[Nut-upsuser] Fwd: Cyberpower usbhid-ups wrong battery voltage
davoremard at gmail.com
Wed Apr 2 14:29:45 UTC 2008
Yes, I mean this is quirk of a certain manufacturer/model
and the fix should be tied to vendor-id product-id.
My proposal is to even when we see product/vendor id, we make
extra sanit(ar)y checks before fixing values just to make
sure cyberpower haven't fixed them in some newer firmware
and we do fix already fixed values
so the condition may be to check battery is 12V and
PbAcid and interpret the unfixed battery voltage. If we see
unfixed voltage greater than let's say 15V we immediately
know we need to fix that value by replacing e.g. min-max
so that in future if the cyberpower fixes min-max in newer
firmwares and doesn't change vendor/product/version id,
we will still have everything working well
On 4/2/08, Arjen de Korte <nut+users at de-korte.org> wrote:
> > So if we see battery voltage 12V and PbAcid chemistry and
> > nonsense PhyMin and PhyMax in respect to current voltage,
> > we might do the heuristic decision to set those values
> > to 9.8 and 14.4 as you suggested and everything would
> > look good?
> First of all, you'd get the same result for 7 and 15 for instance, so the
> values mentioned should be regarded as examples only. Note that PhyMin =
> PhyMax = 0 is a perfectly legal value and quite a number of UPS'es use
> Assuming a first order relationship between logical and physical values,
> you'd need at least two different readings (preferably spaced apart
> widely) to get to the real PhyMin and PhyMax values.
> Second, using observed measurements to decide what to do, has proven to be
> a very support intensive solution in the past, so this won't happen under
> my watch. So basically we can only work with the physical value that is
> reported here and have to work with that. Since there is no guarantee that
> this won't overlap with correct values that don't need fixing, this will
> have to be based on ProductID only.
> Lastly, this problem is due to the UPS firmware being broken, which is the
> responsibility of the vendor (CyberPower). If someone provides me with the
> correct conversion values, I'm willing to include this in the subdriver.
> So far CyberPower hasn't been very supportive when it comes to NUT,
> therefor I'm not going out of my way to fix the problems they created.
> There are UPS vendors that provide much better support/help in this area.
> Best regards, Arjen
More information about the Nut-upsuser