The fault · noises coming from the drive
Take the power off it, and stop there. When the noise begins, what is on the platters is normally still intact; whatever goes after that goes to the retries. A head that has lost its bearings returns to the ramp, launches across the surface, finds nothing, comes back and launches again — over and over, for however long the plug stays in. Every attempt is another chance of touching the platter. Unplug it and the count stops where it is. The first look on the Reading bench costs nothing, a single drive is £300 + VAT once you have the figure in writing, and most of these go home with the greater part of the disk on new media.
An engineer answers, and looking is free
0800 6890668
| What you notice | What is likely wrong | What happens next |
|---|---|---|
| A tick about once a second that will not stop | The arm steers by servo marks laid down at the factory. It cannot pick them up, so it falls back to the ramp and comes round for another go | Cut its power |
| A thin beep, usually out of a 2.5" portable | Silence, nothing turning: a seized spindle bearing, or heads bonded fast onto the platter surface | Do not tap it, do not shake it |
| A harsh rasp, or a grind | A head is down on the platter, and it is doing damage while it runs | Every second of running costs files |
| A short chirp at each start-up | Calibration failing, most often after a fall | Switch it off and let it rest |
| A few clicks, then silence | It gives up, resets, and runs the same sequence again | Unplug it and break the loop |
| It only clicks on certain files | A single head has gone, or one band of the platter is damaged | Stop copying and have it imaged |
A case number goes on your device the day it reaches us, and an engineer then finds the real fault — free, and before anything else happens. You are told plainly what will come back and what will not, with one fixed figure in writing beside it. Nothing chargeable starts until you have read that and said yes.
The mechanism and the platter surfaces get a proper inspection before any power goes near the drive. Where a noisy unit does have to run, an engineer sits with it and gives it seconds at a stretch. It is never left spinning unattended.
Out comes the failed head stack; in goes a donor set matched to your model. A spindle that has locked gets freed. Where heads are resting on a platter, they are raised clear of it without being dragged over anything. Only once all of that is done is the drive asked for one thing: a full image, and nothing besides.
A hardware imager works gently. It steps around whatever stalls, banks the straightforward sectors first, then comes back for the difficult regions until the copy carries as much as the disk is willing to give up. Each step after that is taken against the copy, and nothing is ever written to your drive.
You see the file list first, then decide. Nothing is invoiced until you have said yes, and on most jobs there is no fee at all if the data does not come back. What we pull off goes onto media bought in for your job, and the postage home is ours. The case stays open on the bench until you tell us the files open on your own machine.
The whole page reduces to one word: stop. A head flies nanometres off a platter turning at several thousand revolutions a minute, and nothing engineered that finely repairs itself. On the noisy drives reaching this bench, how fast somebody got to the plug is usually the whole difference between everything back and a partial list. Post it, an engineer has it open the day it arrives, and a single figure comes back to you in writing before anything chargeable begins.
A Tilehurst shop sent one in that ticked at roughly a second's interval, with five years of takings held on it. Donor stack number two held, the image was left to take whatever time it wanted, and the accounts opened again — the VAT return still not filed.
Probably not. A click is the heads failing to find their position, and it tells you nothing about the platters underneath — which is where the files are. Donor heads and a slow, patient image usually return the greater part of a disk. The damage is done by the minutes it spends running once the noise has begun.
It does no good, and it does real harm. The advice predates every drive still in service. Cool a sealed assembly and two things happen together: the metalwork contracts by different amounts, and moisture condenses where nothing wet belongs. Bring it back up to room temperature and you have corrosion on the contacts on top of whatever was already wrong — two faults to work round now instead of one.
The factory lays servo marks down on every platter, and the arm goes hunting for them. When nothing answers it retreats to the ramp and sweeps again; that cycle, repeating, is the sound you can hear. What the sound will not tell you is the cause — a head out of true, a preamplifier failing, a surface carrying contact marks: through a metal case all three are indistinguishable. Separating them means having the cover off on a bench.
None of it helps. Every request the program makes still has to travel through the heads that have failed, and the only effect is to keep the drive doing the one thing wearing it away.
You pay nothing for the look. What comes back is a list — the files that opened, the ones that did not, and one fixed figure to finish the job. Leave it switched off until you have that in writing.