Sporadic power issues with SMT1500I (Smart UPS 1500)

[Anonymous user]

I have a SMT1500I installed at a client since June 2012. It protects one Dell T310 server, a Dell Optiplex, a network switch, WAP, modem, firewall, several USB drives, and a PBX unit. Load is on average at around 15%. PowerChute is installed on the Dell T310 (Windows SBS 2011) and is connected to the UPS via USB - a script is set to run to shut down the server on the Optiplex in the event a shutdown is initialised by PowerChute (after 15 mins on battery power).

For the most part this works well and during power cuts the UPS kicks in and takes over until utility power returns or gracefully shuts down the 2 Windows servers when power doesn't return. However, there have been three or four bizarre issues with this unit over the past 10 months and each time the symptoms are that all connected devices power off at the same time (not gracefully) and remain off. The utility power appears to remain on.

I have not been on site during any of these power issues and am relying on the client to tell me what state the UPS is in when this occurs. The most recent issue occurred today - all devices lost power around 21h Monday night and remained off. The client stated that the UPS display was reading "replace battery" when he arrived and an audible alarm could be heard. He then powered off the UPS and left it off for an hour before contacting me. I advised him then to power on the UPS again and all connected devices then powered back up instantly.

I checked the PowerChute logs and of course there is nothing of interest as the log is lost once the PowerChute host loses power. The only thing that is linking these issues is that each time the last entry in the PowerChute log shows a very high input voltage value - last night it was 252v, which is close to the High Transfer Threshold set in PowerChute at 253v. I am in the UK and utility power is set to 230-240vAC - in PowerChute nominal power output is set to 230v.  When last night's issue occurred, the UPS was in Green Mode.

What should happen if the High Transfer Threshold is attained? I am beginning to think that when this threshold is reached, the UPS powers off completely for some reason. The most annoying part of this is that the attached servers are not being gracefully shutdown - so far I have been lucky and they have all come back up....

Can anyone help shed any light as to what I should do? I have contacted APC support in the past, but they were not able to diagnose anything from the available logs.

[BillP]

How old is that UPS/battery?

If the voltage goes outside of the transfer ranges (where within it, it would normally be using Automatic Voltage Regulation or AVR if it could or run online), the UPS is designed to go to battery power. If there is a problem with the battery (its older or can't hold the load up, disconnected, or the UPS thinks its disconnected), then it could be feasible that the voltage got so high that the UPS tried to go to battery power and then the battery did not support the load and the UPS turned off.

Granted, the UPS self test that by default happens every two weeks and when you turn the UPS on, should catch a bad (connected) battery like this before these things start to happen but depending on battery age, characteristics, and environmental factors, you could encounter this type of situation with a bad or faulty battery.

[Anonymous user]

Hi Angela,

Thanks for the reply. The UPS was purchased in June 2012, so it and the battery are less than a year old. The self test runs weekly, it says it passed on 06/05/2013.

Should I tweak the High Transfer Threshold in Powerchute to set it a bit higher? Over the last 24 hours, voltage has been as low as 235 vAC (during the day), and as high as 249vAC (during the night). From the log (since end of Jan 2013), the lowest ever is 230vAC and highest is 253 vAC.

It really is quite odd behaviour.

Thanks.

[BillP]

Well, the UPS should be going to battery.

Do you have a full data log here? Can the end user or yourself verify what the UPS LCD showed for last 10 transfer reasons and last 10 faults (the menu needs to be in advanced mode for this I believe). Maybe there is a very quick overload causing an issue too here. Or, if communication was lost (which we don't see reported anywhere) maybe it went back and forth to battery and shut down.

I don't suppose you have a spare AP9630/31 to install in this to gather logs if it happens again instead of relying on PCBE which hasnt been helping. I figured no and you shouldn't need to do this but hey I will throw it out there!

[Anonymous user]

I wanted to post back about this issue - I have had a few more identical issues since the last post - this week there have been three. After the latest, the UPS was showing an error "FAULT STATE: AVR Relay Weld, call Tech." - the unit was trimming when it dropped the load and errored.

I have been in contact with APC and happy to say that they are issuing the client a RMA replacement. I have also advised the client to get his circuit and breakers checked out.

[Anonymous user]

Thanks again for your help Angela.

When you say "full data log", what do you mean? Do you mean the PowerChute log?

I will try and get the client to scroll through the LCD Advanced menu logs to see if they show anything (might need an hour on the phone as this client is particularly IT phobic!)

Sadly, no spare network card for the UPS (small business with limited funds).

Will report back.

[BillP · Accepted solution]

Hi again. Yes, I meant the powerchute data log with voltages, etc as you already provided the limited info you saw at that time for the data log. You might also just consider getting or providing the full log too so maybe we can see the history of power issues there a little bit.