[Nut-upsuser] Differences between USB and serial on an MGE Pulsar Evolution 3000

[Arnaud Quette]

Hi Neil,

2011/10/6 Neil A. Hillard <neil at dana.org.uk>:
> Hi,
>
> I've been using NUT for many years to look after the UPS on my home servers.
>  Originally I was using a Compaq R3000 but that was replaced with an MGE
> Pulsar Evolution 3000 about a couple of years ago because the Compaq one was
> just too noisy!
>
> I'm just setting up a replacement server and would like to connect it using
> a serial cable as I've found that the USB connection occasionally drops out
> on the current server.
>
> Unfortunately there appears to be a lot of parameters missing from the
> serial driver (using mge-shut) when compared to the USB driver (usbhid-ups)
> and more importantly, output.voltage.nominal is reported at 72 Volts instead
> of the correct 230 Volts.
>
> I've looked into the source code (I know a little C) but unfortunately I'm
> unable to follow it completely to look at correcting the problem and perhaps
> adding in the missing parameters (if they're available over the serial
> connection.
>
> I've included below a debug output from mge-shut that shows the current
> output voltage being fetched correctly but the nominal voltage coming back
> incorrectly.  I've also included the upsc output for both serial and USB
> connections.
>
> If anyone could give me any advice as to why the nominal voltage is decoded
> incorrectly (or even describe how it's decoded so I can investigate it
> myself) then that would be appreciated.  Advice on how to add the extra
> parameters in would also be appreciated.
>
> Many thanks in advance,
> (...)

I won't go in depth, but this is a known point:
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=411864

This is also why I've developed the 'newmge-shut' driver, which shares
a lot of code with the usbhid-ups one.
FYI, SHUT stands for Serial HID UPS Transfer, which is basically HID
(as in USB, so the same data) but over a serial link.

the newmge-shut take over has been postponed for some time, but it's
only due to a lack of time and formal testing, more than issues with
this driver. And mge-shut is not really maintained anymore...

could you please test again with newmge-shut, and report back?

cheers,
Arnaud
-- 
Linux / Unix Expert R&D - Eaton - http://powerquality.eaton.com
Network UPS Tools (NUT) Project Leader - http://www.networkupstools.org/
Debian Developer - http://www.debian.org
Free Software Developer - http://arnaud.quette.free.fr/