u-boot low-battery handling

Mike Montour mail at mmontour.net
Thu Aug 28 19:11:37 CEST 2008

Joerg Reisenweber wrote:

> I strongly recommend reading of 
> Battery Charging
>     Specification

will do...

> First approach like above seems the right way to go. Though device should 
> *try* to boot up anyway and simply die when current-limiter on USB-inbound 
> current cuts out and we got no battery backup.

That's what breaks the GTA02v5 when the battery cutoff has activated - 
the PMU tries to go into Active state, but this triggers a brownout on 
Vsys when it activates some of the LDOs (based on how Andy explained it 
on the hardware ML), before the CPU can do anything. Not waking on 
insertion allows the charger to bring the battery out of the cutoff 
state, which takes ~10s.

If this problem does not exist on newer models then this behaviour of 
disabling wake-on-USB could be applied only to the older hardware 
revisions (although it would be more confusing to users if older and 
newer units behaved differently on charger insertion).

I suppose we can live with the need for a "jump-start" battery once the 
rest of the power-management code matures to a point where the battery 
will never become discharged far enough to trigger the cutoff.

More information about the openmoko-kernel mailing list