Swapped AP9631 into an RMA'd Smart-UPS SRT 6000VA - Quick amber and unresponsive

[Anonymous user]

Hello, We have several of these Smart-UPS with the additional AP9631 card in them.

So a few days ago one of the units went bad and was boot looping, etc. We had it RMA'd and swapped it last night.

We moved the AP9631 card over but sadly it is unresponsive. If I plug in the console cable I get no feed back. If I hit or hold the reset button then the console/putty will give a garbled filled in box, but thats it. I tried holding reset for 60+ seconds, nothing.

Basically the lower right activity LED just blinks a fast steady amber. If I plug in ethernet, then the left blinks a few times, nothing special.

I noticed the UPS itself has a built in NIC. That was blinking alternate green and amber. I plugged the console into that and it came up right away.

Do I need to do something like disable the internal for the addon card to work? Am I losing any functionality with the internal vs AP9631?

[BillP · Accepted solution]

Can you look at this video and let me know if the flashing looks like this? https://youtu.be/VL-oQ9JKvP8 

If so, the AP9631 is bootmonitor mode and should be recoverable. It just may need a reflash of the firmware as it is in a recovery mode due to corrupt OS that can't boot (or some other similar issue). In this mode, you want to connect at 57600 baud rate instead and press enter a few times. It should say:

Rhodes Bootmonitor

BM>

I wouldn't be surprised when you do that, if you see AOS CRC failure on the screen or something like that.

From there, you'd want to basically follow these instructions: http://www.apc.com/us/en/faqs/FA293874

which you can skip up until 6 since you're in bootmonitor. Step 6 is where it explains how to begin loading the firmware. When you're in bootmonitor, you can do sysinfo and see what it says too and show me the output of that screen and I can tell you more specifically what is required.

All in all, we should be able to get you back up and running with this card I think.

[BillP]

Sorry, forgot to add that the internal NIC is just fine to use. Adding an AP9631 is mostly redundant and you can maybe use this NMC in a different UPS. Major differences I think would be the number of UIO ports for accessories like dry contact or temp sensor (1 port on internal NMC I believe, 2 on AP9631). And I forget if the SRT6K would have a USB port on the back like the AP9631 does which can do firmware upgrade. But that may be a difference too. Other than that, no differences - same firmware, functionality, etc.

[Anonymous user]

Yes, that appears to be the case, it's in bootmonitor mode.

I did a sysinfo,

Model        : AP9631
MAC Address  : xxxxx
Serial Number: xxxxx
CPU Speed    : 200
Link Port    : 3
Warmstart    : YES  NMI Count: 0

*********** FLASH MODULE INFORMATION *****************

Boot Monitor :
  Name       : apc_hw05_bootmon_108.bin
  Version    : 108
  Cookie     : 963200
  Build Date : Apr  8 2014
  Header Crc : A9D5(PASS)
  Module Crc : 9897(PASS)

AOS          :
  Name       : apc_hw05_aos_646.bin
  Version    : 646
  Cookie     : 963201
  Build Date : Oct  6 2016
  Header Crc : 3117(PASS)
  Module Crc : 39EA(FAIL)

APP          :
  Name       : apc_hw05_sumx_646.bin
  Version    : 646
  Cookie     : 963201
  Build Date : Nov  2 2016
  Header Crc : 2603(PASS)
  Module Crc : 70CB(PASS)

MTO          : Not Present

LPK          : Not Present

So by the looks of it the AOS went bad? Updated with 6.56 version and unit came up, configuration still seemed intact.

Thank you.

[BillP]

OK, great! Glad you were able to recover.

Yes, something happened here to cause the AOS to become corrupt since you can see the module CRC as FAIL in the sysinfo. I have seen this before when the UPS the NMC is inside of loses power abruptly, though I have never seen it personally or when removing the NMC from power in trying to simulate it.

[Anonymous user]

Well that is sort of what happened. The main unit died, was throwing some errors, internal hardware failure, etc. It would stay on about 35-40 seconds, make a physical click sound, reboot, and repeat.

So it was put into maintenance mode remotely to stop the alerts, but its possible between then and when we finally removed it, it did cause damage. It was like that approx 14 days before we had the rma unit and had the scheduling to replace it.

[BillP]

Ahh. OK, I don't exactly know why but I wouldn't be surprised of that scenario contributed to the corrupted AOS.