Throughput issues 450b HG PMP

Hello,

Bit of backstory: I am fairly new into working with Cambium radios and am inheriting an existing PMP setup that was not documented well in the past. Please forgive me if my terminology or understanding of some of the functions of the radios!

The setup we have for this sector is a PMP 450m and 3 total 450b’s. The radios are in short range (farthest is 0.5 miles from the AP).

I am experiencing an issue with one of our 450b HG radios that has poor throughput (less than some of our other SMs). I manually adjusted the aim of the radio and was able to jump from 10 Mbps x 2 Mbps to around 80 Mbps x 15 Mbps on average. I believe we can still get more throughput out of this connection but am unsure of what I may be missing. I believe this because the other two 450b radios have much higher throughput (on average about 180 Mbps x 60 Mbps). Just to clarify, I am checking the throughput through the Link Capacity Test tab on the 450b and 450m side.

Compared to the other SMs, the Rx power level is about the same at -52 dBm. This is what is set on the 450m side as well. I have read that this can be too “hot” of a signal and I have tried lowering it to -60 but it did not change the throughput. Please see a screenshot of these below. LUID 6 is the poor performing 450b and the other two are fine.

I also have looked at the alignment tool on the 450b itself and noticed the SNR usually has a yellow or red indicator. I am unexperienced/unsure with how to fully digest this information and know how to improve it, but my assumption would be there is something that is causing noise near the 450b itself? I tried aiming the radio higher and to the side a bit which improved the SNR and the Receive Power did not lower either, but this resulted in even lower throughput then before. Please see a screenshot of the current alignment tab below:

Also, other background: I recently setup an on-premises instance of cnMaestro which has helped us gain more insight into the radios themselves. If there is anything I can add onto this post from there, let me know so I can provide as much information as possible!

To summarize, I am curious to know if there is a way to maximize the throughput between these radios. Thank you for any potential tips/answers!

1 Like

That radio having issues, LUID 6, appears to be experiencing noise, as indicated by the poor SnR ratios compared to the other two. I’m assuming that this is a 5GHz system? The noise is most likely local to that SM, so check to make sure there aren’t other 5GHz systems running on or overlapping the 5GHz channel you’re using. If you can’t find the issue locally, you may need to change the channel of the AP and monitor that SM’s SnR and modulation levels for improvement. Lastly, it’s ok for an SM to hear an AP at a strong signal of up to around -45dBm, above -40 and you may want to turn down the AP’s transmit power or de-tune the SM if you experience throughput and modulation issues. In the opposite direction, on the AP, the recommended SM receive target level is -60dBm.

1 Like

Thanks for the fast response!

Yes, this is a 5 GHz system.

We have a Ubiquiti nanostation on the same pole and pointed about 10 degrees to the left of our 450b that is having these issues. The 450b is using channel frequency 5765.0 MHz and the Nanostation is using 5170 MHz. Would these still be causing interference with eachother?

At 5765 you’re likely looking for a consumer device like a customers router

If it adds any help, the 450b with issues is on a roof and pointed up towards a higher roof where the 450m is. The only building in between these two is a smaller building which is using the PTP Ubiquiti nanostations. It is possible that there is a wireless router in that building. Would your recommendation be to use different frequencies?

I highly doubt that the 450b @ 5765 is causing interference with the nanostation @ 5170. Like I said before, interference is typically caused when competing radios are on the same channel, or overlapping. That being said, radios that have a strong signal that are close to each other, that cannot be sync’d, should have some sort of guard band between their channels. A good starting point is a 20MHz gap between channels. E.G. if the 450b is using a 20MHz channel @ 5765, and the nanostation is using a 20MHz channel, then you’d want to use a 20MHz guard band between them, so the minimum center channel for the nanostation would be either 5725 or 5805, or greater.

I would run a Spectrum analysis on the CPE having issues to find a better channel

1 Like

Sorry for another additional question as I am not very knowledgeable with radios.

The 450b and Nanostation are both using 40 MHz channel widths. So would the solution be to lower the channel width on the Nanostations to 20 MHz, the frequency on the Nanostation to 5805, or both?

The Nanostation and 450M are on opposite sides of the band they aren’t interfering you could lower the channel bandwidth but that’s not going to tell you what part of the spectrum is clean.

Hi @CR_Support, welcome to the community, and thank you for asking such a clear question and providing all that information :+1:

(I’m a software guy, not a radio guy, so don’t treat the following as gospel)

Your Nanostations are using 40MHz channels and their center frequency is 5170MHz. That means they are occupying spectrum between 5150MHz and 5190MHz. Your 450b center frequency is 5765MHz, so with 40MHz channel that means it is occupying 5745MHz to 5785MHz. There is 555MHz of spectrum between the top of the Nanostation’s channel and the bottom of the 450b’s channel, so it is very unlikely that they are interfering.

Thank you for your explanation!

So along with what Jesse had mentioned earlier, I ran the spectrum analyzer and found the 5.4 GHz band was “cleaner” so I switched over the SMs and AP to that band last night. I now see that the two other SMs that were fine throughput wise are now consistently 200 Mbps x 50 Mbps, but the troubled SM is only at 100 Mbps x 15 Mbps. It seems there is still noise even thought he spectrum analyzer told me there was only around -80 dBm noise at every frequency in the 5.4 band.

Please see a screenshot of the current noise and link status below:

  • Reference: LUID 6 is one with throughput issues, 3 and 5 are great *

My first thought is to possibly check our physical alignment/aiming of the radio again as the recieve power level is not as good as it was previously.

I also noticed during the link capacity test that the efficiency for the uplink test is not great. Before the frequency change this was typically 95% or higher. Not sure how this would have changed, please see screenshot below:

What would the best next steps for this be? Could it still be possibly some type of self-interference now that the noise issues are happening at separate bands?

Thank you so much for all that have replied and helped so far, much appreciated!

Jesse,

I did this and changed to a cleaner frequency band, please see me latest response if you would like to look into this! Thank you!

What is the SM target receive level on the AP? For close installs we had to lower it to -66 or lower so as not to interfere.

Correct me if I am wrong, but I believe it is set to -52 (default on these devices to my knowledge). Please see the screenshot below where I am getting this info from:

I did previously set this field to -60 and even -65 but did not see any improvements. This was on the previous 5.7 GHz band though.

Can’t remember what the default is - we had a bunch of SMs very close together and near the AP. Setting this around -66 to -70 fixed the self interference for us. Regarding high gains, we did have a bad batch a few years ago that were junk on arrival - had to RMA them.