[PW] Stuck in boot loop?

[abarthch]

Hi!

I bought a Paperwhite 1st Gen (Model EY21) off a marketplace where the guy claimed it doesn't turn on anymore and probably just needs a new battery. Turns out the battery is actually fine but the PW is stuck in a boot loop! I'm not mad because I only paid the equivalent of 10$...

When I first got it and pressed the power button the Kindle showed this huge battery symbol. So I connected it to my PC and it looked like it charged because the LED was orange and I actually could hear the typical Windows sound when you connect a USB device.

So next I hooked it to a lab power supply to see if it charges sufficiently. It did at the specified rate of almost 500mA (460 exactly). After half an hour the battery symbol went away (just a plain white screen) and the PW continued to charge at a 500mA rate. But at that point a funny thing started to happen. The Kindle would interrupt the charging every minute and reboot. So the Kindle charges at about 500mA for a minute, then drop to 100mA for 5 seconds, then 500mA again in an infinite loop. From the outside you can see that each time when it's at the beginning of the 100mA phase the screen flickers rapidly from white to black twice. W/B/W/B/W within a second.

Also now when I connect it to the PC I don't hear any sound or beep anymore. I have searched the forums but never came across a behaviour like this. Most boot loops I have read about were much longer where you could actually see the picture with the kid under the tree. Mine doesn't show any picture, except the huge battery symbol in the very beginning when the battery was too drained for the Kindle to boot.

In the meantime the battery seems to be fully charged because after the 5 seconds of booting the LED actually turns from orange to green for a minute before the boot loop strikes again. I have tried pressing the Home button for extended times (30s, 2min) but nothing interesting happens, it just continues to be stuck in this boot loop.

My next plan is to hook it up to a wall charger for week, but I don't have a lot of hope because I can see from the flickering of the screen and the LED that it's not repairing a file system or doing anything but just rebooting in the same pattern.

Does anyone know what's going on or have a better plan?

[knc1]

Originally Posted by abarthch

- - - -
My next plan is to hook it up to a wall charger for week, but I don't have a lot of hope because I can see from the flickering of the screen and the LED that it's not repairing a file system or doing anything but just rebooting in the same pattern.

Does anyone know what's going on or have a better plan?

Keep in mind that battery management (and the LED) is under firmware control.
And this firmware is not behaving normally.

Lowest battery charge indicated to highest battery charge indicated is about 1,500 mah.
1/2 hour * 500 mah != full re-charge.
(Post: 3.5 hours * 500 mah == full charge (to maintenace charge rate curve anyway) and I'll believe the LED).

So far all you've shown is the battery has a storage capacity of about 250 mah (normal for a very used / very worn out Li-Ion battery).
If driving an electric car, the same will happen to its $$,$$$ Li-Ion battery someday[**].

It is most likely a corrupted filesystem due to a damaged flash chip.
It isn't doing to get any better until the chip is replaced.
FGA, post #3: https://www.mobileread.com/forums/sho...49&postcount=3

Lets see the boot messages to be sure what is happening:
Open it up -
Connect to serial port (its 1.8 v serial to USB) -
Set Putty[*] to record a log (capture) file -
re-boot.

Then compress and post here as an attachment (use the 'go advanced' button to the advanced editor).

Stay happy, that used, front lighted touch screen is worth a lot more than $10.

- - - -

[*] Why is it Windows users with the problems?
Maybe I'll write a book: "Helpdesk: Attack of the Windows Users"

[**] Early manufacturers had the right idea, they used Nickle-Iron (Edison) batteries - some of the original Edison batteries are still in use 100 years+.

[abarthch]

Thank you for your quick answer.

Originally Posted by knc1

So far all you've shown is the battery has a storage capacity of about 250 mah (normal for a very used / very worn out Li-Ion battery).

I forgot to mention that the first time I got a green light was after about 4 hours of continuous charging (or maybe not exactly continuous, when you consider that the charging was interrupted every minute for a few seconds). So the battery seems to be doing fine so far.

In the meantime I've got some news. I've had it hooked up again on the wall charger since morning and it looks like the boot loop stopped for some reason, because now the display doesn't flicker anymore and the LED stays orange all the time (remember, before it was green with a short orange period everytime it rebooted). I should probably leave it like that for a few days before touching it again.

I have a an old USB to serial/parallel adapter (looks exactly like this one) that I used to communicate with my SAT receiver. Do you think the protocol works with the Kindle when I add some voltage dividers?

[knc1]

Originally Posted by abarthch

I forgot to mention that the first time I got a green light was after about 4 hours of continuous charging (or maybe not exactly continuous, when you consider that the charging was interrupted every minute for a few seconds). So the battery seems to be doing fine so far.
- - - -

Little details like that make a difference.

That's 12 volt, RS-232 serial (similar to a COM port - back in the days when computers had COM ports).

It is unlikely that the Kindle's nano-powered CMOS port can pull the signal line to zero (unable to sink enough current).

The Kindle's port evidently can tolerate 3v3 signal levels, other people have used that rather than 1v8 - it is just that I can't recommend it.
(The hardware reference manual is silent on the subject.)

[abarthch]

I've ordered a new usb-serial converter. I didn't find one with 1.8v within Europe but I hope I can get it to work with a few modifications.

In the meantime the Kindle stopped charging from the wall charger (unfortunately I chose the cheapest charger I could find, it looks like it's dead) and after a day of not charging it was back to that big battery symbol.

So I thought I would connect it again to a PC and apparently the Kindle is still recognizable as a drive. Look at the attached screenshot from USBDeview. Although it doesn't show up in Disk Management, in Explorer it shows up as drive F: but it cannot be opened and there's no disk size next to it and it's greyed out.

As to why the battery is already empty again when the wall charger stopped working for a day or so, I can only speculate. Either
1) the Kindle was in the process of repairing something as I hoped and that process drained the battery, or
2) the Kindle wasn't doing anything in that time and the battery is just garbage

[knc1]

Originally Posted by abarthch

I've ordered a new usb-serial converter. I didn't find one with 1.8v within Europe but I hope I can get it to work with a few modifications.
- - - - -

Set that one for 3v3 if it does not come that way and use it without any other modifications.
Other people have without any trouble.

- - - - -

This is one source of the 1v8 interface level to USB cables (if Norway is your country) -
http://www.digikey.no/

[abarthch]

So the PL2303 has arrived and I've already soldered three wires to the PCB of my Paperwhite:
-Tx to Tx
-Rx to Rx
GND to the shielding on the PCB (since the pin was too small)

Now, the PL2303 has two more pins, '+5V' and '3V3'. To make it work with 3.3v, do I have to short the 3V3 pin to GND?

[knc1]

Originally Posted by abarthch

So the PL2303 has arrived and I've already soldered three wires to the PCB of my Paperwhite:
-Tx to Tx
-Rx to Rx
GND to the shielding on the PCB (since the pin was too small)

Now, the PL2303 has two more pins, '+5V' and '3V3'. To make it work with 3.3v, do I have to short the 3V3 pin to GND?

That is reversed. You need:
Tx to Rx
Rx to Tx
Each device is labeled relative to its own function on each pin.

Probably easier to leave the Kindle board connections in place, and swap the 'large' ends of the wires.

Heck, I can't tell from the chip number on the adapter board.
Consult the directions that came with the device, goodle for the device's use (better find articles with photos, so you can compare your board to the one being written about), or PM someone here that has posted about using the same adapter board.

[abarthch]

Originally Posted by knc1

That is reversed. You need:
Tx to Rx
Rx to Tx

Shame on me

I didn't find a manual with the PL2303 but I avoided that problem by making a level shifter circuit. Now when I connect everything, boot up the Kindle and measure the voltage on the Kindle I get
Tx - GND: 1.8v
Rx - GND: 0.9v
The Rx-LED on the PL2303 flashes every so often.

After that I followed this tutorial on how to set up minicom but where it says:

When the kindle is starting you should see some output from your kindle startup in minicom console.

I don't get anything on the minicom screen. I have set the correct device (on my machine it's also ttyUSB0 like in the tutorial), 115200 bps and no flow control. I have tried with and without external power supply (usb micro).

[knc1]

Originally Posted by abarthch

Shame on me

I didn't find a manual with the PL2303 but I avoided that problem by making a level shifter circuit. Now when I connect everything, boot up the Kindle and measure the voltage on the Kindle I get
Tx - GND: 1.8v
Rx - GND: 0.9v
The Rx-LED on the PL2303 flashes every so often.

After that I followed this tutorial on how to set up minicom but where it says:

I don't get anything on the minicom screen. I have set the correct device (on my machine it's also ttyUSB0 like in the tutorial), 115200 bps and no flow control. I have tried with and without external power supply (usb micro).

Remove the Rx-LED as the warning in the level shifter recommends.

[abarthch]

Originally Posted by knc1

Remove the Rx-LED as the warning in the level shifter recommends.

That worked!

Can you see what's wrong from the capture log?

[knc1]

Hardware failure.
These:

mmcblk0: error -110 transferring data, sector 1054048 nr 8, cmd response 0x900 card status 0x0
end_request: I/O error, dev mmcblk0, sector 1054048
mmcblk0: error -110 transferring data, sector 1054049 nr 7, cmd response 0x200900 card status 0x0
end_request: I/O error, dev mmcblk0, sector 1054049

Indicate that the eMMC chip has died, at least its internal controller has.
Replacing the chip is the only repair.

There was a kernel error in the sd/mmc sub-system that was fixed back in 2011 .. 2012.
The messages above are formatted in the 'after fix' style, so ....

PM me if you have a used PW-1 screen for sale, I have a PW-1 with a broken screen.

# 89 89