Re: envstat wrong: who's at fault?

[Jukka Ruohonen <jruohonen%iki.fi@localhost>]

On Mon, Jul 26, 2010 at 11:59:10AM -0400, der Mouse wrote:
> The last half-hour of this graph makes me go "w t *f*!?", but I've looked
> at the raw logfile manually, and that jump really is there in the envstat
> output.

On Mon, Jul 26, 2010 at 12:53:43PM -0400, Steven Bellovin wrote:
> How old is the battery?  In my experience, as batteries age their
> discharge characteristics change markedly.  The hardware-reported value
> may not account for that properly.

That spike might be also explained by the "granularity" values that ACPI
optionally reports. (I added these some time ago to HEAD.) The specification
defines these as:

        [...]

        Battery capacity granularity between low and warning in [mAh] or
        [mWh].  That is, this is the smallest increment in capacity that the
        battery is capable of measuring.

        [...]

        Battery capacity granularity between warning and full in [mAh] or
        [mWh].  That is, this is the smallest increment in capacity that the
        battery is capable of measuring.  This may be a different value than
        Battery Capacity Granularity 1 to accommodate systems where the
        granularity accuracy may change depending on the battery level.

Apparently there are batteries/chips in the market that may loose/gain some
accuracy when the battery is nearly full/empty. I have no idea how that
might reflect to the individual readings.

But this is all very hypothetical. Does anyone know what kind of chip might
be in a laptop battery? Or how ACPI actually interacts with the battery?

- Jukka.