ST1000DX002-2DV162 – PCB swap + ROM transfer, still shows do

[magnetepazzo]

Hello everyone,

I am working on a Seagate SSHD ST1000DX002-2DV162 with a defective PCB.

I replaced the PCB with a compatible donor board and successfully transferred the original ROM from the patient drive to the donor PCB. The drive powers on, but it still reports the donor serial number and does not initialize correctly. I also see repeated servo reset behavior, so clearly the adaptive parameters are not being applied correctly.

From what I understand, this model uses an internal NAND flash cache, and it seems likely that the drive is still reading identification data and adaptives from the NAND belonging to the donor PCB instead of rebuilding them from the patient ROM and SA.

I tried to clear or initialize the NAND cache via terminal using the command:F3 T>OI
but the drive returns “Invalid Diag Cmd Parameter”, so this command is apparently not supported on this firmware.

At this point, I suspect that the NAND cache needs to be cleared or reset (possibly via enhanced secure erase or a vendor-specific command), but I would appreciate confirmation from anyone who has worked on this exact model.

Has anyone successfully resolved this issue on ST1000DX002-2DV162?
What is the correct procedure to clear the NAND cache or force the drive to rebuild adaptives after a PCB + ROM swap?

Thank you very much in advance for any help or guidance.

[melvin]

Did you try to clear NAND on the donor drive or on the patient? It should be cleaned on the donor.
There's an article about SSHD on ACE's blog.

[SWM]

F3 T>/OI
Or
F3 T>/O
F3 O>I

[magnetepazzo]

Yes, I tried `/o` and then `I` directly on the donor drive and it reported a successful erase. After that, I wrote the patient ROM and moved the PCB onto the patient drive, but it still reads the old serial number and, obviously, it does not align correctly. If I read sector 0, I can also see the partition that was present on the donor disk.

[SWM]

After the successful erase, was the donor immediately powered off? And the board removed?
I have two options:
1. Either the donor was powered on again or switched to ATA from terminal.
2. When the patient was powered on before the NAND erase, it managed to write the donor's cache to the patient's drive.
The second option is unlikely, but who knows what commands were sent to it?.. Maybe the DR gurus can help.

[pepe]

a rom patch is also required imo...

[magnetepazzo]

a rom patch is also required imo...

Ho Pepe,
in MRT there are two different ROM patch types. At first, I used them on the faulty PCB with the damaged NAND and I was able to access the SA successfully. Do you think applying only one of them would be enough?

[magnetepazzo]

After the successful erase, was the donor immediately powered off? And the board removed?
I have two options:
1. Either the donor was powered on again or switched to ATA from terminal.
2. When the patient was powered on before the NAND erase, it managed to write the donor's cache to the patient's drive.
The second option is unlikely, but who knows what commands were sent to it?.. Maybe the DR gurus can help.

If I understand correctly how the PCB works, it keeps the most frequently used data in the 8 GB area and does not write it to the magnetic platters. Writing cache data to a slower medium wouldn’t really make sense.

I ran the command from the terminal and then powered off the donor, waiting just long enough to see that it had completed. After that, I loaded the ROM via boot code with the PCB not connected to the donor, and then I mounted the donor PCB on the patient drive.

When the PCB is not connected to any drive, I can read the serial number from the ROM via boot code, which matches the patient’s serial. But as soon as I mount the PCB on either the patient or the donor drive, I always read the donor’s serial number and its capacity.

[SWM]

These are all the ways in which data from a donor could end up on a patient.
My opinion is that the cache works both ways. Otherwise, why would NAND die?

[SWM]

As far as I understand, one of the ROM patches is designed specifically to wake up the HDD from a coma (LED XX) and clear the NAND directly on the patient (with a dead donor).

[pepe]

pcb boots from nand by default, which contains the donor bootfw and adaptives...

[magnetepazzo]

I applied the patch to the patient’s ROM and installed the PCB on the patient. Then I cleared it with O–I and reinstalled the patient’s original ROM. Now I can read the serial number correctly, but I’m unable to read data from the drive.

Now have cleared ALT because i don't read any sector and I have tryed to rebuild with ALT but see error .
Spin Up
Spin issued

6Gb Max Speed lowered to 3Gb for Intel bug
Send Status: COMRESET seen
TCC:001A

SNOOOZZZZZZEEE Retry
EDAC Value = 0000000000071BA9
SNOOOZZZZZZEEE Retry
EDAC Value = 0000000000071BA9
ABORT!
OpCode = 00000024
SenseCode = 03310000
ReadyStatusFlags = 0002A1A1

ABORT!
OpCode = 00000024
SenseCode = 03310000
ReadyStatusFlags = 0002A1A1

Unrecoverable Read Verify Error Reported
TCC:001A

TCC:001A

TCC:001A

TCC:001A

TCC:001A

ABORT!
OpCode = 00000024
SenseCode = 03310000
ReadyStatusFlags = 0002A1A1

ABORT!
OpCode = 00000024
SenseCode = 03310000
ReadyStatusFlags = 0002A1A1

Unrecoverable Read Verify Error Reported
TCC:001A

TCC:001A

TCC:001A

TCC:001A

TCC:001A

TCC:001A

TCC:001A

[SWM]

Is it possible to see the log on Ctrl X after restart (and "Unrecoverable Read Verify Error Reported" message)?

[magnetepazzo]

Hello everyone, the issue was resolved remotely by the user FWsupport from this forum, whom I thank immensely.