There is no noise to listen for. A solid state drive that has failed gives exactly the same account of itself as one that was never plugged in, which is why owners spend a fortnight on forums before ringing anybody. Flash work shares a building with disk work and almost nothing else: no platters, no head stacks, no filtered air. It runs here every day for households across the city and for the small firms and engineering shops around Loughborough and Hinckley that quietly moved everything they own onto storage with no moving parts in it.
Every SSD that arrives here is examined at no charge. The figure quoted afterwards is put in writing and settled before anybody reaches for a screwdriver: £300 + VAT on any one hard drive or SSD, an NVMe blade counted the same.
Logical recoveries carry no fix, no fee. The named exceptions to it are electronic and mechanical failures, chip-level work, DVR jobs and forensic jobs; physical work takes 50% up front. The five bands are set out in full on the data recovery cost page.
Tracing a symptom back to the failure underneath it is where the real work begins, and these thirty account for all but a handful of the boxes opened at the Cambridge bench. A fault missing from the list is not an unfamiliar one — describe it on the telephone and you will get a straight view of the odds before you spend anything on postage.
This one is worth hearing straight. On a working SSD with TRIM switched on, deleted files are usually gone for good. The operating system tells the drive which blocks it no longer needs, the controller wipes them in the background, and reads of those blocks come back as zeros afterwards. No laboratory anywhere restores an erased NAND page. If the deletion happened minutes ago, shutting the machine down at once is the only useful action. If it has been running since, expect an honest answer rather than a hopeful one.
The management chip goes silent and the drive falls off the bus while every file remains exactly where it was written. What has failed is the part that speaks on the flash's behalf. Supplying that role from outside, through service mode or by reading the packages directly, is the whole purpose of a firmware bench.
Nobody ever manufactured a product with that name. It is a Phison controller stuck in its own boot loader with no working firmware above it, and the label is simply wrong. The data behind the label has not been touched. Service-mode work walks the controller back out and the real identity returns with it.
Those are the default identities a controller falls back on when its translation table collapses. The drive is not shrinking, it has lost the map that tells it what it holds. Rebuilding that map in service mode brings back the real capacity, and the first thing that happens afterwards is a complete image.
Several controllers drop into a protective read-only state when the spare block pool runs out or the wear figures cross a threshold. That is deliberate behaviour and, for once, helpful behaviour: everything on the drive is still readable. Copy it off in one pass, and treat the drive as finished for service afterwards.
Losing power partway through a firmware flash leaves the controller with half of one version and half of another. The flash memory holding your files is unaffected. Getting the controller back to a coherent state is a bench job, and retrying the update on a drive in that condition tends to make it worse.
The board is not answering. That points at a power rail, a failed capacitor, or a controller that will not come out of reset, and none of those touches the NAND. Invisible is a good deal less serious on an SSD than owners assume, because there is no mechanism to have been damaged in the meantime.
Blocks are being read back with more errors than the correction code can fix, and the drive is handing over what it got rather than admitting the problem. That usually means retention loss on a drive that has been left unpowered for a long stretch, or a wear-levelling table drifting out of step. It is imaged first and interpreted afterwards, never the other way round.
SSDs are more vulnerable to sudden power loss than most people believe, because the mapping tables are held in volatile memory and written back in batches. Cut the supply during that write and the drive comes back not knowing where anything is. Enterprise units carry capacitors to survive it. Consumer units mostly do not.
The boot structures live in a handful of blocks and a mapping fault at that end takes the whole volume out of view. Nothing else on the drive needs to be involved. Reinstalling Windows to make the machine usable again writes over the very region the recovery would start from, so the drive comes out first and the reinstall happens after.
Read retries and internal error correction are being run on nearly every request. The drive is working hard to hand you correct data and just about managing it. That is the best possible moment to take a full image and the worst possible moment to keep using the machine for ordinary work.
Consumer drives in machines that never get switched off reach their endurance figure faster than anyone plans for. Once the wear indicator bottoms out the controller starts refusing writes, and shortly after that it starts refusing to come up at all. There is a window in between, and it is the last one available.
Counterfeit drives are sold with the firmware told to report a size the flash does not physically have. Everything appears to work until the write passes the real limit, at which point the earliest data gets overwritten silently. What comes back from one of these is whatever genuinely fits, and the rest was never stored in the first place.
Data on a modern SSD is striped across every package on the board, so losing one is not losing one eighth of the files, it is losing a slice of nearly all of them. Where the controller is still working, the internal error correction can sometimes cover the gap. Where it is not, the reconstruction has to reproduce the interleave exactly.
2.5-inch cases take knocks that the board inside them does not. A cracked substrate or a lifted track leaves the flash entirely intact and the drive completely unresponsive. Microscope rework rebuilds the connection, and where it cannot be rebuilt the packages come off and get read on their own.
Corrosion works slowly and it never stops on its own. A drive that has been wet needs cleaning and drying properly before any power reaches it, because the first thing a shorted rail does is take out components that were fine. Do not attempt to dry it and try it. Send it wrapped and say what happened.
Solder joints and a marginal power rail behave differently at different temperatures, which produces a drive that works for five minutes after a cold start and then vanishes. Repeated cycles wear down the pattern. Imaging in short passes, cooled, gets far more off than one long optimistic attempt.
The partition structures at the front are unreadable and the operating system is proposing to write fresh empty ones. Behind that prompt the pages holding your files have not moved. Accepting it adds a fresh problem, and on an SSD it also invites the controller to start trimming what it now believes is free space.
The partition is visible, the file system inside it is not recognised. On flash storage that generally traces back to the mapping layer handing out inconsistent blocks rather than to file system damage as such. The distinction matters, because repair tools aimed at the file system will make an inconsistent map worse.
Secure erase instructs the controller to change the encryption key it applies to everything on the drive. The instant that completes, every page on the device is meaningless, no matter what physically survives. If the command ran, the answer is no. If it was cancelled early, occasionally there is a partial answer, and it costs nothing to find out.
Many SSDs encrypt everything they store as a matter of course, with the key held in the controller. That is invisible while the drive works and decisive when the controller dies, because reading the flash packages then gives you correctly recovered ciphertext. Where the controller can be revived, the data comes back. Where it cannot, honestly, it does not.
Cloning software that copies partition by partition can silently write an empty structure when it cannot interpret the source. The original SSD is normally untouched underneath. Stop using both, send the one that held the data, and say clearly which of the two is which, because that saves a day at this end.
A quick format on an SSD is worse news than on a hard disk, because the operating system follows it up by trimming the whole volume. Within minutes the controller has begun genuinely erasing pages. If the machine has been switched off since, there is a real chance. If a fresh Windows installation has run since, considerably less.
SSDs in a striped or mirrored set fail the same way as any other member and the array behaves accordingly. What differs is that trimming can propagate across members, so a set left running after a failure loses ground faster than a spinning-disk set would. Power the whole thing down and send every drive, labelled.
Flash cells leak charge, and they leak faster once the drive has been written to heavily. Unpowered storage is not archival storage, whatever the packaging suggested. Some of these come back in full after error correction has been applied on the bench. Some come back partially. None of them improve by waiting longer.
Several drive families have shipped with a counter that stops the drive dead when it reaches a specific figure, and the manufacturer fix is an update the drive is no longer alive to accept. Recognising the model and the hours is most of the work. What follows is service-mode entry and a full image before anything else is contemplated.
Apple used proprietary blade connectors for years and the drives inside them are ordinary flash behind an unusual edge. Adapters exist for the 2010 to 2017 range and those drives are handled normally here. What cannot be handled is a machine where the storage is soldered to the logic board, which covers Apple Silicon and a good number of thin Windows laptops.
An encrypted SSD that is also misbehaving needs the imaging done first, with the volume still sealed, and the key applied to the copy afterwards. Repeatedly typing forty-eight digits into a drive that is struggling spends the reliable time it has left. Encrypted drive work is 400 pounds + VAT rather than the ordinary single-drive figure.
Rare, but it happens to M.2 sticks and to the thin 2.5-inch cases in laptop bags. Whether anything comes back depends entirely on whether the NAND packages themselves survived and whether the break ran through them. The assessment is free and it will give you a plain answer within the two working days rather than a maybe.
No noise, no errors, no warnings, just a folder that used to have twelve years of accounts in it. On flash, that is either a mapping fault the drive is hiding well or a deletion that TRIM has already acted on. The two look identical from the outside and are told apart on the bench, which is exactly what the free assessment is for.
Start with the limit rather than end with it. On a working SSD with TRIM enabled, deleted files are usually gone permanently. The operating system tells the drive which blocks it no longer needs, the controller erases them in the background, and reads of those blocks afterwards come back as zeros. There is no laboratory anywhere that restores an erased NAND page, and any firm suggesting otherwise is selling you something. What that means in practice is a matter of timing. If the deletion happened in the last few minutes, shutting the machine down at once is the only action that helps, and it helps a great deal. If the computer has been running since, expect an honest answer instead of a hopeful one, and expect to get it during the free assessment rather than after an invoice. None of this applies to a drive that has failed rather than deleted something. A controller that has stopped, a translation table that has collapsed, a board that has cracked: in all of those cases the flash is untouched and the odds are good.
An SSD is a controller, a firmware layer and a set of memory packages, and the overwhelming majority of failures are in the first two. The controller stops answering and the drive vanishes from the bus. The translation table, which records where each piece of your data physically lives among the packages, becomes inconsistent and the drive reports itself as 8MB, or 0MB, or as something called SATAFIRM S11, which no manufacturer ever sold. Behind every one of those the memory is intact. There are two routes in. The first is the diagnostic mode the manufacturer uses in production, which most controller families still answer, and it is tried first because it leaves the drive whole. Where the controller will not respond at all, the packages come off the board under controlled heat and are read individually, and the page order, interleave and error correction the original controller used have to be reproduced in software before anything reads as a file. That second route is chip-level work: it takes half the fee up front and it sits outside no fix, no fee, and you are told which of the two you are looking at before agreeing to anything.
Flash does wear out, though rather less dramatically than the early warnings suggested, and a consumer drive in a machine that never gets switched off will reach its endurance figure years before anyone plans for it. The pattern is recognisable: the drive slows down as internal error correction works harder, then it refuses writes and drops into read-only, and then one day it does not come up at all. The read-only stage is a gift rather than a disaster, because everything on the drive can still be copied off in a single pass. Retention is the other half of the story and it gets less attention than it deserves. Flash cells leak charge whether or not the drive is in use, and they leak faster once the drive has been written to heavily. An SSD that has sat unpowered in a drawer since 2019 is not an archive, and a proportion of these arrive with material that error correction on the bench can still recover and a proportion arrive with material it cannot. Nothing about that improves by waiting another year. The price does not change either way: one SSD is £300 + VAT, the same as a hard drive, and the assessment ahead of it is free.
Flash work and disk work happen in the same building and share almost none of the same equipment. There are no platters to open to the air here, so the clean bench sits idle and the rework station, the reader and the terminal do the lifting instead. Everything below is read-only towards your drive.
Controllers from Phison, Silicon Motion, Marvell, SandForce and Samsung all have a diagnostic path the manufacturer uses in production. Getting a stalled drive back onto the bus through that path is the fastest route in, and it is the one tried first because it leaves the drive intact.
Where the controller will not answer, the NAND packages come off the board under controlled heat and get read individually. Chip-level work sits outside no fix, no fee and takes 50% up front, and that is said plainly here rather than in a footnote.
Raw NAND is not files. It is pages, spare areas and a scattering of metadata, and turning it back into a file system means reproducing the exact interleave, page order, XOR scheme and error correction the controller used. That reconstruction is the real work on a chip-off job.
Lifted pads, cracked substrates and burnt rails repaired at magnification, so that a board which was electrically dead can be brought back up long enough to hand over an image. Electronic work is one of the published exclusions from no fix, no fee.
The same Atola and DeepSpar equipment used on mechanical disks earns its place here for a different reason. A flash drive on the edge of failing has a limited number of good reads left in it, and spending them in a controlled order rather than at random is what determines how much comes back.
SATA, USB and M.2 all go through hardware blocking. On flash this matters twice over, because a drive that gets mounted read-write by an operating system can begin trimming blocks the moment it is recognised, and no software setting is trusted to stop that.
Those are the drives that turn up, not a list of what is accepted. The Samsung EVO line and the Crucial MX500 dominate because they were the two obvious upgrades for a slow laptop, and half the machines in Leicester that got a new lease of life around 2018 got it from one of the two. Kingston A400s and the budget badges arrive in the same numbers they were sold in. Anything reporting itself as SATAFIRM S11 is a Phison controller stuck in boot loader mode and is routine work rather than a disaster. One SSD is 300 pounds + VAT, the same as a hard drive, with the assessment free and closing two working days after the drive is booked in. Chip-level work, where the packages have to come off the board, sits outside no fix, no fee and takes 50% up front. Where TRIM has already erased what you are looking for, you get told so plainly and you pay nothing for having asked.
The post office does most of the work of getting a job here. A drive that is already unwell travels better boxed and insured than rattling around a car for a day of errands, and something dropped into a Leicestershire postbox this afternoon is generally logged in at Cambridge tomorrow.
The general rule is the drive travels and the machine stays behind — out of the laptop, out of the tower, out of the iMac, out of the recorder under the counter. This bench does not dismantle equipment, and a repair shop will do it while you wait. Three things are the other way round, and getting them wrong costs you the recovery: an external drive stays sealed in its own case, a NAS comes as a complete unit, and a WD My Passport or My Book travels whole with its cable, because on those the encryption key is held on the bridge board rather than on the disk — separate the two and the data becomes unreadable even to us. A Fusion Mac needs both of its drives, each labelled. The one thing nobody can work round is flash soldered onto the mainboard, as on Apple Silicon machines: if it will not come off, there is nothing to post.
↓ Print the shipping & booking-in form (PDF)
Address it to Cambridge Data Recovery. It is about seventy miles from Leicester if you fancy driving it — M1 south to Junction 19, then the A14 east — and the lab is two minutes off Junction 32 with parking at the door. Posting costs you a stamp and a day instead. Whichever you choose, you hear from us the moment it is booked in, and the free diagnostic closes two working days after that.
Not certain what belongs in the box? Ring 0800 689 0668 before you tape it up, or let the free online diagnostic ask the questions for you.
The examination is free, one written figure follows it, and the band covering this page is £300 + VAT on any one hard drive or SSD, an NVMe blade counted the same.