Flash and SSD · chip-level work · ADR-2025-3468
Nothing Touched It, and It Stopped Answering Anyway.
This card had lived a peaceful existence, and it turned out that this peace was actually the key to resolving the case. She normally stored it within one device and did not usually remove it
; she could remember no physical damage to the card
; but then it just refused to be recognised by four successive devices. She had used four different devices (three more than the initial device) and none of these devices would recognise the card. In addition, there were no secondary copies of the photographs contained on the card.
Same story on yours? Tell us what it does — the look is free.
0203 0868633
What failed, and why.
Where a card fails indicates much regarding why the card failed. Since this card was stationary, located in one device and subject to neither pressure nor abrasion to its electrical contacts, two typical physical causes of card failures were eliminated prior to opening the unit. Thus, we are left with two possible internal causes for the failure. Either the controller has failed, or the memory (which holds the data) has reached the point where additional writing cannot occur. Reading does not consume flash life, while writing consumes flash life. This is why many cards will continue to function for a period after they fail to write data and before failing to return previously-written data. Therefore, if four successive devices are unable to recognise a card, it suggests a failure of the controller rather than a damaged file system. A failure of the controller will prevent any communication with a computer, whereas a file system failure will allow a computer to sense the presence of the card. A silent controller, however, says nothing at all about the state of the photographs that sit behind it.
What the bench did with it.
The order of the job →| Equipment used | Its work on this job | The reason for it |
|---|---|---|
| PC-3000 Flash | The component identified against the library was removed from the board and read independently as a die. | Directly reads the flash off the chips, going around the controller, with a maker-ID library kept updated here |
| Rusolut Visual NAND Reconstructor | Correcting for errors, removing the XOR mask, and placing pages in the correct order (to match the controller's ordering). | Rebuilt raw NAND dump: ECC, XOR mask, page order, files. |
| R-Studio Technician | The FAT (File Allocation Table) volume rebuilt using the ordered pages (reconstructing her original filing structure). | Constructs an array from the images, and reads nearly every file system |
What the bench did with it.
Find out which half of the card had stopped
Since no output occurred when using a reader, the housing of the memory device was removed and probes placed upon the board. The memory responded immediately and its part number was already catalogued in our library. The controller adjacent to it remained silent throughout multiple attempts. That is the whole diagnosis: the storage portion is present and responding to direct access, and the component responsible for transferring it (controller) is non-functional.
Take the pages off first, sort them out after
Data obtained from memory dies do not resemble disk images. Each page retains its respective error-correcting bytes, and the controller created a mask of its own across all bytes, which must be removed to display the content. Additionally, page order is whatever the translation layer decided while the card was still working. To correctly reconstruct images, three operations must be successfully completed: correcting errors, lifting masks and restoring proper order. When all three are completed correctly, the files will appear. If any of these steps fail, what appears to be a reconstructed image may contain large amounts of random noise.
Rebuild the volume, then open the lot
Upon restoration of order, clean FAT tables emerged and her folder structure was visible again. All items contained within her folders were then accessed and displayed on-screen. While counting files provides limited information related to this type of work, an item that refuses to open will remain unreconstructed regardless of how it appears on a file list.
What went home.
Once the memory provided its contents, they went home on new media after first being checked item-by-item. What the read did reach and what it did not reach is reported in detail. It should now be obvious that storing an item on a memory card that represents your only copy is not advisable. This is a lesson learned with each occurrence of this problem.
Begin here if yours is doing the same thing.
More work off the flash and SSD bench.
Yours doing something similar? Power it down first.
Published prices are fixed. Memory cards and USB sticks are £250 + VAT. Any stand-alone drive is £300 + VAT. Drive configurations containing multiple disks (RAID, NAS or server) start at £500 + VAT. First assessments are always complimentary. Thereafter, data recovery charges are also fixed in advance and in writing, so you'll always know how much you owe before paying any money. Shut it down, get a quote first (by phone or with the form), then send it in or drop it off at our Guildford intake laboratory, Monday to Friday, 9am to 5:30pm (no appointment needed). On most jobs you only pay if we recover your data.