Page 1 of 2
WS-8-150-DC fan/temperatures - potential problem?
Posted: Sat Jun 24, 2017 1:48 am
by capinot
1.4.8rc6 4 24V POE devices - total power about 30W for this test. We'll be running at higher loads for installed sites.
Have been running this on my desk as bench test prior to deploying at remote sites. Ambient room temperature has been 65-95F
The fan runs pretty much full-time when above 70F. When running it sporadically sounds as though bearings?? are going.
Since we'll deploy in weather tight enclosures, we're concerned about heat levels and fan failure. We have been running Alix boards with OpenBSD as border routers for some time in these environments with no problems. We're moving to pcengines apu2 boards so we can do mrtg and smokeping monitoring from our border routers. Obviously, these more powerful boards will run hotter and contribute to heat build-up too.
We've been monitoring the netonix switch with mrtg (5 min sampling) and also an adjacent packetflux base unit reporting temperature (reports a bit higher than room ambient, but indicative).
Here are mrtg graphs of packetfulx temp, netonix temperatures and fan speed for the last 30 hours. We had high heat on the previous weekend, so I'm including separate 30hr and 1 week mrtg graphs for the netonix temperatures and fan speed (only stared fan speed previous weekend).
Sooo.... are we likely to see fan failure and subsequent switch failure when deploying?
Re: WS-8-150-DC fan/temperatures - potential problem?
Posted: Sat Jun 24, 2017 10:28 am
by sirhc
The fan has no bearings, it is a magnetic lev fan. The sound is from pulse modulation.
Sometimes the fan screws are too tight which can warp the fan housing but that sound is obvious rubbing of the fan blades.
The fan will run on low speed often as that is how it is designed.
I would not put it in a SMALL non vented box painted black in direct sunlight but it is rated to run in a 55C environment but is the box is small and non vented it might get hotter.
The unit will warn you via SMTP alerts if fan fails.
The CPY and PHYs are rated to run up to 125C.
Put a couple vents in your boxes
https://www.google.com/search?q=nema+bo ... s&tbm=shop
Re: WS-8-150-DC fan/temperatures - potential problem?
Posted: Sat Jun 24, 2017 12:19 pm
by capinot
Ok the fan failed this morning. From the log:
Jun 24 06:04:07 switch[1345]: Fan 1 is failing (0 RPM, cooling level 255, board temp 43 C)
Don't think the sound was blade rubbing, sporadic and more vibration like?
We have temperature monitors in all our enclosures, Highest registered temp (varies by location, sun exposure) is 40.9C
Would never use a black box and do have vents, try to avoid sun exposure too.
Ok, now do we worry about temperatures w/o fan?
We're currently at low ambient room temp.
Below is a screenshot of temperatures and fan speed.
top is packetflux temp
next is board temp (red), DC/DC converter (blue)
then cpu temp (red), PHY (blue)
then fan speed
Re: WS-8-150-DC fan/temperatures - potential problem?
Posted: Sat Jun 24, 2017 1:45 pm
by sirhc
OK, then you do have a bad fan, it does happens, or you did not upgrade the firmware as FALSE fan failure alerts was an issue fixed in firmware?
If you have a DC switch you should be running v1.4.8rc7
RMA it for repair:
viewtopic.php?f=6&t=1259
Re: WS-8-150-DC fan/temperatures - potential problem?
Posted: Sun Jun 25, 2017 1:01 am
by capinot
Well, interesting. As I was preparing a RMA request the fan started up spontaneously (after 12hr).
The temps dropped rapidly after fan startup (within the 5 min snmp sampling for mrtg graphs).
From the log: Jun 24 18:30:09 switch[1345]: Fan 1 is now working
I went ahead and sent the RMA request.
Here's the mrtg graphs showing temp drops and fan start-up
Re: WS-8-150-DC fan/temperatures - potential problem?
Posted: Mon Jun 26, 2017 9:22 am
by sirhc
Did you upgrade to v1.4.8rc7 as I requested?
Re: WS-8-150-DC fan/temperatures - potential problem?
Posted: Wed Jun 28, 2017 1:43 am
by capinot
Yes, had been running v1.4.8rc6 since it's release to take advantage of "- improved fan control on DC models"
I hadn't upped to rc7 since the release notes showed "- Fixed SNMP support for Q-SWITCH-MIB, LLDP and CDP - RC7" which wasn't yet an issue for us.
The upgrade didn't stop the temperature increases until the fan spontaneously started about 3 hr later.
The fan is still sporadically noisy and I do hear "throbbing" when it's running (most of the time). Much louder than first 2-3 months of operation.
Re: WS-8-150-DC fan/temperatures - potential problem?
Posted: Wed Jun 28, 2017 2:38 pm
by sirhc
Your options are to RMA it or buy a replacement fan and put it in yourself.
RMA:
viewtopic.php?f=6&t=1259Replacement fan:
https://www.netonix.com/replacemet-part ... an-51.html
Re: WS-8-150-DC fan/temperatures - potential problem?
Posted: Sun Jul 02, 2017 11:51 am
by capinot
Fan just failed again this morning.
From the log: Jul 2 04:03:06 switch[1345]: Fan 1 is failing (0 RPM, cooling level 255, board temp 44 C)
I did submit an RMA om 24 June, but have not had a reply.
I'm perfectly happy to buy and install a fan, probably quicker than RMA.
What does it do to warranty?
Re: WS-8-150-DC fan/temperatures - potential problem?
Posted: Sun Jul 02, 2017 11:56 am
by capinot
As a backup I was about to order a fan, but the link to the fan description does not list the ws-8-150-dc, is this just an oversight?