SU2200NET Goes Crazy - Serious EEPROM Issues

[BillP]

I have a SU2200NET in my lab that is used for “off-floor” testing. Communication with this unit is done via the serial port. This was a fully functional unit!!!

I forgot to end a terminal session when I went home the other night. When I can in to work today, I found a book resting on my keyboard. I also found what I thought was a battery dead UPS.

The batteries were charged and re-installed into the unit. I turned on the UPS and it went directly to battery mode. I tried several tricks to get this UPS to properly turn on but it keeps returning to battery mode.

I started an rs232 terminal session and I was shocked by the results (no network card, no AC loading):

Old model String:   “Smart-UPS 2200”.

New model String:   “Smart-UPS 450”.

Requests for the “serial number” and “manufacturing date” returned spaces; could be a null value. The correct serial number was “YS0433222060” and the correct manufacturing date was “08/12/04”.

The reported battery voltage should have been “55.65” but now it returns “28.41”.

The reported firmware should have been “80.14D” but now returns "40.14D".

Resting a book on the keyboard bombarded the serial port with erroneous data and this probably drove the UPS crazy. BUT, an end-user should NOT have the ability to change the “model string”, “serial number” and “manufacturing date”; on purpose or by accident.

By some means this UPS no longer thinks that it is a SU2200NET but an SU450. What magic do I need to perform on our SU2200NET to correct this problem??? I do not need this happening to 8 other units.

Thank you in advance for any help offered.

[BillP]

Hi,

You can try to reset the EPROM

1) Turn off and unplug UPS from wall
2) Press and hold the OFF button again
3) You may see a quick flash of the LED's and you should hear an audible "click" from the UPS.
4) Plug UPS back into the wall and turn UPS back on

[BillP]

I already tried what you suggested, Bill, but it made no difference.

A "cold start" and "brain dead" were the first things that I tried. When these failed, I started a terminal session and noticed that the model number changed from a "Smart-UPS 2200" to a "Smart-UPS 450"...

Thanks again.

[voidstar_apc]

This page has some good stuff:

http://lists.alioth.debian.org/pipermail/nut-upsuser/2007-January/002155.html

[BillP]

Thank you for the information provided. I am currently reviewing this information.

I will reply back once I see where things are going.

[BillP]

The link you provided was very much appreciated! It gave me a better understanding of what happen to SU2000NET.

It appears that we somehow managed to trigger a “big reset”. The exact sequence of events that caused this disaster is still very puzzling but we address this issue internally.

One thing is clear; always remove the rs-232 cable when it is not required. None of our units are connected to the Ethernet and all network cards have been removed for security.

I was able to configure the “model String”, “manufacturing date” and “serial number” to their previous values. The battery voltage now reads 54.55.

The ability to calibrate the line voltmeter is a nice option. I noticed that one of our other 2200s is off by 3.1 volts…

[BillP]

The UPS is still not functioning properly.

When I turn on the UPS, it performs a self-test, and then immediately switches to battery to power the load. It will not power the load from the service drop.

The UPS is receiving ac power because it will charge the two battery packs and the fan spins during this event. Also, the UPS returns a line voltage of 122.3, which is very close to the service drop.

Any ideas why this UPS will not power the load from the service drop???

[voidstar_apc]

I'm a bit confused -- I would've expected it to either power the load from the batteries or charge the batteries, but not both.

If you insert a network management card, it might provide more details in the logs. You can also check the transfer points, although 122 seems pretty close to the 120V nominal. Maybe the UPS thinks its a 230V unit?

[BillP]

Your not the only one who is confused, voidstar...

The UPS did not want to switch the load to the service drop. I am wondering if there is a software command that will force the UPS into a battery only mode.?.?.? The software "self test" command might have cause something like this; but I would expect a "cold start" or "brain dead" events to clear any commands in progress...........

The UPS appears to be functioning again but it is going to take time for me to have confidence with its performance. I have no idea why it is now working.

I might be able to generate a procedure on the steps that I took to address this problem but the ending will be a cliff-hanger.

THANKS for all of the help! This was a very educational experience.

A word of advise to all end-users; create a log file that contains all of the software settings for each UPS at your facility/home. There may come a time when this data is needed.

[BillP]

Voidstar, you might have the answer that I am looking for.

The more I fiddle with our SU2200, the more I am convinced that a "big reset" was issued to the unit. Things appear to be back to normal but with one exception, the AP9617 network card is not recognized by the SU2200. There is no LED activity on the back of the card, the LEDS are always off. There is no RS232 communication with AP9617. This AP9617 network card correctly functions in other test units (1400, 2200, 3000).

Is there an undocumented way to enable/disable a network card in the EEPROM?

Best Regards for the Holidays.

[voidstar_apc]

There's no software mechanism to disable the NMC. If it works in other UPSes, then there's a physical problem with the SmartSlot. The LEDs on the network jack should blink if (1) there is a network cable plugged in (2) the card is functional and (3) the card is receiving power. So it seems like the card is not receiving power. Along with EEPROM issues, that suggests the UPS may have absorbed a surge.

[BillP · Accepted solution]

Ya, it is time to put this puppy on the table and open it up. I do not think the network card is getting power.

I just took two of our 2200s out of service because of bad sine wave characteristics; fried resistors. They are also on the list to go in to the ER.