Data Recovery Case File · Solid State & Flash · It Read Something
An Integrity Check Failure Means Data Arrived and Did Not Verify
His enquiry names the error precisely, which makes it useful. A flash drive that stopped working "because of a data error, cyclic redundancy check", and he has been told this means the device is failing. The advice is broadly right and the error is more informative than that: it reports that data was retrieved and did not match its own checksum, which is a different situation from a device returning nothing.
| Media | 32GB USB flash drive — returning an integrity check failure on read; device otherwise presenting |
| Reported situation | Flash drive in normal use · device reporting a data error described as a cyclic redundancy check failure · owner advised the device is failing · specialist assistance recommended · content required |
| Fault class | Retrieved data failing verification — memory cell degradation with error correction exhausted; device still enumerating and partially readable |
| Equipment used | Integrity failure interpreted as retrieved-but-unverified data rather than absent data · device addressed through hardware with imposed timeouts rather than the host stack · controller state assessed in vendor technological modes · memory read past the controller with error correction applied in software · recovered content validated by opening |
The decode: what the error reports, and why it is a middle position
What the check actually is: a calculation performed over a block of data when it is written, and stored alongside it. On reading, the calculation is repeated and compared — if the two differ, something changed between writing and reading.
Why that makes the error informative: to fail the comparison, data must have been returned. The device located the block, read it, and handed something back — which is more than a drive that cannot be reached at all manages.
What it means on flash specifically: memory cells hold charge, and charge leaks. A cell that has drifted far enough returns a different value from the one written, and enough drifted cells in one block fails the check.
Why the drive did not simply correct it: flash devices carry error correction as standard, and it silently repairs a small number of wrong bits in every read. A visible integrity failure means the number of errors exceeded what the correction could handle, which is why it reached him at all.
Why that supports the advice he was given: a device surfacing errors past its own correction is past the point where it copes. It is failing rather than merely inconvenient, and further use is likely to produce more of the same.
Why the position is nevertheless better than it sounds: the failure is per-block. Blocks whose cells have drifted less are read and verified normally, so a device reporting errors on some content frequently returns most of it.
Why an ordinary copy handles this badly: it stops at the first block that fails and reports the whole operation as failed. Everything beyond that point is never attempted, which is why a copy returns nothing while the device holds a great deal.
What is done instead: the memory is read past the controller and correction applied in software, where more computational effort can be spent per block than a device does in normal operation. Blocks the drive gave up on are frequently recoverable given enough attempts.
Why repeated reading can help rather than harm here: a marginal cell does not return the same wrong value every time. Multiple reads of the same block produce slightly different results, and comparing them recovers the intended value — the opposite of the situation on a mechanically failing drive.
What must not happen: nothing written to the device, and no repair or reformat prompts accepted. Writing to flash with degraded cells is what turns partial loss into total.
On the bench
The integrity failure was interpreted as retrieved-but-unverified data rather than absent data — a check calculated on writing and repeated on reading, so failing the comparison requires data to have been returned. On flash, cells hold charge which leaks, and error correction silently repairs small numbers of wrong bits, so a visible failure indicates errors exceeding correction capacity. Memory was read past the controller with error correction applied in software, marginal cells returning varying values across repeated reads.
The outcome
The integrity failure read as retrieved-but-unverified data, the device addressed through hardware, and correction applied in software. 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: that error means data came back and did not match its own checksum, which is better than nothing coming back. It also means the device's own correction has been exhausted, so the advice you were given is right.
An integrity check failure on a flash device
Stop writing to it and refuse any repair or reformat prompt — writing to flash with degraded cells turns partial loss into total. Read the error as a middle position rather than a verdict: the check is calculated when data is written and repeated when it's read, so failing the comparison means data was returned and didn't match. That's better than a device you can't reach. It does mean the drive's built-in error correction has been exhausted, so it is genuinely failing. But the failure is per block, and an ordinary copy stops at the first bad one while the device still holds most of it.
Stop writing to it — call Manchester Data Recovery on 0161 871 0788; integrity failure read as retrieved-but-unverified data, addressed through hardware, correction applied in software.
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.