[Nut-upsuser] Fwd: Cyberpower usbhid-ups wrong battery voltage

Arjen de Korte nut+users at de-korte.org
Fri Apr 11 11:39:44 UTC 2008

> IMHO everything that measures something should be subjected
> to periodic calibration being it a 5-digit precision laboratory voltmeter
> or cheap piece of junk like ups, both can benefit from being
> calibrated with higher precision instrument (not only to improve
> measured values but also to determine the errors and repeatability
> of the measurement)

Your UPS reports a zero based 8-bit value for the battery voltage, which
is about 210 bins at a real value of 13.6V. This gives it a resolution of
almost 65mV. No calibration in the world can improve that. Even when
assuming zero non-linearity error and -tolerance, this is already way too
coarse to use this to base any action upon when it comes to 12V lead-acid
batteries. If users have a clue to decide if a reported battery voltage of
13.5V would be good or bad in the first place anyway.

My point is, it would be measuring without meaning.

> Hopefully the model detection will make it and at least the coarse
> conversion factor would be there for cyberpowers and others that
> need it.

I'll be happy to include this if CyberPower can confirm that this
adjustment is applicable to all 0x0501 productid's, both past *and*
present models.

Best regards, Arjen

More information about the Nut-upsuser mailing list