External drives · bench notes · HUL-2025-1124
I Set DMDE Going at Bedtime and Woke to Grinding.
He took the scan for the cure. Twenty years of work as a mobile DJ sat on that Expansion — every set I've played and every mix of my own
— and Disk Management had started calling the volume unallocated
, with a button there to initialise it. He said no to that, read a forum thread, and set DMDE going right before bed
. By morning the log held thousands of errors, one after the other
, and under it a noise nothing in that drive had ever made: grinding, you could hear it from the landing
. His message finished I had the power on far too long
.
Sounds like yours? Ring us.
0800 6890668
What causes it.
Unallocated is the word that sends people to a scanner, and that is where the damage begins. A drive unable to read even its own partition table has usually stopped reading properly, and worn heads are near the top of the list of causes. A full-surface scan then flies those same worn heads across every sector on the disk, hour upon hour, without a break. Not many failing heads last that trip. At some point in the night the weakest of the four lost the air cushion under it and settled onto a platter. That is the grinding heard from the landing. From then on each pass took off more coating and returned nothing.
The tools this one needed.
The steps we follow →| The gear | What this one did | Why we have it |
|---|---|---|
| DeepSpar Disk Imager 4 | Mapped all four surfaces and fenced off what the scan had wrecked | Grades each head first, reads one surface at a time, and keeps resets, timeouts and power under control |
| PC-3000 Data Extractor | Read in long unbroken sweeps, sound areas taken first | Copying is run at firmware level, with the PC-3000 in charge |
| UFS Explorer Professional Recovery | Rebuilt the volume so each track went back into its own folder | Opens the volume formats most tools cannot: APFS, ReFS, XFS, ZFS, Btrfs |
The stages.
Map the damage before naming a price
With the cover off, the head stack was scrap and both platters were scored — light over most of the disk, far worse where the scan had spent the longest. The job was rescued by the hour at which he pulled the plug. Another night of it instead of a stop at breakfast and this page would have nothing on it.
Heads first, and no figure until the map exists
A donor stack of the correct model got the disk reading again, and the survey that followed showed exactly which areas still responded. Large stretches came back quickly and cleanly. Across the scored bands the signal weakened, then stopped. He was quoted nothing until that map was done.
Take the sound ground first, then work into the scoring
The read ran in long sweeps with the actuator moved as little as the job allowed, since every seek across a scored surface can drag loose debris into ground that is still good. Clean areas transferred first, in one piece, onto fresh media. The banded stretches came later, in short slow passes, handing over whatever they would.
How it finished.
Almost all of it went back into the folder layout he had made himself. What did not come back sat in the banded ground, and the report listed every item: a batch of released tracks he can buy again, and four of his own recordings that are held nowhere else. We wrote that down plainly and put no gloss on it.
What people read after this.
Tagged Desktop externals too.
Is yours doing the same?
Turn it off. Send it to us. The diagnosis comes first. It says which files are still readable and which are not.