On Wed, 15 Jan 2025 10:51:05 -0800, Brian Beezley <k6sti@att.net> wrote:
"Today, with DEMs available to generate the "tiles" for a full 3d model
would be straightforward. That was one of the challenges when Breakall
did his work."
Jim, the data I would need are measured 3D patterns to validate a 3D
modeling program. I thought a drone might generate them, now I don't
think so. There are many sources of error with a drone, some rather
subtle. It might work in certain limited terrain, but not in general.
There's no way one would work at my QTH.
>>> The most significant issue is the polarization. Generating calibrated
>>> signals with a drone (either Tx or Rx) requires a fair amount of work -
>>> you'd need three orthogonal dipoles or monopoles, and then a fair amount of
>>> post processing. I'm actually doing this for a project at JPL (with a
>>> spacecraft, but the problem is the same with a UAV). One could probably
>>> generate the probe patterns fairly accurately with numerical models. If
>>> you look at pictures of the Cassini satellite, there's 3 monopole antennas
>>> sticking out of the side - that's used for a fully polarimetric receiver as
>>> part of what is called Fields and Waves experiments.
>>> Then, it's a matter having your test transmitter cycle through putting the
>>> test tone through various combinations of the probe antennas. You might be
>>> able to be more efficient, but a fully general case would radiate through
>>> each single antenna (3 cases), then through pairs with phase offsets to
>>> generate CP (3 more cases, maybe 6, if you need to generate LCP and RCP).
"I'm not sure an ever increasing model fidelity is useful."
The issue with a radial-only model is that it can be entirely wrong, not
just off a bit. Worse, it gives no indication that the result is
unreliable. I think it's possible to use a radial-only model under
certain circumstances, but you need to carefully vet the terrain. It
definitely wouldn't work in most directions at my QTH. And while some
directions look benign, I'm not sure they really are. It's tricky!
"My understanding is that HFTA is horizontal pol only (the reflection
model is simpler)."
I don't know what HFTA does, but TA used specified ground constants with
Fresnel reflection coefficients for both horizontal and vertical
polarization at all reflection points. Vertical is no more difficult
than horizontal. The equations are just a little different.
>>> Yes, but there's a trick with "any pol" vs H only. I can't remember (maybe
>>> someone can weigh in) but it might have to do with whether tiled flat
>>> plates are an adequate model, not so much with the actual Fresnel
>>> equations. I'm sure there are codes that do it.
Incidentally, after months of making innumerable errors of all kinds, I
think I finally have an accurate stratified ground model. Its
application is rather limited, as is the available stratified ground
data. But it provides some insight into the accuracy of surface ground
probes:
http://ham-radio.com/k6sti/sg.htm
My writeup on the Hagn generic curves, which yield ground constants much
more appropriate at HF than the figures antenna analysis programs
suggest, is here:
http://ham-radio.com/k6sti/hfgc.htm
>>> Yes, Hagn has a lot of useful things to say about soil properties. He spent
>>> quite a while trying to measure them with dipoles at low heights (or more
>>> properly, to get accurate propagation models for low height antennas in
>>> jungles like Vietnam).
Brian
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|