DTIC ADA044772: FM Quieting Curves and Related Topics

Survival, Water, Medical Field Manuals

Military Manuals

Defense Technical Information Center

Document text

Li.  bbOy 


AFCS  TECHNICAL  REPORT 
FM  QUIETING  CURVES 
and  Related  Topics 


SYSTEM  TECHNICAL  APPLICATIONS  FACILITIES 
1842  ELECTRONICS  ENGINEERING  GROUP  (AFCS) 
RICHARDS-GEBAUR  AIR  FORCE  BASE,  MISSOURI 


30  SEPTEMBER  1977 


1842  ELECTRONICS  ENGINEERING  GROUP 
MISSION 


The  1842  Electronics  Engineering  Group  (EEC)  is 
organized  as  an  independent  group  reporting 
directly  to  the  Commander , Air  Force  Communica- 
tions Service  (AFCS)  with  the  mission  to  provide, 
communi cati ons -el ectronl cs -meteorol og i cal  ( CEM) 
systems  engineering  and  consultive  engineering 
for  AFCS..  In  this  respect,  1842  EEG  responsi- 
bilities include:  Qeye.l oping  engineering  apd 
installation  standards  for  use  in  planning, 
programming,  procuring,  engineering*  ins  tailing 
and  testing  CEM  systems,  facilities  and  equip- 
ment; performance  of  systems  engineering  of  CEM 
requirements  that  must  operate  as  a system  or  , 
in  a system  environment;  operation  of  a special- 
ized Digital  Network  System  Facility  to  analyze 
and  evaluate  new  digital,  technology  for  applica- 
tion to  the  Defense  Communications  System  (DCS) 
and  other  special  purpose  systems;  operation  of 
a facility  to  prototype  systems  and  equipment 
configurations  to  check  out  and  validate  ehgi-  . 
neerlng-i nstall ati on  standards  and  new  1 nstalla- 
tion  techniques;  providing  consultive  CEM 
engineering  assistance  to  HQ  AFCS,  AFCS  Areas, 
MAJCOMS,  POD  and  other  government  agencies » 


security  CLASSIFICATION  OF  This  PACE  (Wh»n  Dale  Entered) 

REPORT  DOCUMENTATION  PAGE 


t REPORT  NUMBER 

TR  77-18 

\ 4.  TITLE  (and  Submit, ■ 

/ FM  Quieting  Curves 

& and  Related  Topics  , 


rr  ah  thorc«> 


SAGE  Read  instructions 

BEFORE  COMPLETING  FORM 

J,  GOVT  ACCESSION  NO.  3.  RECIPIENT'S  CAT ALOG  NUMBER 

N/A  N/A 

J.  TYPE  OF  REPORT  A PERIOD  COVERED 


Final  , 


j fyJ  Georee  M./Kizer 


t.  PERFORMING  ORGANIZATION  NAME  AND  ADDRESS 

1842  EEG/EETET 

Richards-Gebaur  AFB,  MO  64030 


«.  PERFORMING  ORG.  REPORT  NUMBER 
N/A 

• ■ CONTRACT  OR  GRANT  NUMBER/*; 

N/A 


QT\  > M\£gGr/u  T£T-TR-77-/r\ 


10.  PROGRAM  ELEMENT,  PROJECT,  TASK 
AREA  6 WORK  UNIT  NUMBERS 


II.  CONTROLLING  OFFICE  NAME  AND  ADDRESS 


1842  EEG/EETET  /'  y \(t /)  AugSB*  W77  f 

Richards— Gebaur  AFB,  MO  64030  \ **  j ) iJT  number  of  pages 

L ; / f 229 

14.  MONITORING  AGENCY  NAME  & AOORESS  (II  dlllerenl  from  Controlling  Olllce)  15.  SECURITY  CLASS,  (ol  thle  report) 

N/A  Unclassified 


is*.  DECLASSIFICATION/ DOWNGRADING 
SCHEDULE 


j 1«.  DISTRIBUTION  STATEMENT  (ol  ible  Report; 

Approved  for  public  release.  Distribution  unlimited. 


17.  DISTRIBUTION  STATEMENT  (ol  Iba  eh.tr  acl.entkred  In  Block  JO*  W dltlera^Nrom  Report) 

n/a 


St  etf 


• a.  supplementary  notes 


1*.  KEY  WORDS  (Cnntlnu a on  rororao  *ld*  II  neceeeary  and  Idtnllly  by  block  number) 

Frequency  Modulation  radio  performance 

Phase  Modulation  radio  parameters 

Noise  FM 

Quieting  Curves  PM 


ABSTRACT  (Continue  on  rovr too  oldo  II  nacaooary  and  Identity  by  block  number ) 

his  report  describes  the  relation  between  readily  measured  frequency  modulation 
radio  transmitter  and  receiver  parameters  and  the  basic  baseland  signal  and 
noise  performance  of  the  radio  transmitter/receiver  combination.  Equations, 
graphs,  and  charts  are  provided  to  facilitate  application. 


DD  1 JANM7I  1473  EDITION  OF  1 NOV  SI  IS  OBSOLETE 


APPROVAL  PAGE 


This  report  has  been  reviewed  and  approved  for  publication  and  distri- 
bution. 


1'  J Lpdt 

ZYGMUND  J.  BARA 
1842  EEG/EET  (AFCS) 
Chief,  Command,  Control 
Division 


Defense  Communication  Systems  Engineering 


WILLIAM  R.  P?'OV7N’~M«ior,  U?Tf  ~ 
1842  EEG/EE1E 

Chief,  Engi/.oering  Services  i ranch 


George  M.  Kizer 

GEORGE  mV  KIZER 
1842  EHG/EETET 

Electronics  Engineer,  Author 


TABLE  OF  CONTENTS 


Para  No  Page 


1 Introduction  1 

2 Basic  Modulation  Theory  4 

3 Wideband  FM  Receiver  Baseband  Signal  and  Noise 

Characteristics  21 

*4  Experimental  Verification  47 

5 Predicting  FM  M/W  Terminal  Thermal  Noise  Performance  109 

6 Baseband  Signal  and  Slot  Noise  Versus  RSL  Using 

Generalized  Charts  119 

7 FM  Microwave  Radio  Terminal  Equipment  Parameters  150 

8 A Digital  FM  System  214 

9 Conclusions  220 

10  Recommendations  223 

1 1 Bibliography  224 

Distribution  List  77Q 


LIST  OF  ILLUSTRATIONS 


Figure  No.  Page 


I FM  RF  Spectrum,  Sinewave  Modulation,  Integer  Betas  6 

? FM  RF  Spectrum,  Sinewave  Modulation,  Carrier  Dropouts  7 

3 FM  RF  Spectrum,  Sinewave  Modulation,  Sideband  Dropouts  8 

^ Slot  Noise  Versus  Received  Signal  Level  22 

5 Typical  Microwave  Oscillator  Induced  Baseband  Noise  38 

6 M/W  Transmitter  Voltage  Controlled  Oscillator  Noise  39 

7 M/W  Receiver  Easeband  Noise  Versus  Received  Signal 

Level  40 

8 Baseband  Noise  Versus  Carrier  to  Thermal  Noise  for 

Hard  and  No  Limiting  44 

9 LOS  M/W  Receiver  Simplified  Diagram  48 

10  LOS  M/W  Receiver  IF  Response  50 

II  TROPO  M/W  Receiver  Simplified  Diagram  51 

IP  TROPO  M/W  Receiver  IF  Response  53 

13  LOS  M/W  Receiver  Baseband  Noise  Versus  Received 

Signal  Level  54 

1M  LOS  M/W  Receiver  Baseband  Noise  Versus  Received 

Signal  Level  55 

15  TROPO  M/W  Receiver  Baseband  Noise  Versus  Received 

Signal  Level  55 


LIST  OF  ILLUSTRATIONS  (CONT.) 


Figure 

No. 

Page 

15.1 

Baseband  Slot  Noise, 

No  Carrier  Present 

58 

16 

TROPO  Quieting  Curve 

, Normal  Configuration 

59 

.17 

TROPO  Quieting  Curve 

, 20  dB  PrelF  Attenuat  ?n 

60 

18 

TROPO  Quieting  Curve 

, <40  dB  PrelF  Attenuation 

61 

19 

TROPO  Quieting  Curve 

, 60  dB  PrelF  Attenuation 

62 

20 

LOS  Quieting  Curve, 

Normal  Configuration 

63 

21 

LOS  Quieting  Curve, 

10  dB  PrelF  Attenuation 

64 

22 

LOS  Quieting  Curve, 

20  dB  PrelF  Attenuation 

65 

23 

TROPO  Quieting  Curve 

, 20  dB  PrelF  Attenuation 

66 

2*1 

TROPO  Quieting  Curve 

, 40  dB  PrelF  Attenuation 

67 

25 

TROPO  Quieting  Curve 

, 60  dB  PrelF  Attenuation 

68 

26 

LOS  Quieting  Curve, 

10  dB  PrelF  Attenuation 

69 

27 

LOS  Quieting  Curve, 

20  dB  PrelF  Attenuation 

70 

28 

LOS  Quieting  Curve, 

30  dB  PrelF  Gain 

71 

29 

Measured  Slot  Noise 

Near  FM  Threshold  Due  to 

White  Noise  Baseband  Modulation 

76 

30 

Quieting  Curves,  Two  Quieting  Sources 

79 

31 

Quieting  Curves,  Three  Quieting  Sources 

80 

32 

Wide  Slot  Quieting  Curves 

82 

33 

Wide/Narrow  Slot  Quieting  Curve  Comparison 

83 

34 

TROPO  Baseband  Signal  Suppression 

86 

35 

LOS  Baseband  Signal  Suppression 

87 

36 

Quieting  Curve,  Quieting  Source  Frequency  Offset 

88 

37 

Quieting  Curve,  Unmodulated  RFI 

91 

38 

Quieting  Curve,  Unmodulated  RFI 

92 

table  OF  CONTENTS  (CONT.) 


Figure  No. 

39  Quieting  Curve,  Unmodulated  RFI  93 

40  Quieting  Curve,  Unmodulated  RFI  94 

41  Quieting  Curve,  Unmodulated  RFI  95 

4?  Quieting  Curve,  Unmodulated  RFI  96 

43  Baseband  Slot  Noise  Versus  Interferring  Carrier 

Frequency  and  Power  97 

44  LOS  Quieting  Curve,  LC-8  98 

45  LOS  Baseband  Noise  Versus  Received  Signal  Level, 

LC-8  99 

46  LOS  Baseband  Signal  Suppression  Versus  Received 

Signal  Level , LC-3  100 

47  Sample  Quieting  Curve  101 

48  Sample  Quieting  Curve  102 

49  Sample  Quieting  Curve  103 

50  Sample  Quieting  Curve  104 

51  Sample  Quieting  Curve  105 

52  Sample  Quieting  Curve  1°6 

53  Sample  Quieting  Curve  1°7 

54  Sample  Quieting  Curve  108 

115- 

54.1  Noise  Equations  Factors  116 

54.2  Noise  Equations  Error  Analysis  117 

55  Theoretical  Quieting  Curve,  Rectangular 

IF  Response 

56  Theoretical  Quieting  Curve,  Gaussian  IF  Response  12i 

iv 


LIST  OF  ILLUSTRATIONS  (CCNT.) 


Figure  No.  Page 

57  Theoretical  Quieting  Curve,  Average  IF  Response  122 

58  Theoretical  Quieting  Curve  Example  126 

59  Theoretical  Baseband  Signal  Suppression  Curve  127 

60  Combiner  Improvement  Curves  134 

61  Power  Addition/Subtraction  Curves  136 

6?  LOS  M/W  Receiver  Quieting  Curve,  No  De-emphasis  140 

63  TROPO  M/W  Receiver  Quieting  Curve,  CCIP  De-emphasis  144 

6R  TROPO  M/W  Receiver  Quieting  Curve,  No  De-emphasis  145 

65  LOS  M/W  Receiver  Theoretical  Quieting  Curve  146 

66  TROPO  M/W  Receiver  Theoretical  Quieting  Curve  147 

67  Generalized  M/W  Transmitter  and  Receiver  151 

68  Example  Transmitter  Configurations  156 

159- 

69  FM  RF  Spectrum,  Sinewave  Modulation,  Various  Dropouts  163 

70  FM  RF  Spectrum,  Distorted  Sinewave  Modulation,  First 

Carrier  Dropout  165 

71  FM  RF  Spectrum,  Distorted  Sinewave  Modulation,  First 

First  Sideband  Dropout  166 

’!?  FM  RF  Spectrum,  Square  Wave  Modulation,  First  Carrier 

DroDOut  167 

73  FM  RF  Spectrum,  Square  Wave  Modulation  168 

7R  CCIR/EIA  Emphasis  Curves  170 

7^.1  REL  Emphasis  Curves  171 

75  Baseband  Noise  Spectrum  Versus  Received  Signal  Level, 

No  De-emphasis  and  CCIR  De-emphasis  172 

76  CCIR/Time  Constant  Emphasis  Networks  175 


V 


LIST  OF  ILLUSTRATIONS  (CCNT.) 


Figure  No. 

77  Serrasoid  Baseband  Circuitry  185 

7s  Simplified  Serrasoid  Baseband  Circuitry  186 

78.  1 Gaus3ian/Rectangular  IF  Power  Response 

79  Peak/RMS  Voltage  Ratios  195 

80  FM  RF  Spectrum,  Heavy  White  Noise  Modulation  189 

81  FM  RF  Spectrum,  Light  White  Noise  Modulation  2no 

8?  Idealized  Thermal  Noise  Power  transfer  203 

83  Idealized  Amplifier 

84  Noise  Figure  of  Caseaded  Devices  206 

85  Noise  Figure  of  Lossy  Network  and  Amplifier  206 

86  Simplified  Noise  Figure  Meter  209 

87  Theoretical  BER  Curves,  Various  Noise  Sources  215 

88  Theoretical  and  Measured  BER  Curves,  Gaus3iar.  White 

Noise  216 

89  Theoretical  and  Measured  BER  Curves,  LOS  M/W  Receiver  219 

90  Typical  Areas  of  Quieting  Curves  Degradation  221 


LIST  OF  TABLES 


Table  No.  Page 


I FM  RF  Spectrum,  Sinewave  Modulation,  Voltage  Ratios  9 

? FM  RF  Spectrum,  Sinewave  Modulation,  Decibels  10 

3 FM  RF  Spectrum,  Square  Wave  Modulation,  Voltage  Ratios  11 

f)  FM  RF  Spectrum,  Square  Wave  Modulation,  Decibels  12 

5 Normalized  Slot  Noise,  Rectangular  and  Gaussian  IF 

23- 

Responses  27 

28- 

6 Normalized  Slot  Noise,  Averaged  IF  Response  32 

7 Theoretical  Narrow  Slot  20/30  dB  Noise  Quieting  34 

8 Theoretical  Narrow  Slot  FM  Threshold  35 

9 Theoretical  Wide  Slot  FM  Threshold  36 

10  Theoretical  Baseband  Signal  Suppression  42 

II  Theoretical  Baseband  Slot  Noise  for  Hard  and  No  Limiting  45 

12  LOS  M/W  Receiver  Characteristics  49 

13  TR0P0  M/W  Receiver  Characteristics  52 

l1!  Measured  FM  Thresholds  72 

15  Measured  20  dB  Noise  Quieting  73 

16  Baseband  Slot  Noise  Near  FM  Threshold  Due  to  White 

Noise  Baseband  Modulation  75 

17  Narrow  Slot  Impulse  Noise  Versus  Received  Signal  Level  79 

18  Measured  Baseband  Signal  Suppression 

19  Slot  Noise  Measurement  Conversion  Factors  123 

vii 


LIST  OF  TABLES  (CQNT.) 

Table  No.  Page 

20  Recommended  Quieting  Curve  Slot  Frequencies  225 

2)  CCIR/EIA  Emphasis  Values  130 

21.1  REL  Emphasis  Values  131 

22  Sample  Quieting  Curve  Data  137 

2 3 Sample  Quieting  Curve  Data  138 

2*4  Sample  Quieting  Curve  Data  139 

2^  Measured  De-emphasis  Values  141 

26  Sample  Quieting  Curve  Data  142 

27  Sample  Quieting  Curve  Data  143 

28  Betas  and  Dropouts  for  Sinewave  Frequency  Modulation  153 

29  Dropout  Power  Levels,  Sinewave  Frequency  Modulation  158 

30  Pivot  Frequencies  182 

31  Emphasis  Approximation  Error  Analysis  183 

32  Rectangular/Gaussian  IF  Responses(deleted-see  Figure  78.1)  

33  RF  Bandwidth  Factors  197 

31)  RF  Bandwidths  for  CCIR  Deviation  Recommendations  198 


viii 


1.  Introduction. 


1.1  With  all  the  articles,  reports,  and  books  written  about  frequency 
modulation,  transmittsrs  and  receivers,  why  write  another  report?  Despite 
the  volumes  written  about  frequency  modulation  equipment,  several  important 
features  have  been  neglected.  Other  significant  features  have  been 
described,  but  are  spread  over  so  many  different  sources  or  buried 
in  so  much  extraneous  material  that  it  has  been  difficult  to  get  the 
important  information  and  put  it  in  perspective.  The  principle  interest 
in  writing  this  report  has  been  to  give  engineers  and  technicians  the 
information  necessary  to  understand  and  predict  the  basic  performance 
of  frequency  modulation  microwave  radio  terminals.  A thorough  knowledge 
of  the  performance  of  communication  systems  is  required  to  adequately 
specify,  develop,  test,  engineer,  install,  and  verify  proper  operation 
of  an  item  of  communication  equipment.  For  frequency  modulation  microwave 
radio  terminals,  this  includes  a knowledge  of  the  baseband  signal  and 
noise  properties  of  the  terminal  transmitter  and  receiver  configuration. 

The  noise  at  the  baseband  of  an  analog  microwave  receiver  is  a complex 
mixture  of  noise  from  such  sources  as  radio  terminal  thermal  and  inter- 
modulation noise,  crosstalk,  and  frequency  division  multiplexer  noise. 

This  noise  can  generally  be  divided  into  two  broad  categories.  The 
first  category  is  noise  which  is  independent  of  the  signal  applied  to 
the  terminal  baseband.  This  noise  includes  the  effect  of  thermal  noise 
generated  within  the  receiver,  the  phase  noise  of  the  transmitter,  cable 
crosstalk  from  other  baseband  signals,  radio  frequency  interference, 
and  multiplexer  tone  leakage,  power  supply  ripple,  and  miscellaneous 
sources.  Since  this  noise  has  no  relationship  to  the  baseband  signal, 
it  can  be  measured  with  the  baseband  signal  removed  from  the  transmission 
equipment.  It  is  called  idle  noise.  The  other  main  noise  category 
includes  the  noise  which  is  directly  related  to  the  frequency  and  power 
level  of  baseband  signal.  Although  this  noise  includes  such  effects 
as  baseband  cable  and  radio  terminal  crosstalk,  this  type  of  noise  is 
loosely  termed  intermodulation  noise.  This  noise  can  only  be  measured 
with  a signal  applied  to  the  transmission  equipment  oaseband.  The 
primary  tool  for  predicting  and  evaluating  idle  radio  noise  is  the 
frequency  moduiation  (FM)  slot  noise  quieting  curve.  The  primary  tool 
for  evaluating  radio  intermodulation  noise  is  the  noise  power  ratio 
(NPR)  bucket  curve.  Bucket  curves  can  be  used  to  predict  and  analyze 
such  noise  factors  as  waveguide  and  free  space  multipath  effects  and 
microwave  terminal  differential  gain  and  phase.  Utilization  of  bucket 
curves  requires  a full  understanding  of  quieting  curves.  Unfortunately, 
present  job  duties  preclude  the  investigation  necessary  to  write  a 
report  on  bucket  curves.  This  report  will  only  cover  FM  aspects  of 
noise  quieting  curves  and  related  topics.  Nevertheless,  quieting 
curves  are  quite  important  and  deserve  considerable  attention.  Idle 
noise  sets  the  lower  limits  on  the  noise  performance  of  an  analog 
digital  microwave  radio  terminal.  The  thermal  noise  component  of  a 
microwave  receiver  is  the  dominant  factor  in  baseband  signal  and  noise 
performance  of  the  radio  terminal  for  low  power  level  radio  signals. 


1 


Therefore,  thermal  noise  of  the  terminal  directly  effects  fade  margin 
and  radio  link  reliability  as  well  as  terminal  noise  performance. 

Although  the  present  report  is  directed  toward  convencional  frequency 
division  multiplexed  analog  baseband  transmission,  the  information  is 
directly  applicable  with  minor  modification  to  the  transmission  of  wide- 
band video  or  digital  quasi-analog  signals  over  analog  microwave  equipment. 
An  example  will  be  given  using  quieting  curve  concepts  to  determine 
bit  error  rate  performance  of  the  three  level  partial  response  quasi- 
digital  FM  radio  terminals  currently  being  installed  in  the  Digital 
European  Backbone  (DEB)  digital  wideband  communications  upgrade. 

1.2  In  the  communications  business,  the  object  is  to  get  information 
from  one  place  to  another  with  the  least  degradation  in  quality.  Infor- 
mation that  must  be  transmitted  quickly  over  a long  diatance  usually 
takes  the  form  of  a telephone  call  or  a data  circuit  processed  for  trans- 
mission on  a telephone  circuit.  These  signals  are  routed  over  cables 
and  controlled  with  various  types  of  processing,  monitoring,  and  routing. 

If  the  signals  are  transmitted  very  far,  they  almost  invariably  make 
their  way  to  a microwave  (M/W)  transmission  system  for  long  distance 
bulk  transmission.  Prior  to  transmission,  as  many  as  several  hundred 
separate  telephone  channels  are  combined  (frequency  division  multiplexed 
or  FDMed  to  produce  a wide  frequency  range  (wideband)  baseband  signal) 
for  efficient  processing.  The  M/W  transmission  system  generally  consists 
of  several  separate  M/W  links,  each  consisting  of  a radio  path  and  two 
M/W  radio  terminals.  At  one  terminal,  the  combined  telephone  circuits 
(baseband)  are  converted  into  a modulated  radio  signal  with  frequency 

as  high  as  11  gigahertz  (GHz). 

1.3  The  modulated  radio  frequency  (RF)  signal  passes  through  nonlinear 
transducers,  imperfect  transmission  lines,  and  i3  transmitted  through 
the  air  where  it  is  susceptible  to  multipath,  interference,  and  fading 
degradations.  After  transmission  over  several  radio  links,  the  baseband 
signal  is  reconverted  (demultiplexed)  into  individual  telephone  circuits. 
While  the  telephone  circuits  exist  as  individual  channels,  they  are 
processed  by  individual  amplifiers  and  cables.  They  are  susceptible 

to  noise  and  other  forms  of  distortion.  The  failure  or  degradation 
of  a single  telephone  circuit,  while  undesirable,  is  not  immediately 
hazardous  to  many  customers.  Failure  or  degradation  of  a single  radio 
terminal,  however,  since  it  carries  many  telephone  circuits,  can  have 
serious  consequences.  For  this  reason,  considerable  attention  is  given 
to  M/W  radio  terminal  performance. 

1.4  As  a radio  terminal  transmits  a baseband  signal  from  one  location 
to  another,  the  signal  is  easily  influenced  by  nonideal  characteristics 
of  the  overall  transmission  medium.  The  effect  of  these  imperfections 
is  to  Introduce  noise  into  the  reconstituted  (demodulated)  baseband 

at  the  last  M/W  terminal.  The  processing  of  baseband  signals  at  a 
frequency  modulation  (FM)  M/W  radio  terminal  is  complex.  The  noise 
degradation  of  the  baseband  signal  can  be  grouped  roughly  into  two 


2 


categories.  The  first  type  is  noise  which  is  primarily  independent 
of  the  baseband  signal  itself.  This  noise  can  be  introduced  by  M/W 
transmitter  phase /frequency  noise  (phase  jitter),  various  forms  of 
interference  and  crosstalk,  and  internally  generated  ("front  end") 
receiver  thermal  noise.  For  simplicity,  these  types  of  noise  will  be 
called  simply  "idle  noise."  The  second  type  is  noise  which  is  directly 
related,  among  other  factors,  to  characteristics  of  the  baseband  signal 
itself.  Sources  of  this  type  of  noise  include  waveguide  echo  and  moding, 
multipath,  differential  gain  and  phase,  amplitude  modulation  to  phase 
modulation  (AM  to  PM)  conversion,  as  well  as  other  forms  of  intermodu- 
lation distortion.  For  simplicity,  this  type  of  noise  will  simply  be 
called  intermodulation  noise. 

1.5  The  primary  method  ol  characterizing  the  thermal  noise  performance 
of  a M/W  radio  is  through  the  use  of  what  is  called  an  FM  slot  noise 
quieting  curve  or  simply  quieting  curve.  The  quieting  curve  is  a plot 
of  noise  in  the  baseband  of  a M/W  FM  receiver  as  various  levels  of 
unmodulated  RF  signals  are  applied  to  the  receiver.  The  baseband  noise 
is  measured  with  a frequency  selective  voltmeter  (FSV)  with  a measurement 
bandwidth  of  3.)  kilohertz  (KHz),  the  nominal  frequency  width  of  a single 
telephone  circuit.  The  FSV  is  tuned  to  different  frequencies  in  the 
baseband  of  the  receiver  while  the  received  signal  level  (RSL)  applied 

to  the  receiver  is  varied.  By  plotting  these  noise  measurements  as 
a function  of  baseband  frequency  and  RSL,  it  is  possible,  given  an 
actual  operational  RSL  of  the  receiver,  to  predict  the  thermal  noise 
due  to  the  M/W  terminal  which  will  be  added  to  a particular  multiplexed 
telephone  circuit.  Of  course,  after  the  noise  at  the  various  baseband 
locations  (slot  noise)  is  demultiplexed,  the  various  slot  noises  will 
appear  in  various  individual  telephone  circuits. 

1.6  This  report  deals  with  the  FM  noise  quieting  curve,  the  various 
factors  which  affect  it,  and  the  parameters  necessary  to  predict  it. 

The  elementary  baseband  signal  characteristics  as  a function  * RSL 
are  also  discussed  briefly.  Due  to  the  wide  range  of  topics  covered, 
this  report  is  limited  primarily  to  formulas  and  results.  Derivations 
have  been  limited  to  those  which  directly  aid  the  understanding  of  a 
concept.  Theoretical  aspects  will  not  be  covered  in  any  depth.  This 
report  will  briefly  cover  the  important  thermal  noise  characteristics 
of  an  FM  radio  and  provide  the  necessary  tools  to  put  that  knowledge 
to  use . 


3 


2.  Basic  Modulation  Theory 


2.1  The  purpose  of  a wideband  microwave  radio  transmission  system  is 

to  transfer  a wide  frequency  baseband  signal.  from  one  location  to  another. 
Invariably,  a sinusoidal  radio  frequency  wave  (RF  sine  wave)  is  modulated 
to  transfer  a baseband  signal  from  one  radio  terminal  to  another.  Ampli- 
tude modulation  is  a simple  modulation  process  to  visualize.  A baseband 
signal  is  merely  transferred  to  radio  frequency.  At  radio  frequency, 
it  may  appear  with  its  normal  low  to  high  frequency  orientation  (higher 
frequency  baseband  signal  producing  a higher  frequency  RF  component 
than  a lower  frequency  baseband  signal).  In  this  case,  an  upper  sideband 
has  been  produced.  If  the  frequency  orientation  of  the  baseband  is 
reversed  at  RF,  then  a lower  sideband  has  been  produced.  Sometimes 
both  sidebands  are  produced  (double  sideband),  sometimes  only  one  (single 
sideband).  A sinewave  (carrier)  is  used  in  the  modulation  process. 
Sometimes  the  carrier  is  eliminated  prior  to  transmission  of  the  RF 
signal  (suppressed  carrier).  A1 though  there  are  several  forms  of  ampli- 
tude modulation,  the  processes  are  similar.  Each  baseband  frequency 
component  produces  one  or  two  discrete  frequency  components  in  the  modu- 
lated RF  signal.  These  components  are  equal  in  amplitude  (except  in 
vestigial  sideband  modulation)  and  are  separated  from  the  carrier  frequency 
by  a frequency  difference  equal  to  the  frequency  of  the  baseband  component. 
The  nature  of  the  modulation  is  such  that  the  frequency  of  the  RF  modulated 
signal  depends  only  on  the  carrier  frequency  and  the  frequency  of  the 
corresponding  baseband  component.  The  amplitude  of  the  RF  spectral 
components  depend  only  on  the  amplitude  of  the  corresponding  baseband 
spectral  component's  amplitude.  The  frequency  of  the  modulated  signal's 
spectral  components  are  independent  of  baseband  signal  amplitude  and 
vice  versa. 

2.2  Instead  of  modulating  the  amplitude  of  the  transmitted  signal, 
it  is  possible  to  transmit  baseband  signal  by  changing  the  angle  of 
the  transmitted  sinewave  relative  to  its  unmodulated  condition.  The 
properties  of  this  angle  modulation  are  considerably  different  than 
amplitude  modulation.  Two  common  methods  of  achieving  angle  modulation 
are  called  phase  modulation  (PM)  and  frequency  modulation  (FM).  The 
United  States  Electronic  Industry  Association  (EIA)  Standard  RS-252- 

A defines  frequency  modulation  (para  2.1)  as  "...  that  process  of  angle 
modulation  in  which  the  instantaneous  frequency  deviation  of  the  sinusoidal 
carrier  is  proportional  to  the  instantaneous  voltage  of  the  modulating 
signal."  The  standard  defines  phase  modulation  (para  2.2)  as  "...that 
process  of  angle  modulation  in  which  the  instantaneous  phase  deviation 
of  the  sinusoidal  carrier  is  proportional  to  the  instantaneous  voltage 
of  the  modulating  signal."  It  should  be  mentioned  that  the  phase  and 
frequency  of  a sine  wave  are  related  to  each  other.  As  Van  der  Pol 
has  pointed  out,  the  definitions  of  phase  and  frequency  are  not  unique. 
Using  Stumper's  "zero  crossing"  definition  of  frequency  has  analytical 
and  heuristic  advantages  and  leads  to  the  frequency  counter  FM  demodulators 
such  as  those  described  by  Latin.  The  most  common  definitions,  however, 
imply  that  phase  is  the  time  integral  of  frequency  and  that  frequency 


4 


is  the  time  derivative  of  phase.  When  a carrier  is  phase  modulated, 
the  frequency  of  the  carrier  is  also  modulated.  When  we  specify  phase 
modulation,  we  are  defining  a specific  relationship  between  phase  changes 
and  the  modulating  waveform.  The  frequency  will  also  change,  but  the 
relationship  between  frequency  change  and  the  modulation  waveform  is 
not  explicitly  stated.  Likewise,  when  frequency  modulation  is  used, 
the  phase  of  the  carrier  is  also  modulated.  All  that  has  been  directly 
defined  is  a particular  relationship  between  the  carrier  frequency  and 
the  modulating  waveform.  Using  integral  and  differential  calculus 
relationships,  the  definition  of  an  equivalent  PM  (or  FM)  could  be 
derived  for  each  of  the  above  definitions  of  FM  (or  PM).  The  advantage 
of  the  proceeding  definitions,  however,  is  their  simplicity. 

2.3  There  are  several  forms  of  practical  angle  modulators  and  demodulators. 
Most  frequency  modulators  are  voltage  controlled  oscillators  (diode 
reactance,  reactance  tube,  or  klystron  type)  or  zero  crossing  type 
(saturable  reactor).  The  diode  reactance  and  klystron  type  are  the 

most  popular  for  wideband  microwave  use.  Phase  modulators  are  the  pulse 
position  type.  A popular  wideband  modulator  of  this  type  is  the  serrasoid. 
There  are  no  practical  wideband  phase  demodulators.  Types  of  frequency 
demodulators  include  the  ratio,  Foster-Seeley,  and  Travis  detectors, 
line  discriminators,  cycle  counters,  and  various  forms  of  phase  locked 
loops  and  frequency  feedback  demodulation  methods.  The  Travis  discriminator 
is  the  most  popular  type  in  current  high  performance  wideband  terrestrial 
communication  systems. 

2.4  Unlike  amplitude  modulation,  the  spectrum  of  the  modulated  signal 
(as  viewed  on  a spectrum  analyzer,  for  example)  is  a complex  function 
of  both  the  frequency  and  amplitude  of  the  baseband  signal.  The  next 
page  shows  the  transmit  spectrum  of  an  angle  modulator  for  a sine  wave 
baseband  signal  of  different  amplitudes.  The  only  difference  between 
the  different  pictures  is  the  level  of  the  modulating  sinewave,  yet 
the  spectruma  are  different.  When  a single  sinewave  is  used  to  angle 
modulate  a carrier,  a definite  relationship  exists  (as  defined  by  Bessel 
functions  of  the  first  kind  of  integer  order)  between  the  modulation 
index  (beta)  of  the  modulator  and  the  amplitudes  of  the  various  sideband 
components.  FM  is  a complex  modulation  process.  It  is  not  easy  to 
determine  frou;  the  signal  Into  the  modulator  what  the  spectrum  of  the 
output  will  be.  In  the  previously  mentioned  pictures,  the  modulated 

RF  spectrum  changed  considerably,  although  the  input  signal  did  not 
change  its  waveform.  Certain  levels  of  sine  wave  modulation  cause 
disappearance  of  certain  RF  modulated  signal  spectral  components. 

Examples  of  this  are  given  on  the  next  two  pages.  The  rolati  .nship 
between  modulator  input  sine  wave  level  and  spectral  component  disappearance 
(drop  out)  will  be  used  later.  The  voltage  and  power  (dB)  amplitude 
of  the  spectral  components  for  a sinewave  modulated  carrier  have  been 
tabulated  on  the  next  two  pages.  For  comparison,  the  various  spectral 
components  for  a square  wave  frequency  modulated  carrier  are  shown  on 
the  following  two  pages.  When  a single  sinewave  is  the  baseband  signal 


c 


6 


First 

Carrier 

Dropout 


f 


Second 

Carrier 

Dropout 


Eighth 

Carrier 

Dropout 


FM  RF  Spectrum 
Sinewave  Modulation 


Figure  2 


btTA 

CARRIER 

SIDEBAND 

SIDEBAND 

SIDEBAND 

SIDEBAND 

SIDEBAND 

I 

2 

3 

4 

5 

0. 

1 . 0000 

0. 

0. 

0. 

0. 

0. 

0 .25 

0.9844 

0.1240 

0.0078 

0.0003 

0.0000 

0.0000 

O.bO 

0.9305 

0.2423 

0.0306 

0.0026 

0.0002 

0.0000 

0.75 

0.0642 

0.3492 

0.0671 

0.0005 

0.0008 

0.0001 

1 .00 

0, 7db2 

0.4401 

0.1149 

0.0196 

0.0025 

0.0002 

1 ,2b 

0.64b9 

0.5106 

0.1711 

0.0369 

0.0059 

0.0007 

i .*y) 

O.bllfl 

0.5579 

0.2321 

0.0610 

0.0118 

0.0018 

1 . lb 

0.3690 

0.6002 

0.2940 

0.0919 

0.0209 

0.0038 

2.00 

0.2239 

0.5767 

0.3520 

0.1289 

0.0340 

0.0070 

2.2b 

0.0827 

0.5404 

0.4047 

0.1711 

0.0515 

•0.0121 

2,b  0 

-0.0404 

0.4971 

0.4461 

0.2166 

0.0738 

0.0195 

2.  lb 

-0.1641 

0.4260 

0.4739 

0.2634 

0.1007 

0.0297 

3.00 

-0.2601 

r .3391 

0.4061 

0.3091 

0.1320 

0.0430 

-0. 3320 

0.241 1 

0.4011 

0.3510 

0.1669 

0.0599 

J.bO 

-0.3001 

0.1374 

0.4586 

0.3868 

0.2044 

0.0804 

S.  lb 

-0.4014 

0.0332 

0.4191 

0.4138 

0.2430 

0.1046 

4 ,00 

-0.3971 

-0.0660 

0.3641 

0.4302 

0.2811 

0.1321 

4,‘/b 

-0.3692 

-0 . 1 556 

0.2960 

0.4341 

0.3169 

0.1624 

4 ,*»0 

-0.320b 

-0.2311 

0.2178 

0.4247 

0.3484 

0.1947 

4.  /v 

-0. 2bb 1 

-0.2092 

0.1  334 

0.4015 

0.3738 

0.2280 

b.00 

-o.  i //  6 

-0.32/6 

0.0466 

0.3648 

0.3912 

0.2611 

b • 2b 

-0.OV31 

-0, 34'jO 

-0,0304 

0.3158 

0.3993 

0.2926 

b.bO 

-0.0060 

-0.3414 

-0.1173 

0.2561 

0.3967 

0.3209 

b . /b 

0,0/60 

-0.31  /</ 

-0.1066 

0.1082 

0.3829 

0.3446 

0.00 

0.1 b06 

-0.2/6/ 

-0,2429 

0.1146 

0.3576 

0.3621 

0.2b 

0.2131 

-0.2207 

-0.2037 

0.0391 

0.3213 

0.3721 

6.  bO 

0.26/»  1 

-0.1 b3H 

-0.3074 

-0.0353 

0.2748 

0.3736 

0.  /V 

0.209b 

-0.0003 

-0.3133 

-0.1053 

0.2196 

0.3656 

/.00 

0,3/101 

-0.004/ 

-0.3014 

-0,1676 

0.1578 

0.3479 

7.2b 

0.2V20 

0.0606 

-0.2731 

-0.2192 

0.0916 

0.3204 

/ , bO 

0.2663 

0 . 1 3b 2 

-0.2303 

-0.2581 

0.0238 

0.2035 

/.  lb 

0. 22b2 

0.1916 

-0.1750 

-0,2023 

-0.0428 

0.2382 

0.00 

0.1/1/ 

0.2346 

-0.  1130 

-0,29! 1 

-0.1054 

0.1850 

II . 2b 

0. 1092 

0.2622 

-0,0456 

-0.2043 

-0. 161 1 

0.1281 

B.bO 

0./14I9 

0.2731 

0.0223 

-0.2626 

-0.2077 

0.0671 

0.  /b 

-0.02b > 

0.26/2 

0.0070 

-0.2274 

-0.2430 

0.0053 

v.oo 

-0,/)V/)3 

0.24b 3 

0.1440 

-0.1009 

-0.2655 

-0.0550 

V,  2b 

-0.1474 

0.2091 

0.1926 

-0.1258 

-0.2743 

-0.1114 

V.bO 

-0, 1939 

0.1613 

0.2279 

-0.0653 

-0.2691 

-0,1613 

V.  /b 

-0.2273 

0. 1048 

0.2480 

-0.0020 

-0.2506 

-0.2028 

10,00 

-0.24bV 

0.043b 

0.2546 

0.0584 

-0.2196 

-0.2341 

10.2b 

-0.2490 

-0.0190 

0.2453 

0.  1147 

-0.1781 

-0.2537 

io.  bo 

-0.2366 

-0.0789 

0,2216 

0.1633 

-0.1283 

-0,26 1 1 

1 0 • 7 b 

-0.2101 

-0.1325 

0.1854 

0.2015 

-0.0730 

—0,2550 

ii.oo 

-0,1/12 

-0.1/6H 

0.1390 

0.2274 

-0.0150 

-0,2383 

j j ,2b 

-0.1226 

-0.2093 

O.OH54 

0.2397 

0.0424 

-0.2096 

1 1 . bO 

-0.0676 

-0,2284 

0.0279 

0.2301 

0.0963 

-0. 1711 

1 1 • 75 

-0.0096 

-0.2331 

-0.0300 

0.2229 

0 # 1 438 

-0,1250 

1 2. 00 

0,0477 

•0.2235 

-0.0849 

0.1951 

0.1025 

-0.0735 

12.2b 

0. 1010 

-0.2004 

-0.1337 

0.1567 

0.2104 

^7.0193 

FREQUENCY  MODULATED  RF  SPECTRUM/SINE  RAVE  MODULATION  INPUT 


(levels  «•  voltage  radios) 
Table  1 


9 


btTA 

CARRIER 

SIDEBAND 

SIDEBAND 

SIDEBAND 

SIDEBAND 

SIDEBAND 

1 

2 

3 

4 

5 

0. 

0. 

-200.00 

-200.00 

-200. 00 

-200.00 

-200.00 

0.2b 

-0.14 

-18.13 

-42.19 

-69.78 

-99.88 

-131  .92 

0-*0 

-0.55 

-12.3! 

-30.28 

-51 .82 

-75.88 

-101 .88 

J.  /5 

-1 .27 

-9.14 

-23.47 

-41.43 

-61 .93 

-84.38 

1 .00 

-2.32 

-7.13 

-18.79 

-34.17 

-52.12 

-72.05 

1 . 2d 

-3.80 

-5.34 

-15.34 

-28.6? 

-44.62 

-62.56 

1 ,bO 

-b  .82 

-5.07 

-12.69 

-24.30 

-38.59 

-54.90 

1 . /d 

-6.66 

-4.73 

-10.63 

-20.74 

-33.59 

-48.50 

2.00 

-13.00 

-4.78 

-9.05 

-17.79 

-29.37 

-43.05 

2.2b 

-21.64 

-5.22 

-7.86 

-15.34 

-25.76 

-38.33 

2.50. 

-26.31 

-o.07 

-7.0! 

-13.29 

-22.64 

-34.20 

2.  Yd 

-15.70 

-7.41 

-6.49 

-II  .59 

-19.94 

-30  .,56 

3. CO 

-11.70 

-V.39 

-6.27 

-10.20 

-17.59 

' -27.32 

j.2b 

-9.56 

-12.36 

-6.35 

-9.09 

‘-15.55 

-24.45 

3.50 

-8.40 

-17.24 

-6 . 77 

-8.25 

-13.79 

-21 .89 

3.  Yd 

-7.93 

-29.57 

-7.55 

-7.66 

-12.29 

-19.61 

4.00 

-8.02 

-23.60 

-6.77 

-7.33 

-11 .02 

-17.58 

4. 2d 

-8.65 

-16.16 

-1^.57 

-7.25 

-9.98 

-15.79 

4.50 

-9.bB 

-12.73 

-13.24 

-7.44 

-9.16 

-14.21 

4.  Yd 

-II .d 7 

-10.78 

-17.50 

-7.93 

-8.55 

-12.64 

b.00 

-15.01 

-9.69 

-26.64 

-8.76 

-8.15 

- 1 1 . 66 

b.  2b 

-20.62 

-9.24 

-28.32 

- i 0 . 0 1 

-7.97 

-10.67 

b.bO 

-43.29 

-9.33 

-18.61 

-11.83 

-8.03 

-9.87 

b.  Yd 

-22.39 

-9.05 

-14.58 

-14.51 

-8.34 

-9.25 

o . 00 

-16.44 

- 1 1 . 1 0 

-12.29 

-18.80 

-8.93 

-8.82 

0.2b 

-13.43 

-13.12 

-10.94 

-28.15 

-9.86 

-8.59 

6.50 

-11.70 

-16.26 

-10.25 

-29.03 

- 1 1 . 22 

-8.55 

6.  Yd 

-10. Y 7 

-21.90 

-IO.OR 

-19.55 

-13.17 

-8.74 

/.  (X) 

-10.40 

-46.59 

-10.42 

-15.52 

-16.04 

-9.17 

/ . 2d 

-10.69 

-23,28 

-II  .27 

-13.18 

-20.76 

-9.89 

/.  50 

-11,49 

-1 7.38 

-12.76 

-11.77 

-32.46 

-10.95 

Y.  /D 

-12.95 

-14.35 

-15.10 

-10.98 

-27.37 

-12.46 

b.00 

-15.31 

-12.59 

-18. 94 

-10.72 

-19.55 

-14.62 

6. 2d 

-1 9.23 

-1 1 .03 

-26.8! 

-10.92 

-15.86 

-17.85 

b.50 

-27.55 

-II  .27 

-33.0? 

-11.61 

-13.65 

-23.46 

a.  Yd 

-31 .72 

-11.46 

-21 .21 

-12.86 

-12.29 

-45.55 

9.00 

-20.88 

-12.21 

-16.78 

-14.85 

-11.52 

-25.19 

y.25 

-16.63 

-13.59 

-14.31 

-18.00 

-11 .24 

-19.07 

9.50 

-14.25 

-15.85 

-12.85 

-23.70 

-II .40 

-15.85 

9.  Yd 

-I2.b7 

-19.59 

-12.08 

-51.21 

-12.02 

-13.86 

10.00 

-12.18 

-27.24 

- 1 1 . 88 

-24.68 

-13.17 

-12.61 

! 0.2b 

-12.0b 

-34.42 

-12.21 

-18.8! 

-14.99 

-II  .91 

10.50 

-12.52 

-22.06 

-13.09 

-15.74 

-17.83 

-11  .67 

10.  Yb 

-13.55 

-1 /.50 

-14.64 

-13.92 

-22.74 

-II  .84 

1 1 .00 

-15.33 

-15. 05 

-17.14 

-12.87 

-36.46 

-12.46 

1 1 .25 

-IH. 23 

-13.58 

-21 .37 

-12.41 

-27.45 

-13.57 

11.50 

-23,40 

-12.83 

-31 .08 

-12.47 

-20.33 

-15.33 

11.75 

-40.34 

-12.65 

-30.45 

-13.04 

-16.84 

-18.06 

12.00 

-26.42 

-13.02 

-2! .42 

-14.19 

-14.77 

-22.68 

12.25 

-19.92 

-13.96 

-17.48 

-16.10 

-13.54 

-34.29 

FREQUENCY  M0UULA1EI)  HF  SPECTRUM/SINE  WAVE  MODULATION  INPUT 

(levels  in  dB) 

Table  2 


10 


BETA 

CARRIER 

SIDEBAND 

SIDEBAND 

SIDEBAND 

SIDEBAND 

SIDEBAND 

1 

2 

3 

4 

5 

0. 

1 .0000 

0. 

0. 

0. 

0. 

0. 

0.25 

0. 9745 

0. 1 f 68 

0.0155 

-0.0165 

-0.0038 

0.0069 

0.50 

0.9003 

0.3001 

0.0600 

-0.0257 

-0.0743 

0.0091 

0.75 

0.7842 

0.4176 

. 0.1283 

-0.0217 

-0.0286 

0.0075 

1.00 

0.6366 

0.5000 

0.2122 

0,0000 

-0.0424 

-0.0000 

1.25 

0.4705 

0.5414 

0.3016 

0.6409 

-0.0509 

-0.0130 

1.50 

0.3001 

0.5402 

0.3858 

O'.  1000 

-0.0491 

-0.0297 

1.75 

0. 1392 

0.4990 

0.4548 

0.1734 

-0.0330 

-0.0469 

2.00 

0.0000 

0.4244 

0.5000 

0.2546 

0.0000 

-0.0606 

2.25 

-0. 1083 

0.3258 

0.5159 

0.3361 

0.0501 

-0.0664 

2.50 

-0.1801 

0.2144 

0.5002 

0.4092 

0. 1154 

-0.0600 

2.75 

-0.2139 

0.1021 

0.4540 

0.4661 

0.1917 

-0.0384 

3.00 

-0.2122 

0.0000 

0.3820 

0.5000 

0.2728 

0.0000 

3.25 

-0. 1810 

-0.0828 

0.291  3 

0.5067 

0.3515 

0.0548 

3.50 

-0.1286 

-0.1400 

0.1910 

0.4848 

0.4201 

0.1236 

3.75 

-0.0650 

-0.1688 

0.0908 

0.4357 

0.4715 

0.2017 

4.00 

-0.0000 

-0.1698 

0.0000 

0.3638 

0.5000 

0.2829 

4,25 

0.0573 

-0. 1 465 

-0.0736 

0.2756 

0.5020 

0.3603 

4.50 

0.  1 000 

-0.1052 

-0.1247 

0.1801 

0.4766 

0.4265 

4.75 

0.1238 

-0.0537 

-0.1505 

0.0853 

0.4257 

0.4748 

5.00 

0.1273 

-0.0000 

-0.15*6 

0.0000 

0.3537 

0.5000 

5.25 

0.1120 

0.0482 

-0.1310 

-0.0689 

0.2671 

0.4991 

5.50 

0.0818 

0.0846 

-0.0943 

-0.  1165 

0.1737 

0.4716 

5.75 

0.0424 

0.1055 

-0.0482 

-0.1405 

0.0821 

0.4195 

6.00 

0.0000 

0.1091 

-0.0000 

-0.1415 

0.0000 

0.3472 

6.25 

-0.0390 

0.0966 

0.0434 

-0. 1223 

-0.6660 

0.2614 

0.50 

-0.0693 

0.0709 

0."765 

-0.0880 

-0.  1 115 

0.1696 

6.75 

-0.0871 

0.0369 

0.0955 

-0.0450 

-0.1343 

0.0800 

7.00 

-0.0909 

0.0000 

0 ,0  99n 

-n ,0000 

-0. 1 350 

0.0000 

7.25 

-0.0811 

-0.0343 

0.0878 

0.0405 

-0.1166 

-0.0641 

7.50 

-0.0600 

-0.06  1 1 

0 .0646 

0.0715 

-0.0839 

-0.1080 

7.75 

-0.0314 

-0.0  112 

0.0337 

0.0893 

-0.0429 

-0 . 1 300 

8.00 

-0.0000 

-0.0808 

0.0000 

0.0926 

-0.0000 

-0.1306 

8.25 

0.0295 

-0.0724 

-0.031 4 

0.0822 

0.0386 

-0.  1127 

8.50 

0.0530 

-0.0537 

-0.0561 

0.0605 

0.0680 

-0.0810 

8.75 

0.0672 

-0.0282 

-0.0709 

0.0316 

0.0850 

-0.0413 

9.00 

0.0707 

-0.0000 

-0.0744 

0.0000 

0.0881 

-0.0000 

9.25 

0.0636 

0.0266 

-0.0667 

-0.0294 

0.0782 

0.0372 

9.50 

0.0474 

0.0479 

-0.0496 

-0.0526 

0.0576 

0.0655 

9,75 

0.0250 

0.0610 

-0.0261 

-0.0666 

0.0300 

0.0818 

10.00 

0.0000 

0.0643 

-0.0000 

-0.0700 

0.0000 

0.0849 

10.25 

-0.0238 

0.0579 

0.0247 

-0.0628 

-0.0280 

0.0753 

10.50 

-0.0429 

0.0433 

0.0445 

-0.0467 

-0.0502 

0.0554 

10.75 

-0.0547 

0.0229 

0.0567 

-0.0246 

-0.0635 

0.0289 

1 1 .00 

-0.0579 

0.0000 

0.0599 

-0.0000 

-0.0667 

0.0000 

1 1 .25 

-0.0523 

-0.0218 

0.0540 

0.0233 

-0.0598 

-0.0270 

1 1 .50 

-0.0391 

-0.0394 

0.0404 

0.0420 

-0.0445 

-0.0483 

1 1 .75 

-0.0207 

-0.0504 

0.021 4 

0.0535 

-0.0235 

-0,061 1 

12.00 

-0.0000 

-0.0534 

0.0000 

0.0566 

-0.0000 

-0.Q642 

12.25 

0.01 99 

-0.0483 

-0.0204 

0.051  1 

0.0223 

-0.0576 

FREQUENCY  MODULATED  RF  SPECTRUM/SQUARE  WAVE  MODULATION  INPUT 

(levels  as  voltage  ratios) 

Table  3 


11 


BETA 

CARRIER 

SIDEBAND 

SIDEBAND 

SIDEBAND 

SIDEBAND 

SIDEBAND 

1 

2 

3 

4 

5 

0. 

0. 

-200.00 

-200.00 

-200.00 

-200.00 

-200.00 

0.2b 

-0.22 

-16.09 

-36.21 

-35.68 

-48.36 

-44.59 

O.bO 

-0.9'. 

-10.45 

-24.43 

-3! .79 

-36.90 

-40.82 

0.7b 

-2.11 

-7.58 

-17.83 

-33.29 

-30.88 

-42.53 

1 .00 

-3.92 

-6.02 

-13.46 

-166.19 

-27.44 

-169.72 

1 .25 

-6.55 

-5.33 

-10.41 

-27.76 

-25.86 

-37.73 

I .bO 

-10.45 

-5.35 

-8.27 

-20.00 

-26.18 

-30.55 

1 .75 

-17.13 

-6.04 

-6.84 

-15.22 

-29. 6« 

-26.57 

2.00 

-154.15 

-7.44 

-6.02 

-11.88 

-163.70 

-24.35 

2.25 

-19.31 

-9.74 

-5.75 

-9.47 

-26.00 

-23.56 

2.50 

-14.89 

-13.38 

-6.02 

-7.76 

-18.75 

i24.43 

2.75 

-13.40 

-19.82 

-6.86 

-6.63 

-14.35 

-28.31 

3.00 

-13.46 

-1 56.65 

-8.36 

-6.02 

-11.28 

-162.67 

3.  2d 

-14.85 

-21 .64 

' -10.71 

-5.90 

-9,08 

-25.22 

3.50 

-17.81 

-17.07 

-14.38 

-6.29 

-7.53 

-18.16 

3.7b 

-23.75 

-15.45 

-20.84 

-7.22 

-6.53 

-13.91 

4.00 

-154.15 

-15.40 

-157.67 

-8.78 

-6.02 

-10.97 

4.2b 

-24.83 

- 10.68 

-22.66 

-11.19 

-5.99 

-8.87 

4.50 

-20.00 

-1 9.56 

-18.09 

-14.89 

-6.44 

-7.40 

4.7b 

-18.14 

-25.41 

-16.45 

-21.38 

-7.42 

-6.47 

5.00 

-17. 90 

-1 55. 74 

-10.39 

-158.24 

-9.03 

-6.02 

5.25 

-19.01 

-26.35 

-17.65 

-23.24 

-II .47 

-6.04 

b .50 

-21  . / 4 

-21.45 

-20.51 

-18.67 

-15.20 

-6.53 

5.7d 

-27.46 

- 1 y.'_>4 

-26.34 

-17.04 

-21 .71 

-7.55 

6.00 

-152.85 

-19.24 

-156.65 

-10.99 

-158.59 

-9.19 

6.25 

-28.18 

-20.30 

-27.24 

-18.25 

-23.61 

-1 1 .65 

6.50 

-23.19 

-22.98 

-22.33 

-21.H 

-19.06 

-15.41 

6.7b 

-21 .20 

-28.06 

-20.40 

-20.94 

-17.44 

-21.94 

1.00 

-20.82 

-1 54.01 

-20.08 

-157.25 

-17.39 

-158.83 

7.2b 

-21 .82 

-29.31 

-21.13 

-27.84 

-18.66 

-23.87 

7.50 

-24.43 

-24.28 

-23.79 

-22.92 

-21  .53 

-19.33 

7.75 

-30.05 

-22.25 

-29.45 

-20.99 

-27.36 

-17.72 

8.00 

-154.15 

-21 .85 

-154.79' 

-20.67 

-157.67 

-17.68 

8. 2d 

-30.59 

-22.81 

-30  XI 

-21.71 

-28,27 

-i8.96 

8.50 

-25.52 

-25,40 

-25.03 

-24.37 

-23.35 

-21 .83 

8.7d 

-23.45 

-30.99 

-22.98 

-30.02 

-21 .41 

-27.67 

9.00 

-23.01 

-1 55.0/ 

-22.57 

-1 55.35 

-21.10 

-157.99 

9.25 

-23. 9j 

-31.49 

-23.52 

-30 ,o2 

-22.1 3 

•:28.59 

9.50 

-26.49 

-26.39 

-26.09 

-25.57 

-24.79 

-23.67 

9.75 

-32.05 

-24.30 

-31 .67 

-23.53 

-30.44 

-21.74 

1 0.00 

-155.04 

-23.84 

-155.74 

-23.10 

-155.78 

-21 .42 

10.25 

-32.48 

-24.74 

-32.14 

-24.05 

-31 .05 

-22.46 

10.50 

-27.36 

-27.28 

-27.04 

-26.62 

-25.99 

-25.12 

10.75 

-25.24 

-32.82 

-24.93 

-32.19 

-23.94 

-30.78 

1 1 .00 

-24.75 

-155.79 

-24.46 

-156.25 

-23.52 

-156.11 

1 1 .25 

-25.63 

-33.22 

-25.35 

-32.65 

-24.46 

-31 .38 

1 I .50 

-28.15 

-28.08 

-27.88 

-27.53 

-27.03 

-26.33 

1 1 .75 

-33.67 

-25,95 

-33.41 

-25.43 

-32.60 

-24.28 

12.00 

-152.85 

-25.45 

-156.38 

-24.95 

-156.65 

-23.85 

12.25 

-34.03 

-26.31 

-33.79 

-25.84 

-33.05 

-24.79 

FRfcOUENCY  MODULATED  RF  SPECTRUM/SQUARE  HAVE  MODULATION  INPUT 

(levels  in  dB) 


Table  4 


12 


to  a modulator,  the  frequency  separation  between  the  various  spectral 
components  of  the  modulated  signal  is  always  equal  to  the  sinewave 
frequency.  When  two  or  more  sinewave  baseband  signals  are  used  simultan- 
eously, the  frequencies  of  the  sideband  components  are  separated  from 
the  carrier  by  all  possible  combinations  of  frequencies  which  oan  be 
obtained  from  sum  and  differences  of  all  harmonics  of  the  modulating 
slnewaves.  For  many  possible  wideband  baseband  signals,  an  exact  deter- 
mination of  the  transmit  signal  spectrum  is  not  practical.  About  the 
only  practical  method  of  analysis  of  the  baseband/transrait  signal  relation- 
ship is  on  a statistical  basis.  There  is,  however,  a general  relationship 
between  the  modulating  signal  and  the  modulated  waveform.  This  generalization 
is  mentioned  by  both  Middleton  and  Giacoletto.  Middleton  calls  it  the 
"Principle  of  Adiabatic  Frequency  Sweeps".  The  principle  simply  states 
that  for  any  general  modulation  waveform  producing  frequency  modulation, 
the  transmitted  radio  frequency  (RF)  spectrum  will  be  symmetric  about 
the  unmodulated  carrier  frequency  if  the  modulating  signal  has  voltages 
which,  when  viewed  in  time,  are  symmetric  about  zero  volts.  Conversely, 
unsymmetrical  modulating  signals  produce  unsymmetrical  FM  modulated 
RF  spectrums.  This  is  a useful  principle.  It  indicates  that  if  an 
FM  modulator  is  driven  by  a sinewave,  the  RF  spectrum  will  be  symmetric. 

If  it  is  not,  the  modulated  wave  has  unquestionably  been  distorted. 

The  FM  receiver  requires  a symmetric  received  RF  spectrum  to  produce 
a symmetric  baseband  signal  like  a sinewave.  It  receives  an  unsymmetric 
RF  spectrum,  it  can  only  produce  an  unsymmetric  voltage  waveform  at  the 
baseband  demodulator  output.  If  a sinewave  (a  symmetric  waveform)  is 
transmitted  and  an  unsymmetric  waveform  is  received,  the  baseband  signal 
has  been  distorted. 

2.5  There  are  several  ways  that  RF  spectrum  can  be  made  unsymmetric. 

At  high  baseband  modulator  input  levels,  nonlinearity  of  the  baseband 
amplifiers  and  the  modulator  itself  are  distinct  possibilities.  However, 
even  at  low  drive  levels,  it  is  possible  to  make  the  RF  spectrum  unsymmetric. 
As  Middleton  and  Panter  have  mentioned,  simultaneous  (correlated)  amplitude 
modulation  and  frequency  modulation  invariably  produce  an  unsymmetric 

RF  spectrum.  It  is  possible  to  have  this  take  place  right  at  the  FM 
modulator.  Most  FM  modulators  are  immediately  followed  by  hard  limiters 
to  avoid  this  possibility.  If  the  RF  circuitry  has  nonuniform  amplitude 
frequency  response,  or  a uniform  but  unsymmetric  (about  the  carrier 
frequency)  frequency  response,  correlated  AM  and  PM  will  be  introduced 
into  the  signal.  In  severe  cases,  either  of  these  can  cause  distortion 
of  the  normal  sideband  amplitudes  and,  therefore,  distortion  in  the 
demodulated  signal. 

2.6  The  modulation  index  of  an  angle  modulated  signal  is  a factor  which 
relates  the  modulating  waveform  to  the  phase  deviation  of  the  modulated 
signal.  This  factor  is  called  beta  (B)  and  has  the  units  radians  per 
millivolt.  For  most  baseband  signals,  a beta  is  difficult  or  impossible 
to  define.  However,  for  a sinewave  baseband  input,  tne  determination 

of  beta  is  straightforward.  For  a phase  modulation  system,  the  beta 
is  just  the  specified  phase  sensitivity  of  the  modulator.  For  a frequency 


13 


modulation  system,  beta  is  the  specified  frequency  deviation  sensitivity 
of  the  modulator  (kilohertz  per  millivolt)  divided  by  the  frequency 
(kilohertz)  of  the  modulating  sinewave.  The  division  of  frequency  deviation 
by  frequency  is  a direct  result  of  having  to  integrate  the  sinusoidal 
frequency  deviation  to  obtain  equivalent  phase  deviation.  As  mentioned 
earlier,  knowing  the  sinewave  beta  of  the  angle  modulation  transmitter 
allows  us  to  predict  exactly  (using  Bessel  functions)  the  RF  transmit 
spectrum  for  sinewave  modulation. 

2.7  Like  amplitude  modulation,  there  are  many  possible  variations  of 
angle  modulation.  As  mentioned  previously,  however,  there  are  two  basic 
angle  modulation  types.  With  amplitude  modulation,  it  is  possible, 
with  suitable  processing  of  the  signals,  to  turn  one  type  of  modulation 
into  another  (for  example,  suppress  the  carrier,  filter  one  sideband, 
etc.).  As  with  amplitude  modulation,  (see,  for  example,  Black  and  Taub 
and  Schilling.)  with  suitable  processing  a phase  modulation  system  can 
be  converted  into  a frequency  modulation  system  and  vice  versa.  Specifically, 
if  the  baseband  input  of  a phase  modulator  is  preceeded  by  a time  integrating 
circuit,  the  combination  will  have  all  the  properties  of  a frequency 
modulator.  Likewise,  if  the  baseband  output  of  a phase  demodulator 
is  coupled  to  a time  differentiation  circuit,  the  combination  will  have 
all  the  properties  of  a frequency  demodulator.  A similar  procedure 
holds  for  frequency  modulation.  A frequency  modulator  with  baseband 
input  preceeded  by  a time  differentiation  circuit  has  exactly  the  same 
properties  as  a phase  modulator.  A frequency  demodulator  whose  baseband 
output  is  followed  by  a time  integration  network  has  exactly  the  same 
properties  as  a phase  demodulator.  Therefore,  it  is  theoretically 
possible  to  mix  and  match  various  modulators  and  demodulators  to  achieve 
a desired  microwave  radio  system.  These  properties  are  not  theoretical 
abstractions.  The  Radio  Equipment  Laboratories  (REL)  military  tropospheric 
scatter  (TROPO)  microwave  radio  system  AN/FRC-39(V)  is  basically  a phase 
modulation,  frequency  demodulation  system.  The  baseband  circuitry  of 
the  modulator  and  demodulator  is  such  that  overall  the  modulator  and 
demodulator  are  completely  compatible.  In  fact,  the  serrisoid  phase 
modulator  u3ed  in  the  AN/FRC-39(V) , when  preceeded  by  the  baseband 
corrector  networx,  produoes  a transmit  RF  signal  which  is  completely 
indistinguishable  from  a comparable  FM  modulator  preceeding  by  the 
appropriate  pre-emphasis  network. 

2.8  Frequency  and  phase  demodulators  have  different  noise  properties. 

A strong  unmodulated  carrier  centered  in  the  symetrir  (if)  passband 
of  a phase  demodulator  produces  a flat  (white)  spectrum  of  thermal  noise 
over  the  entire  output  baseband  if  the  only  noise  generated  in  the 
receiver  is  white  thermal  noise.  In  frequency  division  multiplexing 
(FDM)  used  to  produce  an  analog  baseband  to  apply  to  an  angle  modulator, 
different  3.1  kilohertz  wide  telephone  channels  appear  at  different 
3.1  kilohertz  frequency  positions  (slots)  in  the  radio  baseband.  There- 
fore, for  strong  received  signal  levels,  a phase  demodulator  will  introduce 
the  same  noise  in  ail  telephone  slots.  When  the  telephone  channels 
are  demultiplexed,  the  radio  noise  will  be  the  same  for  all  channels. 


14 


With  a frequency  demodulator  under  the  same  conditions,  the  thermal 
noise  voltage  in  the  receive  baseband  increases  linearly  with  frequency 
(hence,  it  is  often  called  triangular  noise).  Since  power  is  proportional 
to  the  voltage  squared,  the  noise  power  increases  with  the  square  of 
the  baseband  (slot)  frequency.  Therefore,  the  noise  power  increases 
at  a rate  of  6 dB  per  octave  or  20  dB  per  decade.  When  there  is  no 
carrier  present  at  the  receiver,  the  noise  spectrum  (as  Crosby  and 
Downing  note,  for  example)  is  essentially  the  same  (flat)  for  all  baseband 
frequencies  (so  called  rectangular  noi3e).  Under  the  same  conditions, 

(as  noted  by  Taub  and  Schilling,  for  example)  the  noise  voltage  in  the 
baseband  of  the  phase  demodulator  decreases  inversely  with  baseband 
slot  frequency.  Therefore,  the  noise  power  is  reduced  6 dB  per  octave 
(20  dB  per  decade).  Due  to  the  flat  noise  spectrum  in  the  baseband 
of  a phase  demodulator  under  strong  received  signal  conditions,  it  is 
a highly  desirable  type  of  angle  modulation  receiver.  High  performance 
wideband  phase  demodulators  are  difficult  to  construct.  At  the  trans- 
mitter, the  problems  are  more  difficult.  Most  phase  modulators  can 
only  produce  low  modulation  index  modulation.  To  obtain  the  necessary 
modulation  index  with  a phase  modulator,  multiplication  of  the  modulator 
output  is  generally  required.  Using  a multiplier  puts  some  severe  require- 
ments on  the  basic  phase  noise  performance  of  the  modulator's  oscillator. 
Every  time  the  oscillator  output  is  doubled,  the  noise  it  produces  at 
the  baseband  of  a receiver  with  a strong  RSL  (receiver  in  saturated 
region)  i3  increased  6 dB.  From  a practical  point  of  view,  it  is  much 
easier  to  use  a voltage  controlled  oscillator  as  the  modulation  source. 

With  such  modulators,  wide  linear  frequency  shifts  directly  proportional 
to  input  voltage  can  be  obtained.  Little  or  no  multiplication  is  required, 
and  the  modulated  signal  can  often  be  produced  directly  at  RF  without 
the  need  of  an  up  converter.  Using  frequency  modulators,  high  performance 
wideband  microwave  systems  are  readily  achievable. 


2.9  If  we  wish  to  use  a phase  demodulator,  we  must  turn  the  frequency 
modulator  into  a phase  modulator  to  be  compatible  with  a phase  demodulator. 

To  turn  a frequency  modulator  into  a phase  modulator,  the  baseband  input 
to  the  modulator  must  be  preceded  by  a differentiation  network.  Taking 
advantage  of  the  elementary  properties  of  Fourier  transforms,  if  a device 
causes  differentiation  in  the  time  domain,  it  will  cause  an  additional 
term  directly  proportional  to  frequency  in  the  voltage  versus  frequency 
response  of  tne  overall  system.  Since  power  is  proportional  to  voltage 
squared,  the  power  frequency  response  of  a baseba  d circuit  with  a different- 
iator in  it  will  have  a frequency  response  which  increases  with  the 
square  of  frequency.  Again,  this  is  6 dB  per  octave.  This  is  not  an 
undesirable  feature  for  the  telephone  channels  in  the  high  end  of  the 
baseband;  their  level  is  boosted  through  the  differentiation  network. 

However,  relative  to  the  high  end  of  the  baseband,  the  telephone  channels 
at  the  low  end  of  the  baseband  will  be  suppressed.  For  a 600  telephone 
channel  radio  baseband  extending  from  60  kilohertz  to  2660  kilohertz, 
for  example,  out  of  the  differentiation  network  the  highest  telephone 


IS 


channel  would  be  33  dB  greater  in  level  than  the  lowest  telephone  channel. 

To  recover  the  baseband  at  the  receive  end,  the  integrator  network  at 
the  frequency  derrodulator  must  boost  the  lowest  channel  33  dB  relative 
to  the  highest  telephone  channel  so  that  uniform  frequency  response 
is  maintained  in  and  out  of  the  radio  terminal.  In  a noiseless  transmission 
system,  attenuating  the  low  end  of  the  baseband  33  dB  and  then  boosting 
it  33  dB  at  the  receive  baseband  would  be  no  problem.  Unfortunately, 
all  transmitters  use  oscillators  which  have  significant  phase  jitter. 

This  Jitter  causes  a roughly  flat  noise  spectrum  to  be  produced  at  the 
baseband  of  a companion  frequency  demodulator  (if  the  received  signal 
level  is  sufficiently  high  that  thermal  noise  is  negligible  compared 
to  the  transmit  phase  noise).  Because  of  the  significant  amount  of 
transmit  phase  noi3e  (Strictly  speaking,  the  noise  is  equivalent  frequency 
noise  produced  by  phase  noise.)  that  will  appear  at  the  baseband  of 
the  FM  receiver,  attenuating  this  level  of  the  low  frequency  telephone 
channels  by  many  dB  is  out  of  the  question.  The  signal  to  noise  ratio 
of  the  low  frequency  channels  will  have  been  excessively  degraded  at 
the  receiver  by  the  transmitters  phase  noise.  Because  of  these  consider- 
ations, actual  differentiation  and  integration  networks  are  never  used 
with  frequency  modulation  modulators  and  demodulators.  Instead,  dev  ?es 
called  pre-emphasis  and  de-emphasis  networks  are  used.  These  networks 
have  essentially  the  frequency  response  of  differentiators  and  integrators 
over  about  the  top  octave  of  the  radio  baseband.  Over  the  rest  of  the 
baseband,  the  networks  have  essentially  constant  frequency  response. 

Frequency  modulation  radio  terminals  with  pre-emphasis  and  de-emphasis 
are  actually  a hybrid  of  frequency  and  phase  modulation  systems.  Over 
about  the  top  octave  of  the  rad.o  baseband,  they  have  noise  characteristics 
of  phase  modulation  systems.  Over  the  rest  of  the  baseband,  however, 
they  have  the  noise  characteristics  of  a frequency  modulation  system. 

During  the  rest  of  this  report,  frequency  modulation  (FM)  systems  will 
be  considered  exclusively.  However,  it  should  be  kept  in  mind  that 
when  FM  «ystems  use  pre-emphasis  and  de-emphasis  networks,  they  achieve 
some  of  the  thermal  noise  properties  of  phase  modulation  systems.  The 
effect  of  de-emphasis  networks  on  thermal  noise  performance  of  FM  receivers 
will  be  included  in  that  report. 

2.10  At  this  point,  the  character  of  baseband  noise  has  only  been  inferred 
for  high  carrier  to  thermal  noise  ratios  and  for  the  case  where  there 
is  no  received  signal.  As  the  carrier  level  of  the  received  microwave 
signal  is  reduced  from  a strong  RSL,  the  noise  power  in  the  baseband 
of  the  receiver  increases.  As  the  carrier  level  at  the  receiver  is 
reduced  1 d3,  the  noise  at  the  baseband  output  of  the  receiver  increases 
1 dB.  The  frequency  distribution  (shape)  of  the  noise  remains  the  same. 

If  viewed  on  an  osciiliscope,  the  noise  would  appear  as  the  usual  relatively 
smooth  looking  background  thermal  noise.  As  the  received  signal  level 
approaches  low  levels,  the  character  of  the  noise  starts  to  change. 

In  addition  to  the  smooth  noise,  noise  spikes  (impulses^  begin  to  appear. 

As  the  received  signal  gets  lower  and  lower,  the  individual  noise  spikes 
or  clicks  occur  more  and  more  often.  With  continued  reduction  in  received 
signal,  the  clicks  rapidly  merge  into  a crackling  or  sputtering  noise. 


16 


Finally,  the  clicks  murge  to  produce  a continuous  high  level  of  noise 
out  of  the  baseband  of  the  receiver.  As  the  noise  impulses  appear  more 
and  more  often,  the  noise  in  the  various  baseband  slots  starts  to  increase 
faster  than  1 dB  for  1 dB  decrease  in  received  signal  level.  The  received 
signal  level  at  which  the  noise  in  a narrow  measurement  slot  in  the 
baseband  increases  2 dB  for  a 1 dB  reduction  in  carrier  level  is  generally 
called  FM  noise  threshold.  Since  the  impulse  noise  is  essentially  flat 
in  power  spectrum,  the  noise  increase  is  first  seen  in  the  low  baseband 
frequencies  where  the  normal  thermal  noise  is  quite  low.  The  difference 
in  FM  threshold  between  the  highest  and  lowest  3lot  in  a wideband  receiver 
can  be  several  dB.  For  standardization,  FM  noise  threshold  is  normally 
measured  in  one  of  the  highest  baseband  slots.  For  those  interested 
in  why  impulses  occur  for  low  received  signal  levels,  Taub  and  Schilling's 
text  is  highly  recommended.  In  addition  to  the  basic  thermal  noise 
properties  of  FM  systems,  there  are  several  other  properties  of  interest. 

2.11  Because  FM  signals  carry  no  information  in  the  amplitude  of  the 
transmitted  signal,  it  is  possible  to  use  a limiter  in  the  receiver 
to  suppress  any  residual  amplitude  modulation  in  the  received  signal 
before  it  is  demodulated.  The  use  of  a limiter  in  an  FM  receiver  can 
introduce  additional  intermodulation  when  receive  signals  have  passed 
through  certain  types  of  amplitude  nonlinearitie3.  Middleton  has  pointed 
out  that  not  using  a limiter,  however,  leaves  the  signal  prone  to  another 
type  of  intermodulation  distortion.  In  general  limiters  are  used  for 
a couple  of  very  good  reasons.  The  received  signal  and  thermal  noise 
performance  of  the  receiver  is  significantly  improved  near  FM  noise 
threshold  when  the  limiter  is  used.  For  large  received  signal  to 
thermal  noise  (large  carrier  to  noise  ratio)  conditions  (i.e.,  well 
above  FM  noise  threshold),  the  use  of  heavy  limiting  gives  the  FM  receiver 
its  ability  to  reject  unwanted  signals  lower  in  power  than  the  desired 
signal.  This  ability  is  often  expressed  as  capture  ratio.  This  ability 
is  very  important  to  high  quality  microwave  communication  since  it  is 
directly  responsible  for  the  ability  of  the  receiver  to  reject  other 
signals  on  its  frequency  (co-channel  interference),  signals  on  frequencies 
slightly  different  than  its  own  (adjacent  channel  interference),  and 
its  own  signal  reflected  from  stray  sources  (multipath).  Since  all 
wideband  microwave  receivers  use  hard  limiters  (heavy  limiting),  many 
authors  have  stated  or  implied  that  FM  receivers  have  an  inherent  ability 
to  reject  other  signals  of  lower  received  power  level.  This  simply 
is  not  true.  Without  hard  limiting  action,  the  FM  receiver  loses  much, 
if  not  all,  of  its  ability  to  reject  unwanted  signals..  Downing 
specifically  mentions  that  the  ability  of  an  FM  receiver  to  reject  impulse 
noise  and  coherent  interference  i3  directly  attributable  to  the  receivers 
amplitude  limiter.  Corrington  mentions  an  interesting  example.  During 
some  propagation  tests  he  was  conducting,  he  observed  two  distant  FM 
stations.  Both  stations  were  on  the  same  frequency  and  of  nearly  equal 
receive  signal  strength.  However,  each  signal  was  randomly  fading  up 
and  down  in  received  signal  level.  His  receiver  was  such  that  it  achieved 
limiting  for  strong  received  signals,  but  did  not  limit  for  weak  received 
signals.  When  both  signals  were  strong,  the  limiter  was  saturated  and 


17 


only  the  stronger  signal's  program  was  heard.  When  the  other  signal 
became  stronger,  there  would  be  a short  burst  of  noise  and  the  program 
would  abruptly  change  to  that  of  the  stronger  signal.  For  many  minutes, 
conditions  were  such  that  the  programs  changed  back  and  forth  about 
once  every  fifteen  seconds.  Occasionally,  both  signal  levels  fell  below 
the  level  at  which  the  receiver  limited.  During  those  periods  of  time, 
both  programs  could  be  heard  simultaneously. 

?.  12  There  are  two  important  considerations  to  insure  hard  limiting. 

Panter  mentions  that  in  order  to  achieve  high  suppression  of  unwanted 
interference,  the  limiter  and  the  stages  preceding  the  first  limiter 
stage  must  have  essentially  flat  frequency  response  over  the  frequency 
range  cf  the  desired  signal.  Ruthroff  mentions  the  other  important 
consideration.  It  is  well  understood  that  in  limiter  circuits,  a minimum 
level  ir.to  the  limiter  is  necessary  to  achieve  limiting.  Contrary  to 
common  belief,  however,  merely  increasing  the  drive  level  does  not 
necessarily  improve  limiting.  Strangely  enough,  overdriving  a practical 
limiter  circuit  can  seriously  degrade  it.  For  circuits  of  finite  frequency 
bandwidth  and  utilizing  limiter  diodes  with  finite  back  resistance, 
there  is  some  unique  input  level  at  which  limiting  is  best.  The  purpose 
of  Automatic  Gain  Control  (AGC)  in  an  FM  receiver  is  to  keep  the  carrier 
input  level  near  the  value  for  best  limiting. 

One  other  factor  in  good  interference  rejection  is  the  bandwidth 
and  symmetry  of  the  FM  demodulator  (discriminator).  As  Panter  mentions, 
for  best  capture  ratio  the  discriminator  must  be  able  to  accommodate 
the  necessary  amplitude  and  frequency  excursions  which  occur  in  an 
interference  situation.  Leentvaar  and  Flint  have  shown  that  it  is 
possible  to  reduce  or  eliminate  the  capture  effect  of  an  FM  reoeiver 
by  reducing  the  bandwidth  of  an  FM  demodulator. 

2.14  Due  to  the  nonlinear  demodulation  action  of  an  FM  receiver,  limiters 
and  the  general  nonlinear  process  of  FM  generation  and  detection,  the 
noise  produced  in  an  FM  receiver  due  to  an  arbitrary  interfering  signal 
is  difficult  to  predict  with  any  degree  of  precision.  The  articles 
listed  in  the  bibliography  are  available  for  information  on  specific 
interference  situations.  A few  generalizations,  however,  are  available. 
Corrington  suggests  some  effects  due  to  a single  unmodulated  interfering 
signal.  Introducing  an  unwanted  carrier  into  an  FM  receiver  increases 
the  overall  noise  demodulated  at  the  baseband  of  the  receiver.  At  a 
low  level  relative  to  the  desired  signal  (high  carrier  to  interference 
ratio)  the  increased  noise  in  the  baseband  of  the  FM  receiver  is  primarily 
at  low  baseband  frequencies.  As  the  interfering  signal  gets  stronger, 
the  noise  at  higher  baseband  frequencies  increases  faster  than  the  noise 
at  lower  frequencies.  When  the  desired  signal  and  the  interference 
are  about  the  same  level  (C/I=  OdB),  the  baseband  noise  is  quite  high 
and  essentially  the  same  at  all  baseband  frequencies. 


18 


2.15  If  both  the  desired  and  the  interfering  signal  are  unmodulated 
or  only  lightly  modulated  (loaded)  FDMed  FM  signals,  a high  ' . jnsity, 
very  narrow  noise  spike  is  produced  in  the  receiver  baseband.  This 
spike  is  a crossmodulation  (beat)  product  caused  by  the  modulation  of 
the  desired  signal  by  the  interference.  The  baseband  frequency  of  the 
noise  spike  is  the  difference  between  the  frequency  of  the  unmodulated 
signal  and  the  desired  signal.  If  the  interfering  signal  or  the  main 
signal  is  highly  loaded  but  the  other  is  lightly  loaded,  noise  is  produced 

in  the  baseband  and  is  concentrated  around  the  baseband  frequency  corresponding 
to  the  difference  between  the  center  (carrier;  frequency  of  the  desired 
and  the  interfering  signals.  If  both  signals  are  highly  modulated, 
broadband  interference  can  be  expected.  Unlike  the  case  when  the  signals 
are  unmodulated  or  only  lightly  modulated,  the  noi3e  appears  as  broadband 
intermodulation  noise. 

2.16  In  addition  to  interference  considerations,  Taub  and  Schilling 
and  Panter  mention  that  if  the  received  signal  is  not  centered  in  the 
center  of  the  IF  of  the  .•eceiver  or  if  the  IF  frequency  response  is  not 
symmetric  with  respect  to  the  normal  center  frequency,  the  noise  out 

of  the  baseband  will  be  greater  than  normal.  Rice,  Sundie,  and  Taub 
and  Schilling  indicate  that  the  FM  threshold  of  an  FM  receiver  will 
be  degraded  due  to  the  production  of  additional  baseband  noise  if  the 
received  signal  is  modulated.  Middleton  indicates  if  an  FM  modulator 
is  wideband  white  noise  modulated  and  the  signal  is  demodulated  by  a 
receiver  with  no  limiter,  the  receive  baseband  will  experience  considerable 
essentially  flat  noise  even  for  strong  received  signal  levels. 

2.17  One  other  noise  characteristic  of  FM  receivers  is  the  noise  out 
of  the  receive.-  at  received  signal  levels  at  FM  threshold  and  less. 

As  Middleton  has  shown,  the  character  of  the  noise  out  of  the  baseband 
of  an  FM  receiver  changes  dramatically  as  the  limiting  in  an  FM  receiver 
is  changed  from  hard  to  soft  to  no  limiting.  If  limiting  action  is 
reduced  or  lost  in  the  region  of  FM  threshold,  the  noise  in  the  baseband 
increases  dramatically.  The  baseband  noise  is  basically  wide  frequency 
range  flat  spectrum  noise  which  can  significantly  degrade  FM  threshold. 

It  is  interesting,  however,  that  with  no  carrier  present  at  the  receiver 
and  no  receiver  limiting,  the  noise  out  of  the  baseband  receiver  is 
invariably  less  than  the  noise  that  would  occur  with  heavy  limiting. 

The  noise  out  of  the  baseband  of  the  radio  with  no  carrier  present  is 
a direct  function  of  the  amount  of  limiting  in  the  receiver.  The  relation 
between  amount  of  limiting  and  baseband  noise  is  highly  nonlinear,  however. 
Several  dB  of  baseband  noise  is  lost  as  limiting  transitions  from  hard 
to  soft. 

2.18  To  this  point,  no  mention  has  been  made  as  to  how  the  baseband 
signal  is  affected  as  the  received  carrier  i3  reduced  into  the  region 
near  FM  threshold.  With  hard  receiver  limiting,  the  baseband  signal 
remains  constant.  However,  as  the  carrier  falls  below  the  FM  threshold 
level,  the  baseband  signal  is  suppressed  rather  abruptly.  Below  FM 


19 


threshold,  the  demodulated  baseband  signal  is  reduced  2 dB  for  every 
dB  of  reduced  signal  level.  As  with  baseband  noise  near  FM  threshold, 
the  limiter  affects  the  suppression  of  the  baseband  signal.  The  best 
performance  is  obtained  with  hard  limiting.  As  limiting  goes  soft, 
the  signal  suppression  starts  earlier.  Middleton  has  shown  that  the 
effect  with  soft  limiting  is  to  ca'  a the  baseband  signal  level  to  start 
being  lost  at  a much  stronger  received  signal  level.  The  baseband  signal 
level  is  lost  more  gradually.  With  soft  limiting,  the  received  baseband 
signal  level  eventually  drops  ofr  at  the  same  2 dB  per  1 dB  of  received 
signal  reduction  as  occurred  for  hard  limiting.  However,  with  soft 
limiting,  a given  amount  of  signal  suppression  will  occur  at  a stronger 
received  signal  level. 


20 


3.  Wideband  FM  Receiver  Baseband  Signal  and  Noise  Characteristics 

i.l  If  the  noise  in  a single  baseband  noise  slot  were  plotted  as  a 
function  of  RF  Received  Signal  Level  (RSL),  the  plot  would  appear  as 
shown  on  the  next  page. 

1.?  Region  A is  the  region  of  the  quieting  curve  where  the  RF  signal 
has  essentially  been  lost  and  the  slot  noise  is  dependent  on  the  thermal 
noise  generated  within  the  front  end  of  the  receiver,  the  receiver  IF 
response,  and  the  amount  of  receiver  limiting.  Region  B is  the  non- 
linear  region  of  the  quieting  curve.  This  region  is  a complex  function 
of  the  signal  introduced  into  the  receiver,  the  thermal  noise  generated, 
the  receiver  IF  response,  and  degree  of  signal  limiting.  Region  C „'s 
the  linear  region  of  the  receiver.  Here  the  noise  decreases  in  direct 
proportion  to  the  increase  in  signal  introduced  into  the  receiver. 

Region  D is  the  so-called  saturated  region  of  tha  receiver.  Here,  the 
slot  noise  at  the  baseband  of  the  radio  is  independent  of  the  received 
signal  level  and  receiver  generated  noise. 

3.3  Describing  the  noise  in  regions  A,  and  B is  a complex  problem. 
Fortunately,  the  problem  has  been  solved  by  several  people  (e.g.  Middleton, 
Rice,  Stumpers,  and  Wang).  Most  of  the  solutions  are  quite  complicated. 

The  solutions  involve  multiple  infinite  series  and  cros.s  products  of 
confluent  hypergeometric  functions  or  multiple  indefinite  integrals 
of  complicated  infinite  series.  Taking  advantage  of  the  previous  solutions 
to  the  problem,  it  is  possible  to  tabulate  the  slot  noise  for  various 
carrier  to  thermal  noise  (C/N)  ratios  (assuming  hard  limiting).  The 
tabulated  results  on  the  next  few  pages  have  been  normalized.  To  convert 
the  tabulated  noise  values  to  dBm0  noise  in  a 3-1  kHz  baseband  slot, 
add  the  following  factor: 

No  (dBm0)  = +29.8  - 20  log  Af/ch  rm3  + 10  log  B - P 


where 

Af  . , = per  channel  r-ms  deviation  in  KHz 

/ch  rms 

B = receiver  IF  bandwidth  in  MHz 

P = baseband  pre-emphasis  in  dB  relative  to  baseband  pivot  frequency 
(formulas  for  this  are  given  later  in  this  report) 

To  convert  the  C/N  values  to  RSL,  use  the  following  formula: 

RSL(dBm)  = C/N(dB)  - 114.0  + NF  + 10  log  B 

where 

NF  = receiver  overall  noise  figure  in  dB  measured  at  the  same  point 
at  which  RSL  is  to  be  measured 


21 


(0uiap)  J9moj  asiON  }0|$ 

Slot  Noise  Versus  Received  Signal  Level  (RSL) 


Figure  ^ 
22 


Received  Signal  Level  (RSL,  dBm) 


‘t 


C/N 

C/N 

Diff. 

Rect. 

Gaus  • 

(dB) 

(PR; 

(dB) 

Noise 

Noise 

(dB) 

(dB) 

F/B-0.2 

5. 

3.162 

0.6 

-11.7 

-II. 1 

4. 

2.512 

1.2 

-10.8 

-9.5 

3. 

1 .995 

1.6 

-9.6 

-8.0 

2. 

1.585 

1.8 

-8.4 

-6 .7 

1. 

1 .259 

1.9 

-7.3 

-5.5 

0. 

1.000 

1.9 

-6.4 

-4.5 

-1 . 

0.794 

1.9 

-5.5 

-3.6 

-2. 

0.631 

1.9 

-4.8 

-2.9 

-3. 

0.501 

1.9 

-4.2 

-2.4 

-4. 

0.398 

1 .8 

-3.7 

-1.9 

-5. 

0.316 

1.8 

-3.3 

-1.5 

-6. 

0.251 

1.8 

-3.0 

-1.2 

i 

-7. 

0.200 

1.8 

-2.8 

-1.0 

-8. 

0.158 

1.7 

-2.5 

-0.8 

-9. 

0.126 

1.7 

-2.4 

-0.6 

-10. 

0.100 

1.7 

-2.2 

-0.5 

F/B»0. 1 

5. 

3. 162 

1.2 

-14.5 

-13.3 

4. 

2.512 

1 .8 

-13.0 

-11 .2 

1 

3. 

1.995 

2.0 

-11.3 

“9  » 3 

1 

2. 

1.585 

2.1 

-9.7 

-7.6 

1 . 

1.259 

2.0 

-8.2 

-6.2 

0. 

1.000 

1.9 

-6.9 

-5.0 

i 

-1 . 

0.794 

1.7 

-5.8 

-4.0 

-2. 

0.631 

1.6 

-4.9 

-3  • 3 

-3. 

0.50! 

1.5 

-4.1 

-2.6 

• 

-4. 

0.398 

1 .4 

-3.5 

-2.1 

-5. 

0.316 

1.3 

-3.0 

•1*7 

-6. 

0.251 

1 .2 

-2.6 

* 1 » j 

-7. 

0.200 

1.2 

-2.2 

-i.t 

-8. 

0.158 

1 .1 

-2.0 

-0.8 

-9. 

0.126 

1 . 1 

-1.7 

-0.7 

i 

-10. 

0.100 

1 .0 

-1.0 

-0.5 

t 

h/b=o.o 

5. 

3.162 

1 .6 

-16.0 

-14.4 

4. 

2.512 

2.0 

-14.0 

-12.0 

3. 

1.995 

2. 1 

-11  .9 

-9.8 

[ 

2. 

1 .585 

2.0 

-10.0 

-3.0 

1 . 

1.259 

1 .8 

-8.3 

-6.5 

0. 

1.000 

1 .6 

-6.0 

-5.2 

1 

-1  . 

0.794 

1.3 

-5.6 

-4.2 

-2. 

0.631 

1.1 

-4.5 

-3.4 

-3. 

0.501 

1 .0 

-3.7 

-2.7 

-4. 

0.398 

0.8 

-3.0 

-2.2 

-5. 

0.316 

0.7 

-2.4 

— 1.7 

-6. 

0.251 

0.6 

-1.9 

-1.4 

-7. 

0.200 

0.5 

-1.6 

— l.l 

-8. 

0.158 

0.4 

-1.3 

-0.9 

1 i 

-9. 

0.126 

0.4 

-l.l 

—0. 7 

■ 1 
1 
1 

-10. 

0.  IQO 

0.3 

-0.9 

-0.5 

Slot  Noiee 

RtctACguiar  and  G«u»»ian  RK/IF  R«»poo«a 

Table  5 

i 

i 

23 

C/N 

F/B-0.2  <«**> 

20.00 

19.00 

18.00 

17.00 

16.00 
lb.  00 

14.00 

i3.ro 

12.00 

11.00 

10.00 

9.00 

8.00 

7.00 

6.00 
b.00 


F/b*0.  I 


20.00 

19.00 

16.00 

17.00 

16.00 

16.00 

14.00 

13.00 

12.00 

1 1 ,00 

10.00 

V , 00 

6.00 
/ .00 

6.00 

5 . On 

I 

i F/u*o.ob 

■ 20.00 

I V.00 

18.00 

I 1.00 

1 0 . 00 
lb. 00 

14.00 

13.00 

12.00 

II  .00 

10.00 

9.00 

6.00 
/.00 

6.00 
b .00 


C/N 

Diff. 

Rect . 

Gdus . 

(PR) 

(dB) 

Noise 

(dB) 

Noise 

(dB) 

100.00 

0.34 

-28.31 

-28.65 

79.43 

0.34 

-27.86 

-28.20 

03.10 

0.34 

-26.86 

-27.20 

50. !2 

0.34 

-25.86 

-26.20 

39.81 

0.34 

-24.86 

-25.20 

31 .62 

0.34 

-23.86 

-24.20 

25.12 

0.34 

-22.86 

-23.20 

1 9.9b 

0.34 

-21.86 

-22.20 

15.8b 

0.34 

-20.86 

-21 .20 

12.59 

0.34 

-19.86 

-20.20 

10.00 

0.34 

-1 c .86 

-19.19 

7.94 

0.32 

-17.83 

-18. 15 

6.31 

0.25 

-16.70 

-16.96 

5.01 

0.08 

-15.36 

-15.44 

3.98 

0.20 

-13. 7i 

-13.51 

3. 1 6 

0.51 

-11.78 

-11.28 

1 00.00 

o.  07 

-34.33 

-34.26 

79.43 

0.07 

-33.88 

-33.81 

63.10 

0.07 

-32.88 

-32.81 

50.12 

0.07 

-31.88 

-31.81 

39.81 

O.o? 

-30.08 

-30.81 

31  .62 

0.07 

-29.86 

-29.81 

25.12 

0.07 

-26.88 

-28.81 

19.95 

0.07 

-27.08 

-27.81 

15.05 

0.07 

-26.88 

-26.81 

12.59 

0,07 

-25.88 

-25.81 

10.00 

o.OO 

-24.86 

-24.78 

7.94 

0.13 

-23. / 4 

-23.62 

0.31 

0.30 

-22.28 

-2 1 . 98 

5.01 

0.o3 

-20.16 

-19.53 

3.98 

0.98 

-17.43 

-16.45 

3 . 1 o 

1.22 

-14.50 

-13.28 

1 00.00 

0.17 

-40.35 

-40. 

, 18 

79.43 

0. 1 7 

-39.90 

-39. 

.73 

C3.  to 

0.17 

-36.90 

-38. 

,73 

50.12 

0.17 

-37.90 

-37. 

,73 

39.01 

0.17 

-30.90 

-3o. 

,73 

31 .62 

0.17 

-35.90 

-35. 

,73 

25.  12 

0.17 

-34.90 

-34. 

.73 

1 9.  95 

0.17 

-33.90 

-33. 

,73 

15.85 

0.17 

-32.90 

-32. 

,73 

12.59 

0.17 

-31 .89 

-31. 

,72 

10.00 

0.20 

-30.82 

-30. 

62 

7.94 

0.36 

-29.37 

-29. 

,01 

6.31 

0.77 

-26.87 

-26. 

, II 

5.01 

1.17 

-23.2! 

-22. 

,04 

3.98 

1 .40 

-19.23 

-17. 

83 

3.16 

1 .50 

-15.56 

-1  4. 

07 

Sloe  Noise 

Rectangular  and  Gauaaian  RF/IF  Response 


Table  5 
(cont. ) 


24 


C/N 

(dB) 

C/N 

(PR) 

Diff. 

(dB) 

Rect. 

Noise 

Gaua . 
Noise 

(dB) 

(dB) 

20.no 

1 00.00 

0.20 

-46.37 

-46.17 

19.00 

79.43 

0.20 

-45.92 

-45.73 

18.00 

63.10 

0.20 

-44.92 

-44.73 

17.00 

50.12 

0.20 

-43.92 

-43.73 

I0.00 

39.81 

0.20 

-42.92 

-42.73 

lb. 00 

31  .62 

0.20 

-41.92 

-41.73 

14.00 

25.12 

0.20 

-40.92 

-40.73 

13.00 

1 9.95 

0.20 

-39.92 

-39.73 

12.00 

15.85 

0.20 

-38.92 

-38.72 

II  .00 

12.59 

0.21 

-37.89 

-37.69 

10.00 

10.00 

0.31 

-36.61 

-36.30 

9.00 

7.94 

0.73 

-34.10 

- 37 

8.00 

6.31 

1 .24 

-29.63 

-28.39 

7.00 

5.01 

1 .47 

-24.49 

-23.02 

6.00 

3.96 

1 .55 

-10.82 

-18.27 

5.00 

3.16 

1 .58 

-15.88 

-14.30 

tyb*o.oi 


20.00 

100.00 

0.20 

-54 . 33 

-54.12 

19.00 

79.43 

0.20 

-53.88 

-53.68 

18.00 

63.  10 

0.20 

-52.88 

-52.68 

17.00 

50.12 

0.20 

-51.88 

-51.68 

to.  00 

39.61 

0.20 

-50.88 

-50.68 

15.00 

31 .62 

0.20 

-49.88 

-49.68 

14.00 

25.12 

0.20 

-48.88 

-48.68 

13.00 

19.95 

o.2o 

-47.88 

-47.68 

12.00 

15.85 

0.21 

-46.87 

-46.67 

1 1 .00 

12,59 

0.27 

-45.71 

-45.45 

10.00 

10,00 

0.70 

-43.21 

-42.51 

9.00 

7.94 

1.32 

-37.59 

-36.27 

8.00 

6.31 

1.53 

-30.87 

-29.34 

!.()<) 

5.01 

1 .59 

-24.93 

-23.34 

6.00 

3.98 

1 .60 

-20.01 

-18.41 

8.00 

3.16 

1 .61 

-15.97 

-1 4.36 

H/b-'O.OOb 


20.00 

1 00.00 

0.20 

-60.35 

-60.14 

19.00 

79.43 

0.20 

-59.90 

-59.70 

18.00 

63.10 

0.20 

-58.90 

-58.70 

17.00 

50.12 

0.20 

-57.90 

-57.70 

16.00 

39.81 

0.20 

-56.90 

-56.70 

15.00 

31  .62 

0.20 

-55.90 

-55.70 

14.00 

25.12 

0.20 

-54.90 

-54.70 

13.00 

19.95 

0.20 

-53.90 

-53.70 

12.00 

15.85 

0.22 

-52.87 

-52.66 

11  .00 

i 2 ,59 

0.42 

-51,27 

-50.84 

10.00 

10.00 

1.17 

-46.32 

-45.15 

9.00 

7.94 

1.52 

-38.43 

-36.90 

8.00 

6.31 

1 .59 

-31 .08 

-29.49 

7.00 

5.01 

! .60 

-25.00 

-23.39 

6.00 

3.98 

1 .61 

-20.03 

-18.43 

5.00 

3.16 

1 .61 

-15.98 

-14.37 

Slot  Noise 

Rectangular  and  Gaussian  RF/IF  Response 


Table  5 
(cont . ) 

2«; 


I 


C/H 

C/N 

Diff. 

Rect. 

Gaus. 

F/B-0.0025  <«»> 

(PR) 

(dB) 

Noise 

(dB) 

Noise 

(dB) 

20.00 

100.00 

0.20 

-66.37 

-66. id 

19.00 

79.43 

0.20 

-65.92 

-65.72 

13.00 

63.10 

0.20 

-64.92 

-64.72 

17.00 

50.12 

0.20 

-63.92 

-63.72 

16.00 

39.81 

0.20 

-62.92 

-62.72 

15.00 

31  .62 

0.20* 

-61 .92 

-61.72 

14,00 

25.12 

0.20 

-60.92 

-60.72 

13.00 

19.95 

0.21 

-59.92 

-59.71 

12.00 

15.85 

0.25 

-53.81 

-58.56 

II  .00 

12.59 

0.80 

-55.30 

-55.00 

10.00 

10.00 

1.47 

-47.64 

-46.17 

9.00 

7.94 

1.59 

-38.67 

-37.08 

8.00 

6.31 

1 .60 

-31 .14 

-29.53 

7.00 

5.0| 

1.61 

-25.01 

-23.40 

6.00 

3.98 

1 .61 

-20.04 

-18.43 

5.00 

3.16 

1 .61 

-15.98 

-14.38 

F/a-o.oci 

20.00 

100.00 

0.20 

-74.33 

-74.12 

19.00 

79.43 

0.20 

-73.38 

-73.68 

18.00 

63.10 

0.20 

-72.88 

-72.68 

17.00 

50.12 

0.20 

-71 .88 

-71 .68 

10. 00 

3V.8I 

0.20 

-70.88 

-70.68 

16.00 

31  .62 

0.20 

-09.88 

-69.68 

14.00 

25.12 

0.21 

-03.88 

-68.68 

13.00 

1 9. 95 

0.21 

-67.87 

-67.66 

12.00 

15.65 

0.44 

-66.20 

-65.76 

II  .00 

12.59 

1.36 

-58. y4 

-57.58 

10.00 

10. on 

1 .58 

-48.09 

-46.51 

9.00 

7 . 94 

1 .61 

-38.74 

-37.13 

8.00 

6.31 

1 .61 

-31 .15 

-29.54 

7.00 

5.01 

1 .61 

-25.02 

-23.41 

6.00 

3.98 

1 .61 

-20.04 

-18.43 

6.00 

3.16 

1 .61 

-15.  ; 

-14.38 

F/8-0.0005 

20.00 

100.00 

0.20 

-80.35 

-80.14 

19.00 

79.43 

0.20 

-79.90 

-79.70 

18.00 

63.10 

0.20 

-78.90 

-78.70 

17.00 

50.12 

0.20 

-77.90  , 

-77.70 

16.00 

39.81 

0.20 

-76.90 

76.70 

15.00 

31  .62 

0.20 

-75. 9'"' 

-/ 5.70 

14.00 

25 . 1 2 

0.21 

-;h.9o 

-74.70 

13.00 

19.95 

0.23 

-73.85 

-73.62 

12.00 

15.85 

0.83 

-70.65 

-69.82 

II  .00 

12,59 

1,54 

-59.66 

-58.12 

10.00 

10.00 

1 .60 

-48.16 

-46.55 

9.00 

7.94 

I .61 

-38.75 

-37.14 

8.00 

6.31 

1 .61 

-31.15 

-29.54 

7.00 

5,01 

1.61 

-25.02 

-23.41 

o.OO 

3.98 

1 .61 

-20.04 

-18.43 

5.00 

3.16 

1 .61 

-15.99 

-14.38 

Slot  Noise 

Rectangular  and  Gauaaian  RF/1F  Response 

Table  5 
(cont. ) 


26 


C/N 

C/N 

Dlff. 

Ract. 

Gaua. 

F/B-0.000 25  (dB) 

(PR) 

(dB) 

Noise 

(dB) 

Noise 

(dB) 

20.00 

100.00 

0.20 

-86.37 

-86.16 

>9.00 

79.43 

0.20 

-85.92 

-85.72 

18.00 

63.10 

0.20 

-84.92 

-84.72 

17.00 

50.12 

0.20 

-83.92 

-83.72 

>6.00 

39.81 

0.20 

-82.92 

-82.72 

15.00 

31 .62 

0.20 

-81.92 

-81.72 

14.00 

25.12 

0.21 

-80.92 

-80.72 

13.00 

19.95 

0.28 

-79.71 

-79.43 

>2.00 

15.85 

1 .27 

-73.22 

-71.95 

II  .00 

12.59 

1 .59 

-59.86 

-58.27 

10.00 

>0.00 

1.61 

-48.18 

-46.57 

9.00 

7.94 

1 .61 

-38.75 

-37.14 

8.00 

6.31 

1 .61 

-31  .J  5 

-29.54 

7.00 

5.01 

1.61 

-25.02 

-23.41 

6.00 

3.98 

1.61 

-20.04 

-18.43 

5.00 

3.16 

1 .61 

-15.99 

-14.38 

F/8-0.0001 

20.00 

100.00 

0.20 

-94 . 33 

-94. 12 

19.00 

79.43 

0.20 

-93.88 

-93.68 

18.00 

63.10 

0.20 

-92.88 

-92.68 

17.00 

50.12 

0.20 

-91 .88 

-91 .68 

16.00 

39.81 

0.20 

-90.88 

-90.68 

15.00 

31 .62 

0.21 

-89.88 

-89.68 

14.00 

25. 12 

0.21 

-88.07 

-88.66 

13.00 

1 9.95 

0.58 

-86.69 

-86.  1 1 

12.00 

15.85 

1.54 

-74.33 

-72.79 

1 1 .00 

12.59 

1.61 

-59.92 

-58.31 

10.00 

10.00 

1.61 

-48.18 

-46.57 

9.00 

7.94 

1.61 

-38.75 

-37.14 

8.00 

6.31 

1 .61 

-31.15 

-29.54 

7.00 

5.01 

1 .61 

-25.02 

-23.41 

6.00 

3.98 

1.61 

-20.04 

-18.43 

5.00 

3.16 

1.61 

-15.99 

-14.38 

F/B-0.0 

20.00 

100.00 

1.61 

-358.50 

-356.89 

19.00 

79.43 

1 .61 

-354.22 

-352.61 

18.00 

63.10 

1 .61 

-282.77 

-281,16 

17.00 

50.12 

1.61 

-225.92 

-224.31 

16.00 

39.81 

1.61 

-180.65 

-179.04 

15.00 

31  .62 

1 .61 

-1  44.59 

-142.98 

14.00 

25.12 

1 .61 

-115.84 

-114.23 

13.00 

19.95 

1.61 

-92.91 

-91.30 

12.00 

15.85 

1.61 

-74.58 

-72.97 

II  .00 

12.59 

1.61 

-59.93 

-58.32 

10.00 

10.00 

1.61 

-48.18 

-46.57 

9,00 

7.94 

1.61 

-38.75 

-37.14 

• 8.00 

6.31 

1 .61 

-31.15 

-29.54 

7.00 

5.01 

1.61 

-25.02 

-23.41 

6.00 

3.98 

1 .61 

-20.04 

-18.43 

5.00 

3.16 

1.61 

-15.99 

-14.38 

Slot  Nolaa 

Rectangular  end  Causal  an  IF/IF  Kaapooat 

Table  5 
(cont. ) 


33. 


F/B-0.2 


F/B-O. I 


F/B-0.0 


C/N 

C/N 

Diff. 

Avg. 

CdB) 

(PR) 

CdB) 

Noise 

CdB) 

5. 

3. 162 

0.3 

-11.4 

4. 

2.512 

0.6 

-10.2 

3. 

t .995 

0.8 

-8.8 

2. 

1.585 

0.9 

-7.6 

1. 

1.259 

0.9 

-6.4 

0. 

1.000 

1 .0 

-5.4 

-1 . 

0.794 

1.0 

-4.6 

-2. 

0.631 

1 .0 

-3.9 

-3. 

0.501 

0.9 

-3.3 

-4. 

0.398 

0.9 

-2.8 

-5. 

0.316 

0.9 

-2.4 

-6. 

0.251 

0.9 

-2.1 

-7. 

0.200 

0.9 

-1.9 

-8. 

0.158 

0.9 

-1.7 

-9. 

0.  126 

0.9 

-1.5 

-10. 

0. 100 

0.9 

-1.4 

5. 

3.162 

0.6 

-13,9 

4. 

2.512 

0.9 

-12.1 

3. 

1.995 

1.0 

-10.3 

2. 

1 .585 

1.0 

-8,6 

1. 

1 .259 

1 .0 

-7.2 

0. 

1 . 000 

0.9 

-6.0 

-1. 

0.794 

0.9 

-4.9 

-2. 

0.631 

0.8 

-4.  1 

-3. 

0.501 

0.8 

-3.3 

-4. 

0.39B 

0.7 

-2.8 

-5. 

0.316 

0.7 

-2.3 

-6. 

0.251 

0.6 

-1.9 

-7. 

0.200 

0.6 

-1.6 

-8. 

0.158 

0.6 

-1.4 

-V. 

0.126. 

0.5 

-1.2 

•10. 

0.100 

O.S 

-1.0 

0. 

3.  1 o2 

0.8 

-15.2 

4. 

2.512 

: -o 

-13.0 

3. 

1 .995 

1.0 

-10.9 

2. 

'.585 

1.0 

-9.0 

1. 

1.259 

0.9 

-7.4 

0. 

1.000 

0.8 

-6.0 

-1. 

0.794 

0.7 

-4.9 

-2. 

0.631 

O.o 

-3.9 

-3. 

0.501 

0.5 

-3.2 

-4. 

0.390 

0.4 

-2.6 

-5. 

0.316 

0.3 

-2.  1 

-6. 

0.251 

0.3 

-1.7 

-7. 

0.200 

0.3 

-1.3 

-8. 

0.158 

0.2 

-1.1 

-9. 

0.126 

0.2 

-0.9 

■10. 

0.100 

0.2 

-0.7 

Averaged  Slot  Nolae 

Table  6 

F/B«0.2  (dB) 

20.n0 

19.00 

18.00 

17.00 

16.00 
lb.  00 

14.00 

13.00 

12.00 

II  .00 

10.00 

9.00 

8.00 

7.00 

6.00 

5.00 

F/B-0.1 


C/N 

Diff. 

(PR) 

(dB) 

100.00 

0.27 

79.43 

0.27 

63.10 

0.27 

50.12 

0.27 

39.81 

0.27 

31  .62 

0.27 

25.12 

0.27 

1 9.95 

0.27 

15.85 

0.27 

12.59 

0.27 

10.00 

0.27 

7.94 

0.26 

6.31 

' 0.23 

5.01 

0.14 

3.98 

0.00 

3.16 

0.15 

20.00 

19.00 

18.00 

17.00 

16.00 

15.00 

14.00 

13.00 

12.00 

II  .00 

10.00 

9.00 

8.00 

7.00 

6.00 

5.00 

F/B-0.05 


100.00 

0.07 

79.43 

0.07 

63.10 

0.07 

50.12 

0.07 

39.81 

0.07 

31.62 

0.07 

25.12 

0.07 

1 9.95 

0.07 

15.85 

0.07 

12,59 

0.07 

10.00 

0.06 

7.94 

0.04 

6.31 

0.05 

5.01 

0.21 

3.98 

0.39 

3.16 

0.51 

20.00 

19.00 

18.00 

17.00 

10.00 

15.00 

14.00 

13.00 

12.00 
I I .00 
10.00 

9.00 

8.00 

7.00 

6.00 
5.00 


100.00 

79.43 

63.10 

50.12 
39.81 
31  .62 

25.12 
I 9.  95 
15.85 
12.59 

10,00 

7.94 

6.31 

5.01 

3.98 

3.16 


0.02 

0.02 

0.02 

0.02 

0.02 

0.02 

0.02 

0.02 

0.02 

C.02 

o.oo 

0.08 

0.28 

0.48 

0.60 

0.65 


Averaged  Slot  Noise 


Table  6 
(cont . ) 


Avg. 

Noise 

(dB) 

-29.03 

-28.03 

-27.03 

-26.03 

-25.03 

-24.03 

-23.03 

-22.03 

-21.03 

-20.03 

-19.02 

-17.99 

-16.83 

-15.40 

-13.61 

-II  .53 


-34.85 

-33.85 

-32.85 

-31.85 

-30.85 

-29.85 

-28.85 

-27.85 

-26.85 

-25.84 

-24.82 

-23.68 

-22.13 

-19.84 

-16.94 

-13.89 


-40.82 

-39.82 

-38.82 

-37.82 

-36.82 

-35.82 

-34.82 

-33.82 

-32.82 

-31 .81 

-30.72 

-29.1 9 

-26.49 

-22.63 

-18.53 

-U.62 


29 


f-/B°0.02b 


F/B*O.OI 


F/B=0.005 


C/N 

C/N 

Diff. 

Avg. 

Nolae 

(dB) 

(PR) 

(dB) 

(dB) 

20.00 

100.00 

0.00 

-46.82 

i 9.00 

79.43 

0.00 

-45.82 

18.00 

63.10 

0.00 

-44.82 

1 7.00 

50.12 

0.00 

-43.82 

16.00 

39.81 

0.00 

-42.82 

lb. 00 

31  .62 

0.00 

-41 .82 

14.00 

2b. 12 

0.00 

-40.82 

Id.  00 

19.95 

0.00 

-39.82 

12.00 

15.85 

0.00 

-38.82 

11.00 

12.59 

0.00 

-37.79 

10.00 

10.00 

0.05 

-36.45 

9.00 

7.94 

0.26 

-33.73 

8.00 

6.31 

0.52 

-29.01 

7.00 

5.01 

0.63 

-23.76 

6.00 

3.98 

0.67 

-19.05 

b.00 

3.16 

0.69 

-15.09 

20.00 

100.00 

0.00 

-54.78 

19.00 

79.43 

0.00 

-53.78 

18.00 

63.10 

0.00 

-52.78 

17.00 

bO.  1 2 

0.00 

-51 .78 

10. 00 

39.81 

0.00 

-bO.78 

lb.  00 

31 .62 

0.00 

-49.78 

14.00 

2b . 1 2 

n.00 

-48.78 

13.00 

1 9.9b 

0.00 

-47.78 

12.00 

lb. 8b 

0.00 

-46.77 

1 1 .on 

12.59 

0.03 

-45.58 

10. oo 

10.00 

0.2b 

-42.86 

9.00 

7.94 

0.b6 

-36.93 

8.00 

6.31 

0.66 

-30.11 

7.00 

5.01 

0.69 

-24.14 

6,00 

3.98 

0.70 

-19.21 

b .00 

3,16 

0.70 

-15.17 

20.00 

100.00 

0.00 

-60.80 

19.00 

79.43 

0.00 

-59.80 

18.00 

63.  'O 

0.00 

-58.80 

17.00 

50.12 

0.00 

r57 ,80 

16.00 

39.81 

0.00 

-56.80 

lb.  00 

31  .62 

0.00 

-55.80 

14.00 

25.12 

0.00 

-54 ,80 

13.00 

19.95 

0.00 

-53.80 

12. HO 

lb. 8b 

0.01 

-52.76 

1 1 .00 

12.59 

0.  1 1 

-51 .05 

10.00 

9.00 

10.00 

7.94 

0.48 

0.66 

-45.74 

-37.67 

8.00 

6.31 

0.69 

-30.29 

7.00 

b.Ol 

0.70 

-24.19 

6.00 

3.98 

0.70 

-19.23 

5.00 

3.16 

0.70 

-i5. 18 

Averaged  Slot 

Nolae 

Table  6 
(cont. ) 


30 


F/B-0.0025 


C/N 

C/N 

Dlff. 

Avg. 

Noiaa 

(dB) 

(PR) 

(dB) 

(dB) 

20.00 

100.00 

0.00 

-66.82 

19.00 

79.43 

0.00 

-65.82 

IS. 00 

63.10 

0.00 

-64.82 

17.00 

50.12 

0.00 

-63.82 

16.00 

39.81 

0.00 

-62.82 

lb. 00 

31  .62 

0.00 

-61 .82 

14.00 

25. 12 

0.00 

-60.82 

I 3.00 

19.95 

0.00 

-59.82 

12.00 

15.85 

0.02 

-58.68 

1 1 .00 

12.59 

0.30 

-55.40 

10.00 

10.00 

0.63 

-46.90 

9.00 

1.94 

0.69 

-37.87 

0,00 

6.31 

0.70 

-30.33 

7.00 

5.01 

0.70 

-24.21 

6. 00 

3.98 

0.70 

-19.23 

b.00 

3.16 

0.70 

-15.18 

F/B-O.OOI 


20.00 

100.00 

0.00 

-74.78 

IV. 00 

79.43 

0.00 

-73.78 

18.00 

63.10 

0.00 

-72.78 

17.00 

50.12 

0,00 

-71.78 

16.00 

3v,3l 

0.00 

-70.78 

15.00 

31  .62 

0.00 

-69.78 

14.00 

25 . 1 2 

O.no 

-68.78 

1 3.00 

1 9.  yb 

O.oo 

-67.76 

12.00 

lb. 85 

0.12 

-65.98 

1 1 .00 

12.59 

0.58 

-58.26 

10.00 

10.00 

0.69 

-47.30 

9.00 

7.94 

0.70 

-37.93 

8.00 

6,31 

0.70  • 

-30.35 

7.00 

5.01 

0.70 

-24.21 

o , 00 

3.98 

0.70 

-19.24 

5.00 

3.16 

0.70 

-15.18 

F/b«0.0005 


20.00 

100.00 

o.oo 

-80.80 

1 9.00 

79.43 

0.00 

-79.80 

18.00 

03.10 

0.00 

-78.80 

1 1.00 

50.12 

0.00 

-77.80 

16.00 

39.81 

0.00 

-76.80 

15.00 

31  .02 

0.00 

-75.80 

14.00 

25.12 

0.00 

-74.80 

13.00 

19.95 

0.01 

-73.73 

12.00 

15.85 

o.3l 

-70.24 

1 1 .00 

12.59 

0.67 

-58.89 

10.00 

10.00 

0.70 

-47.36 

9.00 

7.94 

0.70 

-37.94 

8.00 

6.31 

0.70 

-30.35 

7.00 

5.0! 

0.70 

-24.21 

6.00 

3.98 

0.70 

-19.24 

5.00 

3.16 

0.70 

-15.18 

Averaged  Slot 

Noise 

Table  6 
(cont . ) 


31 


C/N 

C/N 

Dlff. 

Avg. 

Noise 

<dB) 

(PR) 

(dB) 

(dB) 

20.00 

100.00 

0.00 

-H6.R2 

19.00 

79.43 

0.00 

-85.82 

18.00 

63.10 

0.00 

-84.82 

17.00 

50. 1 2 

0.00 

-83.82 

16.00 

39. HI 

0.00 

-82.82 

1 1> . 00 

31  .62 

•0. 

-81 .82 

14.00 

2b.  12 

0.00 

-80.82 

13.00 

19.9b 

0.04 

-79.57 

12.00 

15.85 

0.53 

-72.58 

1 1 .00 

12.59 

0.69 

-59.06 

10.00 

10.00 

0.70 

-47.37 

9.00 

7.94 

0.70 

-37.94 

8.00 

6.31 

0.70 

-30.35 

7.00 

b .01 

0.70 

-24.21 

0.00 

3.98 

0.70 

-19.24 

b.oc 

3.16 

0.70 

-15.18 

8/8=0.0001 


20.00 

1 00.00 

1 9.00 

/ 9.43 

18.00 

63.10 

17.00 

50.  I 2 

10. 00 

39.81 

lb. 00 

31 .02 

14.00 

25.12 

13.00 

19.95 

12.00 

lb.  Ho 

1 1 .00 

12.59 

10. 00 

10. 00 

9.  no 

/ . 94 

H.00 

0.31 

/ .00 

5.01 

O.00 

3.VH 

b . no 

3.10 

0. 

-94.78 

0. 

-93.78 

0. 

-92.78 

0. 

-PI. 78 

0. 

-90.78 

n.oo 

-89.78 

n.no 

-88.77 

n , | 9 

-86.40 

0.0/ 

-73.56 

0.70 

-59. 1 1 

n.70 

-47.38 

0.70 

-37.94 

0.70 

-30.35 

0.70 

-24.21 

O./o 

-19.24 

0.70 

-15.18 

8/H-0.0 


20.00 

1 00 . 00 

0.70 

-557.70 

19.  on 

/9.43 

0.70 

-353.42 

18.00 

03.10 

0.70 

-281 .97 

17.00 

50.12 

0.70 

-225. 11 

lo.OO 

39.81 

0.70 

-1 79.84 

15.00 

31  .02 

0.70 

-143.78 

14.00 

25.12 

0.70 

-115.04 

13.00 

1 9.9b 

0.70 

-92.10 

12.00 

15.85 

0.70 

-73.78 

1 1 .on 

12.59 

0,70 

-59.12 

10.00 

10.00 

0.70 

-47.38 

9.00 

/ . 94 

0.70 

-37.94 

8.00 

0.31 

0.70 

-30.35 

/,00 

5. 01 

0.70 

-24.21 

6.00 

3.98 

0.70 

-19.24 

5.00 

3.16 

0.70 

-15.18 

Averaged  Slot  Noise 

Table  6 
(cont . ) 


52 


3.4  The  formulas  used  to  generate  the  slot  noise  tables  are  given  later 
in  this  report.  In  some  cases,  there  is  a slight  error  associated  with 
the  equation  approximations.  The  error  limits  are  noted  later.  In 

the  tabulated  results,  F is  the  baseband  frequency  of  the  center  of 
the  noise  slot;  B is  the  IF  bandwidth  C/N  i3  carrier  to  (thermal)  noise 
ratio.  PR  indicates  power  ratio.  DIFF  is  the  difference  between  the 
two  slot  noise  values.  This  gives  a good  check  on  the  dependency  of 
the  slot  noise  on  IF  response.  Gaussian  and  rectangular  IF  responses 
form  reasonable  limiting  cases  which  limit  the  realistic  range  of  IF 
response  shapes  encountered  in  practice.  All  slot  noise  values  have 
been  normalized  to  the  value  of  slot  noise  at  F/B  = 0 and  C/N  = -dB  for 
he  respective  IF  response. 

3.5  Following  this  set  of  values,  the  averaged  values  have  been  computed. 
These  values  are  the  algebraic  average  of  the  values  for  the  Gaussian 

and  the  rectangular  IF  bandwidth  responses.  The  DIFF  column  indicates 
the  difference  between  the  averaged  IF  response  and  either  of  the  Gaussian 
and  rectangular  IF  responses.  This  column  gives  an  estimate  of  the 
variation  to  be  expected  between  averaged  theoretical  quieting  curve 
characteristics  and  actual  quieting  curve  measurements.  The  averaged 
noise  values  have  been  normalized  to  the  averaged  slot  noise  for  F/B 
= 0 and  C/N  = - (dB). 

3.6  There  are  a couple  of  common  FM  receiver  specifications  related 
to  the  slot  noise  performance  of  the  receiver  in  the  A,  B,  and  upper 
C regions.  One  of  the  specifications  is  20  dE  (and  occasionally  30 
dB)  noise  quieting.  The  20  dB  (or  30  dB)  quieting  specification  is 
the  RSL  at  which  the  noise  in  a specified  baseband  slot  is  20  dB  (or 
30  dB)  lower  than  the  noise  in  the  slot  with  no  RSL  present  at  the 
receiver  input.  Theoretical,  20  dB  and  30  dB  C/N  ratios  are  listed 

on  the  next  page.  The  C/N  values  can  be  converted  to  RSL  using  receiver 
parameters  and  the  previous  formulas. 

•j.7  The  other  specification  is  FM  slot  noise  threshold  or  simply  FM 
threshold.  There  is  not  complete  agreement  by  sources  as  to  the  defi- 
nition of  FM  threshold.  The  most  common  definition,  however,  is  the 
RSL  at  which  the  receiver  baseband  slot  noise  has  increased  2 dB  for 
a 1 dB  reduction  in  RSL.  Thi3  is  often  determined  graphically  with 
quieting  curves  by  extending  the  straight  line  of  region  C into  region 
B and  defining  FM  tnreshold  as  the  RSL  at  which  slot  noise  is  one  dB 
above  the  straight  line  extension.  Generally,  FM  threshold  is  measured 
in  a narrow  (nominal  3.1  KHz  wide)  noise  slot.  Theoretical  narrow  slot 
FM  threshold  values  are  listed  on  the  following  page.  Occasionally, 
however,  it  is  desirable  to  measure  FM  threshold  in  a relatively  wide 
slot.  For  example,  if  the  FM  receiver  baseband  contains  a wideband 
video  or  digital  circuit,  the  FM  noise  threshold  over  the  entire  bandwidth 
of  the  baseband  signal  would  be  appropriate.  The  theoretical  wide  noise 
slot  FM  thresholds  are  tabulated  on  the  next  page.  For  the  FM  threshold 
charts,  (f/B)  is  the  narrow  slot  baseband  frequency  divided  by  the  IF 


33 


C/N(dB) 


f/B 

Rect. 

IF 

Gaut. 

IF 

Avg. 

IF 

0.2 

12.6 

10.9 

11.8 

0.1 

7.2 

7.2 

7.2 

0.05 

6.3 

6.5 

6.4 

0.025 

6.1 

6.4 

6.3 

0.01 

6.0 

6.3 

6.2 

0.001 

6.0 

6.3 

6.2 

0.0001 

6.0 

6.3 

6.2 

0.0 

6.0 

6.3 

6.2 

Theoretical  20  dB  Noise  Quieting 
(narrow  slots) 

C/N(dB) 


f/B 

Rect. 

IF 

Gaus. 

IF 

Avg. 

IF 

0.2 

22.6 

20.9 

21.8 

0.1 

15.8 

15.2 

15.5 

0.05 

9.6 

9.6 

9.6 

0.025 

8.1 

8.3 

8.2 

0.01 

7.9 

8.1 

8.0 

0.001 

7.8 

8.1 

8.0 

0.0001 

7.8 

8.1 

8.0 

0.0 

7.8 

8.1 

8.0 

Theoretical  30  dB  Noise  Quieting 
(narrow  slots) 

Table  7 


34 


f/B 

Rectangular  RF/IF 
C/N  (dB) 

Gauasien  RF/IF 
C/N  (dB) 

Averaged  Reaponse 
C/N  (dB) 

0.2 

6.19 

6.69 

6.4*1 

0.1 

7.56 

7.85 

7.71 

0.05 

8.58 

8.79 

8.65 

0.025 

9.39 

9.56 

9.48 

0.0158 

9.85 

10.00 

9.93 

0.01*17 

9.92 

10.07 

10.0 

0.013** 

10.00 

10.15 

10.08 

0.01 

10.26 

10. *40 

10.33 

0.005 

10.8*< 

10.9*4 

10.89 

O.0025 

11.31 

11. *41 

11.36 

0.001 

11.88 

11.97 

11.93 

0.0005 

12.26 

12.35 

12.31 

0.00025 

12.62 

12.70 

12.66 

0.0001 

13-0*4 

13-11 

13.08 

0.0 

Undefined 

Undefined 

Undefined 

s 

1 Theoretical  FM  Noise  Thresholds 

| (.narrow  slots) 

i 

I 

* Table  8 


35 


} 


o 

• 

o 

V) 

I— 1 

r-l 

CM 

CO 

CO 

o 

3 U4 

• 

• 

• 

• 

• 

CO  M 

CO 

CO 

CO 

CO 

CO 

o' 

o 

rH 

r-l 

1-4 

rH 

rH 

II 

X 

8 

■u 

o 

r-l 

CM 

CO 

CO 

/-N 

u u* 

• 

• 

• 

• 

• 

PQ 

o M 

CO 

CO 

CO 

CO 

CO 

erf 

H 

r-l 

r-l 

rH 

rH 

4-1 

> — ' 

T~-< 

a 

w 

O 

o 

r— 1 

CM 

CO 

o 

3 ft. 

• 

• 

• 

• 

• 

• 

CO  M 

CM 

CM 

CM 

CM 

CM 

o 

O 

rH 

r-l 

r-l 

f-H 

f-H 

II 

X 

8 

4J 

O' 

ON 

o 

CM 

CM 

(0 

no 

f-H 

/-N 

U f*-» 

• 

♦ 

* 

• 

• 

PQ 

0)  M 

r*H 

t— l 

CM 

CM 

CM 

Q 

4-1 

Crf 

H 

r-l 

rH 

rH 

rH 

A 

in 

01 

H 

# 

P 

o 

w 

vO 

00 

00 

V 

CO 

*rH 

0 tti 

• 

• 

• 

• 

* 

o 

<B  M 

o 

o 

O 

o 

o 

II 

O 

H 

r-l 

r-l 

rH 

rH 

o 

k 

g 

55 

§ 

z-s 

• 

4J 

CO 

CO 

in 

r*- 

£ 

PQ 

0 f*< 

• 

• 

• 

• 

• 

\ 

0)  M 

o 

O 

o 

o 

o 

f—H 

4-1 

Crf 

r-l 

r-l 

rH 

rH 

rH 

cd 

'w' 

o 

fH 

4J 

r-l 

OT 

o> 

On 

.M 

in 

VO 

<D 

U 

o 

• 

3 U< 

• 

• 

• 

• 

• 

O 

II 

<3  M 

r-» 

00 

00 

00 

4) 

n 

X 

P 

/•V 

u 

v£ 

vO 

o 

JO 

V* 

PQ 

0 

• 

• 

• 

• 

• 

4)  M 

00 

CO 

00 

4-1 

0 6 

N-/ 

(N 

ut 

r- 

00 

CM 

vO 

r>- 

• 

3 U. 

• 

• 

• 

• 

• 

o 

CO  H 

vO 

vO 

r- 

O 

8 

z~v 

■U 

CM 

CO 

00 

CM 

CO 

PQ 

0 b 

• 

* 

* 

• 

• 

41  M 

v£> 

vO 

vO 

h* 

4-1 

B 6 

v-/ 

c 

X 

■f-i 

TO 

O' 

& 

B 

ON 

Z*' 

ON 

ON 

in 

r— 1 

o 

PQ 

PO 

• 

• 

• 

• 

• 

o 

o 

o 

o 

o 

4-1 

4-1 

Table  9 


36 


(wide  noise  slots) 


bandwidth,  (f/B)  max  is  the  highest  baseband  slot  frequency  divided 
by  IF  bandwidth,  and  (f/B)  min  is  the  lowest  baseband  slot  frequency 
divided  by  IF  bandwidth.  Note  that  a wide  slot  FM  threshold  always 
occurs  within  1 dB  of  the  FM  threshold  for  a narrow  slot  which  has 
baseband  frequency  equal  to  the  maximum  frequency  of  the  wide  slot. 

3.8  The  last  region  of  interest  is  D.  By  conventional  linearized 
methods,  (e.g.,  see  Rowe  |nd  Panter)  it  can  be  shown  that  in  the  D region, 
the  noise  is  the  normal  f (6  dB  per  octave)  thermal  noise  plus  the 
time  derivative  of  any  phase  noise  associated  with  the  received  carrier. 

The  differentiated  phase  n^ise  is  called  frequency  noi3e.  It  is  distinctly 
different  noise  than  the  f receiver  thermal  noise.  For  information 
regarding  phase  and  frequency  noise,  see  references  listed  in  bibliography. 
In  the  D region,  the  f noise  is  negligible  and  the  dominate  noise  source 
is  the  frequency  noise  of  the  received  carrier.  At  a sufficiently  high, 
stable  carrier  to  thermal  noise  ratio  with  no  multipath  and  the  carrier 
tuned  to  the  center  of  a symetric  IF,  the  only  significant  frequency 
noise  sources  are  the  various  oscillators  in  the  transmitter  and  receiver. 

3.#  The  frequency  noise  of  an  oscillator  is  a complex  phenomenon. 

In  general,  the  phase  noise  of  an  oscillator  varies  in  power  roughly 
2 

as  1 / f where  f is  tht  frequency  of  the  measurement  slot  away  from 
the  carrier  frequency.  The  noise  at  the  baseband  of  an  FM  demodulator, 
however,  is  the  time  de  'ivative  of  the  phase  noise  (for  high  carrier 
to  noise  ratios).  Sinct.  the  power  response  of  a differentiator  is 
2 

proportional  to  f , the  'requency  noise  appearing  at  the  baseband  of 
the  radio  13  essentially  Hat  (independent  of  baseband  frequency). 

After  review  of  information  regarding  fundamental  frequency  (unmultiplied) 
free  running  microwave  oscillators  from  various  vendors,  it  appears 
that  this  is,  in  fact,  the  case  for  baseband  frequencies  greater  than 
roughly  100  kHz.  Below  this  frequency,  the  noise  gradually  increases. 

The  frequency  noise  within  the  first  few  kilohertz  of  the  baseband  can 
be  several  times  the  noise  in  the  mid  region  of  the  baseband.  As  Baprawski 
et  al  mention,  use  of  properly  optimized  phase-locked  loops  can  be  used 
to  reduce  the  noise  near  the  low  end  of  the  baseband.  Duo  to  feedback 
loop  frequency  response  requirements,  present  phase-locked  frequency 
sources  can  only  be  used  to  reduce  noise  in  roughly  the  first  one  half 
megahertz  of  the  radio  baseband.  Often  the  output  of  an  oscillator 
is  multiplied  to  derive  a higher  frequency.  As  Payne  points  out,  multi- 
plying the  output  of  the  oscillator  increases  the  effective  deviation 
of  the  frequency  noise.  Specifically,  noise  increase  is  20  log  M where 
M is  composite  oscillator  multiplication.  For  example,  doubling  the 
oscillator  frequency  causes  a 6 dB  increase  in  frequency  noise  due  to 
doubling  the  effective  deviation  of  the  frequency  noise  components. 

Payne  also  mentions  that  in  the  gigahertz  region,  the  noise  figure  of 
the  multiplier  diodes  can  be  a significant  factor.  Up  or  down  converter 
mixer  diodes  can  have  the  same  effect.  The  noise  figure  of  the  diodes 
can  cause  a significant  increase  in  slot  noise  abovj  roughly  one  megahertz 
in  the  radio  baseband.  The  following  page  graphs  typical  FM  receiver 
baseband  slot  noise  due  to  a free  running  unmodulated  M/W  oscillator. 


37 


i 

(8P)  J3M0J  asjojj  wj 

; Typical  Microwave  Oscillator  Noise 

Figure  5 

38 


Baseband  Slot  Frequency 


I — -t — h— t + 


i — t. 


! 

! 


! 


I 


I 


Figure  6 


Horizontal:  1 kHz/Div. 


Measurement  Bandwidth: 
100  Hz 


Horizontal:  5 kHz/Div. 


Meaourement  Bandwidth: 
100  Hz 


Horizontal:  20  kHz/Div. 


Measurement  Bandwidth: 
300  Hz 


TROPO  M/W  Transmitter 
VCO  Spectrum 


39 


Vertical:  10  dB/Div. 
Horizontal:  20  kHz/Div 
(0  to  200  kHz) 


Note:  This  LOS  M/W 

receiver  Is  not  the 
one  used  In  the  rest 
of  this  report. 


Vertical:  10  dB/Div. 
Horizontal:  50  kHz/Div, 
(0  to  500  kHz) 


RSLs(dBm)  - -100,  -90, 
-80,  -70,  -60,  -50, 
-40,  -30,  -20 

Measurement  Bandwidth: 
3 kHz 


Vertical:  10  dB/Div. 
Horizontal:  1 MHz/Dlv. 
(0  to  10  MHz) 


LOS  M/W  Receiver 
no  de-emphasis 


Baseband  Noise  Spectrum 
for  Various  RSLs 


Figure  7 


0 


3.10  There  are  several  oscillators  in  a microwave  system.  The  primary 
oscillators  of  interest  are  the  local  oscillators  for  up  and  down  frequency 
conversion  and  the  frequency  modulated  oscillator  at  the  transmitter 
(voltage  controlled  oscillator-VCO) . The  up  and  down  converter  local 
oscillators  can  be  phase-locked  to  reduce  low  frequency  noise.  As  Lawson 
and  Uhlenbeck  have  noted,  certain  hybrid  configurations  allow  the  noise 
sidebands  of  the  oscillator  to  be  suppressed.  These  methods  can  not 

be  used  on  the  modulated  oscillator,  however,  3ince  any  method  used 
to  suppress  the  inherent  noise  of  the  oscillator  will  also  suppress 
the  modulation.  Therefore,  the  oscillator  to  be  modulated  must  have 
inherently  low  noise.  If  the  output  of  this  oscillator  is  multiplied, 
a much  quieter  oscillator  is  required.  It  should  be  mentioned  that 
the  transmit  oscillators  are  usually  frequency  stablized  by  phase  locked 
loops.  The  loop  bandwidths  of  these  systems  are  quite  narrow  (often 
less  than  a Hertz).  These  loops  will  not  suppress  noise  in  the  radio 
baseband  greater  than  a few  Hertz  above  zero  frequency.  It  can  be  antici- 
pated that  the  baseband  noise  of  a voltage  controlled  M/W  oscillator 
used  for  FM  will  be  similar  to  the  noise  of  a free  running  M/W  oscillator. 

3.11  To  point  out  the  complex  nature  of  oscillator  noise,  spectrum  noise 
plots  of  the  sideband  noise  spectrum  of  a microwave  voltage  controlled 
oscillator  (with  Automatic  Frequency  Control  circuitry  disabled)  were 
taken.  The  plots  are  on  the  next  page.  The  sideband  frequency  distribution 
is  a combination  cf  both  amplitude  modulation  noise  (in  phase  noise) 

as  well  as  phase/frequency  noise  (quadrature  noise).  Therefore,  a direct 
relation  between  the  frequency  spectrum  and  noise  in  the  baseband  is 
not  possible.  However,  the  noise  process  is  obviously  complex  in  the 
near  carrier  region. 

3.12  To  illustrate  the  thermal  and  phase  noise  in  a typical  microwave 
receiver,  the  baseband  noise  spectrum  was  plotted  for  various  received 
signal  levels  us.ing  a spectrum  analyzer.  The  pxots  are  on  the  next 
page.  Note  the  oscillator  noise  at  the  low  baseband  frequencies  at 
high  C/Ns  and  the  crossing  of  the  noise  slots  for  low  C/Ns. 

3.13  Having  determined  the  basic  slot  noise  as  a function  of  RSL,  the 
next  item  of  immediate  interest  is  the  performance  of  the  baseband 
signal  versus  RSL.  As  long  as  the  received  signal  is  hard  limited  by 
the  limiter  prior  to  demodulation  and  as  long  as  the  carrier  to  noise 
ratio  (C/N)  is  large,  the  received  signal  at  the  baseband  will  vary 
directly  as  the  transmitted  baseband  signal  and  will  be  completely 
independent  of  RSL.  Part  of  an  FM  demodulator  is  an  envelope  detector. 

All  Incoherent  detectors,  including  envelope  detectors,  exhibit  a baseband 
signal  threshold.  That  is  to  say,  for  low  C/Ns,  the  baseband  signal 

will  be  suppressed.  Several  people  (e.g.,  Middleton,  Rice,  and  Stumpers) 
have  derived  equations  relating  FM  demodulator  input  modulated  signal 
to  noise  ratio  and  output  baseband  signal  suppression  for  a sine  wave 
baseband  signal.  Schwartz  shows  that  the  same  equation  for  baseband 


41 


t 


C/N 

C/N 

Signal 

(dB) 

(PR) 

Laval 

(dB) 

10. 

10.0000 

0.0 

9. 

7.9433 

0.0 

0. 

6,3096 

0.0 

7. 

5.0119 

-0.1 

6. 

3.9811 

-0.2 

5. 

3.1623 

-0.4 

4. 

2.5119 

-0.7 

3. 

1 .9953 

-1.3 

2. 

1.5849 

-2.0 

1 . 

1 .2589 

-2.9 

0. 

1 .00 00 

-4.0 

-1  . 

0.7943 

-5.2 

-2. 

0.6310 

-6.6 

-3. 

0.5012 

-8.1 

-4. 

0.3901 

-9.7 

-5. 

0.3162 

-II  .3 

-6. 

0.2512 

-13.1 

-7. 

0. 1 995 

-14.9 

-r. 

0.1585 

-16.7 

-9. 

0.1259 

-18.5 

-1C. 

0.  1000 

-20.4 

-It. 

0.0/94 

-22.3 

-12, 

0.0631 

-24.3 

-13. 

0.0501 

••26.2 

-14. 

0.0399 

-23.2 

-1b. 

0.0316 

-30.  1 

-16. 

0.0251 

-32.1 

-17. 

0.0200 

-34.1 

-lb. 

0.0158 

-36. 1 

-19, 

0.0126 

-33.1 

-20. 

0.0100 

-40.0 

-21  . 

0,  0u'79 

-42.0 

-22. 

0,0063 

-44.0 

-23. 

0.0050 

-46.0 

-24. 

0.0040 

-43.0 

-2b. 

0.0032 

-50.0 

-26. 

0.0025 

- 52 , 0 

-27. 

0.0020 

-54.0 

-28. 

0.0016 

-56.0 

-29. 

0,0013 

-58.0 

-30. 

0.0010 

-60.0 

-31  . 

0,0008 

-62.0 

-32. 

0.0006 

-64.0 

-33. 

0.0005 

-66.0 

-34. 

0. 0004 

-68.0 

-35. 

0.0003 

-70.0 

-36. 

0.0003 

-72.0 

-37. 

0.0002 

-74.0 

-38, 

0.0002 

-76.0 

-39, 

0.0001 

-78.0 

-40. 

0.0001 

-30.0 

Baseband  Signal  Suppression 
Table  10 

42 


signal  suppression  holds  for  any  arbitrary  baseband  signal.  The  equation 
for  baseband  signal  suppression  will  be  given  later  in  this  report. 

The  relation  between  baseband  signal  suppression  and  receiver  input 
carrier  to  noise  ratio  is  listed  on  the  next  page.  The  dB  factor  listed 
is  (algebraicly)  added  to  the  normal  baseband  dBm  or  dBm0  power  level. 

Receiver  carrier  to  noise  ratio  can  be  converted  to  received  signal 
level  (RSL)  based  on  FM  receiver  parameters  using  formulas  listed  later 
in  this  report. 

3.14  One  of  the  assumptions  so  far  has  been  that  the  microwave  receiver 
was  hard  limiting  the  received  signal  prior  to  demodulation.  This 
assumption  requires  us  to  believe  that  all  microwave  receiver  limiters 
are  driven  to  saturation  with  Just  the  thermal  noise  from  the  front 
end  of  the  receiver.  This  is  a highly  questionable  assumption.  Loss 
of  hard  limiting  effects  both  the  baseband  signal  and  noise  properties 
of  an  FM  receiver. 

3.15  Like  the  problem  of  FM  noise  for  low  C/N,  the  problem  of  demodulated 
FM  signals  and  nol3e  with  arbitrary  or  no  limiting  is  complex.  Middleton 
has  solved  this  problem  for  an  FM  receiver  with  Gaussian  IF  frequency 
response.  His  results  indicate  that  as  limiter  action  transitious  from 
hard  limiting  to  soft  (partial)  limiting  to  no  limiting,  the  slot  noise 
with  no  carrier  present  transitions  to  less  and  less  noise.  Also,  the 
baseband  noise  spectrum  changes  dramatically  as  limiting  is  lost.  Middleton's 
results  for  the  two  extreme  cases  of  limiting  were  used  to  determine 

the  following  normalized  baseband  noise  spectrum  charts.  In  absolute 
power  the  noise  associated  with  "1.0  milliwatt"  on  the  "hard  limiting" 
curve  is  greater  than  the  absolute  power  associated  with  "1.0  milliwatt" 
on  the  "no  limiting"  chart.  The  most  obvious  differences  between  the 
two  curves  is  the  change  in  shape.  This  is  interesting,  but  the  curves 
also  indicate  a more  important  consideration.  For  a given  slot  frequency 
and  given  C/N  greater  than  0 dB,  the  slot  noi3e  is  much  greater  with 
no  limiting  than  it  is  with  full  limiting.  We  can  anticipate  that  if 
a receiver  limiter  action  starts  going  soft  for  low  C/Ns,  the  slot  noise 
will  increase  more  than  usual. At  very  low  C/Ns  however,  we  can  expect 
slot  noise  to  be  lower  than  it  would  be  witi  hard  limiting.  The  slot 
noise  in  the  baseband  of  a M/W  receiver  varies  depending  on  the  amount 
of  limiter  action.  The  baseband  noise  spectrum  for  an  FM  receiver  with 
rectangular  IF  and  no  limiting  has  not,  to  th°  author's  knowledge,  been 
derived.  However,  Stumpers  and  Wang  have  derived  the  baseband  noise 
spectrum  for  a rectangular  IF  FM  receiver  with  hard  limiting.  Their 
results  indicate  that,  when  compared  with  the  Gaussian  IF  results  for 
low  carrier  to  noise  ratios,  baseband  noise  falls  off  more  quickly  with 
increasing  baseband  frequency  and,  for  the  intermediate  carrier  to  noise 
ratios,  the  baseband  noise  peaks  occur  at  lower  baseband  frequencies. 

The  overall  shapes  of  the  baseband  noise  spectra  c~e  similar. 

3.16  Using  the  results  of  Middleton,  Rice  and  Stumper's,  the  slot  noise 

for  hard  limiting  and  no  limiting  were  derived.  These  results  are  summarized 
on  the  next  page.  These  values  are  derived  from  truncated  infinite 


43 


RELATIVE  SLOT  MOISt  fOWER  (MILLIWATTS)  _ IELATIVE  SLOT  NOISE  POWER  (MILLIWATTS) 


£ 


Figure  8 


44 


Rectangular  Response  Gaussian  Response  Averaged  Response 

Hard  Ho  Hard  No  Hard  No 

Ll*itlng  Limiting  Limiting  Limiting  Limiting  Limiting 


§ 


« 

4-i 


HrwicoooNsfoonoti-jO'O 

■ ••••••••••••a* 

OOOOOOHHHHNNin<f-J 
I I I I I I I I I I I i I I 


oo<?»ooooooeoooo\H<n\oo‘o  o\ 

OHHNmsflfl«NONsfNO\H 


OOOOOOOOOHrHH 
I I I I I I I I I I I 


« 

* 

OptS-T'OO'JtMAOnC'C 

OOOOOi-4rHrHCMminr~ 

aaaaaaaaaaaa 

OOOOOOOOCjlpOO 


f-t  CM 

I I 


VO  in 
CM  m 


I I 


I I I I 


« 

OHHN-JvCflOH-JNHN-IOOn 

OOOOOOOrtH<NflM«\OI« 

aaaaaaaaaaaaaaa 

ooooooooooooooo 

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


* 

« 

8 


o o 


m O o 
csimr^ocNinr*. 
O O O /H  rH  rH 


o o o o 
in  o i/i  o 

CM  CM  Cl  CO  <t  sf  VO 


vi 


4) 

3 

rH 

3 

> 

O 

o 

a 

O 

n 

« 

>44 

A 

B 

I 

k 

55 

O 

4) 

.c 


o 

4J 


4) 

3 

rH 

« 

> 

s 

43 

•rl 

<0 

S3 

3 

4) 

o 

00 

a 

•H 

8 g 

i p. 
(0 
4) 
M 
U 
0 
u 


8 

55 

*x- 

O 


3 

3 

H 

4) 

> 

3 

41 

•H 

B 

a 

3 

14 

O 

00 

a 

• -H 
8*0 


43  43 

§£ 

o 

P.  3 
B 43 
43  JG 
U 4J 

U U 

43  43 

Us 

3 M 
3 00 


3 

0 

P. 

a 

43 

P 

M 

O 

u 

4) 

•3 


I 

I 

55 

U 

43 

§ 

O 

a 

SI 

II 

n 

4)  3 
rH  4) 
3 4J 
40  4) 
3 43 
3 P 
bO 


•a 

4J 

u 

3 

o pa 

3 m 

N 

3 TJ 

4a  *3 

Hfliooitn 

00  CO  00  VO 

oo  m 

r-s  in 

•H 

M 

■O  ov  cm  in  cr 

CH  no 

CO  rH 

cr  oo 

r*. 

rH 

rH 

3 >n 

a a a a a 

• • 

• s 

a a 

• • 

3 

3 CM 

J 3 CM 

OOHHH 

<s  cs 

(O  M M i/I 

vO  r* 

B 

X « 

4J  a 

1 1 1 1 1 

1 1 

1 1 

1 1 

1 1 

c 

4J  O 

rH 

3 

•>  B 

rH  B 

rH  *rt 

3 iH 

3 

3 

> 

43 

> 3 

3 3 

3 

3 3 

rH  3 

-P 

rH  rH 

H 

VO  O H Ov  r 

on  m o vo 

rH 

3 

3 3 

ffl^lMOONvJ 

r»  i-i 

oo  m 

3 

3 > 

4J  > 

• • 

a a 

a a 

a a 

> 

4J 

3 

O O O O rH 

rH  rH 

rH  CM 

CM  CM 

co  m 

3 

3 O 

H O 

1 1 1 1 I 

1 1 

1 1 

1 1 

1 i 

03 

rH  O 

o o 

O a 

B a 

B 

B O 

jo  a 

41 

JO 

3 

rH 

3 I 

1 

3 

3 

B 

3 m 

w » 

3 

M 

* 

04 

« 144 

* <44 

►v 

C3 

3 

4) 

3 

cr 

41 

P4  /-N 

4J 

no  3 

§43 
£ 

O 43 
4<  U 

S °* 

« M 
4) 
B mH 
> U 
U 

/-N  3 

*9  ° 

^ o 

3 

3 KS 
3 
•H 

<5 


O 

r4 

CO 


OOOOOOOOOOOOOOO 


Table  11 


4S 


series  summation  of  terms  related  to  the  IF  characteristic.  It  is 
interesting  that  for  the  hard  limiting  case,  the  series  converged  very 
slowly.  The  values  given  represent  250,000  terms  in  the  series.  This 
many  terms  were  needed  to  get  the  answer  accurate  to  within  a couple 
of  units  in  the  fourth  place.  These  results  agree  well  with  the  calculated 
values  of  Rice  and  Stumpers.  Although  the  attempt  was  made,  the  slot 
noise  distribution  for  no  limiting  and  no  carrier  present  could  not 
be  verified  experimentally. 

3,17Limiter  action  also  affects  the  demodulated  baseband  signal.  It 
appears  that  the  solution  for  an  arbitrary  baseband  signal  in  the  presence 
of  arbitrary  limiting  has  not  been  solved.  Middleton  has  determined 
the  effect  of  soft  limiting  on  the  suppression  of  a baseband  test  tone. 

His  results  indicate  that  as  limiter  action  goes  soft,  a given  amount 
of  signal  suppression  occurs  earlier  than  it  would  be  with  hard  limiting. 
Another  effect  he  shows  is  that,  as  limiting  is  lost,  minor  signal  sup- 
pression occurs  at  relatively  strong  RSLs. 


46 


^ . Experimental  Verification 

4.1  Having  established  the  basic  theoretical  characteristics  of  PM 
receivers,  a series  of  experiments  were  conducted  to  verify  the  theory 
and  investigate  effects  not  easily  determined  theoretically.  In  the 
experiments,  two  types  of  PM  receiver  were  used.  One  receiver  was  a 
typical  state-of-the-art  LOS  M/W  receiver.  This  receiver  was  designed 
for  normal  operation  with  typical  RSLs  of  -30  dBm  and  little  fading. 

The  characteristics  of  this  receiver  are  listed  on  the  next  two  pages. 

The  IF  response  data  was  courtesy  of  J.R.  Hammer,  G.L.  Cook,  and  R.J. 

Girvin.  Unless  otherwise  mentioned,  all  data  in  this  report  captioned 
"LOS  M/W  Receiver"  was  taken  using  this  receiver.  The  other  receiver 
used  was  a typical  TR0P0  M/W  receiver.  This  receiver  was  designed  to 
operate  with  a nominal  RSL  of  -60  dBm.  Due  to  the  fading  characteristics 
of  TROPO  RSL,  the  receiver  was  designed  to  operate  satisfactorily  with 
RSLs  approaching  FM  threshold.  The  characteristics  of  this  receiver 
are  listed  on  the  following  pages.  All  data  in  this  report  captioned 
"TROPO  M/W  Receiver"  was  taken  with  this  receiver. 

4.2  The  first  experiment  was  to  see  if  the  previous  baseband  noise 
spectrum  curves  for  hard  and  no  limiting  could  be  used  to  determine 
whether  or  not  the  limiter  in  a M/W  receiver  is  limiting.  Using  a 
spectrum  analyzer  attached  to  the  baseband  of  the  M/W  receiver,  various 
levels  of  unmodulated  RF  carrier  were  applied  to  the  receivers.  The 
baseband  noise  spectrum  as  a function  of  input  carrier  to  (thermal) 
noise  (C/N)  was  photographed  for  both  the  LOS  and  the  TROPO  M/W  receivers. 
The  photographs  (with  image  color  reversed  to  facilitate  reproduction) 
are  on  the  next  three  pages.  The  noise  spectrum  was  plotted  in  millivolts 
rather  than  milliwatts.  However,  since  power  is  just  voltage  squared, 

the  millivolt  display  is  just  compressed  in  vertical  dimension  relative 
to  a milliwatt  display.  The  shapes  of  either  type  of  display  is  identical. 
Comparing  the  photographs  with  the  "Hard  Limiting"  and  "No  Limiting" 
curves  leave  little  doubt  that  the  LOS  M/W  receiver  has  soft  limiting 
for  low  C/Ns.  However,  both  receivers  have  hard  limiting  for  high  C/Ns. 

4.3  Next,  the  theoretical  quieting  curves  were  compared  with  actual 
quieting  curves.  This  was  done  by  measuring  baseband  noise  with  a frequency 
selective  voltmeter  while  different  unmodulated  RF  signals  were  applied 

to  the  M/W  receiver.  Baseband  test  tone  suppression  was  also  measured 
by  applying  a test  tone  (of  frequency  0.05  times  the  IF  bandwidth)  to 
the  M/W  transmitter.  The  output  of  the  transmitter  was  applied  to  the 
M/W  receiver  through  an  attenuator.  Varying  the  amount  of  attenuation 
allowed  various  RSLs  to  be  simulated. 

4.4  The  agreement  between  theory  and  experiment  was  excellent  except 
for  very  low  C/Ns.  As  mentioned  in  the  theory  section,  with  no  carrier 
present  at  the  receiver,  the  slot  noise  will  be  a direct  function  of 

IF  limiting.  Clearly  neither  the  TROPO  nor  the  LOS  receiver  is  fully 
limiting  at  a -10  dB  C/N.  However,  the  TROPO  receiver  limiting  is  more 


47 


4& 


Channel  Capacity:  600  VF  Circuits 

Baseband  (BB)  Range:  60  to  2660  kHz 

Receive  Carrier  Frequency:  5 GHz 

De-emphasis:  None 

Per  Channel  Deviation:  140  kHz  (rms) 

Noise  Figure:  8.1  dB  (overall) 

RF  to  IF  Amplification: 

Overall  Gain:  +19  dB 

Linear  Amplification  Range: 

- oo  to  -17  dBm  RF  Input 

1/2  dB  Gain  Compression: 

-16  dBm  RF  Input 

1 dB  Gain  Compression: 

-15  dBm  RF  Input 

IF  Amplification: 

Maximum  Gain:  +40  dB 

3 dB  Bandwidth:  25  MHz 

LOS  M/W  Receiver  Characteristics 
Table  12 


49 


-0.5  0.0  +0.5 


A 


LOS  M/W  Receiver  IF  Response 
Figure  10 


50 


Frequency  (MHz) 

Relative  to  IF  Center  Frequency 


Channel  Capacity:  60  VF  Circuits 

Baseband  (BB)  Range:  12  to  252  kHz 

Receive  Carrier  Frequency:  400  MHz 

De-emphasis:  CCIR 

Per  Channel  Deviation:  100  kHz  (rms) 

Noise  Figure:  1.5  dB  (overall) 

RF  to  IF  Amplification: 

Overall  Gain:  +40  dB 

Linear  Amplification  Range: 

- oo  to  -50  dBm  RF  Input 

1/M  dB  Gain  Compression: 

-40  aBm  RF  Input 

1 dB  Gain  Compression: 

-37  dBm  RF  Input 

5 dB  Gain  Compression: 

-30  dBm  RF  Input 

15  dB  Gain  Compression: 

-20  dB  RF  Input 

IF  Amplification: 

Maximum  Gain:  +26  dB 

3 dB  Bandwidth:  3-09  MHz 


TROPO  M/W  Receiver  Characteristics 
Table  13 


52 


IF  Noise  BW:  3.286  MHz 
IF  3 dB  BW:  3,087  MHz 


\ ' 


curve  - Actual  IF  Response 
Q - Ideal  Gaussian  Response 


I 


Vertical:  10  dB/Div. 

| Horizontal:  2 MHz/Div. 

( 

I 

I 

TROPO  M/W  Receiver  IF  Response 
Figure  12 


€ 


53 


effective  (less  soft)  in  the  low  C/N  region.  This  is  indicated  by  the 
TROPO  slot  noise  values  being  closer  to  the  theoretical  curves  than 
are  the  LOS  noise  measurements.  This  condition  was  also  indicated  by 
the  previous  experiment.  Slot  noise  measurements  were  also  taken  with 
no  signal  into  the  M/W  receivers.  The  results  are  plotted  versus  theoretical 
results  on  the  next  page. 

4.5  In  addition  to  the  quieting  curves  of  the  normal  receivers,  the 
experiment  was  repeated  with  varying  degrees  of  limiter  action.  To 
reduce  the  limiting  action  in  the  receivers,  various  degrees  of  attenuation 
(attenuator  pads)  were  placed  between  the  receiver  mixer  (down  converter) 
and  the  input  to  the  IF  amplifier.  The  IF  amplifier  is  a variable  gain 
amplifier  which  tries  to  keep  the  input  level  to  the  limiter  section 
constant.  Placing  the  attenuation  (pad)  in  front  of  the  IF  amplifier 
causes  the  amplifier  to  increase  its  gain  to  compensate  for  the  pad 

loss.  As  attenuation  is  increased,  due  to  limited  gain,  the  IF  amplifier 
will  be  less  effective  in  holding  the  signal  constant  into  the  limiter 
section.  As  the  level  into  the  limiter  section  is  reduced,  limiting 
action  is  reduced  (goes  soft).  When  the  pad  is  placed  in  the  receiver, 
not  only  is  limiter  action  affected  but  noise  figure  is  also  increased. 
Compensation  was  made  for  changes  in  noise  figure  in  the  graphed  data. 

Also,  the  TROPO  receiver  data  has  been  corrected  to  delete  the  frequency 
response  of  the  CCIR  de-emphasis  network. 

4.6  Notice  in  the  following  curves  that  as  limiter  action  goes  soft, 
the  noise  for  very  low  C/Ns  is  reduced.  However,  the  noise  in  the  low 
frequency  baseband  slots  start  to  experience  additional  noise  near  FM 
threshold  (approximately  +10  dB  C/N).  For  moderate  amounts  of  limiter 
degradation,  about  the  only  effect  on  noise  in  the  FM  threshold  region 
is  to  cause  the  low  slot  noise  to  increase  more  rapidly  than  normal. 

In  extreme  cases,  however,  even  high  frequency  noise  slots  are  affected. 

Also,  note  that  the  LOS  receiver  is  significantly  more  sensitive  to 
pre-IF  signal  attenuation  than  is  the  TROPO  receiver.  The  previous 
data  has  also  been  plotted  relative  to  the  normal  configuration  data. 

This  shows  the  effect  of  both  limiter  and  overall  noise  figure  degradation. 

4.7  Since  the  data  indicated  that  the  limiter  of  the  LOS  M/W  receiver 
was  soft  at  low  C/N  ratios,  a wideband  amplifier  was  placed  at  the  input 
to  the  IF  amplifier  (the  same  place  the  pads  had  been  placed)  to  drive 

the  limiter  harder  for  the  low  C/Ns.  The  quieting  curve  was  reaccomplished 
and,  as  expected,  FM  threshold  moved  to  the  left  and  slot  noise  with 
no  carrier  present  was  increased.  The  following  two  pages  compare  theoretical 
and  measured  1 dB  FM  noise  thresholds  and  20  dB  noise  quieting  valves. 

4.8  Several  eourc#»  h«v«  indicated  that  if  a received  FM  signal  is  modulated, 
slot  noise  will  increase  (relative  to  slot  noise  produced  by  an  unmodulated 
carrier)  near  FM  threshold.  Middleton  indicated  that  the  slot  noise 

would  be  even  greater  without  limiting.  These  effects  were  investigated 
by  white  noise  loading  the  LOS  M/W  transmitter  with  a Noise  Power  Ratio 
transmitter  connected  to  the  baseband  input  to  the  transmitter.  The 


57 


Figure  16 


♦ 80 


OS* 


64 


+ 80 


n 

•4-i* 

C 

Oi 

• * j-  f 

7t 

w 

•r-i 

♦ » 

4^ 

4* 

7 

co 

o 

' 

00 

4-> 

c 

i> 

i. 

w 

m 

0) 

c 

+. 

> 

T3 

4-5 

0 

* 

cc 

4-5 

a, 

o 

r— 4 

cd 

VO 

• 

o 

fd 

a 

o 

u. 

— 

a 

•H 

h-t 

u. 

i 

2 C 

4-5 

V 

n 0) 

<D 

L 

ft 

1 

E-  > 

L 

a 

Ct 

•f 

W -H 

O 

•o 

4 

M <D 

a> 

arj 

V 

■f 

O 

x: 

•o 

oc 

o 

♦ 

O'  <D 

4-5 

u 

■f 

L 

K 

o 

to 

X 

CO 

1 

CVJ 

< 

I 

4 

n 2 

4 

M V. 

on 

{ 

4 

O X, 

a> 

t 

2 

s- 

V) 

-5  5 


66 


-130  -120 


t 


o ♦->  o 
o *h  l/j 
i »-*  o l 

c/>  z 


Figure  24 


67 


Figure  25 


68 


asp 


Figure  26 


69 


figure  27 


70 


120  -110  -100  -90  -80  RSL  -70  dBm  -60  -5 


issss;iKi55KSKs;;KaK:iKasa:ssMaa:a::»s:Ksxi»iiwgBi3«gK«!!t»»,:ttKngi; 


lBl< 

»! 


■Risen 
HlHSHStH 

|lia MM 

liti 


SUSSSSil 

MMih  iiii 


|lWHgjWU 

ISSSSSSSSSSS 


I5SXS5S255S! 


c 

c 

o 

•H 

o 

4-> 

•H 

03 

4-> 

O 

03 

■H 

3 

k 

3 

Cm 

•H 

> 

tsOr-l 

X 

•rH 

a 

B 

o 

c 

m 

o 

o 

o 

Du 

25  L 

M 

W 0) 

H 

(1) 

H > 

§ 

u 

W -H 

a 

M 0) 

£- 

o o 

o 

in 

O'  0) 

c 

• 

cs 

00 

w 

1 

<\J 

CO  » 
M \ 

n 

, 

o s: 

0) 

25 

> 

n 

CO 

L 

4-> 

ac  o 
E j 

D 

o 

O 

•o 

lililiil 

WNMIMIIIMI 


IIMlin 


r.imtJISHk 


HMiaMiMir>MaMwvR 


pxaia; 

'•HIM 


rjsrjsK 


ViiMnrijr/- 

r ^uMi  »/■  • if- 




aasKaca:asn4:^r:;;v.:.::iSsaa::: 

biS‘S5i§^!!£5a«KU;iSH=£S=; 

■ •••••••  Ml  SHI 

I— —■■■■«■— »-  j 

STo|: 

■ \s2BSSH 

iMMaaT  ,|SS32|Q| 

isssi  JSMSsnni 


IPPlii 

iawiPnwi 

MMiHiaiiaf 


Pi 

Sii 


N N M N N N N 

555335  £ 

KCOt/lOlAO  tS) 

^ M ifl  N h>  04 
v in  o (n  a 

N r-  00 


sj  gTl  BHfwUHgrWff  r~g  cTr?~~?z  n 

mm IS— ■■■*•■■*■! 


issssiJliiifei 


71 


2 dB  io  slot  noise  dB 
obtain  slot  noise  in  dBm0 


TROPO  M/W  Receiver 


f/B 

Gaus.  normal 
IF 

0.1 

7.9  7.5 

0.05 

8.8  8.5 

0.025 

9.6  9.5 

0.01 

10.4  ,9.5 

r Theoretical^ 

C/N 

(dB) 

20  dB 
pad 

40  dB 
pad 

60  dB 
pad 

7.5 

14.5 

12 

9.5 

15.5 

16 

9.5 

18.5 

20 

12.5 

22.5 

22, 

Measured 


LOS  M/W  Receiver 

C/N 

(clB) 


f/B 

Avg. 

IF 

30  dB 
gain 

normal 

10  dB 
pad 

20  dB 
pad 

0.1 

7.7 

8 

7 

7.5 

11.5 

0.05 

8 . 7 

9 

8 

10.5 

14.5 

0.025 

9.5 

11 

10 

12.5 

15.5 

0.01 

10,3 

,13  __ 

13 

13.5 

22.5  i 

^Theoretical  Measured 


FM  Noise  Threshold 


Table  14 


72 


TROPO  M/W  Receiver 


C/N 

(dB) 


f/B 

Gaus . 
IF 

normal 

20  dB 
pad 

40  dB 
pad 

60  dB 
pad 

0.1 

7.2 

8.5 

9.5 

15.5 

16 

0.05 

6.5 

9.5 

11 

18 

19 

0.025 

6. A 

9.5 

11.5 

18.5 

19.5 

0.01 

6.3 

t9.5.  . 

11.5 

18.5 

19.5  | 

I " 1 

1 Theoretical 

V 

Measured 

LOS 

M/W  Receiver 

C/N 

(dB) 

f/B 

Avg. 

IF 

30  dB 
gain 

normal 

10  dB 
pad 

20  dB 
pad 

0.1 

7.2 

10 

12.5 

18.5 

22.5 

0.05 

6.4 

12.5 

16.5 

24.5 

28.5 

0.025 

6.3 

12.5 

1? 

26.5 

31 

0.01 

6.2 

. 13 

17 

26.5 

31.5  , 

Theoretical1  Measured 


20  dB  Noise  Quieting 


Table  15 


73 


modulated  RF  signal  was  applied  to  the  LOS  receiver  using  a variable 
attenuator.  Basic  Intrinsic  Noise  Ratio  (BINR)  and  Noise  Power  Ratio 
(NPR)  measurements  were  taken  using  a Noise  Power  Ratio  receiver  connected 
to  the  baseband  of  the  M/W  receiver  (See  Tant's  book  for  an  explanation 
of  this  measurement.).  The  BINR  and  NPR  measurements  were  converted 
t-o  slot  noise  measurements.  Theoretical  results  were  determined  by 
using  Rice's  results  for  white  Gaussian  noise  loading  and  the  Anuff/Lou 
formula  for  required  IF  bandwidth.  The  results  were  calculated  for 
CCIR  loading  and  a baseband  to  IF  bandwidth  ratio  of  0.1.  The  additional 
assumption  was  made  that  the  additional  noise  due  to  Gaussian  baseband 
noise  loading  was  essentially  flat  spectrum  over  the  radio  baseband 
and  equal  to  the  noise  spectral  density  at  zero  frequency.  The  results 
are  tabulated  and  graphed  on  the  next  pages.  The  increased  slot  noise 
due  to  lack  of  limiting  can  be  observed. 

4tg  From  a theoretical  point  of  view,  it  has  been  shown  by  many  sources 
that  as  FM  threshold  is  approached,  the  baseband  noise  becomes  impulsive 

in  nharaoter.  It  is  not  at  all  obvious  whether  or  not  this  would  cause 
a change  in  the  impulse  noise  in  a narrow  baseband  slot.  In  other  words, 
does  the  impulse  noise  in  an  FDM  channel  become  higher  (relative  to 
the  idle  channel  noise)  as  FM  threshold  is  approached.  This  question 
was  approach<_d  experimentally  by  using  a frequency  selective  voltmeter 
to  convert  the  baseband  slot  noise  to  a 3-1  kHz  channel.  To  avoid  clipping 
the  noise  peaks,  the  FSV  amplifiers  were  operated  at  a very  low  level. 
However,  the  demodulated  baseband  noise  was  always  held  well  above  the 
residual  noise  In  the  frequency  selective  voltmeter  The  results  were 
spot  checked  running  the  frequency  selective  voltmeter  amplifiers  with 
10dB  more  gain.  The  results  were  also  checked  using  a differnt  type 
of  frequency  selective  volumeter  as  well  as  an  actual  frequency  division 
multiplexer  (AN/UCC-iJ).  The  results  were  essentially  the  same  in  all 
cases.  The  impulse  noise  was  measured  using  a conventional  impulse 
noise  measuring  set.  Demodulated  slot  noise  wa3  measured  using  a noise 
measuring  .set  with  no  noise  weighting.  Both  devices  had  IdB  bandwidths 
wider  than  5 kHz.  Therefore,  the  noise  measurement  bandwidth  was  always 
the  3-1  kHz  of  the  FSV. 

4.10  The  peak  value  of  an  impulse  is  difficult  to  define;  it  is  theoreti- 
cally infinite.  Impulse  noise  is  characterized  by  excursions  of  the 
total  noise  instantaneous  voltage  waveform  which  are  much  greater  than 
the  rm3  voltage  (average  power)  level  of  the  noise.  The  definition 
of  peak  level  was  defined  for  purposes  of  this  report  as  the  minimum 
dB  voltage  level  above  the  average  power  level  in  the  3-1  kHz  noise 
slot  (for  the  given  C/N)  which  would  yield  no  indication  on  the  impulse 
level  measuring  set  when  observed  for  a period  of  thirty  seconds.  The 
choice  of  peak  level  was  arbitrary.  If  the  level  had  been  chosen  as 
the  minimum  level  which  would  cause  no  more  than  one  level  indication 
every  5 seconds,  the  peak  level  measured  would  have  been  about  2 dB 
lower  than  the  level  measured  using  the  above  definition.  The  level 
indications  were  essentially  continuous  at  a level  dB  lower  than  the 
level  measured  using  the  above  definition.  Reference  levels  on  the 


74 


C/N 

(dB) 

0.0 

Theoretical 

f/B 

0.0028  0.05 

0.1 

Measured 

f/B 

0.0028  0.05  0.1 

“CO 

0 

0 

0 

0 

0 

0 

0 

0 

0.2 

0.2 

0.2 

0.2 

0 

1 

0 

+2 

0.2 

0.2 

0.2 

0.2 

1 

1 

0 

+4 

0.4 

0.4 

0.4 

0.3 

2 

1 

1 

♦5 

0.5 

0.5 

0.5 

0.4 

1 

1 

1 

+6 

0.6 

0.6 

0.5 

0.4 

1 

1 

0 

♦7 

0.7 

0.7 

0.5 

0.3 

3 

2 

1 

+8 

0.8 

0.8 

0.4 

0.1 

3 

0 

0 

+9 

1.0 

1.0 

0.2 

0 

3 

0 

0 

♦ 10 

1.2 

1.1 

0 

0 

3 

0 

0 

♦ 11 

1.4 

0.7 

0 

0 

3 

0 

0 

♦ 12 

1.7 

0.1 

0 

0 

4 

0 

0 

+13 

1.9 

0 

0 

0 

- 

- 

- 

♦ 14 

2.3 

0 

0 

0 

- 

- 

- 

♦ 15 

2.6 

0 

0 

0 

- 

- 

- 

Baseband  Slot  Noise  Increase  (dB) 
due  to  CCIR  Gaussian  White  Noise 
Baseband  Loading 


Table  16 


75 


76 


Note:  Add  1 dB  to  slot  noise  dB 
value  to  obtain  slot  noise  in  dBm0. 


impulse  measuring  equipment  had  a minimum  step  size  of  2 dB.  This  led 
to  some  granularity  in  the  measured  results.  The  experimental  results 
indicate  that  impulse  noise  peak  to  rms  voltage  levels  in  narrow  slots 
are  completely  independent  of  RSL  or  C/N. 

4.11  It  will  be  noticed  that  the  impulse  noise  peaks  were  higher  in 

the  lowest  frequency  slots  than  in  the  others.  After  taking  the  measure- 
ments, the  frequency  stability  of  the  generator  used  to  simulate  the 
RSL  was  checked  on  a frequency  counter.  The  signal  was  observed  to 

Jump  around  the  nominal  center  frequency  as  much  as  50  KHz.  The  signal 

generator  was  replaced  by  the  actual  microwave  transmitter.  The  actual 
microwave  transmitter  was  frequency  stabilized  and  showed  no  frequency 
jitter  at  a1!.  Tne  results  were  checked  in  the  two  low  frequency  base- 
band slots.  Under  these  conditions  the  results  were  identical  to  those 
of  the  other  slots.  It  is  believed  that  the  unusually  high  peak  noise 

readings  in  the  low  slots  were  caused  by  the  frequency  jitter  of  the 

CW  generator  producing  occasional  high  level  low  frequency  noise  spikes 
(beat  products)  in  the  radio  baseband. 

4.12  The  next  experiment  was  to  investigate  how  closely  a quieting  curve 
taken  with  a typical  continuous  wave  (CW)  M/W  signal  generator  compares 
with  a quieting  curve  taken  with  an  actual  M/W  transmitter.  The  following 
page  shows  the  comparison.  The  M/W  transmitter  used  to  make  the  comparison 
had  a low  power  (J  watt)  output  and  a high  power  (5  watts)  output. 

The  high  power  output  was  just  the  low  power  output  amplified  by  a Traveling 
Wave  Tube  (TWT ) amplifier.  The  quieting  curve  was  taken  separately 
using  each  output  as  the  quieting  source.  There  was  no  difference  in 
the  quieting  curves.  This  was  interesting  (but  predictable)  since  TWTs 
operating  in  the  frequency  range  of  the  LOS  receiver  ( 5 gHz)  have  noise 
figures  of  typically  25  dB.  When  the  quieting  curves  are  compared, 
it  will  be  noticed  that  the  only  significant  difference  is  in  the  high 
C/N  region.  This  is  indicative  of  the  frequency  noise  characteristics 
of  the  two  quieting  sources. 

4.13  Sweeping  generators  (even  in  the  CW  mode)  are  notorious  for  their 
poor  phase/frequency  noise  performance.  Quieting  curves  were  taken 
using  two  different  sweeping  microwave  oscillators  (data  courtesy  of 
H.L.  Bennetts).  Notice  the  wide  range  of  possible  values  for  slot  noise 

in  the  high  C/N  region  when  different  sources  are  used.  Sweeping  generator 
B (the  up  convertor  for  the  HP  microwave  link  analyzer)  gave  the  best 
performance  of  any  of  the  generators  tested.  In  the  high  C/N  region, 
its  noise  performance  was  as  much  as  5 dB  better  than  the  noise  performance 
for  comparable  conditions  with  the  actual  microwave  transmitter.  The 
quieting  curves  illustrate  the  futility  in  using  a quieting  signal  gene- 
rator of  unknown  frequency  noise  performance  to  evaluate  the  high  C/N 
slot  noise  performance  of  a M/W  receiver.  The  inherent  noise  of  the 
receiver  is  sometimes  less  than  the  noise  of  the  quieting  source. 

4 . 14  Next , quieting  curves  were  taken  using  wide  slots.  The  two  relatively 
narrow  bandwidth  slots  were  formed  using  NPR  test  set  filters.  The 


77 


r—i  tO 

o x in  in  in  in  in 

o x • • 

♦ vo^Dinvoa^oooinro 

O in  r—1r~4f-ir—4f—ir-',r-i  <\J  Oj  Oj 

OJ 


N 

in  x 

c\j  x in  m 10  in  in  in  in 

o 

o in  lOvoin^in^t^c^  nco 

O <\J 
VO 


N 

x in  in  in  in  in 

r-l  • • • • • 

o vnvOvcininiTivnioinin 

•O  HHHrHHHHHHH 

o in 

C\J 


\ 


N 

in  x in  in  in  in  in  in  in  in 

C\J  jkl  • • • •*••• 

o vovntninint^nininin 

• in  rHrHr-HrHrHiHfHr-H'HrH 
O 00 
vD 


in  x in  in  in  in  in 

o ? 

♦ in  vo  id  in  io  in  in  in  in  vo  vo 

O 00  HHHrHHHHHHH 
rH 


in 

00 


in  in  in  in  in 

vO  -=T  mOvOvOvO'DiDiO 
«^HrHrHrHrHrH»-Hr-HrHrH 


^ 300  0 000000 

_)  s cuOCT>ooo~vnin.=rronj 

c o fS  r~i  • — < 

XT?  1111)11111 


Table  17 


co 

05 

* V 

co  j 
3 co 

« 05 
t. 

<0  c 
> 0) 
> 

4->  -H 

O tt) 


CO 

■o 

c 

nj 

•O 

0) 

co 

<2 


N O 
X C 
X 

c 

m -h 

o a 2 

o 

C M 


co 


ffl 

■o 


<1) 
O 

c 
<u 

L 
0)  CD 

a.  a 
<v 
s~ 


x 

n) 


0) 

co 

•H 

o 

z 

0) 

CO 


« 

3 TJ 

a 

a o 


no. 


Figure  30 


Figure  29 


Figure  31 


80 


Figure  29 


wide  bandwidth  slot  was  formed  using  the  baseband  low  pass  filter  in 
the  M/W  receiver  itself.  The  noise  was  measured  using  a wide  bandwidth 
true  rms  meter.  The  various  slot  measurements  are  plotted  by  themselves 
and  also  versus  the  normal  narrow  slot  quieting  curves.  For  comparison 
with  the  quieting  curve  data,  10  log  (filter  bandwidth  (kHz)  / 3>1kHz) 
was  subtracted  from  the  wide  slot  data.  FM  threshold  can  be  determined 
for  the  wide  slots  by  integrating  the  narrow  slot  noise  formula  over 
the  frequency  range  of  the  wide  slot.  As  noted  earlier,  this  leads 
to  a result  which  indicates  an  FM  threshold  for  the  widest  slot  should 
occur  about  one  dB  earlier  than  FM  threshold  for  the  2438  kHz  narrow 
slot.  The  experimental  evidence  shows  reasonable  agreement  with  that 
result. 

4.15  After  observing  the  noise  characteristics  of  the  M/W  receiver, 
the  next  set  of  experiments  observed  the  baseband  signal  suppression 
characteristics  of  the  M/W  receiver.  The  modulated  RF  signal  was  then 
applied  to  the  LOS  M/W  receiver  through  a variable  attenuator.  The 
test  tone  suppression  as  a function  of  C/N  (RSL)  was  observed  using 

a frequency  selective  voltmeter  (FSV)  at  the  baseband  of  the  receiver. 
Since  the  noise  power  in  the  FSV  slot  was  significant  at  low  C/N  ratios, 
the  actual  FSV  measurement  was  test  tone  plus  noise  power  rather  than 
just  test  tone  power.  To  correct  for  this,  a FSV  measurement  was  taken 
with  the  test  tone  applied  to  the  transmit  baseband.  Another  measurement 
(of  noise)  was  made  with  the  test  tone  removed  from  the  baseband.  By 
subtracting  the  noise  power  (in  milliwatts)  from  the  test  tone  plus 
noise  (in  milliwatts)  the  FSV  measurements  were  corrected  for  the  effects 
of  noise.  The  test  tone  levels  used  were  less  than  normal  CCXR  loading 
to  avoid  introducing  significant  increases  in  receiver  slot  noise  due 
to  the  modulation  of  the  received  signal. 

4.16  The  first  experiment  was  to  observe  signal  suppression  at  three 
baseband  frequencies.  The  three  frequencies  were  0.0028,  0.05,  and 

0.1  times  the  IF  bandwidth.  The  0.05  frequency  had  the  least  suppression 
of  any  of  the  baseband  test  tone  frequencies.  The  performance  of  the 
other  test  tones  relative  to  the  performance  of  the  0.05  test  tone  has 
been  tabulated  on  the  next  page. 

4.17  The  next  experiment  was  to  observe  signal  suppression  as  a function 
of  baseband  loading.  Using  an  NPR  test  set  white  noise  generator,  CCIR 
recommended  values  of  white  noise  loading  were  applied  to  the  transmitter 
baseband  to  simulate  normal  baseband  loading.  Using  the  NPR  test  set 
bandstop  filters,  noise  was  suppressed  at  three  different  frequencies. 

Test  ton**  were  placed  in  the  transmitter  baseband  at  those  frequencies. 

The  test  tone  suppression  was  measured  with  no  white  noise  loading  and 
was  then  measured  with  full  noise  loading.  The  difference  in  the  baseband 
suppression  for  the  two  conditions  is  shown  in  Tsbls  18* 

4.18  Next,  the  baseband  test  tone  suppression  values  obtained  during 
the  previous  experiments  is  plotted  on  an  expanded  vertical  scale. 


81 


♦80 


o 

to 


o 


o 

m 


o 

vO 


o 


e 

co 

TJ 

o 

00 


CO 

cc 

o 

c> 


o 

o 


Figure  33 


83 


See  note.  Figure  28 


C/N 

(dB)  f/B 

0.0028  0.1 
-3  -0.5  -0.6 

-2  -0.9  -0.7 

-1  -0.5  -0.9 

0 -0.8  -1.0 

♦ 1 0.0  -0.6 

+ 2 -0.8  -0.8 

♦3  -0.3  -0.6 


Baseband  Signal  Suppression  (dB)  at  High 
and  Low  Baseband  Frequencies  Relative  to 


Signal  Suppression  at  the  Middle 
Baseband  (reference  test  tone  at 
ized  baseband  frequency  of  f/B  « 

of  the 
normal - 
0.05) 

C/N 

(dB) 

f/B 

0.0028 

0.05 

0.1 

-3 

0.0 

0.0 

0.0 

-2 

0.0 

0.0 

0.0 

-1 

0.0 

-0.5 

0.0 

0 

0.0 

-0.4 

0.0 

♦ 1 

0.0 

-1.0 

0.0 

♦ 7 

0.0 

0.0 

0.0 

♦ 3 

0.0 

0.0 

0.0 

Baseband  Signal  Suppression  (dB)  with 
CCIR  Gaussian  White  Noise  Baseband 
Loading  Relative  to  Signal  Suppression 
with  no  Baseband  Noise  Loading 


Table  18 


84 


The  graphs  on  the  next  two  pages  show  the  effect  of  changes  in  limiter 
action.  The  C/N  values  have  been  corrected  to  compensate  for  changes 
in  receiver  noise  figure  due  to  the  introduction  of  prelF  attenuation. 

Note  that  as  limiter  action  goes  soft,  baseband  suppression  starts  at 
earlier  C/N  values.  Also,  notice  that  the  LOS  receiver  is  much  more 
sensitive  to  prelF  signal  loss  than  is  the  TROPO  receiver. 

4.19  In  elementary  developments  of  the  theoretical  noise  of  FM  receivers, 
one  of  the  assumptions  is  that  the  received  signal  is  centered  in  the 
passband  of  the  receiver.  The  frequency  response  of  the  passband  is 

also  assumed  to  be  symmetric  about  the  carrier  frequency.  The  unmodulated 
receive  carrier  signal  was  deliberately  tuned  to  one  side  of  the  receiver's 
IF  response  and  quieting  curves  were  accomplished.  Data  was  taken  in 
the  2438,  1248,  and  70  kHz  slots  and  is  graphed  on  the  next  page.  The 
data  reflects  an  extreme  violation  of  the  previous  assumptions.  It 
can  be  anticipated  that  receiver  idle  noise  performance  will  be  degraded 
if  the  received  signal  is  not  centered  in  the  receiver's  (IF)  passband 
response  or  if  the  receiver's  (IF)  passband  frequency  response  is  unsymmetric 
relative  to  center  frequency. 

4.20  Most  theoretical  developments  of  receiver  signal  and  noise  performance 
assume  that  the  noise  in  the  receiver  is  strictly  thermal.  The  noise 
power  is  assumed  to  be  spread  equally  across  the  bandpass  response  of 

the  receiver.  This  is  not  always  the  case  in  practice.  The  most  elementary 
form  of  nonthermal  noise  is  sine  wave  interference. 

4.21  The  next  few  graphs  show  the  signal  and  noise  characteristics  of 
an  FM  M/W  receiv'  ' for  different  levels  and  frequencies  of  unmodulated 
carrier  interfei  ence . The  baseband  test  tone  signal  frequency  was  0.05 
times  the  receiver  IF  bandwidth.  The  tests  were  run  by  mixing  the  outputs 
of  the  unmodulated  M/W  transmitter  and  a CW  signal  generator.  The  composite 
signal  was  applied  to  the  input  of  the  M/W  receiver.  Separate  attenuators 
allowed  the  outputs  of  the  two  signal  sources  to  be  varied  independently,,. 

The  output  of  the  M/W  transmitter  was  considered  the  carrier  and  the 
output  of  the  CW  generator  was  called  the  interference.  The  carrier 
source  was  tuned  to  the  center  frequency  of  the  receiver.  Results  were 
obtained  with  the  interference  tuned  to  two  different  frequencies. 

For  one  set  of  data,  the  interfering  signal  was  tuned  to  a frequency 
differing  by  i MHz  from  the  receiver  center  frequency.  This  frequency 
offset  was  chosen  so  that  the  interfering  signal  would  be  near  the  carrier 
frequency,  but  none  of  the  beat  products  caused  by  crossmodulation  products 
(beat  products)  in  the  receive  baseband  (due  to  the  two  RF  input  signals) 
would  fall  into  the  noise  measurement  slot  skirts.  Also,  the  large 
increase  in  baseband  noise  when  two  RF  carriers  have  essentially  the 
same  frequency,  but  randomly  varying  relative  phase  was  avoided.  The 
other  set  of  data  was  taken  with  the  Interfering  signal  offset  10  mHz 
from  the  receiver  center  frequency.  This  put  the  interfering  signal 
at  the  edge  of  the  receiver's  IF  passband.  Although  the  interfering 
signal  with  the  larger  frequency  offset  caused  more  baseband  noise, 
a spectrum  analyzer  sweep  of  the  radio  baseband  would  not  have  shown 


85 


86 


sc  <u 

■H  i-l 
C/5 


Figure  35 


87 


08+  OL+  09*  OS*  Ofr*  0£*  8P  OZ*  N/D  01*  0 01-  Oi-  0£* 


FM  NOISE  QUIETING  CURVES 
LOS  M/W  Receiver 


88 


See  note,  Figure  28 


the  beat  product  since  it  would  appear  at  a frequency  much  higher  than 
the  cutoff  frequency  of  the  receiver  baseband  low  pass  filter. 

4.22  The  first  set  of  graphs  were  taken  by  holding  the  interfering  signal 
power  at  -60  dBm.  When  the  C/I  ratio  approaches  +10  dB,  the  signal 

and  noise  properties  become  noticeably  affected.  Of  course,  if  baseband 
slot  noise  had  been  observed  at  integer  multiples  of  the  frequency  difference 
between  the  carrier  and  interference,  the  noise  increase  would  have 
been  more  distinct  and  would  have  been  noticed  at  much  larger  C/I  ratios. 

The  capture  effect  of  the  receiver  is  dramatic  as  it  starts  to  quiet 
on  the  interference  rather  than  the  carrier.  The  second  set  of  graphs 
were  taken  with  the  interfering  signal  level  held  at  -90  dBm.  The  third 
set  of  graphs  were  taken  with  the  interference  level  varied  with  carrier 
level  variation  so  that  the  interference  was  always  10  dB  less  than 
the  carrier  level.  Notice  that  in  all  cases  the  slot  noise  was  noticeably 
higher  and  baseband  signal  suppression  earlier  for  signals  farther  away 
from  the  center  of  the  receiver  IF  response. 

4.23  The  last  set  of  interference  data  was  taken  by  holding  the  interfering 
signal  power  constant  but  varying  its  frequency  relative  to  receiver 
center  frequency.  The  receiver  had  an  RF  filter  as  well  as  an  IF  filter. 

The  RF  filter  frequency  response  was  broad  enough  that,  for  the  interfering 
signal  frequencies  used  for  this  test,  the  frequency  selectivity  charac- 
teristics of  the  receiver  can  be  assumed  to  be  entirely  due  to  the  IF 
frequency  response  of  the  receiver.  Clearly  the  noise  process  is  complex 
and  a function  not  only  of  C/1  and  interference  frequency  offset  but 

also  baseband  noise  slot  frequency. 

4.24  It  should  not  be  assumed  that  the  performance  of  all  FM  receivers 
is  the  same.  To  constrast  the  previous  results,  tests  were  performed 
on  another  LOS  M/W  receiver.  It's  nominal  equipment  characteristics 
were  quite  similar  to  the  characteristics  of  the  LOS  M/W  receiver  used 
to  perform  the  previous  experiments.  A quieting  curve  for  the  receiver 
is  given  on  the  following  page.  Note  the  abnormally  high  noise  in  the 
low  frequency  noise  slots.  Also,  note  the  kink  in  the  curves  near  the 
-80  dBm  RSL.  During  the  experiment  it  was  noted  that  the  noise  measure- 
ments in  this  region  were  unusually  erratic.  Although  it  is  difficult 
to  see  on  the  quieting  curve,  two  of  the  low  frequency  noise  slots 
actually  cross  each  other,  a theoretical  impossibility. 

4.25  The  next  page  shows  spectrum  analyzer  photographs  of  the  baseband 
noise  spectrum  for  various  RSLs.  A comparison  with  the  corresponding 
photos  for  the  other  LOS  receiver  shows  considerable  difference.  The 
center  photograph  highlights  the  noise  instability  for  RSLs  near  the 
kinks  in  the  quieting  curve.  Notice  the  crisscrossing  of  the  baseband 
spectrum.  When  this  photo  was  taken,  the  noise  measurements  were  so 
unstable  that  the  same  picture  could  never  be  reproduced.  Every  trace 
of  the  spectrum  analyzer  prruuced  a radically  different  picture.  This 
was  considerably  different  tnan  the  case  for  the  LOS  receiver  used  in 


89 


the  rest  of  the  test  for  this  report.  Results  taken  with  the  AN/FRC-  ' 

157  receiver  were  repeatable  day  after  day.  The  results  taken  on  the 
other  receiver  changed  from  moment  to  moment.  Not  even  the  baseband 
suppression  characteristics  remained  constant  for  the  other  LOS  receiver. 
The  results  of  tests  conducted  at  two  different  tiroes  are  shown  on  the 
following  page. 

4.26  The  FM  quieting  curve  is  a sensitive  indicator  of  noise  performance 
of  an  FM  M/W  terminal.  The  following  quieting  curves  were  taken  of 
FM  M/W  receivers  in  actual  operational  environments.  Although  it  is 
not  always  clear  what  the  problem  is,  the  quieting  curves  do  show  that 
there  obviously  undesirable  conditions  at  the  radio  terminal.  Occasionally 
the  quieting  curves  also  shows  measurement  error  and  measurement  equipment 
problems. 


90 


4) 

O *-*  ^ O 

'C  O i/> 

I •“«  O i 

m 7. 


Figure  38 


92 


Figure  28 


Figure  39 


93 


Figure  28 


94 


See  note.  Figure  28 


laSssaai^^^^MHa^BssfflgsHKgnSSs  Hsl 


la» 


iisissa 

MBBaaMiKataasaBiiiiiiBl 

iHiMNiiMimNmiMMrjaaiSiii 

laMfcaaaiiaiasaaa 5#8Sf«8— ii 

llll|H|nMNiiiiiiNnrj«Mf4«n 

• •••••*•••■'-••  mmm  *mm  ■•  mm  m immm 

bieiinMaM%4s«iV44a«itia««««fHn 

m m\  ;:r 

ajiMMiii  mmm'*mmm»m'*  ■■■■>■■— laiial 
MMM—— ^ ■■■•■rvFJ 

NmuimMunrj/  i 
[**•  '.•#!»#•'>  Bn*  wia»iMiaaM>f.ir4  fl 


lar*.  4«— i^^aiwaaaa^aiaiil 


mmmmmw  • ’*mmm 


.w~* 

niawaM a— aaaaaa 

:s:5^sssar:.^ 

It  .■■■•■•Sfii  ipM8nar'«aflK«i8i*'2aaHawHnl 

acSr'^taHSKSWMi 


%sr 

Ssai 

s:::i 


!9S1^S»S!SSS!!5!SS5«M 
sans  asaatis  atsa:3  K 
■is»ca:sisa:» 


. faSitsss: 

■ ixssm* 

Mmmmmmm*  ^MMaaaaaaaaM 

Inimj 


HUBS! 

asm 


|nmu»  > aaaa 

1 5 mm a mm 

|::nt::ni 


aaottil 

IsSSil 
■Pmuii 

t t a1sjj^s»l 

■SSSmSSSSI 

HSH555B555SI 


SSI 


sis 


95 


96 


-130 


4 


Slot  Frequency 
f/B-0.1 
02.5  MHz) 


Slot  Frequency 
f/B-0.05 
(f-1.25  MHz) 


C/1  - + 10 


C/1-  0 


C/I  - -10 


C/1  - -20 


C/1  - -30 


C/I  - -40 


i C/I  - -50 

I — , — ^ — T I , — j— ( 1 \ 1 — — I 1 

\ T j 0 dB  noise  reference  \.  / 

V Ji  / is  slot  noise  with 

I no  interfering  carrier.  11^ 

Vertical  Scale  : Slot  Noise  Horizontal  Scale  : ’ Tiffset 
(10  dB/div.,  + above  line)  (10  MHz/Plv./ 

Intersection  of  lines  is  zero  on  both  scales. 

LOS  M/W  Receiver 

Quitting  Carrier  Power  (C)  -60  dBm 
Interfering  Carrier  Power  (I)  Variable 
(Beat  products  neglected  on  right  hand  plots.) 

Baseband  Slot  Noise  Va  Interfering  Carrier  Frequency  and  Power 


Figure  43 


97 


LOS  M/W  Receiver 
no  de-emphasis 


Vertical:  mV 
Horizontal:  5 MHz/Div. 
(0  to  50  MHz) 


Vertical:  mV 
Horizontal:  200  kHz/Div 
(0  to  2 MHz) 


Note:  This  LOS  M/W 
receiver  is  not  the 
one  used  i;  ;he  rest 
of  this  report. 


Vertical:  mV 
Horizontal:  500  kHz/Div. 
C 0 to  5 MHz) 


Baseband  Noise  Spectrum 
for  Various  RSLs 


100 


-30  -20  -10  0 +10  RSI  +20  dB  +30  +40  +50  +60  +70  +80 


FM  RECEIVER  NOISE  Of iETING  AGO  OR  I GNAL. 
AND  IF  LIMITING  DATA 


RECEIVER  INPUT  RF  SIGNAL  LEVEL  (RSL)  - dBm 


Sample  Noise  Quieting  Curve 


Figure  48 


102 


SIGNAL  LEVEL  (RSL)  ( - dbm  ) 


Sample  Noise  Quieting  Curve 
Figure  49 


103 


Sample  Noise  Quieting  Curve 
Figure  50 


104 


SIGNAL  LEVEL  (RSL)  (-dbm) 


DATA  SHEET  B . fk  ACC  & FM  C .INC  CURVES  **cc 


Sample  Noise  Quieting  Curve 


Figure  51 


105 


dtu) 


Sample  Noise  Quieting  Curve 
Figure  53 

107 


DATA  SHEET  t-*V  ACC  • FM  QUIETING  CUEVES 


Sample  Noise  Quieting  Curve 
Figure  54 


108 


'HWT  Vf  Sjc.ka  irv‘i  . 


5.  Predicting  FM  M/W  Terminal  Thermal  Noise  Performance 
Symbols  Used  on  the  Following  Pages: 

RSL  = reoelved  signal  level  (dBm)  into  receiver 

NF=  overall  noise  figure  of  receiver  (dB)  measured  at  the  same  place 
at  which  RSL  is  measured 

BTt?=  3 dB  bandwidth  of  the  receiver  IF 
lr 

f = baseband  frequency  of  the  center  of  the  narrow  slot  used  (by  a frequency 
selective  voltmeter)  to  measure  receiver  baseband  noise 

f/B  = f/BT[?  (with  f and  BTI?  converted  to  same  units  (e.g.,  kHz  or  MHz)) 

lr  lr 

= baseband  slot  frequency  compared  (normalized  or  ratioed)  to  the 
receiver  IF  bandwidth 

H = IF  frequency  response  (dB)  at  IF  center  frequency  plus  or  minus 
baseband  slot  frequency  (f)  minus  IF  frequency  response  (dB)  at 
IF  center  frequency 

= normalized  IF  response  of  the  receiver  which  shapes  the  receiver 
thermal  noise 


= normalized  thermal  noise  in  a narrow  slot  at  output  of  IF 
h = normalized  IF  response  of  receiver  (PR) 

= antilog  (H/ 10)  = 10  H/1° 

ftnax  = maximum  baseband  frequency  (excluding  baseband  pilots) 

C/N  = carrier  to  noise  ratio  (dB) 

=power  level  of  received  signal  compared  to  total  noise  power  of 
receiver’s  internally  generated  thermal  noise  which  appears  at 
the  discriminator 


p = carrier  to  noise  ratio  (PR) 
= antilog  ( (C/N)/ 10)  = 10 


(C/N)/ 10 


f = baseband  pivot  frequency  for  the  emphasis  network 
S = baseband  test  tone  signal  power  level  (dBm0)  at  the  receiver 
S = baseband  test  tone  signal  power  level  (dBm0)  at  the  transmitter 

Xj 


109 


N = noise  power  (dBm0)  in  a 3.1  kHz  slot  located  at  a center  frequency 
f in  the  receiver  baseband 

No  = noise  power  (dBm0)  in  a 3-1  KHz  slot  located  at  a very  low  (approxi- 
mately DC)  receiver  baseband  frequency  when  no  carrier  is  applied 
to  the  input  of  the  receiver 

P = baseband  pre-emphasis  (dB)  relative  to  baseband  pivot  frequency 

Pp  = baseband  pre-emphasis  (dB)  at  baseband  pivot  frequency  relative 

to  baseband  gain  (dB)  at  pivot  frequency  with  pre-emphasis  strapped 
out  of  the  baseband  circuitry 

X = emphasis  fime  constant 

Af,  , = per  channel  rms  deviation  (kHz) 

/ch  rms 

= rms  deviation  caused  by  a 0 dBm0  sine  wave  test  tone  (at 
baseband  pivot  frequency  jf  emphasis  is  used) 

PR  = power  ratio  (e.g.,  milliwatts/milliwatts) 

dB  = 10  log  (PR) 

dBm0  = see  next  page 

dBm  = 10  log  (power  in  milliwatts) 

TLP  = see  next  page 

log  (x)  = common  (Bragg's  or  base  10)  logarithm  of  the  number  x 


ir  = pi  « 3. 1416 


e = base  of  natural  (Naperian)  logarithms  = 2.7183 


Noise  Power  Normalization: 

5,1  In  the  pages  which  follow,  all  power  levels  will  be  given  in  arti- 
ficial dBm0  values.  The  desirability  of  the  dBm0  values  is  that  they 
allow  direct  comparison  of  different  equipment.  In  communication  systems, 
power  is  often  measured  in  absolute  power  referenced  to  one  milliwatt. 

This  unit  is  the  dBm.  To  convert  from  dBm0  to  dBm  values  and  vice  versa, 
use  the  following  formulas: 

dBra0  = dBm  - TLP(dB) 

dBm  = dBra0  + TLP(dB) 


no 


where  TLP  is  the  Transmission  Level  Point  in  dB  associated  with  the 
test  point  at  which  the  measurement  is  made. 


Receiver  Baseband  Signal  Level: 


S(dBm0)  = St(dBm0)  + 20  log  (1.0  - e'P) 

C/N(dB)  = RSL(dBm)  + 114.0  - NF(dB)  - 10  log  BIp(MHz) 


P(PR)  = antilog  ( ( C/N  )/ 10)  = 10 


( C/N ) / 1 0 


Receiver  Baseband  Noise  in  3*1  KHz  slot: 

For  C/N  less  than  -10  dB: 

N(dBm0)  = No  + (IF)  - 4.34p 

where  the  "IF"  factor  is  obtained  from  the  graph  on  the  following  page. 
The  formula  is  listed  below  for  certain  values  of  f/B.  The  value  of 
No  is  obtained  using  the  appropriate  formula  listed  below. 

Rectangular  IF 


f/B  - 0.0 

N(dBm0) 

= 

No 

- 4.34p 

max. 

error 

= 0.23dB 

f/B  = 0.1 

N(dBm0) 

= 

No 

- 0.71  - 

4.34p 

max. 

error 

= 0.17dB 

f/B  = 0.2 

N(dBm0) 

= 

No 

- 1.43  - 

4.34p 

max. 

error 

= 0.06dB 

f/B  = 0.3 

N(dBm0) 

= 

No 

- 2.15  - 

4.34p 

max. 

error 

= 0. 12dB 

f/B  = 0.4 

N(dBm0) 

= 

No 

- 2.86  - 

4.34p 

max. 

error 

= 0. 39dB 

f/B  = 0.5 

N(dBm0) 

= 

No 

- 3.54  - 

4 . 34o 

max. 

error 

= 0.74dB 

No(dBm0)  = +29 

.76  - 20 

l0®  Af7ch  rms 

(kHz)  + 10  log  Bjf(MHz) 

Gaussian  IF: 

f/B  = 0.0 

N(dBm0) 

= 

No 

- 4 . 34 p 

max  error  = 

0. 12dB 

f/B  = 0.1 

N(dBm0) 

= 

No 

- 0.03  - 

4.34p 

max. 

error 

= 0.1 IdB 

f/B  = 0.2 

N(dBm0) 

= 

No 

- 0.14  - 

4.34p 

max . 

error 

= 0.06dB 

f/B  = 0.3 

N(dBra0) 

= 

No 

- 0.31  - 

4 . p4p 

max. 

error 

= O.OOdB 

f/B  =0.4 

N(dBm0) 

= 

No 

- 0.54  - 

4.34p 

max . 

error 

= 0.07dB 

111 


f/B  = 0.5 


N(dBm0)  = No  - 0.83  - 4.34p  max.  error  = 0.l8dB 


No(dBm0)  = +29.51  - 20  log  Af.  . (kHz)  + 10  log  BTt,(MHz) 

/ch  rms  Ir 

Averaged  IF: 

f/3  = 0.0  N(dBm0)  = No  - 4.34p  max.  error  = 0.l8dB 

f/B  = 0.1  N(dBm0)  = No  - 0.37  - 4.34p  max.  error  = 0.l4dB 

f/B  = 0.2  N(dBm0)  = No  - 0.79  - 4.34p  max.  error  = 0.06dB 

f/B  = 0.3  N(dBm0)  - No  - 1.23  - 4.34p  max.  error  = 0.06dB 

f/B  = 0.4  N(dBm0)  = No  - 1.70  - 4.34p  max.  error  = 0.23dB 

f/B  = 0.5  N(dBm0)  = No  - 2.19  - 4.34p  max.  error  = 0.46dB 

No(dBm0)  = +29.64  - 20  log  Af/ch  rmg(kHz)  + 10  log  BIp(MHz) 

Note:  The  appropriate  pre-emphasis  value  must  be  subtracted  from  the 
above  values  If  the  terminal  uses  emphasis. 

For  C/N  between  -10  dB  and  +5  dB: 

N(dBm0)  = No  + (A)  + (D)p  r (C)p2 

where  the  ''A",  "B",  and  "C"  factors  are  obtained  from  the  graphs  on 
the  following  pages.  The  value  of  No  is  obtained  using  the  appropriate 
formula  listed  below. 

I Rectangular-  IF: 

f/B  = 0.0  N(dBm0)  = No  - 0.22  - 7.284p  + 0.722p2  max.  error  = 0.35  dB 

f/B  = 0.1  N(dBra0)  = No  - 1.03  - 6.600p  + 0.738p2  max.  error  = 0.37dB 

f/B  = 0.2  N(dBm0)  = No  - 1.91  - 5.036p  + 0.606p2  max.  error  = 0.48dB 

f/B  = 0.3  N(dBra0)  = No  - 2.65  - 3-326p  + 0.390p2  max.  error  = 0.52dB 

f/B  = 0.4  N(dBm0)  = No  - 3-26  - 1.287o  + 0.01  Ip2  max.  error  = 0.49dB 

f/B  = 0.5  N(dBra0)  = No  - 3-63  + 0.399p  - 0.319p2  max.  error  = 0.50dB 

J No(dBm0)=  +29.76  - 20  log  Af/(jh  rmg(kHz)  + 10  log  BIp(MHz) 


112 


Gaussian  IF: 


f/B  =0.0  N(dBm0) 

f/B  = 0.1  N(dBm0; 

f/B  = 0.2  N(dBra0) 

f/B  = 0.3  N(dBm0) 

f/B  = 0.4  N(dBm0) 

f/B  = 0.5  N(dBra0) 

No(dBm0)  = +29.51  - 20 
Averaged  IF: 
f/B  = 0.0  N(dBm0) 

f/B  = 0.1  N(dBra0) 

f/B  = 0.2  N(dBm0) 

f/B  =0.3  N(dBm0) 

f/B  = 0.4  N(dBm0) 

f/B  = 0.5  N(dBm0) 

No(dBm0)  = +29.64  - 20 


= No  + 0.03  - 5.588p 

= No  - 0.01  - 5. 440p 

= No  - 0.13  - 4.91  Ip 

= No  - 0.33  - 4.l63p 

= No  - 0.57  - 3.478p 

= No  - 0.83  - 2.973P 

log  Af , . (kHz)  + 
6 /oh  rms 


+ 0.326p2 

+ 0.392p2 
+ 0.438p2 
+ 0. 391p2 
+ 0.307p2 
+ 0.227p2 
10  log  Bip 


max. error 
max. error 
max. error 
max . error 
max. error 
max . error 
(MHz) 


0.04dB 
0 04dB 
0.03dB 
0.05dB 
0.05dB 
0.06dB 


= No  - 0.10  - 6.426p 

= No  - 0.52  - 6.01 4p 

= No  - 1.02-  4.969P 

= No  - 1.49  - 3.743P 

= No  - 1.92  - 2. 379p 

= No  - 2.24  - 1.282p 

log  Af , . (kHz)  + 
/on  rms 


+ 0.521p2 

+ 0.564p2 

+ 0.521p2 

+ 0. 390p2 

+ 0. 1 58p2 

- 0.047p2 
10  leg  BIp 


max . error 
max . error 
max . error 
max . error 
max. error 
max. error 

(MHz) 


0.17dB 
0. l8dB 
0.23dB 
0.26dB 
0.26dB 
0.26dB 


Note:  The  appropriate  pre-emphasis  va)ue  must  be  subtracted  from  the 

above  values  if  the  terminal  uses  emphasis. 

i.'  Due  to  the  computational  difficulties  involved  in  obtaining  exact 
results  for  low  C/Ns,  the  preceeding  results  were  obtained  by  fitting 
polynomial  curves  to  the  results  of  Stumpers  and  Rice.  The  maximum 
error  in  the  preceeding  results  is  graphed  on  the  following  page. 

For  C/N  greater  than  + 5dB: 


Rectangular  IF: 
N(dBm0)  = -139.0 


20  log  Af/ch  rms(kHz)  - RSL(dBm)  + NF(dB) 


PUB) 


+ 10  log/r(kHz)  + (0.32575)|VF 


(kHz) 


113 


Gaussian  IF: 


N(dBm0)  = -139.2  - 20  log  Af,  . (kHz)  - RSL(dBm)  + NF(dB)  - P(dB) 

2 /ch  rms 

+ 10  logf4-7r(f/B)  f2(kHz)  + (0.450.’ 9)f”  " 2/ 


4P  BIpT (kHz) 


Averaged  IF: 

N(dBm0)  = -139.1  - 20  log  Af 


/ch  rms 


(kHz)  - RSL(dBm)  + NF(dB)  - P(db) 


+10  log/h  f (kHz)  + (0.38797) 


V?  BIp2(kHz) 


for  the  above  formulas 


C/N(dB)  = RSL(dBm)  + 114.0  - NF(dB)  - 10  log  BIp(MHz) 
p = antilog  ( (C/NJ/10) 

Note:  The  above  formulas  do  not  include  baseband  phase/frequency  noise 

associated  with  the  transmitted  carrier.  This  noise  will  dominate  at 
very  high  C/N  ratios. 

The  following  special  cases  apply  when  the  FM  receiver  is  operating 
above  FM  (1  dB  noise)  threshold  (C/N  +13  dB)  but  with  a low  enough 
RSL  that  transmit  phase/frequency  noise  is  not  significant. 

Noise  in  3-1  kHz  noise  slot: 

N(dBm0)  = -139.1  + 20  log  f(kHz)  + NF(dB)  - P(dB) 

-20  log  Af/(jh  rms(kHz)  - RSL  (dBm) 

Noise  in  arbitrary  width  3lot  (no  de-emphasis,  sharp  cutoff  bandpass  filter) 


N(dBm0)  = -148.7  + 10  log  (ffllax(kHz)  - ffllin(kHz)) 

+ NF(dB)  - 20  log  Af.  . (kHz)  - RSL(dBm) 

/ch  rms 

Noise  in  entire  baseband  (no  de-emphasis,  sharp  cutoff  low  pass  baseband 

filter) : 

N(dlim0)  = -148.7  + 30  log  fmax(kHz)  + NF(dB) 

-20  log  Af/ch  rms(kHz)  - RSL (dBm) 


114 


B"  Factor 


Noise  Equations  Factors 


Figure  54.1 
:ont. ) 


16 


-10  dB®*  c/n*z  +5dB 


Nolt*  Equation*  Error  Analysis 

Figure  54.2 


117 


Note  that  although  noise  in  a narrow  noise  slot  increases  with  the  square 
of  the  slot  frequency,  total  baseband  noise  increases  with  the  cube 
of  the  upper  baseband  limit. 

FM  One  dB  Noise  Threshold: 

Narrow  Noise  Slots: 

Rectangular  IF: 

eG/n  (f/B)2  = (1.2580)  (c/n)* 

Gaussian  IF: 

ec/n  (f/B)2  = (1.7386)  e+1T(f/B)  (e/n)* 

Wide  Noise  Slots: 

Rectangular  IF: 

eC/n  ( ( f/B) 3max  - (f/B)3min)  = (3.7741)  (c/n)*  ((f/B)max  - (f/B)min) 
Gaussian  IF: 

eG/n(((f/B)3max  - (f/B)nlin)  - (1.8849)  ( (f/B)Sax  - (f/B)Sin) 

+ (2.1149)  ( (f/B)Zax  - (f/B)iin)  - (1.7225)  ( (f/B)Jax  - (f/B)9min) 

+ (1.1069)  ( (f/B)1max-(f/B)1min))  = (5.2158)  (c/n)*  ( (f/B)max  - (f/B)min) 

Where  the  symbols  are  as  previously  defined  with 

c/n  = carrier  to  noi3e  power  ratio  = antilog  ( (C/N)/10) 
f/B  = baseband  slot  frequency/IF  bandwidth 
(f/B)max  = highest  slot  frequency/IF  bandwidth 
(f/B)min  = lowest  slot  frequency/IF  bandwidth 

Of  the  last  four  formulas,  the  first  three  are  essentially  exact.  The 
fourth  is  based  on  a truncated  MacLaurin  series  expansion  of  e . Maximum 
error  due  to  this  approximation  is  0.5J  at  f/B  = 0.5.  The  error  is  con- 
siderably less  for  smaller  f/B  values. 


118 


6.  Baseband  Signal  and  Slot  Noisr  Versus  RSL  Using  Generalized  Charts 

6.1  Although  quieting  curves  can  be  predicted  using  the  preceeding 
formulas,  the  process  is  cumbersome.  To  facilitate  quieting  curve 
prediction  normalized  charts  have  seen  produced  based  on  various  FM 
receiver  IF  characteristics.  To  predict  the  FM  noise  quieting  curve 
for  an  FM  receiver,  a normalized  quieting  curve  can  be  chosen  based 

on  the  FM  receiver's  IF  characteristic.  If  the  FM  receiver's  IF  response 
is  unknown,  the  averaged  IF  response  curve  is  suggested  as  a choice. 

The  normalized  dB  values  for  slot  noise  and  carrier  to  noise  ratio  (C/N) 
can  be  converted  to  dBm0  and  dBm  values  respectively  using  the  following 
procedures. 

6.2  The  slot  noise  measured  by  the  frequency  selective  voltmeter  (FSV) 
is  plotted  on  the  vertical  scale.  The  noise  corresponding  to  0 dB  on 
the  slot  noise  scale  is  determined  from  the  following  formulas: 

N(dBm0)  = +29.6  - 20  log  £f . . (kHz)  + 10  log  BT_(MHz)  + CF(dB) 

i ch  rms  ir 

CF(dB)  = 10  log  f 3dB  bandwidth  of  FSV(kHz)  \ 

\ 3.1  kHz  / 

= correction  factor  to  allow  for  actual  noise  measurement 
bandwidth  (tabulated  on  next  page). 

6.3  Use  of  the  correction  factor  CF  produces  the  dBm0  noise  values 
whicn  will  be  read  by  the  FSV. 

6.4  The  actual  RSL  is  plotted  on  the  horizontal  scale.  The  RSL  corresponding 
to  0 dB  on  the  C/N  scale  is  found  from  the  following  formulas: 

RSL (dBm)  = -114.0  + NF(dB)  + 10  log  BIp(MHz) 

6.5  The  slot  noise  is  obtained  for  a given  RSL  by  going  up  vertically 
from  the  RSL  to  the  appropriate  (f/B)  line.  From  this  line,  go  to  the 
left  horizontally  and  read  noise  from  dBm0  scale. 

6.6  The  appropriate  (f/B)  valve  is  found  by  dividing  the  baseband  slot 
frequency  (frequency  to  which  FSV  is  tuned)  f (in  kHz)  by  the  receiver 
IF  bandwidth  (in  kHz). 

(f/B)  = Baseband  slot  frequency  (kHz) 

IF  bandwidth  (kHz) 

6.7  If  the  desired  (f/B)  valve  is  between  the  (f/B)  valves  listed  on 
the  curves,  read  the  slot  noise  N for  the  (f/B)  value  which  is  closer 
to  the  desired  (f/B)  value  and  add  the  following  correction  factor  to 
the  noise  valve  read  from  the  graph.  Theoretical  FM  noise  threshold 
has  been  indicated  on  the  graphs  by  short  vertical  lines. 


119 


I 


Theoretical  FM  Noise  Quieting  Curves 


laaaaaaaaaaaa 

■jM 


IiiiiH 

Issssssi 

■ ■■■•Ml 

■ ■••Mad 

I 


Rkss: 


■■•■■■MM 


iBssaKdisdiEiissiuiinsai 


Rss:us:»»; 

EiHM«i«aHiiiB 


*IWMWWI«1 

■ HtN  I B BB  B saDSMBTM i«i^wiw®WBWawawwwtf »='  " uow®"t«ki«! 

pmim'i  i 

PRHSSHSSH5HHHH^BHHHRHB^B5kSm£5X5sj!«b2  5k*2C^^B 

|BMMiBaMBBj»»BBB^aBaMBBBiMMBfrM»B»^gaaB^HiBap^BgHWM«mBBBg 

5iS5K5aw555SSBK5S5SS5?*iiSBW 

aaaaaaaaaa  **faaBaaaaaaB*aattts«aiJM«iaj 
aaaaaaaaaaaaiiaMaaa  ■•■■*■  *aa«Mr3»aa 

lMMMaaaaaaaaaaaaaaawaaaa>-aaaaBa-.<aaaa— — I — — 
|Q— Mw—wti  !■■■■■  nanr  mi  nuaii  n i |iiT-r  r 

aaaaaa  aa  a * aaSa  w ’*  ••  aa  ■' ' 4 mm  m *m  a b*b  *aa  ai  ’*m  6] 


■aaaaaaaaaa 

— ■■■■■aaaaaa 
aaaaaaaaaaaaa 


aaaaaaaaaviaaaa 


uaaRtaaHaaMHWMti'iHaairiRfiBr^aiaaiiiHiatiHiaM 

a— aaaaa  i aaaaaaa a riiawiraMiri— ar^aaaif^iaai— 

■aaaaaaaaaaa  a aaaaaaa  aaia^aaaar^asaaaaaaa  araaaaaaai  wmmmi 
WaMawfcaaiaiiaa^aaaariaaaar.aaaariaanaiaSaaaBt/tdwi*^ 
iaaaaaaaaa  bm  aaaaaa^aaa  aa^aa  aaa*^  aaaa  waaaa  rMfiaaaa^a  a«i  ir.Mj 
m ■■■■■—  aaa  amaa^aM  a a'^aaa  mm ' j ■ ■ mmw*m  mmmm  4 aapSaaaa^ai  iaa>eaa] 
iaaaaaaaaaaaaaaaa~iaaa^a'iaaata'  aat ar  iaaaar  laa^aaaa'iaaaawiftaa 

■kskI 

bBHHMff. 


■ — » 

M MM  IJ«BH  ■ ■■ 

i Saa^aaaSi,<t  aaaaa<Maaaa’*aaaaJ 

k*r4ia«W«mn«  ,m>b«  awlii* 
■■■■■Haauinar^MMriMaat.MaBfi^H 
IHM«BaiBaaaaaMai^aaaaa'vaaaaa^Biaaa4caliaa^aaa«aB 
aiaaaaaa  aaaaaaa  A i'*aaaaa**aaaa  a^aai  aa  mb  aaaaaa  as  aa  a* 
aaaMMBaaaaaaftaaaaaaauraaaa’AMMMajaaaaraaaaaaaa’ji 
■mmmmmMMMMMMHMMH1'''"****  Maaaa*aaaaaaa*M 


liaaSS^^ffl 

I Baaaaaaaaa  RHIIIHHi i ■ l 

^■^■iaaaaaamaaaaaaiMuaniaaaaaUBiraa^tBaaaiaaaatUaaaaaaBwaaBaB'^wiaBaaaaaanHH 
■ ■■■■a<«aaiMaaaaa— <MM><»aM>ia»— ^aaaa»iaaai>iaa— ■>ia»—»  «aaaar<aaaaaarjal 
I aaaaaaaaaatUNiaaaaalaaaaavaaBaa'aaaa  ■■'<■■*■*  *aBaaB’*aaMaaa»'*aBaaB'*aa*aB48aaaaBa’4aafl 
I aa  aiiaaaaa  aaaiaaaaaaa  aaar*aaaa*’«  mm  aaa^a  a aa  a 4 aaa  aa  4M»aaM  »'mb  aa  r.  aa  aa»*aa  aaa  w*m 
fcMMMaaaaaa—aaaaaaaa|rjaaaaa<Baaaaa,4aaaaa  ^aaaar^aaaaaaa^aaaaa^aaaaa^aaaaaar/aaaaa 
^^Kbmi  aaaraaaaaaa  aSaaaap  aaa  mmwmmmmm  « aa  aarjatai  aaa«paaaa#  aaaa  aaaaaa  aaa  aAflamajJ 
|Ha»aHaia»aaa— Miiiwii— af^aa— fiaaaariaaaaaatj«iaai.iaaia»>aaaaia>iiiayii 

l^izziiz^zznxunuunnrAmrnmmimmmmmmmmam^^mmi 


I aaaa  aaaaa  r.*aaaar  aaaa  t .-J"-'  ,.««aaaaaaaBaaaaaaaaBflaaM^^^^^^^^^^^^^^^^^^^^^H^^^H 
Haaaaaar^uf*'  ^aaa»v  ^aaaaaa  aaaaaaaaaaaaiaaaaaaa— aaa— aaaaaaaaaaaaa  aapaaaaaaaaaaaaaaaaa| 
Miii8f<aiftrt«>^BiaaBMiMiiaBMitiaiaBaiiii«MaaaaaaBia«aiiHai»aiaiiaiiaaiBaaaMMMd 


^^■aaBM  aa  aaaaaa  aaaaaaaaaaaa  I 
a aaaa aaaa»a aaaaaaa aaaaaaaaaaaaa! 
(■IBIIBIMHIHIffiaillMail|aB«ll 
i aaaaaaaaaaa aaaiaaaaaaa aaaaaaa* I 


|M*7^a**MaaaaN  a aaaaaaaaaa  aa  aaa  aa 
laafifl#  WfBililHMNttBIMMMMMa 
■ 510' . J**la  aaaaaaaaaa  aaaaaaaaaa  aaaaa 
laaar  # paa^  aaaa  aaa  aaaa  BBPBBaaBBB  aaaaa 
I mb  > jBBilt?ai— b—  ■■— bbbb 99mjmm 

|aa/«  >aaBaa«aaa>. aaaaaa  aaaaaaaaaa 
Bar./  » a aaaa*  aaaaaaaaaa  aaaaaaaaaa 
I a f f iBiAaaaiaaaaaaa.'aa  aaaaaaaaaa 


>aaa  aaaaa  aaaaaa*  aaaaaaaaaa  aaaaa  aaaaa  a aaa  aaa  aaa  aaaa  aaaaaa  aaa  aaaaaaa' 
aaaaa  aaaaaaaaaaMaaaaaa  aaaaaaa  aaaa*  aaaaaaaaaa  liaaaaaaaaa  aaaaaaaauia : 
aaaaMMMaaaaaa>iaaBaaaaa>aBaBBBBBBaaaaBBaaBaaBBiaBiaMBBBBBBBBBM 

aaaa  aaaaaaa! 65 aaaaaaaaaa  aaaaaaaaaa  aaaaaaaaaa  aaaaaaaaaa 


aaaBaaBBB  aaptaBaaBj  aa  BaBBaaBiBapaia 


P— aaaaaa*  aaaa  aaaa  a*  aaaaaaaa  aa  aa*H 
aaaaa  aaaa  a*  aaaa  aaaaaa  aaa  a aaa  aaaa  a aaaa 
aaaaiiaiaBiBaaBMiiMiiaiiiaiBBBBBBBaa 
—6a  a aaaa*  a*w*Maaaaaaaaaaaaaaaaaaaaa 

aaaa  a aaaaaaaa*  a ad 

BBMaBaaaaHMd 


i 


Isssili 


ISS8SS! 

Issnsi 

|*a«Mi 

laBBBB 


imm;sRsn:sKSu:KKKK»|l 

mmmwmmmmrna  ■■■■•■iiaiMill 


EHSUSIB 

jaapaawaaaa 


k2I&SSZS3]2SS«5& 

HuisNRiMaiiiini 

^^■aaipflaiai&aBiiaaaaaaBaifialBHIHi^^HHBHHi^H^H 

••*•••••■■  wHMaaMa*  **■•■■■•■«  mnm'.mmmmw  mm  «if  4m  . 

!SKss:aSs:R!s:!sss!:s»ssi!;iis3;:inKMiiKiiUsi!i 


s:i::ti8iS!ic2sip3ira^iRiK! 


| aaiaai 

Ibbbbbi 

■ ■■■Ml 


I 

■■•Baa 


iiiiinaii  Miirami'iiiai 


BajiinaniiiNniH 

Baa  MtniHaaaMiiuaai  a amaaaa  vaa  an  aar  aaaaar  a iaiaSar  all 

■MuaMiaiaauaiaaiaailMaia.iiiiriaiifiiii^^^Hii^^H 

laaiaaaMaaaaaaaaaaaaaaaaaaaa’aaaa'’  aaaaaaaaas 
1 ■?!?**"* ■ ■> g ■ a aai yatwawh a m^mmmr  mmmmm  <»■■■»  | 

8 

iM 

— laaaaaaaaaaaaaaaa  Sail*  — — — — W W—HB 

fiaaiiaaaaaBmaaaaaa^aaaraaaaa^aaaaa^aaaaa^aaaaaaavtaaaa  jia 
laaaaaaaaaaaaaaaaaSa^aaar  aaaav  ^aaaar  jaaaaa'.aaaaaaavilfiai'aaaaH 
■aaaaaaaaaaaaaaBaaa^aaaraaaaaiaB«aaaaaaaa>«aaaaaa^aaaaaaia5aa| 

'|MNaaiaaaaaaBB>r4Mai’4BBBa'4aBaia4aaaa»’iBiaaaar4a«a«i'4aaBSB>y 
aMaMna««*aiM'iNHr  mmmmm  smmmmm  mmmmmm  *mmmmnmm  *mmmmm  «••••■■ 


mm 


^^^^HMHMNaaaiRaaiaaaiaaar  aaaaa'jaaaaa^aaaaa^aa  aiaaa 

SBaaaaaaanaiiiaaBBM  aaaaaaa  a a a aaaa  aar  aaa  aa  a a ■ a ■ a aa  aaaao  a at  a a a a >!■ 
■aaaaaaaaiaaaeaaaaaaaaaaaaaaaaa'aaaaraaaaa'aBaaaaaaaaaa^aaaaaar^iliial 
laiaBaaBMaaMaaaaaaaiaaaaaaaia^afiaraBaaa'aaaaaaaiaajaa^BBaaaaB'jailiauHHHHHBii 

lBBaaaaaaBaaBBaaBBaBBaaaaaaBBa'aaiaraaiaB'aaaaaBaBB8Ba'4aBa8Ba»aaaBBr^aBBBt'iaBBaBia4 

iMHamMiiRMBiaiRBiBMVBMiiMa'  aaaaaaaaaaaa'.aaaiaaiajaaaaa  ^aaaaa^aaaaaaaaSi 

l«RR»aBRMiaM«iaRM>iaaiaa'4Ratf.«««H4uaa»aaiaaa'iaaBBMa4aBia*iaaaiifMaBiaBi»iaa 
iBaaiBaaaaaRaBaaaaaaBaaaB^aasr^aaarjaaaaaaaaBBa^aaaaaaaiBBBaa ^<«b  aaaaaaa  aaaa#^ 
MiiiaaiaaiaMiaaaaaMBaaaiiaar  aaaafiMaaa  jaaaar^aRaaaa^aaaai  iaaaaaiBaaaaaaH 
■aiaaaaaaaaaaaaaaaaaaiiBiaaar  aaaav^aBaaf  ^aaaaa’iaaaaRBa^BaaBa.aaaBiMaaaaaaa^ 

I aaaaaaaaai  aaa  aaaaaaa  ar^Baar.aaa  arjaaaaraa  aBaa^aaaaaaa^aaaar^a  aaa  aaaaaaa 

■ 4 rnmmw  mmmmm *mimmmm  Mmmmmm'+mmmmmmm  * aaaaa  *aaai  m +mnmmmmm'. aaaaaiaai 

I g^a»g^a<^a>iag^»a»^aaiaauiaMWaaaaaa%aaaiaa»^BBa^^Baaai^aaaaan^ggaj^^M 


«feaaaaMaaaBa4aaa»aaaaa4aaaaa4aaaa0'.aaaMaav4BBB«a«aaaar«aaaaaB'«aa 

::KSt»3sx:RS»i;!»S:r^:s;»:s:v^R>r<r^:::^:»3as 


£i^rms:S!e:!:us"!:»«::s:£!»”“ssnaRK:t:isR:»!«s:::s:::::::K:s: 

|Miy//4lltaaaNBaiaaaaBaB«aaiaaaiaaflMiaaaaiRaaRaaaaRBaBBaBaaaaBaaaaiMaaaaBaBtiBaaBBaBa 


aaa aaaaaaa aaaaaaa aaa | 
aaaaaaaaaaaaaaaaaagil 

aaaaBaaaaaaaaaaRi 


I mmmit  mmmmm mmmmmmmmmm  iBaauaaaa- ■BaaaaaaaaBaBaaaaBaaaBaaaaa 

I aaa'  ^aa  ■aaMBaaBBal,aaiaBRaaaMBia|||||a||iaa||||i|HnBMa|anMd 

I *«  ‘*aaaat;« a»#a»w»»»BBMi#toaa*ta!«®*®«R»(e 

|aa4taaaafiik^aaaaaaaaaaaaa«a8aaaaj|>a 

■ R—w»f>^aaaiaa>— aaaaaaa  aaaSBa— IBM — M— — I— i IB— — WIPBWIlWllllliWIWW 

|:iuBiH::sR:i::R:^u:::3::un::&s::!K3ituu:i::»R::tS3si3s8s:s»!:::sssuss: 

|lMWWS»«WIWHWW*rt«MHIWHIWIIM«limH»MIMMIM«WHMl»»M«IM«aM»m«n»«» 

■■feg^g^JiggggiS^aSiiKigiKIilHillSiKiiiRKilSitiiegiiHSItiiinKM 

pBliiMamiu^B^^^aaaiaiaBBBaaaaaaaaBaaaaiaaaaaaaaaiBaaBBaaaBiaaaBaaaaaaaaaaBai 


|aaaBaaajBaaaaaaB*JJJgp| 

iaaaapafaaa»»  ■■■•••■■^aaBaBaaBaaaaaaaBaaaaaa?! 

aaaBBBaaaaaaaaaaaBaBaR  MMMBMaMMaiiMMaaaaaaaaaaai 


a •••»•» aaaaa 

358383;83333S3Si 


laaaaaaaa  aaaaaaaaai  aaaaaaaaai 

laBaaaBBaiaaaaaaaaiBaBaBBBaaBBBBBaaBaaaBBaBBMaaaiaaBBBaaa 
iaiaaa«BaRMBMBMriiBiaBaBaRaBwaaiMMM|B|a|MMM«MMa 


■ IISMBIM 
aaaaaaaaai 

R^iinaiiiMMi 
BB».,a»paB  aaaaa  papai 


RBBiBBI|BIBI 


aBBaBeiaBaBsaaaaaaeaaaiBaaBaaBBaaEaiSiaaaal 


■iaaBaaaaaaaaaa  aaaaaaa 
aaaaBaaaaaaaaaaaaaaaaa 


>1 


laaBaiaaBsaaBBa 


R]Sh::^::a:!:!88:i:8i::::::::8»3iKK!:s3U»::::ssRsn:; 

bbbbbbM 


■HaBaaaaaaaaBBBk.'BBBaaaaraaBaiaaaa 
BBBBBBBaBaBBBBaaBa^'aaaVBBBBBBBMa 
■laHaaaaiaaaaiaanit.  ^BaaaaMia 


I «i#a*f  a 
Ipiaiaia 


© 


s 


♦80 


Measurement  Conversion  Measurement  Conversion 


rH  (\J  O rH 

On  On  CT>  On 

• • • • 

JT  l*-  rH  JT 

H H N N 

+ + + + 


N 

X 

Ini 

O 

.=r 

t*- 

OK 

o 

o 

o 

o 

' — 

cO 

• 

• 

o 

o 

o 

o 

rH 

in 

CO 

C\J 

rH 

L. 

00 

o 

c 

in 

O' 

oo 

rH 

o 

■=r 

O r-N 

o 

ON 

rH 

oo 

o 

o 

rH 

o 

rH 

-*->  CO 

• 

• 

• 

» 

• 

• 

• 

♦ 

O T3 

0J 

rH 

CO 

\o 

in 

CM 

rH 

o 

o 

(fl  ' — 
U. 

1 — < 

1 

rH 

1 

1 

1 

i 

1 

1 

+ 

Table  19 


123 


(Uniformly  distributed  noise  power  assumed) 


N = 20  log/ (f/B)  desired \ 

c v (f/B)  from  graph  ) 

6.8  To  evaluate  wideband  FM  receiver  performance,  it  is  desirable  to 
measure  receiver  noise  quieting  in  at  least  three  noise  slots.  To 
facilitate  later  comparison  with  Basic  Intrinsic  Noise  Ratio  (BINR) 
and  Noise  Power  Ratio  (NPR)  measurements  taken  with  the  same  receiver, 
it  is  recommended  that  quieting  curves  be  taken  at  the  same  slot  frequencies 
as  those  used  for  BINR/NPR  measurements.  Recommended  frequencies  are 
listed  on  the  following  page.  The  listed  frequencies  were  chosen  to 
conform  to  current  CCIR  recommendations  and  the  recommendations  of  Ta.it. 

If  there  war*  a conflict,  the  CCIR  value*  war*  chosen. 

6tq  As  an  example,  consider  finding  slot  noise  for  an  RSL  corresponding 
to  a +35  dB  C/N.  Refer  to  the  following  graph.  Assume  the  (f/B)  desired 
value  is  (0.0706) . 

N(f/B  = 0.1)  = - 50.0dB 

N( f/B  = 0.0706)  = N(f/B  = 0.1)  + 20  log  0.0706 

0.1 

= -50.0  + (-3.0)  = - 53-OdB 
Alternately 

N( f/B  = 0.05)  = - 56 . OdB 

N( f/B  = 0.0705)  = N( f/B  = 0.05)  + 20  log  0-C706 

0.05 

= -56.0  + (3.0)  = - 53.0  dB 

6.10  The  baseband  signal  level  is  a function  of  RSL  near  FM  threshold. 

The  signal  level  is  found  by  reading  the  correction  factor  from 
the  slot  noise  (dB)  scale  on  the  graph  on  the  next  page.  The  level 
(disregarding  noise)  of  the  received  baseband  signal  S is  found  by  using 
the  following  formula: 

S(dBra0)  = St(dBm0)  + Scp(dB) 


6.11  For  example,  at  an  RSi.  corresponding  to  OdB  C/N,  the  signal  level 
will  have  dropped  4 dB  (S=  S^-4  dB). 

6.12  If  the  receiver  being  tested  has  de-emphasj.3,  the  slot  noise  power 
read  from  the  graphs  must  be  modified  by  the  effect  of  the  de-emphasis 
networks.  To  determine  the  actual  noise  measured  by  the  FSV  after  the 
noise  has  passed  through  the  de-emphasis  networks,  subtract  (algebraicly) 


124 


Number  of 
Channels 

12 

Baseband 

Frequencies 

(kHz) 

12-60 

Slot  Frequencies 
(kHz) 

27  40  56 

24 

12-108 

40 

70 

105 

36 

12-156 

40 

70 

105 

48 

12-204 

40 

105 

185 

•60-252 

70 

185 

245* 

60 

12-252 

40 

185 

245 

60-300 

70 

185 

270 

72 

*12-300 

40 

185 

270* 

120 

60-552 

70 

270 

534 

132 

•12-552 

4 o' 

270 

534* 

240 

60-1052 

70 

534 

1002 

300 

60-1300 

70 

534 

1248 

420 

*60-1796 

70 

534 

1248* 

600 

60-2660 

70 

1248 

2438 

900 

316-4188 

534 

2438 

3886 

960 

60-4028 

70 

2438 

3886 

O 

o 

316-5564 

534 

3886 

5340 

1260 

*60-5564 

70 

3886 

5340* 

*60-5636 

70 

3886 

5340* 

1800 

316-8204 

534 

3886 

7600 

2400 

•316-11404 

534 

3886 

7600* 

2700 

316-12388 

534 

3886 

11700 

Recommendations  conform  to  CCIR  NPR/BINR  slot  frequencies. 
* No  CCIR  recommendations  available  for  these  basebands.  * 

Recommended  Quieting  Curve  Slot  Frequencies 

Table  20 


125 


the  pre-emphasis  effect  (dB)  for  the  slot  frequency  of  interest  f relative 
to  the  baseband  pivot  frequency.  The  formulas  for  the  pre-emphasis 
effect  valves  for  the  various  types  of  pre-emphasis  networks  are  listed 
on  the  following  pages.  For  CCIR  and  EIA  emphasis  networks,  the  pre- 
emphasis valve  may  be  read  directly  from  the  following  table. 


General  Pre-emphasis  Networks: 

CCIR: 

P(dB)  = 10  log  (+0.400  + 1.20  (f/fmax)2  + 0.801  (f/fmax)14 
+ 1.03  (f/frnax)6  - 0.913  (f/fmax)8) 

Pp(dB)  = 0.0  dB 
fp  = (0.608)  fmax 


EIA: 

P(dB)  = 10  log  (0.250  + 2.25  (f/fmax)2) 
Pp(dB)  = 0.0  dB 
fp=  (0.577)  fmax 


Time  Constant: 


P(dP)  = 10  log  (1.0  + 39.48  T2f2)  - 10  log  (1.0  + 13.16 X2fmax2) 

Pp(dB)  s 10  log  (1.0  + 13.l6t2fmax2) 

fp  = (0.577)  fmax 

T = time  constant  (yusec) 

f = baseband  frequency  (kHz) 

fmax  = maximum  baseband  frequency  (kHz) 


REL  Pre-emphasis  Networks: 

72  channel  system: 

(12  to  300  KHz  baseband) 

P(dB)  = 10  log  (0.1743  + 2.477  (f/fmax)  2) 

Pp(dB)  = +7.58 


128 


fp  = (0.577)  fmax  = 173  kHz 

132  channel  system: 

(12  to  552  kHz  baseband) 

P(dB)  = 10  log  (0.1456  + 2.563  (f/fmax)2) 

Pp(dB)  = +8.37 

fp  = (0.577  fmax)  = 319  kHz 

252  channel  system 

(12  to  1052  KHz  baseband) 

P(dB)  = 10  log  (0. 160  + 2.520  (f/fmax)2) 

Pp(dB)  = +7.95 

fp  = (0.577)  fmax  = 607  kHz 

Limitations 

6.13  The  previous  sections  have  indicated  the  slot  noise  produced  at 
the  baseband  of  an  FM  receiver  due  to  the  FM  demodulation  noise  process. 

The  implicit  assumption  was  that  the  demodulator  output  was  only  proportional, 
regardless  of  C/N,  to  the  time  derivative  of  the  phase  of  the  composite 
received  signal  plus  noise  waveform.  It  was  also  assumed  that  the  total 
noise  power  N (producing  the  C/N)  is  due  to  the  thermal  noise  passing 
through  the  IF  amplifier  and  will  be  applied,  without  further  bandlimiting, 
to  the  FM  demodulator.  These  assumptions  are  quite  reasonable  for  conven- 
tional FM  demodulator  with  hard  limiting.  However,  the  assumptions 
do  not  hold  for  FM  demodulators  with  feedback  (FMFB)  and  phase-locked 
loop  (PLL)  demodulators. 

6. 14  Receivers  employing  FMFB  or  PLL  FM  demodulators  have  exactly  the 
same  noise  characteristics  in  regions  C and  D as  any  comparable  conven- 
tional FM  receiver.  However,  the  FMFB  and  PLL  receivers  have  improved 
FM  noise  threshold  performance.  The  C/N  at  which  FM  threshold  occurs 

can  not  be  readily  generalized.  However,  Schwartz  suggests  a simplification 
for  wideband  FMFB  receivers  which  implies  a maximum  attainable  FM  threshold 
improvement  of  7 dB  for  a receiver  with  maximum  baseband  frequency  to 
IF  bandwidth  ratio  of  0.1  (f/B  = 0.1)  relative  to  conventional  FM  demodulation. 

6.15  The  noise  performance  of  an  FMFB  or  PLL  receiver  in  regions  A and 
B is  difficult  to  analyze.  It  involves  not  only  analysis  of  the  FM 
noise  process,  but  also  analysis  of  the  loop  dynamico  of  the  FMFB  or 

PLL  demodulators  under  nonlinear  conditions.  This  problem  is  formidable. 

Enloe  mentions  that  as  of  1962  the  problem  had  not  been  solved. 


129 


f/f 

max 


0. 

0.025 
0.050 
0.075 
0.100 
0.125 
0.150 
0. 1 75 
0.200 
0.225 

0.250 

0.275 

0.300 

0.325 

0.350 

0.375 

0.400 

0.425 

0.450 

0.475 

0.500 

0.525 

0.550 

0.575 

0.600 

0.625 

0.650 

0.675 

0.700 

0.725 

0.750 

0.775 

0.800 

0.825 

0.850 

0.875 

0.900 

0.925 

0.950 

0.975 

1.000 


CCIR 

(dB) 


-3.98 

-3.98 

-3.95 

-3.9! 

-3.86 

-3.78 

-3.70 

-3.60 

-3.48 

-3.35 

-3.2! 
-3.05 
-2.89 
-2.7  i 
-2.52 
-2.32 
-2.  I 1 
-1  .89 
-I  .66 
-!  .43 

-1 .19 
-0.93 
-0.68 
-0.42 
-0.15 
0.12 
0.40 
0.68 
0.96 
! .24 

1 .52 
! .79 
2.07 
2.34 
2.6! 
2.86 
3.12 
3.36 
3.59 
3.80 
4.01 


EIA 

(dB) 


-6.00 

-5.98 

-5.9! 

-5.79 

-5.63 

-5.43 

-5.20 

-4.95 

-4.67 

-4.37 

-4.06 
-3.75 
-3.43 
-3.10 
-2.77 
-2.45 
-2.13 
-I  .81 
-I  .50 
-1.19 

-0.88 
-0.59 
-0.29 
-0.01 
0.27 
0.55 
0.81 
1 .07 
1 .33 
1 .58 

1 .82 
2.06 
2.30 
2.53 
2.75 
2.97 
3.18 
3.39 
3.60 
3.80 
4.00 


CCIR/EIA  Pre-emphasis  Values 
(change  dB  sign  for  de-emphasis  value) 

Table  21 


130 


1 


f/f 

72  Ch. 

132  Ch. 

252  Ch 

max 

(dB) 

(dB) 

(dB) 

0. 

-7.59 

-8.37 

-7.96 

0.025 

-7.55 

-8.32 

-7.92 

0.050 

-7.44 

-8.18 

-7.79 

0.075 

-7.25 

-7.96 

-7.59 

0.100 

-7.0! 

-7.66 

-7.32 

0.125 

-6.72 

-7.31 

-7.00 

0.150 

-6.38 

-6.92 

-6.64 

0.175 

-6.02 

-6.50 

-6.25 

0.200 

-5.63 

-6.05 

-5.84 

0.225 

-5.23 

-5.60 

-5.41 

0.250 

-4.83 

-5.15 

-4.98 

0.275 

-4.42 

-4.69 

-4.55 

0.300 

-4.01 

-4.25 

-4.  1 3 

0.325 

-3.61 

-3.81 

-3.70 

0.350 

-3.21 

-3.38 

-3.29 

0.375 

-2.82 

-2.96 

-2.89 

0.400 

-2.44 

-2.55 

-2.49 

0.425 

-2.06 

-2.16 

-2.  1 1 

0.450 

-1.70 

-1.77 

-1.74 

0.475 

-1.35 

-1  .40 

-1  .38 

0.500 

-1.00 

-1  .04 

-1  .02 

0.525 

-0.67 

-0.70 

-0,68 

0.550 

-0.35 

-0.36 

-0,35 

0.575 

-0.03 

-0.03 

-0.03 

0.600 

0.28 

0.29 

0.28 

0.625 

0.58 

0.59 

0.59 

0.650 

0.87 

0.89 

0.88 

0.675 

1.15 

1.18 

1.17 

0.700 

1 .42 

1 .47 

1 .45 

0.725 

1 .69 

1 .74 

1.72 

0.750 

1.95 

2.01 

1 .98 

0.775 

2.2  1 

2.27 

2.24 

0.800 

2.45 

2.52 

2.49 

0.825 

2.70 

2.76 

2.73 

0.850 

2.93 

3.00 

2.97 

0.875 

3.1  6 

3.24 

3.20 

O.VOO 

3.39 

3.47 

3.43 

0.925 

3.6! 

3.69 

3.65 

0.950 

3.82 

3.91 

3.86 

0.975 

4.03 

4.12 

4.07 

1 .000 

4.23 

4.33 

4.28 

REL  Pre-emphasis  Values 
(change  dB  sign  for  de-emphasis  value) 

Table  21.1 


131 


6. 16 In  general,  the  analysis  developed  in  the  previous  sections  of 
this  report  will  describe  the  noise  performance  of  FMFB  and  PLL  demodu- 
lators in  regions  C and  D.  The  theoretical  FM  threshold  C/N  values 
will  be  pessimistic  and  the  theoretical  slot  noise  for  regions  A and 
B do  not  apply  for  FMFB  or  PLL  demodulator  receivers. 

Combiner  Improvement 

6. 17 To  this  point,  microwave  receiver  noise  performance  has  only  been 
considered  for  a single  receiver.  An  actual  microwave  terminal  normally 
will  have  two  or  more  receivers  with  the  capability  to  contribute  to 
the  baseband  output  of  the  terminal.  The  outputs  of  the  various  individual 
receivers  are  added  (combined)  to  produce  a baseband  signal  which,  if 
the  combiner  functions  properly,  will  have  superior  signal  to  noise 
performance  when  compared  to  the  performance  of  a single  receiver. 

If  the  combiners  do  not  function  properly,  baseband  signal  level  sta- 
bility and  impulse  noise  performance  can  be  seriously  degraded.  In 
general,  three  types  of  combining  are  used:  switching  (selection), 

equal  gain,  and  maximal  ratio  (ratio  squared)  gain.  The  effect  of 
combiners  is  a complicated  subject.  As  Brennan  mentions,  the  relative 
performance  of  combiners  depends  on  the  type  of  RSL  fading  the  microwave 
terminal  experiences.  There  are,  however,  two  cases  of  special  interest. 
The  idealized  situation  for  a LOS  terminal  is  for  all  receivers  to  see 
a constant,  unfading  RSL  which  is  the  same  power  level  at  all  receivers. 

For  post-detection  (baseband)  combining,  the  selection  combiner  will 
give  no  improvement  over  a single  receiver.  The  equal  gain  and  variable 
gain  combiners  will  have  exactly  the  same  signal  to  noise  improvement. 

In  either  case,  the  improvement  will  be  ten  tiu.es  the  common  logarithm 
of  the  number  of  receivers  used  for  combining.  Unfortunately,  this 
Idealized  situation  is  seldom  realized  in  practice.  Diversity  engineering 
sometimes  Intentionally  makes  the  RSL  of  one  receiver  different  than 
that  of  the  other  receiver.  Also,  different  waveguide  run  lengths  and 
waveguide  hybrid  losses  can  also  invalidate  the  idealization.  On  the 
other  end  of  the  scale,  the  TROPO  receivers  generally  experience  randomly 
fading  receive  signals.  If  the  receivers  all  have  the  same  average 
RSL  and  if  the  RSL  are  all  Rayleigh  amplitude  distributed,  then  the 
combiner  signal  to  noise  improvement  will  be,  on  the  average,  the  values 
that  Brennan  has  predicted.  Unfortunately,  in  real  life,  for  various 
reasons,  the  average  RSLs  of  the  various  receivers  is  often  different 
by  a few  dB.  Average  combiner  noise  performance  under  practical  conditions 
is  difficult  to  predict  accurately  without  sophisticated  parameter  measure- 
ment of  the  actual  radio  path. 

6.18  Combiners  serve  to  improve  the  baseband  signal  to  noise  performance 
of  a microwave  terminal.  Since  the  signal  out  of  the  combined  baseband 
must  be  the  same  level  a3  the  signal  that  was  applied  to  the  transmit 
baseband,  the  combiner  will  reduce  the  noise  at  the  combined  baseband 
when  compared  to  the  noise  out  of  a single  receiver.  As  a rough  estimate 
of  the  actual  slot  noise  that  will  appear  at  the  combined  baseband  of 


132 


a microwave  terminal,  the  curves  Brennan  developed  can  be  used.  If 
the  terminal  is  LOS  and  the  RSLs  are  stable  and  equal,  the  following 
conditions  are  approximately  correct.  For  selection  combining,  the 
noise  at  the  combined  baseband  will  be  exactly  the  same  as  for  a single 
receiver.  The  baseband  noise  can  be  read  directly  from  the  quieting 
curve  of  a single  receiver.  If  the  terminal  uses  equal  gain  or  ratio 
squared  combining,  the  noise  at  the  combined  baseband  will  be  less  than 
for  a single  receiver  by  the  amount  shown  on  the  following  chart  for 
variable  gain  combining.  The  slot  noise  at  combined  baseband  will  be 
the  slot  noise  read  from  the  quieting  curve  minus  the  dB  valve  read 
from  the  Brennan  chart.  For  a TROPO  terminal,  if  the  RSLs  are  the  same 
and  all  are  randomly  fading,  then  the  noise  improvement  will  be  found 
by  using  the  curve  on  the  Brennon  chart  which  applies  to  the  type  of 
combining  appropriate.  The  slot  noise  at  the  combined  baseband  will 
be  the  slot  noise  value  read  from  the  quieting  curve  minus  the  dB  noise 
improvement  factor  determined  from  the  Brennan  curve. 

Noise  weighting 

6.19  When  noise  measurements  are  made  on  telephone  circuits,  various 
weighting  filters  ire  used  to  measure  the  noise.  If  a 3 kHz  flat  filter 
is  used  to  measure  the  noise,  the  slot  noise  previously  determined  from 
the  quieting  curve  can  be  used  directly.  If  a C-Message  or  Psophometric 
(CCITT)  filter  is  used,  a correction  factor  must  be  algebraicly  subtracted 
from  the  previously  determined  noise.  This  subtraction  accounts  for 

the  noise  power  lost  by  the  frequency  characteristic  of  the  filter. 

As  Tant.  mentions,  the  theoretical  factor  for  C-Message  is  1.5  dB  and 
is  2.5  for  Psophometric  weighting.  The  use  of  2.0  dB  as  the  factor 
for  either  weighting  characteristic  is  recommended. 

Cable  Noise 

6.20  Cables  connect  the  wideband  M/W  terminal  with  multiplexers.  The 
cables  generally  are  located  near  other  signal  and  power  cables.  The 
other  cables  may  induce  crosstalk  noise.  The  cables  may  have  noise 
induced  by  radio  frequency  interference.  Occasionally,  baseband  noise 
will  be  introduced  by  the  M/W  terminal  itself.  Cable  induced  noise 

is  difficult  to  determine  without  direct  measurement.  With  LOS  M/W 
systems,  however,  the  combined  M/W  terminal  noise  and  cable  induced 
noise  can  be  measured  by  terminating  the  transmit  cable  at  the  multi- 
plexer and  measuring  the  cable  noise  at  the  multiplexer  end  of  the 
receive  baseband  cable  using  a frequency  selective  voltmeter.  Noise 
measurements  can  then  be  taken  with  the  transmitter  baseband  amplifier 
input  terminated  and  the  frequency  selective  voltmeter  located  directly 
at  the  combined  terminal  receive  baseband.  Using  the  two  noise  measure- 
ments and  the  noise  addition/subtraction  curves,  the  cable  noi3e  can 
be  obtained. 


133 


Number  of  Independent  Receivers 
(after  Brennan) 


Combiner  Improvement  Curves 


Figure  60 


134 


Examples 


6.21  To  show  a method  of  comparing  theoretical  quieting  curves  with 
actual  quieting  curve  data,  two  examples  will  be  shown. 

LOS  M/W  Receiver  (without  de-emphasis): 

6.22  The  following  page  lists  various  slot  noise  values  for  various 
RSLs.  The  values  were  taken  in  dBm  at  a -15  TLP.  The  receiver  had 
no  de-emphasis.  Therefore,  to  convert  the  measured  valves  to  dBm0 
values,  +15  dB  must  be  added  to  each  value.  These  valves  are  listed 
on  the  following  page. 

6.23  Following  the  noise  slot  measurements,  a page  is  given  which  lists 
baseband  test  tone  power  level  versus  various  RSLs.  The  test  tone  level 
was  measured  using  a 3*1  kHz  wide  frequency  selective  voltmeter  (FSV). 

At  low  RSLs  the  noise  in  the  FSV  slot  was  significant.  Therefore,  what 
the  FSV  measured  was  test  tone  level  plus  noise.  To  compensate  for 

the  noise,  after  measuring  the  test  tone  plus  noise  (TTL  + N),  the  test 
tone  was  removed  from  the  baseband  of  the  transmitter  and  the  slot  noise 
(N)  was  measured  at  the  same  RSL.  Using  the  dB  subtraction  curve  of 
the  power  addition/subtraction  curves  the  slot  noise  was  subtracted 
from  the  test  tone  level  plus  noise  measurement.  This  value  was  listed 
as  the  test  tone  level  (TTL).  The  proceeding  values  have  been  plotted 
on  the  following  graph. 

TROPO  M/W  Receiver  (with  CCIR  de-emphasis): 

6.24  The  following  page  lists  various  receiver  baseband  slot  noise  values 
for  various  RSLs.  The  values  were  taken  at  a -10  dB  TLP.  The  receiver 
had  CCIR  de-emphasis.  To  convert  the  measured  valves  to  dBm0,  +10  dB 
plus  the  pre-emphasis  value  (or  minus  the  de-emphasis  value)  must  be 
added  to  each  valve  to  arrive  at  a dBm0  slot  noise  corrected  for  de- 
emphasis . 


Slot  Frequency 

TLP 

De-emphasis 

Correction 

(kHz) 

(dB) 

(dB) 

Factor 

(dB) 

308.7 

-10 

-4.7 

+ 14.7 

154.4 

-10 

+0.8 

+ 9.2 

77.2 

-10 

+3-5 

+ 6.5 

30.9 

-10 

+4.3 

+ 5.7 

De-emphasis  values  taken  from  actual  measurement  of  the  receiver  are 
listed  on  the  next  page.  The  values  were  taken  by  running  transmitter 
to  receiver  baseband  frequency  response  with  pre-amphasis  strapped  out. 


135 


i 


[ Pot*r  Addition,  dB  difference  between  two  powers 

I Power  Subtraction 

Com  1:  dB  to  bo  suotrtctod  fro*  larger  power 
Com  2:  dB  difference  between  two  powers 

] 


Power  Addition/Subtraction  Curves 


j Figure  61 

! 


136 


FM  RECEIVER  NOISE  uUlETIMG 


| UMINAHV  DAI  A 


[^)  PINAL  data 


Sample  Quieting  Curve  Data 
Table  22 

137 


FM  RECEIVER  NOISE  QUITTING 


MMfLIMlNAWY  Dm  A 


DATA 


IF  SvO 

2 S M Hz. 

I T A T lOf« 

LOS 

A*  A 4 » * V TONI, 

-1  5. 

V f v I l 

oiCCVAh 

AV/FRC-157 


AGC/i' 

HONAL  VOLTS 
<t)L) 


HO'Sf  POHfi  IM  BAifHAM'jSLOT  (—  J g ^ 


oascbano  Slot  rnr'  ituc  v 


/ i r> 


so 

ISO 

iz5 

12.5 

os 

ekes 

0 o l 

o oo 5 

0.0015 

4.5 


V I 7 


6,  5 I 6. S 


10.5 

1 0 

u.  S 

1 1 

31.5 


33.  5 


27.  S JJ 


a 8.  S 0.S 


a $ 


<3.5  I 13.5 


14,5  I (5 


17  I )7. 5 j |0  j 17. S 


<<1,5 


/2- 


U.  5 13.  S 


?2. 


36 


43.5 


45.5 


41 


41  i S»,  5 51,  S 


-6  1 

1 S 

4 7'  ~! 

5 

5 3,5  I 56.5  ! Si 


/i.Z> 


75.5 


I ->4  5 


60 

54 

63.5 

et.s 

jo  jii  gu.e t inc  mil 


KOI  PMCQt.MHCf 


*50  KHt 


- 8?  S .... 


5( 

6 1.5 

65  5 

7 

tenai 

nn 

70 

‘bos 

BO 

EEKi 

00  70 


NT  HfttlhOl  l)  MSL 


-85 


Sample  Quieting  Curve  De 
Table  23 
138 


I 3.  -j  O xh. 


- 8 


i S)  O Kill 


-1*\ 


Sample  Quieting  Cirve  Data 
Table  24 


139 


Figure  62 


140 


f (kHz) 

De-emphasis  (dB) 

f /fmax 

Measured 

Theoretical 

252 

1.0 

-3.7 

-4.01 

227 

0.9 

-2.6 

-3.12 

202 

0.8 

-1.5 

-2.07 

176 

0.7 

-0.3 

-0.96 

151 

0.6 

+0.2 

+0.15 

126 

0.5 

+2.0 

+1.19 

101 

0.4 

+2.0 

+2.11 

75.6 

0.3 

+3.5 

+2.89 

50.4 

0.2 

+4.0 

+3.48 

25.2 

0.1 

+4.2 

+3.86 

f (kHz) 

f/B 

Measured 

Theoretic) 

309 

0.1 

-4.7 

-4.99 

154 

0.05 

+0.8 

+0.01 

1 

77.2 

0.025 

+3-5 

+2.85 

30.9 

0.01 

+4.3 

+3.79 

t 

Tropo  M/W  Receiver  De-emphasis  Performance 

I 

! Table  25 


141 


FM  R[  CtIVFK  NOISf  Quit  HUG 


'00  Jn„,  -100 


-1 00  ^ 


3 0.°[  m. 


-°ld  _ 


Sample  Quieting  Curve  Data 
Table  26 


142 


Sample  Quieting  Curve  Data 


Table  27 


ft 


144 


-130 


Figure  64 


145 


08*  0L+  09« 


6.25  Notice  that  the  measured  values  are  significantly  different  than 
the  theoretical  values.  This  shows  the  necessity  of  checking  FM  receiver 
de-emphasis  characteristics  if  accurate  comparisons  are  to  be  made  between 
actual  and  theoretical  quieting  curves. 

6.26  Following  the  noise  slot  page  is  a page  listing  baseband  test  tone 
levels  versus  RSL.  Using  the  same  procedure  as  mentioned  for  the  LOS 
receiver,  values  for  test  tone  level  were  determined  from  the  test  tone 
plus  noise  measurements.  The  results  have  been  plotted  on  the  following 
pages.  The  first  graph  shows  the  quieting  curve  witr.out  de-emphasis 
correction.  The  second  graph  shows  the  quieting  curve  after  correction 
for  de-emphasis. 

6.27  Calculations  for  using  the  Averaged  RF/IF  response  chart  with  the 
LOS  receiver  (excluding  effects  of  de-emphasis)  follow: 


Receiver  Specifications: 

IF  bandwidth:  25  MHz 

per  channel  deviation::  140  kHz  rms 

(for  0 dBm0  test  tone) 
noise  figure:  8 dB 

Noise  Slot  Measurement  (FSV)  Bandwidth:  3-1  kHz 

Noise  value  corresponding  to  0 dB  slot  noise  on  chart: 

N = .29.6  - 20  log  Af/(jh  rmg  + 10  log  BJp  ♦ CF 
= +29.6  - 20  log  (140)  + 10  log  (25)  + 0 

= +29.6  - 42.9  + 14.0  + 0 = +0.7  dB  = +0.5  dB 

RSL  value  corresponding  to  0 dB  C/N: 

RSL  = -114.0  + NF  + 10  log  BT„ 

lr 

= -114.0  + 8.0  + 10  log  (25) 

= -114.0  + 8.0  + 14.0  = -92.0  dBm 

The  above  data  is  noted  on  the  following  page. 

Calculations  for  using  the  Gaussian  RF/IF  response  chart  with  the  TR0P0 
receiver  (excluding  effects  of  de-emphasis)  follow: 


148 


Receiver  Specifications: 


IF  bandwidth:  3-09  MHz 

per  channel  deviation:  100  kHz  rras 

(for  OdBm0  test  tone) 
noise  figure:  1.5  dB 

Noise  Slot  Measurement  (FSV)  Bandwidth:  3.1  kHz 

Noise  valve  corresponding  to  0 dB  slot  noise  on  chart: 

N = +29.6  - 20  log  Af/Qh  rms  + 10  log  BIp  + CF 
= +29.6  - 20  log  (100)  + 10  log  (3-09)  + 0 
= +29.6  - 40.0  + 4.9  = -5.5  dBm0 
RSL  value  corresponding  to  OdB  C/N: 

RSL  = -114.0  + NF  + 10  log  B 

= -114.0  + 1.5  + 10  log  (3-09) 

= -114.0  + 1.5  + 4.9  = -107.6  = -107.5  dBm 
The  above  data  is  noted  or.  the  following  page. 

Note:  For  ease  in  comparison  with  actual  quieting  curves,  it  is  suggested 

that  the  above  calculated  valves  be  rounded  to  the  nearest  J dB. 


149 


7.  KM  Microwave  Radio  Terminal  E'  prent  Parameters 

7.1  In  ord'-?r  to  predict  M/W  terminal  performance,  it  is  necessary  to 
determine  a few  characteristics  of  the  terminal  transmitter  and  receiver. 
The  determination  of  these  parameters  will  be  discussed  
…[truncated]