I have a DC HC510 HUH721008ALE604.
As you can see in the picture I removed 1 of 4 diodes, a 2x4mm 5V TVS diode
that was heating as to turn the fingerprint into smoke immediately.
You can see that the 2 contacts in the diode group near the power lines,
in the lower left corner of the PCB, and the upper right corner of this group,
had the 5V TVS removed that prevented it from spinning.
The rest of the diodes in the group are OK too, do not have continuity/short.
I tested it and was shorted in both polarities. I know it is 5V from another
drive with the exact same design of PCB. It is put horizontally. The left/arrow is +
and the right is -, it gives ~4.964-4.965V.
The drive didn’t spin at first, but now it does after removing the TVS.
It it is not recognized. Some enclosures like Seagate 8TB ones immediately
spin it down. Maybe I need to replace the TVS diode, I could buy another
from eBay. Will that make it recognizable again by the computer?
What else can I do to make it recognizable? There are 2 tiny 0-ohm diodes
near the power lines too where these 4 diodes are at the very lower left corner.
They are OK, have continuity.
More photos. There is a 0-ohm resistor near the left top corner, all those on the PCB are OK:
The drive should work without the TVS diodes. Normally I would say that there is a likelihood that the preamp on the headstack has been damaged, but these newer PCBs are much better protected than the older ones. In fact there are two additional e-fuses inside the IC in the top left corner.
https://www.hddoracle.com/viewtopic.php?p=22339#p22339
Measure the voltages at the e-fuses (5Vout and 12Vout).
I took a better photo of the ICs:
It has number 87804C 8320PR, does it have the same pinout?
All 0-ohm resistors are OK, could replacing the diode make it detectable?
I will try to replace it by one of these:
Are there any bridges I can make? I will wait to make sure if the chip’s pinout is the same than the one in the hddoracle link.
I tested those voltages (from the regulator chip you showed me), they are OK.
I tested the 3.3V pins (pin 1, 2, etc.) from the SAS/SATA power
socket at the enclosure/drive and they seem to be disconnected,
they only have noise, but some SCSI SAS/SATA drives seem to have
those lines unused.
This failed drive is SATA, it seems to have those lines unused too.
The drive probably needs the 5V TVS replaced. Maybe the enclosure
that manages to spin it down when sensing something unusual
versus the rest of enclosures tell that the TVS is needed, unless you or
someone has fixed this model and worked without TVS, then some logic
chip could be bad, then I would need to send the PCB for firmware transfer
for $50 to somewhere.
Try replacing the TVS diode with a matching 5V one (like SMAJ5.0CA) to restore protection. If the drive still isn’t detected, check if it appears in BIOS or Disk Management. If not, the issue could be with the PCB or firmware. A firmware transfer to a donor PCB might be the next step.
That would be pointless for the purpose of data recovery. Moreover, it would be inappropriate since the CA version is a bidirectional diode. SMAJ5.0A is unidirectional.
TVS Diode FAQ:
http://www.hddoracle.com/viewtopic.php?f=100&t=86
Have you tested the drive directly via SATA rather than USB?
I have tested it in 3 enclosures, in a Microsemi SmartHBA 2100 4i4e SAS HBA with Windows Server 2019 Datacenter.
It is not recognized. Do you need more photos? Is there something else to test
or procedures to do/soldering to make it recognizable?
If you are getting +5V and +12V at the outputs of the e-fuses, then I can’t see how the symptoms could be due to a PCB fault. ICBW though.
Be aware that your drive is a helium model. Very few, if any, data recovery shops can handle these.
Edit:
Try flowing a blob of solder over each of the fuses (marked with an “S”).
Even if the S resistors/fuses have continuity, would that make a difference?
Are there services capable of offering a generic firmware chip capable of force-reading any drive of a given model?
No. Every ROM is unique, even for the same model.
How did you kill the drive? Was it spinning when it was clobbered? Was it due to a misadventure with modular PSU cables?
I had it plugged for 2 weeks on the ground.
There were like 6-8 organic debris around the right of the main cooled chip.
But it seems that it died only when the debris reached the power diodes,
then it heated while I was using it and stopped spinning, then I removed
the 5V TVS and it worked spinning again but without being detected.
I will try to make a bridge in the S and add a diode to see if it responds.
I could copy data until it died electrically, so it responded well until the end.
Then there could be internal damage, either hardware or firmware. It could also be that the ROM on the PCB has corrupt nonvolatile cache (for power loss data protection), but that’s only a wild guess.
How to fix?
Are there tools with SAS HBAs or Linux?
All the data recovery professionals will tell you that there is currently no solution for firmware problems. That’s because the Russians haven’t yet cracked this firmware architecture. You could try to dump the ROM and copy it to a donor PCB, but I’m not confident that this will work.
Here is a thread where I identified the voltage test points. Many of my links have expired, but you should find them at the Wayback Machine.
https://forums.tomshardware.com/threads/four-of-my-8tb-western-digital-hds-just-had-their-pcbs-fry.3605045/post-21746264
One of the enclosures, an 8TB Seagate that came with a new drive
and I took out to read any disk, spins down the drive after it tries
finishing the initial spin-up and seek. It could hint why the enclosure
spins it down, how does it know that it has something wrong. If the
enclosure could tell somehow what is happening, it could probably resolve
the issue.
Disconnect the SATA data connector from the original Seagate drive and see if the enclosure spins it down.
Alternatively, some enclosures won’t spin up the drive unless they see a USB host.
I’m wondering if the “organic debris” caused some insidious damage. The right side of the MCU connects to the HDA pads (read/write channel), so look for breaks in the traces.
No, this is the only drive that it spins down, from a 5TB, 8TB, 2TB, 1TB, 500GB, etc.
Also, I was thinking to get an identical drive and send both
to hard drive PCB board repair, so they transfer
either the ROM or fix the PCB with the new PCB for replacement parts,
for $50.
Here someone says that this problem is extremely trivial:
Yeah, that was me, but in that case the drive wasn’t spinning at the time of the overvoltage.
Ask your PCB supplier to provide the ROM dump as a file.
Good luck.
