06-07-2018 01:43 PM
I have a Thinkpad Yoga 260, which suddenly stopped working.
The laptop would fail to post, giving me 5 short beeps, pause, 5 short beeps on startup without keyboard backlight, or anything on the screen, while the fan spins, the error code suggesting it's a motherboard failure.
The laptop's warranty has expired, and being an engineer, I took it apart and tried all possible power/reset/clear cmos cycles posted here, with no success.
As I remember updating my bios rather recently, and reading that several users managed to bring their laptop back with a reflash, I tried reflashing the desoldered chip with a CH341a programmer, and trying to clone the bios chip with new 25q128fvsq chips I have bought.
My issue when flashing the bios is, that I am unable to verify it, regardless of what I write to it, and regardless of chip.
I've tried cloning the readout of the original untempered chip, but whenever I write the image to a chip, reading it back I get a mismatch. Furthermore, I've tried flashing the latest bios by extracting and editing the .fl1 image by removing the non-bios hex values from it and pasting it in place of the bios region of the orignal chip readout (as shown here).
The image I've made matched in length and structure to the readout, yet, after flashing I read out a different image than the one I've written.
When trying the new flashes, soldered in, I have the following results:
new chip with old image: keyboard backlight flashes once, fan spins up then shuts down without any other response.
old chip with new (modified) image: same as above happened the first time and won't boot afterwards.
I have also tried removing all power sources (cmos battery too), pressed power button, pressed cmos clear button, but nothing changes. Moreover, I'm certain I've soldered the chip correctly and didn't cause any shorts.
Regarding the bios, I don't understand why I'm unable to verify the images, I think I remember reading that the descriptor and ME regions are write protected, would that be the case, and possibly the cause of not being able to verify it?
Also, when trying to clear the chip, each zero-fill cycle seems to remove more data from it, but I'm not able to get any chip to be completely filled with zeroes (FF).
Do you have an other idea of what else I could try to repair the board?
06-08-2018 01:36 PM
You are right that there are security segments that are not in the fl1 files and you cannot simply copy. That's why you cannot verify, but the serial flash chips are fairly "robust" and the BIOS being sick is probably not the cause of your problem. Patterns starting with 5 beeps are not common and something deep in my memory is telling me that they indicate a CPU error. I always assume the CPU is the part that is least likely to fail, but sometimes it does. I dealt with a failure in the past month. Mine was no big deal because it had a socketed processor. Some of the y-260 machines have soldered RAM (those with WWAN) and some do not. If you have socketed RAM, try another stick. RAM errors, especially those in the low address ranges, can cause any sort of error.
06-08-2018 02:52 PM
Thanks for the feedback!
In the meantime I've found out that the write-protection is only in place when software flashing (not in my case), and that my verification issues were caused by the flashing software. Now I managed to write and verify the original readout and the image I've made with the latest bios, and will be trying them out over the weekend.
As for the RAM, I have a socketed motherboard, and tried a different stick at the very beginning of the troubleshooting, but it didn't fix it.
I've also found out that the service guys who went over the laptop, checked power rails and they were supposedly in order.
If all else fails, I will be reballing the CPU, but I was hoping someone has an idea of what I could still try before the final option.
06-08-2018 03:17 PM
Reballing is up to you, but I always think the CPU is the last thing to suspect. In my case, CPU seemed possible and I could try it easily, since it was socketed.