External drives · bench record · PLY-2025-1124
I Started DMDE at Bedtime and Woke to Grinding.
He took the scan to be the cure. That Expansion carried twenty years of mobile DJ work — the whole of what I've played, and my own mixes as well
— and the volume had begun showing up in Disk Management as unallocated
, a button beside it offering to initialise. He refused that, found a forum thread, and set DMDE running just before I went up to bed
. The log by breakfast held thousands of errors, on and on
, and beneath it a noise the drive had never once made in twenty years: grinding away, audible from the landing
. His message finished: the power was on for far too long
.
Sounds like yours? Talk it through.
0800 6890668
Why it happens.
That word, unallocated, is what sends people to a scanner, and the damage begins there. A drive that cannot read even its own partition table has generally stopped reading properly, and worn heads sit near the top of the list of reasons why. Set a full-surface scan going and those worn heads are flown over every sector the disk has, hour upon hour, with no let-up; few of them come through it. Somewhere in the small hours the weakest head of the four ran out of air beneath it and touched down on a platter. What he heard from the landing was that. Each pass from then on removed more coating and returned nothing at all.
The kit this job called for.
How the job runs →| The kit | What it did here | Why we keep it |
|---|---|---|
| DeepSpar Disk Imager 4 | Mapped all four surfaces and fenced off the ground the scan had ruined | Scores every head before it starts, works one surface at a time, and holds resets, timeouts and power in check |
| PC-3000 Data Extractor | Read in long unbroken sweeps, with the sound areas taken first | The PC-3000 runs the copy through the drive's own firmware |
| UFS Explorer Professional Recovery | Rebuilt the volume so each track came back under the folder it belonged to | Mounts what other software will not: APFS, ReFS, XFS, ZFS and Btrfs volumes |
How it ran.
Map the damage before quoting a figure
Once the cover came off, the head stack proved to be scrap and scoring showed on both platters — light over most of the disk, far heavier where the scan had lingered. What rescued this job was the hour he pulled the plug. Give it one more night rather than a stop at breakfast, and there would be nothing to write here.
Heads first, and no price until there is a map
Reading came back once a donor stack of the correct model was in, and the survey after it showed precisely which ground still answered. Long stretches returned fast and clean. Over the scored bands the signal thinned out and then died. He was quoted nothing at all before that map existed.
Take the sound ground, then move into the scoring
Sweeps were kept long and the actuator moved as little as the work allowed, since a seek across a scored surface can pull loose debris into ground that is still sound. The clean areas went across first, in one continuous piece, onto fresh media. Whatever the banded stretches would give came later, in short slow passes.
How it ended up.
Almost everything returned to the folder layout he had made for himself. The losses lay in the banded ground, and the report names every one: a set of released tracks that can be bought again, along with four of his own recordings held nowhere else. That was written down plainly, with no gloss on it.
Pages people open after this.
Also tagged: desk drives.
Seeing the same thing?
Leave it switched off and post it in. Nothing happens until the diagnosis, which lists what can still be read and what cannot.