current focus

Sean McNeil sean at mcneil.com
Thu Jul 10 02:36:17 CEST 2008


Hi Werner,

Thanks for all the detail. Very interesting. Question:

Werner Almesberger wrote:
> Sean McNeil wrote:
>   
>> How are things with the battery draining issue and phone lockup? This  
>> causes a bad state where the user cannot charge the battery because they  
>> cannot start the phone, right?
>>     
>
> There are two causes:
>
> 1) u-boot exceeds the 100mA current budget and thus can cause the
>    system to die if the battery is too weak. This can be fixed in
>    software. I have a proof of concept hack for this, but it still
>    needs to be turned into something one can actually use.
>   

Can u-boot be modified to bring up the system differently so that it 
doesn't exceed the power budget?

> 2) The inrush current on our power rails, and in particular IO_3V3
>    can be too large for the bypass capacitor on VB_SYS, thus causing
>    the power supply to collapse and the PMU to cycle, unless a
>    battery is present. (*)
>
>    Details are here:
>    http://people.openmoko.org/werner/gta02-chg/
>
>    Some boards have capacitors that may just be sufficent. I'm now
>    making some measurements to determine how much exactly we need,
>    and also what impact this has on the inrush current we inflict
>    on the USB host.
>
> (*) A battery that's discharged deeply enough can go into cut-off
>     mode, causing it to become completely unresponsive until
>     recharged for a bit. "A bit" depends on the battery type: the
>     GTA01 battery leaves cut-off pretty much upon seeing the first
>     electron, while the GTA02 battery wants to have a few millijoule
>     in its belly before it's willing to talk to the world again.
>
> There's also the issue of VB_SYS breaking down when enabling the
> charger while there is no battery. We haven't analyzed that one
> completely yet, but it seems to be caused by inrush current as well.
> Making sure the VB_SYS bypass capacitor is large enough solves this
> as well.
>   

This sounds like it can't be worked around in software. Correct? Can the 
demand on the power rails be mitigated by turning things on over a 
longer period of time?

> - Werner




More information about the openmoko-kernel mailing list