Towertalk
[Top] [All Lists]

Re: [TowerTalk] If you had a choice

To: towertalk@contesting.com
Subject: Re: [TowerTalk] If you had a choice
From: Jim Lux <jimlux@earthlink.net>
Date: Wed, 17 Oct 2012 13:02:01 -0700
List-post: <towertalk@contesting.com">mailto:towertalk@contesting.com>
On 10/17/12 6:33 AM, Joe Subich, W4TV wrote:

near field measurements are full of problems also. besides the
problems of projecting the near field pattern to something that is
useful in comparing in the far field there are all the little
distortions caused by coupling to nearby objects, feedlines, trees,
etc...


I was thinking of a near field range, not just making measurements in-situ in the near field. I have to admit I've never *seen* a HF near field range, but theoretically it should be possible to do it. (there's that Master's thesis thing...). Point the antenna up on top of a flattened off mountain top, figure out how to correct for the partially conducting lossy ground plane, etc.

We actually do almost all of our antenna pattern work at JPL on a variety of nearfield ranges these days, often with scale models. Our 60 foot anechoic chamber now has a huge planar scanner in it. The outdoor ranges mostly get used for doing things like evaluating multipath and scattering from "stuff" on the deck of a rover and confirming what we get from a numerical model of that.

However, "believing" the results of a near field measurement requires that you trust the math, and all those folks out there who don't trust NEC probably wouldn't trust the near field to far field transformation either.

A far field measurement has the advantage of conceptual simplicity, after all.



Ground ranges have been used for years with good accuracy compared to
free space.  With horizontal antennas but the direct wave and the
first lobe have good correlation with the free space pattern.  Tests
like those conducted by Steve and Ward use a consistent range and
single reference so any range distortions fall out when making
comparative measurements.


I don't know about "fall out", but certainly "knowable" and "can be calculated out".. Consider comparing a dipole against something with a fair amount of gain. The "bounce" signal on the dipole is pretty strong, but on the gain antenna, it's suppressed. The classic approach is to move one end of the range up and down at least a wavelength, so you can characterise the bounce. I suspect there are other clever ways to do it by changing the relative heights to a few key points or changing the spacing. Somewhere around I have an IEEE Proceedings from the 70s that has the whole process in it from the folks at NIST (nee NBS)


Mountain top to mountaintop helps, too.. you're less likely to get a specular reflection from something "down in the valley", and even if you do, it's hopefully not in a sidelobe or on the edge of the main lobe. (that's the way the JPL outdoor ranges are set up. Big valley in between source and receiver, and some attention to make sure there's nothing creating a specular reflection.


A fun thing might be to use one of those little quad-copter things
to fly an orbit around an antenna at various altitudes while making
measurements.

I doubt that those devices have sufficient operating height to make
practical far field elevation plots - even at one mile which is still
not completely in the "far field" on some bands.

I have a 1/3 scale R/C powered parachute that I was going to use for this kind of thing: GPS and 3 axis short dipole antennas with a receiver (like RELEDOP from SRI), but it has a whole lot of other practical problems... For instance, the little quad copters (and my powered parachute) can't move fast enough to overcome a decent wind.

Some folks in the UK have done this with flying a probe on a tethered balloon. (and of course, SRI's RELEDOP is a probe towed by a helicopter)



_______________________________________________



_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk

<Prev in Thread] Current Thread [Next in Thread>