A signage and print firm on one of the business parks off Western Road keeps its artwork archive on a 2TB Seagate external: cut files, proofs and customer logos going back years. It had been sounding coarse for a fortnight, in a way the studio manager described as more of a grinding than a click. She bought a new caddy, moved the drive across, and it sounded exactly the same. The repair place they use for printers took one listen and declined.
| Media | A Seagate 2TB external drive that had begun sounding rough in use and has now stopped working. Moving it into a new caddy made no difference to the noise or to the result. |
| What was reported | Seagate 2TB external drive, stopped working, sounded rough while running, a new caddy made no difference, a local repair specialist was unable to help, artwork archive needed. |
| Fault class | Mechanical. The noise came from the mechanism rather than the enclosure: the spindle bearing had become rough and was loading the motor, and one head had made contact as a consequence. The caddy swap had already ruled out the bridge and the power supply. |
| On the bench | Not run further, profiled once on a current-limited bench supply with the spin-up current and the acoustic signature recorded, opened in clean air under laminar flow, motor, head stack and platter surfaces inspected under a stereo microscope, the platter pack migrated to a matched donor mechanism with matched heads fitted and alignment set, imaged on DeepSpar behind PC-3000 UDMA, the volume parsed in R-Studio and artwork validated by opening it. |
The caddy swap was a real test and it gave a real answer. An external drive is a disk inside a shell, and the shell contains a bridge board and a power circuit that fail often enough to be worth eliminating. Moving the disk into a new shell and hearing the identical noise proves the noise belongs to the disk.
Rough is a different noise from clicking, and it points somewhere else. Clicking is a head stack retracting and retrying. A continuous rough or grinding note under load is usually the spindle: a bearing that has lost its lubrication and is no longer turning smoothly. It is a slower failure, which is why it had been audible for a fortnight rather than a day.
A rough spindle drags the heads into trouble. The head fly height depends on stable rotation. A platter pack that wobbles or slows unpredictably brings the surfaces closer to the heads than they should be, and eventually one touches. That is why this job needed both a mechanism and a head stack rather than either on its own.
Migration is a bigger operation than a head swap and it is not a last resort. Where the spindle itself is the problem, the platter pack is moved to a healthy mechanism with its relative alignment preserved. It is careful work, it is done in clean air, and it is the standard answer to a seized or roughened bearing.
It was not run again beyond one profiled start, which recorded a spin-up current well above normal and an acoustic signature with no clean settling point, both consistent with a bearing under strain rather than a head stack hunting.
Opened under laminar flow, the spindle confirmed it: rough to turn by hand, with visible wear. One head showed contact damage and the corresponding surface carried a faint band, which was mapped. The platter pack was migrated to a matched donor mechanism with alignment preserved, matched heads were fitted, and the drive imaged on DeepSpar behind PC-3000 UDMA with the marked band deferred to last. R-Studio read the volume out of the image and a spread of artwork files was opened to prove them.
The archive came back, with a small number of files in the marked band unreadable and named individually rather than reported as a percentage. All of them turned out to be proofs from a job the studio had finished and delivered five years earlier.
One drive, £300 + VAT, migration included. This is mechanical work carried out in clean air, so the published guarantee did not extend to it: 50% before the drive was opened, the balance once the studio had checked the recovered artwork.
Copy what matters off now, in one pass, and then stop using it. A drive that sounds rough is usually degrading rather than dead, which means there is a window, and the window closes. Once it has stopped working, do not keep powering it to see whether it settles: a strained spindle and a touching head do more damage each time they run. Changing the caddy is a fair test to try once, since bridge boards do fail and it costs nothing to eliminate them. Trying six caddies is not a test; it is six more start-ups. And do not put a drive in the freezer, whatever the internet says about seized bearings, because the condensation that follows is a second problem on top of the first.
These files are written up from genuine enquiries taken from households and businesses in Bracknell, across Berkshire and Surrey and into London, with names, firms and anything else identifying stripped out. Each one sets down how the fault was reasoned through, the method it actually needed and the equipment it was worked on. One of them is a job that was turned away, which belongs here as much as the rest of them.
A drive that announces itself to the port but never appears has a mechanism stuck short of ready
A fall damages heads mechanically whether the drive was running or not, and the clicking is the proof
Clicking gets worse with use, and the fortnight between the first noise and the last is spent, not free
Examining it is free and takes 2 working days from arrival, the figure goes to you in writing and does not move afterwards, and a logical fault carries no fix, no fee. Start online or ring the freephone.