[olpc-help] XO shuts down when power cord removed
Richard A. Smith
richard at laptop.org
Mon Jan 14 01:51:57 EST 2008
Andrew Cooper wrote:
> That SO needs to be in an easily searchable FAQ!!! It worked like charm!! Here's what I got...
> Thanks again,
Here's my analysis of your data in time order.
Normal charge for about 35 minutes delivering 738 mA/h into the battery.
The voltage is higher though than I would expect. A deeply discharged
battery should be in the < 6V range.
On the next sample
The voltage jumped up to 7.5V which is above the CC to CV threshold also
notice that your current went from 1.2A to less than 1mA. Very
abnormal. The EC switches from constant current (CC) into constant
voltage (CV), marks the battery as 81% full and watches the current to
be less than 200mA. A few cycles of that and it declares the battery Full.
So 20 minutes later you started the discharge in that space you seem to
have lost 305 mA/h. Odd but ok I guess.
42 minutes of discharge until the voltage nears critical shutoff. 676
mA/h consumed. So 676+305 = 981 mA/h. Previous charge only had 738 mA/h
so that leaves 243 unaccounted for but thats probably close enough to match.
Here's a real indicator of your troubles. When the EC detected the
battery (SOC % is zero so I guess you reset it with batman.fth?) the
Voltage is 7.5V yet ACR shows that only 45 mA/h was added to the
battery. So the battery is still very discharged. The voltage should
be much lower and the current much higher. This condition is what
happens when a battery get close to full. So the EC marks it as such.
The other log files are more of the same. Very short charging period
with an abrupt voltage jump and loss of current.
Because you have one run of 35 minutes where the charging appears to
work correctly I'm inclined to believe that this is a physical
connection issue rather than a battery chemistry issue. I think you
have a bad connection somewhere in your charge path that is creating a
high resistance and faking out the EC charging code.
How close are you to your 30 days? If you are really close then I
suggest you RMA it. If not then I have one more test I'd like to try.
I'm going to hack on batman.fth and add the ability for you to _force_
charging. Then you can run this and watch what is happening while you
prod, jiggle, and torque things around. Then we can see if your
charging current bounces around. Try to see if we can find some sort of
cause and effect. If so then its a bad connection in the charge path
somewhere. If not then something is wonky with the battery cells.
Richard Smith <richard at laptop.org>
One Laptop Per Child
More information about the community-support