Seagate SSHD 1 TB ST1000DX001 (FW: CC41) don´t spin.

[lastmile]

Hello everyone ...

I bought in an auction 100 (yes, one hundred) Seagate SSHD 1 TB hard drives (ST1000DX001 - FW: CC41 - P / N: 1NS162-300 - PCB: 100731495 Rev. B) with exactly the same defect in all of them: The discs don´t spin and nothing on terminal console.

Recorded data is unimportant, what I need is HDs to come back to life.

I did tests on the TVS diodes, 0 ohm resistors, everything is correct.

I put a PCB from an HD donor in some PCB pacient and they spin perfectly, no clicks or weird noises.

I have 6 same hard drives working perfectly and found some different voltages on VCore2 (0.01 V) and VCore1 (0.94 V).

My case is very similar to this: https://forum.hddguru.com/viewtopic.php?f=1&t=29046

What should I do ?

PS.: Sorry, my English is so terrible.

[Masterclass]

What should I do ?

You should take it to nearest skilled DR engineer

[E123]

You should take it to nearest skilled DR engineer

+1
ROM patch is needed to get to the terminal. PC3K can it.

[fzabkar]

@lastmile, can you tell us the voltage measurements on a working PCB for reference purposes? Also, can you at least get a startup message on terminal from a working PCB? Note that your TTL adapter will most likely need to support 1.8V rather than 3.3V. Also, the terminal will be locked, so you won't be able to execute any commands.

[E123]

Here is the same drive:
https://forum.ru-board.com/topic.cgi?fo ... rt=1440#15

[lastmile]

@lastmile, can you tell us the voltage measurements on a working PCB for reference purposes? Also, can you at least get a startup message on terminal from a working PCB? Note that your TTL adapter will most likely need to support 1.8V rather than 3.3V. Also, the terminal will be locked, so you won't be able to execute any commands.

@fzabkar, I attached an image with some voltage measurements for reference. I can access the terminal in good PCB. The console isn´t locked and can execute commands, ex.: V1, V4, V40, etc.

I think there's a short somewhere on the Bad PCB but I don't know where it is.

Attachments

PCB Seagate ST1000DX001 - Working vs Not Working

[lastmile]

@fzabkar, Any ideas ?

[E123]

Well, you don’t want to understand what we say to you ... Your NAND is broken with a probability of 99.99%. NAND needed for the work of the firmware. Firmware can not work with a broken NAND and stops its(f/w) further loading. Therefore, there is no terminal. Therefore, the motor does not work. A short circuit has nothing to do with this. It is necessary to disable the NAND initialization check. Need a patch.
And, by the way, replacing NAND by soldering to another one will not help either. The new one will also need to be reinitialized. And for this, it must pass a test at boot.

[fzabkar]

Firmware can not work with a broken NAND and stops its(f/w) further loading. Therefore, there is no terminal.

I understand that the terminal is locked in some models, so the user cannot input any commands, but shouldn't the POST at least produce an error output?

[fzabkar]

@fzabkar, Any ideas ?

I am wondering about the missing supply voltages.

AIUI, the motor controller generates most of the onboard supplies (Vcore1, Vcore2, Vio and Vneg), and there is an additional linear 3.3V regulator adjacent to the NAND (did you check this?). The MCU requires a Vcore and Vio supply in order to start POST-ing, after (or during?) which time it commands the motor controller to switch on the remaining supplies. In your example the second Vcore is either missing or switched off, and the Vneg preamp supply (-3V) is also missing. This suggests that either the motor controller is faulty (unlikely), or that the MCU has switched off, or not enabled, these two voltages, probably because of a POST error.

You say that you have access to terminal on your working HDDs. Would you be willing to experiment with a good drive, with the understanding that you may brick it? If so, then short pin #1 of the 3.3V regulator IC to ground during power-on. This will disable the 3.3V supply to the NAND IC, thereby simulating a NAND failure. What do you see in terminal after you do this?

[E123]

fzabkar, I specifically gave a link. There is the same disk with the same version of firmware. As for the lack of error messages - this is to the vendor.

[fzabkar]

If so, then short pin #1 of the 3.3V regulator IC to ground during power-on. This will disable the 3.3V supply to the NAND IC, thereby simulating a NAND failure. What do you see in terminal after you do this?

Also, which of the supplies are running and which are disabled, if any?

[pepe]

but shouldn't the POST at least produce an error output?

Short answer: no, it shouldn't.
somewhat less short answer: there is no point in asking what it should and should not do. It should? Who says it should? Yes, it would be a lot more informative, but it should not just because me or you say so.
oops, it got a bit phylosophical
SSHDs are prone to NAND problems and it can result in various outputs or perfectly no output at all. As E123 has already said, 99.99% it has nothing to do with power supplies.
Not to mention, he has 100+ drives with the same issue. It has very low chance they are all having electrical problems. There are a few problems that cannot be diagnosed using a multimeter

pepe

[fzabkar]

but shouldn't the POST at least produce an error output?

Short answer: no, it shouldn't.
somewhat less short answer: there is no point in asking what it should and should not do. It should? Who says it should? Yes, it would be a lot more informative, but it should not just because me or you say so.
oops, it got a bit phylosophical :)
SSHDs are prone to NAND problems and it can result in various outputs or perfectly no output at all.

When I said "shouldn't", I didn't mean to say that such behaviour should be a design requirement (although it really should be, IMO). I was merely asking whether an error message is normal for this particular design whenever a NAND failure is detected. That's why I asked the OP to simulate such a failure on a working drive.

As E123 has already said, 99.99% it has nothing to do with power supplies.
Not to mention, he has 100+ drives with the same issue. It has very low chance they are all having electrical problems. There are a few problems that cannot be diagnosed using a multimeter :(

Did you not read this?

In your example the second Vcore is either missing or switched off, and the Vneg preamp supply (-3V) is also missing. This suggests that either the motor controller is faulty (unlikely), or that the MCU has switched off, or not enabled, these two voltages, probably because of a POST error."

It could also be that the MCU is faulty, but unlikely. Let's see what the OP's multimeter will tell us when he disables the NAND.

[pepe]

No, i did not read all the posts entirely, it is beyond my capabilities. And perfectly unneccessary.
It does not start with 'disabled' NAND. By design. Have you ever heard something like that? I haven't, but that means nothing, i am not on the forums 24/7.
But, instead, i examined how the firmware works on Sg SSHDs and have somewhat deeper insight than the lady at the grocery at the corner.
Probably the same can be said about E123.
Anybody else did research about the fw amongst those who posted in this topic?

pepe

[pepe]

And of course i might be wrong, i am just telling it has very low chance it is not NAND related.
And a few orders of magnitude lower chance that disabling nand supply will produce any output on the diag port.

[lastmile]

@fzabkar, Any ideas ?

I am wondering about the missing supply voltages.

AIUI, the motor controller generates most of the onboard supplies (Vcore1, Vcore2, Vio and Vneg), and there is an additional linear 3.3V regulator adjacent to the NAND (did you check this?). The MCU requires a Vcore and Vio supply in order to start POST-ing, after (or during?) which time it commands the motor controller to switch on the remaining supplies. In your example the second Vcore is either missing or switched off, and the Vneg preamp supply (-3V) is also missing. This suggests that either the motor controller is faulty (unlikely), or that the MCU has switched off, or not enabled, these two voltages, probably because of a POST error.

You say that you have access to terminal on your working HDDs. Would you be willing to experiment with a good drive, with the understanding that you may brick it? If so, then short pin #1 of the 3.3V regulator IC to ground during power-on. This will disable the 3.3V supply to the NAND IC, thereby simulating a NAND failure. What do you see in terminal after you do this?

@fzakbar, I didn´t find the datasheet of Regulator IC near the NAND Flash to short pin#1 to ground. I attached an image of the IC. It seems M441 L112 33.

Additional information: In good PCB, after a few minutes, several components become warm. In defective PCB, only MCU IC get warm, even after 40 mins.

Attachments

IC Nand Flash

[fzabkar]

LDL112PV33R/LDL112PU33R, STMicroelectronics, 1.2A, 3.3V, LDO Voltage Regulator, marking L112 + 33, DFN6:
https://www.st.com/resource/en/datasheet/ldl112.pdf

Datasheet database:
http://www.hddoracle.com/viewtopic.php?f=100&t=1577

[fzabkar]

fzabkar, I specifically gave a link. There is the same disk with the same version of firmware. As for the lack of error messages - this is to the vendor.

Thanks. I did go to that URL, but the "translation" was woeful. Why is it that Google does a good job of translating Russian web sites, but a lousy job when translating Russian forums? Is it because the language in the forums is slang, SMS-ese, or just poorly constructed? Whatever the reason, the amount of useful information is about 100 times what you find in English speaking forums, but the Russian Internet may as well be in a parallel universe if Google's translator is the only portal for non-Russians.

Here is the solution, but the "ROM patch" URL has expired:

https://forum.ru-board.com/topic.cgi?forum=84&topic=5386&start=1480

from the point of view of the complex all three dumps of ROM you have normal.
And in these discs it is not done, without a sense of initialize the cache on the
donor.
The probability that all the biases will match is very small.
Simply because the defects in the chips can be different and the chip banks are differently organized.
Therefore, a patch of ROM is made to get into the terminal, the patient is made inite cache, and voila.
Try
it: https://yadi.sk/d/CXsRGU7acDb9Vg
it's a patient's ROM patch.
Initialize the cache underneath it and then fill the original.
The truth is not sure that without the complex will help and all commands will be available, after the ROM patch should be - unlock tech mode in the lake. to make all teams available.

Rechecked
on PCB donor and ROM patient with jar the patient received an initiation (but the dump ruled

the one

that is higher for unlocking) added:
Gone!!! Works have access to all the information !!!

They've been doing this lately, and they're not alone. )
From his own coolness in the goiter breaths. A common disease of all the big offices.
Did anyone ask the terminal to block if the SSD cache doesn't work?
There's a team like that, but you're going to use it.
And then the hards went, but to drink with grief and generally do not deal with them, there and the complex will not help. A bunch of them are growing, but you can't do anything at all.

Quote:
but only a few of those who have such an opportunity would be able to lay out patched rum.

Well first I immediately removed the file as it downloaded, and then others will start with a fool in their casting, and the patch is individual for each hard.
and to unearth that there patched little is real, yet squeezed into the ROM.
It's got to be disased by the ROM. Who does it and will find it. It's not going to help the rest of us. And the disc is very rare. Only a true fan will buy himself a SSHD.
It's certainly an invention of morons for the same. )

Here is my "dictionary" of pancakes, screws, rabbits, and brooms:
http://www.hddoracle.com/viewtopic.php?f=46&t=2730

[lastmile]

If so, then short pin #1 of the 3.3V regulator IC to ground during power-on. This will disable the 3.3V supply to the NAND IC, thereby simulating a NAND failure. What do you see in terminal after you do this?

Also, which of the supplies are running and which are disabled, if any?

@fzabkar Wow, how much interesting information I am learning from you. Thank you guys!

I made what you requested by shorting pin #1 to ground during power-on in good HD. Here is the terminal results:

Bridge PLL Enable Failed
Bridge PLL Enable Failed
Bridge PLL Enable Failed
ERROR: Flash Led 9C08 - ALF Hardware Initialization Process Failed
Flash HW Init Attempts = 0004

Clump Parametrics:
Clump = F3E3
DataType = 00E0
ModeFlag = 00
EraseCount = 306A5876
EraseFailCount = 06
ProgramFailCount = 07
WeakReadCount = 14
UncorrectableECCError = 00

---SOC PSM Command History---
Cmd#/PsmCmd Clump Offset Length LbaMid LbaLow LbaBits39To32 Options Status DestClump DestOffset
000E F3E3 9970 A650 E661 C3CF 1299 4BB5 0080 C207 7085

010E F3E3 9970 A650 E661 C3CF 1299 4BB5 3051 C207 7085

1BE2 1C1D 4872 2224 9C13 2058 F54E 0CDC FFD6 11AC 9BE6

75C0 1BED 7000 A2F7 292E BA6E 12CA C66E F908 D987 C807

3F6E 2205 D6AE 3C17 5F55 CD7C EF8A 11AF 4A57 B4E8 3BC3

Flash Statistics: Could not be printed

ALFRecorder:
End of ALF Recorder Trace

And hangs terminal. No inputs or more outputs. The motor won´t spin.


After that, I removed the PCB from HD and made the test again. Here the results:

F3 T>
TCC:FFF8

Bridge PLL Enable Failed
Bridge PLL Enable Failed
Bridge PLL Enable Failed
Bridge PLL Enable Failed
ERROR: Flash Led 9C08 - ALF Hardware Initialization Process Failed
Flash HW Init Attempts = 0004

Clump Parametrics:
Clump = F3E3
DataType = 00E0
ModeFlag = 00
EraseCount = 306A5876
EraseFailCount = 06
ProgramFailCount = 07
WeakReadCount = 14
UncorrectableECCError = 00

---SOC PSM Command History---
Cmd#/PsmCmd Clump Offset Length LbaMid LbaLow LbaBits39To32 Options Status DestClump DestOffset
000E F3E3 9970 A650 C661 C3CF 1099 4BB5 0080 C207 7085

010E F3E3 9970 A650 C661 C3CF 1099 4BB5 3051 C207 7085

1BE2 1C1D 5872 2224 9C13 2058 F54E 0CDC FDD6 11AC 9BE6

55D0 1BED 7000 A2F7 292E BA6E 12CA C66E F908 D987 C807

3F6E 2205 D6AE 3C17 5F55 CD7C EF8E 11AB 4A57 B4E8 3FC3

Flash Statistics: Could not be printed

ALFRecorder:
End of ALF Recorder Trace


I measured the voltages on the PCB and the result is identical to the defective board: VCore2 = 0,0 V and VCore1=0,92 V.