Cisco 3750 Power Issues?

[GREENLOANING]

The company I work for has just installed a significant number of Cisco 3750 switches in their wiring closts. These switches are linked to a 6500 core. At the weekend, there was a site powerdown to allow the Facilities department to test UPS. Basically, they wanted to swap UPS supplies (of which ther are three grouped together to form one supply)in an out but maintain a constant UPS feed.

When they did the swap, around 80% of the 3750's appeared to switch themselves off. I am guessing there was some sort of power fluctuation (whether a reduction of some sort of spike). When I looked at the LEDs, none were on initially making me think the switch had either blown or the plug fuse had blown. As it turns out, by popping the power lead in and out, the switches came back up.

We repeated this test with the same result.

Has anyone seen this problem before. I would have thought that the switches would power back on even if there was an initial drop in power. However I am not aware if there is some sort of "protection" on the switches to this type of situation (drop in power or power surge).

 

[ADB100]

There is a known issue with some of the PSU's installed in Catalyst 3560-24PS's whereby the PSU latches up when there is a fluctuation in the input supply. This only affects certain PSU's. The bug only specifies the 3560-24PS but it does sound similar to what you have:

Bug id: CSCeh20636

A small number of customers have reported a power supply
latch-up condition in the Catalyst 3560-24PS that requires user
intervention to recover. We have traced the issue to a condition within
a specific power supply used in the 3560-24PS. Specifically a sub
one second power dip or power outage can cause the power supply to
enter a safety shutdown state requiring a power cycle to resume normal
function. This condition can only happen when the high range input to the
power supply is used. That is, when the 3560-24PS is operating in an
environment where the mains-power is above 200 volts.

All well engineered power supplies have built-in mechanisms to protect
from externally induced mains-power anomalies. Industry standards specify
tolerable ranges for power line anomalies. All power supplies used in the
Catalyst 3560-24PS meet these requirements. In this specific case the
power supply is entering a latch-up condition as it cycles too quickly
through the auto ranging function. The latch-up condition is easily cleared by
power cycling the device.

The best solution for this type of issue is to ensure consistent power is available
to the device, as most equipment will fail or reset with this type of power anomaly.
In this case the undesirable consequence is that user intervention will be required
to clear the condition. Again, this will only happen with power dips of
between 250 msec and 500 msec and the mains-power is 220 to 240 VAC.

http://www.cisco.com/cgi-bin/Support/Bugtool/onebug.pl?bugid=CSCeh20636&Submit=Search

I would raise a TAC case and see if the same issue affects some of the 3750 PSU's.

HTH

Andy