Almost every difference between a lab and a piece of software comes down to one decision: work on a copy, not the original. Everything else follows from that.
A failing drive may only give you one complete read. Spending it on a clone means every later attempt is free, repeatable and risk-free.
Before anything is connected properly, the fault has to be classified: mechanical, electronic, firmware or logical. That determines everything downstream and it is why a diagnostic is not a formality. A drive that clicks needs clean-air work; one that spins silently and is undetected usually needs board or firmware attention; one that mounts but has lost a partition needs neither.
This stage is free here and takes up to 48 hours. It produces a written quote, and it is also where we tell people the answer is no — a crypto-erased phone, a TRIMmed SSD, or an encrypted volume with no key.
The objective is never a working drive. It is a drive that will hold together long enough to be copied. Failed heads are replaced with a matched donor set under filtered air; a failed board is repaired or swapped with the original ROM adaptives carried across; corrupt firmware modules in the service area are repaired until the drive enumerates at its true capacity.
Nothing here is a repair in the ordinary sense. The drive is not expected to work afterwards and usually will not.
The drive is cloned sector by sector on a hardware imager built for failing media — adjustable timeouts, per-head maps, the ability to retry weak areas in short passes and skip genuinely dead sectors rather than stalling. It is strictly read-only, so the original is never altered.
Where sectors cannot be read at all, they are logged rather than guessed, so the map of what is missing is known precisely instead of surfacing later as a file that opens to nothing.
From the image the file system is rebuilt — NTFS master file table, APFS containers and checkpoints, ext inodes, or a RAID assembled virtually from several images at once. Where the structures are beyond repair, files are carved out by signature instead, which returns contents without names or folders.
Then verification: files are opened rather than counted, folder structure is checked against what the customer expects, and anything that fell in an unreadable region is reported by name. Recovered data comes back on fresh media. Prices start at £250 +VAT for cards and sticks and £300 +VAT for a single drive, quoted in writing before any of this begins.
The process of retrieving data from storage that has failed or been damaged, by making the device readable just long enough to clone it and then reconstructing the files from that clone. The defining principle is that the original is read once and never written to.
Five stages: diagnose what failed, repair the drive only as far as imaging requires, clone it read-only on hardware built for failing media, rebuild the file system from that image, and verify the results by opening files rather than counting them.
Because a failing drive may only have one complete read left in it. Spending that read on a clone means every subsequent attempt happens on a copy — repeatable, risk-free and impossible to make worse. Software works on the original, which is fine until it isn’t.
No, and that isn’t the aim. Repairs go only as far as needed to copy the data — a drive fitted with donor heads is not expected to work afterwards and generally won’t. What you get back is your data on fresh media.
From £250 plus VAT for memory cards and USB sticks, £300 plus VAT for a single drive, and from £500 plus VAT for RAID, NAS and servers — fixed in a written quote after a free 48-hour diagnostic, with most jobs no fix, no fee.