Straightforward, device-by-device guidance on what to do — and what to avoid — the instant a drive gives out.
More data is lost in the aftermath than in the failure itself — by carrying on regardless. Stopping is the most useful thing you can do.
Everything else on this page depends on the answer, so start here.
Minutes — the device is making a noise. Clicking, grinding, scraping or a buzz with no spin. Stop reading and switch it off; damage accumulates every rotation and nothing on this page matters more than that. Then read what the noise means.
Hours — files were deleted from an SSD. TRIM erases the blocks in the background while the drive has power. Shut the machine down now; what survives is whatever the drive had not yet reached. More on why SSDs are different.
Days — the drive is slow and throwing errors. Still readable, and that window closes without warning. Copy what matters off now, most valuable first. How to check whether it is failing.
Weeks — deleted files on a healthy hard drive. Nothing happens until something overwrites them, so stop using the drive and you have time to think. What actually works.
No hurry — the drive is dead but silent. A drive that does not spin is not deteriorating. Electronics and firmware faults leave the platters untouched, and there is no clock running.
Rather than a device list — the same four questions apply to a memory card and a twelve-disk array.
The wording is more diagnostic than people realise.
Roughly speaking, software handles what the drive can still read. Everything else needs hardware.
If a drive mounts, spins quietly and simply cannot find your files, that is a software problem — a deleted folder, a lost partition, an accidental format, a corrupt index. Reputable read-only tools installed to a different drive resolve a good proportion of those for a fraction of any lab fee, and we would rather tell you than take the work. Start with what the free tiers actually cover.
The boundary is physical condition, not difficulty. The moment a drive is clicking, dropping off the bus, stalling for thirty seconds at a time or reporting the wrong capacity, software becomes actively harmful: a deep scan is hours of continuous reading on hardware with limited reads left, and drives regularly die partway through one intended to save them. There is no version of that software problem that patience solves.
The four questions that place your case on one side or the other are in software or a lab, and our prices are published so the comparison is a real one.
Each still circulates, and each has a reason it sounds plausible.
This one has a kernel of history — on drives from thirty years ago, cooling occasionally shrank a seized component enough for one more spin. On anything modern it introduces condensation onto the platters as the drive warms, and moisture where the heads fly is destructive. It was marginal then and it is harmful now.
Plausible because the boards look identical and are held on by four screws. Modern drives store adaptive calibration unique to each individual drive in a ROM chip on that board, so a donor without the original ROM contents either does nothing or drives the heads with the wrong calibration.
It genuinely repairs file systems, which is why it is recommended so often. What it does is produce a consistent file system — by deleting or truncating whatever it cannot reconcile, which on a failing drive is precisely the records a recovery needs.
Popular because it appears to work on phones, which mostly survive brief immersion regardless. Rice removes some ambient moisture and nothing at all of the dissolved minerals and salts that actually corrode the board. Sending a drive damp and sealed is genuinely better than drying it.
Every recovery opens with a free written diagnostic. We'll tell you what can come back and what it'll cost before a penny becomes chargeable — and for most jobs, no fix means no fee.