>From: km1h@juno.com (km1h @ juno.com)
To: <topband@contesting.com>
>Date: Thu, 19 Mar 1998 14:40:43 EST
>
>On Thu, 19 Mar 1998 10:23:14 -0700 Eric Gustafson Courtesy Account
><n7cl@sparx.mmsi.com> writes:
>>
>>>From: km1h@juno.com (km1h @ juno.com)
To: <topband@contesting.com>
>>>Date: Wed, 18 Mar 1998 10:25:31 EST
>>>
>>
>>Snip...
>>
>>>Re..your other post and the April '98 QST comments by
>>>N6BV... Does not the military use NEC-4xx to model their field
>>>installed and stealth antennas? The question remains; are those
>>>models accurate to within < 1dB?
>>>
>>
>>In a lot of cases they can be. Just as other modeling based on
>>previous versions of the NEC code can be. In this specific case
>>of correctly accounting for all of the system losses of a
>>vertical with a radial field of varying density under it, at
>>varying distances in close proximity to earth, apparently not. I
>>say this because they report results that do not agree with
>>measurements taken at various times spanning many years in
>>various locations by various researchers.
>
>
>But you said earlier that you have no experience with NEC-4...??
>That was the ONLY program I asked about.
>
I had just read posts to this reflector stating the following
(from two different posts, I believe):
1. EZNEC had a patch that made it similar to NEC4 in its
reported results for antenna systems like we were discussing.
I can only take it on faith that the above statement is true.
I have a great deal of experience with this program and its
predecessors and although I find them extremely useful, I
KNOW they don't get this exactly right compared to a measured
real antenna.
2. NEC4 did not show the improvement (less than 2 dB for NEC4 vs
4 to 5 dB for measurements) that I (and many others) have
made when adding ground screen density and measuring FS
improvement. I must also take it on faith that this
statement about NEC4 is true since I don't believe that
anyone in this discussion is intentionally trying to spread
falsehoods.
So, I (rightly or wrongly) inferred that NEC4 produces results
for this class of antenna system that is at variance with the
measurement experience a number of us have compiled in the real
world.
Here is what we did in the real world that you can do and then
compare with NEC4 to see if it produces similar results:
1. Build a 18 MHz "classic" 5 conductor ground plane (4 radials and
a radiator, all conductors orthogonal).
2. Decouple the feedline (We had a small signal source in a box
at the feed point).
3. Install this assembly on a non-conductive support at a height
of 3/8 wavelengths above grade.
4. Place a FS measurement system out 10 wavelengths from the
radiator at a height that corresponds to a TOA of about 20
degrees (this height isn't critical). Take care to be sure
that the support and power supply wiring do not interfere
with the measurement being made (non-conductive support and
decoupled and carefully routed wires).
Too bad you don't believe that ground level measurements are
valid. They are MUCH easier to do. We had helicopters to do
this with. We stopped using them (for the ground mounted
antennas) once we satisfied ourselves that the field strength
changes were everywhere proportional. I understand that,
nowadays, many broadcast field surveys are conducted from
airplanes.
5. Adjust the radial lengths, and system matching components to
produce the maximum FS that you can get from 1 watt or so of
power. Always use the same amount of true power. Measure it
in the matched feedline to the source.
6. Drop the assembly 0.03 wavelengths at a time and repeat the
adjustment and measurement procedure. Stop after you get the
radials on the ground. Or just do the 3/8, 0.03, and 0
wavelength cases.
7. Compare the results with what NEC4 says you should get at
each height.
8. Let us know how it came out.
If elevated radials do what is being claimed, then the FS should
not degrade as the antenna comes down until the radials actually
touch the ground. If they do what I have observed, then by the
time you get within a foot or so of the earth's surface, you will
be down by 4 or 5 dB.
You can do the same experiment much easier by staying at ground
level and adding pairs of 1/4 (0.28 actually) wavelength radials
until you get up to the 60 to 100 range and see how much the FS
has improved. Heck, just do the 4 and 60 radial cases. NEC4 (I
hear) says it comes up a bit less than 2 dB from 4 radials to
enough to meet the screening criteria. My experience is that it
is closer to 4 or 5 dB (with only 60 radials this will be closer
to 3.5 to 4.5 dB). Do this test with the radials and antenna
feedpoint elevated to 14 inches. That simulates a radial system
elevated to 12 feet at topband. Make sure that you are over
uniform, relatively poor ground.
I know you don't think 18 MHz compares well with 1.8 MHz. In
other work we did, we didn't see a great deal of difference in
the magnitude of ground effects (over relatively poor fairly
uniform ground 3 to 5 mS conductivity) between 18 MHz and 3 MHz
when elevations were scaled for wavelength.
However, if the model at 18 MHz doesn't compare with the
measurements also done at 18 MHz, then we will have a data point
about how well the model represents reality for this class of
antenna system.
73, Eric N7CL
--
FAQ on WWW: http://www.contesting.com/topband.html
Submissions: topband@contesting.com
Administrative requests: topband-REQUEST@contesting.com
Problems: owner-topband@contesting.com
|