Data Recovery Case File · Cameras, Drones & Cards · When Testing Is Genuinely Finished
Dead Card Tested on Everything
His enquiry contained the most thorough elimination in this volume, done before he contacted anyone. A 32GB SanDisk Extreme SD card that "appears to be completely dead. It is not registered in any way in any computer — Mac, Windows or Linux — and it also does not appear to be registering in any camera I have. The card did contain photos." Three separate operating systems and multiple cameras. That is worth writing down because most enquiries arrive after one machine and a spare cable, and because it demonstrates something useful: there is a point at which further testing stops adding information, and he has passed it comfortably. Everything external is eliminated, the fault is located, and the only thing left to gain from another attempt is wear.
| Media | 32GB SanDisk Extreme SD card — not enumerated by macOS, Windows or Linux, nor recognised by multiple cameras; photographic content held |
| Reported situation | Complete non-recognition across three operating systems · multiple cameras also failing to detect the card · no physical damage reported · photographs required |
| Fault class | Controller failure with total non-enumeration — memory typically intact; chip-level access through internal pads required |
| Equipment used | No further host testing · monolith preparation and internal technological pads exposed · PC-3000 Flash Spider Board probing · Rusolut VNR pinout confirmation, XOR descrambling, ECC correction and translation rebuild · photographs validated by rendering |
The decode: what his testing proved, and what it cannot
What three operating systems establish: more than three attempts on one machine ever could. macOS, Windows and Linux are independent implementations with different drivers, different enumeration behaviour and different tolerances for devices that misbehave. All three failing to register the card eliminates drivers, operating-system quirks and configuration entirely. Adding multiple cameras — devices with their own simple, purpose-built readers — removes any question about a particular reader or a particular port.
What that leaves: a single conclusion. When a card is inserted, the host applies power and waits for the card to identify itself, and only after that handshake does anything else happen. A card that no device on earth registers is a card that is not completing that handshake — which places the fault at the controller, the chip conducting the introduction and managing the memory, rather than at the filing structure or the photographs.
Why that is the hopeful answer: the controller is the doorway rather than the storage. The memory holding his photographs sits behind it, and a controller failing does not damage what is written in the NAND. Manufacturers also leave other ways in — internal technological pads beneath the card's surface, behind the failed chip, which speak the memory's raw protocol directly. The data is read through those and reconstructed in software: descrambled for the controller family, error-corrected, and reassembled through its translation layer until a filesystem stands up.
Why stopping now is the right call: the point of this page. Each further insertion applies power to a failed controller and returns the same answer, and on a card that is nearly the only remaining risk he controls. There is a version of this where an owner keeps trying — another reader, another laptop, a friend's camera — over weeks, and the only outcome is a card that has been powered a hundred more times for no diagnostic gain. His testing has already answered the question. The correct next action is to put the card somewhere safe and stop.
The one thing that would make it worse: a device that eventually half-recognises the card and offers to format it. Cards in this state occasionally respond briefly, and accepting a format on a partially-responding card writes over the structures that make a chip-level reconstruction straightforward. Nothing should be formatted, repaired or initialised.
On the bench
No further host testing was attempted, since three operating systems and several cameras had exhausted what that can establish. The card was prepared as a monolith and its internal technological pads exposed and probed on the PC-3000 Flash Spider Board, with the pinout confirmed against the Rusolut VNR database — bypassing the failed controller rather than attempting to revive it. The raw memory was read out and reconstruction ran the full pipeline: XOR descrambling for the controller family, ECC correction, and translation-layer rebuild until the card's filesystem stood up. Every photograph was validated by rendering at full resolution before delivery on fresh media.
The outcome
The contents read at chip level past a failed controller, validated and delivered. Free assessment, one fixed written figure including VAT, 50% of parts and labour upfront with the balance only on successful recovery. The decode, for anyone whose card is dead everywhere: testing on three operating systems and several cameras is a complete elimination — drivers, readers, ports and configuration are all ruled out, and there is genuinely nothing further to learn from another attempt; total non-recognition means the card is not completing the handshake that precedes everything else, which places the fault at the controller rather than at your photographs; the memory behind it is normally entirely intact and reachable through pads the controller never guarded; and the remaining risk you control is the number of times you power it, so stop.
Card that no device will recognise
Once two or three genuinely different devices have failed to see it, stop — you've finished testing and everything after that is just wear. Different operating systems and a camera or two eliminate drivers, readers, ports and configuration completely, and a card that nothing at all registers is failing the electrical handshake that comes before any question of files. That points at the controller chip rather than your photographs, and the memory behind a failed controller is normally untouched and readable directly. So put the card somewhere safe and flat, don't keep trying new readers, and don't leave it in a drawer with loose change or keys. If any device ever does half-see it and offers to format or repair it, refuse — a format accepted on a partially-responding card writes over exactly what makes the recovery straightforward.
Your testing is finished — call Glasgow Data Recovery on 0141 404 0294; read at chip level through the card's internal pads, reconstructed in software, every photograph checked by eye.
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.