Smart-UPS 3000 discharge rate not matching runtime remaining

[BillP]

Hi,

We have a Smart-UPS 3000 with typically a 15% load that started running on battery today.  It was on battery power for approximately 5 seconds, and in this time dropped to 90% capacity.

The UPS still shows a runtime remaining of 36 mins, but if it discharged at this rate then it would last less that a minute.

Is 5 seconds long enough to start calibrating the battery?  What was displaying the wrong value, the 90% capacity after 5 seconds, or the 36 min runtime?

Log:

07/14/2014    06:55:14    UPS: No longer on battery power.

07/14/2014    06:55:10    UPS: On battery power in response to distorted input.

07/14/2014    06:55:10    UPS: No longer compensating for a high input voltage.

07/14/2014    06:54:19    UPS: Compensating for a high input voltage.

07/14/2014    06:54:03    UPS: No longer on battery power.

07/14/2014    06:54:02    UPS: On battery power in response to distorted input.

[BillP]

Hi,

The runtime is dependent on the age of the battery/s and the amount of times the battery has discharged / been in use. The UPS will calibrate when it switches to on battery. What is the exact model Smart-UPS 3000? How old is the battery/s and does the UPS switch to battery often?  With that information we can review the online runtime chat to see if the UPS is within specs.

[Anonymous user]

What happens to the battery charge % readings in the first 5 to 30 or so seconds of when a UPS first goes to battery is not usually indicative of what will happen long term. Depending on load and battery age/health there is usually an initial drop in the first few seconds (that can sometimes be quite large),then sometimes what follows will be a short rebound (batt charge % & run time reported will increase very slightly/briefly), then it will drop steadily/slowly at a fairly even rate until low battery conditions are reached. All of this is governed by what has been termed the "coup de fouet" effect.

That there is sometimes a rebound is somewhat surprising to most people, and it comes in handy if you have a discharged (but not totally dead) car battery (eg left your lights on) if the battery can just crank the engine but not quite start it, turn everything off and wait about 5 to 10 mins then try again it may crank stronger and get the engine started.

[BillP]

Thanks Bill, this battery was installed on January this year and looking at the logs this is the only time it's been running on battery other than it's initial calibration:

01/14/2014	06:26:16	UPS: Runtime calibration passed via internal operation.
01/14/2014	06:01:02	UPS: Started a runtime calibration.

How do I find the exact model?  SNMP just says:

.1.3.6.1.4.1.318.1.1.1.1.1.1.0 = STRING: "Smart-UPS 3000"

[Anonymous user]

Back panel of the UPS should have a Model / Serial No Sticker.

[roger_apc · Accepted solution]

I deal with a large number of UPS units in locations across the US.  Over the past few months I have been dealing with some VERY old UPS units and also some VERY old batteries.  Some of the UPS units were installed in the mid to late 90"s and a few more in the early 2000's.  By far, the majority of our units are from 2006-2007.  Some of the batteries in those units have never been replaced.  We have started to run across units where the run time is listed at 30 min., but they fold after just a few seconds.  The older units don't seem to "self calibrate" very well.

Over the past few months I've tested and calibrated close to 50 UPS units, and here is what I've found.  Even with the short switch to battery power during a self-test, it's not uncommon for the battery % to drop 10-20%.  In this range, there doesn't seem to be any major change in runtime after doing a full runtime calibration.  In fact, some have shown an increase in runtime after calibration. But when that drop is in the 20-30%, the runtime will decrease dramatically.

I just ran a runtime calibration on a unit that was dropping to 69% on the self-test, and it was showing 66 min of runtime.  When I ran the calibration, it dropped to 38% within the first minute, and then ran for about 10 minutes before the calibration completed.  I think the calibration cuts off at around 20-25%.  So, as you can see, the discharge curve is no where close to linier.  NEXT MORNING:  The listed runtime has dropped to 26 minutes.

Just my experience.  YMMV