Hard drives deteriorate audibly. SSDs give nothing — SMART reports flash wear well and controller health barely at all, and the controller is what usually fails. A drive showing 96% life remaining on Monday can be undetectable on Tuesday, which is why the usual backup advice does not apply here.
$ mdr diagnose /dev/nvme0 → Device: Samsung 990 PRO · 2 TB NVMe → Status: NOT DETECTED — controller in safe mode → Client: confidential · Manchester, Greater Manchester $ mdr engineer-working → Controller: reached · garbage collection halted → Translation layer: rebuilt from NAND → Imaging: 1.9 TB / 2 TB · 9sectors mapped $ mdr verify → ✓ documents — 19,400 files → ✓ photos — 22,800 files → ✓ drive recovered — data back
A firmware update rewrites the service area a recovery needs to read, and on an unstable drive an update that stalls can leave it permanently unreadable. Secure erase is irreversible by design. If the drive has started dropping out, stop using it.
The reason SSD failures feel so abrupt, and why the backup advice differs from spinning disks.
Hard drives deteriorate audibly and measurably — reallocated sectors climb, transfers slow, something clicks. SSDs do none of that. SMART on a solid-state drive reports flash wear reasonably well: percentage used, total bytes written, spare blocks remaining. What it barely reports is the health of the controller, and the controller is what actually fails in most catastrophic cases.
So a drive can show 4% wear and 96% life remaining on Monday and be undetectable on Tuesday, with nothing in between. The practical consequence is that “I’ll back it up when it starts playing up” does not work on an SSD, because there is no playing up. That is worth knowing before it matters rather than after.
They present almost identically — the drive vanishes — and the diagnostic separates them.
Either can close a job before it starts, so both are established at the diagnostic rather than discovered later.
Many SSDs encrypt by default — Class 0, TCG Opal, Samsung’s portable T-series — often with BitLocker layered over the top. Without the password or key, healthy NAND yields nothing usable and no laboratory changes that.
If they were deleted from a working drive, TRIM has almost certainly erased them within minutes and there is nothing to read. If the drive failed while holding them, that is entirely different.
Counterintuitively so: a drive that stopped responding cannot run TRIM, so failure preserves data that continued health would have destroyed.
Controller revived in vendor mode where possible — the good path, covering most failures. Direct NAND reads where it cannot, which means de-interleaving, ECC correction and unscrambling to that controller’s arrangement.
The number is simple. What matters more is the conversation before it.
SSD and NVMe recovery is from £300 +VAT whatever the format — 2.5-inch SATA, M.2, NVMe or a portable in a case. Controller-level physical work carries a 50% deposit, with the balance only if it succeeds; everything else runs no fix, no fee.
The part worth saying plainly is that SSDs produce more “no” answers than any other device we handle, and you should hear that free rather than paid for. Encrypted with no key, files TRIMmed from a working drive, storage soldered to a dead board — each of those is a closed job, and each is established during the free diagnostic. We would rather tell you at that point than take a deposit for something with no route through it.
Get the drive to us with a short note on the fault. The diagnostic costs nothing, and the figure that follows is fixed in writing before any work begins.
Nothing starts until the drive to us is on the bench. Pack it properly and put your contact details in with it. The diagnostic that follows costs nothing, and one figure goes in writing before any work begins.
Posting it? Use something tracked and insured — whatever is on the drive is worth considerably more than the postage. Bringing it in? Weekdays, 9am to 5:30pm, and it still wants packing as above for the journey.
Not sure it is worth sending? Tell us about the drive and fault, plus anything already attempted — an engineer reads every one of these and comes back with an honest view of the odds and a price band, before you post anything.
We’ll be in touch shortly. If it’s urgent, call 0161 871 0788.
The technical questions we get most about failed SSDs.
Because SMART reports what is easy to measure — flash wear, bytes written, spare blocks — and barely reports the health of the controller, which is what usually fails catastrophically. There is no gradual decline to spot.
Not necessarily. A drive gone from the BIOS, showing a blank model or an absurd capacity, is the classic controller signature. The flash behind it normally still holds everything and can be reached in vendor mode.
In one important respect: portables run hot in sealed cases with no thermal path, and many encrypt by default. The recovery route is the same, but heat is a far commoner cause and the encryption question matters more.
Probably not, and that is worth knowing before you pay anyone. TRIM erases the blocks behind deleted files within minutes on a working drive. Shutting it down immediately preserves whatever the drive had not yet reached, but the window is very short.
Only while the mainboard functions. Once it has failed — and particularly where encryption is tied to it — there is often no route at all. Send us the model and we will tell you before you post anything.
From £300 plus VAT after a free 48-hour diagnostic. Controller-level work takes a 50% deposit with the balance on success. Where the diagnostic shows a closed job, you pay nothing.
What decides an SSD job is the controller and whether encryption sits behind it — not the badge. Most drives use one of a handful of controller families, and the same recovery route applies across the brands built on each.
Names with an arrow have a guide of their own — brand-specific faults, and what they mean for a recovery.
£300 +VAT flat, whatever the controller turns out to have done. Vendor-mode access first, chip-level reading only where the controller cannot be revived — and a straight answer at the diagnostic if TRIM has already taken what you are looking for.