WS-250-DC dead

[colinhowlin]

Hi

We've installed a number of WS-12-250DC's on our sites as we move to DC power.
Did another yesterday on a water tower.
Deployment went to plan, was passing traffic and everything seemed ok.

Then we noticed some dropped pings and intermittent management access.
All devices connected to switch were accessible no problem - no drops.
Customer traffic was also unaffected.
We had a SIAE AlfoPlus backhaul powered at 48VH, a Netonix MINI powered at 48VH and a number of Cambium/Ubnt radios powered at 48V.
In total, ports 1-12 were all powering a device.
Uplink was through SFP cage on port13.

After a short while trying to figure out why management access was dropping, the switch rebooted.
At this point we went back on site to roll back to the AC switch.
We did try some things before switching back.
We had loaded 1.4.7rc4 before deployment, so we downgraded to 1.4.6 but the issue persisted.
We gradually removed some devices to see if stability could be improved in case of load issue but same problem.
We did notice the input voltage was fluctuating somewhat - normally we see a rock solid line on this but we were seeing variance of maybe .2 or .3 degrees.
Not sure if this was a sympton or a factor but it's not something I'd noticed before.

Back on site today to try another DC switch.
This one loaded with 1.4.6.
Moved backhaul first, then all the customer AP's and finally the Mini.
Not a blip on pings and no issues with management yet.
Input voltage is again rock solid.

I tried to access the switch when back in office but cannot get it to power up.
With a Meanwell RS-150-48 which I used to configure replacement switch we get nothing.
Multimeter at the DC connector points on the switch reads just 2-3V.
Further to this I got the odd spark when reading the voltage.

Any idea what might have caused this?
Dodgy switch or something else?
If you need any other info let me know.

thanks
Colin

[Julian]

Yowza, that's no good. Definitely RMA time.


Looks like the input FETs on your power supply may have shorted.

Not sure what the cause is on this as yet, taken alone they're generally a pretty hearty device.

is it possible to do me this favor?:

once your switch is unplugged, take a multimeter and measure resistance across the DC posts, both ways (red on red, black on black, then red on black, black on red) and let me know what you find.

Like I said, definitely RMA time, regardless of what your readings are, sadly.

Sorry to bear the bad news,

Julian

[colinhowlin]

Hi Julian

Thanks for your response.

Hooking up black to black and red to red, reading jumps about for a bit then settles back on 1.
Black to red and red to black I get a reading of 0.556.

Can you confirm RMA procedure please?

Thanks
Colin

[Julian]

Input looks okay, so that's not it, thanks much though.


RMA procedure can be found here: http://forum.netonix.com/viewtopic.php?f=6&t=1259

Reference this forum post in the last field on the RMA form, please.

Sorry again,

Julian

[colinhowlin]

Hi

We got the switch back today and it powers up ok.
So great RMA service for a start!

Can I enquire what the issue was?
I notice we are still seeing fluctuations of 0.1V on the Input Voltage Graph on Status page.
We usually get a dead steady line on this graph so just curious if this is something to be concerned about.

Thanks
Colin

[Julian]

Hi Colin.

I performed the diagnostics on your switch.

I plugged it into our DC power supply, and it booted right up, amid my vociferous headscratching.

Short answer: I was unable to duplicate your issue. Your switch passed all factory testing with an additional 3(ish) day burn-in test, at which time I noted no abnormalities in switch operation such as management access, powered device operation, input voltage, throughput, etc.

As far as the 100mv fluctuation you describe, here are my best guesses as to what that might be:

1: ADC rounding.. In this case, likely nothing to worry over
2: Common mode coupled noise on DC-: also, likely nothing to get too excited about, better grounding/cable routing can mitigate this
3: Marginal power supply (on your end, not the switch's): Capable of supplying enough current, but barely so.: In this case, can negatively effect service life/reliability of equipment
4: Switching noise from power supply feeding back: Likely, not an issue as the power supply is designed to protect against this

I'm sure I'm missing one or two contingencies, but these would be the most likely to my mind?

Thanks,

Julian