Multiple cards in one board, revisited

Work in progress, not an actual project yet.

Self interference is a major issue. I’ve got several boards out there that I put up when I was green with multiple cards on the same band causing all sorts of issues.

One thing that I have noticed is that the XR5 is not affected nearly as badly as R52H. I am trying to figure out whether it is due to the extra ufl connector w/diversity (only 20dB down when not in use), the difference in pigtails on MMCX or major differences in the cards.

With both cards set to 25dB, if I scan with an XR5 in a board with an R52H, the XR5 reports the R52H at -40dB, the R52H reports the XR5 at -20.

I am considering using the R52Hn as a replacement for a few of my R52H in my backhauls and I’m curious what other folk’s experience is on the issue. Would the R52Hn show the same -40 as the XR5 due to the MMCX pigtail being better, does the secondary port negate that issue, does the secondary port pick up anything if the second chain is disabled?

Have you tried reversing the antenna’s to each radio card & see if signal levels remain the same.

I haven’t. They were in the air before I realized the problem. I have reproduced this on other access points as well.

The XR5 is connected to an omni and the R52H is on a 26 or 29 dBi grid in both cases. The grid is 4 to 8ft below the omni.

The difference in signal levels when scanning from each radio card, is it due to different RF loads (antenna + leads) for each radio card, frequency used, antenna positions, more questions which can only be answered by testing

Yes, I agree, and I’m asking about the internal signals, if anyone else has done that kind of testing, whether the single or dual MMCX makes a difference over UFL, whether it looks like the pigtail, connector or diversity switch seems to be an issue.

Actually, thinking about it this morning (uninterrupted for a change), I have done some testing here when building a backup AP that I didn’t use and had similar results.

The only real way I can think of to test it is to put a 50 ohm dummy load on each connector of a completed AP and see what the scans show internally.