performance of a very long 8 ghz microwave link
TRANSCRIPT
NTIA-REPORT-79-31
Performance of a Very Long8 GHz Microwave Link
J.E. FarrowR.E. Skerjanec
u.s. DEPARTMENT OF COMMERCEJuanita M. Kreps, Secretary
Henry Geller, Assistant Secretaryfor Communications and Information
November 1979
TABLE OF CONTENTS
ABSTRACT
1. INTRODUCTION
2. PATH PHYSICAL PARAMETERS
3. PATH PERFORMANCE ESTIMATES
4. DATA ACQUISITION
5. METEOROLOGICAL ANALYSIS
6. DATA PRESENTATION
6.1. Strip Chart Data from the Zugspitze
6.1.1. Diversity Performance at theZugspitze
6.2. Magnetic Tape Data from Hohenstadt
7. CONCLUSIONS
8. ACKNOWLEDGEMENTS
9 • REFERENCES
APPEr~DIX
iii .
Page
1
1
2
9
9
11
12
13
22
26
35
39
40
41
LIST OF FIGURES
Figure 1. DEB Phase I system map •.
Figure 2. Hohenstadt to Zugspitze profile showingthe ray trajectories for k values of4/3 and 2/3.
Figure 3. North face of the Zugspitze.
Figure 4. Hohenstadt to Zugspitze profile showing theterrain illuminated by the half-power antennabeams. The rays are drawn for a k valueof 4/3.
Figure 5. Hohenstadt tower showing test antennasinstalled.
Figure 6. Zugspitze equipment building showingbEB antennas.
Figure 7. Recording of "free space" propagationconditions made at the Zugspitze.
Figure 8. Zugspitze recording showing scintillationtype propagation conditions.
Figure 9. Zugspitze recording showing well-developedslow phase interference fading.
Figure 10. Zugspitze recording showing diversityswitch activity caused by phaseinterference fading.
Figure 11. Zugspitze recording shOwing rapid phaseinterference fading.
Figure 12. Zugspitze recording showing fast changesamong various fading conditions.
Figure 13. Zugspitze recording showing depressedmedian s LqnaL accompanied by rapidphase interference fading.
Figure 14. Zugspitze recording showing lack ofcorrelation between very deep fades.
Figure 15. Zugspitze recording showing the degreeof protection afforded by the diversitysystem.
iv
Page
3
4
5
6
7
8
14
16
17
18
19
20
21
23
24
Page
Figure 16. Sample computer printout of Hohenstadtfade data. 28
Figure 17. Distribution of single-channel fade durationsin seconds, Hohenstadt data for fades 30 dBbelow free-space conditions. 30
Figure 18. Distribution of single-channel fadedurations in decibels, Hohenstadt data forfades 30 dB below free-space conditions. 31
Figure 19. Distribution of simultaneous fade durationsin seconds, Hohenstadt data for fades 30dB below free-space conditions. 33
Figure 20. Distribution of simultaneous fade durationsin decibels, Hohenstadt data for fades30 dB below free-space conditions. 34
Figure 21. Instantaneous received signal level distribu-tions for all data collected at Hohenstadt. 36
Figure A-I. Hourly histogram data, 1977. 42
Figure A-2. Hourly histogram da.ta., 1977. 43
Figure A-3. Hourly histogram data, 1977. 44
Figure A-4. Hourly histogram data, 1977. 45
Figure A-5. Hourly histogram data, 1977. 46
Figure A-6. Hourly histogram data, 1977". 47
Figure A-7. Hourly histogram data, 1977. 48
Figure A-8. Hourly histogram data, 1977.
Figure A-9. Hourly histogram data, 1977.
Figure A-lO. Hourly histogram data, 1977.
Figure A-II. Hourly histogram data, 1977~
Figure A-12. Hourly histogram data, 1977.
Figure A-13. Hourly histogram data, 1978.
Figure A-l4. Hourly histogram data, 1978.
Figure A-IS. Hourly histogram data, 1978.
v
49
50
51
52
53
54
55
56
PageFigure A-16. Hourly histogram data, 1978. 57
Figure A-17. Ilourly histogram data, 1978. 58
Figure A-18. Hourly histogram data, 1978. 59
Figure A-l9. HouzLy histogram data, 1978. 60
Figure A-20. Hourly histogram data, 1978. 61
Figure A-2l. Hourly histogram data, 1978. 62
Figure A-22. Hourly histogram data, 1978. 63
Figure A-23. Hourly histogram data" 1978. 64
Figure A-24. Hourly histogram data, 1978. 6~
Figure A-25. Hourly histogram data, 1978. 66I
Figure A-26. Hourly histogram data, 1978. 67
Figure A-27. Hourly histogram data, 1978. 68
Figure A-28. Hourly histogram data, 1978. 69
Figure A-29. Hourly histogram data, 1978. 70
Figure A-30. Hourly histogram data, 1978. 71
Figure A-3l. Hourly histogram data, 1978. 72
Figure A-32. Hourly histogram data, 1978. 73
Figure A-33. Hourly histogram data, 1978. 14
Figure A-34. Hourly histogram data, 1978. 75
Figure A-35. Hourly histogram data, 1978. 75
Figure A-36. Hour.ly histogram data, 1978. 76
Figure A-37. Hourly histogram data, 1978. 77
Figure A-38. Hourly histogram data, 1979. 78
Figure A-39. Hourly histogram data, 1979. 80
Figure A-40. Hourly histogram data, 1979. 81
Figure A-4l. Hourly histogram data, 1979. 82
vi
PageFigure A-42. Hourly histogram data, 1979. 83
Figure A-43. Hourly histogram data, 1979. 84
Figure A-44. I-Iourly histogram data, 1979. 85
Figure A-45. Hourly histogram data, 1979. 86
Figure A-46. Hourly histogram data, 1979. 87
Figure A-47. Hourly histogram data, 1979. 88
Figure A-48. Hourly histogram data, 1979. 89
Figure A-49. Hourly histogram data, 1979. 90
Figure A-50. Hourly histogram data, 1979~ 91
Figure A-51. Hourly histogram data, 1979. 92
Figure A-52. Hourly histogram data, 1979. 93
Figure A-53. Hourly histogram data, 1979. 94
Figure A-54. Hourly histogram data, 1979. 95
Figure A-55. Hourly histogram data, 1979. 96
Figure A-56. Hourly histogram data, 1979. 97
LIST OF TABLES
Table 6-1. System Gain Calculations
Table 6.1-1. Zugspitze Fade Data
Table 6.2-1. Ilohenstadt Fade Data
Table 6.2-2. Simultaneous fade Data by ~/1onth
vii
Page
13
25
29
32
/
PERFORMANCE OF A VERY LONG8 GHz MICROWAVE LINK
J.E. Farrow and R.E. Skerjanec*
This paper discusses measurement of signals receivedover a 160-km (IOO-mile) line-of-sight microwaveradio link operating at 8.4 GHz. Signals were recorded at both ends of the path at various timesover a period of about 18 months. The effects ofmacro-scale meteorological conditions are discussed,particularly those which are peculiar to the centralEuropean location of the path. The data obtainedfrom this measurement effort are presented both interms of radio link availability and in terms oflong-term propagation performance. The data werealso analyzed to obtain distributions of fadeduration for single receivers and for bothreceivers in a diversity configuration. The studysupports the idea that it is possible to achievereliable propagation at 8 GHz on a link as long as160 km and with less than normal diversity spacing.
1. INTRODUCTION
From October 1977 through June 1979, the Institute for Tele
communication Sciences of the National Telecommunications and
Information Administration conducted measurements on the Defense
Communication System (DCS) 8-GHz radio link extending from
Hohenstadt to Zugspitze in southern West Germany.
These measurements were made to analyze fading activity and
propagation characteristics on this, the longest line-of-sight
link in the Digital European Backbone (DEB) system.
The intermittent measurements included two autumnal periods
during which anomalous propagation conditions are very common in
central Europe. The interruption of measurements during the in
stallation of the communication equipment precl.uded the collection
of data during the summer of 1978. The warmer time of the year
is expected to produce another period of disturbed propagation
(Vigants, 1975; Barnett, 1972), and this supposition appears to
be confirmed by the limited summertime data available.
*The authors are with the Institute for Telecommunication Sciences,National Telecommunications and Information Administration, U.S.Department of Commerce, Boulder, CO 80303.
2. PATH PHYSICAL PARAMETERS
Figure 1 is a map of the DEB Phase I system currently being
installed in Europe. The Hohenstadt-to-Zugspitze path is seen in
the upper 1/3 of the figure, and is 160 km long.
Figure 2 shows tne path profile for t.hi.svaame ,link~ Ray
tracings are included in the figure for two different vertical
refractive ind~x gradients. The upper set of rays is drawn for a
k of 4/3 and the lower rays for a k of 2/3. Note that clearance
for the center-line rays is available for even the extreme value
of refractive gradient corresponding to a k of 2/3.
The path passes over the rolling farm country and small lakes
of Bavaria and then over the Alps in southern Germany. At about
135 km from Hohenstadt, the terrain rises steeply to the high Alps
as shown in Figure 3, ,a photograph of the Zugspitze.
Figure 4 is a path profile showing terrain illuminated by the
half-power beamwidth rays of the primary antenna, (the lower antenna)
at each end of the path. This figure indicates that for a k value
of 4/3, the terrain is not strongly illuminated by the transmitter
at the Hohenstadt end of the path. The antenna pattern at Hohen
stadt will" also discriminate against~re~lected energy transmitted.i:
from the Zugspitze even though the terrain is illuminated from that
end. This would/make ground reflections an unlikely cause of the
very deep phas~~interference fading observed on this and on other
such paths (Thompson, et al., 1975).
'The radio'equi,pment was configured for dual space diversity
at each terminaV~ At Hohenstadt the diversity spacing of the an
tennas was 25 mwith the ilower antenna 12 m above ground level.
Figure 5 shows a photograph of the Hohenstadt tower with the
test antennas installed. At the Zugspitze terminal the antennas
were separated only 5.2 m vertically because of space limitations
on the mountain top. Figure 6 is a photograph of the Zugspitze
antennas. Th~antennas were stagger~d horizontally to permit them
to fit more easily into the restricted space. Since 12 m between
antennas is the'recommended vertical spacing to avoid simultaneous
2
AdriaticSea
AUSTRIA
GERMANY
Mt.Cimone
Vaihingen
ITALY
SWITZERLAND
Mediterranean Sea
Figure 1. DEB Phase I system map.
3
350121
3000A
((J
Lill
+> 2500(l)
Ev
--.J 200lZJ(J)
2:
W~ > 1500
0m«z 10000H
I--c> 500W..JW
rzJf(J
HST
21ZJ 40 621 80 1aa 120 140
ZUG
DISTANCE (k i. 1 c::::Jrn .E!J. t E3r- e;: )
Figure 2 . IIohenstadt to zugspitze profile showing the ray trajectoriesfor k values of 4/3 and 2/3.
.> 5
3500
3 fZJrZJrzJ
HST ZUG
a o 20 40 60 8fZJ 1fZJfZJ 120 140
DISTANCE (kilometers)Figure 4. Hohenstadt to Zugspitze profile showing the terrain ~lluninated
by the half-power antenna beams. The rays are drawn for ak value of 4/3.
7
8
deep fading on two or more diversity branches (White, 1970), the
degree of diversity protection available at this terminal was
uncertain. The system design called for lO-ft (3-m) diameter an
tennas and 10 watts of transmitter power at each end of the path.
3. PATH PERFORMANCE ESTIMATES
There is no single recognized model to describe in detail
the expected performance of microwave radio links longer than
about 70 km. Development of such a model is underway at this
time at ITS, under the sponsorship of the u.S. Army Communications
Command and the U.S. Air Force Electronic Systems Division. Data
for the model are being acquired on selected DEB links at 8 and
15 GHz (private communication, L. G. Hause).
The problem in selecting the equipment to be used for a
long microwave link is the unknown depth to which the link will
fade and hence the fade margin necessary to provide a given link
availability. Furthermore, there is no well-developed model to
indicate what diversity spacing is required to provide uncorre
lated or independent signals for a protection channel (McGavin
~t al., 1970). The measurement program reported here will
~ddress both the problem of long-term variability and the
question of diversity spacing.
4. DATA ACQUISITION
Two separate data acquisition techniques were used on
this radio link. The measurements taken at the Zugspitze site,
because of space limitations, were made by recording the combiner
control voltage (which is approximately linear in volts versus
received level in dBm) for each diversity receiver on an analog
strip chart recorder. All analysis of the strip chart records was
done by hand. This meant that the zugspitze data could be
analyzed only by inspection; no effort was made to extract long
term signal variability information in t~rms of continuous
distributions of received signal level. The strip chart
9
recorder had a peak frequency response of 50 Hz so that rapid
variations in signal level would be preserved. The charts were
timed with rectangular wave timing signals with half periods of
one minute and one hour. Some of the more recent charts have one
channel that shows which diversity receiver was carrying traffic
so that the activity of the diversity switch could be tracked
and the adequacy of the diversity configuration evaluated.
The signal level information collected at Hohenstadt was
put ~n strip chart also; but in addition, it was fed to a
minicomputer-based data acquisition' system. The system con
tained an analog-to-digital converter board (A/D) which sampled
the combiner control voltage twenty times per second. The
samples were sent to a histogram program which retained every
tenth sample f or binning Lnt,o a 20-level histogram. The histo
gram bin interval was,3 dB so a signal level range of greater
than 60 dB could be accommodated. These histograms were accumu
lated for a one hour period and then written into a permanent
magnetic tape archive record. Each hourly histogram for each
channel contained about 7200 samples which provided good
resolution for level variations.
In order to take particular notice of deep fades, each A/D
sample was also sent to a fade message program. This program
was designed to measure the length 'of time that the received
sLgnalwasbelow ~predetermined threshold for any particular
event and also to note the lowest value of signal observed
during the fade. To do this the program compared two successive
samples, and if the most. vr-ecent; one was below threshold and the
previous sample was above, a fade message was opened with the
start time equal to the clock time of the fLrstsample below
threshold.
To determine the lowest level during a fade, the first
sample was retained and. the next sample compar ed wi th it. (Each
new sample was also comp.ared with the threshold to see .i.f it was
abov~ that level.) The lower value of the two was retained and
compared with the next sample value, and again the lower value
10
retained. In this way, the lowest value occurring during any
given interval would be the one which finally remained. After
the start of a fade when one or more below-threshold samples
were observed, any sample which was again above threshold would
cause the fade event message to be closed at the time associated
with the first above-threshold sample. The signal level value
which was the lowest 6bserved would also form part of the fade
event message. Thus the message would give the start time, end
time, and the lowest level of each fade below the threshold. The
beginning and ending times were known to the nearest 50 milli
seconds which was the.sampl~ngi perLod of the "A!D converter. Since
both received signal level channels were being sampled and the
data handled the same way, it' was fairly straightforward to
analyze the data to determine when either or both of the signals
were below threshold. This permitted us to determine the
distribution of fade lengths at the threshold and also simulta
neous fades on the two diversity branches. Even though data
were available from both ends of the path, the different methods
of acquiring and analyzing the data have made the results some
what difficult to compare directly but the results of both
analyses will be presented later in this report.
5. METEOROLOGICAL,.ANl\LYSIS
Meteorological analyses for mi.cz'owave propagation studies
are usually done during periods of unusually bad propagation
conditions or for periods ~uring which interruptions of radio
link traffic occur. Thecauseso£ ~nterruptions have been
thoroughly discussed Ln it.her Li.t.er-abur-e (Dougherty 1968" White
1970). In general, the conditions that lead to poor propagation
are limited in space and time and can be observed and identified
only if extremely fine-grained meteorological data can be
obtained. The exceptions to this statement are such phenomena
as extremely heavy rains (which lead t.oifLoodLnq and hence con
siderable public notice) or fog conditions associated with
strong stratification (which will ·a.g;ain lead to public notice
11
by interfering with transportation system activity). From the
viewpoint of the student of radio propagation, it is unfortunate
that such catastrophic weather conditions are not only rare in
time but even more so in the space which he may have under
observation. For these reasons, the meteorological analysis
which was done for this path dealt only with macro-scale
phenomena such as could be determined from world-wide synoptic
meteorological charts. Specifically, the type of weather con
dition which has been identified as causing propagation problems
on this path and in fact over many paths in Central Europe is an
autumn condition characterized by a static or slow-moving high
pressure system (Samson, 1970). Such systems have ~een observed
to cause problems with great numbers of military microwave links
in Germany, particularly during October hence the name "Great
October Blackout" which is often used by military personnel in
Europe to refer to the phenomenon. Although this is not a
phenomenon which affects the same links every year, the pattern
is familiar enough to warrant comment.
6. DATA PRESENTATION
Recording was started at Hohenstadt in mid-October 1977 on
a test system which used the DEB antennas and a temporary trans
mitter at the Zugspitze and two smaller, 6-foot (2-m) temporary
antennas and two test receivers at Hohenstadt. This test system
was operated until the end of April 1978 when .the DEB radio
transmitter at the Zugspitze was turned on. At that time, the
temporary transmitter was moved to Hohenstadt and signals were
recorded in both directions on the path for a short time until
1 June 1978 when the temporary antennas at Hohenstadt were
removed to permit installation of the permanent DEB antennas.
On 23 September 1978, recording at both ends of the link
resumed using the permanent DEB equipment. The major part of the
data obtained at the Zugspitze was recorded between 23 September
1978 and 15 June 1979.
12
Table 6-1. System Gain Calculations
Path LengthFrequencyAnt. size ZUG
HST
Ant. gain ZUGHST
Line Loss ZUGLower HSTUpper HST
Free Space LossAtmospheric AbsorptionTransmitter Power
Received Signal LevelLower AntennaUpper Antenna
TemporaryInstallation
159 krn8.3 GHz
10 ft (3 m)6 ft (1.8 m)
~,5. 9 dBi41.5 dBi
2.5 dB5.5 dB7.7 dB
154.5 dB2.6 dB
+42.3 dBm
-35.4 dBm-37.6 dBm
PermanentInstallation
159 km8.25 GHz
10 ft (3 m)10 ft (3 m)
45.5 dBi45.5 dBi
2.7 dB2.0 dB3.0 dB
154.5 dB2.6 dB
+37dBm
-33.8 dBm-34.8 dBm
Table 6-1 gives the values of system gain for each of the
arrangements mentioned. Note the similarity in received signal
level for the two configurations. This similarity lends con
fidence to the expectation that the two sets of data can be
amalgamated to present a longer term data base than either
_.would provide alone.
6.1 Strip Chart Data from the Zugspitze
The strip charts recorded at the Zugspitze were analyzed
primarily to determine the fraction of time that fading occurred.
The fading periods were further segregated roughly by fade rate,
fade depth, and the occurrence of simultaneous fades on both
diversity branches. Primarily, two different types of fading
were considered.
The first type is called scintillation which is the term
used to describe rapid, shallow fading. The signal remained
generally near the unfaded line-of-sight level with only
occasional excursions to 20 dB below this level. An example of
data obtained under the least disturbed conditions is shown in
Figure 7 and an example of scintillation fading is shown in
13
i ri=tzht~~ i i k,~ I ~j- " ~ I i I ii....bl
:" I-
i i:~;;.,:.':. :; ....... :.... : ... j
II
-:-
_.._- -t- __.._.-.._;...--"...-.-" ....., ....,._ ....__.:.- -r- ... '- •. _-'- •. _-_ ..•,,.. .. -- .•.. _ ..
_._.~_.-i.- _
-
--
UO!S!"IO swalsAs IUaWnJI~UI ":)UI PIno!).......~
\IS n '-' pa;U'Jd O!",O PUBieI\3 ::J
l.YVH:>n:>:>v HSnYB
'. . . ..: .. ' ;". . .' '. .: __.. '.•. :.""1 I ~:: M.. _ .. _r-t: :.1"1.. ':'" :f1.=LftJ1Jl!lj::__ .'.1:fiittj:.. '.r--!. r-t: !~.. ::: :.1"'"'1 ...r-"t:
u
J.h1~r-t 1:P r-1~I:..... ~;~O;CO;Du;~·g"u-u-;". -' . '--:' • =UW~4J1J4-J ,0trUn; °mTl Ri rn IAt Ri ,...,. HtM rt"'Lr11 M .rt'"1:r11 -M -r-;---.----------..------------------ ... ---.-------...-----~--
i
~r,ME It.._- I .~
-_.~ -------.- -==r
.-',-.-.--.'--
T-T=+~~~~::r--..~._--_-_---_-" .._-~i.T-~:l-,...J.~.-.L.~-:..-'-.~_---"---L ..-'-~
fT"9St- /. if .1 i;~-+---+--+---~----'-__i_ t. .' I l • •
Figure 7. Recording of "free space" propagation conditions made at theZugspitze.
Figure 8. On these and subsequent figures, the small rectangular
timing signal has a period of 2 minutes, and the larger rectan
gular timing signal has a period of 2 hours. The scintillation
fading-results from the addition of a multitude of small
received signal components of random phase and magnitude to the
main, on-path signal. These components are caused by atmospheric
turbulence and could arise from scattered or weakly reflected
signals originating from antenna side lobes or the edges of the
main beam.
A second type considered was well-developed phase inter
ference fading which results from the addition of two or more
signal components of very nearly equal amplitude. The delayed
component would very likely arise from a strong reflection or be
due to atmospheric multipath associated with atmospheric refrac
tive index layering.
There is a wide variation in fade rates as can be seen from
Figures 9, 10, and 11. Note that Figure 9 shows a period of
very slow fading, apparently resulting from smooth changes in
effective reflecting surface position. Figure 10 shows a more
typical fade tate which could result from more than a single
delayed component and which would in turn imply the existence of
more than one atmospheric component. Figure 11 shows a very hi~h
fade rate, and again the character of signal suggests multiple
delayed components which are changing fairly rapidly in time.
Transitions between fading regimes could be fairly rapid .as shown
by the data in Figure 12. Note that in a period of an hour or
so, the fading changed (reading right to left, in the direction
of increasing time) from scintillation fading to interference
fading to nearly undisturbed line-of-sight conditions.
Periods occurred when more than one fading mechanism
seemed to be involved. An example of this sort of condition is
illustrated in Figure 13 which shows a period of severely
depressed median signal level with superimposed phase inter
ference fading. Periods of this severity were fortunately rare
and close inspection of the data shows that even in these severe
conditions both signals were simultaneously below the -70 dBm
15
......0"\
.---..-.. --111#1,(~f
HtJUlf
~
--i-ISrrr1T1l1I
111
~r\.;' <t
~i
3<D··''C~''''''
fl'(V
""-Tn"II
! I!
,/ME.-I
~ I"
If
nrrt11l1!
tlBt, It-
HIli
,...~~v-
-JiJ_:_rR£c.LLVFll.; IJAL~_lN_E
Co,ofa,'I/C/{l
~~~i~fV'II-·l!J.,vJVi
~o
"40-"A-'~ _ ,n 50~
ReCeLYt:r -Q't:7.'" 60-
-1bAil-
Figure 8. zugspitze recording showing scintillation type propagationconditions.
1·7
---.~ -4· ..-, ....-t-. 1-.. -'- <_.+-- -+- ..-+.~.-+ --+- ... f -
I
~~···'~'-~l.'._-\l--'.':~·l·~:-;'--';g-~~ """,..~._..y- __1[-._•..LA.~
T
--~.::~..... -- - -., _.....•. _.- - .
..................................•................. : .
.... -
bV r-
~1!N1j-~~~i
.-~~rii:
I ~
--.-----1.-.i-.J - .
,....
J
IF':.
1'<: "IV ~,,,,~n
I I I I I I I i I I I I I I I I I I I -4 -4···
~00
.~ ,.",. ..y,.: 1""""\ A I "\1\.,' 1!"'"'\i"'V'\.~ .I-v-'" ~ ~r"" ....... J II"":'V' •...I"\ Y- '_ J"Ii". i('.·"V4,~ ~ : II' It'' \ 1'1 If"I A IH -. 1 N'~W V .". i"'Iw '\ If It" I 1 J"lf:\f "\ I \; Iii. i _r'I
- J III' PI 1.1; \llJ I Y 11' r II: p'f '.1''' U ""'~-f~~~7-t-'-7i7..IY'~±Tt--i~-f--'i-i-t":"90=.-.TI---..... .. 'I'\' lif .• • U il r:
:
'?n ___
.~~Q"""'"
•
l.HYH::>n::>::>Y HsnHS
• • r-:.. -C- ......:..--.;..----i-·---'--'---+--'-i----i---+-......,....-+--'--f---+·--;---+--i--+-r----'---+--;--'--·,-.. ··'-----l.-+-+---+-+_-f----i---+--+--+--+-'--,
i
.... """'T r,,.... M rn r -,r,:,.... """'T rrt ,.... ~ ,..,M ,.., r,,.... !"""'I ,..,,.., r-t ,.., i,.., ...,. r'1 1"" rT rn r !"""'I ri rI -,·m i1 r-T r"i i1 I..•..,. r"1rl P"""'1 M il"""rT r-1,r!"1.: I r I I ~ f J L.l _ f j L."I _ f J L' _ L.J L..!. !...J L! L,.I _ L! L.1 LJ---U.,.l~- L W L1_ IJ L1 _ '..2 !.-l ,_ 1..2.!.. I!... I I J I _ 11 I i_ J 1: 1:1 I: I I
:
• I
I I I I I I I I I I I I I I I I I-+-+-.. . .. +- - ~ .- - _.. - ~ _. - • - • --- - - - - .-. -.. - - - - ._~ .----+--... ---'----- ~.---~ ..+---+-..... ~ ~ '" _. '--'
.
:
. : ;
j •
-: ..V
. . :.~
..IT/~t ~-'~ !
..... . . :
.: I
I ... L-: -,....,.-..~_.,._--c__ ._-.7"'---c._._-7"'- .. --'-----7"'.. --'-----'-:---,---~--- 1---,---,---·-·--·----7-<vf\--'t-· -'---.,--------.:.:------'---,-'----+---'--1-;-L.-+-f-t'-'-+-'---+--+-+-'--t--+-+--+--'--+---'---+-+--+-+-f--,-,f-+-'+-
--+~-'+:--+-'-+--:-'-'-t---+-C+--+-+-'-t--t--+--+-'--+-+-t--t:---+-,+ •.~;
• • :1.
I
--+-+---1-- c-+-+-+--+--+---~-+---+~+--+-+_+-+-- S"9 GS17
Figure 10. Zugspitze recording showing diversity switch activity causedby phase interference fading.
19
20
]
I
• '-··.J-,.--+~-l-:······- :... n_. _ .:.LJ:V!" 11\1: Ill'll, i" !itO vr ~ V
-----.-----'-.-~~•.---'----.~--'-.-'-~_._.-,--._~... ..L.-'-_." .. _'._..._._ .....---.. ,_ ..._.,_.. '-.:~-+A
, '. III If I Ii r ...,... W'~' i "I'll 1 f
• : II' \ I ~"V l:~ V 11: "-ill'"'"\.. ... .4.1\; l'\ :~. _i...A.1¥VY .~
~: ~~=~~r~~T~rTr~~-r-.~.~~i~i'~~~~~~~~~-T~~~~-~~~~~-~---~-~T ---T~~~~ -~.~~~~~~ '1-~~' r
'r~:lIJ
\,,:,j'Ar1~Ji \1\ ~ :1
tn 14-l--rft!t1-'-R=f-i=:-::pJ:~·+-+·+-++-J+~h·...- ~··l----'-L.·~--'-·~~~··-"~·"·~:··~·····i. : .l._.:...L .......:
l.-.l ,_:;'1\ !Ml~ : ~:_L.Jr:";If·1f1,,ll':~ U'il\!.Ii\UVi"~
~~
B
,-1'
-l
A:)~
i 11
r.r: ~
TlTifI:TTT"'f.I1~
UQ,~
I ;;·I.·,W
i i
Gould 'nc., Instrument Systems Division
UO!S!AIOswalsAs IUaWnJJsu\ ":;)U\ Plno~
t\.)
I-'Cleveland. Ohio Printed 'r' USA
v s n u' palcPJd OIUO pUB!aAal:.'J
i
~~~1£~-H~
~I···.I •• ·.I ...1!=FfiffBHlfU'fbflkJ#-PJhCUi~~1J;g~P=~~-EYit:t~fy~~jjjit-i~~+-Yiil:ij-!~+=
;ji•.~~!,r0~"
~;~
1lJ'i
~
_T/~E..........
--t- .-.4- ..... -,. +-- -4----4-
~P\.'"""-_._~__~__~.~'..........._. • ~_~__~ __--"""--':~~--+--':'-._-l--.
Qe-,
~
.~
............ ~t++;.....~
~:~---
~1rb-v_t
Figure 13. Zugspitze recording showingby rapid phase interference
depressed medianfading.
signal accompanied
level (corresponding to peN t.hr-e shoLd) for only a small fractionof the time.
6.1.1 Diversity Performance at the Zugspitze.
When the first charts from the Zugs~p;i.tze were reviewed, it
appeared that the correlation or similarity between the signals
from the two diversity branches might be so high as to reduce
severely the effectiveness of the diversity. The similarity of
fading patterns was thought to be caused by the close antenna
spacing at this site. Figure 14 shows an example of the data
that caused the concern. Note that even~ fine details of the
. signal are seen Ln both channeLs , However, a careful analysis
of later data showed that when deep fades occurred, they were
more often offset in time, as was the case for the deep fade just
before 2300 hours on 270ctoberl979 (Figure 14).
The offset between the minimum values is one small chart
division which is about 15 seconds. This is ample time for the
diversity switch to operate and no interference with traffic
should occur.
Figure 15 is an example of the data obtained after the
switch combiner had been connected to show which receiver was
supplying the baseband signal to the multiplex. Even though the
signals are not so similar visually as on the previous figure,
there is still a very high apparent correlation between the
signals; but the deep fades are themselves uncorrelated, that
is, they do not occur simultaneously on both channels.
From this, we have concluded that the diversity separation
in the vertical direction is ad.equate to provide signals to the
two receivers which are only infrequently subject to simul
taneous deep fades.
22
i
~ +"PTTftf:I
j
~~.~~~~
~I ". I/'
N
+
~
boo
!::!
'.~j
- - f'"'4 :--Kllf'"'ill
·-i
•. _. ~---->-~ ~--+-+--+--~~~ -¥- -, -....
~1
E'" ,., ,., , , '~~~~
,'i~kf!'I'~1~flT~rT'1
i!·:· .•~
1f¥Y, .."-:"t
'"' ft\N
,
IIII I I
i
J
OIl-lO 'pU8IaAaI:)vsn UI palulJd
UOISIAIQ SWaIS~S IUaWnJISUI ":)UI PlnogIII
I I ~I
f\.)
W
'!":
~TljVfG ~
t\s
I. I
i
o V~~' V
Figure 14 • zugspi'tze re-cording showing Lack of correlation between very.deep f ade.s .
~··-~l_ I I-r
_.-
---j-----i--
i i
-.•-:;:t=::
tJ:3}3MARI
O'Z) 3 MAR 11
tv~
.i' 1",
III
•
IS I U pal o'~o·pueI8A818
UO!S,AIO swal.As luawnlisul ··~I Plno9
.1·1••••I••••I•••il·I~'~:.:,:i~.i ... ,:1[11; ~;, ;·'H·'· .Cfm,·:j:·:,::.,::I·:.,· i ; .. ,. i· i··'·li .·1:1·· .....
114m;'-H
~!=~~_tl~·. ~''(LL
~~ COfflfl7t!£R .
,..:.,.,-II~
~.
----_._-_._------
-KAT--~/NI-
.__"_ ..-,-_-,_--, . :3_~. ..__=:j=-=-~":'t~-.-'-.--" ...
--;:::dBiYF --:so:"A~-
._a~Ce;v.<:-,.
11-1-.1 l..~
70_. ··A~·
Figure 15. zugspitze recording showing the degree of protection affordedby the diversity system.
Table 6.1-1. Zugspitze Fade DataSeptember 1978 to June 1979
Number No· Coincident Coincidentof Phase Fades Fades
Hours No Interference Per Per HourMonth/Year of Data Fading Fading >20 dB Month of Data
SEP 78 171 59.1% 78.4% 11 .064
OCT 78 744 38.7% 49 .-9% 61 .082
NOV 78 574 24.7% 32.9% 19 .033
DEC 78 336 84.5% 88.1% 0 0
JAN 79 240 77.9% 93.7% 0 0
FEB 79 360 70.3% 90.9% 1 .003
MAR 79 668 83.8% 94.7% 1 .001
APR 79 611 92.5% 95.6% 0 0
MAY 79 629 85.1% 91.6% 15 .024
JUN 79 336 63.4% 75.6% 12 .036
Table 6.1-1 summarizes the results of the analysis of the Zug
spitze strip charts. The data are blocked off by month with the
second column giving the total number of hours of data available.
The next two columns show the fraction of the data when no fading
occurred and when no fading in excess of 20 dB was present. The
next two columns show the number of coincident fades (simulta
neously observed on both channels) and the normalized coincident
fade rate per hour. This last col~mn is simply the number of
coincident fades divided by the number of hours of data. The
only complete month of data was October; during all other months
interruptions for installation, alignment, and testing led to
the loss of considerable data.
Nevertheless, Table 6.1-1 shows an interesting fact about
propagation over this link at least. The most damaging occur
rences from the standpoint of traffic continuity are the
simultaneous fades below PCM threshold. Although short, these
fades will cause some disturbance. Note the increase of coinci
dent fades per hour from September to October, the decrease during
25
November, and the absence of coincident fades in December. The
-frac t.Lon of time wi thout deep fad i.nq decreases from September
to November but recovers to a high le.vel in December. Thi.s
large percentage of the month without deep fading persists from
December until May. The month of June again shows a significant
decr-ease in the fraction .of .t.Lme which is free from deep fading.
The table shows anot.he.r i r.e.La t.LonehLp which was suggested
by,Vigants (l975), namely, that the warmer time of the year will
be a period of increased fading activity. The marked increase
in fading in June 1979 suggests. that, this may be true ·for this
link. The rate o£simultaneouscorrelatedfading has also
increased during June so thatthe'number of fades per hour is
at ···the highest level since the .autumn. Although the table does
no t; show it, the p'eriodsof severe fading through September,
October and November t.ended to last for a day or even several
days at a t.Lme while the fading periods during June occurred
generally at night and seldom lasted as long as 24 hours. This
difference in the diurnal fading pattern is presumably a reflec
tion of the d i.ffererrt; c auaes of the atmospheric ·layer condi
ions. The situations which produced the autumn fading were
large, stagnan~ high pressure systems which were not chan~ed
much from day to night or from day to day since they gave rise
to 'very stableinversion~typeconditions. The summer effects,
on th~otherhand, resulted~from layering caused by night~
time radiation cooling of the ground and lower levels of the
atmosphere which would be broken up by solar heating during
the~orning. These latter effects could be modified by cloud
cover but would generally' represent typical atmospheric layering
conditions.
6.2 Magnetic Tap~ Data from Hobenstadt
As described previously, the magnetic tapes contained
hourly histograms for each receiver plus data messages giving
the beginning and end times and depth o£ fade for each excur
sion of the signal below threshold. The item of greatest
intere~t which resulted feam the analysis was the measured
26
propagation reliability of the radio link during the period of
investigation. Therefore, q.nalysis·beganwith a . listing and re""
view of the fade messages on each diversity channel separately.
Figure 16 shows part of a typical printout of fade messages. The
type numbers, 50 and 51, describe which of the diversity bran
ches faded1 type 50 refers to signal fades 'on ~he radio receiver
connected to the lower ant.enna ,and 51 refers to signal fades on
the receiver connected to the upper antenna. The ·fade message
printout is organized to show the site (number 1) ,th~ type' (50
or 51) and the message length in. computer words (8, which is
used asa check of data integrity). The next two groups of
data are the start and end,t·imes·of the fade. The Julian date
is given, followed by the hour, minute, ·s.econd, and terts of milli
seconds. This is f o l l.owed vby a value Ln vrnd L'l i.voLtia correspond
ing to the lowest reading of signal level observed during the
fade, and by a duration in days,hours,·minutes, aeconds , and
tens of milliseconds. The 6ccurrence of overlapping type 50 and
51 messages is shown along with the duration of the overlap.
Since the system was being iris taIled and tested during most
"'of the test period, it was necessary to .ed i.t; 'the datara·ther
~carefullyso that theapparent~fadeswhich were caused by equip
ment shutdown or testiIlgwould not'contaminate the results.
The fade messages were analyzed initially by printing each
message out and attempting to eliminate those which appeared to
be ,the result of phenomena unrelatedtb propagation. Aftercon
siderable study, the data 'were edited to eliminate all signal
level excursions below the eof twa.revt.hr-esho Ld if that. excursion
lasted longer than one minute. This was longer by a considerable
margin than the majority of observed fades, and a review of. the
strip chart data showed that virtually none'of the valid data
would be eliminated if this rule was'followed.
The analysis of the fade>message data began with the devel
opment of duration histograms for each received signal level
separately and then for periods. when 'coincident fading was
observed on the two diversity channe.l.s, The time bin sizes
27
~
~~ " ~~.~ "'~
~~ " ... ~ ~
...... s: l ~~ ''4. ~\II .... '-I "" ~ :" It 'il 8 tl~1 ~1~bll215~ Il~ I "t 5b I 231 0 2&20 DURATIONI OIOO~O"'tJO.4S
'il 8 llbl b124t3~ltO 13b' .,.2ll.'"IJO 2A80 DURATION. 01001 0" I no. 20
51 8 13bl b.241~blb~ 13b' *,12~I'iol~'5 2700 l)lJ~ AIILlf\ll 010010"'·"0.25
~1 8 lIb' b12S, h.7~ 1 :S.,I b125. blije; 355'5 I)U~ATJnNI O.OOIOfJIUO.tO
1;0 8 tlol b.2~1 71 0 1.~bt b.2St 711t) ?Q35 DuRATION. 0'0010"'UO.l0
~O ~ tl~. b12512t.70 13&1 hli?5.t't.Qn _,100 l)U~ATlONI OIOOlon,oO.20
-;1 8 t3bl ~13"1321~~ 13.,. hiJuIJ217n ~O35 (}lU(AT ION I 010010"1"0.15
~t 8 tlb. blUI' ~:ln l~bl bll.£11 ~11r; \b~~ O'J~'TI(JNI oI 0 0 I 0 r.JI f' C' • 05
r;1 8 t 3b I h: 43. 1o I 0 116: t\,4\llf')ltr; ~2qo I)lJJ.t ATION t 0.00 I ()'" no. 15
~l 8 1]~: 7111.18110 1J".7r11:tQ.50 3250 l)UHA TION 1 OIOO:O"l u t . 4 0
C;O R 13bl ,.11 :53130 13bl 7:11:-;41015 298'; Ut....IRAT IONt 0100 10 IJ I n t .3S
~l A tlbl 7112'11:15 ,~" I 7112't7t~o ,9HS OUHATJONI O'OO,oo,uC'.45
N t I§O 8 13&' 71tq1471t~ 'J~t 7,'Q,117Il0 320'5 f)lIRAT lO~. OlnOI(),,:no.~,;CO
ISO 8 t 3bl 71211]71('5 1 51)I 7.2 11 3~ I n ?7"o OIJkATIONI 0'0010nl"0.15
fiO A t3b, 7123f 8'75 '3~1 712~1 q'1S ~~r;~ OIJRATln~t OtOOIO n t nO.40
~1 A t3b. 711015at1~ 13"1 7,30:li4:5c; 3525 I)UQAT·IO~. ()I 00 10 C' , f) 0 • fJ0
-;0 e t3bt 7'44:25171 13"1 111J<.I:J1.'Sc; l'1c.10 I)t.H~ ATI 0 ~ I 010010"'11.1.>5
1;0 ~ tlb. 7144151170 t5b: 7al~5:tfl.'8{) ,QS5 tJ1JkATJUN, OIOOIO"I~t.l0
';0 ~ t]blt2'37:21J14~ , 3~ I 121 J1: a.,. b 0 t (.15 lJURAlIl)NI 0'0010nI19.15
1 -;0 8 1381?311oIS01~~ 13AI?3.t na,\11f)'i ?Q70 DUfo(6Tln~1 OtOOIOn,uO.80TypE so STAHT tOlIO~123Ic;t.qo n'j~= .750 TYPE 51 START 101.08,2.J,Sl. Q " oUR: .750 OVERt. AP .75
TOTAL DURATION (SECO~DS)Of OVERLAP: .75
2 OVERLAPPING MeSSAGESE~lEED 1MIN~.NUT P~INTEO
Figure 16. Sample computer printout of Hohenstadt fade data.
chosen were 0.1 s for periods up to 5 s, 0.5 s for periods from
10 to 30 sand 2 s for periods from 30 to 60 s. This selection
provided good resolution for short fade durations and kept the
number of histogram bins from being unmanageably large. These
histograms were transformed into cumulative distributions and
plotted against a logarithmic fade duration axis (on log-normal
probability paper) as shown in Figure 17. Figure 18 shows the
same data for both receivers plotted against the fade duration
in decibels, where T(dB) = 10 10glOt(s) (on linear-normal
probability paper). From Figure 17, a median value of fade
duration to a level 30 dB below the normal unfaded received sig
nal level can be read as 2.2 seconds for the lower antenna and
3.7 seconds for the upper antenna. Both of these results are
consistent with the discussion in CCIR 1978, page 190 which indi
cates that for a 70 km path at this frequency, a median duration
of 5.3 seconds could be expected and that as path length in
creases, the duration decreases but as the path clearance in
creases, so does the duration. These expectations are borne out
by the data,but the large difference in median values for data
from the two antennas is puz z Li.nq since the addi tional path
clearance for the higher path is relatively small as can be seen
from the profile in Figure 2. Table 6.2-1 shows the size of the
data base for these curves.
Table 6.2-1. Hohenstadt Fade Data
Hours of Data Number of Fades
Upper Antenna
Lower Antenna
Both Antennas
4619
4605
4605
2225
1419
98
Fade Rate
.5
.3
.02
Note that the number of fades on the upper antenna was about half
again as large as the number observed on the lower antenna. No
explanation is offered for this phenomenon but the difference in
path clearance might have some influence on the number of fades
which occur.
29
r..---.-r-rr-r- .,.....-r-. w
,---------------~,,~' 1,~
.,'.,",,,'
;;"
-: r,,' i, I
"-:.',",
"""",,~,
;
"-',;;-',
. ,,,<' Lower Antenna"
;;,,",
""",~,
",",
";,,'
Upper Antenna
~
~
ttl
r··r..... i ""1-·····......-.............-' iii i I-·-·-........----.----........-:-r--r-~......----.---·-···-"Trl iii' , ... -
!~i
I
i
L L ,',.,'", I ..._.-t-- ..... ' ",', ,I." I _...,............_.. t t .•.J......-...e....-..--'--....---L---___ I" t , ,,' ,t I,', t t , L •
a,aai 0. 01 a, 1 rae 5 B. 9 S.99 0. 999
Fraction of Fades with Duration Less Then theOrdincteVolue
10-
lm~
0)roc00
w '0)
0 U1
Cerf
e0 1erf
..,>0'-:J
CJ
(I)"'00
lJ...
Figure 17. Distribution of ·single~channelfadedurations in seconds,Hohenstadt data for fades 30 dB below free-space conditions.
• "'" , "I
,-
-------------",'"
",-"
",'"
",'",'"
"-:"
"""""
"",,,,,,,,
",,,""
Upper Antenna /' ,/'/ Lower Antenna,
'I
"""""
",I""""
I'",I'
"I,,'
""f8'-0C0 100CD0)
'-'-iJ
tn0
r-t
W SI-'
....."'wi'
CD-0
C.r4 I
C0....
.fJ0L::l
C
CD-0
CJLL
8. 801' 8. ral f2J. 1 0. 5 S. 9 0. 99 S. 999
Fraction of Fades wi~h Duration Less Than the Ordinate Value
Figure 18. Distribution of single-channel fade, d~rations in decibels,flohenstadt data for fades 30 dB beLow free--space conditions.
22.75ooo
618.86
234.2175.30
173.6510.448.45
o54.1
Month Year Number of FadesOCT 77 52NOV 77 16DEC 77 12JAN 78 6MAR 78 5APR 78 0MAY 78 3
DEC 78 4JAN 79 0FEB 79 0MAR 79 0
98
From Figure 18, an approximate value for the standard devia
tion of the fade duration T,in dB can be read. The value of
6.2 dB compares favorably with the 5.6 dB value offered by CCIR
1978. The observation that the standard deviation of fade dura
tion depends only slightly on path geometrical characteristics
and the climatic conditions encountered seems to be supported by
these data.
A similar analysis was performed for simultaneous fades
below the -66 dBm threshold. The duration of these fades appears
to be log-normally distributed with a median value of 2.5 s.
The standard deviation of the log of the duration is 6.5 dB,
which is interestingly close to the values measured for the data
from each antenna separately. Plots of these simultaneous fade
data are given in Figures 19 and 20 against logarithmic and dB
scales. Table 6.2-2 describes the data base on which these
figures are based. Note that the bulk of the fade data was
collected during October 1977. Unfortunately, signal level
distribution and fade information from Hohenstadt was not avail
able for the September-November 1978 time period because of
problems with the data acquisition equipment, but data were again
being collected by late November.
Table 6.2-2. Simultaneous Fade Data by MonthTotal Duration,
Seconds
Average Simultaneous Fade Duration, 6.3 s.
32
1i1i!1
~l
4~4
1~I
IIi
a, 1
--..,..--~. iii iii ._.. - '-'~'-":-""""-"--r-"'--'r' ..--,.----.. --,---.... -"-r---'''-'''--'T' iii i -r---r--""--·-·rrr.-~..--rr---.
f
rI
~I
~
·r--.--,----.-~
L
. ~t-
~~I
i1iII
L ..-L-., , , , , "I ....~_---J...- .....t..-.-L....-I....l", I _.._.__..L--..---L..-.._---'---_.L.. __..iL--__--'--_~_..__._.~ LL' , · , , ,_-'---~.a.....L._..................J(21. ~~1 0. "']
1
le r~
t~
1"0
0)-0
Coo(I)
(J)
1:.r-f
co......
4->ot.J
o(J)
-0o
u,
ww
Fraction of Fades with Duration Less Than the Ordinate Value
Figure 19. Distribution of simultaneous fade durations in seconds,Hohenstadt data for fades 30 dB below free-space conditions.
I - - .. ~---'--"""-li'i'if .,..-.......--~··_··---r----···---.---·----r--····-r-·_·.,.....···..._·--..--·_·----r-"··-··_·· ... i , , , , • , ,.-- ·.--~-~~··lr-l
, ,., •... , , , ... I "_->--, 1"",.", 1",·""" i'" 1 . 121. 5 B. 9 QI. 99 .rae 999
Fraction of Fades with Duration Less Than the Ordinate Value
~
I
2111~ '" ""'I'
!
18
;//',.•.
;I
/ ...../ .......,)
(",...._.,-',...
ll··
_J.'"
.,........:/_.......' ......
!
~ /~
I..'t I
r (t '!: I, , . , . , "I .
21. "91 B. IZJ 1 1'1. ~
""•-0C00CDCD
'-"-io>
D')0.....
sw ~
~'-"
CD-0
t:....C0....~0'-:J
CJ
CD-00
I.J...
Figure 20. Distribution of simultaneous fade durations in decibels,Hohenstadt data for fades 30 dB below free-space conditions.
Current planning calls for data to be accumulated indefin
itely on this link, and as further data become available, par
ticularly during summer- months, supp.Lement.svtrovt.h.La repor-t will
be issued.
An interesting outcome of this study is that of the
4505 hours of data, a total of only 618 seconds of propagation
outage ,was observed. This results in an overall propagation
reliability of 0.99996. This ~alue includes the effects of
diversity improvement and assumes no bias in the sampling.
Ana;Lyzing the magnetic tape hourly histogram data at the
Boulder Laboratories began with editing the ~ata'to eliminate
periods when the system was out of operation. Most of these
outages were caused by system testing or other disturbances not
related by propagation. The editing was done by plotting the
median and the interdecile range of each hourly distribution.
These plots were grouped by weeks, and since this arrangement
gave a good representation of the radio-link performance, this
grouping was chosen for displaying the data (see Figures A-I
through A-53 in the Appendi~).
From these plots, it was fairly easy to eliminate most "of
the periods during which contaminated data were collected. The
remaining hourly histograms were then summed to obtain one
long-term distribution for each of the two diversity branches
at_Hohenstadt. Figure 21 shows these two distributions of all
·data from Hohenstadt plotted on a Rayleigh distribution grid.
The shape of the measured distributions at the high signal levels
departs somewhat from the Rayleigh distribution, but it is a good
representation of the long-term fade depth as was suggested by
Hause and Wortendyke (1979). These are, of course, single're
ceiver curves and hence do not inqlude diversity improvement.
7. CONCLUSIONS AND RECOMMENDATIONS
The propagation performance of this radio link is of wide
interest primarily because of the length of the path. Relative
ly Li,ttle 8GHz daca for such Lonq links and.;·t"·dver such an
extended period of time have been accumulated to this time,
-35
rtilili1 1''''''- F
ECD-0
C.~
.......G)
>CD
...J
f"-t
ccU)
-,...(f)
-0Q)
>-rotQ)oCD~
..··30
-50
-6~
--7~ -
F-roct i on o~f Time RSL E>-:ceeds Ord i t1ate
Figure 21. Instantaneous received signal level distributionsfor all data collected at Hohenstadt.
36
although a number of long links are currently being investigated.
The lack of a well-developed model for predicting the performance
of such links has made system designers somewhat reluctant to
accept long microwave links as a part of a highly reliable
communication system. At one time consideration was given to
putting an active relay site in the middle of this link, but the
data demonstrate that adequate propagation reliability is
provided without such a relay. In fact, the overall reliability
of communication between the Hohenstadt and Zugspitze terminals
(considering all possible causes for traffic interruption) would
be likely to decrease if a relay were to be interposed between
these sites. The reason for this is that any small gain in
propagation reliability (which could add only about 618 seconds
or 10.3 minutes at the very best during the period of observa
tion) would be offset or far exceeded by time lost to station
equipment malfunctions. While it is true that such links are
possible only in specific physical situations where either one
or both terminals is elevated above the intervening terrain or
where very tall towers can be built, the practicality and
ecomonic utility of such links is more firmly established as
additional measurements are reported.
The conclusions reached as a result of studying the fairly
extensive amount of propagation data available over this link
deserve restating here. One somewhat unexpected conclusion is
that even diversity signals which appear to be highly correlated,
such as the samples shown in Figures 8 and 9, do not necessarily
result in inadequate diversity performance. In the present
instance, even though the signals are visually quite similar at
high levels, the deep fades which depress the receiver input
signal levels below the threshold of adequate performance very
seldom occur simultaneously. This permits the diversity protec
tion switch to maintain traffic continuity even during periods
of severe fading. In the classical sense, it was possible to
achieve good diversity performance in spite of an apparent over
all high correlation between the two signals.
37
The diversity spacing at>e:~ach·'''e·nd of the path was signifi
c an t.Ly idLf fer-errt; .. as was t.he r e La t.Lon..s:t1.~pb~tweenthe instantane-;'0 ", ,-. ':. ... /' .,', . - .. -' ' .. - .. " ,;' ',_ .. ,-~ -«, 1:"" ": ~'_'
ous values ofr;r~.ceivedis i.qnaL .LeveL pai:rsat each end . However,
in spiteof. ;the9~differences, ef f ec t.Lve and nearly equal diver
sity perforIJla:pce.was achieved~
In rev.t;~wi\l)gthe da..ta,.seye~al diff.erent types of fading,,,. II' .. -_,-'.' .•~ ,.,,'. bd C ••::;, j .....i.,/ ./ ...
wer.e observed.~.>(,t~e.... m9s~."~,,l:"~qu~.nt type .... ();f .fadLnq encounte.red
was due to p}:lase int~rf~reJ:)c·e, eit:her two-ray or .multipleJ ,;. ; ..>, .. ,:1< ., ......" .. ,:-':... .. : ~ '-: -,
componerrt, ;!.Tp~/ di.s t.r.i.bu t i.on.jo..f.instantaneous .signal levels
during such p:..eri(),dS\V0UrI~ .. pr'esumabLy...b~. auba tan.t.LaLl.y vdd f f eren t;
from a Ray~e.ig.p. di.st.r.Lb..uti\or-.pu.t the) ~over..a l.], Lonq-rt.e rm distribu
tion is very close to Rayleigh. The values o,frnedian received
signal level (read from Figure 21) are also within one dB of the
predicted "free-space" values shown in Table 6-1.
The analysis of the computer-collected fading data included
a tabulation of the duration of the fades which dropped 30 dB
below the "free space" signal levels and the accumulation of
these data into distributions of numbers of fades for given
duration windows. As was suggested by the CCIR information, the
fade durations are approximately log normally distributed. The
median duration of the fades and the standard deviations of
the fade duration distributions (in dB) of the Hohenstadt
computer-collected data compare favorably with the information
from CCIR documents.
All of the above conclusions support the idea that it is
definitely possible to achieve reliable propagation at 8 GHz on
a link as long as 160 km and with less than normal diversity
spacing. This is not to imply that all such systems will
perform the same way but with careful engineering it is possible
to achieve comparable performance.
38 .
8. ACKNOWLEDGEMENTS
The authors wish to express their appreciation to the
United States Air Force, Electronic Systems Division for their
support of the measurement project. We would also like to
thank the site personnel at Hohenstadt (members of the u.s.Army 252nd Signal Company l vand at the Zugspitze (members of the
U.S. Air Force 1945th Communication Group) for their continuing
assistance in collecting the data. ~ithout their generous help
the project could not have been 'accomplished. Particular thanks
are offered to SGT Peter Moran of 'the 252nd Signal Company who
nursed the computer system at Hohenstadt 'through several
periods of failing health.
39
9 . REFERENCES
Barnett, W.T. (1972), Multipath Propagation at 4, 6, and 11 GHz,Bell System Tech. J., 51, No.2.
CCIR (1978), Recommendations and Reports of the CCIR, 1978 XIVthPlenary Assembly, Kyoto, 1978, Volume V, Report 338-3.
Dougherty, H.T. (1968), A Survey of Microwave Fading Mechanisms:Remedies and Applications, ESSA Technical Report ERL69-WPL4.(NTIS Access. No. COM-71-50288, National Technical Information Service, Springfield, VA 22161.)
Hause, L.G.,and D.R. Wortendyke (1979), Automated Digital SystemEngineering Model, NTIA-Report-79-18.
Mc Gavin, R.E., H.T. Dougherty, and C.B. Emmanuel (1970), Microwave Space and Frequency Diversity Performance 'UnderAdverse Conditions, IEEE Trans. COM-18, No.3, June,261 - 263.
Samson, C.A., ~.A. Hart, and R.E. Skerjanec (1970), WeatherEffects on Approach and Landing Systems, FAA Report No.FAA-RD-70-47.
Thompson, M.C. Harris B. Janes, Lockett E. Wood, and Dean Smith(1975), Phase and Amplitude Scintillations at 9.6 GHz on anElevated Path, IEEE, Vol. AP-23, No.6, November 1975.
Vigants, A. (1975), Space Diversity Engineering, Bell SystemTech. J., 54, No.1.
White, R.F. (1970), Engineering Considerations for MicrowaveCommunication Systems, Lenkurt Electric Company. (GTELenkurt, Dept. C134, San Carlos, CA 94070).
40
APPENDIX
This appendix contains plots of hourly median and inter
decile range values for each of the receivers at Hohenstadt.
The upper trace corresponds to the lower antenna and the lower
trace to the upper antenna in each figure. The ordinate scales
are received signal level in negative dBm corresponding to the
upper edges of the 3 dB wide histogram bins. Apparent discon
tinuities in the early data are artifacts of the data collection
procedure which have proven impossible to remove.
41
Ii5659A2~5~8
51SA
rd 57~ ~ 60~ ~ 65ro (]) G6I ~
.~ 69.... ~ 72~ 75(]) H 78:> (]) 81(]) :3 8AH 0 87H 90..--4rd 93~ 30tJ1 rd 53.~ ~ 56(/) ~ 39
Q) 42ro +l~5Q) ~
:> r::C '8.~ 51Q) H SAU Q) 57Q) P-I 60~ P-I 65::> &&
69121578818l879093
SUU MON TUE WED THU FRI SAT
OCT 9:- 15
Figu:t;e A-I. Hourly histogram dat.a, 1977'.
42
SATFAtB' THU
16"-- 22TUE"
OCTSUN
Jj5659A2A5A851SA5760656669
ii7881SA87909'L- ~-_-.....__....._---- ___.
Ii132iAI51~5760
"666972757881M8790951--__......__.-._..,;;..-__-1-__....... +-__--.. ____
Figure A-~. Hourly histogram data, 1977".
43
SAT~RITUE ED THtJ
OCT 23- 29
coC ~co ~ro (I)I +J
~
" ~~(I) ~
:> (I)(j) ~H 0
H~
CO~tJ"' CO
-M ~(/) ~
(I)
ro +J
~~"(I) ~:> ~
-M(I) ~0 (I)(I) 0...~ 0...
:::>
Figure A-3. Hourly histogram data, 1977.
44
Ji36 ,
~~39A2A5
~~~~~Ju~A851SA
co 57C s:: 60~ s:: 63"d OJ 66I -IJ 69c
.... ~ 72....-4 75OJ ~ 78:> OJ 81(1) :3 SAH 0 87H 90....-4 95cos:: 50tn co I!-H s::U) s:: 59
OJ 42
~~.~.' ~t~\"d -IJ 45OJ s:: 48:> ~
-H 51OJ ~ 5'0 OJ 57OJ o, 60~ 0-& 6'
~ ss69
~7881N879095
SUN "ON ~ e mJ FAI SAT
OCT 30- 31 , NOV 1- 5
FigureA-4. Hourly histogram data, 1977.
45
SAT... ... .
e mJ
6- 12
.. ..'. '."ON Tll
NOV
..SUN
Ii5659A2A5AI51SA5760U69
~7881SA8790956-0- ~.......-~"'-"""""'~.....--------------------I'0'55
"59
i485t5'57II6569
ii7881SA879095....- ......--......~~.......~.,....---- ...........-.--+.......----....----...a
coC ~co ~-o OJI +J
~-.. ~
r-tOJ ~
:> OJOJ ~H 0
H.......I
CO~0' CO
.,-i ~CI.l ~
OJro +JOJ ~
:> F:t:.,-i
OJ ~U OJOJ o,~ 0.-
~
Figure A-5.Hourly histogram data, 1977.
46
B59A2A5A851
m SA~ ~ 57rn ~ 60""d Q) 65I +J 66
~ 69.... ~ 72.-I 75Q) ~
:> Q) 78Q) ~ 81H 0 SA
H 87.-I 90ro 95~ 50tJ1 ro
-..-I ~55
U) ~ 56Q) 39
""d +J A2Q) ~ A5:> ~ AS
-r=I 51OJ ~ SAU Q)Q) P-& 57~ P-& 60
:::> 636669727S788184879095
SUN "ON Tt,E e TN; r:RI SAT
,NOV 1'3- 19
Figure A-6. Hourly histogram data, 1977.
47
"5659A2A5·A851SAco 57
~ ~ 60~ ~ Uro (1)I +J
~ 69.... ~ ir-I
(1) ~ 78:> (1) 81(1) ~ SAH 0 87H 90r-Ico 95~ co IIty\
.,-1 ~ n(J) ~
~(1)
11ro +J(1) ~:> ~ 48
.,-1 51(1) ~ 5'0 (1) 57(1) Pot .0~ Pot 65
~ ;;6972757881SA879095
SUN T\.l WED T~ r:RI SAT
NOV 20- 26
Figure A-7. Hourly histogram data, 1977.
48
I ~~~59A2
.'5A851SA
m 57C c 60~ c 65ro 0) '66I +J 69
.... ~ ~~ 780) ~
:> 0) 810) ~ 8AH 0 87
H 90~ 95m Iic01 m
-ri c nU) c0) iro +J
0) C .q...... ~ 51......
-ri 5&OJ ~ 57o 0)'0) 0.. 60~ 0.. 6'
~ "69
ii7881M87
1&SW ttON TlE e TMJ t:RI SAT
NOV 27- 30, DEC 1- 3
Figure A-8. Hourly histogram data, 1977.
49
Ii5659A2A5AS515'
co 57C ~ 60~ c 63ro OJ 66I +J 69s::
-.. r< 72r-I 75OJ ~ 78:> OJ 81OJ ~ "H 0 87H 90r-Ico 95c Hty\ co
.r-i s:: IICI.l ~OJ 21-o +J
OJ ~ AI:> ~.r-i 51
OJ ~ ~o OJOJ o, U~ o,
::J "69
ii7881..R
UI 1\£ e TJl1 FIt! ~T
DEC 4- 10
Figure A-9. Hourly histogram data, 1977.
50
Jj3659A2A54851SA57&0&5&6&9~7881U879095ijl----------------------------t
ft,<114851sa5760&5
"6972757881U81909!1--__......__.......~-- .......--__I-----....._-__I~-- ...
nl
DECe m1
1J - 17SAT
Figure A-IO. Hourly histogram data, 1977.
51
SUN T\£
DECe 1lIJ
18- 24FRI SAT
Figure A-II. Hourly histogram data, 1977.
52
IiJt~~~"
5659A2A5AS51SA
m 57C ~ 60p:t ~ S5r(j OJ S6
01 +J 69~ 72.... ~
r-I 75OJ H 78:> OJ 81OJ :3 UH 0 87
....::l 90r-I 95m 50~tJl m 55.~ ~ 56Cf.l ~ 59
OJ :iro +JOJ ~ A8:> ~ 51.~ 5£OJ H0 OJ 57OJ ~ II~ o,
~ "697275788tN879095
1\£
DECe mJ
25- 29SAT
Figure A-12. Hourly histogram data, 1977.
53
50555659.&2'5'851
co 5'57E:1 ~ 6O~ ~
'"d OJ 6SI +J 66
~ 69-.. ~ 72
M 7!OJ ~ 78:> OJ 81C) :3 .,H 0
H 87M toco 9S~ SOtJ1 co SS
-r-i ~ .6(l) ~ SIOJ'"d +J 42OJ s:: '1:> ~ ••-r-i 11OJ ~ "0 OJ 17OJ o, CO~ o, IS::J
"It72717181..8710IS
TUE .a TNU ntI SAT
~AN 1- 7
Figure A-13. Hourly histogram data, 1978.
54
5055565942A5A851!.t
co 57S s:: 60
CO s:: 65res OJ 66I -W 69s::--. ~ 72
......-I 7!OJ H 78:> OJ 81OJ :3 SA...:l 0 87
...:l 90......-I 95cos:: SOty\ co 55
-..-I s:: 56U) c 59
OJ A2'"d +J .5OJ s:: AI> ~ II-..-I
"OJ H0 OJ 17OJ ~ 80~ c, 85
::J 818'727S71II..179015
SUN TUE "!D mJ 'AI SAT
~AN 8- lA
Figure A-14. Hourly histogram data, 1978.
55
5055..- lp5659
MA2A!AI51!SA
co 57c-'
~ 10l-I
P:l ~ ISro Q)
"I +J I'~ 72~ ~
rl 71Q) ~ 71:> Q) IIQ) :3 letH 0 17
H 10rl ISco 10~b1 co II
.r-1 ~ •U) ~ II~ tQ) Qro +J 41
Q) ~ • t,:> ~ II.r-1 Mill ~ 170 Q)Q) o, 10~ o, II
::J II
"727171II..•710IS
~ ~ • mJ "'1 SAT
~AN 15- 21
Figure A-IS. Hourly histogram data, 1978.
56
505S• ..5942'5..51
"co 57~ s:: 10~ s:: IS'""d (l) 81I -IJ n
~ 72.... ~ 71r-I(l) ~ 71:> (l) I.(l) ~ ..
J--=l 0 17J--=l 10
r-I IIco SO~tJ1 co n
•...-1 ~ ..(,t) ~ II
OJ Q •'""d -IJ •OJ ~ •:> ,::J:: II-...-I ..
OJ ~ 17o OJOJ o, .0c~ o, II
::J ..II721171II..1710• ... ... . ... ... ...
IUN lION TUI MID TIll 'II SAT
'-'AN 22- 28
Figure A-16. Hourly histogram data, 1978.
57
5055565942.IS...51SAco 57
C ~ 60~ ~ 65ro OJ 66I 4J
~ C9" ~ 72
M 7!OJ ~ 71~ OJ IIOJ :3 1£H 0 17H 10Mco IS~ SOty\ ro SS
-r-f ~ •U) ~ nOJ 42ro 4J .IIOJ~ •~ II-r-f
OJ ~ let0 OJ 17OJ o, -60~ P-t IS
0 IIn727171IIlet1710IS
111111111111.11
. . -. -. . .. ~
G "., r:R1 SAT.... 'tON 1\£
JAN 29- 31' ~EB 1- 4,
Figure A-17. Hourly histogram data, 1978.
50
50555659 ..A2'5...51N
ro 57C ~ 80CO ~ 8SrQ Q)
"I +> 81~ 72.... ~
r-I 71Q) ~ 71:> Q) IIQ) :3 ...-:I 0 17
H "r-I IIro~
10tJ1 ro II
-r-I ~ II(/) c II
Q) QrQ +> •OJ ~ •:> ~ 11-r-I ..Q) ~ 17o Q)OJ 0.. .0~ o, IS
:::> II.1721171II..1710I' .. . .. ..
SUN .. T\E • ,.., ,..1 SAT
~EB 12-- 18
Figure. A-18. Hourly histogram data, 1978.
59
5055S659.t2'5.q51
co !'S s:: 57~ s:: 60-o OJ 65I 4J "s:: 69
r< 72.-I 7!Q) ~ 78:> Q) 81Q) ~ ..~ 0
~ .7.-I 10co ISs:: SOtJ"' co SS
-M s:: ..(J) s:: IIOJro 4J QQ) s:: •:> ~ •-M 11(1) ~ ..U Q) 17C) 04 10p~ 0... II::> II
II721171II..•710IS
SUN. ... ... ...... '- -
ntl SATT\E tIED TIll
~EB 26".. 28, MAR 1.. 4
Figure A-19. Hourly histogram data, 1978.
60
50555S ...59~.&5...51
co 5'57S ~ 10CO ~ IIra OJ
I +J "~ II... ~ 72M 7!OJ ~ 78~ OJ 81OJ ~ It~ 0 87HM 10co II~ SOb"' co SS
-,-1 ~ IIU) ~ II '.11'11111
OJ Qr"(j +JOJ ~ .1~ ~ ..-~ 11OJ ~ Mo OJ 17OJ o, 10~ ~ IS:::> II
II721171I'..1710IS
q .' "., "1 SAT
MAR 12- 18
Figure A-20. Hourly histogram data, 1978.
61
coC ~~ ~-o Q)I +J
~... ~r-I(I) ~
:> (I)Q) ~H 0
Hr-I
CO~t;'\ CO
en ~U) ~
(I)
ro +J(I) ~:> ~en(I) ~
0 (I)(I) Pot~ Pot
::J
SUN ~
MAR• 11lJ
19.. 25"1 SAT
Figure A-21. Hourly histogram data, 1978.
62
505·5
~S859~
'5..51
"ro 57S C 10~ C ISro OJ "I -+J 19C
-.. ~ 72~ 7SOJ ~ 78:> OJ 81OJ ~ ..H 0 17
H 10~ ISroC SOb1 ro SS
-n C H(J) c II
OJ .t2ro -+J .1OJ C •~ i<C 11-nOJ ~ ..U OJ 17OJ o, 10~ o, II
:::> IIn72117.II..1710IS
SUN 'tON
MAR1\& lIED
26- 31,mJ
APR 1SAT
Figure A-22. Hourly histogram data, 1978.
63
cdS s::r::Q s::ro OJI +J
s::" ~
..-IOJ H:> OJOJ ~H 0
H..-Icds::t;l cd
.r-I s::U) s::
OJro +JOJ s:::> t<
.r-IQ) HU OJOJ ~p:: o,
p
T\a
APR..,2 ..
SAT
Figure A-23. Hourly histogram data, 1978~
64
5055581942A5.q515'
co 57~ s:: 60~ s:: 6Sro OJ 68I +J n
-. ~ 72M 71OJ ~ 71:> OJ IIOJ ~ ..H 0 17
H .0M IScos:: 1001 co II
-,.-i s:: ..U) s:: II
OJ 42ro +J .-OJ s:: •:> r:::C 11-,.-i ..OJ ~
U OJ 17OJ ~ 10~ o, IS
::> ..II721171II..17to..
~ • "., "1 SAT
APR 9- 15
Figure A-24. Hourly histogram data, 1978.
65
SAT
-
11111111111.1111
• TIll
16- 22T\E
APR
soSSS859Q
'5,.51
"co 57E: s:: SOCQ s:: SSro Q) •I +J ns:: 72.... r<: 7S....-IQ) 5--1 78:> Q) IIQ) ~
.,...:J 0 17
...:J 10....-I ISro 10s:: II0" ro-M s:: IIU) s:: ..
Q) Q'"d +J •Q) s:: ,.:> .<: 11
-M ..Q) 5--1 17o Q) 10Q) o,~ o, IS
::> IIII727S71II.,.790IS
SUN
Figure A-25. Hourly histogram data, 1978.
66
50 •55 •5851A2'5AI515'
co 57~ ~ SO~ ~ SSra Q) 18I +l 'I~ 72.... e<:
.-I 71(]) ~ 71~ (]) IIQ) ~
.,H 0 17
H .0.-I 'Sco SOCt;l co IS
-..-l C II ...-..•(J) c It(]) Q
'"d +l .1OJ C ..~ ~ 11
-..-l ..(]) ~ 17o (])(]) P4 10P:: P4 IS
p IIII721171II..1710
" T\E • TIll ,ntl SAT
APR 23- 29
Figure A-26. Hourly histogram data, 1978.
67
5055565.Q••51Set
co 57S ~ SO~ ~ SS""d OJ 18I 4J
~ a-.. ,::x:: 72
r-t 71OJ ~ 71:> OJ IIOJ ~ ..H 0 17H 10r-Ico IS~ 10tJl co II
.r-I c ..Cf.) ~ II
OJ Q""d 4J .-OJ ~ •:> ,::x::.r-I II
OJ ~ No OJ 17OJ 0-1 10
c.::: 0-1 I::J II
..APR
~ • mJ
30- 30, MAY"1 SAT
1 - 6
Figure A-27. Hourly histogram data, 1978.
68
10ISII
- 19.t2'1...II!'
co 17C ~ 80CO ~ 8Sres (1) 88I +J n
~~ ~ 72
r-f 7S(1) ~ 78:> Q) II(1) ~ ..H 0 87
H 10r-f ISco~ SOtn co SS
.r-I ~ ..(J) ~ II
(1) Qro +J 41(1) ~ ..:> t==:C II.r-I ..(1) ~
U (1) 17(1) P-i 10P:: P-i II
::J IIII721171II141710IS
T\& G TIll ,..1 SAT
MAY 7-- 13
Figure A-28.Hour1y histogram data, 1978.
69
50555659.&2A!5AI!5t!SA
co 57S s:: 60~ s:: 65ro (1)
"I +J 69s:: 72, ~ 7!M 78Q) H:> Q) 8tQ) ~ 8AH 0 87
H .0M .1co 10s:: IItJ" co•..-1 s:: •(J) s:: II
Q) Qro +J atlQ) s:: ...:> ~ 11
•..-1
"Q) H 17o Q)10Q) o,
p::; o, II::J "It
721171.1let17to'I
1\E 11lI r:RJ SAT
MAY 20
Figure A-29. Hourly histogram data, 1978.
70
SO515851.t2'5AI515'
co 57S s:: 10t:Q s:: IIro OJ "I +J ns:: 72-. ~ 7SH 71OJ H:> OJ 81OJ :3 ..H 0 87
H .0H I.m SOs:: I.ty\ co.~ s:: •Cf) s:: II
OJ 42ro +J •OJ
~..
:> II.~ ..OJ H 170 OJ .0OJ o,~ ~ a
:::> ..II721171II..1710I.
~ .. TMJ ,.1 SAT
MAY 201" 27
Figure A-3D. Hourly histogram data, 1978.
71
50555851A2£5.t851!£
ro 57S ~ 10~ ~ IS-o (J)
"I +J II~ 72--. t<
rl 7!(J) ~ 71:> (J) II(J) :3 Pt-=l 0 17
t-=l .0rl ISro~ SOtyl ro IS.~ ~ 18U) ~ 59
(J) Qrc:s +J £1(J) ~ •:> ,:::C 11.~
M(J) ~
U (J) 57(J) 0.. 10~ 0.. IS
::J IIII727S78II..17toIS
SUN T\E .. ,...,"I SAT
MAY 28- 29
Figure A-31. Hourly histogram data, 1978.
72
i~56 1.59
~AS
m ~S ~
57co ~ro (1)I -t-J
~... ~r--t(1) ~:> (1) 1(1) ~H 0 7Hr--t 0m 9~b1 m
o,-f ~
~.~U) ~(1)
ro -t-J .t(1) ~:> ~ AS
or-f(1) ~0 (1)(1) ~~ o,
::J
TLE
NOVlIED
12-mJ
18SAT
Figure A-32. Hourly histogram data, 1978.
73
I ~~co
S c~ cro (1)I oW
C... t<CrlQ) ~
:> Q)Q) ~H 0 7HrlcoC ,.C1 co.,.., cU) c
Q) t-o oWill C:> F::t:.,..,Q) ~
0 Q)(]) o,~ 0-1
t:J
• mJ
19- 25SAT
Figure A-33. Hourly histogram data, 1978.
74
I ~."i
cd~ ~CQ ~-o (1)I -W
~... ~~
(1) ~
:> (1)(1) ~H 0
H~
cd~tJ1 cd
-ri ~(J) ~
(1)
ro -W(1) ~
It:> ~
-r-i(1) ~
0 (1)(1) ~p::; o,
~
78
I'OJ TlE B nIJ At! SAT
NOV 2S- 30, DEC 1.. 2
Figure A-34. Hourly histogram data, 1978.
75
m8 ~p:) ~'""d illI +>
~.... ~~
OJ ~
:> ill(]) :3H 0
H~
m~bJ ro
-.--I ~U) ~
ill'""d +Jill ~:> r<
-.--Iill ~
0 ill(]) o,~ o,
t::J
IiIi
5tH T\.E
DECnu
9Atl SAT
Figure A-35. Hourly histogram data, 1978.
76
co8 ~~ ~'"d (1)I +J
~--. r<C
r-t(JJ ~
:> (JJ(]) ~H 0
Hr-tco~tJI co
.r-! ~U) ~
(1)
ro +J(1) ~:> ~
.r-!(]) ~
0 (1)(]) o,~ o,
~
I
1\£
DECB mJ
24- 30SAT
Figure A-36. Hourly histogram data, 1978.
77
nj
c ~~ ~ro Q)I -W
C, r<r-IQ) ~
:> Q)Q) ~H 0
Hr-Inj~b1 nj
.r-! ~vI) ~
Q)ro -WQ) C:> t<t:
.r-IQ) ~0 Q)Q) o,~ o,
::J
T\.E
DECe31
SAT
Figure A-37. Hourly histogram data, 1978.
78
!~~",llHtNlftIlIINIHIIflIINllfIttIIHHIt~~'~ss
39A245485154m 57C ~ SOt:Q ~
.~'"d Q)I +J
~ 69-.. r<C 72~ 75Q) H 78:> Q) 81Q) :;: SAH 0 87H~ 90m 93~ IItJ1 m
.,.-1 ~(J) ~
Q)
~'"d +JQ)
~:> 48.,.-1 51Q) H
~U Q)Q) c, 60~ o, 63
~ 6669
~7881SA87
~SUN MON· 1\£ 'lEO mJ F'RI SAT
JAN 1- 6
Figure A-38. Hourly histogram data, 1979.
79
mS ~~ ~r(j Q)I ..IJ
~-.. r:::C
r-IQ) ~
:> Q)Q) ~H 0
Hr-Im~OJ m·n ~U) s::
Q) ir(j ..IJQ) ~ 48:> r:::r:·n 51Q) ~ I;0 Q)Q) o,
p.-:; o, II~
~78818'
ti~ ED
7-1llJ
13SAT
Figure A-39. Hourly histogram data, 1979.
80
i~5659
!i1
ro
I8 ~~ ~ro OJI -JJ
~.... F::t: ~r-I
OJ H 78:> OJ
ItOJ ~~ 0
~r-Iro~bJ ro.,.,
~U) ~
OJro -JJ .tOJ ~ Ai:> F::t:
.r-fOJ H0 OJOJ o,~ o,
~
~H
1\E e TMJ r:RI SAT
JAN 14- 20
Figure A-40. Hourly histogram data, 1979.
81
co~ ~P=l ~ro OJI +J
~--. ~
r-tOJ ~
Ii:> OJ(1) ~H 0
Hr-t
CO~tJ1 CO
-..-I ~CJ) ~
OJro +JOJ ~:> t<I::
-..-IOJ ~ 70 OJOJ o,~ o,
::::>
78
11H
TlE e 00 FRI SAT
JAN 21- 27
Figure A-41. Hourly histogram data, 1979.
82
j~5639
"~485154
rd 57~ ~
~co ~rcj OJI +J S9
~
~.... ~~<J) ~ 78:> OJ H1OJ ~H 0 16H..-I
Ird~OJ rd.~ ~(J) ~
OJ~"d +J
OJ ~ 48:> ~ 51.r-I
~(]) ~
0 OJOJ O-! 6~ P-i
:::> 66S9
~78
H1
tis THJ r:RI SAT
310 , ~EB 1- 3
Figure A-42. Hourly histogram data, 1979.
83
cd~ ~
P-1 ~r(j (J)I +J
C,~
r-f(J) H:> (J)(J) :3H 0
Hr-fcd~b1 cd.,..,
~(J) ~
(J)r(j +J(J) ~:> ~.,..,(J) H0 (J)(J) O-!~ O-!
::J
I9
1148
~
I78
11ti-----+----of-----+----+----I------Io----J
ED
4-THJ
10SAT
Figure A-43. Hourly histogram data, 1979.
84
fi5659
i48
~co Is ~
co ~ro Q)I +J
~ i--. ~r-IQ) H:> ,Q)
IiQ) ~H 0
Hr-I 9co~Q"1 co
-r-f ~(J) ~
Q) ,ro +JQ) ~ AS:> ~
-r-fQ) H0 Q)Q) O-!~ O-!
:::>
i
ITMJ r:RI SAT
17
Figure A-44. Hourly histogram data, 1979.
85
coS c~ C'"d OJI 4J
C"- F:C
HQ) ~
:> Q)Q) ~H 0
HHCO~ty\ CO
-r-1 CU) C
Q)'"d 4JQ) C:> ~
-r-1Q) ~CJ Q)Q) 0-4P:: o,
~
~
itIi
e 00 r:RI SAT
18- 24
Figure A-45. Hourly histogram data, 1979.
86
Figure A-46. Hourly histogram data, 1979.
87
J~5S39A2454851
rd5457
S s:: 60r:o s:: !ares Q)I +J
;a 69.... 72
....-t 75Q) ~ 78:> Q) 81(l) ~ 84H 0
H 87....-t 90rd 95~ 50tJ" rd i-M ~(J) ~ 39Q)res +J A2Q) .~ 45:> 1':4 48
-M 51(1) ~ 54o Q) 57(1) 0.- SO~ 0..
:::> S3S6S9
~788:84879C95
SJ; Mar; T'i ED' ~ FR: SAT
MAR 18- 24
Figure A-47. Hourly histogram data, 1979.
88
jj3639A245485154
ro 57C ~
H~ ~'"'d OJI +J
~
~, ~HOJ H 78:> OJ 81(1) ~ 84H 0 36HH 93ro~ Itn ro t---r-i ~
(.1) ~OJ
~'"'d +JOJ ~ 48:> r:::c: 51-r-iOJ H ~o OJOJ o,
~~ ~::J
S9
~78818487§i
SJ; lEO ~ r:R: SAT
MAR 25- 31
Figure A-48. Hourly histogram data, 1979.
89
!~5659
~485154
co 57~ ~ 60(1:) ~
~ro Q)I +J 69
.... ~ 72~ 75Q) H 78:> Q) 81Q) ~ 84H 0
~6~r-Ico 93~ Ity\ co
.r-1 ~CJ) ~
Q)
~rcJ +JQ) ~:> r< 48
.r-1 51Q) H 5lo Q) 57Q) o,
~~~ o,::J 66
S9
~78818.4~79!
SJ; !1Or~ ~ lED ';'~ FR! SAT
APR 1- 7
Figure A-49. Hourly histogram data, 1979.
90
coC cco cro (1)I +J
C-- ~
,....-4(1) H:> (1)(1) ~
I--=l aI--=l
,....-4cocOJ co
-..-I CU) C
(1)I"'{j +lQ) C:> .<:
-..-IQ) H0 (1)Q) c,~ c,
~
,.,~
APRtEO mJ
8- 14SAT
Figure A-50. Hourly histogram data, 1979.
91
!~5659
i48
~co 57
~ ~ I,:Q ~"d OJI +l- ~
~... ~,...-I 78(I) H:> OJ 81(1) ~ 84H 0
~6~,...-I 95ct1s::01 co
-r-I s::(/) ~
OJ"d +J ,(I) ~ 48:> .:::r: 51-r-I(I) ~ HCJ OJ
I(]) 0.,~ 0..
p
~18
11Ii
S); ~,; £ ~. ~: SA~
APR 15- 21
Figure A-51. Hourly histogram data,1979.
92
t'"~57co
~C s::~ s::'U OJI -iJ
~s::.... ~~ 78OJ f...4 91:> OJ(1) s:...:I 0 36H~ 9cos::ty\ co
-.-I s::U) s::
OJres -iJ ,OJ s:: AS:> ~ 51-..-I
~OJ f...4U C)OJ o,
II~ o,::>
~78
11~
SJ; ~t: Tt-t: EO TJ.l' ~': SAT
APR 22- 28
Figure A-52. Hourly histogram data, 1979.
93
F"RI SAT
1- 5T1~ WED iHJ
29- 30, MAYMON
APR
j~36394245485154
co 5760
~ c s~ cro OJI +J 69
c 72, ~ 75r-I 78OJ ~
:> OJ 81OJ ~ 84~ 0 87
~ 90r-I 93co
IctJ1 co
. ..-1 CU) C
OJ A2-o -iJ' 45OJ c 48:> r:::C
-..-I 51Q) ~ 54o OJ 57OJ o, 60rt.: o, 63
::::>- 6S69
~78818487§~
SJ~
Figure A-53. H~ur1y histogram data, 1979.
94
mC ~t:Q ~'U (I)I +J
~.... r=::t:
MQ) ~
:> OJQ) :::t-=I 0
t-=IM
CO~C"' m-n ~U) ~
Q)'U +JQ) ~:> ~-nQ) ~U Q)Q) c,
(Y; 0..:..::>
TlE
MAY 6- 12SAT
Figure A-54. Hourly histogram data, 1979.
95
jj5639
~48515457co 60S s::~~ s::ro Q)
I +J 69s:: 72... t=:I:: 75
r-f 78Q) ~ 81:> Q)84Q) ~
H 0 ~6t-=l~ 93co
Ds::b1 co
.r-I s::(J) s::
~Q)ro +JQ) s:: 48:> t=:I:: 51.r-IQ) ~ ~o Q)Q) ~ !it.Y: o,
:::> SS69~788184
IiSv"; 'MOtJ TLE e mJ r:RI SAT
MAY 13- 19
Figure A-55. Hourly histogram data, 1979.
96
Figure A-56. Hourly histogram data, 1979.
97
FORM OT-29(3-73)
u.s, DEPARTMENT OF COMMERCEOFFiCE OF TELECOMMUNICATIONS
BIBLIOGRAPHIC DATA SHEET
1. PUBLICATION OR REPORT NO 2. Gov't Accession No.3. Recipient's Accession No.
NTIA Report 79-31
4. TITLE AND SUBTITLE
PERFORMANCE OF A VERY LONG 8 GHzMICROWAVE LINK
s. Publication Date
November 19796. Performing Organization Code
NTIA/ITS
9103479
10. Contract/Grant No.
9. Project/Task/Work Unit No.7. AUTHOR(S)
J.E. Farrow and R.E. Skerjanec8. PERFORMING ORGANIZATION NAME AND ADDRESS U. S. Departmentof Commerce, National Telecommunications andInformation Administration, Institute for Tele-~----------------------~
communication Sciences, 325 Brdwy.,Boulder, Colorado 80303
11. Sponsoring Organization Name and Address 12. Type of Report and Period Covered
Electronics System DivisionHanscom AFB
114. SUPPLEMENTARY NOTES
NTIA Report13.
15. ABS!RACT (A 2?O-word or less factual summary of most signiti cent information. If document includes a significantbi llz()~ra(Jhy at lit ereture survey, mention it here.)
This paper discusses measurement of signals received over a l60-km(IOO-mile) line-of-sight microwave radio link operating at 8.4 GHz.Signals were recorded at both ends of the path at various timesover a period of about 18 months. The effects of macro-scalemeteorological conditions are discussed particularly those whichare peculiar to the central European location of the path. The dataIobtained from this measurement effort are presented both in termsof radio link availability and in terms of long-term propagationperformance. The data were also analyzed to obtain distributions offade duration for single receivers and for both receivers in adiVersity configuration. The study supports the idea that it ispossible to achieve reliable propagation at 8 GHz on a link as longas 160 km and with less than normal diversity spacing.
16. Key works (Alphabetical order, separated by semicolons)
17. AVAILABILITY STATEMENT 18. Securi rv Class (This report) 20. Number of pages
~ UNLIMITED. Unclassified 106
o FOR OFFICIAL DISTRIBUTION.
19. Security Class '(This page)
Unclassified
21. Price:
l'SCOMM-OC 297 t C5-P73
* U. S. GOVERNMENT PRINTING OFFICE 1979 - 682-2T2/170 Reg. 8