RB5009 PoE out constant cycling

RouterOS General
1

I have a RB5009UPr+S+ with two RBcAPGi-5acD2nD Access points and a GPON smart SFP stick.
All have the latest firmware and software 7.11.2 .

Log message: ether7 detected poe-out status: (cycling between on | disabled | power_reset)
Log message: ether8 detected poe-out status: (cycling between on | disabled | power_reset)

I have tried the 24 V 2.5A| 28 V | 48V 2.0A power supplies.

All devices were on a CRS112-8P-4S-IN with the three power supplies without issues.

Went with the RB5009 to implement Docker.

What could be causing this on the RB5009?

2

So absolutely nothing has changed except the RB5009?
How did you configure POE? auto on/forced on? If not the latter, can you give that a try?

3

Did try forced on. Have it currently off along with a power injector supplied with the AP

4

/interface/ethernet/poe> print
Columns: NAME, POE-OUT, POE-VOLTAGE, POE-PRIORITY, POE-LLDP-ENABLED, POWER-CYCLE-PING-ENABLED, POWER-CYCLE-INTERVAL

NAME POE-OUT POE-VOLTAGE POE-PRIORITY POE-LLDP-ENABLED POWER-CYCLE-PING-ENABLED POWER-CYCLE-INTERVAL

0 ether1 off auto 10 no no none
1 ether2 off auto 10 no no none
2 ether3 off auto 10 no no none
3 ether4 off auto 10 no no none
4 ether5 off auto 10 no no none
5 ether6 off auto 10 no no none
6 ether7 off auto 10 no no none
7 ether8 forced-on auto 10 no no none

5

Try bisecting the problem. Unplug ether5-8; if that makes the symptom go away, either the problem is in that group or the PoE subsystem is overloaded. If it doesn’t help, plug those four back in and unplug ether1-4.

Bisecting then means you cut the test sets in half and go again. If you isolated the symptom to the first half, you test 1 and 2, then 3 and 4. Else, you test 5 and 6, then 7 and 8. After that, you’re down to one culprit in one pair.

Max tests needed for a single failure in n elements is log₂(n); 3 tests for n=8.

6

Is there a system schematic showing the bridge chip(s) and power distribution somewhere?
I can try say ports 4 and 8.

7

Good call!

So one of these APs is on a long ethernet cable the other very short 6ft.
Having on port close to each other such as 7 and 8 was causing some issue. Moving to 4 and 8 seems to be stable.

THANK YOU

8

If cable length was the problem, I would have expected the 48V PSU to fix it.

As far as I’m aware, there’s nothing in the documentation saying why port 4 should behave differently from port 8. The only guess that comes to mind is power dissipation and cooling, where moving the cable means you’ve reduced the ability of one IC to heat a neighboring IC past its limit. That’s not a great solution since the air inside the case is basically static in this model, lacking active cooling, so that over time, the problem may recur. You might want to contact support about it.

(This forum is user-to-user, not a primary channel for getting the attention of MikroTik support staff.)