DTIC ADA063798: Analytical Determination of the Interference of Commercial FM Stations with Airborne Communication and Navigation Receivers and Experimental Verifications.

Survival, Water, Medical Field Manuals

Military Manuals

Defense Technical Information Center

Document text

AD-A063  798  OHIO  UNIV  ATHENS  DEPT  OF  ELECTRICAL  ENGINEERING  F/6  17/7 

ANALYTICAL  determination  OF  THE  interference  OF  COMMERCIAL  FM  S— ETC<U> 
MAR  78  J E ESSMAN.  T LOOS  D0T-FA77WA-3932 

UNCLASSIFIED  EER-32-1  FAA-R-6050-1  NL 


1 OF  2 

AD 

A063798 

■ 

O 

% 

■ 

HI 

a 

m 

11 
l 1 

MICROCOPY  RESOLUTION  TEST  CHARI 

NAIIONAl  BUM  All  Of  ! ANDARDS  ]9(  • A 


AD  AO  63  798 


ANALYTICAL  DETERMINATION  OF  THE  INTERFERENCE  OF  COMMERCIAL 
FM  STATIONS  WITH  AIRBORNE  COMMUNICATION  AND  NAVIGATION 
RECEIVERS  AND  EXPERIMENTAL  VERIFICATIONS 


CL_ 

o 

C3 


LU 


GS3 

5 


MARCH  1978 


FINAL  REPORT 


D D C 

@f?nnoia 

JAN  29  1979 


K^Einnsl 

er  b 


~ DISTRIBUTION  STATEMENT  A 

Approved  tor  public  rclcoMj 
Distribution  Unlimited 


U.S.  DEPARTMENT  OF  TRANSPORTATION 
FEDERAL  AVIATION  ADMINISTRATION 
Airway  Facilities  Service 
Washington,  D.C.  ?0590 


kmMiiMantiai 


Terhnicul  Report  Documentation  Page 


Apaiauauta  Canvaraiaaa  fits  Manic  Maaiata 


TABLE  OF  CONTENTS 


SUMMARY 
INTRODUCTION 
RECEIVER  MODEL 

MEASUREMENTS  OF  MODEL  PARAMETERS  AND 
APPLICATIONS 

A.  Input  Filter  Characteristics. 

B.  Determination  of  Model  Parameters. 

ELT  (EMERGENCY  LOCATOR  TRANSMITTER)  MODEL 


PAGE 


APPLICATION 


Calculation  Signal  Strengths. 

Calculation  of  Intermodulation  Contribution. 
Cross- Modulation  Considerations. 

Combined  Effects  of  Intermod  and  "Brute-Force" 
Interference . 


EXTENSIONS  OF  THE  TECHNIQUES 


CDI  MODEL 


A.  Analytical  Model. 

B.  Modeling  Receiver  Filters. 

C.  Effects  of  High  Level  Modulation  on  CDI  Reading. 
COMPUTER  PROGRAM 

CONCLUSIONS  AND  RECOMMENDATIONS 

A.  Statistical  Model. 

B.  Model  Improvement. 

C.  Standard  Test  Procedure  . 

D.  Interference  Profile  Mapping  of  Potential 
Airport  Faci  lities . 


63 

65  fa 

66  White  Section 
Cuff  Section  □ 

66  □ 
3H 


67  — 


jisRM|wrjrco3Es_ 

On'.  AV.V.l  frua^.v  SPiClST 


XI 


ACKNOWLEDGMENTS 


PAGE 


XII 

XIII 

XIV 


REFERENCES 

BIBLIOGRAPHY 

APPENDICES 

A.  Theoretical  and  Experimental  Measurements  Using 
Model  and  an  FM-Interfering  Signal. 

B.  Fi  Iter  Characteristics  . 

C.  Interference  Due  to  ELT. 

D.  Measurement  of  RMS  Modulation  Due  to  Intermod . 

E.  Noise  Equivalent  Bandwidth. 

F.  CDI  Interference. 

G.  User  Manual. 

H.  Subroutine  Descriptions. 

I.  Computer  Program  Listing. 


LIST  OF  FIGURES 


PAGE 


Figure  1.  (a)  Simplified  Block  Diagram  of  Receiver,  (b)  RF-Amplifier 

Model,  (c)  Overall  Model  from  RF-Input  to  IF  Output. 

Figure  2.  RF  Input  Fi Iter  Characteristics. 

Figure  3.  Relative  IF  and  Detector  Response. 

Figure  4a.  Variation  of  3 K^/2  Kj  as  a Function  of  Intermod  Generating 
Frequencies. 

Figure  4b.  Plots  of  Effective  3K^/K^  as  a Function  of  Interfering  Signal  Level. 
Figure  5.  Typical  Applications. 

Figure  6.  IF  Characteristics. 

Figure  7.  Signal  Levels  for  No  Interference. 

Figure  8.  Regions  of  Interference. 


Figure  9. 


’Skirt"  Characteristics. 


Figure  10.  Signal  Spectrum. 

Figure  11.  Birmingham,  Alabama  Area  Airport. 

Figure  12.  Location  of  FM  Broadcast  Stations  in  Birmingham  Area. 

Figure  13.  Map  Showing  Locations  of  FM  Stations  and  Locations  at  Which 
the  FM  Station  Interference  is  Analyzed. 

Figure  14.  Block  Diagram  of  Audio  Processing  Circuit  of  NAV  11 . 

Figure  15.  Simplified  Schematic  of  Comparator  Circuitry  of 

NAV  1 1 Receiver. 

Figure  16  . 90/150  Filter  Responses. 

Figure  17.  Normalized  Detector  Filter  Characteristics. 


v 


PAGE 


f 


Figure  18.  Audio  Filter  Characteristic*.  58 

Figure  A-l  . Interfering  Signal  Spectrum;  Modulation  Index  6; 

Bandwidth  '200  KHz.  69 

Figure  A-2  . Spectrum  of  Desired  Signal  (108.5  MHz)  at  RF-Amplifier 

Output.  Signal  - 17  dBm  70 

Figure  A-3 . Spectrum  of  Desired  Signal  (108.5  MHz)  at  RF-Amplifier 

Output.  Signal  -47  dBm  71 

Figure  A-4.  Detector  Output  with  No  Interference.  71 

Figure  A-5.  Detector  Output  with  Interfering  FM-Signal  (0  dBm)  at 

105.5  MHz.  72 

Figure  A-6.  Detector  Output  Using  Input  Signals  Described  in 

Figure  A-2  , with  10%  Modulation  on  Desired  Signal.  72 

Figure  A -7.  Detector  Output  Desired  Signal  Level  =-70  dBm,  No 

Interfering  Signal.  73 

Figure  A-8.  Detector  Output.  73 

Figure  B-l  . Measurement  Procedure  for  Obtaining  a(u).  74 

Figure  B-2  . Simplified  Schematic  of  RF  Front  End  of  NAV  11  Receiver.  74 

Figure  B-3 . Test  Procedure  for  Determining  a(u)  Using  Cross- 

Compression  Methods.  75 

Figure  C-l  . Spectra  of  the  Reradiated  Signals  at  the  Output  of  ELT.  78 

Figure  C-2  . Test  Procedure  for  Spectral  Measurements  for  Figure  C-l  . 83 

Figure  C-3.  ELT  Model.  84 

Figure  C-4 . Experimental  Procedures.  85 

Figure  C-5.  IM  Effects  of  Two  Closely  Aligned  Frequencies.  88 

Figure  C-6  . ELT  Output  Voltage  as  a Function  of  Input  Frequency.  90 

Figure  C-7.  k^  as  a Function  of  Intermod  Frequency  for  Varying  FM 

Band  Generating  Frequencies.  91 

Figure  C-8  . Variation  of  k^  for  Different  Signal  Levels.  92 


VI 


PAGE 

93 


\ 


Figure  C-9.  Antenna  Coupling  Measurement. 

Figure  C-10.  Antenna  Locations. 

Figure  C-l  1 . ELT-Communi cation  Antenna  Coupling. 

Figure  C-12.  Interference  Model. 

Figure  C-13.  Coupling  Between  ELT  Antenna  and  Communication 
Antenna  as  a Function  of  Frequency  22"  ELT  Antenna 
Length. 

Figure  C-14.  Normalized  (50  ohms)  Self  Impedance  at  Various 
Frequencies  of  the  ELT  Antenna  (22")  on  Board 
N1200U. 

Figure  C-l  5.  Normalized  Self-Impedance  of  ELT  Antenna  (16" 
Modified  Antenna)  on  Board  N900U. 

Figure  C-16.  Normalized  (50  ohms)  Self  Impedance  at  Various 
Frequencies  of  the  Communications  Antenna  on 
Board  N1200U. 

Figure  E-l  . Noise  Equivalent  Bandwidth  of  a Filter. 

Figure  G-l  . Flow-Chart  of  FM  Interference  Program. 

Figure  H-l  . Subroutinesand  Functions  Called  by  the  Main  Program. 

Figure  H-2  . Assumed  Attenuation  Characteristics. 

Figure  H-3  . Flow  Chart  of  Subroutine  CONVOL,  Section  Where  AXB 
is  Convolved  with  Station  " K"  . 


94 

95 
98 

101 

102 

103 

104 
110 
115 

120 

123 


LIST  OF  TABLES 


PAGE 


Table  1 . 

Possible  Intermod  Components  (IM). 

4 

Table  2. 

3K  /2K  Determined  by  Cross-Modulation  Measurements 
wifn  10  KHz  AM  Tone  Modulation  on  Interfering  Signal. 

15 

Table  3. 

Parameters  of  Model  Determined  by  Intermodulation 

Technique. 

16 

Table  4. 

Parameters  of  Model  Determined  by  Cross- Compression 
Measurements. 

17 

Table  5. 

Summary  of  Results. 

17 

Table  6. 

2 

Coefficients  of  the  FM  signals  e^  (t)  and  ec  (t)  for 

R]  =2.4,  f]  = 10.42  KHz,  p2  = 2*4  and  f2  = ,2‘5  KHz. 

24 

Table  7. 

IM  Spectrum. 

25 

Table  8. 

IM  Components  Out  of  IF. 

27 

Table  9. 

Summary  of  Measured  and  Calculated  Results. 

32 

Table  10. 

FM  Broadcast  Stations  in  the  Birmingham  Area. 

40 

Table  1 1 . 

Airport  Facilities. 

40 

Table  12. 

Distance  of  FM  Broadcast  Stations  from  Localizer. 

40 

Table  13. 

Printout  of  Signal  Level  Computations. 

41 

Table  14. 

Signal  Levels  at  Location  (1). 

44 

Table  1 5. 

Cross-Modulation  Due  to  Interfering  Signals.  Location  (1). 

44 

Table  16. 

Intermodulation  Interference.  Receiver  at  Point  (1)  location. 

45 

Table  17. 

Signal  Levels  at  LOM  Location. 

45 

Table  18. 

Cross-Modulation  Due  to  Interfering  Signals.  LOM  Location. 

46 

Table  19. 

Intermodulation  Interference.  Receiver  at  LOM  Location  2. 

46 

Table  20. 
Table  21 . 
Table  22. 
Table  23. 
Table  24. 
Table  25. 
Table  26. 
Table  27. 
Table  28. 


Table  B-l 


Table  B-2 . 
Table  E-l . 


Table  F-l 


Table  F-2 , 


Table  F-3. 


Table  F-4 . 


Table  H-l  . 
Table  FI-2 . 
Table  FI-3 . 


Signal  Levels  at  Location  (2)  in  Figure  13. 

Cross- Modulation  Due  to  Interfering  Signals.  Location  (2). 
Intermodulation  Interference.  Receiver  at  Location  (2). 
Signal  Levels  at  Location  (3). 

Cross-Modulation  Due  to  Interfering  Signal.  Location  (3). 
Intermodulation  Interference.  Receiver  at  Location  (3). 
Measurements  of  CDI  Voltages. 

Reference  Notes  Relative  to  Measurements  in  Table  26s 


Summary  of  RMS  Modulation  Measurements  and 
Calculations. 


Comparison  of  Results  Obtained  Using  Cross-Compression 
and  RF  Measurements. 


Variations  in  a (u)  as  a Function  of  Receiver  Frequency. 


Comparison  of  Theoretical  Modulation  Factor  Ratios 
from  Computer  Program. 


Effects  of  Detector  Clipping  on  CDI  (Normal  Reading  of 
CDI  = 60  pA). 


Effects  of  Detector  Clipping  on  CDI  (Normal  Reading  of 
CDI  = 30  pA) . 


Effects  of  Detector  Clipping  on  CDI  (Normal  Reading  of 
CDI  = 0 pA) . 


Effects  of  Interfering  Signal  Frequency  on  CDI  (Normal  Reading  of 
CDI  = 60  pA). 


Variable  Names  and  Their  Meaning. 

Edit  Mode  Commands. 

Variable  Names  for  Subroutine  XMOD. 


IX 


PAGE 


4~ 

47 

47 

49 

49 

49 

53 

54 


59 


75 

76 


111 


112 


112 


113 


113 

121 

122 

126 


Jl 


I 


SUMMARY 


This  paper  summarizes  the  work  performed  under  Contract  FA77WA-3932  sponsored  1 

by  the  Federal  Aviation  Administration  (FAA),  Airway  Facilities  Service.  The  primary 
objective  of  the  research  sponsored  here  is  the  determination  of  an  analytical  model 
capable  of  predicting  the  effects  of  strong  interfering  FM  broadcast  stations  on  airborne 
navigational  (specifically  the  lo.jlizer  receiver)  and  communication  receivers. 

An  analytical  model  of  the  RF-amplifier  stage,  using  a third-order  nonlinearity, 
of  airborne  communication  and  navigation  receivers  capable  of  predicting  the  effects  of 
multiple  interfering  FM  stations  is  developed.  The  effects  of  receiver  input  filters  and 
IF  filtering  are  modeled.  Measurement  techniques  useful  in  determining  the  various 
parameters  of  the  model  are  described  and  evaluated.  The  effects  of  different  signal 
powers  and  different  signal  frequencies  on  the  parameters  are  investigated.  Limits  are 
postulated  on  the  accuracy  of  the  model  and  regions  of  validity  defined. 

"Brute-force"  interference  due  to  strong  interfering  FM  stations  which  saturates  the 
front  end  and  intermodulation  distortion  due  to  the  interaction  of  two  or  more  FM  stations 
are  considered. 

Various  signal  characteristics  are  considered  for  evaluating  the  degree  of  distortion 
introduced  by  interfering  FM  stations.  The  effects  of  filtering  due  to  the  receiver  are 
considered  and  the  resulting  FM-to-AM  conversion  analytically  described.  Signal  char- 
acteristics considered  for  evaluating  distortion  include  peak  (cross-)  modulation  indices 
and  RMS  output  of  an  envelope  detector.  The  effects  of  receiver  desensitation  due  to 
saturation  of  the  front  end  by  the  interfering  signals  are  considered.  A substantial  amount 
of  time  and  effort  was  expended  in  experimental  testing,  on  the  bench,  and  the  models 
were  developed.  Although  all  the  experimental  results  are  not  reported  here,  the  most 
important  results  are  documented  here. 

A computer  program  was  developed  and  is  described  here  which  theoretically 
predicts  the  effects  of  multiple  interfering  FM  stations.  Input  data  include  FM  station 
frequencies,  powers,  location,  and  other  signal  parameters,  and  calculations  are  made 
of  the  RMS  AM  modulation  at  the  receiver  filter  outputs.  Examples  are  presented  which 
illustrate  the  use  of  the  techniques  developed.  Tradeoff  studies  indicating  the  effects  of 
the  creation  of  additional  FM  stations  end/or  the  results  to  be  expected  when  one  or  more 
stations  change  their  power  levels,  location,  antenna  patterns,  etc.  are  illustrated  by 
example.  The  number  of  FM  stations  that  can  be  considered  is  limited  only  by  the  com- 
puting facilities  available  to  the  user. 

Decision  criteria  that  can  be  reliably  used  in  predicting  potential  interfering 
sources  are  considered. 

The  results  of  the  interference  on  the  Course  Deviation  Meter  (CDI)  of  a localizer 
receiver  are  investigated  using  single  tone  FM  modulation  of  the  carriers.  Due  to  the 


aasasss 


9 


narrow  bandwidths  of  the  90  and  150  Hz  filters  the  CD!  response  due  to  interference  would 
be  extremely  dependent  on  the  modulation  tone  being  assumed.  To  alleviate  this  and  make 
the  observations  more  general,  "white  noise"  with  a constant  spectral  density  is  used  to 
evaluate  the  /~DI  response. 

The  results  are  applied  to  specific  airports  where  interference  problems  have 
occurred  and  the  results  obtained  by  the  theoretical  model  are  analyzed. 


II  INTRODUCTION 

Interference  has  always  been  an  important  problem  area  in  communications  systems. 
With  increased  power  levels  of  many  transmitters  and  the  solid  state  design  of  modern 
receivers,  interference  due  to  high  power  stations  driving  receivers  into  nonlinear  operation 
is  an  increasing  problem.  This  has  been  the  case  recently  where  there  have  been  many 
reported  cases  of  commercial  FM  broadcasting  station  interference  with  airborne  commun- 
ication and  navigational  receivers.  The  Federal  Aviation  Administration  (FAA)  has 
become  concerned  with  the  increased  number  of  reports  of  interference  with  airborne 
receivers.  The  research  described  in  this  report  summarizes  the  efforts  in  developing  a 
useful  analytical  model  of  a localizer  receiver  having  the  following  two  properties: 

(1)  The  parameters  required  to  specify  the  model  should  be  easily  measured 
or  otherwise  capable  of  being  determined  from  manufacturers'specifications . 

(2)  The  analytical  model  should  be  as  simple  as  possible;  however,  it  should 
be  useful  in  predicting  potential  interference  problems  due  to  multiple  interfering  sources. 

In  essence  the  objective  is  to  develop  an  analytical  model  capable  of  predicting  possible 
interference  problems  due  to  existing  FM  stations,  additional  sources  beyond  those 
currently  in  existence,  changes  in  existing  sources,  etc. 

Techniques  for  describing  nonlinear  distortion  characteristics  were  emphasized 
in  the  early  1950's  when  cable  television  (CATV)  engineering  had  its  beginning.  Further 
emphasis  in  this  difficult-to-handle  area  were  spawned  by  the  increased  use  of  satellite 
communications  systems,  low  cost  solid  state  receiver  circuitry,  etc.  Many  investigations 
have  represented  the  nonlinear  elements  as  having  zero  memory;  - however,  recently 
interestiias  been  revived  in  the  volterra  series  which  was  introduced  by  Wiener  around 

1942. 

For  the  particular  investigation  described  here, it  was  decided  to  use  a simple 
power  series  for  the  transfer  characteristic  for  the  active  device  in  the  RF  amplifier,  i .e 


K.  e.  + K0  e.  + K0  e. 

i £■  i o I 


( 1 ) 


where  e-  is  the  instantaneous  input  voltage  to  the  active  device,  eQ  is  the  output 
voltage  of1  the  nonlinear  element,  Kj  , and  are  complex  numbers  describing  the  gain, 


-2- 


phase  shift  and  distortion  characteristics  of  the  device.  Although,  in  general,  any  phase 
angle  can  be  specified  by  a proper  choice  of  the  constants  Kj  , and  K^,  for  this  work 
we  assume  the  phase  angle  on  and  to  be  0°  and  the  phase  angle  on  to  be  either 
0°  or  180°.  These  assumptions  give  a simple  worst  case  model  in  which  the  parameters  of 
the  model  are  easily  determined  by  measurements.  Procedures  are  described  which  can 
easily  be  applied  to  most  receivers  for  determining  the  needed  parameters.  This  simplified 
model  was  deemed  to  be  adequate  in  view  of  the  wide  variations  in  receivers  from  manu- 
facturer to  manufacturer  and  even  from  receiver  to  receiver  of  the  same  type.  In  addition 
characteristics  of  a particular  receiver  will  change  significantly  with  age,  temperature, 
AGC,  etc.  Measurements  indicate  that  even  with  all  these  possible  variations  the  results 
can  be  accurately  reproduced  within  approximately  3 dB  which  is  assumed  to  be  of  suffi- 
cient accuracy  in  practice. 


The  terms  generated  which  are  of  concern  can  be  identified  by  considering  the 
first-and  third-order  terms  given  in  (1),  i.e., 


(1 ) Linear  gain  term 

e . = K,  e. 
ol  1 


(2  ) 


(2) 


Third-order  term 

3 


'o3 


K3  e. 


(3  ) 


The  distortion  terms  can  be  illustrated  by  assuming  the  input  to  be  a desired  signal 
(e.  = A cos  a t ) with  angular  frequency  a and  two  interfering  signals  of  angular  frequencies 


b and  c respectively  (e^  - B cos  b t and  e.^  = C cos  c t),  i.e.. 


e.  (t)  ~ A cos  a t + B cos  (b  t + <t>  ) + C cos  (c  t + <t>  ) 

i I l 


( 4a  ) 


Using  the  inputs  given  in(4q)  each  of  the  signals  can  be  assumed  amplitude  frequency  or 

phase  modulated  by  letting  the  amplitudes  A,B,C,  the  frequencies  a,b,c  or  the  phases 

and  functions  of  time  respectively.  Whichever  happens  to  be  the  case, we  can 

assume  the  bandwidths  of  the  signals  to  be  B , B,,  and  B respectively. 

a b c 


Using  (4a)  the  possible  intermodulation  terms  of  interest  are  listed  in  Table  1 . For 
this  table,  the  maximum  FM  station  bandwidth  is  assumed  to  be  240  KHz.  In  addition  to 
the  distortion  terms  listed  in  this  table,  the  third-order  nonlinearity  will  contribute  self- 
compression or  self-expansion  terms  at  the  desired  frequency  (a)  depending  on  whether 
is  negative  or  positive,  i.e., 


3/4  K3  A~ 


cos  at 


(4b) 


-3- 


Similarly  cross-compression  or  cross-expansion  terms  generated  at  the  desired  frequency 
occur  from  the  following  terms: 

3/2  AB2K  cosat  t3/2AC2K  cos  at  ( 4c  ) 

O j 

All  other  terms  are  assumed  filtered  out  by  the  receiver.  It  should  be  noted  rhat  even 
if  no  intermodulation  components  generated  are  within  the  passband  of  the  receiver  as  given 
in  Table  1 , there  still  existsa  possibility  of  interference  due  to  "brute-force"  interference  as 
indicated  by  (4b)  and  (4c)  . This  type  of  interference  shows  up  as  cross-modulation  which 
is  a phenomenon  where  the  modulation  of  an  interfering  carrier  is  impressed  upon  another 
carrier.  This  "brute-force"  type  of  interference  can  be  important  when  the  receiver  is  in 
the  presence  of  strong  interfering  signals. 

Interference  of  the  types  discussed  above  can  also  have  the  effect  of  desensitizing 
the  receiver.  This  is  the  result  of  the  AGC  reacting  to  the  self-and/or  cross-compression 
of  the  carrier . 

Ill  RECEIVER  MODEL 

In  general  it  is  the  RF  amplifier  stage  where  most  of  the  nonlinear  distortion  is 
generated;  however,  it  is  possible  and  in  the  presence  of  extremely  strong  interfering  signals 
it  would  be  expected  that  the  mixer  stage  would  contribute  to  the  nonlinear  distortion.  In 
this  investigation  we  assume  the  RF-amplifier  to  be  the  significant  source  of  nonlinear 
distortion;  therefore,  an  adequate  model  of  this  stage  is  necessary  and  sufficient  for  our 
purposes.  Figure  la  indicates  a block  diagram  of  a typical  receiver  and  Figure  lb  the  model 
of  the  RF-amplifier  used  in  this  work,  where  the  nonlinear  device  input-output  is  specified 
by  the  transfer  characteristic  given  in  (1).  The  input  filter  block  is  assumed  to  represent 
the  effects  of  the  input  circuitry,  while  the  output  filter  block  represents  the  effects  of  the 
output  circuitry  of  the  RF-amplifier.  In  the  frequency  domain  input  and  output  attenuations 
are  represented  by  a(u)  andp  (u»)  respectively.  If  one  assumes  all  the  nonlinear  terms  to  be 
generated  in  the  RF-amplifier,  then  the  model  shown  in  Figure  lc  can  be  assumed.  Using 
this  model  the  input  is  assumed  to  be  the  RF-amplifier  input  (voltage  at  the  output  of  the 
receiver  antenna)  and  the  output  is  the  IF-amplifier  voltage  output.  In  this  case  the  output 
filter  will  include  the  attenuation  properties  of  the  RF-amplifier  output  circuitry,  the  mixer 
and  the  IF-amplifiers.  Since  no  distortion  terms  are  assumed  generated  with  the  mixer 
stage,  the  effects  of  this  stage  is  assumed  to  be  a linear  frequency  translation  and  all 
calculations  and  modeling  can  be  done  at  either  the  IF  or  RF  frequency,  whichever  is 
convenient.  This  representation  has  significant  advantages  in  practice  in  determining  the 
parameters  experimentally.  Methods  were  developed  which  allow  the  overall  parameters 
specified  in  Figure  lc  to  be  determined  by  simply  monitoring  the  AGC  voltage  with  an 
interfering  and  desired  signal  present. 


-5- 


RF  INPUT 


RF  INPUT 


RF 

MIXER 

IF 

AMr 

INPUT 

FILTER 

a(u) 


4 

L 

< 

. i 

V .. 

N N 


_RF_-_AMP 

(b) 


NON  LINEAR 
DEVICE 
«!,  K3 


INPUT 

Li:  ;E  \R 

DEV 

OUTPUT 

FILTER 

a(u) 

Kr  ' 3 

FILTER 

P(u) 

RF  OUTPUT 


OUTPUT 

FILTER 

P(«) 

IF  OUTPUT 


Figure  1.  (a)  Simplified  Block  Diagram  of  Receiver. 

(b)  RF-Amplifier  Model,  (c)  Overall  Model 
from  RF-Input  to  IF  Output. 


Using  the  model  specified  by  Figure  lb  requires  measurements  of  the  RF-amplifiers 
output,  which  in  general  are  very  sensitive  measurements  due  primarily  to  loading  effects. 
Both  methods  were  investigated. 

IV  MEASUREMENTS  OF  MODEL  PARAMETERS  AND  APPLICATIONS 

Application  of  the  receiver  model  requires  that  the  signal  level  at  the  receiver 
input  be  known.  In  addition  the  following  information  regarding  the  receiver  characteristics 
is  needed: 

( 1 ) Input-filter  characteristics  a(u). 

( 2 ) Kj  and 

( 3 ) Output  filter  characteristics  from  the  RF  to  the  output  of  the  IF- 
amplifier  stages.  With  this  information  one  can  calculate  the  resulting  distortion  terms  at 
the  output  of  IF-amplifier  stages. 

With  this  information  the  model  can  be  used  to  predict  the  output  from  the  IF- 
amplifier  and  the  output  of  the  envelope  detector.  In  a later  section  of  this  report  an 
analytical  model  of  the  CDI  circuitry  is  described  and  developed.  Since  the  interest  in 
this  research  was  primarily  concerned  with  the  interference  of  commercial  FM  broadcasting 
stations  with  aircraft  localizer  receivers,  the  examples  and  applications  are  in  this  area. 

The  receiver  used  for  experimental  work  was  Navll  localizer  receiver  manufactured  by 
NARCO  Corporation.  The  frequency  range  of  the  localizer  receiver  is  108  MHz  to  118 
MHz,  while  the  commercial  FM  broadcasting  band  is  from  88  to  108  MHz.  As  a result  of 
the  adjacent  location  of  these  two  frequency  bands, both  intermodulation  distortion  and 
"brute  force"  type  interference  must  be  considered. 

A.  Input  Filter  Characteristics.  Figure  2 indicates  a plot  of  the  receiver  RF- 
amplifier  input  filter  function.*  TFe"  receiver  is  tuned  to  108.5  MHz.  The  effects  of  the 
output  circuitry  of  RF-amplifier  can  be  combined  with  the  IF-amplifier  and  detector 
characteristics  so  that  one  overall  attenuation  function  is  all  that  is  required  and  since  most 
of  the  band  limiting  is  done  in  the  IF-amplifier,  these  characteristics  are  the  most  important. 
These  characteristics  are  normally  available  from  manufacturers'  data;  however,  if  not, 
they  are  easily  measured. 

Figure  3 i llustrates  the  overall  attenuation  characteristics  of  the  NAV  1 1 localizer 
receiver  from  the  IF  input  to  the  detector  output.  Some  discrepancies  were  observed  between 


*The  measurement  techniques  used  to  obtain  these  results  are  described  in 
Appendix  B. 


-7- 


Relative  Response 
in  dB 


to 

m 

m 

m 

in 

in 

in 

in 

O' 

O 

o 

o 

o 

8 

CO 

o 

s 

m 

o 

o 

o 

I I I I l I I I 


Receiver  Tuned 
to  108.5  MHz 


(A 

/ 0S 


/ \ 

,®  BW  = 4 MHz  © 


-10  - 


Characteristics  taken  using  Vector 
Voltmeter.  Assuming  constant 
incident  voltage  and  measuring  voltage 
at  RF  AMP  input  at  Mosfet  Gate  input 
as  function  of  frequency. 


Figure  2.  RF  Input  Filter  Characteristics. 


-8- 


cfor  Response. 


measured  values  and  theoretical  values  using  an  Ideal  Bandpass  Charcteristic . To 
alleviate  this  a higher  order  approximation  of  the  IF  characteristics  was  used.  In  par- 
ticular a triangular  approximation  was  developed  and  is  shown  in  Figure  3. 


Since  the  interest  in  this  research  was  primarily  with  regard  to  commercial  FM 
broadcast  station  interference  with  navigational  receivers  and  due  to  the  fact  that  the 
navigational  information  is  conveyed  via  AM  techniques,  the  attenuation  characteristics 
of  the  receiver  are  important.  This  is  due  to  the  fact  that  the  effect  of  bandlimiting  the 
FM  interfering  signal  is  to  introduce  AM  modulation  on  it  which  causes  significant 
interference  at  the  output  of  the  AM  detector.  A considerable  amount  of  experimental 
data  was  collected  during  the  execution  of  this  project.  Appendix  A gives  representa- 
tive spectral  plots  illustrating  these  effects. 


B.  Determination  of  Model  Parameters.  Using  the  analytical  model  requires  that 
the  parameters  and  1C  be  determined.  Actually  it  is  only  necessary  to  determine  the 
ratio  Kg/K^  • Three  methods  were  used  to  determine  these  parameters. 

(1)  Gain  Compression  Technique  (two-carrier  method)  - in  this  scheme  two 
sinusoids  are  applied,  one  at  the  desired  receiver  frequency  (a)  and  the  other  an  inter- 
fering signal . 

Using  (1)  and  (4c)  (with  c = 0)  the  third-order  term  predicts  a cross-modulation 
term  at  the  desired  frequency  which  is  proportional  to  the  product  Kg  AB^a  (b  ) where. 

Kg  = parameter  of  the  third-order  model 
A - amplitude  of  the  desired  signal 
B = amplitude  of  the  interfering  signal 
c{u)  = input  filter  characteristic 

2 

The  desired  signal  amplitude  A is  assumed  small  compared  with  AB  so  that  the 
contribution  due  to  the  AJ  term  in  (4b)  can  be  neglected.  The  measurement  of  the  ratio 
using  this  technique  was  performed  as  follows: 

(a)  Measure  the  AGC  voltage  with  only  a desired  signal  input, 

e.  = \/~2  A cos  at  (A  is  RMS  Voltage)  ( 5 ) 

Neglecting  the  self- compress  ion  term  in  (4b),  the  output  is  approximately 

e = ^7  AK.  cos  at  ( 6 ) 

o I 

(b)  Applying  an  interfering  signal  along  with  the  desired  signal.  The 
result  of  the  interfering  signal  is  to  compress  the  desired  signal  or  expand  the  desired  signal 
depending  on  whether  Kg  is  negative  or  positive. 


J 


Using  (4c)  and  choosing  those  terms  of  interest  at  the  receiver  frequency  we  have, 

e.  v/2  A cos  a t i-  v 2 B cos  b t ( 7 ) 

i 

e v/2  AK,  0 i 3B?a2(b)  K_/K,)  cosat  ( 8 ) 

(c)  Increase  the  desired  signal  level  until  the  AGC  voltage  is  the  same 
as  in  step  (a).  The  amount  of  dB  increase  in  desired  signal  required  to  bring  the  AGC  voltage 
back  to  the  value  in  step  (a)  is  the  amount  of  gain  change  of  the  desired  signal  level  due  to 
the  interfering  signal.  Hence,  using  (6)  and  (8)  the  ratio  is  easily  determined. 

(2)  Two-carrier  method  with  modulation  - this  method  uses  an  interfering 
AM-modulated  signal  along  with  a desired  CW  signal  as  the  input.  By  measurement  of 
the  cross-modulation  obtained  the  ratio  of  K^/Kj  can  be  determined.  The  steps  in  the 
procedure  are  as  follows: 

(a)  With  a desired  signal  and  no  interfering  signal  applied,  measure 

the  AGC  voltage. 

(b)  Applying  a combination  of  a desired  signal  and  an  interfering 
signal  with  a known  percent  modulation, observe  the  AGC  voltage. 

(c)  Increase  or  decrease,  whichever  is  required,  the  desired  signal 
level  until  the  AGC  voltage  is  the  same  as  in  step  (a). 

(d)  Measure  at  the  IF-amplifier  output  the  AM  sidebands  due  to  the 
cross-modulation  of  the  interfering  signal  on  the  desired  signal.  This  gives  the  amount  of 
cross-modulation  obtained  which  can  be  used  to  determine  K^/Kj  (note  that  cross- 
compression of  the  carrier  must  be  taken  into  account). 

Mathematically  the  procedure  can  be  described  as  follows:  Let  the  input  be 

e.  = v/5  A cos  a t + v/5  B (t)  cos  b t ( 9 ) 

where  B (t)  = B,  |1  + m cos  u t|  ( 10  ) 

1 m 

m = percent  modulation 

w = frequency  of  modulation 
m 

2 2 

Considering  the  third-order  term  and  neglecting  the  terms  containing  m and  B 
the  result  is 

K3  o 2 K3 2 2 

e = AK.  (1  + 3f^T  o (b)  B.  + 6-ir-a  (b)  B.  m cos  <j  t)  cos  at  (11) 

o I I I 1 ' 1 I m 

(3)  Intermod  Method  - this  method  is  based  on  simply  measuring  the  level 
of  an  intermodulation  component  generated  by  two  interfering  signals  of  frequencies  such 


-11- 


that  they  produce  an  intermod  component  at  the  desired  frequency.  These  measurements 
can  be  accomplished  as  follows: 

(a)  Apply  a desired  signal  of  known  amplitude  and  measure  the  AGC 

voltage . 

(b)  Without  a desired  signal  apply  the  two  interfering  signals  and 
adjust  the  amplitudes  such  that  the  AGC  voltage  is  the  same  as  in  step  (a). 

(c)  The  desired  signal  level  in  (a)  is  then  equal  to  the  intermod  level. 

The  three  methods  above  can  also  be  modified  such  that  the  AGC  voltage  is  held 
constant  and  the  levels  at  the  IF-amplifier  output  are  measured  using  a spectrum  analyzer. 

Specifying  a receiver  by  such  a model  brings  up  the  question  of  how  much  variation 
in  ratio  K^/Kj  is  there.  Most  past  work  has  been  concerned  with  the  modeling  of  wideband 
amplifiers  where  problems  with  variations  in  the  AGC  characteristics  are  not  present.  In 
addition,  the  ratio  K3/K]  could  be  a function  of  input  signal  levels  and  input  signal 
frequencies,  etc.  A considerable  number  of  experimental  measurements  were  obtained 
using  the  NAVll  localizer  receiver  to  investigate  these  characteristics.  Figure  4a  shows 
the  results  obtained  using  the  intermod  method.  These  results  give  an  indication  of  the 
variations  in  3 K^/  (expressed  in  dB)  as  a function  of  the  intermod  generating  frequencies. 
Indicated  on  this  figure  are  the  signal  levels  of  the  two  interfering  signals.  Figure  4b 
indicates  the  results  obtained  as  a function  of  interfering  signal  level. 


If  the  model  is  consistent  all  three  measurement  techniques  should  give  relatively 
the  same  results  for  K^/Kj  . Comparison  of  the  results  obtained  using  the  three  different 
measurement  techniques  are  given  in  Tables  2,  3 and  4.  These  results  again  indicate  that 
the  ratio  K_/Kj  can  be  assumed  essentially  constant  in  regions  where  the  third-order 
model  can  be  assumed  valid. 

Although  the  amount  of  data  given  in  these  tables  is  limited,  some  important 
observations  and  conclusions  can  be  made  and  are  given  in  Table  5. 

The  measurements  seem  to  indicate  that  there  are  variations  which  can  probably  be 
attributed  to  the  following: 

(a)  The  sensitivity  of  the  measurements. 

(b)  The  receiver  characteristics  changing. 

(c)  The  order  of  the  model  not  being  sufficient. 


Signal  levels  required  to  give  the  data  3.  Filter  function  used  to  calculate  31^/2  Kj 

points  are  given  in  parentheses.  The  first  is  shown  in  Figure  2 . 

is  the  b frequency  signal  level  in  dBm 
incident,  the  second  is  the  c signal  level. 


Figure  4a.  Variation  of  3K^/2K|  As  a Function  of  Interi.  Generating  Frequencies. 


□DK 


Desired  Signal 


Interfering  Signal  Gain  Chang 

Level  (-dBm)  B GC(-dB) 
105.5  MHz 


AGC 

(Volts) 


Comments 


1 .2  1 .727 


1.1  1 .498 


0.9  1 .376 


1 .0  1 .299 


1 .218 


-1.8  1 .198 


-1.2  1.177 


None  measured  by 
this  method . 

Gain  expansion- 
model  not  valid 

Same  as  above 


3/2  K3/K]  = 20  log  (1  - GC')  - 2BdBm  - 2a  (b)dB  - 6dB 

d B 

where  GC’  = 10GCdB/20 

Localizer  signals  of  these  levels  would  be  unusual. 


Table  4.  Parameters  of  Model  Determined  by  Cross-Compression 
Measurements . 


Gain 

Compression 

Technique 


Cross  - 
Modulation 
Technique 


Inter- 

Modulation 

Technique 


Average  Value 
3K3/2K  j (dB) 


Maximum  deviation 
from  the  mean  (dB) 


Table  5.  Summary  of  Results 


The  following  conclusions  can  be  drawn  from  the  data  obtained: 


(a)  The  model  is  useful  for  localizer  signal  levels  at  -30  dBm  or  lower. 

(b)  The  model  is  useful  for  "brute-force”  interfering  FM  signal  levels 
of  - 10  dBm  or  lower. 

(c)  The  model  is  useful  for  interfering  signals  of  the  appropriate 
frequencies  such  that  an  intermod  is  generated  with  strength  of  -30  dBm  or  less.  From 
Table  3 it  is  observed  that  this  requires  input  signal  levels  less  than  -7.5  and  -10  dBm 
(referenced  at  the  receiver  input  terminals). 

(d)  The  value  of  K^/K^  can  be  assumed  essentially  constant  for  the 
regions  specified  by  (a),  (b),  and  (c). 

V ELT  (EMERGENCY  LOCATOR  TRANSMITTER  ) MODEL 

During  the  course  of  this  investigation,  it  was  found  that  a considerable  interference 
problem  was  encountered  on  aircraft  carrying  certain  ELT's.  Substantial  effort  and  time  was 
expended  investigating  these  effects  and  in  the  modeling  of  this  device.  The  results  of  these 
investigations  are  documented  in  Appendix  C. 


VI  APPLICATION 

To  illustrate  the  procedures  involved  in  determining  possible  interference  problems 
using  the  model  proposed,  consider  the  simple  case  where  there  exist  only  two  interfering 
FM  stations.  The  geometry  of  the  problem  is  assumed  to  be  that  shown  in  Figure  5. 

The  analysis  of  the  problem  can  be  separated  into  three  specific  topics,  i.e., 

(a)  Calculation  of  signal  strengths  at  the  receiver  input  terminals. 

(b)  Calculation  of  "brute-force"  interference  contributions. 

(c)  Calculations  of  intermodulation  distortion. 

A.  Calculation  of  Signal  Strengths.  The  free  space  attenuation  between  lossless 
isotropic  antennas  is  given  by 

a (dB)  = 36.3  + 20  log1()f  + 20  log1()d  (12) 


-18- 


r 


where  f is  the  frequency  in  MHz  and  d is  the  distance  in  miles.  Using  (12)  the 
path  attenuations  can  be  calculated  as: 

(a)  Path  attenuation  for  station  1:  a^  100  dB 

(b)  Path  attenuation  for  station  2:  ~ 91  dB 

(c)  Path  attenuation  for  station  3:  a ^ ' 91  dB 

The  maximum  power  available  at  receiver  antenna  is  given  by 

P = (ERP).  - (Attenuation).  (13) 

ai  i i 

where  (ERP).  is  the  effective  radiated  power  from  station  1 and  (Attenuation), 
is  the  attenuation  ot  the  i signal.  For  the  example  here,  we  have, 

P = 80  - 100  = - 20  dBm  (power  available  from  station  1) 
al 

P =80  - 91  = - 11  dBm  (power  available  from  station  2) 
a2 

P .=42  - 91  = - 49  dBm  (power  available  from  station  3) 
a3 

Applying  the  model  requires  a knowledge  of  the  voltage  input  at  the  receiver 
input  terminals.  Therefore,  the  power  input  to  the  receiver  from  signal  i is  given  by 


P.  = P . - L. 

i ai  i 


(14) 


where 

L.  = A.  + C.  - D.  (total  loss  due  to  antenna  cabling  system  on 
board  the  aircraft.  (15) 

A.  = loss  due  to  aircraft  antenna. 

C.  = loss  due  to  cabling,  etc. 

D.  = directive  gain  of  antenna. 

In  general  A.  and  C.  would  have  to  be  estimated  or  measured  for  a particular 
application.  For  our  example  we  use  measured  values  obtained  using  a VOIVNAV 
antenna  on  a Cessna  150  aircraft,  i.e.. 


A,  + C1  = 9 dB  (102.5  MHz) 
A2  + C2  = 7 dB  (105.5  MHz) 
A3+C3=3dB  (108.5  MHz) 


(16a) 

(16b) 

(16c) 


The  directive  gain  of  the  dipole  antennas  are  assumed  to  be  6 dB;  therefore,  the 
signal  levels  at  the  input  to  the  receiver  are: 


j 


-20- 


(17a) 

(17b) 

07c) 


FM1:  Pj  - 20  - 9 +-6  =-  23  dBm  (17a) 

FM2:  P2  = - 11  - 7 +-6  = - 12  dBm  (17b) 

Localizer  : P^  = - 49  - 3 + 6 = - 46  dBm  (17c) 

With  this  information  the  next  step  is  to  calculate  the  effects  of  these  signals  in 
the  receiver. 


B.  Calculation  of  Intermodulation  Contribution  . For  the  example  here,  assume 
the  FM  signal  characteristics  given  below: 


FM  1 (Signal  B) 

Tone  modulation  10.42  KHz 

P|  - 2 .4  (Modulation  Index) 


FM  2 (Signal  C) 

Tone  modulation  12.5  KHz 

=2.4  (Modulation  Index) 


Mathematically  the  interfering  signals  can  be  described  by  either  the  time  domain 
or  frequency  domain  descriptbns.  For  example  for  the  B signal  (FM  1)  the  expressions  are: 

e^  (t)  = B cos  [ b t + pj  sin  u^t)  (18a) 

uj  = 2tt  (10,420) 

N 

e,(t)  = BE  J (B,)  cos  fb  t + moj.  t|  (spectral  representation 
m _N  of  the  FM  signal) 

(18b) 

After  passing  through  the  input  filter  the  results  are, 

e/  (t)  = B i n (b  + mu,)  J ( p, ) cos  fb  t + mu,  t| 

b I m i I 

m=  -N 

where  N spectral  components  are  assumed  adequate  for  the  representation  and  the 
input  filter  is  assumed  to  have  an  attenuation  of  a (u)  at  frequency  u and  no  phase  shift. 

Si mi  lari  ly  for  FM  2: 

ec  (t)  = C cos  [ c t + p^  sin  u^t  1 0 9a) 

where  =2w  (12,500) 

M 

e'c  (t)  = Cj  a (c  t-k^)  (P2)  cos  lc  f +l<U2tl  (19b) 

k = -M 

Using  the  Table  1 the  term  of  interest  in  this  example  is  (the  RF  filter  removes 
the  rest  of  the  terms): 


-21- 


* 


.1 


t 


3/4  BC2  cos  [ (2  c-b)  M-  2<t>2  - ^ | (20) 

(2  c-b)/2ir  = 2 (105.5)  - 102.5  - 108.5  MHz 
Using  (18b)  and  (19b)  the  intermodulation  term  in  (20)  can  be  written: 

N MM 

IM  = (3/4) K„  BC7  Z I I a (b  + mu,) 
m=  -N  k=-M  I = -M 

a (c  + ku2)  a(c  + 1^)  (P2)  J|  (f*2)  Jm  (Pj) 


cos  f (2  c-b)  t + (k  + I)  ~ mu^t|  (21a) 

where  K3  is  a parameter  of  the  model  and  a (u)  is  the  input  filter  attenuation. 

The  IF  frequency  characteristics  are  assumed  given  byy  (u);  hence,  the  output  of  the  IF  is 

N M M 

IM=3/4K  BC2I  I Z a(b  + mu.)  a(c + kuJ 

J m=  -N  k=  -M  1=  -M 

a(  c + l2  ) y | 2 c-b  + (k  + |)  - mu^|  (21b) 

Jk  (P2  ) J|  (P2  ) Jm  (P])  cos(  (2  c-b)  t + (k  + I)  ujt-mu^l 

In  this  equation  the  number  of  terms  that  must  be  considered  in  each  summation 
depends  on  the  modulation  index  and  the  AM  detector  output  depends  on  the  number  of 
terms  passed  by  the  IF-amplifier,  which  depends  on  the  bandwidth.  The  method  used  to 
specify  the  number  of  terms  considered  is  the  "rule  of  thumb"  quite  often  used  in  practice 
to  specify  the  bandwidth  of  an  FM  signal/  . e . , 

Bandwidth  = 2 f (1  +p) 
m 

where  f is  the  modulation  frequency  and  p is  the  modulation  index.  The  number 
of  terms  in  the  series  expansion  for  the  FM  signal  that  must  be  considered  is 

No.  of  terms  = 2 fm  (1  + P)  / 2 fm  = 1 + P 

In  particular,  for  the  case  being  considered,  i.e.,  P^  =2.4,  f^  =10.42  KHz 
(?2  = 2 .4  and  f 2 - 12.5  KHz,  the  number  of  terms  that  should  be  retained  is 

N ~ 3-4  and  M = 3-4 

Further  modifications  of  the  spectral  components  will  occur  due  to  the  RF  and 
IF  characteristics  of  the  receiver. 


-22- 


The  results  given  in  (21a)  can  be  simplified  by  assuming  the  input  filter  a (u) 
a constant.  This  will  have  very  little  effect  on  the  results,  since  most  of  the  AM 
distortion  (due  to  FM-to-AM  conversion)  results  from  the  band  limiting  through  the 
IF-amplifier.  With  this  assumption  the  squared  term  (e  ^ (t)  ) can  be  written  in  the 
time  domain  as 


e ^ (t)  = C^f  cos  ( c t + p sin  u,t  | | 2 
c z 


e ^ (t)  = / 2 f 1 + cos  [ 2 c t + 2 p2  sin  t|  1 (c) 

where  a (c)  has  been  assumed  constant  over  the  bandwidth  of  the  signal.  The 
term  in  (22)  which  contributes  to  the  intermodulation  distortion  is 


( / 2) cos  [ 2 c t + 2 p^  sin  (c) 


(23  ) 


Note  the  term  in  (23)  is  simply  an  FM  signal  with  a carrier  frequency  and 
modulation  index  twice  that  of  the  original  waveform. 

Using  (23),  the  intermodulation  components  contained  in  the  term 

3 e,  e ^ k„  are: 
b c 3 

IM=  3 a (c)  a (b)  lc„  BC^  I 1 J (P.)  J.  (2P0)  cos  | (2  c-b)  t + ku0t  - mu  tl 
o . m I k z z 

k m 


Table  6 lists  the  coefficients  of  the  FM  signals  e,  (t)  and  e (t)  which  are  above 
.01  in  magnitude  for  the  specific  case  of  p.  =2.4,  f = 10.42  Kfiz,  R„  =2.4  and 
f2  = 12.5  KHz.  1 

Table  7 lists  the  various  intermodulation  terms  along  with  the  components  of  the 
original  FM  signal  which  generates  them.  Only  terms  with  amplitudes  above  .01  are 
listed  to  show  how  each  term  is  generated . 

For  example,  the  first  listing  in  Table  7 is  generated  from  the  interaction  of 
7th  order  term  of  the  squared  signal  e£  (t)  and  the  1st  order  term  of  the  e^  (t),  ? .e. , 


k = -7  of  e (t):  J 7 (2-2.4)  cos  |2ct  - 7-2tt  (12,500)  t|; 
m = 1 of  e^  (t):  (2  .4)  cos  [ b t + 2ir  (10,420)  t | 

Amplitude  of  IM  relative  to  unmodulated  IM  = (J  y (4.8)  ) 


I 


-23- 


J M .8)  • J (2  .4)  (.0429)(.520)  = .00223 

Spectral  frequency  = (-7)  (12.5)  - 10.42  = - 97.92  KHz 

For  this  example,  the  signal  levels  are 

B - signal  power  = -23  dBm 

B =15.8  mv 
rms 

C - signal  power  = - 12  dBm 

C =56.3  mv 
rms 


Order 

n 

J (p.)=J  (2.4) 
n 1 n 

Amplitude  Coeff. 

of  e^  (t) 

J (2pJ  = J (4.8) 
n rZ  n 

Amplitude  Coeff. 

of  e ^ (t) 

c 

0 

0.00251 

-0.24043 

1 

0.52019 

- 0.29850 

2 

0.43098 

0.11605 

3 

0.19811 

0.39521 

4 

0.06431 

0.37796 

5 

0.01624 

0.23473 

6 

0.00337 

0.11105 

7 

0.00059 

0.04290 

8 

0.00009 

0.01408 

Table  6.  Coefficients  of  the  FM  signals  e,  (t)  and  e (t) 

b c 

for  |3j  = 2 .4,  f^  =10.42  KHz,  =2.4  and 
f2  = 12.5  KHz. 

Considering  the  effects  of  the  intermod  only  and  assuming  that  the  intermodulation 
center  frequency  is  the  same  as  the  receiver  frequency,  i .e. , 2 c-b  = a,  the  effects  can  easily 
be  evaluated.  Combining  (24)  with  the  desired  signal  the  output  can  be  written 

v.  (t)  = A K cos  at+3a(b)a  (c)  k,,  BC^  I Z J (B.)  • 
i — , — o i n 

4 k n 


(2Bj)  cos  f a t + k u>2  f ~ "XJj  f| 


(25) 


-24- 


Output  Intermod 
Freq.  in  KHz 
(Relative  to  Intermod. 
Center  Frequency) 


-97.92 
-85.42 
-95.84 
-72.92 
-83.34 
-93.76 
-60.42 
-70.84 
-81  .26 
-91.68 
-47.92 
-58.34 
-68.76 
-79.18 
-35.42 
-45.84 
-56.26 
-22.92 
-33.34 
-43.76 
-54.18 
-10.42 
-20.84 
-31.26 
-41.68 
2.08 
• 3.34 
-18.76 
-29.18 
14.58 
4.16 
-6.26 


Output  Intermod. 
Level  (Relative 
to  Unmodu  lated 
Intermod . ) 


-0.0223 
0.0578 
0.0479 
-0.1221 
-0.1012 
-0.0465 
0.1966 
0.1629 
0.0749 
0.0243 
-0.2056 
-0.1703 
-0.0783 
-0.0254 
0.0604 
0.0500 
0.0230 
0.1553 
0.1286 
0.0591 
0.0192 
-0.1251 
-0.1036 
-0.0476 
-0.0155 
-0. 1 553 
-0. 1286 
-0.0591 
-0.0192 
0.0604 
0.0500 
0.0230 


Order  of  Used 
to  Generate 
Intermod. 


Order  of  B 
to  Generate 
Intermod. 


Table  7.  IM  Spectrum. 


-25- 


Output  Intermod. 
Freq.  in  KHz 
(Relative  to  Intermod. 
Center  Frequency) 


27.08 
16.66 

6.24 

-4.18 

39.58 

29.16 

13.74 
8.32 

52.08 

41.66 
31.24 
20.82 

64.58 

54.16 

43.74 

77.08 

66.66 


Output  Intermod. 
Level  (Relative 
to  Unmodulated 
Intermod. ) 


Order  of  Used 
to  Generate 
Intermod. 


0.1629 

0.0749 

0.0243 

0.1221 

0.1012 

0.0465 

0.0)51 

0.0578 

0.0479 

0.0220 

0.0223 

0.0185 


Order  of  B 
to  Generate 
Intermod. 


Table  7.  IM  Spectrum  (Continued). 


•-  • ■ 


This  signal  is  further  filtered  by  the  output  stage  of  the  RF-amplifier  and  the 
IF  -amplifier  stages.  For  simplicity  assume  the  filtering  due  to  the  IF  to  be  ideal  with 
characteristics  shown  in  Figure  6. 


Figure  6.  IF  Characteristics. 


The  IF-bandwidth  is  assumed  to  be  50  KHz  i .e  . , Au  = ( 2ir)  ' ( 25  ) • (10). 

With  this  assumption  only  those  components  within  25  KHz  of  the  carrier  will  be  passed  by 
the  IF-amplifier . These  are  listed  in  Table  8. 


IM  Component 

Frequency  Relative 
to  Carrier 

Amplitude  Relative 
to  Unmodulated  IM 

2.08 

0.16 

4.16 

.05 

4.18 

.03 

6.24 

.08 

6.26 

.02 

8.32 

.02 

8.34 

.13 

10.42 

.13 

14.58 

.06 

16.66 

.17 

18.74 

.07 

18.76 

.06 

20.82 

.02 

20.84 

.10 

22.92 

.16 

Table  8.  IM  Components  Out  of  IF.  Only  Those  Components 
Above  the  Carrier  Frequency  are  Listed.  Those 
Below  are  the  Same  . 


23MM 


The  effects  of  the  IF-amplifier  are  represented  by  selecting  the  terms  in  the 
summation  given  in  25)  which  are  within  the  receiver  bandwidth,  i.e.. 


v 2 (t)  - Kj  A cos  a t i- 


P.  t 1 2P  t 1 

2,  J I 


Terms  in  3 BC  K a (b)  a (c)  , _ , .»  _ n 

IF  Bandpass  | ^ * ’>  &S  '> 


J (P.)  J.  (2P0)  | cos  | at  •-  nu0t  - ku  t 
m I k Z Z 


(26  ) 


Rewriting  (26)  in  terms  of  an  amplitude  and  a phase  the  results  of  passing  this 
signal  through  an  envelope  detector  can  easily  be  determined,  i.e.. 


Kj  A [ 1 t-  d1  (t)  ICj/  A Kj| 


where  d.  (t)  = Terms  within  f (3/4a(b)  c>2(c)  B C2  I I J (P,) 
IF  passband  k n 


(2p2)  cos  | (ku2-  ncj| ) t|  | 

For  the  example  being  illustrated  here  we  have  the  following: 


(27) 


a (b)  - a (102.5)  = - 9 dB 
a (c)  = a (105.5)=  - 4.5  dB 


(.36  linear  magnitude) 
(.6  linear  magnitude) 


A = 1 . 123  m volts 
rms 


Max  ! d.  (t)  | =3  BC2  c(b)  a2  (c)  F =2N/5  (4.87X  10~6  ) F 
1 4 


where  F is  a factor  introduced  that  accounts  for  the  IF  filtering  and  the  phase 


differences  between  the  spectral  components.  The  maximum  value  of  the  ratio  d^  (t)/A 


is, 


K.d  (t)  - 

Max|^-i ) = 20  (8.67X  10  F)  = .173F 


where  = 20  (linear). This  value  would  indicate  a significant  interference 

potential. 


This  model  can  be  used  in  several  other  ways.  For  example,  the  following 
question  might  be  asked:  What  minimum  distance  must  be  maintained  between  the  localizer 
receiver  and  two  100  kw  FM  stations,  operating  at  frequencies  such  that  an  intermod  at 
the  localizer  frequency  is  produced,  to  avoid  IM  distortion.  The  distortion  criteria  is  based 
arbitrarily  on  the  following: 


1 


-28- 


If  Max  | (d,  (t)/A)~rr—  F|  N .1  distortion  potential 

1 K, 

■s 

If  Max  | (d,  (t)  / A)—p — F|  < .1  no  distortion 

1 "S 

The  arbitrary  decision  criterion  listed  above  assumes  that  the  receiver  would 
capture  the  stronger  localizer  signal;  however,  there  would  be  notable  output  when 
there  is  no  desired  signal  present  at  interference  signal  levels  below  these  values. 

Note  here  we  have  arbitrarily  taken  the  Max|  d^  (t)  F/A  Kj|  as  the  statistic. 
Whether  or  not  this  is  a good  statistic  needs  further  investigation. 

For  the  example  calculation  we  have 

Max  d1  (t)  «3  3 BC2  F^a  (b)  a (c) 


( B and  C are  peak  values).  If  a decision  boundary  of  0.1  is  used  and  the  specific 
values  for  this  example  are  used  the  equality  in  (28)  holds  . 

B (dBM)  <-2  C (dBm)  + A (dBm) -20  log  F -20  log  IC/K  - 5.5 

J ( 28  ) 

We  have  used  the  attenuations  a (c)  = -4 .5  dB  and  c^b)  = -9  dB. 

The  region  of  allowable  combinations  of  signal  levels  B and  C which  would  pose 
no  interference  potential  under  our  criterion  is  illustrated  in  Figure  7. 


B (dBm) 


C (dBm) 

-5.5  -20  log  F -20  log 

/ 1^1  + A (dBm) 


No  Interference 


Figure  7.  Signal  Levels  for  No  Interference. 


-29- 


The  developed  equations  can  be  combined  to  place  restrictions  on  the  minimum 
allowable  distance  from  the  localizer  receiver  for  the  commercial  FM  antenna.  The 
power  input  to  the  receiver  in  dBm  (assuming  (ERP).  expressed  in  dBm)  is 

P;  = (ERP).  -36.3  -20  I °g ] 0 f " 20  l°910  d - L (29) 

Using  this  relation  the  signal  level  in  dBm  available  at  the  input  to  the  receiver 
is  (using  the  B signal  as  example) 

B (dBm)  = (ERP)b  -36.3  -20  log1Q  b-20  log1Q  db  -l_b  ( 30  ) 

Putting  in  the  specific  values  for  this  example  the  result  is 

-20  log  d <-42.4 -K./K  (dB)  - 20  log  d +40  log  d (31) 

D u q q 

The  value  of  3 K /2K  has  been  given  in  the  previous  section  for  the  NAV  1 1 
receiver  as  approximately  4 dfe;  hence  ^/Kj  = .5  dB.  Assuming  F = I,  (31)  becomes, 

20  log  db  > 42 .9  + 20  log  dy'd^  or  db  d ^ >140  d ( 32  ) 


Assuming  a critical  distance  of  5 miles  from  the  localizer  (32)  indicates  there 
would  be  a potential  interference  if  the  distances  are  such  that 

2 

db  dc  > 698  ( 33  ) 

Using  (33)  the  regions  of  interference  can  be  identified  as  illustrated  in 
Figure  8 . 


d 

c 


26  1 5 

18.7  2 5 

15.2  3 5 

13.2  4 5 

11.8  5 5 

8.3  10  5 

5.9  20  5 


Figure  8.  Regions  of  Interference. 


Assuming  F to  be  unity  gives  a pessimistic  result  in  general.  Another  statistic 
which  has  been  used  is  the  RMS  output  of  the  detector  which  is  very  simple  to  calculate 
and  measure.  Experimental  and  theoretical  results  using  this  as  the  statistic  have  been 
shown  to  agree  very  well  and  since  it  is  easier  to  specify,  is  probably  the  better  statistic. 
Using  the  RMS  value  at  the  output  of  the  detector  we  define  the  modulation  factor  (MF) 
as, 

Mp  _ RMS  Detector  Output  with  FM  Modulated  Signal  ^ ^4  ) 

RMS  Detector  Output  with  FM  Carrier  Only 

Experimental  and  analytical  results  obtained  are  given  in  Table  9.*  In  order  to 
obtain  the  experimental  results  a true  RMS  reading  meter  and  a full  wave  rectifying 
averaging  AC  digital  voltmeter  were  used  to  measure  the  detector  output.  It  is  seen 
that  the  experimental  and  calculated  results  agree  very  well. 


C.  Cross-Modulation  Considerations.  As  indicated  earlier,  the  terms  contributing 
to  cross-modulation  effects  result  from: 

3 

( 1 ) e (generally  small,  so  this  term  can  be  neglected) 
a 

2 2 
( 2 ) 3e  e,  and  3e  e 
a b a c 

It  is  easy  to  show  that  there  would  be  no  cross-modulation  distortion  produced 
other  than  a change  in  carrier  level  if  all  frequency  components  are  passed  with  the  same 
attenuation  a.  Since  the  frequencies  of  the  interfering  FM  stations  are  located  in  the 
"skirts"  of  the  localizer  receiver  characteristics,  each  frequency  component  will  be 
attenuated  differently.  The  different  attenuation  of  each  of  the  frequency  components 
results  in  AM  modulation  which  appears  as  cross-modulation  on  the  localizer  signal. 

To  account  for  these  effects,  it  is  necessary  to  have  available  the  receiver 
selectivity  characteristics.  A very  simple  compact  expression  can  be  used  to  calculate 
these  effects.  Considering  one  interfering  signal,  i.e., 

e^  (t)  = B cos  | bt  + pi  sin  om^t  ] 

The  equivalent  Fourier  series  representation  is 
ao 

e,  (t)  = B I J (fO  cos  f bt  + mu.t]  ( 35  ) 

b ml 

m = -00 

Assuming  the  "skirt"  characteristics  of  the  input  filter  of  the  RF-amplifier  can 
be  approximated  by  a linear  amplitude-frequency  characteristic  and  zerophase  shift  as 


* Details  of  the  measurement  procedures  and  results  are  given  in  Appendix  D. 


-31  - 


shown  In  Figure  9,  the  attenuation  characteristics  can  be  written 


a (b  + kuj)  = a (b)  [ 1 + kuj6  | ( 36  ) 

where  6 ~(a(b  1_Au>)-a(b))/A<j 

and  is  the  slope  of  the  input  filter  characteristics.  Assuming  an  FM  signal  input  to  a 
filter  such  as  described  above,  the  output  can  be  shown  to  be 

v = a (b)  (1  t-  B 5u,  cos  u.t)  B cos  (b  t + B sin  u,t)  ( 37  ) 
o II  I 


Figure  9.  "Skirt"  Characteristics. 


Figure  10.  Signal  Spectrum. 


It  is  seen  that  the  Frequency  modulation  of  the  output  is  the  same  as  the  input, 
but  in  addition  there  is  amplitude  modulation  which  is  the  same  as  the  original  modulating 
signal.  The  percent  AM  modulation  produced  by  this  attenuation  function  is 

% modulation  = pS  Uj  X 100  ( 38  ) 

The  cross-modulation  effects  are  given  by 

3 AB^  a (b)  I 1 + pj  cos  t|  ^ cos  (at)  ( 39  ) 

2 


and 

3 AC^  a (c)  [ 1 + p„  6<j0  cos  t ] ^ cos  (at) 

(40) 

Assuming 

p 6 small,  then  equations  (39)  and  (40)  become. 

3 KgAB^  a (b)  | 1 + 2 p^5*co | cos  Ujt|  cos  (at) 

2 

(41  ) 

and 

0 0 

3 K-^AC  a (c)  | 1 + 2 P26u2  cos  ujH  cos  (at) 

(42  ) 

2 


Combining  these  with  the  desired  signal  the  result  is 

(43) 

2 2 ^ ,xo  2 m2 

a (b)2PjSuj  B cosujt+^—  a (c)  2p^  C cos  ^ *1  cosat 

The  spectrum  of  this  wave  form  is  shown  in  Figure  10. 

Assuming  these  components  within  the  IF  Bandpass,  the  output  of  an  AM  envelope 
detector  will  be  within  a multiplicative  constant 

3ICj  2 o 2 2 

2fc~  I 2 a (b)  B cos  t + 2 a(c)  ^2^2  ^ ^"cos  Uj  H ( 44  ) 

Equation  (44)  specifies  the  resulting  output  of  an  envelope  detector  due  to  "brute- 
force"  interference  of  two  FM  stations  operating  at  frequencies  b and  c respectively. 

For  the  specific  example  being  considered  in  this  report  specific  numerical 
calculations  of  these  effects  can  be  made.  For  example,  the  peak  value  of  the  percent 
AM  modulation  due  to  the  interfering  stations  can  be  calculated,  i.e.. 


e = K.A 

O ! 


3K3  2 2 3K3  2 2 3K3 

I *2|<7  8 a (b)  a (c)  + 2^ 


'] 


-34- 


% AM- Mod  = 
peak 


K 

2 R~  ^2n  (b)  f}|  S<->  j B2  i 2a^  (c)p2  C^6u2 

3K 

1 + ~J-  | B2  a2  (b)  + C2  a2  (c)  | 


x 100 


2K 


(45) 


For  the  example  being  considered, 

a(b  = 102.5)  = - 9 dB 
a(c  = 105.5)  = - 4.5  dB 


3K 


2K, 


— = 31  .7  linear 


(.355  linear) 
(.6  linear) 


3|  =32  =2.4 
Ul  - (2ir)  (12,500) 
u2  - (2ir)  (10,420) 


.0612 


.103 


as 


8’  - TMHz  ’ s"  = Tmhz  (see  Fi9ure  H~2) 

B = -23  dBm;  B = 15.8  mv  (RMS) 

C = -1 1 dBm;  C = 56 .3  mv  (RMS) 

Using  these  (45)  gives  the  percent  AM  modulation  due  to  "brute-force"  interference 
% AM  modulation  = .04% 


The  above  percent  AM  modulation  due  to  "brute-force"  interference  seems  minimal 
when  compared  with  that  due  to  intermodulation  as  given  in  the  previous  section.  This,  of 
course,  is  due  to  frequencies  being  chosen  so  as  to  produce  an  intermod  at  the  frequency  the 
receiver  is  tuned  to  and  having  the  modulation  indices  low. 

D.  Combined  Effects  of  Intermod  and  Brute-Force  Interference.  Although  calculations 
for  the  example  given  here  indicate  that  the  "brute-force"  interference  is  minimal  when 
compared  with  the  intermod,  it  is  desirable  to  indicate  how  these  two  effects  can  be  combined. 

The  RF  output  resulting  from  the  desired  signal  and  an  intermod  is 
Kj  A cos  at  + dj  (t)  cos  (at  + <t>)  + d2  (t)  sin  (at  + O) 


(46) 


where  it  has  been  assumed  that  2 c-b  - a . If  this  is  not  the  case,  then  the  proper 
frequency  will  have  to  be  specified  in  the  last  two  terms  of  (46).  Using  (43)  and  (46)  the 
combined  effects  are  given  by 

3K3  2 2 3K3  2 2 

Distortion  = K.A  [ 1 + — B a (b)  + — C a (c)  + 

2K]  2K1 

3K3  2 2 3K3  2 2 

2 a (b)  p.  5u,  B cos  u,t  + 2 a (c)  p_  5u„  C cos  u_t  + 

2Kj  2K1 


K3  K3 

d.  (t)  cos  at  + d (t)  sin  at 

KjA  K] A 


(47) 


Equation  (47)  has  made  some  simplifying  assumptions  which  could  be  considered 
worst  case.  For  example,  it  has  been  assumed  that  all  components  are  in  phase,  which  is 
not  true  in  general.  This  assumption  will  provide  a worst  case  analysis;  however,  the 
technique  can  be  extended  to  account  for  different  phases. 

The  output  of  an  envelope  detector  would  be  (using  (47)  ). 

KjA  M + — (|B2  a2  (b)  +|C2  a2  (c)  + |2a2  (b)  p,  gU]  B2 

K1 

cos  Ujt  + J 2a2  (c)  P2  S"u>2  C2  cos  u2  + d1  ^ )2  + d22(t)  1 


2 2 
K1  A 


(48  ) 


This  expression  can  be  approximated  by 


3K_  _ 0 ^ 00 

KiA  M+  2irB  a (b)+?  i^-c  ° (c) 


Carrier  compression  term  due  to 
"brute- force" 

+ | [ 2 a (b)  p^  6!j^  B2  cos  Uj  + 2 a (c)  P2  S"u2  C2  cos 


AM-modulation  term  due  to  "brute-force"  interference 
(cross-modu  lati  on) 


d,  (t) 


(49) 


In  ter  mod 
Term 


-36- 


In  obtaining  (49)  the  squared  terms  have  been  neglected.  It  has  been  assumed  that 


K1  A 


d2  (t)  « 1 


( -j^-)2  B2  a2  (b)  + 2 B2  2 a2  (b)  p2  cos  ^t  + | 2 02  (c) 

C2  COS  u„t  + ^ )2  <<  1 

Z A 

These  conditions  hold  except  in  cases  of  extreme  interference. 

VII  EXTENSIONS  OF  THE  TECHNIQUES 

The  extensions  of  the  techniques  to  handle  cases  where  there  are  multiple  inter- 
fering sources  are  conceptually  very  simple;  however,  the  calculations  required  would 
be  lengthy  and  involved.  A computer  program  was  written  which  would  perform  the 
required  computation  for  any  specified  number  of  interfering  stations.  As  an  example 
we  evaluate  the  interference  potentials  in  the  greater  Birmingham,  Alabama  area 
airport.  This  region  was  chosen  due  to  the  fact  that  there  have  been  reported  cases  of 
interference  with  airborne  communications  and  navigational  receivers. 

Figure  11  shows  the  area  of  interest  qnd  Figure  12  shows  an  expanded  view  of  the 
area  along  with  the  various  FM  broadcast  stations  of  interest. 

Table  10  lists  the  various  FM  broadcast  station  call  letters,  frequencies, 
coordinates  and  powers  and  Table  11  lists  the  localizer  frequency  (110.3  MHz)  and  its 
coordinates,  along  with  the  outer  marker  location  and  the  VOR  frequency  and  location. 

Using  the  data  the  approximate  distances  of  each  FM  station  from  the  localizer 
aro  given  in  Table  12. 

Table  13  gives  a listing  of  the  various  signal  levels  at  the  receiver  assuming  a 
normalized  distance  of  1 mile  from  each  radiating  source  and  the  receiver.  In  order  to 
apply  the  tabulated  results  they  must  be  adjusted  by  a correction  factor  which  would 
account  for  the  proper  distance  from  the  FM  source  to  the  receiver.  This  correction 
factor  is  20  log^d,  where  d is  in  miles. 

Assuming  a localizer  power  of  40  watts  and  a normalized  distance  of  one  mile 
from  the  receiver  the  following  results  for  the  localizer  are  obtained: 

Frequency  = 110.3  MHz 
Miles  from  localizer  = 1 mile 
Free  space  attenuation  = 77.2  dB 


-37- 


Call 

Letters 

Freq.  In 
MHz. 

Coordinates 

Power 

WBHM 

90.3 

330-29,-19"/86°-47"-58" 

50  kw 

WDJC 

93.7 

33  - 26  -36  /86  - 52  - 50 

100  kw 

WAPI 

94.5 

33  -29  -26/86  -47  -48 

100  kw  1 

WQCZ 

96.5 

33  -29  - 02  /86  - 48  - 21 

50  kw 

WVOK 

99.5 

33  -26  -28/86  - 55  -00 

100  kw 

WZZK 

104.7 

33  -29  - 02  /86  - 48  - 35 

100  kw 

WKXX 

106.9 

33  -29  -19/86  -47  -58 

100  kw 

WENN 

107.7 

33  -29  - 02  /86  - 48  - 35 

100  kw 

WBRC  CH:  6 TV-100  kw  also  near  l 

Table  10.  FM  Broadcast  Stations  in  the  Birmingham  Area. 


BHM-VOR 

114.4  MHz  . 

33-40-12/86-53-59 

LOM 

(Outer  Marker) 

33-30-40/86-50-44 

LOC 

110.3  MHz. 

33-34-17/86-44-33 

Table  11  . Airport  Facilities. 


Station 

Distance 
in  miles 

WBHM 

6.1 

WDJC 

11.8 

WAPI 

5.9 

WQCZ 

6.4 

WVOK 

12.0 

WZZK 

6.6 

WKXX 

6.0 

WENN 

6.6 

Table  12.  Distance  of  FM  Broadcast  Stations  from  Localizer. 


-40- 


Power  in  dBm  = 46 . 

Antenna  loss  = 0 
Input  Fi  Iter  loss  = 0 

Resultant  desired  signal  level  at  RF  amplifier  input  = -25.1  dBm.  This  value 
must  also  be  corrected  for  the  distance. 

Several  calculations  were  made  to  investigate  the  interference  problems  relating 
to  this  area  and  are  given  below 

(a)  Receiver  at  location  (1)  in  Figure  13 

Referring  to  Figure  13  this  calculation  assumes  the  aircraft  at  position  (1)  closest 
distance  to  the  cluster  of  FM  stations. 

Table  14  gives  a computer  printout  of  the  signal  level  computations  for  each 
station  and  the  various  attenuation  factors.  Also  included  in  Table  14  is  the  calculation 
of  the  localizer  signal  level  at  position  (1).  The  predominant  type  of  interference 
expected  in  this  case  is  of  the  "brute-force"  type  as  indicated  by  the  signal  levels  in 
Table  14.  Table  15  gives  the  cross-modulation  resulting  from  each  interfering  station 
and  also  the  resultant  cross-modulation  due  to  the  combined  effects  of  the  interfering 
stations.  For  these  computations  a frequency  deviation  of  40  KHz  was  used  and  the 
slope  characteristics  of  the  receiver  front-end  (RF  amplifier)  were  modeled  by  the 
following: 


L 


.15  dB/100  KHz  for  frequencies  less  than  8 MHz  awoy 
from  the  receiver  frequency 


. ,075dB/ 100  KHz  outside  the  range 


Using  this  data  the  resultant  AM  modulation  due  to  cross-modulation  of  the 
interfering  signals  was  computed  using  the  techniques  given,  which  have  been  expanded 
to  include  more  than  two  stations.  The  results  of  these  calculations  are  given  in 
Table  15. 


For  these  calculations  no  modulation  was  assumed  so  that  the  maximum  inter- 
ference power  is  at  the  desired  localizer  frequency. 


The  statement  at  the  bottom  of  Table  15  indicates  that  there  is  significant  "brute- 
force"  interference  in  this  case.  As  a matter  of  fact,  for  this  location  the  amount  of  cross- 
modulation calculated  which  results  in  compression  of  the  desired  signal  is  in  excess  of 
1;  hence,  the  third-order  model  would  not  be  valid.  However,  this  is  of  no  consequence 
since  there  will  definitely  be  high  interference  levels  observed  at  this  location. 


Table  16  gives  the  resulting  intermodulation  results  due  to  the  interfering  signals. 
It  does  not  consider  the  effects  of  cross-compression.  In  any  case  the  significant  inter- 
ference at  this  poinf  is  due  to  "brute-force"  interference;  however,  the  resulting  com- 
pression of  the  desired  signal  would  allow  the  relative  weak  intermod  component  to  feed 
through  and  cause  problems. 


-42- 


PRINTOUT  OF  SIGNAL  LEVEL  COMPUTATIONS 


STATION 

MILES 

STATION 

FREE 

NAV 

RCVR  INPUT 

SIGNAL  LEV 

FREQ 

FROM 

POWER 

SPACE 

ANTENNA 

FILTER 

AT  RFAMP  IN 

(MHZ) 

RECEIVER 

IN  DBM 

ATTEN-DB 

LOSS-DB 

ATTEN  DB 

DBM 

90.3 

2.9 

77.0 

84.7 

18.2 

19.2 

-40.9 

93.7 

6.9 

80.0 

92.5 

14.8 

17.2 

-39.8 

94.5 

2.9 

80.0 

85.1 

14.0 

16.7 

-30.9 

96.5 

2.9 

77.0 

85.2 

12.0 

15.5 

-30.6 

99.5 

6.9 

80.0 

93.0 

9 .0 

13.7 

-30.1 

104.7 

2.9 

80.0 

85.9 

3 .8 

6.7 

-12.1 

106.9 

2.9 

80.0 

86.1 

1 .6 

4.1 

-6.8 

107.7 

2.9 

80.0 

86.2 

0.8 

3.1 

-4.9 

LOCALIZER  SIGNAL  LEVEL  CALCUIATIONS: 

110.3 

6.7 

46.0 

93.7 

0.0 

0.0 

-41.7 

Table  14.  Signal  Levelsat  Location  (1). 


L. 


*********  **CR0SS  MODULATION  CALCULATIONS  ********** 


STATION 

COMPRESSION 

FREQ  DEV. 

% AMMOD 

FREQ 

OF  LOC.  SIGNAL 

FORXMOB- 

CAUSED 

(MHZ) 

IN  DB 

CALC.  (KHZ) 

BY  STAT. 

90.3 

-0.0 

40. 

0.00 

93.7 

-0.0 

40. 

0.00 

94.5 

-0.0 

40. 

0.00 

96.5 

-0.0 

40. 

0.00 

99.5 

-0.0 

40. 

0.00 

104.7 

-1.6 

40. 

.11 

106.9 

-7.6 

40. 

1 .08 

107.7 

-21.3 

40. 

9.33 

****CROSS  COMPRESSION  GREATER  THAN  1 , THIRD  ORDER  MODEL  OF  RECEIVER  NO 
LONGER  VALID.  TOTAL  PERCENT  MOD.,  NOT  CONSIDERING  CROSSCOMPRESSION= 


.86  COMPRESSION  FACTOR  1 .679  DESIRED  SIGNAL  COMPRESSION  IN  DB***** 
TOTAL  PERCENT  AM  MODULATION  = ***** 


Table  15.  Cross- Modulation  Due  to  Interfering  Signals.  Receiver  at  Point  (1)  Location. 


-44- 


RECEIVER  FREQUENCY  IS  110.30  RECEIVER  BANDWIDTH  IS  40.0  KHZ 
FREQUENCIES  WHICH  PRODUCE  INTERMOD  INTERFERENCE  ARE: 


FI 

F2 

F3 

IM  TYPE 

IM  CENTER  FREQ 

93.7 

106.9 

90.3 

FI  + F2  - F3 

110.30 

IM  LEVEL  IN  DBM,  NO  FM  STAT.  MODULATION 

-81 .5 

Table  16.  Intermodulation  Interference.  Receiver  at  Point  (1)  location. 


PRINTOUT  OF  SIGNAL  LEVEL  COMPUTATIONS 

STATION 

MILES 

STATION 

FREE 

NAV 

RCVR  INPUT 

SIGNAL 

FREQ 

FROM 

POWER 

SPACE 

ANTENNA 

FILTER 

LEV  AT 

(MHZ) 

RECEIVER 

IN  DBM 

ATTEN-DB 

LOSS-DB 

ATTEN  DB 

RFAMP  IN  DBM 

90.3 

3.5 

77.0 

86.3 

18.2 

21  .0 

-42.5 

93.7 

5.5 

80.0 

90.5 

14.8 

18.4 

-37.8 

94.5 

3.5 

80.0 

86.7 

14.0 

17.8 

-32.5 

96.5 

3.5 

77.0 

86.9 

12.0 

16.3 

-32.2 

99.5 

5.5 

80.0 

91.1 

9.0 

14.1 

-28.2 

104.7 

3.5 

80.0 

87.6 

3.8 

8.4 

-13.8 

106.9 

3.5 

80.0 

87.8 

1 .6 

5.1 

-8.5 

107.7 

3.5 

PO.O 

87.8 

0.8 

3.9 

-6.5 

LOCALIZER  SIGNAL  LEVEL  CALCULATIONS: 

110.3 

8.5 

46.0 

95.7 

0.0 

0.0 

-43.7 

Table  17.  Signal  Levels  at  LOM  Location 


(b)  Receiver  at  the  LOM 

Similar  results  are  given  assuming  the  aircraft  at  the  Localizer  Outer  Marker 
(LOM)  as  noted  in  Figure  13.  Tables  17,  18  and  19  give  these  results.  Again  it  is 
observed  that  a significant  interference  due  to  "brute-force"  is  expected. 

(c)  Receiver  at  Point  2 on  Figure  13 

Tables  20,  21  and  22  give  results  when  the  aircraft  receiver  is  at  location  2. 


****************CROSS  MODULATION  calculations**************** 


STATION 

COMPRESSION 

FREQ  DEV. 

% AMMOD 

FREQ 

OF  LOC.  SIGNAL 

FOR  XMOD 

CAUSED 

(MHZ) 

IN  DB 

CALC.  (KHZ) 

BY  STAT. 

90.3 

-0.0 

40. 

0.00 

93.7 

-0.0 

40. 

0.00 

94.5 

-0.0 

40. 

0.00 

96.5 

-0.0 

40. 

0.00 

99.5 

-0.0 

40. 

0.00 

104.7 

-1  .1 

40. 

0.07 

106.9 

-4.5 

40. 

0.52 

107.7 

-8.6 

40. 

1.48 

******CROSS  COMPRESSION  GREATER  THAN  1,  THIRD  ORDER  MODEL  OF  RECEIVER 
NO  LONGER  VALID.  TOTAL  PERCENT  MOD.,  NOT  CONSIDERING  CROSSCOMPRES- 
SION = .40  COMPRESSION  FACTOR  1 . 1 55  DESIRED  SIGNAL  COMPRESSION  IN 
DB  = ********TOTAL  PERCENT  AM  MODULATION  = ***** 


Table  18.  Cross-Modulation  Due  to  Interfering  Signals.  Receiver  at  LOM  Location. 


RECEIVER  FREQUENCY  IS  110.30  RECEIVER  BANDWIDTH  IS  40.0  KHZ 
FREQUENCIES  WHICH  PRODUCE  INTERMOD  INTERFERENCE  ARE: 

FI  F2  F3  IM  TYPE  IM  CENTER  FREQ 

93.7  106.9  90.3  FI  + F2  - F3  110.30 

IM  LEVEL  IN  DBM,  NO  FM  STAT.  MODULATION 

-82.8 


Table  19.  Intermodulation  Interference.  Receiver  at  LOM  Location. 


-46- 


PRINTOUT  OF  SIGNAL  LEVEL  COMPUTATIONS 


STATION 

MILES 

STATION 

FREE 

NAV 

RECVR  INPUT  SIGNAL 

FREQ 

FROM 

POWER 

SPACE 

ANTENNA 

FILTER 

LEV  AT 

(MHZ) 

RECEIVER 

IN  DBM 

ATTEN-DB 

LOSS-DB 

ATTEN  DB 

RFAMPIN  DBM 

90.3 

5.7 

77.0 

90.5 

18.2 

21.0 

-46.7 

93.7 

3.2 

80.0 

85.8 

14.8 

18.4 

-33.1 

94.5 

5.7 

80.0 

90.9 

14.0 

17.8 

-36.8 

96.5 

5.7 

77.0 

91  .1 

12.0 

16.3 

-36.5 

99.5 

3.2 

80.0 

86.4 

9.0 

14.1 

-23.5 

104.7 

5.7 

80.0 

91  .8 

3.8 

8.4 

-18.0 

106.9 

5.7 

80.0 

92.0 

1 .6 

5.1 

-12.7 

107.7 

5.7 

80.0 

92.1 

0.8 

3.9 

-10.8 

LOCALIZER  SIGNAL  LEVEL  CALCULATIONS: 

110.3 

11.5 

46.0 

98.4 

0.0 

0.0 

-46.3 

Table  20.  Signal  Levels  at  Location  (2)  in  Figure  13. 


*************** ***£^OSS  MODULATION  CALCULATIONS****************** 


STATION 

COMPRESSION 

FREQ  DEV. 

% AMMOD 

FREQ 

OF  LOC.  SIGNAL 

FORXMOD 

CAUSED 

(MHZ) 

IN  DB 

CALC.  (KHZ) 

BY  STAT. 

90.3 

-0.0 

40. 

0.00 

93.7 

-0.0 

40. 

0.00 

94.5 

-0.0 

40. 

0.00 

96.5 

-0.0 

40. 

0.00 

99.5 

-0.1 

40. 

0.00 

104.7 

-0.4 

40. 

0.02 

106.9 

-1.4 

40. 

0.14 

107.7 

-2.3 

40. 

0.27 

TOTAL  PERCENT  MOD.,  NOT  CONSIDERING  CROSSCOMPRESSION=  0.6  COM- 
PRESSION FACTOR  0.488  DESIRED  SIGNAL  COMPRESSION  IN  DB=  5.2  TOTAL 
PERCENT  AM  MODULATION  = .43 

Table  21  . Cross-Modulation  Due  to  Interfering  Signals.  Receiver  at  Location  (2). 


RECEIVER  FREQUENCY  IS  110.30  RECEIVER  BANDWIDTH  IS  40.0  KHZ 
FREQUENCIES  WHICH  PRODUCE  INTERMOD  INTERFERENCE  ARE: 

FI  F2  F3  IM  TYPE  IM  CENTER  FREQ 

93.7  106.9  90.3  FI  + F2  - F3  110.30 

IM  LEVEL  IN  DBM,  NO  FM  STAT.  MODULATION 
-83.5 


Table  22.  Intermodulation  Interference.  Receiver  at  Location  (2). 


From  these  results  it  is  seen  that  the  resulting  cross-modulation  due  to  "brute- 
force"  interference  is  approximately  .43  percent,  assuming  the  frequency  deviation 
and  receiver  front  end  characteristics  given  earlier.  Whether  or  not  this  creates  a 
particular  significant  level  of  interference  would  be  a matter  of  judgment  and  more 
investigation  of  this  important  decision  criterion  is  required. 

(d)  Receiver  at  Location  (3). 

The  results  to  be  expected  when  the  receiver  is  at  location  (3)  are  given  in 
Tables  23, 24  and  25.  As  in  the  previous  case  whether  there  is  an  interference  problem 
at  this  position  is  a matter  of  the  "definition  of  interference" . Since  the  resulting  AM 
modulation  due  to  cross-modulation  resulting  from  the  interfering  signals  is  only  .16 
percent  and  the  intermod  level  is  -89  dBm,  it  is  expected  that  there  would  be  minimal 
interference  at  this  location. 


VIII  CDI  MODEL 


A.  Analytical  Model.  In  order  to  predict  the  effects  of  interfering  FM  stations 
on  the  navigational  aids  it  is  necessary  to  have  an  analytical  model  of  the  CDI  Circuitry. 
Thus  far,  the  model  has  been  capable  of  predicting  responses  throughout  the  receiver  to 
the  AM-detector  output  circuitry. 

Figure  14  shows  a block  diagram  of  the  audio  processing  circuitry  of  the  NAV  1 1 
receiver,  which  is  assumed  to  be  typical  of  those  to  be  encountered  in  practice.  Figure  15 
shows  a more  detailed  schematic  diagram  of  the  circuitry  of  interest. 

After  rather  detailed  and  involved  investigations  the  CDI  response  was  modeled 
by  a relative  simple  relationship,  given  in  (50)  below 


CDI  = k (VB,  - 0 Vffi  ) 


where 


CDI  = the  measured  voltage  “ ^£407  as  indicated  in  Figure  15 

(measurements  are  made  with  a DC  DVM). 

VB1  = RMS  voltage  at  the  base  of  Q 409 

VD_  = RMS  voltage  at  the  base  of  Q 410 
B2 

D = Ratio  of  the  RMS  voltage  at  the  base  of  Q409  to  the  RMS  voltage 
at  the  base  of  Q 410  with  CDI  = 0 p amp 

d-Vb1  I 

U I 1 I r\  i a t C.  1 \ 


CDI  =0. 


(51  ) 


-48- 


PRINTOUT  OF  SIGNAL  LEVEL  COMPUTATIONS 


STATION  MILES 

STATION 

FREE 

NAV 

RCVR INPUT 

SIGNAL 

FREQ  FROM  POWER 

SPACE 

ANTENNA 

FILTER 

LEV  AT 

(MHZ)  RECEIVER  IN  D3M  ATTEN-DB 

LOSS-DB 

ATTEN  DB  RFAMP IN  (DBM> 

90.3  8.C 

77.0 

93.5 

18.2 

21.0 

-49.7 

93.7  3.1 

80.0 

85.6 

14.8 

18.4 

-32.8 

94.5  8.0 

80.0 

93.9 

14.0 

17.8 

-39.7 

96.5  8.0 

77.0 

94.1 

12.0 

16.3 

-39.4 

99.5  3.1 

80.0 

86.1 

9.0 

14.1 

-23.2 

104.7  8.0 

80.0 

94.8 

3.8 

8.4 

-21.0 

106.9  8.0 

80.0 

94.9 

1 .6 

5.1 

-15.6 

107.7  8.0 

80.0 

95.0 

0.8 

3.9 

-13.7 

LOCALIZER  SIGNAL  LEVEL  CALCULATIONS: 

110.3  14.4 

46.0 

100.3 

0.0 

0.0 

-48.3 

Table  23.  Signa 

1 Levels  at  Location  (3) 

• 

******************crqss  MODULATION  CALCULATIONS****************** 

STATION 

COMPRESSION 

FREQ  DEV. 

% AMMOD 

1 

FREQ 

OF  LOC.  SIGNAL  FOR  XMOD 

CAUSED 

(MHZ) 

IN  DB 

CALC.  (KHZ) 

BY  STAT. 

90.3 

-0.0 

40. 

0.00 

93.7 

-0.0 

40. 

0.00 

94.5 

-0.0 

40. 

0.00 

96.5 

-0.0 

40. 

0.00 

99.5 

-0.1 

40. 

0.00 

104.7 

-0.2 

40. 

0.01 

106.9 

-0.7 

40. 

0.06 

107.7 

-1  .1 

40. 

0.12 

TOTAL  PERCENT  MOD.,  NOT  CONSIDERING  CROSSCOMPRESSION=  0.01  COM- 
PRESSION FACTOR  0.235  DESIRED  SIGNAL  COMPRESSION  IN  DB-2.3 
PERCENT  AM  MODULATION  0.16 


Table  24.  Cross-Modulation  Due  to  Interfering  Signal.  Receiver  at  Location  (3). 

Receiver  frequency  is  110.30  receiver  bandwidth  is  40.0  kh2 

FREQUENCIES  WHICH  PRODUCE  INTERMOD  INTERFERENCE  ARE: 

FI  F2  F3  IM  TYPE  IM  CENTER  FREQ 

93.7  106.9  90.3  FI  + F2  - F3  110.30 

IM  LEVEL  IN  DBM,  NO  FM  STAT.  MODULATION 
-89.1 


Table  25.  Intermodulation  Interference.  Receiver  at  Location  (3). 


-49- 


To  determine  the  various  parameters  in  (50),  a 150  Hz  and  90  Hz  signal  which 
would  result  in  a 0.0  microamp  deflection  is  used  as  the  input  at  point  E 403  (see  Figure  14) 
of  the  receiver  and  adjustments  are  made  so  as  to  give  0.000  volts  D.C. , i .e. 

V - V = 0*  The  data  obtained  is  given  in  Table  26  (see 

measurements  a,  d and  g). 

Using  this  data  the  following  results  are  obtained  for  the  factor  D in  (50): 


Measurement  a 
Measurement  d 
Measurement  g 


D = .701/. 640  = 1 .095 
D = .658/. 604  = 1 .084 
D = .638/.  585  = 1.090 


Using  the  average  of  these  as  the  value  for  D we  have, 

D = 1 .09 


(52  ) 


Similarily  using  (50)  and  the  data  given  in  Table  26,  k can  be  determined  i .e. , 

(53  ) 


“!=VB!-DV82 

k 


Measurement  b 


CPI  = .77 4 - (1 .09)  (.563)  = .160 

nr 


Using  the  above  the  measured  value  of  k is 


k measured  - 


CD  I meas  (i.e.  V£404  - VE407) 
VB1  -°VB2 


.064 

.160 


(54) 


= .400 


Similarily  using  measurements  c,  e and  f one  can  obtain  several 
independent  values  of  k,  i.e., 


Measurement  c 
Measurement  e 
Measurement  f 


k meas  = .416 
k meas  = .412 
k meas  = .4005 


The  data  given  under  measurement  h was  obtained  by  adding  an  assumed 
interference  signal  with  frequency  of  105  Hz.  Using  the  model  above,  the  desired  CDI 

response  calculated  by 

' CDI  calculated  = .4  (.744  - (1  .09)  (.594)  ) = .0385  volts 

The  measured  value  is  .040  which  gives  less  than  4%  error  between  the  measured 
and  calculated  result. 

For  completeness  several  points  relative  to  these  measurements  are  given  in  Table  27. 

-52- 


(1)  AC  voltage  measured  with  HP  3476B  DVM 

(2)  AC  voltage  measured  with  HP  3476  DVM 

(3)  DC  voltage  measured  with  Digitek  DVM 

(4)  Measured  with  50  p amp  deflection  on  ratio  generator 

(5)  Measured  with  105  Hz  tone  added  to  90/150  Hz  signal 
given  in  measurement  (g) 


Table  26.  Measurements  of  CDI  Voltages. 


Reference  Notes 

(1) 

Measurement  h is  obtained  taking  the  average  of  the 
maximum  and  minimum  voltage  readings.  The  variation 
of  the  voltage  is  due  to  the  beat  between  the  105  Hz 
and  the  90  Hz  signal. 

(2) 

For  all  measurements  a 200  (jf  capacitor  was  placed 
across  C 409.  This  makes  the  voltages  at  the  emitters 
of  Q409  and  Q410  the  halfwave  rectified  output  of 
the  signals  appearing  at  their  bases. 

(3) 

The  capacitors  C407  and  C408  which  couple  the  signals 
from  TP  402  and  TP  403  into  the  comparator  circuit  have 
significant  impedance  at  90  Hz  and  150  Hz.  Therefore, 
to  correctly  calculate  the  CDI  deflection,  AC  voltage 
measurements  must  be  performed  at  the  bases  of  Q409 
and  O110. 

(4) 

CDI  output  was  measured  at  the  output  of  comparator 
circuit  to  prevent  errors  arising  due  to  the  nonlinearity 
of  buffer  circuit. 

Table  27.  Reference  Notes  Relative  to  Measurements  in  Table  26. 


B-  Modeling  Receiver  Filters.  In  order  to  predict  the  CDI  response,  it  is 
necessary  to  model  the  output  of  the  ?0  and  150  Hz  filters.  Likewise  to  predict  the 
amount  of  audio  interference  caused  by  FM  stations,  the  audio  filter  characteristics 
must  be  known.  Figure  16  shows  the  measured  characteristics  of  the  90  and  150  Hz 
filters.  The  audio  characteristics  are  shown  in  Figure  17  and  the  detector  characteristics 
in  Figure  18 . 

Since  the  output  of  the  90  and  150  Hz  filters  would  be  highly  dependent  on  the 
properties  of  the  modulation  on  the  FM  interfering  signals*,  the  filters  were  modeled 
using  the  following  technique: 

(1)  Assume  the  input  to  the  90  and  150  Hz  filters  to  be  "white"  noise, 

i .e. , the  modulation  is  assumed  to  have  a constant  spectral  density  over  the  passband 
of  the  filters. 

(2)  Determine  the  noise  equivalent  bandwidth**  of  the  90  and  150  Hz 
filters.  For  applying  the  model  it  would  be  necessary  for  the  user  to  obtain  or  measure 
this  information . 

(3)  With  the  information  contained  in  (1)  and  (2)  the  RMS  outputs  of  the 
90  and  150  Hz  can  be  determined,  which  allows  us  to  compute  the  modulation  factors. 

Figure  17  shows  the  filters  with  the  same  noise  equivalent  bandwidth  as  the 
90  and  150  Hz  filters  used  in  the  computer  program. 

Table  28  compares  the  results  obtained  from  the  computer  program  to  measured 
results  with  IM  present  and  with  desired  signal  noise  modulated.  (Details  in  Appendix  D) 

These  results  were  compared  with  those  obtained  using  a tone  modulated  intermod 
formed  from  the  FM  signal  as  before. 

The  results  show  the  good  correspondence  between  the  modulation  caused  by 
IM  and  that  caused  by  a noise  source,  justifying  the  analysis  of  IM  using  a noise  source 
model.  It  also  shows  that  the  computer  program  accurately  predicts  the  amount  of 
interference . 


*See  Appendix  D,  Section  2 

**See  Appendix  E for  details  in  determining  the  noise  equivalent  bandwidth. 


-55- 


ITSSE." W81 * 


I 


T3 


^djnQ  pazjiDLUJOfsj 


Figure  16.  90/150  Hz  Filter  Responses. 


jnc^nQ  pazijDuiJO^ 


Figure  18.  Audio  Filter  CharacterisH 


Fi  Iter 

Measured 

Theoretical 

% RMS  Modulation 
with  desired  signal 
noise  modulated 

0) 

% RMS  Modulation 
due  to 

IM  Measured 

(2) 

% RMS  Modulation 
due  to  IM 
Theoretical  results 
from  Computer  Program 

Detector 

12.1% 

12.1% 

12.1% 

Audio 

5.0% 

4.7% 

5.7% 

150  Hz  Filter 

2 .4% 

2.5% 

2.4% 

90  Hz  Filter 

2.0% 

2.0% 

1 .8% 

(1)  Desired  signal  present  only,  desired  signal  AM  modulated  by  General  Radio  Model 
1381  Random  Noise  Generator  . 

(2)  Desired  signal  present,  IM  formed  by  2 FM  tone  modulated  stations. 

Table  28.  Summary  of  RMS  Modulation  Measurements  and  Calculations. 


C.  Effects  of  High  Level  Modulation  on  CPI  Reading.  During  the  experimental 
verification  of  the  model  several  difficulties  were  encountered  which  needed  study. 

One  particularly  important  problem  was  that  of  detector  overload  (or  clipping).  The  net 
result  of  these  effects  is  a desensitization  of  the  receiver. 

The  procedure  for  determining  these  effects  was  to  input  a desired  signal  of 
108  .5  MHz  modulated  with  a standard  90/150  localizer  signal  and  another  signal  at 
lower  amplitude  with  a frequency  slightly  offset  from  the  108.5  MHz.  In  this  experiment, 
108.501  MHz  was  used  which  results  in  an  additional  IKHz  sideband  in  the  IF  stage. 

This  1 KHz  signal  was  detected  along  with  the  90  and  150  Hz  signals.  This  interfering 
1 KHz  signal  caused  the  detector  to  overload  and  the  detected  signal  was  clipped.  The 
result  is  that  the  1 KHz  sideband  signal  simulates  the  effects  of  a high  level  intermod.  In 
theory  this  should  have  no  effect  on  the  receiver;  however,  due  to  the  overloading  of  the 
detector  it  does  cause  a significant  desensitization  problem. 

7 fie  90  and  150  Hz  filters  have  very  narrow  bandwidth,  which  eliminates  a 
considerable  amount  of  the  distortion.  In  order  to  obtain  significant  intermod  interference, 
it  is  necessary  that  the  intermod  level  be  significantly  high  so  as  to  cause  overload 


-59- 


distortion  at  the  detector.  From  previous  measurements  (see  Appendix  D),  it  was  determined 
that  an  intermod  carrier  level  which  is  10  dB  less  than  the  desired  signal  (which  results 
in  a modulation  factor  of  .45  and  a percent  modulation  of  approximately  12%  at  the 
detector  output)  causes  significant  interference.  In  this  case  the  voltages  at  the  outputs 
of  the  90  and  150Hz  filters  were  measured  to  be  .1  volts,  which  corresponds  to  an 
equivalent  modulation  of  about  2%.  If  the  intermod  level  were  increased  by  three 
times  to  give  an  RMS  modulation  of  36%,  the  equivalent  modulation  at  the  output  of  the 
90/150  Hz  filters  would  be  approximately  6%. 

As  an  example  of  the  use  of  the  techniques  consider  the  following: 

(a)  Assume  the  localizer  signal  such  that  it  produces  60  pA  CDI  response, 
then  DDM  = 60  (.155/150)  = .06  or  6% 

(b)  In  this  case  the  receiver  would  see  23%  modulation  at  the  90  Hz  filter 
output  and  17%  modulation  at  the  150  Hz  filter  output. 

(c)  Assuming  an  intermod  present  which  results  in  an  equivalent  of  6% 
modulation  at  the  90  Hz  filter  output,  the  modulation  would  be 

J ( (,06)2  + (.23)2  ) x 100  = 23.8% 

In  a similar  fashion  the  150  Hz  filter  output  would  be 

/(  (.06)2  + ( . 1 7)2  ) x 100  = 18.1% 

Using  these  results  the  CDI  with  the  assumed  interference  present  would  be: 

CDI  = (23.8-18.1)  (150/15.5)  = 55.1  pA 

This  corresponds  to  an  8%  reading  change  from  the  CDI  reading  with  no  interference 
present,  which  would  be  a relatively  small  change  in  CDI  reading.  On  the  other  hand 
with  a modulation  of  40%,  the  CDI  (assuming  60  pA  without  interference)  changes  by  20% 
or  gives  a reading  of  48  pA  due  to  limiting  or  clipping  of  the  detector  output.*  Therefore, 
limiting  due  to  overloading  is  the  significant  problem  affecting  the  CDI  reading. 


IX  COMPUTER  PROGRAM 

In  Section  VI  an  illustrative  example  of  the  calculations  required  assuming  only 
two  interfering  FM  signals  was  given.  If  there  exists  more  than  two  interfering  signals, 
the  amount  of  calculations  required  becomes  excessive  for  hand  computations.  A computer 
program  was  written  which  is  capable  of  making  the  required  computations  for  any  number 

*See  Appendix  F for  detailed  data. 


-60- 


of  interfering  signals  (limited  only  by  the  computer  facilities  available  to  the  user). 

The  program  performs  the  following  tasks: 

(1)  Calculations  of  the  signal  levels  at  the  RF  amp  input  due  to  FM  stations. 

(2)  Cross-modulation  and  cross-compression  calculations. 

(3)  Printout  of  intermodulation  producing  stations. 

(4)  Calculations  of  % AM  modulation  due  to  IM. 

(5)  Computation  of  the  effects  of  interference  on  CDI. 

(6)  Combined  cross-modulation  and  intermodulation  interference  calculations. 

Appendices  G and  H describe  the  computer  program  in  detail  and  in  Appendix  G 
is  a sample  printout  from  the  program. 

X CONCLUSIONS  AND  RECOMMENDATIONS 

The  research  reported  herein  has  culminated  in  analytical  techniques  capable  of 
predicting  possible  interference  of  airborne  communication  and  navigation  receivers  due  to 
commercial  FM  broadcast  stations.  Approximate  analytical  models  of  the  receiver  were 
dt .eloped  which  could  be  specified  by  rather  simple  measurements. 

Both  the  effects  of  "brute-force"  interference  and  intermodulation  distortion  can 
be  readily  handled  by  the  model.  It  was  found,  as  was  expected,  that  the  majority  of 
the  nonlinearities  were  due  to  saturation  of  the  RF  amplifier.  A relative  simple  analytical 
model  using  a third-order  nonlinearity  was  found  to  be  satisfactory  for  interference  levels 
of  approximately  0 dBm  and  below.  For  higher  signal  levels  than  this,  the  model  does  not 
account  for  limiting  and  other  effects;  however,  for  interfering  signal  levels  of  this  strength, 
it  is  expected  that  an  interference  problem  would  exist. 

Measurement  techniques  involving  only  a monitoring  of  the  AGC  voltage  were 
developed  which  allows  for  the  determination  of  the  various  parameters  of  the  nonlinear 
model.  It  was  found  that  most  of  these  parameters  could  be  assumed  relatively  constant 
over  a wide  range  of  input  signal  levels.  In  addition  to  the  determination  of  the  various 
parameters  of  the  nonlinear  model,  the  following  information  must  be  known: 

(1 ) RF  frequency  response 

(2)  IF  frequency  response  characteristics  (bandwidths  and  skirt  attenuation 
characteristics) 


-61- 


With  this  information  the  signals  at  the  output  of  the  IF-amplifier  can  be  predicted. 

In  most  cases  it  was  found  that  the  best  way  to  handle  the  above  characteristics  was 
to  specify  the  overall  frequency  characteristics  in  terms  of  an  input  filter  skirt  slope  attenua- 
tion characteristics  and  an  overall  frequency  response  characteristic  of  the  RF-amplifier 
output  circuitry  and  the  IF-amplifier.  For  the  latter  both  ideal  bandpass  and  triangular 
bandpass  characteristics  were  used.  Using  the  above  analytical  model,  signals  at  the 
input  to  the  audio  detector  can  be  predicted. 

Models  of  the  audio  detector  circuit  and  the  90  Hz  and  150  Hz  filters  in  the  case 
of  navigation  receivers  were  developed.  This  requires  the  frequency  characteristics  of 
the  audio  circuit  and  the  90  Hz  and  150  Hz  filters  either  be  known  or  determined  by 
measurements.  It  is  found  that  the  interference  at  the  output  of  the  90  Hz  and  150  Hz 
is  rather  small  due  to  very  narrow  bandwidths  of  these  filters.  An  analytical  model  of 
CDI  (course  deviation  indicator)  was  determined  which  could  predict  the  effects  of  the 
interfering  signals. 

Investigations  were  performed  for  determining  the  signal  levels  at  the  input  to 
the  receiver  given  the  station  power  and  it  was  found  that  a satisfactory  model  to  use  was 
the  free  space  attenuation  formula,  along  with  an  empirical  correction  factor  to  account 
for  the  vehicle  antenna  loss.  The  empirical  correction  used  was  to  assume  1 dB/l  MHz 
attenuation  for  signals  below  108.5  MHz. 

A computer  program  using  the  developed  models  and  techniques  was  written  which 
can  calculate  and  predict  interference  due  to  any  number  (limited  only  by  the  user  compu- 
ter facilities)  of  FM  stations.  The  program  uses  the  measured  characteristics  of  the  vehicle 
receiver  to  predict  the  amount  of  interference  due  to  high  power  FM  stations.  The  program 
assumes  sinusoidal  modulation  on  the  FM  signals  and  that  the  interference  is  due  to  3rd 
order  non  linearities  in  the  RF  amp  of  the  receiver. 

The  user  must  supply  the  following  information  to  analyze  an  interference  situation: 
Number  of  FM  stations,  receiver  frequency,  distortion  parameter  3K^/2K  . For  each 
FM  station,  the  following  information  is  required:  Station  frequency,  modulation  fre- 
quency and  modulation  index,  distance  from  the  receiver  and  station  power  level.  Loca- 
lizer distance  from  the  receiver  and  power  level  is  also  needed. 

Although  the  program  was  written  as  general  as  possible,  there  are  some  specific 
limitations  and  these  are: 

(1)  A modulation  index  of  75  is  a practical  limit  on  the  maximum  modulation 
index  that  the  program  will  accept.  Modulation  indices  higher  will  cause  extremely  long 
program  execution  times.  A modulation  index  of  75  corresponds  to  a maximum  frequency 
deviation  of  75  KHz  with  a modulation  frequency  of  1 KHz. 

(2)  The  program  assumes  that  the  modulation  frequencies  of  the  FM  stations  are 
not  harmonically  related.  For  example  frequencies  of  1000  Hz  and  400  Hz  are  related 
whereas  1000  Hz  and  398  Hz  are  not.  The  intermodulation  due  to  stations  with  harmon- 
ically-related modulation  frequencies  has  components  which  occur  only  at  certain  frequen- 


-62- 


'Mgr 


cies  relative  to  the  IM  carrier  and  does  not  exhibit  the  "noise-like"  (realistic)  properties 
which  are  assumed  for  IM  calculations . Harmon!  cal  ly-related  modulation  frequencies  will 
be  treated  as  though  they  were  not  related  by  the  program.  Therefore,  when  using  harmon- 
ically-related frequencies  such  as  1000  Hz  and  400  Hz,  it  is  advisable  to  check  the  results 
to  see  if  the  same  results  are  obtained  as  with  frequencies  of  1000  Hz  and  398  Hz. 

Although  satisfactory  analytical  models  and  techniques  have  been  developed  in 
this  research,  there  were  certain  key  items  identified  during  the  program  which  are 
recommended  for  further  research. 

(1)  The  determination  of  reliable  statistical  criteria  which  would  allow  for 
realistic  decision  making  regarding  interference  potentials  and  which  can  be  adequately 
substantiated  thus  providing  a strong,  defendable  technical  basis  for  future  judgment  and 
decisions  concerning  station  allocations. 

(2)  Receiver  modeling  improvement,  either  by  extending  the  current  model  or 
the  use  of  more  sophisticated  models  than  are  currently  being  used. 

(3)  A written  specification  and  verification  of  field  testing  procedures  for 
evaluating  interference  problems. 

Although  the  above  list  is  not  an  exhaustive  list  of  identifiable  items  requiring 
further  investigation,  those  items  on  the  list  are  believed  to  be  of  immediate  concern  and 
importance  for  the  proper  enhancement  of  the  results  reported  in  this  research.  Brief 
descriptions  of  the  items  recommended  for  further  research  are  given  below. 

A.  Statistical  Model. 

1 . Type  of  Modulating  Signals  to  be  Assumed.  The  research  reported  herein 
has  used  primarily  single  tone  modulated  FM  signals  in  the  development  of  the  model . 

To  a limited  extent,  "white"  noise  as  a source  of  modulation  has  also  been  considered.  Since 
the  CDI  in  a navigation  receiver  is  preceded  by  rather  narrow  bandpass  filters,  the  type 
of  modulation  will  have  a considerable  bearing  on  the  CDI  response. 

Since  the  receivers  will  be  operating  in  a real  world  environment,  there  is  a 
need  for  a statistical  study  to  determine  a model  which  would  be  truly  representative  of 
real  world  modulations.  Without  this,  one  is  left  with  only  the  option  of  talking  about 
probable  "worst"  case  effect,  which  would  be  hardly  a convincing  argument  in  the  real 
world . 


2.  Statistical  Basis  for  Interference  Decision.  The  RMS  value  of  the  output 
of  an  envelope  detector  due  to  the  interfering  signals  was  used  as  a parameter  to  deter- 
mine a potential  interference  problem.  It  is  possible  that  other  parameters  (statistics)  might 
be  a better  indicator  of  potential  interference  problems.  For  example,  peak  modulation 
might  be  a more  realistic  parameter  on  which  to  base  decisions  regarding  potential  inter- 
ference problems. 


-63- 


There  is  a need  to  investigate  and  determine  which  parameter  or  parameters  are 
useful  and  reliable  for  basing  predictions  of  potential  interference  problems.  It  is  proposed 
that  various  parameters  be  investigated  and  evaluated  and  that  these  be  related  to  real 
world  occurrences.  The  analytical  methods  and  parameters  would  be  substantiated  by  experi- 
mental documentation. 


3.  Decision  Criterion.  For  any  mathematical  or  analytical  model  to  be 
useful,  in  a practical  sense,  one  needs  to  know  the  implications  of  the  results  obtained,  and 


exactly  how  realistic  they  are.  To  i llustrate  this  rather  involved  statement,  consider  dis- 
cussed criterion  used  in  deciding  a potential  interference  problem.  The  means  of  deciding 
whether  or  not  interference  would  occur  was  to  simply  postulate  the  following  decision  rule 


using  the  RMS  output  of  the  envelope  detector  (V  ) as  the  statistic  (parameter): 


Decision  Rule 


If  V (RMS  output  of  envelope  detector  due  to  interfering  signals)  > T decide 

«-  O.  f ii  • ■ • i .1  . .1  • r 


in  favor  of  an°interference  problem,  i.e 
problem . 


decide  that  there  is  a potential  interference 


If  V < T decide  that  there  is  no  significant  potential  interference  problem.  The 
consequences°of  the  above  are  summarized  below: 


Actual 

Condition 


Interference 


No  (minimal) 
Interference 


V0  >T 


Correct 

Decision 


Incorrect 

Decision 


Incorrect 

Decision 


Correct 

Decision 


Summary  of  the  Results  of  Decision. 


The  peak-modulation  index  and  the  RMS  output  of  an  envelope  detector  were 
used  as  the  parameter  (statistic)  on  which  the  decision  was  made.  In  particular,  for  checking 
the  model,  T was  arbitrarily  set  at  some  constant  value.  It  was  possible  to  provide  a "profile" 
of  potential  interference  problems  in  terms  of  FM  station  (or  FM  stations)  locations  and 
power  relative  to  usable  navigation  signals.  This  allows  for  predicting  the  effects  of  adding 


A 


-64- 


one  or  more  FM  stations  in  the  near  vicinity,  or  the  results  to  be  expected  if  signal  powers 
are  changed . 

In  order  to  provide  realistic  and  convincing  arguments,  regarding  potential  inter- 
ference problems,  two  areas  need  further  study,  i.e.,  what  decision  rule  and  what  statistic 


Additional  research  is  recommended  which  would  result  in  the  determination  of 
useful  and  reliable  decision  rules  and  parameter  (these  may  be  different  for  navigation  and 
communication  receivers)  for  predicting  potential  interference  problems.  The  results  are 
to  be  presented  in  terms  of  a statistical  model  using  probability  measures  of  the  errors  associ- 
ated with  the  decision  rules.  This  could  be  accomplished  by  specifying  a "confidence" 
interval  describing  the  results  of  decisions  regarding  interference  problems  and/or  probabilis- 
tic measures  of  incorrect  decisions. 

B.  Model  Improvement.  Although  results  obtained  here  have  shown  that  the 
analytical  techniques  used  for  modeling  the  propagation  and  the  receiving  of  interfering 
and/or  desired  signals,  in  conjunction  with  the  all-important  model  of  the  airborne  naviga- 
tion and  communication  receivers  in  the  presence  of  a multi-signal  environment  are  useful 
and  applicable,  there  is  a need  for  model  enhancements  and  improvements.  Enhancements 
and  improvements  are  needed  in  the  two  areas  of  concern,  i .e.,  (1)  propagation  model  and 
(2)  receiver  model . 

(1)  Propagation  Model.  There  is  a need  for  providing  an  improved  model 
of  the  propagation  of  both  the  desired  and  interfering  signals  that  would  include  more 
realistically  the  effects  of  transmitting  and  aircraft  receiving  antennas.  The  problems 
associated  with  modeling  an  aircraft  antenna  are  of  considerable  difficulty  and  the  solutions, 
at  most,  only  approximate  in  nature.  This  complication  is  due  in  large  part  to  the  presence 
of  the  aircraft.  It  is  suggested  that  investigations  be  undertaken  which  would  result  in  a 
more  realistic  model  of  the  receiving  antennas  on  board  selected  aircraft.  For  these  inves- 
tigations it  is  suggested  that  selected  "reference"  aircraft  be  identified  and  that  each  of  these 
be  considered  and  the  antennas  and  aircraft  be  modeled  as  a system.  Such  a model  would 
include  coupling  characteristics  between  key  antennas  on  the  aircraft.  It  is  expected  that 
numerical  moment  method  techniques  performed  with  the  aid  of  a digital  computer  will  be 
required  to  obtain  these  results.  The  end  result  of  such  investigations  would  be  a computer 
program  which  would  be  capable  of  modeling  certain  selected  aircraft  to  a high  degree  of 
accuracy.  Theoretical  results  should  be  verified  by  experimental  documentation. 


should  be  used  to  provide  realistic  and  convincing  arguments  regarding  potential  interfer- 
ence problems . 


A second  phase  of  this  investigation  would  involve  a study  which  would  enumerate 
the  resulrs  and  expected  consequences  of  simplifying  the  propagation  model.  Trade-off 
studies  involving  complexity  vs.  accuracy  would  be  performed. 

The  results  of  the  previously-mentioned  investigations  would  allow  for  the  realistic 
and  accurate  modeling  of  the  signal  propagation  from  the  transmitting  to  the  receiving 
antenna . 


t 


r 


(2)  Receiver  Model.  It  has  been  demonstrated  that  the  third-order  receiver 
model  developed  for  predicting  interference  effects  is  useful  in  predicting  the  effects  of  multi- 
ple interfering  signals  on  aircraft  navigation  and  communication  receivers;  however,  it  would 
be  desirable  to  improve  the  accuracy  of  these  predictions.  It  is  recommended  that  considera- 
tions be  given  to  enhancing  the  existing  receiver  model  in  order  to  improve  the  accuracy  of 
the  predictions,  along  with  investigations  of  other  modeling  schemes. 

In  particular,  it  is  suggested  that  enhancement  of  the  model  be  made  so  as  to 
be  able  to  consider  the  effects  of  higher  signal  power  levels  either  due  to  more  interfering 
signals  or  higher  radiated  power  levels.  It  has  been  shown  that  the  accuracy  obtained  using 
the  simple  third-order  model  is  limited  at  power  levels  of  0 dBm  or  higher. 

It  is  recommended  that  research  on  the  receiver  model  be  continued  and  the  best 
way  to  enhance  and  extend  the  regions  of  validity  of  the  receiver  model  be  determined. 

In  particular,  investigations  using  higher  order  terms  in  the  existing  model  and  a comparison 
of  these  techniques  in  terms  of  complexity  and  accuracy  with  other  models  should  be  made. 

In  addition  to  the  above  important  areas,  effects  of  impedance  mismatches,  re- 
flection coefficients,  etc.,  should  be  considered. 

Completion  of  these  investigations  along  with  the  results  obtained  here  would  make 
a complete  package  which  would  allow  the  user  the  flexibility  of  using  a model  of  just 
sufficient  complexity  to  fulfill  needed  requirements.  In  some  cases  a very  simple  model 
could  be  used  to  obtain  results  within  the  necessary  requirements,  while  on  the  other  hand 
it  may  be  necessary  to  use  a rather  complex  model  to  obtain  reliable  results. 

C.  Standard  Test  Procedure.  Efforts  reported  herein  have  been  specifically  for 
the  purpose  of  development  and  experimental  testing  of  analytical  models.  A "Standard 
Test  Procedure"  should  be  defined  which  would  allow  the  engineer  to  set  up  an  experimental 
test  using  signal  generators  with  specified  modulations  and  antennas  that  would  simulate 
potential  interference  problems. 

It  is  suggested  that  further  research  be  performed  which  would  identify  such  things 
as  signal  levels,  modulating  levels  and  signals,  antennas,  etc.  This  would  allow  an  engineer 
to  set  up  an  experimental  test  which  would  realistically  simulate  the  potential  interference 
problems  being  considered. 


D.  Interference  Profile  Mapping  of  Potential  Airport  Facilities.  It  is  recom- 
mended that  further  work  be  performed  on  detailed  investigations  of  certain  selected  existing 
and/or  proposed  new  airport  facilities  as  identified  by  FAA  regarding  the  potential  of  an 
interference  problem  existing  and  possibilities  of  such  a situation  developing.  Typical  ap- 
plications would  be  to  consider  a specific  airport  and  all  possible  interfering  signal  sources 
in  the  near  vicinity  along  with  their  respective  power  levels,  and  signal  frequencies,  and 
using  this  data,  calculations  would  be  made  to  determine  a "profile"  of  the  interference 
potentials  existing.  These  "profiles"  would  be  given  in  terms  of  minimum  distances  for  the 
establishment  of  additional  FM-stations,  effects  of  increased  power  levels,  etc.  These 


"profiles"  would  be  an  indicator  of  how  potentially  close  a particular  facility  is  to  having 
an  interference  problem. 

Similar  techniques  as  described  above  could  be  applied  to  proposed  new  facilities 
as  identified  by  FAA  personnel.  Such  results  would  be  valuable  in  determining  new  sites 
and  the  consequences  of  the  selections  could  easily  be  determined. 

XI.  ACKNOWLEDGMENTS 

The  authors  appreciate  the  encouragement  and  the  many  useful  contributions 
to  the  research  reported  here  by  the  Project  Director,  Dr.  Richard  H.  McFarland,  and 
Dr.  Robert  W.  Lilley,  Assistant  Director,  Avionics  Engineering  Center.  Several  other 
people  deserve  a special  word  of  thanks  for  their  technical  contributions.  They  are 
Dr.  Raymond  Luebbers  and  Professor  G.  E.  Smith.  In  addition  thanks  are  given  for  the 
valuable  discussion  with  FAA  personnel.  In  particular,  recognition  goes  to  Mr.  Gerald 
Markey  and  Mr.  Reuben  Michaeles. 

A special  word  of  thanks  goes  to  Mrs.  Shirley  Wilson  for  her  excellent  typing 
of  the  involved  technical  monthly  reports.  In  addition  thanks  go  to  Ms.  Hope  Mills  for 
her  coordination  in  the  typing  of  the  final  report.  Without  these  two  peoples'  contribu- 
tions the  research  reported  here  would  lack  proper  documentation  and  reporting. 


XII.  REFERENCES 

ri]  Chang,  K.Y.,  "Intermodulation  Noise  and  Products  Due  to  Frequency  Dependent 

Nonlinearities  in  CATV  Systems",  IEEE  Transactions  on  Communications,  Vol. 
com-23,  No.  I,  pp.  142-155,  January  1925. 

T2]  Bennett,  W.R. , "Cross- Modulation  Requirements  on  Multi-Channel  Amplifiers 

Below  Overload",  Bell  Syst.  Tech.  J. , Vol.  19,  pp.  587-610,  October  1940. 

r3]  Simons,  K.,"The  Decibel  Relationships  Between  Amplifier  Distortion  Products", 

Proc.  IEEE,  Vol.  58,  pp.  1071-1086,  July  1970. 

T4]  Glave,  F . E . , "A  Statistical  Approach  to  the  study  of  Intermodu lation  Noise"  , 

Bell-Northern  Research,  Rep.  TM  8260-1-66,  July  1966. 

f5]  Wiener,  N.,  "Non-linear  Problems  in  Random  Theory,"  Cambridge,  Mass.: 
Tehcnology  Press;  and  New  York:  Wiley,  1958. 


-67- 


XIII.  BIBLIOGRAPHY 


Bedrosian,  E.,  and  S.  O.  Rice,  "Distortion  and  Crosstalk  of  Linearly  Filtered 
Angle-Modulated  Signals",  Proc . IEEE,  Vol.  56,  pp.  2-13,  January  1968. 

Bedrosian,  E.,  and  S.  O.  Rice,  "The  Output  Properties  of  Volteria  Systems 
(Nonlinear  Systems  with  Memory),  Driven  by  Harmonic  and  Gaussian  Inputs"  , Proc.  of 
IEEE,  Vol.  59,  pp.  1688-1707,  No.  12,1971. 

Mircea,  A.,  E.  Bedrosian,  and  S.  O.  Rice,  "Further  Comments  on  Distortion 
and  Crosstalk  of  Linearly  Filtered,  Angle-Modulated  Signals",  Proc.  IEEE  (letter), 

Vol.  57,  pp . 842-844,  May  1969. 

Narayanan,  "Transistor  Distortion  Analysis  Using  Volterra  Series  Representation", 
Bell  Syst.  Tech.  M.,  Vol.  46,  pp.  991-1023,  1967. 

Narayana,  "Application  of  Volterra  Series  to  Intermodulation  Distortion  Analysis 
of  Transistor  Feedback  Amplifier",  IEEE  Trans.  Circuit  Theory,  Vol.  CT-17,  pp.  518  — 

527,  November  1970. 

Rice,  S.  O.,  "Second  and  Third  Order  Modulation  Terms  in  the  Distortion 
Produced  When  Noise  Modulated  FM  Waves  Are  Filtered,"  Bell  Syst.  Tech.  J.,  Vol. 

48,  pp.  842-844,  January,  1969. 


XIV.  appendices 


A.  Theoretical  and  Experimental  Measurements  Using  Model  and  an  FM-Interfering 
Signal.  Theoretical  and  experimental  measurements  were  obtained  using  the  proposed 
model  for  the  RF-amplifier . For  these  investigations,  the  desired  localizer  signal  was 
assumed  to  be  a 108.5  MHz  signal  which  was  amplitude  modulated  in  some  cases  by  a 
1000  Hz  tone  to  allow  for  the  determination  of  distortion  at  the  output  of  the  detector. 
The  interfering  signal  was  assumed  to  bean  FM  modulated  signal  of  various  signal 
strengths  operating  at  105.5  MHz.  This  signal  was  modulated  by  a 10  KHz  sinusoid 
producing  a modulation  index  of  6,  resulting  in  an  FM  bandwidth  of  approximately  200 
KHz.  The  interfering  signal  spectrum  is  illustrated  in  Figure  A-l  . 


Vertica  I 

Linear 

Scale 


105.5  MHz  50  K Hz/di v 

Figure  A- I . Interfering  Signal  Spectrum;  Modulation 
Index  6;  Bandwidth  200  KHz. 

Using  this  as  the  input  signal  and  assuming  no  distortion  due  to  input  filtering 
of  the  FM  spectral  components  it  can  be  shown  that  the  cross -modulation  terms  located 
near  105.5  MHz  will  result  in  only  a compression  term  at  the  carrier  frequency  of  the 
desired  signal  (108.5  MHz).  Considering  the  effects  of  the  input  filter  (.2  dB  '100  KHz 
attenuation)  on  the  FM  spectral  components  theoretical  calculations  indicate  the  ratio 
of  the  first  sideband  to  the  carrier  term  to  be  -44  dB  compared  to  a measured  value  of 
-42  dB.  For  the  second  sideband  theoretical  results  predicted  -92  dB  which  could  not 
be  measured.  These  results  are  illustrated  in  Figure  A -2. 


Fic^ireA-2.  Spectrum  of  Desired  Signal  (108.5  MHz)  at  RF-Amplifier 
Output.  Desired  Signal  Strength  -17  dBm  (.03  volts); 
FM-Interfering  S ignal  (105.5  MHz,  P = 6;  Bandwidth  = 

200  KHz)0  dBm  (.2  volts). 

Figure  A-3  shows  the  spectrum  at  the  RF  amplifier  output  with  the  desired  signal 
at  -47  dBm  signal  level  with  the  interfering  signal  level  the  same  as  given  for  Figure  A-2  . 
The  sidebands  are  not  visible  because  of  the  residual  noise. 

Qualitative  investigations  of  the  distortion  produced  at  the  output  of  the  detector 
were  performed.  Figure  A— 4 shows  the  detector  output  using  a 1 KHz  modulation  signal 
with  10  percent  modulation  and  no  interfering  signal  present. 

Figure  A-5  shows  the  detector  output  when  the  desired  signal  is  the  same  as 
described  in  Figure  A -4  and  the  interfering  FM  signal  of  0 dBm,  as  described  in  Figure 
A-l , is  added. 

Figure  A-6  shows  the  detector  output  when  a desired  signal  is  reduced  to  -47  dBm. 

Figure  A- 7 indicates  the  results  obtained  using  a desired  signal  strength  of 
-70  dBm  (.1  mv)  and  no  interfering  signal.  Figure  A-8  shows  the  results  of  an  interfer- 
ing signal  of  0 dBm  (.23v)  on  this  weak  input  desired  signal.  A significant  amount  of 
distortion  due  to  the  FM-interfering  signal  is  observed. 


-70- 


Figure  A-3.  Spectrum  of  Desired  Signal  (108.5  MHz)  at  RF-Amplifier 
Output.  Desired  Signal  Strength  -47  dBm  (1  mv);  FM- 
interfering  Signal  1 105.5  MHz,  P - 6,  Bandwidth  = 200 
KHz)  0 dBm  (.2  volts  ^ . 


5 msec/div 


Figure  A-4.  Detector  Output  with  No  interference.  Carrier 
frequence  - (08.5  MHz;  Modulation  frequency 
1 KH?  10  Modulation. 


71 


5 rrv;ec  di 


Figure  A-5.  Del-,-.  t \.'l' 
(0  dBrr '•  ul  10: 


!csireri  S 


VfflJkf’Cty 

7 ■ ■ 1 ■ 

'. 'j4>  '•  £>y  f*viw  TKtf 

[tS'SsMrHrcfr  '?‘wl  vA'x  .•&{$  i 

. 5 msec  div 

Figure  A-',  Defe  jrrai  Level 

• 0 dr.-  . '\lo  inferfei  ng  S igna  I . 


■ *"?  v-it: 


B . Filter  Characteristics 

1 . Measurement  of  RF  Amplifier  Input  Filter  Characteristics  ci(u)-  The  test 
setup  to  measure  the  RF  ampl  ifier  fi  Iter  characteristics  0(0)  is  shown  in  Figure  B-l. 


Figure  B-l.  Measurement  Procedure  for  Obtaining  a (w). 


The  voltage  at  the  input  to  the  gate  of  dual  gate  MOSFET  of  the  RF  Amplifier  was 
measured  as  a function  of  Wavetek  frequency.  The  voltage  at  the  RF  amplifier  input  was 
measured  using  a H P 8405A  vector  voltmeter . 

The  AGC  voltage  of  the  receiver  was  fixed  and  the  signal  ievel  from  the  Wavetek 
3000  was  held  constant. 

The  resultant  filter  characteristics  measured  were  found  to  be  shifted  in  frequency 
from  the  receiver  frequency  by  2 MHz.  It  was  assumed  that  this  was  due  to  the 
capacitive  loading  of  the  vector  voltmeter  X10  probe  used.  Therefore,  the  characteristic 
was  shifted  over  2 MHz  so  that  the  resonant  frequency  of  the  filter  is  the  desired  frequency 
108.5  MHz.  Figure  B-2  gives  a simplified  schematic  of  the  Nav  11  front  end. 


Figure  B-2.  Simplified  Schematic  of  RF  Front  End 
of  NAV  11  Receiver. 


- 


-74- 


In  order  to  verify  that  the  RF  Amplifier  input  filter  characteristics  were  correct, 
cross-compression  and  intermodulation  measurements  were  made  varying  the  frequency 
of  the  interfering  signals. 

The  intermod  measurements  given  in  Figure  4 of  the  main  text  shows  that  the 
a(u)characteristics  measured  result  in  constant  K^/K.  (within  experimental  accuracy). 
The  test  procedure  illustrated  in  Figure  B-3  was  used  to  measure  cross-compression  of 
the  desired  signal  as  a function  of  interfering  signal  frequency  and  the  results  give  the 
following  for  a(cj). 


Figure  B-3.  Test  Procedure  for  Determining  a (u)  Using 
Cross -Compress  ion  Methods. 


Interfering 

Frequency 

MHz 

Equivalent  a ( u ) 
Calculated  from 
Measured  GC* 

e(u)  Found 
Probing  RF 
Front  End 

107.5 

- 1 dB 

-1.2  dB 

106.5 

-3.1  dB 

-2.7  dB 

105.5 

- 5 dB 

-4.5  dB 

104.5 

-6.6  dB 

-6.2  dB 

*Gain  Change 

Table  B-l . Comparison  of  Results  Obtained  Using  Cross- 
Compression  and  RF  Measurements . 


-75- 


Because  cross -compress ion  is  a function  of  a (u)  it  was  not  possible  to  obtain 
enough  interfering  signal  level  with  the  above  test  procedure  to  measuie  1 dB  of  GC 
for  frequencies  below  104.5  MHz. 

2.  Measurements  of  Variations  in  a (u)  as  a Function  of  Receiver  Tuning. 
Investigations  of  the  variations  in  a(o)  as  a function  of  frequency  were  made  by  tuning 
the  receiver  to  different  frequencies  while  keeping  the  interfering  signal  frequency  at 
a fixed  3 MHz  below  the  receiver  frequency  and  measuring  the  resulting  gain  change 
(GC).  The  variations  in  a(u)  with  receiver  tunings  are  given  in  Table  B-2. 


Receiver 

Frequency 

(MHz) 

Interfering 

Frequency 

Calculated  cr(u) 
Assuming  K^/K  ^ 
Constant 

108.5 

105.5 

-4.5  dB 

110.5 

107.5 

-4.2  dB 

112.5 

109.5 

-4.0  dB 

114.5 

111.5 

-3.6  dB 

116.5 

113.5 

-3.4  dB 

117.5 

114.5 

-3.2  dB 

Table  B-2.  Variations  in  a(u)  as  a Function 
of  Receiver  Frequency. 


-76- 


c. 


Interference  Due  to  ELT 


].  Introduction.  There  have  been  numerous  reports  where  particular  ELT's  have 


been  shown  to  aggravate  the  interference  problem  by  reradiating  signals  generated  by  the 
nonlinear  mixing  in  the  ELT  output  transistor  stage.  It  is  believed  that  the  transmitter 
(inactivated)  output  circuit  acts  as  a diode  mixer  generating  interfering  signals  in  the  lo- 
calizer and  communication  bands,  particularly  in  the  communications  frequency  band  (118- 
136  MHz). 


The  interference  frequency  is  the  result  of  a third-order  nonlinearity  term.  For 
example,  an  interfering  signal  at  122  MHz  can  result  from  the  mixing  of  two  FM  stations 
at  108  and  94  MHz  (2  fj  - ^ ) term,  due  to  a third-order  nonlinearity).  It  has  also  been 
reported  that  it  is  possible  to  observe  enchancement  of  the  interference  problem  due  to 
the  ELT  as  a result  of  the  combining  of  an  ILS  , VOR  and  FM  station.  For  example,  an 
ILS  frequency  of  329  MHz^TVOR  of  113.8  MHz  and  FM  station  at  94  MHz  could  mix 
producing  an  interferin^signal  at  121.2  MHz  ( lj^  - ^ “f2term)- 

A third-orde/model  has  been  investigated  for  modeling  the  effects  of  ELT  in  the 
interference  problerr^.  It  was  expected  such  j model  would  be  adequate  for  predicting 
interference  levels  and  because  of  its  simplicity,  it  could  be  utilized  effectively  in  prac- 
tice. Figure  C-l  illustrates  the  results  obtained  when  two  interfering  signals  are  directly 
coupled  into  the  ELT.  The  experimental  procedure  for  obtaining  these  results  is  shown  in 
Figure  C-2.  Figure  C-l  illustrates  the  results  obtained  for  various  signal  levels.  The  third- 
order  term  (102.5  MHz)  clearly  shows  up  in  the  spectral  plots  shown  in  this  figure.  In  ad- 
dition, it  is  observed  that  at  extremely  high  signal  levels  4 5 dBm  and  above  additional 
intermod  terms  appear  (see  Figures  C-l  e,j,k  and  I)  which  would  tend  to  indicate  that  a 
higher  order  model  may  be  required. 


Since  the  primary  interest  is  in  the  analytical  modeling  of  the  ELT's  role  in  the 
interference  problem,  it  is  necessary  to  be  able  to  predict  the  signal  levels  coupled  from 
the  EL7  antenna  to  the  communications  and/or  navigational  antennas.  Measurements  were 
obtained  and  a worst  case  theoretical  analysis  was  performed. 


2.  ELT  Model.  A proposed  analytical  model  representing  the  ELT  is  illustrated  in 


Figure  C-3a.  Mathematical ly  the  model  may  be  interpreted  in  two  ways: 


(1)  The  voltage  e can  be  represented  in  terms  of  the  received  voltage  by 


. .2,3 

e - k.  e.+  k.  e.  + k„  e. 
a I i z i 3 i 


(C-l) 


(2)  A second  and  somewhat  more  satisfying  way  to  visualize  the  model  is  illus- 
trated in  Figure  C-3b.  In  this  representation  the  incident  voltage  v+  is  the  received 


-77- 


Figure  C-l.  Spectra  of  the  Reradiated  Signals  at  the 

Output  of  ELT.  Vertical  = 10  dB/1  .5  cm; 
Horizontal  scale  = 2 MHz/cm. 


-78- 


Figure  C-4.  Experimental  Procedures. 


Wavetek  Model 
3000 


Bi-directional 

Coupler 


Wavetek  Model 
3000 


(b) 


Figure  C-4.  Continued 


-86- 


I 

I 


AD-A063  796  OHIO  UNIV  ATHENS  DEPT  OF  ELECTRICAL  ENGINEERING  F/6  17/7 

ANALYTICAL  DETERMINATION  OF  THE  INTERFERENCE  OF  COMMERCIAL  FM  S-ETC(U) 
MAR  78  J E ESSMAN*  T LOOS  D0T-FA77WA-3932 

UNCLASSIFIED  EER-32-1  FAA-R-6050-1  NL 


2 of  2 

AO 

A063798 

■ 

*<-'  <• 

! (• 

(• 



END 

DATE 

FILM  D 

3-79 

DOC 

signal  and  the  reflected  voltage  is  the  reradiated  signal , i.e., 


v = kj  (v+)  + \<2  (y4)  + (v  (C-2) 

It  can  be  shown  that  both  of  these  models  are  essentially  equivalent,  the  basic 
difference  being  in  the  interpretation.  Neither  model  postulated  takes  into  account 
frequency  selective  characteristics  of  the  ELT  output  circuitry  and  both  are  limited  to 
third-order  nonlinearities.  Also,  the  high  frequency  characteristics  of  the  p-n  junction 
are  ignored  in  this  preliminary  investigation. 

Using  the  models  specified  above  the  two  experimental  procedures  illustrated  in 
Figure  C-4  were  used  to  determine  the  parameter  k^  for  the  model.  Although  a consid- 
erable amount  of  data  relating  to  self-and  cross -compress ion  was  obtained,  the  primary 
important  data  is  relative  to  the  intermod  frequency.  In  this  section  summary  results  are 
given  in  graphical  form.  The  test  procedure  illustrated  in  Figure  C-4a  was  used  to  obtain 
the  results. 

Figure  05  il lustrates the  effects  of  two  closely  aligned  frequencies.  In  particular 
this  curve  was  obtained  using  two  signals  separated  by  1 MHz  and  extending  over  the  FM 
band.  This  curve  tends  to  indicate  that  whenever  two  closely  aligned  frequencies  interact 
the  coefficient  k^varies  approximately!  5 dB  from  a reference  value  arbitrarily  chosen  at 
the  ELT  frequency  of  121.5  MHz.  Some  of  these  effects  can  be  explained  by  considering 
the  frequency  characteristics  of  the  ELT  output  stage  which  are  in  Figure  C-6.  In  this  the 
corrected  voltage  measured  at  the  ELT  input  with  a - 10  dBm  incident  signal  is  plotted  as  a 
function  of  frequency. 

A rough  approximation  which  ignores  both  frequency  effects,  (capacitance  of  the 
junction),  signal  levels,  etc.,  is  to  use  the  low  frequency  model  of  the  p-n  junction,  i.e., 

I - I.  ( e ‘>vAT-l) 

where 

lG  = reverse  current  in  amperes 
k = Boltzmann's  constant 
T = temperature  in  °k 
v = applied  voltage 
q = electron  charge  in  coulombs 

Using  this  model  would  imply  that  the  intermod  term  would  have  an  amplitude  given 
approximately  by  (A  + 2 B + 15) 


Strong  Frequency  1 MHz  less  than  intermod 
Weak  Frequency  2 MHz  less  than  Intermod 


FM  -an  Communications  ran 


where 


A = amplitude  of  fhe  weak  signal  (for  the  case  shown  in  Figure  C -6  = -10  dBm) 

B - amplitude  of  the  strong  signal  (for  the  case  shown  in  Figure  06  = 0 dBm) 

The  15  dBm  is  the  result  of  expanding  the  exponential  in  a series.  For  the  case 
considered  here  this  would  imply  that  the  intermod  term  would  have  an  amplitude  of 
approximately  -25  dBm.  Using  Figure  05  this  result  seems  to  be  reasonable  when  the 
intermod  frequency  is  approximately  121 .5  MHz;  however,  it  deviates  somewhat  from  this 
value  for  different  frequencies.  It  must  be  stressed  that  the  equation  used  to  describe  the 
p-n  junction  is  by  no  means  accurate  at  these  frequencies;  however,  bounds  on  the  problem 
can  be  obtained.  Further  investigations  are  required  in  order  to  arrive  at  a more  accurate 
description. 

A large  amount  of  data  was  obtained  which  tended  to  indicate  that  the  coef- 
ficient k^  can  be  assumed  relatively  constant  over  wide  frequency  ranges.  Although  the 
data  obtained  is  too  numerous  for  inclusion  in  this  brief  report,  a summary  of  the  results  is 
illustrated  in  Figures  C-7  and  C-8.  The  data  was  obtained  using  the  test  procedure  illus- 
trated in  Figure  C -4a.  Figure  C-7  indicates  the  variation  in  k^  as  a function  of  the  inter- 
mod frequency  as  indicated  on  the  graphs.  These  curves  were  obtained  with  the  two  signals 
held  constant  at  -10  and  0 dBm  respectively.  These  results  indicate  that  the  coefficient 
k^  is  constant  within  ± 5 dB  of  the  value  at  121.5  MHz  (ELT  frequency). 

Figure  C-8  shows  the  variation  in  k^  as  a function  of  signal  amplitude.  For  signal 
levels  of  0 dBm  and  belovjk^  can  be  bounded  by  approximately  ± 5 dB.  Results  indicate 
that  k^  deviates  more  for  signals  with  frequencies  in  the  lower  part  of  the  FM  band  than  for 
frequencies  in  the  middle  of  the  band. 

3.  ELT  Communications  Antenna  Coupling.  In  order  to  determine  the  effects  on 
communication  receivers  of  an  intermod  signal  reradiated  by  the  ELT,  it  is  necessary  to  have 
some  measure  of  the  coupling  between  the  ELT  and  communications  antennas.  Experimental 
measurements  were  performed  on  a Cessna  150  commuter  (N1600U)  using  the  test  procedure 
illustrated  in  Figure  C-9.  The  relative  locations  of  the  communications  and  the  ELT  antenna 
are  illustrated  in  Figure  C-10.  Figure  C-ll  shows  the  results  obtained. 

Although  the  mathematical  calculations  for  aircraft  antennas  are  very  complex  and 
can  only  be  approximated  very  crudely,  it  is  of  interest  to  achieve  a worst  case  model. 

The  relations  between  a set  of  coupled  antennas  can  be  given  by 
n 

v;  = *j  =|  (2  ij)lj  i = 1 to  n (C-3) 


-89- 


Frequencies  generating 
Intermod  term 


intermod  Frequency  in  M'’z 

Figure  C-7.  as  a function  of  intermod  frequency 

"or  varying  FM  band  generating  frequencies. 


Figure  C-9.  Antenna  Coupling  Measurement. 


co-^ 


Top  View 


ELT  Antenna 


/\ 


Communications  Antenna 


Figure  C-10.  Antenna  Locations. 


-30 


where 


Zii  = self  impedance  of  antenna 


Z ij  = Z ji  = mutual  impedance  between  antenna  i and  antenna  j . 

For  simplicity  the  following  assumptions  are  made: 

1.  The  ELT  antenna  (transmitting)  has  a 50  ohm  self-impedance. 

2 . The  communications  antenna  (receiving)  has  a 50  ohm  self-impedance. 

3.  Both  antennas  are  matched. 

4.  Both  antennas  are  \/4  monopoles  separated  by  \/2  (actually  the  ELT 
antenna  was  measured  to  be  16  inches  and  the  communication  antenna 
was  measured  to  be  23  inches). 

5.  Both  antennas  have  infinite  ground  planes. 

6.  Antennas  are  parallel  and  on  same  plane. 

Using  the  above  relations  and  the  curves  of  mutual  impedance  for  two  parallel 
antennas  as  given  in 

"Antenna  Engineering  Handbook"  by  Henry  Jasik , McGraw-Hill, 

the  coupling  can  be  calculated  as  - 16  dB,  a considerably  more  pessimistic  value  than  the 
measured  value  . Figure  C — 1 1 indicates  the  maximum  measured  coupling  was  approximately 
-28  dB.  Further  investigations  for  refining  this  model  are  required;  however,  the  calcu- 
lations could  be  considered  a worst  case  analysis.  For  a more  realistic  value  of  the  coupling, 
the  measured  values  could  be  used. 

ELT's  used  on  many  Cessna  aircraft  seem  to  be  more  susceptible  to  third-order  inter- 
modulation phenomenon.  Reportedly,  the  Cessna  Company  has  partially  solved  this  problem 
by  certain  modifications.  Mr.  Gordon  Wood  of  the  Cessna  Corporation  in  Wichita,  Kansas, 
was  contacted  regarding  this  and  he  said  that  the  problems  of  interference  are  greatly  re- 
duced (to  tolerable  levels)  if  the  ELT  antenna  is  shortened  from  the  original  22"  to  16"  and 
adjustments  in  the  output  stage  and  cabling  for  matching  and  maximum  power  are  made.  The 
effect  of  this  would  seem  to  be  a reduction  in  the  coupling  between  the  ELT  and  communi- 
cation antennas.  Ohio  University  has  available  several  small  aircraft  instrumented  with  the 
Share  7 ELT  which  allowed  practical  measurements  on  both  aircraft  with  modified  and  non- 
modified  antennas.  These  aircraft  along  with  the  antenna  lengths  are  listed  as  follows 


-96- 


for  reference: 


Cessna  150  Commuter  N1600U 

Antenna  Length 

- 16 

N 900U 

M II 

16 

N 1200U 

II  II 

= 22 

II 

II 

II 


In  addition  to  the  decrease  in  antenna  length  from  22"  to  16",  the  locations  of  the 
22"  and  16"  antennas  on  the  aircraft  were  slightly  different.  Although  no  evidence  is 
present  which  would  indicate  this  to  be  a significant  factor,  it  is  possible  that  the  location 
was  moved  so  asto  decrease  the  coupling  between  the  communications  antenna  and  the  ELT 
antenna.  It  is  unlikely  that  the  reduction  in  coupling  obtained  is  totally  due  to  decreasing 
the  length  of  the  antenna. 

4.  Summary  and  Applications.  In  summary  and  as  an  illustration  of  the  techniques 
and  models  developed  here  a worst  case  analysis  (very  approximate)  is  given  illustrating  the 
role  of  the  ELT  in  the  interference  problem.  For  the  example,  two  100  kw  FM  stations  are 
assumed  - one  located  15  miles  from  the  receiver  and  the  other  5 miles  from  the  receiver. 
The  geometry  is  illustrated  in  Figure  C-12.  Only  the  interference  caused  by  the  ELT 
generated  intermod  is  considered  in  this  example.  Using  the  information  and  data  given  in 
this  report,  the  following  step-by-step  computations  can  be  made: 

a.  Aircraft  with  16"  ELT  Modified  Antenna 

I.  Coupling  between  ELT  and  communications  antennas  at  120  MHz 
(intermod  frequency) 

From  Section  3 of  this  report 

Theory  (worst  case)  = - 16dB 

Measured  = - 35dB 

The  large  discrepancy  between  theory  and  measurement  is  probably  due  to  many 
factors  such  as: 

1)  The  modified  ELT  antenna  of  length  16"  is  not  a X/4  antenna  as  assumed  in 
the  calcuations. 

2)  The  possible  location  change  of  the  ELT  antenna. 

3)  Communications  and  ELT  antennas  not  being  parallel,  as  assumed. 

4)  ELT  and  communications  antennas  being  in  different  planes. 

II.  Calculation  of  the  incident  voltage  at  the  ELT  assuming  the  following: 

(a)  Directive  gain  of  the  ELT  antenna  = 6 dB 


(b)  Attenuation  equation  between  lossless  isotropic  antenna 
holds,  i.e.  , 

a (decibels)  = 36.3  + 20  log^  f + 20  log^d 

where 

f is  the  frequency  in  MHz 
d is  the  distance  in  miles 

FM  station  ^1  (108  MHz) 

P ^ (power)  = 80  dBm  (50  ohm  reference) 

o ] = 91  dB 

FM  station  ^ 2 (96MHz) 

Pj  = 80  dBm 
a = 100  dB 

The  power  at  the  ELT  is  obtained  using  the  following  equation: 

P'  = gain  (antenna)  + ERP  - attenuation 
llT 


P'  (ELT)  = -5  dBm 


FM  *2 

P'2  ELT  - -14  dBm 

III.  Intermod  Calculation: 

3 

Using  the  value  of  2 k^  = - 8 dB  given  in  this  report  (Figure  C-8),  the  reradiated 
signal  from  the  ELT  is  given  by 

3 

RC1  _ = A + 2B  + 7 k 
ELT  3 


where 


(A)  = amplitude  of  station  1 (weak  station)  in  dBm 


-99- 


(B)  = Amplitude  of  station  2 (strong  station)  in  dBm 


Therefore, 


(reradiated  signal  power)  - 32  dBm 


IV.  Calculation  of  intermod  input  at  the  receiver  terminals. 

(1)  Using  the  worst  case  coupling  of  - 16  dB 

Intermod  = - 48  dBm  ( ~ 900  qv) 

(2)  Using  the  measured  coupling  of  - 35  dB 

Intermod  = - 67  dBm  ( ~ 100uv) 

b.  Aircraft  with  22"  ( Non-Modified)  Antenna. 

Investigations  and  inquiries  as  to  the  possible  modifications  of  the  calculation  pro- 
cedures to  include  the  effect  above  were  made.  These  findings  indicate  that  numerical 
procedures  should  be  possible  which  would  allow  for  a more  accurate  specification  of  the 
coupling  between  antennas  located  on  aircraft.  Such  an  undertaking  is  beyond  the  work 
defined  under  the  current  contract  and  would  consist  of  a rather  significant  program  in 
itself.  It  would  appear  in  view  of  the  increased  possibilities  of  interaction  between 
electronic  devices  on  board  aircraft  (one  example  being  the  ELT  and  communications  an- 
tennas), it  would  be  beneficial  for  the  FAA  to  consider  supporting  investigations  in  these 
areas  either  as  an  extension  of  the  scope  of  the  current  study  or  possibly  an  additional  project 
in  the  future.  Although  neither  time  nor  funds  have  been  allocated  on  the  current  contract 
for  these  investigations,  the  analytical  techniques  which  have  been  determined  are 
believed  to  indicate  worst  case  conditions. 

Results  given  in  Section  A indicate  theoretically  a coupling  of  -16  dB  between 
the  ELT  and  communications  antennas,  while  measured  values  for  16"  ELT  antenna  gives  a 
coupling  of  -35  dB.  Further  measurements  were  made  on  aircraft  Cessna  N1200U,  150 
commuter  equipped  with  a Share  7 ELT  and  a non-modified  antenna  of  22"  in  length  and 
another  aircraft  Cessna  N900U,  Commuter  150,  with  the  modified  antenna.  These  results 
are  summarized  in  Figures  C-13,  C-14,  C-15,  and  C-16. 

Figure  C-13  shows  the  coupling  between  the  ELT  antenna  (non-modified  length  -22") 
and  the  communication  receiver.  These  results  were  obtained  by  exciting  the  ELT  antenna 
and  measuring  the  voltage  at  the  output  of  the  communications  receiver.  The  results  of 
Figure  C-13  indicate  a measured  coupling  approximately  -16.5  dB . Comparing  this  with  the 
calculated  value  of  -16  dB  given  above,  the  agreement  is  very  good.  This  is  the  result 


gp  u;  6uj|dno^) 


101 


Figure  C-13.  Coupling  Between  ELT  Antenna  and  Communication  Antenna  as  a Function  ot  Frequency 
22"  ELT  Antenna  Length. 


C-16.  Normalized  (50  ohms)  Self  Impedance  at  Various 
Frequencies  of  the  Communications  Antenna  on 
Board  N1200U. 


-104- 


of  the  geometry  of  the  ELT  system  on  board  N1200U  being  more  representative  of  the  worst 
case  system  used  in  Section  A.  The  calculated  intermod  input  at  the  receiver  was  - 48  dBm 
( <*>  900  |jv)  using  a coupling  of  - 16  dB.  Such  a level  would  result  in  an  extremely  severe 
interference  problem  which  would  be  of  considerably  more  importance  than  either  the 
"brute-force"  or  the  normal  intermod  (generated  by  the  interaction  of  two  or  more  FM 
stations  of  appropriate  frequencies)  types  of  interference.  For  reference  Figures  C-14,  C-15, 
and  C-16  are  included  which  indicate  the  self-impedance  of  the  22"  ELT  antenna  the  16" 

ELT  antenna  and  the  communications  antenna  respectively.  It  is  noted  that  impedance  of 
the  modified  16"  antenna  is  somewhat  independent  of  frequency  and  far  from  the  assumed 
50  ohms  used  in  the  calculations. 

In  summary,  results  have  indicated  that  the  analytical  model  agrees  with  the  22"  ELT 
antenna  measured  results  rather  well  and  may  be  considered  a worst  case  situation;  however, 
it  does  not  agree  with  the  measured  values  obtained  using  the  modified  ELT  antenna.  Since 
this  was  not  a specific  delineated  task  on  this  contract  investigations  for  refining  the  ELT 
model  were  suspended  at  this  point. 


■ 


-105- 


D. 


Measurement  of  RMS  Modulation  Due  to  Intermod 


1.  Measurement  Technique.  To  measure  the  amount  of  RMS  modulation  due  to 
intermod  the  following  set  up  was  used: 


Desired  Signal 


The  outputs  of  three  RF  generators  are  combined  through  bidirectional  couplers 
and  inputted  to  the  NAV  11  receiver.  Two  of  the  generators  are  FM  modulated  by  audio 
sine-wave  generators.  The  other  generator  is  the  desired  signal  tuned  to  the  same 
frequency  as  the  receiver.  A true  RMS  voltmeter  is  used  to  measure  the  receiver  filter 
outputs . 


Using  two  FM  signals  with  characteristics  given  below  the  detector,  the  audio 
filter  output,  the  90  Hz  and  150  Hz  filter  outputs  are  measured,  i .e.. 

Desired  Signal:  f = 108.5  MHz,  Carrier  only,  -60  dbm  signal  level 

FM1 : f = 102.5  MHz,  fm  = 400Hz,  p = 30,  signal  level  = -18  dbm 

FM2 : f = 105.5  MHz,  fm  = 1 KHz,  p = 20,  signal  level  =-17  dbm. 

Measured  IM  Carrier  level  = -70  dbm. 


-106- 


Detector  output  voltage  = .155VRMS 
Audio  output  voltage  = .305  VRMS 


150  Hz  output  voltage  = .11  VRMS 

90  Hz  output  voltage  = .10  VRMS 

In  order  to  compute  the  necessary  constants  the  following  normalizing  constants 
were  determined: 

(1)  Detector  voltage  with  20%  AM  modulation  output  was  measured  as  .250  volts 
volts  with  fm  =20  Hz 

(2)  Audio  voltage  with  10%  modulation  at  1 KHz  = .665  volts 

(3)  90  Hz  filter  output  voltage  with  20%  modulation  at  95  Hz  = .891  volts 

(4)  150  Hz  filter  output  with  20%  AM  modulation  at  145  Hz  =.999  volts 

Using  the  normalizing  constants  calculations  of  % modulation  can  be  made  as 
shown  below: 

Calculating  Equivalent  RMS  Modulation  of  Filter  Outputs 
.155 

Detector  output  = ,^q  x 20%  = 12.1%  modulation 

305 

Audio  output  = ' ^52  x 10%  ~ 4 .7% 

150  Hz  filter  output  = -gg-j  x 20%  = 2 .5% 

90  Hz  filter  output  = .10  x 20%  = 2.0% 

It  is  of  interest  to  note  that  the  theoretical  results  taken  from  output  of  the  computer 
program  given  in  Table  28  were  obtained  with  an  IM  carrier  to  desired  signal  ratio  of 
-7.7dB.  The  measured  IM  to  signal  ratio  was  -lOdB  using  AGC  method  and  -8  .3  dB  by 
measuring  detector  voltage  output  with  IM  carrier  and  desired  signal  only  and  comparing 
it  to  a detector  voltage  with  known  modulation. 

2.  "Flutter11  of  IM  Relative  to  Desired  Signal.  Several  problems  were  observed 
during  the  experiments  due  to  the  narrow  band  widths  of  the  90  and  150  Hz  filters.  One 
was  the  "flutter"  which  is  investigated  using  an  intermod  carrier  and  a desired  signal 
as  inputs  to  the  receiver.  The  detector  voltage  was  monitored  with  an  oscilloscope  and 
its  frequency  measured  with  a frequency  counter.  The  resultant  frequency  was  measured 


-107- 


to  be  approximately  100  Hz.  The  waveform  observed  was  a sine  wave  with  varying 
frequency  due  to  slight  frequency  fluctuations  in  the  desired  frequency  and  IM 
frequency.  The  average  frequency  drift  was  estimated  to  be  approximately  100  Hz. 

Since  the  90  and  150  Hz  filter  functions  have  very  sharp  bandpass  characteristics  on 
the  order  of  30  Hz,  the  100  Hz  IM  "flutter"  becomes  very  significant. 

For  calculation  purposes  an  approximation  of  the  filter  characteristics  was  used 
in  the  program  to  compute  modulations  factors  for  the  90  and  150  Hz  filter  characteristics. 

These  factors  vary  as  much  as  25%  with  the  modulation  frequencies  varying  as  little  as 

a few  hertz.  The  theoretical  modulation  factorsdid  not  agree  well  with  the  measured  < 

modulation  factors  which  were  found  to  vary  little  as  the  modulation  frequencies  were 

changed.  In  order  to  obtain  useful  theoretical  modulation  factors  for  the  90  and 

150  Hz  filters,  ideal  bandpass  filters  with  the  same  noise  equivalent  bandwidth  as  the 

actual  filter  functions  were  used.  The  ideal  bandpass  filters  have  enough  bandwidth 

so  as  to  get  a good  sampling  of  the  intermod  generated.  This  is  evidenced  as  observed 

in  the  good  agreement  between  theoretical  results  and  measured  results.  This  method 

as  given  in  Table  28  in  the  text  seems  to  work.  With  low  modulation  indices  and 

higher  modulation  frequencies  (when  the  intermod  does  not  have  enough  components 

which  land  in  the  receiver  passband  so  as  to  make  the  intermod  "look"  noise-like) 

results  will  not  agree  as  well. 

Although  the  90  Hz  filter  has  a half  power  bandwidth  of  approximately  35  Hz, 
its  noise  equivalent  bandwidth  (found  by  integrating  the  measured  filter  characteristic 
numerically)  is  168  Hz.  This  is  because  the  skirts  of  its  characteristics  do  not  fall  off 
rapidly. 

Similarly  the  -3  dB  bandwidth  of  the  150  Hz  filter  is  55  Hz  and  its  noise 
equivalent  bandwidth  is  195  Hz. 

3.  Measurement  of  Modulation  Factors.  Modulation  factors  due  to  intermod 
are  measured  by  inputting  the  lM  carrier  and  desired  signal  into  the  receive  jnd  measuring 
the  filter  output  voltage.  For  the  detector  filter,  the  IM  carrier  frequency  is  set  to  the 
desired  signal  frequency.  For  the  audio  filter,  the  IM  frequency  is  set  to  1 KHz  above  or 
below  the  receiver  frequency. 

The  interfering  signals  are  then  FM  modulated  (with  the  IM  carrier  frequency 
equal  the  receiver  frequency)  and  the  RMS  output  of  the  filter  is  measured. 

The  result  of  measuring  the  IM  carrier  filter  voltage  output  is  to  provide  a 
normalizing  voltage.  Variations  in  gain  of  various  parts  of  the  receiver  are  effectively 
divided  out  by  using  modulation  factor  measurements.  This  is  especially  true  for  the 
audio  filter  which  has  a gain  determined  by  the  setting  of  the  volume  control.  (For 
measurement  purposes,  the  volume  control  was  replaced  by  a 10  turn  trim  pot  and 
adjusted  for  1/5  full  volume.) 


-108- 


_ A. 


i 


Measured  values  of  modulation  factors  were  found  to  agree  very  well  with 
theoretical  values.  Measurements  of  the  actual  IM  carrier  level  found  by  comparing 
the  filter  output  voltage  to  an  equivalent  filter  voltage  with  the  desired  signal  AM 
modulated  with  a known  percentage  do  not  seem  to  agree  as  well.  A disadvantage  of 
the  MF  method  is  that  the  90  and  150  Hz  filter  modulati  on  factors  cannot  be  measured 
this  way  since  the  IM  carrier  frequency  cannot  be  adjusted  to  90  or  150  Hz  away  from 
the  desired  frequency.  Also  it  was  found  that  the  audio  output  overloads  with  more 
than  20%  modulation. 


E-  Noise  Equivalent  Bandwidth  The  noise  equivalent  bandwidth  is  defined  as 

the  bandwidth  of  an  ideal  filter  with  the  same  RMS  output  voltage  as  the  actual  filter. 
These  concepts  are  illustrated  in  Figure  E-l , where  B is  defined  as  the  noise  equivalent 
bandwidth,  n.  is  the  input  spectral  density  of  a "wRite"  noise  source  and  S (f)  is  the 
output  spectral  density.  ° 


Figure  E-l  . Noise  Equivalent  Bandwidth  of  a Filter. 


The  RMS  output  voltage  is  given  by 


(N  ) RMS 
o 


/°°  I H $w)  I2  df 
/ -oo 


The  noise  equivalent  bandwidth  is  then  defined  by 
(N°)  RMS  = ' "I  1 H 0 <*>)  I2  df  = ' H('U)  max  '^i 

and  I H (|  u)  I2  d f 

B = 

" I H (j  u)  I2 

M max 

It  can  be  shown  that  the  (MF)  Ratio,  i.e.. 

Modulating  Factor  (MF)  of  Audio  /B  Audio 


(MF)  Ratio  = 


Modulating  Factor  (MF)  of  Audio  ^B  Audio 

= n 

Modulating  Factor  (MF)  of  Detector  X/B  Detector 

n 


depends  only  on  the  filter  functions  or  the  noise  equivalent  bandwidths  of  the  filters. 
Several  calculations  were  made  to  verify  these  facts  and  the  results  are  given  in  Table  E-l 
Note  a very  high  modulation  index  and  relative  low  modulation  frequency  are  used  to 
obtain  a noise-like  spectrum. 


Station  1 

Station  2 

Carrier 
Freq . i n 
MHz 

mm 

Modulation 

Index 

Carrier 
Freq . i n 
MHz 

Modulation 
Freq . i n 
KHz 

Modulation 

Index 

Detector 

MF 

7 

Audio 

MF 

MF 

Ratio 

105.5 

1 .000 

30. 

102.5 

.742 

30. 

.240 

.113 

.471 

105.5 

1 .250 

24. 

102.5 

1 .042 

24. 

.241 

.107 

.444 

105.5 

.906 

55.2 

102.5 

1 — 

.543 

55.2 

.184 

.085 

1 

.462 

L 


Table  E-l 


Comparison  of  Theoretical  Modulation  Factor  Ratios 
From  Computer  Program, 


Determination  of  the  actual  noise  equivalent  bandwidth  B of  the  receiver  filter 
(audio,  detector,  90  and  150  Hz  filters)  requires  numerical  integration  to  find  the  area 
under  the  actual  filter  response  curves.  These  characteristics  were  obtained  for  the 
NAV  11  receiver  The  results  obtained  are: 


Audio  Filter  (see  Figure  18  of  the  text  for  response  curve): 


B =1.08  KHz 
n 


Detector  filter  (see  Figure  1/  of  the  text  for  response  curve): 


B =4.93  KHz 
n 


Using  the  above  data  the  MF  ratio  is  found  to  be 


yi  .08  KHz 

(MF)  ratio  = = .468 


. '4.93  KHz 


(E-4) 


Comparing  this  result  with  those  given  in  Table  E-l  it  is  seen  that  the  MF  ratio 
determined  by  filter  characteristics  and  those  from  IM  calculations  using  the  computer 
program  agree  very  well.  Similar  techniques  can  be  applied  to  the  90  and  150  Hz  filter 
functions.  Because  of  the  good  agreement  of  the  "noise"  model  of  the  IM,  the  90  and 
150  Hz  filter  functions  were  modeled  in  the  computer  program  by  a filter  with  the  same 

noise  equivalent  bandwidth.  See  discussion  in  Appendix  D. 

-Ill- 


F.  CDI  Interference.  Figure  18  in  the  text  shows  the  test  setup  used  to  determine 
the  effects  of  high  level  modulation  on  the  CDI  response.  These  results  are  obtained  by 
injecting  a 108.501  MHz  signal  along  with  a desired  signal  of  108.500  MHz.  The 
108.501MHz  signal  results  in  an  injected  1 KHz  sideband  signal  in  the  IF  which  is 
detected.  Tables  F-l , F-2,  and  F-3  give  the  results  obtained. 

The  results  given  in  Table  F-4  show  the  CDI  variations  as  a function  of  sideband 
frequency . 


Interfering  signal  level 
relative  to  desired  level 

Measured 

CDI 

Percent  change  in  CDI 
reading  due  to  sideband 

-15  dB 

58  pA 

3% 

-12  dB 

55  pA 

8% 

-10  dB 

51  pA 

15% 

-8  dB 

48  pA 

20% 

- 6 dB 

42  pA 

30% 

- 3 dB 

31  pA 

48% 

Table  F-l . Effects  of  Detector  Clipping  on  CDI  (Normal  Reading  of  CDI  - 60  pA) . 


Interfering  signal  level 
relative  to  desired  level 

Measured 

CDI  (pA) 

Percent  change  in  CDI 
reading  due  to  sideband 

-15  dB 

30.7 

3 .4% 

-12  dB 

28.9 

9.1% 

-10  dB 

27.7 

12.9% 

-8  dB 

25.8 

18.8% 

- 6 dB 

21.0 

34  % 

- 3 dB 

18.4 

42  % 

Table  F-2.  Effects  of  Defector  Clipping  on  CDI  (Normal  Reading  of  CDI  =30  pA). 


-112- 


Fixing  AGC  at  its  value  with  no  interfering  signal  present  and  introducing  1 KHz 
sideband,  gives  the  same  CDI  deflections  as  before;  therefore,  the  CDI  interference  is 
not  due  to  AGC  changes  of  the  receiver. 


Table  F-4  shows  the  CDI  reading  as  a function  of  sideband  frequency.  Interfering 
signal  level  is  fixed  at  -3  dB  lower  than  desired  signal  level  (70%  modulation). 


Sideband  frequency  in 

KHz  above  desired  level 

Measured 

CDI 

1 KHz 

31  .2  pA 

2 KHz 

29.0  pA 

4 KHz 

34 .4  pA 

6 KHz 

38  pA 

8 KHz 

37.7  pA 

10  KHz 

32.5  pA 

12  KHz 

26.9  pA 

15  KHz 

49.7  pA 

Table  F-4.  Effects  of  Interfering  Signal  Frequency  on  CDI  (Normal  Reading  of  CDI  - 60  pA) . 

Note  that  the  CDI  values  as  the  measured  IF  frequency  response  varies  (see  IF/ 
detector  characteristics,  Figure  3).  The  CDI  reading  is  lower  at  12  KHz  because  of  the 
peak  in  the  side  of  the  IF  response.  Likewise  at  15  KHz,  CDI  is  closer  to  its  actual 
value. 


•yrasagzgsrTwi  ' 


G . User  Manual 

1 . FM  Interference  Program  Users  Instructions.  The  FM  interference  program 
uses  the  measured  characteristics  of  the  Nav  11  receiver  to  predict  the  amount  of 
interference  due  to  high  power  FM  stations.  The  program  assumes  sinusoidal  modulation 
on  the  FM  signals  and  that  the  interference  is  due  to  3rd  order  nonlinearities  in  the  RF 
amp  of  the  receiver. 

The  user  must  supply  the  following  information  to  analyze  an  interference 
situation:  Number  of  FM  stations,  receiver  frequency,  distortion  parameter  3/2 
K3/K1  . For  each  FM  station,  the  following  information  is  required:  Station  frequency, 
modulation  frequency  and  modulation  index,  distance  from  the  receiver  and  station  power 
level.  Localizer  distance  from  the  receiver  and  power  level  is  also  needed. 

Most  parameters  for  the  program  are  entered  in  the  edit  mode.  In  this  mode  the 
user  can  enter  parameters  or  modify  them  by  typing  commands  at  a remote  terminal. 

The  user  may  also  exit  the  edit  mode  with  the  ' I M'  command  and  begin  intermodulation 
calculations,  or  begin  signal  level  computations  with  the  'MXDB'  command  or  start 
cross  modulation  calculations  with  the  'XMOD'  command.  "STOP1  will  cause  the 
exiting  of  the  edit  mode  and  normal  program  execution.  These  commands  allow  flex- 
ibility in  altering  parameters  and  seeing  their  effects  on  the  interference  situation. 

A summary  of  the  program  limitations  are  given  below. 

Program  Limitations: 

(1)  A modulation  index  of  75  is  a practical  limit  on  the  maximum  modulation 
index  that  the  program  will  accept.  Modulation  indices  higher  will  cause  extremely  long 
program  execution  times.  A modulation  index  of  75  corresponds  to  a maximum  frequency 
deviation  of  75  KHz  with  a modulation  frequency  of  1 KHz. 

(2)  The  program  assumes  that  the  modulation  frequencies  of  the  FM  stations 
are  not  harmonically  related.  For  example  frequencies  of  1000  Hz  and  400  Hz  are  related 
whereas  1000  Hz  and  398  Hz  are  not.  The  intermodulation  due  to  stations  with  harmonically 
related  modulation  frequencies  has  components  which  occur  only  at  certain  frequencies 
relative  to  the  IM  carrier  and  does  not  exhibit  the  "noise-like"  (realistic)  properties  which 
are  assumed  for  IM  calculations.  Harmonically  related  modulation  frequencies  will  be 
treated  as  though  they  were  not  related  by  the  program.  Therefore,  when  using  harmon- 
ically related  frequencies  such  as  1000  Hz  and  400  Hz,  it  is  advisable  to  check  the  results 
to  see  if  the  same  results  are  obtained  as  with  frequencies  of  1000  Hz  and  398  Hz . 

Before  using  the  program,  it  is  suggested  that  the  user  read  the  program  subroutine 
descriptions  given  in  this  report.  Also,  the  flow  chart  given  in  Figure  G-l  is  very  useful 
in  understanding  the  program.  A sample  terminal  session  is  given  on  the  following  pages. 


-114- 


Enter  Edit  Mode 
I.0.,  Coll  Subroutine 
Change  Input 
Parameter*  Such  As 
Power  Levels, 
Distances,  etc. 


Call  Subroutine 
MU  ED  B 

Calculates  Signal  Level 
Due  to  FM  Stat  at 
RF  Amp  Input 


Cdll  Subroutine  XMOD 
Calculates  " Brute- Force" 
Interference  - Cross-Mod 
and  Cross- Compress! on 


No  - To*ol  Restart 


Calculate^ 
. Intermod? ) 


Call  IMCALC 
Calculate  IM 


Calculate  % RMS  AM 
Modulation  o»  Rereiver 
Filter  Output* 


Calculate  Effects 
of  Both  IM  and 
Brute  Force  Combined 


Restcrt  Mod  if  y Parameters 
Program?  *0^ 


Figure  G-l  . Flow-Chart  of  FM  Interference  Program 


r 


Sample  Session  of  Interference  Program 


load  interfer  ( start  nomap 
EXECUTION  BEGINS  . . . 

START  OF  PROGRAM 

***  FM  INTERFERENCE  PROGRAM  *** 

Program  is  executed  at  a remote  terminal 

THIS  PROGRAM  CALCULATES  THE  INTERFERENCE  DUE  TO  HIGH  POWER  FM  STATIONS 
AND  ITS  EFFECTS  ON  A NAV  11  RECEIVER. 

READ  IN  NUMBER  OF  FM  STATIONS  (FORMAT  NN) 

User  enters  no.  of  interfering  stations 

.02 

TYPE  STATION  FREQUENCIES  IN  MHZ,  MODULATING  FREQUENCY  IN  KHZ, MODULATION 
INDEX  (FORMAT  XXXXXYYYYYZZZZZ) 

The  station  frequency,  modulation  frequency  and  modulation  index  is  inputted  for  each  station 
.105.512.50  5.0 
.102.510.42  5. 

TYPE  RECEIVER  CENTER  FREQUENCY 

.106.5  User  enters  receiver  frequency  which  is  the  same  as  the  localizer  frequency 

TYPE  3/2  K3/K1  IN  DB 

.3.  The  distortion  parameter 

***  PROGRAM  PARAMETER  EDIT  MODE  *** 

FOR  INSTRUCTIONS  TYPE:  INST 

Edit  mode  is  entered.  The  distance  from  station  1 is  set  at  5 miles.  From  station  2 is  4 miles. 

.MILE  01  5. 

.MILE  02  4. 

Power  level  for  both  stations  is  50  Kw.  Localizer  power  is  set  to  40  w. 

.POWER  00  50. 

.POWER  20  . 040 

Distance  to  localizer  is  2 miles. 

.MILE  20  2. 

.TYPE  00 

FM  station  parameters  are  listed 

STATION  STATION  MODULATION  MODULATION  MAX  FREQ  LEVEL  MILES 


NUMBER 

FREQ 

FREQ 

INDEX 

DEVIATION  AT  RF  AMP  FROM 

(MHZ) 

(KHZ) 

(XMOD  CALC)  INPUT=DBM  RECEIVER 

1 

105.5 

12.500 

5.00 

0.0  0.0  5.0 

2 

102.5 

10.420 

5.00 

0.0  0.0  4.0 

-116- 


.TYPE  20 

Localizer  parameters  are  listed 

STATION  STATION  MODULATION  MODULATION  MAX  FREQ  LEVEL  MILES 
NUMBER  FREQ  FREQ  INDEX  DEVIATION  AT  RF  AMP  FROM 

(MHZ)  (KHZ)  (XMOD  CALC)  INPUT=DMB  RECEIVER 

20  108.5  0.0  0.0  0.0  0.0  2.0 

.STOP 

End  of  edit  mode  command 

TO  COMPUTE  SIGNAL  LEVEL  AT  RF  AMP  INPUT  TYPE  1 
OTHERWISE  VALUES  OF  LEVEL  ( I ) MUST  BE  INITIALIZED  IN  THE 
EDIT  MODE  AND  USED  AS  THE  SIGNAL  LEVEL  AT  THE  RF  AMP  INPUT 

.1 

Signal  levels  at  RF  amp  are  calculated. 

PRINTOUT  OF  SIGNAL  LEVEL  COMPUTATIONS 


STATION  MILES  STATION  FREE  NAV  RCVR  INPUT 

FREQ  FROM  POWER  SPACE  ANTENNA  FILTER 

(MHZ)  RECEIVER  IN  DBM  ATTEN-DB  LOSS-DB  ATTEN  DB 


SIGNAL 
LEV  AT 

RFAMP  IN  (DBM) 


105.5  5.0  77.0  90.7  3.0  4.5 

102.5  4.0  77.0  88.6  6.0  9.0 


-15.3 

-20.6 


LOCALIZER  SIGNAL  LEVEL  CALCULATIONS: 

108.5  2.0  46.0  83.0  0.0  0.0  -31.0 

TO  USE  MAXIMUM  FREQUENCY  DEVIATION  SPECIFIED  BY  MODULATION  FREQUENCY 
AND  MODULATION  INDEX  TYPE  1 , OTHERWISE  FQDEV  ( I ) MUST  BE  SPECIFIED 
.1 

Cross  mod  and  cross  compression  calculations  made 
*********  CROSS  MODULATION  CALCULATIONS  ********* 


STATION 
FREQ  (MHZ) 

105.5 

102.5 


COMPRESSION 
OF  LOC.  SIGNAL 
IN  DB 
0.8 
0.2 


FREQ  DEV. 
FOR  XMOD 
CALC.  (KHZ) 
63. 

52. 


% AMMOD 
CAUSED 
BY  STAT. 
0.12 
0.02 


TOTAL  PERCENT  MOD.,  NOT  CONSIDERING  CROSSCOMPRESSION  =0.01 
COMPRESSION  FACTOR  =0.109 
LOCALIZER  SIGNAL  COMPRESSION  IN  DB  = 1 .0 
TOTAL  % RMS  UNFILTERED  AM  MODULATION  = 0.12 


*************************  *******  ******  *****  ************************************* 


Filtered  AM  modulation  due  to  crossmod  is  printed. 
TOTAL  RMS  AM  MODULATION 
DUE  TO  CROSSMOD  INTERFERENCE 
CROSSCOMPRESSION  EFFECTS  INCLUDED 


DETECTOR  OUTPUT  AM  MODUIATION:  0.1  % 
AUDIO  OUTPUT  AM  MODUIATION:  0.0  % 

150  HZ  FILTER  OUTPUT  AM  MODULATION:  0.0  % 
90  HZ  FILTER  OUTPUT  AM  MODUIATION:  0.0  % 

Re-entry  of  edit  mode  by  program 

***  PROGRAM  PARAMETER  EDIT  MODE  * * * 

FOR  INSTRUCTIONS  TYPE:  INST 

. / End  of  edit  mode  - no  parameters  altered. 


Printout  of  intermod  generating  frequencies: 

RECEIVER  FREQUENCY  IS  108.50  RECEIVER  BANDWIDTH  IS  40.0  KHZ 
FREQUENCIES  WHICH  PRODUCE  INTERMOD  INTERFERENCE  ARE: 

F1  F2  F3  IMTYPE  IM  CENTER  FREQ  IM  BW  IN  KHZ  IM  LEVEL 

IN  DBM 

105.5  102.5  2F1-F2  1 08.50  425.  -48.5 

TO  CALCULATE  INTERMOD  SPECTRUM  TYPE  0,  TYPE  1 TO  STOP,  TYPE  2 TO  RESTART 
PROGRAM,  TYPE  3 TOCHANGE  PARAMETERS  AND  RECALCULATE  INTERFERENCE 
.0 

IM  is  calculated 


***************** ********************************************************J 


INTERMOD  AM  MODUIATION  CALCULATIONS: 


CALCULATION  PARAMETERS  : 
INTERMOD  TYPE:  2F1-F2 


STATION  FREQ 
(MHZ) 


MODUIATION 
FREQ  (KHZ) 


MODULATION 

INDEX 


FI:  105.5 

F2 : 102.5 


12.500  5.0 

10.420  5.0 


INTERMOD  LEVEL  (NO  FM  STATION  MODUIATION)  = -48.1  DBM 


DETECTOR  MODULATION  FACTOR  = 0.156 
AUDIO  OUTPUT  MODULATION  FACTOR  = 0.032 
150  HZ  FILTER  MODULATION  FACTOR  = 0.0 
RMS  AM  MODUIATION  AT  DETECTOR  OUTPUT  =2.2% 
RMS  AM  MODUIATION  AT  AUDIO  OUTPUT  = 0.4% 

RMS  AM  MODUIATION  AT  150  HZ  FILTER  OUTPUT  = 0.0% 
RMS  AM  MODULATION  AT  90  HZ  FILTER  OUTPUT  = 0.0  % 


-118- 


FOR  PRINTOUT  OF  TOTAL  INTERFERENCE  TYPE  1 

.0 

Typing  a '1'  will  print  out  total  interference  due  to  intermod,  CDI  interference  and 
combined  crossmod  and  intermod. 

************************************** *********************************** 

TO  STOP  TYPE  0,  TO  MODIFY  PARAMETERS  TYPE  1,  FOR  TOTAL  RESTART  TYPE  2 

.0 


R; 


End  of  program  execution 


H . Subroutine  Descriptions.  In  this  Appendix  we  briefly  describe  each  subroutine 
used  and  its  purpose.  Figure  H-l  shows  a block  diagram  of  the  various  subroutines  and 
functions  used  in  the  interference  program.  Each  of  these  will  be  explained  in  detail 
in  this  report.  Figure  G-2  in  Appendix  G gives  a flow-chart  of  the  main  program. 

Table  H-l  gives  a listing  of  the  variable  names  used  and  each  of  their  meanings. 

1 . Subroutine  CHANGE.  This  subroutine  is  entered  in  the  parameter  edit 
mode.  The  program  parameters  are  changed  by  typing  the  proper  instruction.  The  instruc- 
tions can  also  cause  the  program  to  exit  edit  mode  and  continue  normal  execution  or  jump 
and  perform  certain  tasks  such  as  calculate  IM  immediately,  calculate  signal  levels  at 
RF  amp  input,  etc.  A list  of  instructions  and  what  action  they  take  is  given  in  Table  H-2. 


Figure  H-l  . Subroutine  ana  Functions  Called  by  the  Main  Program. 


-120- 


VARIABLES: 

NSTAT  IS  THE  NUMBER  OF  FM  INTERFERING  STATIONS 

FREQ  (I)  IS  THE  FREQUENCY  IN  MHZ  OF  THE  ITH  FM  STATION 

FMOD  (I)  IS  THE  MODULATING  FREQUENCY  FOR  STATION  I 

BETA  (I)  IS  IT'S  MODULATION  INDEX 

LEVEL  (I)  IS  THE  SIGNAL  LEVEL  OF  THE  I TH  FM  SIGNAL. 

IT  IS  THE  SIGNAL  LEVEL  AT  THE  RF  AMP  INPUT  I.E. 

IT  IS  THE  LEVEL  AT  THE  RECEIVER  INPUT  TERMINAL  MINUS 
THE  ATTENUATION  OF  THE  INPUT  FILTER 

MILE  (I)  IS  THE  DISTANCE  FROM  THE  RECEIVER  TO  THE  ITH  STATION  IN  MILES 
POWER  (I)  IS  POWER  LEVEL  OF  THE  ITH  STATION  IN  KWATTS 

FQRCVR  IS  THE  CENTER  FREQUENCY  THAT  THE  RECEIVER  IS  TUNED  TO,  GIVEN  IN  MHZ 

BW  IS  THE  BANDWITH  OF  THE  RECEIVER  IN  KHZ.  I.E.  THE  RECEIVER  BANDWIDTH. 

IT  IS  ASSUMED  TO  BE  IDEAL  BANDPASS 
IT  IS  GIVEN  THE  VALUE  OF  40  KHZ 

RK3K1  IS  THE  DISTORTION  PARAMETER  3 K3/2  K1  IN  DB. 


Table  H-l 


Variable  Names  and  Their  Meaning 


IN  PARAMETER  EDIT  MODE  ALL  COMMANDS  ARE  TYPED  AS: 


to 

<r  to 

! § 


uj  £ 2 

x < O 

•-  £ <J 

- 5 3 

T LL-  CO 

<J  X Co 

§<  o 

> LU  Cl. 


Hz 

o o Z 

I—  *“  • 

Z Z H- 

UO 


oo  < £*= 

oo  z 
o o > 

C£  _ LU 

u <^>  u 


Z X LU 

UJ  UJ  X 

0^5 

lu  i 

a:  7 O 
n “ 
z2“- 


3 O £ 

n^Q 


Q-  LU 
1 1 1 AT 

m 3 8 

t!2t 

0- 2 

o;  _ 
00  a.  ^ 
• — , O 

z ' z 

5|u 

Q-  Z • 

1—  UJ  9: 
10  Q.  O 

— 00 


Q LU  LU 

Z o o 

< z z 

x < < 
U x x 
, <J  U 

>-  ' 1 

Z Z z 
Qzz 

0<^ 
2 LU  ^ 

u.  m i 


3a:- 
UJ  O > 

£°-S 

~ 1 z 

fz^ 

O O 3 
00  •—  o 

<T  LU 

^a: 
X I/O  LL. 

i/o  1/0 

1/0  LU  LU 

Iss 

x u u 

(J 

I I 


>^Q 

iOO 


IM  - INTERMOD  INTERFERENCE  IS  CALCULATED 

XMOD  - CALCULATIONS  OF  CROSSMOD  INTERFERENCE  ARE  MADE 


OdB  4- 


dB  Attenuation 


-8  MHz 


Slope  1.5  dB/MHz 

/ linear  slope  = 6 - (.1  726) 
for  crossmod 
calculations 


-.075-  |f| 

10 


+8  MHz 


Slope  .75  dB/MHz 
linear  slope  = 8 = .04326 
for  crossmod  . iq-.0375  -|f | 
u!ni  lotions 

.f 


Frequency  away  from 

receiver  center  frequency 
Receiver  center  frequency  in  MHz 


Figure  H-2.  Assumed  Attenuation  Characteristics  . 

2.  Subroutine  MILEDB.  MILEDB  calculates  the  signal  level  at  the  RF 
amp  input  of  the  receiver  using  the  free  space  attenuation  formula.  It  takes  into 
account  NAV  antenna  loss  (1  dB/MHz  below  1 08. 5 MHz)  and  the  RF  amp  input 
filter  characteristics.  The  free  space  attenuation  formula  is 

a (attenuation  in  dB)  =36.3  + 20  log  (freq.  in  MHz)  +20  log  (d  in  miles) 

The  signal  level  in  dBm  at  the  RF  amp  input  due  to  station  I: 

Level  (I)  = Power  (I)  in  dBm  - Free  Space  Atten.  (I)  - NAV  Antenna  Loss 
- Loss  Due  to  RF  Amp  Input  Filter  + 6dB 

(The  6dB  is  due  to  the  directive  gain  of  the  NAV  11) 

The  loss  due  to  the  RF  amp  input  filter  characteristics  is  calculated  by 
REAL  FUNCTION  RF  FILTER.  The  attenuation  characteristics  of  the  NAV  11 
receiver  are  approximated  by  the  characteristics  given  in  Figure  H-2. 

3.  Subroutine  TEST.  This  subroutine  checks  to  see  if  there  is  3rd  order 
IM  due  to  stations  I,  J,  K which  lands  within  the  receiver  passband.  The  program 
uses  the  relation  given  below  to  calculate  the  bandwidth  of  FM  signal  spectra  of 
stations  I,  J,  K . 


-123- 


Bandwidth  2 (R  • 1)  fm 


\ 


where  R - modulation  index  and  fm  - modulation  frequency. 

Any  intermod  component  formed  from  a spectral  component  from  each  FM  station 
is  given  by: 


IM  component  = FREQ  of  component  from  I + FREQ  of  component  from  J - 
FREQ  of  component  from  K 

To  see  if  any  of  these  IMs  are  within  the  receiver  BW  the  lowest  and  highest 
frequency  IM  component  are  calculated. 


IM  = FI  . , t-  fm.  (1  t p ) + F2  . , +fm  (1+P) 

max  carrier  freq  1 1 carrier  freq  2 2 

- F3  + fm3  ( 1 + p3)  ( H-l) 

Likewise, 

IM  . - FI  . , - fm,  (1  + p.)  + F2  . , -fm  (1+pJ 

min  carrier  freq  1 1 carrier  freq  2 2 


- F3  - fm^  ( 1 + R^) 


(H-2) 


The  receiver  frequency  is  denoted  by  FQRCVR  in  the  program  and  if  there  is  IM 
interference,  then 


IM 


freq  min  — 


FQRCVR  ±20  KHz  < IM 


— Freq  Max 


For  these  calculations,  the  bandwidth  of  the  receiver  is  assumed  to  be  40  KHz. 

4.  Subroutine  XMOD.  XMOD  calculates  the  amount  of  cross-compression 
and  cross-modulation  due  to  interfering  FM  stations.  The  program  uses  either  the  maximum 
frequency  deviation  calculated  from  BETA  (I)  and  the  modulation  frequency  FMOD  (I)  or 
FQDEV  (I)  (the  maximum  frequency  deviation  is  entered  in  edit  mode).  The  user  can 
specify  which  to  use.  Entering  the  maximum  frequency  deviation  for  crossmod  calculations 
allows  the  user  to  analyze  interference  situations  easier  without  specifying  a BETA  or 
FMOD  for  each  FM  station.  FQDEV  =40  KHz  might  be  a "typical"  value  for  example. 


The  gain  change  (GC)  of  the  desired  signal  due  to  station  I is  calculated. 


GC  = 


Desired  signal  amplitude  with  interference 
Desired  signal  amplitude  without  interference 


= 1 - 3K./K  Bj 

J ( H-3  ) 


where  B is  the  signal  level  of  the  Ith  FM  station  at  the  RF  amp  input.  At  this  point  it 
is  assumed  that  the  RF  amp  input  filter  attenuation  has  already  been  taken  into  account. 


-124- 


( H-4  ) 


The  gain  change  in  dB  is: 

GCdB  = 20  log  f 1 - ('10rr3K3//2l<l )dB  + 2BdBm  + 6)/2° , 
Total  GC  due  to  all  stations  is  given  by 


3K„  N 

GSo,al=  ,r=1 


Bi 


.2 


(H-5) 


where  Bi  is  the  signal  level  due  to  the  Ith  station  at  the  RF  amp  input. 
The  amount  of  cross-modulation  due  to  station  I is: 


% Am  mod  - Max  Freq  Dev  of  FM  Station  • Slope  of  Input  RF  Amp  Filter 

•6K,j/Kj  • B?  x 100%  ( H-6  ) 

(not  considering  cross-compression) 


Equation  (FI. 6)  does  not  consider  cross-compression  of  the  carrier.  If  one  wants 
to  consider  cross-compression,  the  equation  used  is 

% Ammod  without 
% Am  Mod  with  _ Cross-Compressi on 
compress?  on  3 K~ 


1 - 


K 


B 


The  total  RMS  Am  cross-modulation  due  to  N stations  is  given  by: 


N 

Total  RMS  % Am  = [ .7  (f  max.  • 5- 

i-l  i i 


6K 


3 _ 2 2 1 


B.  ) ] 


K, 


I B. 


( H-7) 


( H-8  ) 


where  f max;  - maximum  frequency  deviation  of  station  I 

5.  = slope  of  input  filter  characteristics  at  frequency  of  station  I 


Subroutine  XMOD  calculates  % Am  mod  at  the  receiver  filter  outputs  also.  It 
uses  the  same  equation  as  above  except  that  each  term  is  weighted  by  y (f  mod;  ),  where 
y is  the  appropriate  normalized  filter  function  and  f mod;  is  the  modulation  frequency 
of  the  Ith  interfering  FM  station.  Note  that  if  FQDEV  (I)  is  only  specified,  the  filtered 
% cross-mod  calculations  will  not  be  accurate  since  f mod;  must  be  specified.  The  % 
modulation  present  at  the  receiver  filter  output  due  to  interfering  signals  is: 


-125- 


•rns".  ssr 


Total  RMS  % Am  Modulation  at 
a Receiver  Filter  Output 


N 6K3  2 2 

f I (f  max.  ‘ 8.  • -p—  B.  • y (f  mod.)  ) ) 
i=l  ' ' K1  ' 


1 - 


577 

Ki 


( H-9) 


Table  H-3  gives  a listing  of  the  variable  names  used  in  Subroutine  XMOD. 


. 


VARIABLE  NAMES: 

CROSSL-LINEAR  COMPRESSION  FACTOR  FOR  ONE  STATION  ONLY  (ITH  STATION) 
I.E.3  K3/K1  B**2  WHERE  B IS  THE  SIGNAL  LEVEL  OF  THE  ITH 
STATION  AT  THE  RF  AMP  INPUT. 

PERCNT  - % CROSSMOD  OF  LOCALIZER  SIGNAL  DUE  TO  STATION  I,  CROSS 
COMPRESSION  NOT  CONSIDERED 

PERCOM  - % AM  MOD  DUE  TO  STATION  I CONSIDERING  CROSSCOMPRESSION 
SUMC  - COMPRESSON  FACTOR  DUE  TO  ALL  STATIONS.  (RESULTANT  LOC.  LEVEL 
IS  MULTIPLIED  BY  (1-SUMC) 

SUMCDB  - COMPRESSION  OF  LOC.  SIGNAL  DUE  TO  ALL  FM  STAT.  EXPRESSED  IN 
DB.I.E.  SUMCDB=20  LOG  (1-SUMC) 

XMOD1  - TOTAL  % RMS  AM  MOD  AT  THE  DETECTOR  OUTPUT  DUE  TO 
CROSSMOD  FROM  ALL  STAT.  (IT  INCLUDES  EFFECTS  OF 
CROSSCOMPRESSION  ONLY  IF  SUMC  IS  LESS  THAN  1) 

XMOD2  - % RMS  AM  MOD  AT  THE  RECEIVER  AUDIO  OUTPUT 
XMOD3  - % RMS  AM  MOD  AT  150  HZ  FILTER  OUTPUT  OF  RECEIVER 
XMOD4  - % RMS  AM  MOD  AT  90  HZ  FILTER  OUT 


Table  H-3.  Variable  Names  For  Subroutine  XMOD. 

5.  Subroutine  IMCALC.  This  subroutine  calculates  the  AM  modulation  due  to 
intermod  interference. 

IMCALC  checks  all  combinations  of  3 stations  (stations  I,  J,  K)  to  see  if  the  3rd 
order  intermod  causes  interference  at  the  receiver  frequency.  This  is  done  by  calling 
Subroutine  TEST  which  checks  the  3 freqs  and  returns  with  ITEST  = 1 if  they  produce 
interfering  IM.  Subroutine  CONVOL  is  called  which  does  the  convolution  of  the 
3 FM  spectra  and  returns  with  the  modulation  factors  of  the  receiver  filter  outputs  (called 
SUMDET  , SUMAUD,  SUM150,  SUM90).  IMCALC  uses  the  modulation  factors  to  calculate 
the  % RMs  AM  modulation  at  the  filter  outputs. ( called  PER  1 , PER  2,  PER  3,  PER  4) 

Total  % AM  modulation  due  to  the  IM  from  all  interfering  stations  is  summed  (called 
TOTAL  I,  TOTAL  2,  etc.).  The  CDI  with  interference  is  calculated  by  calling  subroutine 


-126- 


CDI.  The  above  calculations  do  not  consider  effects  of  brute  force  interference.  The 
combined  effects  of  cross-mod,  cross  compression,  and  intermod  are  added  and  printed 
out.  CDI  calculations  are  again  made. 

A listing  of  the  variable  names  used  in  the  IM  % AM  modulation  are: 


TOTAL  1 is  the  total  % RMS  AM  mod  due  to  all  FM  stations  at  the  output  of  the 
detector  fi  Iter 

TOTAL  2 is  % at  the  audio  output 
TOTAL  3 is  the  % at  the  150  Hz  output 
TOTAL  4 is  the  % at  the  90  Hz  filter  output. 

PER  1 is  the  % AM  mod  due  to  IM  from  stations  I,  J,  K only  at  the  detector 
fi  iter  output. 

PER  2 is  the  % AM  mod  due  to  stat.  I,  J,  K at  the  audio  output. 

PER  3 is  the  % AM  mod  at  the  150  Hz  filter  output. 

PER  4 is  the  % AM  mod  at  the  90  Hz  output. 


The  IM  level  due  to  2F1-F2  IM  is  given  by 


IM  . . 

carrier  leve 


Ko  2 

3/2  -j-r-  IT  C 


( H-10  ) 


where  B and  C are  signal  levels  at  the  RF  amp  input  of  FM  station  1 and  FM 
station  2.  The  IM  level  due  to  FI  + F2  - F3  IM  type  is: 

K3 

IM  = 3 -±  BCD  ( H-l  1 ) 


where  B,  C,  D are  signal  levels  of  stations  1,  2,  and  3 respectively. 


The  % Am  modulation  due  to  intermod  carrier  only  is  (no  FM  modulation  on 
the  FM  stations ): 


% Am  mod  due 
to  IM  carrier 


- (Localizer 
level  in 
1 0 dBm 


IM  carrier  ) / 20 
level  in 
dBM 


100°i  H_12  ^ 


The  RMS  AM  Modulation  due  to  the  IM  with  modulation  is 


% RMS  Am  modulation  =%  Am  modulation  due 

to  carrier  alone 


Modulation  factor 

( H— 1 3 ) 


-127- 


Total  RMS  % modulation  due  to  all  IM  is  given  by 


2 1 2 

Total  RMS  % mod  = ( I % Am  mod  due  to  ) 

all  IM  from  3 stations 
intermods  ( H-14  ) 


Including  effects  of  cross-compression  and  cross  modulation  as  well  as  IM 
results  in  a total  RMS  percent  modulation  given  by. 


Total  RMS  % mod  due 
to  IM  and  brute  force 
with  cross  compression 


2 2 

(total  RMS  % Am  mod  due  to  IM  + Total  Am  mod 
due  to  brute  force,  AM  mod  from  cross-modulation 
without  cross-compression)  1/2 


where  GC  — 1 - 3 


B.2 

I 


GC 


( H-15  ) 
( H-16  ) 


which  is  the  gain  change  due  to  "brute  force"  interference  due  to  all  FM  stations. 


6.  Subroutine  CONVOL.  Subroutine  CONVOL  calculates  the  3rd  order 
intermod  due  to  interfering  tone  modulated  FM  stations.  There  are  two  types  of  intermod 
possible,  2F1  - F2  type  and  FI  + F2  - F3 . 

The  2F1  - F2  intermod  is  due  to  2 FM  stations  with  an  IM  center  frequency  = 

2 FI  - F2,  where  FI  is  the  carrier  frequency  of  the  one  FM  station,  F2  is  the  carrier  level 
of  the  other . 


FI  + F2  - F3  intermod  is  due  to  3 FM  stations  and  has  a carrier  frequency 
FI  + F2  - F3.  The  subroutine  CONVOL  handles  each  type  separately. 


a.  2F  1 - F2  IM  Term.  For2Fl  - FI  intermod,  the  IM  is 

Jm  (2  f2)Jk  °V  ( H"17) 


2 P2+  1 


P1  + 1 

!M  _ 3 .2  _ . .2  ' 

. rm<:  - B c c<b)  a (c)  I I 

f 2 k = -(pj+1)  m=-2  P2+l 

■ cos  2irt  (2b  - c + m -frr^  - k*fm^) 

Thus  the  ( m,  k)  th  intermod  term  is 


IM 


m,k 


= (Pj)  J (2p2)  cos  [ 2b  - c + mfm2~  kfrr^  ] (H  -18  ) 


Am,  n 


L 


-128- 


3 


The  modulation  factor  is  calculated  by 

MF 


( I Am,  l<  *|  y (01  ) 

all  terms  with 
frequencies  wi  thin 
receiver  passband 


1/2 


( H-19  ) 


where  f =IM  carrier  frequency  - Receiver  frequency  + m -fm^  - k-fm^  . 

In  this  equati on,y(f)  is  the  normalized  filter  function  of  one  of  the  receiver  filter 
outputs  and  f is  the  frequency  of  the  IM  term  relative  to  the  receiver  center  frequency, 
y (f)  is  assumed  symmetrical  with  respect  to  frequency.  The  maximum  value  of  the 
modulation  factor  is  1 and  this  occurs  only  if  all  IM  terms  are  received  in  the  receiver 
passband  and  y(f)  = 1 . 

b.  Fl  + F2  - F3  IM  Term  . The  FI  + F2  - F3  intermod  term  is  given  by: 

IM  = 3 B C D a(b)  a(c)  a (d)  I I l J.  (B,)  J.  (P2)  Jk  (P3)  cos  2nt 

terms  ' i k 

( H-20  ) 

( FI  + F2  - F3  + ifm1  + jfm2  - kfm3  ) 

For  convenience  let 

A i j k = J.  (P,)  J.  (P2)  Jk  %) 

Using  (H-21)  the  modulation  factor  (MF)  is  given  by 


( H-21  ) 


where 


Modulation  = ( I 

factor  all  IM  terms 

that  pass  through 
receiver  bandwidth 

f = FI  + F2  - F3  + ifrr^  + jfm2  - kfm3 
- Receiver  center  frequency 


A i i kf  y (f)l  ) 


1/2 


( H-22  ) 


( H-23  ) 


The  program  stores  in  Array  A X B and  array  A X BFQ  the  intermediate  result 
of  convolving  two  stations  together.  In  addition,  A X B holds  the  amplitude  and  A X BFQ 
contains  the  relative  frequency  of  the  intermediate  result. 


For  a two-station  case  the  intermod  is 

A X B (I)  = JUM  (2  P}) 


( H-24  ) 


-129- 


and 

AX  BFQ  (I)  ( L - I - 1 ) ‘ fm1 

where  I ranges  from  1 fo  2L  - 1 
and  L = integer  value  of  I (2 1 

Si  mi  lari  ly  for  a three-station  case  the  IM  is 

A X B (I)  = JL  ( B1  ) Jn  ( P2) 

A X BFQ  (I)  = L • fm  + m • fm2 


H-23 


( H-26  ) 
( H-27  ) 


where  A X B and  A X BFQ  have  been  sorted  in  order  of  ascending  frequency  and 
. and  M are  indices  which  range  from 


INT  [ - (P1  + 1 ) ] L ' INT  C P,  f 1] 
INT  [ - (p2  H ) ] < M < INT  rp2+  1] 

( Note  - INT  stands  for  take  integer  value) 


( H-28  ) 
( H-29  ) 


c Convolution  of  AXB  with  the  Spectrum  of  Station  K.  A flow  chart  of 
the  program  part  of  subroutine  tU  convolves  the  intermediate  resuTTAXB  w.th  spectrum 
of  F3  (or  F2  Fn  the  case  of  2F2  - F2  IM),  is  given  in  Figure  H-3.  To  save  computation 
Hme  in  calculating  which  intermodulation  components  fall  into  the  rece.ver  passband 
the  program  uses  the  fact  that  AXB  is  ordered  as  to  increasing  frequency.  The  mtermod- 

ulation  frequency  is  given  by: 


IM  component  - IM  carrier  t- 
frequency  frequency 


A X BFQ  (I)  - K*  fm3 


( H-30  ) 


The  program  starts  with  K = 1 and  begins  to  search  through  the  array  AX  BFQ  for 
on  intermodulation  component  that  lands  just  inside  the  low  end  of  the  rece.ver  possband. 
, , ” , , the  Val„e  | is  stored  so  that  when  K is  incremented  after  going  through 

re\"nM,e  X « BFQ  a sl^ch  starts  at  that  value  of  I.  colled  UST  in  the  pr^ram, 
ft  the  next  va)ue  of  IM  which  just  falls  within  the  receiver  possband . Note  that  w en 
the  IM  components  begin  to  fall  outside  of  the  high  end  of  the  rece.ver  possbund,  £«  ° 
need,  no  longer  to  be  seorched  since  oil  other  components  will  pcocktM  IM  I arms h g 
in  frequency.  So,  K is  then  incremented  and  search  begins  through  AX  BFQ  again. 

This  program  calls  subroutine  SORT  which  is  o bubble  sort  program  to  , >ut  the 
elements  of  AX  B and  AX  BFQ  in  order.  It  is  colled  only  for  FI  + F2  - IM- 


-130- 


Real  Functions:  FILDET,  FILAUD,  FIL90,  FI  LI  50  are  straight  line  approximations 
of  the  measured  filter  outputs  of  the  receiver.  They  are  normalized  and  are  symmetrical 
with  respect  to  frequency  above  and  below  the  receiver  frequency.  For  IM  calculations 
real  functions  EQ90  and  EQ150  are  used  which  are  bandpass  filters  with  the  same  noise 
equivalent  bandwidth  as  the  90  and  150  Hz  filter  functions. 

Subroutine  CONVOL  uses  Real  Function  BESSEL  to  calculate  the  bessel  function 
of  a given  argument  and  integer  order.  Real  Function  BESSEL  allows  negative  orders 
to  be  used  and  calls  the  IBM  Scientific  Subroutine  Package  (SSP).  Subroutine  BESJ  is 
used  to  calculate  the  actual  bessel  function. 

7.  Subroutine  CDI  . This  subroutine  calculates  the  CDI  with  interference 
present.  The  % modulation  at  the  90  and  150  Hz  filter  outputs  are  inputs  to  the  program. 

The  subroutine  first  calculates  DDM  (difference  in  the  depth  of  modulation)  for 
a CDI  reading  without  interference. 

DDM  = CDI  • 15.5/150%  ( H-31  ) 

Let  % modulation  90  denote  the  % modulation  at  the  90  Hz  filter  due  to  localizer 
signal  with  no  interference  present,  i.e., 

% mod  90  = 20%  + DDM/2  ( H-32  ) 

Likewise,  for  the  150  Hz  filter  output, 

% mod  150  = 20%  - DDM/2  ( H-33  ) 

If  interference  is  added  with  a certain  percentage  modulation  the  resultant 
modulation  is  given  by: 

% mod  90  new  = (%  mod  90^  + % mod  interference  90^  ( H-34  ) 

Si  mi  lari  ly, 

% mod  150  = (%  mod  150^  + % modjnterference^ 

new  at  150  Hz  out)  ' ( H-35  ) 

The  resulting  CDI  response  is  given  by: 

CDI.  = (%  mod  90  - % mod  150  ) • 150/15.5  ( H-36  ) 

NEW  new  new 

In  this  section  the  basic  operation  of  the  computer  program  written  for  considering 
cases  of  multiple  interfering  signals  was  described.  Complete  listing  of  these  programs 
is  given  in  Appendix  I. 


-132- 


THIS  PAGE  IS  BEST  QUALITY  PRACTICABiifi 
FROM  COPY  EURlilSHEX!  TO  DOC  


Computer  Program  Listinc 


M I T FRF  ERENC  V POnG««M 


THIS  BOGrAVI  C«LCIJL*  Tr~  T H „ I N T ~ . F ■-  =>  r rjr  . DU  - v.Ul  T T • >L  r 

FM  SIGNALS  AND  ITS  -FFI'T'  TH  <*  MAV  1!  Fi'TIVF.-. 

THF  INTEPF^t?CMCF  IS  ">IJ-  'D  ■»“  R •?r-l  MDML  I Nr~  A ■"  T t y j-j  yn* 

NAV  RSCEIVFR 

THF  ORT3RAM  CALCULATES  = e~T£  anD  C'-CSS  1.  D I L — C’ c t:  ‘ M C : 

DUE  TO  TONE  M O D J L A T 60  FM  ST  A- I DUS 


written  ty  -r.  loos 

a VI  TNI  CS  ^NCINFtr.  IHS  C .-:T 
D-JIP  UNIV.,  ATH"N=  ~M. 
JAN*  I 9 7 S 


(•*«*<:*  <.-**ti***«****t  * * F * < 


**«**«-***«?. 


C**$*  *****  ***< 

C*  **  * ***  **  ***  ■ 

c 


:********<******************************** 


DC  CG  •’  AM  V A * I A 1 J . 


NS  T A T is  th:  NT-*  I' 


•A  f-T  --•Ft  ■>  in 


0(1)  IS  T ! I “ F -T  >“»! -V  ! I • H 7 -h  TF-*  I"H  F " ' T * T I " . 


-M"r  ,r,  IS 


A'UUL  ATI!"  F !U-'‘ifY  ci;  ^ t 4 t i nj  j 


D " A ( I ) IS  I 


F>UI.  A T I f N I NO'  X 


Li  V = L ( : ) IS  THT  cIGnAL  L V“L  IN  rrr  ~F  Th'  I T J * c"r  A T : . ;t  * 
~F  A «o  IND'JT  I.f.  L'V'L  AT  THF  -t.fr-  I Vs-  A I NPl.iT  ' ! J,  I * 

T|-  STTf'N'.)4T  I Y*  D'l--  - i nf  .-Mr  ]ni>jt  f 1 L T 7 T , 

V!ur(  I)  IS  Th:  DICTa  NC  - th--  .'‘hf  IVfD  T)  TH:  [ TH  STA”I',,J  It 

VI  Ln  c . 


*■’  lw  F r.  ( i ) 1 


LcVtL  F Tpr  J Tti  : T T I ' N IN  K •>  t ~ T ■ 


F'JFCV'.  IS  THL  --C'lV  ■ C'NT;- 


OU'NCY  IN  "'V. 


’*  IS  THe  f’ANJv*ITH  ’f  -H- 


I v - ; 1 n Kt-1 7 , 1 . - . r n‘  r-  C-  I v 


£?■  SSSSSS^ST^11* 


c 

c 

c 

c 

c 

c 


c 

c 


HANDWIOTH.  IT  IS  ASSUNFD  TO  R r IDFAl  SANTfUS^ 
IT  IS  GIVFN  THC  VALUE  T A3  KHZ. 

^K’Kl  IS  THE  D I S T 39  T I ON  PAPAMFTFP  ">  K3/2  K|  IN  DH  . 


' ** 

■A****************** 


n I MENS  I CM  F REQ  (501  . r«nO(  5 3 ) . FSET  A ( *»  0 > 

pp  AL  *ILE{  23  ) . P.Twf.'A  { 33  ) . FOOEVF  ? 1 ) 

PEAL  LEV  EL (23) 


c »VM?N  PQ^CV  •)  . X‘A"m  , XV  ir>  , xwcr  ’.XM^TL  suvr.  ,0  v , -K  -•«  : 
CTX'MC’J  FREQ, LEVEL.  FMDP  .BETA 
COMMQm  mi l= , PO wE  P . FQOE  V , N ST  A * 

C....  P r A 0 IN  OF  PP'Or.PAV  .... 

C 

PRINT  R 


a 

P~PV«AT(  //,  • tu  F M IN-p: 

■ 

E NCF 

iprGR  a x 

* . //. 

c 

* THIS  PPOGPAv 

CftLCUL  A TP  3 

T Hi*' 

PIT- 

^-N[  N c r 

DJ 

r to  mi  S'4  p 

c 

T A T I 3 NL  • ./  . • AN 

0 ITS  “pPP. 

:ts 

M A 

N *V  1 l R 

- c - 

1 V . 1 , //  ) 

i 

P=INT  3 

p 

F .3  R M A T ( • P P A 0 IN 

> IJM1SP  IF 

"TATI 

T*  S (p  -,J 

A r 

N Ml  • I 

ci}-'!  FJ 


RcAO  1 3 . N S T A T 
10  FQ»VAT(I?) 

PRINT  11 

1 : «-PRM  A”(  • Ty  r-  £ 5t*tI  ON  r"r  1 |PN  I IN  UI-.Z,  M lOJLA  T I NO  s>‘.  ) lr  N *V 

C'IN  KHZ. WPOJLA'I  If.  INP-X*,/,*  (phc.  \1A^  X X X X X Y V Y V V ? Z ? 7 Z ) • ) 

00  12  1=1  . NS  TAT 

p £ A D 1 ii(  I ) .P  < ■ 3 ( I ) , J.  ~ a ( I ) 

12  C'R  T !’'Jr 

1 x = .1  r<  H .-.  = ( F " . 1 , t F - . ) 

PR  I \ T 1 A 

1 A F1PHTO  TY3.  C-  IV"C  .T  - TNCY*  > 

a o i ;,F(F£v' 

1 e n‘.YAT(F13,A) 

RP£0(  . ))  =F  3P ' VP 
P:-’  I N r ISA 

h;  H»MAT(I  Ty,i:*  >/  ' K -»/K  1 IN  P->  •) 

■-~A~)  IS9,PK.3<1 

IE’  ■"  ’ ;m  JT(  . , ) 

L > r X =1 

£ "r  T T t ’ In  R;.P£IVP"  UN"HpTH  I'  x-(Z 

P.V  = A £ . 
r 

o. . . . I.  - ~n  I ” .... 

»qpp  r all  £HA».Gr<  J J 'P  ) 

T-  | A 33  J . A 3 JO  I a 001  . A oo  ~ ) . JU  -,r> 
r 

*....  CALCULATING  .IUN-1  I.  V*L'  &T  -IF  AVI’  UHMT  .... 


■ 134- 


*HIS  PACE  IS  BEST  QUALITY  PRACTICABLE 
*ROK  COPY  PUKNISHED  TO  DOC 


4030  CALL  MLfc'rm 

c 

C....  CALCULATING  'STUTC  F<3?Cr  I NT  fp'  = i ~ NC  € .... 

c 

4001  CALL  X"JD 
C 

C....  FNTSPING  LDIT  M'3'.Tf  .... 

C 

CAL  L CHANGE  ( JU'^PI 

GO  TO  1 4003. 4000 ,4031  . A 33  2 > . J' > wp 
4003  HNTnur 

c 

c 

c 

C....  PRINTOUT  3F  I NTF  CMn0  r,  ■ M • FkFOUFNCI-?  .... 

c 
c 
c 
c 
c 
c 


c 

c 

c 

c 

c 

c 

c 

c 

c 


POINTING  ->F  HFADING' 


CALL  -f 3T( 0.  . 3 
C 3K  1 ) 


3 . • 3 • 


,o.ij..Fj-cv-,L*,r 


TF  ST  1=  DF"='5"-3  T n S I = JvTs;vnO  SPFCTP'.JM  P-’oruC- O 

er  thnff  fptoufncifs  will  ;tv-  a c '>*jt  i juti  cn  at  th- 
F'FCzIV~3  =P=0Ur.NCY 


DA 

5?  3 

1 = 1 

. » 3 T A T 

0 3 

r 2 l 

J=  I 

. N tT  A T 

DO 

S2  ? 

<=  1 

•nftat 

1 


c 

c 

c 

c 


SHOP  OUT  I NS  Tf  37  Pl  I NTS  3U  7 F“  ~ y J F\>  C I FS  wHL 
■A  I THIN  HZ  3tCSI  Vr  F AW 


C 

c 

c 

call  tt  <jt(  r j~.3(  I ) . F - “Of  J)  , = r,-  )(  y ) , f • «.JD  < 2 ) . F^  V-,  ( J > , <1  , 

C TFT  A < I t ,f<-.TA(  J ) , T A I < ) ,L  r V _ ( I > iL’V’Ll  J)  , L _ V * L ( < ) 3 ACV - . n* . 

C I T* ST , 1 , PK3K 1 ) 

C 

C 

C 

5?  ? CTiT  IN  )" 

5?  1 C ’NT  I N »- 

S?0  CV.TINIJ; 


-135- 


* 


UHIS  PAGE  IS  BEST  QUALITY  PRACT1 CALLS 
PKOM  COPY  FUKlilSHEL  TO  LUC ^ ' 


C 

4002 

577 


C 

c 

c • • . 

1 Oc 
1 Ol'i 

1 1 00 


1 l 0 


COnt  I NUT 
PL  I NT  6 "*7 

FO=:VAT  ( ///•  TO  CALCULATE  !NT-‘3MV  Q TyDC  J 

CP.  TY°-  2 Tj  jRFSTAO’-  PPi:,f<4  M' , / , > TyPC  ■*  T D Ch4NV 
C ®EC  AL CU  L At  £ I NT  EPFrPfNCF ' ) 

’EAD  17,  ICALO 
FORMAT (111 

IF( I CALC.^Q. T ) GO  TO  TQoq 
IF ( I CALC-1  ) 1 OP. 1 1 0, 1 


CALCULATION  OF  INTCRMPO  .... 

CALL  IMCALC 
PPINT  1000 

FOPVAT(//,ix,-’-’(  ),//) 

OPIN’"  1 t 00 

FQCVIAT(«  to  STOP  TYnF  0.  T n ‘‘flOIPy  CABiv'T' 
C ‘TOTAL  RESTART  TYP-  ? •) 

READ  17,  ICALC 
I F ( ICALC-1 )1 10.3P9O, 1 

STOP 

END 


. TYPE. 

BinA'C 


T - 5 - -j 

">e  AND 


T Y3;.  l.FpP 


C = = 


SUB5  C UT  I NS  T5ST(  -P“D‘  .^DE>2  .“O-IQ'.F  MODI  .fyqo2,  3ETA1 . 3FTA2. 

C npT4  7.R.EV1  , CL  - VS  » "L’  V 7,  r JRCVP  .R  I TF  jt  . IPr-  I JT,r<3Ki) 

C 

THIS  SUOOPUTIfir  T £ c T E'  FM  fts/jc  : F~  r Jl  ,FPFO?  , Fo:  U>  ” o 3?- 
Ic  TOT  F1+F2-F?  I A*  IS  v I T MI  • i the  3ECFIVFS  PW  wJTh  PECEIV'T’ 
FPrOlFQPCVP.  IF  T H r-  •»'  ” IS  , I T.S  T S-  T to  1 , r THr  o 'M  I S c S ,‘T  tj  j 
I PO  I NT  = 2 .POINTS  7|jt  M^ADINGS 


c 

c 

c 

c 

c 

c 


c 

C 


C 

c 

C 

c 

c 

1 

60  0 


60  1 


I T ES  t = j 

? U *-*L  K V IS  THE  NO  MC  •VILA”  I CM  ! N T T •- v 00  LlV  "L  iE  D" 
FPfQl  IS  NOT  T Hz.  SAwr  A “3  Ft-JS 
SU‘-1Lr-  V-»Lr  VI  ♦ «L f V2  ♦ -’L  r-  V 3+  5 K ‘ K 1 
I f < rc  --  01  • N"  , F£  “Q  * ) S 1 IML  " V=SU  ML~  V . 

I f ( I AD  I N T — 1 ) ? ,2 , 1 


A..'):.D  ir 


OP  ! NT I No 


HEATING’  *iF*|  I P":  I NT  =2 


I NT-.  )0  . foSCYo.hw 

f-ipvit  (////,«  .‘.Fti-IV-.fi  F ‘ l.lfM’Y  I*-  '.o  .-1 

C IE.  * , F 6 . 1 , ‘ K -H  7 • > 

P«  I NT  e ) 1 

n^'tf  (•  - : _ 0 J~NC  I rr  yHI’H  ujTDt'O;  i * j T E S m * p IYT 


V :W‘-  » Da  1 r>T  i 


A 


-136- 


THIS  PAGE  IS  BEST  QUALITY  PRACTICABLE 
PROM  COPY  EURN.ISHED  TO  ULC  , 


60c 

C 

c 

c 

c 

c 

c 

c 

c 

c 

c 

c 

c 

c 

c 

c 

c 

c 

c 

c 

c 

c 

2 


c 


c 

c 


* f / 0 


23 

c 

c 


52  11 

C 

C I 

52  0? 


P^INT  n32 

F^OMAT  ( 3x  • 1 "1  • if  X , • c ? • , 5X  . • F~  • , RX  < • I M T Y HE  • . 2 X , • I M F ■<  < O', 

C JXiM"  DW  1*0  KHZ  Iw  Lf  V~L  IN  r«  r~AT.  M'~  OUL  ATI  CN  • , / > 

FH  TUPN 


T-ST  IS  P<~  BF  TtfM=D  TP  e-r-f  j-  [‘jTr.jyn.,  cp-p-fA-Jv  =>i  r"»UC"  O 
ny  Thp:S  ccc  jijlnci -5  will  r,  i v"  a C r nt i hut I r_  n a t th.. 
-HEC5IV-?  FFUOUHNCY 


FMIN  IS  THr  MIN  FCF)  _-r  jmtfpw  'O 
FJAX  IS  MAXIMUM 

FUL5  0-  THUMB  Fir  TH~  =)AN0wIDTH  jF  A F M SIGNAL!"  TZ 

CALCULATE  WHLTHES  I NTrH'M'T)  Tc  # I TH I N F«,:C  ~ I VC  3 PA'S’UTJ 
F.TT  ! 8 W=2*Fh  5 QMOOIJL  A t I NGC  1 +MUDUL  A -I  UN  I ND-rx  ) 


T”  ST  T*i  S=f  Ic  I NT  HP  vr  r,  j c n;,at-j  within  r<SC;iV  ' 

0r  LT  A = = MODI  * ( I . +BE  T A1  ) ( 1 . + BET  A£  I+I-  M-)0  3$  ( 1 . + -v:t  A T ) 

FFMiN-p^rai  hfi-  eo ■'-f~  op cv- 

FF'M  X = F F M I N 

/ 

FN*IN  AND  FV  AX  a r- - C "IN V~ ° T ) Tp 
FMtN=cFvlN*l 0 0 J • -DELTA 
F‘<  A X = rrM  A X*  1 0 0 0.  +1r  L TA 

CHECK  T ■)  gc  1 tF  FMI‘I  ‘ * A x F A L L c « I THIN  —_Cr  I V „ 

IF  ( ABS(  - MAX)  . LT.  ( ° W/Z  - A . AN-,  ( F»I  -:)  T . (PW/2  .))  c,r,  TJ  - 2 J j 

CHUCK  TO  S'-  T-  TNT  3 un;  t.TJ  ~ -CV» 

IF  ( F v In.L  ~ . ( -'l#/  - . ) . AN-.  - ma  x .C.  . ( 3 a/  ’ . ) ) GO  TO  - ’00 
GT  TC  520? 

F ;UMzF':-.il  +F  --o-'-h-  - ; i 
f ? a = D“LTA*2. 

I F ( h - 5 0 1 . NT  . = ?•-  or  . AND.  I ~ I*  I T . J . \ ) P-  I r • T ' 3 , F •;  5 0 ' “ 0 - . - S J 

C M , FF-w  , ="  JML  f V 

nr'MtT(-)FH,l  ,=X,  1 KHF’-M  ,,;X,FV,JX,Fr,J>!:X,”,') 


Ic(cr  JJl.iG.FP-0  :,X'n.IAt.!-:’,-l|.1)o  ! N”  r J 0 1 . F - II  ,r.  ; J • , F;  j c j„ 

C . S UML"V 

F IBM  at  ( 2F  s.  I ,1  OX  , • ?f;_t  1 , 5X.c-h' , H, -c  , j,  ! cX.F‘.l  I 


u^Tljr  »j 

'*  IS  Nr  T IN  Th  c~r.  T v"F  < * 
I TES  T=  ) 

‘NO 


C 

c 

C* ****************************************************** ************** 


-137- 


l 


c 

C THIS  SU-IB^UTI  NS  CALC'JUT‘iS  THF  AMT  ,1F  CFJSS  MOOJLA’I^N  DJ~ 

C TO  INTERFERING  "m  S t AT  ! >'  S.  AL  S 0 COMPUTES  THE  Avr  'if  CPJSE 

C COMPASSION. 

c 

DI  M'NS  I ON  FREQ ( 20  > ,FMOD(  ? 0 > . SETA  < 2 0) 

C 

PEAL  MXLFI  20  ) , PJW^Pt  >0),FQ">cV(  ">0  ) 

PEAL  L-VF.H20) 

C 

COMMON  FQ?fi/<,X‘-IJPi  . XMOOr  . X M-10  ’ . XMOlOA  . SUMC  i JW  »F<  2K  1 
COMMON  CP"()  , LS  VE  L . FMOO  , 2F  T A 
COMMON  MIL^.POvxEP.FaOEV.NFTAT 

c 

c 

c 

c 

C 1=  THB  am  MODULATION  IS  TO  COMPUT'D  USING  the  F p C OJt  NCI'S  OF  *'ODU. 

C MODULATION  S PFC  I F I FO  F CALCULATIONS  A ° E MADE 

c 

c 

c 

of INtt 

l FOPMAM*  t0  USE  maxIM.jm  FP  E iU~NCY  DEVI  ATI  r'  \j  C^cCtcIED  OY  M TOUL  A • 

C » * T I ON  F Pc  OUT  NC  Y •,/,«  and  mdu  J L ATI  ON  INDEX  TYO  1,  0 ’ HE  P W I S ,£ 

C FODEV(I)  moot  Sd:CIFI~D») 

p E AO  El . I Fl  AG 
51  FOPMtT(M) 

C 

C 

C 

C 


C » ; I NT  ! NO  ■"  f M-«HMGr  

C 

C 

Prj  I N T ' J 

20  FDEMATI*  a AA  *A  A * A*  '•r.n  ? r VCPULA’IDN  CALCULAT!  NS  < a A A * A A A A A A A • ) 

DP  I NT  21 

21  c-iVM  ATI  • STATION  r '.MPPr  SS  I 'N  FFFQ  O'- V . \ \ vv  "I  • , / , 


C*  F — CO  T LHC .SIGNAL  ftp  XV  CD  CAUSED'./. 

C • ( MH l i I N DP  FAUC.IKH/)  Hy  STAY.',/) 

c 

C 

c 

C 0 - -Ar-I  ST*  T I N (.4  L C'Jt  a T I mg  CP  '.c  c . ■ i"  ||.  A T I AN 

C 

r 

C V A • 7 A ' ‘L  0 N A M . c ; 

r 

c 

c r-  ■ ' OL-l.  I U"  A COMPP‘-«»AN  ‘ »rT  T f r>  ; PNC  STATION  j'.'LTt  ITH  c'rATID\> 

-138- 


o o o o o 


c 

c 

c 

c 

c 

c 

c 

c 

c 

c 

c 

c 

c 

c 

c 

c 

c 

c 


THIS  PAGE  IS  BEST  QUALITY  PRACTICABLE 
FROM  COFY  FURNISHED  TO  DOC  

I 3 K3/KI  *1**2  «HF.op  n IS  TH~  SIGNAL  L'lV  L OF  tHE  ITh 
ST  AT  ION  at  t hf  »‘<p  IN^UT. 

Pr  PC  NT-  Si  CPOSSMOD  '1  - LOCAL  I2"p  SIGNAL  DUS  TP  ',TlT  In‘J  [ ,C  '1SS 
COMDR'SSI JN  NOT  CuNSIPFPUD 

PER  COM-  X AM  MT!>  DU*  T)  STATION  I c OUS  I Dr  C I NG  IN  j'6  T<3»  Z c S I rN 

SL)MC-CONPPcSSON  FACT^p  FHJf  1)  ALL  S T AT  t Ot''S  . ( : If  JL  T ANT  LDC.  L'-v/f. 
IS  MULTIPLIED  MY  (1-StjMP) 

SUMCDU-COMWRiSSI  ON  f'c  LOG.  SIGNAL  DUr-  Tf  All.  rv  ST  A T , " xp:  . S ' - ' IN 
03. I. E.  SU  MCU  3=2  0 LOG  C-SUMC) 

XMODt-TOTAL  X WM$  AM  V(V)  AT  -i-|r  OrT^^Ty.  r'UTr>')~  nj.  tp 
CPCSSAOD  FCDM  ALL  STAT.(  jt  IMCLUOF.S  EFFECT  r-  1.  f 
GPOSSCOMP  vHSS  ION  ~NLY  I F SU«C  IS  LSS3  Than  \ 

XMPP?-%  C NS  AM  MOO  at  THC  JECc  IVF‘<  AUDIO  1UTOUT 
XMODt-tj  9>'S  Am  MOO  AT  ISO  N 7 FILTpF  OUTPUT  OF  •‘-C  IVT- 
XMODA-  x 'MS  AM  MOD  AT  A)  02  Fi^-pp.  r,UT 


TO  TAL =0. 

X M 0 D l = o . 

X M " 0 2 = 0 , 
xm;d j- j, 

X MTI  d a - , 

SUMC  = 0 • 

DO  50  I=1,NSTAT 


f 3MPJT  AT  I ON  OF  TU-  Cp  >*  ,M*,0  "S:  I "■!  DU=  T.t  T 3 * J --  STATION, IN  03 

C 
C 
C 

C^niSiROKH'.Al.’V  L ( I ) 4-P  . 

c 

C GONV-  ' -I  NG  -•■'MM  r;p  TP  Ll’j*'.  ■ 

c!’:?ri.=  io.*i‘('' Nc-/n.  > 
c 

C CALCULATION  of  r n-o: <r;  - \ t C'L1’‘- 

C '•TATI-  I 

C SITING  C H M P 7 S T } A LAf'GF  VALUe  Ti  'JIN'  > - 

C 1MD“S=1 000 . 

T p ( ( 1 . -C-  '•SSL  ) .GT  . J . )CDVP-  ? =■-'  0 . A A l 0010(1. 

c 
c 

C '••JV'i  OS  C ~ C S r CN-rL-LMM  'U  T. . A j l TA'l  N 

-•  JMP-r  Si  1 *c  t-C  J •"  3 

c 

c 

C CALCULATING  max  pr  ) P YT  ",  t t \ 1-  . .1  i(  c[  j]  * j -4  - V , ' - N T I.  •'  40 

C IN  FS  0*1  AMOVE 

c 

c 

C 

-139- 


IHIS  PAGE  IS  BEST  QUALITY  PRACTICABLE 
FROM  COPY  FURiilSHED  TO  DDC  «•- 


IF<  I^LAG.tO.  1 ) FQD~V(  I ) =<3F  TA(  J)  *c«aD(  I ) 

r 

C THE  R F AMP  INPUT  FILTER  SLOPE  IS  APPROXIMATED  TC  BE  .15DB  /100KHZ 

C FOR  FREOS  LESS  THAN  8 MHZ  FFCM  RECFIVFF  FREQ.  ,.07SD8  / 1 COK  HZ 

C OTHERWISE  * REVISED  JUNE  78 

C 

C COMPUTATION  OF  SLOPE 

SLOPE3 ( ( 2.  302)  / 20.)  *(.  15/10  0.) *10** (-  { (FQRCVR-FREQ <T))*1. 5/20. ) ) 
IF  ( (FORCVP-FREO  (I)  ) .GT.  8.)  SICPE=.2511*(.C7  5/  100.  >*(2.303/20.)* 

C 10**  (-(  (-8. ♦FQRCVR-FREQ (I)) * .7  5/2 C.) ) 

C CALCULATION!  OF  PERCENT  CROSS  v JP  OUT  TP  STATION  I 

C 

C 

FHOCNT  = 3LOP>E*FQDtV(  I )*?.*rn«TL*  11). 

C 

c 

C SUMMING  PMS  am  mod  DUE  TO  ALL  STATIONS 

C 

C 

C TOTAL  IS  THE.  UNFILTEPFD  TOTAL  % am  mtp 

T1T4L  = A£CCNTt*  ?tTTry_ 

c 

C 

C 

C CALCULATING  THE  5v;  \ am  MiO  AT  T-T-  >JTPUT  1"  TF  :.'IVC) 

C f ! LTES; 

C 

c 

XMOO  I =XM0D1  F ( PFPCNT*ri  LO=  -T(  p*.,r  oil)))**? 

XM3D2  = XMOO?F  ( PER  CNT*3IL  4JT(  I ) ) ) **  ? 

X»»CjO  3=XMOD3  + ( PF“CN‘T*  F I L 1 s O ( =V->D(  1 > > ) *** 

XVO  J A = XMCDAF  ( P£A  COT*  3 I LOO  ( F M ',0  ( I ) ) ) * *2. 

C CALCULATION  of  3ftCf’lT  M-OJLAT  ICM  C INS  ID*"-  I N » C ~ 'SE  :i  <p?-a;,I  1\ 

c pu?  n:  stati jn  i 

f £&cr v=pf«cmt/( i . -cpos  - l ) 

c 

SO  PPINT  lJO.r?r1(!  I.CI^n-5,1-  J^ryj  n,P---C:iW 

130  t-noyAT(Tx,F5.1.pX.FF.i,t-!X.rA.  ) , tx  , f S.  2) 

r 

c 

c. ca.cjlati'ns  ’p  - ital  o-oss'or  

c 

c 

c 

C I F T jTAL  c pMp.v-r  " r V -F  L'jr  • SIGNAL  ^ J ~ ALL  S T A T I P N S I 

C IS  GREATER  than  1 THJFD  OLNL-  4-jr>e  L OF  Pr  C.~  IV‘»  D*'r  E NOT  APPLY 

C 

IF((1.-AUMC).LT.0.)C,  ) t-<  >-)0 

S' )Mpp  --?o.  *alogi  o ( 1 ,-Tjr  ) 

N'T  TO  ->10 
30  0 PL  I NT  ’01 

301  *M*T4  jP  S G.  c S c 1 r ’<  ,L  Aft1  THAN  ’ , THU’S  '.P'T>  i, 

r .y  - r,  *.p  3«rrj  vr-  •('  LfN  ■ - V»L  I .»  *♦♦*♦**  • > 


140. 


SHIS  PAGE  IS  best  QUALITY  practicable 
from  cof*  furnished  to  ddc 


c 

C Setting  DUMMY  VALUF  eod  total  rCMBpf  $?i  on.  I . c.  1 UMCDt1 

C TO  PRINT  ASTERISKS  WHEN  S'JMCDH  I?  P^tNTcD 

c 

c 

SUMCD^-IOOOO. 

c 

c 


C CALCULATING  nT4L  MODULATION  in  ff  " >Nf  , CCN?  IDE.R  !f"C.  "OM3  ;r  =■  3 I UN 

C C>e  tup  DFSIRFD  SIGNAL 

c 

c 

c 

313  SUMM=sqqT(  TOTAL) /( 1 . -5'JMC  ) 

PS  I NT  4 00  • TOT  AL  , SUMC  . SUM  CD  3 , SUM  m 

400  FO  smaT  ( # TOTAL  sfrcpnT  MOD. ,NCT  C ON  A I DLL  I N G CR  "S  f C LVOi  S 1 1 ON  = • . 

C -S.n./.*  CJMOprSSI0N)  FACT-P-  I, F3. ->,/,•  L C'C  AL  I 7ra  F I5N  AL  C'USPE 
CION  IN  OBs'.F’.l  ,/,  • 'T'<L  y.  am-  ij'(F  IL  T t P FD  AM  M.  DUL»TI  TJs  • «F 

C?  ) 

C TOTA_  X ptlt-c-o  a v MOD  'A  LCUL  ATI  

C 

C 

C !c  THE  T0TtL  C»0SSC0MPPH<5SI  fN  DU?  T ’ ALL  F T AT  I r N “ I?  '.R*-  a-  ; - 

C THAR  1 ,TH-DI  THE  PFF'FTS  np  CL'S?  C CMPPESS I C-N  AS;  N T INCLUDED 

C IN  THE  X A M M.DO  C-*- 1 E'JLAT  I ON  S , t Tin;  W I ■>£  IT  I?  CCN5I0*-  -:.;0. 


C 

c 

r = 1 . 

I P7  ( S U • L T .1  . )C=1  .-?'|MC 
C 

c 

X*"*01  -?OFT  ( X m:  !D!  )/‘~ 
fed  ’ = : )st(  xmom-  > /c 

XMCD  ! = - OCT ( X '"9;  )/C 

XND94  -30°T(  X'LjOI  I /C 

c PRINTOUT  OF  TH-  " ra|_  AMOUNT  r AM  < SOUL  ~ I ON 

C 

c 

c 

npivT  ’JOJ 

3000  F 1PM  A T (//,  1 X .■''»(•*•)./.  ’x  .*  Tr  T AL  -M-  AM  M-'D  JLA-l  ON*  . 
C /,  ax,  'JUe  T / CR"  3SMTD  L-  • ) 

IE  ( FUMF.l  I . ) PS  INTO  JJ- 

30  0*3  FDPM  AT  < ix  » • D -0?SC0mp.t:  c AI  >»•  j "F'-CT?  INCLUD'D'  ,/) 

I r ( *u  .;c . Gr: . i . ) ps  i nt  to  ) - 

30  0 3 (X.'D.d^ccvP'J-'F'oi  TE  •>.  Tt  Nr>-  I »:C|  ,/> 


M.,  I KjT 

•»  0 01  . XMUD1 

PRINT 

•’302, XMOO’ 

I ■ | T 

■"103.  X NOD  3 

p-  I NT 

'OJA  , X M JDA 

30o:  f'jivuri  ix,'j:Tr:cL  dut->ut  a ■ i m if'ui.  a*  i on  : • «r  a . : , • .„•,/) 

300.0  F V- ma- ( 3X.  • ;jr y 11  IT  a ■ j r V MODULATIONS  '04,1,'  ■.',/) 

3003  F'»M.1T{3X,'1-J  H / f IL'-s  DiTnijT  am  'O  OiJL  A T I >j  S S',/) 

30  0 A H4M4T|1X,  • MO  MX  E n T -r  -|jr-''UT  am  MODULATIONS  ' ,-*.1  ,' 

F-FTU-  N 


141 


BBS  PAGE  IS  BEST 

rat*  COP*  FURNISHED  TU  D*>C 


SUBROUTINE  MIL  ED  n 

THIS  SUBROUTINE  CONVERTS  THE  DISTANCE  FROM  THE  -M  STATION 
TQ  SIGNAL  LEVEL  IN  DOM  t T THE  R'FAMR  INPUT. 

THE  PROGRAM  TAKES  INTO  ACCOUNT  THF  OF  AMP  FILTF®  CHJ  3»CTt-  ICHCS 
AND  THF  ATTENUATION  dF  THE  NAV  A NT  CNN A . 

US'S  FREE  SPACE  ATTENUATION  FORMULA 

DIMENSION  FR'Q(’J)  , PMODI?  0 ) . HFT  A ( «- J ) 

REAL  LEV  EL (20) 

rF  AL  mIlEIEJI.POWERI’OI.^OOEVI  ^D  ) 

COMMON  FQ»C  VR  . XMJDI  , XM->d.->  ,x  MOD  7,XMC->A.SUMC.Bl*,R<3Kl 
c DM  MON  FREU  , Lf.VF.  L . RM.OD  , ^T  A 
COMMON  MIL"*  PGWE R ,FODr V ,NST » " 

PR  I NT  S 

FDPMJTC  T o COMPUTE  SIGNAL  LEVEL  AT  CF  AMD  ]'PJT  TYPJ  1* 

C,/,*  GTHFR  WISE  VALUES  DF  LrV'L(I  I MUST  FF  INITIALIZED  IN  Tur  • 

C ./.  • EDIT  NODE  and  HScD  as  T|f  SIGNAL  LEVS  L AT  THE  - F AM’.»  IN  »JT  • ) 

c-ac  e.  I flag 

F D " M A T ( I 1 > 

I r ( I FLAG  .NT.1I RETURN 
PRINT  40 

FORMAT! • PRINTOUT  fF  SIGNAL  L'V'L  C DM PUT  A T I ON S • • / ) 

POINT  41 

FORMAT! • STATION  MIL'S  STATION  FREE  NAV  ACV  = IN*. 

C ‘PUT  SIGNAL'./.  • FPrO  PRIM  PPACc  AN', 

C 'TF  IMA  FILTER  LrV  -A T ') 

PRINT  42 

FJRMAT(«  (MH7)  R "OF  IV  IN  Dl<m  a T T N -R  J bT5>iV,  A T T " N DR', 

c • RP  AMP  in  ( d <M  » • , / ) 

DO  S3  1 = 1 .DO 

IF  { I . GT.NSTAT.  and.  I , N- . ->  ) )G  3 T i GJ 
R::  SPAC=  ATTENUATION  C ' VP  U"  A T I C’l 

At"pn=36  . ’t’O  . * AL  DC,  I D(  pR"  J<  I))  + ’J.*ALQG1J(  * I L’  ( I ) > 

R«rv*P  R = 6 0. 4-  l D.  *A  LCC.l  3 ( P OW  -’(I)  ) 

COMO  JT  AT  J ON  OF  LOSS  DU"  TC  t NT  * NN  A CHA  ' AC  TE  '•  I ST  I •“  5 I OR  PER  MHZ 
nrL’1H  I3a.:'!H? 

A L D S S = J . 

! F ( P - -()(  I ) ._  T . ID  R.S  ) f.L  J = S=  1 3 i.  -5-FR  F 0 ( I ) 

OF l c S S = «FP  Ilt  { FQC  CVO  ,F---D  ( r ) ) 

* nFi  GAIN  IS  ADDED  FTP  r,|  d=ctI  VE  GAIN  PR  NAV  AM‘'\‘;S 
L"  v/FL  ( I | = r.  n I*p  - ATT  ON-  AL  ">S  S LO  SS  ♦-  . 

I p ( I . :'3.?J)RR  INT  t 0 ) 

PRINT  I 00  . F'RP  1(  I ) . M I L-  ( I ) ,'P-wFR  , A TTEN,  AlDSS  . RF  . "S  - . LrV-_  ( * ) 

F 3 = MAT(Tx,FS,1  ,'tX.FA.I  ,'X,'  . 1 . S X . F S . I , ■>  ( 4 X . F E . 1 ) . X , p " . 1 ) 

CONTI  NUE 

PDPMATI/,'  LDCAlIZ"  SIGNAL  L E V f L CALCULATIONS:*./) 

Rr  TURN 
PNO 


REAL  FUNCTION  ^MLMS  - rrv.P^N'u) 


-142- 


IMt.ttUUUC 


non  n nn  nnnnnr  ->  -»  on 


THIS  PAGE  IS  BEST  QUALITY  PRACTICABLE 

7ROU  COPY  FURNISHED  TO  DDC 


THIS  FUNCTION  COMPUTES  THE  A HEN  U A TIC  N OF  THE  RF  AMP  INPUT  FILTER 

FOR  THE  NAV  11  RECEIVER  * RFVTSED  JUNE  78 

RFFIL?=0. 0 

DELTA=PRORCV-FMFFFO 

IF  (DELTA. LT. 0 . ) RETURN 

IF (DELTA. GT. 8. )00  TO  10 

RFFILT=  1.5*DELTA 

RETURN 

0 RFFILT=12. ♦ (DELTA-8. )*. 75 

RETURN 

END 

*« •* A******************************  ***************************  **<  * * * * 

SUBROUT  INF  CHANGE! JUMP) 

»<.«*********************************•**«************•*********** 


THIS  fCOG^A-4  MA<  :.S  CHANG*  « 1\  PC  (X,c  AM 


PE  4L  NAME  ( 1 5 ) / »K-iFO  • ( ' - ML  D*  , • 7 r.T  A • , • LSTV*  • , ■ *■*!  L • . • P'n;  • , 

C • , • I NST*  . • T YPC  • . • Srn=  • , • MX03  • , 'XW  I M 

01  MENS  ! ON  F«*  ()(.">  ’j  ) ,F'nn(  ? 1 ) ,“-ta  { ->3) 

■';l  ‘i i Lf  ( 30)  , piwcf  < 7 0)  so ) 

fpal  l::vfl<20) 

C!"M?N  FQOCV^  . XV  joi  , X‘"~r'>  , <*»(  U4  . SIJMC.  • )W  . 7K  l 

r o M ON  F p nU . L = V z L < f v DC*  » U : T A 
CO'*  VC  N MIL”,  PJ  W:  f i r O r>  F V . \ <iT  A T 


04c  avc-t  = F JJ  I"  T 9 C = TU  - NF  O «I7f  A VALU1-  V-ltCH  T"L!_E  XH?C 
CONTF  “L  is  71  >)“  TCAN'F^C  J«-n  . JJ  '>P:  ) !c  N1CMA. 


JUVD=; 
D?  j NT 


• ( • ***  PFPGPA  11 

FTC  INST '-JCT  I r NG 


* A* • . / , 


TYP'-:  I Nc T 


c ~ t > i n sr  c )>•  v a mp 

c - AS1  1 , o A-1 A , M,  X 

>-  7 ~ “ A T ( A 4 , 1 X . I ' . 1 1 . •'•>  ) 

)F  ( p AC  A . FO.  NAMC(  1 0 ) ) f,l  T ^ ■'J'" 


FLAG  I 


1 W 


I &G  -0 

I F ( N . - 1.0)  I F 1_  AG-  I 
I r ( N . 1 . 0 ) N = 1 


143 


r 


mis  PASS  is  BEST  8<JttI « raACIICABU 

JBOM  copy  FURNISHED  ro  HDC  . — - 

c 

c 


c 

rcTcpMI'JING  WHICH  C 0 MW  A NO  T ^ li^AO  I N QY  COMPARING  PARA  T r 

NA'»F(  I ) 

c 

• 

c 

DO  £0  1=1,15 

IC(PAR'A.EQ.NAWE(  I ) ) G O TO  «3 

20 

CONTINUE 

POINT  ?1 

21 

FORMA T(  • ERROR  IN  INSTRUCTION  .RETYPE  •) 

GO  TO  1 

C 

c 

c 

CONTROL  IS  TRANSFERRED  TO  THE  PROPER  STATEMENT  ACCORDING 

T C 

c 

THE  INPUTTED  COMMAND 

c 

c 

c 

30 

GO  TP  (41, 42,43,44,45,45,4"’', 40,40,50, 51. 52. 53. 5 4, 5 5).  I 

♦ 1 

FREU( N )=X 

GO  TO  ? 

42 

FMTD ( N >=X 

co  n ’ 

43 

Q=TA( N)=x 

GO  TO  2 

44 

LE  VEL ( N) =X 

GO  TO  ? 

45 

NILE( N)=X 

GO  TO  2 

* 

45 

POWER ( N ) =x 

GD  T1  ? 

[ 

47 

FODEV  ( N)  = x 

CO  TO  2 

45 

PRINT  4=1 

PR  I NT  432 

GC  TO  1 

49 

I F ( I FL  AG.c3.  1 . ANP.N  t -5  . I P AO  . “Q.  3)  SD  TJ  44«- 

50  TO  405 

49^ 

PR INT  491 

49  ! 

roPVAT(t  FTATIj-4  ST4~I  )N  ■<  O'JL'T  I IN  ‘4  ISOLATION  PAX  F- 

C L NILES  STATION') 

- 0 

L - Vr 

PR  I NT  49  2 

49  1 

CIU>4T(I  BJJNOOP  FO^q  =t-E  ) (K!J(  ) IN^tx  V I A - 

I 1 N 

AT  3 

CP  AMD  .“COM  POw'C-  t<P*l 

4 

P? int  ,q3 

49  ’ 

r jONA-(  1 ox  , 1 ( «H2  ) • . 2 -X  . • ( X wop  CALC)  INPUT  =*>3T  ~ if  C f I V • J * 

• 

r • ( Km ) • ) 

4=' 

PRINT  I 30 . N,  RREUl N ) , FW TD( N)  ,l  ‘ T A ( N ) , FO 35VI  N) ,Lr V _(N), 

r «-n.  _ ( n)  ,pow"p<  n) 

10  0 

rORM4^(3x.I7.5X,4  4 , 1 , 5X  . r - , ',H,FP.’,7X,t‘i,|  , A X , c ^ . 1 , 4 X , 

,7X  ,- 

C7.p| 

ro  TP  2 

50 

5 J TO  39 

5 1 

JU^Dp1 

c,r  T o so 

-144- 

- 1 


UHIS  PAGE  IS  BEST  QUALITY  PRACTICABLE 
FROM  COPY  FURNISHED  TO  DDC 


52 

JUMP: T 

GO  TO  1 

51 

JJMP  - 1 

GO  TO  05 

54 

JU*AP  = 4 

C.O  TO  0 4 

55 

JUMP:  1 

GO  TO  49 

C 

c 

C I F IFLAG  =1  THFN  TH~  C’MMAm  ) IS  >■  T D FT  C AT*-  ?TAT  1 p.N 

C (EXCtPT  FTP  "T
…[truncated]