DTIC ADB010373: Telemetry System Parameters and Bit Error Performance of NRZ and DM PCM/FM

Survival, Water, Medical Field Manuals

Military Manuals

Defense Technical Information Center

Document text

UNCLASSIFIED 


AD  NUMBER 

ADB010373 

LIMITATION  CHANGES 
TO: 

Approved  for  public  release;  distribution  is 
unlimited. 


FROM: 

Distribution  authorized  to  U.S.  Gov't,  agencies 
only;  Test  and  Evaluation;  29  MAR  1976.  Other 
requests  shall  be  referred  to  Pacific  Missile 
Test  Center,  ATTN:  Code  4250,  Point  Mugu,  CA 
93042. 


AUTHORITY 

PMTC  ltr  dtd  10  Apr  1986 


THIS  PAGE  IS  UNCLASSIFIED 


I 


I 


TELEMETRY  SYSTEM  PARAMETERS 
AND  BIT  ERROR  PERFORMANCE 
OF  NRZ  AND  DM  PCM/FM 

(AIRTASK  A5355352  054D  5W47410030, 
Work  Unit  A5352 10000002) 

By 

D.  A.  KING 

Instrumentation  Development  Division 


29  March  1976 


DISTRIBUTION  LIMITED  TO  U.S.  GOVERNMENT  AGENCIES  ONLY; 
TEST  AND  EVALUATION  INFORMATION;  29  MARCH  1976.  OTHER 
REQUESTS  FOR  THIS  DOCUMENT  MUST  BE  REFERRED  TO  THE 
COMMANDER  (CODE  4250),  PACIFIC  MISSILE  TEST  CENTER, 
POINT  MUGU.  CALIFORNIA  93042. 


PACIFIC  MISSILE  TEST  CENTER 

Point  Mugu,  California 


■'•■■"  — ■ uj|--  • f '..•.  ' ■:-:  ”••  ■ ‘ J.'f**!*? 


mmmmmmmmmmms. 


DEPARTMENT  OF  THE  NAVY 

COMMANDER 

PACIFIC  MISSILE  LEST  CENTER 

POINT  MUGU.  CALIFORNIA  93042 


IN  REPLY  REFER  TO 


4250/me 
Ser  700') 

29  March  1976 


From:  Commander,  Pacific  Missile  Test  Center 

To:  Distribution  List 


Subj.  PMTC  Technical  Publication  TP-76-6  “Telemetry  System  Parameters  and  Bit  Error  Performance 
of  NRZ  and  DM  PCM/FM"  by  D.  A.  King,  did  29  March  1976;  supplement  to 


/ 


I , The  following  should  be  added  to  page  3:  bit  synchronizer  A was  the  EMR  720  and  bit  synchro- 
nizer B was  the  Monitor  335. 


2.  This  information  was  not  included  in  the  text  of  the  report  to  avoid  the  possibility  of  construing 


f the  results  as  a comparison  of  the  two  bit  synchronizers. 


’ ,/-<  v / "'J'  ‘ / 

I ' _ y/'  - 


W.  J.  KIRKPATRICK 
By  direction 


U 

\ 


Cl 

% 


’'n,  rtf6 


Itfurwniy  n 


...  - >»,  ■>/ v*1,,  W-  ^ f 


UNCLASSIFIED 

SECURITY  CLASSIFICATION  of  THIS  page  r»Fhen  Data  Entered) 


REPORT  DOCUMENTATION  PAGE 


READ  INSTRUCTIONS 
BEFORE  COMPLETING  FORM 


H Report  number 


TP-76-6 


TITLE  (and  Subliti*) 


REPOHya  PERIOD  COVERED 


TELEMETRY jySTEM PARAMETERS  AND^IT  ERROR/ 
PERFORMANCE  OF  NR?  AND  DM  PCM/fM  - ^ / 


6 PERFORMING  ORG,  REPORT  NUMBER 


17.  author^; 


I • CONTRACT  DR  GRANT  NUMBERf s) 


^f))PMrC-TP-7L-i  1 


19  PERFORMING  ORGANIZATION  n4 


Pacific  Missila  Test  Center 
Point  Mugu,  California  93042 


“ROGRAM  ELEMENT.  PROJECT,  TASK 
AREA  A WORK  UNITwUMBERS  ; , 

AIRTASK,  A5356352/054)b/ 
5W4^pU0p30,  Work  Unit 
A5352 10000002 


FlT  CONTROLLING  OFFICE  NAME  ANOAOORE55 


Naval  Air  Systems  Command 
Washington,  DC  20361 


29  Mar  Mi  76 


U MONITORING  AGENCY  NAME  A ADORESSfi/  different  from  Controlling  Office)  I IS  SECURITY  CLASS,  (of  thie  report) 


UNCLASSIFIED 


'/2JS2P- 


1S«  OECL  ASSI  FlC  ATION/ OOWNGRAOING 
SCHEOULE 


16-  ^eTDiBn^iijtfyFMrh:  ^ 1 ■ " 

Distribution  limited  to  US.  Government  agencies  only;  tast  and  evaluation  information;  29  March 
1976.  Other  requests  for  this  document  must  be  refarred  to  tha  Commandar  (Code  42501,  Pacific 
Missile  Tast  Center,  Point  Mugu,  California  93042. 


I 17.  OIST RIBUTION  STATEMENT  (of  the  abstract  entered  In  Block  20,  if  different  from  Report) 


0QA535-  Vg/z 


19  KEY  WOROS  (Continue  on  reveree  eide  if  necessary  and  identify  by  bfoch  number) 

Non-raturn-to-zero  RF  bandwidth 

Delay  modulation  (Miller  code) 

PCM/FM  optimization 
Bit  error  probability 


^^STRlkCT  (Continue  on  revaree  eide  ff  n«c«»e*fy  and  Identify  by  bfock  number) 

«Vhe  work  reported  herein  involved  the  axperimantal  determination  of  optimum  telematry 
systam  paramaters  for  tha  transmission  of  NRZ  and  DM  PCM/FM.  The  objactiva  was  to  compara 
the  efficiency  of  the  two  PCM  formats  and  to  evaluate  rules  of  thumb  for  detarmining  systam 
paramaters.  A PCM/FM  telematry  system  was  simulated  and  optimum  raceivar  IF  bandwidth,  RF 
transmitter  deviation,  and  pramodulation  filter  bandwidth  for  the  transmission  of  NRZ  and  DM 
wera  determined  so  that  NRZ  and  DM  performance  could  be  comparad  on  an  equivalent  basis.  ' — 

(Continued) 


00  ,^nR“s  1473  EOITION  OF  1 NOV  6S  IS  OBSOLETE 


UNCLASSIFIED 

SECURITY  CLASSIFICATION  OF  THIS  PAGE  (When  Omlm  Entered) 


UNCLASSIFIED 

SECUWTV  CLASSIFICATION  OF  THIS  P«OEflWi«n  P«U  Bnlmil) 


20.  ABSTRACT  (Concluded) 

RZ  and  DM  rules  of  thumb  were  examined  to  determine  the  system  performance  loss  due  to  non- 
optimum operation. 

The  experiment  verified  other  reported  conclusions  that  show  NRZ  to  be  3 dB  better  than  DM 
for  equicalent  bit  rates  under  their  respective  optimum  conditions.  Thus  DM  is  not  recommended 
for  epplications  of  maximum  data  transfer  in  e bandtimited  RF  system  where  noisy  signals  may  be 
received. -Use  of  rules  of  thumb  in  setting  system  parameters  will  generally  result  in  less  than  3 dB 
degradetiylin  BEP  as  long  as  the  rules  fell  within  certain  bounds  about  optimum: 

1.  The  receiver  IF  bandwidth  should  be  at  least  twice  the  optimum  (fR  for  NRZ,  2fR  for  DM) 

for  data-recording  purposes.  The  equivalent  bandwidth  of  the  prerecording  and  post-recording 
IF  combinetion  should  be  close  to  f_  but  less  than  2fR  for  NRZ  and  close  to  2fR  but  less 
than  4fB  for  DM.  B 

2.  The  peak-to-peek  RF  transmitter  deviation  should  lie  between  0.6fR  and  03fR  for  NRZ  end 
between  1 2fR  and  1 BfR  for  DM . 


3.  The  premodulation  filter  bandwidth  should  fall  between  0.5f_  end  1 ,0fR  for  both  NRZ  end 
DM. 


ACKNOWLEDGMENT 


The  author  wishes  to  thank  F.  R.  Hartzler  and  E.  L.  Law  for  their  assistance  throughout  this 


project 


1 


CONTENTS 


Page 


ABBREVIATIONS v 

SUMMARY 1 

INTRODUCTION 3 

TEST  EQUIPMENT 3 

TEST  METHODS  AND  RESULTS 4 

RF  Transmitter  Deviation 4 

IF  Bandwidth 9 

Premodulation  Filter  Bandwidth 11 

RF  Bandwidth 14 

NRZ  AND  DM 20 


SUMMARY  OF  RESULTS 20 

REFERENCES 25 

TABLES 

1 . Equipment  List 4 

2.  NRZ  Optimum  Deviations  for  Various  IF  Filter  Bandwidths 9 

FIGURES 

1 . System  Block  Diagram 4 

2.  BEP  Variations  With  RF  Deviation  for  NRZ  PCM/FM 6 

3.  BEP  Variations  With  RF  Deviation  for  DM  PCM/FM 7 

4.  NRZ  RF  Signal  Spectra 8 

5.  DM  RF  Signal  Spectra 10 

6.  NRZ  Bit  Error  Sensitivity  to  IF  Filter  Bandwidth 12 

7.  DM  Bit  Error  Sensitivity  to  IF  Filter  Bandwidth 13 

8.  NRZ  BEP  Variations  With  Premodulation  Filter  Bandwidth 15 

9.  DM  BEP  Variations  With  Premodulation  Filter  Bandwidth 16 

10.  NRZ  BEP  Sensitivity  to  RF  Deviation  and  Premodulation  Filter  Bandwidth 17 

1 1 . DM  BEP  Sensitivity  to  RF  Deviation  and  Premodulation  Filter  Bandwidth 18 

12.  Bit  Rates  Containable  in  1 MHz  and  3 MHz  RF  Channel  Bandwidths  by  Band 

Limiting  With  a Premodulation  Filter 19 

13.  PCM  Signaling  Formats 21 

14.  NRZ  PCM  Video  Spectra  of  a Pseudo-Random  Pattern 22 

15.  DM  PCM  Video  Spectra  of  a Pseudo-Random  Pattern 23 

16.  Comparison  of  Optimum  NRZ  and  DM  PCM/FM  Bit  Error  Probabilities 24 


m 


PRECEDING  PAGE  BLANK. NOT  FILMED 


-‘t-MfiTM  4'  / ' • **"  ' 


ABBREVIATIONS 


Automatic  gain  control 
Bit  error  probability 
Bandwidth 
Decibels 

Decibels  with  reference  to  1 milliwatt 
Delay  modulation  (or  Miller  code) 

Bit  rate 

Frequency  modulation 
Filter/sample  bit  detector 
Frequency  shift  keying 
Gigahertz 

Integrate  and  dump  bit  detector 
Intermediate  frequency 
Inter-Range  Instrumentation  Group 
Kilobits  per  second 
Kilohertz 

Megabits  per  second 
Megahertz 
Non-return-to-zero 
Pulse  code  modulation 
Peak-to-peak 
Radio  frequency 
Signal-to-noise  ratio 


FRSCCDIN3  PAG*  BLANK.NOT  FILMED 


PACIFIC  MISSILE  TEST  CENTER  TP-76-6 

Point  Mugu,  California  29  March  1976 

TELEMETRY*$YSTEM  PARAMETERS  AND  BIT  ERROR 
PERFORMANCE  OF  NRZ  AND  DM  PCM/FM 

(AIRTASK  A5365352  0540  5W47410030, 

Work  Unit  A535210000002) 

By 

D.  A.  KING 


SUMMARY 

The  work  reported  herein  involved  the  experimental  determination  of  optimum  telemetry  system 
parameters  for  the  transmission  of  NRZ  and  DM  PCM/FM.  The  objective  was  to  compare  the  efficiency 
of  the  two  PCM  formats  and  to  evaluate  rules  of  thumb  for  determining  system  parameters.  A PCM/FM 
telemetry  system  was  simulated  and  optimum  receiver  IF  bandwidth,  RF  transmitter  deviation,  and  pre- 
modulation filter  bandwidth  for  the  transmission  of  NRZ  and  DM  were  determined  so  that  NRZ  and  DM 
performance  could  be  compared  on  an  equivalent  basis.  NRZ  and  DM  rules  of  thumb  were  examined  to 
determine  the  system  performance  loss  due  to  non -optimum  operation. 

The  experiment  verified  other  reported  conclusions  that  show  NRZ  to  be  3 dB  better  than  DM  for 
equivalent  bit  rates  under  their  respective  optimum  conditions.  Thus  DM  is  not  recommended  for  ap- 
plications of  maximum  data  transfer  in  a bandlimited  RF  system  where  noisy  signals  may  be  received. 

The  experimental  optimum  values  of  RF  transmitter  deviation,  IF  filter  bandwidth,  and  premodu- 
lation filter  bandwidth  were  found  to  be: 


P-P  RF  Transmitter  Deviation 
IF  Filter  Bandwidth 

Premodulation  Filter 
Bandwidth 


NRZ 

F/S  I/D 

0BfB  0SfB 

lX)fB  U)fB 

0.5fBtolX)fB  05fBtolj0fB 


DM 

F/S  I/D 

lj6f-  l-8f„ 

0ifBtolj0fB  0.5  f„  to  1.0fB 


Use  of  rules  of  thumb  in  setting  system  parameters  will  generally  result  in  less  than  3 dB  degradation  in 

BEP  as  long  as  the  rules  fall  within  certain  bounds  about  optimum: 

1 . The  receiver  IF  bandwidth  should  be  at  least  twice  the  optimum  for  data-recording  purposes.  The 
equivalent  bandwidth  of  the  prerecording  and  post -recording  IF  combination  should  be  close  to  fR 
but  less  than  2fB  for  NRZ  and  close  to  2fB  but  less  than  4fB  for  DM. 

2.  The  peak-to-peak  RF  transmitter  deviation  should  lie  between  0.6fB  and  0.9fB  for  NRZ  and  be- 
tween 1.2fg  and  lBfR  for  DM. 

3.  The  premodulation  filter  bandwidth  should  fall  between  0.5ffi  and  1 jOfB  for  both  NRZ  and  DM. 
(U)  Publication  UNCLASSIFIED. 


Distribution  iimitsd  to  U.S.  Government  agencies  only; 
tost  end  evaluation  information;  29  March  1976.  Other 
requests  for  this  document  must  be  referred  to  tha 
Commander  (Code  4280),  Pacific  MtasRe  Test  Center, 
Point  Mugu,  California  93042. 

1 


INTRODUCTION 

An  experiment  was  conducted  on  a simulated  RF  telemetry  link  to  compare  the  bit  error  perform- 
ance characteristics  of  two  PCM  formats,  delay  modulation  (DM  or  Miller  code)  and  non-return-to-zero 
(NRZ).  A valid  comparison  required  that  the  telemetry  system  be  operated  in  an  optimum  manner  such 
that  bit  errors  for  both  PCM  signals  were  minimized  for  equivalent  data  transfer.  Consequently,  this 
report  is  also  concerned  with  establishing  optimum  values  for  NRZ  and  DM  system  parameters.  There 
were  four  parameters  under  control  in  the  simulated  telemetry  link  with  which  to  minimize  the  bit 
errors:  the  PCM  format,  the  premodulat'on  filter  bandwidth,  the  RF  transmitter  deviation,  and  the  re- 
ceiver IF  bandwidth.  All  other  system  parameters  were  held  constant  during  the  experiment. 

The  criterion  for  optimality  was  that  combination  of  parameter  values  for  each  of  the  two  codes 
which  minimized  their  bit  error  probability  (BEP).  However,  optimum  parameter  values  for  minimizing 
BEP  may  not  be  optimum  from  the  viewpoint  of  RF  bandwidth  considerations.  To  remain  within  an  RF 
channel  assignment,  the  trading  of  more  bit  errors  for  a narrower  RF  signal  bandwidth  may  be  necessary. 
This  tradeoff  can  be  accomplished  by  decreasing  the  RF  transmitter  deviation  and  premodulation  filter 
bandwidth  from  their  optimum  values.  Both  RF  bandwidth  and  minimization  of  BEP  were  considered 
in  this  experiment.  The  work  was  performed  under  AIKTASK  A5T55352  0J4D  5W4l4l00Jl>,  Missilt 
Flight  Evaluation  Systems,  work  unit  AS 3 52 10000002,  to  provide  analytical  support  to  the  Telemetry 
Group  of  the  Range  Commanders  Council. 


TEST  EQUIPMENT 

The  simulated  telemetry  system  and  associated  test  equipment  are  shown  in  figure  1 . The  equip- 
ment Is  listed  in  table  i 

The  EMR  721  test  set  served  as  both  a bit  error  rate  detector  and  an  NRZ  and  DM  signal  generator 
with  variable  bit  rate.  The  NRZ  and  DM  signals  were  pseudo-random  sequences  of  2,047  bits.  The  pre- 
modulation filter  was  a four-pole,  linear  phase  filter  with  adjustable  bandwidth.  RF  transmitter  devia- 
tion and  RF  attenuation  were  adjustable  on  the  FM  signal  generator  which  was  operated  at  a carrier 
frequency  of  approximately  1.48  GHz.  The  receiver  IF  bandwidth  was  selectable  from  the  following 
fixed  units:  100  kHz,  200  kHz,  500  kHz,  750  kHz,  and  1 .0  MHz.  The  receiver’s  video  filter  was  by- 
passed and  the  video  signal  applied  directly  to  one  of  two  bit  synchronizers.  Bit  synchronizer  A con- 
tained buth  a U.75fB  Mirtehed  HH<-f/s<n,iple  tit  detector  and  tn  irtfcgtaw  and  dump  dclaclix  Bil 
synchronizer  B contained  only  an  integrate  and  dump  detector.  The  selected  bit  synchronizer  returned 
the  recovered  NRZ  and  DM  bit  streams  to  the  EMR  721  for  error  detection.  RF  bandwidth  was  mon- 
itored on  a spectrum  analyzer. 


hbcedihj  pahe  blamunot  tombd 


HP  8555A 
SPECTRUM 
ANALYZER 


HP  3205A 
FM  SIGNAL 
GENERATOR 


RF  SIGNAL 


VIDEO 

SIGNAL 


SA  410 
RECEIVER 


BIT  SYNCHRO- 
NIZER B 

INTEGRATE 
AND  DUMP 


FILTERED 

PCM 

SEQUENCE 


ROCKLAND  MODEL  400 
4-POLE.  LINEAR 
PHASE  FILTER 


EMR  721  TELEMETRY 
. BIT  ERROR  RATE 
DETECTOR 


— 1 — 1 1 

BIT  SYNCHRO-  j 

BOONTON 

42BD  POWER  U-U- 

NIZER  A | 

METER  1 

1 

FILTER/SAMPLE  1 

1 

| 

L. 

J 

RECOVERED  PCM  SEQUENCE 

Figure  1.  System  Block  Diagram. 

Tibia  1.  Equipment  Lilt 

Bit  error  rete  detector,  EMR  721 

Bit  synchronizers  A end  B,  two  st»te-of-the-ert  bit  synchronizers 

Premoduletion  filter,  Rockland  model  1200 

FM  signal  generator,  HP  3205A 

Receiver,  Scientific-Atlente  series  410  WA 

Oscilloscope,  Tektronix  type  564 

Spectrum  Anelyzer,  HP  8555A 

Microwett meter,  Boonton  Electronics  42  BD 


TEST  METHODS  AND  RESULTS 

The  system  parameters  for  NRZ  and  DM  were  first  optimized  for  minimum  BEP  and  then  RF 
bandwidth  considerations  were  examined.  The  initial  parameter  values  were  set  according  to  the  follow- 
ing commonly  used  rules  of  thumb : 


P-P  RF  Transmitter  Deviation  0.7fB  1.2f0 

Premodulation  Filter  Bandwidth  0.6ffl  0.6ffi 

Receiver  IF  Bandwidth  f0  2.0ffl 

where  ffi  is  the  bit  rate. 

RF  Transmitter  Deviation 

The  optimum  RF  transmitter  deviations  for  NRZ  and  DM  were  found  by  fixing  the  attenuation  of 
the  RF  signal  such  that  the  BEP  was  approximately  10"4.  The  deviation  and  BEP  were  recorded  as  the 


deviation  was  incrementally  varied  at  the  transmitter  so  tha)  BEP  variations  were  adequately  defined.  In- 
creasing the  RF  attenuation  in  2 dB  steps  and  repeating  the  deviation  and  BEP  measurements  resulted  in 
a family  of  curves  from  which  the  optimum  deviation  could  be  determined.  Figure  2 for  NRZ  and  fig- 
ure 3 for  DM  show  the  BEP  variations  with  RF  transmitter  deviation  for  fB  equal  to  a 500  kb/s  rate. 

This  data  shows  that  the  optimum  deviations  are  ±200  kHz  at  the  500  kb/s  rate  for  NRZ  and  ±400  kHz 
for  DM.  NRZ  bit  rates  of  200  kb/s  and  750  kb/s  were  also  investigated  and  their  optimum  deviations 
were  found  to  be  ±80  kHz  and  ±300  kHz,  respectively.  DM  bit  rates  of  100  kb/s  and  375  kb/s  were 
investigated  and  their  optimum  deviations  wtte  also  found  to  be  ±b0  kHv  anJ  1 10U  k Hi  rtai^rtrvely 
In  general,  the  data  indicate  that  the  optimum  peak-to-peak  RF  transmitter  deviations  are  0.8fB  for 
NRZ  and  1 .6f„  for  DM  with  the  other  parameters  at  their  initial  values. 

A research  of  available  literature  on  optimum  deviation  for  NRZ  PCM/FM  shows  a variation  rang- 
ing from  0.7f  to  0.9f_.  Kotel’nikov  (reference  1)  and  Smith  (reference  2)  derive  the  optimum  devi- 
ation to  be  0.^1 5fB  for  FSK.  Experimentally,  Aeronutronic  (reference  3)  found  the  optimum  devia- 
tion to  be  0.75fB  as  did  a study  by  Electro -Mechanical  Research  (EMR)  (reference  4).  At  the  other  end 
of  the  range,  Shaft  (reference  5)  calculated  0.796fB  and  experimentally  found  0.84fB  as  the  optimum 
values. 

Perhaps  one  of  the  more  likely  reasons  for  the  variation  in  optimum  deviation  is  the  method  of 
bit  detection.  In  this  experiment  a 0.75fB  matched  filter/sample  detector  in  bit  synchronizer  A resulted 
in  optimum  deviations  of  0.8fB  for  NRZ  and  1 jWb  for  DM.  An  integrate  and  dump  detector  <squwe 
PCM  matched  filter)  also  in  bit  synchronizer  A was  tested  for  NRZ  and  gave  0.9fB  as  optimum.  Bit 
synchronizer  B with  a square  PCM  matched  filter  detector  was  also  tested.  Bit  synchronizer  B resulted 
in  an  optimum  deviation  of  0.9fB  for  NRZ  and  1 JBfB  for  DM.  Aeronutronic’s  optimum  of  0.75fB 
(NRZ)  was  with  an  integrating  detector.  They  also  used  a sampling  detector  that  resulted  in  an  optimum 
of  0.9f  (NRZ).  Kotel’nikov  and  Smith’s  optimum  was  derived  for  coherently  detected  FSK,  whereas 
Shaft’s  optimum  was  determined  using  discriminator  detection.  The  reason  for  the  variation  of  optimum 
deviation  with  bit  ikttelot  *n  tml  iHwsti|ptedr  but,  as  shown  in  figures  2 and  3 and  discussed  later,  the 
degradation  in  BEP  due  to  a non-optimum  deviation  setting  is  not  severe  if  maintained  near  optimum. 

It  was  found  that  the  optimum  deviation  is  independent  of  the  premodulation  filter  bandwidth  but 
dependent  upon  IF  bandwidth  and  IF  signal-to-notee  ratio  (SNR).  From  figures  2 and  3 tt  appein  that 
the  optimum  deviation  increases  roughly  10  percent  at  low  IF  SNR  due  to  AGC  action  and/or  changing 
IF  fitter  characteristics  (not  rtecWtw  exhibit  this  phenomena)  However,  tMi  charge  causes  a rela- 
tively insignificant  increase  in  BEP  and  can  probably  be  ignored. 

The  dependence  of  optimum  deviation  on  IF  bandwidth  can  be  related  to  the  RF  signal  and  noise 
power  spectrums.  Figure  4(a)  shows  the  RF  signal  spectrum  with  optimum  RF  deviation  for  a 500  kHz 
IF  filter  bandwidth  (fH  = 500  kb/s).  The  IF  bandwidth  was  doubled,  and  the  optimum  deviation  was 
found  no  longer  to  be  0.8fB  but  increased  to  1.09fB  as  shown  in  figure  4(c).  Doubling  the  IF  band- 
width doubled  the  noise  power  to  the  demodulator  (assuming  white  noise),  but  as  shown  in  a 1 MHz 
bandwidth  of  figure  4(a),  the  signal  power  contributed  by  the  introduction  of  the  second  sidebands  into 
the  IF  passband  did  not  double  the  total  signal  power  to  the  demodulator.  Therefore,  as  a result  of  de- 
creased IF  SNR,  the  BEP  increased.  However,  the  doubled  IF  bandwidth  then  allowed  the  deviation  to 
be  increased  up  to  1.09f„  without  BEP  degradation  due  to  intermodulation  distortion  (filter  phase  non- 
linearities)  and  forced  FM  thresholding  (signal  amplitude  limiting  by  IF  filter  skirts).  Increasing  the 
deviation  improves  the  receiver’s  video  SNR  and  decreases  the  BEP.  It  does  not  increase  the  IF  SNR. 

In  fact,  due  to  constant  transmitter  power,  the  spreading  of  the  RF  spectrum  by  increasing  the  deviation 
wM  limy"  the  IF  SNR  by  removing  signal  power  from  the  passband.  This  power  loss  is  not  as  significant 
as  the  improved  video  SNR,  so  consequently  the  net  result  is  to  lower  the  BEP  as  the  deviation  is  in- 
creased to  1.09fB.  Beyond  1 ,09fB,  the  BEP  begins  to  increase  again  due  to  intermodulation  distortion, 
signal  power  loss,  and  forced  FM  thresholding. 


BIT  SYNCHRONIZER  A,  F/S 
500  kb/s  NRZ 

500  kHz  PREMODULATION  FILTER 

200  kHz/DI  VISION 

0.8  fg  PEAK-TO-PEAK  DEVIATION 


BIT  SYNCHRONIZER  A,  F/S 
500  kb/s  NRZ 

500  kHz  PREMODULATION  FILTER 

200  kHz/DI  VISION 

0.53  fg  PEAK-TO-PEAK  DEVIATION 


Figure  4.  NRZ  RF  Signal  Spectra. 


iMHMlH , 


BIT  SYNCHRONIZER  A,  F/S 
500  kb/s  NRZ 

500  kHz  PREMODULATION  FILTER 

200  kHz/DIVISION 

1.09  fD  PEAK-TOPEAK  DEVIATION 


Figure  4.  ( Concluded /. 


Obviously,  if  the  IF  bandwidth  is  halved  from  the  500  kHz  of  figure  4(a),  the  BEP  will  increase 
because  of  IF  filter  phase  nonlinearities  and  forced  FM  thresholding.  Reducing  the  deviation  to  remove 
these  effects  will  lower  the  BEP.  The  experimental  NRZ  data  in  table  2 shows  the  optimum  deviation 
for  various  IF  bandwidths.  All  that  remains  in  order  to  specify  the  optimum  deviation  is  to  determine 
the  optimum  IF  bandwidth. 


Table  2.  NRZ  Optimum  Deviations  for  Various  IF  Filter  Bandwidths 


IF  Bandwidth 


In  the  experiments  conducted  by  Aeronutronics  (reference  3)  and  EMR  (reference  4),  the  optimum 
receiver  bandwidth  for  NRZ  was  investigated  and  found  to  be  equal  to  the  bit  rate,  fg.  An  unpublished 
report  on  DM  by  Dr.  W.  R.  Hedeman  of  Aerospace  Corporation  indicates  an  optimum  bandwidth  of  2fg 
Their  experiments  were  not  repeated  for  this  report. 

A verification  of  these  optimum  bandwidths  was  conducted  by  visually  examining  the  RF  spec- 
trums  of  NRZ  and  DM  at  various  RF  deviations  and  with  the  premodulation  filter  bandwidth  equal  to 
the  bit  rate.  The  RF  signal  spectra  of  figures  4 and  5 were  taken  with  fg  equal  to  500  kb/s.  Figures 
4(a)  and  5(a)  show  RF  signal  spectra  with  optimum  deviations  for  the  optimum  RF  bandwidths  reported 
above.  The  spectra  are  roughly  flat  and  constant  over  a bandwidth  equal  to  the  bit  rate  and  signal 
power  drops  abruptly  outside  this  bandwidth.  The  3-dB  points  of  the  optimum  filter  are  thus  located  at 


Bit  Rate 
(kb/s) 

IF  Filter  Bandwidth 
(kHz) 

Peak-to-Peak 
Deviation  Ratio 

Bit  Synchronizer 

100 

100 

0.77 

A (F/SI 

100 

200 

033 

A (F/SI 

100 

300 

1.38 

A (F/S) 

100 

500 

3.37 

A (F/SI 

750 

500 

0.75 

B (l/DI 

750 

750 

050 

B (I/D) 

750 

1,000 

1.09 

B (1/DI 

BIT  SYNCHRONIZER  A.  F/S 
500  kb/s  DM 

500  kHz  PREMODULATION  FILTER 
200  kHz/DIVISION 


1.6  f„  PEAK-TO-PEAK  DEVIATION 


BIT  SYNCHRONIZER  A,  F/S 
500  kb/s  DM 

500  kHz  PREMODULATION  FILTER 
500  kHz/DIVISION 


1.04  fg  PEAK-TO-PEAK  DEVIATION 


Figure  5.  DM  RF  Signal  Spectra 


BIT  SYNCHRONIZER  A,  F/S 
500  kb/s  DM 

500  kHz  PREMODULATION  FILTER 

500  kHz/DIVISION 

2.17  f„  PEAK-TO-PEAK  DEVIATION 

D 


Figure  5.  (Concluded!. 


a bandwidth  equal  to  the  bit  rate  for  NRZ  and  twice  the  bit  rate  for  DM;  in  this  case,  500  kHz  about 
center  frequency  for  NRZ  and  1 ,000  kHz  about  center  frequency  for  DM.  Both  bandwidths  encompass 
the  first  sidebands  of  the  spectra  where  most  of  the  signal  power  is  concentrated.  Larger  bandwidths 
may  allow  more  noise  power  than  signal  power  into  the  IF  passband  and  narrower  bandwidths  may  un- 
necessarily restrict  the  signal  causing  intermodulation  distortion  and  forced  FM  thresholding.  Optimizing 
the  RF  deviation  for  a non-optimum  IF  bandwidth  results  in  RF  spectra  such  as  those  in  figures  4(b) 
and  5(b)  (narrow  .’r  IF  filter  than  optimum)  and  4(c)  and  5(c)  (wider  IF  filter  than  optimum).  These 
spectra  are  not  as  optimally  distributed  in  the  IF  passband  as  in  figures  4(a)  and  5(a)  and  result  in  a 
higher  BEP. 

To  illustrate  the  differences  in  BEP  between  optimum  and  non-optimum  IF  bandwidths  and  devi- 
ations, BEP  variations  with  RF  power  were  recorded  and  plotted  in  figures  6 and  7 for  NRZ  at  750  kb/s 
and  for  DM  at  375  kb/s.  BEP  measurements  were  made  using  a 500  kHz,  750  kHz,  and  1 .0  MHz  IF 
bandwidths  with  the  RF  deviation  at  optimum  for  750  kHz  and  also  with  the  RF  deviation  optimized 
for  the  nonoptimum  IFs.  The  results  show  severe  BEP  degradation  for  the  narrower-thanoptimum  500 
kHz  IF  filter  and  an  approximate  0.5  dB  degradation  for  the  wider-thanoptimum  1.0  MHz  filter  for 
both  NRZ  and  DM.  Slight  improvements  in  BEP  were  made  by  optimizing  the  RF  deviation  for  the 
nonoptimum  IF  filters,  but  the  minimum  BEP  was  still  produced  with  the  750  kHz  IF  filter  at  a de- 
viation of  0.9fg  (bit  synchronizer  B,  I/D). 

For  data  recording,  the  receiver  IF  bandwidth  should  be  at  least  2fg  for  NRZ  and  4fg  for  DM  to 
ensure  no  loss  of  signal  because  of  IF  clipping  by  signal  drift.  In  such  cases,  the  optimum  RF  deviation 
is  set  according  to  the  total  effective  bandwidth  of  prerecording  and  post-recording  IF’s.  Optimally,  this 
effective  bandwidth  is  set  as  close  as  possible  to  fg  for  NRZ  and  2fg  for  DM. 

Premodulation  Filter  Bandwidth 

After  setting  the  RF  transmitter  deviation  and  receiver  IF  bandwidth  to  the  previously  determined 
optimum  values,  the  premodulation  filter  bandwidth  was  adjusted  for  minimum  BEP.  As  expected,  the 
optimum  bandwidth  was  infinite  for  both  NRZ  and  DM  since  signal  energy  per  bit  increases  with 


L 


»"*  ',*•*  -w-ar*  < *•.-  #“—•*•  «*■  WIWI 


rnmmum m 


-■'•'■fcMgaafet.i: 'ti. 


w* 


'kl'Afy  * 


bandwidth.  However,  the  primary  purpose  of  a premodulation  filter  is  to  limit  RF  spectral  occupancy 
by  attenuating  the  tails  of  the  RF  signal  spectrum.  Therefore  a tradeoff  of  BEP  for  RF  bandwidth  is 
necessary.  Figures  8 and  9 show  the  variation  in  BEP  with  premodulation  filter  bandwidth  for  ffi  at 


500  kb/s.  Both  figures  suggest  that  the  premodulation  filter  bandwidth  should  be  set  between  0.5ffi  end 

are  due  to 


1.0fn 


.g.  The  relatively  small  losses  in  BEP  for  premodulation  bandwidths  as  narrow  as  O.5f0 
the  band  limiting  of  the  IF  filter.  Since  the  sidebands  of  the  RF  signal  spectrum  are  generally  rejected 
by  the  IF  filter,  the  premodulation  filter  should  have  little  effect  on  BEP  as  long  as  its  bandwidth  is 
greater  than  0.5fg  and  the  IF  filter  bandwidth  is  at  optimum.  Bandwidths  less  than  0.5fg  will  cause  a 
loss  of  signal  power  in  the  baseband  that  begins  to  severely  degrade  the  BEP.  A wider  bandwidth  than 


1 ,0f„  will  result  in  a relatively  insignificant  decrease  in  BEP  because  of  the  optimum  IF  filter’s  band 


limiting  but  will  increase  the  RF  bandwidth. 


Figures  10  and  11  illustrate  the  difference  in  BEP  between  a premodulation  filter  set  at  O.Sfg  and 


1.0fg  for  fg  equal  to  500  kb/s  and  with  optimum  IF  bandwidths  and  RF  deviations.  The  BEP  varia 


tions  with  RF  power  were  determined  by  incrementally  varying  the  FM  signal  generator’s  attenuator  and 
recording  the  attenuation  and  the  BEP.  RF  power  was  calibrated  to  the  attenuator  by  measuring  high 
RF  power  levels  at  the  receiver  input  with  an  RF  power  meter.  There  was  a 0.4  to  0.6  dB  improvement 
in  BEP  for  NRZ  using  a 500  kHz  premodulat’on  filter  over  a 250  kHz  premodulation  filter  and  a 0.4  dB 
improvement  for  DM.  At  the  non-optimum  deviations  in  figures  10  and  11,  the  improvement  was  1 dB 
for  both  NRZ  and  DM. 


RF  Bandwidth 


IRIG  document  106-73,  Telemetry  Standards,  lists  three  RF  channel  bandwidths  in  the  L-  and  S- 
band  frequency  ranges;  they  are  1 .0  MHz,  3.0  MHz,  and  10.0  MHz  in  width.  These  channel  bandwidths 
are  equivalent,  as  defined  by  IRIG,  to  RF  signal  bandwidths  of  1.2  MHz,  3.2  MHz,  and  10.2  MHz,  re- 
spectively, where  the  signal  is  60  dB  down  from  the  unmodulated  carrier  at  the  band  edge.  At  each  of 
several  bit  rates  (with  pseudo-random  data),  the  premodulation  filter  was  varied  and  the  RF  transmitter 
deviation  held  fixed  at  optimum  to  find  the  maximum  allowable  premodulation  filter  bandwidth  that 
would  still  restrict  the  RF  signal  bandwidth  to  within  one  of  the  IRIG  channels.  Figure  12  defines 
approximate  maximum  premodulation  filter  bandwidths  for  various  NRZ  and  DM  bit  rates  such  that  the 
RF  spectra  remain  within  a 1.0  MHz  or  30  MHz  channel.  Bandwidth  limitations  on  the  premodulation 
filter  and  the  modulation  section  of  the  RF  signal  generator  did  not  permit  10.0  MHz  channel  measure- 
ments. The  RF  signal  spectrum  width  will  vary  with  the  NRZ  or  DM  formatting  and  shift  with  trans- 
mitter drift;  thus  the  premodulation  filter  was  always  set  such  that  the  RF  signal  was  down  60  dB  at  the 
band  edges  of  the  IRIG  channels.  This  left  100  kHz  of  bandwidth  on  either  side  of  the  signal  spectrum 
to  allow  for  spectrum  variations  and  drift. 


The  lower  diagonal  line  in  figure  1 2 represents  the  lower  limit  on  premodulation  filter  bandwidth 
as  defined  by  Q.5fg.  Thus  the  approximate  highest  bit  rate  in  which  the  RF  spectrum  will  remain  in  a 
1.0  MHz  channel  is  roughly  350  kb/s  for  pseudo-random  NRZ  and  220  kb/s  for  pseudo-random  DM. 

For  a 3.0  MHz  channel,  the  highest  bit  rates  are  approximately  1 .0  Mb/s  for  NRZ  and  720  kb/s  for  DM. 
Higher  and  lower  bit  rates  may  be  possible  depending  on  the  data  sequence.  For  lower  bit  rates  than 
these  maximums,  it  is  recommended  that  the  premodulation  filter  be  set  as  wide  as  possible,  up  to 
1.0ffl,  without  the  RF  signal  spectrum  exceeding  an  IRIG  channel.  While  larger  premodulation  band- 
widths  are  possible,  they  do  not  give  a significant  decrease  in  BEP  for  the  increase  in  RF  bandwidth.  As 
an  example,  consider  a 500  kb/"  NRZ  signal,  figure  12  indicates  that  the  premodulation  filter  can  be  set 
from  250  kHz  to  850  kHz  and  still  remain  in  the  3 MHz  channel.  Figure  10  shows  a 0.4  to  0.6  dB 
improvement  in  BEP  with  a 500  kHz  filter  over  a 250  kHz  filter  but  no  improvement  over  the  500  kHz 
filter  with  a 869  kHz  filter. 


If  for  a particular  bit  rate  the  RF  spectrum  is  marginally  within  its  RF  channel  and  the  premod- 


ttlalkji!  fitter  ii  at  0 5frt  llterr  ft  h puutblv  to  deCtcaW  thr  PF  bandwidth  by  <kftci»ir.g  the  RF  (t.wii- 


mitter  deviation.  By  using  figures  2 and  3,  limits  can  be  set  on  a permissible  range  of  RF  transmitter 


14 


rJ- — 


' ' >>$ 


A 


BIT  SYNCHRONIZER  A,  F/S 

500  kb/»  NRZ  PSEUDO-RANDOM  SEQUENCE 

500  kHz  IF  FILTER  BANDWIDTH 


PEAK  TOREAK  DEVIATION 
PR EMODU LAT ION  FILTER 

PEAK  TOREAK  DEVIATION 
ANO  1.74  fq  PREMODULATION  FILTER 


PEAK  TOREAK  DEVIATION 
PREMOOULATION  FILTER 


NOTE;  CURVES  INDICATED  BY  CIRCLES  ANO 
SQUARES  EACH  REPRESENT  DATA  UNDER 
TWO  SETS  OF  CONDITIONS  SHOWN  ABOVE. 

*0.6  l0  PEAK-TOREAK  DEVIATION 
0.6  f0  PREMODULATION  FILTER 


RF  POWER  (dBm) 


Figure  10.  NRZ  BEP  Sensitivity  to  RF  Deviation  and  Premodulation  Filter  Bandwidth. 


deviation;  the  proposed  limits  are  0.6fB  to  0.9ffi  for  NRZ  and  1.2ffi  to  1.8ffi  for  DM.  Within  these 
limits,  the  BEP  has  been  slightly  degr^d  from  that  at  the  upper  limit  for  a small  decrease  in  RF  band- 
width. Exceeding  the  upper  limit  inr  iCases  not  only  the  BEP  but  the  RF  bandwidth  too,  whereas  drop- 
ping below  the  lower  deviation  limit  begins  to  significantly  increase  the  BEP  for  any  further  bandwidth 
savingt  For  the  deviation  ranyei  p#opoi*d  above,  an  RF  bamKHtMl  savings  J lUO  to  2UU  kHz  tor  NRZ 
and  200  to  400  kHz  for  DM  can  be  realized  if  required. 

Figures  10  and  11  show  the  tradeoff  in  BEP  that  results  when  decreasing  the  RF  transmitter 
deviation  to  save  RF  bandwidth.  With  the  premodulation  filter  at  0.5fp,  there  was  a 1 .0  to  1.2  dB 
improvement  in  NRZ  BEP  when  using  the  optimum  deviation  of  0.8fB  compared  to  using  the  lower 
deviation  limit  of  0.6fB  and  a 0.4  to  0.6  dB  improvement  with  the  premodulation  filter  at  1.0f  With 
the  premodulation  filter  set  at  0.5L,  there  was  a 0.6  to  0.8  dB  improvement  in  DM  BEP  using  the 
optimum  deviuk*)  <J  1 6I"B  h-stradof  the  lower  deviation  limit  of  ! ,2ffi  and  a 0.2  dB  improvement 
with  the  premodulation  filter  at  1 0f„ . 


NRZ  AND  DM 

A description  of  the  two  PCM  formats  is  given  in  figure  13  (see  reference  6).  Because  of  mid-bit 
transitions,  DM  is  at  twice  the  clock  rate  of  NRZ.  The  optimum  system  parameters  reflect  this  differ- 
ence in  KF  JcvkUkn'  and  IF  tn.m!wklth  which  for  DM  are  twice  those  of  NRZ;i.e., 


F/S 

I/D 

F/S 

I/D 

P-P  RF  Transmitter  Deviation 

0.8ffi 

0.9fB 

1.6fB 

1.8fB 

IF  Filter  Bandwidth  (or  equiv- 

alent pre-  and  post -recording 

IF  BW) 

iofB 

1.0fB 

2.0fB 

2.0fB 

Premodulation  Filter 

Bandwidth 

0.5 fB  to  1.0fB 

0.5fB  to  1.0fB 

0.5fB  to  1 .0fB 

0.5 fB  to  1.0fB 

Figures  14  and  15  compare  the  video  spectra  of  two  PCM  formats,  and  figure  16  compares  the  BEPs  of 
500  kb/s  NRZ  and  DM  under  their  optimum  conditions.  Figure  16  shows  that  the  BEP  for  NRZ  is 
approximately  3 dB  better  than  for  DM.  This  result  agrees  with  those  results  reached  by  Dr.  W.  R. 
Hedeman  of  Aerospace  Corporation  in  an  unpublished  report  and  by  Dr.  W.  C.  Lindsey  of  Southern 
California  University  in  reference  6.  Both  reports  conclude  that,  for  a selected  BEP,  a given  telemetry 
link  can  wppocfttfcta  the  hil  me  when  NR?  n umJ  » cn.'panx!  to  the  use  of  DM  This  is  expert 
mentally  verified  by  figures  6 and  7 where  NRZ  is  twice  the  DM  bit  rate. 


SUMMARY  OF  RESULTS 

Optimum  conditions  for  transmission  of  PCM/FM  in  general  depend  on  the  PCM  code,  the  bit  rate, 
the  RF  channel  bandwidth,  and  system  equipment.  Optimum  conditions  are  those  system  parameter 
values  that  minimize  the  BEP,  subject  to  RF  bandwidth  requirements  and  system  tolerances  (i.e.,  trans- 
mitter and  receiver  drift).  It  was  found  that  the  optimum  IF  filter  bandwidth  is  equal  to  the  bit  rate 
fct  NRZ  md  iwtee  tire  bM  sate  fbr  DM.  Huswh,  Wider  "ither  than  laboratory  conditions,  a ■wijer=tiian- 
optimum  IF  bandwidth  may  be  necessary  because  of  the  standard  fixed  IF  filter  sizes  or  because  of 
transmitter  and  receiver  drift.  In  fact,  it  is  recommended  that  for  tape-recording  purposes,  the  receiver 
IF  tindwid!1’  be  set  whhrf  then  Jptfcifthn  to  avoid  IF  dipping  Dorn  signal  drift  and  that  a narrower 
filter  bandwidth  be  used  for  data  playback. 


! 1 i 0 ! 1 

i i i 


i i 
i i 

i i 


1 I ° I ° I ° I 1 ! 1 ! ° • 1 • 

iii  i 

i i I I1 

III  I . 

I I I I, 


I I 
I I 


I I 


I I 
I I 
I I 


I I 
I I 
I I 


I I 
I I 
I I 


I I 
I I 

I I 


I NRZ-LEV EL  {OR  NRZ  CHANGE). 

'ONE''  IS  REPRESENTED  BY  ONE  LEVEL. 
"ZERO"  IS  REPRESENTED  BY  THE  OTHER 
! LEVEL. 


I I’ 

I I 

I I 

1 i' 

I h 

i r 

I I1 

i i 

I I 


NRZ-MARK  (DIFFERENTIAL  ENCODING) 
"ONE"  IS  REPRESENTED  BY  A CHANGE 
IN  LEVEL. 

"ZERO"  IS  REPRESENTED  BY  NO  CHANGE 
IN  LEVEL. 


I I I I I I I I I I I I 

I I I I I I I I I I I I 

ifhniliTuuhji 

i iii  i i( 

i iii  ii 

njinnjuirLr 

uuinrLruuLfi 


NRZ -SPACE  (DIFFERENTIAL  ENCODING) 
"ONE"  IS  REPRESENTED  BY  NO  CHANGE 
IN  LEVEL. 

"ZERO"  IS  REPRESENTED  BY  A CHANGE 
IN  LEVEL. 


"ONE”  IS  REPRESENTED  BY  A HALF-BIT 
WIC  - PULSE. 

"ZEh  " IS  REPRESENTED  BY  NO  PULSE 
CONDITION. 


I I I I 

I I I I 

I I I I 

till 
I I I I 

I I I I 


i i i i i i r 

I I I I I I I 

I I i i i i i 

i i i i i i i 

i i i i i i i 

i I I 1 I I l 


BI-PHASE-LEVEL  (SPLIT  PHASE  OR 
MANCHESTER  CODE). 

"ONE"  IS  REPRESENTED  BY  A HALF-BIT 
WIDE  PULSE  OF  ONE  POLARITY. 

"ZERO”  IS  REPRESENTED  BY  A HALF-BIT 
WIDE  PULSE  OF  THE  OPPOSITE  POLARITY 

BI-PHASE  MARK. 

A TRANSITION  OCCURS  AT  THE  BEGINNING 
OF  EVERY  BIT  PERIOD. 

"ONE"  IS  REPRESENTED  BY  A SECOND 
TRANSITION  ONE-HALF  BIT  PERIOD  LATER. 
"ZERO"  IS  REPRESENTED  BY  NO  SECOND 
TRANSITION. 


uiruumruui 


i i 

i ! 


i i 

i i 


BI-PHASE  SPACE. 

A TRANSITION  OCCURS  AT  THE  BEGINNING 
OF  EVERY  BIT  PERIOD. 

"ONE"  IS  REPRESENTED  BY  NO  SECOND 
TRANSITION. 

"ZERO"  IS  REPRESENTED  BY  A SECOND 
TRANSITION  ONE-HALF  BIT  PERIOD  LATER 

DELAY  MODULATION  (MILLER  CODE). 

"ONE"  IS  REPRESENTED  BY  A SIGNAL  TRAN- 
SITION AT  THE  MIDPOINTOF  THE  BIT  PERIOD. 
"ZERO"  IS  REPRESENTED  BY  NO  TRANSITION 
UNLESS  IT  IS  FOLLOWED  BY  ANOTHER  ZERO. 
IN  THE  LATTER  INSTANCE,  A TRANSITION  IS 
PLACED  AT  THE  END  OF  THE  BIT  PERIOD  OF 
THE  FIRST  ZERO. 


Figure  13.  PCM  Signaling  Formats. 


->•'  '***»  *»*“'**<» 


500  kb/s  NRZ 

500  kHz  PREMODULATION  FILTER 
200  kHz/DIVISION 


500  kb/s  NRZ 

250  kHz  PREMODULATION  FILTER 
200  kHz/DIVISION 


Figure  14.  NRZ  PCM  Video  Spectra  of  a Pseudo-Random  < it  tern. 


500  kb/s  DM 

500  kHz  PREMODULATION  FILTER 
200  kHz/DI VISION 


500  kb/s  DM 

250  kHz  PREMODULATION  FILTER 
200  kHz/DI  VISION 


Figure  15.  DM  PCM  Video  Spectra  of  a Pseudo-Random  Pattern. 


BIT  SYNCHRONIZER  A,  F/S 
BIT  RATE  = 500  kb/« 

IF  FILTER  - 500  kHz  FOR  NRZ;  1 MHz  FOR  DM 


I DM  «1.2fB  PEAK  TO-PEAK  DEVIATION 
I 0.5  fB  PREMODULATION  FILTER 

- DM  41.6  f8  PEAK  TO-PEAK  DEVIATION 

0.5  fB  PREMODULATION  FILTER 

,0-5 NRZ  40.6  fB  PEAK  TO-PEAK  DEVIATION 

- 0.5  fB  PREMODULATION  FILTER 

I NRZ  iOB  fB  PEAK-TO-PEAK  DEVIATION 
0.5  fB  PREMODULATION  FILTER 


-9B  -97  -96  -95  -94  -93  -92  -9 

RF  POWER  (dBm) 

Figure  16.  Comparison  of  Optimum  A IRZ  and  DM  PCM/FM  Bit  Error  Probabilities. 


:?$>••••**<*•■ 


The  optimum  deviation  is  that  deviation  which  minimizes  the  BEP  within  RF  channel  constraints. 
The  optimum  peak-to-peak  RF  deviation  was  found  to  depend  upon  IF  filter  bandwidth  and  bit  detec- 
tion equipment.  Theoretical  and  experimental  data  (see  references  1 through  5)  indicate  that  the  op- 
timum deviation  ranges  from  0.7fB  to  0.9fB  for  NRZ  with  an  optimum  IF  filter  bandwidth  and  may  be 
higher  for  a wider-than-optimum  IF  filter. 

The  optimum  premodulation  filter  bandwidth  is  a value  which  minimizes  the  BEP  and  keeps  the 
RF  spectral  occupancy  within  requirements.  Results  show  bandwidths  less  than  0.5fB  begin  to  severely 
raw  the  BEP  and  that  bandwidth*  6r<-aiet  than  1 .0ffi  offer  very  little  gam  in  terms  ot  btP.  Because  the 
premodulation  filter’s  primary  purpose  is  to  bandlimit  the  RF  signal  spectrum,  it  should  be  associated 
with  RF  channel  bandwidths  as  well  as  the  bit  rate. 

Obviously,  optimum  conditions  for  data  transmission  can  take  on  a wide  range  of  values.  Rather 
than  attempting  to  achieve  maximum  performance  from  a system  by  determining  optimum  conditions 
for  every  requirement,  rules  of  thumb  for  setting  system  parameters  car  be  established  for  all  systems 
in  noncritical  applications  with  only  a moderate  loss  (~2  dB)  in  BEP.  That  is,  many  telemetry  system 
applications  must  tolerate  non-optimum  operation  to  some  degree.  For  these  applications,  rules  of 
thumb  may  be  used  to  facilitate  the  determination  of  system  parameters.  Such  rules  were  given  earlier 
in  this  report  under  Test  Methods  and  Results.  In  general,  the  parameter  values  given  by  various  rules 
of  thumb  should  fall  within  the  following  ranges: 

1 . The  receiver  IF  bandwidth  should  be  at  least  twice  the  optimum  for  data-recording  purposes.  The 

equivalent  bandwidth  of  the  prerecording  and  f ost-recordii*  IF  combination  ih.uL'  be  ck«  lu  f 
but  less  than  2ffl  for  NRZ  and  close  to  2ffi  but  less  than  4ffl  for  DM.  B 

2.  The  peak-to-peak  RF  transmitter  deviation  should  lie  between  0.6f„  and  0.9fo  for  NRZ  and  be- 
tween 1 ,2fB  and  1 ,8fB  for  DM.  B 

3.  The  premodulation  filter  bandwidth  should  fall  between  0.5fB  and  1.0fB  for  both  NRZ  and  DM. 

This  investigation  verifies  the  conclusions  of  Dr.  Lindsey  (reference  6)  and  Dr.  Hedeman  that  for 
equivalent  bit  rates  under  their  optimum  transmission  conditions  NRZ  is  3 dB  better  than  DM;  for 
equivalent  BEP,  the  bit  rate  of  NRZ  is  approximately  twice  that  of  DM.  Thus  DM  is  not  recommended 
for  applications  of  maximum  data  transfer  over  a bandbrntii-J  FT  tyueiu  wbeie  tioity  signals  may  be 
received 


REFERENCES 

1.  Kotel’nikov,  V.  A.  “The  Theory  of  Optimum  Noise  Immunity.”  McGraw-Hill,  N Y , '960 

2.  Smith,  E.  F.  “Attainable  Error  Probabilities  in  Demodulation  of  Random  Binary  PCM/FM  Wave- 
forms; IRE  Transactions  on  Space  Electronics  and  Telemetry.  Vol.  SET-8  (Dec  1962)  pp.  290-7. 

3.  ‘Telemetry  System  Study”  Final  Report,  Aeronutronic  Publication  U-743  (Dec  1959);  U.S.  Army 
Signal  Research  and  Development  Laboratories  Contract  No.  DA-36-039  (SC-73182)  Proj.  No. 

3 "1 6-00-300  • 

4.  Electro-Mechanical  Research  Inc.  “Experimental  Determination  of  Signal-to-Noise  Relationships 

in  PCM  FM  and  PCM  PM  Transmission,”  by  L.  R.  Brown.  NASA  Contract  NAS  5-505,  20  Oct  61. 

5.  Shaft,  P.  D.  “Error  Rate  of  PCM-FM  Using  Discriminator  Detection,”  IEEE  Transactions  on  Space 
Electronics  and  Telemetry:  Vol.  SET-9  Dec  1963.  pp.  131-7. 

6.  Naval  Missile  Center.  Bit  Synchronization  System  Performance  Characterization,  Modeling,  and 
Tradeoff  Study,  by  W.  C.  Lindsey.  Point  Mugu,  California.  4 Sef  1973  (Technical  Publication 
TP-73-iS)  UNCLASSIFIED. 


25 


INITIAL  DISTRIBUTION 


EXTERNAL 


Copies 


Commander 

Naval  Air  Systems  Command 
Headquarters 
Washington,  DC  20361 
Attn:  AIR-954 
AIR-6502 


2 

1 


Defense  Documentation  Center 
Cameron  Station 
Alexandria,  VA  22314 
Attn:  TIOG 


Commander 

Naval  Sea  Systems  Command 
Headquarters 
Washington,  DC  20362 
Attn:  SEA- 653 
SEA- 6 54 3 


1 

1 


Commanding  Officer 
U.S.  Naval  Station 
Atlantic  Fleet  Weapons  Range 
Box  545 

FPO  New  York  09551 
Attn:  H.  T.  Herring 


Commanding  Officer 
Naval  Avionics  Facility 
21st  and  Arlington  Avenue 
Indianapolis,  IN  46218 
Attn:  Code  906 

Code  814  J.  Brining 
Code  814  K.  Burgess 


Off icer-in-Charge 
Fleet  Analysis  Center 
Naval  Weapons  Station,  Seal  Beach 
Corona,  CA  91720 
Attn:  Code  851  1 

Code  8513  1 

Code  8522  1 


Commanding  Officer 
Naval  Ship  Weapons  Systems 
Engineering  Station 
Port  Hueneme,  CA  93043 
Attn:  Code  4920  (C.B.  Tendick)  1 


EXTERNAL 


Copies 


Commander 

Naval  Weapons  Center 
China  Lake,  CA  93555 
Attn:  Code  373  1 

Code  374  1 

Code  533  1 

Code  3742  (R.E.  Rockwell)l 

Code  451  (K.D.  Cox)  1 


Commander 

Naval  Air  Test  Center 
Telemetry  Data  Center  TDS/TSD 
Patuxent  River,  MD  20670 
Attn:  H.O.  Norfolk,  Jr.  4 


Commander 

U.S.  Naval  Ordnance  Missile 
Test  Facility 

White  Sands  Missile  Range, 
NM  88002 

Attn:  Code  503  (W.W.  Bohn) 


DDR&E  AD(TR) 

The  Pentagon 
Washington,  DC  20301 
Attn:  J.  A.  Webster 


Director  of  Navy  Ranges 
and  Targets 
Point  Mugu,  CA  93042 
Attn:  AIR-630 
AIR-6302 
AIR-6302D 


1 

1 

1 


Commanding  Officer 
Yuma  Proving  Ground 
Yuma,  AZ  85364 
Attn:  STEYP-MDP 

(H.  S.  McElfresh) 


\ 


Commanding  Officer 
Ballistic  Missile  Defense 
Systems  Command 
PO  Box  1500 
Huntsville,  AL  35807 
Attn:  BMDSC-RD  (G.E.  Wooden)  1 
BMDSC-ROO  (D.E.  Parker)  1 


27 


\ 

\ 


G 


JPRSCEDIN3  PAGE  BLANK- NOT  FILMED 

■■  i i 1- JL 


,4-W  'll  Til  -IqbiiJ 


external 


Copies  EXTERNAL 


PICATINNY 

Bldg  63,  SMUPA-RX 

Dover,  NJ  07801 

Attn:  Telemetry  Division,  TSD 

(J.  C.  Greenfield)  1 

Commanding  Officer 
U.S.  Army  Missile  Command 
Redstone  Arsenal,  AL  35809 
Attn:  AMSMI-RTA,  Bldg.  7855 

(R.  D.  Bibb)  ^ 

Commanding  Officer 
U.S.  Army  Electronic 
Proving  Ground 
Fort  Huachuca,  AZ  85613 
Attn:  STEEP-MT-I  (W.T.  Lyon)  1 

commanding  General 
Department  of  the  Army 
White  Sands  Missile  Ranee 
NM  88002  ’ 

Attn:  STEWS-ID-IT 

(B.  Chin)  x 

(J.  D.  Cates)  i 

STEWS-ID-S 

(B.  E.  Norman)  1 

(E.  Bejarno)  2 

ADTC 

Eglin  AFB , FL  32542 
Attn:  ADTC  - TSGGS 

(E.  J.  Poschell)  1 

ADTC  - TSGPG 

(B.  L.  Culberson)  1 

Vitro  Services 
(R.  L.  Manley)  2 

AFSWC 

Holloman  AFB,  NM  88330 
Attn:  6585  TG/TKIA 

(J.  A.  Haden)  2 

AFSWC 

Kirtland  AFB,  NM  87115 
Attn:  AFSWC  - TED 

(CAPT  R.  j.  Skelton)  1 

4900  TG/FTEI 

(J • C.  Scott)  2 


AFCRL  - LCS 

L.  G.  Hanscom  Field 

Bedford,  MA  01730 

Attn:  C.  H.  Reynolds  ] 

SAMTEC 

Vandenberg  AFB,  CA  93437 
Attn:  SAMTEC  - ENIT 

(D.  K.  Manoa)  2 

SAMTEC  - R00E 
(K.  0.  Schoeck)  2 

AFETR 

Patrick  AFB,  FL  32925 

Attn:  AFETR  - ENIL  (F.  Mann)  l 

AFFTC 

Edwards  AFB,  CA  93523 

Attn:  AFFTC  - DOESD  (J.  Ramos)  1 

Sandia  Laboratories 
Kirtland  AFB  East 
Albuquerque,  NM  87115 
Attn:  Division  1251 

(R.  S.  Reynolds)  2 

Division  9486,  Box  5800 
(P.  L.  Walter)  2 

Division  9421,  Box  5800 
(H.  0.  Jeske)  2 

AFSCF 

Sunnyvale,  CA  94086 
Attn:  Det  1,  AFSCF  - D0ZR 

(G.  V.  Kreider)  2 

AFSCF 

Aerospace  Corporation 
P.0.  Box  92957 
Los  Angeles,  CA  90009 
Attn:  C.  Nakamura 

Bldg.  115/1341  2 

National  Bureau  of  Standards 
Electronic  Division 
Bldg.  225,  Room  A109 
Gaithersburg,  MD  20760 
Attn:  GSA-FSS 

(J.  J.  Fiori)  2 


"V  - »• 


28 





wmm* 

S!3r51!!2-3in!S 


pppws* 


EXTERNAL 


Copies 


National  Bureau  of  Standards 
Washington,  DC  20234 
Attn:  Div.  425.03 

(P.  S.  Lederer) 


Aerospace  Corporation 
Suite  4040 

955  L'Enfant  Plaza  S.W. 
Washington,  DC  20024 
Attn:  Dr.  W.  R.  Hedeman 


National  Aeronautics  and 
Space  Administration 
Goddard  Space  Flight  Center 
Greenbelt,  MD  20771 
Attn:  Code  730.4 

(W.  B.  Poland,  Jr.) 
Code  734 
(R.  M.  Muller) 

(C.  Trevathan) 


1 

1 


National  Aeronautics  and 
Space  Administration 
1520  H Street  N.W. 
Washington,  DC  20546 
Attn:  Code  MF 

(W.  E.  Miller,  Jr.) 


Marshall  Space  Flight  Center 
Huntsville,  AL  35812 
Attn:  Code  S&E-ASTR-I 
(W.  Threlkeld) 


Boeing  Company,  ASG,  The 
Systems  Instrumentation  and 
Telemetry 
Seattle,  WA  98]«|4 
Attn:  Org.  2-5180 
MH  8H-07 
(V.  A.  Jennings) 


Data  Control  Systems,  Inc. 
Commerce  Drive 
Danbury,  CT  06810 
Attn:  M.  Pizzuti 

D.  A.  Zeller 


1 

1 


Data  Control  Systems,  Inc. 
East  Liberty  Street 
Danbury , CT  06810 
Attn:  T.  Castellano 


EXTERNAL 


Copies 


EMR  Telemetry 
P.0.  Box  3041 
Sarasota,  FL  33578 
Attn:  S.  Bass 

G.  Tremain 


1 

1 


Federal  Electric  Corporation 
Vandenberg  AFB,  CA  93437 
Attn:  R.  Streich 


Grumman  Aerospace  Corporation 
Instrumentation  PLT  6 
Calverton,  NY  11933 
Attn:  R.  S.  McElhiney  1 

E.  Char land  1 


Harris  Electronic  Systems  Div. 
P.0.  Box  37 
Melbourne,  FL  32935 
Attn:  C.  Curry 


IED 

7440  Ronson  Road 
San  Diego,  CA  92111 
Attn:  C.  Brewster 


Jet  Propulsion  Laboratory 
4800  Oak  Grove  Drive  (156-142) 
Pasadena,  CA  91103 
Attn:  R.  Piereson 


Martin-Marietta  Corporation 

P.0.  Box  179 
Denver , CO  80201 
Attn:  MS  8200 

(C.  M.  Kortman) 

(S.  Hurst) 


McDonne ll-Douglas  Astronautics 
5301  Bolsa  Avenue 
Huntington  Beach,  CA  92647 
Attn:  A3-833-BBS0 

(D.  R.  Andelin) 
A3-250-AG30 
(R.  M.  Frahm) 


Microcom 
1115  Mearns  Road 
Warminster,  PA  18974 
Attn:  C.  Rosen 
T.  Grady 


29 


IL  III  IJJIJI  JIM 

f -^T  ® 


EXTERNAL 


Copies 


Monitor  Systems 
401  Commerce  Drive 
Fort  Washington,  PA 
Attn:  D.  Mahoney 


19034 


Sandia  Corporation 
Box  5800 

Albuquerque,  NM  87115 
Attn:  Division  9421 

(C.  S.  Johnson) 


Teledyne  Telemetry  Company 
1901  South  Bundy  Drive 
Los  Angeles,  CA  90025 
Attn:  H.  F.  Pruss 


University  of  Michigan 
Computer,  Information  and 
Control  Engineering 
East  Engineering,  1520 
Ann  Arbor,  MI  48104 
Attn:  L.  L.  Rauch,  PhD 


University  of  Southern 
California 

Olin  Hall,  University  Park 
Los  Angeles,  CA  90007 
Attn:  Dr.  W.  C.  Lindsey 


Mr.  E.  J.  Baghdady 
21  Overlook  Drive 
Weston,  MA  02193 


Mr.  L.  W.  Gardenhire 
906  Whitmire  Drive 
Melbourne,  FL  32935 


Mr.  E.  H.  Straehley 
Consulting  Services 
1631  Platevna  Road 
Santa  Barbara,  CA  93103 


EXTERNAL 


Copies 


30 


; ; • :■  "X  )/f  i 

'ti* 


mm 


INTERNAL 


Copies  INTERNAL 


Technical  Director 
Code  0002 
W.  L.  Miller 


Project  Management  Group 
Code  0100 

LT  COL  C.  L.  Phillips 
Code  0150 
M.  H.  Cain 


1 

1 


Systems  Evaluation  Directorate 
Codo  1001 
CAPT  D.  D.  Buck 
Code  1002 
R.  S.  Nelson 


Weapons  Systems  Test 
Department 
Code  1142 
C.  N.  Gorevin 


RF  Weapons  Systems  Division 
Code  1262 
D.  H.  Barthuli 


Weapons  Evaluation  Department 
Code  1300 
Dr.  M.  A.  Garcia 


Fleet  Weapons  Engineering 
Directorate 
Code  2001 
CAPT  D.  DeWitt 


Range  Directorate 
Code  3001 

CAPT  G.  M.  Dempsey 


Capabilitities  Development 
Department 
Code  3100 
W.  C.  Christensen 
Code  3101 
J.  A.  Means 


1 

1 


Development  Management 
Division 
Code  3120 
H.  J.  Crawford 


Copies 


Measurement  Systems 
Development  Division 
Code  3140 
T.  R.  Carr 
Code  3142 
R.  W.  Lytle 
C.  G.  Ashley 
J.  Anderson 
Code  3143 
M.  A.  Bondelid 
Code  3145 
Z.  H.  Blankers 


Control  Systems  Development 
Division 
Code  3150 
J.  R.  Scott 
Code  3151 
A.  H.  Weishaar 
J.  L.  Weblemoe 
Code  3153 
D.  E.  Ogden 
Code  3154 
P.  M.  Newton 
C.  E.  Ohlen 


System  Development  Support 
Division 
Code  3160 

E.  F.  Mutz 
Code  3161 

R.  L.  Nifong 
Code  3163 

F.  L.  Allard 


Data  Collection  Division 
Code  3410 
R.  P.  Carpenter 
Code  3411 
R.  J.  Roanhaus 
R.  H.  Wymore 
J.  T.  Dickman 
W.  T.  Norton 
Code  3412 
F.  D.  Leiblein 


Data  Control  Division 
Code  3440 
W.  F.  Tolman 


1 

1 

1 


1 

1 


1 

1 


1 

1 


1 

1 

1 


1 

1 

1 

1 


% 


31 


""  'iwwpw^- 


INTERNAL 


Copies 


Engineering  Applications 
Directorate 
Code  4001 
N.  F.  Rohn 
Code  4002 
LCDR  R.  P.  Evans 


Equipment  Support  Division 
Code  4131-1 
R.  W.  Kenny 


Technical  Publications 
Division 
Code  4250-1 
M.  F.  Hayes 
Code  4251 
D.  S.  Walker 
Code  4253-3 

Technical  Reports  Library 


Instrumentation  Department 
Code  4300 
V,  E.  Orris 


Instrumentation  Development 
Division 
Code  4320 
K.  L.  Berns 
Code  4321 
F.  R.  Hartzler 
D.  R.  Hust 

D.  H.  Rilling 
M.  V.  Wechsler 

Code  4322 

E.  R.  Hill 
D.  R.  Knight 
D.  A.  King 

D.  L.  Laubacher 
Code  4323 

E.  L.  Law 

B.  E.  Bishop 
E.  T.  Kimball 

C.  M.  Kaloi 


10 


1 

1 

1 

1 


1 

1 

1 

1 


50 

1 

1 

1 


INTERNAL 


Copies 


Instrumentation  Engineering 
Division 
Code  4330 

R.  P.  Tegt  1 

Code  4331 

M.  A.  Beckman  1 

T.  P.  Waddell  1 


Instrumentation  Support 
Division 
Code  4340 
J.  D.  Martin 
Code  4341 
E.  R.  Sandy 
G.  J.  Harbold 
N.  R.  Dragoo 
D.  J.  DeJong 


1 

1 

1 

1 


Office  of  Patent  Counsel 
Code  PC 

Dr.  J.  M.  St.  Amand 


32 


****3?0W| |W*%' 


piil',' 


THIS  REPORT  HAS  SEEN  DELIMITED 
AND  CLEARED  FOR  PUBLIC  RELEASE 

UNDER  DOD  DIRECTIVE  5200.20  AND 

NO  RESTRICTIONS  ARE  IMPOSED  UPON 
ITS  USE  AND  DISCLOSURE. 

DISTRIBUTION  STATEMENT  A 

APPROVED  FOR  PUBLIC  RELEASE j 
DISTRIBUTION  UNLIMITED.