PB TL4 does not turn on (sort of)

[koraks]

Hi community,
New user here; sorry if I missed anything essential in posting this thread. Likewise, if my problem has already been discussed, kindly link me to the threads/sites in question so I can do some reading.

The device is a PocketBook Touch Lux 4 which has been in daily use for the last 5 years or so. The problem description is that the device won't turn on, which is to say:
* The device does respond to the power button: power button light comes on, backlight comes on, Vcc comes on to ca. 3.2V (measured at test points on PCB). Once turned on, the device can also be powered down by keeping the power button pressed for 15 seconds or so, after which Vcc drops to zero.
* The display is unresponsive. It shows the "Touch Lux 4" screen with the marketing specs below. Display does not change, does not show any artefacts etc.
* When plugged into USB on a computer, I don't get device enumeration in Windows and no extra drive pops up in explorer etc.
* Battery voltage looks normal (ca. 4.2V). Device appears to charge just fine. Vcc looks normal (ca. 3.2V) as described above.
* On the test points for the display, I get some of the voltages I'd expect for an eInk display.

I've tried starting the device with just the battery and the display connected to the main PCB, so no backlight and touchpad controller. The behavior is the same.

I have not (yet) hooked up a computer to the UART test pads to see what happens. I do notice that the voltage on TX drops briefly in the seconds direct after power-up, which suggests the device is outputting something to UART.

Is this a known failure mode of this device? If so, what are the probable causes?

I am reasonably well-equipped with a scope, SMD soldering equipment and decent hobby-level working knowledge of embedded systems. The main factor is how much time I want to spend on this, but I hate throwing stuff away even if it seems 'broken' (my wife, whose device this is, has of course already ordered a new one, so there's no pressure...)

Thanks for sharing any thoughts you may have!

[nhedgehog]

Just to give you the feedback from another amateur that you are listened too:

From what you have written I come to the same solution as you probably are:
Its either the screen (most probably from my experience) or something on the mainboard. You need another working mainboard to be sure of it. Ebay is full of broken screen pocketbook devices, sometimes it needs a little time to find the right one.

No, this is not a known failure of the device, normally we see either broken screens due to probable mishandling of the device or sometimes some problems with the internal clock or charging failures.

[koraks]

Thanks so much @nhedgehog for chiming in and confirming that we're thinking in the same direction. The suggestion of getting another spare/donor device is an interesting one; I'll consider it.

I think the next thing I should try is hook up the device to a computer using the UART connection and see if I can pry some diagnostics information from it that way.

As to the potential for a display failure, I might dive into that a little deeper. There's a TI chip that generates the various voltages the display needs and while some of the voltages did check out, I think I encountered one or two test pads that didn't show the voltages that are printed on the silkscreen. This may or may not be a clue. There's also the possibility of something obvious like breakage of a conductor in the flat cable that I haven't spotted so far. I think I might be able to disassemble the display module one step further.

Any further ideas/suggestions remain very welcome!

[nhedgehog]

Yes a failure in the voltage creation for the eink display could be one of the most plausible mainboard failures that results in your error descriptions too.
Please keep me informed, sounds like I could learn a lot about by your project.
Eventually this could be helpful too:
https://www.synacktiv.com/en/publica...-reverser.html

This too:
https://linux-sunxi.org/PocketBook_Touch_Lux_3

[nhedgehog]

And don't forget to look at this:
https://www.mobileread.com/forums/sh...ART+Pocketbook

[nhedgehog]

There was a website which got deeper into the Pocketbook UART connection, htouh i seems offline.
https://cod-box.net/pocketbook-conne...via-uart-hack/

Wayback gives us some of it back, but I had problems to see the whole site including pics.:
https://web.archive.org/web/20240522...via-uart-hack/

[koraks]

@nhedgehog thanks so much for the links; they're super-useful! And yes, I've read the forum post about the hwconfig file, and indeed I had also recovered the info about UART access. Fortunately that part seems bog standard UART communications (although perhaps flow control is non-standard) with 115200/8N1 over the clearly labeled test points on the mainboard. As I understand, UART will simply give terminal access, so from that point onwards, all the information about the linux part should apply.

Thanks again; this helps!
So far I've not had time to work on the device and it's kind of on the backburner, but I hope to get back to this.

[koraks]

I've done a little rooting around last night; hooked up the e-reader to a computer via UART and played with it for a bit. It seems to work just fine apart from the unresponsive display. The voltages generated by the TPS65185 seem OK as well and this seems like a low-risk part to begin with in terms of defects.

I've tried to disassemble the front panel a little further to have a closer look at the display, but since it's all glued together, I didn't get very far. I'd have to heat it up to pry the display loose, but I think I saw a crack in the glass somewhere, which may just be the whole problem to begin with.

Not sure what I'll do with this; maybe just keep around the mainboard for a future project that calls for a SoC with WiFi.

[nhedgehog]

OK, so it was the screen but broken at an unusual part preventing it from showing those classic artefacts we know so well from broken screens. Just to learn from it, did your wife use the Pocketbook in case or did she use it with out it?
Disassembling displays is something I wouldn't consider even with the right equipment.

[nhedgehog]

I find Aliexpress a good source for PB screens, though they are not cheap:
https://aliexpress.com/item/4000469717370.html

[nhedgehog]

But consider this too:
https://www.mobileread.com/forums/sh...d.php?t=356344

Screen voltage is set via the value in the vcom file in the boot partition, which is normally not accessable.

[koraks]

Originally Posted by nhedgehog

OK, so it was the screen but broken at an unusual part preventing it from showing those classic artefacts we know so well from broken screens.

That's my hypothesis at this point, yes, although I can't draw a definitive conclusion.

I've looked at replacement screens briefly, but the cost difference with a brand new device in a comparable league is too small to justify spending any time on replacing it. Note that with the replacement screens you linked to, it's still necessary to remove the battery from the screen and the screen from the metal frame & plastic fascia, then remount (glue/tape) these items back onto the new screen. It's feasible, but quite a bit of work especially if it's going to be a one-off repair.

Thanks for the hint on the drive voltage; I doubt this has much to do with it because the malfunction was sudden ("last night it worked fine, but today it won't turn on"). Moreover, there have been no issues with changes in contrast etc which would signify drive voltage problems, and there's no sign (or likely cause) of any defect that would have caused a sudden change in drive voltage. Not to mention that in such a case, the fix would be to remove that underlying cause instead of adjusting the voltage. Given that the voltages generated by the TI chip look entirely plausible, I don't think I'll spend time exploring into this direction.

I did check with my wife and she confirmed she had carried the reader in a different spot in her bag where it's more likely to have been compressed. My guess is that the device flexed a little despite being inside a protective case, breaking the glass at the backside of the screen where it's barely visible.

Here's some shots of the supposed crack in the glass:
Overview with main PCB removed:


A little closer, looking across the flex cable with the touchpadcontroller:


Closer still from the other angle:


What I don't know is (1) whether this really is a crack in the glass; I'd have to remove the metal frame to be sure. (2) If it's a crack, whether or not it happened while I was working on the device - I didn't notice anything, but glass can crack unnoticeably. (3) If a crack in that place would explain total display malfunction. I can see how it might be a good explanation if it extends across all data lines, but I don't know the physical layout of the screen itself.

So some uncertainties left, although I'd say with 95% confidence that this is the cause. In which case, it's not really viable to repair this unit unless I can find a replacement screen for like €10 somewhere.

[nhedgehog]

Its definitely a crack. The place of it explains the failures you described perfectly. On the picture you can even see the broken electrical conductors. Vcom info was only if you decide to change the screen and use a ED060XH7 instead of the original EX060XCD one. I go with you, it isn't worth the costs replacing the screen of this modell.
The location of the crack is strange. I hope is is not due to internal strains because of design flaws. I heard rumours but those cases never hold up to the claim since the structure damage led to more probable different causes. Battery looks good too, so it didn't cause the crack by slowling blowing up.

[koraks]

Yeah, I'm pretty sure the crack is simply due to the device getting squashed in wife's bag. Too bad. Well, maybe one day I'll think of something fun to do with that A5 SoC!