Data Recovery Case File · Solid State & Flash · It Was Getting Worse
Intermittent Is Not Mostly Working, It Is Failing With Gaps
His enquiry describes a trajectory reported as a symptom. A card used in a phone for years holding photographs and video, where "recently the phone randomly stopped and started detecting the card. Now" it does not at all. The intermittency was the warning and the present state is where it was heading — which means the useful moment was during the weeks when it still worked sometimes.
| Media | microSD card in continuous phone service over several years — enumeration becoming intermittent and subsequently ceasing; photographic and video content held |
| Reported situation | Card in phone service over several years · storing photographic and video content · host detection becoming intermittent over a recent period · card no longer detected at all · contents required |
| Fault class | Progressive controller failure following spare block exhaustion — intermittent enumeration as an intermediate stage; memory contents unaffected by controller state |
| Equipment used | Intermittency treated as a progression rather than a fluctuating state · no formatting or write attempts permitted · device addressed through hardware with imposed timeouts · chip-level read past the controller with translation layer reconstructed in software |
The decode: why intermittent devices are on their way somewhere
What intermittency actually is: not a stable condition. A device that sometimes works is a device whose controller sometimes completes its initialisation — and the proportion of successes falls over time rather than fluctuating around a level.
Why that matters for how people respond: intermittent reads as "mostly fine, occasionally annoying." It is closer to "failing, with gaps that are closing" — and every week of continued use is spent while the gaps are still there rather than after they have gone.
What produces it in a card of this age: flash memory wears with writing, and the controller replaces worn blocks from a reserve held for that purpose. When the reserve is exhausted, the controller can no longer maintain its own working structures reliably — including the ones it needs to start up. So initialisation succeeds sometimes and fails increasingly.
Why a card in a phone reaches that point faster than most: phones write constantly. Applications, caches, thumbnails, message attachments and logs all write in small amounts continuously, which is a far heavier workload than a camera card sees and consumes the reserve steadily over months.
Why the contents are nevertheless likely intact: the failure is in the controller's ability to present the storage, not in the storage. The memory holds charge exactly as it did, and the photographs written to it years ago are unaffected by whether the controller can currently introduce itself.
What the route is: the memory is read directly at chip level, past the controller entirely, and the pattern by which the controller distributed data across it is reconstructed in software from the contents themselves. The failed component is bypassed rather than repaired.
What must not happen: no formatting, and no tools claiming to repair the card. Anything that writes to exhausted memory makes the position worse, and a card that will not enumerate cannot be repaired by software in any case.
The habit worth taking from it: when a storage device starts behaving intermittently, that is the moment to copy everything off — not when it stops. The window is open and closing.
On the bench
Intermittency was treated as a progression rather than a fluctuating state — a controller sometimes completing initialisation and doing so less often over time, which follows spare block exhaustion once the reserve used to replace worn blocks is consumed. No formatting or write attempts were permitted. The device was addressed through hardware with imposed timeouts, with chip-level reading past the controller and the translation layer reconstructed in software from the contents.
The outcome
Intermittency read as a progression, no writes permitted, and the memory read past the controller. 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: intermittent is a direction rather than a state. A device that works sometimes is one whose controller sometimes completes its start-up, and that proportion falls — so the weeks when it still worked occasionally were the window.
Storage that has started working only sometimes
Copy everything off now, while it still works sometimes — that's the whole lesson, and it applies to any device. Intermittent reads as mostly fine and occasionally annoying; it's closer to failing with gaps that are closing. A card or stick that sometimes isn't detected is one whose controller sometimes completes its own initialisation, and that proportion falls rather than fluctuating. On a card that has lived in a phone for years, the cause is usually exhausted spare blocks, since phones write constantly through apps, caches and thumbnails. The contents are almost certainly fine — the failure is in presenting the storage, not in the storage.
Don't format it — call Oxford Data Recovery on 01865 593000; intermittency read as a progression, addressed through hardware with imposed timeouts, memory read past the controller with the translation layer rebuilt.
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.