Display units with large tap-insensitive margin

[agent008 243]

34 minutes ago, EskeRahn said:

I'm a little confused here. Why do you need the OTA update? Isn't the issue fixed with flashing newest stock directly? Do it need to be flashing an older version with successive OTA update to be fixed?

I flashed the 20200825 fastboot firmware from the Back to Stock guide thread: https://community.fxtec.com/topic/2466-guide-restore-to-stock-firmware-using-fastboot-method/?do=findComment&comment=33838

 

[EskeRahn 5,471]

9 minutes ago, agent008 said:

I flashed the 20200825 fastboot firmware from the Back to Stock guide thread: https://community.fxtec.com/topic/2466-guide-restore-to-stock-firmware-using-fastboot-method/?do=findComment&comment=33838

But there have not been an update to 20200825, so even if OTA was up, that would not offer anything newer.

[VaZso 1,998]

49 minutes ago, VaZso said:

Anyway, Pro1's touch driver seems to be called "gt1x" and it is an I2C device.
There is a related code at GitHUB which also has an update code.

Also, i2cdetect shows two buses on Pro1, i2c-6 and i2c-7.
In former, there are devices at address 0x28 and 0x58 and in latter, at address 0x08, 0x09 and 0x14 where all devices are seem to be used by a driver except 0x09.

However, based on gt1x_generic.h, I2C address of touch panel should be 0x5D (if that is the same device).
There may be other I2C buses which are not accessible through dev filesystem...

Also, i2c-7 has an owner of system.audio

Anyway, device 0x58 is in use by aw9523b which basically is the keyboard, 0x28 is used by nq-nci which is the NFC controller, 0x08 is used by i2c-pmic, there is a 0x38 which is used by fts, 0x4a which is used by mxt-ts and 0x14 which is used by gt1x module.

So it seems touch panel is connected to bus i2c-7 at address 0x14.

Edited May 3, 2022 by VaZso

[agent008 243]

14 minutes ago, EskeRahn said:

But there have not been an update to 20200825, so even if OTA was up, that would not offer anything newer.

I see. I might try an older stock then, if OTA is still up i might be lucky. I only found older stock firmware at @tdms link but they are EDL files and his Linux tool didn't work for me. I'll try the windows tool. EDIT: or if anybody has the older fastboot image I'll be thankful if you share it with me. Fastboot works well with my Linux box.

Edited May 3, 2022 by agent008

[VaZso 1,998]

3 minutes ago, agent008 said:

but they are EDL files and his Linux tool didn't work for me. I'll try the windows tool.

I think Windows makes it harder. 🙂

Your problem may be potentially related to the USB port you use.
Try to find a port which does not have other devices attached as EDL is really picky for timings.

Edit: I mean in your computer there are USB HUBs which have more or less devices attached and usually not all ports on your PC are connected to the same root HUB.

Devices connected to a root HUB (like internal web camera, WLAN or other components) may generate interrupts which delays communication with the phone and may cause update process running into  an error.

Edited May 3, 2022 by VaZso

[VaZso 1,998]

19 minutes ago, VaZso said:

Anyway, device 0x58 is in use by aw9523b which basically is the keyboard, 0x28 is used by nq-nci which is the NFC controller, 0x08 is used by i2c-pmic, there is a 0x38 which is used by fts, 0x4a which is used by mxt-ts and 0x14 which is used by gt1x module.

So it seems touch panel is connected to bus i2c-7 at address 0x14.

Okay, so it seems 0x14 is the default address for gt1x touch panel instead of 0x5D, see here:

Quote

"reset pin output 0 -> INT pin output 1 if i2c address is 0x14, or output 0 if i2c address is 0x5D -> reset pin output 1 -> delay
(...)
To solve this problem, we fix the I2C address in the firmware to 0x14, and 
discard the operation on INT GPIO in our touch driver."

So basically gt1x codes which use address 0x5d seem to speak about the same thing.

Edited May 3, 2022 by VaZso

[agent008 243]

23 minutes ago, VaZso said:

I mean in your computer there are USB HUBs which have more or less devices attached and usually not all ports on your PC are connected to the same root HUB.

 

I used two different ports from my laptop, will try a third one. Thanks

[EskeRahn 5,471]

1 minute ago, agent008 said:

I used two different ports from my laptop, will try a third one. Thanks

Try an external hub also!
It helped here using the front USB on a Lenovo Thunderbolt-4 docking.

[agent008 243]

15 minutes ago, EskeRahn said:

Try an external hub also!
It helped here using the front USB on a Lenovo Thunderbolt-4 docking.

Thanks. With my Samsung laptop, it doesn't work. Or it might be my Pro1's already a bit worn out USB port?

The saga continues...

[Rob. S. 1,664]

1 hour ago, agent008 said:

2 hours ago, EskeRahn said:

But there have not been an update to 20200825, so even if OTA was up, that would not offer anything newer.

I see. I might try an older stock then, if OTA is still up i might be lucky.

If I understand this correctly, if the freshly flashed 20200825 ROM doesn't do the touch margin adjustment, neither would an OTA update from an earlier version – and you might be having a completely different problem. "Screen completely unresponsive, 100% not working" would also make it seem so.

But just to be sure, did you actually boot the 2020825 stock ROM after flashing it? It needs to start up once to do the touch margin adjustment trick.

 

[agent008 243]

29 minutes ago, Rob. S. said:

But just to be sure, did you actually boot the 2020825 stock ROM after flashing it? It needs to start up once to do the touch margin adjustment trick.

Thanks. Yes, I did. In fact the phone is on right now in hopes the OTA might come yet. haha

If I remember correctly, last time it took an OTA to solve the screen too.

However, I will reopen the phone this evening and recheck all connections. Maybe something is not right in there. I also can try the old screen and a third new one I have still on the box.

Thanks for everyone's inputs!

Edited May 3, 2022 by agent008

[Hook 3,043]

@agent008 I have never heard of an OTA being needed or used if the 20200825 version of stock was flashed.  It should have the fix.  There was a file floating around with files from BOE that could fix the margins of the new display on Lineage 16-17, but I suspect, as @Rob. S. suggested, you are dealing with a different problem.

Last November, my original, factory installed  screen stopped responding to any touch.  I checked the connections. They seemed good. So I had a local shop that is very good replace the screen  with a new AlieExpress screen.  Still, no response to touch. That tends to rule out the display itself.

Since this occurred 3 weeks from the end of my 2 year warranty period, F(x) Tec agreed it was a warranty repair and I sent it to them knowing it would be forever before I got it back (I assumed they would have no parts until manufacturing was underway for the Pro1x as I would guess they ordered bulk parts at the same time to save money. I knew I wasn't technically knowledgeable enough to chase the problem down myself.

The point is, I don't think the margin correction has fai;ed.  I think you, like I, have a deeper problem that is preventing qany response from the screen. Maybe you, unlike me, can chase it down.

As for me, just curious which I will get first, my Pro1 back or my new Pro1x,  I'm fine with the wait for now.  They are small and I'm sure everyone's energy is on the Pro1x manufacturing.

[VaZso 1,998]

1 hour ago, Hook said:

The point is, I don't think the margin correction has fai;ed.  I think you, like I, have a deeper problem that is preventing qany response from the screen. Maybe you, unlike me, can chase it down.

It may happen - @agent008, maybe it worth a try to check if i2c-7 has a device address 0x14 using i2cdetect, so if touch controller replies to its address or it is simply inaccessible.

However, it needs root privileges.

It should look something like this (if querying all addresses):

~ $ sudo i2cdetect 7
Probe chips 0x03-0x77 on bus 7? (Y/n):y
     0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
00:          -- -- -- -- -- UU 09 -- -- -- -- -- --
10: -- -- -- -- UU -- -- -- -- -- -- -- -- -- -- --
20: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
30: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
40: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
50: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
60: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
70: -- -- -- -- -- -- -- --

The value at cross of row 10 and column 4 is in question.

UU means it is currently used by a driver so practically it does not test it but a driver was loaded before which should only happen if it is accessible...

[agent008 243]

Okay, before opening the phone again to check the screen connector etc I'll flash Lineage or AICP and Magisk for root then I'll run i2cdetect and report back here. Tks

[agent008 243]

@VaZso : bingo! I don't have any drivers using address 0x14 (just "--" instead of "UU" on row 0x10, column 4). So I'm now suspecting a hardware issue, either flat cable misconnected or something else. Did I understand this right? Any other testes I could carry out to confirm there is nothing at address 0x14? Thanks

Edited May 3, 2022 by agent008

[VaZso 1,998]

7 minutes ago, agent008 said:

@VaZso : bingo! I don't have any drivers using address 0x14 (just "--" instead of "UU" on row 0x10, column 4). So I'm now suspecting a hardware issue, either flat cable misconnected or something else. Did I understand this right? Any other testes I could carry out to confirm there is nothing at address 0x14? Thanks

Do you see other devices on the bus?
So in my case, also 0x08 and 0x09 exist, how about you?

Right, it may happen a simple disconnection or an unfortunate event your display has a faulty touch controller, but I²C may also stall and that is why it is interesting if you see other devices on the bus.

[agent008 243]

2 minutes ago, VaZso said:

So in my case, also 0x08 and 0x09 exist, how about you?

Yes, I do. the same as yours, 0x08 and 0x09.

[VaZso 1,998]

14 minutes ago, agent008 said:

Yes, I do. the same as yours, 0x08 and 0x09.

Then I would think of a discontinuity somewhere which may be a not properly plugged cable (between main board and display) or really a touch controller (practically display) fault.

You may also check visually if there are any problems with display cable or its connectors (like a deformed pin).
So connections, connectors, display cables and also display can be the potential cause of your problem.

[agent008 243]

I checked the connectors and flat cables. They all seem fine. Both display connectors & flat cable, and the keyboard connector & flat cable.

I checked i2c bus 6, and I also do not have the keyboard present ("--" on row 0x50, column 8). So that is why my keyboard is not working either...

[VaZso 1,998]

31 minutes ago, agent008 said:

I checked the connectors and flat cables. They all seem fine. Both display connectors & flat cable, and the keyboard connector & flat cable.

I checked i2c bus 6, and I also do not have the keyboard present ("--" on row 0x50, column 8). So that is why my keyboard is not working either...

Interesting... it sounds really strange you have more parts which are not working... do you also have other devices on I2C-6 what I wrote earlier?

Naturally I don't know Pro1's schematics but another idea may be a missing power line somewhere - so at least for touch panel, I²C line seems to be working. It uses two wires, an SDA (data) and SCL (clock) signal... device may have other pins like interrupt but basically two other pins are needed which are VCC and GND.

So if these parts are otherwise working and SCL/SDA lines are connected, then either power does not reach these ICs or GND is missing somehow.

Do all your sensors / interfaces work on your phone?

Anyway, in an unfortunate event, it may happen a thin wire of PCB may break as a cause of an accidental short-circuit somewhere. As you wrote keyboard is also not present and it is at the bottom side of the phone, I think a potential issue may be at a common part of the power supply of aw9523b and gt1x.
Such a problem may happen if power line can supply relatively high amount of current but these ICs are consuming a minimal current in usual case, thus, they don't need much power so wire width of PCB trace can be narrow.

So, theoretically a short-circuit of power line during display replacement may cause something to break at main board if there is the weakest point, be it a fuse or a very narrow PCB trace.

I did not find the working voltage of gt1x but AW9523B needs a voltage between 2.5 and 5.5V.
I think it will be somewhere between 2.5 and 3.0V (most likely below the minimal working voltage of battery).

So if this is the case and really power line is missing then it would be good to find the power wire at main PCB side and reconnect it to the right power rail.

[EskeRahn 5,471]

5 hours ago, VaZso said:

So if this is the case and really power line is missing then it would be good to find the power wire at main PCB side and reconnect it to the right power rail.

Is the pcb a plain old two sided or multi layer print? In the later case it can be hard to spot/fix, without a diagram, It might only leave a slight miscolurisation around an area where the fault is internally. But tiny lane, tiny burn, tiny spot...

[VaZso 1,998]

1 hour ago, EskeRahn said:

Is the pcb a plain old two sided or multi layer print? In the later case it can be hard to spot/fix, without a diagram, It might only leave a slight miscolurisation around an area where the fault is internally. But tiny lane, tiny burn, tiny spot...

Definitively not.
It should be a multi-layer board and a discontinuity simply can happen at an inner layer.

However, connector may be accessible (not easily but may be soldered) and an appropriate power wire could be picked but without documentation it is not easy, also BGA packages do not help at all.
Maybe one usable 3rd-party info would be a pinout of display's wiring itself - it may happen it may be reachable somehow and may be checked if appropriate power wires are present, also it may happen power wire of display part may be used to powering touch, but these are only tips from far away and without checking the board...

[raymo 179]

Out of subject, but I install the kraken apk who ask every time fingerprint confirmation at start, and (strangely for me) it works perfect every time.
Fingerprint when phone is locked works one in ten.
It could mean there's also a low level software possibility for this problem.

We definitely need a retro-engineering work on this phone 😄

[agent008 243]

11 hours ago, VaZso said:

Do all your sensors / interfaces work on your phone?

The fingerprint sensor works. I have been using the phone with a USB -C hub with a mouse connected to it, that is how I can navigate the system and how I was able to install Termux and run i2cdetect.

Every time after flashing I use the mouse to pair a bluetooth keyboard+touchpad, it's fun to use an Android device without touch, haha. 😪

Apart from the fingerprint, I have devices 0x08 (microphone?) and 0x09 (unknown device?) on bus 7, and on bus 6 I have 0x28 (NFC).

I spent 2½ days troubleshooting this and now don't have much time to look further into it until the weekend. If I have some time today I will connect the broken screen (which had a broken glass but was working, both screen and touch) and see if it still works. If it works, then I have 2 problems: new screen's touch digitiser is dead, and the keyboard isn't working. If the old screen doesn't work anymore, then I suspect one problem is the cause of both malfunctions and that might be something on the i2c bus (maybe solvable only by changing the motherboard 😪)

Edited May 4, 2022 by agent008

[VaZso 1,998]

3 hours ago, agent008 said:

If it works, then I have 2 problems: new screen's touch digitiser is dead, and the keyboard isn't working.

When has your keyboard stopped working anyway?
Did it work before replacing the screen?

3 hours ago, agent008 said:

If the old screen doesn't work anymore, then I suspect one problem is the cause of both malfunctions and that might be something on the i2c bus (maybe solvable only by changing the motherboard 😪)

If I²C was not working, other devices on the bus were also missing.
Any devices can stop the whole I²C bus easily by pulling down data or clock wires but that would also stop other devices on the bus to work.
So the bus itself (that part of the SoC) is working.

Anyway, have you used a metal tool to disconnect connectors?

Naturally, replacing the motherboard would help but even F(x)tec does not have replacement motherboards for Pro1 as far as I know.

My suspect is somehow you made a short-circuit and now power of these devices are missing, so it would be good to check that somehow and find a way to power them... but that is not something an ordinary user can do.

 

Edited May 4, 2022 by VaZso