Hi all,

I am trying to help a friend recover some photos off a WD10TMVW drive that is not being recognized properly when attached to windows. It is one of the drives with the USB PCBs. I work as an IT admin so I'm reasonably comfortable running software but I am not comfortable doing anything requiring soldering etc to the drive (especially not while there's data that they really want to recover from the disk).

Symptoms:
-On initial power on, the drive does make strange noises, but not the usual hard click that I am familiar with. It sounds more like its scrubbing across the drive? It isn't constant and it does eventually stop.
-On windows the drive is usually recognised but inaccessable - just hangs.
-On linux, after what I assume is a timeout period, I am able to read data with ddrescue.
-ddrescue is painfully slow, even with some tweaking of the parameters (--force -n -N -c32 -R -vvvv)
-ddrescue is throwing ocassional read errors ~1000 every 24 hours of running time.

Progress So Far:
After a few weeks of running ddrescue I was able to get about 80% of the data written to an img file on another drive, but the remaining 20% is crawling along at ~1-2 kb/s so would take literally years to complete at this rate.

Questions
1. Given the symptoms, does this sound like a hardware issue alone? Or could a sata / firmware patch help the situation?
2. Are there any freeware tools that can test if the heads are OK?
3. If there's a bad head, is it possible to do anything in software to disable it and have another head compensate for it?
4. Any other non-destructive tips / tricks that I can try?

Thanks in advance!

I have a bad feeling the head might have crashed and that scraping noise you are hearing is the head chewing up the disk. You might not get more than you recoved. Hopefully most of the dats is in that image file.

4 heads on this, is it 80% or 75%.

I'd go with dead, or if it's still reading good data, a very weak head. You have 2 choices - replace the heads - not really a DIY job or sit it out and pray you're not turning data into dust. The value of the data will help you make the decision.

hddsuperclone will give you a better idea of what's going.

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

Having read your post fully now you're probably looking at platter damage too, so even a head swap may not get you much more back.

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

The drive does have periods where it sounds pretty normal... is there a possible reason for that? I would assume a crashed head would be consistently scraping on the platter?

They tell me the drive hasn’t been dropped or anything... I will try to capture some audio of the disk both when it’s normal and abnormal sounding.

I’ve grabbed hddsuperclone and the analyse reports 37.5% dead or weak head.

Sadly it also reports 100% bad.

Regardless, when I attempt a recovery, hddsuperclone flips between broken sounds and skipping sections, and quietly running with 15mb/s copy rates.

With ddrescue, it did seem to very slowly be recovering data (on phase 3)

I’ve attached a recording of the drive when it is noise / broken sounding. Is it seeking and then resetting the heads?

Running a drive that makes a unusual sound might kill your friends data in seconds. Running a drive without locating source of error involves a very high risk for data loss. Step one before applying power is to open that drive and check heads under microscope, will reweal course of error and what trouble to expect in the recovery process.

_________________
Digitalsupport Data Recovery
https://digitalsupport.dk

It's looking more and more like a single dead/weak head and platter damage.

Symptoms point towards it taking a knock whilst spinning.

Constant scraping would only be head if the head was in contact with the platter constantly, your occasional sounds (and the recording isn't showing) is when the head is trying to be used.

You've done all the DIY you can without replacing the heads. There are no magic fixes for hardware failure I'm afraid.

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

Could seem to get the sound file to upload direct to the forum so it’s here: https://sndup.net/88y2

I’ll check with them if they are happy with the files recovered (and if they’re ok to let me play around with the disk some more) or if they want to pay for the recovery now from someone with the right expertise and tools.

If they let me keep fiddling I’ll try leave it run with hddsuperclone for some time to see if it will eventually recover the disk, even if it’s incredibly slow!

Last question... I let it run overnight and hddsuperclone has completed 124gb (~12.5%) of the drive. I noticed that the data preview seems to always show the same data (either all 0 when it’s struggling) or a similar random string when it’s working.

I also stopped the clone and tried to run photorec over the image file it’s creating and it has got some files that it’s found... should I just ignore that the preview strings aren’t different?

Use the hddsuperclone log file viewer http://www.hddsuperclone.com/sitev1/hddscviewer to see what's going on. Post the log file and if I get the chance I'll take a look. Don't worry about the preview, is as much to show it's still working as anything else.

Maximus (who wrote hddsuperclone) is normally lurking somewhere - try saying he name 3 times in a mirror, see if he appears.

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

What I meant with the preview was that its either:
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................

OR

81 fa 24 4f a5 19 9a cd 42 d4 24 4e aa 6b b4 a1 ..$O....B.$N.k..
81 fa 24 4f a5 19 9a cd 42 d4 24 4e aa 6b b4 a1 ..$O....B.$N.k..
81 fa 24 4f a5 19 9a cd 42 d4 24 4e aa 6b b4 a1 ..$O....B.$N.k..
81 fa 24 4f a5 19 9a cd 42 d4 24 4e aa 6b b4 a1 ..$O....B.$N.k..

never ever anything but those two options. Just seemed odd to me that's all.

The viewer graph has some nice regularly spaced non-tried / non-trimmed in amongst the mostly green finished blocks. It looks a bit like a set of concentric black semi circles son the left panel over the green background. I'm doing the scan on an offline standalone box so cannot copy the logs over to post ATM.