Data Recovery Case File · Case 1600 · Three Decisions, Three People

Three Things Happened to This Drive and He Was Present for None of Them

This enquiry is the most honest in the archive because the answer may be no. A 500GB external drive holding a local society's records — images and documents, the accumulated work of a group of people. "It was dropped while spinning. Before I got it, an amateur attempt was made to recover it; it was opened, and is in a bad way with scratches on the platter." He did not drop it, did not open it, and is now the person responsible for the outcome — which was largely settled before the drive reached him.

The decode: three harms, in the order they happened

The first — dropped while spinning. This is the worst version of an impact and the difference is not marginal. A drive at rest has its heads parked clear of the platter, mechanically secured, which is why drives survive being posted. A drive that is running has its heads flying nanometres above a surface moving at speed. An impact then drives them into it.

The second — opened outside filtered conditions. Somebody tried to help. A sealed assembly opened in ordinary room air admits particulate that settles onto the surfaces within seconds, and closing it again encloses rather than removes it. Every subsequent rotation grinds whatever landed into the platters.

The third — time. However long the drive sat in that condition between the attempt and reaching somebody equipped to assess it.

Now the thing worth saying plainly: nobody here was careless. The person who dropped it did not intend to. The person who opened it was trying to help a society recover its records and had no reason to know what filtered air is for — that knowledge is not general, and nothing about a hard drive announces that opening it is different from opening anything else. The delay was ordinary.

And yet. No single one of those decisions was fatal. Together they were. A drop while running might have left a recoverable head fault. An opening might have been survivable on an undamaged drive. What removed the margin was the sequence — each step taken reasonably, each one narrowing what the next could achieve.

The honest part: scoring is the one thing that cannot be undone

Almost everything in this archive is recoverable in principle. Boards are repaired or replaced. Heads are replaced with matched donors. Firmware is rebuilt, structures are reconstructed, controllers are bypassed, memory is read directly. The recurring message across sixteen hundred cases is that the situation is usually better than it looks.

Scoring is the exception, and it should be said without softening. Data on a platter is held in a magnetic recording layer a few atoms thick. A scored track is one where that layer has been physically removed — scraped away by a head driven into the surface. There is nothing there. Not damaged, not encrypted, not unreadable pending better equipment: absent. No technique recovers information from material that is no longer present, and none ever will.

There is a second cost that matters practically: a scored surface sheds debris, and debris damages the heads passing over it. So attempting to read across a scored region can spread damage into regions that were intact — which is why a drive in this condition must not simply be spun up and imaged.

What remains genuinely possible

Scoring follows tracks, not whole surfaces. A head travels a circular path, so damage tends to fall in bands rather than across everything. Regions outside those bands are frequently untouched.

A drive has more than one surface. Platters have two faces and there may be several platters, each with its own head. Damage caused by one head affects that head's surface, and the others may be entirely clean — which on a multi-platter drive can mean most of the capacity is unaffected.

So the route is selective: every surface mapped under filtered air before anything turns, scored bands identified and excluded from head travel, donor heads fitted where the originals are damaged, and imaging confined to what can be read without crossing damage. What comes back is real, and how much there is cannot be known before that mapping.

Why documents and images help here: both are small. A file sitting entirely within an undamaged band comes back whole, and a society's records are mostly many small files rather than a few large ones — which means a partial recovery is likely to return complete documents rather than fragments of everything.

The doctrine of case 1600: a device arrives carrying everybody's decisions

He is not being asked what is wrong with this drive. He is being asked what has been done to it. Those are different questions, and on any device that has passed through more than one pair of hands the second matters more — because the fault is what happened, and the handling is what determined whether it still matters.

Which means the useful moment is always earlier than the moment somebody asks for help. By the time a drive reaches a bench, its prospects have usually been set by decisions taken in the first hours by people who did not know they were deciding anything.

So, for anybody who inherits a device: establish the history before the symptoms. Was it running when it was damaged? Has it been opened? Has anything been run on it? Who else has had it? Second-hand information about what was done is worth more than first-hand observation of what it does now — and the person handing it over is the only source of it.

And one thing for the society itself: its records existed in a single place, and everyone involved assumed somebody else had a copy. That assumption is the commonest form of shared-custody loss there is — because no individual felt responsible for a second copy, and no individual was wrong to think it belonged to somebody.

On the bench

The full handling history was established before any assessment — impact while rotating driving heads into a moving surface rather than leaving them parked and secured, and subsequent opening outside filtered conditions admitting particulate that reassembly encloses rather than removes. The drive was treated as contaminated throughout. Inspection ran under the laminar flow bench with surface mapping across every platter face before any rotation, scored bands mapped and excluded from head travel since a damaged region sheds debris that damages heads passing over it, and selective imaging of unaffected surfaces in one supervised session. The honest position was stated before any charge.

The outcome

The full history established, every surface mapped before any rotation, scored bands excluded from head travel and unaffected regions imaged selectively. Free assessment, one fixed written figure including VAT, 50% of parts and labour upfront with the balance only on successful recovery — and no charge where the answer is genuinely no. The decode, and the doctrine of case 1600: scoring is the one damage that cannot be undone, because the recording layer has been physically removed and there is nothing left to read. But it falls in bands rather than across everything, and other surfaces are frequently clean. And more than that: a device arrives carrying every decision anybody made about it. You are not asked what is wrong with it so much as what has been done to it — and by the time it reaches somebody who knows, the outcome may already be settled. Nobody in this case was careless. No single decision was fatal. Together they were.

Our case files are drawn from genuine enquiries received by our laboratory over the past ten years, anonymised to protect client confidentiality. Each one describes the diagnostic and recovery procedure our engineers apply to that fault, using the equipment listed.