Documented SSD Recoveries
Dying NVMe SSD: NAND Repair with PC-3000
Transcend 512GB NVMe (SM2263XT) | Corrupted service area / bad NAND sectors
Drive showed up as 'ready' in PC-3000 but returned no valid device ID. Service area full of bad sectors. Recovery required save mode activation via pin shorting, firmware injection for the SM2263XT controller, translator rebuilding from surviving NAND metadata, and heat-assisted reading of degraded NAND cells. Final result: 42GB recovered with only 184MB inaccessible (99.6% of customer data).
Full case study →Samsung 970 EVO: Magician Said It Was Fine. It Was Not.
Samsung 970 EVO 2TB | Bad NAND sectors masked by firmware
Samsung Magician reported SMART health as 'Excellent' and speed as 'Excellent.' The drive had 9MB of scattered bad sectors that Magician's sampling never hit. After the drive stopped showing up entirely, PC-3000 sector-by-sector scanning revealed the damage. Three extraction passes with increasing retry counts recovered 42GB of customer data with only 38KB inaccessible (non-critical system cache).
Full case study →MacBook SSD Catches Fire: Can the Data Be Recovered?
MacBook Air (Soldered Apple SSD) | Short circuit / thermal runaway
MacBook Air produced visible smoke from the SSD area on power-up. Multimeter confirmed 0.4 ohms across SSD pins: a direct short causing thermal runaway. Video covers damage assessment, T2/Apple Silicon encryption implications, and the only viable recovery path: repairing the logic board at the component level so the Secure Enclave can power on and decrypt in place. NAND desoldering produces only AES-256 ciphertext without the board functioning.
Full case study →MacBook A1706: Corroded Resistor Kills SSD Detection
MacBook Pro A1706 (Soldered SSD) | Liquid damage / corroded resistor R9350
MacBook A1706 boots to question mark folder after liquid spill. SSD physically intact; the problem was corroded resistor R9350 on the SSD power delivery circuit. The resistor (1.3 megohm pull-up on the IUVD pin of U9330 Piccolo PMIC) read open line. Replacement restored SSD detection. NAND was undamaged; only macOS installation needed reinstall.
Full case study →A1706 Probe Point Corrosion: Same Board, Same Spot, Every Time
MacBook Pro A1706 (Soldered SSD) | Liquid damage / corroded probe point on V2.7 NAND rail
Liquid enters through the keyboard and flows to the same corner where a probe point sits on the V2.7 NAND power rail. The gold-plated surface looks intact, but the copper trace underneath corrodes and disintegrates. The Mac boots, finds no storage, shows the question mark folder. Fix: jumper wire bypass from a good PPBUS source to the buck converter input. Data fully intact after repair.
Full case study →Every SSD recovery case we handle is affected by TRIM status. Here is how different operating systems handle TRIM, and why it matters for your recovery chances.
TRIM Reality Check: What Happens to Your Data by Operating System
| Operating System / Filesystem | TRIM Behavior | Data Recovery Impact | Risk Level |
|---|---|---|---|
| Windows 11 (NTFS) | Enabled by default. Executes on delete and periodically via scheduled optimization. Can be disabled: fsutil behavior set DisableDeleteNotify 1 | Deleted file blocks are zeroed within minutes to hours. Recovery window is narrow. | HIGH |
| Windows 10 (NTFS) | Same as Windows 11. TRIM enabled by default since Windows 7. | Identical to Windows 11. Slightly older optimization scheduler but same destructive result. | HIGH |
| macOS Sonoma/Ventura (APFS) | TRIM runs aggressively and automatically with no user-facing disable option for internal drives. Third-party NVMe drives also trimmed. | Blocks zeroed almost immediately after deletion. Effectively zero recovery window on APFS volumes. | CRITICAL |
| macOS (HFS+) | TRIM support added in El Capitan via trimforce enable. Not automatic on all drives. | Slightly longer recovery window than APFS, but still destructive once enabled. | HIGH |
| Linux (ext4) | Two modes: continuous (discard mount option) or periodic (fstrim.timer systemd service). Neither enabled by default on most distros. | If neither mode is active, recovery window can be indefinite. If discard is set, same as Windows. | MEDIUM-HIGH |
| Linux (Btrfs) | Supports TRIM via discard mount option or fstrim. Not enabled by default. | Similar to ext4. Check mount options and fstrim timer status. | MEDIUM-HIGH |
| ChromeOS | TRIM enabled by default, no user control. | Effectively zero recovery window for deleted files. | CRITICAL |
The takeaway: If your SSD is still detected by your operating system but showing symptoms of failure, UNPLUG IT IMMEDIATELY. Every second it remains connected, your OS may be executing TRIM commands that permanently zero your data blocks. Do not run Disk Drill, TestDisk, or any other software. Ship it to a lab that can image the raw NAND before TRIM finishes the job.
Related SSD Recovery Services
SSD Data Recovery Hub
Full pricing, equipment list, and recovery process for NVMe, M.2, and SATA SSDs. Starting at $200.
Firmware Corruption Recovery
SATAFIRM S11, 0GB capacity bugs, and translation layer failures fixed with PC-3000 firmware tools.
Chip-Off NAND Extraction
When the controller is dead beyond repair on an unencrypted drive, we desolder NAND chips and reconstruct data from raw flash. Not applicable to Apple T2/M-series hardware.
NVMe & PCIe SSD Recovery
M.2 NVMe drives fail differently than SATA SSDs. Thermal throttling, PCIe lane failures, and controller burnout.
SSD not showing up? We record every recovery.
Free evaluation. Published pricing from $200. No data, no fee. Ship your drive or walk in to our Austin lab.