Single AP keeps crashing (unstable power?)

[sbyrd]

I have a WS-12-250AC with 9 radios attached to it. 6 are Rocke-5ac Prism, 2 are Rocket2AC, and one is an AF5X. Recently it has been single digits with wind chills below zero and one of the 2ACs has been rebooting seemingly in the evening and early morning when it is coldest.

The AP is actually crashing as I log into it it has a crash report saying it was due to watchdog airMAX - Troubleshooting Unexpected Reboots – Ubiquiti Support and Help CenterUI support said it can be due to unstable power.

When I looked at the Watts used by that port sometimes I would see it go as high at 16W (port max is 18). I was never to actually see it right as it crashed so I do not know if that has anything to do with it.

Anyway I thought it could be a power draw issue so I turned off the Airview radio (spectrum scanner) and lowered the TX power of the AP in an effort to lower the W used. It seemed to have had a positive affect as it reboots less now, but it still is happening just not as often.

Now according to the UI spec sheets the AP is only supposed to draw around 8.5W. This AP only has 6 clients and is very low average usage.

My question is how is this AP sometimes drawing 12-16 watts and maybe more? Can the extreme cold cause devices to draw more power?

[sbyrd]

Here is an example of the crash history on the first day it started happening. It was also the first day we had temps this cold. That day it stopped as the sun came out and the temps warmed to the 15-25deg F range.

[mike99]

I would test in this order
Connector
Cable
Defective radio
Defective port on the Netonix

But since there 2 free ports on the Netonix, the easiest would be to fist connect to one of the 2 free port to make sur it's not an deflective port issue.

[sbyrd]

What I've done today is I switched the AP from the netonix switch to a standalone UBNT 24v 1amp power supply. My thoughts are if the reboots stop I know it is not the cable or the radio and is either the switch or something related to the power draw of the AP on the switch.

Right now Ubnt support is saying that the 2AC radios have an internal heater the comes on when needed and the AP needs a full 1amp power supply. However, he is checking with the hardware team to confirm as I told him my other more active 2AC on the same tower and switch is not having any issues.

[lligetfa]

Right now Ubnt support is saying that the 2AC radios have an internal heater the comes on when needed and the AP needs a full 1amp power supply. However, he is checking with the hardware team to confirm as I told him my other more active 2AC on the same tower and switch is not having any issues.

Maybe the more active one is staying warmer and so the heater is not kicking in?

[sbyrd]

Maybe the more active one is staying warmer and so the heater is not kicking in?

Possible I guess. Here is what I just got from UI support.

Hello,

I just received confirmation that the R2AC can use up to 1Amp when its heater is engaged.

If possible, could you try moving this unit to a port that supports 1Amp and let me know if you are still seeing reboots?


What is odd to me is they seem to be implying the R2AC can under certain conditions use 24watts (24V x 1amp)? Their device spec sheet only lists the device at max 8.5W. Also what if I was using the UI Edgepoint switches which are also only 24v .75a per port?

Also does not entirely explain why the other 5 R2AC radios I have in my network all powered by the same 24v .75a Netonix ports do not have this issue at all.

[sbyrd]

Well the AP has crashed/rebooted another 7 times since I moved it off the switch to a 1amp power brick so the switch is not the issue.

[mike99]

The radio draw too much power, their was few chance that the source of the problem is the device that provide provider.

[JustJoe]

One would think that the first thing a design team of engineers would put into their product's "development specs" would be minimum voltage required at the input and absolute power consumption under any circumstances.

Just from years of experience with ... that ... company, those always seemed to be things they tried to determine after the product was being manufactured ... Often seeming to wait until reports of users came in from the field. :(

All I can add is that I feel your pain. :(

[sbyrd]

Well I replaced the AP 3 days ago and also put it back on the netonix Poe. So far no more reboots even though it is even colder than it was last week.