Opinions sought on a way to fix a failed IBM Deskstar HDD

Hello

My 100GB IBM Deskstar ("DeathStar") drive (type IC35L100 series ) packed up
a few days ago. It was being used in an IDE- USB enclosure and basically ,
XP failed to recognise
it as a drive. I tried it in the computer's BIOS which also fails to
recognise it.

Judging by the sound it makes, the heads are tracking across the platters OK
but failing to calibrate or locate track 0. Internet research suggests an
NVRAM failure ( the HDD BIOS may be corrupted).

Having found a potential donor 120GB Hitachi drive (Hit. took over IBM's HDD
business), I plan to get a known working controller PCB with a similar
Machine Level Code (MLC) revision, and substitute it for the board on my
failed drive. I aim to recover some/all data this way. Again, internet
research shows some success ... but only "some".

Questions then:

1) Does anyone know what the NVRAM or BIOS actually does on a HDD that might
change between revisions of the same HDD, or between generically similar
versions of the same HDD?

2) How likely do you think I might be in succeeding, whereby I use a similar
MLC coded PCB but not identical? A daft question I suppose in that the
changes per MLC release may be significant.

It's not a subject I am particularly familiar with so if someone has some
suggestions I'd be happy to learn. Unfortunately, a data recovery company
will charge many hundres of pounds so not viable in this case.

Thanks for any useful tips or experiences you can share; I don't know a fat
lot about the component-level functionality of how a hard drive works so I'm
keen to learn a little.

DDS

Previously Duncan Di Saudelli wrote:
> Hello

> My 100GB IBM Deskstar ("DeathStar") drive (type IC35L100 series ) packed up
> a few days ago. It was being used in an IDE- USB enclosure and basically ,
> XP failed to recognise
> it as a drive. I tried it in the computer's BIOS which also fails to
> recognise it.

> Judging by the sound it makes, the heads are tracking across the platters OK
> but failing to calibrate or locate track 0. Internet research suggests an
> NVRAM failure ( the HDD BIOS may be corrupted).

This would be very surprising. Typically systems check their
firmware checksums before doing anything else and refuse to
start if the firmware is corrupt.

> Having found a potential donor 120GB Hitachi drive (Hit. took over IBM's HDD
> business), I plan to get a known working controller PCB with a similar
> Machine Level Code (MLC) revision, and substitute it for the board on my
> failed drive. I aim to recover some/all data this way. Again, internet
> research shows some success ... but only "some".

More likely sucees on misdiagnosis.

> Questions then:

> 1) Does anyone know what the NVRAM or BIOS actually does on a HDD that might
> change between revisions of the same HDD, or between generically similar
> versions of the same HDD?

It may change even if the drives look exactly the same from the
outside, except for the serial number.

> 2) How likely do you think I might be in succeeding, whereby I use a similar
> MLC coded PCB but not identical? A daft question I suppose in that the
> changes per MLC release may be significant.

No idea.

> It's not a subject I am particularly familiar with so if someone has some
> suggestions I'd be happy to learn. Unfortunately, a data recovery company
> will charge many hundres of pounds so not viable in this case.

I would think that the most likely outcome is that you will
have wasted the money for the replacement PCB.

> Thanks for any useful tips or experiences you can share; I don't
> know a fat lot about the component-level functionality of how a hard
> drive works so I'm keen to learn a little.

I had the complete service manual for an old RLL disk at one time.
Really, really complex. Should have gotten even more complex by now.

Arno