test result of battery current against display brightness and GPS power mode

[Rask Ingemann Lambertsen]

On Mon, Mar 30, 2009 at 02:41:54PM +0200, Richard Kralovic wrote:
> Rask Ingemann Lambertsen wrote:
> > On Sat, Mar 28, 2009 at 07:32:24PM +0100, Richard Kralovic wrote:
> >> It may be the case that the fixes for the current leak were introduced
> >> in devel branch (linux-openmoko-devel). On kernel 2.6.29-rc3, my tests
> >> show a drop from cca 80mA to cca 47mA.
> > 
> >    Do you know which git revision that kernel is? Alternatively, where did
> 
> I am using gitr1e257a0e99817a338e3706708ebb5036518e46d8, I compiled it
> myself.

   OK, I think I've found something. If I boot up Debian with LXDE (with
xscreensaver and pcmanfm uninstalled) and just run the test there, I once in
a while get 53 mA, but usually it's 55 mA. Now I stopped some userspace
stuff:
# /etc/init.d/apmd stop
# /etc/init.d/nodm stop

$ sleep 120 && cat
/sys/class/power_supply/battery/{status,current_now,voltage_now,capacity}
Discharging
47250
4151000
100

   I.e. 8 mA lost because of something in userspace. :-(

   I'm working on CPU frequency scaling support. Slowing the CPU to 100 MHz
and 1.1 V core power supply gives:

$ sleep 120 && cat
/sys/class/power_supply/battery/{status,current_now,voltage_now,capacity}
Discharging
39750
4148000
100

:-)

-- 
Rask Ingemann Lambertsen
Danish law requires addresses in e-mail to be logged and stored for a year