Providing high current on battery pin from USB gets it booting

[Joerg Reisenweber]

Am Mo  30. Juni 2008 schrieb Andy Green:
> Somebody in the thread at some point said:
> | Andy Green wrote:
> |> Anyway the main battery + pin shows no signs of any dents or load and
> |> sits there at about 0.4V itself before during and after the fun on Vsys.
> |> ~ I don't think this can be it.
> |
> | I added an 100uF bypass to VB_SYS and a 470uF capacitor where the
> | battery ought to be. That monster didn't impress the system at all,
> | and it happily continued to run.
> 
> Now my backup battery has charged up (to ~1.3V), the behaviour changed,
> I couldn't hear the ticking sound from the mic any more but it is still
> spinning with Vsys trouble.  However the Vsys broken part is much more
> variable and can oscillate -- the attached pic shows a 250us outage
> happening at low probability.
> 
> | I then just shorted the battery. That got me some "motorboat"
> | effect. So the battery isn't completely ignored.
> 
> I also tried to hook VBUS (USB +ve after the host switch) to the Main
> battery +, it starts beautifully.  But a LOT of current is pulled that
> way, hooking it to the battery + via 5R did not help the symptom, a
> chunk of wire did.  So the system tries to pull >2A I would guess from
> Vsys starting the whole disaster.  Where is all that current going!
> 
> It's a strong indication USB power is disconnected from Vsys by the PCF
> due to USB overcurrent as suspected, we fill in the hole from the
> "battery" and we go on.
> 
> | I'm confused :-(
> 
> It all comes down to where is all the current going at startup.  We'll
> get unconfused real quick when we work that out.  The regulators are
> current limited during this phase and shouldn't be able to add up to
> ~10W that is suspected.  It seems the culprit is via Vsys path not via
> battery directly.

The auto_converter powering IO_3V3 is designed to deliver high current, and in 
step-down mode probably even might draw way more than 2A, to charge C1707 
(47u) and dunno what else. I wouldn't rely on current-limiter of this 
regulator on startup.
To test it might be a simple setup to feed some 3V to IO_3V3 from external 
powersupply, thus precharging all these Cs - dunno if this would terribly 
spoil anything on startup. However the test is so simple it might be worth 
it.

/jOERG
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 194 bytes
Desc: This is a digitally signed message part.
Url : http://lists.openmoko.org/pipermail/openmoko-kernel/attachments/20080630/40f61602/attachment.pgp