Flash & SSD · case notes · PLY-2025-0939
Five Machines Tried It. Not One Answered.
It worked at one pond and did nothing at the next. An ecology consultancy in Tavistock carried it round the office — two desktops, three laptops, and finally a powered hub of its own — and got not a flicker out of any of them
. Nothing chimed, nothing appeared under Disk Management, and, as the surveyor put it, you might as well be pushing in a pebble
. A full season of newt survey photographs sat on that stick.
Sounds like yours? Talk it through.
0800 6890668
Why it happens.
When six hosts in a row say nothing, the computers stop being suspects and the fault moves inside the plastic. There are only two candidates in there. The metal connector is soldered to the board on pads that flex a little each time the stick is pushed into a port; alongside them sits the controller, which does the talking to a host on the memory's behalf. Not one photograph lives in either part. Both merely stand between you and the pictures, which are held in the memory itself — and memory keeps what was written to it long after anything is left that can ask. That is where a lab picks the job up.
The kit this job called for.
How the job runs →| The kit | What it did here | Why we keep it |
|---|---|---|
| PC-3000 Flash | Read the memory package on its own, with the controller taken out of the path | Bypasses the controller and reads the raw NAND, against a maker-ID library we keep up to date |
| Rusolut Visual NAND Reconstructor | Reassembled the dump as a logical image — ECC first, then XOR, then page and block order | Turns a raw NAND dump back into data: ECC, XOR, page order, then reassembly |
| IR preheater & BGA rework station | Took the BGA off undamaged and dressed its pads to sit in the socket adapter | Hot enough to lift the NAND and reball it, and no hotter than the data will stand |
How it ran.
Start with the repair that costs an hour
Under the stereo microscope first. A hairline crack in the solder under the connector is bench work — an iron, an hour, and the stick reads again. There was none here. Every joint was whole, every track continuous. On the bench supply the controller drew current as it should while its data lines said nothing, which is what the office had been seeing.
Straight to the memory; the controller stays out
Off came the package, its pads were cleaned, and it was read raw through a socket adapter. The rest of the job stands on a single point: a failed controller takes nothing away with it. What was written stays in the silicon, in whatever order that particular chip happened to use at the time.
Recreate the arrangement and the folders return
Rusolut was set on the dump and worked back through it: correction codes run, the XOR worked out, and pages and blocks returned to the places they once held, until what the failed controller used to present existed a second time. A filesystem mounted, with the survey folders inside it.
How it ended up.
The photographs went back on new media, three frames short, with the licence deadline still ahead. The three that were missing all came off the last visit to the ponds, all three returned damaged, and the report says so.
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.