Lenovo NM-D562 v2.0 main EC programming help needed

[questore]

And the crazy part is that a binary diff of the dump(s) reveals several differences, I attached three of them as an archive.

I have found a full pack from a working system. You can try this EC firmware while i wait for the programmer hardware

[davidebeatrici]

Thenk you for sharing. I think I order XZZ IT8390 and give it a try.

The tool is amazing, it just works and the price is okay.

I originally bought it to experiment on a cheap Chuwi laptop, but the EC was partially shorted and unreadable in that case.

Maybe you also have a BIOS problem too.
Can you read the BIOS with a direct clip or is yours the same as mine, that i have to desolder the IC to read and write?

I never encountered a problem reading an SPI with the direct clip on this board. Are you sure yours is working properly?

This one I have to figure out myself also since my board have came to me after a repair attempt and all the BIOS info were empty (not even sure this is the correct one)

Interesting, also in my case the DMI region was empty.

I posted the dumps here: https://www.badcaps.net/forum/troubl...46#post3717446

[davidebeatrici]

I have found a full pack from a working system. You can try this EC firmware while i wait for the programmer hardware

No luck unfortunately.

[m1ch43lzm]

You can try BIOS from the last page of https://www.badcaps.net/forum/troubl...request/page11
Either the one by hoaca388 or qayyum786, both have DMI inserted and are latest version GKCN65WW extracted from Lenovo website
If you want EC from that version, open BIOS with hex editor and select from 21000h-60FFFh, copy to new file, this is 256KB, you can try trimming the end if you want 192KB

It appears for both of you, that the laptops were messed by other repair shops before and the first thing they tried was BIOS

Or it may be a cracked solder ball in one corner of the CPU, right where the black glue is..., check this video (in Spanish), another Lenovo board NM-D563 but also AMD CPU, skip to 1h:08
https://youtu.be/OBtX26N6tuw?t=4099
At 1h17m that guy is probing in diode mode on the oscillator pins, and when he heats up that corner the multimeter gives a diode reading (was OL before), not a permanent fix, as that connection broke again at 1h30m

[davidebeatrici]

I flashed the BIOS from qayyum786 and the EC firmware I extracted from it, no change whatsoever. I had already tried several dumps before.

As for the oscillator: according to my multimeter there is no continuity to ground on any of the two pins, on both YC1 and YC2.
However, since one of the 4 pads under the CPU is near the border (AY1) I managed to confirm continuity between it and one of the two pins on the 32K crystal.
My oscilloscope with AC coupling shows some sort of waveform on both pins of both oscillators. My multimeter reads 0V DC on all of them.

Keep in mind I have already replaced the CPU:

I think I have reached your situation now.

After removing the CPU (and cleaning all pads) CPU_RST_R_N and CPU_PWROK became high (1.8V) and VCORE increased from ~0.8V to ~1.1V.
I then installed a 100-000000295-40 (ES 5800H) and it gets warm as opposed to the original CPU which stayed completely cold.

Unfortunately still no POST whatsoever and no activity on SPI_CS for BIOS and VBIOS.

Four pads were ripped off thanks to the black glue, but fortunately they're all NC (AU2, AV3, AY7 and AW7).

[questore]

No luck unfortunately.

what is CPU_PWROK voltage on yours?

[davidebeatrici]

1.8V, same as CPU_RST_R_N.

[davidebeatrici]

I'm starting to think the GPU has an internal short: after leaving the motherboard on for two minutes or so to take measurements on the other side of the board I noticed it became very hot to the touch.
I immediately investigated with my thermal camera and noticed the GPU reached the insane temperature of 144°!



I decided to perform a reflow and remove the power coils at the same time:



Still no POST and no CS activity on the BIOS chip. I suspect removing power to the GPU is not enough, especially if the platform expects it to come up before the CPU.

Restoring power to the GPU results in it still heating up without doing anything (as far as I know):



At this point I don't know how to proceed. I have already ordered a replacement EC chip just because it's cheap and I want to keep a spare around, but I'm pretty sure it's not the issue here.

Any input/opinion is well appreciated. Thanks to m1ch43lzm we now know how to extract the EC firmware, making sure that's definitely not corrupted. The bytes that change after the flash are probably internal flags that are set on first boot and afterwards.

[questore]

I'm starting to think the GPU has an internal short: after leaving the motherboard on for two minutes or so to take measurements on the other side of the board I noticed it became very hot to the touch.
I immediately investigated with my thermal camera and noticed the GPU reached the insane temperature of 144°!



I decided to perform a reflow and remove the power coils at the same time:



Still no POST and no CS activity on the BIOS chip. I suspect removing power to the GPU is not enough, especially if the platform expects it to come up before the CPU.

Restoring power to the GPU results in it still heating up without doing anything (as far as I know):



At this point I don't know how to proceed. I have already ordered a replacement EC chip just because it's cheap and I want to keep a spare around, but I'm pretty sure it's not the issue here.

Any input/opinion is well appreciated. Thanks to m1ch43lzm we now know how to extract the EC firmware, making sure that's definitely not corrupted. The bytes that change after the flash are probably internal flags that are set on first boot and afterwards.

What is your GPU resistance on coils? (Mine are 0.31 ohms )
What is strange in yours that in 110 Degrees celsius The board should shut down i think.

[questore]

You can try BIOS from the last page of https://www.badcaps.net/forum/troubl...request/page11
Either the one by hoaca388 or qayyum786, both have DMI inserted and are latest version GKCN65WW extracted from Lenovo website
If you want EC from that version, open BIOS with hex editor and select from 21000h-60FFFh, copy to new file, this is 256KB, you can try trimming the end if you want 192KB

It appears for both of you, that the laptops were messed by other repair shops before and the first thing they tried was BIOS

Or it may be a cracked solder ball in one corner of the CPU, right where the black glue is..., check this video (in Spanish), another Lenovo board NM-D563 but also AMD CPU, skip to 1h:08
https://youtu.be/OBtX26N6tuw?t=4099
At 1h17m that guy is probing in diode mode on the oscillator pins, and when he heats up that corner the multimeter gives a diode reading (was OL before), not a permanent fix, as that connection broke again at 1h30m

On mine board I have voltage on both sides of the 32Khz and 48Mhz crystal as well.
But have noticed that CPU has connection on the a SPI_CLK. It could be pulled low if the EC firmware or communication is not correct on my case. Maybe this is why sometimes i can connect to BIOS onboard and sometimes i dont (cracked CPU connection balls under)
(Im still waiting for the programmer to arrive. 3/5 preprogrammed EC on my case is arrives with bad programmed firmware unfortunatelly)

[davidebeatrici]

What is your GPU resistance on coils? (Mine are 0.31 ohms )
What is strange in yours that in 110 Degrees celsius The board should shut down i think.

0.00Ω! I don't think it's a problem of the multimeter because it correctly measures ~3.4Ω on the CPU coils when it's cold.

Surprisingly, injecting 0.8V right after the coils results in a power consumption of ~1.3A.
I expected much more if it's really a short, but the behavior makes sense because the GPU doesn't get extremely hot right away.
Also, the current consumption slowly increases, but I didn't wait to see how high it would go.

On mine board I have voltage on both sides of the 32Khz and 48Mhz crystal as well.
But have noticed that CPU has connection on the a SPI_CLK. It could be pulled low if the EC firmware or communication is not correct on my case. Maybe this is why sometimes i can connect to BIOS onboard and sometimes i dont (cracked CPU connection balls under)
(Im still waiting for the programmer to arrive. 3/5 preprogrammed EC on my case is arrives with bad programmed firmware unfortunatelly)

Fingers crossed it's not a CPU problem because replacing it on this board without ripping pads is very hard.

[questore]

0.00Ω! I don't think it's a problem of the multimeter because it correctly measures ~3.4Ω on the CPU coils when it's cold.

Surprisingly, injecting 0.8V right after the coils results in a power consumption of ~1.3A.
I expected much more if it's really a short, but the behavior makes sense because the GPU doesn't get extremely hot right away.
Also, the current consumption slowly increases, but I didn't wait to see how high it would go.

This makes sense if GPU shorted. I have seen many of them going to afterlife like this.

Fingers crossed it's not a CPU problem because replacing it on this board without ripping pads is very hard.

Yeah lenovo makes it hard to repair.
I dont have the hardware to do BGA-s, except EC and RAM reball. I only remove them and repair pads but after that it needs and expert with proper infra heater to not damage the new chip when soldered on.

[questore]

Furthermore I have measured some things around the very hot mosfet in the GPU line.
Turns out the mosfet had a whole in it. Fused itself to the board and it was very hard to remove.
After removing, the GPU coils not getting hot at all. only GPU core is rising slowly until about 60celsius degree, but still NO POST.
SPI_CLK is still not rising... 0V
waitng for programmer to verify the firmware.

[davidebeatrici]

This makes sense if GPU shorted. I have seen many of them going to afterlife like this.

Do you perhaps know if this platform can boot without the GPU and, if yes, how I can disable it?

Yeah lenovo makes it hard to repair.
I dont have the hardware to do BGA-s, except EC and RAM reball. I only remove them and repair pads but after that it needs and expert with proper infra heater to not damage the new chip when soldered on.

I bought a SoFix S-BGN2 specifically for this board!

Furthermore I have measured some things around the very hot mosfet in the GPU line.
Turns out the mosfet had a whole in it. Fused itself to the board and it was very hard to remove.
After removing, the GPU coils not getting hot at all. only GPU core is rising slowly until about 60celsius degree, but still NO POST.
SPI_CLK is still not rising... 0V
waitng for programmer to verify the firmware.

If the chip stays around 60° I would say that's fine. Also my GPU coils are not getting hot at all, but the GPU reaches the insane temperature that can be seen in the thermal picture I sent above.
Core voltage stuck at 0.76V.

The CPU coils, instead, reach 70° with the chip only being warm.
Core voltage stuck at 1.1V.

I hope it's just an EC firmware problem in your case.

[davidebeatrici]

I checked the corresponding oscillator (YG1) with my oscilloscope and saw a somewhat sinusoidal wave:



The two oscillators for the CPU, instead, show very faint activity, probably just interference at this point.

I decided to remove the GPU anyway, but no changes in the behavior. Quite a few ripped pads (once again, thanks to the black glue), all NC except two that I'll have to repair somehow:



I was thinking: is there a chance I was sold a defective CPU?

[questore]

I
I was thinking: is there a chance I was sold a defective CPU?

- Do you have proper oscillation on YC1 and YC2 ? (32khz and 48Mhz)
- Did you checked the ram slots with ram tester if any connections are faulty?
- You could also check if there is any activity or signs of error when there is NO RAM installed, (mine are not showing missing ram error that is why i suspect firmware issue, but I dont know how this particalar modell shows missing ram)

[davidebeatrici]

No oscillation on neither of the two. Unfortunately I don't have a RAM tester, but I can buy one.

The behavior is the same with and without RAM.

[questore]

No oscillation on neither of the two. Unfortunately I don't have a RAM tester, but I can buy one.

The behavior is the same with and without RAM.

Strange... not likely that both crystals are bad...
What voltages do you have on YC1 and YC2 sides?

[davidebeatrici]

Mostly 0V, on all of them. It varies slightly probably due to interference.

[questore]

Mostly 0V, on all of them. It varies slightly probably due to interference.

0v not OL ? then My oppinion is that either some joint under CPU is not connected properly or you have a bad CPU. There should be like 0.6-0.9Voltage
I'm just poking here but on intel chipsets i think RSMRST must be present for CPU to turn on the oscillation.
So hard to guesss because there is no startup sequence for this modell