I've never seen anyone try to plot the pattern of their antenna. I've only seen computer generated plots. So I decided to give it a try for the four 91XGs. See my previous post for a photo.
Attached are two plots of the horizontal pattern. I should point out that arranging the antennas into a square produces the same horizontal pattern as having two side by side. It narrows the vertical pattern but you'd never notice that unless you could tilt the antennas up and down.
The first plot is of KMMW on RF 28. I use that as my reference since it is LOS and just 14 miles away. The signal almost never varies. The plots are a little jagged because I took data points every 5 degrees. One degree increments would be better but it's too hard to position the rotor in such small steps.
You can clearly see the main lobe, the first nulls, the first lobes, and a little less obvious the second nulls and lobes. The front-to-back ratio (F/B) looks to run around 25 dB. You can see a lobe that points at 95 degrees. I have a strong reflector off the mountains there and all signals from the west have a peak there. Even a 147 MHz ham repeater above Palo Alto has a strong peak there. I suspect the antenna pattern is a little better than what it looks like at 95 degrees.
Now take a look at the antenna pattern for KOVR. I picked KOVR because it was close in frequency to KMMW and only 2 dB weaker. The F/B looks awful. This is because the mountains east of me are LOS to Walnut Grove and my antennas have 2 ridges to look over. I estimate those 2 edges attenuate the signals by 15-20 dB.
I need to find a way to increase the F/B on my antennas. I'm wondering if larger reflectors on the 91XGs would help? Taking an extreme example, a 20' x 20' sheet of metal would certainly block the reflections 100%. How big of a screen would be needed to block another 10 dB? I haven't found any good information on this.
I noticed in these plots that my antennas were pointing 3 degrees too far CCW. KOVR is 270.3 degrees, not 267.5 degrees as shown in the plot. I fixed that.