The Dead Hard Drive: What Is Recoverable and What Is Not
A hard drive rarely dies all at once. It gives a warning that most people do not recognise as one, and then, one morning, the computer will not boot or the external drive is not there any more. What happens in the next ten minutes matters more than anything that happened in the years before, so this article is written for that moment.
The short version: most dead hard drives are recoverable, a minority are not, and the difference is usually decided by what the owner does before the drive reaches a lab. The rest of this piece explains how to tell which situation you are in.
What is inside, and why it matters
A hard disk drive is a small precision machine. One or more glass or aluminium platters spin at 5,400 or 7,200 revolutions a minute. A read head, thinner than a human hair, floats above each platter surface on a cushion of air about three nanometres thick. The heads are moved by an actuator arm, the whole thing is driven by a circuit board on the underside, and a tiny area of the platter holds the firmware the drive needs to understand its own layout.
That gives four separate things that can fail: the electronics, the firmware, the mechanics, and the magnetic surface itself. Each one produces a different symptom, and each one needs a different kind of recovery. Working out which you have is the whole game.
The symptoms, and what each one means
Clicking, ticking or a repeating scrape
This is the sound everyone has heard about, and it is the most serious one. A regular click is the actuator arm sweeping across the platter, failing to find the servo tracks it needs to orient itself, and snapping back to its rest position to try again. The heads are either damaged or reading nothing, and every sweep is a chance to scrape the platter surface.
A clicking drive is a mechanical failure. It cannot be fixed with software, and it cannot be fixed by any method that involves leaving the drive powered on. It needs the heads replaced in a clean room, using a matching donor drive, before the platters are read at all. Done properly, recovery rates from clicking drives are high. Powered on repeatedly at home, the same drive turns into an unrecoverable one in an afternoon. If you hear this sound, unplug it now and read the rest later.
Spinning up, then spinning down, or a beeping sound
A drive that starts, stalls and tries again is usually a motor or head problem. A high-pitched beep, especially from a portable drive, is very often the heads stuck to the platter surface, so the motor cannot get it turning. Both are clean-room jobs. Both are usually recoverable if the drive is not forced.
Completely silent, no spin, no lights
If nothing happens at all, the first suspect is the circuit board. Power surges, a wrong power adapter on an external drive, and plain component failure can kill the electronics while the platters and heads are untouched. This is one of the better outcomes: the board is repaired or swapped, the firmware chip is moved across so the new board understands the drive, and the data reads normally afterwards.
Check the obvious first. A different cable, a different port, and for an external drive a different power supply, rule out the failures that cost nothing. Do not open the enclosure or the drive.
Detected, but painfully slow or freezing
A drive that shows up but takes minutes to open a folder, or locks the whole computer up while it thinks, almost always has bad sectors: small areas of the platter that can no longer be read reliably. The drive keeps retrying them, which is where the freezing comes from.
This one is deceptive because the drive looks alive. Copying files off it yourself is the natural instinct, and it is often the wrong move, because a normal copy hammers the bad areas over and over and the drive gets worse while you wait. The right approach is to image the drive with hardware that reads the good areas first, skips and returns to the bad ones, and never gives up on a sector until it has been tried gently from several directions. That image is then the thing the files are recovered from, and the failing drive is never asked to work again.
Detected, but showing the wrong size, no partitions or asking to be formatted
The drive is talking, but its map of itself is damaged: either the firmware area on the platters or the file system. Firmware faults are common on some drive families and are repaired in the lab by working with the drive's own service area. File system damage is a software job, and recovery rates are excellent, provided nobody clicks Yes on the format prompt. That prompt is not a repair. It is the fastest way to make the next step harder.
Dropped, knocked, or had something spilled on it
A drop while the drive is running is the classic cause of a head crash: the head touches the platter and gouges it. A drop while powered off usually damages the heads without touching the platters, which is far better news. Either way, do not power it on to check. Liquid is a different problem; the electronics are the concern rather than the platters, and a drive that has been wet should go straight to a lab without any attempt to dry and test it.
The one rule that decides most outcomes
If a drive is making a noise it did not make before, or has failed and come back more than once, stop powering it on. Every start-up is another opportunity for a damaged head to score the platter, and a scored platter is the thing we cannot recover from. The drive with the best chance is the one that has been turned on the fewest times since the fault appeared.
This also rules out the folk remedies. Freezing a drive does nothing useful for a modern one and introduces condensation. Tapping it, opening it to look, or swapping the circuit board with one bought online without moving the firmware chip across, each takes a recoverable drive some distance toward an unrecoverable one. We see the results of all of these regularly, and the recovery that would have been routine becomes a long, uncertain one.
Software recovery versus lab recovery
There is a genuine place for recovery software, and it is narrower than the advertising suggests. Software works on a healthy drive with a logical problem: files deleted, a partition removed, a quick format, a file system that has lost its table. In those cases the data is physically intact and the software rebuilds the map.
Software cannot help a drive that clicks, will not spin, is not detected, or is throwing read errors. Run on a drive with bad sectors, it does the same damage as a normal copy, only more thoroughly. The honest test is simple: if the drive is behaving strangely in any physical sense, or the operating system cannot see it as a normal drive, it is past the point where software is the tool.
Our hard drive data recovery service covers both ends of this. The logical cases are quick and inexpensive. The physical ones go through imaging hardware and, where the heads or motor are the problem, a clean room with donor parts. What you do not get from us is a technician running consumer software on your failing drive and calling it a recovery.
What is genuinely not recoverable
It is worth being straight about the limits, because a business that promises everything is not one to trust with a drive.
- Heavily scored platters. A head crash that has been allowed to continue leaves grooves in the magnetic surface. Where the surface is gone, the data is gone. This is almost always the result of a clicking drive being powered on repeatedly, which is why the earlier rule is repeated so often.
- Overwritten data. If a drive was working and the files were written over, by reinstalling Windows onto it, restoring a backup to it, or using it for months after the deletion, the old contents are not underneath the new ones. Magnetic recording does not keep a history.
- Encrypted drives without the key. A BitLocker or FileVault drive can be recovered physically, but the result is encrypted data. Without the recovery key or password, it stays that way. Keep those keys somewhere other than on the drive they protect.
- Drives that have been through a previous failed attempt. Not always unrecoverable, but every earlier attempt by someone without the right equipment narrows what is left. If you are going to send a drive to a lab, send it first.
What an assessment actually involves
When a drive arrives with us it is not plugged into a computer. It is inspected, the electronics are checked, and it is listened to on a controlled power supply for a few seconds at most. From that we know which of the four failure types we are dealing with, whether donor parts are needed, and roughly how long the imaging will take. That becomes a written quote with a realistic estimate of what will come back, and nothing further happens until you approve it.
The assessment is free and takes one to two business days on average. Where a business cannot wait that long, the emergency data recovery option puts the drive at the front of the queue for a fixed fee. Either way, the recovery fee is only invoiced once we have confirmed that your data is recoverable, and if it is not, you do not pay for the attempt. A quote that asks for the full amount before anyone has looked inside the drive is a quote to be careful of.
One practical note: tell us what matters. A recovery that gets back the tax folder and the photographs is a success even if the downloads folder is lost, and knowing the priorities lets us spend the drive's remaining life on the right areas first.
Solid state drives are a different story
Everything above is about spinning drives. An SSD fails differently: no noise, no warning, and usually a sudden disappearance caused by a controller or firmware fault rather than physical wear. The recovery techniques are unrelated, involving the flash chips and the controller's mapping rather than platters and heads, and the outcomes vary more by model. If the drive in question is an SSD, or you are not sure, the SSD data recovery page explains how those cases are handled. The one rule that carries over is the same: stop using it.
Afterwards
A recovered drive is not a repaired drive. Whatever we recover is delivered on new media, and the original is returned to you as a dead part, because a drive that has failed once and been opened is not something to trust again. That is also the moment to fix the thing that made this an emergency in the first place. A drive holding the only copy of anything is one fault away from this article, and the automated backup that would have made the failure an inconvenience costs less than a single recovery.
If you are reading this with a drive on the desk that has just failed, the sequence is short. Unplug it. Do not run software on it. Write down what you saw and heard. Then submit a job, and let it be assessed properly before anything else touches it. Most of the drives we see come back. The ones that do not were usually lost before they reached us.
