Embed Size (px)
Citation preview
SELECTIVE FADING ON 8 GHZ LONG PATHS IN EUROPE.(U)
UUCMAR 80 L 6 NAUSE DCFR-04022WECLASSIFIED DCEC-R-040022 NL
11111.2
MICROCOPY RESOLUTION TEST CHARTNATIONAL BURA 1 IAlNDARD,; iAbA A
R-040022
DEFENSE COMMUNICATIONS ENGINEERING CENTER
REPORT NO. 040022
SELECTIVE FADING ON 8 GHz
LONG PATHS IN EUROPE
MARCH 1981 EAA 2
8 APPROVED FOR PUBLIC RELEASE: DISTRIBUTION UNLIMITED
PREPARED FOR THE DEFENSE COMMUNICATIONS AGENCY BYTHE INSTITUTE FOR TELECOMMUNICATION SCIENCES, NATIONAL
TELECOMMUNICATIONS AND INFORMATION ADMINISTRATION,U.S. DEPARTMENT OF COMMERCE0 31
8-1 5_26_00
I:S
S
S
gCFt~~6//Lv *'* / ~ 2I' /1
/4-
- - -- -r>V -
UNCLASSIFIED March 1980SECURITY 9j F4(ATION OF T4
0ASE (W',.i Daea Entered)
REORDCNPAGE I READ INSTRUCTIONSl/Z/REPORT DOCUMENTATION PAGE BEFORE CO1PLETING FORMR1POF~r11U&AACCFSSION No. 3. PFC'0r . CATALOG NUMBER
O .. TTLEfan Subitl) 5 TYFE aF R.EPOR'T G PERIOD COVEREDt Long Paths in =Europev 02/01/80-12/30/80
g.ahs Ere 6. PERFORMING ORG. REPORT NUMBER
7. THOR(.) ," . CONI RACT OR GRANT NUMBER(s)
LaeranseJ " DCFRMr.229. PERFORMING ORGANIZATION NAME AND ADDRESS I0P ,-G.'AM FLEMENT. PROJECT. TASK
ITS/NTIA/ .DoC A;" A A WORK UNIT NUMBERS
325 BroadwayBoulder, CO 80303I IF1. CONTROLLING OFFICE NAME AND ADDRESS
Defense Communications Engineering Center /1 Mar8OB7Transmission System Development Branch, R220 - ,s.. WMrWGES
1860 Wiehle Avenue, Reston, VA 22090 9574. MONITORING AGENCY sAME 6 ADDRESS(If different from Controlling Office) IS. SECURITY CLASS. (of this report)
UnclassifiedIS. DECLASSI FICATION/DOWNGRADING... SCHEDULE . -
16. DISTRIBUTION IMT./ol this Report),. 1 I ," I " 2,-7.' . / 31 " ( ...
Approved fo pub]4 etes ;-d rt iit-intt. .
17. DISTRIBUTION STATEMENT (of the abstract entered in Block 20, It different from Report)
1I. SUPPLEMENTARY NOTES
This is Report No. 3 in a series on Frequency Selective Fading beingpublished by DCEC.
19. KEY WORDS (Continue on reverse side if necessary and identify by block number)s
selective fading; multipath distortion; microwave radio; digital radio;diversity
L 20. ABSTRACT (Continue on reverse side it necessary and identify hy bI.ck number)
....-- This report presents the description, analysis, and results of a set ofmeasurements made to obtain statistics about the distortion of the frequencyspectrum amplitude across the IF band of a 3-level-partial-response, digitalline-of-sight, microwave radio system. The length of path was about 90 km.Results show that for line-of-sight systems:
(1) Large values of distortion (1 dB/MHz) were observed during multipathfading eventsj -.-> (Abstract cont's on reverse side)
DD I FONM, 1473 A,-EDITION OF I NOV 65 IS OBSOLETE UNCLASSIFIED ........ ____
UNCLASIFIE
UNCLASSIFIEDSECURITY CLASSIFICATION OF THIS PAGE(n1hm, Date Entemd)
i42) Diversity reception looks very promising as an effective tool incounteracting these distortion effects.
/
(3) Multipath received-signal-level statistics can be used to predictthe frequency and severity of this type of distortion on a line-of-sight path i, j
4) The 3-level-partial-response modulation system used here was veryrobust in terms of its susceptibility to the severity of amplitudedistortion observed on this path during the measurements.
LA
UNCLASSIFIEDSECURITY CLASSIFICATION Or THIS PAGEf'tMen Doe Entered)
SELECTIVE FADING ON 8 GHz LONG PATHS IN EUROPE
by
Laurance G. Hause
NATIONAL TELECOMMUNICATIONS AND INFORMATION ADMINISTRATIONINSTITUTE FOR TELECOMMUNICATION SCIENCES
BOULDER, COLORADO
Acaessinn ForlNTIS GFA&I
DTIC TABUnannco,.:-,c cd
JU, a' ir ton -
By -Distrit-ation/Availability Codes
Avail and/orDist Speoial
a. .. , -._
PREFACE
This publication* presents results of an R&D project funded by DCA andcarried out by the Institute for Telecommunication Sciences (ITS), Boulder,CO. This effort involved the collection and analysis of propagation datafrom several digital line-of-sight microwave links in the Digital EuropeanBackbone, Stage I. Concern for the presence and effect of frequency selectivefading on DCS digital links prompted this project, as well as a series ofother related projects.
Readers who have comments or questions regarding this report areencouraged to contact the author directly at ITS or Dr. David R. Smith atDCEC, Code R220, (Commercial) 703-437-2085, (VON) 364-2085.
* This publication is the third in a series of DCEC reports on frequencyselective fading and its effects on DCS digital transmission systemperformance.
TABLE OF CONTENTS
Page
ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I
1. INTRODUCTION ...... ... ............................ .. 1
2. PREVIOUS WORK ...... .... ............................ 4
3. INSTRUMENTATION AND TEST OPERATIONS ...... ................. 4
4. DATA DIGITIZATION AND ANALYSIS ....... .................... 15
5. RESULTS ....... .. .. ............................... 17
5.1 Heavy Fading Period ...... .. ...................... 79
5.2 Moderate Fading Period ..... ..................... .81
5.3 Light Fading Period ...... .. ...................... 82
6. CONCLUSIONS ...... .... ............................. 82
7. REFERENCES ...... ... ............................. .86
LIST OF FIGURES
Figure 1. Line-of-sight microwave links converging at Mt. Corna, Italy. . 2
Figure 2. Chronological occurrence of each data category for the pathfrom Mt. Paganella to Mt. Corna ..... ................ 5
Figure 3. Chronological occurrence of each data category for the pathfrom Mt. Venda to Mt. Corna ...... .................. 6
Figure 4. Chronological occurrence of each data category for the pathfrom Mt. Cimone to Mt. Corna ...... .................. 7
Figure 5. Instrumentation used to measure and record selective fading atMt. Corna ....... .. ........................... 9
Figure 6. Fade depth distortion comparison for time period of29 May 1980, 0100-0200, Mt. Venda to Mt. Corna ........... 18
Figure 7. Fade depth distortion comparison for time period of29 May 1980, 0200-0300, Mt. Venda to Mt. Corna ............ 19
Figure 8. Differences between reference sweep and the sweep valuefrom May 29, 02-18-10 to 02-19-20 .... ............... .20
Figure 9. Typical analog representation of the IF spectraldensity function ........ ........................ 21
Figure 10. Fade Depth Distortion Comparison (Slopes only) fortime period of 29 May 1980, 0100-0200, Mt. Vendato Mt. Corna ....... .. .......................... 22
Figure 11. Fade Depth Distortion Comparison (Slopes only) fortime period of 29 May 1980, 0200-0300, Mt. Vendato Mt. Corna ....... .. .......................... 23
Figure 12. Correlation of spectrum amplitude distortion to fadedepth for 29 May 1980, 0100-0200, Mt. Venda toMt. Corna ....... ........................... .25
Lminnw~miii
PageFigure 13. Correlation of spectrum amplitude distortion to fade depth
for 29 May 1980, 0200-0300, Mt. Venda to Mt. Comna. .. .. ... 25
Figure 14. Correlation of spectrum amplitude distortion to fade depthfor 28-29 May 1980, 2300-0400, Mt. Venda to Mt. Cona ... ... 26
Figure 15. Correlation of spectrum amplitude distortion (slopes only)to fade depth for 28-29 May 1980. 2300-0400.Mt. Venda to Mt. Comna.. .. ... .... ... ... ..... 27
Figure 16. Fade Depth Distortion Comparison (Slopes only) for timeperiod of 13 May 1980, 0100-0200, Mt. Venda to Mt. Comna.. . . 28
Figure 17. Fade Depth Distortion Comparison (Slopes only) for timeperiod of 13 May 1980, 0200-0300, Mt. Vendato Mt. Comna. .. .. ... ... ... ... .... .......29
Figure 18. Correlation of spectrum amplitude distortion to fade depthfor slopes only, 12-13 May 1980, 2300-0400, Mt. Venda toMt. Comna... .. .. ...... . ... ... .... . .... 30
Figure 19. Correlation of spectrum amplitude distortion to fadedepth for 13-14 May 1980, 2200-0300, Mt. Venda toMt. Cona ... .. ....... .... . .. ... ...... 31
Figure 20. Cumulative distortion distribution (slopes only),Venda-Corna, 28-29 May 1980, 2300-0400 hours .. .. .. ..... 32
LIST OF TABLES
Table 1. Instrumentation Purposes .. .. . ... . ...... . ......10
Table 2. Instrumentation Identification. .. .. ... .... . ......11
Table 3. Major Instrumentation Interconnections .. .. . ... ...... 12
Table 4. Spectrum Analyzer Setting for Normal Operations. .. ...... 14
Table 5. Typical Spectral Density Representation. .. .... ... ... 16
Table 6. Cumulative Distortion Distribution For 29 May 1980,2300-0400. Mt. Venda to Mt. Comna, Standard SpectrumSweep No. 1 24 8 (Distortion Values Include Hulls) .. .. . ... 33
Table 7. Cumulative Distortion Distribution For 28 May 1980,2300-2400, Mt. Venda to Mt. Comna, Standard SpectrumSweep No. 1 24 8 (Distortion Values Include Nulls) .. .. . ... 34
Table 8. Cumulative Distortion Distribution For 29 May 1980,0000-0100, Mt. Venda to Mt. Corna, Standard SpectrumSweep No. 1 24 8 (Distortion Values Include Nulls) .. .. . ... 35
Table 9. Cumulative Distortion Distribution For 29 May 1980,0100-0200, Mt. Venda to Mt. Comna, Standard SpectrumSweep No. 1 24 8 (Distortion Values Include Nulls) .. .. .... 36
Table 10. Cumulative Distortion Distribution For 29 May 1980,0200-0300, Mt. Venda to Mt. Comna, Standard SpectrumSweep No. 1 24 8 (Distortion Values Include Nulls) .. .. . ... 37
Table 11. Cumulative Distortion Distribution For 29 May 1980,0300-0400, Mt. Venda to Mt. Comna, Standard SpectrumSweep No. 1 24 8 (Distortion Values Include Nulls) .. .. . ... 38
iv
Page
Table 12. Distribution of Distortion Event Durations and IntervalsBetween Events for the Primary Receiver using StandardSweep 1 24 8, Mt. Venda to Mt. Corna, 28-29 May 1980 .. ..... 39
Table 13. Distribution of Distortion Event Durations and IntervalsBetween Events for the Diversity Receiver using StandardSweep 1 24 8, Mt. Venda to Mt. Corna, 28-29 May 1980 ..... 43
Table 14. Cumulative Distortion Distribution (slopes only) For 28-29May 1980 2300-0400, Mt. Venda to Mt. Corna, StandardSpectrum Sweep 1 24 8 ....... ..................... 46
Table 15. Cumulative Distortion Distribution (slopes only) For28 May 1980 2300-2400, Mt. Venda to Mt. Corna,Standard Spectrum Sweep 1 24 8 .... ................ ... 47
Table 16. Cumulative Distortion Distribtuion (slopes only) For29 May 1980 0000-0100, Mt. Venda to Mt. Corna,Standard Spectrum Sweep 1 24 8 ...... ................ 48
Table 17. Cumulative Distortion Distribution (slopes only) For29 May 1980 0100-0200, Mt. Venda to Mt. Corna,Standard Spectrum Sweep 1 24 8 ...... ................ 49
Table 18. Cumulative Distortion Distribution (slopes only) For29 May 1980 0200-0300, Mt. Venda to Mt. Corna,Standard Spectrum Sweep 1 24 8 .... .. ................ 50
Table 19. Cumulative Distortion Distribution (slopes only) For29 May 1980 0300-0400, Mt. Venda to Mt. Corna,Standard Spectrum Sweep 1 24 8 .... .. ................ 51
Table 20. Distribution of Distortion Event Durations andIntervals Between Events for the Primary Receiverusing Standard Sweep 1 24 8, Slopes Only .... ........... 52
Table 21. Distribution of Distortion Event Durations andIntervals Between Events for the Diversity Receiverusing Standard Sweep 1 24 8, Slopes Only .... ........... 55
Table 22. Event Occurrences during the 28-29 May 2300-0400Data Period ........ .......................... .. 57
Table 23. Sweeps Showing Distortion Greater than 0.5 dB/MHzon the Primary Receiver ...... .................... .. 58
Table 24. Sweeps Showing Distortion Greater than 0.5 dB/MHzon the Diversity Receiver ..... ................... ... 59
Table 25. Sweeps Showing Flat Fading Greater than 35 dB ... ......... 60
Table 26. Cumulative Distortion Distribution (Slopes only) for12-13 May 1980, 2300-0400, Mt. Venda to Mt. Corna,Standard Spectrum Sweep 1 0 15 ...... ................ 61
Table 27. Cumulative Distortion Distribution (Slopes only) for12 May 1980, 2300-2400, Mt. Venda to Mt. Corna,Standard Spectrum Sweep 1 0 15 .... .. ................ 62
Table 28. Cumulative Distortion Distribution (Slopes only) for13 May 1980, 0000-0100, Mt. Venda to Mt. Corna,Standard Spectrum Sweep 1 0 15 .... .. ................ 63
v
PageTable 29. Cumulative Distortion Distribution (Slopes only) for
13 May 1980, 0100-0200, Mt. Venda to Mt. Comna,Standard Spectru Sweep1 0 5 ... . ... ... ... .... 64
Table 30. Cumulative Distortion Distribution (Slopes only) for13 May 1980, 0200-0300, Mt. Venda to Mt. Comna,Standard Spectrum Sweep 1 0 15. .. . ... ... ... .... 65
Table 31. Cumulative Distortion Distribution (Slopes only) for13 May 1980, 0300-0400, Mt. Venda to Mt. Comna,Standard Spectrum Sweep 1 0 15. .. . ... ... ... .... 66
Table 32. Distribution of Distortion Event Durations and IntervalsBetween Events for the Primary Receiver using StandardSweep 0 1 15, Mt. Venda to Mt. Comna, 12-13 May 1980(Slopes only). .. .. ... ... ... .. ... ... .... 67
Table 33. Distribution of Distortion Event Durations and IntervalsBetween Events for the Diversity Receiver using StandardSweep 1 0 15, Mt. Venda to Mt. Comna, 12-13 May 1980(Slopes only). .. .. ... ... .. ... . .. ... .... 69
Table 34. Event Occurrences during the 12-13 May 2300-0400Data Period .. .. . ... .. ... ... ... . . ... ... 71
Table 35. Sweep Showing Distortion Greater than 0.5 dB/MHz onthe Primary Receiver. .. . ... ... .. ... ... .... 72
Table 36. Sweeps Showing Distortion Greater than 0.5 dB/MHz onthe Diversity Receiver. .. . ... ... ... .. ... ... 73
Table 37. Sweeps Showing Flat Fading Greater than 25 dB .. .. . ......74
Table 38. Cumulative Distortion Distribution for 13-14' May 1980,2200-0300, Mt. Venda to Mt. Comna, Standard SpectrumSweep 0 0 54 (Distortion Values Include Nulls). .. . .. . .. 75
Table 39. Distribution of Distortion Event Durations and IntervalsBetween Events for the Primary Receiver, Using StandardSweep 0 0 54, Mt. Venda to Mt. Comna, 13-14 May 1980 .. .. .. 76
Table 40. Distribution of Distortion Event Durations and IntervalsBetween Events for the Diversity Receiver, using StandardSweep 0 054, Mt. Venda to Mt. Comna, 13-14 May 1980 .. .. .. 77
Table 41. Event Occurrences during the 13-14 May 2200-0300 DataPeriod .. .. . .. ... .. ... ... ... .. ... ... 78
vi
SELECTIVE FADING ON 8 GHz LONG PATHS IN EUROPE
Laurance G. Hause*
This report presents the description, analysis and results of aset of measurements made to obtain statistics about the distortionof the frequency spectrum amplitude across the IF band of a 3-level-partial-response, digital line-of-sight, microwave radio system. Thelength of pdth was about 90 km. Results show that for line-of-sightsystems:
1) Large values of distortion (1 dB/MHz) were observed duringmultipath fading events.
2) Diversity reception looks very promising as an effectivetool in counteracting these distortion effects.
3) Multipath received-signal-level statistics can be usedto predict the frequency and severity of this type ofdistortion on a line-of-sight path.
4) The 3-level-partial-response modulation system used herewas very robust in terms of its susceptibility to theseverity of amplitude distortion observed on this pathduring the measurements.
Key words: selective fading; multipath distortion; microwave radio; digital radio;diversity
1. INTRODUCTION
Multipath fading in LOS links is known to cause short term outages when the
signals from both the primary and diversity radio fade below the bit error rate
threshold for flat fading. In addition, outages have been observed which occur
at relatively high signal levels. These outages are often attributable to in-band
(selective) fading.
To obtain statistics on selective fading and relate them to flat fading infor-
mdaLon, the Defense Communication Engineering Center (DCEC) provided Institute for
Telecommunication Sciences (ITS) with tasking. The project consisted of a two month
test by ITS to assist in determining the effect of selective multipath fading on dig-
ital line-of-sight microwave links. Data was collected within the Digital European
Backbone Stage I (DEB I) on LOS links (see Figure 1) on which ITS had instrumenta-
tion installed for another test program. Data collected on this selective fading
project was analyzed at the ITS laboratory. These results are needed for improving
performance and design criteria for wideband digital LOS links.
*The author is with the Institute for Telecommunication Sciences,National Telecommunications and Information Administration,U.S. Department of Commerce, Boulder, CO 80303.
" , ,.. -- - - - - - "-- .- r -- - - -" _ -" .. . . . . . . . . .. .. ... . . . . . . .. . .
Hahefistadt
1~ GERMANY
I" I
L-
~i. /oaelaAi
/4~****ITALY
cog t
'-9 Zuppitztie Se
Mi SerraCAUSTRIA
Fiur 1 in-o-sgt icowv lnk onerin t t.Con, tay
SWITERLND NdmaGallnG2
Specific purposes of this selective-fading measurements program was to measure
the amplitude distortion and other related parameters in order to answer the follow-
ing questions:
1. How often does the amplitude distortion occur?
2. What is its range of severity?
3. What effect does the distortion have in producing errors in
the DEB I system (12.6 Mb/s)?
4. Can the year-long, flat-fading statistics (recently measured for these
paths) be quantitatively related to the in-band fading statistics?
In order to answer these questions, the following tasks were completed:
1. We conducted a literature survey of previous work accomplished
in the area of selective fading on LOS links, including work
published by Bell Labs, Bell Northern Research, and Collins Radio.
2. ITS ins'illed, tested and calibrated an amplitude distortion measurement
system at the Mt. Corna site.
3. ITS obtained in-band fading data over a two-month period on the
primary and diversity radio pairs and measured various signals at
selected times during periods of multipath fading. We continued
to monitor and record the distributions of received signal levels
that ITS had been previously monitoring on 8 GHz.
4. The signal spectrum data were analyzed to obtain the different
values in dB across the signal frequency spectrum using a standard
spectrum obtained during a quiet received signal level (RSL) period
as a reference. In order to obtain distributions of the values of
spectrum slope and in-band null depths, data searches were made
on the basis of frame/format events and maximum spectral density
differences across the band.
All signal digitizing was done at the ITS Boulder Labs.
The data was analyzed to obtain the following statistical parameters:
4.1 Distribution of spectrum amplitude distortion observed during
periods of flat fading.
4.2 Distributions of selective fade durations.
4.3 Distributions of rate of occurrence of selective fading.
4.4 Correlation of flat fade depth with the presence of
frequency selective fading.
4.5 Correlation of frame loss and format violation events with
selective fading events.
3
2. PREVIOUS WORK
Much work has been done to investigate the effects of selective fading ondigital, LOS, microwave systems, for example (Dougherty and Hartman, 1977), (Ander-
son, et al., 1978), etc. The effort has resulted in combiners and equalizers
which have greatly alleviated the effects of the distortion. A brief description
of the mechanisms at work on LOS systems is provided by Smith and Osterholz (1979)
pp. 15 - 26. The model presented by Smith and Osterholz indicates that the
distortion will be particularly bad for long LOS paths for two reasons. The first
reason is that the direct-path RSL is faded substantially below the median for a
much larger fraction of the time than it is for short LOS paths. The second reason
is that the potential for time dispersion between signal components is much greater
on long paths than on short ones.
3. INSTRUMENTATION AND TEST OPERATIONS
The purpose of this section is to provide a description of the instrumentation
system used to measure and record spectrum amplitude distortion and other relatedsignals of interest. The mode in which the instrumentation was used is also
described.
The three links converging at Mt. Corna (Cimone-Corna, Venda-Corna, and
Paganella-Corna) operate in a space diversity mode. The ITS instrumentation was
all located at Mt. Corna and it had the capability of looking at only the two
receivers associated with one of the paths at any given time. The observations on
the Paganella-Corna path were very brief since very little multipath fading is
observed on this path and for this reason it was abandoned for testing (Fig-
ures 2, 3, 4). The two receivers associated with each path are designated the"A" and "B" receiver. For these paths, the "B" receiver is the primary one and
it is connected to the lowest antenna.
The Cimone-Corna path was monitored for approximately three weeks (the lastpart of April and the first part of May 1980). On May 8, 1980, the instrumentation
was connected to the Venda receivers. The instrumentation was switched only once
C"certain commercial equipment, instruments, or materials are identified inthis paper to specify adequately the experimental procedure. In no casedoes such identification imply recommendation or endorsement by theNational Telecommunications and Information Administration, nor does itimply that the material or equipment identified is necessarily the bestavailable for the purpose."
4
:::h.=. Raina . It P th
Fcadsng
M I tia Pth 4Fading sithPa-. Fadsng
M-dt&-p-tl 3.1th High
Ca.di t ian
Ect.P-t GM. PRIMARY RECEIVER DURAING MAY IM6
MalV.nt.an
Pa... Fading 6-.tha-t Rain. Mw.ltipath
Rain 5AttenationFading
Fading ai.thPaa- Fading
Multi-path 1.ith High
Equipeen G z PRIMARY RECEIVER DURING APRIL I M
DAYS OF THlE MONTH
Figure 2. Chronological occurrence of each data category forthe path from Mt. Paganella to Mt. Comna.
5
P Fading
Or M4ut , path
Ran a
Fading
N.Ita-poth 4Faing with
Po,,.- Fadang
kwiti-Pth Cl Ca a a a a a.Lth H-agh
"ads-
LJ.,di .t-bod 2 _______= C= = = -a C= =
C.W: A On.
Condat a aon
Eqa.apl.at 1 GHx PRIMARY RECEIVER "URING MAY eeMai f-,mt -o
Po.wr Fodang 5.atlao .t R.anar Mltpath
Ra. 5Att-~toa
Muati-path 4Fadang wathPar Fading
Wo]ta*-path 3.ath High
Undaet- .bd 2 c '
S GHz PRIMARY RECEIVER OURING APRIL 196Eqa.pa .t.1Ha f,..nat tan
3$ 23' 36
DAYS OF THE MONTH
Figure 3. Chronological occurrence of each data category forthe path from Mt. Venda to Mt. Corna.
6
Pow..- Foding6
o.- MaI t %pPth
Rain a
M.ats-apt 4Fading wathPow.- Fading
M..ItsaP.i 3 . a - .-.At, High,M.
Unda..a-biid 2 C= = 49C== cc=Cand.a aw
Eq.. -t 18 GNU PRIMARY RECEIVER DURING MAY 1BIa
Mwl F-nt.ao
P-ww Fadinsg6
oadsoaaL Ranar M.1tapota
Rain 5
Fading
M..ats-ath 4Fodang withs
Pa-w Fadinag
MwIt.ath 3a -
wad, Mgl
Undawt..b.d 2Condaitiaon
Ew. p t 1 GHw PRIMARY RECEIVER DURING APRIL 1I3M
MoI F.noe. .C
DAYS OF THE MONTH
Figure 4. Chronological occurrence of each data category forthe path from Mt. Cimone to Mt. Comna.
7
instead of frequently between paths (as was originally planned) because of the
difficulty in verifying connections as well as the upsetting of communications oper-
ations. As seen in Figure 4, there was little fading during the last two weeks in
April on the Cimone-Corna path. No significant data were obtained on the Cimone-
Corna path until May 7, 1980. This 7-hour (2000-0300) data set is available but it
was not digitized and analyzed in detail since it consisted of short periods of
moderate fading. Resources of time, money, and specialized computer system access
were exhausted even before completing the analysis of all of the Venda-Corna
magnetic tapes. Visual observation of the May 7-8 data showed 2 hours of signif-
icant multipath amplitude distortion (sl.:.es across the IF band greater than
0.2 dB/MHz). The preponderance of slopes was negative as in the case with the
Venda-Corna path. Maximum multipath fading depth for the period was 30 dB. No
diversity switching was observed during the 7-hour period.
The instrumentation is shown by block diagram in Figure 5. The purpose of
each major instrumentation module is shown in Table 1; each major item is identi-
fied in Table 2. Table 3 lists the major instrumentation interconnections.
The equipment which ITS had been operating for monitoring RSL on the three
8 GHz links during the previous year were kept in operation for the two month
duration of these tests so that a relationship between the long-term flat-fading
statistics and the short term in-band fading statistics might be established. On
May 26, the strip chart recorder used for RSL measurements failed and was not
available in June.
The instrumentation system is very simple in its configuration. An AGC IF
amplifier is used to keep power levels constant into the spectrum analyzer in order
to prevent rapid flat fading from appearing as amplitude distortion. A complete
sweep through the IF spectrum takes a minimum of 1/2 second if the spectrum adjust-
ments are set so that the spectrum analyzer remains in calibrated operation. The
settings used during the distortion measurements are shown in Table 4.
Calibration was done once each day on the channels corresponding to the spectrum
analyzer, the normalizer and discrete events. Daily operations were as follows:
1. The spectrum analyzer and the normalizer channels were calibrated
during a multipath quiet period of the day, usually (1400-1500),
and the standard spectral density function envelope was set into
the normalizer memory.
2. The recordings were started between 1800 and 2000 and the playback
output of each channel was checked to see that all channels were
recording properly.
8
.2 V
"B" 86Hz Primary Receiver *W
IF eer, AmC a..-13FLA..
140 dB Linear Amp -2
IF Ap wth AGC *24Str-ipChartRooorder
To *A'Speo.Desk Top
d 4 AnalPamogCam
te
Anolo 9 o
i. aGn
Th Ioef orte A acistesoedein
FigurP5. nstuttion- use tomaur and
record ~ RS seletiv fOdngtetoa
9-o
Table I. Instrumentation Purposes
BlockDiagram No. Name Purpose
I Radio Receiver Provide long-path radio signal parametersduring multipath conditions
2 40 dB Linear Increase the IF level obtained from theAmplifier sampling point
3 IF Amplifier Normalize the spectral density functionwith AGC
4 Spectrum Analyzer Provide frequency reference (saw-tooth) andthe spectral density function envelope
5 20 dB Attenuator Provide isolation between the SpectrumAnalyzers
6 Normalizer Provide a tool for data checking and editingin the field on the basis of distortion level
7 Marker Signal Provide a frequency marker for data pro-Generator cessing use
8 Potentiometers Condition signals for compatibility with themagnetic tape recorder
9 DC Amplifier and Condition signals for compatibility with theZero Offset magnetic tape recorder
10 Analog Magnetic Record the IF Spectrum Envelopes and theTape Recorder various signals associated with amplitude
distortion of the spectrum
11 Time Code Generator Record "Z" time, for data analysis, checking,and Reader and editing purposes
12 Storage Scope Provide a tool for data checking and editingin the field
13 Pulse Multiplexer Condition signals for efficient tape channeluse
14 Pulse Stretcher Condition 3-level-error and reframe signalsfor recording system requirements
15 8-Port Mux Provide 3-level-error and reframe indicatorsignals
10
Table 2. Instrumentation Identification
No. Name Mfg. and Serial No.Model No. "A" "B"
1 Radio Receiver Collins MTC 019APAG OlOA
AN/FRC-165(V) MTE 01A
2 40 dB Linear Avantek 12 13Amplifier UT8-2211M
3 IF Amplifier Collins 741 750with AGC 22E4G-MW
4 Spectrum Display H-P 141T 1337A- 1320A-Analyzer IF Sec. 8552B 08018 05582
RF Sec. 8553B
5 20 dB Attenuator Narda
6 Normalizer H-P 946A- 2005A87A 02316 02356
7 Marker Signal 4o1 1431A02425Generator 86408 (common to
A and B)
8 Potentiometers NTIA/ITS
9 DC Amplifiers NTIA/ITSand Zero Offset
10 Analog Magnetic Honeywell 102347- 102358-Tape Recorders 5600C FA77 FA77
11 Time Code Gen. Datametrics 760and Reader SP-425-A/B (Common to
A and B)
12 Storage Scope H-P 1926AO09371744A (Common to
A and B)
13 Pulse Multiplexer NTIA/ITS
14 Pulse Stretcher NTIA/ITS
15 8-Port Mux VICOM
AN/FCC-97
11
4.)
go., r- (a s-
ww- a,0
ONI )0 4A >E ~ ~ ~ 0 (A '
C~ ~ ~ ~ ~ L I~~~.-. ~ 0
0:C 4.J-.
r- S-.4 e.Im5 - >I O =1 4) in a) to ~
Q Iw to IA +,f -cc. =Q- +0 L
~ 4. 0 0 0 0 0 4.. 4.(J 0 0 0C
0 5a. )4-4. 4-A
4-u
cm. 6-. , 0> Z' 0 4- d)
a) CL- mU 4) cm 0 :P 0 c4. E =c 43!1t LL. 4 0- r_ 4. 4.) ee LL o
0a Li.L4- 0 c
to 4 .)4 -4 .)
4.) I 4- CCC. 4-~ 4-) M-
-nC4) ~C' (L C- (D CI (1) VC 0- .0 4..- 4.) L. 4. S- 4J 0- 0 1.
( C C, C U4- 1044) 4.-.)
0- in C.JN 0 ~ 0i 0 (1).., . .
EU 4 I4 ..J I . Q).j 0 4-0 4j 4..I\J J C\ 1 *aJ cc ClJ C-. $-IW CLJ 'C..L-. ~ 0 0 1 4
S- (- 4-4 J )4
CD C> CC S. 43 C C4 CaL 4*I-~V 4L3 CL 43J . . 3 N I 3 I
u 0j 0l .-- cm- 0 -. s
43 i CD3 4D 'o CD as u . C L L- .)
u (U a ) 4
>. o4)c c .
4) 4- 04 41 0 4 - 4- 4- 4 4- 04 J .44
u 430
+- 44434- 4 1 U 4. CL ~ 404- N) C4 CD 4- Q 4 04-' 340 ND 4- 4 4)D1
to- OLL 4J N34 W7.4 4-b S-a .. C 0 O a.~~~~~~fl~~~~~~~~ L. 0 ~ n.J ~ /.. ' 4 0 ~ 4I- 3' ~ 1
12
2 m0) *
0 4.>
cn.. C1d)o o 41CO
1~4 04- 0100 4CC -)M q-) Ja
C) c 0 (A~ 0 (a (A 44.M- toEO ~ 4.8 I c. - CIt It If > 40 .- -0) ~0 L. I U (-> IC.C)L> - 11 > > 4-'0 z u
cm =.I 00 +0 4=10
ca ~ 0 0l 000) -0 LLO e 0UIn In wf q D .4-19- 0 ,Ag C..)C
Ca >.- . 'to..- 4+)0to 4-- C2 ..C 04 o" 0 D C-O r- 0 0.1 (1). 4)4 o:s -
.. j C%j C r- 11 a I I ++ I CL o LCL ( 000 CI41 0D
r- r- 4-J CA 41 (A
4JC a CC J 4- (1 4.)C S- CD s-O t.. 00 Q0
u >L- >L/.-0- 4.1 0Cu0- ~4- 4-J 0 0 CA0 n
CA 0 0 r.O .. 41 ~ 4J 4 'o - (Al
c OC r- I 3 0) uOI 4)t 0 u CD 00 47 (aa C.m OO_ Wo > L- 3 > 6- o> 0
>- 'Cr- C - *-0 IL to ~ 3 C.W)3 I-C/ lAc J Lf -.j CM- Cu In 4- 0
0 4.
0
C Ul)
C"I
Ml C
1..1
Cl S
S- I- &
0x 1--
IL.r10 4-
4 I- S- S. S-* =- (A
Ca 4.) OI.
U) ao0 0 fa 0 1 0. 0 0
4- S0 I, In a A>3N0 % .- >
0. > I. to -j >$- j - - -41 a) o. $.- C-Iu ulc LI ( < .1 S-
13-
Table 4. Spectrum Analyzer Setting for Normal Operations
Name of Setting Value
Frequency 70 MHz
Bandwidth 300 MHz
Scan Width 2 MHz/Div
Input Attenuation 10 dB
Base Clipper "0" 0
Scan Time 0.1 s/Div
Log Ref. Level -40 dBm
Log Linear Select Switch -6 dB
Video Filter 100 Hz
Scan Mode Internal
Scan Trigger Auto
Writing Speed Standard
14
3. Recordings during the night usually lasted from 2000 the previous
evening until 0400 the following morning. In a few cases, power
outages during the night prematurely terminated the recordings.
4. In the morning, the previous night's recordings were played back at
64 times real time to determine maximum distortion on the tape,
3-level-error or reframe events, and RSL. If the maximum levels
of distortion observed were of the same order as routine spectrum
variations observed in the commnunications system (less than
0.2 dB/MHz), the information was noted in the log book and the
tapes were erased and reused. If the distortion was significant,
the reels were labeled and kept.
The distortion recording channel was used only for field data editing and was not
used in the digitized data analysis except as a comparative check.
4. DATA DIGITIZATION AND ANALYSIS
The analog data was sampled and digitized at 200 samples per second per data
channel. Each IF spectral density sweep (Table 5) was given a number consisting
of the hour, the minute and the sweep within the minute, for example 1 0 34. A
number of average values of spectral density corresponding to discrete frequencies
within the IF band were calculated for each spectral density sweep (see Table 5).
This table shows values before normalizing them by subtracting them from standard
sweep values. These values were obtained by averaging samples in the neighborhoods
corresponding to the various frequencies. Of the 200 samples per second digitized
from the spectrum sweep channel, 100 were ignored due to the 1/2 second period
between sweeps. Of the remaining 100, only the even numbered samples were saved
leaving 50. The first 9 and the last 9 of these samples were ignored in order that
the analysis would be made on the more slowly changing part of the spectral density
function. The remaining 32 samples were grouped into sets of 4 which were averaged
to provide 8 points corresponding to each spectral density function sweep. In order
to calculate distortion values from the data, it was necessary to determine a set of
points representing a standard spectral density function. This standard was usually
selected by printing out the sets of points representing the sweeps from the first
minute of each hour. From these values a set of points representing the standard
sweep was derived for each day's data.
The selection of the points representing the standard sweep was found to have
a pronounced influence on the distortion values calculated. If a standard sweep
had a slope bias, of course all the distortion values were biased by that amount.
More importantly, however, the noise on the standard sweep adds substantially to the
15
%A w wVnw v w v w w wwwwwwwwwwwwwwwwwwwww
go 0
NS 0% 71i0%07%. 7ew %o 0%0%4'li7 7oo 1: 77os 7 1! 7 71% s000- . . . . .
01 I r*ft 7 i 71 G" 1 ! 1 !II : iw 1:1 io i7l 7' !1 i771 - l
v: Im vn c wOf'- inw
>- I co
0j '49
w 0
cu~.s N 9 N 9 N 9 ~ .. 4 N 99 w.q~~ 4N9 44~ft9 w99 ~lf
4Im D 0.WQ 49. DI- LIn) 1. 4
In0www ) -( z~0
O O 0 0 0
O Oa 0
OQ4m Q mm Qmw 0
mvOom0
W Ww v - - -O-mm mOO oQ
40~
Lc; Ow 0 0 0 0 0 0 0 0 0 00 0 0 0 0 0 0 0 0 0 0 0 0
4- c5.45ea- ; '0 -a0 - -- n4 Q o0wwww av wWWmWQ ww 4m5 Wt Wm 0t
to >
a% 40.)1 c c 1.09w 0: 00 0 0 0 0 0 0: 00000000000 c000 0 0 0 00 0
- M 4
CA1 G
range of apparent distortion. This effect was investigated by obtaining time dis-
tributions of distortion using several sweep value sets as standards. One effect
observed was the creation of a narrow vacant zone of distortion values that appears
at times in the correlation plots. Similar vacant zones were also caused by digitiz-
ing status programmring error. Each sweep was normalized by subtracting it from the
standard sweep. From this set of differences, a minimum value was selected from the
differences corresponding to the two edge frequencies (67.7 and 72.9 MHz). A maximum
value was obtained from the other difference values for the sweep. The difference
between maximum and minimum was then divided by the corresponding differences in
frequencies to yield the distortion in dB/MHz.
Minimum values of RSL (Table 5) were collected during each spectral density
sweep period. In order to convert RSL into fade depth below the long-term median,
each RSL value was subtracted from the long-term median RSL value.
If 3-level error events (format violations) or reframe events occurred during
a sweep period, the events were counted and stored in correspondence with the
sweep. The receiver-on-line status (primary or diversity receiver) was also stored
with each sweep. The status at the end of the sweep was stored as being the
receiver-on-line. From these values of distortion, fade depth, 3-level-error
events, and reframe events, the distributions and correlation functions were
obtained which are presented in the results section.
5. RESULTS
Approximately 50 hours of recordings were returned to the ITS Boulder labora-
tories for analysis. Of the 50 hours, approximately half were for quiet hours
showing no significant distortion or fading. Since careful editing of each channel
was necessary after digitizing the data in order to eliminate data anomalies caused
by software and hardware, 15 hours of data analysis are presented here. This data
includes the hours of deepest multipath fading observed during the total recording
period. The results are presented in terms of time distributions of amplitude
distortion, flat fading, distortion event durations and intervals between events.
Four hours of the flat fading and amplitude distortion are presented as time func-
tions. Correlation plots of flat fading and amplitude distortion are also presented.
All of this information is presented in Figures 6 through 20 and Tables 6 through
41.
The results are presented in three sections. Each section represents five hours
of data. The first section is for a period of heavy fading; the second section is
for a period of moderate fading and the third is for a period of light fading. Onc
17
{ (Prtn mc.-y Roc.,6.vr
9 2 3 .5872 4 1 16 17 10 19 U 21 U 21 24 25 M2 W 1 3W
6~ ~ ~~~1 11 31 I 2 23 14 1. a7 1. 6 9 221 22-2 24 25 221 193 1 3
I
N, 6
D..-It
-I, ! L I I .6 6 .. ...
9 2 3 95 69 81 1231 156171 19 1M21 22 23 24 n20222 3 W
Figure 6. Fade depth distortion comparison for time period of29 May 1980, 0100-0200, lt. Venda to Nlt. Corna.
18
L 0
-l-
OtvraY R.- Iv
1 2 2 A S 0 7 9 9 0I 2 19 14 15 1- 17 IS IS SO 21U 0 0 * N
Ot-Ir.ty R ... I-
a 1 2 3 4 9 * 0 If It 12 12 1. IS 10 17 19 19 MI 22 N 324 n S N SIt N
Sasmple N~mbwr Hunrdreds )
Figure 7. Fade depth distortion comparison for time period of29 May 1980, 0200-0300, Mt. Venda to Mt. Comna.
19
~ ~4)
1-- '4.-
(1)
.44-
Lnn
L.. .- A L -- lr.
too
(epa) GiC-- 04S
20(J
-4-
(am)
0 4
IC a
4, Ina)-
VT~T1 21
21 -e
1 P2 3 r to I I n Z ZS n ? N . 1
So
* 2 3 4 5 5 7 9 A it 44 2 134 15 10 17 10 40 252 r a)2 2 29 P M n)411 2
-. 5
. .... -.... 3 5 a 9 9 9 44 42 43 4 5 49 41 49 49 2 21 2 23 2 25 .f 2' M 20 30 31 3
,e -..-. N H
S~'.pIe Nubl C H.rd~odI )
Figure 10. Fade Depth Distortion Comparison (Slopes only) for
time period of 29 May 1980, 0100-0200, Mt. Venda
to Mt. Corna.
22
7 24 .... ..
to 11 I Q 9 2 . 11 2 L? 1.0.
9 2. 3 4 S4 0 7 0 4 19 11 12 I 43 IS 4 17 1 1 n 2 t 22 22 2 2 25 24 27 20 20 3 SS4 3
Pr o.ty R.o.. vor
... ... . 0 It 1 5 4 2 13 4 15 10 17 46 40 20 24 22 23 2024 5 1 f 2 O 2
.2G
" ; . . . .. . . . . . . . . . ...... R e. ...r
2 I 2 l 3 4 5 0 6 3 40 44 42 43 43 14 40 47 46 40 1 24 2 2 23 23 25 5 27 0 0 U 34 32I
0 n
0
3.
-Ug
LLrot R-o.4v-
a 4 2 3 3 5 6 7 3 0 43 $1 12 43 43 I4 18 17 14 0 M 21 24 f n 2 n 20 IT 07 U It U
Somple Nwmber < Hnd 7 .de )
Figure 11. Fade Depth Distortion Comparison (Slopes only) fortime period of 29 May 1980, 0200-0300, Mt. Vendato Mt. Coma.
23
I.I
2.I
1.5
N
E
I
,*-4
MO0 Vo 44) -- 51-Lo4.)
o 1-1.5
SPr-iniary Rec i ver
-2. I ..L . 1--L . _L± L-I . . a . t u t- aJ- J., .t t-.t_.±. I .J L -L.L . . J .-- J-J J LI_ J-L.-.-- ..J.-A-- --4 -4 4 I-29 -1i 0 10 26 36 46 5d8
2.0
1.5
'N i
~.. ..............:......... .... ,........
0
4.)
0
-1.5 .
Diversity Receiver
-2.0 -1t . - t i i A JIt. t 1 = .- A.L i.- .L.L.k.... .a. . . ±-20 -1 0t 20 36 4095
Fade Depth (dB)
Figure 12. Correlation of spectrum amplitude distortion to fade depth for
29 May 1980, 0100-0200, Mt. Venda to Mit. Corna.
24
- -- _-. -_
2.11
N .N[ i
N Im .s[.
"Io
Primary Receiver
-2. ,, ~ ..L.-.'a..!.L. 1.. ..LLL .j.... .. ::;.:.." ..., . .I,.
-20 -to 0 Is 23 36. 405
21
1:5
N i
MU
'") -1
Z)
I
4.)
D Piversity Rec ivor
-2.n L L. . J,,,J.L .A.LL.±-.. JLL. 4.L-Jt. .LJ.... L C...L .Lt_, J, .. L LA L, . . j_ g t ,k
-2. -A I II 2 3 99 4I
Fade Depth (dB)
Figure 13. Correlation of spectrum amplitude distortion to fade depth for29 May 1980, 0200-0300, Mt. Venda to Mt. Comna.
25
-- -
N
CO .2 I
N I.
o .
0
Primary Receiver-
2.8La
0
-2.DiOversity~ Receiver-
............................................. ... ............-23 -13 8 to. . .. ..
Fade Depth (C.18)Figure 14. Correlation of Spectrum amplitude distortion to fade depth for
28-29 May 1980, 2300-0400, Mt. Venda to fit. Comna.
26
2.9 -
1.5 -
N i.uT
M .5_
L0
(0
'-4
-1.5 . Pg-i mary. Receiver-
-2.0 L . . . .a i .. . . . . . .. . . . . . .a . a .a a . n _j p / .., . .. . J..J .. .-21 -10 0 to 29 36 48 58
2.0 .
1.5 _
NN iLe.T
MT .5.T
"0 .5
0
.-
L0
4-10
.5 Diver-eit.y Receiver
-2.0 L , . . . . . . . . . , , . . . . . . . . . ,-, 1 1 a.LL a a a t ,J. - L -L .L a. a LL J. A-20 -10 0 1 29 39 41 59
F de Depth (dB)
Figure 15. Correlation of spectrum amplitude distortion(slopes only) to fade depth for 28-29 May 1980.2300-0400. Mt. Venda to Mt. Corna.
27
I '
1 i~~.. .. ..... .......... .... ..
2 3 33 3 3 :7 36 I 33 3 0 3 2 3 3 2 2 2 20 3 33 3
Figure 16. Fade Depth Distortion Comparison (Slopes only) fortime period of 13 May 1980, 0100-0200, Mt. Vendato [it. Corna.
28
.1
.1
a " .j
' 3 t 33 22 2 2. 24 * " M 2 3 1 2
..... .2 .. .,.
Figure 17. Fade Depth Distortion Comparison (Slopes only) fortime period of 13 May 1980, 0200-0300, Mt. Vendato Mt. Corna.
29
201 Z5
-0
0.0
N0 0 03 .,
U 30
2.3
1.5
N i.@
m .5-
0
L0
4
1.5
Pr- mar-y Reei ver-
-2.03 .. . .. . . I .. .. L ...... ...... I...... . . .. L
-.620 -1 3 1 3 2 3 3 3 4 3 5 3
1.5
N .
IL
0
L0
o 1.51
OivereSity Receiver
-4.0 L I -L. . . . ..L. . .. . . . . . L-20 -13 3 a3 23 3435
Fade Depth (de)
Figure 19. Correlation of spectrum amplitude distortion to fade depth for13-14 flay 1980, 2200-0300, Mt. Venda to Mt. Comna.
31
Li
0
C
0
L 30
0.
P iary io ofiimerBlo tve
Fiue2. Cmltv(itrio9itiuin(lpsol)
.4.)Cona 282 a 90 20-40hus
132
MH21 Mo iit ':tl 10086!~* m -l.l... - . -2 -
h0 L) 4A00SO
S- 41i c",,Lj-g
0I
4L, ,I,-) X 3 2,,,-lo g o co 00 0
0 L h i iiloiii
,. > --t +.o ,.,+
C 0.2 -,000
01 0
1 a ,
I<. . -CL-----
.I1
*!- ----- -- - 4 - --C
."k.
.0 m
..- 0 S... * ...
04~10
C'U'fl
oneaC;li-ii a
Pal *son-aso
CD S-e
ON UE1 I
CLA
C 4)
jj~*£/ ~-a- 00050oo ;* 00000 @ @ O
0
*,- Q)co
C IA
- Co 0
(o~ C ' a.L J+L!
7 0 . . . . . . . . .
*r+JCl O96
.04
ji.12-
-
L0 4Jgin, ~ ex
4j64 1* -4M 0 .W_ Ln I I
SE
-0
09
Sb. 4ot
0a A
to C)
4)L
0000
00*--
1;11
U3
3 NON
oo0-1 C.
3- "
12~~~ *N*wI.0- *Coco*-. tfafHiHII 00
* h.0
4. s-~ll5f tol~ce~. r
L 0X.4. N46OO flove-000 Oo*OoeoooS
S- (U
4-3
4flt.5
>0
C) 0
S- 3 3 N- oem
u- -u V)
363
S. 6m
x
*-N
00 4J. (Ae A, a~ . -~ tN
to C .0 000 *0@0
Z i
(A X 4aNSj
in 0q 0~@,@ mnn0u o.e,4J 41li0 4 -laao. - ~ O
(A U
o u Zo 041UC 0 * l0.00 *t-4fgl o u...
> ~ -*.e@ oe * e -- Ue~oo
Iw CL-4- D W.F_( DW00
cm .j 16! **4.4.b) V) 5
C;Z4' .
*s(D
000
CU~ 37
c U
000000 . 00 NM M :t soo
0 (a F. ift
I. Is 000 00*.****
a--
cm 0-o a..!
00 ccC
4-)h.052.c
4jU
o ML ai
> ~ IOM 'co
4- R3.CDuO)k0l 00KUS0KHKKK
C- -. HKCK .H H fl -4C1C*
arag A~ .
o u C) L
~~cogG.-
000
4.)4J
C U38
M V
C ~ be.- 04 . C
=, =
06S- %I-
4b4J CU ~ la -0000-N*U> -031*06
0 m*
M4) 0 Q %0--4-
CM. u) S... teI. 0 ~ z,.A 0e@ *
> S-4
0
0 -4J )4- 6l)
S-w 00a66. 0 30. 0 00 Co" = a a .
CM01
4-3U4-W
C =S
> 6r
in~ co V
1.0
1. 0
39
. . . . . . . .
s-a ltj a t -C3 o0Nt N 21 4M a.0% DDt * tc c l
r- 00 cc0M-,-O-9 - cooo o %0 01 V7 0
- ~ ~ ~ NI0 C410 0C 0o~o~~ce
.O0C4 U U
060
F a~o o3 a
0 0 l
- 0 a
0c.e -C4I) 0M 1 O
.ou (A-cqp
0 3VCl
4-oJ -- -
4--l
.0 to0.xu.I-u 0 > laoV Cs- s- ) 04 IU
41 ( a
4- -- f
C A ar.
0 U a
00 0c
.2
- 01
40
. . . . . . . .
0 U
NL 00 03' 0 .O cn
o1 9. --@'. cc
00
4AW W -)-A
o cm -0
V) 0-0
>~ je.OU-~ 000
4j (D CCO
40..
o S*l
N-o
'41
16004
- r. iz~000oooo~ o0xm~co~
414
to ~ ~ -- $ *ael 3'
im ~ a 0.rai.
4J d.) . ;90
= 41
U S- ao 6 0--0- , c
c-- 0 iCI
+4-) an0-
S- c 0 co
4-) ';- C 0 0 . 0
-F. he= -
4- di -OCo 3:
4J ecoo
Cn 2)C) e043 CaO
10 Ln emW-
s- S.. i Cal
cmm
I-0 4) bb
- ; ai 0 3t C 010
- C
1. C ' a'
42
0
(A -0S-. Soo
to am ;.
4A 04
S- u
LCO
.l flJ a *.. cm 0- A- Io
S- W10J N.
Am-0
4- .~0 om
4J LZ
-I-'~~ 0 8-
- -
~~00
U 9
o~n l43
u o aGo 000 . t.,, 0,, to !mto) l
- . ,,a" ,,. m+I2L+~IC-0 1. U Coc b
,..~ ~ r..J . *''
S-
-0
P 0 e+' -C1 0 .,. te
. .
t00 a)
S'- (a
> .
-)4> 0 ovol10 I
a) o o ,10 •
,0 c 0 '
C 44-043V e 'aC
- -o 444
13- 0
(v 4nl0) 000wCoc4) Cj - 0%ato =z ~0 20
4)
> on
Cma 0 - -c in n-eIa4
U .
o 'aa
00 00
44
cu .
to = em r 4
0C0
cm cm w 0-03te e CAtoo
a
LJS
o, co
C 4-
4.J 4.. (710
0~ ~ ~ -L Utom 7
4- 4) -Cmo -4ii440S-~~* 0070
0 Va
>oco04
o 4-).. 22: bon~
c-a Ca% l c- JVC
h o ueooU
>~c 0* 0.~ *-9
@A45
aa V-
,Cl-a S_o -0sssenq....-oeoe
.a @0 00 .. .. . .. . .. .004. .h. a_ w ss o
4~J 4) a 00
:C040400 0 £30000pn *@0#IS-0.a-ew1-0x 00
00
L. a=SC
a., Cj V)
00 000
InI
a a6o .a a*U-00000 00P£3-000 4£4C40006
-- 1-46
h ,.
S-
L .t.t it .. tl!.2tt!t55 I-- -II IIIII -- III
Unee
CL.
*i.. 0 0 0
- 41
0&
to--41-.-. ..-- *.... m-- * 11U 0 o-94Na,9o -cl ouo,,o
4 -3 FS
0 * -
D . Jig
• 1.. #l .
04) co 55
If4J L.
00 -
I
000 '-"*04*0
i'-
.. i
o :lb
I/I I-) .! *4
>, "** -"
Ui. I -
I--,
I - I-
I A I
47
HUS, SCSI=QI I!ll lO
I i
:A
°* . -.:::g--"o
b 4 O O 00.I *- - ;4OO *000O40POO
ta
al 00"l00e
es
0.f
r_ 4 1
4-o 0- U44 V-0 4.0 c- 0
L. 0i
0 4j
L gl".00
CD. 7
,A.
0 C
>. CJ
to U
f41
L> 4n
O . o00000 -- 4000 ~t+O.0
A 8~ A 0C
L0
s~g~48
A
." .. ;4 O " ~ O *49900-054 :J Iiliiiiiii Ii IiiUh1I m . 0 @iii " "i el
21'vo OUR
(Lal d -- es
., 0 t4.-
.- I A
0.04.)4
00
49
@ 0 . ......... n.' .: e.e. .e 9.ee .e ..e . .:iiBM 0fe.n *1 ..- 022f
0 c 0
h 4B
°.
S° cli 11 i ti !i i
(AA
. W.
0 -
LI.* **in*. r-- .0
•n -- i 4 J!.
4j S-
to 00 4.;-(71 -1 Aa i
I.I
S. .. . . .. .. .. . . . . . . . .
ot 0
-- B
I-.- 06000
.. .. 0 50
. ..... .... .... ".....i i ii ii
. -. !*
I&AA1 6=3r
LL..
CAu
. .. . .
c o.
o
.A2 W
G) +ooooooo 00oooooooo ooooooooo
L. *~
4-x,*i" -&
c.m CL°
* a (A0 0 00 0 0 0
-)0 .. ." *
0
4 -- CD S_ 0:
(- 00 V"0 0
4.1.
A 16.43 6
5 0O
Il
0 0 3 . .,.;... . .. . .e . . :; .;-
-eltl0
LIE t': i.i i ,"0" I
4 ~~E J1
4A1
.0
O' Nl 1 o F02% N0 03I
09-3Oe Sa*
4-) 0) 04
a.J S-. cc;q
(A 06
Ln LL ul
'0 X0sU
41 01 -f"
.4. ..) 'to (A - a
0 4-) _ ;
LLIL ~ 41 41
o *co
o~0 13 u--Ai .coi-e
4152
4e04
.h~. . .........COC C~
cr.
doU
CL= , q1 3ot o IW
*,- L -
I .- = *U*aa
-- A- 1. 00 0
C~04
o -U
OW 4) 090 co oN.
. -
a ; u0 0
go4 U
0 2fCc
05
LmU
~. .
00
* a Ile
:. 0300N O- - - - - - - - -*0 4)- - ;w
---~m
4- J
im CL 4
CL
o2 %Dc o;0 o 2% k ;e~ ~UC 10:; a01 * 4
t.LJ Cs
0~ aW*e (A C'.j4-J 4.)
(A£ L P.- f-
..- x "- see as
4-3 -. h0 bogn0
an- .m
- CMa i
4J () C Q 0
U,
L O . oo
. wa.
00
(-54
of a•
b.100004
o ~ ~ .,- ... .........16-- X
too no""'
e as
-------
I10 -0
I::!
0 dwo
4- .Ln 0 i"
to ~ ~ (U D
• .. ,
cU 0
'- 0 o
UJ - -u t '000 a A!0
C 4- @u 0 0O0g1-
16@ OOt.
4 -1
- 41 U
41 4. C 1 C
106@UColl
0)~
$5
4.) 41 C
S_ C
4-a.
)flj up U-it;
>a 00a30
:3.v.. 4J
0 0
0 4)
C
C4 4. 4
4J03C O03
~- cca.M
.0 o co
a56
I z
1 I Il N l 0 I
CL.C
C) ,; ...
'-no
. . . . . . . . . . . .. . . . .... 4.
C1 U. K
U)0
>1 41
.0 v
CAICJDU "~O
C4
Ix- -
01 0 . .o- l'.-~l3' - ChC al
0 Mccoo
.. .
or -
C ClIII C
..> .
CII NDo I w% 9%
I ~ ~ ~ ~ ~ r I IC1 1NN*1 19
-1 - ----O
@i ... a57
4 1 14N41
coco
C) mo
0a1
p I = OW0C0NU'N0ONNOC.CI: 0N: U':N WNNOCCNC: 40WN: M00Nr.U''.:.iNO:00
I% CL aU 'M C 6C0 C6U CC C C O ' U U U C ' ' ' CL CU UC ' ' ' ' t O C 0.CLI 9
i*iC* i" i * * *I * * I* i " I i I"i i " I I i i"
In0
0.- -
• lN N.. . . . . . . . . . . . . . . . . . . . . . . . . N. . . . . . . . . . . . . . . . . . . . . . . .
L _ *---------- -- --- --- --- --- - . . llr=
V) 0C .
C ti E ,Z
~~~~~~~0~~~~~ -00 0 -00 0 0 0 0 0 -OO O-0 0 0
-0 0M- O0-0 00-M0-
oI2 In 1, mII I 0 na Um 0 do 0. 0 Pal ,Ic '', 0 ,0
~~~~ 9NNN NN 4N.,; N N N N N NN N N N N N MMNNNNNNN MMMM; ,
,.~ ..~ MMM NMM MMM NMM MMM r.M MMM #~. M~r .MM CCC M M NSC C(OwlM
-- - - - - -
o5
ai @%o mfN~ mN m ~N 1N 10W~~ N a0 0Im m w
* ts
I4
CA 1 .' 0 N. NN N . .- . . . .4. .*.
:41
LOu
V -
WI.
S- ~ ~ Ij 4c 40
41 Z co 21O O M C N N N Q
0 >
"0 I... ... W
4-) cc I 0
=n 0.- . . . . . . . .
f- f.I M " m r .
-- , - -
- - - - - -- - - - -
~ 59
no
o a
.V c
C!
LO 0 . rm
ca ccWc)c n 4c nmwmmm4vmmmMw cmc~cN #4a00 0 m
V2 C
T 0
-0 a Nca:
4J ,a ,a ,a ,a ,a ,a ,a , ~~,a ,a ,a ,a ,a -a ,a , ,a ,a ,a ,a ,a ,a ,a ,a
iC 00 0 C!00000a a 1.a0 k 'c 00 00
WC
a-r, a-
ot oc=zw rr- o - cww r w
- '00a-a-a ~ ~a-.4a.W-a-a'a'a-~..ao..4a'au'a'aa'a'4, ~ ~ ~ ~ ~ ~ I -. .. . . . . . . . . . . . . .~~~~~~~~~w 0-9fI aaaaaaa W a ----- w 'aa0 -aa-0aaa- WW WW WW
- - - - - - - - - - - - - - - -
C ~C*~ 60
As
Aa,4-d
S-ie
* 0-
CL00
. U4J
00 U 000 pamciO 0 Vgf00OO0000
4jOU0 0!~i~i iiII
4JA C;. **3
0 4j0 0 gfc...0 Q.***
-003UC0(L
.... .. . .. . ......
0Q~96
~~61
0a I-
0 a-
0 39
4J I
CLIn '
1. I
a 9
al62
I ro . ... .. . oo~e000 @0o~~o @oeo~.e.oeo
loa
0:b
o s.. .
-041-
.,-.
= ad ,: 0..
..-.S- C
4o 0- AI
CL a4 o -V)
>0
6-
00 -; 4i0 0-
- k3 000 00 46#000
bo
o a
.0S- .0 * *
4J 4eJ. mO*U*
+1 ab.-
m>o w -
~fl) - ~ @-00 0 l- 0000 eo a 0006
4.64
0N@000000 .N 0300 000000
vism
400e
bC 00 0000 000
al4&
to
faa O .... uf .... .. ..
a
4.J 16
~ 65
hoe00o@S O oo
0 S(A., , S.:
4-. 1 .
aj
o CD.U. 3
t33
W V)
,.) I--.
tia g,
= ooo.** ooo* e-oo~e.+o { -**0.00$
Iii
S.1
! -6
-- . . . . . . . . . .-. . . . . . . .0 "-.... . .*+ .. .
C..) -CA - IN-
i + I ,
- S6
6 -0% 3O c00000 k0 4
I. :: 0 r ! 04 t. ;:, ; 0a me ' 0 *. o-%D-- -atc
4- a- Not N N of--f- -*@-
LImm
L - C
> 0~a-
4- = Ug
L4- 0
00411o *0
4. . . .. .i .4 ~ ? . . . . . c
4--As 4--.JSgo -
r_ > *
00 rd c 4
C C)f >
S.. = 00.(4-J 4)0
.- (U 3 u
o U
'-a a 0N oC 4
o.0
I 67
8
o' O.. I OI~mJi+'II€C~lCI€ ¢4 O@,.OO00000000- . . . . . . . . . . . .. . . . . . . . . . . ..-- - - - - - -
S- to a. ajcqqD~oo ~ OME .
.CU I
S- 4 go
S.. ..'-O a.m> C -
*- .-. . .
-- J4>
'L .,- e
S-
U aa
4U 4-O
4J >
o1 L aJ
CD a.4-) 48 0)
000
d) 0 a -a .q-O hi OO.-
m ".. a 0 O - coae o;N0' lI(U 0. I 048* 0
r4 I
00
48~68
U . g o .............. . . . . . .10
. *0
> 410 0
0, ON
4. o co pp,4.
(n -
oUo
0 .(-,- >. 'cI-
• - u cc r.
CC1 03 -
toa - 10 co 02
S- a) a la .
0@O
oo *€
4C 4.)
4-
64J
iU
.6 L>0 0-~ *0 1 k)
4J4.0 W a
(4-j 4)
C~CO CIs 10Am
0 0.- 000 cm> C '4* 4
4.) LIO @ - 0k~t. 0. 69
OH Ia
-- - - - - - - - -
- . is W.
C V)- .0 1
.V- ;p
4uJ X; C -a v o Dolco %V C c C
4) - C: 0-0001Nv
0 ar
4- >)
4)c. >d 4.
S.- d) a
44-) .d)
4-
4.)
4-J t UJm
4) O 4 CU~- -
4 C4 *to2 %oNCI
.1 a -fC 0, o%0e -ccC oio - 03 o
456
- 0,0
-. co
s. o cc
C^ 00
07
%a a *. % % D D%
4, C
*0
0 4,
Ini 8
= _: 2 w c
00 I C-. 01
W. 0u
En C6
o .71
14 4
o -'.-o ~ .' - -- er I I I~ I I I I I I I
410
0- ~0,,
'a
0 4
L31 1 1 N 1( 1 1 1 I
4n) >-- I,
o '
I ~ ~ ~ ~ ~ C CN..- C!04 (.'ION .Iv
3= 0 I W4..INn
2 w0 EO 01ONNNN0 1 10 1' 1010 10 'CID D I
.~146
t2 --
~7
0i
4. C C! ei 1 a N1o 0 % 0
45-. ON'O DU N 0 'a 10 a'
- 04
~r4 0
C6 Q 0.. 0 a 0 C6.'- a0. Ix NCLa 0 .960
s-. CU
4-J 0 O.J
04 0 4 0 000U0 00 - 0 - 0
413
to4,
0> --,
----
(flW 73
4zg
- b.
cm
-8 -8*0i -
vi-4 0444 I r .-4 .~ l . 49...4
u .0
(a V,
1j -Z I -
0 0 0 W 0 N' W0 '8 N0 '8 9c 'a m 4'4 .9n.-. 4
NIII
20 10I
00
-. - -80---- ---
CL .c
411 c..~ W 4 4 4 4 1
ILI 0 0
.0 r
.I I. U 0
.a. . . . . . . . .
10 W I1 a
U- 1420 1 -N I-~ 0 0 0 0 00 0 0 0 0
c 01
-- - -- - -------
1974
I
0A
- ~ 0t0N 00 0--0 Li'i 000Cl000
Do 2 ,i03li9t 00 000I g o oooo~oooeeooeo.oooo ooqoooOOoQ
_ 0 0 . ....
"ti .010W
1-;;:-. ' 00 00
"" 0
* - S. r-ee'-. =.--.^l0
(.*,Iu=" "
CIi .....o~e o, eo..o... 0011.....000S--
ets 1 a
a
W ;3 0 00 oo gooO 000000 0
0 0 6ss
o-) 4-; W.Ldi~le 0
.- 4.
s- ~ ~ a)(
00 !,' 04-) C1 CD.e0
T C,.
i (nl
M too -- ---
- 0
eaa
---- 0 - -- 0 -- -0
-II
C-75
ChC
au 0j
4) 0 L. 0 ) 041A- Q: C@NO)0 Oa a
::
L
.~ 0::= e-S_. 10 -- a0 I
o c Das 1. -
- >4- W '
0 L4LJE
4)LJ
-a.a.43 'Up.0j
41 03 ~ O - O~~l.41.O -O j' 1-
Q) 3': oo~
C0C
03. 10 -C)D ln - l O I 3-
> L. C0 l -C
L. Go to a vo loC
0N c
Oi
76
m VI=
>c--o -------
) 4,- .0,
-S-
c S- 1..j 0 a 2
. ? ~ ~ ~ ~ ~ ~ ~ 0 0o0 m 0 000000030% elcvC3%C4-J -; r. C4-r oo* e 0 00 0 0 00
4-- >-, U-" . ....
4) 1: UO
I I 3 0
c- > 0 00-a - 7
4-J 00%con 0
o = c .4; pl
0 1A -
41 ; . 0 x
a 5 .-. 00 0 04k
0 0l OC4- 0: M 0N~
4-) W
V Wafl O-.'(4 0~CC 03rca00 0a c
.- aj
scm (-
OmO
-~ 04 0 CIOOOS'I0NC
'77
CL anm t.
+j i.
en
gl
*i= '.4t04
0
4-a Mo. 0
0 00 A
re.0
2Z.
780~
o
part of the data is in terms of amplitude distortion slope across the bond %617
to 72.9 MHz). The slopes were calculated by obtaining the difference of spectral
density amplitude corresponding to the two end points of the IF band and dividing
that difference by the difference in the corresponding IF frequencies. This anal-
ysis shows most of the distortion (null information is lost) while eliminating much
of the test instrumentation noise from the data. For this reason, the heavy fading
period data is presented in both ways (maximum distortion measured during a sweep
and distortion measured across the band).
5.1 Heavy Fading Period
The first results presented are the time functions of flat fading and distor-
tion obtained for early morning hours from the Venda-Corna receivers (Figures 6 and
7). From these figures, it appears that nulls in the frequency spectrum often
correspond to the flat fading nulls since the slope of the distortion in dB/MHz
often changes sign at the sample corresponding to the flat fading null (Figures 6,
7 and 8). Figure 8 shows a detailed example of this change in sign. The null
itself, however, occurred between sweeps.
An important observation obtained from these functions is that while some of the
distortion seems to be of a continuous nature, another component seems to be discon-
tinuous (at least on a second by second basis). This characteristic, as well as the
fact that the selected sample sweep affects the average amplitude of this "discon-
tinuous" distortion, leads one to suspect that the more slowly varying short-term
average better represents the amplitude distortion characteristic of the path as a
function of time. A large number of analog sample playbacks of the recorded
envelopes of the IF spectral density functions show noise on these functions (Fig-
ure 9). For this reason, the data was reanalyzed, ignoring nulls, measuring only
slopes across the IF band. As seen from Figures 10 and 11, a function is obtained
which is very much like the short-term average of distortion values based on maximum
slopes (Figures 6 and 7).
From Figures 6, 7, 10, and 11, it becomes clear that diversity switching has the
potential for being very effective in counteracting the effect of slope distortion aswell as nulls. Therefore, it is very important that the switching threshold power
ratio of the diversity switch be kept small and that switching be permitted at rela-tively high levels of received signal level which is not now the case for the DEB
receivers. Table 6, 14, 26, and 38 show quantitatively the amount by which the DEB
I diversity system reduced the distortion of the receiver-on-line compared to the
distortion from a single receiver. This reduction can best be observed by comparing
79
the distortion distributions for the primary and diversity receivers with the one
for the receiver-on-line (especially the column for number of samples greater than
or equal to particular absolute value of distortion). This comparison shows that
the DEB I diversity system provides fairly good protection against large values of
distortion but very little protection against the smaller values of distortion.
Using information from the radio manuals, the DEB I diversity system using
the FRC-162 radios has the following characteristics:
1. The switching threshold power ratio (hysteresis) =5 dB.
2. The system will not switch radios until the receiver-on-line fades
down to approximately -65 dBm (about 30 dB below the median level)
and then only if the receiver-off-line is at a 5 dB higher level.
3. The switch response time (the period between an event which will cause
switching and the time until the switch starts to react) is approximately
2 milliseconds. (This information was obtained from Mr. James Hefner of
Collins Corporation.)
Fade levels corresponding to the various switching events in Table 22 confirm
operation roughly consistent with the combiner characteristics given in 1 and 2
above. Consideration of the combiner properties and careful observation of Fig-
ures 10 and 11, using a straight edge to line up concurrent fading and distortion
events, shows how little of the potential diversity improvement of on-line-receiver
distortion avoidance is presentlY being re;alized. It must be stated here, however,
that for this system (12.6 Mb/s) the additional distortion avoidance hardly seems
needed considering the very few frame loss events shown in Table 22. Figures 10 and
11 show that the distortion occurs in events (similar to fades). If the combiner
switches in a manner such that the receiver-on-line is always the one with the great-
est RSL, the effect on traffic of many of the distortion events will be eliminated
entirely and at least parts of the others will be reduced in magnitude. If the
receiver-on-line is to be the one with the greatest RSL, certain conditions are
necessary: switching must not be disabled at high RSL levels, the switching thresh-
old power ratio must be close to 0 dB, and the switch response time must be very
short (a few milliseconds).
Figures 12, 13 and 14 are plots of distortion values at the fade depth at
which they occurred. These three figures include nulls and noise with a small band
of points at relatively high signal levels that lie above a small "forbidden" zone.
The points that lie above the "forbidden" zone and at high RSLs were introduced by
a digitizing error introduced by an optional audio signal used by the digitizer
operator to determine that data were being digitized. These figures as well as
80
Table 25 show that the largest values of distortion are associated with the deep
RSL fading nulls. Figure 15 is the same type of diagram as Figures 12, 13 and
14 but shows slopes only. It is clear from Figure 15 that the distribution of
distortion amplitudes widens rapidly for decreasing signal levels. It is alsoclear that for this 5-hour period the distortion slopes are primarily negative
and generally increase in amplitude as signal level decreases. The reason for the
preponderance of negative slopes is not clear. We are confident that it is a
measured effect on this path and not the result of a bias introduced by test equip-
ment operation or the digitization of the analog data. A preferred atmospheric
structure seems to exist during multipath conditions causing a preferred range ofamplitudes and rf phase delays.
The data set for May 28, 1980, 2300 hours to May 29, 1980, 0400 hours is also
presented in terms of cumulative distributions of distortion values, fade depths,
distortion durations and intervals between distortion events. Tables 6 through
13 present these distributions with noise and nulls. Tables 14 through 21 present
the data for slopes only.
Selected sets of values for this period are presented in terms of maximum
distortion. The first set (Table 22) lists the sweeps during which events occur.
The event types are a change-of receiver-on-line status, a reframe event, or a 3-
level-error event. A 3-level-error event was recorded only if a minimum of 0.002
seconds had passed since the start of the previous one. The pulse had to be
stretched to this period so that the recording electronics would be able to detect
it. The same conditions applied to reframe events. The second set of values
(Tables 23 and 24) list the sweeps during which distortion exceeded 0.5 dB/MHz. The
third set of values (Table 25) is a presentation of sweeps during which flat fading
exceeded 35 dB.] 5.2 Moderate Fading PeriodMuch the same type of results presentation is made for the moderate fading
period (May 12, 1980, 2300 hours to May 13, 1980, 0400 hours) as was made for the
heavy fading period. Two hours of time functions of flat fading and distortion are
presented (Figures 16 and 17). These figures presented distortion in terms of
slope across the IF band. The same is true for Figure 18 which is a correlation
plot of distortion in terms of flat fading. These three figures are consistent with
the h eavy fading results. Distortion is somewhat less because there is less flat
fading. These three figures show that the flat fading channels of the recording
system become unstable at high signal levels. This is particularly apparent in the
plots for the diversity receiver in Figures 16 and 18. The narrow vertical strips
81
in Figure 18 are caused by the flat fading digitizing granularity. These strips
also appear on the other correlation plots and vary somewhat depending on the
slope of the particular calibration.
Tables 26 through 33 are time distributions of event durations, distortion
values, and fade depths. In the distributions related to distortion, distortion
is calculated in terms of slope across the IF band. Tables 34 through 37 are
selected sets of values for the moderate fading period and are presented in terms
of maximum distortion. The data set types are the same as those described in the
heavy fading section.
5.3 Light Fading PeriodThe light fading period was from 13 May 1980, 2000 hours to 14 May 1980,
0300 hours. The results for this period are presented in terms of maximum distor-
tion within the band instead of slope of distortion across the band. The manner
in which the distortion is presented makes little difference since very little
distortion was observed. Figure 19 is a correlation plot of distortion and flat
fading depth. As was the case for the heavy and moderate fading periods, distortion
slopes were predominately negative and the distribution of distortion values widens
as flat fading depth becomes greater. Tables 38, 39, and 40 are the time distribu-
tions of distortion values, flat fade depths and event durations. Table 41 indi-
cates that there were no 3-level-errors or reframe events during this period but
that there were receiver switching events.
6. CONCLUSIONS
1. Consideration of the data in Figures 15, 18, and 19 shows a consistent statis-
tical relationship between the depth of RSL fades and the amount of amplitude
distortion. This relationship is one in which the probability of large distortion
values increases with increasing fading depth. This is of particular interest
since the data in Figure 15 shows depressed median signal levels (approximately
6 to 10 dB) on both primary and diversity receivers. (See Figures 10 and 11.) This
consistency leads one to conclude that a useful relationship between calculated or
measured estimates of the time distributions of multipath fading and distributions
of amplitude distortion can be made. This conclusion holds for relatively narrow
band systems (up to approximately 50 MHz) since ultimately, for very wideband sys-
tems, there will not be any significant RSL multipath fading. This relationshipshould contain the same type of occurrence factor as is used in the multipath fading
prediction models. The nature of the distribution within a multipath fading period
for the 8 GHz band and for distortion slopes less than 0.6 dB/MHz is considered to
82
be log linear (see Table 14). A rough cut at an empirically derived relationship
is:
Pm= Q() (d3" 5 ) 1056" (1)
This expression uses the occurrence factor suggested by Morita (1970), page 810.
The occurrence factor is less than 1, such that:
Q 2 (d35 (2)
Pm = the fraction of time that the distortion is greater than a given
value of 6 during the worst fading season
6 = the distortion in dB/MHz. (0<6<0.6)
d = the path length in km. (d>50)
f = the frequency in GHz. (l<f<50)
Q = 2 x l0-9 over mountains
Q = 5.1 x 10 average terrain
Q = 3.7 x l0- 7(1/h)0 .5 over water and coastal areas
h = average path height above ground in meters.
An estimated occurrence factor value for a path may be derived from the distribution
of flat fading data obtained during the worst fading month. The estimate is ob-
tained by observing the fraction of time that flat fading exceeds 20 dB during the
worst fading month and then dividing that value by 0.0069. See Hause and Wortendyke
(1979), p. 36. The slope of the log linear distribution, 10-56, whose probability
of occurrence is predicted by the Morita model, was selected from the distortion
(slope) distributions in Table 14. The fraction of time during which greater than
a given absolute value of distortion was observed is plotted for both the diver-
sity and primary receivers (Figure 20). The data population used for selecting
this distribution is small. A much larger data base is required to make an
estimate of the distribution characteristics in which one can legitimately place
a high degree of confidence. With these qualifications in mind, application of
the distortion prediction equation to the Venda-Corna path on 8.3 GHz for the
worst month estimates the single-receiver probability of distortion greater than
0.2 dB/MHz to be: )1.2 23 5 0502
Pm = 5.1(10)'9 ( i) (90.2) 3. 10-5(0.2). (3)
Pm = .0085 of the worst month, but without diversity improvement as mentioned
above.
83
2. We conclude that for DEB I, amplitude distortion for this narrow band(approximately 14 MHz) system is not a significant factor in its performance.(See Tables 22, 34, and 41.)
Out of 15 hours of data, 6 reframe events are recorded (much less than 12
milliseconds outage time). The multiplex has a reframe recovery time less than
0.5 millisecond. All six of these reframe events occurred when the receiver-on-line
indicated a distortion value equal to or less than 0.25 dB/MHz and four of them
occurred when the distortion value was equal to or less than 0.1 dB/MHz, which in-
diates that these events may not be related to the distortion level at the time of
the event since higher distortion levels are observed when no reframe or 3-level-
error events occur (Tables 23 and 24).
3. Diversity switching is a very effective way of reducing the effects of ampli-
tude distortion (Section 5.1). Figures 6, 7, 10, 11, 16, and 17 show how effective
diversity can be if 1) the switch controller is properly aligned; 2) the switching
threshold differential is low enough (<3 dB) and, 3) the switch is set to operate
at high as well as at low signal levels. Diversity switching is especially useful
in reducing the effects of distortion nulls since these nulls correspond strongly
to the RSL fading nulls. See Figures 6, 7, 10, and 11.
4. The distortion changes at a slow enough rate so that the combiner response
time presently available is sufficient (Section 5.1).
5. Future testing needs to be done to advance development of performance predic-
tion models. In order to carry out this testing and development, several things
should be done.
a) Obtain a large population of differential distortion values so that a
better estimate of the characteristics of the short-term time distribution
which pertain to multipath fading periods, can be obtained.
Some of the things which can be done to maximize the amount of this data
while minimizing the test duration are:
1) Obtain data using a radio with a wide frequency spectrum that is
fairly flat across the band.
2) Measure long links (80 km or greater) with sufficient terrain
clearance in reasonably warm, humid climates where the difference
in antenna heights above mean sea level at each end of the path
is small enough such that the absolute value of the antenna
elevation angle is less than 0.8 degree at each end of the path.
(Draft report, Fading on long LOS 8 and 15 GHz paths, L.G. Hause.)
84
7. REFERENCES
Anderson, C.W., S. Barber, and R. Patel (1978), "The Effect of Selective Fadingon Digital Radio", IEEE International Conference on Communication 78,Toronto, Ont.
Barber, S., and C.W. Anderson (1977), "Modulation Considerations for the RD-391 Mb/s Digital Radio", IEEE International Conference on Communications,ICC 77, Chicago, IL.
Dougherty, H.T., and W. Hartman (1977), "Performance of a 400 Mb/s System Overa Line-of-Sight Path", IEEE Trans. Comm. COM-25, No. 4.
Hause, L.G., and D.R. Wortendyke (1979), Automated Digital System EngineeringModel, WTIA-Report-79-18.
Morita, K. (1970), Prediction of Rayleigh fading occurrence probability ofline-of-sight microwave links, Rev. Elect. Comm. Lab., NTT, Japan, 18,11-12.
Smith, D.R., and J. Osterholz (1979), Assessment of Frequency Selective Fading onDCS Transmission System Performance, Defense Communications EngineeringCenter Engineering Publication No. 5-79.
86
3) Measure the amplitude of the whole spectral density function ormany discrete values between the edges of the spectrum (not just
the edge values).
4) Measure both the primary and diversity receiver spectral density
function envelopes.b) Obtain data that will result in a prediction method for estimatingthe reduction of receiver-on-line distortion by diversity performance.
To achieve this end, the following steps are recommnended:1) Obtain time corresponding values of fade depth and distortion
on primary and diversity receivers.
2) Monitor both receivers on a path configured with a typicalspace diversity system.
3) After digitizing the various data and time channels, usecomputer programs to obtain distributions of distortion forthe receiver-on-line for various values of diversityswitching threshold differential.
c) Determine the sensitivity of the radio to distortion, by monitoringreceiver switching, frame error and format violations or events to determinethe susceptibility of the particular radio system to distortion.
85
AD0A099 490 NATIONAL TELECOMMUNICATIONS/INFORMATION AOMINISTRATIO--ETC PIG 17/2.1SELECTIVE FAOING ON 8 0HZ LONG PATHS IN EUROPE.dUIMAR 80 L B HAUSE DCFRO040022
UNCLASSIFIED OCEC-R 040022 NLLcEEE Ij
15
.2
1111I IIII15 I11
MICROCOPY RESOLUTION TEST CHART
NATIONAL BUREAU OF STANDARDS-1963-A
DISTRIBUTION LIST
STANDARD:
RIO0 - 2 R200 - 1R102/RI03/R1O3R - 1 R300 - 1R1O2M - 1 R400 - 1RlO2T - 9 (8 for stock) R500 - 1R104 - 1 R700 - 1R11O - 1 R800 - 1R123 - 1 NCS-TS - 1R124A - 1 (for Archives) 101A - 1
312 - 1
* R1O2T - 12 (Unclassified/Unlimited Distribution)
DCA-EUR - 2 (Defense Communications Agency European AreaATTN: Technical DirectorAPO New York 09131)
DCA-PAC - 3 (CommanderDefense Communications Agency Pacific AreaWheeler AFB, HI 96854)
DCA SW PAC - 1 (Commander, DCA - Southwest Pacific RegionAPO San Francisco 96274)
DCA NW PAC - 1 (Commander, DCA - Northwest Pacific RegionAPO San Francisco 96328)
DCA KOREA - 1 (Chief, DCA - Korea Field OfficeAPO San Francisco 96301)
DCA-Okinawa - 1 (Chief, DCA - Okinawa Field OfficeFPO Seattle 98773)
DCA-Guam - 1 (Chief, DCA - Guam Field OfficeBox 141 NAVCAMS WESTPACFPO San Francisco 96630)
US NAV Shore EE PAC - 1 (U.S. Naval Shore Electronics EngineeringActivity Pacific, Box 130, ATTN: Code 420Pearl Harbor, HI 96860)
1843 EE SQ - 1 (1843 EE Squadron, ATTN: EIEXMHickam AFB, HI 96853)
DCA FO ITALY - 1 (DCA Field Office Italy, Box 166AFSOUTH (NATO), FPO New York 09524
(continued)
DISTRIBUTION LIST (cont'd)
USDCFO - 1 (Unclassified/Unlimited Distribution)(Chief, USDCFO/US NATOAPO New York 90667)
SPECIAL:
RADCATTN: DCLDGriffiss AFB, NY 13442
CommanderESDATTN: DCF-l (Capt. Jack Machuzak)Hanscom Field, MA 01730
CommanderESDATTN: WE (Mr. Owen Cote)Hanscom Field, MA 01730
CommanderAFCCATTN: 1842 EEG (Mr. Dennis Buck)Scott AFB, IL 62225
CommanderU.S. Army Communications CommandATTN: CC-OPS-SM (Mr. Stan Braun)Ft. Huachuca, AZ 85613
COMNAVTELCOMATTN: Code 621B (Mr. Bill Addison)4401 Mass. Ave. N.W.Washington, DC 20390
Officer In ChargeU.S. Naval Shore Electronics Engineering ActivityBox 91FPO San Francisco 96651
Officer In ChargeU.S. Naval Shore Electronics Engineering ActivityBox 5FPO Seattle, WA 98762
CommanderHeadquarters USACEEIAATTN: CCC-TD - CCC-CED-XEM - CCC-OPS-CFt. Huachuca, AZ 85613
DISTRIBUTION LIST (cont'd)
Officer In ChargeU.S. Naval Shore Electronics Engineering ActivityBox 194FPO San Francisco 96630
Mr. Paul HartmannRockwell InternationalP.O. Box 10462Dallas, TX 75207
Mr. Tom GiuffridaBell LabsHolmdel, NJ 07703
Mr. Stephen MatsuuraUSACEEIAATTN: EMEOFt. Huachuca, AZ 85613
Mr. Andrew E. HooperNUWESKeyport, WA 98345ATTN: Code 8013A
Mr. Henry C. MerhoffPACMISTESTCENPt. Mugu, CA 93042
Mr. Adrian EleyATTN: Code R153NSAFt, Meade, MD 20755
Mr. Steve BarberBell Northern ResearchOttawa, Ontario, Canada
Mr. Arvids VigantsBell LabsHolmdel, NJ 07703
Mr. James L. WeblemoePACMISTESTCENPt. Mugu, CA 93042
Mr. Jyoti S. SharmaWestern UnionOne Lake StreetUpper Saddle River, NJ 07458
DISTRIBUTION LIST (cont's)
M%. Ed LyonsNSAFt. Meade, MD 20755ATTN: Code R153
Mr. John K. WebbMITREBedford, MA 01730
Mr. Gayton YancyRaytheon1415 Boston-Providence TurnpikeNorwood, MA 02062
Mr. Gerry BlakefieldTRWOne Space Park - Building R2, Room 2178Redondo Beach, CA 90278
Dr. Phil BelloCNR, Inc.220 Reservoir St.Needham, MA 02194
Dr. Ronald L. MitchellNEC America, Inc.9218 Markville DriveDallas, TX 75243
I
MED
