WD800 problem with MHDD & R-Studio

[Kilroy2k1]

Hello,
I have a WD800 that died and I would like to try and recover the data from it if possible but have run into a few issues and I hope someone can help me decypher the scan results i am getting.

First a few details
OS is Windows XP Home edition
Drive is set as secondary master, only device on that ide channel
all unnecessary processes disabled
The WD800 drive was the primary OS drive and just died suddenly.
Chkdsk shows many errors and stops with a cannot check this drive message
With the drive attached to the system it drastically slows just about any operation down to a crawl..
I have a fair bit of electronics and programming experience so I'm not afraid to swap chips/eproms if necessary..

I have read many posts from here and on the recommendation of a few posters I tried the demo version of R-Studio to scan the drive and see if it was possible to make an image of it to do the actual file recovery on (1 time lic is quite reasonably priced). R-Studio seems to get upto a certain point then I get either a bluescreen error in XP or the machine just outright crashes, never gets past about 30%.

Next step was to do a scan with MHDD with the cd version.

The WD800 is now plugged in as primary master, MHDD identified the drive properly and I initiated a non destructive scan.
The scan starts out fairly well but eventually gets to areas that are >500ms test times in an odd pattern.
When the scan hit about 59,000,000 blocks the pattern of the slower block sections become X unrecoverable blocks.

What I am trying to wrap my head around is why the pattern in the corruption. The scan shows about 140 blocks with fairly normal results, then 140 blocks of slow/Unrecoverable, then 140 normal and this pattern just repeats.

I am still learning how these drives work and have messed with MHDD on a few older non important drives before but this time I would actually like to get the data from the drive so Its important I understand exactly where the failure is before I try anything on it.
Here are some of my thoughts of possible causes or faults, please correct me if i am wrong.

Could it be a weak or failing head as the slow/dead areas are in a fairly consistant pattern?
Or failing control chip or even a corrupt SA area?

I hope I covered enough information
Any ideas would be appreciated.
Thanks for your time
Tom S.


(edit) Yeah i cant spell..

[pcimage]

Yes, could be a weak head if the pattern is all the way across the drive.

Or media degradation.

Either way, you're better off cloning with a non-windows based imager (i.e DeepSpar, Media Tools Pro, CopyR etc)

Don't let windows access the drive any more, it'll kill it.

[Kilroy2k1]

Thanks for the reply pcimage,

Looking through the MHDD scan I believe your pretty accurate on the weak head thing. I checked into those programs you mentioned and they look good, but i am afraid to run any more demos on this for fear of trashing the data. The hardware solution is definately out of my price range. I'll do a little more research on imaging before trying anything more on it. once I am sure of what i am dealing with I'll hopefully choose the proper course of action, or at least get more advice..

Thanks again.
Tom S.

[Kilroy2k1]

Little update on this one thats baffleing me..

A friend of mine came over today and brough a cd with all sorts of diagnostic and recovery utilities on it and he ran prosoft media tools to try and make an image of the wd800 drive to a wd1600. Well the program estimated 36000 hours to make an image, speeding up and slowing down on different areas. I expected to see a consistant pattern to the slow areas like what MHDD was showing but what i see while its imaging i cant to make much sense of.
After letting it run for over 6 hours and watching the head and sectors when the drive pauses it seems random. On one cyl it'll pause on head 0 sect 1, then after a few min count up fast to head 5 sect 63, pause there then pause on head 6 sect 1 then count fast upto head 11 sect 63 then..
Watching it long enough and you'll see as it reads through cylinders at one time or another every head will eventually have a pause at either the beginning sector or the end sector, and some cylinders will have more pauses then others.

If it were 1 or 2 heads failing wouldnt i see some sort of pattern of heads/sectors where it hangs up and pauses?
Could a bad PCB or firmware cause this kind of strange issue?
Can anyone tell me how the heads are actually numbered in this WD800JB drive? I know there are not 16 actuall heads but I havent had one of these exact models apart to know how many platters and heads there are.
I'm still learning and figure this is a good project to get my feet wet on.

Thank you for any assistance with this.
Tom S.

[pcimage]

If you're in UK, I can help you out.

[Kilroy2k1]

Unfortunately I am in Canada. SW Ontario region.

Tom S.

[Kilroy2k1]

Ok now i am stumped,
Every imaging program dies out and locks up the system right around the 18gb mark. each one reads pretty fast upto about 16gb then gets pregressivly slower untill it eventually just locks up.
Does this sound more like a controller failure rather then a head failure now? I have used my friends media tools pro, copyR and ddrescue with the same results on each one.

Thanks again
Tom S.

[pepe]

Hi,

have u tried to image in reverse direction?
Or simply set the start of imaging to let's say 20G or 25 for the first try, just to skip a damaged area.

pepe

[Kilroy2k1]

Yes I have tried media tools in reverse, it just kicks out and stops after 3 sectors. MHDD (f4) scan did have a hard time around that area but it managed to get all the way through.

Tom S.

[Shortscurcuits]

Hi try removing the PCB and cleaning the contacts where the heads to PCB contacts are and feel the controller chip area of the PCB when it's running. With a lot of WD bad controller chips and bad PCB contacts can do this also. Clue is reads get worse as board gets warmer.
Could be heads degradation or media problem but try this first.

[Kilroy2k1]

Hi Shortscircuits,
When i first started the task I had pulled the board off and checked for physically damaged chips and cleaned all contact points, it made very little difference.
It took well over a week, but ddrescue managed to get about 90% of what was on there.
I learned a lot from this and plan on possibly expanding out onto the field if i can learn enough to be confident and competent at it. I love a challenge.
Thank you all who replied and offered help, It was really appreciated.
Tom S.