[Nut-upsuser] Help needed: LIEBERT PowerSure Personal XT - USB interface

[Arjen de Korte]

Ariel Dembling wrote:

> a) The battery.type attribute is indeed a string, as Charles Lepple 
> guessed in the commented-out code (PbAc = lead acid).

I think I added that line in r1043. We'll uncomment it... :-)

> b) The voltage (both nominal and real) is wrong, but not by a factor of 
> 10 as Charles mentioned on the driver:

That comment is incorrect, UPS.PowerSummary.(Config)Voltage usually
deals with the battery (it does so for most other HID drivers as well).
In that case, it is probably spot on. In the HID PDC specification
(which most of these devices seem to follow) the examples listed also
mention that this is the battery voltage, not the AC mains.

> I have 220 volts nominal here,
> and so probably did Alain Williams, who first reported this last year, 
> since he's from the UK according to his email address. I measured the 
> real voltage with a voltmeter and I got 213 volts here both for the UPS 
> input and for the UPS output.

Could you post the output of the driver with debugging enabled?

	usbhid-ups -DD -a <upsname>

See the man page for additional comments on running this driver in debug
mode.

> c) I can confirm that the ups.load value goes up a little when running 
> on battery (it went from 25 to 35). No big deal. On the other hand, the 
> battery.voltage attribute went down from 13 to 12 when on battery.

That's not uncommon.

> d) I don't know what to make of the unmapped attributes, but I don't see 
> a way to get the UPS to turn the load off during shutdown.

There probably is none. I have looked at the output from another device
and found no evidence that it supports shutting down. This is not the
best UPS for (unattended) server applications I'm afraid... :-(

> I find the 
> unmapped.ups.powersummary.capacitymode attribute intriguing, though.

That's a rather uninteresting value by default. You can use it to change
the way how the UPS reports the charge and capacity left, but we don't
use this in NUT.

> e) I don't know if this is of any importance to NUT, but in case anyone 
> cares let me point out that this UPS does not have a voltage stabilizer 
> built-in.

That's not important to NUT.

> f) The battery.runtime value is reasonably accurate (estimated in seconds).

There is probably nothing we could do if it wouldn't be... :-)

> g) I have replaced the SHUTDOWNCMD with the following custom script in 
> order to hibernate the system instead of shutting it down. It seems to 
> work for me on CentOS 5.1:
> 
> --------
> #!/bin/bash
> echo disk > /sys/power/state
> sleep 10
> service ups start
> --------

This will only work if you don't need (or can) shutdown the UPS. The
problem is that the kernel doesn't provide any hooks to insert a script
that can issue a shutdown command to the UPS. Since yours probably
doesn't support this, this is of no importance to you. On systems that
do, shutting down the UPS prevents nasty situations with the power
returning when the system has already started a shutdown (in which case,
the UPS cycles the power).

Best regards, Arjen