Seagate Desktop SSHD bad PCB

[pcn]

Hi all,

I hope someone can give me some hints on this one: I got a Seagate hybrid harddrive that does not spin anymore (Seagate Desktop SSHD 1 Tbyte) and does not give any terminal output. The HDD comes from a Desktop-PC.

Things I tried so far:
- The PCB looks fine, no burnt components as you can see in the images attached.
- I tested the TVS diodes (marked red) and they seem to be ok (diode test: no beep).
- So I tried swapping the winbond ROM chip (8 legs) to a new drive (same model number, same firmware, same site-code, same date-code).

After ROM chip transfer the disk spins up fine, detected in BIOS, capacity matched but I was unable to Image:

- First tried with Acronis (maybe would have been the fastest way) - did not work bc HDD not recognised.
- So tried with DDI. DDI could not head map, forward Imaging failed completely, backward imaging failed with UNC

So I soldered back the ROM, would be better to fix the PCB itself, anyone successfuly fixed those Seagate hybrid pcbs?

Thank you all,

@pcimage: You posted a similar case http://forum.hddguru.com/viewtopic.php?f=1&t=27790. Have you been able to fix that drive?

Known working PCB

Original BAD PCB

[Tomset]

Need a log terminal at startup.

[fzabkar]

Need a log terminal at startup.

I got a Seagate hybrid harddrive that does not spin anymore ... and does not give any terminal output.

[fzabkar]

So I soldered back the ROM, would be better to fix the PCB itself, anyone successfuly fixed those Seagate hybrid pcbs?

I can see at least 5 voltage test points.

[Tomset]

fzabkar,
Log after:

- So I tried swapping the winbond ROM chip (8 legs) to a new drive (same model number, same firmware, same site-code, same date-code).

After ROM chip transfer the disk spins up fine,

NVC with another's data. Need initialization NVC

[pcn]

Thank you for your replies

@Tomset: What exacly do you mean with initialization of NVC and how can this be done? I thought the disk is not readable because of the ssd-part / caching.

@fzabkar: My knowledge in electronics is limited, it would be kind if you or some other of the hddgurus could give me some hints on that board. I measured voltage on many testpoints on the pcb and attached a picture showing the voltages.

I think all orange points are OK at 12.34 V / 0.94 V
The blue one is 5 V should be OK

All the yellow voltages are not OK:
0.01 V should be 1.05 V
11.86 V should be 12.34 V

So what component could be damaged and what can I do to fix that PCB?

Voltages

[fzabkar]

I can see that some voltages are present and at least one other is missing. I don't know whether this constitutes a fault, or whether the MCU is switching one or more of the voltages off.

The attachments are my guesses as to the location and identity of various test points. A comparison with your working board would be helpful.

I expect that the ASIC would require Vcore and Vio voltages of its own. It may or may not share these with the main MCU.

The serial flash memory and SDRAM appear to require 1.8V.

The NAND flash requires 3.3V, probably supplied by the LDO regulator.

ST1L05CPU33R, ST Microelectronics, LDO regulator, 3.3V, 1.3A, marking 1L05C33:
http://www.st.com/web/en/resource/techn ... 186925.pdf

W9751G6KB-25, Winbond, 8M × 4 BANKS × 16 bit DDR2 SDRAM, 1.8V:
http://www.winbond.com/NR/rdonlyres/548 ... 51G6KB.pdf

W25Q80BWIG, Winbond, 8Mbit, serial flash memory, SPI, B = 1.65V - 1.95V:
http://www.nexflash.com/NR/rdonlyres/A4 ... 5Q80BW.pdf

Attachments

Vneg.jpg (75.74 KiB) Viewed 23380 times

Vio_1V8.jpg (51.09 KiB) Viewed 23380 times

preamp.jpg (103.12 KiB) Viewed 23380 times

3V3_LDO_reg.jpg (65.85 KiB) Viewed 23380 times

[pcn]

@fzabkar: Thank you for your help with this one. I checked those test-points (good & bad pcb):

3V3_LDO_reg: These are correct: 5V and 3.3V
preamp: gives me 0V 5V 0V
Vcore: On the bad pcb Vcore2 is 0V, Vcore1 is 0.94V / Good pcb Vcore2 is 1.05V an Vcore1 is 0.94V
Vio_1V8: Is correct with 1.8V on good & bad pcb
Vneg: All 0V on good & bad pcb, but on the left side of the black chip should be 12V I think (good pcb) vs 0V (bad pcb)

I will check the docs you attached tomorrow.

[fzabkar]

Could the absence of any terminal output be due to an inability of your equipment to handle the 1.8V LVTTL signal levels?

My next thought would be whether there is a short circuit on the Vcore2 output. Can you measure the resistance between this test point and ground?

The absence of the negative preamp supply appears to be due to it not being enabled by the SMOOTH IC. This is normal behaviour in some designs. That said, I am not certain that I have correctly identified the corresponding DC-DC converter circuitry. Perhaps the regulator in the top left corner is a buck boost converter that generates the boost voltage (18V ?) for the VCM circuit during normal operation and which is switched off when the spindle motor is not spinning?

Anyway, ISTM that the primary cause for concern should be Vcore2. If this supply powers the main LSI MCU, then its absence would explain why the PCB is brain dead. At the very least, this MCU should be receiving a Vcore and Vio supply from the SMOOTH controller. Once the MCU powers up, it can then tell the SMOOTH chip to switch on the other supplies, or to leave them in power saving mode. Alternatively, if Vcore2 powers the ASIC, then it could be that the MCU is switching it off. That said, if the NAND flash is receiving its 3.3V supply, then it wouldn't make sense to switch of the ASIC's Vcore supply on its own.

I would determine which of the Vcore supplies is powering the MCU and which is powering the ASIC. To this end, I would measure the voltages across the small bypass capacitors that surround each of these two controllers. Alternatively, you could just measure the resistances between the Vcore test points and each of these capacitors.

[fzabkar]

Here is a board (from ST1000DM003 or ST2000DM001) with the same SMOOTH IC and similar switchmode supplies:

http://ep.yimg.com/ay/yhst-444636922352 ... 7658-4.gif

It appears that the two Vcore regulators are for a dual-core MCU. It could be that they also power the ASIC.

The 6R8 coil does appear to be part of the negative preamp supply.

Would it be possible to use such a board to recover the SSHD?

[Tomset]

fzabkar
IMHO, toil nonsense.

firmware is able to interrogate the heads or controller NVC
See that there is no order. Immediately turn off HDD.
Not even attempting to start the engine.

[fzabkar]

fzabkar
IMHO, toil nonsense.

firmware is able to interrogate the heads or controller NVC
See that there is no order. Immediately turn off HDD.
Not even attempting to start the engine.

What is "toil nonsense"?

How do you know whether the missing Vcore supply is dead or has been switched off (I suggested both possibilities), and how do you know whether the other Vcore supply is sufficient on its own to keep the MCU alive in the idle state? If the Vcore2 supply is being switched off, then why is the 3.3V supply also not switched off? If this latter supply had been missing, then that would have confirmed the idea that the supplies were intentionally being shut down.

As for the donor PCB, ISTM that one would need to ensure that the data in the donor's NV cache does not contaminate the patient's data. To this end, I would list the drive's terminal command set. Hopefully there will be commands to clear or disable the NAND cache. Alternatively, I would dump the drive's ConGen parameters and look for something there.

[fzabkar]

fzabkar
IMHO, toil nonsense.

firmware is able to interrogate the heads or controller NVC
See that there is no order. Immediately turn off HDD.

Both the donor and patient PCBs appear to have been measured while disconnected from the drive. This means that neither will have been able to see the preamp, yet only the patient PCB is missing Vcore2.

[pcn]

Hi all,

first of all I must say I appreciate the feedback and help I got for this case from the forum, especially thank you fzabkar for your great help! Keep up the great work.

Back to the broken harddrive: I just switched back the ROM chip to the donor and tried to access ROM / SA with DFL-ST. That was impossible too, so I think there is something messed up badly or some adaptives are stored somewhere else.

As the client does not want to go any further, he aborted this case. So thanks to the forum members for the help. If there are similar cases with these hybrid drives, it would be nice if someone can share some insights because I think these drives seem to be a bit rare in the DR field atm.

Best regards,

[hamed_crazy]

Hi
i have sshd. my sshd not spin. after serching i see your coment
my pcb such as your pcb.
all voltage is good but motor voltage is 1.6 . in other hdd is 4.4 or 5v
with multimeter diode test I tested all resistor/Capacitor around SMOOTH ic .
i saw a short circuit. in dont know what is that!?
if you have this pcb plz test and tell me

Attachments

[hamed_crazy]

some one help me plz!!!

[labtech]

Looks like a resistor. Any marking on it?

Use the search function on the forum for PCB repair, voltage testing, etc. So much info..

[hamed_crazy]

fzabkar, can you help me plz?

[fzabkar]

fzabkar, can you help me plz?

sorry i dont have that pcb try puis

[labtech]

fzabkar, can you help me plz?

Is this your own personal drive? Did you notice anything out of ordinary before drive failing?