RB-750UP w/5.21 and POE Firmware 2.0 Weird issue

Background:
We recently upgraded a production RB750UP to 5.21 and also updated the POE Firmware to 2.0

Prior to upgrade, the 750UP was powering two Ubiquity Radios one on port 5 and the other on port 2.
Working without any issues. :slight_smile:

After the upgrade, the POE Firmware upgrade, the router keeps power-cycling the POE on port 5.


We replaced this router with another one … (did not do any firmware upgrades), and brought this one back for bench testing.

Observed behavior:-
…Router works as expected, until a cable is plugged into the Port 3.
regardless if it is a device that needs poe or just Ethernet)
does the same behavior after a factory reset.
does the same regardless of POE being off
does the same even if the Ethernet Port is disabled.
Winbox shows that RB750 is putting out some current on port 3 regardless of anything plugged in, and or the POE being turned off.

However if ROS is downgraded to 5.16, it behaves normally.

Question:-
How can we Downgrade poe firmware to 1.0 ?

Is this possibly a problem with the Firmware ?

Any suggestions for further troubleshooting ?

Thanks.
Faisal

please contact support@mikrotik.com for v5.22rc1 as that version has fixes for port5 problems.

Deployed a 750UP powering 3 x RB411AH and one Ubiquiti 3.65 Nano 10 days ago. The POE feed line to the 750UP is about 220 feet. Running off a PAC Wireless 24Vdc injector. Power level (system>health) at the 750 is reported as 23.4 to 23.6 Vdc. The jumpers from the 750UP to the connected devices are 10 to 20 feet. Every jumper going to the RB’s has two PAC Wireless TP-ESP-1000 Ethernet surge suppressors; one at the 750 and one at the RB. The Jumper to the Nano has 1 TP-ESP-1000 Suppressor on it.

Yesterday one of the RB411’s, on Ether5, and the Nano, on Ether2, started rebooting. At the same time the RB411 on Ether4 shut down completely.

Turning on only one port at a time and Looking at the POE output the Output current spikes over 1.2 Amp, and continuous is around 600ma on either Eth2 the Nano port or the Eth5 RB411 port. The RB411 on Ether4 will come up for a very short period of time and then reboot again, but only if the POE on all the other ports if OFF. Auto doesn’t work at all now (it did). I must use Forced on.

Anywhere else that I have tried something similar I see the nano drawing less than 80ma, and the RB411 typically 120ma up to maybe 200ma under full load.

This all worked fine for 9 days, the past 4 days with full load. POE Long cable was previously disabled. Tried enabling it and no change.

So I Just tried 5.22 and the results are actually worse - now I can’t power up any of the devices on any ports.

1.2A is too much and it will definitely turn off the port.

Also, with 5.22 has new poe controller software revision 2.1, that resolves some issues seen with 2.0

you could attempt to get higher voltage PSU (28V 29V) with that you might have better luck, since amps will be lower giving more space fore measurement error.

It is nice, but why is not it written in the changelog???

But a Nano doesn’t draw 1.2 amps. The RB is reporting that it does, but they are less than 200 ma under full load. I’ll try 2.1 firmware.

2.1 firmware installed and the three RB411AH’s powered up again.

But why does ether 5 show 510ma but the other two range between 80 and 200 ma ? There’s no load on any of them (No associations) and the Three units are configured exactly the same.

How can one draw 2x the current of the other TWO combined ?

Thx,
Z

Bad cable crimping = more resistance = more current being pulled?