Flash & SSD · job notes · PLY-2025-0639
The Flash Was Healthy. Nothing Could Find It.
For a month now the Dell Latitude had been seizing up. It goes solid for a minute or so, then picks up again as though nothing had happened
. She researches for a living in Truro, and four weeks is a long stretch to work around. Come Tuesday morning the only thing it produced was Boot Device Not Found
. The SSD went into her partner's machine next, and then into a USB caddy, and it turned up in neither of them
. IT at work offered a rebuild and that was all
, which does nothing whatever for an MA that is not finished and the interview recordings underneath it.
Sounds like yours? Talk it through.
0800 6890668
Why it happens.
Every SSD contains a small processor with two jobs on its hands. One is presenting the drive to whatever it is plugged into. The other is maintaining a table showing where, among the flash pages, any given logical block is being kept today. The controller is that processor, and as a rule it fails long before the memory in its charge is anywhere near used up. Those minute-long freezes were the table falling apart beneath her, though nobody could have known so at the time. She had narrowed things down herself before the parcel arrived: where nothing at all will list a disk, the trouble lies past the connector, and an SSD offers very few parts to blame. Working on flash with the controller dead resembles a disc job in no respect. Nothing turns. There is no mechanism. The SATA connector ceases to matter.
The kit this job called for.
How the job runs →| The kit | What it did here | Why we keep it |
|---|---|---|
| PC-3000 SSD | Held the chip in its factory service state throughout | Built for SSD controllers alone; there is nothing mechanical about it |
| PC-3000 Flash | Matched the V-NAND against the chip database and read out the raw image | Bypasses the controller and reads the raw NAND, against a maker-ID library we keep up to date |
| UFS Explorer Professional Recovery | Recovered folders, chapters and recordings from the reordered image | Mounts what other software will not: APFS, ReFS, XFS, ZFS and Btrfs volumes |
How it ran.
Worn-out memory, or a controller that has died
No host port went near it after that. Everything took place at controller level on the bench, and each attempt to start failed at the same stage: loading the translation tables, which a disk does before it will identify itself to anyone. Then the wear counters, and there lay the one encouraging finding of the day — plenty of life remained in the memory itself.
Reaching the factory's own service mode
What the controller needed was technological mode. That is a factory condition, one an ordinary computer would never think to ask for, and vendor-specific commands are what get you there. Put into that state the chip hands over its raw contents, and nothing short of it will do. Skip the step and the drive is scrap. Manage it and the drive goes home.
Put the pages in order before hunting for anything
A raw dump arrives scrambled, and that is by design. Wear levelling drops each successive piece of a file wherever a package happens to have room at the time, so the mapping wants rebuilding before a single document reads as one. Once it was worked out, NTFS mounted without complaint — folders in their proper places, chapters carrying the names she had given them.
How it ended up.
Complete, the thesis went back on fresh media, along with all the interview recordings and her whole reference library. Nothing had ever been wrong with the Latitude itself: a new SSD went in and it works perfectly well. From now on, she says, the finished work is going to sit in three separate places. We intend to check up on that.
Pages people open after this.
Other SSD & flash jobs.
Seeing the same thing?
Leave it switched off and post it in. Nothing happens until the diagnosis, which lists what can still be read and what cannot.