PMP450i 3,5 GHz, problem when frequency sharing

Hello,

we have site, where we plan to run 4 sectors in ABAB 4x90° configuration. Curently there are 3 AP (2xPMP450 + PMP450i). One PMP450 and one PMP450i are on the same frequency (back to back, +- 2m horizontal distance between them), all synced with CMM4. Same "Frame Configuration" on all APs.

When I put them on the same frequency, PMP450 behaves as expected, SNR drops little bit, but modulations are ok, 8xMIMO-B mostly. On PMP450i all clients dropped to 1xMIMO-A, allmost all clients show SNR 14 (probably error, they are in different places and I dont believe they really have same SNR). Linktests are really bad, AP throughput drops from 60 Mbit to 10 Mbit. It was in the evening and I kept the configuration in spite of it was running badly.

The next day when I logged into the PMP450i, everything was sudennly fine, most clients 8xMIMO-B SNR 25+, everything running as if nothing happened the day before. No restarts, nobody touched configuration.

This lasted about a day and today in the middle of the day (again no confing change,  no restart) all clients back to 1xMIMO-A SNR 14.

Reebooting the AP changes nothing, changing TX power and SM target level has verry little effect. Spectum analyzer shows no outside interference. There was no change in in the second AP performace when the PMP450i was runnig badly, suddenly fine and than badly again.

Because the PMP450i runs OK when on different frequncy, HW should be OK. It was able to run for more than a day on the shared frequency with good speeds, so radio environment and interference is also OK I gues.

Is there any other explanation than a firmware bug?

Really like to know the outcome of this. i have seen the same issues and ended up staying with 10mhz channels on the tower for now and waiting for spring to tackle it again.

We had a similar issue with the 450i 900mhz when they first came out with modulation issues and a firmware update resolved it.

I have a feeling the 450i and the 450 platform have slightly different timing. i have towers with 4x450 365 AP in abab config no issues. Its only when mixing the platform's on the tower i've had issues.

I want to swap out the 450 off of that tower and see what happens once we have 4x450i's on the tower.

Thanks

Brenton Kidd

Hello Peca,

I would like to take a look at the configuration you are running on all 4 APs on the tower you are having issue with. Can you please send me complete CNUT captures from all APs on this tower at dhaval.chaudhari@cambiumnetworks.com

Also can you tell us a bit about Sync sources you are using on these APs? Are they all using same type of sync source? What software version they are running? When you are talking about throughput drops, is that happening in Downlink or Uplink or Both? If you reboot the AP when it is in BAD state, will it resolve the issue? All of this information will help us to understand the issue. 

Thanks

Dhaval

Thanks for the reply. I configured them on the same frequency again and currently it runs OK, I will send you CNUT captures as soon as it happens again.

There are currently 3 APs, 4th is not installed yet. All synced from CMM4. APs have software 15.1.1, clients mostly 15.1.1 few have older but all above 15. When the performance on 450i goes down, it happens on both download and upload, but it is worse on upload.

 The picture shows linktest to one of the clients on 450i running on shared frequency. You can see, that when it is working , there is no radio environment issue.

It took less then 3 hour and the problem reappeared, linktest to the same client now looks like this:

Rebooting the AP doesnt help. Performance of the other sector (plain 450 non-i) on the  same frequency is completely unaffected by all this.

I owe you information what happened with our problem:

After lot of testing and trying I looks to me, that there in no problem with Cambium HW or AP synchronization. I discovered, that the problem affectc olny north and nort-west direction and only part of available frequency spectrum. If I set the frequency low enough or high enough, it works.
Problem apears only somewhere between 3460-3540 Mhz (roughly) and only sometimes. I wasnt able to prove any interference from spectrum scans, my best bet so far is some kind of weird signal reflection that happens only on some condition. One theory is that the reflections happen when it rains or snows and everyhing i wet, but it is just a theory, this site is in different town than my office and I am only guesing weather there.

1 Like

Thanks for the update.  Appreciate the follow up.