Cambium ePMP 4600 – intermittent disconnection issue at longer distances (PTMP)

We are operating several Cambium ePMP 4600 base stations in PTMP mode and have encountered a difficult-to-isolate issue.

At subscribers located farther from the base station (around 15–17 km, approximately -68 dBm signal level), we observe the following:

  • On the CPE web interface, no disconnection is visible; the session time continues without resetting.

  • However, CnMaestro reports a disconnection event.

  • The customer experiences a real service interruption with no data traffic.

  • When a Mikrotik router is connected behind the CPE, the connection does not automatically recover. It only comes back after rebooting the Mikrotik router.

  • At sites without a router behind the CPE, short, intermittent drops occur.

We are running the latest firmware on all devices.
We have tried disabling the Hardware Acceleration Engine on both AP and CPE sides, but the issue persists.
Router replacements were also performed at the affected sites to rule out CPE-router compatibility problems, but the issue remained.
At some locations, a Tarana link was previously operating without any problems, and the issue appeared only after replacing Tarana with Cambium.
We have now seen the same behavior at two independent sites, which rules out a local network issue.
If the router is removed, the connection becomes stable and stays up.

Common characteristics at affected sites:

  • Distance typically 14–18 km

  • Signal level around -67 to -70 dBm

  • PTMP mode

  • Client devices are mixed: ePMP 4625 and ePMP 4600c

  • VLAN traffic is typically used through the radios

  • When the link is up, throughput is excellent (300–400 Mbps)

  • The problem appears intermittently and unpredictably

Has anyone experienced something similar?
Could it be that at certain distance/SNR thresholds the radio silently drops sessions without re-registration?
Or could this be related to TCP session handling, keepalive behavior, or an incompatibility caused by hardware acceleration or VLAN traffic behind routers?

Any feedback, logs, or workarounds would be appreciated.

Hi @creed23 ,

At first sight it looks like interference, but it strange that Mikrotik is causing problems with link recover.

I suggest you to Raise Ticket in the top right corner of this page and attach Tech Support Files from AP and CPE right after the connection problem occur again. Before Tech Support Files download, please enable and execute Spectrum Analyzer.

Raise a support ticket is free. Our Engineers will investigate and help with your issue.

We have already opened a support ticket, and the last advice we received was to enable the watchdog (which I didn’t really understand). My colleague reopened the ticket this morning. In terms of interference, the situation doesn’t seem hopeless, and it looks suspicious that the affected endpoints are specifically the ones farther away.

Can you give us the ticket number? More importantly this is an interesting issue and I would like to have our Engineers get on the call with you and troubleshoot. Please send me an email with your contact info : Sakid.ahmed@cambiumnetworks.com

This almost sounds like a DHCP issue… the wireless link stays up, but data stops flowing until the router is rebooted. No idea why distance would be a factor in DHCP though. :person_shrugging:

Thank you for shared details, I’m already working on your ticket, and going to help you solve your issue.

1 Like

we can see the mikrotik router on L2 level, but it’s becoming unreachable by mac telnet.