Case file · Desktop externals · SDR-2025-1124
I Ran DMDE and Then It Began Grinding.
Three terabytes, 95% full of music
, showing up as unallocated in Windows
, so the owner turned to software and left DMDE running for a few hours
. The result it returned was too many corrupt
, and then the hardware started telling him something much louder: the disk makes grinding noises
. At that point, as he put it, it was time to ask a pro
.
Seeing the same thing yourself?
0800 6890668
The translation.
That the drive showed up as "unallocated" indicated that this was the beginning of problems for the failing drive. What followed was the expensive part: a full-surface scan makes an unhealthy mechanism hunt forward and backward for many hours and heads on their way out almost always do not survive a session like this. Grinding noises are much worse than clicking noise because they indicate a head contacting the platter, which changes what would be a recoverable situation into a rush to find anything that is readable on the remaining surface.
Kit used on this job.
What happens in a case →| Which platform | Its role in this case | Why we use it |
|---|---|---|
| DeepSpar Disk Imager 4 | Mapped the heads to see which surfaces the long scan had spared | Head-map-driven imaging, with per-head timeouts, resets and power-cycle control |
| PC-3000 Data Extractor | Careful imaging, healthiest areas first, with seeking kept to a minimum | The PC-3000 platform's imaging unit, commanded by its firmware unit |
| UFS Explorer Professional Recovery | Reassembled the volume and brought the library back inside its folders | Handles the difficult filesystems better than most — APFS, ReFS, XFS, ZFS, Btrfs |
In the lab.
Work out what the scan had cost him
The head stack shows how worn down the heads are from running a failing drive heavily, and the platters show very light contact marks across much of the surface but heavier contact in areas that were scanned for longer periods of time. Stopping as soon as he did was also correct, because a couple more hours of operation would have significantly changed what data could be recovered.
Get it reading again, then map before committing to anything
Replacement matched donor parts allowed the drive to read its information again. A head map clearly showed which surfaces were good enough to recover data from and which were bad. While this was an uneven recovery instead of a completely lost cause, most of the drive was readable and some clear bands were completely unreadable.
Read in the gentlest mode, with as little head travel as possible
Reading favoured long sequential passes over jumping around, because every move of the actuator across a contact-damaged surface risks carrying debris further. The healthy bands were read first, in their entirety; the marked regions were read last, slowly and with no promises attached.
The result.
Most of the music library came back, with a straight count given against a straight situation. The missing albums were traced back to the bands the grinding had claimed. What survived did so because he stopped when he did.
Similar cases on the index.
Also from Desktop externals.
Does this sound like your own drive?
The rule holds as in every case above: switch it off, and let a free diagnosis come before any decision.