If your USB flash drive was recognized until yesterday and now shows nothing at all, or reports zero capacity, or throws Code 43 in Device Manager, and the connector on the drive looks perfectly fine — you’re looking at the failure mode where chip-off recovery pulls the data off the memory directly, bypassing the failed component entirely. Chip-off is not connector repair. It is not a fix. It is a controlled extraction of the NAND flash memory chips off the failed drive’s circuit board so that we can read the memory contents on isolated recovery hardware and reconstruct the file system in software. Gillware has been performing chip-off recovery on USB flash drives since 2004 in our ISO 5 Class 10 cleanroom in Madison, Wisconsin. Every case starts with a free in-lab evaluation.

Start a free chip-off recovery evaluation →

What chip-off recovery does

A USB flash drive has three main functional components on the PCB inside its housing: the USB connector (the physical bridge to the computer), the controller chip (the logic that translates USB requests into NAND memory operations), and the NAND flash memory chip or chips (where the actual data lives). When the connector fails, micro-soldering can restore it. When the controller fails, or when the PCB itself is damaged in ways that leave the NAND intact, connector repair does nothing — the physical bridge is fine but the drive still can’t deliver data through the USB interface because the internal logic isn’t responding.

Chip-off recovery goes around the failure. The NAND chip is desoldered from the drive’s PCB and placed in a socket adapter that connects it directly to specialized hardware capable of reading raw NAND memory outside the context of the drive’s own controller. The raw NAND contents are captured to a file. Then, in software, we emulate what the original controller would have done — parsing the flash translation layer, mapping logical addresses to physical pages, correcting bit errors, and reconstructing the file system exactly as it appeared to the operating system when the drive was working.

Chip-off works on traditional-construction drives that use discrete NAND and controller chips on a PCB. It doesn’t work on monolithic drives (Ultra Fit, iXpand Mini, FIT Plus, most compact USB-C drives) where the NAND and controller are combined on a single silicon die — those drives have no chip to desolder and follow a different recovery path involving direct access to test pads on the monolith package.

Failures we recover with chip-off

Failed controller, intact NAND

The most common chip-off scenario. The drive plugs in, the connector is intact, but the operating system doesn’t recognize the drive at all — or recognizes it with zero capacity, or shows a Code 43 error in Device Manager. Under magnification the controller chip is often visibly intact. What’s failed is the controller’s internal logic: sometimes the firmware has corrupted, sometimes the controller has taken electrical damage from a power surge, sometimes the controller has simply aged out after long high-temperature deployment.

The NAND flash memory is unaffected by controller failure. Data on the NAND continues to hold its charge regardless of what the controller is or isn’t doing. Chip-off removes the NAND from the failed drive, reads it on our own hardware, and reconstructs the file system in software. This is the case with the highest chip-off success rate: the controller was the failure point, the data was never touched, and recovery is largely a controlled sequence of well-understood procedures.

Destroyed or unreadable PCB, intact NAND

The drive has been physically damaged in a way that leaves the PCB unusable but leaves the NAND chip itself intact. Common scenarios: a drive crushed hard enough to fracture the PCB substrate, a drive burned in a fire where the plastic housing melted but the ceramic NAND package survived, a drive with the PCB traces so corroded from moisture ingress that no amount of cleaning can restore continuity between the connector and the controller.

NAND flash packages are surprisingly resilient to physical trauma. The memory die inside is protected by an epoxy encapsulation and by the metal or ceramic package itself. Drives that arrive at the lab with the PCB in fragments frequently have NAND chips that read normally after being desoldered and placed on the socket. Recovery success on these cases depends entirely on the NAND survival, not on the state of the rest of the drive.

Prior DIY repair damage

The drive arrived at the customer’s workstation with a snapped connector, was worked on with a general-purpose soldering iron, and is now in worse condition than when it started. Common failure signatures: pads lifted by an oversized iron tip and torn away from the PCB, solder bridges between adjacent data pins, heat damage that has shifted or destroyed the controller, and flux residue eating at the copper traces.

When the connector-side repair has been made too damaged to salvage, chip-off bypasses the damaged region entirely. The NAND is desoldered from the still-functional side of the PCB and read directly, ignoring the damage on the connector end.

Fire, water, or corrosive events

The drive has been through a scenario that would kill most electronics: pulled from the ashes of a house fire, retrieved from a flood, submerged in seawater, dropped into a chemical spill. What matters for chip-off recovery is not the outward condition of the drive but the survival of the NAND package. NAND flash chips are enclosed and can survive events that reduce the surrounding PCB to unusable condition. We’ve pulled data from drives whose exterior plastic melted, whose PCBs corroded down to unreadable fragments, and whose USB connectors are functionally gone — because the NAND package itself was still intact underneath.

Recovery from these cases involves careful pre-cleaning to remove contamination that could damage the NAND during handling, controlled desoldering with attention to any heat damage the NAND may have already sustained, and often extended raw-read sessions to work around bit-level damage in specific memory cells.

ESD (electrostatic discharge) events

The drive was working normally until someone reached for it after walking across carpet in a dry room. Static discharge through a USB connector can destroy the controller chip in a fraction of a second, leaving the drive dead to the operating system but with the NAND memory unaffected. ESD damage typically kills the controller entirely rather than partially: the drive doesn’t enumerate at all after the event.

Chip-off recovery on ESD-killed drives is well-understood work with high routine success rates. The NAND survived because the discharge went to ground through the controller, protecting the memory itself.

When chip-off isn’t the recovery path

Chip-off is powerful when it’s the right tool. It’s not always the right tool. Situations where recovery goes elsewhere:

Connector-only physical damage where the drive’s internal logic is fine — the connector is snapped, but the controller and NAND are undamaged. Chip-off would work, but it would be unnecessary work. Recovery goes to micro-soldering instead: reattach the connector to the PCB and image the drive normally through USB. Same result, less invasive procedure.

Monolithic drive construction where the NAND and controller are combined on a single die inside one small package (Ultra Fit, iXpand Mini, FIT Plus, most compact USB-C drives, most sub-32GB modern flash drives). There’s no discrete NAND chip to desolder. Recovery goes through the gold contact fingers on the top of the monolith package using a fine-probe station.

NAND-level physical destruction where the memory package itself has been cracked, shattered, or heat-damaged beyond function. Chip-off can only work when the NAND is still readable. When the NAND itself is destroyed — a rare event, requiring severe crushing force or extreme heat — no recovery path exists.

The full explanation of when each path applies is on our USB flash drive recovery pillar page.

What happens at the chip-off bench at Gillware

1. Cleanroom disassembly and NAND identification. The drive is opened in the ISO 5 cleanroom. The engineer identifies the specific NAND chip on the PCB — manufacturer, package type (TSOP, BGA, LGA), pin count, and the controller family the drive uses. Every NAND package has slightly different desoldering requirements and different socket-adapter compatibility.
2. Controlled NAND desoldering. The NAND is removed from the PCB using controlled hot-air rework, with temperature and duration calibrated for the specific package type. TSOP packages come off with a temperature-controlled hot-air pencil directed at the package leads. BGA packages come off with a preheat station followed by targeted hot air. LGA and other surface-mount packages require their own procedures. Under-temperature desoldering damages the pads on the PCB (which we won’t reuse but which affects the recovery timeline). Over-temperature desoldering damages the NAND die itself — the case Claude cannot come back from.
3. Pad cleaning and NAND inspection. The desoldered NAND package is inspected under microscope for pad damage, corrosion, or physical trauma to the ceramic or plastic encapsulation. Any solder residue on the NAND pads is cleaned. Any obvious contamination on the package is addressed before the chip goes into the socket adapter.
4. Socket adapter and raw NAND read. The NAND is placed in a socket adapter matched to its package type and pin count. The adapter connects to specialized NAND-reading hardware that operates outside the context of the original drive’s controller — it reads raw physical pages from the NAND array, ignoring whatever the controller would have translated. The read produces a raw binary image of the NAND contents. On drives with multiple NAND chips (higher-capacity Extreme Pro drives, for example), each chip is read separately.
5. Controller emulation and FTL parsing. The raw NAND image is not a file system — it’s the physical memory laid out as the drive’s controller organized it, which is not the same as how the operating system saw it. Our in-house software (HOMBRE) emulates what the original controller would have done: applying the flash translation layer for the specific controller family (SanDisk, Kingston, Longsys, Silicon Motion, Phison, and others we’ve reverse-engineered over the years), mapping logical addresses to physical pages, applying wear-leveling reversal, and correcting bit errors using the ECC information stored alongside each page.
6. File system reconstruction. Once the logical drive image is reconstructed, we parse the file system on it — FAT32, exFAT, NTFS, HFS+, ext4 as appropriate — and extract individual files. Damaged or partial files are flagged and repaired where possible.
7. Data return. Recovered files are returned on new media or transferred securely. The original NAND and PCB are not returned in operational condition — chip-off is by nature destructive to the source drive.

Why DIY chip-off almost always destroys the NAND

Chip-off videos on the internet make the procedure look approachable. In the lab, it is emphatically not. The specific reasons DIY chip-off attempts fail:

• Hot air temperature and duration. NAND flash memory dies are damaged by prolonged exposure to temperatures well below the temperature required to melt the solder holding the package to the PCB. The window between “solder is molten” and “NAND die is damaged” is narrow, and it varies by NAND generation and package type. A hot-air gun set too hot or held too long against the package destroys the memory before the solder releases.
• Mechanical stress during removal. A NAND package that’s only partially released from the PCB will crack if lifted with normal force. The cracking is often invisible externally but severs the internal bond wires that connect the die to the package pins — the chip looks fine but reads no data.
• Contamination on the pads. Solder flux and rework debris left on the desoldered NAND pads prevent the chip from making reliable contact in a socket adapter. Every pad has to be clean and flat before the chip goes into the reader.
• Wrong socket adapter. NAND packages come in many pin counts and pinouts. A TSOP-48 socket won’t work with a TSOP-56 chip. A generic adapter without correct routing to the reader hardware won’t transfer data reliably. Getting this wrong doesn’t just fail to read — it can burn the NAND from applied voltage on the wrong pins.
• No way to interpret the raw read. Even if the NAND is successfully desoldered and read, the resulting binary is not the file system. Without controller emulation, the data appears as scrambled fragments — wear-leveling has moved logical blocks around, error-correction data is intermixed with user data, and blocks are ordered by physical position rather than logical file layout. Making sense of the raw read requires the same in-house software labs like Gillware have built over years of reverse-engineering controller behavior.

If your drive has been through a DIY chip-off attempt and stopped responding, tell us that when you submit the case. The failure signatures are visible under the microscope and we’ll assess whether the NAND is still recoverable.

Why Gillware for chip-off recovery

ISO 5 Class 10 cleanroom. Chip-off desoldering releases microscopic amounts of solder, flux vapor, and package debris. Doing this in a controlled particulate environment prevents contamination of the NAND pads and preserves the drive’s recovery prospects.

Full controller-family reverse engineering. Chip-off is only half the job. The other half is emulating the specific flash translation layer of the controller family the drive used. Our in-house software (HOMBRE) has controller emulation for the major USB flash drive controller families used across SanDisk, Kingston, Corsair, PNY, Lexar, Samsung, and OEM-branded drives.

Full socket-adapter library. We maintain adapters for every mainstream NAND package used in consumer flash drives — TSOP-48, TSOP-56, BGA-63, BGA-100, BGA-132, BGA-152, and the various LGA variants.

Multi-chip drive handling. Higher-capacity flash drives (Extreme Pro, Kingston DataTraveler Ultimate, Samsung BAR at larger capacities) use multiple NAND chips in parallel for performance. Chip-off recovery on multi-chip drives requires reading each chip separately and reconstructing the interleaved data pattern. This is standard work in our lab.

More than two decades of USB chip-off. Gillware has been performing this work since 2004 in Madison, Wisconsin. Your drive does not leave the country.

Pricing and engagement

The evaluation is always free. After our engineers inspect the drive and determine whether chip-off is the correct path (or whether the case moves to micro-soldering or monolith recovery instead), you receive a firm written quote — not a range, not an estimate that grows — before any recovery work begins. You decide whether to proceed.

Standard chip-off recoveries operate under our “no data, no charge” engagement: if the recovery is unsuccessful, you don’t pay for the work. That covers controller failures with intact NAND, standard PCB-damage cases, and standard ESD-killed drives. Cases involving significant additional engineering — drives with heat-damaged NAND, drives with multiple failure modes, drives through prior DIY chip-off attempts, or multi-chip drives requiring extensive interleave reconstruction — are quoted individually before work starts. More on data recovery pricing →

Start your chip-off recovery

If your USB drive has stopped being recognized and the connector looks physically fine, or if the drive has been through fire, water, corrosion, or a prior DIY repair attempt, the next step is to stop plugging it in and start a free evaluation. We’ll receive the drive, inspect it in the cleanroom, tell you exactly what path applies (chip-off, micro-soldering, or monolith recovery), and quote you a firm price before any work begins.

Start a free chip-off recovery evaluation →

Prefer to talk to someone first? Call 1-877-624-7206 during business hours (M–F 8 am–7 pm, Sat 10 am–3 pm Central), or schedule a 15-minute consultation with a client advisor. For related recovery scenarios, see our USB flash drive recovery pillar, our page on micro-soldering for connector-side failures, or brand-specific pages for SanDisk, Kingston, Corsair, PNY, Lexar, and Samsung.