Question about Spectum Analysis

We hung a connectorized ePMP 1000 and Cambium 90º sector on one of our hilltops to test against some UBNT gear and I had a question about the spectrum analysis in these units.

I scanned the band, and it seems the very lowest noise 'average' it hears is -78dBm in the public safety band!  I expected to see the lowest at 5Ghz in the -90's at least with spikes where spectrum was being utilized, but not mid -70's.

I've attached a couple band scans from the Cambium, as well as two airview scans from Ubiquiti Prism AC equipment, one hung on a 60º, and one hung on a 45º facing the same way as the Cambium.  

Any help would be greatly apprecaited!

Well as far as the spectrum results says.It seems that the environment has high noise or there may be some other vendors too who are using the 5GHz band.For noisy environment there is nothing much which you can perform. But But.......We should focus on our primary goal (as a Network Expert) of "network performance". How we can achieve it even though there is noise on the floor.

There are few but very very important stuff you can perform to avoid noise even if it is there.So the steps are :-
1.) First of all go to https://support.cambiumnetworks.com/files/epmp/#r3 this link and download the Capacity planner tool..Fill the details in this macro sheet as per your device details and configuration.Below are the guidlines to help you on this :- 

"""""

Region :As per device configuration.
Frequency band :As per device configuration.
Fade Margin: Not really required.
Channel Bandwidth: Fill it as per device configuration.
AP type:Fill it as per device hardware type.
AP antenna system: Fill it as per device hardware type.In your case 90degree.
3rd party connectorized SMs in sector?: Yes or No as per your deployment.
AP 3rd party antenna gain: If any
AP 3rd party antenna cable loss: If any
3rd party connectorized SMs in sector? : If any
3rd party conn. SM antenna gain: If any please mention.
3rd party conn. SM antenna cable loss : If any.
Environment : Line Of Sight or near Line Of Sight or Non Line Of Sight.
Range unit : As per device configuration.
Max range : As per device configuration.
Frame duration : As per device configuration.
Number of SMs/sector : As per deployment requirement.
Duty cycle : As per device configuration.
Max Combined AP Tx power (min MCS): As per device configuration.
Interference measured? Yes
SM antenna for measuring interference : Enter as per SM antenna type ?
Downlink interference level: AP side noise level on perticular band.
Uplink interference level: SM side noise level on perticular band.
% of SMs with uneven paths (single stream MCS): not really required.
SM antenna for displaying data: SM hardware type Integrated or Connectorized ?

"""""

2.)So the catch here is if you enter all realistic information (as per current site environment conditions)on this tool you would be able to know about the details like :-

a.) LINK BUDGET FOR INTEGRATED ANTENNA

 This contains AP/SM's Rx Power and MCS levels at each condition.
b.) SYSTEM GAIN

  This will give you the AP/SM's complete system gain at each MCS level.

c.) COVERAGE AND CAPACITY.

  This is last but not least and it is very very important. The complete coverage and throughput of your 90degree or 60degree or 45degree sector at each and every MCS level.

I would like to repeat and give a temporary but permanent fix for Noisy environment.

1.) Either you should use high transmitt power (Please do not exceed regulatory boundaries as per your region,Cambium do not emphasis this practice or neither suggests).Meaning use max Tx power on AP as per your regulatory standard.

2.)If still issue persists or you still see the performance issues: Then use reflector dishes or connectorized or Force110/Force180 High gain Subscriber Modules at customer site.

3.)Always keep an eye on MCS value.If device achieve higher MCS value like MCS15 or MCS14.This will give you max throughput.Rest you can do R&D in Capacity Planner to understand the catch.

NOTE: Always remember that the noise would always not be at AP side,It can be at SM side as well. So we should consider both values and enter them in Capacity planner as mentioned above in the steps.

I hope this solution will help you understanding the "Noise" related issues for future as well.

1 Like

The unusually high (and flat) noise floor in PSB is very odd.  My first suspicion in the upper part of the band would have been that you were seeing the Ubiquitis, but that wouldn't explain 4.9.  It really does look like the noise floor across the entire supported band starts about -75 instead of about -100, and for a radio right next to it to NOT see that suggests it's internal on the radio that DOES see that.

SO I'd have to suspect a problem with the ePMP AP - do you have another you can swap into place?  Just transfer the config and replace the radio onto the same sector antenna...

j


@newkirk wrote:

The unusually high (and flat) noise floor in PSB is very odd.  My first suspicion in the upper part of the band would have been that you were seeing the Ubiquitis, but that wouldn't explain 4.9.  It really does look like the noise floor across the entire supported band starts about -75 instead of about -100, and for a radio right next to it to NOT see that suggests it's internal on the radio that DOES see that.

SO I'd have to suspect a problem with the ePMP AP - do you have another you can swap into place?  Just transfer the config and replace the radio onto the same sector antenna...

j


This is what makes me think it could possibly be the AP, because out of the 1000+ radios I have deployed (almost all UBNT), I have never seen this across the entire spectrum like that.  It's as if the radio has a pre-amp in it that is ~ 20dBm and any noise in the band is killing the receiver.  I have a couple other connectorized radios I could try also when I head up to the hilltop next week.  

Hi everyone,

I'm seeing exactly the same in one new ePMP 1000 GPS radio today!

Look at these screenshot!

I can see about -67/-70 noise floor accross all the visible spectrum (4900-5970), and it's very, very strange.

I'm going to try to swap this radio with another one...

Well I would suggest rather replacing the AP.We should first try to upgrade the software version on radio. Because some time if nothing works on technical troubleshooting then "WE" as a troubleshooter move to software upgrade. Just to make sure that all bug fixes can be applied through new software. If that doesn't work,I would recommend to perform hardware swap.

Hope this information will help you in future as well,as a best troubleshooter.

1 Like

@Deepesh Sharma wrote:

Well I would suggest rather replacing the AP.We should first try to upgrade the software version on radio. Because some time if nothing works on technical troubleshooting then "WE" as a troubleshooter move to software upgrade. Just to make sure that all bug fixes can be applied through new software. If that doesn't work,I would recommend to perform hardware swap.

Hope this information will help you in future as well,as a best troubleshooter.


The person above is also on 2.6.1.  Isn't this the most recent version?

Just opened another non-GPS unit and attached it to a UBNT Titanium Sector here in the office and loaded 2.6.1.  It's spectral graph looks as it should.  Guess I got unlucky and got a bad radio then?

Wow, took that replacement connectorized non-GPS unit up to the hill and swapped it with the one I thought had a problem.  This one hears the same nonsense.  It's almost like the co-located UBNT Airmax radios are overdriving the receiver and putting a -75dBm noise floor across the band, even in areas nothing is transmitting.  Interesting.  Radio to radio test to my dad's gets 90-95Mbps to his house, and 12-16Mbps back up to the AP at 20Mhz, flexible TDD.

I'm still following up on this. I have the GPS unit and haven't seen that yet. I'll steer away from non-GPS AP's for now until  I see there's a solution for it.

I have a GPS version and my problem is the same, but I didn’t solve it by changing the AP.
I think the problem is the antenna positioning near a wall. I’ll change the mount position next week!

67firebird455, your upload is low for a flexible setting, and I’m sure it’s related to that noise floor.
Did you try another antenna mount or another antenna? Maybe UBNT spectrum analyzer works in another way so you don’t see the same noise even if you have…

It would make no sense as to why the cambium is hearing ridiculous noise across the entire band, including public safety.  Only time I've ever seen something like this is when a radio is poorly shielded and getting hammered by an airfiber or something.  

I didn’t have any shielding issue on the radio even in very crowded towers with tons of other equipment.
In the place I’m having this issue I have a wall behind the antenna, and maybe the signal is reflected on the wall so I can see it as noise on the front even if the Front to Back ratio is very high (> 34).
The same radio doesn’t have the same issue in another environment.
I’m trying the radio and the same antenna in the same place but not in front of the wall, I’ll keep you updated.

Any update?  The ePMP 1000 is very fast to the client, but having a heck of a time upstream from an SM, no matter what channel.  We're seeing 85-95Mbps down, and max of 16Mbps up on flexible TDD mode to a single station on a 20Mhz channel.  The fact that I swapped radios, and still see the noise like I originally did, something is going.  It's like the UBNT radios are desensitizing the receiver in the Cambium because these units don't hear this type of noise in our office connected to a sector, and no way is the public safety band (unused) at best, -78dBm 'average'.  Something is very wrong here, and I'd like a little help getting to the bottom of it.