UPS switching to battery on computer boot

[Anonymous user]

I just purchased a APC Back-UPS XS 1500 (BX1500G) to provide backup power for a FreeNAS server.  I have the UPS connected to the FreeNAS server via the USB cable supplied with the UPS, the FreeNAS computer connected to one of the UPS's battery backup outlets (not the master outlet, however), and the UPS to AC power (supplying ~122 VAC).  When I boot the FreeNAS server with the USB cable connected between the UPS and FreeNAS server, the UPS switches to battery within 30 seconds, and cuts pawer to the battery backup outlets within 60 seconds.  If I boot the FreeNAS server without the USB cable connected between the UPS and the FreeNAS server, the UPS stays on AC power and operates as expected.  I can re-attach the USB cable after FreeNAS boots up, and the UPS stays on AC power, and supplys the expected UPS power event notifications to FreeNAS via the USB cable.  I've performed a self test on the UPS unit and it reports that it passed the test and the battery is 100% charged.

I've attached the technical datat output about the UPS from PowerChute in case this is helpful.

Any thoughts on why this is happening and how to fix it?

[BillP]

FreeNAS is a Linux based thing, correct? We don't have PowerChute Personal Edition for Linux so did you get the data from a Windows machine you attached at some point? Just trying to understand the setup is why I ask and how you got that data.

Do you run anything on the FreeNAS server that'd be running and potentially talking over the USB cable when the computer/server boots up? Any native OS support or something like apcupsd third party software?

Is there any log on the FreeNAS server that we can explore in the OS or any software, like apcupsd, that shows if a command is being issued to the UPS?

The other thing we'd need to you to validate is the USB cable - is it a purchased APC cable or was it made? An incorrect or maybe faulty cable could be issuing a command. The cable you should be using with this UPS would be part number/model AP9827 or 940-0127 - both mean the same cable. I'd expect it to be printed on the jacket of the cable.

[Anonymous user]

FreeNAS is a FreeBSD-based network attached storage operating system.  Yes, I got the data from a Windows 8.1 computer running PowerShute Personal Edition.

The FreeNAS server doesn't finish booting up before the power is cut, so I'm not sure if anything is communicating over USB when this is happening.  It's possible, but I'm not sure.  When is is running, is uses the FreeBSD usbhid-ups driver and NUT (network UPS tools) program to receive notifications from the UPS over USB.

No log, since the server soedn't get far enough along in its boot up process to log anything.

I'm using the USB cable that was shipped from APC with the UPS.  It has "940-0127EFCI" embossed on the USB end of the cable.

I tried usig a Windows computer to see if the same thing happens when it boots.  The UPS seems to operate normally when the Windows computer boots.  However, after an AC power loss and subsequent AC power restore, the UPS does not behave as it should.  Here's what happened:  The UPS is connected to the Windows 8.1 computer via the USB cable supplied with the UPS, the Windows 8.1 computer is connected to one of the UPS's battery backup outlets (not the master outlet, however), and the UPS is connected to AC power (supplying ~122 VAC).  When I boot the Windows computer the UPS operates as expected.  When the UPS is disconnected from AC power, it switches to battery and the Windows PC displays a notice that the UPS is no longer receiving AC service.  After 1 minute, the Windows PC shuts down.  When the UPS is reconnected to AC, it remains on battery rather than switching to AC.  When I power the Windows computer on, it boots up.  After after about 1 minute the UPS cuts power to the battery backup outlets and the PC powers off without notice.  The UPS display goes black, then a few moments later lights up again and appears to be on AC power.  When I boot the Windows PC, everything works as expected.  This doesn't seem to be correct behavior for the UPS. Shouldn't it return to AC power when it is plugged back in to AC, and why would it ever just cut power to the battery backup outlets when the battery is fully charged?

Thanks,

Randy

[BillP]

Randy,

For your test you should not unplug the UPS from the wall outlet. By doing that ground is lost / broken. When the UPS is plugged back in it may not react as expected. To test you should plug the UPS into a grounded power strip and using the switch on the strip cut power to the UPS.

When PowerChute reacts to a power issue and commands the OS down it sends a shutdown command to the UPS at the same time. The UPS will count down approximately 2 minutes from the time the command is received and if AC is not restored will cut output to the outlets and enter sleep mode. If AC is restored after the command has been received but before the UPS has entered sleep mode it will continue the 2 minute count down then cut power for 3-4 seconds and then restore power. It does that because most system are configured to auto boot when power is restored so the UPS cuts power to the outlets momentarily to allow the system to register the lose and restoration of power.

[Anonymous user]

I'll try that.

Is there any way to control the 2 minute count down/power cut behavior?  I don't want that to ever happen in my case.  It seems presumptive to assume that that is always desired by your customers, so I would expect it to be configurable.

Thanks,

Randy

[Anonymous user]

Bill,

I just tried what you suggested about using a power strip instead of unplugging the UPS and got the same bahvior as before.  After switching the power strip back on, the UPS remained on battery power until it suddenly power itself off and then back on again without notification or warning.  In the meantime, when the power strip was switched back on, thje UPS restored power to my PC and it booted (because it's configured to start up after a power loss when power is restored).  Then, my PC was abruptely powered off by the UPS without any warning or messages from PowerChute.  This seems crazy to me.  Why would the UPS purposely do the one thing it is supposed to prevent from happening?  Is this really by design?

-Randy

[BillP]

Randy,

On 4/28/2015 4:09 PM, Randy said:

 After switching the power strip back on, the UPS remained on battery power until it suddenly power itself off and then back on again without notification or warning.

That is expected behavior. Once PowerChute sends the shut down command to the UPS it will wait 2 minite and then power off. This happens whether or not AC has been restored during those 2 minutes.

On 4/28/2015 4:09 PM, Randy said:

In the meantime, when the power strip was switched back on, thje UPS restored power to my PC and it booted (because it's configured to start up after a power loss when power is restored).

As long as the UPS remains on battery for the 2 minutes power will not be cut to the PC so it should not restart. If the PC is restarting then there is an issue with the PCs power supply.

Test the system by turning the power strip off. Wait for the PC to power down and then wait for the UPS to go into sleep mode (LED flashing). Sleep mode should start approximately 2 minutes after the OS start the shut down process. Once the UPS is in sleep mode wait 8 seconds then turn the power strip back on. The UPS should turn back on, restore power to the outlets and the PC should start.

[Anonymous user · Accepted solution]

Hi Bill,

Thanks.  Since I can't control how long a real power loss will take, waiting for the UPS to finish its cycle seems to be a test scenaro that isn't very useful.  So, I did more testing with my PC and the UPS and I think I know what's happening.  I thought the UPS was cutting and restoring power due to the PC automatically starting up.  However, it appears that it was the UPS sending information to the PC over the USB cable that was causing the PC to automatically start up during the 2 minute cycle you described.  To test this, I turned off the BIOS options to wake the computer via keyboard and/or mouse and the behavior stopped happening.  When I reenable those options, the behavor happens.  Aparently, the PC is interpreting the UPSs USB communication the same way it interprets keyboard and mouse signals.

So, I have a few questions:

1. Is there a way to prevent the UPS from communicating over the USB cable during the 2 minute power cycle?
2. Is there a way to stop PowerChute from hibernating my PC, and have it just shut it down (see attached image of the wanring PowerChute displays before hibernating my PC)?  I don't think USB communicaiton will wake the PC if it isn't hibernating or sleeping.

Thanks,

Randy