Successfully using PMP450 3GHz/CBRS LTE collocation tools

Recently a PMP450m 3GHz/CBRS AP started experiencing interference and despite quite a bit of troubleshooting, including swapping out the AP, I was unable to resolve it. As a last ditch effort, I decided to try Cambium’s LTE collocation tools.

Symptoms were:

  • Poor uplink modulation to a high number of SM’s, especially during peak hours
  • Downlink modulation was actually pretty good for most SM’s

Here’s what it looks like when a PMP450m 3GHz AP can experience from a nearby LTE eNb. AP-> Tools → Link Status page:

Here’s an aggregate speed test at non-peak hours:

Here’s an SA from the AP:


NOTE, we’re using a 40MHz carrier at the very bottom of the band with a center at 3570. The AP is seeing noise at about -60 in that part of the band.

What’s really strange is that all of the SM SA’s show very little noise, most are similar to this or just a tad bit noisier, but still very quiet by most measures:

We ran some scans in the area with a 3GHz LTE UE but it didn’t seem to find anything. Running a scan with an LTE UE can be useful because it will typically reveal what TDD config type and special subframe type the eNb is running. Here’s an example of what you can find:


NOTE! This is for a B41/2.5GHz eNb, but a 3GHz/CBRS eNb’s scan will look similar. The EARFCN can be inputted into a converter like this to give you the eNb’s center channel.

As a last ditch effort we decided to play around with Cambium’s LTE collocation tools. We reviewed the guide and downloaded the tool. In the collocation tool there are a number values to consider, and after playing around with all of them the most important setting appears to be the PMP450 “Downlink Data” value. The values we ended up choosing we believe are a good fit for our WISP and most WISP’s and give the greatest chance for timing success without knowing exactly what timing settings the LTE operator is using.

Most 4G LTE operators are going to use a frame configuration of 2 (aka TDD config 2), while some will choose a s frame configuration (aka SSF configuration) of either 5, 6, or 7… with 7 being the most likely. The good thing about the config we chose is that it will have proper timing with SSF’s of 5,6 and 7. Here is a shot of the settings we ended up using:

Here’s what that looks like on the AP → Configuration → Radio tab:


NOTE! you have to use 5ms frames!

After implementing these changes were were delighted to find that the uplink modulations for all SM’s had improved dramatically, along with some minor improvements for downlink modulations as well!

Aggregate speed tests are looking quite a bit better as well!

Here’s what latency/jitter and packet loss look before and after:

In closing, I wish I would have tried out Cambium’s LTE collocation tools sooner! For some reason, because we weren’t really seeing horrible noise and the LTE scans we ran didn’t come up with anything, we kind of ruled out these issues being caused by LTE interference. This should have been something I tried very early on :grin:

This write up is dedicated to @LuciaCambium

5 Likes

Even after all this time, there are still many small operators using LTE who are not running on CBRS, but rather just parked in the old 3650 band.

Whole-heartedly, I second Eric’s use of the co-location tool to resolve LTE interference issues. Make sure your GPS is working properly on your sectors, then apply the recommended settings from the tool, and UpLink (SM to AP) interference is reduced by a signifcant margin, and performance will increase.

1 Like

Thank you for posting this!! Huge help and insight!

1 Like

Thank you Eric! We enabled LTE colocation in a couple areas a few months ago and saw significant improvements in some areas, but not much in one area. I wasn’t sure about the S frame config and you’ve cleared that up for me. The one area we are still having some issues is more urban and I’m wondering if we need to try 5G colocation.

2 Likes

I have thankfully not had to deal with any 5G 3GHz/CBRS radios, and so I don’t really have any advice for anyone as to the best settings to use. If you end up having to implement something and it’s successful, please report back.

We have been advised by our SAS to try this in an area where the modulation has been changing regularly. Spectrum Analysis shows almost the entire CBRS spectrum at a -60 for a South facing sector, but the North/East/West sectors are not seeing the same noise. Per the SAS, we have an LTE operator in the area that is within a mile or so of our affected AP. We had been dealing with the issue by changing channel on the affected AP, which helps for a few days, then the modulation tanks again and we are forced to move channels again. We are not seeing any issues with the grant. From the tool, it looks like we would need to adjust our Downlink Data from 75% to 80% as well as enabling Co-Locate TDD 2 in order for the frame gaps to line up.

A few questions:

  1. The Co-locate guide refers to a geographical region, but does not specify what sort of distance that region refers to. 2 miles? 15 miles? Asking for clarification because we have some sites that are as close as 1 mile apart with 4 sectors of 3G CBRS on each.

  2. Would it be detrimental to just enable co-location on all of our sites? Due to our density and the “geographical area”, we can foresee a cascading effect where we have to enable it for this site, that is close enough to that site, that is close enough to this other site.

Any insight would be greatly appreciated

I believe the geographical region it refers to is whatever region the interference is being experienced… which can vary, depending on environmental factors, EIRP of the operators, etc. From a PMP450 optimization standpoint, your max distance setting should only be 1-2mi over your farthest SM.

In regards to enabling co-location across all of your sites, I’d recommend this, and this is something that I’ve done on my network to not only provide sync with my PMP450 AP’s, but also to be prepared for instances when LTE providers light up new towers.

1 Like