A USB drive is, in miniature, an SSD: a NAND memory chip driven by a controller, all wired to a USB connector. This apparent simplicity hides the same difficulty as an SSD — translation table, possible encryption, cell wear — plus a mechanical fragility: the connector, the most frequent breaking point.
This chapter is for anyone who wants to understand what's recoverable before handing over a drive. For terms, see the USB & memory card service.
1 · Recognize the failure
Broken or torn connector
The most common mechanical failure: a drive left in a port takes a knock, the connector desolders or the traces break. The memory is intact; the electrical connection must be repaired.
Not recognized / zero capacity
The drive shows 0 bytes, asks to be formatted, or no longer appears: the controller's translation table is corrupt. Data intact but unaddressable.
Dead controller
No response, overheating, short after a surge. Reading then goes through the NAND chip itself (chip-off).
Monolith drive
On many recent drives, NAND and controller are embedded in a single resin block. No chip to desolder: the internal contacts must be exposed.
2 · The laboratory process
Step 1 — Diagnosis & identification
Inspection under a microscope, testing the connector and controller, identifying the NAND type. We determine the route: electrical repair, controller access, or direct memory read.
Step 2 — Connector repair
If the connector is at fault, micro-soldering of the traces and the USB port to restore power and dialogue, long enough to read the content.
Step 3 — Controller route (reverse FTL)
If the controller responds, we put it into factory mode and rebuild the translation table from residual metadata, without the original firmware.
Step 4 — NAND route (chip-off or Spider Web)
If the controller is dead, we desolder the NAND chip (chip-off) to read its raw content. On a Monolith drive, there's no separate chip: we expose the internal contact points by controlled abrasion and micro-solder wires onto them — the Spider Web technique. In both cases, we then reconstruct the interleaving, scrambling and ECC specific to the controller.
Step 5 — Extraction & VeriFiles
File-system reconstruction, data carving if needed, then the VeriFiles list approved before payment. Return on a new device.
3 · Success rates by scenario
- Logical failure (format, deletion) — 82%
- Broken connector (micro-soldering) — 80%
- Corrupt controller (reverse FTL) — 75%
- Monolith drive (Spider Web) — 65%
- NAND cells destroyed — around 10%
4 · The mistakes that destroy data
What you must never do to a failing USB drive
- Keep using it — every write overwrites recoverable data.
- Accept the OS format prompt — badly complicates recovery.
- Force a bent connector into the port — risk of tearing off the NAND or breaking traces.
- Run repair software repeatedly — it writes to the drive.
- Try to open a Monolith drive yourself — abrasion without tooling destroys the internal contacts.
Logical case only. If the drive is healthy and it's a recent deletion, an experienced user can attempt a read-only recovery with reputable software — never writing to the drive.