Case record · Portable drives · SDR-2025-0925
Rattling More Than It Should, and Now Invisible.
The drive still went through the motions — the lamp lit exactly as before, the light still comes on
— but it had gained a rattle: a loose sound when it spins up
. It showed up nowhere: nothing in Finder or Disk Utility
, and a colleague's Windows laptop agreed. Spare leads had already cleared the cable: I've tried three different cables
.
Seeing the same thing yourself?
0800 6890668
The translation.
A hard drive is built to fine tolerances, and it never makes a new noise for decoration. A rattle is a different animal from a click or a beep: it says some part now has room to move that the design never allowed it — a head stack out of its seat, a fastening gone slack, or a broken limit stop on the actuator. Trying it under two operating systems with a known-good lead had already ruled out everything outside the casing, which is exactly the sifting worth doing before anything is posted anywhere.
Kit used on this job.
What happens in a case →| Platform | Its role in this case | Why we use it |
|---|---|---|
| PC-3000 Data Extractor | Read the drive once the mechanism held steady, zone by zone under control | The PC-3000 platform's imaging unit, commanded by its firmware unit |
| DeepSpar Disk Imager 4 | A head map showed which surfaces the loose stack had spared, and which not | Head-map-driven imaging, with per-head timeouts, resets and power-cycle control |
| UFS Explorer Professional Recovery | Reassembled the Mac-formatted volume from the image, folders and all | Handles the difficult filesystems better than most — APFS, ReFS, XFS, ZFS, Btrfs |
In the lab.
Work out what is actually loose
Once it was open, the noise turned out to come from the head assembly's mounting, not from anything rolling around loose inside — the stack had shifted off its seat, which accounted for the rattle and for the drive never finishing its start-up routine. The platters were examined first, because a head stack with play in it can score a surface on every rotation.
Put the mechanism right, then map before you read
Matched parts were fitted, the seating reset, and the drive brought back up under controlled conditions. The head map was less encouraging than hoped: five surfaces clean, the sixth lightly scored across its outer tracks, earned during the fortnight it stayed in use after the rattle began.
A single full pass, and no second go
With the mechanism steady, the whole read was completed in one sitting. You do not keep switching a drive with a rebuilt head stack on and off to check on progress; it is read once, carefully, and then taken out of service.
The result.
The five sound surfaces gave up everything. The scored one cost part of one project, and the missing files were named in the report. Two drives replaced it, one kept off site.
Similar cases on the index.
Further cases in Portable drives.
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.