Under a desk in most houses there is a tower holding every photograph the family owns, and in plenty of small firms the same box carries the books. When one stops, everybody assumes the disk went with it, and nine times out of ten only the computer did. Getting from a dead PC in Coventry, Southam or Atherstone to a padded envelope takes a side panel, four screws and two plugs. From there it is off your hands.
Looking at a desktop computer 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.
Power, lights and fans prove the computer works. Arriving at a black screen or a firmware message afterwards proves only that nothing on the disk looked like a place to begin, which is normally damaged boot code, a missing partition description or a drive that has gone quiet on the bus. The files are almost always still exactly where they were. A dead-looking machine and a dead disk are separate findings with separate answers.
Inside a steel case a failing disk sounds like a bearing somewhere else entirely, which is why the fan gets the blame for a fortnight first. What is really happening is the head assembly sweeping back to its stop and setting off again because it cannot work out where it is. Every start-up after that costs a little more of the surface.
Nothing is answering on the bus, and there are four ordinary reasons for it: a component on the board has failed, a diode has shorted deliberately to protect what is behind it, the motor has seized, or the heads are lying on the platter instead of flying above it. Each wants a different repair, and trying other cables and other ports tells them apart not at all.
A spike travels along the power lead and exits through whatever is plugged in at the far end. The drive's board takes the hit, very often at the protection diode, which shorts itself on purpose to save everything behind it. That is the best outcome a surge offers and it is repaired at a rework bench in a short session.
Water leaves salts and acids behind and they keep eating into a board for weeks after everything feels dry to the touch. A disk that has been soaked is stripped and cleaned before it is offered any power at all, since a wet board that is switched on takes healthy components down with it when it shorts. The radiator and the hairdryer are what finish these off.
Windows arrives with a clean set of structures and the old ones simply stop being referred to, though what they pointed at is still lying on the platters. How much of it survives depends almost entirely on how hard the machine has been used since, because every fresh file has to be put somewhere and the front of the disk is exactly where setup planted itself.
Mirrors on a workstation are meant to be watched and hardly ever are. One disk stops taking part, no warning is ever read, and the machine carries on with the remaining member for another year before that one gives up too. What is left is two copies of the same volume with a year between them. Both are imaged, and the most recent readable version of each file is picked out of the pair.
Either the boot sector has become unreadable or the table listing the partitions has been corrupted. Both live in the first few kilobytes of the drive, and neither has anything to say about the hundreds of gigabytes sitting behind them. Rebuilding them from an image is ordinary work.
The platters get up to speed, the controller goes looking for its own firmware in a reserved area of the disk, cannot get a clean copy of it, and parks itself rather than carry on blind. Mechanically the drive may be in perfect order. Getting back in needs a service terminal talking to it in a mode no operating system can reach.
An all-in-one hides the whole computer behind the panel, so reaching the disk means taking the machine apart rather than sliding a side off. The drive itself is a standard 2.5 or 3.5-inch unit once it is free. If the back resists, ring rather than lever, because a broken screen costs a good deal and contributes nothing at all to the recovery.
Supplies do not always fail quietly. Some put twelve volts where five ought to be on their way out, and the small parts on a drive board are the first things to give. The mechanism behind them is generally untouched, so a donor board with the original adaptive data carried onto it brings the disk back onto the bus. That is electronic work, half up front.
Cold, damp and stillness are hard on a mechanical disk. Bearings stiffen, heads can bond to the surface, and the board picks up condensation. That drive is inspected on the bench before it is offered power, because a single hopeful start is what converts a recoverable disk into a scored one.
A perfectly healthy data drive with a damaged file system presents itself to Windows as a disk that needs formatting, and the dialogue is worded as though agreeing is the sensible thing to do. It is not. Decline it, unplug the drive, and leave the volume in the state it is in.
Windows offers to check the disk on the next start and most people say yes. On a healthy volume with an untidy file system that is exactly right. On a disk that is failing physically it writes corrections onto weak sectors, keeps the heads working across damaged ground for hours, and moves what it cannot repair into numbered fragments with the names stripped off.
RAW means Windows can see a partition and can no longer identify a file system inside it. The boot sector, the master file table or the file system header has been damaged. The files themselves are usually all present behind that damage, and the correct next move is to stop using the machine rather than to accept any offer to repair it.
Initialising writes a fresh partition scheme onto a disk that already has one, which is why the prompt appears at the worst possible moment. Windows is offering it because it cannot read the existing structures, not because they have gone. Cancel the dialogue and take the disk out.
Firmware updates change how the board enumerates devices, and a module that was working before the update can drop off the list afterwards. Sometimes it is a setting, sometimes the controller has genuinely stopped. The memory behind that controller is unaffected either way, which is why these are worth looking at properly rather than writing off.
A machine that takes four minutes to reach the desktop and hangs for half a minute at a time is nearly always a disk relocating sectors it can no longer read on the first attempt. That is the most valuable warning a hard drive ever gives. Copy what matters off now, or send it now, because the state it is in today is the best state it will be in.
Deleting removes the pointer and leaves the clusters where they were, so a file deleted this morning is very often still complete this afternoon. What removes it for good is the operating system putting something else in the same place. Every hour the machine stays switched on lowers the odds, which is why the useful advice is to shut it down properly and stop there.
Towers on the floor of a garage, a unit or a fabrication shop breathe in whatever is in the air, and a clogged heatsink lifts the temperature of the disk with it. Drives that have run hot for years develop read errors long before they make any noise. These arrive from the trade estates around the city in a fairly steady stream.
Some high-capacity desktop drives write their tracks overlapping, which is efficient for storage and unhelpful under sustained writing. A big transfer runs at speed and then slows to a crawl as the drive reorganises itself. That is normal behaviour rather than a fault, but it hides genuine faults, and it makes imaging one of these slower than the size alone suggests.
Drive letters get reassigned, a storage driver changes, or a volume comes back marked offline. The data has not moved. This is one of the cases where the answer may cost nothing at all, and it is worth describing on the telephone before anybody pays for postage.
Everything is scrambled where it stands and there is usually nothing whatever wrong with the machine holding it. The outcome depends on the strain, on whether the restore points were destroyed as well, and on how much fresh data has landed since. It is charged like any other single disk at 300 pounds + VAT. Nobody here calls it forensic work or prices it that way.
Adding a second operating system is where partition tables get rebuilt, and one wrong choice in a setup screen removes the entry for a volume that is otherwise entirely complete. Nothing inside the volume has changed. The table is rebuilt from the file system structures further in, which is routine work.
Long sessions, high temperatures and a case full of components all pushing air at each other put a mechanical disk under conditions it was not sold for. The failure usually shows first as corrupted game files and stuttering, then as a volume that will not mount. The pattern is recognisable and the disk is generally readable if it is stopped early.
A reset that removes everything writes a fresh installation and, on some machines, runs a cleaning pass over the free space as well. Where it only reinstalled, a great deal comes back. Where it wiped, considerably less does. The difference is decided in the first minutes of the diagnostic and you are told which one you are looking at.
Sage, QuickBooks and the rest keep everything in one large file, and a machine that loses power mid-write leaves that file truncated or inconsistent. The rest of the disk may be perfectly healthy. Repairing a database file from an image is different work from recovering a volume, and it is worth saying on the booking form which of the two you need.
A damaged journal, a lost superblock or an interrupted resize. Backup superblocks and the journal generally allow the volume to be rebuilt from an image with the directory structure intact. LVM and RAID layers underneath are unwound in the same way, and none of it is treated as an exception here.
An encrypted volume is recoverable with the key or the credentials and not without them, and that is the whole point of the encryption. The key usually sits in a Microsoft account, in Active Directory, or on a printed sheet somebody filed away. Encrypted drives are the 400 pounds + VAT band. Find the key before the drive is posted.
Plenty of towers arrive here in working order simply because they are too slow to keep and hold twenty years of photographs, letters and scanned paperwork. The disk comes out, the contents come off, and they go back on fresh media. It is the cheapest job on this page and one that nobody should be talked out of.
A tower is the easiest machine in this trade to open, which matters because nothing else inside it is wanted here. Take the side panel off, undo four screws, unplug two cables, and it is in your hand. Posting the whole computer means paying a courier to protect a steel box wrapped around something the size of a paperback. All-in-ones and the small boxes that clip behind a monitor follow the same principle with more fiddling, and both are worth a telephone call before force meets a plastic clip. Where a machine held two disks, or a small SSD alongside a larger drive, send the pair and label them: on a number of models the two were tied together as a cache, and neither half means anything by itself. Put a sheet of paper in the box as well, with a name, a mobile number and a sentence on how the trouble started. That sheet saves more time at this end than anything else in the parcel and it costs nothing to write.
The first is a surge. It comes in on the mains and leaves through whatever is plugged in, and on a hard drive it usually stops at a small protection component on the underside board that exists to short itself out and take the hit. A drive killed by a storm is therefore often a short job at a rework bench with the mechanism behind it entirely untouched. The second is a repair shop reinstalling Windows to cure a machine that had got slow, which puts a new file system over the beginning of the disk and leaves the earlier one unaddressable while the old files sit intact further in; how much comes back depends on how heavily the new installation has been used since, and there is a page here given over to that exact situation. The third is a mirrored pair in a workstation where one member dropped out months ago, nobody noticed, and now the survivor has gone as well. Both disks get imaged and the newest good copy of each file is assembled from the pair, which usually ends better than the owner expects.
Windows makes three suggestions at the worst possible moment. It offers to format a disk full of finished work, which happens because a header will not parse rather than because the disk is empty, and agreeing lays an empty structure over the top of the old one. It offers a check-disk pass at the next restart, which is a tool designed for healthy hardware with untidy filing and does real harm on a disk that is failing physically; there is a page here about undoing what one of those passes removed, and the short version is to decline. And Disk Management offers to initialise a disk with years of work on it, which writes a new signature and an empty partition scheme. Refuse all three, shut the machine down, take the disk out. Everything those prompts would have destroyed is still there while it is left alone, and finding out costs nothing: one drive or SSD is £300 + VAT and the examination in front of it is free.
A tower gives up a bare disk and nothing else, and from that point it is treated exactly like any other piece of storage in the building. These are the benches a desktop drive tends to pass through on its way to an image.
Head stack replacement, platter work and spindle repairs are done in filtered air with the drive levelled and clamped. Desktop mechanisms are larger and heavier than notebook ones, which makes alignment more forgiving and the consequences of getting it wrong rather less so.
PC-3000 access into the service area for translator rebuilds, defect list repairs and modules read straight off the platters. A drive that reaches speed and then stops itself, or reports a capacity nobody has ever sold, is usually a firmware problem with a perfectly sound mechanism behind it.
Atola and DeepSpar hardware reads head by head, with timeouts enforced by the imager rather than requested politely of the drive, and current draw watched throughout. Cooperative regions are banked first and the difficult ground is left until last, which is the opposite of what a copy-and-paste attempt does.
Heads and boards have to match the family and often the exact production run, not simply the model number on the label. The shelf here is organised that way, and where the right donor is not on it, one is sourced. Donor parts are bought for your specific drive, which is why physical work takes half the figure up front.
Shorted protection diodes, burnt regulators and cracked tracks repaired at magnification, with the drive's own adaptive data carried across whenever a board is replaced. Fitting a board from another drive without moving that data across produces a disk that spins happily and reads nothing.
A surprising number of desktops turn out to hold two disks tied together by the motherboard's own controller, and the owner had no idea. Members are imaged separately, the layout is worked out from the images, and the volume is assembled from copies. Nothing is ever rebuilt onto the disks that arrived.
Nothing in this trade gives up a disk more easily than a tower, and that counts for something because the disk is all that has to go anywhere. One side off, four screws, two plugs, and the computer stays under the desk. A single disk or SSD is 300 pounds + VAT. The examination costs nothing, closes two working days from the drive being logged in, and the recovery that follows normally takes two to four working days. A mirrored pair with only one failure is still priced as single-drive work; a striped pair is an array and begins at 500 pounds + VAT, climbing with the member count. Logical faults are covered by no fix, no fee, while mechanical and electronic repairs take half up front and appear in the published exclusions. PATA drives are still read here, which comes up more than you might expect, because a fair number of the machines pulled out of Warwickshire garages were built before SATA existed.
The business half of this bench is mostly Dell OptiPlex and Lenovo ThinkCentre, for a plain reason: they were bought a dozen at a time and they reach the end of their working lives a dozen at a time as well. Inside them is whatever the supplier could get by the pallet that quarter, which in practice means a WD Blue, a Seagate BarraCuda or a Toshiba P300. Self-assembled machines are a different matter and often an easier conversation, since whoever built one can usually name every component in it from memory. Alongside those, Fujitsu Esprimos, Compaqs, Packard Bells and Gateways keep coming out of spare rooms and back offices around the county, and their disks read no differently from anything made this year. What is printed on the front changes nothing about the price.
Some desktops have no side to take off. A machine built behind its screen wants prising tools, patience and somebody's video of the same model first, and it punishes force with a broken panel. The very small boxes that bolt to the back of a monitor are simpler than either, because there is usually one M.2 module inside and it lifts out under a single screw. Where a case feels as though it needs a shove, put it down and telephone instead.
Part the storage from the machine before the box is made up, because a bare drive is all that should reach Oxford. Beaten by the screws? Ring 0800 689 0668 and somebody will walk you through it, or let a repair shop on almost any high street free it for a few pounds.
Only the disk needs to make the journey. Towers are not stripped at this end, nothing but a bare drive gets logged in, and posting a steel case means paying a lot to protect an object weighing a few hundred grams. On almost any desktop it is a thumbscrew or two, the side panel off, and a power and a data lead unplugged. If that goes badly, 0800 689 0668 is answered on weekdays between 9:00am and 5:30pm, and any repair shop in the city will do it over the counter in a few minutes. Address the parcel, tracked and insured, to Oxford Data Recovery, John Eccles House, Oxford Science Park, Robert Robinson Avenue, Littlemore, Oxford OX4 4GP. Driving it down is about fifty-five miles of M40 and roughly an hour, with parking at the door and a reception taking hand deliveries on weekdays. There is no van, no collection anywhere in this network and no counter in Coventry, so however it gets there, it gets there through you. Write your name and a mobile number on a sheet in the box, and where two disks are going, mark which port each came off.
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.
Where an old profile ends up, and the installs that keep none of it
An emptied bin, a tidied share and a lost folder are three problems
Which tower disks can be held stable for a single copying run
Windows offers to mend the volume; accepting costs you clusters
Two disks striped under one desk, and the day the pair gives up
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.