Upload
clifford-stone
View
146
Download
8
Embed Size (px)
Citation preview
UNCLASSIFIED
AD NUMBER
LIMITATION CHANGESTO:
FROM:
AUTHORITY
THIS PAGE IS UNCLASSIFIED
AD263873
Approved for public release; distribution isunlimited.
Distribution authorized to U.S. Gov't. agenciesand their contractors;Administrative/Operational Use; JUL 1961. Otherrequests shall be referred to BallisticResearch Laboratories, Aberdeen Proving Ground,MD 21005-5066.
BRL notice dtd 23 Sep 1965
BRL 1136 c. 2A s
REPORT NO. 1136 JULY 1961
F l N A L T E C H N l C A L S U M M A R Y REPORT
P e r i o d 2 0 J u n e 1958 - 3 0 J u n e 1 9 6 1
B R L - A R P A DOPLOC S A T E L L I T E DETECT1 O N C O M P L E X
ARPA Satellite Fence Series I
I;'::),.: I !:.?E 6 a t l
A. H. Hodge
Report No. 25 i n the Series
Department of the Army Project No. 503-06-011 Ordnance Management Structure Code No. 5210.11.143
BALLISTIC RESEARCH LABORATORIES
BERDEEN PROVING GROUND, MARYLAND
ASTIA AVAILABILITY NOTICE i
Qualified requestors may obtain'copies of this report frum ASTIA.
Foreign announcement and dissemination of this report by ASTIA is limited.
B A L L I S T I C R E S E A R C H L A B O R A T O R I E S
REPORT NO. 1136
JULY 1961
FINAL TECHNICAL SUMWIY REPORT
Period 20 June 1 9 9 - 30 June 1961
BRL-ARPA DOPLOC SATELLITE DETECTION COMPLM
ARPA S a t e l l i t e Fence Series
Report No. 25 i n the Series
Department of the Army Project No. 503-06-011 Ordnance Management Structure Code No. 5210.11.143
Shepherd Frogram, ARPA ORDER 8-58
A B E R D E E N P R O V I N G G R O U N D , M A R Y L A N D
B A L L I S T I C . R E S E A R C H L A B O R A ' T O R I E S , ,
. .
REPORT no. 1136 . .
J=w8e/l~ AbarQsa Proving ((round, W. m y 1961
FINAL TECHNICAL SUMMARY REPaRT
Period 20 June 1958 - 30 June 1961
BRL-ARPA DOPLCC SATEUITE 'DETECTION COMPLEX
ABSTRACT
This report consists of a technical summary of the research and
development program sponsored under ARPA Order 8-58 which led t o the
proposal of a system for detecting and tracking non-radiating orbiting
sa te l l i tes . An interim research fac i l i ty consisting of an illiumlna-
ting transmitter and two receiving stations with fixed antenna arrays
was f i r s t established t o determine the feas ibi l i ty of using the Doppler
principle coupled with extremely narrow bandwidth phase-locked tracking
f i l t e r s as a sensor system. The interim system established the feasi-
b i l i t y of using the Doppler method and resulted i n real data for which
ccrmputational methods were developed which produced orbital parmeters
fran single pass data over a single station. To meet the expected
space population problem of detection and identification and t o obtain
a maximum range with practical emitted power, a scanned fan beam
system was proposed i n which a transmitter and a receiver antenna
separated by about 1,000 miles would be synchronously scanned about
the earth chord joining the two stations t o sweep a half cylinder of
apace volume 1,000 miles low end having a radius of 3,000 miles. An early
order stopphg further developmant work on the project prevented fFeld
testing of the proposed scanning system in a scale model. The engineering
study indicates that the proposed DOPLoC system or sl ight modifications
thereof may offer scrma uniqw, advatages over other systems for surveil-
lance type of detection, tracking, identification and cataloging of both
active and non-racUating sate l l i tes .
TABLE OF CONTENTS
Page
. . . . . . . . . . . . . . . . . . . . . . . . . . . . I . INTRODUCTION 15
I1 . EVALUATION OF DOPLOC RASSIVE TRACKING SYSTEM . . . . . . . . . . . . 19
A . The Basic ~ 0 . ~ ~ 6 2 stem . . . . . . . . . . . . . . . . . . . 19
. . . . . . . . . . . . . . . . . . . . . 1 . DOPLOC ~ e r i v a t i o n 19
. . . . . . . . . . . . . . . . . . . . . . 2 . . Tracking F i l t e r s 19
. . . . . . . . . . . . . . . . . . . . . . . . . 3 Sensi t ivi ty 20
. . . . . . . . . . . . . . . . . . . . . . . . . . '4 . 'Antennas 21 . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Output 21
. . . . . . . . . . . . . . . . . . B Application To Passive Tracking 23
. . . . . . . . . . . . . . . . . . 1 . Transmitter power. Pt 24
. . . . . . . . . . . . . . . . . 2 . Antenna Gain. Gt and Gr 25
. . . . . . . . . . . . . . . . . . . . . 3 Operating Frequency 26
. . . . . . . . . . . . . . . . . . . . a . Radar Equation 27
. . . . . . . . . . . . . . . . . . b . Antenna B e w i d t h 28
c . Cosmic Noise . . . . . . . . . . . . . . . . d . Bandwidth . . . . . . . . . . . . . . . . . e . Mlnimum Detectable Signal . . . . . . . . .
\ f . Conclusions Relative t o Choice of Frequency
4 . Receiver Power Sensi t ivi ty . . . . . . . . . ~ . 5 . S a t e l l i t e Reflection Cross-Section . a . ~ . . . 6 . Reflections from Other Objects . . . . . . . . . 7 . Solutions of Orbit Parameters . . . . . . . . a
A . Bsckgmund Ewperlence and Developments . . . a . q . . . . . 44
. . . . . . . . . . . . . . . . . B . Interim BLparlmental Fence 45
1 Radio Trsnemlttsr . . . . . . . . . . . . . . . . . . . 50 . 2 DOPLOC Receiving Station Interkp .
Red10 Frequency Circuits . . . . . . . . . . . . . . . 57
3 . Automatic Search and tock-&I (A.L. 0 . ) . . . . . . . . . 59
. . . . . . . . . . . . . 4 . The Phase-Lock Tracking F i l t e r 63
5 . TFming . . . . . . . . . . . . . . . . . . . a s . . . . 70
. . . . . . . . . . . . . . . . 6 . Data Recording E q u i p n t 78
. . . . . . . . . . . . . . . . . a . Digital Pr int Gut 78
. . . . . . . . . . . . . . . b . St r ip Chart Recorders 81
. . . . . . . . . . c . MagneticTapeRecording . . . a 81
. . . . . . . . . . . . . . . . . d Doppler Data Digitizer 82
. . . . . . . . . . . . . . . . . . . . . . e Dsta Formst 82
. . . . . . . . . . . . . . . . . f . Dsta hansmission 87
. . . . . . . . . . g . Data Transmission Accuracy Test 91
. . . . . . . . . . . . . . . . . . . . . h Dsta Samples 91
IV . TIIE FULL SCAIE.DOPLOC SYSTEM CONCEPT . . . . . . . . . . . . . . 104
A Introduction 104 . . . . . . . . . . . . . . . . . . . . . . . . . . . B . The ~ n t e r i m DOPLOC System . . . . . . . . . . . . . . . . . . 104
TABLE OF COIWERIB ( ~ o n t . 'd)
Page . . . . . . . . . . . . . . . . . . . . . . . C Design Objectives ; 105
. . . . . . . . . . . . . . . . . . . . . . . 1 Detection Range 105
. . . . . . . . . . . . . . . . . . . . . . . . 2 -Target Size 105
. . . . . . . . . . . . . . . . . 3 W t i p l e 0 b j e c t ~ ' a ~ a b i l i t i 105
. . . . . . . . . . . . . . . . . . .. 4 Detection Probebility 105
. . . . . . . . . . . . . . . . . . . . 5 Period Calculation8 105
. . . . . . . . . . . . . . . . . . . . . . 6 orbit . . ~nclinat ione 105
. . . . . . . . . . . . . . . . . . . . . . 7 Time of Arrival 105
. . . . . . . . . . . . . . .. 8 Accuracy, Azimuth and Elevation 105
. . . . . . . . . . . . . .9 Time for F i r s t Orbit Determination 105
. . . . . . . . . . . . . . . . . . . . . . 10 Identification 105
. . . . . . 11 Adequate Data for Correlation'with Known Orbits 106
. . . . . . . . . . . . . . . . . . . 12 Counter-~ountermeaeure 106
D Target Size . . . . . . . . . . . . . . . . . . . . . . . . 106 . E Antennae . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 . F . SignalIjynsmics . . . . . . . . . . . . . . . . . . . . . . . 108
. . . . . . . . . . . . . . . . . . . . . . . . . G AntennaGain 111
. . . . . . . . . . . . . . . . H Received Signal Power Required 113
. . . . . . . . . . . . . . . . . . . . . . . 1 External Noise 113
. . . . . . . . . . . . . . . . . . . . . . .. 2 1ntepa.l Nolee 114
. . . . . . . . . . . . . . . . . . . . . . . . . 3 Bandwidth 114
. . . . . . . . . . . . . . . . . I . Bystem Gearnatry and COM- U.5
1 . hgnsmltter Power Requirement8 . . . . . . . . . . . . . 122
, . .,.
TABLE;:^, comnm. (c.yt. la);
Page
V , ~ . S c A l w I N o ~ ~ , B y 6 ~ i P R O p o s A L ' . : . ...;;. e. .;:. 0, . .., 124.
ha.. Requests .. . for C.ontrqct,or. Prop.os@. .: ., ., . . . . . . ." . : . . . , . . ~ . . 125
. .. . . . . . . . ... , . :
, . KC. s&T&-sR~,IE~.~ . : ., .... . . . ., .: . . . . .. . , .. . . ., a. 131
A. Scanning Ag++nne., Cm.r~ge . . . . , . . , . . ., ., . : . . . .. . . . 131 B. Comic Noise q@'+terference a t &59,,39 m.c . ., . ., .. . . .. ! 139
$, circulating &mow ~iltq . . i . ., . . . . . . . . . . . . 139
1 . . The ~ g i ~ Frequency .R* O q t (pR1 ., . '., ., . ., . . . 140
2, Mgi- Data Re$q+ipg and %tp, Raea qf. Cp. . . . .. . .. 143
3. C q ' DOPLg Receivy ~ ~ c i f i c s , t i , o q e . . . ., . . . . .. . . . ., 143.
a . Receiver Bandwid.+ ;. . . . ., .. . ., + . ! . : ., .. 143
b. F F q F n c Y S q q . . . , . ., t ' : : 3 . : : * , . : . % . . 144
C . . Dynamic R q g q . . . . . . ., . . . . . . . 144 d. ' Receiver +in &d Val-e Lev+! . : . ? . . . : . 144
. ..
e . ' Timing Pulses ~ & i l l + r y Qutputs . , . . .. . . . 145
f. Automatic Noise Level Control i . . ! . . . . . . . . 145
D. Circulating Mmory F i l t e r Specifications . . . . . . . : 1. Spectrum to be Analyzed . . . . . . . . . . . . . . . . .
. . . . , .
2. Input Resis$ance . . . . . '.. .. . . ? . : . . . : . . . . 3. Effective Input Noiae V+t+ge ' . . . : : : . . . ! . : , . 4. 1 P ~ n a l W I n ~ ~ t S i g n s l * ~ 1 : ~ ! , . . : , : . ! . . . . 5. IVopld C*rf3pt +tegration T* . . , . . . : c . . . 6 . Integration start . . . . .: . . ,. . . ; . . . . : . . . . 7 . DeedTim . , . . . c : . . ' . . . . : . . : . . . ! . . c
TABU OF c m (~ont.'d)
pass
. . . . . . . . . . . . . . . . . .U near @prating Range . ~ 147
. . . . . . . . . . . . . . . Frequency Coverage Ripple 147
. . . . . . . . . . . . . . . . . . . . . . . . . . 8electivity 148
. . . . . . . . . . . . . . . . . . . . Signal Enhancement 148
. . . . . . . . . . . . . . . . . . Spurioue . Response Levels 148
S igna l Limiting AGC . . . . . . . . . . . . . . . . . . . . . . . 148
. . . . . . . . . . . . . . . . . . . . . . VideoGutputs 151
. . . . . . . . . . . . . . . . . . Display Period Star t 151
. . . . . . . . . . . . . Threshold Levels and Stabil i ty 151
. . . . . . . . . . . . . . . . . . . Output Data Pulses 152
. . . . . . . . . . . . . . . . 18 . Output Impedance Levels 152
. . . . . . . . . . . . . . . . . . . . . 19 . External Terminals 152
. . . . . . . . . . . . . . . 21 . CMF Iaformation Available 153
. . . . . . . . . . . . . . . . . . . . . . 22 . Installation 153 .
. . . . . . . . . . . . . . . . . . . 23 . Power Requirements 153 . .
. . . . . . . . . . . . E . Studies Relative t o DOPLOC Rediver 153
1 . ERL Report No . 1093. he Dymdc Characteristics of . . . . . . . . . . . . . . . . . . . ' Phase-Uck Receivers" 153
2 . BRL Technical Note No . 1345. "The DOPLoC Receiver for . . . . . . . Use with the Circulating Mamory ~ i l t e r " 15.5
3 . ERL Technical Note No . 1 9 5 . "Parametric Pre-Amplifier . . . . . . . . . . . . . . . . . . . . . . . ~ e e u l t s " 155
J. Comb Filter bnlaplaant . . . . . . . . . . . . . . . . . 163
VII. . SATELLIZPE w m , -TION FROM DOPLOC DATA :, . . . . . . . 174
B. Deecription of (hmrmd Data . . . . . . . . . . . . . . . 175
D. Init ial Appraxfaationa for the Sin@e.&or solution . . . 189
E. w t i p l e Paea solution . . . . . . . . . , . . . . . . . . 198
0. Conclueion . . ... . . . . . . . . . . . . . . . . . . . . 208
VIII. POTENTIAL OF DOPLOC S Y S m IN ANTI-SATELUTE DEFENSE . . . . . . . A C r n O W L E ~ r n . . . . . . . . . . . . . . . . , . . . . . . . . . ; . . 212
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
Fi@xa 1. Minhm Raceivsr power as a Function of Frequency fo r Constant Beamwidth Antenna
Figure 2. Relation Between Black Body Temperature, Comic Noise, Power Denalty and Frequency
Bi- 3. Receiver B i g n a l Input Power and Tranmit te r Power versus Frequency
Figure 4'. Location of DOPLOC Tracking Station8
Figure 48.. DOPLCC Tran8mitter Building . . . . . . , . . . , . . . . . .
FIF 5. DOPLCC System 50 kw Radio Transmitter and Frequency &aswing Equipment
Figunr 6. Photograph af Frequency Generating and Egaauring Equipment
Figure 7. View of Front of 50 kw Transmitter
Figure 8. V i e w of 50 kw DOPLOC Transmitter. 50 kw Cubicle i s i n Foreground
Fig- 9. Orientation of Transmitting Autennas
Figure 10. Block D i s g r r u ~ of R.F. Section i n POPLOC Stat ion Passive Tracking a t 108 mc . ,
Figure 11. Block D i a g r a m of R.F. Section f o r Active Tracking
Figure 12. Block DFagFam of Automatic Lock-On Syetem 4
Fl.@xe 13. Block Diagram of Phaee-Lock Tracking F i l t e r In ters ta te Model N
Figure 14. S/N Improvement by BRL Tracking F i l t e r
Figure 15. Graph of Trsckiw F i l t e r Output Bandwidth versus Step- Bcceleratlon
Figure 16.. Phase-Lock Tracking F i l t e r
Figure 17. Block Diagram of Radio Frequency Oecillator, Type 0-471
. . . . ' . . F i m 18b. Internal . . ~ ipw . . . . , . . :of &;;bncy : S%andard ' ,
. . . . .
Figure 19. . Block Dia&&a q f ' D ~ ~ " b i & . ,. 8yn- . . , . ,.
. . . . . . . . .. , . . .,, . . . , . .
20. . : =idck. .&:,&$edi: T u f i g s y 8 d . . .. . , .;., . . , . . . . . . . . ~. .. ,
Figure 21. Block Diagram of Doppler Data Reeding 8~nt.PI fo r a Single Channel
. . ,>< ,,
Figure 22'. Block M ~ ~ ' of Singl=Chennsl DoPPier Ihtr Digitizer . " . . - .
F i d 23. :simplified N.&k DI*@ -' Systmm DataEsndling' :
Figure 24. Data Digitigidg E q u i e t . . . . . . . ,
~ig t i re 25. Punched Paper Tape 90-t ; ', , . . . .~ >
Figure 26. Teledata Transmitter Receiver Unit
Figure 27. DOPLOC Data Transmission - C~~rmunication Network
Figure 28. Block Diagram of Equ iwn t for Frsqwncy MBasurernent Tests , , ,, r . ' .
F ~ W ' 29.: ..DOPLOC Doppler Record :if ;60 Delta ,. pev. . 140;, Forrest 'city, Arkansas Station . ,. .
. .
Fi- .30, , ' . Typical DOPLOC Data Output . *
Fi- 3l.s. Dopplei. Rscord of ' ~ c t l e " $ ~ Traclri-& 19% Zeta (~iscoverar V I ) ,Rev. 1% a t APO, W. . :.. , ' ..
B i g w e 31b. Doppler Frequency Record 'oi Active N-8 of 1959 Zeta (Macoverer VI) Rev. 1 , a t AW, MY.
=(pw Za. Doppler Record of Active S-I Track of 1959 Kappa (Discoverer VII) Rev. 7, a t AFG, Mi.
Figure 32b. Doppler Frequency Record of Active 8-1 Track of 1959 Kapp (Discoverer V I I ) Rev. 7 a t A X , Md.
Figure 33. Doppler Frequency Record of BFU Ianoeghere Probe (Strongam I) Launched a t Wallops Island, Va. and Tracked a t AFG, Ma.
Figure 34. '~appler Freq&ncy and 6 Strength Records of the Launch of Explorer I. (1958 Alpha) Tracked a t AW, Ma. ..
LIST OF F I G W . ( ~ o n t . 'd)
Figure 35.
Fis~re 36.
Mgure 37.
Figure 38.
Figure 3 b .
Figure 42.
Figure 43.
Figure a. Figu-8 45.
F- 46.
Figure 47.
Figure 48.
Figure 49.
Figure 9.
Fi- 51.
Fisura 52.
Figure 53.
Scanning Beam Traneverae Coverage . .
Area Coverage i n Plane of Fan Beme
Volume Coverege of One Transmitter Receiver Pair
Tantative Tracking Station ~ o c a t i o n e
p~ Bi-s tat ic R a d a r Attenuation or - ve Range sum ( R ~ + R2)
p~
Proposed Antenna Design
Cwposite S a t e l l i t e Chart f o r One 4 X 40 Degree Scanning Antanna Base Line. Aberdeen Proving Ground - C q Blanding Florida.
Crmpoeite S a t e l l i t e Chart f o r Cne 4 X 40 Degree Scanning Antenna Base Line. Shove timt In beam and number of h i t s at scan rates of 15 and 7.5 seconde per scan.
S a t e l l i t e Pasees as a Function of Altitude f o r Selected S a t e l l i t e s
Same as Figure 41 Bxcept f o r Different Data
Similar Data t o Figure 42
: Similar Data t o Figures 4 1 and 43
An Implementation f o r the WR
Typical CMF Response t o a Whum Amplitude Sinewave (with Time Weightid)
Signal AGC Requirements
Schem8tic Block Diagram ofReceiver
Optimum Noise Level f o r M%ximum Sensitivity
Sensitivity C h a r t
Dynamic Range Chart
AGC Action Chart
Figure 54. Two Crossing Signale k i t h V8rFabI.e Sweep Rates .., , . ,
P"l- 55. 60 Dalta Rev. 172 ' ",: . . ': . . .
~ i & e 56. 60 Delta ~ e v . 165 . . . . . . . . . . . . . . . . . . . . . : . , : , . . . .
, . . . . . : . . . . . . . . , . . . , . Fi- y ;':' 9 Delta Re+ :' 124 : : '. .;
. . . . Bigur~B 58.. ~ d p p l e r ' Frequency TI& C m s
3. . . . Figure 59. Problem Ceomtzy
Figure 60. ' Coordinate Bystams . . . . . . : . .
Figure 61. G e e t r y fo r Determining the Initial Appraxiwtione
Figure 62. DOPLOC Frequency and R a t e of Chaw of Frequenoy as a Fiiuction of Position i n the YZ-Plane (for 80° inclination)
Figure 63. ARPA-BRL DOPLOC ~o$~ler:,Recofd of 58 Delta Rev. 9937, Forrest City, Arkansas . . . . . . .
. .
' ~ i g u r e 64. DOPLOC Observations for Rev. 30,124;: 140, 156, 165 and 172 of Discoverer X I . . . .? . . , . . . ..& ;.
Figure 65. Doppler Frequency ~ x t r a = k frcen ARPA-BRL DOPLOC Record fo r 1958 'Rev. . No. 9937 ; ::
Early in 1958 the AaVanoad Research Projects Agency (ARPA) took orer
the meponsibl l i ty f o r and directian of a committee established by the
Director of Gulded Missiles for the purpose of investigating the technical
pr&&m8 involved in a national s a t e l l i t e tracking network. The committee,
chaired by Captain K. M. Oentry, USN, establiehed working groups i n the
amas of detection and tracking, c ~ u u i c a t i o n s , data processing, United
Statee programs and fofeign progrspre. An ad hoc camittee, comprised of
tha chairmen of the f i v e working groups, w s s formed when ARPA requested
that immediate at tent ion and emphasis be directed toward solution of the
problem of detecting non-radiating or "dark" s a t e l l i t e s .
Following a number of meetings of the working groups, the ad hoc
~Wttee, and the Gentry codmuittee, a proposal f o r an intekim solution t o
t8a dark satellite problem was prepared and presented to ARPA i n a brief-
by .the Gentry cammitt-. Tha committee proposed the use of s a t e l l i t e
tracking techniquee developed by the Naval Research Laboratory (Minitrack)
and the Army Ordnance Ba l l i s t i c Research Laboratories (DOPLOC), each
modified t o permit operation i n the passive (non-radiating s a t e l l i t e )
mode.
It was tentat ively established that a national passive detection
network ahould be able t o detect objects having effect ive radar ref lect ion
cross sections of 0.1 square meter or greater a t a l t i tudes betwee 100
and 2,000 miles. It was further deemed desirable t o be able t o determine
the o rb i t parameters, i n a relat ively short time and preferably within
one o rb i t period, and t o establ ish h e d i a t e l y that the object detected
was i n f a c t a new object or t o identify it as a previously detected and
known s a t e l l i t e .
The br ief ing t o ARPA indicated t h a t the state-of-the-art i n the
tracking f i e l d would not result i n a system capable of fully meeting the
desired characteristics. Specifically, it was pointed out that the tech-
n ica l l imitation8 vould reduce system capabili ty t o tracking a one square
mtur tesgat t o 1,000 miles a l t i tude or less. The ef fec ts of meteors,
aircraft, etc . , on the tracking systems could not be ccmapletely predicted.
Camputationrrl programs f o r o rb i t determination had not been developed. It was estimated that a fence spanning the United States i n East-West
15
d iawt ion a t about LBtitUas 32' and composed af two Minitrack and O n e
DOPLGt! complex could be insfelled and placed i n initial. operation within
s i x months.
On 20 June 1958 the ARPA issued Order 8-58 t o the Ba l l i s t i c Research
Laboratories. The order s ta ted i n par t :
"Pursuant t o the provisions of DOD Directive dated
7 February 1958, the Secretary of Defense has approved and
you a re requested t o proceed at once on behalf of ARPA t o
es tab l i sh a Doppler system complex. Thie system i s t o be
i n accordance with EiRL l e t t e r t o ARPA of 8 May 1958, and
w i l l , in cooperation with the Minitrack s a t e l l i t e tracking
system produce the capabili ty of detecting, identifying,
and o rb i t predicting of non-radiating objects in space. I 1
Using state-of-the-art techniques and essent ial ly available equipment
BRL, with some help from the Signal Corps, USASEA, proceeded w i t h a crash
program and within s i x months, by 20 December 1958 had ins ta l led a three-
s ta t ion complex consisting of a 50-KW illuminator radio transmitter a t
Fort B i l l , Oklahoma, and receiving s tat ions a t Forrest City, Arkansas
and White Sands Missile Range, New Mexico. Three Minitrack type antennas
were used at each s tat ion. Detection w a s by ref lected signal frm the
transmitter v ia the s a t e l l i t e . The axes of the beams of the receiving
and t r a n m i t t i n g antennas were oriented t o in te rsec t i n space a t a dis-
tance of 900 miles fruu the stations. Of the three ants- a t each
s tat ion, one was directed ver t ica l ly above the great c i r c l e joining the
etatione, one was directed 20 degrees above the horizon t o the north and
the other 20 degrees above the horizon t o the south. In t h i s manner the 76 x 8 degree fan-shaped beams were designed t o detect the approach of a
aatellite as it cane over the horizon, intercept it again overhead, and
again as it receded toward the horizon. This system was placed i n
operation i n January 1959.
BRL personnel, r ed i e ing from the beginning of tha program that the
interim system would have limited range and coverage, urged ARPA t o supply
funds and to authwice a raseerch grogram designed to delralop a realistic apsclficaticm far a eyatanr capable of edequ~tely maating all of the orig-
inal performance specif ications. Amendment No. 3 t o ARPA order 8-58 provided funds t o extend the system and t o undertake an extensive research
prcgram, investigating the problem areas which were responsible fo r the
limited performance of the interim system. Other amendments established
reporting procedures and provided funding. BRL was la te r relieved of the
rasponsibility of routine tracking for the purpose of supplying data to
the f i l t e r center.
An extensive engineariag study of the system m q w t a w a 6
in i t ia ted by &, and taatative conclusioas were o u t a d to ARPA repre-
8antetives at a meeting at BRL an 14 A p r i l 1959 88 briefly described in
BRL Rehnicsl Rote Wo. 1266, "An Apprcach to tbe Doppler Dark Sate l l i te
Detection Problem", by L. G. deBey, July 1959. The idea of using several
pairs of scanned, co-plannar, fan-shaped beams t o cover the volume above
the United Statea was presented t o ARPA in the Second Semi-Annual Summary
Report for the Period 1 January t o 30 July 1959, BRL Memorandum Report No.
1220, by L. 0. &My, IV. W. Rlcbard and R. B. Pattan, under section V,
entitled, "A New System Concept". By l e t t e r dated 19 November 1959,
subject "Second Generation DOPLOC R & D Frogram" BRL submitted t o ARPA
a proposal covering a program of research and d e v e l o p n t work designed
t o validate the conclusions stated in BRL Memorandum Report No. 1220,
and t o test the perfommce of the proposed scanning beam DOPLOC system
through the installation and operation of a scaled-donn model.
This proposal was not approved. On 13 April 1960 Amendment No. 5 t o ARPA order 8-58 was received. It stated "Aa a result of technical
evaluatiws of the SHEPHERD Program, it has been determined that the
DOPLOC system w i l l not meet the immediate objectives of the s&tellite
detection and tracking program assigned t o the Advanced Research Frojects
Agency. Accardingly, further research on the DOPLOC system i a not
required". The order further requested that plans fo r closing out the
project be submitted t o ARPA. The cessation of work ceme juet
DOPLW had demonstrated its ab i l i ty t o detect, track and cmpute renab le
orbits within minutes a f t e r a single passage of a non-radiating sa te l l i t e
o m a single station. It had the highest resolution or discrimination
against meteors and other extraneous signal sources of any known elec-
tronic system of s a t e l l i t e detection. W i t h the scanned fan shape beam
t y ~ e system proposed t o ARPA, acquisition would be automatic and a
minimum number of low cost stations would insure s sneak-proof coverage.
Multiple target capability would be adequate for any forseeable density
of sa te l l i t e s . The range and sensitivity of detection would be equiva-
lent t o or better than those of any other known system. The cost of
providing volumetric coverage of the space above the United States t o
any given al t i tude would be less then, or a t most equivalent to, the
cost of a l ine type fence having coverage t o the same al t i tude. Casnputing
time required t o produce an orbit ranged f r m f ive t o 25 minutes on a
slow computer of the ORWAC t y p . . W i t h a modern high epeed computer
thase times could be reduced t o one tenth t o one hundredth of the above
values. The problem of data digit isat ion and transmission i n rea l time had been solved. The system would have a limited applicability t o anti-
s a t e l l i t e and ICBM defense.
11. EVOLUTION OF DOPLOC PASSIVE SAlELLITE TRACKING Sy8!lIEM
A. lZIs Basic DOPZDC System
1. DOPLOC Derivation. The DOPLOC (Dower w e e LOO^) system i s a
s a t % l i t e tracking system using the phase-locked Doppler frequency tracking
f i l k technique. Historically it evolved from DOVAP and PARDOP Doppler
tracking systems developed by the Ballistic Research Iaboratories. DOVAP
(Doppler Velocity And position) i s an "active" Doppler tracking system,
i.e., one which uses a transponder in the missile t o return the signal t o * ground.. PARDOP (Phsive Ranging ~ 0 ~ p l e r ) is a reflection Doppler tracking
system. DOVAP and PARDOP ut i l ize the Doppler principle for the accurate
determination ofmlasi le trajectories and other f l ight phenomena. AEI
originally conceived, DO- was t o be used in conjunction w i t h an upper
atmosphere ionosphere axparbent t o provl.de a high quality Doppler fre-
quency recerd frasa a sa t e l l i t e carrgiag an ultra-stable frequency, lowJ]pwr
radio transmitter. Single station, single pass operation was envisioned as
a desirable goal and served as a basis for design and choice of system - parameters. DOPLOC i s a reflection Doppler tracking system consisting of
one or more sa te l l i t e illuminatingradio transmitters and one or more
Doppler frequency receiving radio receivers. By utilizing extremely
narrow band f i l t e r s in the receiving system it i s possible to operate with
the extr-ly low signal levels that are reflected from 'the sa te l l i tes .
By wing proper station geometry it i s possible t o compute orbits of the
sate l l i tes from single passes over a single pair of stations, i.e..trsns-
m i t t e r and recelver . 2. Tracking Filters. The Doppler frequency phase-locked trackLng
f i l t e r i s the heart of the DOPLOC system, providing a greatly improw
signal-to-noise ra t io of the output data and therefore i n c r e w d range.
The system is designed t o take f u l l advantage of the narrow bandwidth
in signal-to-noise impromment. Redetection f l l w perPomima3 i s
realized, .taking f u l l advantage of product detection prewded by l i n e a r amplifiers. The tracking f i l t e r is a phaae-locked, audio frequency,
electronic bandpass filter that continually follows or tracks the Doppler
Dtveloped for the Ballistic Research Laboratories by the Ralgb M. Parsons company, Pasadena, California, under contract no. DA-04-li95-CWD-483.
frequency automatice;lly. Its function i n the D(mL0C recelviag s t a t ion is
t o BE&@ possible the datection and accurate frequency measurement of
a i g m l ~ which are much weakar then the noise a t the output of the receiver.
of the Doppler frequency FB a c e ~ i e h e d a u t m t i c a l l y by means of a tight, phase-locked fe& back loop, thereby enabling the uae of an
w-ly narrow filter bandwidth t o obtain a large signal-to-noise imgroYe- * m n t . A uniqw feature of t b Ballistic R e s e a r c h Laboratories1 Interstab
tracging Wtsr Fs the succaesful employment of a third order servo control
circuit. This c l r c u l t gives the filter an ef fec t ive and "aided
tracglng" type of operation which maintains good lock-on t o the ai@ under widely varying dynamfc flight condltiona. Thus the w r o w e s t pos8ible
bandwidth consistent w i t h the radial ~ ~ c e l 8 r a t i o n component of lnotion can
be uaad in curdar t o achieve a maximum signal-to-nolee h-t.
Bandwidths manually adjuetsble f ran 1.0 t o 50 cgs available i n
tb(spre88nt models. The narrowest bandwidth gives tha DOPLOC receiving
etat ion a capabili ty of dstect ingand xa%intaining lock-on t o a e i @ which is 38 db below the noiee a t the receiver output, i .6 . the noise
powm. w i t h i n the receiver output bandwidth is 6300 timee greater than the
aimel mr. This is possible w i t h the 1-cps f i l t u r bandwidth and 16 kc m i m r bsndvidth. A 16 kc bandwidth is used i n ordsr t o p e e a D o p p l e r
frepwncy a h i f t of 32 dc and leave a 2 kc eafety =gin a t each edge of the geesband.
3. Beneitivity. The DOPLOC receiver sya tm uaed vacuum tube m- amplifiers with a noise figure (m) of 2 t o 3 db at 108 mc. !tke in te rna l
phase noiee i n the receiver local osc i l la tors was low enough t o permit opemtion down t o 1 cps bandwidth (B) . The tracking filter requires a minimum of 4 db output signal-to-noise r a t i o (s/N) to maintain lock-on.
Based .an these paminetere t he overal l received p o r n sens i t iv i ty ( P ~ )
18 camputed as follows: Pr = X?l?xKTxBx9/~
= 2 x.4 x x 2.5
= 2 x watts
where K is Balt&mannls constant, and T i s temperature.
* This tracking f i l t e r was designed f o r BRL by the In ters ta te Engineering Corporation, Anaheim, California, on contract DA-04-495-ORD-822.
' 20
With a steady, unmodulated signal of 2 x watts into the
receiver, the signal-to-noise ratio of the output of the tracking filter corresponds to an output phase jitter of 36' rme or 0.1 cycle. A signal
about 3 db stronger is required to lock on initially with the automatic eyetern. When signal freqwncy and amplitude dynamics require a larger
bandwidth, the threshold received pwer level to maintain lock-on
increases linearly with the bandwiath used. The exact bandwidth required
to maintain lock-on is a function of the received signal strength, the
rate-of-change of frequency and the type of modulation that may be &lib-
erately or indirectly imposed on the signal. Ballistic Research Labora-
tories Menorandun Report No. 1173, "DOPLOC Tracking Filter" by Victor W.
. Richard, contains a discussion of the bandwidth selection problem and gives the basic equations and graphs of the relation between bandwidth,
rate-of-change of frequency and signal-to-noise ratio.
4. Antennas. With the extreme sensitivities realized by the
optimum use of very narrow bandwidths and the low noise receivers it is
generally possible for tracking active satellites to use antennas that
have sufficiently broad radiation patterns, that the antennas need be
oriented only in the general direction of arrival of the expected signal.
The direction in which the antennas are pointed need not be known
accurately. Accurately known and controlled radiation phase patterns
are not neceseary. The only requirement on the antennas is that they
deliver sufficient signal, preferably as nearly constant as possible,
irrespective of the state of polarization of the incoming signal. As
further discussion will show, the success of the passive DOPLOC system
is dependent upon the most effective utilization of the best antennas
known to the state-of-the-art. In addition to the three Minitrack
type receiving antennas available at each receiving station, flat top,
corner reflector and dipole antennas were also available.
2. Data Output. The basic data output of the receiving station is
Doppler frequency correlated with time. Since very stable beat frequency
sources can be used, the Doppler frequency record can be interpreted to
be a very accurate record of received carrier frequency. Tha Doppler
frequency output of ths traEking filter is s u f f i c i m t Q Clean, if
filw is locked on a t a l l , t o permit d i rec t e lectronic ~0untiXX and
reed-out or print-out i n any desired form such as arabic numerals,
punched tape in binary code, IBM cards, wide paper analog frequency,
m8gnetic tape, e t c . These data can be used with appropriate cmpu-
t s t i onb l program t o obtain the o rb i t a l parameters desired. The 1
accuracy of the solution i s determined largely by s ta t ion geometry
with respect t o the s a t e l l i t e o rb i t and not by electronic e q u i p n t
l imitat ions and er rors . Propagation e f fec ts must be taken i n t o account
f o r maximum accuracy. When locked, the tracking filter output i s within
90' or 1/4 cycle of the input, thus i n d i c ~ t i n g a. capa t i l i t y of providing
a Doppler frequency accuracy of b e t t e r than one cycle. An accuracy of
one pa r t i n ten thousand is obtained with multistation missile tracking
Doppler systems such a s DOVAP. Because of the lack of any r ea l ly
dependable system t o ca l ibra te the accuracy of the DOPLOC system, the
o r b i t determination e r ro r i s not known with a high order of precision.
However, based on the agreement between the orb i t a l parameters determined
by DOPLOC data and Space Track data, the DOPLOC data appear t o be as good
a s the Space Track data.
A prime feature of the DOPLOC system is the t ranslat ion o r hetero- 7
dyning of the input radio frequency s ignal down t o a frequency spectrum
low enough i n frequency t o be recorded on magnetic tape. This trans-
l a t i o n i s done without degradation of the signal-to-noise r a t i o o r l o s s
of intel l igence from the or iginal s ignal out of the antenna. Thus the
Doppler frequency from any pass can be stored and played back through
the tracking filter and en t i r e data processing system i n the event t h a t
any of the equipment items f a i l e d during r e a l time.
A discussion followe of the problems of modifying the basic DOPLOC
system Just described t o provide a system with the anticipated capabi l i ty
of detecting, identifying and o r b i t predicting of nan-radiating objects
i n space.
A paasivi tracking eyattm w i t h the capabiSity specified by AWA
mqulres the very beet of a+lable technique6 and demands components of
the hlgtiest performance attainable within econcaaic feasibil i ty. The
pftrfonaance characteristics and parmeter values proposed for the DOPLOC
p a s i r e s a t e l l i t e tracking system, baaed on the state-of-the-art w i l l be
treated in this section.
The baeic relation between ran& and the other pareaeters of a
reflection system can be expressed aa r
w h e r e : R = -> Pt = tmusuiitter pouer, Pr = receiving p o w , Gt and
a,, are transmitter and receiving antenna power gain, A = w a v e Length and
a = efYective reflection ecias section.
It ie of interest to salve for the range of a system using all high
performance, state-of-the-art ccmponents for which known techniques of
fabrication e x b t even though they may not be econdca l ly feasible or
attractive a t this time. Opt& parameters w i l l be used in order t o
obtain a theoretical msxinrum perfonaance f r a which perfonaance in practice
can be predicted by applyingexpected degradation factors. Under these
conditions a transmitter of one megawatt i s possible; transmitting and
receiving antenna gains of 20 db each are reasonable since they are
associated with moderately wide coverage, shaped radiation patterns of
about 6' x 60'. A threshold received power requirement of 2 x watts
w i l l be assumed, although seldm possible t o attain in practice because of
ex-al noise ~OUPOBB, vier bandwidth requiremate and Fnit ial lock-on
requirements. NaxUnum range w i l l be computed with three target sizes,
0.1, 1.0 and 10 6quare meters, covering most of the exgected target siees.
For a ~'0.1 sq. meter
= 2330 miles
For a = 1.00 sq. meter
R = 'fl x 2330 = 4140 miles
Bor o = 10 sq. aeters
R = 4fi x 4140
=. 7370 miles.
As previously stated these ranges are the maximum ones theoretically
possible under optimum conditions. In practice sevaral econamic and
technical. factors enter to modify the parameter valuas attainable. A
diecussion follows on the nature and megnitude of some of these factors
as they apply to the DOPLOC Passive Rgcking System.
1. Transmitter Power - Pt. The use of one-megawatt transmitters at
108 to 150 mc at this time appears to be well within the state-of-the-art and limited only by economic considerations. However, such transmitters
have to be specially built and require up to 24 months lead time at a
cost of at least $1,000,000 each, for the basic transmitter without the
antennas, shelter, installation and power facilities. Since the fence
was to be installed on a six-month crash program it was necessary to
procure an existing transmitter. The largest available unit located was
a 50-KW f.m. broadcast transmitter. Since the reduction in power from
one megawatt to 50 KFI reduces the range by only a factor of 2.1, the
50-KW transmitter was considered to be adequate to demonstrate the feasi-
bility of the DOPLOC system in an interim installation.
2. Antenna Gain - Gt and Or. In the original planning it appeared
t o be technically and economically feasible t o utilizeaatennas with about
20 db gain and 6' x 60' fan-shaped besms. The closest thing immediately
avai lable was the Minitrack type antenna. It produced approximately the
deeired beam shape and was rated by the manufacturer at 18-db gain. The
reduction f r o m 20 t o 18-db gain reduced the range by a fac tor of 1.25.
Unfortunately, i n pract ice in the f i e l d , the realieed antenna gain was
s t i l l somewhat l e s s , apparently'nearer 16 db. A f a c t o r which must be
considered i s t h a t the rated gain of the antennas could not be realized
i n practice due t o pohr i za t ion ro ta t ion and misalignment'between the
antennas and the ta rge t . An average lo s s of 4.3 db has been estimated
f o r t h i s factor , reducing the range by 1.27. It i s a l so impossible t o
or ien t the transmitting and receiving antennas and t o shape the i r patterns
t o overlap a t optimum gain over a la rge volume of space. This condition
i s caused by the geometry involved, which i s due i n pa r t t o the curvature
of the ear th . The antenna or ientat ion i s t reated in d e t a i l i n BRL Report
No. 1055 October 1958, en t i t l ed "Doppler Signals and Antenna Orientation
f o r a DOPLOC System" by L. P. Bolgiano. Three geographic locations along
a great c i r c l e were chosen. Fort S i l l , Oklahcrma, was chosen f o r the
transmitter, and White Sands Missile Range and Forrest City, Arkansas,
f o r the receiving s ta t ions. The distance between Fort S i l l and WSMR is
480 miles and frm Fort S i l l t o Forrest City, 433 miles. The site
locations are a s follows:
WSMR FORT SILT., FORREST C I T Y
Latitude 33'42 9 1 8 " 34O39'20" 35°00t54"
Longitude 106'44 16" 98'24 920" 90'45 I 50"
Great Circle Bearing 79°51'16" 84'32 129" 89'01 '23"
Three antennas were located at each s ta t ion . They are designated a s
the center, north and south antennas. The center antenna a t Fort S i l l
had Its ground plane horizontal and i t s long ax is i n an approximate north-
south direction. This placed the broad dimension of the antenna beam i n
an approximate east-west direction. The north and south antennas were
25
oriented t o d i r e c t the fan-shaped beams 20' above We horizon. A t the
z?eceiving s ta t ions, the three antennae were directed t o have the axee of ths bums in tersec t a t a distance of 900 miles inmn the s ta t ions . The
receiving antennae were therefore t i l t e d toward the transmitting anteXUIW.
The above mentioned report (by ~ o l g i a n o ) contains a detai led computation
of the antenna orientation angles and reduces the ins ta l la t ion ine t rwt ions
t o a s e r i e s of rotetione about eas i ly defined axes. The types of "9" curves
t o be expected from s a t e l l i t e s flying various courses re la t ive t o the detec-
t i on system are plotted. Doppler s h i f t and rate-of-change of Doppler s h i f t
a r e computed and graphed f o r 108 megacycles per second signal ref lected
from a passive s a t e l l i t e . Orientation angles t o converge the beam patterns
of a transmitting and receiving antenna, each with a 76' x 8' beam, a r e
computed and l i s t e d i n tabular form. Computational procedures are described
i n d e t a i l t o f a c i l i t a t e fur ther investigation. It i s shown that a one
squaw meter cross section s a t e l l i t e within the antenna bem patterns w i l l
be detectable t o a range of about 1,000 miles. In f i e l d practice t h i s
Value of 1,000 miles range was found t o be too optimistic fo r emall s ize
s a t e l l i t e s . The ac tus l achieved range was nearer 500 miles. Bolgiano
a l so shows t h a t tracking f i l t e r bandwidths greater than 10 cps are required
only f o r the c loses t approach of very close s a t e l l i t e s
3. Operating Frequency. The choice of an optimum frequency fo r a
DOPLW passive s a t e l l i t e tracking system required considerable study. The
frequency of 108 mc w a s available and had been chosen f o r the ear ly U. S.
s a t e l l i t e work. Equipment was readily available in t h i s frequency range.
Thus it appeared expedient t o s e t up the interim fence on a frequency of
108 mc. However t h i s was not believed t o be an optimum frequency or
r ea l ly even a good choice. The cosmic noise a t 108 mc had been measured
t o be large en- a t cer ta in times t o reduce the receiver system sensit iv-
i t y by a s much a s 6 db, i . e . by a factor of 4 in power of 1.4 i n range.
The average cosmic noise leve l has been measured t o be about 2 t o 3 db
above receiver noise with some periods of very l i t t l e or no cosmic noise.
A t about the time of the i n i t i a t i o n of work on the fence, the opinion of
many radio engineers appeared t o favor a much higher frequency than 108 mc.
This opinion waa 'based on the f a c t that as the frequency of operation is
increased the external noise, such as cosmic noise decreases. A more
intensive investigation of t h i s subject showed that, of the several fac tors
having parameters which vary with frequency i n the 100 t o 1,000 mc range,
the 8 d w t a g e s and disadvantages a s one increases frequency almost exactly
balance each other out t o such an extent that it was found t h a t the bes t
frequency range for a re f lec t ion DOPLW system was between 100 and 150 ~ c / s .
A s applied t o the s a t e l l i t e fence program, the optimization of the frequency
problem is one of choosing a frequency which w i l l give a maximum probabili ty,
within technical and economic 1Mtaticols of detecting non-radiating satel-
lites. The maximum amount of transmitter power available f o r radiation is
assumed t o be fixed. The principal fac tors which a f f ec t the selection of
the optimum frequency include the following:
1. R a d a r equation.
2. Antenna beam width requirements.
3. Cosmic and galact ic noises.
4. Bandwidth
5. Minimum detectable signal.
The way in which these fac tors a f f ec t the choice of frequency t o be
used is discussed b r i e f l y in the following paragraphs.
a. Radar Equation. The radar equation f o r the ref lect ion type Doppler
signal i s the combination of two equations, each of which presents the
inverse-sqimre power lo s s i n one leg of the path from transmitter t o sa te l -
l i t e t o receiver. It may be writ ten i n the form:
where Pr i s the receiver power, K is an a rb i t ra ry constant, Gt i s the gain
of' transmitting anterina, Pt i s the transmitted power, As i s the effect ive
sca t t e r a rea of the satellite, Ar i s the effect ive area of the receiving
antenna, and rl and r2 a re the distances fmm the satellite t o the trans-
m i t t e r and t o the receiver, respectively. 27
The value of As i s lFmited in its maximum value by the physical
projected area of the sa te l l i te . Consequently the rrrtio of As t o rl 2
ha8 a limiting value s e t by the physical size of the sa te l l i t e and by
the range. Sh i l a r ly , on the receiving path, the eolid angle within
which energy i s received from tha sa t e l l i t e is determined by the ra t io 2
oi tbe effective antenna area t o the square of the elant range, r2 . The larger the effective area of the receiving antenna, therefore, the
larger i s the magnitude of the signal received a t the receiving s i t e
for eny given magnitude of transmitted power and mtenna gain a t the transmitter and receiver. The effective capture area of a receiving
antenna, of fixed gain, decreases as the frequency increases. Therefore,
aa the operating frequency is increased, the received power w i l l decrease, based on capture area considerations.
b. Antenna Beemwidth. The beam angular dimnsione for both the
trcmclmitting and the receiving antennae for a sa te l l i t e detection system euch as DOPLOC must be established on the baeie of the orbital paremetere
of the sa t e l l i t e and the minimum data requirements for an orbital solu-
tion. The eolid angle within which the system i e to receive energy from
a s a t e l l i t e must intereect the eolid angle through which energy frcrm the tranemitter i s radiated, and it must be of such dimneione that the probability of recognition of a sa te l l i t e pseeege i s high. The numerical e ~ c i f i c a t i o n e of a-r dimsneione explicitly determine the gain snd the dimensions of tha antanna in wava lengthe. To obtain a maximum
capture area from euch an antenna, it 10 wain deeirable t o uee tha l m e t poeeible frequency. Otherwise, maximum energy c a p t w i s not achieved, and the objective of maximum range i e defeatma. Increasing the antenna area in square wave lengthe and using higher frequencies (keeping the
physical area approximately consfant) produces narrower beem widths than
those required t o permit the to ta l required mount of data, with the
resul t that the received signal w i l l not have eufficient duration. Beam- width considerations therefore also favor low frequencies.
c. Cosmic NcrLse. Cosmic and galact ic noise i s made up of electro-
magnetic disturbances o r signals having character is t ics similar t o white
noise. When, however, these signal6 are examined over a wide frequency
range they lose some of the similar i ty t o white noise. Consequently, the
equivalent black body temperature of the radiation varies with the fre- quency of measurement, decreasing from between 2,000 t o 20,000 degrees
Kelvin a t 64 mc/s, depending on the sky area, t o l e s s than 100 degrees
a t 910 mc/s. (True white noise character is t ics would r e su l t i n a constant
temperature a s a function of frequency.) The effect ive sky temperature 1" has been reported by W o n to change at different r a t e s with respect t o
frequency, depending on the area of the sky which i s being observed. The
frequency and the temperature have an inverse exponential re lat ion, the
exact value of the exponent depending on whether the measurement i s made
i n the direction of the galactic plane or not. The equation applying
when looking toward the center of the galaxy is given as:
where Tw Is the equivalent black body temperature, K i s the constant of 8
proportionality f o r the galaxy, and f is the frequency i n mc. A similar
equation i s given f o r outer cosmic areas:
where Kc is the constant of proportionality of cosmic space. A single 2 exponent of -2.3 is given by Cottony of the National Bureau of Standards
which gives an overall average relationship of
Superficially, the rapid reduction of Tw with frequency would make it
appear that the higher the frequency used, the be t t e r the results. This
i s not necessarily the case, since it has been shown that the desired
s i g n a l w i l l a lso decrease as the frequency increases. Since the received
signal f a l l s off as the inverse square of the frequency, the new increase . .
i n signal-to-noise ratio varies exponentially w i t h frequency between -6.1
9 Superscripts r e fe r t o references at the end of the report.
29
and -0.7 based on the cosmic noise consideration alone. There i s a lso
f i l t e r bandwidth e f f e c t t o be considered which requires an increase i n
bandwidth by the squere m o t of the frequency increase which tends t o
b a I w e out the cosmic noise fal l -off , depending upon which exponent is used. Figure 1 shows the variation of minimum detectable s a t e l l i t e
Signal (SIN 0 1 ) as a function of frequency, aasuming s value o f 100 mc
ss a reference, both f o r areas of the sky toward the galact ic center
(tovard sagl t ta r ius) and toward the pole of galaxy (extragalectic). The
minimum leve l i s shown f o r two cases, constent signal bandwidth and
bandwidth proportional t o the ca r r i e r frequency. ".. The minimum signal requirement i n the presence of cosmic noise i s
derived by converting the cosmic noise level , which i s given i n equiva- >.
l e n t sky brightness or black body temperature i n degrees Kelvin, t o
power density i n watts per square meter. This conversion, based on the
Rayleigh-Jeans law, i s a s follows:
where p is i n watts per square meter, K i s the Boltzman constant, f is
i n cps, B i s bandwidth i n cps, Tw i s the black body temperature i n
degrees Kelvin, and C i s the velocity of propagation. Black body radiation
is randomly polarized, with equal power i n each plane, therefore a l inear ly
polarized receiving antenna w i l l intercept only one half of the power given
by the above expression. It i s s ignif icant t o note with regard t o the
above equation t h a t the noise power density i s proportional t o the square
of the frequency. The net e f f ec t of increasing the operating frequency
i s the combination of t h i s factor with the fa l l -of f of cosmic noise w i t h
frequency as follows:
These relationships between black body temperature, power density and
frequency are shown i n Figure 2 using measured values a t 108 mc as
COSMIC NOISE VS. FREQUENCY
FIGURE 2 . 32
r,eference points. These p lo t s show t h a t there is a qu i t e small decrease
i n cosmic noise power density with frequency. A ten-to-one increase i n
frequency reduces the cosmic noise power density by only a fac tor or two.
In summary, the exact manner i n which cosmic noise var ies a t t he
output of the antenna or the receiver a f t e r passing through the bandwidth
l imi t ing c i r cu i t s , i s strongly dependent upon the basic type of t racking
system under consideration. The interim DOPLOC system required a fixed
antenna beam width and increasing bandwidthwith frequency, therefore t h e
optimum frequency based on cosmic noise consideration i s a low frequency.
Figure 3 s u m r i z e s the e f fec t of cosmic noise with a p l o t of the r e l a t i o n
between receiver s ignal power required, t ransmit ter power, and ca r r i e r
frequency. m e saving i n t ransmit ter power by ope?ating a t low frequency .
i s shown i n t h i s graph. The major opposing fac to r not shown on the graph
is the increasing height of the antenna s t ruc ture a s the frequency i s
decreased. This can be a c r i t i c a l fac tor since the cost of unusually high
s t ruc tures increases roughly a s the cube of the height. k d. Bandwidth. The information bandwidth required f o r tracking the
Doppler frequency signal ref lected from a s a t e l l i t e i s proportional t o
the square root of the c a r r i e r frequency. Since the signal-to-noise r a t i o
i s the c r i t i c a l fac tor i n s a t e l l i t e detection, both a minimum bandwidth
and minimum operating frequency consistent with 'other conditions must be
accepted f o r optimum resu l t s . Fundamentally, the f i l t e r s f o r a l l t racking
systems must have su f f i c i en t bandwidth t o encompass the random phase noise
and higher derivat ive s ignal components. For t h i s reason, the lowest
frequency which has negligible ionospheric noise should be used. 2.
e. Minimum Detectable Signal. The introduction of narrow band
tracking f i l t e r s and of new types of low noise receivers has stimulated an
in t e res t i n determining what physical parameters r e a l l y l i m i t the detect ion
of a coherent s ignal . Qualitative arguments had been advanced suggesting
t h a t quantum l imitat ions might e x i s t a t power levels within the capabil-
i t i e s of modern radio receivers. Accordingly it was considered of i n t e r e s t
t o invest igate whether quantum mechanical considerations s e t l imi ta t ions t o
33
RECEIVER SIGNAL INPUT POWER AND TRANSMITTER POWER VS. FREQUENCY
FIGURE 3 3 4'
achievable sensitivities and to evaluate their engineering importance. A
emall contract designed to support an investigation of the probable quantum
limitation of detectable signals was awarded to the Electrical Engineering
Department of the University of Delaware, Newark, Delaware, with Dr. L. P. Bolgiano, Jr. and Prof. W. M. Gottschalk as the principal investigators.
This work was reported in "Quantum Mchanical Analysis of Radio Frequency
Radiation" by L. P. Balgiano, Jr. and W. M. Gottschalk, Report No. U in the BRL DOPLW Satellite Fence Series, June 15, 1960, Electrical Engineering Deparbwnt, University of Delaware, Newark, Delaware. In this report an
intuitive account of thermal radiation, making complementary use of classical
wave and particle descriptions, is used to give a qualitative introduction
to the nature and magnitude of the effect which might be expected. Quantum
mechanical computations of the phenomena obtained in specific detectors and
preamplifiers are reviewed and found to support the obsenrability of
quantum fluctuations. It is found that the random signals observed in wide
band radiometers are sufficiently weak to make these quantum fluctuations
of comparable magnitude to classical wave fluctuations if receiver noise
is ignored. It is also shown that the specific power level at which quantum
levels becaane important may depend strongly on the information to be
extracted from the signal. For the purpose of this discussion relative to
frequency selection for DOPLW, it should suffice to say that if the quantum
mechanical limitation should prove to be a limiting factor, it can be
expected that the minimum detectable signal power will be proportional to
the square of the frequency, and will thus mitigate in favor of a low DOPLOC
carrier frequency.
f. Conclusions Relative to Choice of Frequency. The considerations
presented above indicate that the optimum frequency for the DOPLOC system
should be as low as is consistent with the bandwidth required to accammodate
ionospheric noise. Experience with the analysis of DOVAP (Doppler) and
satellite tracking records at BRL indicates that, became of the ionospheric
backgmund noise modulation on the signal, the probable minimum bandwidth
setting on a tracking filter operated at 76 mc is approximately 5 cps, where aa for nipale at 108 mc it is approximately 2.5 cps. Since the maxFmum
35
bandwidth, from consideration of missile signal dynamics, which might be
required to lock on to the satellite at an altitude of 500 miles is 4.5 Cps, and at 2500 miles, 1.4 cps, and after lock on, a m one or two cycles would be required to track, it is evident that operation at 76 mc shows the presence of an excessive amount of noise, whereas operation at 108 mc is
perhaps an excellent compromise. Actually, as indicated above, a reason-
able broad band of frequencies may be considered essentially optimum since
the loss of sensitivity with frequency is shown by the Figure 3, page 34, to vary at a rate of foa3 to f0'7 relation. On the basis of the data shown,
the range within which the frequency should be selected extends from 100 to
150 mc with 175 mc as the highest permissible value. Thus 108 mc was quite
suitable for the interim fence system. For the ultimate DOPMC system
which was proposed but never instrumented, a frequency of 150 mc was chosen.
Receiver Power Sensitivity. 4. An effort was made to design the receiver circuits to provide a signal sensitivity of 2 x 10'~' watts when
used with a 1 cps bandwidth. Under controlled conditions and with the
antenna pointed in low noise directions, this sensitivity was actually
obtained. However it is not practical to attain such high sensitivity and
nesrar bandwidth in tracking earth satellites at 100 to 1,000 miles altitude
and at 108 mc frequency. As previously noted, cosmic noise parer will on
tho average double the required power for reception. Also passive satel-
lites at 100 - 1,000 miles altitude require more bandwidth than 1 cps to acccmmodate only the dynamics of the problem. The highest sensitivity is
needed when searching at low elevation angles to detect satellites as they
emerge over the horizon. This ie compatible with the information bandwidth
required since the rate-of-change of frequency is then smallest, permitting
the use of the narrowest bandwidth. Wider bandwidths are required to
maintain lock-on as the satellite passes by the station, particularly at
the lower altitudes. This again is compatible since the distances are
shortest and sensitivity requirements least at closest approach.
Any automatic iock-on system for use with the tracking filter requires
cisignal about three db greaterfor capture or lock-on than is required to
hold lock and track once the signal is locked on. This is because of the 36
f a s t ra tes a t which the lock-on device must sweep in frequency i n order t o
cover the spectrwn t o be searched. Thus a t o t a l of approximately 6 db
increase i n signal may be required i n practice above that assumed i n previous
range calculations. This can cause a further range reduction of apprax-
imately 1 .4 . Zn summary, the received power required for locking onto a
signal, under different conaitions of bandwidth md cosmic noise is tabulated
as follows:
Received 8ignal Power Required Bandwidth, cps Cosmic Noise Level
watts (-190 dbw) 2.5 Negligible
2 x lo-19 watts (-187 dbw) 2.5 3 db above receiver noise
lom1' watts (-180 dbw) 25 Negligible
2 x watts (-177 dbw) 25 3 db above receiver noise.
5. Sa te l l i t e Reflection Cross Section - a . Since the DOPLW system
had been i n i t i a l l y designed for use with radiating (active) sa te l l i t es , there
was real ly no problem in obtaining the required signal strength from very
small missile borne transmitters over the alt i tude range' and from horizon t o
horizon. To adapt DOPLOC t o the small signals which would be reflected from
small bodies of 0.1 t o 10 square meters effective reflection cross section
required a careful study of signal levels t o be expected from practical
transmitter powers and realizable sa te l l i t es . Body configurations, dimen-
sions, ,spin, tumble, and extent of cmouflage a l l affect the reflection
cross section of a body. The t e r m "effective" cross section i s used t o
mean the value of a i n the radar equation that muet be used t o determine
the actual range of propagation with the particular s a t e l l i t e target under
coisideration. The effective reflection cross-section area i s strongly
affected by the alignment of the target ' s reflection amplitude pattern and
polarization with respect t o the illuminating and receiving station antennas.
By using the equations available i n the l i terature for the monostatic case,
i.e., transmitter and receiver coincident, the effective dLmensions of
simple spheres and cylinders were calculated. The values thus obtained
were for peak reflection cross section with optimum orientation of the
target on the maximum of the reflection pattern lobe. Calculated values
for spheres and cylinders are as follows:
Max. Cross Section Sphere Dia. Cylinder Length
0.1 sq meters 1.5 fee t 2 1/4 f ee t
1.0 4.2 3 112 f e e t
10.0 11.7 14 t o 50 fee t .
The areas for the cylinders were computed for a length-to-diemeter r a t i o
of 15. A range of 14 t o 50 f ee t was estimated for a cross section of 10
aquare meters because of the effects of resonance. If the effective
length of the target happens t o be an exact multiple of a half wave length,
it w i l l have a much larger effective area than i f it i s some random length.
For the b i s t a t i c case, where the transmitter and receiver are widely sepa-
rated compared with the range of tracking, the angle between the incident
and the reflected rays further modifies the effective cross section area.
Parer and effective croso section area calculations s r e treated i n more
de ta i l i n BRL Report No. 1330, ent i t led "DOPLOC Observations of Reflection
Cross Sections of Satell i tes" by H. Lootens. Using the DOPLOC data and
the equation nde2 and the basic radar equation, effective cross = - A
section areas were computed for a number of s a t e l l i t e s that were tracked.
The computations are based on actusl, measured data. To give an idea of
the magnitude and ranges of effective cross section area, the data below
are taken from Lootens1 report. In the original data effective cross
section areas are camputed for the center, north, and south antennas when
sufficient data &re available. Slnce the principal in te res t here i s t o
show the extent of the variation i n reflection cross section, data from
the center antenna only are given.
TABU I1 - 1 Effective Cross Section Area for Selected Satell ites and Revolution Numbera
SATEUITE Revolution Ao. (cross Section Area (sq. ft.)) Meaeured in Center Antenna
'58 Delta. . ,
SATELLITE
' 59 zeta
Revolution No. (Croas Bection Area (sq. f t . ) ) Meaaured i n Center Antenna
TABLE 11 - 1 (Cont. 'd)
SAT!EUI'l!E Revolution No. (cross Section Area (sq. ft.)) Measured i n Center Antenna
'60 Epsilon 1 165
386
522
'60 Epsilon 2
6. Reflections from other Objects. It was in i t i a l l y expected that
both airplanes and meteors would produce spurious reflections which might
be troublesame t o eliminate from the DOPLCC data. Since it was known that
Stanford Research Inst i tute personnel had made an extended study of reflec-
tions frcrm meteors and meteor t r a i l s , advantage was taken of an existing
contract with SRI t o have th i s problem studied. The effects of meteors and
meteor t r a i l s are discussed i n de ta i l in a report "DOPLCC System Studies"
by W. E . Scharfman, H. Rathman, H. Guthart, and T. .Morita (Final Report,
Part B, Report No. 10 i n the BRL DOPLOC Satel l i te Fence Series), Stanford
Research Insti tute, e n l o Pmk, California.
In practice, meteors enter the atmosphere with velocities between 10
and 75 kilometers per second. Although some meteors are decelerated
noticeably i n the atmosphere, most of them evaporate and deposit their
ionization along the i r path a t a nearly constant velocity. While a meteor
i s coming in, an echo i s received from the elemental scattering lengths of
ionization formed along the straight-l ine path of the t r a i l . This "Doppler
echo" has the range and velocity of the meteor a t the start of the ioni-
zation t r a i l , and shows a considerable Doppler sh i f t which depends on i t s
velocity and on i t s orientation with respect t o the transmitter and :
receiver. The'echo from the meteor i t s e l f i s normally toosmall t o be
directly detected, so that the detected signal i s due t o the reflection
from a trail which i s increasing i n length with time. The mplitude of
the echo r i s e s t o a large value a t the point ( i f any) where the angle formed by e ray from the transmitter and the normal t o the trail axis is
equal t o the angle formed by a ray from the receiver and the normal, t o
the t r a i l axis. After the meteor has passed th i s point, referred t o as
the apecular point, the received signal i s composed of a Doppler echo
superimposed on a continuing epecular echo from the body of the t r a i l near
tills point. The "body echo" decays while remaining a t a more or l e s s fixed
range, and shows only a s l ight Doppler s h u t caused by ionospheric winds.
Thus the head echoa ware at very high Dopglsr frequenciee an8 wsre unable
t o pass tha tracking f i l t e r , w h l h the echo6 fian the trails showed 1ittI.a or no Dopplar elope. Tham coub¶ easily be i&ntrtiied se trail echo0 and
presant8d no data d u e t i o n probhm.
During the en t i r e obmrv~ t i an m o d apprcuhately 10 so call&
unidentified bodies were indicated i n tha DOPLOC data. These data had
the proper Doppler frequency change for sa te l l i t es , but did not correspond
t o any known orbi ts . They may have been caused by meteors or could have
resulted frm equipment, malfunctions, or possibly even unidentified bodies.
No data are available t o indicate the seriousness of the problem of diecrim-
inating against meteor signals when the scanned beam antennas are used as
proposed for the full scale system. However it appears that t h i s problem
can be easi ly solved by limiting the range of Doppler frequency which can
be locked i n on the tracking or circulating memory f i l t e r
7. Solutions for Orbit Parameters. When the interim DOPLOC fence was
f i r s t conceived, the only problem which had not already been essentially
e i ther solved or for the solutlon of which known engineering techniques
were available, was tha t of determining the orbi ta l parameters from DOPLCC
data. The three or four s ta t ion DOVAP problem had been solved with
consistency and accuracy for years. But i n the DOVAP problem the missile
carried a transponder and a minimum of three stations with reasonably good
s ta t ion location geometry were available. The goal s e t for the DOPLCC
42 .I' L
system was t o determine an o rb i t with suff ic ient accuracy t o e i the r
ident i fy a satellite as a known one or a s a new one from data obtained
from a single pass over a single receiving s t a t ion and i n a time a f t e r
detection of l e s s than one pass or o r b i t period. The f i r s t e f fo r t s
seemed t o be ra ther hopeless, but by the end of the second six-month
reporting period a method was suff ic ient ly well developed t o be included
i n the second technical summary report. More recently, an extensive
e f fo r t , both in house and by contract, has resulted i n the development of
sa t i s fac tory methoda of solution. These methods w i l l be t rea ted i n smne
detail l a t e r i n t h i s report.
111. DOPLOC EXPERDENTAL FACILI'PY
A. Background Experience and Developments
Since 1945, when BRL first established th? DOVAP instrumentation at
White Sanaa Missile Range, New Mexico, the Ballistic Measununent Laborrrtcuy
had been actively working t m d extending the r u e and dependability of
trajectory data gathering equipnt. Through a contract with C o n ~ i r the
DORAN equipnt was developed in an effort to remove the ambiguity from the
Doppler data. By contract with Ralph M. Parsons Company a reflectam Doppler
syetem was developed for measuring the velocity of missiles, specifically
artillery shell, at short ranges. This was a strictly passive system and
had inherently many of the problems that were later to be encountered in
DOPLOC. Another trajectory data gathering system which was developed and
demonstrated, but which has never been used extensively, was called
SFTSfRZDOP. In addition to the system develo1;rment work, BRL had devoted
considerable effort to the development of swcial narrow band tracking fil- .
tars, high gain rsdio receivers and stable oscillatore or constant frequency
eourcee. The problem of receiving low energy signals in the presence of
high noise levels had been studied extensively. Such problems as attempting
to eliminate the effects of spin and attitude change from the DOVAP data had
led to special techniques in signal reception.
Within a few hours after the Russian6 announced the launching of the
first Sputnik satellite, BRL had established a tracking station and wae supplying data to other interested agencies. These data were of course Only
active data from the satellite's own radio transmitter. However narrow band
techniques and many of the features later built' into the DOPLOC system were
employed. Thus, when ARPA asked for a satellite tracking facility, BRL had
already developed a considerable competence in the detection of low energy
signale in the presence of noise.
In b y 1958 BRL proposed to establish, in co-operation with the Naval
Research Laboratory, a satellite detection and tracking facility. This
total facility would extend across the southern United States in an east
to west direction from coast to coast at about latitude 32' N. The Naval
44
Research Laboratory would provide sections of what was later called a fence,
extending from the east coast to about the Mississippi River, and from the
west coast eastward to near White Sands Missile Range, New Mexico. BRL
would provide the center section extending from White Sands Missile Range
to Memphis, Tennessee. This proposal was approved by ARPA and resulted in
ARPA Order 8-58 dated 20 June 1958. This order directed BRL in co-operation
with the Minitrack satellite tracking system to produce the capability of
detecting, identifying and orbit predicting of non-radiating objects in
space. The only specific direction contained in the order relative to the
nature of the facility was that BRL would establish a Doppler system complex.
The method of reporting to ARPA and an indication of the funding to be
expected were contained in the order.
BRL personnel realized that no facility or developed technique.within
the United States existed which was capable of meeting the proposed perform-
ance specifications suggested by the Gentry committee. Briefly this speci-
fication called for detecting and tracking all objects having effective
reflection cross section areas greater than 0.1 square meter to altitudes
of 2,000 miles. The plan of attack was therefore to establish an experi-
mental facility with a minimum range capability and to establish engineering
studies and research projects as required to develop a specification for a
full scale DOPLOC system. Following this plan BRL launched a crash program
to install in the field an interim facility within a period of six months
beginning approximately 1 July 1958. Innnediate action was taken to collect
together the applicable equipment that was available in BRL and to place
orders for such equipment aa could be bought directly "off-the-shelf". As
fast as the problems could be analyzed sufficiently to prepare scopes of
work, negotiations were started toward contracts with other research
facilities.
B. Interim Erparimental Fence
In making plans for the interim fence, the choice of an ogerating
frequency was one of the first actions required. Since 108 mc was avail-
able for the Minitrack equipment under the Geophysical Year Program, and
since a q u i a study showed this f~equency to be clom t o optimum, and with
that equipment aLteady designed, 108 mc wae chosen for the interim fence.
!l!hwr&ical and prsctical considerations leading t o and confinnine the
wiedam of this choice of frequency were e~nnmarized a t sane length in section
I1 of th i s report.
Another early problem concerned the choice of station s i tes . The upper
end of White Sands Missile' Range and Fort S i l l , Oklehom, were found t o be
close t o the &wired great circle path and about the proper distance apart
t o covar t h e distance between W8MR and haphis with o m trcmaplitter for
illumination and two lacoivers f o r detection. Since land aud many other
facilities w e r e laadily available on the Dovarnment w e d and operated fail-
i t i e s , it was decidad t o uss WSMR and Fort Sf11 as two of the station6 e d
t o locata a thiz-3 near Mmphis, Tennessee. A great c i rc le from WSMR t o near
Mmphia through Fart S i l l ran through F m s t City, Arkansas, and just south
af Xau~phis. Since most of the land just south and southwest of Memphis is
e i ther heavily populated o r e lse very l o w r iver bottom land which is subject
t o flooding, the Forrest City area w s s chosen as a probable, good location
for one receiving station. A high section of land known a s Crawly's Ridge
runs just east of Forrest City, Arkansas. A hil l top one m i l e from the town
proved t o be an excellent location. It was close t o power and water, yet
f a r enough from develowd areas and traveled roads t o be quiet from the
point of view of electromagnetic noise. The radio transmitter station was
then located a t Fort 9111 and the other receiving station near Stall ion
S i t e near the upper end of WSMR. A s it turned out, the Fort S i l l and
Forrest City s i t e choices were excellent. The White Sands Missile Range
s i t e l e f t much t o be desired. This location was too isolated, being twelve
miles from Stall ion over undeveloped desert t r a i l s . Stall ion was i t s e l f
35 miles from the nearest villsge where housing could be obtained. Although
the s i t e was probably quiet in so far as man-made electr ical interference
was concerned, it was plagued with sand storms, sand s t a t i c and power failures. Long power and signal l ines were required v i a the attendant
l i ne failures. Sand leakage i n the building also caused am? damsge t o
moving e q u i w n t , tape recorders, relays, e tc . Sand also tended t o foul
up the cooling system and prevent proper cooling of electronic e q u i p n t .
-. 46 .
As a research station and a quick and ready place to try out ideas
experimentally, BRL retained and improved the active satellite tracking
station at Aberdeen Proving Ground. This station was moved from a
temporary shack to a permanent type building located on Spesutie Island.
This facility has proven to be very useful both to BRL personnel working
on the DOPLOC system and to the missile firing agencies working at
Wallops Island and at Cape Canaveral. Figure 4 is a rough outline of the
United States and indicates the positions of the DOPLCC stations.
The design of the station buildings and antenna fields was turned over
to the Little Rock Office of the U. S. Army Engineers. Using drawings from
Little Rock, the Albuquerque and Tulsa Offices of the U. S. Army Engineers
contracted for and supervised the construction of the WSMR and Fort Sill
stations. The Little Rock Engineers contracted for and supervised the
construction of the Forrest City station. The two receiving stations were
housed in prefabricated metal buildings. These buildings were 24 x 100 feet
single story structures, set on concrete slabs. The transmitter occupied a
],refabricated metal building which was 20 x 80 feet, also on a concrete slab. IA addition a 20 x 12 foot garage building of sheet metal was provided for
an auxiliary generator at the transmitter building. Since the generator was
never procured, the garage was utilized for a storage facility. Figure 4a
shows an exterior view of the transmitter building.
Since it was contrary to BRL policy to assume routine operational , :.
responsibility for a project such as the fence, the Signal Corps was requested
to plan to assume responsibility for the operational phase when and if a
suitable installation became available. In addition the U. S. Army Signal
Engineering Agency was requested to install the radio transmitter at Fort Sill.
After installation by USASEA the operational phase was to be turned over to
the U. S. Army Communications Agency. This agency assumed control of the
transmitter station and furnished a station supervisor and teletype operators.
At the receiving stations USACA furnished station supervisors and teletype
operators. since the equipment at the stations later became a research facil-
ity, USACA withdrew from the operation when it became apparent that no routine
operational phase would develop. Except for the station supervisors and
teletype operators the stations'were manned by Philco Technical Representa-
tive contract employees. 47
DOPLOC STATION COORDINATES
APQ YdJlYUID IO(MILST CITY, ARKANSAS We 45' 50. VEST AW 36- 00' 54. Y n T Y
m. S ~ L , O K U ~ W O I " 24' 20. WEST MiD M0 8s' 90. m u IMITL M O B , NEW @Elm 8 0 0 ~ 44' 40. WEST UD SSO 4s' 18. mt(( P D I S T M E E f s f f E W STAfDYS
L P . 0 . To - aw: w m u s C-ST uw JD roar su: 4os r l u s
mw s1u m earn .yoe: *a mlLLs L *
FIGURE 4, . LOCATION OF DOPLOC TRACKtNG STATIONS
Figure 4a DOPLOC Transmitter Building /
1. Radio Transmitter. In the DOPLOC satellite detection and tracking system, a high power radio transmitter is uscd to excite a ~ v r w beam
antenna which in turn illurninstas a satellite to cause radio frequency
energy to be reflected back to a receiving antenna. TO maat the six month installation period, it w e neceseary to utilize existing equipfent, since
the lead time for fabrication of any reasonably high power transmitter would
exceed six months. A used cmmercial 50-KW transmitter, manufactured by
Westinghouse Corporation wae locatad in storage. Since it had been built
several years previously, a contract was arranged with the Gates Radio Corpo-
ration to remove the f .m. modulation equipnt and to completely refurbish
the transmitter. Since the DOPLOC system requiree a very stable frequency
source for determining the transmitter frequency, the original frequency
. standard or oscillstor and the low power fregwncy multiplier and amplifier
stages were replaced with a high stability frequency standard, phase stable
multiplier and low power amplifiers. This exciter equipent was engineered
and fabricated by the U. S. Army Signal Engineering Laboratory, Fort Monmouth,
New Jersey.
Since the frequency standard which served as the basic frequency source 9 for the transmitter had an instability of one part in 10 or less over a
period of thirty minutes to one day, a precise method of frequency measure-
ment and monitoring was required. The frequency measuring equimnt at W e
transmitter station consisted of a second precision frequency standard, the
radio receiving and frequency comparison equipment required to compare the
standard's frequency with that of the National Bunau of Standard$' standard
frequency transmissions, a time code generator, a real-time counter and a
ten-megacycle-per-second counter. Figure 5 is a bloak diagram of the trans- mitter and frequency measuring system. Figure 6 is a photograph of this equipnent. Figure 7 and Figure 8 are two views of the transmitter instal- lation.
The code generator and real-time counter served as an electronic clock
indicating hours, minutes and seconds in digital form with the accuracy of
the basic frequency standard. Using code pulses and a calibrated oecillo-
graphic sweep, the time counter was synchronized with WWV signals to plus . , 50
- . . .
FREQUENCY STANDARD SOKW RIDID FREQUENCY. MULTIPLIER C
TRANSMITTER CO-AX , , TO I08 MC. WESTlWNWSE .8wlTCH
>
10 WATT AUPLIFIER A . ,
ANT EIIIU
MULTIPLIER v 1- TO 5 YC.
DIGITAL PRINTER E. E. CO. REAL
L A
H-P 560A - I t - CODE - A - - I ?!YE
GEWERATOR COUUTER .
BORG FREQUENCY STANDARD .
FIGURE 5 DOPLOC SYSTEM 5 0 KW RADIO TRANSMITTER AND FREQUENCY MEASURING EOUIPMEWT
Figure 6 Photograph of Frequency Generating and Measuring Equipment
Figure 7 View of Front of 50 KW Transmitter
or mlnu on0 acllliucond. me counter18 output war pin- out on tho
first rir v W r of a Emwlett-Fackard, d l 36-A d ig i t a l printer onco
uch srcond, or oiteilar, a t tho choice of th. 0pMtm.
Tb output of tha froquamy rt.adacd, a b l i b n t d with WLN, n r m d
u r tim bua for the ton-mgscyclr-per-wmd omtar , H.P. aodal 7219. An r.f. mob. frua tho output amplifier of the t r ~ r m i t t o r lupplied a 10%- rignal which wu fod t o t h i s cycle countor t o obtain a frequency
.muuranent. Tb. output of the first fivu digit8 of tho cycle counter wu
fed to tho remaining five digit w h e e l 8 of the printer w b n it w a s printad
out on c w by control pulses frm the tLp. c o b gonorator. Thus, tho
printed record contained the transmitter froqwncy t o an accuracy of p l u
or minua one cycle, w i t h real time data, a t int.rva.8 an ravrll as o m
second. A hall-duplex, leered, cammercisl taletype line connected the tM.- mitting station t o a l l the receiving stations and t o the canputation centor
a t the B a i ~ i s t i c Research Laboratories, allowing either manual or tape trans-
mission of the transmitter frequency data.
The output of the transmitter was fed t o a system of motor driven,
remota1.y controlled, coaxial cable switchor which providrd for switching
the transmitter power t o any of the throe high gain antennas. The antenna8
were located 200 f ee t fran the transmitter and fed through 6 1/8" d i m t o r
coaxial cable. Each antenna radiated a fan shaped beam having a solid angle
coverage of 76 x 8 degrha. Orientstion of the antennas is shown i n Figure 9. The beam frm the center antenna waa directed vertically with i t s 76-degree
dimension i n the east-west direction. The center of the north antenna beam
was rotated t o 20 de@-ees above the horizon with the 76-degree dimension in
the north-east t o north-wet sector, w h i l e the south antenna was similarly
inclined t o the south.
This arrangement was designed t o powi t acquioition and tracking of a
~ a t e l l i t e soon siter it appeared above the horizon, Md then switching t o
the other two antennas in euccession for tracking through their respective
be- t o obtain three segnvnts of the typical Doppler "S" curve. Tiltiw
end rotational adjurtmnts wore provided i n both the transmitting and
nceiving antennas t o provide for aligning the bow8 to,view and illuminate
the s m e space. 55
Figure 9 Orientation of Transmitting Antennas
<', s.,
In order t o provide complete test f a c i l i t i e s and a l s o t o g e t the
transmitter i n t o operation as soon as possible, a single dipole antenna
was mounted over a concrete service p i t on a copper sheet ground plane.
This antenna was w e d f o r a few weeks before the high gain antennas were
delivered. However, it had l imi ted power handling capacity of about 25 KW
because of an improperly designed sect ion. The dipole was never used a f t e r
the high gain antennas became avai lable. Extra precaution was taken t o
provide an adequate ground plane system and good ear th grounds i n both the
t ransmi t te r bui lding and the antenna f i e l d . A spec ia l duct and blower
system was provided t o cool the t ransmit ter i n summer and t o conserve some
of the t ransmit ter heat t o heat the building i n winter.
2 DOPLOC Receiving Stat ion Interim - Radio Frequency Circui ts . When
t h e interim DOPLOC fence equipment was being planned, radio receivers which
operated d i r e c t l y a t 108 mc were not immediately avai lable. A system of
preamplifiers and converters was therefore used t o reduce the frequency t o
t h e tuning range of a standard type R-390A radio receiver . A BRL b u i l t
preamplifier operating a t 108 mc was connected between the antennas and
a Tapetone Model TC-108 converter. Early in the program t rans is tor ized ,
s tabi l ized , c r y s t a l o sc i l l a to r s were used a t 61.2 mc and the frequency
doubled t o 122.4 mc t o convert t o 14.4 mc a t the input t o the R-390A receiver.
Mother standard frequency source was used f o r the conversion frequency i n
the receiver. The frequency biased Doppler s igna l was then fed t o the
tracking f i l t e r and t o the magnetic tape recorder. Later i n the program,
Borg frequency standards became avai lable a s frequency sources. Rohde and
Schwarz frequency synthesizers and decade frequency mul t ip l ie rs became
avai lable f o r conversion purposes. These were very convenient since any
frequency over extremely wide ranges could be obtained t o an accuracy of 9 1 cycle per second and an i n s t a b i l i t y of l e s s than one p a r t i n 1 0 .
Figure 10 is a bas ic block diagram of the r. f . sect ion used i n the interim
DOPLOC s ta t ion . A 108-mc crystal-controlled s ignal generator was loosely
coupled t o the antenna or connected t o a dipole antenna f o r a b u i l t - i n
s e n s i t i v i t y ca l ibra t ion .
I4.4'C + ? I C I FREOUUCl R C S C I I
1 W U U I E I ( F S W U
(UP s u m )
C W U T U S W R E I
t o w n ISIII
r n E a u c n n (QIITLR W P SC481
- uumu
Amlo-.o.lmo Iwm ram D I C l L U r n
(1110.. -1 - m l * l E . I u l T l l r n
n.*IC I D U I UL..nn
I 1m.c
musun *uOU11S1
I
FIGURE 10 BLOCK DIAGRAM OF R.F. SECTION IN W U K : STATION PASSIVE TRACKING AT I O B Y C
O s c I u A m .1.115 I M L )
W U L E n - COI*CRT€R IOs'C 14r'c 14.4 'C + DO'PLLR
AUPLIFLI w.1 me m r z . r r c 1 IlAPCTO*CI
(.RLI 1.2 IDS I
Figure 11 chows . . . - . a . . . . block disgrarP of tha equiprunt wed f o r active
tmcklng, (Radiating sa ta l l l tus ) . F& channel6 of -iving equipment
were available i n each station. Cme channel could work directly into tha e n c a r s and teletype equipment thence be.ck t o the computing
laboratmy. Othar mcorde which occurred during record tranamlssion
were aagnetically recorded and transmitted sequentially. Each receiving
station wss also provided with receiving equigrment capable of receiving.
frm active s a t e l l i t ee over a wide range of frequencies, actually from
20 mc t o 960 me. In general just one channel of equipent was available
a t these special frequencies. mL Report No. 1123 ent i t led "DOPLOC
Instrumentation System for Sa te l l i t e hacking" by C. L. Adem describes
i n de ta i l the mechanics of the instrumentation.
For extending the frequency range from 55 t o 9 0 mc, a Nems-Clarke
radio reciever and a range extension unit, model REU-300-B were used.
3. Automatic Search and Lock-on. One of the most ingenious develop-
ments of the DOPLOC system was an automatic signal search and lock-on
&*ice which automatically placed the tracking f i l t e r on any signal
occurring within the search range of the equipment. The energy reflected
from a sa t e l l i t e is weak and uaually imbedded i n noise. For a passive
system t o perform defense surveillance requires 24-hour-a-day operation
with maximum detection sensit ivity and the shortest possible reaction
time. Manual searching requiring constant operatar at tention would c a w
excessive operator fatigue and greatly reduce the dettrction capability of
the system. The short duration and low signal strength of the signals
would l imit the operation of the tracking f i l t e r s . Signals f a r down in
the noise would require excessive time t o identify and lock on manually.
Therefore an automatic search and lock-on system was found t o be necessary.
The automatic lock-on (ALo) described below overcame the limitations of
manual operation and provided a ful ly automatic search fac i l i ty . The
following basic requirements were eetablished for the ALO:
a. The chosen audio frequency band, in th i s case 12 kc must be
searched i n a minimum time.
&.* Dl - POLE CQIER REFLECTOI * '3'. +E ga, 2 - M MC +#)-LEI IW UC +DOPPLER
1.- - TEST SIBNAL
S H U C Oe*El.m(l (Y m.0)
TO RECORDIN6 *EIL TIME TO RECCOID(N6 SYSTEM SYSTXM
FIGURE 11 BLOCK DIAGRAM OF R.F. SECTION FOR ACTIVE TRACKING ( 2 CHANNELS OF 4 AVAILABLE AT -6. NO. )
b. The system must be capable of finding the smallest theoretical
signal determined by the search bandwidth.
c. The time (known as acquisition time) required t o recognize and
use the signal .%hat is present i n noise must be held t o a minimum.
d. The system must position the tracking f i l t e r oscil lator t o the
desired frequency and place the f i l t e r i n a lock condition.
A block diagram of the ALO system i s shown i n Figure 12.
To meet the requirements of minimum search time, ten f i l t e r s of fixed
bandwidth were placed i n the desired audio frequency band and the signal
frequency swept across the f i l t e r s . The ten fixed f i l t e r s were each spaced
100 cycles per second apart t o cover a one-kc band, compatable w i t h tracking
f i l t e r bandwidth characteristics. These f i l t e r s were then switched twelve
times t o cover the ful l12-kc band. The ra te of switching determined the
time available t o find a signal as well as the t o t a l sweep time. Three
switching ra tes were used, 2.5, 5 and 10 cps. When the 10 cycles per
second rate was used, the time for each 1-kc band was 100 milliseconds and
for the whole band 1.2 seconds. The mixing and switching techniques employed
i n the ALO removed the requirement for 120 individual oscil lators or ten
switchable oscil lators normally needed for coverage a t 100-cps bandwidth from
2 t o 14 kc.
Each f i l t e r consisted of a mixer (which had the output of an oscil lator
for one input and the input signal plus noise for the other), a low pass
filter and a control relay. The low pass f i l t e r s were variable i n fixed
steps fran 0.5 t o 50 cps. !The part of the frequency spectrum viewed by each
f i l t e r was the oscil lator frequency plus or minus the bandwidth of the low-
pass f i l t e r . Shifting the oscil lator frequency allowed swltching of the
f i l t e r t o cover another frequency band. The oscil lators were placed 100 cps
apart and covered a 1-kc band. The oscil lator frequencies were switched i n
groups of ten from two t o three kc, three t o four kc and so on t o U t o 14 kc.
Mri.83. pulses necessary for gating and switching were derived f r m a counter
with a count cycle of 12. A frequency generator chassis using 1 kc and
100 cps inputs, obtained from the stations timing system, provided the
multiplication and mixing t o produce 10 simultaneous output frequencies.
Each pulse from the counter changed the eignals being mixed and shifted
the range by 1 kc. If a signal was present in the noise i n the frequency
range from 2 t o 1 4 kc, of sufficient amplitude, and within the bandwidth
of one of the oscil lators, it activated one of the f i l t e r s .
After the presence and the frequency of a signal was indicated by the
ALO, the phase-lock tracking f i l t e r had t o be correctly positioned. In
the track position, an internal oscil lator was locked t o the input signal
by the use of a feed back loop. The automatic lock-on circui ts used an
external loop around the tracking f i l t e r in the s e t position t o lock the
oscil lator near the input signal. The fixed f i l t e r , upon receiving a
signal, closed i t s control relay wi th one se t of its contacts, putting the
oscil lator frequency of the fixed f i l t e r into a c i rcui t termed the "set
frequency" control. This c i rcu i t was part of the external loop and compared
the fixed f i l t e r oscil lator frequency with the tracking f i l t e r oscil lator
frequency and generated an error voltsge which was applied t o the tracking
filter and controlled its oscil lator t o produce a phase-lock between the
two oscil lators.
In the automatic system, the "set track" switch i n a standard tracking
f i l t e r was replaced by a rotary solenoid. A t the time the fixed filter
oscil lator was applied t o the "set frequency" control, voltage was also
applied t o the solenoid. The tracking filter was positioned t o the correct
frequency before the solenoid moved t o the track position. In the'track
position, the tracking f i l t e r loop uas closed and a phase lock was obtained
between the internal osciUator.and the true input signal.
4. The Phase-lock Tracking F i l te r . The phase-lock tracking f i l t e r
used in the interim DOPLOC system was the Model N as supplied by the
Interstate Electronics Corporation, Anaheim, California. This tracking
f i l t e r was the resul t of several BRL developmental contracts start ing with
the development of PARDOP by the Ralph M. Parsons Company, Pasadena,
California. It was an electronic bandpass f i l t e r whose center frequency
was made t o track automatically the frequency of the input signal. This . .
was accomplished with a v e r y t i g h t , phase-locked, servo-controlled c i rcu i t .
A very large signal-to-noise improvement was obtained by the use of an, =-trem-py narp~.; filter ban&jidth relatpre to the iw2t c(rcai+_ ha&vi&+_h;
Figure 13 i s a block diagram of the $racking f i l t e r system with the basic
closed loop system sham i n heavy l ines .
The input Doppler s ignal was fed t o the input mixer where it was
subtracted from the VCO output frequency t o r e s u l t i n a 25-kc signal . This
was amplified i n a tuned 1.f . stage of about 250-cps bandwidth. The ampli-
f i e d signal was then fed t o the main phase detector where it was compared
i n phase with the o u t g t of a 25-kc fixed frequency c rys ta l osc i l l a to r .
Ut i l iz ing a cross-correlation type detector and equalizer network, the osc i l -
l a t o r was controlled t o follow the variat ions of frequency and phase of the
input s ignal . This technique allowed smooth tracking and extrapolation of
the output phase i n the event of shor t periods of s ignal drop-out by including
an effect ive acceleration memory. Noise pulses or other s ignal interruptions
could r e s u l t i n the loss of s ignals fo r a short time during tracking missions.
The memory feature allowed the filter t o operate through these in tervals with-
out losing lock o r introducing e r r o r s i n the data. Output s ignals from the
voltage controlled osc i l l a to r were translated down t o the audio frequency
rmge i n the output mixer. In the model IV filter, this range was 100 cps
t o 20 lsc. The output Doppler s ignal from the mixer was i n phase with the
input Doppler s ignal t o the tracking f i l t e r because the inherent 90° phase
s h i f t i n the basic tracking loop was campensated f o r by the phase-shift net-
works i n the c rys ta l osc i l l a to r c i r c u i t . The output mixer fed an output
8 .c . amplifier through a low pass filter which removed modulation products
a t the output terminals.
In addition t o the main tracking phase-lock loop, a closed loop auto-
matic gain control (A.G.c.) sect ion was b u i l t in to the f i l ter . This included
tin auxi l iary phase detector, d.c. reference signal, d.c. amplifier and 1.f.
amplifier. The measured amplitude of the auxil iary phase detector was sub-
. t k c t e d from the d.c. reference signal and any difference was amplified and
applied as bias t o the 1 . f . amplifier. '.
64 :+p3
.~2:.
a,..:. ZC -., ."L...
- CORRE - D.C. LATIOM
OUTPUT
DOPPLER / 4 SIGNAL
.a, INPUT I F MIXER
DOPPLER
,b' INPUT AMP. CI.
SIGNAL
A OUTPUT
FREQ. A FREQ. L
ANALOG FREQ.
AMP. OUTPUT
FIGURE 'D BLOCK DIAGRAM OF PHASE-LOCK TRACKING FILTER INTERSTATE MODEL Z
PHASE + ERROR
OUTPUT
I .
pm PHASE ERROR
0.c. _AMP. +
. . . . . . .
In addition t o the deai&d f i l t e r e d Doppler a*gnal, other o u t p u t ~ ~ w e r e
made available both ' t b f ront pan.1 indicating metera and t o output terIDirial8. . . . . These were the correlat ion output, the.din pM.6~ aetector output a n d t h e
-log frequency output. Two factors:which were . . imgortant i n the operation
and application of the tracking filter were iiignal-to-nofse r a t i o and si&nal
dynamics.
The signal-to-noiae r a t i o improvement by , the tracking -filter was t h a t . ,
of the r a t i o of input noiss,.bsndwidthto f i L t e r bandwidth. Line A o f Figure
1 4 shows the re la t ion o f the input noise-to-signal (N/s) i-atio i n db t o the
bandwidth fo r which the outgut S/N w i . 1 1 be unity, wh6n We input bandwidth
is 10 kc. The input N/8 conditian was convenientlj expressed i n %enus of
r a t i o s i n db above unity N/S since input S)N r a t i o s ,Bo niuch knialler than one
.are. ordinar i ly used. Figure 14 a l so shows the i i e a g ~ k d rn&ik.m input N/S t o
maintain lock-on f o r filter bandwidth set t ings between 0.5 and 50 cps.
Tests indicated the capabi l i ty of . . the tracking filter t o lock on t o a aignal
'which i s buried i n nbise 6300 t o 1 (38.db) when the nois i input bandwidth
was 16 kc and the' f i l t e r bandwidth was one cycle per second. For a receiver . ,
.with a 3-db noise f igure, the.38-db value i s equivalent t o an input power
of' 2 x watts. In te rns of receiver performance with conUb6nly usid
uni t s , t h i s is a s e n s i t i v i t y o i -197 dbi o i0 .001 uiicl'ovolts across 50 ohms. . . . . . . 3
This gives & '4-db output S/N, correspo~diiig t o about 36' o r 0; i djrcle per .. , . . second rms phase J i t t e r .
The r a t e of change of received s ignal frequency due t o r ad ia l components
of velocity, acceleration and r a t e of change of acceleration defines the
s ignal dynamics. For optimum information reception, the proper bandwidth
can be selected t o match the s ignal dynamics a s well a s the S/N r a t i o of the
s ignal received. Figure 15 ahowe the re la t ion of the tracking f i l t e r band-
width t o the maximum r a t e of frequency change and t o the equivalent acceler-
a t ion i n g 's . Additional technical specifications fo r the tracking f i l t e r
a r e contained i n BRL Report No. 1123. Figure 16 is a photograph of the f ront
panel of the tracking filter.
5. Timing. Any system of satellite tracking is dependent upm
having bq+ a universal time standard and a locel: time or &B~encY standard. For the DOPIAX iiiterfm system the National Bureau Of Stan-
standard frequency emissions constituted the standard of time. LOC~. oscillators of the stabilized quartz frequency controlled type served as
the 1- source of frequency.
The local oscillator or frequency standard in each station was a type
0-471 manufactured by the Borg Corporation. It had outputs of 100,000 and
1,000,000 cpa. This equipment was temperature controlled and kept in
constant operation. The basic crystal frequency was actually five mga-
cycles per second. Stand-by batteries were built into the oscillator
cabinet and sen& to supply power over brief periods of power makns failure.
Figure 17 is a block diagram of the radio frequency oscillator, 0-471(~~-1)/~. Figures 18 a and b are photographs of the frequency standard. This standard \
9 provided a frequency which had an instability of less than 1 part in 10 over
periods of 30 seconds to 1 day.
Various frequencies are derived from the Borg oscillator 100 KC/S
output by divider and multiplier circuits. One type of divider circuit is
a commercially available Time Code Generator, Model ZA1935 manufactured by Electronic Engineering Company.
This circuit served a dual purpose. It produced outputs of 10 kc, 1 kc,
one per second pulses,. one per minute pulses and one per 10 minute pulses.
In addition to these outputs the unit generated a time code suitable for
recording on magnetic tape, and displayed this code on front panel decimal
indicators in increments of hours, minutes and seconds. These decimal
indicators, appearing as two vertical columns for each unit of time, changed
as the code was generated, and the visual display could b6 pre-set to any
desired time in a 24-hour priod, allowing the coded time to correspond with
the time of day. In most operations it was convenient to set the clock type
indicators to Greenwich Mean Tine. Figure 19 is a block diagram of the DOPLOC timing system.
:I:
t i ...
- - - REFERENCE
I AC POWER SUPPLY VOLTA6E
1 . . -
CONTROL, SATMA= -*CONTROL, TRAMSISToR -* RECT. . AMPLIFIER
REACTOR i
115 V POWER CONTROLLED , - POWER . " TRANSFCMYER BRID6E 27 VDC ,27 VAC RECTIFIER REBLAm
- - RELAY0 027 VAC
BATTERY -1- C O w ~ T o r \ - -
STUIDBY BATTEW . l.m r- 7 V NI. CAD BATTERY VOLTA6E
DISCOUMfCT
- . OSCILkATm --- I mc
ourcur 1
TRANSFER RELAY 43 I
ma OWN *HEATER WIND= HEATER ONLY OVEN
DC -- FR-Y FaEalmm m*C p T R O L L E R 5 %
-&m DIVIDER DIVIDER REGULATOR OSCILLATOR -
AND f 5 i 10 REG. 20 V DC INNER a OVENS
WNVERTER 4- 5 UC TO I YC l YC TO 1 0 0 1CC OVEN a OVEN
TRANSISTOR CONTROLLER POWER
\ B P M ~ ? € R J L W ~ V ~ E ~ ~ ~ - ~ j i 100 VDC REGULATED PLATE SUPPLY - - - - - - - - 1 - I - * - - -
. FIGURE ..17-.. RADIO..FREOUENCY OSCILLATOR, 0-471 (XN - I ) /Us BLOCK DlAGRAY I
4 - -
Harmonics of 100 kc and one per second pulse output were used to the time output with signals received from the National Bureau
of Standards station WWV. Provision was made on the rear of the time code
generator chassis t o couple loosely the harmonics of the basic 100-kc
frequency t o the antenna of a W receiver such as Specific Products Can-
pany I s type W C . This receiver .provided several bands for reception of
WUV time signale. In this locality 5 mc provided the best signal. Repeated
tusts conducted over a 24 hour geriod indicated that the instabil t ty in 8 the basic 100-kc frequency was approximately 2 parts i n 10 .
Since the one-per-second output from the time code generator could be
shifted in small time increments, th i s signal could be accurately synchro-
nized with the one per second tone bursts f r m WWV. This WWV signal was
dieplayed on the ~ r t i c a l curis of a monitor scope whose horizontal sweep
was triggered from the station-generated one-per-second pulse. Proper manip-
ulation of either the "Advance" or "Retard" button enabled the operator t o
establish coinci&nce of the start of the tone burst w i t h the start of the
scope sweep t o an accuracy of one millisecond. Sychronization of the
station.'s timing system occurred a t leas t once each day and was always
checked just prior t o each tracking operation. Clock time of day was
encoded on the basis of a 24-hour clock w i t h sFx code groups neceesary t o
display a 24-hour interval. The code was generated in a 4-2-2-1 binary
coded decimal form, w i t h each time indication identified by a 20-digit code.
The output was a 100-cycle carrier, modulated a t a 25-pulse-per-second rate,
upon which a 1000-cycle signal was superimposed. Although not ueed for seakch purposes, the code was suitable fo r magnetic tape search.
Figure 20 indicates how the outputs derived Pram the ttnre co&e &urer- ator c i rcui t were distributed via switch circuits atad cathode followers t o
the various timing and control circuits required i n the system. 8tsnderd
pulse shapers and inverters produced the desired polarity, duration, and
amplitude of signals for the timing applications.
Timing mar& were producad on one edge of the stripchart record on both the Varian and Banbarn recorders at the rates of ane per second and
one p r minute. In addition, one-per-tan eecond markers could be super- imposed directly on the date tracp of the Varian recorder. The 1000-cycle
output af tha tims coda generator was ueed to produce calibration freqwn- cies in 3 kc steps o m r the range of 3 to 15 kc for calibrating the Varian recorders.
By front panel switching for each of four reoording channels, a print-
out selector circuit allowed proper data and timing inputs to be switched
to the pre-set and display counters for either period or frequency measure-
ments. A low frequency pulse generator operating from the one-per-second
input also allowed switching the print rate of the digital recorders from
one-per-second to once per 2, 4, 8 or 10 seconds.
The digital clock and the real time counters were driven f r m the
basic one-per-second pulee derived from the time code generator. The real
time counters operated in a manner similar ,to that of the decade counters
in the time code generator described above. Time wae advanced in one-per-
second increments and displayed in hours, mbutes and seconds, with the
outputs being a discrete voltage level for each digit displayed. A Hewlett
Packard type 560A digital printer was used to print out this information.
The digital clock also advanced in one-second increments, and time was
displayed in hours, minutes and seconds with the output in the form of
contact relay closures for each digit indication. Modification of the
56W digital printer was necessary before it would accept these data. The
data digitizing also required time indications and these were produced in
a binary time code generator by counting the one-per-second time pulses in
a straight binary form and displaying the output with front panel indicators.
Command pulees, generated external to the data digitizer and synchronieed
to the time pulsee, shifted the time data out of the code generator into
the data digitieer for read-out.
. . . . .
A &p recorder tk d o n W c i r c u i t was y e d t o begin time pulse . . . . . . . . : . . , .
recording at a one-per-second rate t o the n e e s t $en-second r e a l gee . . . . . .
indication. This tin% was manually recorded and served as a known start
time fo r se t t ing up indications f o r play-back of tape recorded data-. The
recorded one-per-second pulses were then used ae the timing source f o r all
play-back requirements.
6. Data Recording ~ ~ u i g r n e n t . The basic Doppler informetion, available
a t the output of the phase-locked tracking filter, was a constant amplitude,
varying frequency s ine wave. Other data available from the f i l t e r were the
correlation signal, the phase e r ror output, analog frequency i n terms of a
d.c. voltage and the autamatic gain control voltage. Recording was acccsn-
plished in both s n u g and d i g i t a l form u t i l i e ing magnetic tape recorders,
s t r i p chsrt recorders, d i g i t a l pr in ters and =per tape punches. Figure 21
i s a block diagram of the =cording section used in the DOPLOC receiving
s ta t ions . Each t;ype of recording is b r i e f ly *scribed i n the following:
a. Digi tal Pr in t Gut. Pre-set counters and display counters provided
readings of either period o r frequency of the tracking s ignal fr& the out-
put Of the tracking f i l t e r . The type of recording u t i l ieed , period or
frequency, was determined by the required accuracy of a specif ic tracking
operation. When frequency output was desired, a known number of cycles of
the external timing frequency obtained from the precision frequency standard,
was counted by a pre-set counter a f t e r proper se t t ing of the multiplier d i a l s .
A start pulse derived at the beginning of t h i s count and a s top pulse derived
a t the end of the count, were used as gate controls f o r a second counter
ca l led the display counter. The signal of unknown frequency was fed t o the
gate of the display counter, where it was gated by the control gate pulses,
t o the counter decades f o r counting and displaying. For exmple if a 10-kc
time base was used and the mult ipl ier d i a l s of the pre-set counter were set t o read 10,000 then the gate would be open f o r one second, and a 5-kc fre- A . * quency would be d i rec t ly read out as the number 5000. . .i -
When period output was desired, the signrrl frequency wa8 fed t o the
input of t h e p r e - s e t counter, and the multiplier d i a l s were set t o count
the"dee1red number of cycles of t h i s frequency. The display counter then
counted the timing pulses fo r the interval of a k n m number of cycles of
the signal frequency and displeyed the count. For oxauiple, assuming the
Doppler frequency t o be 2 kc/s, the multiplier dials of the pre-set counter
might be s e t t o 1000 to allow the display counter gete t o be open for one
half second. I f the timfng frequency were 100 kc/s, the dlsplqy counter ---.. 3 s ---> LL- -.._L-- -a A- --- &-.- ---- 1 - ~n AM mbaa w u u u rrou me rrwu- w g u r u a +n wc nnu m c c s m u us a1w uuuuro. * r u m
thus produced a period measurament which is the inverse of the previouely
&scribed f'requency measurcrmant.
The period measurement obviously provided the highest accuracy. For
improved accuracy of measurement a 10-mc counter was used t o obtain the
period count. The recorded data were printed out on the Iiewlett Packard
model %0A printer. This printer operated with counters using the binary
decimal counting system and produced a specific wl tage level for each
displayed number. Thie voltage level was fed to the corrr~ponding digi ta l
place i n the printer which sensed the voltage an8 printed the number a s - played on the counter. The printer accepted data entereb i n paral lel snd
a l l columns were printed simultaneously on command. The capacity of the
printer was eleven 6 ig i t a i piaces. For fiequeiicy, or imer resOiu%iOii
period measurements, the f i r s t f ive digi ts were used for the Doppler data,
and the l a s t s ix were fo r time indication. For period measurements Of
higher resolution, using the 10-mc counter, the first seven digits were
used for the Doppler data, and the remaining four f a r tima indications.
The time data presented to the appropriate p r in t wheels could be one
of two kinds, depending on the circuitry used t o generate them. I f the
data were generated from modified counter decades with a specific voltage
level produced for each decimal number in the decade, then no internal
modification of the printer was required. Barever th i s requireu a separate
rea l time counter fo r each digi ta l printer used. I f aeveral channels of . ..
data were being printed, the equivalent number of rea l time counters
required became cumbersome and d i f f i cu l t t o synchronize with WWV time
sigqals.
The second kind of rea l time data was generated from a d ig i t a l clock, g . %.
Chrono-log model 2600-4, the output of which was i n relay contact closures. -:? '
~odi f ica t ions t o the d igi ta l printer were required for those positions where
- *i ' 80
tima was recorded. The advantages of using this clock were the ease of
se t t ing up the real time information each day, the ease of synchronization
w i t h WHV time signals and the f a c t that one clock could drive many pr inters .
A l l of the printed time data were ident ical and properly correlated. Tne
disadvantage was the modification of the pr in ter t o accept the time inform-
t ion frm the d i g i t a l clock. The pr in ter became a hybrid instrument with
p a r t of the p r i n t wheels operating from voltage levels and pa r t from contact
closures. Once the modification had been made the r e l i a b i l i t y and ease of
operation were greatly improved.
b. Steip Chart Recorders. Two types of s t r i p chart recorders were
a l so used t o record the output of the tracking f i l t e r . The first was a
single channel Varian recorder, model 0-11-A, u t i l i z ing a f ive inch wide
recording paper. Timing marks were placed on the edge of the paper by an
independent s tylus and a lso superimposed on the data stylus. The Doppler
data w e r e fed t o an integrator where the frequency was converted t o a d.c.
voltage signal. Frequent frequency versus deflection calibrations were
made.
The autamatic gain control and correlation outputs of the tracking
f i l t e r were recorded on a two-chamel Sanborn model 152-100B, normally a t
1 mm per second chart speed. Timing marks were recorded on the edge of the
paper. Calibrations were applied after a s a t e l l i t e pass.
c. Magnetic Tape Recording. Since provision was made f o r handling
only one channel of data through the r e a l time encoder f o r transmission
over the teletype l i n e t o the computer, some provision had t o be made f o r
storing the data from the other three channels. Back up data were a l so
desirable i n case of other equipment f a i lu re or errors i n data trans-
mission. Therefore, all data from the tracking filters were recorded on
magnetic tape. An Ampex FR-114 fourteen-channel recorder u t i l i z ing one-
inch tape was used a t each receiving s tat ion. Tape speeds from 1 7/8
inches per second t o 60 inches per second were available. On most tracking
operations tape recordings w&e made of receiver output, tracking filter
output, s ta t ion t h i n g s ignals and WWV time signala. Spare channels were
tbsn available f o r recording any other data desired. In the reproduce
mob four channels were available f o r simultaneous playback of any four of tbr switch selected fourteen channel6 of data.
81
An Ampex model FR-1100 magnetic tape recorder, using 112" tap, and
recording at speeds of 3 314 t o 30 inches per second was used t o recard
the binary output data i n d i g i t a l form a t the output of the serial shift regis te r .
d. Doppler Data Digitizer. To permit rapid data handling and real
time transmission back t o the computer it was necessary t o d i g i t i z e and
ehcode the Doppler data a t the stat ions. Two basic functions of the
d ig i t i za t ion system w e e t o produce in binary form real time data at the
beginning of each recorded output and period data of the Doppler signal.
The f i r s t function was accomplished by the binary time code generator a t
each tracking s t a t ion and the second by the Doppler Data ~ i g i t i z e r (DDD)
f o r each recording channel i n the stat ion. Standard t ransis tor ized
building block logic was used i n both equipments. The DDD was bu i l t t o
BRL specifications by the Hoover Electronics Company of Baltimore, Mryland.
This equipment i s described i n d e t a i l i n BRL Report No. 1123 by C. L. Adams.
Figure 22 i s a block diagram of a single channel Doppler Data Digitizer.
Figure 23 i s a simplifies block diagram of the interim DOPLOC Data handling
System. Figure 24 is a photograph of the data Digitizer. The d ig i t i ze r
operated a tape punch drlyer which i n turn operated a Friden paper punch
Model 2 t o produce the binary data i n the proper format.
e . Data Format. The punched-paper-tape type of output data were
produced i n s t r a igh t binary form because of the input requirement of the
BRL ORDVAC computer used i n the calculation of the s a t e l l i t e o rb i t a l
parameters. Each recording channel i n the s ta t ion produced data i n stand-
& f ive l eve l teletype tape i n the format shown in Figure 25.
A camplete data block, or word, was contained in seven rows of the
five-channel tape with t h e and Doppler data in sFx r o w s of the f F r s t
three2,channels. A t the beginning of each w a r d , in pa ra l l e l with the d
binary data i n row one, a block or t i m e marker b i t was produced in channel
four. This occurred at a one-per-second r a t e and w a s correlated with the
d i g i t a l pr in ter readout s ignal . Modification i n the d ig i t i ze r c i r cu i t s
w a s mde t o allow a one-per-minute b i t to be punched in row two of c h a ~ e l
four to . fur ther correlate timing data. An end of word marker, acting as
a .caiiputer control, w e s prOauced i n a l l five :channels of row seven. , ... ...& . . 3:. 82
FIGURE.22 BLOCK DIAGRAM OF SINGLE CHANNEL DOPPLER DATA DIGITIZER
- a.
. GATED DOPPLER BINARY TRANSFER
INPUT FROM- COUNTER CONTROL PRE-SET COUNTER 2 lo
BINARY TlYE COM GENERATOR
T4YE INPUT - I PER S,FCWO
t ,I - - - - -
r - I I A
TO 3 OTHER - - I - - I 1 - 1 ~ I K E CHANNELS
- , ,
W
MAGNETIC TAPE
RECORDER OOPPLER
GATE
I I - I I t U C E DELAYED TIME OUTPUT
I
TIME GATE
DOPPLER GATE
TlME GATE
SHIFT M I S T E R + CONTROL UNIT
FOR
CONTROL 1 - SERIAL REGISTER
(4 CHANNEL) E S l W cwwul
I , I I I ( ~ a
BUFFER CONTROL -
TAPE wmn
TAPE PUNCH DRIVER
. PARALLEL
BUFFER REGISTER
- SPROCKET LIKE CHANNELS SERIAL SHIFT. "OR"
AN0 U I I BLOCK UARKER
GATE
i
I I GENERATOR
ORlVER
I t + - - - MAGNETIC
TAPE FROM 3 OTHER
- t RECORDER LIKE CHANNELS 1
I I I
L ,,------,---------- -
FRIOEN PAPER TAPE
PUNCH -
FRIOEN PAPER TAPE wrrcn
* - --J
I COWUTING DATA TAPE
DIGITAL - - . - - - MQITIZER - RECORDER * INDICATOR (7 CHANNEL) (DECIMAL)
PAPER TAPE RECORDER #2
(7 CHANNEL)
REAL TIME .
TIME CODE. PLAYBACK COUNTER
GENERATOR REGISTER (SYNCHRONIZED (BINARY) (BINARY) wlrn wwv) i
TIM~~NG PULSES PUNCHED TAR!
TRANSMITTER RECEIVER
PAPER TAPE PUNCH
L,-4---
i RECEIVING SITE , , - i COMPU~NG SITE& L - - - - - - - TELETYPE
LINES PUNCHED TAPE TRANSMITTER
+ . OUTPUT DATA
SIMPLIFIED BLOCK DIAGRAM - DOPLOC SYSTEM DATA HANDLING., . FIGME 8,- . 84 3.y
Figure 24 Data Digitizing Equipment
85
Manual operation of a start switch at the console, or automatic start switching from the autcmnatic lock-on circuit6 when the tracking f i l t e r locks
on a received signal, began the tape punchi* sequence. The first data word
contained universal time data and a l l subsequent blocks contained Doppler
data. The d a t a were punched i n parallel, one row a t a time a t a ra te of 60
lnillieeconds per row. After each end of row marker i n row seven, the tape
was at rest fo r appraxlrnstely 0.5 second while the counting process in the
d igi t izer took place. The om-per-second timing signal triggered the readout
systeia and the count jus t accumulated was punched out.
f . Data Wansmission. A retry irportant part of any satellite tracking
net is the data tr'ansmisaion equipment. If data are to be transmitted i n
real time and &bits camputed i n minutes a f te r a s a t e l l i t e pass, the trans-
mission system m u s t be reasonably f a s t j accurate and dependable. In
designing the DOPLOC interim e q u i s n t it was determined that satisfactory
tx-ansmieriion could be obtained using a 60-word-per-minute ccanmercial tele- graph channel, i f connected t o terminal equipmint that contained a system
of parity checking of transmitted data. A commercial paper tape t ransmi t t e~
receiver called Teledata, manufactured by Friden, Incorporated, was chosen.
The model 7-B Teledata shown i n Figure 26 was used with five channel tape
and consisted of a reader-punch chassis and a transmitter receiver chassis.
One Teledata unit was required a t each f i e ld station location and one for
each stat ion 's output a t the computer location, connected by a full duplex
telegraph line. This type of connection allowed simultaneous transmission
i n both directions.
Any type of f ive b i t code read f r m the transmitting tape was trans-
mitted as a seven b i t code by adding a redundant b i t t o each of two groups
of the f ive tape code b i t s t o farm two parity groups, one group containing
four b i t s and the other group containing three b i t s . The sequential trans-
mission cycle was so arranged that b i t s were sent alternately from the two
parity groups. The received code was checked for accuracy i n the tape punch section by independently checking of each group of b i t s . Operation
of an even number of punch pins i n either group was registered as an error.
The pins associated with the redundant b i t s do not punch the tape so that
the received tape was in the identical f ive channel form as the transmitted
tape.
In order t o u t i l i ze the f u l l capabilities of the f u l l duplex l ine,
provision was made t o switch the l ines a t the ends of the links from the
Teledata t o cammercial teletype machines, thereby providing simultaneous
two-way teletype communication during periods when the data tapes were not
being transmitted. In addition t o th i s combination communication-data l ink
between the receiving s ta t ioss and the BRL computer, a standard half duplex
teletype l ink t i ed a l l stations together i n parallel . A separate 100-word-
per-minute teletype l i ne linked the Laboratories1 canputing center w i t h the
National Space Surveillance Control Center, Spacetrack, New Bedford,
Wssachusetts.
During the Vanguard project, another half-duplex teletype l ink existed
between the computer center and the Naval Reaearch Laboratory in Washington
D. C . The tracking station at AFG and the BRL computing center each had a
commercial half-duplex teletype machine which could be connected on c a l l t o
other similar machines throughout the nation. It was necessary t o operate
w i t h t h i s variety of communication equipent because of the large number of
agencies involved in launching the various types of s a t e l l i t e s and probes
which were t o be tracked. Figure 27 shows the data transmission and
communication network.
A series of t e s t s were run on the data links t o determine the accuracy
of bata transmission. Both Teledata and teletype equipments were adjusted
t o optimum performance and a punched tape with a known code format contain-
ing 4000 data b i t s was transmitted from BRL t o each receiving station.
Then, over a period of several days, t h i s t e s t tape was transmitted back
t o BRL from the stations a t selected times. A s t a t i s t i ca l record was made
of errors. The same tape was transmitted over the links without the
Teledata parity check. This resulted in roughly twice a s many errors as
were made by the Teledata machine. The teletype equipment also showed no
indication that an error had been made. With the Teledata equipment, an
LIKE MACHINES BALLISTIC RESEARCH LABS. COMMUNICATION-COMPUTER
CENTER
WHITE SANDS Fy NOTES . - .,.. .
1 INDICATES TELETYPE MACHINE ON 0 FULL IDUPLEXI LINE - 2 INDICATES TELEOATA TRANSCEIVER UNIT. 6, SWITCHED TO FULL DUPLEX LINE
'TRANSMITTING SITE
3 INDICATES MULTI-POINT TELETYPE 0 MACHINE ON HALF DUPLEX LINE - @INDICATES ON-CALL TELETYPE MACHINE
@INDICATES FULL-TIME LEASE TELETYPE MACHINE
ALL LINE!$ AND MACHINES ARE FOR 60 W.P.M. TRANSMISSION,EXCEPT AS NOTED. f ' F : FIGURE 27 DOPLOC DATA' TRANSMISSION - COMMUNICATION NETWORK
er ror caused the machine t o stop transmitting and thus permitted er ror
location and correction. Data transmitted over a 2000-mile l ink using
60-word-per-minute telegraph l ines showed an er ror ra te of only one par t
i n 60,000. This m e quite adequate f o r the interim DOPLCC system data
handling requirements. A more rapid r a t e of data handling was planned
f o r the f i n a l equipment.
g. Data Transmission Accuracy Tests. An analysis of data received
from tracking both active and passive s a t e l l i t e s showed an occasional R b S
s ca t t e r of as much as three cycles per second. This resulted i n some
suspicion of the performance of the phase-locked tracking f i l t e r . A ser ies
of tests were made at the Laboratories' tracking s tat ion on data simulating
those obtained on actual satellite passes. The test results showed a
random sca t t e r i n the data corresponding as closely as could be determined
t o the precision of measurement, thus remwing the suspicion from the
track* filter. The random sca t te r was thus at t r ibuted largely t o propa-
gation phencrmena. The satellite spin and ins t ab i l i t y of the satellite
transmitter i n active s a t e l l i t e s w i l l introduce a limited mount of data
sca t t e r also. Efforts t o analyze the data f o r systematic e r rors as w e l l
as random er rors reeulted in an increase i n the precision of the period
measurements by using a ten megacycle counting r a t e instead of the one
hundred kc rate. After an extended ser ies of tests using simulated data
mhed with noise and various tracking filter bandwidths it was determined
that the peak frequency sca t te r was 0.16 cycles per second and the nus
frequency e n o r was 0.052 cycles per second. Figure 28 is a block diagram
of the equipment used f o r the frequency measurement tests.
h. Data Samples. Examples of the types of recorded data are shown
inF igures 29 - 34. Figure 29 is a dual channel, s t r i p chart record of
data obtained on pass 140 of s a t e l l i t e 1960 ~ e l t a isco coverer XI) a t the
Forrest City, Arkansas, s ta t ion. Operation was i n the passive mode. The
upper record of the chart indicates frequency as a function of time and
shows three d i s t i n c t segments of received Doppler frequency. These seg-
ments correspond t o the times of passage o f ' t h e s a t e l l i t e through the beems
1 START -I E D c a w r 1 0 -
10 I C DIOITAL FREWEWCY - DHIITAL - - RDKAK'R COUNTER f a l U l C R svwTnESIZER
11
- - - I
. - 7 -
8 I
t 1 T I C vacua=* DRL COI
Y I - - - - rUCIU0 CLOCK - MIXER OEIIRATOR *
Faa.IR.t'l FlLlCl l m*YI -
cQl(nl0, START .oD(IRD
10 # D I O I T U D I O I T U P R l a l L R
m w u m STOP
FIGURE 28 BLOCK DIAGRAM OF EQUIPMENT FOR FREQUENCY MEASUREMENT TESTS .. ,
-
ONIVlAMVYY 'ON~OMS 9NlAOUd N 3 3 O M 3 8 V 1 V +GI 'A3M (m M3M3A03Sla) V 1 3 Z 61461 J O M 3 V t l l
S - N 3A113V J O Otl033M A3N3nO3MJ M 3 l d d O O (IT£ a~n6l!d
Figure 32b DOPPLER FREQIUENCY RECORD Of' ACTIVE S-N TRACK OF 195!9 KAPPA (DISCOVEIRER PII) R E V 7 AT ABERDEEN PROVING GROUND , MARYLAND
of the three receiving antennas. The s tep wave forms show the functioning
of the automatic search and lock-on system used f o r acquisit ion of the
ref lected signal. The frequency range of the automatic search equipment
was changed a f t e r s a t e l l i t e passage through one antenna beam, and search
over a higher frequency range continued before s a t e l l i t e a r r i v a l a t the
next antenna be&. The lower channel indicates signal strength data
received a t the time of s a t e l l i t e passage a t each antenna position. C a l i -
bration data f o r each channel are niarked a t the r igh t of the record, and
timing marks are indicated on the lower edge of the record.
Two types of recordings of Doppler data output are shown in Figure 30.
The first is the Doppler period output recorded i n binary code on standard
five-level punched paper teletype tape. The f i r s t data block contains
Greenwich Mean Time a t the beginning of the recording run. Subsequent data
blocks, recorded a t the rate of one per second, contain the Doppler period
measurement. The second type of recording i s the printed record obtained
from a Hewlett Packard, Model 560A d i g i t a l pr in ter . Two examples are
shown. Each printed l i n e contains Universal Time data and the Doppler
data a t p r in t out time. When frequency i s recarded the f i r s t slx d ig i t s
indicate time and the remaining f ive d i g i t s indicate Doppler frequency.
When high r e s o l u t i ~ period measurements are desired, only four d i g i t s of
time, minutes and seconds, a re printed, allowing the remaining d i g i t s t o
indicate the period measurements.
Figures 31 and 32 a and b a re typical examples of single and dual
channel s t r ip-chart records obtained a t the DOPLOC stat ions during a
tracking operation. The dual channel records contain signal strength data
and Doppler frequency data, each indicated a s a function of the . The
marks are on the lower edge of the chart . Calibration data fo r each
channel are recorded at the beginning of each record just pr ior t o the run.
In these Figures the calibration data were transcribed from the or iginal
record and placed near the pertinent section of the record fo r c la r i ty .
The single channel records a l so contain the Doppler frequency as a f u n c t i w
of time but allow higher frequency resolution because of the increased
chart width.
101
The two similar s e t s of data a re included t o show the d i s t i n c t
differences i n the records for two s a t e l l i t e s . Figures 31 a and b show
the data obtained during the tracking of 1959 Zeta isco coverer VI) which
psssed almost d i r ec t ly over the receiving s ta t ion , indicating a re la t ive ly
s teep frequency vs time curve (usually cal led the "St' curve). Figures
32 a and b show similar data recorded during tracking of 1959 Kappa
(Discoverer VII) indicating a lower slope of the "S" curve because the
s a t e l l i t e passage was a t some distance from the receiving s ta t ion .
Figure 33 shows the s t r i p char t data recorded during the ear ly portion
of f l i g h t of the BRL ionosphere probe (strongarm No. 1 ) on 10 November 1959.
Data were obtained f o r approximately 24 minutes and included information
over nearly the complete t r s jec tory .
Figure 34 i s an example of s t r i p c,hart data recorded during a rocket
launch. The Doppler data record ok'tained on 1958 Alpha ( ~ x p l o r e r I ) indi-
ca tes s tep l i k e s h i f t s i n frequency which correspond t o rocket acceleration
due t o f i r i n g of successi.ve stages. This type of data is very useful i n
determining the success of a rocket launching and the subsequent orbi t ing -. of a s a t e l l i t e .
Sample o rb i t a l parameters from the DOPLOC system are shown below.
The f i r s t s e t i s from passive tracking of revolution 124 of 1960 Delta
and the second is from act ive tracking revolution 34 of 1960 Garmns 2. The
data are compared with the published r e su l t s of Space Track, Bedford,
Mrtssachusetts.
1. 1960 Delta (Discoverer X I ) Rev. 124 (passive)
DOPLOC SPACETRACK DL-WRENCE (Statute Miles )
Semi maj. ax is 4115 4121 - 6
Eccentricity .0198 . 0177 .0021
Inclination 80. jiO 60. ioo 2i0
.Right Ascension of : . Ascending Node 207 .'joO 207 .2T0 .23O
'Psgument of Perigee 97-49 125.22 -27.73' .: 'Mean Anomaly a t Epoch 40 .17~
102
2. 1960 (hama 2 ( m e i t D) ~ e v . 34 (Active)
m m B P A C E W DIFFEI($RCE (atetuta mlse
Eccentricity .02U 0394 - ,0181
Inclination 51.40' 51.22' .18'
Right Asoansion of Ascending Node 267.21' . 266.75' 0.46'
Argument of Perigee 321.17' 277.15O 43.62O
Mean Anamaly a t Epoch lbi.07' . .
BRL R e p & a. L123 ~0ntsine 8 8- 0f the tracking 8ctivitieS of the BRL DO- etetioas f o r both active and psseive type tracking.
IV. THE FULL SCALE DOPLOC SYSmM CONCEPT
A. Introduction
Amendment No. 3 to ARPA order 8-58 provided funds for a supporting research program with a goal of arriving at reccomnendations and specifi- cations by July 1960, for an "ultimate" full scale Doppler type satellite detection and tracking system. The tentative conclusions of Systems
engineering studies were outlined to ARPA representatives at a IEeetinI3 at
BRL on 14 April 1959 and briefly described in BRL Technical Note NO. 1266, "An Approach to the Doppler Dark Satellite Detection ~roblem", by L. G. deBey.
This desciption did not attempt to justify the concept presented nor
to derive quantitative parameter specifications. In the interest of pub-
lishing the general concept at an early date, the result of a number of
studies which had been under way for some time were largely omitted. This
section of the technical summary report presents the proposed "ultimate"
DOFT.0.2 fence in somewhat greater detail and includes changes in design
parameters which resulted from continuing studies.
B. The Interim DOPLOC Complex
The interdm fence, which was designed and installed before the R & D
program was initiated, made use of essentially off-the-shelf items of
technical equipment and was intended to provide a capability of detecting
one square meter targets at msximum altitudes of 500 to 1000 miles. This
system employed antennas having east-west fan beams Of 8 x 76 degrees (16 db gain) directed to produce one vertical beam and two beams at eleva-
tion angles of 20°, one toward the north hind the other toward the south.
It can be shown that to meet the full ARPA design objectives the transmitted
power required with these antennas would be about 30 megawatts, assuming
parametric receiver front ends (1 db NF), cosmic noise level 3 db above . t 7 ,
receiver noise, 10-cps bandwidth, 10-db output signal-to-noise and an
operating frequency of 108 mc. Such power levels in a single,illuminator
installation would be difficult to obtain. Also several such installations
would be required to cover the space volume above the United States. This
situation led to an attempt to find a mesns of improving-the.gain of the
"~texhas and of using them more efficiently. -I.-< ff: 104
C. Design Objectives
Finn requirements of system performance or specifications were not
issued by ARPA during t h i s design period. Instead, a number of design
objectives were informally stated in various technical conferences dealing
with the dark s a t e l l i t e fence problem. These objectives, when viewed from
the atandpoint of a system employing Doppler techniques, gave r i s e t o
certain specific system requirements which more nearly represent specifi-
cations. The following tabulations contain lists of these objectives and
8pecFfic Doppler system requirements around which the f u l l scale Doppler
s a t e l l i t e passive detection and tracking system was planned.
1. Detection Range. A l l objects i n orbit between altitudes of 100
and 2000 statute miles. It is now believed that a l l objects having effec-
t ive reflection areas greater than 1 square meter can be tracked over the
range of 100 t o 3000 8. miles.
2. Target Size. A l l objects below 2000 s. miles whose effective radar
cross section is 0.1 square meter or larger.
3. Multiple Object Capability. System should be capable of handling
multiple objects. Maximum number not specified. Design objective i s
capacity fo r the order of 100 - 1000 objects i n orbit over the United States
a t one the.
4. Detection Probability. The system should not permit "sneak"
satellites t o pass; the fence should be t ight .
5. Period Calculations. Accurate t o within 1 - 5 seconds.
6. Orbit Inclinations. All.
7. Time of Arrival. 1 - 5 seconds.
8. Accuracy, Aximuth and Elevation. 0.1' t o 0.5'. Resolution
500 - 5000 feet .
9. The Required for Fi rs t Orbit Determination. Less than 20 minutes.
10. Identification. Indication of she and direction of mtion. Would
monitor eJniaslone snd fndlcate tumbling or 8tabFlity of f l ight .
iog
11. Adequate Data for correlation with Gown Orbits. . .
12. No counter-countermeasure proyisions except for extreme narrow
banding of receiving equipment.
Based on the use of Doppler techniques and the mathematical solution
required to derive orbital parameters from such data, the following Specific
system requirements are postulated:
1. A minimum of twelve Doppler frequency measurements must be obtained
for low altitude satellites and twenty-five or more frequency measurements
for medium to high altitude satellites.
2.. The system must track the satellite wer a segment; of its orbit
carresponding to at least 150 seconds of flight time for lar altitude satel-
lites and for 600 seconds of flight the for maximum altitude satellites.
This tracking need not be continuous, but may consist of a number of short
segments.
I). Target Size
The objectives outlined above werenot based on a f o m l investigation
of & potential threat created'by the technical feasibility of orbiting. . .
.&k saieliiies. Eather they constitute good guesses as io ine pro-r
order of magnitude. Doubt exists that the value of the minimum detectable
target size is realistic. It is difficult to imagine that a satellite
about 20 inches in diemeter, which would have an effective reflection cross
section area of 0.1 square meters at 100 mc if uncamouflaged, would' pose a
serious threat, from the reconnaissance point of view, at an altitude of
2000 miles. It is conceivable, however, that satellites of larger physical
size could be camouflaged to make them appear as 0.1 aquare meter targets,
the degree to which this can be accomplished being dependent upon size,
camouflage weight, frequency-bandwidth considerations, effect on flight
performance and economic factors. Sincearbitrary specifications of a tar-
get %wo?or three times s&llep in effective cross section than that which
can.'b& expected to perform. a useful reconnaissance function would require i,: 2 2 2
antenna, gains or transmitter pover larger by these same fac tors (involving
cost increases of millions of dol la rs ) an econolnically sound proposal f o r
an "ultimate" system could not be assumed. Thus an investigation t o a i d
i n establishing a r e a l i s t i c t a rge t size was deemed t o be required. Such a
study was in i t i a t ed within the Bal l i s t i c Research Laboratories but did not
progress suf f ic ien t ly t o jus t i fy a report at the closing of the project.
E. Antennas
A number of beam configurations and combinations were studied and
discarded because they did not appear t o meet a l l of the system requirements.
Fixed fan beams spaced every f e w degrees would provide enough data points
but would require large numbers of antenna arrays and independent high
powered driving transmitters w i t h attendant high cost. High gain pencil
beams, scanning to give adequate volume coverage, would reduce the power of
individual transmitters, but would require large numbers of cost ly arrays
and transmitters in order t o provide a t i g h t fence (each pencil beam trans-
mit ter receiver would have t o t rack a given s a t e l l i t e t o obtain enough data,
thus leaving the fence "open" f o r other s a t e l l i t e s ) . Furthermore, w i t h a
b i - s t a t i c system i n which transmitter and receiver pencil beams emanate
f rom different surface locations, the beams in tersec t only at a small volume
or c e l l i n space a t any instant . To provide the required c-lege coverage,
one beam would be required t o acan the length of the other w h F l e it waa
scanning, and thus would use too much time, permitting other areas to go
unmapped, or s a t e l l i t e s t o cross other areas undetected.
As a result of the above considerations it was decided that a fan
shaped beam was required. The dimensions of the beam were then t o be
determined. Because of the curvature of the ear th it is not possible
t o have more than one transmitter and one receiver scan a given area i n
a coplanar manner such that the receiver w i l l always see all objects
illuminated by the transmitter. To scan even one pa i r of antennas together
requires tha t the scan axis be the chord of the ear th c i r c l e between the
two antenna s tat ions. The wide dimension of the fan i s determined prima-
r i l y by %he distance between the two stat ions, which in . turn i s dictated
by the minimum radiated power required t o detect a minimum size target a t
2000 t o 3000 miles al t i tude. The dimensions of the United States,
indicated that a base l ine of approximately 1000 miles was about correct.
A t greater distances, low al t i tude sa te l l i t e s would not be seen a t a l l
points between the stations, while for smaller separations of stations,
ful l a d ~ t a g e would not be taken of available power. Early e0timateS
indicated tha t the beam dimensions should be about 60' by lo. Further
study resulted i n the adoption of a 60 x 1 .4 degree fan beam, scanned
from horizon t o horizon.
F. Signal Dynamics
The maximum Doppler frequency, using 108 mc for a s a t e l l i t e i n orbi t
a t 100 miles al t i tude approaching a DOPLOC complex i n a plane of a great
c i rc le through the stations is 6000 cps. The m i m u m first derivative 2
f d of the f requency- th curve is about 100 cps . For a circular orbi t 2
a t an al t i tude of 2000 miles fd w i l l be approximately 9 cps . The minimum
information bandwidth required t o acccmodate Doppler signals w i t h these
rates of change of frequency is given by:
For h = 100 miles, Bi = 10 cps; while fo r h = 2000 miles, Bi = 3 Cps. . .
Since the received sibal is expected .to be less a t the higher alti tude,
it i s fortunate tha t a smaller information bandwidth is tolerable since it i s extremely important t o u t i l i ze the smallest practical bandwidth.
BRL experience has shown 2.5 cps t o be about the minimum useful bandwidth
for 108 mc. This lower lhit i s due t o phase pertubations introduced by
the antennas and the propagating medium.
The maxFmum signal processing or integrating time is given by: . i '-
Ti ranges from 0.1 second when 8 10 cpe f i l t e r is used t o 0.33 seconds 3 4
when a 3 cps f i l t e r is used. Gabor and Woodward derive an expression
f o r the lincertainty of frequency measurement.
where s is the power signal-to-noise r a t io . For a phase locked tracking
filter of the type wed in the DOPLCC fence the minimum usable signal-to-
noise r a t i o has been experimentally shown t o be about 4 db. Since other
types of narrow band frequency measuring equipment6 might require a greater
signal-to-noise r a t io , a safety margin was provided by assuming a required
output S/N = 7 db. The frequency uncertainty (Af ) i s thus dependent upon
integration time assumed and w i l l be e i the r 1.4 cps or 4.5 cps f o r integra-
t ion times of 0.33 and 0.1 seconds, respectively. The r e su l t of o rb i t
computations containing random er rors of frequency measurement of these
same general magnitudes show t h a t the accuracy of the solution i s not
seriously degraded by such er rors .
Shorter integration times result i n higher noise-to-signal ra t ios ,
deter iorate the accuracy of frequency measurement and reduce detection
range. A heavy penalty is paid i n terms of t o t a l transmitter power required
by even a twofold increase in bandwidth. The minimum signal duration
(integration time) cannot be appreciably l e s s than 0.1 seconds. Conversely
an increase i n the integration time above the value of 0.33 seconds would
reduce the bandwidth below 3 cps.
For the scanning antenna system proposed, the maximum signal processing
time i s equivalent t o the time required f o r the narrow dimension of the
scanning fan beam t o pass over the target , the time on target ( tT ) and i s
equal to:
Beamwidth (Q t~ = Scan Angle Veloci!y (wl
Conversely f o r a given tT, the beemwidth is:
The angular scan velocity i s determined by the number of data points required
and the total time the target is within tracking range. These factors vary
as functions of target altitude. At low altitude the slope of the Doppler
=?&we is high P Q ~ fe~er &t= pojnts a_-* .reguireB. Fewer than six points On
he Doppler curve will not yield a solution, even in theory. Experience has
shown 12 - 15 points to be adequate. With close approach satellites at high
altitudes, the slope is low and the "S" curve has much less character,
requiring more data points to obtain a solution.
The time the target is within tracking range, T, is summarized in
Table I, and ranges from 192 to 1000 seconds. The angular scan rate, W, is
given by
where ND = number of data points required. @ = scan angular coverage.
The lowest value of total time for tracking, T, must be used since
this will determine the maximum scan rate. Thus for:
= 160'
. . - -0 I w = l2 160 = LLJ second 192
One scan period is thus Ts = @ = * = W 10 16 seconds. The number of data points which can be obtained for targets at higher
t altitudes may be obtained through the use of the expression ND = - .
Ts Representative values of ND for selected altitudes are given in table IV - 1
MINIMUM MJMBER OF DATA POINTS VERSU8 ALTITUDE*
Altitude V S T Number Miles W S ~ C . Miles Sec . Data Points
*For circular polar o ~ b i t
V = Sate l l i te velocity a t al t i tude specified.
S = Length of observed s a t e l l i t e path in antenna coverage volume.
T = Total time during which s a t e l l i t e may be detected.
0 . Antenna Gain
The minimum effective narrw dimension of the composite fan beam
eyetem, camprised of two coplana~ fan beams, can now be determined by
the relation:
0 = W t T
The minimum value of 0 w i l l be obtained when tT is smallest. The
time on target, tT, was shown previously t o have values between 0.1 and
0.33 seconds depending on the beamwidth required by signal dynamics.
Thus the minimum effective beamwidth is:
0 = 1 0 ~ 0 . 1 = ldegree
Although larger values of tT would allow the use of greater beamwidths
the resultant lower gain of the antenna system would force the uee of
higher transmitter' powers or would reduce the output S/N rat io. On the
other hand, the fac t tha t the reciprocal of the narrowest bandwidth
(3 cpe) ' a l lws an integration time as great a s 0.33 seconds does not
mean that t h i s f u l l time must be used unless the lowest value of frequency
uncertainty must be achieved. It has been shown t h a t the uncertainty
equivalent t o integration times of 0 .1 seconds is adequate for the dark
s a t e l l i t e probIem. It is thus advantapeous i n establishing the remaining
system parameters t o assume 0 = 1 degree (effect ive) . The effect ive value
of Q i s stressed intentionally. Each of the two coplanar beam introduces
a 3 db pat tern fac tor at the design beamwidth and the t o t a l loss due t o
pat tern factor i s 6 db. In order t o achieve an effect ive beamwidth of one
degree at the 3 db point the individual beams must each introduce a pattern
loss of only 1.5 db at a beamwidth of 1.0 degree. A n t e m s having 1.4
degree beamwidth at the three db points w i l l y ie ld an effect ive system beam-
width of 1.0 degree. The fan beams required would thus have cross sections
of about 1.4 x 60 degrees a t the 3-db The gain of an antenna i n
terms of i t s beam dimensions i s given conservatively by:
It should be noted tha t economic signal dynamic, and data quantity
considerations determine the minimum tolerable beam dimensions and hence
the maximum antenna gain which may be used. The proposed DOPLOC s a t e l l i t e
fence was thus an antenna gain l imited syatem. This fac tor had an impor-
t a n t bearing on the choice of optimum operating frequeccy.
Having determined the anten,m beamwidths and gain, the next problem
was t o f ind a design which would be economically and physically prac t ica l
i n s i ze and cost, and which would pro-?ide f o r the necessary angular scan
without serious degradation of the beam ohape. This problem resulted i n
an extensive engineering stud'j both by BRL and by prospective contractors
who were requested t o submit bids on a scaled down model of the scanning
antennas and transmitter for the "ultimate" DOPLOC system. On f i rs t
consideration it appearea tha t a Wullenweber type antenna design mounted
with i t s normal diameter i n a v e r t i c a l plane would provide fo r the necessary
scan:operation. The f i r s t idea was t o feed the various trans.ndtting elements .?"i'l.
? 1 112
of the WuXIenmbar antem fram cammutated low power circuits in such a manner that the r e ~ u i r e d number of antenna ehmente was fed eegwntially
around tba rim of tho aataana. Further etudy ah& h o w that uhen the
bsclpre from tha Wullenweber antanne. are tilted off the exis of the antenna
eye- and than ecanned, the beam breaks down into a conical shape and will
not perform the required scan at low angles. This subject will be treated
in more detail in connection with the proposal submitted to ARPA. Suffice
it to say here that satisfactory antenna designa were derived.
B. Received Signal Power Required. . .
The minimum usable received signal power which will provide suitable
DOPLOC data is determined by the following factors: 1) External noise,
2) Internal noise, 3) Bandwidth, 4) Output signal-to-noise ratio. These
factors are discussed in relation to the "ultimate" DOPLOC system.
1. External Noise. Noise sources external to the DOPLOC system are
5 expected to provide the major noise contributions for a carrier frequency
in the range of 100 to 150 mc. These sources are extraterrestrial such as
cosmic and galactic as well as of local origin such as automobile ignition,
diathermy or other man made interference. Cosmic noises from the galactic
plane and extragalactic regions are the most continuous sources of extra-
terrestrial noise, with occasional periods of solar noise when the sun is
excited. High intensity, discrete radio star cosmic noise sources, most
of which have angular dimensions less than lo will produce a strong pulse
of noise when the high gain receiving fan beam sweeps past them. These
noises did not prove to be serious since they could at worst only interfere
with a few data points.
The amount of extraterrestrial noise available at the input to the
receiver'is a function of the antenna pattern width and the region of the
sky in the beam at the moment. As the beam sweeps across the galactic
sector which is about 15 to 30 degrees wide it crosses the region of highest
cosmic noise density. Therefore it was necessary to design the system to
operate in the presence of this noise. The relation between the amount of
noise picked up and the antenna gain and beemwidth ie a camplex function of
the orientation of the antenna and its gain level. When an antenna is
oriented toward the galactic plane, its noise output will initially incream
as its beam width is decreased. However, a limit ia resched where a further
declrctaee in antenna beamwidth (corresponding to an increaee in gain with a
,point source) will not appreciably increase the comic noise received. mi8 limit is between 10 and 15-db antenna gain. Whan an antenna is directed of?
the galactic plane and the noise is uniformly distributed in space, the comic
noise received is independent of antenna beamwidth or gain. The proposed
1.4 x 60 degree antenna with 25-db gain is considerably over the 15 db gain limit for increasing cosmic noise. The noise measurements made with the
interim DOPMC antennas, which had 16-17 db gain, should apply directly. Tlr
cosmic noise measurements made with the intertm DOPLOC center antennas shoved
a maximum noise temperature of less than 1160' K, 99 per cent of the time. This 'temperature corresponds to a noise power of 1.6 x watts per cps
bandwidth.
Atmospheric and local noise sources that have been observed include
lightning, ignition, power lines and electrical equipment. The noise from
these sources is generally of the impulse type, consisting of vary high
intensity, short duration pulses that contain very little energy and are
not expected to seriously affect tracking for an appreciable portion of time.
Sand static can also be a problem in desert areas where the antennas are
exposed to blowing sand.
2. Internal Noise. A preamplifier with a 2-db noise figure is attain-
able in the 100 to 150 mc region with vacuum tubes. The internal noise for
a 2-db noise figure is 2.4 x watts per cps bandwidth. This does not
include the source resistance noise which for the DOPLOC system is the
antenna with its associated noise temperature. The botal of the antenna
noise plus the receiver noise is, 1.6 x + 0.24 x = 1.84 x
watts per cps bandwidth. . .
3. Bandwidth. The bandwidth required in the receiving equigaent is
dependent on both the orbital parameters of the satellite and on the design
of thel'receiver. The method and amount of bandwidth limitation prior to
f i na l detection is also &pen&nt on the characteri.8tice of tbe detection
eyetam wed i n the receivar. Ths RF sectian of a WPMC raeaiver be
relatively conventional. Tho ruepame ehwld be Unsrv 111th input signal
for the range of signals aad noise which must be received. !the receiver should be free of regemratica and oscillation tendencies, sud its phacre characteristics should be relatively unaPfected by variation of amplifi-
cation. The output bandwidth of the IF amplifier should pass the maximum
Doppler frequency, and should preferably be adjustable t o quite narrow
values when it i s possible t o eliminate the noise by narrow banding tech-
niques. The interim DOPLCC receivers used about 15-kc bandwidth i n the
i.f. amplifier section, but through the uee of phase-locked tracking f i l t e r s
i n tbe output circuits obtained the effect of pre-detection bandwidth
reduction. The trackisg f i l ter operated a t bandwidthe of 2.5 to 10 cps.
I. Sya- Geometry and Coverage
The purpose of Wm original DOPUX prop- was t o provide sa t e l l i t e detection coverage t o the volume of space above the United States. No
counter measures were t o be ini t iated. The principal goal was to detect
and identify every object passing over the protected area w i t h a known
orbit, or t o compute a new orbit i f the pass were the f i r s t one for a given sate l l i te . The spacing between the transmitter antenna and the receiver
antenna was determined primarily by the altitude limits over which the
system vae t o openrte. Ae indicated above the two antennas were to either
physically rotate or t o cause their beams t o sweep around an axis formed
by an earth chord by drawing a straight l ine between the two stations i n the plane of the great circle containing the two stations. me range along
the earth's surface t o the sub-target point, for a target a t 100 miles al t i tude i s 420 miles when the angle between the horizon and the l ine of
sight t o the target i s 10 degrees. Based on these data it would seem t o be
necessary t o place the stations every 420 miles or less apsrt. Greater spacing would permit volumes of space in which 100 mile targets would not
be seen even though they were between the stations. However, because of tbe beem broadening and side lobe effects, and also because of the proximity
Of 100-mile objects t o the system, the distance between the stations can be
axfended t o 1000 miles without leaving serious dark areas for low flying objects. 115
It i s of course not sufficient marsly-to &teat t h ~ object. It
be tracked over a large enough s e m n t of i t s p t h t o pannit tbs orbit
parameters t o be determined. The time the target i# within tracking raw3 i s a function of its altitude and the volume covarsge of the tracking
system. The volume coverage, conversely, can be determined by the minimum
permissible time the target must be tracked a t a given altitude. For
example a t an al t i tude of 2000 miles th i s minimum tim,e for sa te l l i t es in
circular orbits is 610 seconds while a t 100 miles this time is about 150
seconds. A t the higher al t i tude the sa te l l i t e velocity i s about 4 miles
per second and the distance travelled along the orbit during the 610 seconds
i s 2440 miles. The bi-stat ic DOPLOC system m u s t then have a capability Of
tracking the target along th i s segment of its p t h .
Figure 35 shows the transverse coverage of the nodding or scanning
antenna. The coverage radius (2237 miles) intersects the 2000-mile altitude
surface a t points 2440 miles apart thus providing the capebility of tracking
the target along the required segment of the i rb i t . Table IV - 1, page 111,
shows the length of orbit segment for various values of altitudes between
100 and 2000 miles and the times required fo r sa te l l i t es i n circular orbits
t o traverse these diatances.
The proposed patterns and coverage of two antennas of the nodding
type in a plane containing the base line, are shown in Figure 36. The area
of coverage i s approximately bounded by the earth's radi i extended through
the two stations, these radii lying in the plane of coverage shown in the
previous figure. One range, transmitter t o target, has been determined
above as being 2237 miles maximum. The other range, target t o receiver, i s
approximately 2450 miles maximum.
The to t a l volume of space swept out by one of these nodding antenna
systems, and thus the volume coverage of one transmitter receiver pair, i s
shorn in Figure 37. Total coverage would depend ugon the number of tram- m i t t e r receiver pairs employed and upon their deployment. . .
Figure 38 i l lus t ra tes a possible location of stations within the United
S'tetes. A pair of transmitters, TI and T2 would be located a t about 40 i: " r degrees north lat i tude and 98 degrees west longitude. Receivers R1 and R2
VOLUME COVERAGE OF ONE f RANSMITTER-RECEIVER PAIR' FIG. 37
TENTATIVE TRACKING STATION LOCATIOF(S
FIG 38
81 -STATIC RADAR
ATTENUATION OR VS RANGE' SUM (RI + R2 )
would be located about 1000 miles e a s t andwest along the 40 degree l a t i t ude
c i r c l e . In principle, these s ta t ions would yield suf f ic ien t cwerage t o
permit obtaining enough data from which o rb i t s could be computed. In general
only data from one base l i n e would be received f o r each s a t e l l i t e pass f o r
s a t e l l i t e s i n near polar o rb i t s and from two base l i nes for satellites i n
low incl inat ion orb i t s . The improvement i n strength of the orb i t solution
and the additional course angle data which could be obtained through the use
of additional s ta t ion pairs with similar nodding antennas, makes it appear
desirable t o add two (or possibly more) base l i n e s t o the network. Error
propagation considerations make it desirable t o i n s t a l l these base l i nes a t
r i gh t angles t o the or iginal base l i ne . Figure 38 shows base l i n e s R 3 - =3 and R4 - T located about 125 miles eas t and west of the location of trans- 4 mit ters T1 - T2. This i s not necessarily the optimum location i n the east-
west direct ion but i s approximately correct. Targets passing through the
coverage volume of the system would be intercepted by the nodding beams of
one or more traiismitter-receiver pairs .
During the time on ta rge t interval , tT = 0.1 sec., the system must
Locate the Doppler frequency in the range of frequencies from 2 kc t o 1 4 kc
and measure the frequency t o an accuracy determined by the system output
signal-to-noise r a t i o and t The interim DOPLCC automatic lock-on and T'
tracking f i l t e r required a minimum of 1.6 seconds t o accomplish t h i s sequence
of operations. It i s thus not adequate fo r the scanned beam system. The
circulat ing memory f i l t e r (CMF) developed fo r the ORDIR radar by Columbia
University i s well sui ted t o t h i s application. The CMF w i l l be discussed i n
more d e t a i l under the section of t h i s report devoted t o the research program.
Other approaches using comb f i l t e r s were investigated within BRL and an
experimental comb f i l t e r was developed. This w i l l a l so be discussed i n more
d e t a i l elsewhere.
1. Trailsmitter Power Requirements. The t o t a l f r e e space attenuation
f o r the b i - s t a t i c radar case i s given by:
r Attenuation (db ) r
This function is giMn in Figure 58 a for the case where R1 and R;! are the variables and: Qr = Ot a 25ab (1.4' x 60' beauwidth)
x = 9.12 ft (108 mc)
a = 1 BQ. ft. (0.1 2 ) In the figure for the case of R1 = 2237 miles and R;! = 2450 miles the
rat io: V p t = 1.7 x or -247db = F(A).
'r The transmitter power required is given by: Pt = i q ~
The received power required has previously been determined as being - -
2.8 x 10-l9 watts. Hence P = 2.8 x 10'19 6 = 1.65 x 10 watts.
1.7 lo-*?
Although th i s i s a high CW power level it can be achieved wi tha t e -o f -
the-art components ei ther by using a multiplicity of power amplifiers, each
feeding one or more elements in the transmitter array, or by paralleling
several high power transmitter tubes.
V. SCAI8 MODEL S-G REAM SYSTEM PROPOSAL
The full scale or improved DOPLOC system as proposed f i r s t i n BFU
borandum Report No. 1220 is based on the use of the bi -s ta t ic , continuous
wave ref lect ion principle and u t i l i z e s f o r the basic detection and tracking
u n i t a high power radio transmitter and a high gain, narrow-band radio
receiver, each coupled t o high gain antennas driven through synchronizing
equipment. Thin, fan shaped, coplanar beams are scanned about the axis
formed by a s t r a igh t l i n e between the s tat ions t o cover a space volume
consisting of a half cylinder whose length i s equal t o the distance between
the s ta t ions and whose radius i s 2000 miles. Computed parameters, based on
the radar equation and ver i f ied by experience with the interim fence equip- 2
ment, indicate that t h i s system would detect and track objects having 0.1 m e f fec t ive cross sectional area over an a l t i tude range of approximately 100
t o 2000 miles when a two-minion-watt output power i s used at the trsnsmitting
antenna. Engineering considerations indicated tha t such a system was feasible
with exis t ing state-of-the-art techniques and equipment. The basic cost of
auch a system, while reasonable i n comparison t o the coat of other systems
of similar proportions, represented a large investment in both funds and
engineering e f fo r t . Therefore since one of the goals of the reoriented DOPLOC
program was the demonstration of the f e a s i b i l i t y of the concept of determining
o r b i t parameters by passive continuous wave Doppler techniques, it was proposed
t o instrument two 700-mile base l ines with scaled-down models of the proposed
ultimate equipment.
One of the base l ines was chosen t o u t i l i z e the Aberdeen Proving Ground
s t a t ion as the principal research receiving s tat ion, using exis t ing equipment
augmented with equipment from White Sands M s s i l e Range receiving s ta t ion .
A transmitting s ta t ion having two 100-KW transmitters and ecanning antennae
was t o be located a t Camp Blanding near Jacksonville, Florida. With one
antenna beamed toward Aberdeen, Maryland, and one toward Forrest City, Arkan-
eas, two base l ines a t approximately r igh t angles t o each other would provide
geometrically strong data f o r o rb i t computation and would have permitted
u t i l i z i n g most effect ively the technical s k i l l of the BRL persmnel without
extensive travel and l o s t time i n se t t ing up experiments a t remote f i e l d r - ~ . L
rzt? 124
u t i &&;a..t
stations. As an alternate location it was proposed that the scale mcdel
systam of two base lines could be located to fill the gap in the fence in
the southeastern section of the United States between Forrest ~ity,"&kansas
and WSMR.
The scaled d m system would have employed 100-KW radio transmitters,
4 x 40 degree scanning beam antennas, and base lines of approximately 700 miles. Low altitude coverage would have extended down to approximately
125 miles above the surface of the earth, while the high altitude 1 M t
would have been about 650 miles. Such a system would have provided adequate
detection capability to thoroughly explore the computational and data han-
dling techniques and would have permitted the test and development of any
improved techniques found necessary to adopt the DOPLCC concept to the
scanning beam methods.
A scanning system of W e type would have offered significant improve-
ment over the flxed beam system in detection capability. Instead of relying
on having all factors affecting detection suitable during a short period of
time that a satellite falls within a severely restricted angular segment of
btection volume, this system would have continuourtly scanned a 160' sector
of space volume and have made repeated observations of every object in the
scanned volume. With the a priori knowledge that satellites must follow
closely the classical laws of orbital motion, the multiplicity of DOPLCC
observations on each satellite pass greatly simplifies the problem of
distinguishing between satellites and other sources of reflected signals
such as meteor trails, aircraft, ionospheric clouds, etc. In addition, it
is this multiplicity of observations over a length of arc during a single
par which gives the DOPLOC system the capability of determining orbital
pammters vithout waiting several hours for the satellite to complete one
or more revolutions along its orbital path.
A. Request for Contractor Proposals
BRL ~pecificition No. 16-9-59, entitled "Supplemental Information to Oral Briefing Pertaining to Requirements for Scsnning Antenna ~ y s t k " ,
was prepared and an open briefing was held by BRL and the local procurement
o f l i c e on 16 September 1959 t o acquaint prospective bidders with the a
requiremanta of the propoaed scaled down DO- system and to s o l i c i t bid9
covering the furnishing and in s t a l l a t ion of one and two baseline systema 8s .
had been proposed t o ARPA. Tb basic bid covered the supplying Of the C W Q
Blanding - Aberdeen base l ine , less the receiving equipment. !This included
a prefabricated instrument she l te r at Camp Blending, a t r a n m i t t e r , two
scanning beam antennae, and the antenna synchro~lieing equipwnt. As an
alternate bid, proposals were a l so requested covering the basic bid, and
i n addition, a second transmitter housed in the common instrument she l te r
and two scanning antennas with synchronizing equipment f o r the Camp Blanding - Forrest City base l i ne . Approximately 25 contractor f a c i l i t i e s were repre-
sented a t the briefing. !i%elve proposals, some of which were joint e f fo r t s ,
were received. Prices ranged fram l e s s than a million dol la rs per base l i n e
t o approximately f i v e million dol la rs per base l ine . In general the lowest
priced systems proposed using mechanically ro ta t ing l i nea r antenna arrays.
Other system proposals were based on re f lec t ing type, horn fed antennas
scanned by organ pipe scanners, o r modified Wullenweber type antennas fed
with organ pipe scanners or electronic scanning systems. Several interest ing
r e su l t s came from the analysis of the bids received.
Six of the proposals contained antenna designs based on scanning the
fan beam by ro ta t ing the antenna array s t ruc ture , Seven proposals contained
antenna designs based on stationary antenna s t ructures with the beam scanned
by electronic methods. The proposals were requested f o r two frequencies,
150 mc and 216 mc since these two frequencies had been suggested a s possible-
assignments by the FCC. It was a l so considered of value t o know the cost of
the system f o r what might be considered the lowest and highest p rac t ica l
operational frequencies. A t the lower frequency (150 mc) the tracking range
is greater f o r a given antenna power but the antenna s i ze is la rger . A t the
higher frequency (2i6 mc) the antenna s i ze is reduced by one th i rd but the
tracking range is reduced and twice the transmitting power i s required
compared with t h a t required a t 150 mc t o maintain a 2000-mile a l t i tude
tracking capabili ty. Soon a f t e r the contractor br ief ing the FCC authorized
the use of 150.79 mc f o r DOPLOC dark s a t e l l i t e tracking. Therefore, eubsequant
plaMing was based on the use of the 150.79-mc frequency. It is of in t e res t
t o note that the 216-mc system antenna and 100-KW transmitter cost estimates
were only between f i v e and eight per cent lower thaa f o r the 150-mc systems.
Considering that increased transmitter power would be requlred at the higher
frequency t o obtain the desired range, the higher frequency system would
actually be more expensive t o i n s t a l l .
During the analysis of the proposals, an important disclosure was made
regarding w h a t appears t o be a theoret ical l imi t t o the use of a c i rcu lar
array or Wullenweber type antenna where the wide dimension of the fan beam
muet be t i l t e d away from the plane of the array. The e f fec t is that the
t i l t e d swept beam degenerates a t low elevation angles. When this e f fec t
was cal led t o the a t ten t ion of the contractors who had proposed using a
circular scanned array it stimulated a rather extensive theoret ical study
of t h i s problem. The general conclusion i s t h a t f o r a 4 x 40 degree beam
such as was required f o r the scaled down model the pattern degeneration
could be tolerated. However f o r 1 x 60 degree beams an extreme amount of
compensation would be required or the beam would degenerate from a l i n e in to
8 "bow t i e " e f f ec t which would destroy its low angle coverage. Since one of
the conditions of the specification required that the design be capable of
scaling up t o the two megawatt 1 x 60 degree, 1000-mile base l i n e system,
the circular array antenna proposals were excluded from the f i n a l consider-
a t ion. In general these designs were chwacterized by considerable complex-
i t y , high cost and some new component development requirements.
Ultimately, the cost and available funds r ea l ly determined the choice
of antenna system proposed f o r the scaled down model. The cheapest antenna
consisted of a linear array of Yagi antennas mounted on a rotat ing boom
73 feet i n length. Sixteen Y a g i s , each having seven elements, were used t o
form the proposed 4 x 40 degree beam. Two arrays, pointed 120 degrees
apart , were mounted on the boom t o provide continuous scanning of the fan
beam with continuous rotation of the boom. One array was t o be switched t o
the feed l ine when i ts beam was ten degrees above the horizon, then, a f t e r
scanning t o within 10 degrees of the opposite horizon, it was switched off
and the other array switched on. Four scans par minute, all in the same direction, were to be made by a rotational rate of two revolutions per
minute. A l l polarizations (circular right or left hand and linear vertical
or horizontal) were to be Drovided by the use of crossed Yagi elements.
The desired polarization was to be selectable by changing feed connections
to the Yagi's.
Synchronization of the transmitting and receiving antenna beams was to
be accomplished by a servo motor controlled drive system which was synchro-
nized to a precision frequency standard at each station. Synchronization
was to be maintained under winds up to 80 m/hr. There was some question as
to whether a rotating boom type of antenna could be scaled up to a 1 x 60 degree beam capable of handling two megawatts of power. One contractor did
submit a mechanical design for such an antenna.
A more expensive antenna, but one which seemed more m i l y adaptable
to scaling to the ultimate system requirements consisted of two complex
curved surfaces arranged back to back. Each curved surface was illuminated
by a set of horns to produce an eighty-degree scan from ten degrees above
the horizon to vertical or just beyond vertical. The horns were fed in
sequence from an organ plpe scanner. h e set of horns and its reflector
scanned from 10 degrees above the horizon up to vertical. The power would
then be fed to the other set of horns and the scan continued from the vertical
down to 10 degrees above the opposite horizon. This constituted a basically
new approach where instead of sequentially feeding a large number of radiating
elements located on the periphery of a circle, only a relatively small number
of elements are fed and the c m d reflecting surfaces form the narrow beam.
The resultant reduction in complexity is considerable and a much desired
degree of versatility is gained in'control of beam pattern shape and polari-
zation. Figure 39 shows a drawing of the propoeed 8ntenna design.
. + This reflecting type antenna appears to be the best design for meeting
the DOPLCC requirements. The scanner provides accurate angular information
relative to the position of the beam once the system has been calibrated.
It also has the ability to be scaled to the 1 x 60 degree beam with the two
"THE INFORMATION HEREON IS PROPRIETARY AND THE FACT THAT A PERSON MAY BE GIVEN ACCESS THERETO UNDER GOVERNMENT PROCEDURES IS NOT DEEMED A LICENSE TO USE SAME WITHOUT THE EXPRESS WRITTEN APPROVAL OF MELPAR."
Figure 39 Proposed Antenna Design Melpar, Inc.
megawatts parer handling capabilities. The organ pipe scanner and horn
feed system have been designed and tested out in other installations.
Therefore the engineering information is available. In view of the larger
cost of the horn fed reflecting antenna and the desire to obtain as much
information as was practical f r m the scale model installation, one pair
of organ pipe scanned antennas was recommended for the Aberdeen - Camp Blanding base line and one pair of mechanically rotated Yagi arrays was
recommended for the Camp Blanding - Forrest City base line. In fact, when
the future of the program was in doubt,several research programs ranging in
amount of effort from the installation of a single base line in accordance
with the lowest priced proposal to two base lines of the scanned reflector
antenna system at several times the minimum cost were proposed to ARPA.
All of these proposed programs were directed toward arriving at a set of
specifications for an ultimate or improved DOPLOC system. As a byproduct
of these programs, the application of DOPLOC to the anti-satellite defense
problem was to be studied. Although tentative plans were made to install
the scaled down base lines and a site was obtained at Camp manding, Florida,
ARPA did not approve this pnrt of the program. The project was therefore
deactivated wlthout field testing the scanning beam concept.
V I . SuPPClRm 8mEB
A. Soenukrg Antsm Corn-'
. A study WM canduatod to btarmlne the number of r a t e l l i t e paeeee
.crorr orch of the two proposed bare l ime , Clrprp Bl8ndln&t - Aberdeen and
Can@ Bland- - Forreet City, atad th6 of tlrm that tha sa t e l l i t e
would be in tho antanas be- during each psre. ! b i r r was accaupliehed by
wing s mp of th. Mitad 8ta-8 a d mpremntiag tho 4 x 40 , O e antenna pettome a s two space volumer, each having a width of 800 mlles, a height
of b00 miles and a length equal t o the distance bstwwn the stations. Thls
distance for the Aberdsen - C a u g B l a n d i n g base l ine is 730 miles, while the
Farcast City - Cemg Blanding baee l ine 16 610 milee.
Ue- eatel$ihe predictlone furnished by Spce Track Control Center,
each satellite w e examined dvrLag the period indicatud and the to ta l
number. of w e e e croeeing sach base. l ine . vae recorded. 'The length of time-
in-bsam for each pees WEW calculated u t i l i t ing the sub-satellite trace and
sssllming a nominsl ea ta l l i t e velocity of five rmilee per second. A tabula-
t ion of reeulte 18 ehown in Table VI - 1 and a graphicel representation i e dieplayed in Figuree 40 t h r o w &5.
.. '5 '" u, ,O"N ",HI W l d O l X O L
'03 H31L131C 3h31nP i t l d v d Udva3 ti231131U 01. 0 1 C "*i
J
,., -9
. . . . . . . . - . . . . . . . . . .
. i::.;.j - : : 1 I : : : : ...
. . . . . I . :
. . . . . . . . . . . . . . J . . . .
Abe-n Proving Clround, lvsryland - Camp BLsnbbg, B l o r i d 8 (AW) Forrest City, Arkansas - Cemp Elanding, Florida (FCY)
Over lap-Area encompassed by both of ebove pa tbme.
2. Satel l i tes and Period Examfned
1958 Delta, 1 M a y 1959 - 1 Dectmbsr 1959 1958 Epsilon, 1 Ma3r 1959 - 23 October 1959 (Dew) 1959 Epsilon, 13 A m s t 1959 (Lsunch) - 23 Beptamber 1959 (Decay)
1959 Zeta, 19 Auguet 1959 (Launch) - 20 October 1959 (~ecay )
1959 Iota, U October 1959 - 1 December 1959
1959 Happa, 7 November 1959 (Launch) - 26 November 1959 (Decay)
3. Altitude Distribution APG FCY OVERLAP
Total Passes 814 750 447 Passes Below 401 Miles 400 377 221
Percentage of Total Below 401 Miles 49.1 50.3 49.4
4. Time-In-Beam Distribution ..;. !.
For APG base l ine 53.0 $ of passes.are i n the beam from 150 - 179 sece.
For FCY base l ine 54.6 $ of passes are i n the beam from 120 - 164 sece.
For overlap pattern 39.8 $ of pa'saee are i n the beam from
120 - 164 secs.
It can be seen from these data that, of the two proposed base lines,
the Aberdeen - Camp Blanding base l ine was the most promising, both fram
the standpoint of to ta l number of passes and from the tims-in-beam of these
passes. The long base l i ne accounts for the larger t o t a l number of passes,
while the geographical location of Aberdeen and Camp Blanding i s such that
a sa te l l i t e having an orbital inclination of fram 50 t o 90 degrees crosses
the pattern nearly diagonally and therefore, is within the beam a greater
length of time. The data also ahow that it w i l l be necessary t o decrease
the antenna scan time f r a 15 seconds to 7 1/2 seconds per scan t o provide
10 - 20 DOPLOC hi ts per pass for the majority of passes crossing the baee l ine . , ,
5 r. . ...
'.< .: 138.
B. Cosmic Noise and Interference a t 150.79 mc
Following the assignment by the Federal Communication Commiesion of
150.79 mc f o r the DOPLCC frequency a study was made of the radio in te r fe r -
ence occurring over the frequency band of 148 - 153 me. A standard dipole
antenna, a preamplifier, a sensi t ive receiver and a decibel meter were used
i n the study. Graphic recordings of signal leve l were a l so made. Taxi
service dispatch s ignals were observed at 149 and 153 mc. Except f o r cosmic
and receiver noise the band from 150 - 151 mc was clear . Calibrating the
db meter t o an a rb i t ra ry scale reading of 10 when the preamplifier was
terminated with 50 ohms showed a cosmic noise ranging from 0.5 t o 1.75 db
above the reference point. S ta r t ing a t 0800 hours loca l time the noise
gradually increased t o a peak at around 1200 hours and then decreased toward
evening. Using the same equipment and reference point but tuned t o 108 mc
the noise leve l ranged from 1.5 t o 3 db above the reference w i t h a similar
amplitude d is t r ibu t ion as a function of time between 0800 and 1630 hours.
No l o c a l interference was observed except when vehicles passed within about
200 yards of the antenna.
C. Circulating Memory F i l t e r
In order t o obtainthe maximum benefi t of narrow band technique8 i n the
interim system, phase-locked tracking f i l t e r s were ueed. These f i l t e r s
were supplied by the In te rs ta te Electronic Corporation, Anaheim, California.
The f i l t e r s were designated Phase Locked Tracking F i l t e r M d e l IV, and had
been developed by In te ra ta te f a r BRL under a aeries of research contracts.
When the scanningbeam antenna concept w a s developed it became apparent
t h a t the time on target would be qui te &ort and that a Large amount o f d a t a
could be generated by multiple targets. Only appraximately 0.1 second of
time would be available t o lock onto a signal and t o measure the frequency.
This type of performance seemed t o be beyond the capabi l i ty of any type
r i l t e r which operated on the principle of the In te rs ta te un i t .
The Electronics Research Laboratories of Columbia University were
known t o have developed a frequency estimator using coherent integration
detection methods f a r the CEDR radar. Contract DA-30-069-509--2748
U9
was therefore negotiated with Columbia University to cover a study of
the applicability of the ORDIR system techniques. to the tracking of passive
satellites. Final Report No. ~1157, No. 14 of the BRL DOPLOC Satellite Fence Series entitled "Summary of The Prelimimry Study of the Applica-
bility of the ORDIR System Techniques to the Tracking of Passive ~atellites",
Columbia University, School .of ~n~ineerin~, resulted from this study. !Phis
report contains the results of a study into the problems of coherent inte-
gration and frequency estimation in the DOPLOC CW Doppler earth satellite
tracking system.
Specifically the report discusses the incorporation of the coherent
signal proceeser known as the Circulating Memory Filter (cMF) into the
satellite tracking system. After the parameters of the system are reviewed
the theory of operation of the CMF is presented. The coherent integration
of signals and the resultant improvement in the signal-to-noise ratio are
discussed and the problems involved in distinguishing signals from noise by
means of a threshold test are investigated, all with particular reference
to the CMF. A technique for estimating the frequency of a sinusoidal input
to the CMF and an implementation of the technique are also investigated.
Finally the requirements imposed by the CMF on the remainder of the satel-
lite tracking system are discussed. These include requirements for the
receiver which furnishes the CMF input and for the data recording, trans-
mitting and processing equipment which makes use,of the CMF output.
The detailed specifications for the CMF recommended for use in the
satellite tracking system are extracted from the Columbia University report
into this report. so inciuded are specifications for a frequency esti- mator to accept the CMF output and furnish digital data to a computer and
for a receiver which would present acceptable input'algnals to the CMF. .. . ~ , ,,. See.' Section VI - D.
.- 1. The Digital Frequency Readout (DFR~. The basic video signal pr&c$uced in the CMF represents, in analog form, a frequency analysis of
the CMF input signal. The time positions of the peaks in the video wave
form are measures of the discrete frequencies present in the CMF input
ai&XL. The use of a threshold crossing device in the CMF results in a
convumion of the video analog data into a d iscre te form. More specif ical ly ,
a pulse w i l l be generated a t the instant the video increases above the appro-
p r i a t e threshold and another pulse w i l l be generated a t the ins tan t the
video decreases below this threshold. Such a pa i r of pulses w i l l occur fo r
each freqwncy component present a t the CMF input and of suf f ic ien t smpli-
tuie t o exceed %be h w c t i o n threshoid.
The object of tha DFR is to convert the data (frequency) pulses at the
output of the CMF un i t i n to binary numbers which r e p r e e t a unknown
frequencies. One method of accomplishing t h i s i s shown i n Figure 46. Enter-
ing on the l e f t of t h i s f igure is the output of the CM!? uni t . For the
purpose of t h i s explanation, it i s assumed t h a t three unknown frequencies
are present and aU..of these w i l l be read out.
The f i r s t pulse of a pair of cal ibrate pulses causes the t r a i n of clock
pulses t o appear a t l i nes la, 28 and 38 at the respective outputs of the
logic blocks. These pulse t r a i n s are applied through binary stages ( f l i p -
f l ops ) t o the counters i n sections 1, 2 and 3. The second ca l ibra te pulse
causes the t r a i n of clock pulses a t l i n e s l a , 28 and 3a t o be shif ted t o
l i n e s lb , 2b and Jb, respectively, thereby bypassing the binary stages.
The e f f ec t of t h i s i s t o divide the number of clock pulses exis t ing between
the pair of ca l ibra te pulses by two. Hence any count regis tered later on
a counter w i l l indicate the number of clock pulses occurring from the
average time ins tan t between the cal ibrate pulses t o the ins tan t when the
counter was stopped. The average time between threshold crossings is a
good measure of the applied frequency since t h i s method e s t h e t e s the time
of occurrence of the peak i n the video signal. After the second ca l ibra te
pulse has occurred, then, the clock pulses a re entering all three counters
along l i nes lb , 2b and 3bJ respectively. When the f i r s t data pulse of the
f i r s t pair of data pulses occurs, the clock pulses a re switched so tha t the
counter i s counting a t half its previous rate. The second pulse of the
f i r s t pa i r of data pulse6 causes the counter of section 1 t o stop. The
count thereby regis tered is the number of clock pulses between the center
of the cal ibrat ion pulses and the center of the f i r s t pa i r of data pulses.
When the counter of section 1 is stopped, the logic un i t of section 1
sends a s ignal t o the logic uni t of section 2 i n order t o prepare it f o r
responding t o the next pa i r of data pulses. The second pa i r of data
pulses then causes a count t o be regis tered on the counter of section 2
i n a manner sMla r t o that described above. The th i rd pair of data
pulses causes the counter of section 3 t o stop.
A clock pulse r a t e of about 10 mc i s reasonable since t h i s has a
period of 0 .1 ps, which corresponds t o a frequency quantization of about
2 CpS (using the conversion fac tor of 20 cps/ps). If a 10-mc clock r a t e
were used, the maximum count possible i n 1 m s i s 10,000. This requires a
1 4 b i t reg is te r . The implementation presented here i s only one of a
number of poss ib i l i t i e s and i s given a s an i l l u s t r a t i o n of the general
nature of the implementation problem.
2 . Digital Data Recording and Data Rates f o r CMF. In t h i s study it
was assumed t h a t the following data would be recorded whenever a "frequency
readout" occurred:
a . Time (of occurrence of a frequency readout).
b. Antenna beam elevation angle.
c. Echo frequency count (up t o three counts).
To determine the average data transmission r a t e it was assumed tha t :
1. On every scan of the antenna beam, one ta rge t readout occurs.
2. On every scan of the antenna beam, one meteor echo i s observed.
3. he meteor and ta rge t readouts d o not occur simultaneously.
Based on a reasonable f a l s e alarm probabili ty, the data r a t e derived has
an average r a t e of about 18 h i t s per second. A s an indication of chamel
capacity, a 7-channel teletype operating a t 65 wards per minute has a
capacity of about 18 h i t s per second. Thus the recording rate required
was not found t o be excessive.
3. CMF DOPLOC Receiver Specifications.
a . Receiver Bandwidths. The receiver pre-detection bandwidth should
be large enough t o accommodate the 24 kc Doppler frequency range, but not
so large a s t o enable extraneous echoes t o be processed. Although the CMT
provides large rejections for signals outside of its 12-kc input fI'equencY
range, such signals might produce undesirable inter-modulations or cause
temporary saturation of the CMF input circuitry. A 1 db pre-detection
bandwidth which is no larger than 20 kc and no smaller than 15 kc was spci-
fied for the 1.f. stage which feeds the CMF. The receiver pre-detection
band pa88 characteristics must be maintained flat over the entire 24-kc
Doppler frequency range. Since, the CMF rejection characteristics are
specified to be flat to - + 1 db (maximum) withln its 12-kc frequency corer- age, the receiver characteristics apould be roughly .. + .1 db to prevent further deterioration.
b. Frequency Scan. The receiver will scan the twenty-four kc Doppler
frequency range in two steps (12 kc per stage). Each frequency scan will
be actuated by a command pulse which is generated by the CW at the termi-
nation of its integration period.
c. Dynamic Range. The receiver must meet the following dynamic range
specifications:
1. For a noise free, sinusoidal input voltage of amplitude 55 dbYrms atany frequency in the 24-kc frequency coverage, all receiver output
spurious responses shall be more than 40 db (with respect to the input) below the output peak response.
2 For a noise free input signal consisting of two slnusoids each of
amplitude, 55 db nus and at different frequencies within the 24-kc coverage all receiver output spurious responses shall be more than 40 db (with respect to the input) below the output peak responae. Spurious responses
generated in the receiver outside the 1 kc channel containing the signal
must be kept 55 db below the signal peak.
d. Receiver Gain and Voltage Levels. As a compramise between CMF and
~QC,.,quirements and lose in detectability due to the addition of CMF and
receiver noise, the receiver gain was set at such a level that the sum of
the CW and receiver noise would attain a level of about lldb. This . . - 4 , . I
carresponde to a loss in &etectability of about 0.4 db and to a maximum CMF'AGC requirement of about 15 db, that is, to a maximum signal of about 70 db.
<
,?;...;;1,The db values quoted are relative to the usual reference level of CMF - - internal noise referred to the CMF input.
T%?3 '
e. Timing Pulses and Auxiliary Outputs. The receiver must also
supply (in binary form) all remaining information not supplied by the
CW that is either necessary for computing purposes (receiver oscillator
settings, absolute time, etc) or desirable to record (gain settings,
antenna positions etc ) . f. Autamatic Noise Level Control. During the antema.scan period,
cosmic noise fluctuations may cause the receiver noise level to vary by
several db. Such fluctuatians could cause a large increase in the system
false alarm probability and might therefore result in a temporary satu-
ration of the data transmission equipment. To control these fluctuations
an automatic level control technique which changes the gain setting only
when changes in the noise level have occurred is required. A parallel controlschene with the following design factors is recaamnended:
1. An accurately calibrated receiver gain bias characteristic of
sufficient dynamic range to allow correction for anticipated input noise
variations.
2. A gain stability in both the receiver and estimator cha~els
sufficient to maintain the above calibration.
The overall response of the DOPLOC receiver shall be such as to
supply the following to the input of the Circulating Memory Filter:
1. Carrier Frequency - 11 mc . 2. Bandwidth (1 db points). No -larger than 20 kc and no smaller than
15 kc.
3 F S Voltage Levels. 11 mv + 1 mv minimum detectable signal, - 560 mv + 6 mv maximum processed signal to 4 volts maximum output - signal.
4. Receiver Gain. The receiver shall have a gain such that the rms
value of its output noise (after AGC) shall be nominally 3.6 mv - + 0.5 mv i n 8 10-cps band.
5. Receiver Noise A N . The receiver shall have a noise actuated AQC
such that the receiver output noise remains constant to within
+ 6.2 db. -
6. Frequency Stability. All lkal oscillators shall have a stability
or be so recorded that the output frequency uncertainty is less
than 2 cgs (standard deviation) and 2 cps man.
7. Output Resietsnce. 100 ohms to feed the 100 ohm input resistance
cf L h ? C.T.
8. Fzquency Scan. The DOPLCC receiver shall scan the 24-kc expected
frequency range in two steps (12 kc per step).
9. Timing Fulses. The DOPLOC receiver ehsll supply "Integ~ate Start"
pulses to the CMF st 8 prf of 10 per ssc. These shall be positlye
pulse8 aud shall have arqlxlltudes of at least 5 volte, into a 100-ohm load, rise aad fall tbes (10 per cent to 90 per cent of
peak amplitude) of not more than 112 wec., and duration (between
go per cent points) of not less than 1 gsec nor more than 3 vsec.
D. Circulating M3mory Filter Specification
The folloving specification is extracted from Columbia University
Report "Summary of the Preliminary Study of the Applicability of the ORD3R
System Techniques to the Tracking of Passive Satellites", 11 February 1960, Appendix B. These specifications are quoted here because they do not other-
wise appear in a BRL report. Readers not interested in detail of the CMF
specification are advised to omit this section.
1. Spectrum to be Analyzed (Frequency Coverage):
2. Input Resistance. 100 + 2 ohms *'! ;; ,,:. . - , : 3. Effective Input Noise Voltage. The noise generated by the CMF
shall be such that if a 1.0 mv m s noise voltage per 10-cps bandwidth is
applied to the input, the output noise shall increase by no less than 3 db. - The nominal value of effective input noise (self-generated) equal to 1 mv rms
gr'l0-cps bandwidth is thus defined. All subsequent voltage levels shall be
*seed in db above 1 mv (CMF noise in 10-cps bandwidth is at 0.0 db). :;.: '
146 !?L,%*: $
4. Nominal W input Signal Levels. (db above CMP noise in a
10 cps band)
a. Receiver plus CW noise per 10 cpa band . . . 11.0 b. F+eceiver noim? per 12 kc6 band . . . . . . . 41.8 c. Miniam detectable signal . . . . . . . . . . 21.4
d. kclmum (unlimited) s igna l . . . . . . . . . 70.0
5. Nominal Coherent Integration Time. Variable from 20 msec t o 110 msec
i n approximately 1 msec s teps . A l l specif icat ions s h a l l be met a f t e r inte-
grating during a nominal integrat ion time of 93 msec.
6 . Integration Start. The CMF shall begin t o integrate an input * voltage not more than 1.0 laa after receiving an in tegra te start pulse. This
pulse w i l l have the fallowing ~ t e r i a t i c s :
a. polar i ty: pos i t ive
b . amplitude: 5 vo l t s or greater across 100 ohms
c . duration: between 1 and 3 psec
d. r isetime: 0.5 Mec (10 per cent t o 90 per cent)
e. period: nominally one pulse every 100 as.
The repe t i t ion r a t e acceptable t o the CMF shall be from g pulses per sec t o
40 pulses per sec.
7. Dead Time. The nominal dead time s h a l l be within 1 msec of the
difference between the input pulse period and the s e t integration time but
s h a l l never be allowed t o become smaller than 4 msec.
8. Linear Operating Range. For any s ingle sinusoidal, noise free
input s ignal within the frequency coverage, having an rms amplitude between
16 db and 36 db there s h a l l e x i s t a l i nea r r e l a t i on between rms input
voltage amplitude and peak video output voltage. L h e a r i t y is here defined
t o be within 5 per cent f o r the bes t linear f i t .
9. Frequencycoverage Ripple. For an input sinusoid of constant
amplitude between 16 db and 36 db, t he CMF video output s h a l l be constant
t o within + 1 db f o r a l l input frequencies within the frequency coverage. - * Externally supplied
147
10. Selectivity. For a sinusoidal, noise free input signal at any
frequency within the frequency coverage and of amplitude no greater than
55 db and no less than 16 db, the output selectivity shall be as shown in Figure 47. All side lobe responses must be at least 40 db below the effective peek response for all inputs up to 55 db, All specifications pertain to that part of the characteristics which lies below 36 db (linear operating range ) ,
11. Signal Enhancement. The selectivity characteristic as specified
shall be obtained without causing the signal enhancement to fall more than
1.7 db below that obtainable from an ideal coherent integrator, for an input signal not more than 30 db.
12. Spurious Response Levels.
a. For a noise free, sinusoidal input voltage of amplitude 55 db rms at any frequency in the frequency coverage, all output spurious responses
shall be more than 40 db (with respect to the input) below the output peak response.
Far a noise free Consisting of two ainiisoids of
amplitude 55 db rms and at different arbitrary frequencies within the coverage, all spurious responses shall be more than 30 db below the main
responses (with respect to the input). This voltage level shall be called
the dynamic processing limit.
I , Signal Limiting AGC. The CMF shall supply a signal limiting AGC
which doasnot significantly deteriorate the specified selectivity charac-
teristic and spurious response levels (items 9 - 12). There ahall be an
independent AGC for each time multiplexed channel. Each AQC shall be
designed so as to meet the minimum specifications shown in Figure 48. In this figure, both the input to the AGC and its output have been normalized
with respect to the input terminals of the CW. !The absolute gain and
signal levels that are processed by the AGC shall be chosen so as to best . . ~. -. compliment CMF design. The salient features of the AGC &e:
<L'',! a. For all inputs between 10 and 40 db the output-input character-
? % .." $
ls'tic shall be linear to within 5 per cent (best linear fit). c 5.r!i;,, 1 . .. ! -. . , 148
OUTPUT IF SATURATION DID NOT OCCUR
OUTPUT WITH NORMAL SATURATION
MAXIMUM SIDELOBE
Z . - 2 0
f 0 f ,', FREQUENCY - fl : f8- f* '
I I
f.1;' Afs - - 2 f 0
DB MAXIMUM MAXIMUM BELOW EFFECTIVE FREQUENCY ASYMMETRY
(UNSATURATE01 SPREAD Af CPS PEAK 1, C P S
0 0 0 - I 0 . 5 2 1.0 - 2 1 3 . 0 1 .5 - 3 17 .0 +_ 1.5 - 6 2 1 . 5 2 2 . 0 - l o 2 8 . 0 2 2 . 0 - I 5 3 2 .O 2 2 . 5 - 20 3 7 . 5 + 3 . 0
A-157 - S - Q O 5 l , " % - " 7yrypical C p z Respan== to a :.:axiiil'rirr?i A?iP::tiiee L Y . - 1
Sinewave (with Time weighting)
8 db.
R M S INPUT ( d b l
NOT MEASURABLE DUE TO C M F NOISE.
ACCEPTABLE REGION OF AG C ACTION.
Fig. 48 Signal AGC Requirements
b. For all inputs between 40 and 70 db there is an acceptable region of AGC action as shown in Figure 48 but un& no circumatancea shall the
output exceed 55 db fm a l l input signals q to 70 db.
c. For all inputa below 10 db the AM: output ehould be such as to not
degndt system performance.
d. The AGC gain must be reset at least once during each integration
time for each multiplexed channel i.e., it neednotbe continuously in oper-
ation during the integration the.
14. Video Outputs., !I% laat circulation time of the integration time * shall be called the display period. The thirteen the intervals during the
display period which containthe thirteen time: multiplexed channels shall
be called the channel display times. The following video outputs with
appropriate oscilloscope synchronizing pulses shall be provided by the CW:
a. Entire integration time.
b. Gated video during display period.
c. ~a&d video during each channel display time.
All Qideo outputs shall 'have an amplitude of at least 40 db peak into 100 ohms
for a noise free input signal of 20 db.
15. Display Period Start. A synchronizing pulse shall be provided to
indicate the start of the display period. This pulse shall occur approx-
imately 1 psec before the beginning of the display period and shall have
the electrical characteristics specified for the timing pulses described in
Section 6.
16. Threshold Levels and Stability. The C W shall provide a video
threshold which s h a l l be continuously variable from 15 to 35 db with a 'drift of less than + 0.2 db during any t W interval no shorter than - 0.1 sec. and no longerthan the time between calibrations. The nominal.
threshold level shall be 24.4 db.
This threshold shall operate on all video output signals which result
from input signals less than 40 db. For video output signals which result
from input signals greater than 40 db, this threshold shall operate on the video after it has been attenuated by 10 db. This operation shall be automatic.
* Twelve signal channels and one calibration channel
151
17. Output Data Pulses. Esca time during the display period that
the vldeo output of the CMF is increasing and is equal to the threshold
dbecribsd in Section 16 ths CMF shall produce a pulee. !l%a time occur- * mnce of this pulse shall have a m a n value within 0.05 - of the time thst the threshold is equal to the video output and ehall have a standard
deviation of not more than 0.1 wec. These pulses ahall be positive and
have en amplitude of, at least 2 volts, rise and fall times (10 per cent
to 90 p r cent of peak amplitude) of no more than 0.05 psec, and a duration
(bet-n 90 per cent points) of not more than 3 psec nor lese than 0.2 vsec.
Each tlme during the display period that the video output is decreasing
and is equal to the threshold describeb. in &ction 16 above, the CMF ehall
produce a pulee. The time of occurrence of this pulee shall have a ~ a a n
value within 0.05 psec oP the time that the threshold is equal to the video
output and shall have a standard deviation of not more than 0.1 psec. The
electrical characteristics'of these pulaes shall be the same as those for
the data pulses described above.
The pulses marking the positive going threshold crossings shall be
provided at a separate output terminal from the pulses markingthe negative
going threshold croasinge.
18. Output Impedance Levels. All video outputs, synchronizing pulses,
and output data pulses shall be capable of driving 100 ohm im-cerr.
9. External Terminals.
a. Chassis ground
b. Display period synchronizing pulse
c. Output data pulses (two terminals as provided in Section 17)
d. AGC voltages
e. External time weighting function generator with disconnect for
built-in t b e weighting function
* The %be of occurreace of the data output pulses shall be measured with respect to pulses originating fram the calibration signal contained in
.':::-the calibration display channel. The time of occurrence specifications shall apply to all input signals of amplitude between 21.4 db and'
:-.:;-- 55 db and a threshold setting of 24.4 db.
f. AI1 necessary.monitoring functions for alignment and operation
as determined by the contractor.
20. Acceptance Tests. All items specified shall be checked for . .
conformity with the specifications in the presence of BRL representatives
and a reliability run shall be made such that the equipment shall be
observed to operate satisfactorily for five consecutive days at eight hours
per day. During this five-day period, preventative maintenance and check-
out shall be allowed during the period in which the equipment is turned off.
21.. CW Information Available. The contractor shall supply an
instruction report sufficiently comprehensive to allow an engineer to
perform preventative maintenance and generally service the equipment. The
report shall cover theory of operation of the equipment, block diagrams,
detailed circuit diagrams, checkout and preventative maintenance procedures,
servicing of breakdowns, etc.
22. Installation. Installation and initial. checkout of the equip-
ment at the radar site shall be supervised by a contractor representative.
23. Power Requirements. The equipment shall op&ate from any constant
frequency constant amplitude power source whose rms amplitude is greater
than 105 volts but less than 120 volts and whose frequency is greater than
50 cps but less than 65 cps.
E. Studies Relative to DOPLCC Receivers
1. BRL Report No. 1093, "The Dynamic Characteristics of Phase Lock Receivers" by K. A. Pullen. The first receivers used in detecting satellite
signals were ccmercial versions of the R-390 A/URR as sold by Collins Radio
Company under* model number 5-51. These receivers were used to receive
the signals from the first Russian Sputniks. The same receivers along with
type R-~~OA/URR were later used for the interFm DOPLCC stations. Since
these receivers would not tune above 32 mc modified Tape-Tone converters
were used to operate at 108 mc. These circuits were improved through the
use of very stable frequency standards and frequency synthesizers as
heterodyning frequency sources. Except for some preamplifiers and special
mixing c i r cu i t s the interim DOPLOC receiving equipment used essent ia l ly
off-the-shelf i t e m s . The phase locked tracktug. f i l t e r s were b u i l t t o
specifications but ear ly models wereavailable when s a t e l l i t e s became a
r e a l i t y .
With the development of the concept of scanning beam antennas fo r
the DOPLCC system, it became obvious t h a t more sophisticated receivers
would %e required. The f i r e t study made resulted i n BRL Report No. 1093,
January 1960, e n t i t l e d "The Dynamic Characterist ics of Phase-Lock Receivers",
by Keats' A. Fullen, Jr. This was Report No. 8 i n the ARPA S a t e l l i t e Fence
Series. The report t r e a t s of the general theory of phase locked c i r c u i t s
88 applied t o small amplitude signal detection. Except f o r the following
conclusions r e l a t ive t o the receiver, the reader is referred t o the or iginal
report .
a. "Constant s ens i t i v i ty t o phase changes requires the use of a
phase detector i n the signal c i r c u i t and a cross correlation type detector
i n the gain control c i r cu i t , so t h a t t h e g a i n action is controlled by the
desired s ignal alone and not the s ignal and noise together."
b. "The l imitat ion of phase e r ror requires e i the r the use of a caammon
i:f. system for a pa i r of signals or the use of considerable loca l feedback
on a l l of the r.f. and 1.f . amplifiers. The use of a c i r c u i t without l oca l
feedback around the 1.f. amplifiers does not introduce phase s tab i l iza t ion ;
it only introduces a compensating phase in to the feedback path. Since the
phase of the s ignal i n the feedback path is used as a measure of the
received phase, no removal of in te rna l phase e r rors r e su l t s unless l oca l
feedback over the in&P~ridunl atages i f i used."
c . "Adjustment of the phase-lock loop fo r minimum position, velocity,
and acceleration e r ror requires the use of a rather different type of . , .. c i r cu i t ry than has been used in general i n t h i s type of equipment. The
presence of frequency compensating c i r cu i t s between the phase detector and
t h e : ~ a c t a n c e modulator introduces neither posit ion nor velocity e r ro r
correction, but contributes only t o the correction of the acceleration . ..,. . .-
er ror . Correction f o r posit ion and velocity e r ro r must be accomplished
iiietiie feedback loop ." .'.P :"., .-.. . ., 1.54
? i i .rp. . '.. ".$&,.
d. "The introduction of f u l l correction f o r veloci ty e r ror requires
the use of a phase network in the feedback path of ra ther unusual charac-
t e r i s t i c s . Further investigation of t h i s portion of the c i r c u i t appears
t o be justif ied. ' '
2 . DOPLCC Receiver Design f o r CMF. BRL Technical Note No. 1345, August 1960 e n t i t l e d "The DOPLOC Receiver f o r Use w i t h the Circulating
kmory ~ i l t e r " by Kenneth H. Patterson. Report No. 18 in the ARPA Satel-
l i t e Fence Series, discusses a radio receiver which was proposed f o r use '
with the Circulating W o r y F i l t e r . Figure 49 is a schematic block diagram
of this receiver. Basically it consists of an r . f . amplifier operating a t
150.79 mc and three i.f. amplifiers w i t h automatic gain control and the
s tab l i ized mixer frequency sources. The ls t , 2nd and 3rd i.f. amplifiers
operate at 19.217296, 2.770708 and 10.994002 mc respectively. The
receiver provides the specified 11 mc signal output t o the circulat ing
memory filter. The required frequehcy s t a b i l i t y i s provided by bringing
the required mixer frequencies from frequency standards and synthesizers.
The receiver a l so supplies t o the f i l t e r the "start integrate" pulses. For
a more detai led discussion reference i s made t o the above mentioned report .
3. Tests of Parametric Pre-amplifiers. BRL Technical Note No. 1354,
October 1960 en t i t l ed "Parametric Pre-Amplifier Results" by K. H. Patterson.
Report No. 19 , ARPA S a t e l l i t e Fence Series describes the r e su l t s of a
s e r i e s of t e s t a w i t h some parametric amplifiers supplied by Microwave
Associatee, Inc. These amplifiers were rather simple i n so far as physical
construction was concerned. , They consisted of a coaxial cavity rdth an
inject ion and a pickup loop. The 50-ohm loops appeared t o be ident ical .
Each was provided with a standard type N coaxial cable f i t t i n g . A coarse
and Pine tuning adjustment were provided. The latter was actual ly a
modified micrometer mechanism. Provision was a l s o made f o r mounting a
varactor diode and 'an adjustment t o vary a capacitor connected i n series
w i t h the diode.
F i r s t msults w i t h the parametric amplifiers w e r e very disappointing.
It was then discovered t h a t the recommended pump frequency was the wrong
value f o r the amplifiers. L a t e r laboratory t e s t s showed the amplifiers
to be practical-and to give a reduction of about one db in the noise
figure of the receiver system. However, later attempts to use these
amplifiers atthe field stations gave unsatisfactory results. The
high "Q" circuit seemed .to be sensitive to temperature drift. This
sensitivity combined with the narrow bandwidth seriously changed the
overall performance. Since the improvement over vacuum tubes was only
about one db and since the external noise was frequently higher than
the receiver input noise, no further attempt was made to overcome the
problems associated with the parametric amplifiers.
F. Study of Orbital Data Handling and Presentation
BRL Technical Note No. 1265, June 1959, entitled "Orbital Data Handling and Presentation" by R. E. A. Putnam, Report No. 3 ARPA Satel- lite Fence Series, discusses the problem associated with handling the
data for the DOPLCC system. This is a preliminary study, the purpose of
which is to present the relative merits and limitations of various alter-
nates proposed; to indicate the particular methods worthy of serious
consideration and further study; and to outline the time factors and
development costs likely to be involved in procuring prototype equipment.
Discussion is limited to a general, rather than to a specific treatment
of the subject, because numerous pertinent details of instrumentation and
orbital computational procedures we- not stabilized at the time the
report was written. The subtitles treated are "preliminary data editing,
orbital data catalog, orbital comparisons, visual data displays". The following is a summary of concluaians reached.
a. Preliminary data editing by visual inspectlon of the Doppler "St'
curve is believed to be essential, at least during the early stages of the
project.
b. Orbital garmeter data should be catalogued at the data analysis
center in printed tabular form for operator reference, also in digital
form far satellite identification and plotting board input. The choice
of digital storage medium will be governed by facilities locally available,
equipment,ccnnplexity, ccmrmaercial availability and cost factors.
157
c. A plotting board display appears to be justifiable for providing
a visual check on satellite identification in doubtful cases. ,
d. Orbital charts produced on plotting board and distributed in
advance to ground receiver stations may sene a useful purpose in alerting operators to the Fmminent approach of known satellites and the* dinction
of approach.
e. It is recommended that prototype items of equipment hardware be
aesaplb&ed to demonstrate their suitability in handling actual orB'lital data
before gttempting to prepare final apeoifications for a field installation.
G. Synchronization of Tracking Antennas
BRL, Technical Note No. 1278, September 1959, entitled "Synchronization of Tracking Antennas" by R. E. A. Putnam, Report No. 6, ARPA Satellite Fence Series, considers the problems associated with synchronizing scanning
antennas.
Synchronizing a two-station complex to within + 1 degree, relative to - each other, requires that each individual station be synchronized to within
+ 0.5 degree, relative to a ccormron reference standard. In the table which - follows are presented estimates of the relative component and total error
magnitudes which might be expected from each of the three types of scanning
antenna designs capable of meeting performmice requirements.
Estimated Error Usgnitudes for Various Antenna Types
Error Source
1. Station Timing
2. Antenna Orientation
3. Antenna Pattern
4. Phasing
5. Structural Deformation
6. Scanning Beam Drlve
Totals .I .
Rotatable Element Switching
+ - 0.03~ + 0.02 - + 0.05 - + 0.10 - + 0.05 - + 0.10 - + 0.35' -
Controlled Phase Shift
+ 0.03" - + 0.02 - + 0.05 - + - 0.35 + 0.05 - + 0.00 - + - 0.50'
It is to be noted that error ratings of the three antenna systems
are identical for the first three error sources, a d are believed to be
reasonably representative of values that can be maintained under ordinary
operating conditions. Incidentally, they add up to 20 per cent of the
total error permissible. Inasmuch as error magnitudes from almost any
source can be progressively reduced by concentrating on equipment and
circuit refinement, it is reasonable to suppose that all three antenna
type8 under discussion can be developed to the point where they meet the
necessary performance requirements for this proposed project, if cost is
not a factor. However comparative figures tabulated above strongly
indicate that the switched element type of antenna has the distinct advan-
tage of a lower error susceptibility, hence the probability that it will
carry the lower price tag and a simplified maintenance program. Other
conclusions reached are:
a. Sky-scanning antennas, synchronized to within + 1 degree, appear - to be technically feasible for station spacings up to, and somewhat in
excess of 1000 miles.
b. Each station should include a precision time generator to permit
continuing operation for a reasonable period following interruption of the
external source of synchronizing signals.
c. Economic and other considerations indicate the desirability of
synchronizing each station directly with W timing signals.
d. Sky-scanning rates Wluence station synchronization procedures
an8 data transmission problems. It is desirable that a scanning rate be
selected such that the resulting number of scan cycles per minute becomes
a whole number.
e. It is not.essentia1 that the angular scan rate be completely
unlform throughout the sweep range, but it is important that any non-
linearities be identical for both stations of a pair.
f. Sky-scannFng afiangements with a reciprocating a m p may be
eubject to non-linear caaponents of motion in one direction that a'e not
duplicated in the reverse direction, thus cmnplicating synchr~nizati~n
procedures. For this reason, bema that sweep contFnuously in One direction
are to be preferred.
g. A fixed antenna structure, where in a scanning beam i6 genersted
by prbgresaive switching of radiating elements, appears to be the most
praC%ical antema type.
5 . Same moderate decrease in beam widtb, or some increase in beam
scanning rate, appear to be feasible, if desired. However, such changes
tend to impose increasingly severe requirements on the synchronization . prop?m, and at some point a limit will be reached beyond which further
changes will be impractical. Whether synchronization requirements can be
met with beemwidths of only one degree is problematical.
E. Erecision Frequency Measurement of Noisy Doppler Signals
BRL Report No. 1110, June 1960, entitled "PrecisLon Frequency Measure- ment of Noisy Doppler Signals", by William A. Dean, Report No. 15 in the BRL DOPLOC Satellite Fence Series, gives the rerults of a brief study of
the uncertainty expected in measuring the frequency of a radio Doppler signal
from the data gathering equipment of the DOPLCC satellite tracking station.
Two methods of measurement are alternately uaed depending qon the cycle
count and the desired precision. One method simply counts Doppler cycles
from the tracking filter for a given time period, for example, one second.
The second method is indirect and integrates a chosen number of Doppler
cycles over which period a high frequency, say 10 mc standard frequency is
counted. Since the resolution of the counter is - + 1 cycle, yielding a counting precision of +G - cycles per second when both ends of the interval
are concerned, the precision of the 10 mc count is considerable improved
over that of a Doppler count of from zero to a few thousand cycles/second.
me random scatter varies with propagation phenomena and the angle of elevation of the satellite with respect to the observation station, but V2.l
may be as much as 3 cycles per second. This scatter is of such character
that it does not a f f e c t the operation of the tracking f i l t e r . Typical
conditions are plo t ted i n the repor t where a t an input frequency of 10 kc
an input s ignal t o noise r a t i o of -18db, and a tracking f i l t e r bandwidth
of 1 0 cps, peak excursions of 0.16 cps were observed and the ws frequency
error w a a 0.652 cps . I. Contract DA-04-200-m~-674, Stanford Research Ins t i t u t e , System Studies
Under contract DA-04-200-ORD-674, Stanford Research I n s t i t u t e conducted
a DOPLCC! system study which lead t o two reports . The first of these is
e n t i t l e d F ina l Report, Par t A, Sta t ion Geometry Studies f o r the DOPLCC
System by W. E. Scharfman, H. Rothman and T. Morita. This report i s number 9
i n the BRL DOPLCC s a t e l l i t e fence se r ies . It per ta ins en t i r e ly t o t he interim
type system which u t i l i z e s fixed antenna beams oriented i n such a manner as
t o obtain data f o r segments of an "St' curve. An analysis was made of the
r e l a t i on between the layout of the receiving and transmitting s ta t ions of
the BRL CW re f lec t ion Doppler system ( fo r the detection of non-radiating
s a t e l l i t e s ) and such f ac to r s a s antenna gain, pat tern shapes, time of t r a n s i t
of a beam by a s a t e l l i t e and coverage f o r fixed-beam antennas.
The r e su l t s of the study show the regions i n space where suf f ic ien t data
would be obtained t o determine the s a t e l l i t e ' s o r b i t a l parameters, and the
complexity of the required antenna system for various s t a t i on geometries and
s a t e l l i t e a l t i tudes . This report is c l a s s i f i ed confidential and is therefore
not: abstracted i n d e t a i l here.
The second SRI report i s e n t i t l e d "Final Report, Par t B, DOPLOC System
Studies" by W. E. Sharftnan, H. Rothman, H. Guthart and T. Morita and is
Report No. 10 i n the BRL DOPLW S a t e l l i t e Fence Series.
Par t B of t h i s f i n a l report shows t h a t because of the var ia t ion i n
re f rac t ive index of t he ionosphere, an e r ro r is introduced i n the determina-
t i o n of the radial veloci ty of a moving target . This e r ror r e su l t s because
the direct ion of the refracted ray at the t a rge t d i f f e r s from the line-of-
s igh t direct ion. The Doppler e r ro r introduced a r i s e s from conditions a t the
ac tua l instantaneous posit ion of the s a t e l l i t e r a the r than from the inte-
grated e f f o r t of conditions along the e n t i r e propagation path. This e r ro r
is greatest when the satellite is at the level where the ionization is
nost dense, at the msximm of the F2 layer (300 KM). Here the error amounts
to about 0.25 per cent of the Doppler frequency for a carrier frequency of
100 mc.
The Faraday effect causes a rotation of the plane of polarization Of
a radio wave passing through an ionized medium in the presence of a megnetic field. As a typical example the number of rotations that a 100 mc vavc
might incur in a double passage through an ionosphere 260 KM thick, at a
height of 150 KM, varies from about 20 at zero degrees elevation to about
5 at 90 degrees elevation.
This report shows that the Luxembourg effect, or ionospheric cross-
modulation effect, can be neglected for frequencies above 4 mc. Hence
cross-modulation at 100 mc shourld be extremely small. It is also shown that
the atmospheric noise generated in the troposphere is at least 40 db less than cosmic noise at 100 mc and can therefore be neglected.
Forward-scatter propagation involves scattering either by ionospheric
or tropospheric inhomogeneities in the conmnon volume of the transmitting
and receiving antenna beams. For frequencies above 100 mc and separation
less than 600 miles, the tropospheric transmission component appesrs to dominate. Typical features of the tropospheric scatter are that it
decreases rapidly with increasing distance, and decreases very slowly with
increasing frequency. For a station separation of 450 miles, the average signal level with optimum antenna orientation will be about 90 db below that expected at the same distance in free space with the same power and
same antennas. Free-space attenuation at this frequency and separation
is 130 db. This received signal level is considerably greater than the
received signal expected from forward scattering of the satellite. The
actual tropospheric signal received is further below the free space signal
since the side lobes and not the main beams will be forming the common
valume . -
, s
--t ,<
SRI spec i f ica l ly investigated the question, a s t o whether the noise
f igure could be reduced by reducing the temperature of the antenna or of
any reasonable volume surrounding t h e antenna. A t 100 mc the background
noise temperature is about 800 degrees K, while some areas of the sky
(notably m e ga lac t ic cen te r j nave temperatures up t o thousands of degrees
Kelvin. It i s shown t h a t i f a receiver has a 3-db noise f igure and an
apparent antenna temperature of 800 degrees K, the operating noise figure
w i l l be 5.5 db. It is therefore concluded tha t no reduction i n noise
f igure would be achieved by cooling e i t h e r the antenna or the volume
immediately surrounding it.
The e f f ec t s of meteors and meteor t r a i l s have been discussed b r i e f ly
elsewhere i n t h i s report . SRI presents an extensive analysis of the e f f ec t s
t o be expected from both meteors and meteor t r a i l s . This treatrhent i s too
extensive t o abs t rac t here except f o r t he following remarks. The specular
echo count f o r meteor t r a i l s with i n i t i a l l i n e densities of 1014 electrons
per meter i s d i r ec t ly proportional t o t he square root of the transmitter
output power and the receiver s ens i t i v i t y . The interim DOPLOC system was
sens i t ive t o t r a i l s with l i n e densi t ies of the order of 1012 electrons per
meter. If the system were t o become sens i t ive t o l i n e densi t ies of l e s s . -10 - .
thm lo electrons per meter, the count might become d i r ec t iy proportionai
t o these system fac tors , and possibly even proportional t o t h e i r square or
cube. Various methods such a s increasing the frequency, the incorporation
of f i l t e r s , e t c . a re available t o discriminate against meteor t r a i l s ignals .
The remainder of Part B of the SRI report t r e a t s the e f f ec t s of
antenna phase pat terns and suggests antenna systems f o r the DOPLOC system.
For this treatment the reader i s referred t o the basic report .
J. Comb F i l t e r Development
Since the lead time f o r the development of the c i rculat ing memory
f i l t e r was known t o be a t l e a s t 12 months and the cost approximately
$250,000, it was decided t o look f o r a second method of obtaining narrow
band f i l t e r act ion on the data from the scanning antenna DOPLOC system.
The search and lock-on time f o r the phase locked tracking f i l t e r s was too
163
long t o permit these f i l t e r s t o operate with scanning antennas. Therefore
a progr~m was i n i t i a t e d t o develop ~ ~ u l t a n e o u e l y with t h e c i r c u h t i n g
m+oq filter a comb type filter.
Star t ing with a 10-cycle f i l t e r pass band and a possible 12-kc band
of Doppler frequencies t o be covered, it is a t once apparent that 1200
individual f i l t e r s would be required t o cover the en t i r e band. It was
therefore decided t o build only 8 section of the f i l t e r f o r test purposes.
The esult was one hundred and eighty individual ten cycle f i l t e r s spaced
20 eps apart t o cover a t o t a l of 3600-cps bandwidth. This input was t o
be switched three t b e s t o cover the 12-kc band. The r a t e of change of
Doppler frequency varies with a l t i t ude of the s a t e l l i t e from approximately
25 t o 100 cps. This maximum r a t e of 100 cps determines the minimum usable
bandwidth of 10 cps. This i n turn s e t s the maximum time available t o
recognize the Doppler signal a s 100 milliseconds. This maximum t i m e deter-
mines the maximum scan r a t e f o r a given antenna beamwidth. Since twelve
data points a re desired t o insure suff ic ient data f o r a convergent solu-
t i on from a single pass, there i s a lso a minimum scan r a t e associated with
the low a l t i t ude passes f o r R scan r a t e tha t would give 12 data points on
a pas6 and an integration time of 100 m . s . , a beamwidth of approximately
1 degree i s indicated.
Describing the f i l t e r action br ie f ly , the output of the receiver i s
fed through a noise actuated automatic gain control c i r cu i t , thence t o
the f i l t e r comb, each f i l t e r of which has an amplifier, AGC, and threshold
control c i r c u i t which feeds a multi-pen recorder. A record indicates a
s ignal of frequency corresponding t o the f i l t e r frequency present at the
input t o the comb. This comb filter has been completed and tes ted on
both r e a l time and simulated tracking data and found t o perform satis-
f s c t w i l y . It is described i n fu l l d e t a i l i n BRL Memorandum Report No.
1349 en t i t l ed "A Comb F i l t e r f o r Use i n Tracking Satel l i tes" , by R. Vitek.
Figures 50 t o 57 inclusive a r e photographe of actual recordings made
a t the output of the comb f i l t e r . Figures 50 t o 54 ahtm the performance
charac ter i s t ics of the f i l t e r with simulated date.. Figurea 55 t o 57 ara
OPTIMUM NOISE LEVEL
FOR - MAXIMUM SENSITIVITY
VARIABLE NOISE
vs SIGNAL LEVEL
- 2 ,,
1 c 8 NOISE LEVEL : I 0 0 MV RMS ' I . :'!. 'i ! ..., SWEEP RATE : 80 CYCLESISEC.~ . .
:. .. SWEEP LEVELS : AS SHOWN
FREQUENCY SCALE - C.P.5
Figure 50 Optimum Noise Level for Maximum Sensitivity (Variable Noise vs Signal Level)
.......... - ....... .- . . . . . . . . . S E N S I T I V I T Y CHART
VARIABLE NOISE AND SIGNAL L E V E L T O
D E T E R M I N E OPTIMUM S E N S I T I V I T Y
, NOISE LEVEL : 80 MV RYS ' SWEEP RATE : 100 CYCLESISEC.' . .... .... SIGNAL LEVEL : AS SHOWN , ...... .. '. ... . -.._
"""' ......... : ! , , . , . '
' . j ' ' " ' . ' 1 : : . , ....... I . . :
! I.
NOISE LEVEL : 7 0 UV RMS SWEEP RATE ! 100 CYCLESISEC.' .- , ....... .;. SIGNAL LEVEL : AS SHOWN .... .... -..
-2,*)l - I . , . ' * . .-., : t .. I -Z.dD *... ..... !
! ' . . . -2ldb
FREQUENCY SCALE - C.P.S. . -- ". ". .
Figure 51 Sensitivity Chart
166
DYNAMIC RANGE CHART
RANGE OF MAXIMUM TO MINIMUM LEVELS
GIVING SINGLE PEN OPERATION
FREQUENCY SCALE - C.P.S.
Figure 52 Dynamic ~ange 'char t
167
i . : : . . . . I :.. I l:
I : NOISE LEVEL : I 5 0 MV RMS . !..I j . ! '! !., . ? SWEEP RATE : 80 CYCLESISEC.~ ' ." P 3
,,, : j , I . .I I. jl. ,'-;: :::I ., SIGNAL LEVELS : AS SHOWN ,, ; I ,I . . 1,. : , : ;I:!. 1 . k i :. ...... ..
I .. : ... .i , , I 9 . . : . . . ," : :. '.'- '. >:, :;., , n , , . - 3 0 d b ,.? .. , I , .I ::' . . .. . , , . .
.I I;, I :, , , , 1.: ,t a 1' . .iL :'.: 1. ., ,I .. . . . . . C ' . . . . .!: i , . 6 : ; . . . . ' i - ' . : . ' ! . . . I
. : . . . 1 , 6, . -; :*.. ; ! i ' , . ; ! ~ l ' ,.. ! . . I ; . ' ... ! * .... 1: 0 6 , . : . I j f . . : 8. i . : s:. . . .. ! : 1 , . . : j ,
, . . : , ! ' I .. , : . . , ,,I ! .... ,..'.. .:';.I . : . . . . . , , , t . , # j l. I ;.I;
% :;,;,;,!;!' ! I ., ,
. i ' . ( . I *
, , { j : ! . ~ . . * . . I i
. ; .I 4 ;. I , . , .
' I I, .: I . I .
I.: _ ; . .. >. P ,, ., # 8 . ,
: . . . ... I. :. 3 :
. i!:!.. I , . * i, 1 : ' ; , ! , :, ! . ( . ' ...k 1 . t 8 ' ' ( " '1 ' ! ., ; 1s
I . / I ,
AGC' ACTION C H A R T
LARGE AMPLITUDE F IXED FREQUENCY SIGNAL V S
SWEEP SIGNAL L E V E L ' .
.;, . . . / I . . . . .'. ! I ... I : . ' NOISE L E V E L .: 150 MV R M R . : .'. 2 1 . : . . ; ,:. , 8:.
:;, 1 . ... ;I: ' " , . F I X E D S IGNAL : - 6 d b , . . ; . .I . ! I , . , .. . '
j ! : SWEEP RATE : 77.5 C Y C L E S / S E C . ~ , . i:' ,;. ;;' :: .
SIGNAL L E V E L S : AS SHOWN I . :': . . :, ! . . :. . , .
: . i ,
. . . ' . . ?
. -:,
/ . . . . , . : . ! :. , a ,
, . . .. !
. . . .
! 1 F R E Q U E N C Y S C A L E - C.P.S.
:. . ' ~ . . .. . . . ~~ . .
Figure 53 AGC Action Chart
TWO CROSSING SIGNALS *;:. , ..., :I .! 1:: .:,. .i a ; . . 9 W I T H VARIABLE SWEEP RATES 'I
,,. :. :: I:. . .
FREQUENCY SCALE - C.P.S.
Figure 54 Two Crossing Signals wi th Variable Sweep Rates
................... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .-. 7 60 DELTA '; REV. 172
i NORTH -CENTER -SOUTH ANTENNAS NORTH- - SOUTH PASS
DOPPLER FREQUENCY - C.P.S
Figure 55 60 Delta Rev. 172
DOPPLER FREQUENCY - C.P.S.
Figure 56 Delta Rev. 165
. . - .. . . . d .. .- .- 60 DELTA R E ~ 124
NORTH -CENTER - S O U T H ANTENNAS NORTH - SOUTH PASS
. . . . DOPPLER FREQUENCY - C.P.S..
. - ..
Figure 57 60 Delta Rev. 124
copies of s a t e l l i t e pass data from magnetic tapes recorded on the
interim DOPLOC system a t Forrest City, Arkansas. Chart records a re
a l so available of the performance of the ALO - tracking f i l t e r equip-
ment on these passes. The arrows on the records indicate the frequency
l i m i t s of the tracking f i l t e r records. In general the comb filter
tracked the signal fo r a longer period than did the tracking f i l t e r .
This was expected since the bes t s ens i t i v i ty of the ALO was 21 db. It
could not therefore pick up the s ignal as soon a s the comb f i l t e r .
After a lock was obtained the tracking f i l t e r was independent of the ALO
and i t s sens i t i v i ty was ident ica l t o t h a t of the comb f i l t e r . The t e s t
records of simulated and actual s a t e l l i t e signals are indicative of a
high l eve l of performance. The comb f i l t e r prototype met a l l of the design
requirements. The uni t i s f u l l y capable of being expanded t o the fu l l
complement of 1200 f i l t e r s and of providing an effect ive tracking capa-
b i l i t y f o r the proposed DOPLOC scanning system. It could provide an
excellent means of tracking any short duration signals having a high r a t e
of frequency change. The frequericy accuracy is limited t o + 2 cps with - the 10-cps bandwidth f i l t e r elements.
VII. SATELLITE ORBIT DFLTERMINATION FROM DOPLOC DATA
R. B. Patton, Jr.
A. Introduction
Computing methods have been developed which provide the DOPLOC system
with the capability of detecting and identifying non-radiating satel4itee.
While these methods have been derived specifically for Doppler-type tracking
syst,ems, the computing procedures are sufficiently flexible that, with minor
modification, they may be applied to observations from any satellite or ICBM
tracking system. Moreover, they provide relatively accurate orbital deter-
ininations when the tracking observations are limited to a few minutes of data
recorded at a single receiver in the course of a single pass of a satellite.
This capability is fundamental for a detection system. In addition, the
computing program has been extended to provide a highly accurate solution
when the input data are increased to include observations from more than a
single pass of the satellite or from more than one receiver.
A prime requirement for a satellite detection system is the capability of establishing an orbit within minutes after the first:pass over the tracking
system. The computing program presentedhere provides the DOPLOC systep with
the capacity for orbit determination within one to five minutes after satel-
lite passage. Further,.theae methods function effectively with as few as
6 to 12 observations recorded at a single receiver over a 2 to 3 minute interval. It is doubtful whether any other satellite detection systemcan
provide equivalent accuracy for first pass detection. Moreover, many have
no capability whatever for orbital determination from single-pass observa-
tions.
Computing programs for multiple-pass solutions have been under consid-
eration for sometime. Although the development of these methods has not
been completed, a relatively simple procedure is emerging which provides
sufficient accuracy for positive identification. In addition, these results may be used to predict position with enough precision to be used as an aid
in future acquisition. Although this work is not complete, the method
appears to be both practical and useful. Moreover, the initial results
from solutions with actual field data are very promising and ha&, therefore,
been included in the section, "Computational Results".
In practice, orbital determination from DOPLOC observations is accom-
plished in three phases:
1. Initial approximations for position and velocity are established
from frequency and slope measurements near the inflection point of the
Doppler frequency time curve for each set of observations.
2. These results are used as input to obtain single-pass solutions for
individual sets of observations. Such computations yield moderately accurate
Keplerian orbital parameters to be used for initial identification.
3. The single-pass results are used to initiate a multiple-pass solu-
tion which simultaneously combines observations from several'passes to
compute a perturbed orbit with sufficient accuracy to provide both positive
identification and practical prediction.
The first two steps in the computing procedure have been completely
developed and thoroughly tested with numerous sets of actual field data.
Development of the multiple-pass solution is nearing completion. Limited,
but successful, testing has been accomplished with field observations.
B. Description of Observed Data
Doppler observations consist of recordings of Doppler frequency as
a function of time. Here the Doppler frequency is defined as the frequency
obtained by heterodyning a locally generated signal against the signal
received from the satellite followed by a correction for the frequency bias
introduced as a result of the difference between the frequency of the local
oscillator and that of the signal source. The Doppler frequency is defined
to be negative when the satellite is approaching the receiving site and
positive when it is receding. If the Doppler frequency, as defined, is
plotted as a function of time, one obtains curves of the form shown in
Figure 58, usually referred to as "St' curves. The asymmetry of the curves
is typical for a tracking system with a ground-based transmitter and a
receiver separated by an appreciable distance. Only for a Satellite,
whose orbital plane either bisects or is orthogonal to the base line,
will the Doppler data produce a symmetrical "s" curve with a reflection system. With a satellite-borne transmitter, the "St' curve is very nearly
s~metrical, being modified slightly by the earth's rotation and the
refractive effect of the ionosphere. If continuous observations are made
and sampled at frequent intervals, such as once per second, Figure 58 (a) illustrates an analog plot of the data available for computer input.
However, with a ground-based transmitter it is necessary to limit the
number of observations in order to minimize equipment cost and complexity.
For example, the three beam antenna complex of the interim DOPLOC system
provides three sections of the "St' curve as shown in Figure 58 (b). Another
possibility is the use of a scanning antenna beam to provide discreet obser-
vations at regular intervals as shown in Figure 58 (c). Such data could
be obtained by an' antenna with a thin, fan-shaped beam which scanned the
sky repetitively.
Any of these sets of data may be used readily as input for the computing
procedures. Whenever possible, this input consists of the total cycle count
rather than the Doppler frequency, i.e., the area under the curves or arcs
of curves presented in Figure 58 (a) and (b). Hence, this method of solution
is based, in a sense, upon every available observation rather than upon a
selected few of the total observations (which would be the case if the com-
puting input were limited to a representative number of frequency measure-
ments). Experience has shown this to yield a very significant gain with
regard to the accuracy and convergent properties of the solution.
C. The Single-Pass Solution
The method of solution consists of a curve-fitting procedure in which
a compatible set of approximations for the orbital parameters, are improved
by successive differential corrections. The latter are obtained from a
least-squares treatment of an over-determined system of equations of
condition. The imposed limitation of single-pass detection permits several
assumptions which considerably simplify the computing procedure. Among these
176
(c)
Figure 58 (a) (b) (c) -Doppler frequency-time curves.
. . A .
' 2 e L
0
. . . . . . . . . . . , TIME
is the assumption that the Earth may be treated dynamically as a sphere
while geometrically regarding it as an ellipsoid, in addition, it iu
assumed that no serious loss in accuracy will result if drag is neglected
as a dynamic force. With these assumptions, it is apparent that the setel-
lite may be regarded as moving in a Keplerian orbit. An additional simplification in the reduction of the tracking data is warranted if the
frequency of the system exceeds 100 megacycles; for it then becomes fea-
slble to neglect both the atmospheric and ionospheric refraction of the
transmitted signal.
In formulating the problem mathematically, it is helpful to regard
the instrumentation as an interferometer. In this sense, the total number
of Doppler cycles observed within any time interval will provide a measure
of the change in the sum of the slant ranges from the satellite to the
transmitting and receiving sites. If. h is the wavelength of the radiated
signal and N the total number of Doppler cycles resulting at the ith i5
receiver from themotion of the satellite between times t and t 5 J +l J
it follows that v . hNIJ is a measure of the total change in path length i 5
from the transmitter to the satellite to the ith receiver between times
t and t5+l. 5
Let gi5 be defined as the actual change in path length. It fcll~ws frcz Pig=e 56 that:
where T is the transmitting site, R the location of the ith receiver, i
S the position of the satellite at time t5+l, 5 +l
and S the position J
of the satellite at time t In the event that the observations consist 3'
only of discrete measurements of frequency, the same definitions will
apply, but the time interval from t to t will be limited to one second. 'I 'I +l
The solution consists of improving a set of position and velocity
components which have been approximated for a specific time. The latter
will be defined as to, and as a matter of convenience, it is generally taken
as a time near the inflection point of the "St' curve. Although the location
of.the coordinate system in which the position and velocity vectors are
defined is relatively immaterial from the standpoint of convergence,
a eyetem fixed with reepect to the Earth's surface doee provide two
dietinct advantagee. Firet, the function g ie eomewhat simplified iJ
since the coordinates of the instrumentation sitee remain fixed with
respect to time. Of greater significance, however, is the fact that the
method may be altered to accept other types of measurements for input by
merely changing the definition of the finction g and correspondingly, ij
the expressions for its derivatives, with no additional modification to
the balance of the procedure which in fact, constitutes the major portion
of the computing process.
The set of initial approximations for the position and velocity
components are defined for time to as (x0, yo, z0, Go, iO, iO). The
reference frame is the xyz-coordinate'system which is defined in the
following section. If second and higher order terms are omitted from the
Taylor expansion about the point (x0, yo, z , , , io), the equations
of condition may be mitten in matrix form,
where
for all values of i and j. Hence, J is a matrix of order (i. jx6), AV a
179
matrix of order (i* jxl), and m a matrix of order (6x1). Since there are
six unknowns, a minimum of six equations are required for a solution. In
practice, sufficient data are available to provide an over-determined system,
thus permitting the least squares solution,
m = (J* J)-~J*AV, (4)
where J* is the transpose of the Jacobian J. Finally, improved values for
the initial conditions are obtained from
x + m,. where
As a matter of convenience, no subscripts were introduced to indicate
iteration; but at this point, the improved values of X are used for the
initial point and the process is iterated until convergence is achieved.
Since the function g cannot be expressed readily in terms of X ij
directly, the evaluation is obtained implicitly. An ephermeris is computed
for the assumed values of the position and velocity vectors at time to.
Then using equation (1) gij may be evaluated for all values of i and j.
In the process of this evaluation, it is expedient to use two
rectangular coordinate' systems in addition to the Earth-bound system whose
origin is at the transmitting site. Referring to Figwe 60 these coordinate
systems are defined as follows.
Figure 59 - Problem geometry.
181
COORDINATE SYSTEMS
Figure 60 - Coordinate systems.
1. The xyz-coordinate system is a right-hand rectangular system
with the origin on the Earth's surface at the transmitter. The x axis
i~ positive south, the y axis is positive east and the z axis is normal
to the Earth's surface at the transmitting site.
2. The x'ylz'-coordinate system is a right-hand rectangular
system with the origin at the Earth's center. The xt axis lies in the
plane of the equator and is positive In the direction of the vernal
equinox, while the positive a' axis passes through the north pole. The
y1 axis is chosen 60 as to cmplete a right-hand system.
3. The x"y"z"-coordinate system is likewise a right-hand rectangular
system with the origin at the Earth's center. The xl'y" plane lies in the
orbital plane of the satellite, with the positive x" axis in the direction
of perigee. The positive z" axis forms with the positive z' axis an angle
equal to the inclination of the orbital plane to the plane of the equator.
The y" axis is chosen so as to complete a right-hand system.
The first step in this phase of the computation consiets of computing
values for the initial position and velocity components in the x'y'et-coor-
dinate system. This may be achieved by one rotation and one translation
which are indepenhent qf time, and a second rotation which varies with
time. Iat the notation ~ ~ ( c r ) indicate the matrix performing a rotation
through an angle a about the ith axis of the frame of reference such that the angle ie positive when the rotation is in the right-hand direction
and i is equal to 1, 2 or 3, according to whether the rotation is about the x, y or z axis respectively. The desired transformation follows,
where
Q j ) when Q 6 ( -0 ) the time derivative of R ( - 3 0 . = O0'
0 = the r i g h t ascension of the transmitting s i t e a t time t J - J'
fl r the geodetic l a t i t ude of the transmitting s i t e ,
p# z the radius vector from the Earth 's center t o a point on the
Earth 's sea leve l surface a t the l a t i t ude 6 , A = the difference between the geodetic and geocentric l a t i t udes
a t the transmitt ing s i t e .
.me next s tep i n the computation involves the evaluation of the follow-
ing o r b i t a l parameters: . .
a semi-major axis ,
e eccentr ic i ty ,
o E mean anomaly a t epoch,
i incl inat ion,
R E r i gh t ascension of the ascending node,
o 2 argument of perigee.
The evaluation of these o rb i t a l parameters i s obtained from the following n
equations :
* Derived by D r . B. Garfinkel, Ba l l i s t i c Research Laboratories, Aberdeen Proxring Ground.
184
where g is the mean gravitational constant and R is the radius of the
Earth, which is assumed to be spherical in the development of the
equations,
where
o = Eo - e sin Eo - nto , hl = ylil
0 0
Ztkr - X ~ $ , r % = 0 0 0 0 '
1 - = cos -5 , whereOTi zn, 3 - x = a(cos Eo - e ) ,
where 1 - ~ g n [k; + S(x;k2 - Y&
P = 2
9 -1 0 = (-1) cos N1 9
where 9 =
Having obtained values f o r a, e, a, i, n, and w, gij may be
evaluated f o r each time of observation. The f i r s t s tep in the computing
procedure consists of solving f o r the eccentric anomaly E i n Kepler's .I
equation,
The posit ion of the s a t e l l i t e as a function of time is determined i n the
x"y"z"-coordinate system. -
r~ = a(1 - e cos E
y" = r s i n f . j J .I
z" is zero according t o the def ini t ion of the xny"z"-coordinate system. .I
A transformation t o the x'y'z' coordinates can be achieved by three
rotat ions as follows:
Finally the position i n the xyz-coordinate system may be obtained by
two additional rotat ions and a t ranslat ion.
p s in A ,'!:I [ ~ ] - [ ~ ~ o s ~ - (381
Referr- t o (1 1, gij may be then exyresaed i n tern of the sa te l l i te ' s
positions a t time t and t . J 9+1
where (xi, yi, zi) i s the surveyed position of the i t h receiver. Finally,
the residuals Av may be determined from i j
Having evaluated the vector AV, there remains the problem of
Cie?e.mlni~g +he : T R c o ~ ~ R ~ .Tc necessary dif ferentlation may be carried
out numerically, but the computing time w i l l be reduced an& the accuracy
increased i f the derivatives are evaluated from analytical expressions. =,
Recalling that fo r all values of I and j,
l e t
where Bl0' " " gij, . . .
J 1 =( Xj+l' Y j + l ' Zj+l' , y , z j
The orders of the above matrices are (I* jx6) for J1, and (6 x 6) for J2, J and J,+. A distinct advantage of this method of evaluating J results 3 from the fact that the function g appears only in J1. Hence, if the
iJ solution is applied to other types of measurements, only J1 needs revision
lor the appropriate evaluation of J. This is trivial compared to the effort
required to derive the derivatives contained in J2 and J The expressions 3' for many of the elements of these Jacobian matrices are rather long and
involved, and therefore will not be presented here, but the complete results
are reported elsewhere. 5
D. Initial Approximations for the Single-Pass Solution
Convergence of the single-pass computation rests primarily upon the
adequacy of the initial approximations for position and velocity. It has
been established that, for a system consisting of a single receiver and
an earth-bow& transmitter at opposite ends of a 400 mile base line, conver- gence is assured when the error in each coordinate of the initial estimate
is not in excess of 50 to 75 miles and the velocity components are correct to within 112 to 1 mile per second. However, if single pass measurements
are available from two or more receivers, the system geometry is greatly
strengthened. Cocvergence can then be expected when the initial approxi-
lnations are within 150 to 200 miles of the correct value in each coordinate
and 1 to 2 miles per sxond in each velocity component. Larger errors may
occasionally be tolerated, but the figures presented are intended to specify
limits within which convergence may be reasonably assured.
Therefore, it hes 'ken rlecessary to develop a supporting computation
to provide relatively accurate initial approximations to position and
velocity for the primary computation. Several successful methods have been
developed for this phase of the problem; but discussion will be confined to
a few applications of a differential equation which approximately relates
the motion of the satellite to the tracking observations. The following
definitions will be useful in the derivation of this equation..
Ex;, 4, $1 _P tllft psitien of ths txaumaitting site.
(9, .jrl 0 ) = the pasitixm d' the atellib3 a t t h e tj. j 3
PA 2 the astan~l, fzm the +zammitting site to the
satellite at the tj.
; the distance Pnap the i th receiver t o * sate l l i te
a t th? t J * H E the Pltituda t b sate l l i te &we the Earth's
v = the valocitg of the Satellite.
To smliip the problem, certain a@flsaptiona have been made:
1. the satellite mom Fn a c h u l a r Keplerian orbit,
2. $ is ~ l a t i p e l y aaall throughout the period of obsemtion,
3. the Earth i s not rotating.
A number of useful relationships BBJ be derived as a result of these
a a s ~ t i o n s .
where H la constant.
The reiarenca fmw for thLe derivation ia the x"y"a"-aaxaiwte
-tea which has been defined ma. It fallowe frat the
definition of pj that
Miferentisting twice with respect t o t h e yields
which be almpllfied to
R-i 4
Z- . P
It followa that
A expression may be derived for p Recalling the defhition for gij, M C M C ~ U ~ ~ that
i J ' -
. 2 2 ! A - Pi, A - PJ
.\r B,, , P~
, "LJ
where
k t , us define a r i g h t - W d r e c ~ a c coordinate system, as
n h m L~I Figure 61, with the origin at the transmitter and the z-axis
posit ive i n the direct ion of the ver t ica l . The y-axis i s formed by
the interoaction of the tangent p lme &t the theranamitter with the plane
detem.ber! by the transmitter, the i t h receiver, and the Fkuth's center.
me yeceiver w i l l then be a t the known point (0, yi, zi). If the
variable p o b t (x , y , z ) indicates the poeition of the satellite, the J J J almt ranges f r a the t r a n d t t e r and the i t h receiver are respectively
given by
Prom which it follows that
In the three-beam mode of operation, the s a t e l l i t e w i l l be approximately
i n the YE-plane a t to, which i s aefFned as the time halfway between the
in i t i a t ion and termination of tracklng i n the center beam. Let the
s a t e l l i t e ' s posit ion and velocity a t this time be defined as (xo, Yo, zo)
and (ko, $o, i 0 ) , respectively. Obviously, xo may be approximated by
zero and we may safely assume that the ver t i ca l component of velocity i s small and can well be approximated by zero. The last pa i r of equations
then reduce t o
TO EARTH'S CENTER
GEOMETRY FOR OET ERMlNlNG THE INITIAL APPROXIMATIONS
Figure 61 193
Let fio and iio be the Doppler frequency and rate of change of frequency
fnr the i t h receiver a t to. It follows that
From equation (481 wc conclude
Fxpressing equations (55) and (56) i n terms of the position coordinates
a i d mlociei caqcntat:: of the =~t;-plli(P at time, to, yields
yo + (yo - yi)
J(y, - yiJr + (z o - z112
Let ue assume a speclfic orbi tal Inclination. With our previous
assmption of circular motion, ii may readily be cmputed as a function
of yo and zo. Then equations (57) and (58) w i l l lFlrewise provide fio
and iio a~ functions of position i n the yz-plane. Thus, fo r a given
inclination, families of curves may be computed and plotted i n the yz-plane
fo r both f,, and i,,,. Figure 62 presents such a plot, for an inclination A" J."
of 80°, with the transmitter and receiver separated by 434 miles and with
both located 35' off the equator. To at tain symmetry and s ~ l l f y the
construction of such charts, zi was assumed t o be zero, which is a
reasonable approximation fo r this approach t o the problem. If similar
charts are prepared for a number of inclinations, satisfactory i n i t i a l
appraximations may be rather quickly and easi ly obtained by the following
operations:
1. Assume an inclination. This, of course, is equivalent to selecting
a chart. Accuracy is not essential at this atage since the
estimate may be in error by 15' or more without preventing conver-
gence.
2. Enter the chart with the observed values of fio and Pie to determine an appropriate position within the yz-plane.
3. Approximate the velocity components. These should be consistent
with the assumption of circular motion, the height determined in
step 2, and the assumed inclination.
4. Determine the position and velocity components in the coordinate
system for the primary solutionby an appropriate coordinate
transformation.
In addition to the graphical method, a digital solution has been devised for eqs. (57) and (58). Aa in the previous development, we have two measw-
ments available and desire to determine three unknowns. In this approach,
one unlrnom is determined by establishing an upper bound and assuming a value
which is a fixed distance from this bound. The distance has been selected
to place the variable between its upper and lower bounds in a position which
is favorable for convergence of the primary computation. In this method,
we chose to start by approximating eo. It may be obeerved in Fig. 62 that,
for larger values of fio, the maximum value of zo occurs above either the transmitter or receiver while, for smaller values of fio, the maximum value
of zo occurs over the mid-point of the base line. The first step in the
computation is to determine a maximum value for zo. To this end, io is eliminated from eqs. (57) and (9) to yield an expression which varies only
in yo and zo. A appears in this expression, but it is also a function of these variables. The resulting equation may be solved by numerical methods for so with yo = 0 and then, solved a second time for zo with yo = (1/2)yj.
The larger of these results is to be used as a value for ( z ~ ) ~ which is
defined to be the maximum possible value of zo. Assuming the altitudes of
all satellites to be in excess of 75 miles, we may conclude, from the general characteristics of the family of curves for ? in Fig. 62, that the 10
TRANSMITTER RECEIVER Y - A X I S (EAST IN MILES)
DOPLOC FREQUENCY AND RATE OF CHANGE OF FREQUENCY AS A FUNCTION OF
POSITION IN THE Y Z - PLANE (FOR 80' INCLINATION
Figure 62 196
satellite's altitude will differ from ( z ~ ) ~ by no mre than 100 miles.
Since an error of 50 miles may be tolerated in the approximetion for
each coordinate, EzO)M - 5 q is a suitable value for zo. With the altitude thus determined, we may solve eqs. (57) and (58) for yo and
to. In the process, A and hence the velocity, will be determined. With i assumed as zero, ko may be readily evaluated to complete the
0 initial approximations which consist of the position (0, yo, zo) and the
velocity (io, io, 0). It is worth noting that there are a pair of
solutions for yo and $o. Further, the method does not determine the sign
of Po. If, in addition, we accept the possibility of negative altitudes for the mathematical model at least, we arrive at eight possible sets of
initial conditions which are approximately symmetrical with respect to
the base line and its vertical bisector. It is an interesting fact that
all eight, when used as input for the primary computation, lead to
convergent solutions which exhibit the same type of symmetry as the
approximations themselves., Of course, it is trivial to eliminate the
four false solutions which place the orbit underground. Further, two
additional solutions may be eliminated by noting that the order in which
the satellite passes through the three antenna beams determines the sign
of Go. In the two remaining possibilities, io is observed to have opposite signs. Since the y-axie of the DOPLCC system has been oriented
*om west to east, the final ambiguity may be resolved by assuming an
eastward component of velocity for the satellite - certainly a valid assumption to date. In any event, all ambiguity may be removed from
the solution by the addition of one other receiver. Moreover, this
would significantly improve the geometry of the system and thereby strengthen
the solution.
These two methods, for obtaining initial approximations, haw been
developed for a system which provides observations of the type displayed
in Fig.S%(b). If (very') minor modifications are made in the procedures, both
methods may readily be applied to data of the type presented in Fig. 5 8 ( d ) .
Indeed, with any tracking system that provides observations of satellite
w?lr?clty components, eq. (48) furnishes an adequate base for establishing an approximate orbit to serve as an initial solution which m y be refined
hy more sophisticated methods.
E. Multiple-Pass Solution
Single-pass orbit determination is unquestionably a primary requirement
for any satellite detection system. Of no less importance is the capability
of the system to generate suff'i~ient accuracy in the computed orbit to
permit pooitive identification and predictions which are adequate for later
acquisition. Imp~oved accuracy is most readily obtained by increasing the
number of observations and lengthening the time interval over which data are
caliected. Both of these objectives may be realized by simultaneously using
data fram several passes in an orbital determination. Further improvement
will result if the system may be expanded to include additional receivers.
The single-pass solution has already been programed to accommodate several
receivers; but it is inadequate for a solution with input data from more
than one pass. Therefore, a separate solution for dealing specifically
with t M s phase of the problem has been under development for sometime.
The derivation of the computing procedure is essentially complete; and in
fact, a few encouraging results have already been obtai.ned. However, the
method still requires the introduction of a few minor refinements as well
as extensive testing to verify its compatibility with a wide range of
input data.
The multiple-pass solution was designed to allow a major portion of the
computing procedure for the single-pass method to be used in the actual
computation. It is assumed that the satellite moves in an unperturbed, or
Keplerian, orbit for those short intervals of time that it is under obser-
vation by the DOPLOC System. Otherwise, the satellite is considered to
move in a perturbed orbit. The latter is determined by thirteen parsmeters
from which six time-varying Keplerian orbital parameters may be readily
derived. For a given period of observation, the Keplerian parameters are
determined for to, a fixed time within the tracking interval. These
parameters are then assumed t o determine the o rb i t f o r the few minutes
that the satellite remains w i t h i n the volume c k e g e of the DOPLOC System.
Thus, if data are available f o r Q revolutions, Q se t e of Keplerian o r b i t a l
parsrpeters, each constent within a given in te rva l of observation, w i l l be
expmesed i n terms of the thi r teen unknowns. The l a t t e r JII~Y then be
adjusted by a se r i e s of d i f f e ren t i a l corrections t o obtain a least-squares
f i t t o all observations f o r two o r more revolutions. No attempt has been
made t o determine par t icular perturbing elements euch a s drag. Rather, it
has been assumed t h a t an empirical determination of the thir teen unknowns
w i l l adequately r e f l e c t the t o t a l e f f ec t of a l l perturbations upon the orbi t .
To conserve computing time, it i s considered desirable t o use input that
consists of the integrated Doppler frequencies over time intervals of several
seconds as recamended fo r the single-pass solution.
Let the subscripts i a n d j be defined a s before. A t h i r d subscript,
q, w i l l be introduced t o sp&clfy t h e s a t e l l i t e revolution number. The
perturbed o rb i t a l are expressed i n terms of ah where h ranges
from one through thir teen.For theqth revolution, we define the unperturbed
parameters a s follows:
where
90 : the minimum value of q ,
toq 5 the time fo r which the unperturbed parameters are computed for the qth revolution,
- AYq : Tq -
, : y90 '
y9 the angular distance along the o rb i t from the equator t o the
satellite a t time t . oq
199
From eq. (62), we conclude that
The set of unperturbed paraetere for qth revolution is completed with the
camputation of a from the equations 9
7 may be evaluated with sufficient accuracy from the single-pass solution 4 far positian at the the, t 1 0 , oq, e , a m d a constitute a set
aq' q' q 4 9 'a of unperturbed parameters for each revolution q for which tracking observa-
tions are available. Uther, they are functionally related to %. Hence,
the observed data may be axpressed as a function of the set of parametera
5,. As in the single-pass method, the solution is obtained by using least-
squares methods to compute a series of differential corrections to initial
approximations for the set of parameters, 4. For each value of q, eqs. (32) through (40) may be used as before to compute the residuals, aVijq. The least-squares solution is obtained from the matrix equation
where AV and Crr are column matrices whose element8 are respectively the
residuals and the differential corrections for the set of unknowns ~ 6 . - J is a Jacobian defined as follows:
J = ,&) , - . . , . . gi . . . . . Jq =pi2, . . . . . . !q.,. . 5 3 (68) where
The definition for g is similar to Chat for g of the single-pass iJs ij
solution. Let
where
It will be noted that Jlq and J have definitions which are similar to 2q
J and J2 of the single-pass solution. Hence, the analytical expressions 1 for the elements of these matrices are identical to those for J1 and J2
which have been reported upon previously? The elements of Ja have not
been published but are easily derived, and therefore, will not-be presented.
To summarize, initial appraximations for q, are derived from the aingle-paas solutions for each value of q. We assume the perturbed param-
eters a eq, oq, co , 0 and i can be adequately expressed in terms of 9' 9 9
% by eq. (59) throw 763). Further, it is assumed that they remain
conetant during each short period of observation, but otherwise vary in
accordence with this same set of equations. Hence, we have sets of Keplerian
parameters for each period of observation. They are used to compute the
elements of the AV matrix for the least-squares solution to determine
appropriate differential corrections for the estimated values of ah. The
process is iterated until convergence is achieved. The resulting set of
provide perturbed orbital parameters ae a function of time and revolution number. The results may also be used to predict the time and
position of future passes of the satellite as an aid in later acquisition.
F. Computational Results - Numerous convergent solutions have ljgen obtained with both simulated
and actual f i e l d data oerving as computer input. Since the l a t t e r are of
major in te res t , -the discussion w i l l be r e s t r i c t ed t o r e su l t s obtained from
r s a l data. The present results were obtained from observations which were
recarded during a period when the WSIa receiver was inoperable and, there-
fore , are derived from data recorded by a single receiver. Included with
the DOPLOC results m e o r b i t a l parameters which were determined and pub-
l i shed $;y the National Space Surveillance Control Center (Space Track). As
an a i d 4.n camparing the two s e t s of determinations, the Space Track param-
e t e r s have been converted t o the epoch times of *he DOPLOC reductions.
The i n i t i a l s~iccessfdl reduction f o r the Fort S i l l , Forrest City system
was achieved for Revolution 9937 of Sputnik 111. The DOPLE observations,
ao well a s the r e su l t s , are presented i n Fig. 63. hkasurements were
recorded fo r 28 seconds i n the south antenna beam, 7 seconds i n the center
beam, end 12 seconds i n the north beam with two gaps i n the data of 75 seconds each. Thus, observations were recorded f o r a t o t a l of 47 seconds
within a time in t e rva l of 3 minutes and 17 seconds. Using the graphical
method described i n the previous section t o obtain i n i t i a l approximations,
convergence was achieved i n three i t e r a t i ons , on the f i r s t pass through the
computing machine. In comparing the DOPLCC and Space Track solutions, it
w i l l be noted t h a t there i s reasonably good agreement i n the values fo r a ,
e , I, and 0, par t icu la r ly f o r the l a t t e r two. This i s charac te r i s t ic of
the single-pass solution when the eccent r ic i ty i s small and the computa-
t i ona l input i s l imited t o Doppler frequency. Since the o rb i t i s almost
c i rcu la r , a and u, are l e s s s ign i f ican t than the other parameters and l ike-
wise, a r e more d i f f i c u l t f o r e i t he r system t o determine accurately. How - ever, as a r e s u l t of the small eccent r ic i ty , the sum of u, and o i s a ra ther
good approximation t o the angblar distance along the o rb i t from the nodal
point t o the posit ion of the s a t e l l i t e a t epoch time and a s such, provides
a bas i s of comparison between the two systems. In the DOPLCC solution, 0
(u, + a ) = 32.66' while the Space Track determination yields a value of 35.73 ,
I I
--tRBO -- --
Ni$7AN+Nru I -
ARPA-BRL DOPLOC DOPPLER RECORD OF Jurrr ltym I
58 DELTA REV 9937 FORREST CITY. ARKANSAS I I - - -p- . - - - . - -p- . - - . . . - . - ( 1 I I I
I I B n
~ 0.. . . . ~ ~ ~ ~ ~ . ~~ ~~ ~ ~ ~ - . 1 ; ..(w+-+7L -r DOPLOC 4149mi 0.0153 65.37' 178.24O 104.62' 288.04' 1 132.66.
-?. ~ E - . T R M ; ~ - - ~ ~ 41 1 I m i 0.0130 65.W0 178.22O-- . t37.9S0 257778'--I---- -- f - 35-7SC- I
DIFFERENCE I f
381111 0.0023 0.31° 0.02. -33.33. 30.26. 1 I
1 -3.07. - . - - ~ . ~ . ~ . . . ~ ~ I
1 I ~ ~ . .
I I I
- 0 0 . .- .. ~ ~ ~ - 1 I ~. . 0 0 .I I I
I I 77Wb -~~ :' CSKE!? &??TEHY& I ! ; . . - ~ - ~
Z w
. l SLOPE 89 CPS/S I 3 I I
7200 I 1 I I IV K
I Y I j I I
IS) I I 6700 ny I 1 I
A a I I I I I a I I
I I 6200 g I I
I I I I 2 8 I 7 5 ' 7-75-12J
5700 I I
MEASURED 0643:OO Z. PREDICTED 0 6 4 2 Z AiTiTUDE is6 Yi iES. 372 Yi iES EAST Fi. S i i i SOUTH -CENTER - NORTH ANTENNAS. S - N PASS BFO 1 0 8 , 0 0 7 , 0 0 0 CPS TRANSMITTER 108.000,000 CPS
i - 3200.- . ... .. .. . (.+ -. . .. .. ~ . . . -
@.+--- - /" SOUTH ANTENNA - 1 - . .. sLom 23 CPSE . . ~~. . -. - . .... - . ~ - ~ . . . - .
. ~~ . ~~ .. Z . TIME . . . -. . - . :s4. 31 MAR. osii:m 60 :,, - .. - -
I 2 4 :Y :44 :54 0 6 4 2 : ~ :I4 : 24 3 4 :44 2 4 :34 144 :S40644:W:14 :e4 :H :44
Figure 63
a differeuce of 3.07' between the two sets of results. To summarize, when
lFmitqd to single-pass, single-receiver observations, the DOPLOC system
provides an excellent determination of the orientation of the orbital plane,
a good determination of the shape of the orbit, and a fair-to-poor determi-
nation of the orientation of tt-e ellipse within the orbital plane.
Occasionally, exceilent results have been obtained for both a and u;
but in general, the interim DOPLOC system with its present limitations fails
to provide consistently good single-pass evaluations of these two quantities.
Therefore, in presenting the remaining single-pass DOPLOC reductions, a and o, have been eliminated from further consideration. Results have been indicated
in Table I for six revolutions of Dihcoverer XI, including number 172 which
was the last known revolution of this satellite. As a matter of interest,
the position detenined by the interim DOPLOC system for this pass indicated
an altitude of 82 miles as the satellite crossed ihe base line 55 miles west of Forrest City. To provide a basis for evaluation of the DOPLOC results,
orbital parameters, obteined bj converting Space Wac& determinations to the
appropriate epoch times, haye been included in the table. Table I also
contains a listing of the amount of data available for each reduction in
addition to the total time interval within which the observations were
collected. The observations recorded for the Fort Sill, Forrest City complex
are plotted in Fig. 64 for the six revolutions of Discoverer XI which are
surmnarized in Table I. In Fig. 65, results are tabulated for a reduction
based on only seven frequency observations from a single pass over the
interh DOPLM: system. These have been extracted from the complete set of
observations previously presented for Sputnik 111. They were selected to
serve as a crude example of the type of reduction required for the proposed
DOPLOC scanning-beam system. The example shows that the method is quite
feasible for use with periodic, discrete measurements of frequency. Of
course, the proposed system would normally yield several more observations
than were available in the example.
There has not been sufficient time thus far to completely evaluate the
computing methods for the multiple-pass solution. However, one set of
reeults has been obtained for observations from revolutions 124, 140 and
TIME - SECONDS
Figure fA
TINE - SECONDS
Figure 65
t + .
0 I
l70.QS0 178.2Z0 -0.13O
A.
+
0
1 DOPPLER FREQUENCY EXTRACTED FROM ARPA- BRL DOPLOC RECORD FOR
0.0221 65.36. 0.01 58 65.06O 0.009 1 0.30.
1958 DELTA REV. NO. 9937
TRANSMITTER AT FORT SILL RECEIVER AT FORREST CITY
a
4105 ml. 4 I I l ml.
-6 mi.
100 SEC.
WPLOC SPACE TRACK DIFFERENCE
7
+
+
DOPL(X: Space Track
Difference
Difference
DOPLOC Space R g c k
Dlff erence
Daaoc Space aaclr
Dlfference
DOPLOC Space Track
MfY erence
D O P m Space Track
'6 Dlfference --.
Comparison of Single-pass DCPLOC snd Space Ilack Results for Discoverer XI
Total Amount Interpal of Revolution a e 1 n of Data (Ibeervation .Number (miles) (dew=-) (aegrees) (seconds) ( e c m )
156 of Discover XI. These are presented in Table I1 where, as before, 8
comparison is made with Space Track results for the same re~0l~ti0nS.
*era11 agreement is excellent, and in general, the differences between the
two determinations are of the same magnitude as the error to be expected
in either' set of parameters. Disagreement is excessiye only in a, the mean
anomaly at epoch. For this particular parameter, the major portion of the
error must be in the Space Track results, for the DOPLOC reduction is
designed to force the satellite's calculated position to lie in the narrow
east-west vertical beam of the tracking system at the actual tima of
~bservation. The resulting error in the DOPLOC determination of satellite
passage is certainly less than one second. Hence, the angular error in
position is relatively insignificant.
As an indication of the feasibility of using the multiple-pass
solution for prediction, the DOPLOC results in Table I1 were used to
compute the times of satellite passage through the mid-point of the
vertical beam for revolutions 165 and 172. These results were compared to
the observed thes. The predictions were found t.0 be two minutes early
for 165 and seven minutes early for 172. Since the latter was the final
revolution for this satellite, these predictions covered a period in which
drag was increasing rapidly and extrapolation was rather hazardous. These
results are considered reasonable for this portion of the orbit and would
certainly have been adequate as an aid in acquisition.
G. Conclusion
These methods of solution have been shown to be both practical and
useful by numerous successful applications with real as well as simulated
data. Although results have been presented for passive data only, the
computing procedure has been altered slightly and applied to active data
with considerable success. Computing times are reasonable since convergent
solutions have required from 20 to 40 minutes on hhe ORDVAC with the coding
in floating decimal, whereas more modern machines would perform the same
computation in a fraction of a minute. These methods allow the determi-
nation-'of a relatively accurate set of orbital paremeters with as little - %..
DOPLOC Space Track
Difference
Space rack 0 \O
Difference
DOPLOC - Space hack
Difference
Camparison of Space 'R9ck and Multiple-pass DOPLOC Results for Discoverer X I
Revolution a e i R a 0)
Number (miles) /
(degrees) (degrees) ( d e p e s ) (degrees)
as 1.5 to 3 minutes of intermittent observations from a single receiver
when the signal source is a ground-based transmitter. While multi-receiver
systems provide numerous distinct advantages, it has been shown that it is
possible to obtain, routinely, quite satisfactory results with observations
from a single receiver. However, it should be noted that a system with two
receivers will reduce error propagation in the computations to approximately
one tenth of that to be expected from a single receiver system.
&tension of the computational methods to include multiple-pass
solutions results in a well-rounded and complete computing program for
satellite detection and: identification. Not only are the orbital elements
much Setter determined, but the increased accuracy is sufficient to permit
adequate prediction for future acquisition.
It should be emphasized that solutions have been obtained for actual
field data, and further that such solutions were completely independent
of other measuring systems. Results from the latter were used only for
comparison and did not enter any phase of the computations leading to these
solutions.
In conclusion, the methods are quite general and need not be confined
to either ~oppler measurements or Keplerian orbits. It is merely necessary
to modify J1, J , g.. and g.. to allow these methods to be applied to 1J IJq
observations from any satellite or ICBM tracking system.
V I I I . POTENTIAL OF DOPLOC SYSTEM IN ANTI-SATELLITE DEFENSE
If it can be assumed that within a reasonable period of time the
population of earth bound artificial satellites will reach the predicted
number of 10,000, it csn be assumed with equal certainty that any adequate
system of satellite detection, tracking and orbit cataloguing equipment
will utilize the Doppler principle and at least some Qf the DOPLOC tech-
niques. DOPLCC was certainly the first and probably is still the only
system to have demonstrated its ability to consistently compute orbits
within minutes after first detection by a single station from single pass
data. The performance capability of the DOPLCC system is equal to that
of any other known system. Continuous wave bi-static Doppler Radar and
narrow band reception techniques combine to provide a uniquely economical
utilization of the physically limited power and receiver sensitivity. The
DOPLOC system has its greatest potential as a surveillance system for
cataloguing and identifying satellites. Its application as an anti-satellite
or ICBM defense system is limited in its present form. Either as a surveil-
lance system by itself, or as a member of a family of sensors, it offers
a low cost dependable method of detecting and cataloguing satellites over
an extremely large volume of space. With modifications to the scan
configuration the system may have application to the ICBM and the anti-
satellite defense problem. An early warning feature would need to be
incorporated since DOPLOC is presently effective primarily after one pass
over the sensor.
The author wishes t o ac1;nowledge t h a t most of the work done under the
DOPLCC pro jec t (ARPA Order8:-58) was directed and supervised by Mr. L. G.
de Bey, recent ly deceased. A s indicate& by the t i t l e , t h i s report i s
intended t o be a technicai summary covering the e n t i r e project . The material
p~eeen$ad was therefore compiled, from various reports and eourcee and repre-
sents the work of most of the technical s t a f f of the Electronic,kasurements
Branch of t he B a l l i s t i c MEasureroent Laboratory over t he period July 1958 - 30 June 1960. Section V I I , "Sa t e l l i t e Orbit Determination from DOPLOC Data"
was authored en t i r e ly by M r . R. B. Patton, Jr. Others who made major contri-
butions t o the program were V. W. Richard, C. L. Adams, K. A. Richer,
K. P.. P l l e n , J r . , R . E. A. Pntnum, Vitekj FI. Lnotenn, C : ShaPP-r and
K. Patterson. Material extracted from the work of contractors i s referenced
i n the tex t .
A . H. HODGE
REFERENCES
1,. Muon, L. A. Variation of Cosmic Radiation with Frequency. Nature, Vol. 158, No. 4021: 758 - 759, November 1946.
2. Cattony, H. V. and Hohler, J. R. Cosmic Radiation in W e V.H.F. Band. Proc. IRE: 1053 - 1060, September 1952.
3. Gabor, D. Theory of Communications, Part I. Journal of Institute of Electrical Engineers: 429, 1946.
4. Woodward, P. M. Probability and Information Theory with Applications to Radar. Permagon Press, 1955.
5. Patton, R. B., Jr. A Method of Solution for the Determination of Satellite Orbital Parameters from DOPLOC &asurements. Aberdeen Proving Ground: BRL MR 1237, September 1959.
1. Patton, R . B., Jr. Orbit Determination from Single Pass Doppler Observations. IRE Trat~ssctions on Military Electronics, Vol. MIL-4, Numbers 2 and 3 : 335 - 344, Apri l - July, 1960.
2 . Pnt$onn, R. B., Jr. A Method of Solution f o r the Ihtermination of ;Sa t e l l i t e Orbital Paraneters from CCPLCC Measurements. Aberdeen Proving ,Gr.omd: BRL MR 1237, BRL DOPLCC S a t e l l i t e Fence Series, Report No. 7 i n the Ser iesp September 1959.
3. Guier, W. H. and Weiffenbach, G. C . Theoretical Analysis of Doppler Hadio Signals from Earth Sa t e l l i t e s . Applied Physics Laboratory, S i lver Spring, . : Bumblebee Report No. 276, April 1958.
4. Richards, Paul E. Orbit b te rmina t ion of a Non-transmitting S a t e l l i t e Using Doppler Tracking D a k . Space Sciences Laboratory, General E lec t r ic Company, Philadelphia, Pa.: BRL DOPLOC S a t e l l i t e Fence Series, Report No. 12 i n the Ser ies , October 1960.
5 . Besag, P. L. Polystation Doppler System. Western Development Laboratories.,. Philco Corporation, Palo Alto, Cal i f . : BRL DOPLCC S a t e l l i t e Fence Series, Report No. 11 i n the Series, May 1960. c
6. Davis, R . C. Techniques f o r S t a t i s t i c a l Analysis of Continuous-Wave Doppler Data. U. S. Naval Ordnance Test Stat ion, China Lake, Cal i f . : NAVORD Report 1312, April 1951.
7. Maulton, F. R . An Introduction t o Celes t ia l Mechanics. New York: Macmillan and Company, Ltd., 1902.
8. Richard, V. W . DOPLCC Tracking F i l t e r . Aberdeen Proving Ground: BRL MR 1133.
9. Bolgiano, L. P., Jr. Doppler Signals and Antenna Orientation f o r a DOPLOC System. Aberdeen Proving Ground: BRL R 1055, BBL DOPLCC S a t e l l i t e Fence Series, Report No. 1 i n the Ser ies .
10. Maxon, L. A. Variation of Cosmic Radiation With Frequency. Nature Vol. 158, No. 4021: 758 - 759 , November 1946.
11. Cattong, H. V. andJohler , J. R. Cosmic Radiation i n the V. H. F. Band. Proc. IRE: 1053 - 1060, September 1952.
12. Balgiano, L. P., J r , and Gattschalk, W, M. Quantum Mechanical AnaLyeis of Radio Frequency Radiation. E lec t r ica l Engineering Dept., University of Delaware, Newark, Del.: BRL DOPLCC S a t e l l i t e Fence Ser ies , Report No. 13 in the Ser ies , 15 June 1960.
BIBLIOGRAPHY ( m n t . 'd)
13. Lootene, H. DOPE Observations of Reflection Croes Sections of S a t e l l i t e s . Aberdeen Proving Ground: BRL R 1330, BRL DOPLOC Satellite Fence Series, Report No. 22 i n the Series.
14 . Scharfman, W . E., Rothman, B. , Guthart, A. and Morita, T. DOPLOC System Studies, Stanford Resemch Ins t i t u t e , Menlo Park, C a l i : ' Final Report Par t B, BRL DOPLOC S a t e l l i t e Fence Series, Report No. 10 i n the Ser ies .
15. Adams, C . L. DOPLOC Instrumentation System f o r S a t e l l i t e Tracking. Aberdeen Proving Ground: BRL R 1123, BRL DOPLOC S a t e l l i t e Fence Series, Report No. 21 i n t he Series.
16. Summary of the Preliminary Study of the Applicabil i ty of the Ordir System Techniques t o the Tracking of Passive S a t e l l i t e s . School of Engineering, Columbia University: Final Report No. ~1157, BRL DOPLOC S a t e l l i t e Fence Series, Report No. 1 4 i n the Ser ies .
17. de Bey, L. G . , Richard, V. W., Hoage, A. H., Patton, R. B. and Adams, C . L. F i r s t Semi-Annual Technical Summary Report for the Period 1 July 1958 - 31 December 1958. Aberdeen Proving Ground: BRL MR 1185, BRL DOPLOC S a t e l l i t e Fence Series, Report No. 2 i n the Series, January 1959.
18. Putnam, R. E. A. Orbital Data Handling and Presentation. Aberdeen Proving Ground: BRL TN 1265, BRL DOPLOC S a t e l l i t e Fence Series, Report No. 3 i n the Series.
19. de Bey, L. G. An Approach t o the Doppler Dark S a t e l l i t e Detection Problem. Aberdeen Proving Ground: BRL TN 1266, BRL DOPLOC S a t e l l i t e Fence Series, Report No. 4 i n the Series, July 1959.
20. de Bey, L. G., Richard, V. W. and Patton, R. B. Second Semi-Annual Technical Summary Report f o r the Period 1 January 1959 - 30 June 1959. Aberdeen Proving Ground: BRL MR 1220, BRL DOPLOC S a t e l l i t e Fence Series, Report No. 5 i n the Series, July 1959.
21. Putnam, R. E. A. Synchronization of Tracking Antennas. Aberdeen Proving Ground: BRL TN 1278, BRL DOPLOC S a t e l l i t e Fence Series, Report No. 6 i n the Ser ies .
22. Fullen, K. A. The Eynamic Characterist ics of Phase-Lock Receivers. Aberdeen Proving Ground: BRL M R 1093, BRL DOPLOC S a t e l l i t e Fence Series, Report No. 8 i n the Series, March 1960.
23. Scksrfman, W. E. , Rothmsn, H. and Morita, T. Sta t ion Geometry Studies f o r t he DOPLOC System. Stanford Research Ins t i t u t e , h n l o Park, C a l . : F inal Report Par t A, BRL DOPLOC S a t e l l i t e Fence Series, Report No. g i n t he Series, SRI Control No. FO-0-93, July 1960.
BIBLIOGRAPHY (~ont. 'd)
Dean, W. A . Precision Frequency Measurement of Noisy Doppler Signals. Aberdeen Prcving Ground: BHL R 1110, BRL DOPLOC Satellite Fence Series, Report N3. 15 in the Series, June 1960.
Third Technical Summary Report for the Period July 1959 thtough:June 30, 1960. Aberdeen Proving Ground: BRL MR 1287, BRL DOPLCC Satellite Fence Series, Report Xo. 16 in the Series.
Kreindler, E. The Theory of Phase Synchronization of Oscillators with Application to the DOPLW Tracking Filter. Columbia University: TN T-1/157, BRL DOPLOC Satellite Fence Series, Report No. 17 in the Series, August 1, 1959.
Patterson, K. H. The DOPLOC Receiver for Use With the Circulating Memory Filter. Aberdeen Proving Ground: BRL TN 1345, BRL DOPLOC Satellite Yence Series, Report No. 18 in the Series, August 1960.
Patterson, K. H. Parametric Pre-Amplifier Results. Aberdeen Proving Ground: BRL TN 1345, BRL DOPLOC Satellite Fence Series, Report No. 19 in the Series, October 1960.
Vitek, R. A Comb Filter For Use in Tracking Satellites. Aberdeen Proving Ground: BRL MR 1349, BRL DOPLOC Satellite Fence Series, Report No. 24 in the Series.
Lootens, H. T. Satellite-Induced Ionization Observed With the DOPLOC System. Aberdeen Proving Ground: BRL MR 1362, BRL DOPLOC Satellite Fence Series, Report No. 23 in the Series.
:de Bey, L. G., Berning, W. W., Reuyl, D. and Cobb, H. M. Scientific Objectives and Observing Methods for a Minimum Artificial Earth Satellite. Aberdeen Proving Ground: BRL R 956, October 1955.
32. Patterson, K. H. The Design for a Special Receiver for Satellite Ionosphere Experiments. Aberdeen Proving Ground: BRL TN 1158, April 1958.
33. Patterson, K. H. A Pre-Amplifier Design for Satellite Receivers. Aberdeen Proving Ground: BRL MR 1154, July 1958.
34. Katz, L. and Honey, R. W. Phase-Lock Tracking Filter. Interstate Engineering Corp., Anaheim, Cal.: Final Engineering Report for BRL, 22 M Y 1957.
35. Bentley, B. T., Prenatt, R. E. and de Bey, L. G. Summary of Doppler Satellite Tracking Operations for Soviet Satellite 1957 Alpha 2 and 1957 Beta. Aberdeen Proving Ground: BRL MR 1174, October 1958.
BIBLIOGRAPRY (~ont . 'd) Bentley, B. T., Prenatt, R. E. and de Bey, L. G. Summary of Doppler Satellite Tracking Operations During the First Week After 1958 Alpha Launch. Aberdeen Proving Ground: BRL TN 1224, October 1958.
Instrumentation for Tracking Satellite Ionospheric Payloads. Ballistic Measurements Laboratory, Aberdeen Proving Ground: BRL TN 1267, July 1959.
Cruickshank, W. J. Instrumentation Used for Ionosphere Electron Density Measurements. Aberdeen Proving Ground: BRL TN 1317, . k y 1960.
Zancanata, H. W. ~alliitic Instrumentation and Summary of Instrumen- tation Results for the IGY Rocket Project at Fort Churchill. Aberdeen Proving Ground: BRL R 1091, January 1960.
Marks, S. T. Summary Report on BRL-IGY Activities. Aberdeen Proving Ground: BRL R 1104, April 1960.
Radio Receiver, R-~~~A/uRR. Department of the Army Technical Manual TM - 85&, Jan~ary 1956.
Phase-Lock Tracking Filter Model 1V. Operating Manual, Interstate Electronics Corporation, Anaheim, Cal. : October 1958.
Radio Receiver, Model 2501. Operating Manual, Nema-Clarke Company (subsidiary of Vitro corporation), Silver Spring, Ma.
Fre-Amplifier, Model Pr-203. Operating Manual, Nems-Clark Company.
Frequency Synthesizer, Rhode and Schwarz Type XUA, BN - 444463, Operating Manual.
Schwab, D. and Taneman, H. D. 108-M: Precision Frequency Generator. U. S. Army Signal Research and Development Laboratory, Fort Monmouth, N. J. : USASRDL Report No. 2166, November 1960.
Department of the Army Washington 25, D. C.
1 Commanding Officer 1 Diamond Ordnance Fuze Laboratories ATTN: Technical Information Office
Branch 0l2 Washington 25, D. C.
2 10 Commander
Armed Services Technical Information Agency
A'ITN: TIPCR Arlington Hall Station Arlington 12, Virginia
10 Cammander Air Force Systems Command ATPN: SCTS Andrews Air Force Base Washington 25, D. C.
1 Ccmunander Electronic Systems Division L. G. Hanscom Field Bedford Massachusetts
The entagon Washington 25, D. C.
Commanding General White Sands Annex - BRL White Sands Missile Range New Mexico
Commanding General Army Ballistic Mlesile Agency AWN: Dr. C. A. Lundquist
Dr. F. A. Speer Redstone Arsenal, Alabama
Director Advanced Research Projects Agency Department of Defense Washington 25, D. C.
Director National Aeronautics & Space Administration
1520 H Street, N.W. Washington 25, D. C.
1 Coimnander Air Proving Ground Center A m : PGTRI E g i i n Aii' Force Base, Flori&&
- .-. - - - - .- --. .-. .- ~. p r y l ~ ~ s g p o S ~ c d ~ ~ 1 9 5 8 - Pnul B Y X U E A L KERlRT ' Rriod 20 JM 1958 - J o ~ l % l B B I d B P d ~ ~ ~ ~ C O W P I E I 30 Jlme 1961 -A DJPLOC sAZELLI1S ISVSCTIM COI(PIEI[
AEBlSatelUtuleaeSeNs r e p n - t m . 8 i n t a s 8 e r i ~ ~ ARPA satellite Rncc Series RCpcDt 110. 4 in the S v i e ~ A. H. Bstellltca - Detst i rm A. 8. Hcdm & U t e s - Dctcetirm
AD -1m b. LlEIAS3lFmD AD Accession no. InCIASSIPIED
AD dscarsia Ilo. uuxaEam A,,=ession W. Ballldicbaea~ch-- U s t i c Research w t m i c e --smUWrElUUtf IcrioaZQJanc1958- 1 r' PnullEmDKc&aanaRI- R r i o d 2 0 J l m e l 9 5 8 - 3 3 ~ 1 9 6 l - o o P u x : M e n r s n r l E ¶ m m E C c l r w E x 30 Jlmc 1961 BRL-ARPA DOFl.CC -08 C w A I m 6 a ~ ~ ~ ~ r e p n - t . 0 . 2 5 i n t h S e r i c s ARPA Sate l l i te Pence Series Report Ila. 4 in th & N s A. E. Bbdeo 8ateu.tt.e~ - atteetion A. x. Hodgc satellites --l - Qtection
rmuistlc - - Ballistic ucvlrch 1Tab-tarie. I ' - -- -
8ar.BePn-tEa.46 Wl96l Ex& gte2- m ROJ m. y~3-06-oll, QB Cod+ m. Plo.ll.143 D=s* ---
m m - i s t s d a t e r h n i u r l - o f t b e r e d a n d d e v c l ~ p ~ m t rmdcr Aim (Wn a58 vhich led to tbe pmposal of a system for
dew* cud tncLine -ndi.tiag orbiting satel l i tes. An interim research hut, of 81 Uh&mtb transmitter and two receiving statione with flnd en- - rss first established to detelmine the feasibil i ty of using tar - prineiplr - with d r e n l y narrow bandvidth phase-locked tracking film as a scnsm satem. interim system established the feasibil i ty of miag tbe Ibpplu nrtbDd apd -tad in real data far which computational mcthode wen dne- which aoQna mbital parameters n-m s W pass data aver a ein&le station. m ret tbc -tad s- population problem of detection and ldentlficatim ard to a a m a d m m range with p r a e t i a l emitted power, a
fm beam saten - proposed in - a trmsdtter and a receiver a n t e m -ted by - Im, ailcs soula be -ly scarmed about the earth chord J- tba aa at ions to - a u CJ- of saee vahme i0a) miles lous aPb2mvi.lEa-cU3omnFles.
- 8ar.RoPlPtm.46 W196l - - satellib m ~cpmt BO. uJ6 JulJ 1961
Radar - Sate l l i tes DrrpplutracLinssyatema- m 1 e r tmcnlu system -
nn RUJ m. ~o~-c~-ou, am c=de m. 5uo.ll.10 D=3* M ~ o j m. 5 0 ~ - 0 6 - ( ~ 1 , as codt m. mo.n .143 Den* W B C p l P L UEIASsIPIED BePn-t
- ERL Repor t So. llJ6 JulJ 1%1 Radar backiw - sa&tes
mlu tracuse SF- - M proj NO. ~ o J - o ~ o ~ , 016 codt W. zlo.n.143 Den* u r n -
Zhis report consistll of a tec4olcal of th research and d m l w pmgram sponsored under ARPA CTder 8-58 which led to the pr- of a system far detecting md treeking non-radiating orbiting sa te l l i tes . An interin research f ac i l i t y consisting of an illmimating tmmdtter and two rea iv lng stations with fixed antenna arrays -as f i r s t established to determine the feas ib i l i ty of using the Doppler principle cmpled with extremely narrw bandwidth phase-1- tnrebin( f i l t e r s as a sensor aystem. The interim system established the f e a s i b U t y cb using the Doppler method and resulted in real d a t a Xcn which ccmputatian~l metho& were developed which produced orbital parameters *am si&e psss data over a single station. To meet the expected space population problem of detection and identificaticm and to obtain a mar- range with -tical mlttcd p e r , a scanned fan beam syntem -as proposed in which a tmamdtter and a receiPcr an- separated by about 10a) miles vovld be s~nebr~nously scanned about the carth ehmd joinins t h e two stations to -P a bsli cy-r or spw r ~ l t p ~ 10a) riles lols and ha* a -us of 3COO miles.
-
l Q U r r p m + - i s t s d a t e c h k n l ~ o f t b c r e s e a r c h a n d d n r r l D p m m t propam rmdv AEBl (Wn 8-56 which led to the propxal of a system fcn dew- and --a orbit* satel l i tes. An interim rese-h h u t y -isting d an lll-ullg M t t u and tuo receivFng stations with - m- - - iirsr ~ l l s h e d to &- the I-mfiity of using tbc PoppLr prtmziph - with -4 Darros haivldth rhaue-loebcd tmckjng rum as a aemax -tea. The i n t e r i m s a t e n est&m&d the fuuribillty cb us- tas n~pplcr - and -tea in rrsl aata for which computatirmal mthods mere dmlopa which aoQna orbital pmmetue fx-m single p s data opu a single station. m ret tbc eqected space poplstim @Icm cb detection and idcmtiiieatim ard to obtain a rezse vltb suactical emittad m, a sewmed 1- bs8 qatP - pcp3sd in a t a a u a d t t a d a receiver antem4 sepanted by aaoot lorn dlcs mmld be scanned about the earth chrPd J- tas aa statlms to nrpp a bsli EJ- of space wahm 1000 miles lcms cud2mvi.lEamzUnadJOQ)ILlcs.
5m rrport carslsts of a technical mmary d the research and &d-t pmgran m o r e d lmdv ARPA (Wer 8-58 which led t~ the proposal cb a m t e m for detecting and tmckiq nm-radiatiag orbiting ~atellltes. ~n interim re-ch fac i l i ty ccmsisting of an illumlnatiag traPsmitter end two reeeirb@ stations with r i n d ~ ~ n t e m a arraJs vas first established to dcte-• tbe f i eas ibu ty of using the mlu principle CmPlCd atb -4 -0y bandwidth phase-lacked trmking film as a sensor system. 'Ih interim Byatem established the feas ib i l i ty of using th Doppler metbod and rrsulted in real data rm which ccqutatimml rcthods YM dewlOpCd which orbital paramctus tr(l single pass data orrr a sin& station. m meet the u p c t e d nzacc popvlatlrm @la of detection and ~ t i f l e a t i m aob to Dbtain a mslmm with m-aetical d t t ed -, a scannd fan hem -ten uae PnmC553 in which a ta'8bmdtte.r and a r r e e i n r antema -tcd W a b d 1000 miles mmld be -4 scanned about tba - chord Joining tba two Stationa to -P a u OX space vo1- lQlO riles 1- and har insa rad ius af JOOOnFles.