Re: APC UPS CS500 won't switch on again
On Sat, 26 Feb 2005 13:00:17 +0100, Alexei Chetroi
<alexei.chetroi@debian.uniflux-line.net> wrote:
> On Sat, Feb 26, 2005 at 11:28:49AM +0100, Matthijs wrote:
> > Date: Sat, 26 Feb 2005 11:28:49 +0100
> > From: Matthijs <vanaalten@hotmail.com>
> > To: debian-user@lists.debian.org
> > Subject: APC UPS CS500 won't switch on again
> >
> > I recently bought an APC CS500 UPS to protect my Debain mailserver
> > from powerfailures. I installed apcupsd, configured it and tested it.
> [snip]
> > I don't know if it's relevant, but when I execute apcaccess to get the
> > statusinfo, the following is reported:
> > LINEV : 000.0 Volts
> > LOADPCT : 21.0 Percent Load Capacity
> > BCHARGE : 095.0 Percent
> > TIMELEFT : 28.2 Minutes
> > MBATTCHG : 20 Percent
> > MINTIMEL : 10 Minutes
> > MAXTIME : 60 Seconds
> > OUTPUTV : 230.0 Volts
> > DWAKE : 000 Seconds
> > DSHUTD : 000 Seconds
> > LOTRANS : 000.0 Volts
> > HITRANS : 256.0 Volts
> > (only part of the report copied&pasted)
> >
> > I find the report for LINEV and LOTRANS odd. I've adjusted the
> > sensitivity to 180/266V, HITRANS is then changed to 266 but LOTRANS is
> > always 0V.
> Is this report when ups has power applied to it?
> I wonder what capacity your ups has? By it's name I can guess it's
> about 500VA, but in that case TIMELEFT value is too long IMHO. May be
> you should run ups calibration test.
It's indeed a 500VA capacity UPS, but the TIMELEFT is probably quite
accurate :-)
The server is a VIA EPIA M10000 mainboard/processor (consumes about
17watt when idle, which it is most of the time), has a 5400rpm
harddisk connected to it (5 watt when idle) and an unused
CD-ROM/floppy connected to it. All in all it's probably less then 25
watt, so half an hour is not that long. I expected an even longer
battery time...
> [snip]
> > Possible explanations:
> > a) I misunderstood everything, the UPS is not supposed to switch
> > automatically back on again; (I'm definitely a newbee on UPS-es so I
> > keep my options open)
> > b) UPS is defective;
> > c) USB cable is defective;
> > d) configuration of apcupsd not completely correct;
> >
> > Sorry for the lengthy message, but better complete than short. Anyone
> > who can comment on this?
>
> Cannot tell you about apcupsd, because we use nut for apc ups
> monitoring. But these things comes to my mind:
> 1. iirc ups can be configured not to switch on imediatly when power
> applied, but only after battery are charged to some level.
Sounds reasonable, but when it's going down because of 60-second
TIMEOUT, then the server has only been running on battery power for 60
seconds before shutdown, so the UPS will have *plenty* of juice left.
Also, I waited for 15 minutes - that should be good enough to
compensate the 60-second battery run time.
> 2. ups was told to shutdown, not to switch of its load.
Could be the cause. However, in the manual of apcupsd, I haven't found
a configuration setting for this. Anyone?
> I'd suggest you to run ups calibration, so ups may estimate more
> precisely battery run time under current load. Charge batteries.
I'll skip this, since the runtime is accurate enough.
> Disconnect usb cable and leave ups on battery until it switches off by
> himself. Connect it to the power again and chech if it will power on
> load. So you can tell wether ups doesn't bring on load or it told to do
> so.
Sounds like a good test - I'll see if I can perform this. Disadvantage
is that the server will not shutdown in a proper manner, but for one
time the chance of damage is acceptable.
Thanks for the advice!
--
Matthijs
vanaalten@hotmail.com
Reply to: