Problem with a new site - PMP 450 Medusa

[Tuvix_IT]

Hi, my company has installed a new site with 2 new AP PMP 450 Medusa because we have another site (with 6 AP Medusa) overloaded and we wanted to move some customer in the new location.
The problem is that the two new APs seem to work badly.
As you can see from the two scans it seems that there is availability to transmit without problems on a 20 Mhz channel for both APs

AP124

AP125

The download of many users is really bad despite having a good signal, hardly go over 20 Mb in download while in upload they easily reached 40Mb.

These users, if put on the old site with the 6 Medusa, are ok. If put on the new one, 80% will have a very very low download.

LINK

The values of the column "Signal to Noise Ratio (dB)" are very strange...

Here the main data of the AP:

Here a customer test:

If the upload of the AP customers were interfered , I know that it would be enoug to change the frequency of the Access Point to solve the problem.
But how to do when the download is interfered like this?
I can't think that it is a local interference of all customers, it would seem really strange to me because these same customers work well on other location with the 6 AP Medusa that is very close to the new one.

So I ask Cambium how to move when there are these kinds of problems, I need some urgent help 😥

[MW_WISP]

Do the new and old site shares the same parameters for "radio settings" (max distance, channel bandwidth, downlink data)?

[Tuvix_IT]

---

@MW_WISP wrote:

Do the new and old site shares the same parameters for "radio settings" (max distance, channel bandwidth, downlink data)?

---

Yes, configuration exported and imported. Changed only the IP, name, color code and frequency.

[Cambium_AndyR]

Hi Tuvix_IT

Did you set both AP to 5715 MHz ?

The spectrum analyser was run at the same time (AP124 14:48:15 on 10/12/2019, AP125 14:48:11 on 10/12/2019) - so they would not have seen each other (no transmit during analyser).

If you have them the same frequency and not back to back this may cause the results you are seeing.

best regards,

Andy.

[Chris_Fikert]

Hello, here's a few observations and questions.  The poor quality is definitely related to the low SNR on the downlink side (uplink looks good).

- 5715MHz is in the 5.4GHz frequency band.  Not sure where you are located, but that is a very high EIRP for any 5GHz band.  http://community.cambiumnetworks.com/t5/PMP-Best-Practices-and-Examples/Leveraging-the-FCC-DFS-bands-5-2GHz-amp-5-4GHz-with-PMP450m/td-p/96143

- The max range value is very high.  Typically we recommend a value that is slightly longer than the farthest SM on the sector.  However, there are considerations for synchronization and you should use the colocation calculator to ensure that the settings amongst all of your AP's in the same frequency band are agreeable.  https://support.cambiumnetworks.com/files/colocationtool/#r3

     - This value heavily influences the gap time between frame transmissions.  Setting it to a large value will decrease the overall sector capacity, as well as throughput per SM.  http://community.cambiumnetworks.com/t5/PMP-FAQ/Throughput-vs-Range-An-Explanation/m-p/37199#M269

     - http://community.cambiumnetworks.com/t5/PMP-FAQ/Recommended-Contention-Slots-and-quot-Max-Range-quot-setting/m-p/67960#U67960

- Ensure that you are using v16.0.1.1 or v16.1 (if no PMP430 SMs) on both the AP's and SM's.  It appears that some of your SM's aren't on 16.x (which is why the MU-MIMO fields are showing N/A).

- Ensure optimal physical installation parameters for the PMP450m (and SM's).  There are 2 degrees of electronic downtilt built into the PMP450m antenna.  Providing coverage (in general) at longer distances (i.e. 7-10 miles) typically requires physical uptilt on the antenna (based on the AP and SM having similar elevations).  The beamforming on the PMP450m also has some requirements for optimal performance to the SM's.  Providing 'spotlight' coverage to the SM's is recommended.  LINKPlanner is a great tool for determining optimal tilt (and performance).

     - Not sure what type of SM you are using, but higher gain SM's are recommended for most, if not all, installations.  Optimal alignment (mostly for ideal beamforming performance) is recommended.  LINKPlanner is a great tool for comparing a baseline of expected performance vs. actual results.

- Even though your signal looks good, the signal to noise ratio is going to determine the overall performance per SM (and per sector).  It appears that there is isolated interference at particular SM's (versus across the entire coverage area).  This could be caused by: other AP's operating in the same area (on same or adjacent channels, possibly your own if sync isn't working properly), local Wi-Fi AP's operating on the same channel, or other RF interference sources.  Running a spectrum analysis on the problem sites would help (best to also not be transmitting from the AP - initiate a remote SA, then do a local SA on the AP for a period longer than the remote SA).  The goal would be to see other RF interference (otherwise you will see the AP on the SA).

     - Side note, it is normal to see the SU-MIMO and MU-MIMO modulations not be equal.  Typically the MU-MIMO is a modulation less due to the extra 'noise' of having multiple beams in the same coverage area (it's not really noise, but we measure it as noise).  Also, if the SM's aren't transmitting data, the contol messages go out at the lowest modulation, so running a link test will show more realistic modulation levels (something else is going on in your case).

Next steps:

- Upgrade firmware on AP's and SM's to v16.0.1.1 or v16.1

- Verify/modify configuration for optimal values (synchronization/frame parameters and max range, specifically)

     - Should be done on all AP's operating in the same frequency band in the same general area

- Run Spectrum Analysis at the SM's for other sources of interference (ideally with curent AP in SA mode, or turned off)

- Compare current performance with expected performance within LINKPlanner

     - physical installation of AP and SM's

     - physical obstacles as well as RF noise floor locally

- Open a support case and provide output from http://ipaddressofPMP450m/field_diags.cgi 

[rbroch]

hello, 

Can you share how you solved that problem?

Thanks