Data recovery for hard drives

A modern hard drive spins its platters at 5,400, 7,200 or 10,000 rpm with the heads suspended a few nanometres above the magnetic surface. That proximity is what allows densities of hundreds of gigabits per square inch — and also what turns any dust particle into a projectile capable of scratching the platter and permanently destroying tracks. That's why any physical intervention takes place in an ISO Class 5 cleanroom: no more than 3,520 particles larger than 0.5 µm per cubic metre of air. A regular hospital operating room is several orders of magnitude above that threshold.

When the failure is mechanical — damaged heads, seized motor, burned preamp — the only path is to transplant components from an exact donor drive. Same model isn't enough: it must be the same capacity, the same firmware revision, ideally the same date and manufacturing plant. Tiny variations in actuator geometry or preamp calibration make seemingly compatible heads produce unstable reads or no data at all. We keep an active stock of Seagate, WD, Toshiba and HGST donors classified by family.

On modern drives the electronics add another layer of complexity: the PCB carries a ROM or NVRAM with parameters unique to the heads and platters of that specific drive — defect maps, translation tables, preamp adjustments. Just swapping the PCB hasn't worked for over fifteen years: you have to extract the original ROM (sometimes by soldering a BGA chip) and transplant it onto the new PCB. Then it's verified that the drive's service area reads correctly before initiating the image clone onto a destination device.