A Drive That Has Started Clicking

Power it off, and do not plug it back in until you have decided what you are doing. A clicking mechanical drive is a mechanical fault in progress, and every spin-up spends recovery odds that do not come back.

The instinct is to plug it in and investigate. That instinct is the reason a good proportion of these end badly.

A note on this guide: this rests on how mechanical drives are constructed and on how the failure progresses, rather than on drives failed deliberately here. See how we test.

The problem, as it appears

A rhythmic click, every few seconds, from a machine that was fine last week. Sometimes with the drive still visible in the operating system, sometimes not. Sometimes it works for ten minutes and then stops responding.

The uncertainty is the dangerous part. A drive that half-works invites experimentation, and experimentation on a failing mechanical drive is expensive.

Why the obvious first steps disappoint

Because almost all of them involve running the drive, and running the drive is the thing that is consuming what you have left.

Diagnostics read the entire surface. A repair utility reads the entire surface and then writes to it. Copying everything off in one large clone starts at the beginning and works through, which is fine until it fails at eighty percent — and you have spent the drive's remaining life retrieving things in the order they happened to be stored rather than the order you cared about.

The second reason is that "it seems to be working now" is not information. Intermittent function is characteristic of this failure, and reading it as recovery is how people postpone the decision until it is made for them.

What actually matters

Decide what the data is worth before you touch anything

This single question determines everything that follows, and it is not really a technical one.

Irreplaceable — photographs, documents, anything with no other copy. Stop now. Do not power it on again. The next step is a professional recovery quote, and the reason to stop immediately is that every additional attempt reduces what they can do.

Inconvenient but replaceable — a games library, media that can be redownloaded, anything already backed up elsewhere. Then the drive is a component that has failed, the data is not the issue, and you can be as casual as you like.

Most drives contain a mixture, and people treat the whole drive as the second category because most of it is. It is worth naming the small part that is the first one.

One attempt, made properly

If it still mounts and the data is worth having, you may get one usable session. Spend it on the irreplaceable things, smallest and most important first — not on a full clone.

Copy to a different physical drive, never to another partition on the same one. Do not run repair tools first. Do not defragment. Do not let the operating system "check and fix" the volume when it offers, because that writes to a drive you are trying to read.

If it does not mount, stop there. Repeated mount attempts on a drive with a head problem are the mechanism by which a recoverable drive becomes an unrecoverable one.

The freezer trick belongs to older hardware

It circulates because it occasionally worked on drive construction that is now decades old.

On a modern sealed assembly, condensation is the outcome, and it is not something that reverses. It also contaminates the internals in a way that makes professional recovery harder and sometimes impossible — so the cost of trying it is not zero, it is the option you had left.

An SSD fails differently and gives less warning

No moving parts, so no clicking. A failing SSD typically goes read-only, disappears from the system entirely, or reports a rapidly falling health figure.

Read-only is the good version: the data is usually still readable, and it is a clear instruction to copy it off now. Sudden disappearance is the bad version, and there is often no preceding symptom at all — which is precisely the argument for backing up on a schedule rather than on a symptom.

The real lesson is upstream

A clicking drive is a bad time to discover your backup arrangement. Everything above is damage limitation, and the amount of it available is decided by choices made months earlier — backing up more than game saves is the version of this written before the emergency.

The options, and who each suits

This page contains affiliate links. If you buy through them, Ugly Fixes may earn a commission at no extra cost to you. Products are selected independently.

Stopping and getting a quote suits anyone whose drive holds something irreplaceable. It costs nothing to ask, and it is the only path that does not spend the remaining margin.

A dock or adapter to read it once suits a drive that still mounts, holding things you would like but could live without. Connecting it externally means the machine is not depending on it to boot, which is one less thing running.

A replacement backup drive suits everyone reading this, and it is the purchase that makes the next occurrence uneventful rather than serious.

A NAS suits a household with several machines and a genuine need for a copy that is not in the same box. It is a larger commitment and a real answer, provided nobody mistakes redundancy for backup — a mirrored array faithfully mirrors a deletion.

The mistake to avoid

Running it to see whether it gets better. It does not get better, and the intermittent stretch where it seems to be recovering is part of the failure rather than a sign of the end of one. Each power cycle in that stretch is spending the same finite resource — head passes over a surface that is being damaged.

The second mistake is doing a full clone instead of prioritising. If the drive dies at eighty percent, what you have is whatever happened to be stored in the first eighty percent of the addresses, which has no relationship to what you would have chosen. Copy the things that matter first, in order of how much you would mind losing them.