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

davor emard davoremard at gmail.com
Thu Apr 10 22:07:21 UTC 2008


HI

They are still processing my request (it takes time),
in the meantime I have an idea. Suppose that each
ups voltage measurement could have some measurement
error. It could be used to correct some minor measurement
errors for any ups, but also could be used in the cyberpower
case

Let's introduce a calibration polynomial like
Ucalibrated = c0 + c1 * U + c2 * U^2 + ... cn * U^n

This will allow practically any type of conversion
including high-order nonlinearities.
in the config file we enter first n coefficients of the
polynomial, the rest c are considered as zero.

In cyberpower case it is c0 = 0, c1 = 1/1.53

Ucalibrated = 0 + 0.6536 * U

and this would be defined as ups.conf like

voltage.calibration = 0 0.6535

So everybody can calibrate his own ups battery measurement
make the polynomial fit and just enter the coefficients in
ups.conf or just use recommended values for some models
that are know to have systematic error

what do you say?

Best regards, Davor


On 4/8/08, davor emard <davoremard at gmail.com> wrote:
> I guess it's tied to vendor/product id probably all models e.g.
> 400/600/800 differ only in battery capacity and run the same
> firmware but I will ask them
>
> On 4/8/08, Arjen de Korte <nut+users at de-korte.org> wrote:
> > > I just got reply from Cyberpower R&D department about the battery
> > > voltage correction formulae
> > >
> > >> responsibility of the vendor (CyberPower). If someone provides me with
> > >> the
> > >> correct conversion values, I'm willing to include this in the
> subdriver.
> > >
> > >  Our software group provided the following
> > >
> > >  Battery voltage =  (Hid battery voltage) / 1.53
> > >  for example:
> > >     21.3 /1.53 = 13.9 (measured 13.6)
> > >
> > >  Note, the tolerance is possible for each unit of UPS.
> >
> > It is not clear to me what the line above means. Do they mean all units of
> > the same model, or all CyberPower devices with ProductID=0x0501. We can
> > only do this in the latter case since the device provides no information
> > about the model, so all we can work with are VendorID and ProductID.
> >
> > Best regards, Arjen
> >
> >
>



More information about the Nut-upsuser mailing list