Soldering a SMOOTH chip wouldn't be easy for a novice. Then you would need to consider cost, especially if you don't have a hot air station.

_________________
A backup a day keeps DR away.

Well done !

The OP did a good job. I was ready to give up, be he hunted down the resources and was inspired by Spildit's tutorial at the HDD Oracle.

We had several disappointments along the way. The first was our inability to retrieve a full set of SA resources from the patient using the donor PCB. We needed modules 0x102, 0x103, 0x104, 0x105, 0x106, 0x107 and 0x109, but were only able to recover 0x102, 0x103 and 0x105. However, this was enough to rebuild ROM modules 0x0A, 0x30 and 0x47.

The second disappointment was a bug in WDMarvel (demo) which produced a corrupt ROM when we attempted to regenerate a ROM from the above modules (plus the missing modules from a donor). We then had to patch a hybrid ROM by hand. Fortunately this hybrid ROM was good enough to gain full access to the UA and nothing more needed to be done.

_________________
A backup a day keeps DR away.

A note to the above,

At first, the malfunctioning PCB had the Vcore (1.2V) and the Vio (3.3V) supplies. And therefore the ROM could have been read with a USB-TTL adapter.

However, I made the mistake of measuring the voltages too many times. Since in order to do this, I powered up the PCB. This will surely contributed to the PCB failing even more than before. And in the end the voltages of the Vcore (1.2V) and the Vio (3.3V) were missing.

In short, if you have a failing PCB, it is advisable to power up as few times as possible. Because due to its being a faulty item, there is the serious risk of its failing completely if you do too many tests.

The conclusion is that it is not a good idea to do so many tests (as I did) to a failing component.

I'm still trying to figure out how the drive came ready and any of the SA was readable without the -5v rail. Can't really expect DIYers to swap the smooth or create hybrid roms, DR tools would have made it easier but at a cost. OP has their data and hopefully has now put a good backup solution in place so all is good.

_________________
Data Recovery Services in the UK.
https://www.usbrecovery.co.uk/

In order to retrieve the SA modules from the damaged HDD., I replaced the faulty PCB with a working PCB of the same exact model.

For this, I purchased an HDD which was the donor. This donor HDD is very similar to the damaged drive. Apart from its having the same exact PCB model, it also has the same capacity (500 GB) and its firmware is very close to the patient HDD.

(Firmware of the Donor HDD is 005C003H. And the firmware of the patient drive is 005C003G).

I placed the PCB of the donor HDD into the patient HDD. And in this way it is how I retrieved 249 modules from a total of 348 modules. (Therefore, 99 modules were missing).

Finally, I am using the Donor HDD as a replacement of the damaged HDD. Because I have already thrown away the patient HDD.

@fzabkar
Do SA modules 0x102, 0x103, 0x104, 0x105, 0x106, 0x107 and 0x109 correspond to ROM modules: 000A, 000B, 000D, 004F, 020B, 0030, 0047?

0x109 is the ROM plus a 0x200 byte "header" which needs to be stripped off.

0x102, 0x103, 0x104, 0x105, 0x106 and 0x107 correspond to 0x0A, 0x0B, 0x0D, 0x30, 0x47 and 0x4F, but not necessarily in that order (I don't have my other computer in front of me). You need to strip off the 0x18 byte header from each SA module and paste the appropriate header from the ROM module. Then you need to recalculate the 32-bit checksum.

_________________
A backup a day keeps DR away.

0102=Copy of 0A
0103=Copy of 47
0104=Copy of 0D
0105=Copy of 30
0106=Copy of 4F
0107=Copy of 0B
0109=Copy of Flash code

_________________
A backup a day keeps DR away.

Am I allowed to make a suggestion?

Since this method seems new. And there are some people interested in it.

I think it would be a good idea to do a tutorial on how to rebuild a ROM when you have some of the SA modules (but not all of them).

If you wish, you could use any of the module files I sent you or even the rebuilt ROM which has proved to be a working ROM.

Most people would be using PC3000 or MRT Lab. Both tools can regenerate WD ROMs. WDMarvel is far less popular.

That said, I was in fact thinking about producing such a tutorial at the HDD Oracle. I was even thinking about writing a tool to automate the process.

_________________
A backup a day keeps DR away.

I forgot that I had already written a tutorial.

Regenerating a WD ROYL ROM from SA MODs:
https://www.hddoracle.com/viewtopic.php?f=59&t=761

_________________
A backup a day keeps DR away.

Thank you for the link.

The info is interesting.

By the way, there is one thing that I do not understand.

For example, my PCB HDD had a Marvell MCU. However in the SA modules it reads ROYL and not Marvell.

And when reading the ROM with the WDR Demo. I did so by selecting MARVELL first and then I selected the ROYL option. But both options generated exactly the same ROM. Then, why does it differentiate between ROYL and MARVELL?

Both the "Marvell" and "ROYL" firmware architectures are base on Marvell MCUs. The first MCUs were WDC branded, then came Marvell and Marvell ROYL. Yours is the latter.

_________________
A backup a day keeps DR away.

@fzabkar
I've noticed that for this PCB in this thread and the one you referenced in your HDD Oracle post about regenerating ROM from SA, both of these PCBs did NOT have U12 external ROM chip.
I tried to open the SA modules in my WD1001FALS-40Y6A0 which has an onboard ROM, but all I got was essentially empty ROMs:

It was "empty" for SA modules: 104 (0D), 106 (4F), 105 (30), 103 (47) as well. Is the information not replicated in the SA region for these PCBs with ROM chips?
In other words, if we have a HDD/PCB with a dead ROM chip, how can we extract the proper information from the SA regions if they aren't populated. Is there another area to look in? Thanks!

As long as the SA has no defects, then module 0x30 from a donor can be used in the patient. 0x0B would be present in 0x109, as would a template for 0x0D. Module 0x0A has a head map, so you could use the same module from a donor that has the same heap map.

Module 0x47 is the most important. Its unique data can be recovered from module 0x40.

_________________
A backup a day keeps DR away.