Netonix Forums

We installed 3 WS-8-150-AC units at a small hotel to power their APs (xClaim X3 models) at 48V.

They have been running for about 4 weeks with no known issues. Last week we replaced one switch, due to it constantly rebooting. I found that when we had the 6 APs connected to the switch and POE turned on the switch would reboot within 2 minutes of powering on, then continue to cycle through rebooting. Once I turned off POE on 2 ports the switch was stable for 2 days, turned the POE back on and the switch started back into the reboot cycle. This starts happening after turning on the 5th POE.

Yesterday we have a second switch starting to show the same symptoms.

EAch switch has 6 APs installed. All cables and APs have been tested and working properly.

Suggestions??

Do you know how many peak watts the switch was supplying?

Did you have this plugged into a UPS?
I have seen people using too small of UPS that will cause this, what size UPS is it?
If so try a larger UPS or remove the UPS for test.

Another thing I have seen that can cause this is using Ethernet Surge Suppressors that clamp POE wire to ground which will cause the switch to reboot as this is a dead short and can damage the switch as well. You can not use Ethernet Surge Suppressors that clamp load to ground when using "passive" POE switches.

Are you using Ethernet Surge Protectors?
If so try removing them.

If there was no UPS, Ethernet Surge Suppressors, or power issue then this is not a software issue and is obviously a hardware issue so there is no hidden command to fix this.

You will need to RMA these with specific details on how to recreate this issue in our shop.

RMA instructions are located here: viewtopic.php?f=6&t=1259

Using a 700A/400W UPS only this Netonix and an unmanaged 48 port switch are attached.

Not using Surge Supressors

Is the peak Watts shown on the bar under the graphical interfaces of the gui or somewhere else?

Yes, the bar is the total watts

Keep in mind watts reported are averages over seconds and not actual spikes which can be much more.

Believe it or not that UPS is not really sufficient and over time it will damage the UPS and these things can occur.

I tell people 1200 to 1500 watt UPS is sufficient under most cases for our 150-250 watt switches and larger for our 400 watt units depending on POE load.

Could you try plugging the switch into just a surge strip and not the UPS.

The WS-8-150-AC onsite that has not had any issues yet is currently showing 30.5W, we replaced a failing one with a WS-10-250-AC until we could get new units in and it is showing 33.1W, I have not seen anything spike over 35.0W during the time I watched them. Can this me monitored via SNMP?

What is the reason the UPS I am using is not sufficient? We will be in the area of this site next week and will have to remove the switches from the UPS at that time for a test. I have one that was failing on my bench and can work on performing tests to replicate.

The xClaim APs we are using have the following power specifications:

Power over Ethernet 802.3af/at
• Idle: 4W
• Typical: 5.95W
• Peak: 10.5W

Same behavior when removing the UPS

For clarification: you aren't seeing this with the 10-250-AC, am I correct?

the 8-150-AC has several on-board power supplies: in this case, I believe 3.3v is the issue. There was an issue where component tolerance was causing this weird oscillation on the 3.3v rail (after a long period of time) while multiple (5+) PHYs are linked. When 3.3v dips below 3.1v momentarily, the reset controller thinks there's a problem (which there is), and the device reboots. Apparently, you have some of the very few units built before we made that particular catch, I'm sorry to say.

It's a simple hardware fix, but unfortunately I do need you to RMA the unit(s) to put it in place. I apologize for the inconvenience.

Julian, we have these ready to send back. Teo from our office will be calling for RMA

soundwavz wrote:Julian, we have these ready to send back. Teo from our office will be calling for RMA

Don't call follow these instructions
viewtopic.php?f=6&t=1259

I got your switches in, and I've had some time to 'poke around'.

The 8's stability issue does seem to be caused by that bad reel of caps we have rearing its head, those were easy.. What i'm not getting is what's up with the 10-250.

So far, it's passed every test I can think of to throw at it. Is there something in specific it was doing?

Thanks