Nut upsdev - Sep 2007 - Tripplite Smart 2200 and Battery Voltage

For my Smart 2200, a charged battery string is anywhere between 25.6 and
26.8 volts. I'm seeing nut report 13.0 volts, which seems suspicious. Even
if the UPS were reporting half of the actual battery voltage to keep
protocol compatibility with models having a 12V battery system, the value
of exactly 13.0 seems unlikely. 

According to the page at http://kursknet.ru/~boa/ups/rtinfo_command.html,
the :B command returns "relative battery voltage", not any absolute
voltage.

On 9/4/07, Brian Ristuccia <brian at ristuccia.com>
wrote:
> For my Smart 2200, a charged battery string is anywhere between 25.6 and
> 26.8 volts. I'm seeing nut report 13.0 volts, which seems suspicious.
Even
> if the UPS were reporting half of the actual battery voltage to keep
> protocol compatibility with models having a 12V battery system, the value
> of exactly 13.0 seems unlikely.
Do you happen to know the protocol version for your Smart 2200? It is
displayed by some versions of the Windows monitoring software. I
haven't had a chance to backport some of the changes from
tripplite_usb to the serial-only tripplite driver because of a lack of
systems to test with.
> According to the page at http://kursknet.ru/~boa/ups/rtinfo_command.html,
> the :B command returns "relative battery voltage", not any
absolute voltage.
> From looking at the source, the calculations seem to be based on an actual
> reading in decivolts or something similar. It also calculates relative
> capacity based on a maximum battery voltage of 13.0 (in nut 2.0.1) and 13.4
> (in nut 2.2.0). The upscmd.cpp in TL_los_cs also uses a similar algorithm.
That could probably be merged with the code here:

http://boxster.ghz.cc/projects/nut/browser/trunk/drivers/tripplite_usb.c#L1245

http://boxster.ghz.cc/projects/nut/browser/trunk/drivers/tripplite_usb.c#L491

... which uses part of the V response to figure out the number of batteries.
> Has anyone looked at this closer already, or should I do a rundown with an
> external voltmeter connected and figure out what the returned values
> actually mean?
Never hurts to double-check with a voltmeter. You may not need to run
it down all the way, either.

-- 
- Charles Lepple