The sound is the head assembly being driven back to its reference position and trying again, over and over. It is the one symptom where the cost of ignoring it is measured in minutes. A worked example: an IBM Deskstar with a degree on it.
The drive cannot read its servo information, so it resets the heads and retries. If the heads are damaged, every one of those passes drags them across the surface holding your data.
Platters carry servo information — positioning data the drive uses to know where the heads are. Every read begins by locating it. If the heads cannot, the drive retracts the assembly to its stop, recalibrates and tries again, and it is that retraction hitting the limit that you hear. The clicking is not damage occurring; it is the drive attempting recovery. The damage is what happens alongside it if the heads are physically compromised.
Which is why this symptom is urgent in a way that slowness is not. A drive with bad sectors deteriorates over weeks. A drive with damaged heads deteriorates over minutes of runtime.
A steady click every second or two, indefinitely is the classic head or servo failure. Three or four clicks then silence and spin-down usually means the drive gave up and parked — often a firmware or service-area problem rather than physical head damage, which is comparatively good news. Clicking that began after a drop is almost certainly a head crash. A scraping or grinding note mixed in is worse again: that is contact with the platter surface, and it is doing permanent damage with every rotation.
None of these change the action. All of them mean power off.
Freezing. The theory is that contraction frees a stuck component. In practice condensation forms as the drive returns to temperature, and moisture where heads fly at a fraction of a hair’s width is destructive. It has never been reliable and it makes professional recovery harder.
Tapping or shaking it. Occasionally a drive that has been struck will spin once more, which is exactly what makes this dangerous — it encourages another attempt on a mechanism whose heads may now be sitting on the platter rather than above it.
The drive is opened in a filtered clean-air environment and the head assembly replaced with a matched set from a donor of the same family, firmware revision and head configuration. If the platters have been scored, that damage is permanent and defines the ceiling on what can be recovered — which is precisely why runtime after the clicking starts matters so much.
Then the drive is imaged in short controlled passes, prioritising the areas holding what you need. A single drive is from £300 +VAT after a free 48-hour diagnostic; donor work carries a 50% deposit with the balance payable only on success.
Because the heads cannot locate the servo information they need to position themselves, so the drive retracts the head assembly, recalibrates and tries again. The click is that retraction hitting its stop. It means the drive is failing to read, and if the heads are damaged each attempt causes more harm.
Not in any way that leaves it usable, and not at a desk. Recovery means opening it under filtered clean air and fitting a matched donor head assembly, then imaging it in short passes. The drive is not expected to work afterwards — the objective is your data, not the disk.
No. Condensation forms on the platters as the drive returns to room temperature, and moisture where the heads fly is destructive. It has never been a reliable technique and it actively complicates professional recovery.
You shouldn’t use it at all. Unlike bad sectors, which worsen over weeks, head damage worsens over minutes of runtime — every rotation with damaged heads can score the surface holding your data. Power it off and leave it off.
From £300 plus VAT for a single drive, fixed in a written quote after a free 48-hour diagnostic. Clean-air and donor-part work carries a 50% deposit, with the balance due only on a successful recovery.