AFAIK someone mentioned on this very forum, that using test pads for both data and +5V power source, phone can get flashed without battery. Seems logical, cause proper tool for using test pads doesn't leave place for battery. Not tested by me, though.

---

dr_frost_dk, nice find. ho ever, I wonder why they mention "upstream" so many times - does it mean it will interfere with hostmode (aka using port as downstream)?

Don't know, that was just a 5minute google for ESD circuit/chip, a little more effort and we should be able to find just the right one that's in the N900 just for total "compatibility"

True, but just keep in mind that we would also need +5V/ground circuit protection. To get comparable quality (or, even better, higher, reminding You broken micro-fuse after switching + and ground of +5V...) we would need - probably - a little bigger one, than ESD protection. Just assuming, I hope I'm wrong.

Because epoxy was never used on small devices like the N900 is, solder is ALWAYS stronger and lasts longer under these conditions.

Not a wise move to mix glue alongside small smd's unless your aim is to hide design work and stop any repairs being done.

What is the problem with replacing the smd fuse? it is easy to get too easy to take off the board and even easier to solder back in place?.

I never could understand why people attempt a workaround especially when the workaround involves extensive work and the use of "alien" components when the obvious choice is to simply replace a broken component, an smd fuse is normally a straight forward either end solder job and especially in this case as it is sitting so easily for removal and replacement, also very easy to solder a link wire on to.

Remember one thing here.... the N900 works well as it is "standard" , try to modify away from the original design then you will encounter problems in the future.

The test pad mod is absolutely very ugly and should never be attempted, should only be used with the proper module that is sprung loaded to make contact with the TEST pads for testing and programming purposes NEVER should you solder on to these test pads !!!!..

Stupid thing about all this is the answers to a proper repair are ALL on the thread

http://talk.maemo.org/showthread.php?t=75920

i have some kind of weird problem, when i wan't to charge battery it shows on screen message "charging" but instead of charging, it drains my battery all down to zero!?
i can get usb connection and transfer files but this....
i make connection like on picture i attached: connected + and - data, +5v and GND, but only what i get when connect charger is draining battery
if anyone have success in charging and have usb connection working let me know which connection to use

I think You should not use battery - as 5V line ground (picture is obviously wrong in this), but let's wait for joerg_comment, he's expert on this.

I'm not going to comment on it, all said already
http://talk.maemo.org/showthread.php?t=75920
/j

and i didn't use battery gnd as -, i use one from round pinouts, and when try to charge phone it shows message "charging" as i said in post before.
and another interesting thing: when the charger is disconnected from wall the battery animation is still on, like it is on charge, but when micro usb connector is disconnected from phone then message came up to take out charger from power source
as i said before that this way is my only option to make charging and usb connection because i don't have so thin pin for soldering.

edit: i found on n900 schematic that +5v on usb pin is actually vbus, or how i find on cpkb.org usb power detector. so, if i connect the wire to the pinout that is described as vbus or usb power detector, does it means that it is the same thing as on the origginal usb connection?

here is the picture of schematics

(*sigh* who said I'm not going to comment? ;-P)
dvijetrecine, probably your wiring is not OK. I can't think of any other explanation than poor solder joints or too thin cables leading to too low voltage on VBUS +5V when N900 charger chip bq24250 starts to draw charging current from USB. The chip will notice that problem and throw error "poor VBUS" or sth like that and stop charging all together. I don't know if bme will detect this condition and pop up a notifier "no sufficient power for charging!" (OWTTE), or simply will happily continue to signal "charging" to battery-applet, but anyway quite obviously when the charger chip stops charging then there's no charging no matter what the GUI is signalling. Anyway when no charging is going on there'll be no load on VBUS and thus the wallwart fastcharger can keep voltage for up to several minutes when unplugged from mains - the gear detecting this voltage is unrelated to the bq24150 charger chip and thus your device doesn't notice you unplugging the charger from mains (unless you wait for long enough)
You possibly could confirm what's going on when you use shadowjk's charger script instead of bme, or simply do a `watch i2cget -y 2 0x6b 0x00`
Come here with the hex values to get them decoded and interpreted, I'm too lazy to transcript the datasheet's error state table here for now.

[edit2] here's an excerpt of charge21.sh_
# Register 0x00
# 8: Read: OTG Pin Status
# Write: Timer Reset
# 4: Enable Stat Pin
# 2: Stat : 00 Ready 01 In Progress
# 1: : 10 Done 11 Fault
# 8: Boost Mode
# 4: Fault: 000 Normal 001 VBUS OVP 010 Sleep Mode
# 2: 011 Poor input or Vbus < UVLO
# 1: 100 Battery OVP 101 Thermal Shutdown
# 110 Timer Fault 111 NA
so $((i2cget -y 2 0x6b 0x00 & 0x07)) = 3 means "Poor input or Vbus < UVLO"
ps[edit2]: any other error here is also stopping charging, so esp 001 VBUS OVP comes to mind which would mean you applied >6V VBUS voltage.
[/edit2]


[edit] please refer to http://talk.maemo.org/showpost.php?p...8&postcount=20 [/edit]
/j