Battery charge and Q2E42

[Richard A. Smith]

On 02/17/2010 12:25 AM, Mikus Grinbergs wrote:

> As I said, it may be the battery.  But I am surprised at how RAPIDLY
> that battery went from nearly fully charged (at the beginning) to nearly
> empty (with Q2E42, etc.), and how RAPIDLY that battery went from 18%
> charged (on Q2E42b) to 80% charged (on Q2E41, different machine).

The EC code between q2e41 and q2e42* are binary identical so its not 
anything introduced by new firmware.   I'm not going to introduce any EC 
instability into the mix until later.

State of charge has a lots of different ways that it can get out of sync 
with reality.  There are some sync points where the EC will try to 
correct the % value to specific values based on various criteria.

A month or so ago while in China I fixed a EC XO-1.5 bug where on 
battery recognition after a EC reset it can mess up the SOC %.  This bug 
is also present in current XO-1 firmware.  You don't see it much since 
you don't reset the EC much.  Flashing firmware resets the EC.  So its 
possible you hit that bug.

Rapid changes in SOC are usually a sign that the EC detected that SOC is 
out of sync with the battery and is trying to correct.

If you want to investigate further then you can use olpc-pwr-log in a 
terminal to look at whats going on.

More info here:

http://wiki.laptop.org/go/XO_power_draw#olpc-pwr-log

-- 
Richard A. Smith  <richard at laptop.org>
One Laptop per Child