[TowerTalk] More Interference Pattern

[David Gilbert]

I did two runs in the YouTube video, one using my drone with a 28.26 MHz 
signal source riding on it and the other using WWV at 15 MHz. The two 
antennas were short (10 ft) horizontal dipoles at the same height above 
ground (18 feet) spaced 12 feet apart.  In both cases the antennas were 
broadside to the source.  Calibration to account for variations in 
feedline length and receiver phase lock was done using a source dipole 
between the two sense dipoles.   Keep in mind that the RF ground plane 
is almost certainly not at the surface of the physical ground.

The run using the drone would only have ground reflections to worry 
about, but my terrain is very irregular within just a few wavelengths of 
the antennas.  The WWV test clearly was experiencing significantly 
variable multipath propagation as well as the ground reflections ... 
from the multipath signals as well.

I've done EZNEC modeling with antennas each at a different height above 
ground.  I guarantee that gives inaccurate results.

73,
Dave   AB7E



On 4/14/2026 4:21 PM, Jim Lux wrote:
> I did a simple simulation to get the "difference in phase between two 
> antennas" as a function of incoming elevation angle.
> I assumed 20 meter wavelength, 10 meter antenna height, and two 
> configurations, one where the second antenna is 5 meters higher, the 
> other where the second antenna is 5 meters horizontally displaced, in 
> the direction of propagation.
>
> And for 4 reflection coefficients: -1 - (typical for low angle H-pol), 
> -0.5 (lossy version of previous), 0 (no reflection), and +1 (flipped 
> reflection phase)
>
> Here you go - I can post the matlab code if anyone wants it.
>
> https://drive.google.com/file/d/19LO0YacHbloWAfboGQq_3SmFTWrBKXtE/view?usp=sharing
>
> There was a bug - the incremental phase between the two antennas just 
> due to displacement is partly wrong, but the general shape of the 
> responses should be ok.
>
> If Dave posts his frequency, height, and spacing, I can run it for that.
>
>
> On Tue, 14 Apr 2026 13:30:22 -0700, David Gilbert via TowerTalk 
> <towertalk at contesting.com> wrote:
>
> I decided that the SVG version of the moving interference pattern I
> previously linked had some flaws.  The motion was at a constant rate and
> the direction of the motion of the combined diffraction pattern should
> have been in any direction.  Codex had difficulty resolving those
> problems within SVG so after many iterations I told it to switch to a
> Canvas implementation within JavaScript/HTML.  I think the result is
> much more accurate and is located here:
>
> http://www.ab7e.com/Interference Pattern.html
>
> Notice that the red and green arrival angle lines now shift in phase
> relative to each other in sync with the displayed phase shift ... which
> also now goes both positive and negative.
>
> The interference pattern of the combined waves is now the much broader
> shaded light colored pattern and it now shifts in multiple directions.
> The is a more accurate display compared to the SVG version ... which
> turned out to be more illustration than representation.
>
> A really key point is the comment made by Gemini (another AI agent) that
> my good friend Bob, K7ZB used to analyze what is going on. And that is:
>
> What we really measure when we compare the phase of the signals from two
> antennas is the instantaneous GRADIENT of the interference pattern that
> results from multiple signal paths (both relative angle and relative
> phase) due to propagation and various local reflections (ground, nearby
> terrain, other structures).
>
> Not sure if anyone else finds this stuff as interesting as I do, but at
> least it explains what I empirically found when trying to use my
> ArrivalAngle application.
>
> 73,
> Dave   AB7E
>
>
> p.s.  For reference:
>
> (http://ww.ab7e.com/ArrivalAngle.html).
> (https://youtu.be/MxRaYSRMPE4)
>
>
>
>
>
>
>
>
> _______________________________________________
>
>
>
> _______________________________________________
> TowerTalk mailing list
> TowerTalk at contesting.com
> http://lists.contesting.com/mailman/listinfo/towertalk
>