Thinkpad X280 working fine but not negotiatiating PD (stuck at 5v)->no battery charge/power-throttled performance

[ErParucca]

you'll notice that the SPI flash chip is "worn out" if it fails to verify when programming
Run the Thunderbolt firmware update afterwards if it works (negotiates 20V)

https://support.lenovo.com/us/en/sol...dates-thinkpad

I had various verification failures but I don't know if that was because the clip moved or the chip failed. What I can say is that despite a few of these failures, I have been able to successfully write and verify the EEPROM. I already ordered a bunch of replacement ones so despite the effort of desoldering/resoldering, it costs nothing to try, hoping I won't phisically break anything in the process

Any way to verify if the current EEPROM is working or not?
mon2 thanks for the caps tip but should they be missing or defective, I would have no way to replace them

[m1ch43lzm]

Any way to verify if the current EEPROM is working or not?

Read it again, and compare its contents with hex editor with the file you flashed onto it, or a file comparison tool, I use WinMerge for that
If you see random bits being changed in some random places, likely it's faulty

Also check BIOS setup if Thunderbolt BIOS assist mode is enabled; disable it
Some people claim that after writing zeros to the chip they managed to update the firmware, make a file filled with 00 with hex editor, length 0x100000 and program that to the chip

[ErParucca]

thx, the EEPROM works fine (erase, read, write, verify) with the CH341a, I just wanted to know if with that I can assume it works fine also when called by the mobo.

[ErParucca]

forgot to be more specific: sometime the verification fails. I have no way of telling if it's because contact is lost (clip moving, etc.) or if the chip is failing. Rephrasing my question: if the chip is broken, does that imply that not a single complete write-and-then-verify cycle should succeed and by corollary that if it succeeds then it's useless to replace the chip? Thx

[ErParucca]

m1ch43lzm speaking of:

Run the Thunderbolt firmware update afterwards if it works (negotiates 20V)

Do you know which steps are supposed to fix the 20v negotiation? Would:

1. Open laptop
2. Remove battery
3. Disconnect CMOS battery
4. Push reset button
5. Flash the Winbond with padded tbt.bin
6. Reconnect battery
7. Power On and wait for the CMOS/Config error
8. Power off
9. Connect charger

be enough for the 20v negotiation to take place? Because that's what I've done after multiple verification of the EEPROM's content and that doesn't work so I wonder if the process is OK or if something else (for ex. TB software in windows) has to be done.

Thanks!

[m1ch43lzm]

https://gitlab.com/MobileAZN/lenovo-...firmware-fixes
https://libreboot.org/docs/install/t...ernal-flashing

Those pages recommend flashing a "null" (1MB filled with zeros) file to the chip first, power on the laptop (with the "null" file programmed), wait until it boots, power off, then flash the "real" tbt.bin (padded)
Use the X280 specific tbt.bin of course

[ErParucca]

same outcome

[ErParucca]

do you have any opinion on the software versions mentioned in this reddit post ? For what I know that shouldn't be relevant at least for the 20v negotiation as at this point the OS hasn't even started loading but I don't know if the drivers/software may somehow "initialize" the hardware. And that's why I am trying to understand what the process is between PD/thunderbolt/TB firmware: mon2 I think that we can skip the (still helpful) part about the schematics and TB controller being detected by the OS as the PD/20v negotiation should be indipendent from that, right?

PS: there was a sale and I bought an i5-8550U 16GB motherboard but I would really like to fix this one: X280 motherboard with 16GB are rare and still very much useful.

Thanks!

[m1ch43lzm]

That's done at a firmware level, if the Thunderbolt SPI flash is corrupted then:
1. It won't negotiate 20V from the charger, either won't charge, or charge very slowly when off, limited to 5V, max 15W, the charger IC (BQ25700A) can boost 5V to the appropriate battery voltage
2. Thunderbolt controller won't show up in Device Manager, as it needs to be initialized by firmware

The software versions mentioned on that reddit post are only relevant for updating to prevent the problem from happening, but as it already happened, you need to flash the Thunderbolt SPI ROM manually

As that chip is shared between the Thunderbolt controller and USB-C PD controller, it affects both

https://www.badcaps.net/forum/troubl...33#post3273033

I correct solder joints and measured resistance to ground. It was same as other board with working 20V.
i dumped from OK board U15 chip and just reflash it to old board. Verification was OK, but board was still not working.
I tried to read dump to compare and it give errors during verify.
After soldering new W25Q80BV all went well and i have 20V 😀

This post (on a T480) suggests to replace the SPI flash chip with a new one anyway

[ErParucca]

That's done at a firmware level, if the Thunderbolt SPI flash is corrupted then:
you need to flash the Thunderbolt SPI ROM manually
https://www.badcaps.net/forum/troubl...33#post3273033

This post (on a T480) suggests to replace the SPI flash chip with a new one anyway

I've done it multiple times, trying multiple versions (various releases of TBT.bin and copying the content from another working X280). The flash operation goes well: after getting a new clip, verification is always fine even after disconnecting and restarting so the problem is not the same of rbvndonh who experienced verification errors.

I still can't find an answer to my specific question: at what time does the thunderbolt controller and the charger IC read the firmware? Given the Winbond chip behaves as expected to flash, how to diagnose the rest? I have a few faulty motherboard and use as donors but I need to know what exactly to replace in case of HW failure as I can't do it myself and that's not for free.

Thx

[m1ch43lzm]

Did you solder a new W25Q80DV/JV? Here's an experiment you can try:
After flashing (and verifying) the chip, unplug it from your programmer, wait a bit, read the chip again and compare the contents with the file you flashed using a hex editor/hex comparison software
If the contents are different, that means the chip is worn out, it can't "remember" the contents properly

I still can't find an answer to my specific question: at what time does the thunderbolt controller and the charger IC read the firmware? Given the Winbond chip behaves as expected to flash, how to diagnose the rest? I have a few faulty motherboard and use as donors but I need to know what exactly to replace in case of HW failure as I can't do it myself and that's not for free.

Thx

When the board first receives power and it's on S5 (power off) state, the USB PD (power delivery) controller U192 TPS65988 reads the firmware/configuration from U184, Thunderbolt (U182) doesn't matter at this point I think
Also check if you have 3V at pin 8 of U184, there's a diode D420 there, measure voltage on both sides of the diode

[ErParucca]

I mentioned earlier that I already made multiple verification tests all with consistent positive results.

Do you know which steps are supposed to fix the 20v negotiation? Would:

1. Open laptop
2. Remove battery
3. Disconnect CMOS battery
4. Push reset button
5. Flash the Winbond with padded tbt.bin
6. Reconnect battery
7. Power On and wait for the CMOS/Config error
8. Power off
9. Connect charger

be enough for the 20v negotiation to take place? Because that's what I've done after multiple verification of the EEPROM's content and that doesn't work so I wonder if the process is OK or if something else (for ex. TB software in windows) has to be done.

I've done it multiple times, trying multiple versions (various releases of TBT.bin and copying the content from another working X280). The flash operation goes well: after getting a new clip, verification is always fine even after disconnecting and restarting so the problem is not the same of rbvndonh who experienced verification errors.

I know we are on badcaps but unfortunately I am not very skilled/have the right tools to safely replace the EPROM/test voltages without risking of incurring on (further?) damages:

I need to know what exactly to replace in case of HW failure as I can't do it myself and that's not for free