Wisp switch temperature

[highlands]

How can we verify this?
So far this situation matches a temperature issue, based on time of day and recently adding load to the WS-12-DC, the fact it is in the smallest enclosure we have in our network.
The WS-12-DC stays off line for about 10 to 15 minutes and then come back on line with us doing nothing. We can not get to the site that quick, but it seems like that is enough time for it to cool down. Today it did it twice, Saturday once @ 3:47 to 4:03 and once on Thursday for 15 minutes.

Remember the image I sent was after it had come back on line.

Some questions;

Do these units have internal temp sensors that tell it to shut down when the temp is too high?

If this happens does it send a SNMP alert about the high temp and does it do it BEFORE the shutdown? (we do not use log servers only SNMP)

You provide 3 temps on this type on unit, what is the real meaning of each one and where does each on go into a critical temperature.

What can we look for to t-shoot this.

If we RMA this unit does Netonix have a 'torcher chamber' it uses to test with and can this specific unit be tested?

John

[Dave]

our switches all go through design temperature testing at 55 degree C, with a full POE load.

please upgrade to 1.4.3rc7 as we had several bugs that might be related to your issue.

[highlands]

How can we verify this?
So far this situation matches a temperature issue, based on time of day and recently adding load to the WS-12-DC, the fact it is in the smallest enclosure we have in our network.
The WS-12-DC stays off line for about 10 to 15 minutes and then come back on line with us doing nothing. We can not get to the site that quick, but it seems like that is enough time for it to cool down. Today it did it twice, Saturday once @ 3:47 to 4:03 and once on Thursday for 15 minutes.

Remember the image I sent was after it had come back on line.

Some questions;

Do these units have internal temp sensors that tell it to shut down when the temp is too high?

If this happens does it send a SNMP alert about the high temp and does it do it BEFORE the shutdown? (we do not use log servers only SNMP)

You provide 3 temps on this type on unit, what is the real meaning of each one and where does each on go into a critical temperature?

What can we look for to t-shoot this.

If we RMA this unit does Netonix have a 'torcher chamber' it uses to test with and can this specific unit be tested?

John

We have continued to let this situation soak for a bit as we see what happens. We do not want to do more than one change at a time or we will not know what the real cause is/was. Since we live in sunny NV and I do not have to worry about rain for a bit longer, we cracked the door to the enclosure and have had no issues since, however the PHY Temp is going from 57 at night to 64 during the day at that one site. This weekend I plan on closing the enclosure and seeing (again) what happens when the temp hits 71 or higher, since that seems to be where the unit shuts down. This could be a single unit that is temperature sensitive, I do not know. I will say we have other locations where the PHY Temp goes to 75 to 80 and we have not had an issue.

If this weekend during the test it does shut down, then in my opinion, I will fully believe it is a temperature issue, although we are getting a little cool off this weekend and the outdoor temps will be about 10 to 15 degrees less than the prior few weeks. Assuming the unit faults off-line the first step I will take is the UG to 1.4.3rc7 to see if that will affect a change.

Also could you please answer the above questions?

John

[sirhc]

PLEASE UPGRADE TO v1.4.3rc7

This could be your issue, I am not sure what more to tell people to convince them that v1.4.2 has some BAD bugs.

ALL THAT IS IN v1.4.3rc7 IS BUG FIXES FOUND IN v1.4.2 and OLDER AND SOME ARE BAD AND CAN CAUSE ISSUES LIKE THIS.

No this is not a temperature issue, these units have an SOC temp rating of 125C.

[lligetfa]

No this is not a temperature issue, these units have an SOC temp rating of 125C.

That is like saying that "you are paranoid, nobody is out to get you". Someone could be out to get you even if you are paranoid.

What if the switch has a defective SOC and not able to withstand the rated temp? Is there anything else inside the box that could go squirrelly and spew some unintended RF to affect the switch?

This thread has gone 6 pages now but I think highlands just recently jumped in so I don't recall if the switch has been swapped out yet to rule out a bad unit? After putting RC7 on it, if it still has issues then I would swap it out with a known good spare.

[sirhc]

I agree it is not impossible but unlikely, I would like this done first if at all possible.

First please upgrade to v1.4.3rc7

If it happens after that then swap it out and if it never happens again I would buy it is a defective unit but I doubt this.

[highlands]

Update;

We closed the enclosure on Friday due to some rain coming our way. We had no issues during the time the week plus that enclosure was open and the PHY Temp did not get above 65C according to SNMP gets we were doing every 15 minutes. The failure that did occur before was occurring at about 70 to 71C which would lead toward a heat issue, however we have had other units get to 75c and have not had an issue.

We are quickly turning to fall here in northern Nevada and even have a freeze warning in affect at altitude the next few days. I believe that we may not know the real status of the situation till spring/summer. We will wait another week or so and then upgrade to v1.4.3rc7 as we do not like doing too many changes at once or else you lose out on what the real root cause is. As I assumed you noticed (see link sent before) the enclosure is a small one and I did worry from day one that this small of an enclosure would be too small (although it is such a perfect fit like the one for the mini 6 you have for sale), but I was also hoping that it being aluminum would be of help in dissipating heat generated inside the enclosure to radiate it outside without having to vent the enclosure, which would have been near impossible due to its size. In order to vent we would have to add a bigger enclosure with venting.

We do greatly like the Netonix product line. Please do not consider any of this anything but feedback and information gathering. Speaking of which can the previous questions get answered?

Questions;

Do these units have internal temp sensors that tell it to shut down when the temp is too high?

If this happens does it send a SNMP trap or alert about the high temp shut down, and does it do it a few seconds BEFORE the shutdown? (we do not use log servers only SNMP)

You provide 3 temps on this type on unit, what is the real meaning of each one and at what temperature does Netonix considered a critical temperature?

If we RMA this specific unit does Netonix have a environmental testing chamber it uses to test with and can this specific unit be tested?

John

[sirhc]

Do these units have internal temp sensors that tell it to shut down when the temp is too high?

It does not shut itself down no, it will run to 125C PHY/CPU temp with no problem. It will then run until it fails, no shut down

If this happens does it send a SNMP trap or alert about the high temp shut down, and does it do it a few seconds BEFORE the shutdown? (we do not use log servers only SNMP)

Does not shut down

You provide 3 temps on this type on unit, what is the real meaning of each one and at what temperature does Netonix considered a critical temperature?

The PHY and CPU temps are the Switch Core itself and are internal sensors. The Chip is rated to run at 125C all day long.

The board temp is an external sensor and the board should be fine up above 70C

If we RMA this specific unit does Netonix have a environmental testing chamber it uses to test with and can this specific unit be tested?

We do have a chamber in our LAB in Canada but not here in PA but I am considering getting one soon.

The easier test next year would be that if it happens swap it out and if the replacement does not do the same thing then you know it has a thermal issue.

[jdiggity]

I have a straight out of the box new WISP- mini powered by an WS-12-250-AC that experiences the same problem As soon as the sun comes out the switch will completely stop working. The lights are still working but it requires a reboot to bring radios and switch back online. After the sun went down it would be OK until the next morning. We swapped out with another Mini and all problems are resolved.

Not sure what firmware was on Mini. This was mid September when we had problem. We will RMA mini back.