NTIA-REPORT -79-14

Aircraft Obstruction of
Microwave Links

R. E. Skerjanec
R. W. Hubbard

u.s. DEPARTMENT OF COMMERCE

Juanita M. Kreps, Secretary
Henry Geller, Assistant Secretary
for Communications and Information

January 1979
 TABLE OF CONTENTS
Page
ABSTRACT 1

1. INTRODUCTION 2

2. PHYSICAL DESCRIPTION 3

3. MEASUREMENT CONFIGURATION 10

4. RECEIVED SIGNAL LEVEL ANALYSIS 20

5. IMPULSE RESPONSE ANALYSIS 29

5.1. Hartsfield-Atlanta International Airport 34
5.2. O'Hare-Chicago International Airport 38

6. CONCLUSIONS AND RECOMMENDATIONS 44

7. ACKNOWLEDGMENTS 47

8. REFERENCES 47

APPENDIX - Explanation of Impulse Test Data 48

REFERENCES 57

iii
 LIST OF FIGURES
Page
Figure 2.1. Plan view of Hartsfield Atlanta
International Airport. 4

Figure 2.2. Atlanta transmitting antennas. 5

Figure 2.3. Atlanta receiving antennas on control

tower. 6

Figure 2.4. Profile of Atlanta path. 7

Figure 2.5. Plan view of O'Hare Chicago International

Airport. 8

Figure 2.6. Chicago transmitting antenna configuration. 9

Figure 2.7. Chicago receiving antennas. 11

Figure 2.8. Chicago microwave path. 12

Figure 2.9. Profile of Chicago path. 13

Figure 3.1. Block diagram of the PN probe trans~

mitter at 8.6 GHz. 15

Figure 3.2. The theoretical line spectrum of a

pseudo-random binary signal. 15

Figure 3.3. The measured power spectrum (in dB) of the

transmitted signal of the PN probe operating
at a clock frequency of 150 MHz. The
center frequency is 8.6 GHz, and the
frequency scale is 50 MHz per division. 17

Figure 3 .4. Block diagram of the PN probe receiver. 19

Figure 4.1. Aircraft fades for select periods. 21

Figure 4 . 2 . Aircraft fade (Atlanta) . 23

Figure 4 .3. Aircraft position in microwave path. 24

Figure 4 .4. Aircraft position above microwave path. 25

Figure 4.5. Aircraft position below microwave path. 26

Figure 4 . 6 • Long duration fade at Chicago. 28

Figure 4 .7. Aircraft orientation at Atlanta. 30

iv
 Page
Figure 4.8. Multiple short duration fades at
Chicago. 31
Figure 4.9. Multiple'mixed fades at Chicago. 32

Figure 4.10. Time expanded fade at Chicago. 33

Figure 5.1. A typical response measured over the

reflected (Channel B) link at the
Atlanta airport without aircraft
interference. 35

Figure 5.2. The impulse response measured at the

Atlanta airport with a DC9 aircraft in
the microwave beam. The time separa-
tion in (a) is 100 ms. The data rate in
(b) 50 impulses per second. 35

Figure 5.3. Examples of the impulse response measured

at the Atlanta airport with 727 aircraft
intercepting the microwave beam. In (a)
the aircraft was approaching Runway 27R,
and in (b) it was approaching Runway 27L.
The time between responses is 100 ms in
each case. 37

Figure 5.4. A sequence of impulse responses measured

over the Chicago-O'Hare RML link with
no aircraft interference. 37

Figure 5.5. A typical response measured over the

Chicago-O'Hare link with an aircraft
obstruction. 40

Figure 5.6. A sequence of impulse responses measured

during an obstruction fade. Multipath is
seen in the fourth trace from the top. 40

Figure 5.7. Multipath observations over the Chicago-

O'Hare link. 42

Figure 5.8. An obstruction fade without multipath. 43

Figure A-I. Time and frequency domain displays. 49

Figure A-2. The theoretical transfer function of a

multipath channel. 52

v
 AIRCRAFT OBSTRUCTION OF MICROWAVE LINKS

R.E. Skerjanec and R.W. Hubbard*

Conversion to digital transmission has renewed the

concerns about what effects aircraft obstruction of
microwave links have on user quality. This is of
particular concern where it is necessary to install a
telecommunication system that crosses runways and
taxiways where the frequency of obstruction may be
great.

A limited measurement program at 8 GHz at Atlanta

and Chicago Airports was undertaken to determine
if a condition existed that could cause excessive
error rates on digital systems. Measurements were
made of the received signal level together with
the impulse response of the transmission medium.

Measurement results indicate that during takeoff and

landing, aircraft can cause signal level fades to
20 dB. Modern system margin is usually sufficient
to cope with such fades.

The impulse response measurements at Atlanta did not

reveal any delayed or distorted pulses that would
indicate excessive multipath and frequency selective
fading. However, slight distortion from taxiing
aircraft at Chicago was observed. The implication
of this detected distortion is reflected in the
potential of a distorted asymmetrical frequency
response within the 15 - 20 MHz passband of a micro-
wave digital receiver.

Key words: Microwave propagation; impulse response;

digital transmission; aircraft
obstruction

*The authors are with the Institute for Telecommunication

Sciences, National Telecommunications and Information Admin-
istration, u.S. Department of Commerce, Boulder, Colorado 80303.
 AIRCRAFT OBSTRUCTION OF MICROWAVE LINKS

1. INTRODUCTION

Recurrent situations arise when it is necessary for a

variety of reasons to establish a line-of-sight microwave link
at or near airports. Engineering questions often arise because
of this situation. Do aircraft, when they are near the micro-
wave link, cause a disturbance in the transmission of informa-
tion? If so, what are the engineering trade-offs available to
the designer? The superficial answer may be that there is
little or no effect on analog information transmission when the
user terminal has a longtime constant or persistence or the
refresh rate is not critical. This .type of effect may not be
important but because of the apparent lack of published data, it
is quite possible that expensive overdesign has resulted.
Recent trends in wideband microwave transmission have been
toward digital transmission. Data rates of interest are in the
12 to 50 Mb/s range. If such a system were implemented across
a runway,the expected short disturbances could have a more
profound effect than they might have with analog information.
The Army Communications Command contracted the Institute
for Telecommunication Sciences to perform a limited measurement
program to obtain data on an established system and to form
some conclusion on whether a design problem exists and what its
magnitude may be or whether it could be ignored. The basic
question to be addressed was whether a large aircraft in its
normal operation was sufficient to cause near total microwave
beam blockage or sufficient multipath to cause phase and
frequency distortion to result in a significant disturbance
in the information baseband.
A measurement program was undertaken at two locations.
The international airports at Atlanta, Georgia and Chicago,
Illinois both had microwave links crossing active runways and
taxiways. Permission was granted by the Federal Aviation
Administration Regional Offices to utilize these links.

2
 2. PHYSICAL DESCRIPTION

The microwave links of interest were established to remote

airport surveillance radar (ASR-7) video to the control tower
for use in airport control. A plan view of the microwave links
at the Atlanta airport is shown in Figure 2.1. The direct path
goes from the ASR-7 radar site to the control tower and is
0.8 miles (1.3 km) in length. The indirect route goes from the
ASR-7 site to a passive reflector on a water tower at a height
of 40 ft (12 m) and then to the control tower. Those distances
are approximately 2.6 (402 km) and 2.3 miles (3.7 km) respec-
tively. The transmitting antenna is approximately 19 ft (5.8 m)
above ground level and the receiving antenna on the control tower
is approximately 265 ft (80.8 m) above ground level. A view of
the transmitting antenna is shown in Figure 2.2 and the receiv-
ing antenna in Figure 2.3. The relative locations between the
runways and taxiways and the microwave terminals are shown in
Figure 2.4 for the link that goes directly from the radar to the
control tower. In examining this figure, one should be aware of
a large vertical exaggeration and in fact the left-most runway
is only 90 ft (27.4 m) below the microwave beam. For perspec-
tive, the tail height of a-Boeing 747 aircraft is 63 ft (19.2 m)
above ground.
The physical orientation of the two links is such that
the probability of the same aircraft intersecting both micro-
wave routes is nearly zero. When aircraft land from the east,
the approach may intersect the indirect route but aircraft
taking off from the east are normally not airborne at the direct
microwave route. If operations are from the west, the landing
aircraft are normally down and rolling below the microwave beam,
whereas the aircraft taking off will on occasion intersect the
direct microwave beam.
Similarly, the Chicago, Illinois airport is shown in
Figure 2.5. The transmitting antennas are shown in Figure 2.6.
These antennas have a horizontal separation of approximately
95 ft (29 m) providing horizontal space diversity rather than

3
~

SCALE 1: 24 000

=-=;;
l
e-+3 e=-+3
0 1 MILE

Figure 2.1. Plan view of Hartsfield Atlanta International

Airport.
U1

Figure 2.2. Atlanta transmitting antennas.

Q'\

Figure 2.3. Atlanta receiving antennas on control tower.

 1~250 ---".,
CONTROL
TOWER

120B --

1150
-1
(J)
~

I
>
0 11 r2Ilt?J
.0
I)

-.J ."
4.
(
.~

105~
t:
0
...,
orf

•> 10210
.....
W
R T
R-Runway
--
R T
---
R T
T-Taxiway

9521 21 0.5 1. fZJ 1. 5 2.0 2.5 3. fZJ 3.5 4. fZJ 4.5 5. f(J

Pc~h Di.~ano. X 100~f~

Figure 2.4. Profile of Atlanta path.

 o
9l MM

RTR-ORO e

022R lOC

). ~A~S(R. · ·0· · . ·
27ll0C

./_ "'O~ 7"'""w,

'£~~ -- -

Figure 2 .5 Plan v·lew~ of O'Hare

8
~

Figure 2.6. Chicago transmitting antenna configuration.

route diversity described above for the Atlanta airport. The
receiving antennas are the two side-by-side dishes near the top
of the tower in Figure 2.7. A view of the path is shown in Fig-
ure 2.8 from ground level near the transmitting terminal. The
radar microwave link (RML) tower can be seen behind the center
of the aircraft. A profile view is given in Figure 2.9 showing
the relative locations of the runways, taxiways, and antenna
terminals. The runway on the left of the figure is approximately
40 ft (12 m) below the microwave path. This provides the oppor-
tunity for the tail sections of the jumbo-jet class of aircraft
to intersect or partially obstruct the microwave path.
The path is so oriented that it crosses the runway near
the middle as shown in Figure 2.5. Operations from either
direction would affect the system in a similar way. That is,
aircraft taking off would normally be airborne by the time they
reach the microwave path. They normally have sufficient height
to clear the path, although a small percentage do intersect. On
the other hand, aircraft landing have touched down by the time
they reach the microwave path but the clearance is such that
there would be some effects from the aircraft tails.

3. MEASUREMENT CONFIGURATION

With permission and access received from the Federal

Aviation Administration, measurements were made at the Chicago
and Atlanta commercial 'airports. The measurements were made
from a twofold approach. The first was to gather path amplitude
variability data from the AGC circuits on the existing RML-6
systems. The purpose of this segment was to obtain the path-
loss data over a dynamic range commensurate with a standard
microwave receiver and to obtain these data within the limita-
tion of the AGe response time of the receiver system.
The second approach was to utilize a channel probe that is
a separate stand-alone element containing the transmit and
receive functions and requiring only the use of the transmis-
sion lines and antennas of the existing FAA system.

10
Figure 2.7. Chicago receiving antennas.
11
 00

12
 92J~

I
in
~
850

•>
0
..f1
0
..., 8~0
ta-
RML
....c I TOWER
~
u-:s c
....0
~
0
750

W
•>
.-f
~

R-Runway
T-Taxiway
7~0

R T T
-..

650 l
0
I
t

1 2 3 4 5 6 7 8
Pa~h Di.~ano. X 1000~t

Figure 2.9. Profile of Chicago path.

 The received signal level data and the baseband noise were
recorded on an FM analog tape recorder and monitored by an
appropriate strip chart recorder. The tape recorder was capable
of recording changes of 625 Hz. This can be translated to a
time constant of 1.6 ms for full scale changes. Measurements
were to be made for a one-week period at'each of the two air-
ports. Recording of the parameters was accompanied by visual
observation of the various aircraft at Atlanta. The recording
facilities at Chicago were beyond sight of the runways so visual
observations could not be made.
A wi~eband channel probe system developed at ITS for eval-
uating multipath conditions in microwave and other radio propa-
gation links was used in the test program (Linfield et al., 1976).
Th~s system transmits (in the normal operating mode) a
wideband pseudo-random binary bit stream as a test signal,
which results in a radiated signal at 8.6 GHz that resembles
psuedo-random noise (PN). The binary code is generated from
a 9-stage shift register with feedback, providing an optimal
length psuedo-random stream of 511 bits. The bit stream is
clocked at a rate of 150 MHz.
The following is provided to convey the probe operational
details, including power budget, spectrum and frequency assign-
ment considerations.
Figure 3.1 shows a block diagram of the psuedo-random noise
(PN) probe transmitter. The reference source is a 5 MHz stable
oscillator that is used to develop both the IF (600 MHz) and
the rf radiated signal.
The 600 MHz IF signal is bi-phase modulated with the PN
binary sequence, which is clocked at f == 150 ~1Hz rate.
The 5 MHz reference frequency is multiplied to an rf of
8 GHz, and is mixed in a double-balanced circuit with the
600 MHz IF signal. This produces a do~ble~sideband suppre~sed

carrier signal with the two sidebands centered at 7.4 GHz and
8.6 GHz. The rf filter shown in the block diagram of Figure 3.1

14
 {~.• n",n.d-paU ~....
8.0 Glfz.
.·TW,T..... . 8.6 GH'I
Fl1ter .

600 MH~

150 MHz

Figure 3.1. Block diagram of the PN probe transmitter

at 8.6 GHz .

...
. ',~ "368.t I·4U(
'\

\
,

Figure 3.2. The theoretical line spectrum of a pseudo-

random binary signal.

15
is used to pass only the upper sideband, and reject the lower.
Therefore, the radiated signal is centered at the 8.6 GHz
frequency.
The spectrum of a PN binary signal is a line spectrum,
with power concentrated at frequencies given by K/N6Twhere
N = length of PN sequence in bits
6T clock period (l/f)
c
K = harmonic number.
From this, it is seen that the power occurs at intervals
of 1/N6T. The envelope of this line spectrum ha~ the shape of
(sinx/x)2, with nulls occuring at intervals of 1/6T. The
primary power is thus concentrated in the first lobe of the
envelope, with the first null at

1/(6T) =
c
= 150 MHz.
f
This spectrum function is illustrated in Figure 3.2. The PN
signal bi-phase modulates the 600 MHz IF signal, producing the
spectrum of the PN signal centered around 600 MHz. The spectrum
of the radiated signal is the double-sided spectrum of that
illustrated in Figure 3.2, centered at 8.6 GHz.
The rf filter is designed to shape the final spectrum
around 8.6 GHz, attenuating the secondary lobes of the
2
(sin x/x) envelope. The spectrum as measured over a trans-
mission link, is shown in the photograph of Figure 3.3, taken
from the display scope of a spectrum analyzer.
The maximum power capability of the probe transmitter is
that of the TWT amplifier used in the final stage. This is
rated at approximately 15 watts. However, this power level
was not required- on any of the test links. The transmitter
power was held to the minimum level that was commensurate with
the path loss and the receiver sensitivity for each link.
The radiated signal of the probe system resembles a ran-
dom Gaussian noise with fairly uniform power density. The
power is distributed over an approximate bandwidth of 300 MHz.
Therefore, if this signal were detected by a narrow-band system,

16
Figure 3.3. The measured power spectrum (in dB) of the
transmitted signal of the PN probe operating
at a clock frequency of 150 r1Hz. The center
frequency is 8.6 GHz, and the frequency
scale is 50 MHz per division.

17
it would not normally be a serious source of interference. It
would result only in a slight increase in background noise
level in the narrow-band system.. For example, .assuming complete
uniform power density over the proposed 300 MHz bandwidth, and
the maximum power from the transmitter of 1.5 watts, the power
per unit frequency would be approximately:

Trans. Power Density 1500 mW!300 MHz

-6
5 x 10 mW/Hz.
Using the gain at 8.6 GHz for a 6 £t (I.E m) parabolic antenna
(41.5 dB), the above power density would result in an ERP
density of:
ERP Power Density = 5 x 10~6 x 14,125
:::: O. 07mW/Hz.
The re,ceiverof th,epr,obe system is sh'own in Figure 3.4 in
bl,Q·ck ,dia,gram form. The re;ceived signal is correlated with a
CO~ and quad-phase replica of the transmitted signal, at the
IFof 600 MHz. 'The relative phas,esof the r·eceivedcomponents
are then determined from the quadrature signals developed by the
:correlation detectors. Th·ese qu.adra·ture fun·ctionsare s,quared
and combined in the receiver to provide a directcrutput of the
impulse magnitude. The output of the receiver is the impulse
response of the transmission channel. The 600 MHz IF signal
ismonit"ored with a power meter, and is record,ed through
a log converter amplifier for received signal level (RSL)
measurement. All four of thes·eoutput signals are recorded on
analog magnetic tape for £uture analyses. The quadrature or
magnitude impulse response functions ar·emonitored in real
time on an oscilloscope during measurement. In addition, the
RSL is recorded on a strip-chart recorder fbr direct compari-
son with the impulse data.
The receiver is designed for self-calibration~ A locally
generated 8 ,.6GHz PNmodulated signal is fed to the rf head of
the syst,ern, .andmea sure,dw i tha powe r meter a t the ;600 MH z IF

18
 Power Log
Mt~teI' Amplifier RSL
8.6 GHz

)-. H.F
Arnpliiier
IF
Amplifier
Integrator 81 ~ ... hi(t)

2
Ih(tll

Integrator L - - - 4 - - -__- - -
I "
1"__ hI' (t)

~
~

600 11Hz

:t .... Sync.

5 },,'U-Iz Frequency PN
StClGle Synthesizer Generator
Clock

~ ---
1~)O MFlz
!

Figure 3.4. Block diagram of the PN probe receiver.

for a reference level (both RSL and impulse magnitude). Step
attenuators in the rf head are then used to calibrate the
receiver over the dynamic range. Transmitter power is measured
directly with a power meter at the 8.6 GHz output of the TWT
amplifier. All line losses at both the 8.6 GHz rf and 600 MHz
IF were measured prior to link measurements, and accounted
for in the system cali~ration.

The recorded data included the following:

1. Co-phase impulse
2. Quad-phase impulse
3. Power impulse
4. Sync. pulse
5. Time code
6. IF RSL
These data were recorded on magnetic tape for playbackapd
analyses in the ITS laboratories.

4. RECEIVED SIGNAL LEVEL ANALYSIS

As detailed in the previous section one phase of the data

acquisition was the recording of the AGC voltages on the FAA
RML-6 receivers. The segments of strip-chart recordings in
Figure 4.1 are a time compression of select periods during data
acquisition at Chicago and Atlanta. This figure is intended to
illustrate the density of aircraft-microwave beam intersections.
The Atlanta data indicates about four occurrences per hour as
typical and Chicago data indicate a frequency of up to eight
per hour. Further, in the same figure one can get a feel for
the variance in the depth of fades.
As a first step in the analysis, the strip-chart recordings
were reviewed and each case of a significant fade was time
tagged and the depth of fade hand scaled from the chart. The
depth of fade in this case is defined as the difference in
signal level between that immediately prior to the aircraft
detection and the minimum signal level.

20
 I' I'

~ t-:-;:.::;~~~.:...:.4~~+L-~!..f'
Of
~ +-1 --+~--+-+-+---+--~--t-----t~

~ lr ~:-:::-:i--:-:-:-r:r:-:-:-:r-:-~~~~.
~
F=-!
t-l:_:+--+--+-......--+-+--..--~--t----t--
r:
......

~.;;: .

r--· L. tT ;, ~ ~'- f:
r---+---+----r----r--+--~---. ---t .. l .... 1 ; .i ( ~, . ~,., "'..;

I, I' 'i Ii •
Time (hrs)
I' " i! " it r , ~ I

o~ -+--~--+--++--+--+-+-l--t-+-
:::> r .-.-+---+-o--......-...;-~-+---+---+--
H-
E-t Ii:.
Htt-----+--+--+-----+-----+-?----+--

H +----r---+--+-----+----+-~_+____
P4
~ +----+--+---+---+--+---+----+---

A _.- ~\ ~\ -r" !\
1 L.._ I \, -j i r.
I :

Ii

t·:.-·
oI:LI " --+--+--++-':_:

~,l' .
H) .
H --+----+-'---+---+-+--+---+-
+-;:..
P4 "
Ic:::"'f +--+--+----+---+---+-

~ -+----+------t-----1~-+__+----t-
.... "'1
I ; I c I' Ii • • •
Time (hrs)
Figure 4.1. Aircraft fades for select periods.
21
 For the recording period between 11 July 1978 and 14 July
1978 at Atlanta Hartsfield Airport, 87 cases of fades caused by
aircraft were identified on the strip-chart recording. Of
these, 38 were detected on the A Receiver, the direct link
between the radar and tower. The median for this data group is
6 dB and the mean is 7 dB. The deepest fade detected was on the
order of 18 dB.
In addition to recording the AGe voltage, the combiner con-
trol voltage driven by a baseband noise sensor in the RML-6
equipment was monitored. No direct means was readily available
to calibrate this function but as a function of received signal
level, output changes were not detected until the input signal
level dropped to -45 dBm to -50 dBm. This is used as an indica-
tion of excessive noise in the baseband. Only on the occasion
of the three deepest fades was any measurable change noted, and
then only approximately a 5-dB increase in baseband noise. The
general appearance for the Atlanta configuration is that fades
do occur but at moderate depths and with little disturbance to
the baseband signal-to-noise ratio.
Figure 4.2 shows the characteristic of the deepest fade
(18 dB) that was recorded at Atlanta. The A Receiver designa-
tion is for the direct link from transmitter to receiver. The
duration is about 0.5 s with the bottom displaying two distinct
nulls. These null pairs will be shown later to be common and
possibly suggestive of separate fuselage and tail obstructions.
The leading edge of the fade shows about 13 dB change in 1/8 s
suggesting an approximate fade rate of 100 dB/s.
Figures 4.3 through 4.5 are taken to illustrate three
aircraft positions with respect to the microwave path. The
view is from the control tower toward the transmitting antenna
marked by the arrow in all three figures. In Figure 4.3, the
aircraft is completely obstructing the transmitting antenna
causing one of the observed fades. Figure 4.4 illustrates an
aircraft near but above the microwave path resulting in a
limited shallow fade. The L-l011 jumbo jet in Figure 4.5 can

22
 ;
;
It:
.j .j

~ :: ;::~ ~ ~~L: ::t r !::: ';'~ :~ ~ :~ ;~ . :i::

•. j
.;. :\:; ~lt J~i :: ~:J . .. ... 'I, ... ! ...•..~+.;.I

!
~

:; .. '
tll :::: l:i: .. ".... !.',.l!, .•• H- ..... ! t ' , . l I

i f ' · :;:; .. ,.: I . i'j'd' 'i::' J~J ;ti~i:J~ i J

T I ., ~ ::: ~ ::: T
,; I!:t.:n LL .. ·;j:·..... i •• H ••. ! .•• i-l--t ..... 1

t j i .... ,ll :;t. ;+-1 ,.

::±

! •.•/o! ... , Ii· '.~·.+;d+ .. 'I.. • .• '!'" ..I· '·dIH ' I . l

i
1
, i -t· I ;~. j ;

. ~i..
.
Lf :j: I
l)
: j:l
:+~t
:t
' j
~ .: ;:; :t~:
:r 11-1.
·l.;
:HF1. ::. it:.!
•. ·· j
r:r .,." .
"',
1 ~ ..:'

:i: t
I
1I
••• ;·.;\.',·11>·.- ..... ,,···,
t:; 1, ... 1 j . . ·1·~·iH·
I
1I
i
i
j ·1·;·+ :::: . :t l , , ! d . I ! •. .;.! .•. ,1
.+ I ' ··i -~ , ...

i;:' ; ! it· :::: t: : ::1:: t: :11 ~ : :tt; :~t: : :: I

::;: ,t':t:i : ; : : ;.. :
I t Ii
.;,.1 : I
! . .. i l .--!-.• '. I~ •• j tI . • 1
~
..-. ~::: f+ ....
.;:; ,++...;
.~. t·,

:~ Li: ~i~if:H ::::

N
.. 1.
.. :U ~::: :::: :::: ;
:.: ~ i .. :::: . :.:::11
I· ,q-~ ...... ! .... j

:.'.'::.'i.' :,':'" :'. :.I! :.;::::

-, , .. 1·+ f ... :;:;
.
: ::: h. !::: 1::: .l. i 1_::::
:::: i . t ....

• I • ,,:.±~: ~. : ....
j
.T ........
n= .'" . ....
I
: :;::
I.t lt ., ••
.., . . ::: .l

r :::: 1::: 11.:::: t: ;: ~+1: ::t~::: I

.. " j .

,tf
::::! I' ,

·t ;::; ::::,
I-
:~~. ::1: HI ..
+t ;. :::;
:;J:' :t ::;:
.. ,.

I, .\ .. :::: :J:tr::

23
24
N
U1

Figure 4.4. Aircraft position above microwave path.

 ----.....---r~

. .. .. .. . ....
...
. ~

:: .~. :::: .... ::::

~.
t::: .. i .

: :~:.- :': ..
~ ...

:11::: •.•. fl::: t ..... tt.••.•.t , ... ! ...• !

.::r~:t?.=# , , .. 1
:~ ••. r::"- ...•..••. ~H +tsl~~)
··1.-·.

~l ~ .:;. :::: ····-t···· jl· H.q::-.:±H

\.~::: :::: :::: : : :: :::: ~.:-;" :;::

:::,'
. n'."
..
~ .

.. - .
• 1-' .•• j ••
..... .... ,.,.

:r i~ :~ .... .,.. ....

. .....
.~:
....... :... . ,

~ -+-+__I---+----+--4-.---+-_+__~
:..ci. .. ~: ::::

l.t :;h
:~ .

'-',::'1:..
:.~ ~:r:r-:-:-:--r:·:j
;.:.1:. ..

~ fi i~ ~ ~:_~~ ~ ~ :~ ~ :~ ~ ~ ~:: ~ :: ~ ~ :.1 H ....

~ l~ ~ ~~': ~ :~ ~: :: ~: :~ :~ ~::; ;:;: ~; ~: ::.~
't:. ::::::::
...... : ......

.. :1 ...•.. ,.
::r: :::: .
...,.;

..,-r.::..:::: ::::: ..
r ~ ~: :::: :::: :::. I

....... i
J :::: :0~2~:: :~~-~ ::;; .: :E~ ~~:~ :~:~ I
;~>?~: ..,. .... .... .... .. ..
.... :::: :::::::: ;.:Y--J. .
. .... . :: :;:; :::: ::: :t:::. :::: ~:::
: ::: ~ I ...~ :::: : ..
;::: i ....... , :::: :::: :::: .. :: ·-·t. ::::
.... :::::;; ::::i
;;: :::; t·" ....... I
• I·
..•
- 'H+ .t.t·
. -t t::: :::: ::::
...
1
"1
: :: :/:~ i::' .: ~·:i ::: ~ .:::: ' ..
:::~ :: :;: .:: ::! .::~ : ::::
J
1
I
'::. ::::
t •••. I ••• +·I~.·.. f .... ::::

....
:r~
,j

... :::: .. :: -t.:::: , ..

.
L;

32
 i
I
I I I t I I ! I I I r ! 1 lit t I I J r t t! I I t I r ~ i I I 1 I I I I 1- i
t -25dB
_. H +. ,w;w:. JtlA.
.: ~T :: L.~ q:::.f: : q:::
d: .::t ~ ::d::~ d: ::d: .:d:.. f: HI ~: t. :r~~~ q: ::: d::kL :. . .
", t ~ :~: :~,:~t::::L:: :t::::t. i ~~L~: q:: d:: ::::1:: ::f::: d:· :::f::: l::':r::d::::t: :t> :N.:: .I> :q>: :: :t/~ d: <n: :.r~~t: q:~: :r:ffftnf:l ~ ~~rt·:tFt:::!::: 0:F::
·---1· .. t.:::!: :l~::: ·:::J:~nl: ~d01::' '::~t~~~L' :1::::1::~ . : .
l-~+
~--+- ...... -..::~ ~r:FTITrq ~.~ ~ :!:;: d::\V~ ~~h ~ :JI:: ;q~ ~ ~ d~ ~ ~q: ~ ~ d: ~ :d~~ ::I: ~: q::::!~ :::t ~ ::d: ~ ~~h:: d~.~ :d ~i,~~ p~~ ~ t:~: d: ;:
ill~E ~:-f++"
:: tilE I:,: ::t,::~L:: :t~!I:i. _.. . . . . .. :: 1: : : : i ::: f : : i : :: : f : : : : i : : :: f :.: : : f ~ : ~i 17! ~ :~l~~~~ f:~ :~f ~; H~ : ~ ~: 1::;:1: ::1::: ~ 1:::: t: : :: L : : : ! : : .. + •••••••••

·t····i·.:....:.
.i.j~ii:r~~ : .. ::L:~q: ::!:·::r.::: ::q~~: ......
~rtnftfHl~J±ili~
.,

~~ J.: ::-
..
-:':~.'."'~~~:-
:-:··t +. _. +.. - _+-+ -.._
w::: .~-
.. ::~±i ....................... .. ..

····H+-: ... 1--'.: :::;1:::: . ..... .... .. .

~ . . u~ ~ f ;: ~ I ~ ::!::::t : : : ~ i: ::J:~ :;+: ::':1: :::+: :::1:::
-
-- h: 1-· rrr-- t:-r. ::.:: t++ .... iA.··t·.·.~.··.~.:~.:;' .. ··t·.··.·.t:;..;~~·· f····
...... :::1::::; :\:t :::tt':~'-4el1:r\t:J;:IJ:j:::: I::::

1 1 "1 I
~

I I I 1
1~ I I :=n---I-t- I I I

nii1i :: :..t~li
'~h~' ·tH.:;.·:t~·: .t*:~L~:t:: ~ :1::::
.. 'h~ .t· .~ 1:: ::1: ~:: 1::::1::: ~!::l:~,a::::
' . . . . . . . . • • . • . . . . . ++:- ... ~·N···· i---. :~~;:+. .. ~. : :~~~Pi1-~~p!~ij,A~4~~I~~~~
tN
tN .. f+ ··t+-:-:- .
: .::
. •.•• , :t:g:' -:-:: : : :.L'.:
::--h···~::f:
•
: :: ~ .:
• ••••
~... ~:::
:;T::t::::l:·:.;.~t:;.¥
\",'" • • • • . . • • . -J ••
: ~.: :: : .-+
F.+:: ~,-:,:~r:~'l.- ~ ~ ~ ;!:::.~ i ~ :~ ~ f ~ : : :i ~ :: .t ; ~ ; : t : ~ ~ ~. i;~ : :::1:
::m~t~: ::r ::=h: . ~ ~r'~~ d~ ~ ~4~ i; ;[-i\i ~ t~: ;~ Iii ~ :WEb: i 4~'~ ~.~ tTS1~i:]i i! :1::t-::::.. :t:+::f::.::.... -.. :: ::1:~ ::I::::!.
~--Hffif:: ::-1;;: ~f::: q::
~:f':: ~~p; ~J~ ~: ;t; ;~~ht~~ i iH; ~ ~p~TnETnT~ ~r ~ ........ : ::::t~.-+ .•....••.... :·r::
~, . :.:It i : ::1 :t! :f ~ : H: ~ : Ii I::: ~ I;;;J ~ ; ~ J~ ~ ~ iit ~ i j ~ ~ : ~ f:: ~ :1::: ~ I: ~ ::I ~ r :: . ~ ~F~~ilL:;
'r' ·'t--; "r
.(:

.....
t-i-= ~"t.:j;J.;..;..::-~-~: ...·f···r"'r
"'1';"
t--+-~~t.,....._·;························.······_·····.·.·-tt
~ ::,:: ::~:- "~r"
::.:.: :::: :::::::: ::::'1":::: :·t:::t~::: r
1 .• ··t····!····!····1········
.::: .. ':::1:::: ::1. ...
~-l.,.;.!.-:-:'m""m1ti"~l'
lit
t !"t.. ~ ,. ±.t...r .. J-il' '~~::
... !..
•. .I:":' +1 .. ::L1 :::: : ::·t-::::::::
...
::.:: ::
t • •~
:::: :::: :::: I::: ::::,. :.:'.
•• }
:...
t· .. ·t·-_.. i' ... t-·· '1' .. ;l-..
.:.: : : : ' . ' : :: :::: ':':''':.:'':...:'' .'-
::::1·::: .• : 1 :~..:.: 1: :

F1T'l"'!""'l""~+ +-+-
·---!-·;i
+-- ~~H 'r
Fff+.:.
.'. :: ::1:.·· ~- +:: ; i 1: :: : ::::F::-:E~:' •. ..:-.~ .. t
i.++--
.-'1' .... t": Hi-1 :::
·t'

~ ....
;;fut~'~T 1 : : ::1:: :: l~:; : 1: : .. • .... 0-+' . . . ::::t.,;,,;,..: ............ f .... i ... : t· . :: i E~ ~
......-... ,;
.. __..:::: f r
~
,+-t-r
tt •··.. 1·-
..10 .t1(.'"
1 ':- :8""
." e'. ''-'~:~a:' .
..:.-+; i;: I~+t~i..·· . . ·~~"t
._·-.:tkt~.......---
·-·f

----=t---t= t=--r=-- r- t=- -t..==t -t =+--=t= . t=t --i=-=t -r- t--=t-----t=·_" - =+=.=t --1-"- t =t r---=i i-"-i -t'-"F =t- ==t .=1-- ·-r ==t i -i= t-·-.t- =i -t --+ -1-=4=' -,

Figure 4.10. Time expanded fade at Chicago.

 5.1. Hartsfield - Atlanta International Airport

The Atlanta airport has two microwave RML links as illus-

trated in Figure 2.1. The direct path crosses both runways on
the south side between the radar site and the control tower.
Due to a potential interference between the FAA operating
frequency and the 8.6 GHz test frequency of the probe, it was
not possible to instrument this link. Very little data could be
obtained from this configuration in any event, since the arriv-
ing and departing aircraft were rarely in the microwave beam.
They either passed under the beam on landing, or passed above
the beam on takeoff.
The probe was installed on the reflected (Channel B) path,
and the obstruction measurements were obtained from aircraft
intercepting the reflected path between the water tower reflec-
tor and the control tower (see Fig. 2.1). This path proved to
be almost ideal for the experiment, as landing aircraft approach-
ing Runway 27R intersected the beam approximately 0.5 miles from
the end of the runway. Occasionally, an aircraft approaching
Runway 27L could also be observed, but these events were rare.
Estimates of aircraft position, elevation, and speed on approach
to Runway 27R would place them directly in the microwave path.
This was also observed visually for most of the data obtained.
The traffic pattern on 14 July 1978 was such that most of
the measured responses could be directly identified as to the
type of aircraft.
All of the data for 14 July 1978 were played back from the
magnetic tape recordings and scanned in the laboratory. No
multipath was ever observed in the impulse responses from this
configuration. Each aircraft caused only a power-fade response,
with depths as reported in Section 4. The link itself displayed
some slight ground reflections which are shown in Figure 5.1.
The first surface reflection is seen as a distortion in the
trailing edge of the response in this figure, and a smaller
reflection is seen beyond the main response. The delay times
are on the order of 3 ns and 7 ns respectively, as the base

34
 TIME
Figure 5.1. A typical response measured over the reflected
(Channel B) link at the Atlanta airport
without aircraft interference.

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

.::'1'.:.1
::::1:: .'

"·:1

l.r! rl!,:il:tt.llt:t:tL:: . :1:: :·;:L: : :::·Ll:i~:i I: tl ~l·"

'I r1 I: 1Lt lllll~ t I :11 J1 J1: 1L If +1~ 1J t111 ti t t: 11 : J
:;:: :·1 >:!:. ':,:::'

1~4~-----1 SEC

( a) (b)

Figure 5.2. The impulse response measured at the

Atlanta 'airport with a DC9 aircraft in
the microwave beam. The time separation
in (a) is 100 ms. The data rate in (b)
50 impulses per second.
35
width of a clear channel response would be 13.6 ns (1 em in the
figure). The response is linear in magnitude, so that the two
reflected components are approximately -6 dB and -20 dB with
respect to the direct response. These reflections are seen to
be very stable in all of the following figures for the aircraft
obstruction measurements.
A typical response with an aircraft intersecting the re-
flected path (Channel B) beam is shown in Figure 5.2. Part (a)
of this figure shows a time sequence of the impulse response
observed when a DC9 aircraft was approaching 27R for a landing.
The data rate was 10 impulses/s, so that the time between each
response of this figure is 100 ms. A flat fade in the response
is seen, which is on the order of 15 dB between the top trace
and the fifth trace down. Note that the typical double null
discussed in Section 4 is also visible in this response sequence.
Part (b) of Figure 5.2 is a slow playback on a strip chart
recorder of another measurement made later the same afternoon~

The aircraft was again a DC9, and the envelope of the impulse
response is seen to display the double null characteristic.
The fade depth is less in this example than that in Part (a),
and the data rate for this measurement was 50 impulses/s.
Figure 5.3 presents two more examples of the impulse
response with obstructing aircraft in the beam. Part (a) is an
example of the response where a 727 aircraft was on a landing
approach to Runway 27R. Part (b) is an example of the response
for a 727 aircraft that was on a landing approach to Runway 27L.
In each case, we note the absence of any multipath with the
exception of the surface reflected components noted for Fig-
ure 5.1. The sequence is not long enough in these photographs
to illustrate the extent of the power fade in time. However,
the duration of the fades were noted and discussed in Section
4. Note in example (a) however, that the double null is again
evident from the sequence.
In all of the responses seen in Figures 5.2 and 5.3 that
the magnitude relationship between the direct and surface

36
 Time
(a) (b)
Figure 5.3. Examples of the impulse response measured
at the Atlanta airport with 727 aircraft
intercepting the microwave beam. In (a)
the aircraft was approaching Runway 27R,
and in (b) it was approaching Runway 27L.
The time between responses is 100 illS in
each case.

Time
Figure 5.4. A sequence of impulse responses measured
over the Chicago-O'Hare RML link with no
aircraft interference.
37
reflection remains relatively fixed. This fact indicates that
the surface reflection must lie between the radar site and the
passive reflector since the aircraft would not normally disturb
the reflection if it were on the reflected leg of the path.
From these measurements, we conclude that the obstruction
of aircraft within the geometry for the AtlantaRML causes only
a power £adeto the microwave system. Theabsenc~ of ~ny multi-
path from the aircraft indicates that no frequency selective
fading will result, and the performance of any digital system
could be predicted on the basis of flat fading over the duration
of the obstruction.

5.2. O'Hare - Chicago International Airport

The two microwave links at Chicago O'Hare airport are side-

by-side, providing a space diversity path as discussed in
Section 2. The PN probe was installed on the northern-most
link (Channel A) to minimize the potential interference to the
FAA operating frequency. The most significant impulse response
data were recorded on 19 July 1978, when the air traffic was-
operating on Runway 14R toward the southeast. The likelihood of
obstruction by the entire aircraft on these links is much lower
than that for the Atlanta airport, as noted in Section 2. Many
more smaller disturbances were observed, probably due to the
tail structures of aircraft penetrating the microwave beams when
taxiing on the surface or when rolling on Runway 14R. However,
a deep fade in the signal was observed occasionally. These
deep fades have been interpreted as cases where the full air-
craft has intersected the beam. Direct observation of the
activity on the runway and taxiways was not possible due to the
link configuration. This area was blocked from view at the
receiver site (see Fig. 2.5).
The microwave path was essentially clear of any surface
reflections, as shown in the sequence of responses in Figure 5.4.
These responses are typical of a clear channel response, again
with a base width of 13.6 ns and a time separation of 100 ms.

38
 A complete scan of the impulse data recorded on 19 July 1978
was made in the laboratory. The deep-fade obstructions were
found to display results quite similar to those observed in
Atlanta. In general, the impulse response did not show any
multipath when the aircraft was in the beam. A power fade with
depths of the same order of magnitude as observed in the Atlanta
data was the significant feature. However, occasionally a re-
sponse was observed that contained a strong multipath component
during an obstruction fade. This observation is discussed in
more detail below, and a conclusion is given based on other
observations from the data.
Figure 5.5 presents a typical response measured at Chicago-
O'Hare. Part (a) presents a sequence of impulse responses
recorded during the aircraft blockage shown in the accompanying
record of the RMLsignal fade shown in Part (b). The impulse
data correspond to the Channel A (upper) RML record. The fade
is seen to extend over a period of approximately 100 ms, which
is consistent with the impulse record where the separation
between responses is 100 ms.
An example of a deep obstruction fade where multipath was
observed is shown in Figure 5.6. The fourth response from the
top of the photo displays a strong multipath component. The
delay is between 3 and 4 ns, and the cusp in the response
indicates that the component is near opposing phase with respect
to the direct response. These features are discussed in greater
detail in the appendix to this report. Similar cases were seen
fora few other obstructions measured at O'Hare. To attribute
the multipath to the aircraft causing the obstruction is
contrary to all of the observations made at Atlanta, and to the
majority of the obstructions measured at Chicago-O'Hare. In
addition, the configuration for the Atlanta measurements would
be rnoreconducive to multipath from the obstructing aircraft
since the flight path and the microwave path intersect at a
more oblique angle. The possibility of a reflection from the
vertical tail of the aircraft would be greater in the Atlanta

39
 .;~

~~~W

,-1-' -+-"... --"

t + +··t -
H-

·-tt ::1 +:r ."

1- ~
~ S-+:
.m: .

Time
":

(a) Impulse response :. :~ -+-

at 100 ms intervals. "'~----....-.(-ol •5 SEC
......

(b) FAA/RML signal records.

Figure 5.5. A typical response measured over the Chicago-

0' Hare link '{vi th an aircraft obstruction.

Figure 5.6. A sequence of impulse

responses measured
during an obstruction
fade. Multipath is seen
in the fourth trace from
the top.

Time

4-0
case than in the O'Hare configuration, where the intersect
angle is nearly 45°. For these reasons, we do not believe
this multipath was caused by the same aircraft that caused
the deep-fade obstruction.
In order to determine the possible cause of the multipath,
the data for the entire recording period at Chicago-O'Hare were
played back from the magnetic tapes. The RML signals were tran-
icribed on a strip-chart recorder, and the impulse response
data were monitored on a storage oscilloscope. In this
analysis, there were many cases of severe multipath observed
in the impulse reponse at times when there were no (or only
minor) disturbances seen in the RML records. Periods of
observed multipath were flagged on the ID1L chart record using
the event marker pen of the recorder. An example of this
process is shown in Figure 5.7. Part (a) of this figure is a
short section of the strip chart record showing periods of
strong multipath that occurred both before and after a deep
obstruction fade. Part (b) of the figure is a sequence of
impulse responses measured approximately 15 s prior to the
deep fade shown in (a). -This sequence is typical of the
multipath observed at other times, and the detail of the
response is seen to be similar to that found in the one
response shown in Figure 5.6. A close examination of the
records for all of the O'Hare data indicates that any multipath
seen during an obstruction fade was correlated with periods of
multipath observed before and after the obstruction. For
example, another event is illustrated in FigureS.8. Note that
very little multipath has been flagged on the ~ML record for
this eyent, and the impulse response shown in (b) for the
obstruction fade indicates no multipath.
As a result of the above analysis, we conclude that the
multipath seen at O'Hare must be from aircraft moving on the
surface of the airport, and not from the same aircraft that is
directly intersecting the microwave beam. Using a mean delay
of 4 ns for the primary multipath component, the ellipsoid for

41
 --.iii- • • •
I I . , I I '

.- -+ +

-
-_.
~ ....-- --7"--
i-

I I I
t f I I I I

I '":lUI Plno9
·::t:
..j..., : ..•..
. .. ..,

"'-

~ ,
Q '.
..

P : .... :
: . +
8 +. J.
H
...:i :.t ,.
::
:·f"';'··
... :
:

~
,:::C
'":--
r" •.~ . :
.-
.. r+
~. f't.
l~~~:Tl.+tIr·-- t ;: -
H· n

(a) Sample of the RML records where the times of observe·d

multipath is indicated by the event marker (top trace) .

(b) An impulse response

sequence measured
at the time indicated
by the arrow in (a).

Time

Figure 5.7. Multipath observations over the Chicago-

O'Hare link.

42
 II
I 1 T I r r

".'1-'" . - ~.~
t-H--

I I I I I I I
I r I I 1 T T r r r

-. .., ::::h
""" T LlL ! .. : : ~ ~.~,.+ : ::~ •T

.. ~.

.-- ..
1----,.-
.:':
--; H-.i:
f-';' ++

-~I 20 SEC ~ I

(a) A sample of the RML record where very little multipath

was observed. Multipath was seen at times shown by the
event marks in the top trace.

(b) The impulse response

sequence measured
,during the deep
obstruction fade in (a).

Time
Figure 5.8. An obstruction fade without multipath.

43
the locus of the reflecting points can be drawn around the path
shown in Figure 2.5. In the horizontal plane, the ellipsoid
is on the order of 120 to 125 ft (36 to 38 m) on e~ther side of
the path at the location of the runway and two taxiways (see
Fig. 2.9). It is noted that the tails of large jet aircraft
are high enough to penetrate the . Im~c. ~~a1ec . b.e. aI\1 a~.any of these
locations. The most likely Orig~9J\q~/~1;Ah~,\~~~ot~~I?at~woUldb~/"
from aircraft on the inner taxiway and the terminal ramp area,
as the orientation of the tail structure is more likely to be
nearly parallel to the path. However, aircraft turning from the
runway onto the two taxiways near the microwave beam could also
produce the multipath. These conclusions are consistent with
the data, but, of course, cannot be considered to be final or
absolute.
The most important aspect of the multipath seen in the
O'Hare measurements is its effect on digital transmission
over a microwave circuit of this type. The multipath is severe
enough to cause significant frequency-selective fading in the
transmission channel of high data rate systems, with the
associated degradation of the bit-error-rate (BER) performance.
These matters are discussed in more detail in the Appen-
dix. The conclusion to be drawn is that extreme care must be
taken in the design of such a microwave link. Even though
approaching or departing aircraft on the runway appear to cause
only a flat power fade when intersecting the beam, other aircraft
on the surface of the airport could cause very significant
problems to a digital transmission. The eng~neering for a link
of this type must include adequate clearance for all surface
aircraft and other structures.

6. CONCLUSIONS AND RECOMMENDATIONS

The primary objective of this program was to determine if

aircraft obstruction of a microwave link was sufficient to
cause a break in service to the subscriber. This program
utilized two 8 GHz systems that were established by the FAA at

44
Atlanta and Chicago airports. The configuration examined was
for aircraft in motion; i.e., take-off, taxi, or landing. The
case of stationary aircraft was not addressed. The intent of
the program was to determine if a problem may exist and to
what degree rather than being a comprehensive, detailed measure-
ment and analysis program in a controlled environment.
The results of this limited effort indicate that in a
normal operational mode such as take-off, it is possible for
aircraft to intersect a microwave beam and further to provide
a degree of blocking sufficient to cause an observable signal
fade. However, in the examination of nearly 200 cases, no
fade was detected to exceed 20 dB and there was no indication
of an enhancement of noise in the baseband to be of any concern.
Systems designed with a 30 dB to 40 dB fade margin should more
than tolerate the observed disturbances with no detectable
degradation in customer service.
Having observed no pulse distortion at the Atlanta airport
during the aircraft-caused fades indicates no frequency selec-
tive components within a 300 MHz bandwidth. This provides
substantial support for the observed fades as being character-
ized as "flat" fades where the necessary band fades uniformly
and equally. This fact minimizes any potential for baseband
distortion or data stream disturbance.
There were also no observations at Atlanta of delayed
pulses that would indicate discrete, strong multipath charac-
teristics. This type of distortion could potentially cause
second order effects on the data formats. Since there were
none detected, confidence builds in stating that aircraft
obstruction is not a significant problem. The test results
indicate no significant first order effects that could cause
baseband distortion. This does not preclude the existence of
higher order, more subtle effects. This program did not address
such problems nor did it eliminate the possibility of their
existence.

45
 On the other hand, some observations at the Chicago airport
indicate a potential delayed component possibly caused by air-
craft taxiing through a critical point on a system with low
ground clearance. The detected delayed pulses were in the 3 to
5 ns range and relatively fixed. Two simple interpretations may
be approached. As discussed in the appendix, the effect on a
digital system operating with a 15 to 20 MHz bandwidth may be
the distortion of the amplitude response to the degree of some
asymmetry on the received spectrum. The effect of this asym-
metry on digital system performance is not fully understood yet
and requires additional experimental effort (discussed in more
detail in the appendix.) A second simple view assumes that
digital system intersymbol interference becomes excessive with
delays greater than some fixed percentage of a bit interval
(OT Tech. Memo 74-182, Farrow and Skerjanec, limited distribu-
tion). For the sake of this discussion consider a quarter of a
bit interval to be the limit. With the 5 ns delay, the highest
supportable data- rate translates to 50 Mb/s. If the distortion
criteria were one-half of a bit interval, the usable data rate
becomes 100 Mb/s. This is to say that with the fixed delay
detected at the Chicago airport and with the above assumptions,
it may be possible to achieve a satisfactory system with data
rates up to 50 Mb/s. These cases are for fixed phase differences
and for a specific multiplex timing recovery system. These
assumptions may be voided for variable phase cases and for other
clock recovery and decision making technique.
From an engineering point of view certain criteria should
be set in order to assure that the design choices are made to
minimize the probability of intercepting an aircraft. Typically
microwave links that traverse runways are short to the degree
that variable atmospheric effects are negligible. The choice
of terminal placement should be considered in light of minimiz-
ing the probability of intersection with aircraft movement.
The terminal ends should not be closer than necessary to the
nearest runway. The best choice would be across the middle of

46
a runway and avoiding paralleling of taxiways. Further, one
should avoid crossing areas where aircraft stop while waiting
for takeoff, gate access, etc. Finally, the terminal antennas
should be sufficiently high to clear the tail section of the
largest aircraft expected to operate at that airport. This is
to exercise the available choices to assure the smallest
probability of aircraft path and microwave path intersection.
System availability should not be affected by the shallow
fades caused by aircraft. If the path is sufficiently short to
not be affected by the atmospheric variability, diversity pro-
tection need not be considered for other than equipment
availability.

7. ACKNOWLEDGMENTS

The authors wish to recognize the u.S. Army Communications-

Electronics Engineering and Installation Agency for financial
support for the program and in particular the personnel in the
Electromagnetics Engineering Office at Ft. Huachuca, Arizona.
We wish to thank the Federal Aviation Administration for grant-
ing access to and use of their communications systems. The
cooperation and assistance were greatly appreciated from the
Great Lakes Region and Mr. A. Qualiardi, O'Hare Radar/ARTS
Section Chief and his personnel at Chicago and the Southern
Region and Mr. J. Bryant, Manager, Airways Facilities Sector
and his personnel at Atlanta.

8. REFERENCES

Linfield, R.F., R.W. Hubbard, and L.E. Pratt (1976), Transmis-

sion Channel Characterization by Impulse Response
Measurements, OT Report 76-96, August.

47
 APPENDIX - EXPLANATION OF IMPULSE TEST DATA

This section is provided to aid in the interpretation of

the impulse test results and in the translation of these results
to other systems. The primary objective of the ITS probe is to
transmit a band of frequencies sufficiently wide to encompass a
significant portion of a microwave band to allow detection of
many of the shorter delays and distortions. The receiver
portion of the system indirectly detects the amplitude response
of the same frequency band.
In the ideal case, it would be desirable to transmit an
infinitely short pulse such that its transform includes all
frequencies as shown in Figure A-I(a). Figure A-I contains in
the right column a series of short pulses presented in the time
domain. In the left column is the frequency spectrum corre-
sponding to each pulse or pulse pair in the right column. The
amplitude, frequency, and time scales are not shown since the
figure is for illustrative purposes only. If the two pulses
maintain a fixed delay but their amplitudes vary with respect
to each other, the depth of the null or nulls in the spectrum
will vary as displayed in Figure A-l(b). As the difference, A,
increases, the spectral null becomes shallower. If, on the
other hand, there are two pulses present separated by some time
delay T, the spectrum will be sinusoidal in shape as shown in
Figure A-I(c). If the relative amplitudes are fixed, but the
time delay, T, allowed to change, the null positions change
and become more dense as the time delay increases. This is
illustrated in Figure A-I(c).
The discussion to this point has been for very wide band-
widths. Even for a pulse with a finite width such as that
realized from the ITS probe, the bandwidth is approximately
300 MHz. The question at this point is what such a measurement
would tell about a system operating with a bandwidth of 10% of
this value or less. This may be illustrated by the frequency
bands superimposed on the spectra in Figure A-I(b) and (c). If

48
 Frequency Time

j
fl f2 f3 f4 Time

Frequency
(b)

fS f6
T
Frequency

Fre9uency Domain Time Domain

Figure A-I. Time and frequency domain displays.

49
the operating digital system were constrained between f and f ,
l 2
the system would experience fading that would have the appearance
of flat fading. The amplitude response in this case can be seen
to be flat, or nearly so, and symmetrical. If the system limits
were between f 3 and f 4 , the fading would not be as deep as the
previous case but the pass band would contain a sloped response
as a function of frequency and would be asymmetrical. Finally
in the third case, Figure A-l{c), a system operating between f
S
and f would experience sloped amplitude responses with changes
6
in direction of the slope as well as change of direction within
the band. In the latter two cases, the fading is frequency
dependent in that all frequencies are not attenuated uniformly.
Finally a fourth case can be imagined that consists of varying
time delays and relative amplitudes so that one has a combina-
tion of those effects illustrated in Figure A-l(b) and (c}.
This in fact may be the case in most physical situations when
multipath is a problem.
The effect of this frequency variable response within the
passband of a receiver and detector is not fully understood yet.
Experimental efforts are underway at Bell Labs and Collins Radio
(Barnett, 1978; Rockwell International, 1978) to determine a
quantative measure in terms of bit error rate performance as a
function of the slope of the frequency response. Considerable
experimentation and analysis is still necessary to provide
sufficient engineering guidelines to cope with such problems.
In order to provide some indication of the impact of the
multipath observed at Chicago-O'Hare airport, we present the
following brief analysis. In this development, we will refer
to the generalization presented (where possible) with values
from the measured data.
The mean value of the delay of the multipath component was
measured to be 2 to 4 ns. Also the delayed signal was approxi-
mately of the same magnitude (and sometimes greater in magni-
tude) as the more direct signal. As an illustration, we choose

so
to model this transmission channel as a two-path channel with
equal magnitudes for the two paths. Thus, the impulse response
would be that illustrated in Figure A-l(b) where A=O and the
delay T=2 to 4 ns. The ~ourier transform of the impulse re-
sponse yields the frequency transfer function H(f) of the
channel. The transform in this case is a sinusoid as illus-
trated in Figure A-I with a frequency period of 500 MHz for
T=2 ns and 250 MHz for T=4 ns. On a logarithmic power scale,
this transfer function is similar to a cycloid with deep nulls
along the frequency axis as illustrated in Figure A-2. The
location of the nulls is given by

f
n = niT (A-I)

where n is an integer (>0), and is odd for a reflection with

phase reversal and even for a delayed component in-phase with
the direct path. For the case considered here, where the
reflection is assumed to be from the surface (or tail) of an
aircraft on the airport ramp or taxiways, the reflection would
have a phase reversal, and n would be an odd integer. The
spacing between frequency nulls is given simply from

6f
n = liT. (A-2)

Let us assume an operating frequency for the microwave link

to be 7.4 GHz. We first need to determine where the frequency
nulls of H(f) fall with respect to the operating frequency.
Using (A-I) for a value of f
n = 7.4 GHz and; a delay of 2 ns,
we find that n must have a value of 15. Thus, the nearest null
to the operating frequency is found as

f
n
= ~
2 ns
= 7.5 GHz, (A-3)

and subsequent nulls will occur at 500 MHz intervals. The

value of n corresponds to the Fresnel zones; in other words,
the reflection at 7.5 GHz in the channel would be from the
15th Fresnel zone ellipsoid.
In a similar manner, we find that for a delay T=4 ns, the
transfer function H(f) has a null at f
n = 7.25 GHz (n=29).
51
 l--l--~___+_,--:· I l.l I~ 1=1 I~
I f't1 ~ ~~ 1- • I :,rr~
o I i I
r'l I I I I I 1>1..:: I I I I I I I : ! I I: I Iii I I

--.t---,.. -

-lOW
~
"""""""
.. ..

~
ro
I
J:iI
(l) -20
Z
0
Pol
Cf)

U1
tv
~

"_F' WI_ _
~ -30
~
H
E-f
_ r m r _ij'.-u
m - 1
-. ++
+--tj----
' .
_.
-I
LL
, '- -
- J-+--
. -+-~"j t..- ---+---
! ..
. --
. i _.
..
I
---+---+-. -
~
H
+= L~'r - ...~ . ..t~-T-=~ -" +__ ~ ~ = - . ,_.t ~t. -t~.-
tlli±'!'t'
-
~
I 1 .. . , -' .. -. •
I
- .. -·i-J···r- t -'- - - -' --l·
-H-- --f- , , -+-, ----. ·····t-t-r·J -+-~-t+l--r-+- -.,..+ ·__L : +-"t--+---t-
-40

+
r
. I
-50

- 60 d i I I i I ! " ,I , i I I I I, I • i ' , . ,. I 'I' ,! I !, , , ! , I'" ' , ., ! I I ! ,

i.O 7.2 7.4 7.6 7.8

FREQUENCY - GI-Iz
Figure A-2 o Tlle tl'leoretical transfer function of amul tipath channel.
This transfer function has a null-spacing of 250 MHz. Both of
these channel functions are plotted in Figure A-2, identified
by the value of the integer n and delay time T. Also drafted
on this figure for illustration is the signal spectrum of the
microwave system at the 7.4 GHz operating frequency. We have
assumed a spectral bandwidth (BW) of 30 MHz, typical of a
digital transmission system operating in this frequency band.
At first glance, one might conclude that the multipath
structure considered here, which results in the frequency trans-
fer functions of Figure A-2, would have little or no effect on
the operating system. In a static situation, this conclusion
would be correct. The transmission channel in both cases would
cause only minor distortions across the signal spectrum. The
operating system is well removed in frequency from the deep nulls
in the transfer functions, and the frequency selective fading
within the signal BW is at most 2 dB across the BW. This aspect
of performance degradation in a digital system does not become
critical until the spectral slope distortion becomes consider-
ably greater. This matter is discussed in more detail below.
The most critical consideration in the multipath structure
which we have depicted is a result of the dynamics of the chan-
nel. For example, the physical situation we are modeling here
involves reflections from moving aircraft. Thus, the channel
transfer function is not static, but becomes a time-variant or
dynamic function; a time-variant filter in our communication
system. For the conditions shown in-Figure A-2, let us assume
that the reflection point (on a moving aircraft) has changed
positions over a short time span such that the original delayed
component shifts from a delay of 4 ns to 2 ns. The frequency
transfer function in the dynamic view becomes very difficult to
depict; however, it can be visualized as "elastic". For exam-
ple, consider the functions in Figure A-2 as stretching or
relaxing in an elastic fashion. The nulls of the function thus
change position along the frequency axis, and the shape of the
function changes. As can be seen in Figure A-2, the slope of
53
the function as it approaches the cusp of a null will also
change. In addition to this dynamic change, we must keep in
mind that the value of the integer n is also changing as the
time delay changes. The dynamics of this change can be visual-
ized by considering the entire H(f) function shifting along the
frequency axis either to the right or left. This change is
most critical in considering the dynamic effects on system
performance. For example, if we assume a slight change in
delay-time from 4 to 3.8 ns, the frequency transfer function of
the channel shifts between the two dashed-line functions in
Figure A-2. Note that the null corresponding to n=29 in these
two functions has moved from 7.25 GHz to 7.63 GHz. Thus, this
null has passed through our system pass~band as the time delay
changed. In the case we are modeling here, we note that the
latter change will cause a frequency null to pass through the
operating system BW a total of 8 times as the delay changes
from 4 to 2 ns. This will obviously cause severe frequency
selective fading within the signal bandwidth, as well as power
fading across the complete BW during times when the null or cusp
of the filter characterstic passes our operating frequency.
The rate at which the distortions impact on the signal BW is
a direct function of the rate at which the time-delay between
the signal components changes. However, it is also important
to observe that even if the time-delay change is linear with
time, the rate of change in the frequency transfer function is
non-linear. For example, if the time delay is changing from
2 ns to 4 ns (increasing) the initial nulls that pass the opera-
ting frequency will do so at a higher rate per unit time than
the later ones.
The actual impact on (degradation of) a microwave digital
transmission system, caused by the multipath structure consid-
ered above, is very difficult to assess analytically. To begin
with, the resulting bit error rate (BER) will depend somewhat
on the type of system used and on the actual digital transmis-
sion rate. A simplified view of the transmission rate

54
dependence can be gleaned by considering a 5 MHz signal BW in
Figure A-2 in comparison with the 30 MHz shown. It is obvious
that the frequency selective distortions would be less severe in
a 5 MHz BW, and the frequency nulls would cause fading across
the signal BW that more nearly resemble power or attenuation
fading. Most digital systems will perform differently under
.thes~ two classifications of signal distortions. The most
straightforward techniques that can be used to evaluate system
performance under varying channel conditions, and different
system characteristics and parameters, would be through labora-
tory simulation processes. Without benefit of such stu~ies, or
actual digital system tests in the multipath channel, it is
almost impossible to predict the performance under the dynamic
conditions measured or postulated for the microwave link at
Chicago-O'Hare airport. However, the multipath observed over
this link occurred a significant amount of time, and the impact
on a digital transmission link can be gleaned from the discus-
sion given below.
Based on the measurements made and presented in this report,
the most important recommendation that can be made is that the
proposed microwave links be engineered and tested so that they
exclude any possibility of multipath interference on the airport
surface. Measurement results from the Atlanta airport indicate
that multipath is not a problem with aircraft in the microwave
beam; at least within the geometries of the two test configura-
tions. Thus, the important criteria is to avoid the surface
type multipath observed at Chicago-O'Hare.
There have been some recent efforts to determine the
degradation caused by frequency selective fading to a digital
system. Even though these measurements are made only in a
static or semi-static sense, they are significant enough to
report here. With respect to actual over-the-air performance
tests, Barnett (1978) has reported results that convey the
importance of frequency selective fading, or what Barnett terms
the in-band linear amplitude dispersion. In his measurements,

55
Barnett used a number of narrow-band (slot) filters within the
signal BW of a 78 Mb/s, 8 PSK system, operating at 6 GHz.
Observing the power in the two filters at the extremities of
the pass-band provides a rough measure of the amplitude dis-
persion caused by the characteristic of the frequency transfer
function of the channel. He reported that a slope on the
order of 0.2 dB/MHz across the BW of his system was sufficient
-3
to cause the BER performance to degrade to 10 ,or below this
threshold value. An in-band slope of this value (or greater)
would occur for the cases shown in Figure A-2 below the levels
marked with the short cross-lines. Thus, for example, if the
transmission of interest were operating at 7.45 GHz in a static
channel that has either of the response characteristics shown in
Figure A-2, the performance would be degraded to the order
reported by Barnett. It thus becomes clear that the performance
in a dynamic channel can only be treated on a statistical basis.
The measurements performed in the current project at Atlanta
and O'Hare airports do not form a statistical sample that is
sufficient to predict long range performance. The fact that a
dynamic multipath structure was observed is, however, sufficient
to raise a genuine concern, and to serve as a cautionary example
in the design of future digital systems in an airport environment.
It should be pointed out here that the synopsis given in
this Appendix applies only to a single non-diversity link, and
one which uses no other adaptive process. It has been shown
(Anderson, et al., 1978; Rockwell International, private commu-
nication, 1978) that both space diversity and some form of
adaptive equaliz~tion in the microwave receiver can significantly
improve the system performance in multipath environments.
However, detailed studies need to be made under controled
dynamic conditions (preferably simulated) to document the poten-
tial improvements that have been observed under quasi-static
conditions.
A step beyond understanding the foregoing effects is the
need to know the frequency of occurrence, the duration of such

56
occurrences, and its total impact on system availability. The
effects discussed above are not continuous occurrences but
rather appear somewhat infrequently and are of short duration.
In short, there is a time factor that weighs heavily in the
determination of system impact.

REFERENCES

Anderson, C.W., S. Barber, and R. Patel (1978), The effect of

selective fading on digital radio, International Communi-
cations Conference, Toronto, Ontario, Canada, June.

Barnett, W.I. (1978), Measured performance of a high capacity

6 GHz digital radio system, International Communications
Conference, Toronto, Ontario, Canada, June.

57
FORM OT-29 u.s. DEPARTMENT OF COMMERCE
(3-73) OFFICE OF TELECOMMUNICATIONS

BIBLIOGRAPHIC DATA SHEET

1. PUBLICATION OR REPORT NO. 2. Gov't Accession No.3. Recipient's Accession No.

NTIA Report 79-14

4. TITLE AND SUBTITLE 5. Publication Date
Aircraft Obstruction of Microwave Links Januarv 1979
6. Performing Organization Code

7. AUTHOR(S) 9. Project/Task/Work Unit No.

R.E. Skerjanec and R.W. Hubbard 9103533
8. PERFORMING ORGANIZATION NAME AND ADDRESS 9103534
National Telecommunications and Information
Admirlistration, Insti tute for Telecorrrrnunication 10. Contract/Grant No.
Sciences, Boulder, Colorado 80303

11. Sponsoring Organization Name and Address 12. Type of Report and Period Covered

USACC/CEEIA NTIA Report

Fort Huachuca, AZ 13.
Final
14. SUPPLEMENTARY NOTES

15. ABSTRACT (A 200-word or less factual summary of most significant information. If document includes a significant
bibliography o-t literature survey, mention it here.)
Conversion to digital transmission has renewed the concerns about
what effects aircraft obstruction of microwave links have on user
quality. This is of particular concern where it is necessary to
install a telecommunication system that crosses runways and taxiways
where the frequency of obstruction may be great.

A limited measurement program at 8 GHz at Atland Chicago Air-

ports was undertaken to determine if a condition existed that could
cause exces error rates on digital systems. Measurements were
made of the received signal level together witthe impulse response
of the transmission medium.

(Continued)
16. Key Words (Alphabetical order, separated by semicolons)

Microwave propagation; impulse response; digital transmission;

aircraft obstruction

17. AVAILABILITY STATEMENT 18. Securi ty Class (This report) 20. Number of pages

[K] UNLIMITED. Unclassified 68

19. Security Class (This page) 21. Price:
D FOR OFFICIAL DISTRIBUTION.
Unclassified

L'SCOMM-DC 29716-P73
15. ABSTRACT (Continued)

Measurement results indicate that during takeoff and landing, air-

craft can cause signal level fades to 20 dB. Modern system margin
is usually sufficient to cope with such fades.

The impulse response measurements at Atlanta did not reveal any

delayed or distorted pulses that would indicate excessive multipath
and frequency selective fading. However, slight distortion from
taxiing aircraft at Chicago was observed. The implication of this
detected distortion is reflected in the potential of a distorted
asymmetrical frequency response within the 15 - 20 MHz passband of
a microwave digital receiver.

* u.s. GOVERNMENT PRINTING OFFICE: 1979-0-678-478/298