[Gta04-owner] Phoenux drained and resuscitated

[Dr. H. Nikolaus Schaller]

Hi Neil,

Am 22.10.2011 um 11:45 schrieb Neil Jerram:

> Interesting...
> 
> A few minutes ago my Phoenux wouldn't boot, and I started writing an email to say that I thought the battery was drained, and to ask what to do about that.  I know there was a recent email from Nikolaus that covered this in detail - and which left me with the impression that it was not an easily recoverable situation - but I can't seem to find it.
> 
> But I left the phone plugged into my laptop USB anyway, and after a few minutes, it started a series of reboot attempts, with each attempt apparently getting slightly further than the previous one.  (Through the sequence of (1) initial white screen, (2) Golden Delicious big penguin screen, (3) Blank screen with small penguin in top left corner.)
> 
> On about the fourth attempt, the boot reached stage (3), and then continued successfully.  So it seems that if it gets as far as booting Linux, it's over the hump.

Yes! Linux initializes the USB charger in one of the earlier steps and then the system comes up.

As long as your PC supplies enough current (>500 mA) and you don't switch on too many high
power consumers (WLAN, WWAN), there should be some remainder for charging the battery.

Maybe we can improve that by enabling USB charging already in MLO i.e. some seconds before (2).
It is a little tricky since MLO does not have an elaborated I2C infrastructure to access the PMU. And
is quite difficult to debug. So never touch a running system :)

> Then I immediately ssh'd in to look at the battery status:
> 
> gta04:~# cat /sys/bus/platform/devices/twl4030-bci-battery/power_supply/twl4030_bci_battery/uevent
> POWER_SUPPLY_NAME=twl4030_bci_battery
> POWER_SUPPLY_TYPE=Battery
> POWER_SUPPLY_STATUS=Charging
> POWER_SUPPLY_ONLINE=3
> POWER_SUPPLY_VOLTAGE_NOW=3222
> POWER_SUPPLY_CURRENT_NOW=1615
> POWER_SUPPLY_CAPACITY=20
> POWER_SUPPLY_TEMP=29
> 
> I guess - but don't really know - that these figures (voltage, current, capacity) are low enough to support the theory that the initial failure to boot was because of battery drainage.

The level where the system starts is 3.2 V. If the voltage stays above 3.2 V it keeps running. Then, it should slowly go
up to 4.2 V (battery full).

What I am not sure about is if the value of current_now is correct. It is the same range as I have seen,
but I suspect some bug in the calculation in the driver.

The capacity is not a valid value - the twl4030 can't read that.  We need the HDQ interface to ask
the BQ27000 chip inside the battery. David has it working in the 3.1 kernel.

The temperature is also not valid since there is no NTC in the GTA02 battery.

>  Does that sound right?
> 
> Anyway, at the end of the day all is well, and it looks as though battery drainage is not as unrecoverable as I had thought it might be.

It may not be reliable or not all devices may be able to overcome the hump.
But generally this is good news!

Maybe others can share their experiences.

Nikolaus