I have an RB433 + XR5 that has been losing power just about every night, goes down and comes back up after a couple of hours. We tried switching out PoE, moving it to a different port on the switch, tried a new patch cable. That still didn't work, it did it the next day, well yesterday we changed the RB433+XR5 plus Ethernet cable connecting the PoE to the RB433. I didn't get a down notification all night so I thought everything was good, but come in this morning and log into the radio and it has an uptime of 9mins. Log says Router was rebooted without proper shutdown (cause 1), says the same thing as the previous RB433. Previous RB433 was running v4.9, current RB433 is running v5.6.

I'm kind of a my wits end because everything has been swapped out, power wise we have moved it from a UPS to a power strip that other devices are on and aren't losing power. I really need some help on this guys, if I could get some help it would be greatly appreciated.

First, check the watchdog timer and disable it.

First, check the watchdog timer and disable it.

I've never enabled that, so if its a default setting it is set on all my routerboards, yet this is the only one that has been giving me this issue. I'll give it a try though.

I disabled it last night but when I came in this morning it had lost power and rebooted about 1.5hrs ago. The good thing is that it seems to be "losing power" not proper shutdown and coming back up, previously we would lose it for hours then it would come back and have that error. Now its just "losing power" rebooting itself and coming back online, but obviously this isn't ideal I need a solution, everything has been swapped I really don't know what else the problem could be.

Hi,

When you say 'switched out PoE' did you replace with an identical one?
It might be a good idea to change one with different voltage/power and see what happens then (don't go over 24vdc).

Rgds,
Mark.

Hi,

When you say 'switched out PoE' did you replace with an identical one?
It might be a good idea to change one with different voltage/power and see what happens then (don't go over 24vdc).

Rgds,
Mark.

Yes we did swap PoE with an identical brand new out of the box PoE, it had been working for a year on an 18v and we swapped it with a brand new 18v. Also used a brand new 18v PoE to configure the current RB433 that is in use now, I don't have any 24v PoE's on hand so swapping it would take some time. But again the 18v has been working for quite some time also the other RB433 thats on that tower with 2 XR5's is using an 18v as well.

Hi,

What you might need a higher power PoE unit as in delivering Amps. This is where the power draw is. If your RB and the XR cards have high usage, then you are going to draw a bit more power than is normal. What are your 18vdc units rated at?

Rgds,
Mark.

Hi,

What you might need a higher power PoE unit as in delivering Amps. This is where the power draw is. If your RB and the XR cards have high usage, then you are going to draw a bit more power than is normal. What are your 18vdc units rated at?

Rgds,
Mark.

18v 1amp.. I took some customers off that sector and moved them over to a stable RocketM5 sector and so how there are only 13 users connected and the max the bandwidth has hit is 5Mbps over the last 24hrs..

Well it did it again this morning, came in and it has an uptime of 1:40:22.. so it "lost power" about 7:50am, I checked every device on that tower that are plugged into that UPS and surge protector, none of those devices lost power. Guys I am really pulling whatever hair I have left out, I can't understand why its rebooting, same symptoms over two RB433, 2 XR5's, 2 Ethernet runs, 2 PoE's, 2 patch cables, 2 different ports on the switch, 2 different ports on UPS and Surge protector. I have 11 routerboards spread across my network and this the first time I've ever seen this.

Get a 24v POE.

You are probably losing too much voltage over that long of a run. XR5 is a power hungry card. I ran into similar issues with the SR9 and 133C combo. One side ran fine 90% of the time, the other (my side, luckily), had some quirks.

Bottom line, if you run any high powered UBNT cards, use a good 24v POE, otherwise you will run into problems.

It could be partially environmental, higher temps cause for more voltage loss, the power feeding the building may not be as reliable (higher swing between 100-130v), any of a dozen different possibilities. 18v at 1A is just not good enough for those devices.

Get a 24v POE.

You are probably losing too much voltage over that long of a run. XR5 is a power hungry card. I ran into similar issues with the SR9 and 133C combo. One side ran fine 90% of the time, the other (my side, luckily), had some quirks.

Bottom line, if you run any high powered UBNT cards, use a good 24v POE, otherwise you will run into problems.

It could be partially environmental, higher temps cause for more voltage loss, the power feeding the building may not be as reliable (higher swing between 100-130v), any of a dozen different possibilities. 18v at 1A is just not good enough for those devices.

I hear you and I will try it out but out of 11 RB's on the network I would say 8 of them are running off 18v PoE and these are the PoEs the distributor recommend. I do use some 24v PoE's on a few RB's but also have had issue sometime where and 18v will work but a 24v wont work and vice versa its happened maybe 3 times. The ethernet run going up to the routerboard is maybe 65ft max, its not a long run and its only running 1 XR5 with 14 customers connecting to that XR5. I will change out the PoE the first chance my climbers have time to do so, in the meantime is there anything esle I can check out on the board config wise or supout wise?

Hi,

I would enable some of the logging options and also save them using to disk option and not echo.
That way you don't lose all the info on reboot and you could maybe see what is happening before the unit actually 'reboots'.

Rgds,
Mark.

Hi,

I would enable some of the logging options and also save them using to disk option and not echo.
That way you don't lose all the info on reboot and you could maybe see what is happening before the unit actually 'reboots'.

Rgds,
Mark.

Ok I set them all to disk, hopefully this gives me some more info, thanks for the tip.

Ok so I was up late last night so I decided to watch the sector see what it was doing, I would be logged in and I would lose connection, after it came back up the uptime had reset like it had rebooted, in the logs however it just says interface info: ether1 link up (speed100M, full duplex). Didn't say reboot although all signs point to it rebooting or "losing power", I tried a few things, disabled the wireless for about 10mins and the unit stayed up, enable wireless and it reboots within 6mins, change frequency no change, removed a client but still rebooted after a few mins. All in all looking at the logs it rebooted about 11 times between 1am to 8am, also the weird thing is that it won't give me any problems until maybe 10pm, so during the day is fine but at night it goes to junk.

What about wireless card config? Are you using high tx-power? If yes try to set it to 17 or default.

What about wireless card config? Are you using high tx-power? If yes try to set it to 17 or default.

Yea it is set to default and I guess I spoke too soon about it never giving me a problem during the day because its rebooted 5 times since I logged in at 9:54am.

I was just checking the RB using the dude and saw its cpu was at 27%, seemed a little high for what is now only 9 customers. I was wondering could that be causing the problem, getting "overloaded", if so how would I be able to solve that?

I was just checking the RB using the dude and saw its cpu was at 27%, seemed a little high for what is now only 9 customers. I was wondering could that be causing the problem, getting "overloaded", if so how would I be able to solve that?

check the "tool profiler" menu to see which process is causing the CPU load

I was just checking the RB using the dude and saw its cpu was at 27%, seemed a little high for what is now only 9 customers. I was wondering could that be causing the problem, getting "overloaded", if so how would I be able to solve that?

check the "tool profiler" menu to see which process is causing the CPU load

Ok thank you, looking at it now idle is like 90+ for the most part..

idle means unused, ie. "free" resources.

I've pretty much come to my rope's end on this one guys, I've already ordered a Ubiquiti sector to put up instead. I have 9 customers on there that can't get consistent internet so I have to make a change, thank you for the suggestions though guys.

I have had problems with these motherboards, the cause of problems like yours are filters that are damaged, I have changed and the problem is resolved. Usually these filters are 680UF and 6.3V.

I was just pulling customers off that sector on to other ones when I saw your response, looked at the old unit that came down and sure enough they were expanded and one looked to have had a leak. Thank you so much for the info, I am going to order some and replace these, see if it does the trick. The ones in purple are the ones you changed out? I also see the bigger one in purple did you replace that as well?

if you can change them all, then exchange them. do not know why these cards have poor quality filters, because this problem only I've seen in these models .

if you can change them all, then exchange them. do not know why these cards have poor quality filters, because this problem only I've seen in these models .

I never got to thank you for your help on this issue, out of the 11RBs on the network we've had to replace 4 since you recommended this solution. We're even thinking of replacing the rest before they start giving us headaches.

I just replaced the Caps on 3 RB411 boards with bad caps (the green/tan ones). Now i have 3 working boards. I plan to burn them in for a while before using them in a working CPE.