Antenna Tests and OEM Practices
Consultants Also Comment on Mobile Field-Strength
Plots as a Measure of Performance
by Richard J. Fry
How can you get the most out of your FM antenna system?
To provide some insight into this subject, three prominent broadcast
consulting firms were asked to respond to 15 topics about FM antenna
performance. Their earlier responses can be seen in previous issues
of Radio World and online at under the tab "Better
FM Coverage."
The participants are Ben Dawson of Hatfield &
Dawson; Bob Culver of Lohnes & Culver; and Don Markley of D.L.
Markley & Associates.
Note that any consultant would need specific site
and application information to provide an accurate recommendation
for a given situation.
Topic: Comment on the validity of using one- or
two-bay tests to extrapolate the final azimuth patterns of a complete
array containing more bays.
Dawson: Such measurements are precise only
when each element has the same mounting conditions and tower geometry
within a wavelength or so. On small cross-section towers, the extrapolation
of the data is, in the world of decibels, as accurate as needed
for most situations.
Culver: If you have looked at the horizontal
(azimuth) pattern variation vs. the mounting position along a support
structure, and there is no significant variation of maxima or null
position or depth, then the composite vertical array horizontal
pattern will be the same.
A good way of estimating the final pattern would be
to make several azimuth pattern measurements with the elements slid
along the mounting support in several small steps (1/3 spacing,
for example). The final pattern for a large array will be the average
of all these horizontal patterns.
Markley: We have no problem using one or two
bays to determine an antenna pattern. One bay is satisfactory if
the pattern of a single bay is symmetrical about its axis. Many
antennas are not quite symmetrical. For this purpose two bays can
be used with one reversed and the phase controlled either by bay
spacing or interbay circuitry.
Topic: Comment on antenna OEM standard practice,
and results to the user/ listener, of measuring only H and V fields
on a CP antenna, omitting all other polarization planes.
Dawson: Ideally the phase relationship should
be determined, but it's a price market.
Culver: The H and V field values are at their
maxima, rather they deliver the maximum energy for a like polarized
receiver antenna. There is very little possibility of de-polarization
of the signal over the distance and propagation normal to FM radio.
Of course the vertical polarized component is reversed in polarity
(180 degree phase shift) upon reflection from an object.
The circular polarization field is generated by the
90 degree offset between the H and V polarized fields, which generate
a rotating field orientation along the direction of propagation.
It is "circular" only if the relative fields are equal.
There is never a circular field, just a constantly changing linear
field with position.
Measuring these two fields independently, H and V,
is fine. Measuring the relation between the two with a constantly
changing (rotating) linear receiving antenna vs. azimuth pattern
angle will yield some additional information.
They have done this for C-pol TV antennas and call
it the axial ratio, this is the ratio between the H and V polarization
as the azimuth pattern is swept. By the nature of the display of
such measurements, a sinusoidal wavy line around the azimuth, it
is illustrating the independent H and V values, only reaching a
maximum when the rotating receiving antenna is aligned with the
transmitted field polarization.
I have done similar measurements but by quickly switching
between the cross-polarized receiving antennas.
Markley: It might be desirable to take more
than just horizontal and vertical fields. However, there would seem
to be less and less to be gained as the number of such measurements
is increased.
In any case, the polarization is going to become more
and more random as the signal is filtered through trees and buildings.
Topic: Comment on the value of ground- and air-based
mobile field-strength plots as a measure of the pattern performance
of an installed antenna.
Dawson: Measurement data is really only valid
as relative data, and it's not always easy to wash out the ground
reflection problems from aerial measurements. Both types can be
useful for problem solving.
Culver: GI-GO: Garbage in, garbage out.
Now to explain this. You can measure the full-scale
antenna to perhaps +/- 0.5 dB on a good range. For scale measurements
the accuracy may be +/- 0.8 dB or so. Others may argue that the
results are much better or worse, but how much, another 0.2 dB?
My reason for saying this is the following.
You can measure the differential field from
two antennas in the field with ground-based measurements to perhaps
+/- 2 dB. That assumes you return to precisely the same measurement
path under similar environmental conditions, and work to minimize
all controllable variations and the only error is in making the
measurement run and calculating the average field.
By differential I mean that two transmitting systems,
nearly co-located can be measured to find the difference between
them. You can make absolute measurements of one such source to less
precision, the reason being that you may not be aware of time, environmental
or other variations that would be disclosed in the differential
measurements.
You can extrapolate these measurements back to the
antenna performance by working the field-strength estimation process
in reverse. Why do they call them F(50,50) curves? Because they
are statistical averages and they cannot tell you what the
exact field may be at any particular time or location.
If you can use other estimation processes, you can
get a different answer. Better? Perhaps; but what error will be
built in? At least 5 dB, most likely 8 to 10 dB, especially over
moderately long paths. (If you are close you have the vertical pattern
to deal with.)
The same thing applies to airborne measurements but
with some added problems. First, how do you know exactly where you
are? OK, GPS is accurate to a few meters but how fast does it update?
For that matter how fast does your meter and computer
record the values and how far have you moved between the GPS record
and the field record? We ran into this problem in doing some digital
audio field tests and eventually came up with a system that made
and recorded the field measurements in a few milliseconds so the
measuring vehicle moved only a few centimeters at highway speed.
We made sequential but virtually simultaneous H and
V pol measurements on the fly (so to speak). What are you actually
measuring, what does your measurement antenna see, ground reflections?
OK, you fly at an altitude equal to the antenna (500
feet minimum rural altitude, and worse in urban and controlled air
space) and a distance calculated to keep the antenna null(s) at
the reflection angle.
There is a way around all this so that you might
be able to just fly anywhere around the antenna and automatically
make measurements and reject ground reflections and know
precisely where you are in three-dimensional space so as to pattern
the antenna on its mounting system. But it requires using a special
test signal and active components on both the tower and in the aircraft.
They do something similar in audio. No one has done it at RF yet
(haven't checked the DOD or NSA projects).
Measurements at ground level, 6 to 8 feet not 30 feet,
can be done very fast and takes into account all of the local propagation
effects that are part of the reception environment. This was done
for some of the digital audio field tests.
Markley: We place no value at all on ground-based
measurements to determine pattern performance. The effects of the
terrain on the signal strengths are so significant as to eliminate
any value of the measurements.
Carefully performed air measurements can provide good
information regarding the pattern. To show the pattern itself, these
measurements need to be done close to the antenna. We prefer a distance
of two or three miles and to do the measurements with a helicopter.
Measurements at a greater distance are more affected
by terrain and serve more to show the areas serviced than to actually
determine the antenna pattern.
Richard Fry is a retired FM applications engineer
with almost 35 years of service with major U.S. broadcast transmitter
and antenna manufacturers. Reach him via e-mail to rfry@adams.net.
|
