Case record · Logical & corruption · SDR-2025-0311
A Cyclic Redundancy Check, and My Life's Work.
The Toshiba desktop drive gave out mid-use. It vanished from File Explorer
, and Disk Management met every attempt with the same cyclic redundancy check
message. He had run Windows error-checking twice before ringing. His summary wanted no translation — everything I have ever drawn is on there
— and he had no idea whether the fault lay in the disk or the software.
Seeing the same thing yourself?
0800 6890668
The translation.
A CRC error means the data got through and then failed the checksum riding with it — the drive produced something it could not stand behind. Two faults do that. One is a broken filesystem returning rubbish from sectors that are perfectly sound; the other is a surface that has started to fail. Telling them apart matters, because the second worsens with every retry — and Windows, left alone, retries until something gives.
Kit used on this job.
What happens in a case →| Platform | Its role in this case | Why we use it |
|---|---|---|
| DeepSpar Disk Imager 4 | Read the disk with the retry count held down, logging weak sectors instead of battling them | Head-map-driven imaging, with per-head timeouts, resets and power-cycle control |
| R-Studio Technician | Rebuilt the NTFS structures once the image was complete | Wide filesystem support, with RAID rebuilds you can trust |
| ACE Lab PC-3000 Express | Checked SMART and the defect list at firmware level before a single imaging pass | Talks to drives using maker-specific commands beyond the ordinary ATA instruction set |
In the lab.
Work out which of the two faults this is
The disk was questioned at firmware level before anything was read off it — SMART counters, the grown defect list, the reallocation history. The answer left no room: reallocations rising, pending sectors waiting. Hardware wearing out rather than a filesystem in a muddle, and the correct response to each is the opposite of the other.
One gentle pass, with the retries held down
The drive went on the imager with short timeouts and a firm ceiling on retries. Regions that failed were noted and stepped over rather than pounded, then approached again at the end from a different angle. The reasoning is plain: on a surface that is breaking down, the fiftieth go at a bad sector kills the drive far more often than it wins the sector.
Mend the filesystem where mistakes cost nothing
With the image safe, every repair was carried out on it and never on the drive. The master file table had been damaged across the failing area, and the two rounds of Windows error-checking had already thrown parts of it away. Records that survived were rebuilt and the truly unreadable areas mapped, so he could be told exactly what was missing.
The result.
Nearly all of it lifted off the image — the drawing archive whole, the surveys whole, and a written list of the few files that did not survive. Most had been cut short by the error-checking, not the disk. All of it checked against his priority folders before any money changed hands.
Similar cases on the index.
Other cases in Logical & corruption.
Does this sound like your own drive?
The rule holds as in every case above: switch it off, and let a free diagnosis come before any decision.