APC 2200 Question

[BillP]

Hopefully this is an easy question for the board.

I have a 2200 (SUA2200R2X147) UPS that hosts two of our three 3750 -E Core switches, 1 2960 G Access switch and 1 SonicWall. Over the last two months we have taken two power hits that have caused this particular UPS to have to switch to battery power. When the UPS does so, the two core switches drop and reboot. The other two devices plugged into this UPS remain online and operational. I called and spoke to the tech support from APC yesterday and they recommended doing a brain dead test on the UPS. I did so this morning then took a spare 2960 and a fan and plugged both devices into the UPS and began testing. I switched the Cisco switch to test out all the plugs in the back of the device and the end result is the UPS works with the 2960. At no time did this switch reboot or lose power. Is there a known issue with the 3750s and this particular model? I read the boards and have seen that the reaction time of the devices is anywhere from 2-4 ms...is there a way to change this setting if in fact this is the reason we keep rebooting those devices? Or sadly will a replacement UPS need to be bought to support these particular devices, seeing as how the current UPS is not able to keep the power going to them?

Thanks in advance for any help given.

Jeff

[TheNotoriousKMP_apc]

The SUA family of Smart-UPS exhibit a 2-4 millisecond transfer time to battery and back online. It seems as if this is definitely the problem with the other components dropping during that transfer. In order to rectify this problem, you would need a Smart-UPS with a lower transfer time, and the only family line that covers that is the Smart-UPS RT.

[BillP]

hi jeff,

you are indeed right. the transfer time to battery is 2-4ms and it is not adjustable because it is a characteristic of the UPS topology, in this case, line interactive. we have UPSs that use a double online conversion which has no transfer time - Smart UPS RT. any of those UPSs or anything larger, should have no transfer time to battery which is needed for some sensitive equipment/servers, etc as you've discovered.

i personally have never come across this particular switch to say i've seen it before

[here is a link to a document that explains UPS topologies|http://www.apcmedia.com/salestools/SADE-5TNM3Y_R5_EN.pdf]

[here is a link to the Smart UPS RT family|http://www.apc.com/products/family/index.cfm?id=223]

[BillP]

Before automatically assuming it is in fact a transfer issue, I would recommend doing some research about the power draw of the devices you've mentioned.

Your testing was done using a 2960 which you said never dropped in the first place. If at all possible I'd recommend doing the test with a 3750. I looked up the 3750 and it seems the power supply can differ depending upon how it was purchased, so I can't safely guess how much power it might draw.

Looking at an image of the back of this unit it looks like there are seperate breakers for the outlet groups. Do you currently have the access switch and the sonicwall on the top two outlets and the 3750s on the bottom two?

[BillP]

All,


It looks as if the sensitivity setting on that particular UPS was turned off, which may be the culprit of this issue. Also, the third 3750 is plugged into another UPS, this one an older SU2200RM3U which did not drop the last time we had a power spike. I will be doing some testing to see if in fact the sensitivity setting was the reason that those devices were being rebooted, if not I will come back and post additional information in regards to this issue.



Thanks for all the suggestions and help.


Jeff

Message was edited by: sysadmin

Message was edited by: sysadmin

[BillP · Accepted solution]

All,
It looks as if the sensitivity setting on that particular UPS was turned off, which may be the culprit of this issue. Also, the third 3750 is plugged into another UPS, this one an older SU2200RM3U which did not drop the last time we had a power spike. I will be doing some testing to see if in fact the sensitivity setting was the reason that those devices were being rebooted, if not I will come back and post additional information in regards to this issue.

Thanks for all the suggestions and help.

Jeff