Installed antenna array ignores phase shifts

Consider a simple two element linear array with 300 MHz dipoles spaced 2m apart. Suppose that I now add this array as an installed antenna using MATLAB's glider.stl example platform. Meshing this we get the arrangement below.

So far, so good. Now, let's just calculate the azimuth pattern for the bare array.

patternAzimuth(arr,300e6)

Let's follow this up by adding a 90° phase shift to one element and plot it again.

arr.PhaseShift = [90 0];
patternAzimuth(arr,300e6)

We get the two plots below:

The pattern becomes asymetric as we would expect.

Now, let us plot the pattern for the array when it is installed on the platform.

patternAzimuth(gld,300e6)

As we can see, the pattern has become distorted due to the presence of the platfform, but there is very little effect in the forward direction as we would expect from the geometry.

Now, for the final test let us add 90° to one element of the installed array and plot that.

gld.Element.PhaseShift = [90 0];
patternAzimuth(gld,300e6)

I put it to you that this has made absolutely no difference whatsoever!! I have tried doing this several times with variations such as adding the phase shift to the array before adding it to the platform and I get the same answer every time. No matter what I do the array seems to ignore element phase shifts once it is installed onto a platform.

I should perhaps add that I am using the "MoM" solver for this relatively electrically small problem.

I also tried adding two separate antennas to the platform, but I couldn't get them to act as an array at all.

What am I doing wrong?

Answer 1

[Richard]

Actually, I'm about to answer my own question since I just spotted where I had been going wrong.

It turns out that the right solution is indeed to add the two dipoles as separate elements and this does work. Not sure why it didn't work before, but it might be due to me using the MoM-PO solver. This seems to have problems with multiple elements.

It seems that adding the elements separately, using the MoM solver and the installed antenna "FeedPhase" setting fixes the problem.

It still seems to me that setting phase offsets inside an array should work since otherwise there are few benefits to being able to add an array to a platform, but at least I seem to have a solution for now.