[Nut-upsuser] Battery test feature of APC Back-UPS RS series

[Jean Delvare]

On Sat, 10 Oct 2009 15:57:24 +0200, Arjen de Korte wrote:
> Citeren Jean Delvare <khali at linux-fr.org>:
> 
> > Now the good news: the battery test command appears to work!
> 
> That's good to hear.
> 
> > The test.battery.start.quick command behaves exactly as when using
> > APC's own software. The battery is being tested for a few seconds, then
> > the UPS returns online, and ups.test.result is set to "Done and
> > passed". Perfect.
> 
> Indeed.
> 
> > I am a little skeptical about the test.battery.start.deep command. When
> > I run it, the UPS switches to battery. After one second, nut tells me
> > I'm on battery, and one second later that I am back online. However,
> > the UPS is _still_ on battery, I can see it on the front panel (and I
> > can hear it, too.)
> 
> While this may seem confusing, this is actually what it is supposed to  
> do. After all, the OL/OB flag only tells you something about the line  
> status and when you run this command, most likely that is not  
> changing. So during the test, it is actually correct that the OL flag  
> is set (so 'online' would be correct).
> 
> > Then I waited some more and nothing changed. After a
> > minute I had to issue a test.battery.stop command to return online.
> 
> You probably need to be more patient. It may take quite a while for  
> before the battery reaches the battery low charge.
> 
> > It worked OK, but now ups.test.result is set to "Aborted".
> 
> Very good, so the test.battery.stop command also works. This seems to  
> be a resounding success... :-)
> 
> > So maybe I wasn't supposed to stop it? How long is it supposed to last?
> > I couldn't find documentation on how the deep test is supposed to behave.
> 
> Usually until the battery charge reaches the 'low' level, at which  
> time the UPS should stop the test and resume to normal. Depending on  
> the amount of time since the test was started and the load present, it  
> may also inform you about the health of the battery.

Ah, OK. Then indeed I have not been patient enough, as my UPS has 33
minutes of runtime when the battery is full (which it was) and
battery.charge.low is 10/100 so I guess I should have waited for about
30 minutes before the low limit would have been hit.

> > After that I couldn't issue a test.battery.start.deep command again,
> > nothing happened when I tried. test.battery.start.quick still works
> > though.
>
> I just recalled that some UPSes will refuse to run a deep discharge  
> test when the battery isn't fully charged and/or need to be charging  
> continuously for an 'X' amount of hours (with 'X' typically being 12  
> to 24). So chances are this will work again if you wait a day or so.

Yes, that could be what happened, as the battery was still recharging
after the first (aborted) deep test when I started the second one.
Makes sense.

> PS  Also, be advised that running deep discharge tests wear out your  
> battery, so you should only run these once every three to six months  
> (at most).

Thanks for the hint. I didn't intend to use it again any time soon
anyway. I presume it is essentially useful when you suspect that the
battery needs to be replaced. Which I hope I won't have to do before 4
years or so.

Thanks a lot again for your excellent support :)

-- 
Jean Delvare