[TowerTalk] Symptoms of antenna interaction?

[Lux, Jim]

On 3/24/21 6:58 PM, Jim Brown wrote:
> On 3/23/2021 12:30 PM, Al Kozakiewicz wrote:
>> Is this symptomatic of interaction, i.e. feed point characteristics 
>> change while resonance frequency remains the same?
>
> If you can do modeling, try modeling dipoles for the bands where you 
> are concerned about interaction, with a generator in only one of them. 
> Open the View Antenna window. Then compute the far field response, and 
> on the View Antenna window, go to View, then Objects, and check the 
> box for Currents. You can check other boxes if you want to see those 
> things. I like to show Axes and Segments to help me tweak the model. 
> Then in the View Antenna window, move the Zoom Current slider to make 
> the currents large enough to see current distribution in the dipole 
> with the generator, and look for current in the one without the 
> generator. If there's no current, there's likely to be little 
> interaction.
>
> Repeat this process as you vary the length of the antenna you're not 
> feeding, then move the generator to the other antenna and repeat the 
> process.
>
> This is not as accurate as full models, which are FAR more 
> complicated, but it will tell you whether to look further. A dipole in 
> NEC is simply a wire with a generator in the middle of it. Always use 
> an odd number of segments for a dipole so that the generator can go in 
> the center. Wires are entered by their x, y, and z coordinates.
>
> 73, Jim K9YC 

This is exactly the process we are using at JPL for evaluating "is it a 
potential problem" before going to a more high fidelity model.  Build 
your models with simple wires - don't try to taper or model mounting 
hardware. You could model the Yagi as 4 wires.

With modern tools like EZNEC, 4nec2, or AC6LA's it's pretty fast to put 
the geometry in (and, in 4nec2, which I use, you can define symbols, to 
make it easier)

I don't know that you really care about resonant frequency in this sort 
of analysis 5% is probably fine - what you're really doing is looking at 
the *magnitude* of the coupling, and that doesn't change very quickly if 
you change the length of the elements.

As Jim says, you visualize the currents on the other antennas, and if 
it's "big" (where "big" is more than, say, 1% of the current at the 
feedpoint), there's a potential problem.

This is the HUGE benefit of modern tools on modern PCs.  You can run the 
iterations quickly, so you can get a more intuitive "feel" for what's 
going on.  You can, for instance, change the skew angle between two 
dipoles or antennas and see where the coupling starts to come up (yeah, 
it's basically cos(theta), but that's not instantly visualizable for 
small changes).

R.W.Hamming said: "The purpose of computing is insight, not numbers"