Re: Weird battery (?) problem bricked N900

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



[Dawid Lorenz Di  13. April 2010]:
> [...] He said that during normal
> use, just while fiddling with menu/apps, suddenly a "charging error" message
> appeared (*without* a charger connected!), notification light turned red and
> device has switched off by itself, never booting back again.

No idea yet what to make out of that to come up with a reasonable story. 
*could* be a short in  battery, heating up the cell and same time losing 
capacity and tripping the temperature sensor. It could be a wacky contact in 
battery's BSI. Could also be a hardware break in device itself, or even a 
software glitch (in bme?). Anyway there's no means the red indicator lights up 
without software commanding the LP5523 LED controller chip to do so (remember 
that fact for later explanation)
 

> As soon as I've got his N900 in my hands, I've started with simply trying to
> switch it on. It does indeed switch on, however it gets stuck on white NOKIA
> splash screen for about 20-30 seconds and then shuts down again. Once
> connected to the charger, it gets into infinite reboot cycle, still not
> getting any farer than NOKIA splash screen. 

This can be caused by system init or bootloader finding some fishy details in 
battery and power state. Also there are other pitfalls that can cause infinite 
boot loop. Flashing a new fiasco image should cure most of them. Nevertheless 
see below about "Houston, we got a problem" bootcounter issue.


> Note that notification light is
> *solid* amber while hooked up to the charger (usually blinks slowly). 

Though you'll probably hear other tales, it's actually a fact the N900 *can* 
charge without system booted up, I.E. without any software support from 
system. The BQ24150 USB battery charger controller chip is able to charge 
battery fully autonomously (though initially with as little as 100mA only), 
and it signals this charging state with a hardwired yellow indicator. Later on 
during *normal* boot process this hardwired yellow light eventually is 
switched off when bme daemon takes control over the charger chip, and also bme 
signals mce daemon to do the yellow flashing same moment.

> Also,
> what's probably worth noting, while powering up notification light doesn't
> turn on all, while this light in my N900 is always turning white for 1-2
> seconds before actually switching display on with initial NOKIA splash.

This initial bluish white indicator ramp-up is done by NOLO bootloader, and I 
don't know exactly what to make out of the fact it doesn't show on the 
defective device. NOLO might decide not to light it up on certain unusual 
conditions (e.g. battery too low(1) ), or NOLO is broken, or the whole system 
doesn't start CPU due to battery too low (GAIA power management chip won't 
start the system when battery voltage is too low(2) - so the basic charging by 
charger chip will slowly charge up the battery and system doesn't draw more 
power than the 100mA) - but AIUI you see this missing white indicator even 
with your good fully charged battery.
[too_low(2) < too_low(1)]

> Anyhow, since device actually does switch on, I've tried to reflash a
> firmware, discovering shortly after that device doesn't enter usb mode while
> pressing "u" key and plugging cable in. Just powers on and nothing happens
> at all. Crap. :(

When system powers up, that means NOLO works, and if NOLO detects "u" hold, it 
should enter flashing mode - unless low battery.
So with the considered dead or discharged original cell this behaviour is 
somewhat expected.

 
> Now very interesting thing happened when I've swapped battery from my own
> N900. On that battery it still didn't boot fully (dying on NOKIA screen),
> HOWEVER I was able to enter usb flash mode and reflash firmware. Also, while
> connecting wall charger it does display usb icon in screen corner for few
> seconds. Unfortunately, despite reflashing both OS and eMMC images, it still
> doesn't work, simply shutting itself off after 20-30 seconds. :(

There's been an issue with some counter tripping after 30(?) failed boots. 
Maybe somebody else more savvy (jebba?) on that topic can take over to check 
if failing boot might be caused by this. You might want to check bugtracker 
for this issue.

 
> Nonetheless, I was very suprised to see that device was acting a little
> differenty with other battery in it. Does that give any clue? 

First of all it seems to indicate the original battery is near flat. You 
should try (to charge) that considered defect battery in your known good N900, 
and if that succeeds then do the check again to see if it still acts 
differently than with your known good battery.


> Is there any
> chance to ressurect my friend's N900? Or a warranty repair is the only
> reasonable option here?

Too early to do any honest judgement.
The sudden red indicator is somewhat scary, and it doesn't indicate anything 
if you don't see it anymore now - see above for explanation why a working 
system is needed to light up that red light.

Good luck
jOERG

Attachment: signature.asc
Description: This is a digitally signed message part.

_______________________________________________
maemo-users mailing list
maemo-users@xxxxxxxxx
https://lists.maemo.org/mailman/listinfo/maemo-users

[Index of Archives]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [XFree86]     [Big List of Linux Books]    

  Powered by Linux