It won't shut down!

[slacker20012]

I guess this is better than it not switching on, but I can't get my pandora to stay switched off! Sometimes it will go off but then it will power itself on again - the only surefire way I have of switching it off is removing the battery but I'm not sure this is entirely healthy for the unit. Has anyone else experienced this issue?

 

[meandu229]

Is it plugged in? As the charging circuit boots up the pandora if its plugged in.

 

[slacker20012]

d'oh! Blonde moment, ignore me!


It was plugged in lol

 

[DaMummy]

yeah um, dont remove/insert the battery when the pandora is plugged in

 

[jorlinn]

I still think it's quite silly to autoboot whenever you insert the charging cable.

I would even go so far as to call it a design flaw.

 

[EvilDragon]

That design flaw happens to be with every device using a Cortex A8 CPU - or any other modern SoC as well.

They all start up when you connect the charging cable.

However, they got a minimal OS in the bootloader that simply charges and switches on an LED (or sometimes displays something on the screen).

We don't have that minimal OS, but everyone is free to code such thing, and I'll gladly include it in a firmware update.

It's no design flaw, it's only the software which hasn't been written to do that.

 

[quartercast]

A minimal OS/charging screen... Would that be trivial for someone with the know-how to code?

 

[WizardStan]

A minimal OS/charging screen... Would that be trivial for someone with the know-how to code?

No. If it were trivial it would have already been done.

 

[EvilDragon]

It's trivial, just a proper way to implement it needs to be found.


You should even be able to use a modified version of Pandora Systeminfo.

 

[Stephane Hockenhull]

pseudo code for shutdown script:

if charging_cable_plugged then

    display "charging.png"

    while charging_cable_plugged do wait;

end if

for init script:

if charging_cable_plugged then

    display "charging.png"

    while charging_cable_plugged && !power_button_pressed do wait;

    if power_button_pressed then

       continue_boot

    else

       screen off

       init 0

    end if

end if

anyone good with linux init scripts ?

 

[EvilDragon]

That would halt a reboot, which is bad for devs who connect to their Pandora from work via SSH and reboot.


If the unit is being rebooted, it should not show the charging thing.


So you would have to create a temporary file on reboot (not on halt) so that the charger knows it's a reboot and not a fresh start.

 

[Stephane Hockenhull]

or put a setting to disable the wait script for the few devs who remote to their pandoras ?

would be better to have the devs disable it than have everybody else's pandora do a write/delete temporary file on each reboots.

OMG THE NAND!!!!1111!!!!

devs know what's up with the charger

as a quick hack we could also simply put up a "charging.png" for 3 seconds on boot if the charger is plugged in. "oh! right! charger!"

or display a different loading image.

 

[WizardStan]

It is possible to disable the boot-on-charger setting. It requires installing the i2c applications from the Angstrom repo though.

 

[SNESFAN]

Is there a way to tell if the unit was powered on by the charger rather than flipping the switch?

If so, have it boot to a minimal OS on charger initiated boot. If charger is removed, then shut down.

If power switch initiated the boot, then power on normally. If power switch is flipped while in charging OS then restart to normal OS.

If shut down while in normal OS with charger plugged in then boot charger OS. If restart in normal OS, then restart into normal OS.

 

[Neelix]

It is possible to disable the boot-on-charger setting. It requires installing the i2c applications from the Angstrom repo though.

What would that do in terms of charging if you then plugged in the charger while the pandora is switched off?

 

[urjaman]

What would that do in terms of charging if you then plugged in the charger while the pandora is switched off?

My educated guess is that it wont charge. Atleast as long as the battery has enough charge to keep the power chip in standby. Eventually it will propably transition to trickle-charging the by-then close to dead battery (having forgotten settings, etc) and will then after the 3.2V threshold is crossed boot up again.

My take on this is that if somebody would pay me to do the elegant charging system, I'd do it, but the current behaviour is not annoying enough to warrant me spending my free time on it. (Note that most phone manufacturers had paid-for programmers that were told to do it, too )

 

[notaz]

How much do you cost? B)

 

[WizardStan]

My educated guess is that it wont charge. Atleast as long as the battery has enough charge to keep the power chip in standby. Eventually it will propably transition to trickle-charging the by-then close to dead battery (having forgotten settings, etc) and will then after the 3.2V threshold is crossed boot up again.

That's the proper way to do it. The improper way is to tell it to charge anyway, but then there's no indication that it's charging, and when it's full it trickle charges. We've seen what happens to trickle charged batteries: puffy batteries.It's too bad there isn't a way to get it to turn on when it detects the battery is full. It'd be pretty simple for the boot process to recognize it was turned on by the power supply, detect the battery is at 100%, tell the TWL to stop charging, and then just shut itself back off again.

 

[quartercast]

How much do you cost? B)

How much do you... charge?


Get the pun? Ho ho ho