SNR sawtooth effect

Hi all,

Has anyone seen this issue, and more importantly found a cause/solution?

SNR (uplink) bounces between 2 values giving a sawtooth effect.

In this case it is a 450m and a 450B-LR but I see it on all combinations. This is affecting sumbunall SMs on the AP. It seems to be transitory and in one case was due to 2 visible AP with same Color Code but this SM only sees 1 viable AP (2 others <-80) and no Color Code overlap.


I'm having this issie with a PMP450i 900Mhz...

Did you figure out what was causing this?

In 450, the SNR is an estimation using the modulation of the link and measuring the data transmitted. It needs data moving (and enough data) for accurate measurements. It is possible the modulation is moving between two values here, which will immediately move the SNR.

I recommend moving to EVM, which was added in 16.1 and a better statistic for measuring the quality of the link.

1 Like

As an update to what I was seeing with Sawtooth like appearance of SnR graphs…
After concluding this had to be un-timed interference, I found that a local fishing resort had set up a small 900Mhz Point-to-Multipoint backhaul for Wifi access-points around their very rural property. They were approximately 1 KM from our tower, and were using Ubiquiti 900Mhz (airmax???) AP, and multiple Nano-stations.
Once we assisted them with replacing this equipment with something suitable, we were able to turn off the Ubiquijunk and the Interference immediately cleared up. Our RF performance, as seen on stable SnR graphs, is 100% better. No more sawtoothing.

Just thought it would be worth updating in case anyone else has seen this.


Thanks for the follow up Neil… this may indeed help others that might experience similar issues!

1 Like

Thanks Neil, it may indeed be interference, these are in 5GHz and it is noisy out there…

I have seen this on a number of SMs on our network, and was unsure if it showed an issue. At this point, I’m leaning more toward it’s just graphing that way due to lack of data/traffic flowing…

I am seeing this on many of our 3.6 SMs running on SAS. SMs with steady traffic even. They are running mostly 16.2.3. The spectrum sweeps indicate clear last time run. Any ideas?