A clicking hard drive is a drive whose read/write heads can no longer read the firmware stored on the platters: the circuit board then returns the heads to the stop of their parking zone to retry the read, and this back-and-forth produces the clicking. The noise itself is not a breakage; it is a recalibration. But if a faulty head touches the surface, each cycle of clicks grinds away a little more of your data. The right course of action fits in one sentence: unplug the drive immediately, and limit yourself to one or two connection attempts in an external USB enclosure before having the device analysed in a laboratory.

That is the essential point. The five minutes of reading that follow explain what really happens inside, because what you do over the next few hours will probably decide the fate of your data.

What happens when a hard drive starts up

To understand the click, you first need to understand the normal start-up of a hard drive.

When powered on, the circuit board sets the platters spinning, those mirror-polished metal discs that carry your data. Once cruising speed is reached, the read/write heads leave their rest position and move over the platters. Their first mission is not to read your files: it is to read the drive's firmware, the internal program that describes to the controller how the data is organised on the platters (tracks, sectors, locations).

Contrary to what one might imagine, a hard drive's firmware is not stored in a chip, but largely on the platters themselves. On current drives, it amounts to about 100 to 200 MB. Without it, the circuit board is unable to know where anything is located.

Start-up therefore follows this sequence: the platters spin, the heads position themselves thanks to markers they read on the surface, they load the firmware, the drive initialises, and it is finally ready to respond to the computer, which in essence asks it: "identify yourself." Only after this handshake do your files become accessible.

Why it clicks: the mechanics of the clack

What happens when the firmware read fails?

The circuit board itself only has a tiny program stored in its ROM (the equivalent, for a hard drive, of a computer's BIOS). Faced with the failure, it applies the only strategy it knows: it returns the read/write heads to the parking zone, the area where the heads rest and are stowed when the drive is switched off, then sends them back to attempt to read the firmware at the intended location. And it starts again, over and over.

The famous "clack" is the sound of the heads returning to the stop of the parking zone. The clacking is therefore not a part breaking: it is a recalibration movement. The drive hopes that, by starting again from a neutral position, the next read attempt will succeed.

Depending on the model, this behaviour differs. Some drives perform cycles of three to five clicks, then stop of their own accord, in a kind of safety mode. Others start over indefinitely, and can click for days if no one unplugs them.

One and the same noise, four possible causes

The click is a symptom, not a diagnosis. Behind the same noise lie very different failures, with very different prognoses.

Cause	What happens	Recovery prognosis
Faulty read/write heads	The heads can no longer read the firmware, the board loops	Good: head replacement in a cleanroom, then extraction of the data
Firmware corrupted (logically)	The content read is unusable, the drive does not initialise	Variable: workarounds possible by loading other firmware into the board
Firmware physically damaged	The area of the platters carrying the firmware is damaged in its substance	Limited: sometimes bypassable, otherwise recovery impossible
Combined case: heads + surface	Broken heads scrape the platters and also destroy the firmware zone	Poor: if the platters are too damaged, recovery is impossible

First case: one or more faulty read/write heads. This is the most favourable scenario: a head replacement in a cleanroom, with compatible parts, makes the drive functional again long enough to extract the data.

Second case: corrupted firmware. The heads work, but the content of the firmware comes back corrupted, unusable by the circuit board. Depending on the extent of the corruption, certain problems can be worked around: loading other firmware into the circuit board to access the readable areas and reconstruct the information that says where the data is located.

Third case: physically damaged firmware. It is sometimes possible to bypass the problem; when that is not possible and the firmware cannot be repaired, recovery becomes impossible.

Fourth case, the most formidable: the combination. Faulty heads that touch the surface damage the platters, including the firmware zone. If the damage is too extensive, replacing the heads is no longer of any use: the new heads would in turn break on a devastated surface.

A few nanometres between your data and disaster

Why does a faulty head cause so much damage? Because of a feat of engineering that is also a fragility.

In normal operation, the read/write heads never touch the platters: they fly over them at a few nanometres. The rotation of the platters creates an airflow at the surface, and this "wind" carries the heads as if on an air cushion. A head literally surfs above your data.

A broken head loses this lift. It touches the surface, and at several thousand revolutions per minute, it does not scratch: it scrapes, it grinds. With each rotation, it tears away a little of the magnetic layer that carries your files. Every restart "to see if it works" is Russian roulette: the more a drive clacks, the more attempts you pile up, the more the surface is eaten away by the heads.

A case handled in our laboratory illustrates how far this can go. A RAID system was installed in a rarely visited and hard-to-reach technical room. First one drive failed, then another; by the time a technician arrived on site, the drives had kept spinning, and clicking. On opening them in the cleanroom, the drives were destroyed to a degree rarely seen: the mirror surface of the platters had completely disappeared, ground to dust inside the enclosures. The platters had become transparent. Data ground to dust cannot be recreated: there was nothing left to recover.

What you must not do (even if the Internet advises it)

Faced with a clicking drive, the forums overflow with recipes. Almost all of them make the situation worse.

Opening the drive "just to look inside." Useless and illusory: there is nothing to see that could help, and everything to lose. Outside a cleanroom, the slightest speck of dust that settles on a platter becomes an obstacle on which the heads will catch.

Swapping the circuit board. On modern drives, the board contains adaptation data unique to each unit. Swapping the board with that of an identical drive does not resurrect the drive, and wastes time during which further start-ups are attempted.

Hitting or tapping the drive. No.

The freezer. The legend dies hard, so let us be precise. Cold does indeed cause materials to contract and expand. But above all it produces condensation inside the drive, because a hard drive is not completely airtight, and that is deliberate: breather vents ensure atmospheric pressure balance. Condensation means water on your platters. You do not cure a failing drive by filling it with moisture.

Leaving the drive running in hope. You now know the mechanism: each cycle of clicks is one more opportunity, for a faulty head, to grind away the surface. Drives that click indefinitely can self-destruct within a few days.

Attempting a head replacement yourself. Impossible to carry out cleanly without specialised tooling and without a cleanroom. And even if successful, the work is only beginning: a drive with replaced heads remains unstable. It must be stabilised, its firmware worked on, and the data extracted through progressive bit-by-bit copies, sector after sector. It is laboratory work, not screwdriver work.

The only thing you can try

✅ Allowed	❌ Forbidden
Unplug the drive immediately	Open the drive "to look inside"
One or two connections in an external USB enclosure, unplug immediately if it clicks	The freezer (condensation = water on the platters)
Copy the most important files first if the drive mounts	Swap the circuit board
Have the device analysed in a laboratory	Hit or tap the drive
	Leave the clicking drive running
	Attempt a head replacement yourself

There remains a reasonable margin of manoeuvre for a private individual, only one.

Take the drive out of the computer and place it in an external USB enclosure. Connect it: if the drive mounts and your files appear, copy the most important ones immediately. If it clicks, unplug at once. You can make a second attempt, but not a third. One or two connections at most, and then stop.

If the click persists, the rest plays out in a laboratory. Each additional attempt reduces the chances of recovery.

Frequently asked questions

Why does my hard drive go click click?

Because its read/write heads can no longer read the firmware stored on the platters. The circuit board then returns the heads to the stop of their parking zone and retries the read: this repeated back-and-forth produces the clacking.

Is a clicking hard drive dead?

Not necessarily. If the cause is a faulty read/write head, a head replacement in a cleanroom often makes it possible to recover the data. If, on the other hand, the heads have scraped the surface of the platters or destroyed the firmware zone, recovery can become impossible. Only a laboratory analysis can settle it.

Can I still connect my clicking drive?

One or two attempts at most, in an external USB enclosure, unplugging immediately if the clacking resumes. Each additional cycle of clicks gives a faulty head the opportunity to grind away a little more of the surface that carries your data.

Can the freezer repair a clicking hard drive?

No. A hard drive is not completely airtight (breather vents balance the atmospheric pressure): cold creates condensation inside, that is to say water on the platters. The freezer makes the failure worse instead of repairing it.

How long can a clicking drive hold out?

Some drives stop of their own accord after cycles of 3 to 5 clicks; others click indefinitely, for days. A drive whose faulty head touches the platters can self-destruct within a few days of operation, to the point of grinding the magnetic surface to dust.

What will the laboratory do with my drive?

First an analysis: opening in a cleanroom to establish the cause of the click and the actual state of the platters, then a report and a quote. If the surface is sound, a head replacement offers good prospects; if the damage is too great, an honest laboratory tells you so before you spend anything at all.