Wondering if anyone else has seen this. Got a curious but not alarming "Battery Disconnected" warning from PowerChute Business Edition. Went on site to discover the unit almost too hot to touch, batteries bulging. PowerChute reported only 41 C internal temp, and by default doesn't alarm unless temp reaches 70 C. This on a DLA1500, Dell equivalent of an SUA1500. Additional detail and logs here:
Hi mcbsys, the internal temp reading typical is of the unit, not the batteries. The batteries appeared to have gotten warm at some point which can cause them to expand. When the happens the ups will typically say that batts are disconnected. I would like to get some more info on your powerchute setup if possible as well so can better answer some of those questions. are you running powerchute network shutdown or powerchute business edition? and which version of each on what OS? Hope this helps.
PowerChute Business Edition 18.104.22.1686 on Windows Server 2008 R2 SP1.
There's "warm" and then there is "Holy cow is this thing about to catch on fire? I'm lucky it's not outgassing." As you can see in the data log (in the blog article), temp rose from 30C at 7:30am to 41C at 4pm. Temp was exactly 40C when the batteries were disconnected (2:39pm). Temp continued to rise after that until I got on site and opened the case. Batteries were still warm two hours after removing them from the UPS. The scary thing is that there was no event indicating this as an emergency overheating situation, as in "get those batteries out of the building NOW." I hate to think what would have happened if I hadn't decided to mosey down there and check out the cause for the innocuous "battery disconnected" message.
23 Apr 2009: Purchased (so less than 2.5 years old)
05 Sep 2011: Self-test passed
08 Sep 2011: Performed okay during long power outage - provided power for 40 minutes
21 Sep 2011: Overheated. "Battery Disconnected" event.
There was no self-test performed between Sep 8 (your outage) and Sep 21 (when the batts were overheating) so that last outage probably did them in. Actually, with the default 14 day cycle for the self-test, it should've performed another test on Sep 18 - 2 weeks after the last test passed on Sep 5. Did you change those settings by any chance? Environmental conditions also play a significant role in battery health and longevity. If you've got time for spare reading, [this kbase|http://nam-en.apc.com/app/answers/detail/a_id/8975] has a lot of information regarding conditions of thermal runaway.
If there had been another self-test in there after the outage, it should've recognized the issue at that point. You'd see a "replace battery" error after a failed self-test, since that's when the batteries are actually analyzed for health conditions. Instead it let you know there was a problem through the "batteries disconnected" error message.
No, didn't change the self-test cycle. Maybe it resets after going on battery?
Would like to read about "thermal runaway" but the link tells me, "This answer is no longer available."
Any thoughts on the temp threshold setting? The default 70 C seems awfully high when 27-32 C is normal operating temp. In fact, if the software were a little smarter, it could alert on unexpected temperatures, e.g. more than 1 C increase per hour, or if temp at 7am today is 2 C above temp at 7am yesterday. For now I've set the absolute threshold to 35 C.
Thanks, that is an interesting white paper. So I guess a SmartUPS uses a Valve regulated (VRLA) battery? Re. thermal runaway, they say (p. 7):
"The primary hazard of thermal runaway with VRLA and MBC systems is the emission of hydrogen (a flammable gas) and hydrogensulfide gas (an irritant). It is possible for these batteries to enter a state in which heat is generated faster than it can be dissipated during charging. The increasing battery temperature results in more current being drawn from the charger, which in turn further raises the battery temperature. The cycle continues until build-up of internal battery pressure causes the vents to open."
They go on to quote [this article|http://www.concordebattery.com/otherpdf/TB2.pdf] by Concorde Battery about thermal runaway in airplane batteries. "The battery will reach a moderate internal temperature (approximately 260 °F / 126 °C) at which point the water in the electrolyte vaporizes and the battery vents steam. As the separator is glass, it is unaffected by this temperature. The loss of water caused by the venting reduces the conductivity between the battery plates and the battery ceases to accept further charge. The battery slowly cools.”
They don't quote the next paragraph about the Induced Destructive Overcharge Test (IEC standard 952-1:1988) and how flame and explosion must be contained inside the battery. I wonder if APC batteries meet that standard.
It is interesting, and a little reassuring, that Concorde considers 260 F / 126 C to be a "moderate" internal temperature. Forty-eight degrees F above water's boil point seems pretty hot to me, but if a VRLA battery can withstand that and only vent steam before failing and cooling down, that's good.
The APC white paper also doesn't explain _how_ a battery can get into a heat-generating state. The Concorde test forces thermal runaway by charging at 3 volts per cell "and continuing to charge until the battery fails." So maybe faulty charge regulation caused the failure?