Platters, heads, boards, bridges
What is actually inside an external hard drive. And why the box around it matters as much as the drive.
An external hard drive is two things: a hard disk, and a box with a USB board in front of it. The disk holds your files magnetically on spinning platters and reads them with heads flying nanometres above the surface. The box supplies power and translates USB into the disk's own language, and on many drives the two have been merged so that the USB socket sits on the disk's own board. Nearly every failure is in one of five places: the heads, the disk's board, its firmware, the bridge, or the file system on a healthy disk. This page says what each one is, plainly, for anyone whose drive has just stopped and who wants to understand what the bench is about to do to it.
Rather talk it through? An engineer answers the bench line
0800 6890668
Platters, heads and the spindle.
Inside the sealed chassis are one or more platters, glass or aluminium coated with a magnetic layer, spinning at 5,400 or 7,200 revolutions a minute on a spindle motor. Your files are patterns of magnetism on those surfaces. An arm carries a stack of read-write heads, one per surface, which fly over the platters on a cushion of air thinner than a hair and never touch them while the drive is running. When the drive is off, the heads park on a ramp beside the platter or in a landing zone near the centre.
That geometry explains most of what goes wrong. A drive dropped while spinning pushes its heads into the surface; a head that has failed or weakened cannot read, and the arm sweeps, fails and slams back with the click you hear; heads that have come to rest on the platter stick there, and the motor's straining against them is the beep. Every one of those is mechanical, and every power-on of a drive in that state is a failed head dragged across the platter that holds your files.
The circuit board, and the ROM on it.
Under the drive is a circuit board with the motor driver, the preamplifier link to the heads, and a small ROM chip holding data unique to that drive: how its heads are tuned, where its firmware lives, how its surfaces are mapped. A board from an identical drive carries different adaptive data and will not read this one, which is why the online advice to swap the board rarely works and why the bench moves the original ROM to a donor board instead.
Firmware and the service area.
Only part of a hard drive's software is in the ROM. The rest lives in a reserved area on the platters, the service area, and the drive reads it with its own heads every time it starts. If the heads cannot read it, or a module in it has become corrupt, the drive spins up and never announces itself, or reports 0 bytes, or a factory name. Seagate's slim drives lock their diagnostic terminal, and the lock has to be undone before the service area can be reached; on any drive the modules are backed up before a command is sent, because regenerating the translator without a backup is how a recoverable drive becomes a lost one.
The translator.
The drive maps the addresses your computer asks for onto physical tracks and sectors, skipping the ones it has retired. That map is the translator, and on a shingled drive it is doing far more work, because overlapping tracks cannot be rewritten in place and the drive keeps a media cache to buffer the shuffle. A drive with weak heads and a translator out of step with its cache is a drive that crawls, disconnects, and shows a wrong size. It is imaged slowly and deliberately, with the firmware stabilised first.
Bridge boards, and native USB.
A hard disk speaks SATA. The box speaks USB. In a plain enclosure a separate bridge board translates between them, and when it fails the drive comes out and reads directly on its SATA connector. On many portables made since about 2012 — WD Elements and My Passport, Seagate's slim drives, most Toshiba Canvios — the bridge has been merged onto the disk's own board and the USB socket soldered to it, so there is no SATA connector at all. A snapped socket or a dead USB chip is then repaired on the drive's own board, or that board's ROM is carried to a compatible SATA board.
WD's encryption.
Western Digital's My Passport, My Book and easystore go one step further: the USB board carries an encryption engine that scrambles every sector with a key generated when the drive was made and kept nowhere else, whether or not you ever set a password. The password only locks the key. The drive read without its board is noise, which is why the case of a dead WD drive is never thrown away, and why a board swap is the wrong repair. The bench repairs the board so its key survives, or reads the key material out and applies it during imaging. Elements is the exception: no encryption engine.
Why clicking means stop.
The click is the arm giving up and parking, and each attempt before it is a failed head over a surface it can score. Freezing the drive adds condensation; tapping it drags stuck heads across the platter; opening it lets in dust that a flying head cannot survive. None of those was ever a repair. A drive that clicked once and was left alone is a good job; one that was tried thirty times has a ring in the platter where the files used to be.
Head swaps and clean air.
Failed heads are replaced with a matched set from a donor drive of the same family, model, firmware generation and head count, in air filtered clean enough that nothing lands on the platters while they are exposed. The heads are lifted from a stuck platter with the right tools, the new stack fitted with the platters never turned by hand, and the drive closed and brought up once. It is not repaired for use again; it is made to read one more time.
Imaging with head maps.
The drive is then read sector by sector on equipment that knows which head reads which surface, reads the healthy surfaces first at full speed, and comes back to the weak one last, in short passes with strict limits on time and retries, until it gives up what it holds or is marked as lost. Nothing is done to files on the original. The file system is put back together on the image, encrypted volumes are unlocked there with your key, and what comes home is checked by opening it, not counting it.
Where that leaves your drive.
Nearly every external hard drive that reaches us has failed at the heads, the board, the firmware or the bridge, with the platters intact, and nearly every one is read once and its files sent home. A smaller number arrive with platters scored by attempts made before the parcel was packed, and those recover less. The free look tells you which yours is, and what it will cost, before anything chargeable happens.
The questions that come up first.
Can data be recovered from a dead external hard drive?
Usually. The heads, the board, the firmware or the bridge have failed, and the platters underneath are intact. Each is dealt with on the bench and the drive read once.
What does a head swap involve?
A matched donor head stack fitted in clean air, the platters never turned by hand, and one careful image afterwards. The drive is not repaired for use; it is made to read one more time.
Why can't I just swap the circuit board?
The board carries a ROM with this drive's own adaptive data, and on WD's encrypted drives its key. A board from an identical drive will not read it. The bench moves the original ROM to a donor board instead.
Why does my WD drive read as blank outside its case?
Because the case's board encrypts everything it writes, password or not, and holds the only key. Put it back and send the unit whole.
Does any of this cost me anything to find out?
No. The free look identifies the fault and one figure follows in writing. £300 + VAT for one drive.
Now you know what the bench is about to do.
Send the form with the model number and what it sounds like, and the first look tells you which of these your drive needs, and what it would cost.