage is given by
(a-W)e, Wk3ei where
(7)
the main input of
c v- -• 1
ei is the referenec input voltage, a is a complex scale factor that relates
the ICS to its reference input, W is the complex value assumed by the complex weight, and ks is the coefficient of the cubic term that describes the
weight nonlinearity
The reference is composed of two tones
e. = VTcoswTt + V cosu t,
where Vy is the amplitude of the transmitted signal tone UT is its radian frequency, Vp is the amplitude of the pilot tone and (jon is its radian frequency
(8)
a n d
39
..... . . . ,
PILOT AT SFR OUTPUT WITH SFR LOOP OPEN
Horizontal Scale 50 kHz/division
Vertical Scale: 10 dB/division
ioa/ me fa
'V (:■%%'
PILOT AT SFR OUTPUT WITH SFR LOOP CLOSED
7is i "> -)■ n w FIGURE 21
SFR PILOT OUTPUT SHOWING EFFECT OF NOISE MODULATION OF RF WEIGHT
40
.,(-■ I
' , _ I '. :, ...''t.-U+m. - i — ■
ro T-
CO
i
a: UJ
OJ
t o
CD i—i
X UJ + / 3
/ 1 *" en UJ Q
o i
(-
CO -r- 0) CO
cc 1—1
X h-
ra u_ D- + CO o
t—* i— CM h-
3r UJ a:
UJ
^C ID U- n o
CD 1—I
U-
UJ
LL. o O0 1—<
00 >- —1 X
UJ h- < Ö
_l rr. 3 Q. cs ct
SL 1—t
o uj D.' u 3
UJ o
i i i k Lu
z t- 00 LU rs
< u- u. UJ z 1—1 u. »—t h- UJ Z a: UJ
•t- ) (J0 in UJ
c J s: */ » =>
o o
II
41
„ wi
Insertiug (8) into (7) gives
ec = (a-W) (VTcosü)Tt+V COSüJ t) - Wk3(VTcosü)Tt+V cosw t)
3v; = [(a-W)VT - -^Wk^
3VTV
3VTV Wk, T p 3-, . 2^ J COSWrpt
3V^V J-£k3Wcos(2ooT-a) )t ^k3Wcos(2a) -ü)T)t
3V' + [(a-W)V - ^Wk.
3vjv I>i
p 4 -3 —T^Yl) C0Swpt
It may be seen from (9) that the ratio of the amplitude of cos^xt to the amplitude of coscüpt has been changed from Vj/Vp by an amount determined by the third-order coefficient k^.
C9)
Ws may simplify (9) by assuming that the transmitted signal is much stronger than the pilot,
VT » Vp (10)
and by defining the amplitude of the stronger third-order inter- mod to be
VI =
3vjv T P k ~^r- k3 We have
[(a-W)V V^jW
T ■ —y—] cosaiTt - VTWcos(2w -a'T)t p J^ P *
+ [(a-W)V -2VTW] cosa) t 1 v ^ p I J p For perfect cancellation of the pilot we have
2V, ot-W I W
(ID
(12)
(13)
which gives V V T I ex T I
c ~ r+2V /V '■T— COS^T1 " V cos(2a) -ü3T)t]
I D P '
For ' >> VT » we hav6
e - af-^ c l V
VTVI cosoojt - VT cos (2a) -(jüT)t]
(14)
(15)
42
From (15) it can be seen that the amplitude of the trans- mitted signal cancellation residue
aVcT = aVjOyVp) (16)
is equal to the output amplitude of the stronger third-order intermod scaled by the input ratio of transmitted signal-to-pilot. For example, if the output intermod is at a level corresponding to 100 dB of cancellation, and VT/Vp = 10 dB, then the transmitted signal cancellation is limited to 9u dB.
Recent measurements of the third-order intermodulation performance of the RF ICS complex weight have shown that with the transmitted signal output level at +28 dBm and the pilot output level at +16 dBm, the third-order intermod generated by the RF ICS complex weight appears at the IF ICS output at a level of -68 dBm. According to the above theory, the transmitted signal cancellation residue at the same point should be -56 dBm (12 dB higher). This agrees with the level of the cancelled transmitted signal measured at the IF ICS output, as shown in the lower photograph of Figure 17.
It should be noted that the initial data taken on a bread- board of the RF complex weight showed a third-order intercept approximately 8 dB higher than is currently being measured. A check of the RF ICS complex weight circuit showed it to be operating properly, except for the third-order intermod perfor- mance. Further work is needed to reestablish the intermod per- formance to previously measured levels.
*Abrams, B.S., et al, "Adaptive Same Frequency Repeater (SFR) Study," Final Report RADC-TR-76-78 on Contract F30602-74-C- 0242,'March 1076, p. 62.
43
____
._. fW. >'■ ■■ ■ - .;• ". "" ■••■■ii'-iTiftiii'i ii ^~ .
SECTION VI
RESULTS WITH OTHER PILOT WAVEFORMS
The bulk of the results reported herein were achieved vrth a pilot generated with biphase PSK modulation of a carrier by a pseudorandom noise source. The pilot was then confined to a 100 kHz band by crystal filtering. Other pilot types were tried briefly with varying results. These results are presented in this section.
1. FM PILOT
A sinusoidally frequency-modulated pilot with 100 kHz bandwidth was tested briefly. It was found that can"J^tion of the pilot and the transmitted signal was always worse than with the PSK pilot. If the modulation rate was high, then the pilot spectrum contained few lines, each with a significant faction of the power. If the transmitted signal frequency were located on one bf those lines, it would cause ICS loop instability far worse than when the pilot energy was more evenly distributed across the band. If the modulation rate was lowered so that the pim spectrum contained more lines, each with a lower power level, then the ICS'swould tend to optimize cancellation at the instantaneous frequency, with cancellation ^f^V™. at some, other frequency. The PSK pilot has the advantage °f close spec tral lines without an instantaneous frequency, so that cancel- lation is optimized over the entire 100 kHz band.
2. STRADDLE PILOT
In order to remove the interaction problem between the pilot and the transmitted signal, a pilot structure was tried in Hid the pilot straddles the signal band. The Pilot was gener- ated as shown in Figure 23. In order to make use of this pilot, Se 100 kHz wide crystal pilot filter in Box M was bypassed.
Some difficulty was experienced in getting the second and third !ooJs in the notch filter IF ICS to oP^V^iT* ?his pilot. When they did operate, transmitted signal cancella- tion was about the same as with the PSK pilot At the time the limitations in the RF weight noise m?d^at^" ^"^^I0™ nnniinearitvl were not recognized, and the straddle pilot was no? J^sued It is felt thlt further work with the straddle pUot may be fruitful if the other performance limitations can be overcome.
44
■ 1 .v....ta.Ü.„.- ^
■ * ■ ■ m
O _J
a.
N
O TJ
CO
&-0 +
©
Q CO Q
on Ui I—
ZD CD U- I-H O
2
CD
OC SO z U. (— o
•a: N I-H N _J X H- X-l-^ <
o o > O V) I-H LLl CO O + Q
45
.". ^d : . . _____._________^
: ■ ■
■ ■
^ „ ,,,.; . &CJ» ■ ■ ■
3. SEPARATE RF AND IF PILOTS
Some tests were also run with separate RF and IF pilots. The PSK pilot was kept for the IF ICS, but a CW pilot was used for the RF ICS. The CW pilot was kept outside the 100 kHz band to prevent interaction with the transmitted signal in the RF ICS
The breadboard modification used to implement a separate RF pilot is shown in Figure 24 with a block diagram of Box M'. The upper section of the figure shows the present setup in the system using a single pilot. The lower section of the figure show? the additional components used to add the second pilot.
Cancellation tests run with the separate RF and IF pilots showed performance similar to that with a single pilot. The same limitations of noise modulation of the RF weight and third- order nonlinearity in the RF weight apply to both cases.
46
i
*
0 o IS> a: <_> < ►—i
—•, . u. 5@ 1—1
CO o o fl »— 1— to
1- 1- x: =3 =J h-Q. CL
H- 1- Q = ZJ Ul o o ^ •-•»- H- 00 O O X:-J —1 o •-• •—i o a. a.
J «t Z I O QC O
«/) a z i-
^ CO »—i
cc
»-M i3 »—i
a. « Q »- ^ => o 1- o •-• O
—I I- u- »—i
OQC a. -4 •-IO o: a. »- u.
00
u^
a.
O
to
o z
CM
HH a. U. UJ
to
u. o z o I—I H- ««
a. a. o--» _i^
•-<< •-I «o a. a.
-JO ^-^ <co -j z< «tH- o Z=> •-« x: a: a. to o UJ z
«X »-•-• X u. X
• o O r-t
+ 1 o CO
47
._..._ . , -
^..r^ipp'ipfginippipmp^r^ppifB' ,
'
SECTION VII
RECOMMENDATIONS FOR FURTHER WORK
The experiments on the modified SFR breadboard have un- covered areas in which further work would result in improved SFR performance. They are: reduction of circuit noise in the RF ICS weight control unit, linearization of the RF ICS weight, and further investigation of the use of a straddle pilot. Recommended approaches for effecting these improvements are out- lined in the subsections below.
It is believed that these improvements will allow the present 102 dB of forward gain to be achieved with less than 2 dB ripple across the 100 kHz band and without the momentary lapses in cancellation. It should also be possible to extend the for- ward gain capability to 110 dB.
Following these improvements, further laboratory tests should be run to establish the new performance levels. At that point it is recommended that field tests be conducted on the SFR with a real antenna in a real terrain environment.
Another area of investigation is also proposed. That is, widening the SFR bandwidth to allow relaying of more narrowband channels or relaying spread spectrum signals,
1. CIRCUIT NOISE IN THE.RF. WEIGHT CONTROL UNIT
The itost likely circuit components to be generating the control unit circuit noise are the opeational amplifiers. There are improved versions presently on the market which offer much improved low frequency noise characteristics (e.g., products made by Precision Monolithics, Inc.). Some are pin-for-pin replacements of the currently used operational amplifiers, and do not sacrifice other performance features.
2. LINEARIZATION OF THE RF ICS WEIGHT
Two areas should RF weight linearity. ;Fi tested in order to selec nonlinearity. Second, t should be reexamined exp drive currents into the vide minimum third-order mance measured with the an achievable target for
be investigated in order to improve the rst, individual PIN diodes should be t those with the minimum third-order he design of. the weight driver circuit erimentally to determine if the two DC PIN diode network are optimized to pro- nonlinearity. The third-order perfor- original RF weight breadboard provides this experiment.
48
.. .^«.ä,/...,.:-..,...^ ^ I..-.. ^....L.^., -— '—
3. STRADDLE PILOT
Further experimentation with a straddle pilot should be conducted. Because this pilot straddles the transmitted signal band, the interaction between the pilot and the transmitted signal in the ICS control loops is greatly reduced. This allows the loop bandwidth to be widened, providing greater ability to track out the circuit noise effects.
The straddle pilot experiment discussed in Section VI.2 used FM sidebands. Figure 25 shows a means of implementing a straddle pilot with the two sidebands originating from a PN- BPSK source.
4. FIELD TEST
After the above improvements have been incorporated into the SFR, laboratory tests should be run to establish the new performance levels. At that point, the SFR should be field- tsted with a real antenna in a real terrain environment.
A field test diagram is shown in Figure 26, with spacing between antennas of 50 or 100 meters. Using these relatively short transmission distances allows one test team to operate all the equipment. Path loss attenuation is simulated by coaxial attenuators in line with the source transmitter antenna and the fi :ial receive antenna. The objectives accommodated by this test would be:
a. Measure forward gain and demonstrate suitability for AM voice channels in outdoor environment.
b. Measure effects of reflecting terrain on GFR performance.
c. Test SFR use for multiple channel relaying.
d. Demonstrate dynamic range of SFR.
5. INCREASED BANDWIDTH
Increasing the SFR bandwidth from the present value of 100 kHz will allow greater utility in relaying more channels or accommodating spread spectrum signals. The pilot bandwidth would be similarly increased, as would the IF filter bandwidths and the bandwidths of the notch filters in the IF ICS.
As bandwidth is rxreased, the achievable SFR forward gain inareal environment decreases.* For example, with 100 kHz
*See Section II of RADC-TR-76-78, "Adaptive SFR Study," March 1976
49
_^ -"-+-' -
-
1
i I
Ix. N Q. X CO -^ o
_i CO < ?l f- 3 X 03
'
o
^4
o
00 Q ■;.'_
cc CO UJ Q
o 00 N
x:
00
00
^
CO o co
a
I—
FIGURE 25
ADDLE PILOT WI
N a-, i— CO
o <C en
o
L'->
5 0
■ ■^- -■-^.^.■■■i^^^>^..^^.aM^M^....^ - .^^.^...^ ^_^___^__
...
.0 SFR WITH OMNI ANTENNA >
CALIBRATED REFLECTOR
100M 100M
>
ARC-164 RX
50M
FIGURE 26 SFR FIELD TEST
51
—■ - ■ ■'-■ - -'■'■■■i---" - '•■■ ■ — U
-ii.5>,T-'.^••;.«:.--T v,,,, .,:,,,-.,,„-, ,;rT.!.„^a7-,,. .,,,.„...,„^,,^, ■
:Xm n1 ii lii^m
*
bandwidth, the transmit-to-receive isolation is limited to 116 dB by ground clutter effects. For 1 MHz bandwidth, the Tx/Rx iso- lation drops to 96 dB. Since only 96 dB of forward gain is achievable, the ICS cancellation requirements on the ICS are not as great as would be needed to achieve 110 dB of forward gain with 100 kHz bandwidth.
52
!!WSPlflp»Ri!ipi!Wpp|BWlBfPWPW!pi ■
SECTION VIII
SUMMARY AND CONCLUSIONS
Improvements were made in the SFR breadboard model to increase its forward gain from 67 dB to 102 dB. These improve ments were in the nature of circuit modifications. The basic system structure, and the concepts on which it was founded, remained intact.
were The modifications incorporated into the SFR breadboard
a. IF notch filters inserted into the RF front end to prevent IF ring-around,
b. Isolation of the transmitted signal from the pilot lines to ensure that the ICS's are controlled strictly by the pilot,
c. Redistribution of gain in the IF ICS control loops to reduce the cancellation limit imposed by DC offsets,
d. Decreasing the AGC time constant so that All signals mav be relayed without distortion,
e. Improved isolation from power supply ripple of the drivers for the RF complex weight,
f. Redesign of the notch filter chains for the IF ICS to narrow the notch bandwidth from 2 MHz to 300 kHz without sacrificing notch depth.
Investigations were carried out to determine those factors which limit the SFR forward gain to its present level of 102 dB When used with a common antenna for transmission and reception. The limiting factors are those which limit the cancellation per- formance of the ICS's. The first is the third-order nonlinearity of the RF ICS complex weight which slightly changes the ratio of the transmitted signal and the pilot waveforms passing through it. Thus, even with perfect cancellation of the pilot, some residue of the transmitted signal would still remain uncancelled. The second is circuit noise in the RF ICS control loop which imposes a low level noise modulation on the signals passing through the RF ICS complex weight. The IF ICS is not fast enough to follow this noise modulation because its closed loop bandwidth is constrained to be narrow to prevent an interaction between the pilot and transmitted signal in the ICS loop which would severely degrade cancellation.
53
•^^ei^^^-^ -_^ MHliiiü _ mmmm mm
. inpmqjinj, ■ WSSW^PfKSWWSJHl'': s^ppppp^p; ^ppp^pp^j ... ■ ./■ . üüeL. mitUSI
Further modifications to the breadboard to overcome these limitations have been outlined. After further performance re- finement, the SFR breadboard should be field-tested,- where natural reflecting terrain will provide a far more realistic performance test than can be accomplished in the laboratory.
54
: __._ iüJ
■' ■ ,., . _ j,' ..*:.:.- ..^:-.'aa..-ii.t»i. ■ -i , .. . ' / i
ACRONYMS AND PRINCIPAL-SYMBOLS
BrT = ICS closed loop bandwidth
B = pilot bandwidth P BPF = bandpass filter
C = control unit for complex weight
e- = reference input voltage
e = voltage at ICS output
F(S) = bandpass filter transfer function
GSFR(dB) = SFR forward gain in dB
H(S) = notch filter transfer function
ICS = Interference Cancellation System
K = gain factor in notch filter circuit
k, = coefficient of the cubic term in a power law expan- sicn describing the weight nonlinearity
PrFN(dBm) = generator output power level in dBm
P = power level of the pilot signal
P^.CdBm) = power level into the spectrum analyzer in dBm
p = power level of the .transmitted signal
Q = loaded quality factor of a resonator
Q = unloaded quality factor of a resonator component
SFR = Same-Frequency Repeater v = amplitude of cancelled transmitted signal
V.. = amplitude of the pilot
VT = amplitude of the transmitted signal
W = complex weight value a = complex scale factor that relates the main input
level of the ICS to its reference input level
w = center frequency in radians per second of the notch 0 filter
a) = pilot frequency in radians per second
w = transmitted signal freqeuncy in radians per second
55
.,_ i.— ,....' . _ ___ „ _■...._. _^„t.._ ... _,_. _„ HuU
_ , _. .'....J - . ~~.:..'m..^d,tm ■ I.~*~J J -'. •*
■
APPENDIX A
CIRCUIT SCHEMATICS
56
...-.J„^-^~^-»..^-..-.^...J.^-.
-WTTBr ' Sri.,'film,
r • -^
UJ —> OC 1— Q. U. Z Z lO «t •-«
*i z :r: I-H
oo >- n: . < o
tH _i l— r» üj o CM O z
(*>^ ®-vMy—:
ID
00
I
N
vxAxy
3:
O i CO
*>i\
CVJ vl- 00
00
C£>
cc
o
CSi
UJ 3: o: CD
CS UJ
LL.
X UJ
o o
o
u.
oe i u a' lO ■3t UJ l-t
OQ » Q. •-. U
> C0 M -O Q ac «£»
OQ JT1-! IJD TJ
Ü. UJ Z /| ^ K
57
ii_ ■ .• .„ . _. ': i «.,:ii „..i : __, : ^ ^ : ..m..«.«.^-^«!.». ..--.. .^^--.■-^^-»■'^■.
CONTROL INPUT
CR1 CR2 * W- *
.33
1800
. 33 CR5 CR6 300
-\(r
.33 T2 RF
OUTPUT H M * onnr^ :R7 CR8 rf? CR7
T1-T2 CONSIST OF 3 TURNS*J0 TWISTED PAIR ON Ql CORE
CR1-CR8 ARE UNITRODE UM9303 PIN DIODES
ALL CAPACITANCE VALUES ARL IN PICOFARADS ALL INDUCTANCE VALUES ARE IN MICROHENRIES
FIGURE 28 RF BIPOLAR ATTENUATOR CIRCUIT
58
! . : J.
■
< CM CD SO» T3 t-t ir> CJ
0 a:
CM
OC O =) Q
U. Z < a.
a.
59
—"'■"^"-■-- ■"■■"■ - i ■'-•-^ •^i^-.^.v..:..^-^..:^.^:.. - "'■ v- ^...»VB....
z 3 _l o cc h z o o t- I
o 10 UJ [£
0 o Z(-
a: 3
is a •-
H^ o ^
a. u
/
60
,
VALUE COMPONENT
RF CORRELATOR IF CORRELATOR
C1 300 pf 1000 pf C2 .15 uf .15 uf C3 .15 uf .15 uf C4 .15 uf .15 uf CS 300 pf 1000 pf C6 .15 uf .15 uf C7 300 pf 1000 pf
CR1 - CR4 ARE FAIRCHILD FH1100
5IK WV——I+15V
PILOT INPUT
8 CHOPPED -O——O ERROR
SIGNAL
-15 V
FIGURE 31
RF AND IF CORRELATORS (USED IN RF AND IF WEIGHT CONTROL UNITS)
61
I
- *—Vs/v--« •■
K o o o IO
U 9 o
3 3 o
a
K
a to in IO
> o
K Ul
d u o z
u
u. u. N
1
Hh-^
^--^
Z j -i j 5 < < <
CONTROL CURRENT TO RF BIPOLAR ATTENUATOR
DRIVER iOK INPUT > Q ^
2200
+ 15V + 15V CONTROL CURRENT
TO RF BIPOLAR ATTENUATOR
UNLESS OTHERWISE SPECIFIED: ALL RESISTANCE VALUES ARE IN OHMS, +5%.ijM ALL INDUCTANCE VALUES ARE IN uh, +10« ALL CAPACITANCE VALUES ARE IN uf. +10«
FIGURE 33 RF BIPOLAR ATTENUATOR DRIVER
63
1 ^biWt ■■ ■'-■:-- • ■ ^^^A..^^^.^.^...^^^^.-»..^^^».*-"---»--^» ^lM^M^.,»y^A^^s.,......,.r|.|f ^.fJiliigM^
MCL-PSC-3-1 THREE-WAY COMBINER
«r
rv
♦ 15V
»3.4.7.
,1 uf
a. t^«
f
1 2\
12
IF INPUT
MCL PSC-2-1 TWO WAY
COMBINER
51
T
'"I .1 uf
®£ 3 I u
OPTIHAX AH-592 6 = 12 dB
2.3,4,7,8 OPTIMAX AH-592 G » 12 dB
MERR MAC QHF-2A-30
QUADRATURE HYBRID
4.7.X »
UNLESS OTHERWISE SPECIFIED:
ALL RESISTANCE VALUES ARE IN OHMS tS%, liW ALL CAPACITANCE VALUES ARE IN PICOFARADS
MCL PSC-3-1 __ HREE-WAY COMBINER [DJ
0 MCL PSC-3-1 THREE-UAY DIVIDER
+ 15V
i;22
MCL PSC-2-1 TWO WAY DIVIOEP.
!r ^:.iuf
+ i 5V
MCL SRA-1 MIXER
\6 ♦ i3rj4 J /pZ40^ J.. HDTIMAV * ' i I OPTIMAX
AH-592 G > 12 dB
455 kHz CW ^ T8 FROM N
30 MHz BPF BW=3 MHz
MERRIMAC QhF-2A-3ü
QUAD HYBRID
1 3 3 pf
20 {-.12
43,4,7,8 17
OPTIMAX AH-592
G * 12 dB r
L11 ^8,10
11, L2 22T #32 ON MICROMETALS T30-6C0RE
1 JT i—i
PIEZO 4102F MONOLITHIC
CRYSTAL FILTER fo ■ 30 MHz B » 100 kHz
■I 3s .8-l(
e fIF OUTPUT
^l 1 "■ '
1 ♦ 3,4.7,8
MCL PSC-3-1 THREE-l'AY DIVIDER
30 MHz CRYSTAL FILTER
FIGURE 34 SFR IF CANCELLER AND ERROR SIGNAL DISTRIBUTION
64
mam timmm
1
I^^^IL
: ._ , s -- ■. ■■-;'■*.. :.A
u.a. u. a.
®Hi' SHi'
CO 2:
«3 X
1—1 UJ LU
tCu.
or la- to
@
E
66
1 ■ ' / '
t/5 00 LU Q •-•
o: z <C Lü U. X o o o cc *-* o
ex: ui
UJ U)
ui O UI z o «CZ
»-II- <_> o
O. Q «c z
QL I—
O
U>
«O-rvw^,
r^o coi-
«c uirs a:z ^ui CD»-
u.«c
oc «c _i o D- t—i
CO
67
L ■ — -- — - - - -
O. QC «t H- u o z- rs o ii: CD o. 3: D: »-< z: CJ bJ IT) I—i
00
UJ
o
er
er
OC'— a: o o-
—i «t o => o: Q. z UJ
00 (— t-i (- D:
o cc: —i z o '-1
O I— O ti- o: o z n UJ < «^ oh- v> z: UJ <c >• Q o-
O oL
^
o H <
«t -J o D- i—i CD
O
<
tn
o
o
er er- O o
o: O'
—I et O ID I o. z i »—i LU ; CQ (— I
o
a:
OQ; _| Z o t-i
0:0z ^ LU <: i^> <->(_ 00 ZLU <:
to 3
o
eau
^
D:.- O
o o
^T
u>
et: LU r-i Q 1 t-t (SS >• I *-* t_) Q «^
Q. o: 1 LU _J 3 O os:
UJ
0 B
00 z
o: o
es t— »•H Z U- o
o
co
ro m
CM
o
o
o n
a. i-H
68
.^^l«Mi^,1nj1Wari|ife -^ „*_
■-l»PW»»p«W!!WF^W^ wmtimirmrww^! . /
5ww « IU Q
nil i5| = UJ
_._. - f-,..^^^:^.,a^.^.i^e „■. ■,.:^&.^
_______ / _;.. it 1 • ■ ■
a&iti«it»iViii' ■"' .. .
•!
S" 70
_^ .
Br- '
SH Q < "^
(sH
71
j
- ■.i^ro>^iw^TW7;,w^JW7^;,-p-':
O 1/1
in a 5 o
O i-i o
UJ
z o O ^ «/)
a: to -i i-i _i «t 3 «t >
OJ UJ UJ ro z: O =«= o z
o: < t- o i- (\J HI I—I CM Z U
< a. <
^- o -j
i 00
:5iK^
n
CO
I CO
u. uj N n: C\J »— 3: ^c O _J^ f—I l-H O ^J- u. o o
fvl < H II UJ (- 1-4 U1 O a. >- «♦- CG
o:
oc
I 00
o
O—J ivi < UJ (-
o. >-
o
:is1M
^Ife^
(*H f— (_> UJ r» ta HH a. < u_ as i— t-iaE -J
OQ- u. a: x M u. <c
>-
z o
»—> cr t-
•—i
o CM »3- H-
O UJ _l cy t—i Z3 a.
»-1 Q
to IO < Q- Q Z «t
GQ co x -o <M U- z: CTl CM »—• •-e in r-l h- t Qi a. JZ II u. o< CO
CB
72
__________
HF .. , ■
tn IS) LU O !-• «t Q: ex z «C LU U. X
•> c o ÜO o oc x: l-H «J X Q- •-< o j^31
z >-" z • • h-4 ►—1
Q LU
LU LU de. LU i—l ac «t a: U- <c < H-l </i
<_) m LU I/O u-i
LU LU =3 LU Cd
Q- =3 _l =) o l/l _l «I -J 1-
«t > «C (-3
UJ > > CO LU <•) LU H- t- t—* LU O LU ra LU ^ O z o a. Q
oe z < z ^- UJ < (- «c o ui 3: i- •-" t- o =) l- 00 o <_> O
o 1—( 3 «a: = LU Z IA) ^l- o. o > o
in LU ~^ «c z < o: CO OH t-i (_) I-« 3 X
UJ O
_l _l W _l _I LU z z _l in _J —1 Qt >- Z3 < M «t «t ■t to
=3 cro toi-
(SH
CM CM
O LU
|H SH SH SH
►-wv-^
o ( IT) !
<: CM CM CM CM
CC
o I-
l^
rs to
1~ LU
O Q. O
O O SH cn CM
CM
►-rron—^
o
H(-6 o in
oo
Mvvvv^
S HM
H(-e e
ov o mV ir>
LO
CM CM CM CM Z CM
HM Hf*:.
LO
W> C\J
CO
N —I X
I- z ro LU • Z IT) et in
z =3 II U-
>- Q O LU Z Q; LU ■=) =)
to cy «C LU LU QC z: u.
w CM
CM
Qi o (- <: Q: LÜ z LU
ro O
s: LU CC oco ZD ü- O LU •- >■ Ll_«C
h4 X
in in
to
in i-H
I
I i
o o oo CM ^t
^ o 3» CM
HH
73 in
i
