Data Recovery Case File · Solid State & Flash · A Half-True Answer
Why Removing the Memory Works on a Card and Not on This
His enquiry repeats something he has been told that is true of one thing and not another. A 1TB solid-state module: "it stopped working. I've tried everything — reseating it, swapping slots — I have no luck whatsoever in getting it to connect and am pretty sure it's a total bust. But I was told people can take the flash memory off sometimes and still get it to work." That advice is accurate for memory cards and simple sticks, and it is where this archive's chip-level work happens. On a drive like his it usually does not apply, and the reason is worth understanding before anyone pays for it.
| Media | 1TB solid-state drive module — not enumerating; reseating and slot changes attempted without effect |
| Reported situation | Drive ceasing to function · reseating attempted · alternative slots attempted · no connection achieved on any attempt · owner enquiring about direct memory removal |
| Fault class | Controller failing to complete initialisation — memory not independently reassemblable where mapping and encryption reside in the controller |
| Equipment used | Assessed on a direct adapter with initialisation state read directly · vendor technological and safe modes attempted (PC-3000 portfolio, SSD support) · firmware area and translation layer assessed · position stated before any charge |
The decode: why it works on a card, and what an SSD adds
Why it genuinely works on cards and sticks: those devices hold a single memory package with a relatively simple controller. The data is written in a straightforward pattern, and the translation between logical addresses and physical locations follows a scheme that can be worked out from the memory contents themselves. Read the raw memory, descramble it, correct the errors, and reconstruct the mapping in software — and the filesystem stands up. That is exactly what this archive does with failed cards, and it is why a controller failure on a card is usually recoverable.
What a solid-state drive adds, and this is the difference. Three things, and each is enough on its own. The drive spreads data across many memory packages simultaneously and interleaves it between them for speed, so no single package holds anything coherent. The mapping between logical addresses and physical locations is far more complex, is maintained by the controller, and is not derivable from the memory alone. And most modern drives encrypt everything as it is written, with the key held in the controller — so raw memory read separately is ciphertext even if the first two problems were solved.
What that means in practice: removing the memory from a drive of this kind returns data that cannot be reassembled without exactly the information that died with the controller. It is not impossible in every case — some designs and some faults permit it, and specialist work exists — but it is not the routine fallback his advice implies, and anybody promising it as a general answer is overstating what the technique does.
What actually is the route, and it is a good one: manufacturers build technological and safe modes into their controllers precisely for controllers that cannot complete a normal start-up. Those allow the firmware area to be assessed, damaged modules repaired and the translation layer rebuilt — after which the drive addresses its own media normally and images conventionally, with its own mapping and its own key intact. The aim is to revive the controller rather than to bypass it, which is the opposite of the card approach and the reason the two cases are answered differently.
Why his own testing was worth doing: reseating and trying different slots eliminates a poor connection and a faulty slot, both real causes on modules held by a single screw. Having done that, the eliminations are complete.
The honest position: where the controller cannot be addressed at all, this is one of the harder outcomes in the archive, and that is said before any charge rather than after.
On the bench
The drive was assessed on a direct adapter with its initialisation state read directly, establishing how far start-up progressed before failing. Vendor technological and safe modes were attempted through the PC-3000 portfolio's SSD support — the aim being to revive the controller rather than bypass it, since a drive of this class interleaves data across many packages, maintains its mapping in the controller and commonly encrypts to it. The firmware area and translation layer were assessed and damaged modules repaired, and the position was stated before any charge where the controller could not be reached.
The outcome
The drive assessed on a direct adapter, vendor modes attempted and the position stated plainly. Free assessment, one fixed written figure including VAT; where a chip has to be removed, 50% of parts and labour is payable upfront with the balance only on success — otherwise no recovery, no fee. The decode, for anyone told the memory can simply be taken off: that is true of cards and simple sticks, where one package holds the data in a derivable pattern — which is why card recoveries usually succeed past a dead controller. A solid-state drive interleaves data across many packages, keeps its address mapping in the controller, and generally encrypts to it. So the route is reviving the controller, not bypassing it.
Told the memory chips can just be removed
That's true of memory cards and simple USB sticks, and not usually of a solid-state drive. On a card there's a single memory package and a relatively simple controller, and the pattern linking addresses to physical locations can be worked out from the memory itself — so reading the raw chip and rebuilding in software genuinely works, and it's why card recoveries often succeed past a dead controller. An SSD is different in three ways, any one of which is enough: it spreads data across many packages and interleaves between them, so no single chip holds anything coherent; its address mapping is maintained by the controller and isn't derivable from the memory; and most encrypt everything as written, with the key in the controller. The route is reviving the controller instead.
Ask about the controller, not the chips — call Newcastle Data Recovery on 0191 406 1051; assessed on a direct adapter, vendor technological modes attempted, firmware and translation layer assessed.
Request a quote online →
Our case files are drawn from genuine enquiries received by our laboratory over the past ten years, anonymised to protect client confidentiality. Each one describes the diagnostic and recovery procedure our engineers apply to that fault, using the equipment listed.