The drive knocked off a desk is always the one holding the only copy. Portable units, mains-powered ones, every badge in the shops: they turn up here ticking, beeping, blinking or entirely silent, and they go back as files on fresh media. One instruction on this page counts for more than the others, and enough people get it wrong that it belongs at the top. The unit travels whole, in its own case, with the lead. Do not open it up first.
Looking at an external drive costs nothing. What comes out of that examination is one written figure, agreed between us before anybody picks up a screwdriver: £300 + VAT for one hard disk or SSD, an NVMe blade counting as a single drive.
Logical work is no fix, no fee. Four things sit outside that: electronic and mechanical failures, chip-level work, DVR jobs and forensic work, and anything physical is half up front. All five bands are laid out on the data recovery cost page, while data recovery services follows a postal job from the parcel to the answer.
Tracing a symptom back to whatever caused it is where the job genuinely starts, and these thirty cover very nearly every box opened on the Oxford bench. A fault that is not on the list is not an unfamiliar one: describe it over the telephone and you will get a straight reading of the odds before you have spent a penny on postage.
What you can hear is a start-up attempt failing and then being repeated. Two quite different things produce it. On a unit drawing its power from the socket it plugs into, the port may not be delivering enough current at the instant of spin-up, and a rear socket on the machine itself, or a hub with its own supply, is one fair test of that. Where the spindle has genuinely seized in its bearing, repetition changes nothing and every cycle is wasted.
A portable drive that falls while its platters are turning is the classic head crash, and the ticking is the heads searching for a reference they can no longer find. Note the fall when you send it, because it changes what the bench does first and there is nothing to be gained by leaving it out. What does harm is powering it up again to see whether it survived.
Power is reaching the unit and no storage is being presented. That pattern favours the bridge board inside the case rather than the disk, which is much the better of the two outcomes. It is also indistinguishable from a dead disk without opening the case, which is why the diagnostic exists and why it costs nothing.
Two changes are worth making before anything is posted, and both are free. Swap the lead for a known-good one, because worn micro-USB sockets and cheap cables cause a surprising share of these. Then try a machine running a different operating system, since a unit invisible to Windows sometimes appears perfectly in Disk Utility on a Mac.
It appears, takes a letter, reports no usable size and asks to be formatted before use. Decline. That message means the file system header could not be read, which is a different matter from the files having gone, and accepting writes a fresh empty structure over the map describing where everything lives.
A copy starts, reaches a particular file and stops, and the whole machine becomes unresponsive with a greyed-out window. The disk has met a region it cannot read and is retrying it for as long as anyone will wait. Left running it will spend the rest of its readable life on that one region.
The socket on the drive has worn or its solder joints have cracked, which is common on the small connectors used by portable units. It is a repair rather than a recovery, but it is done here as part of the job, and it is worth stopping immediately rather than nursing it through one more backup.
Desktop units that need their own power supply fail at the supply far more often than at the disk. Where it goes wrong is the replacement: an adaptor of the wrong voltage or the wrong polarity will put a healthy disk straight onto the bench as a board repair. Check the figures printed on the case before plugging anything else in.
No noise at all after a fall is not better news than clicking. It generally means the head assembly has come to rest on the surface, or the spindle has been knocked out of true. Both need filtered air and correct tooling, and both get considerably worse if the unit is switched on again in the meantime.
The disk attempts to start, fails a check, and repeats. It may be the firmware area rather than the mechanism, in which case nothing is physically wrong and the answer is a service terminal. It may equally be a motor that cannot reach speed. From outside the case the two look identical.
The commonest logical fault on this page by a wide margin. Pulling the cable while the operating system still has writes outstanding leaves the file system half updated, and the visible result is empty folders, a format prompt or a volume that mounts as RAW. The data is generally sitting untouched behind all three.
Backups made over a long period build a complicated structure of hard links, and damage anywhere in that structure can make the whole history unusable even though the underlying files are present. These are recovered by reading the volume as an image and extracting the files directly rather than by persuading the backup software to cooperate.
The partition is still declared and its contents can no longer be identified. Behind that is usually a damaged boot sector, a partition table edited by an installer, or an unreadable region sitting exactly where the file system header lives. All three are worked from a copy, and none of them should be handed to a repair utility first.
Taking a disk out of its case is popular and it is not always a step you can simply undo. A disk written while it sat behind a bridge presenting an unusual sector size has to be read back through something presenting the same thing, or every offset in the volume lands in the wrong place and healthy data looks like rubbish. Keep the case and the board, and put both in the parcel with the disk.
Bus-powered units are working to a tight power budget, and a disk that is drawing more current than it used to as it ages will trip that budget at spin-up. Sometimes it is the port. Often it is a mechanism starting to bind. A powered socket is worth one attempt; a repeated cycle of attempts is not.
A hard disk in an enclosure does not usually receive the background erase instruction that an internal solid state drive gets, so deleted files often remain in place until something writes over them. Unplug it and leave it unplugged. Most of the harm in these cases is done in the hours afterwards, by people carrying on as normal.
Enclosures give a disk almost no cooling, and desktop units in cupboards and portable units under paperwork run hotter than anything ever intended. Heat shortens a mechanism's life and produces exactly the intermittent behaviour that people mistake for a cable fault. If a unit is too hot to hold, it has been telling you something for a while.
Rugged portables, d2 towers and the twin-bay boxes all house perfectly ordinary mechanisms behind the aluminium, and a healthy share of them arrive with a dead interface board or a failed supply rather than a failed disk. Where the electronics turn out to be the answer, the job is quick and what comes back is everything. The dedicated LaCie page works through the range in detail.
Somebody changed the file system so that a Mac could write to it, which discards the index rather than converting it. If nothing has been written since, the recovery is usually straightforward. If a month of work has gone on top, it is not. Stop using the drive from the moment you realise.
An external kept permanently connected as a backup is treated by ransomware as just another volume, which is why it is so often encrypted along with the machine. This is charged as ordinary media at 300 pounds + VAT rather than as forensic work. What can be recovered depends on the variant and on what shadow copies survived.
The interface disappeared, not the data. FireWire 400 and 800, eSATA, early USB and the proprietary connectors two or three makers used for a season are all still handled here. A drive that cannot be plugged into any modern computer is not a failed drive, and it does not need to be treated as one.
Dual-bay desktop units were frequently configured for speed rather than safety, which means one failed member takes the entire volume with it. Both disks are needed, both are imaged before anything is assembled, and the job is priced as an array from 500 pounds + VAT. The page on striped sets explains why one lost member ends the volume.
Damp, cold and long periods standing still are hard on a mechanism. Heads bond to the surface, boards corrode, and lubricant migrates. Units that come out of storage around the Warwickshire villages are usually recoverable, and the one thing that reduces the odds is plugging them in repeatedly to see whether they wake up.
That is not an improvement. A drive that clicked and then went quiet has generally stopped trying, which usually means the heads are no longer in a state to attempt a read at all. The window for the least invasive work has closed, and what remains is bench work with donor parts.
A deep scan reads the entire surface repeatedly, which is the worst possible use of a failing disk's remaining hours. It is not fatal and it is worth declaring, because it tells the bench how much reading the drive has already endured. What genuinely harms the job is software writing its output back onto the drive it is reading.
A unit presenting the wrong size has usually lost its translator or is being read through a bridge doing something unexpected with sector sizes. The disk itself is often untouched. Do not accept any offer to initialise or repair it, because both write to a disk whose existing structures are merely unreadable.
Roughly half of these units turn out to be the case and roughly half turn out to be the disk, and when it is the disk it is exactly the same work as any internal drive: heads, boards, firmware, donors. The price does not change with the answer. One external drive is 300 pounds + VAT whichever of the two it was.
One very widely sold family cracks where the USB socket is anchored to the board, which severs the connection and takes the unit off the bus without any warning at all. Nothing has spun, nothing has crashed into anything and no noise was ever made. The memory is normally in good order, the repair is done at the microscope, and one clean image is all that is asked of the unit afterwards.
Consumer devices frequently reformat a drive into a layout only they understand, and some encrypt it to the device. A drive prepared by a set-top box or a recorder may be perfectly healthy and unreadable on a computer. Say what it was attached to, because that single line usually settles the method before anything is opened.
This is the story behind most of the units on this bench: the files were moved off a laptop to free up space, and the drive they were moved to is the one that failed. A single external holding the only copy of anything is not a backup, it is a second single point of failure with a longer cable.
Almost everything else on this site is posted as a bare drive. An external unit is one of the two exceptions on Oxford's shipping form, and there is a practical reason behind the rule. An external is two things in one box: a disk, and an interface board that translates between that disk and the USB lead. Either can stop, and a fair proportion of the units arriving here have a perfectly healthy mechanism sitting behind a board that has died, a failed power supply or a damaged socket. Working out which of the two failed takes ten minutes with both parts in front of you and considerably longer without. Take the disk out yourself and you throw away half the evidence, and on some units you throw away more than that, since a good number of enclosures apply their own handling to the disk behind them without saying so on the box. So the unit travels as it stands, with the lead and the mains adaptor if it has one. That also means no screwdriver, no prising and no weekend spent proving that a disk taken out of its case reads as nonsense.
Portable units are 2.5-inch drives powered down the USB lead. They get carried, so they get dropped, and they are almost always spinning when it happens. They are also sensitive to a tired lead or an underpowered port, which produces the classic intermittent unit that works on one machine and not another. Desktop units are 3.5-inch drives with their own mains supply, and those fail differently: heat inside a sealed plastic shell, a supply that ages before the disk does, and enough weight that a fall from a desk is decisive. Portable solid state drives are a third case and sit in the same £300 + VAT band, with the usual TRIM caveat on anything deleted. Two things move the price. A dual-bay box configured as a stripe is an array, so it starts at £500 + VAT, since one missing disk takes the volume with it and both members have to be copied and put back together. Chip work on a portable SSD takes half up front and sits outside no fix, no fee. Old interfaces are no obstacle: FireWire, eSATA and the PATA-era boxes are still read here.
A repeating tick says the heads have lost their reference on the platter surface, and every switch-on to see how bad it is spends a little more of what could have been recovered — that is how a head replacement becomes a scored disk. A beep at one-second intervals is a different fault carrying the same instruction: the motor is trying to turn and cannot, because either the spindle has seized in its bearing or the heads have settled on the platters and are holding everything in place. Neither improves with persistence, and neither responds to any of the folk remedies that circulate — the freezer, a sharp tap on the desk, one more lead. Both are repaired properly in filtered air with a matched donor stack, and how much comes back depends almost entirely on how many times the unit was asked to try again before somebody stopped asking. Unplug it. If the badge on the case is LaCie there is a page here devoted to the range, because the name on the box and the mechanism inside it come from different companies.
There are two machines inside every one of these cases, and the first question is always which of the pair fell over. The benches below settle that question without a single byte being written to your storage, and then handle whichever answer comes back.
USB 2.0 through USB-C, FireWire 400 and 800, eSATA, Thunderbolt, and the proprietary connectors a few makers used for a season. A unit that will not talk to a modern computer very often talks perfectly well to something on this shelf, and establishing that takes minutes and costs nothing.
Desktop enclosures fail at the power supply more often than at anything else, and the wrong replacement adaptor is one of the quickest ways to convert a healthy disk into a board repair. Voltage, polarity and current are checked against the case markings before a unit is powered at all.
Head replacements, freeing a stack that has settled onto the surface, and platter transfers all happen inside a filtered enclosure. The disk inside a portable case is an ordinary laptop mechanism and it is treated as one, with the same tooling and the same donor requirements.
Copies are taken head by head with current draw watched throughout and timeouts enforced at sector level. A drive that would lock up a file manager for twenty minutes on one region is skipped, logged and revisited at the end, once the straightforward bulk is already safe on fresh media.
Working inside the part of a disk the machine is never shown: rebuilding a collapsed translator, repairing defect tables, pulling modules out and writing corrected ones back. Plenty of units written off as dead hold a disk whose firmware fell over rather than a mechanism that failed, and there is no way whatever to tell those apart from outside a sealed case.
Burnt regulators, shorted protection diodes and cracked tracks repaired under a microscope, with the disk's own adaptive data moved across wherever a board has to be changed. The board that came in the case stays with the disk it came with throughout the job.
A disk written behind a bridge that presented 4096-byte sectors has to be read back the same way or nothing lines up and a healthy volume looks like rubbish. Identifying the original translation is a routine step here and an impossible one at a kitchen table, which is another reason the case comes with the disk.
Put the entire thing in the parcel as it stands, lead and mains brick included, and leave the screws alone. Deciding whether the fault lies in the box or in the disk accounts for a good half of the diagnosis, and it is a five-minute question when both parts are on the bench together and a much longer one when they are not. Oxford's shipping form arrives at the same place from the other direction: a laptop, a desktop, a Mac or a recorder posts its drive alone, whereas an external in its own enclosure comes whole, and so does a NAS. One external is 300 pounds + VAT, and a pocket SSD is charged the same way. Put two disks in one case and configure them as a stripe and it becomes an array instead, from 500 pounds + VAT. Looking at it costs you nothing and finishes on the second working day after the unit is logged in. Logical faults are covered by no fix, no fee; anything mechanical or electronic takes half the figure in advance and is named in the published exclusions.
More units arrive wearing this badge than any other, and they divide neatly. My Passport is the 2.5-inch bus-powered drive that lives in a bag, gets carried between home and work, and gets knocked off desks; inside it is a laptop mechanism working to a tight power budget with almost no cooling. My Book is the 3.5-inch desktop unit with its own mains adaptor, and on those the adaptor fails at least as often as the disk. Elements is the plainer line and the same mechanisms turn up inside it. Whichever shape yours is, it comes here whole and sealed with its cable and its power supply, because the board in the case is part of the diagnosis and half of these units turn out to be the case rather than the disk. The WD ranges have a page of their own on this site that goes through them properly.
The pattern here is simpler than on some badges: put a shell around an ordinary mechanism and sell it. Fail one and the odds tip towards the disk rather than the interface, so it goes onto the mechanical bench and is treated as any bare drive would be. Bags are what shortens their lives. The slim 2.5-inch units go everywhere with their owners, and the small mechanisms inside are built to notebook tolerances with thin platters that do not forgive being dropped onto a hard floor. Once one of these has begun to tick, it needs a donor stack matched by family and revision, fitted in filtered air. That is 300 pounds + VAT like any other single drive, with half of it settled before the bench starts, since the parts are bought specifically for your example and are of no use on anything else. The internal ranges are set out at greater length on the Seagate page.
Two decades of selling into photography, film and design work show up in what arrives: these units are usually somebody's business rather than a family album, and the owner is on the telephone the same morning. The orange rubber-cased Rugged is the one that travels in a camera bag; the d2 sits under a desk as a tower; the 2big holds two disks and was configured for speed far more often than was wise. What is underneath the aluminium is an ordinary Seagate or Toshiba mechanism. The encouraging part is how frequently the fault is not the mechanism at all but the interface board, the Thunderbolt controller or the supply, which is the cheapest and quickest ending available here. A dedicated LaCie page on this site takes the range apart model by model.
Anyone who carries work between sites has probably stopped using a portable hard disk, and the Samsung and SanDisk pocket drives that replaced them break in ways the old ones never did. No spindle to seize, no head to land on a surface, no noise of any kind. Three things go instead: the bridge that presents the drive over USB, the controller that maps the flash, and, on one very widely sold family, the join where the board meets the socket, which fractures and takes the unit off the bus in an instant. In all three the memory itself is normally in good order. Work of this kind happens on the flash bench, at the ordinary single-drive figure of 300 pounds + VAT, and where the packages have to be taken off the board it becomes chip-level work: half up front, and outside the no fix, no fee guarantee.
Almost everything worked on here arrived in the post. A drive that is already in trouble has an easier time boxed, padded and insured than it does being carried round in a bag for an afternoon, and a parcel handed over in Coventry today is normally booked in at Oxford tomorrow morning.
As a rule the storage comes out and the machine stays where it is. That applies to a laptop, a tower, an iMac and to the recorder sitting under a counter. Taking equipment apart is not something this bench does, and a repair shop will free a drive in a few minutes. Two things go the other way: an external drive stays sealed inside its own case, and a NAS travels as a complete unit with its disks still in their bays. A Fusion Mac is a third case — both of its drives come out and travel together, each one labelled. The single situation nobody can work around is memory soldered flat onto a mainboard, which is how Apple Silicon Macs and a good many slim laptops are built: if the storage will not unbolt, there is no parcel to send.
↓ Print the shipping & booking-in form (PDF)
Put Oxford Data Recovery on the label. From Coventry it is roughly fifty-five miles straight down the M40, about an hour if you would rather drive it in than post it. Either way you are told the moment it is logged, and the free diagnostic finishes two working days later.
Unsure what ought to go in the box? Ring 0800 689 0668 before you seal it, or work through the free online diagnostic and let it do the asking.
What repair can mean on a portable, and where it stops meaning much
Four suspects — lead, socket, bridge, disk — and how to narrow them down
Elements and My Book units, and the chip between USB and the platters
A beep every few seconds is a spindle that will not turn over
TeraStation and LinkStation boxes, and the file system found inside
Take the shell off and it becomes a straightforward disk job
Looking at it costs nothing, one written figure follows, and the band governing this page is £300 + VAT for one hard disk or SSD, an NVMe blade counting as a single drive.