Wired Intelligent Edge

We have fourteen of the 3500yl-24G-PWR (J8692A) model switches in use in our environment.  We are planning on deploying six more.  We are not using the PoE features.

In the past year or so we've had five of these switches fail in some way, each time related to problems with the power supply.  In each instance HP replaced the switch with little to no questions asked.

We had another failure over the weekend, the switch spat out a bunch of errors over the time span of eight hours then eventually rebooted.  Here is a portion of the logs captured via syslog:

2015-02-08 12:43:31.197	 00569 chassis: Ext Power Supply 1: -50V Fault Failures: 2391.
2015-02-08 12:43:30.095	 00573 chassis: Ext Power Supply 1 disconnected.
2015-02-08 12:43:29.213	 00569 chassis: Ext Power Supply 1: -50V Fault Failures: 2390.
2015-02-08 12:43:22.094	 00573 chassis: Ext Power Supply 1 disconnected.
2015-02-08 12:43:19.444	 00071 chassis: RPS Supply Removed Supply:EXT
2015-02-08 12:42:47.702	 00071 chassis: RPS Detected Supply: EXT /n
2015-02-08 12:42:47.702	 00071 chassis: Power Supply Failure: Supply: EXT, Failures: 2278
2015-02-08 12:42:46.645	 00071 chassis: RPS Supply Removed Supply:EXT
2015-02-08 12:42:30.250	 00071 chassis: RPS Detected Supply: EXT /n
...
2015-02-08 04:54:23.017	 00573 chassis: Ext Power Supply 1 disconnected.
2015-02-08 04:54:22.019	 00569 chassis: Ext Power Supply 1: -50V Fault Failures: 2091. 

 We've been told in the past that error messages like these are related to the PoE features of the switch.  We are not using PoE so we have "no power-over-ethernet" set for each interface.  That setting doesn't appear to help.

Does anyone know what's causing this?  Is this something that can be prevented with a specific configuration or firmware version?  For this specific switch we are running firmware version K.15.10.0009 (date 02/08/13).

Thanks.

I used to manage about 40 of these switches and I don't recall ever encountering terminal power issues with any of them.

The power supplies we used were 240V.

Our power was very clean, but we were frequently struck by lightning and the switches were not on UPS, so several times in the year about 1/3rd of the building would be blacked out by a power surge. I don't recall any switch power supply failures, though.

Here's the kicker: we're not using RPS/EPS.

I finally got some very helpful information from HP.  Apparently this is a known manufacturing defect for several switch series.  I'll include some of the information below.

Unforunately something like half our 3500yls are suspect to this failure.  We just replaced our sixth switch tonight and we have another one in the death throws.  I'll be asking HP to just replace any remaining switches that fall within the manufacturing range (can be identified by serial number) to get ahead of this problem.

---

DESCRIPTION
Some HP 2610, 2910, 3500 and 6200 series switches may encounter Power Supply Unit (PSU) failures due to loose fitting pins on PSU connector to main board.

SYMPTOMS
Visual symptoms are as follows (further detail about items 4, 5, and 6 may be found below).
Switch may appear to be powered down with no signs of life
Switch may get into repeated reboot cycles (might appear that the power is ON, but keeps rebooting).
Switch may appear to turn ON for a while, then power down, then power back ON. The whole cycle may repeat.
Fault LED along with the front panel EPS and / or RPS LED blinking amber*
Fault LED along with the front panel PoE mode LED blinking amber (only applicable for PoE products)
Fault LED blinks
*Logs associated with item 4 above (these logs might appear even if NO external EPS / RPS 630, J9443 OR EPS / RPS 620 J8696A is hooked to the system).

ROOT CAUSE
Some PSUs were manufactured with crimp on pins in the connector that mates with the main switch board.