Desktop externals · case notes · MHD-2025-7124
I Left DMDE Running, and It Began Grinding.
The scan was meant to be the rescue. A mobile DJ had two decades of work on the Expansion — everything I have played out, and everything I recorded
— until Disk Management reported the entire volume as unallocated
and proposed initialising it. He declined, read a forum thread, and set DMDE running before I went up to bed
. By morning the log showed thousands of errors
, and the drive had a new noise: grinding, loud enough to hear from the hallway
. His message ended I should have pulled the plug hours sooner
.
Much the same fault? Call us.
0800 6890668
In everyday terms.
What that word does is invite precisely the wrong response. A drive that has lost sight of its own partition table has usually lost the ability to read reliably, and worn heads are the common reason; a full-surface scan then asks that same weakening mechanism to cross every sector on the disk, hour after hour, without pause. Few failing heads survive that errand. Somewhere before dawn the weakest of them dropped out of its air cushion and began touching the surface it was built to fly above, and grinding is what that contact sounds like from the next room. Every further minute spent platter and returned nothing.
What the bench used here.
How a job is handled →| Tool | Why we used it | What it brings |
|---|---|---|
| DeepSpar Disk Imager 4 | Mapped each surface, marking off the ground the overnight scan had ruined | Head map first, then head-by-head reads, holding resets, timeouts and the power line steady |
| PC-3000 Data Extractor | Ran the recovery in long continuous sweeps, healthy regions first | The PC-3000's imaging module, working under firmware-level control |
| UFS Explorer Professional Recovery | Rebuilt the file system and returned each track to the folder it came from | Reads the awkward filesystems others stumble over: APFS, ReFS, XFS, ZFS, Btrfs |
The lab work.
Count the cost of the night before making promises
Under inspection the head stack was finished, and the platters carried banding — light across much of the disk, considerably heavier wherever the scan had dwelt. What saved the job was the moment he pulled the plug. Had the drive been granted a second night rather than switched off at breakfast, there would be nothing here to write up.
New heads first; numbers only after the survey
A donor stack matched to the model restored the drive's ability to read, and the survey that followed drew the boundaries honestly. Large stretches came back at speed and without complaint. The scored bands did not; across those, the signal thinned and then disappeared. No figure was offered to him until that picture existed.
Bank the healthy ground, then approach the damage gently
Reading was done in long sweeps with as little actuator movement as could be arranged, since every seek across a scored platter risks dragging loose material into ground that is still clean. The undamaged areas were lifted entire and written to safe media before anything else; the banded regions followed, in short low-speed passes taking whatever they would give.
How the job closed.
The bulk of the collection returned with its folders as he had arranged them. Everything missing sat inside the banded regions, and the report set it out item by item: a run of commercial albums that can be bought again, and four of his own recorded sets that cannot. The account was given as it stood, not rounded in our favour.
More of the jobs we've closed.
Filed beside Desktop externals.
Does any of that match your case?
Nothing needs deciding until the diagnosis is back: switch the device off, send it in, and let the findings settle it.