Data Recovery Case File · Desktop Externals & Aging Drives · The Drive That Keeps Hanging Up

The bus divorce, decoded: why a drive "keeps disconnecting" (it stalls, the computer hangs up), why the videos he wants are exactly the reads that trigger it — and the managed image that held the line to the end

The enquiry was the modest, practical kind: "My old HDD keeps disconnecting. I have installed a new SSD, but I need some files from the old HDD. Are you able to assist? They are just video files that I need — not the full HDD." He has already done the sensible half (the system migrated to healthy storage) and needs only the awkward half: files off a drive that won't stay in the room. The decode he's owed: a drive doesn't "disconnect" the way a cable falls out — what happens is a stall-and-hang-up: the drive hits trouble (weak sectors, a firmware struggle), goes unresponsive for longer than the computer will tolerate, and the operating system divorces it from the bus — dropping the device rather than waiting forever. The drive is still powered, still spinning, still trying; the computer has simply hung up on it. And the cruel specificity of his case: the video files he wants are precisely the reads that trigger it — long, sustained, sweeping across whole regions of the disk, guaranteed to visit whatever weak ground the drive has been developing. The road, as ever, inverts the failing approach: stop asking the drive to serve files across a line it keeps losing, and image it once on equipment that never hangs up.

The decode: the hang-up, the trigger, and the selective ask honoured

What "disconnecting" actually is: the operating system runs every drive on a patience budget: issue a read, wait so long, and if the drive stays silent past the limit, declare it gone — the device vanishes from the system mid-session, then often reappears on its next successful breath, only to vanish again. From the user's chair that looks exactly like a loose cable; from the bench it looks like what it is: a drive whose reads are stalling — weak sectors demanding retry storms, firmware wading through its own degrading bookkeeping — for seconds at a time, longer than any host will hold the line. "Keeps disconnecting" is therefore a health report, not a connection report: the cable is innocent, and the drive is telling you, in the only vocabulary the interface gives it, that parts of it now read very, very slowly.

Why the videos are the trigger: and the files he needs are the workload the fault punishes most. Small files are quick visits; a large video file is a sustained expedition — hundreds or thousands of consecutive reads sweeping contiguous regions — statistically certain to cross whatever weak ground exists, and long enough that a single stall inside it kills the whole transfer. Worse, the failure mode compounds: a drop-off mid-copy abandons a half-written destination file (the truncated-video species this archive's transfer cases decode) and, on reconnection, the copy restarts from scratch — re-reading the same weak ground, re-triggering the same stall, the drive spending its decline on a loop that can never complete. Every "let me just try copying them once more" is another lap of exactly that circuit.

The inversion, and his selective ask: professional imaging never gets hung up on, because it never waits on the operating system's patience budget: the drive is read under instrumentation with per-region timeouts owned by the equipment — a stalling read is cut short in milliseconds, logged, deferred; the sweep continues; the weak ground is revisited at the end on its own terms. The line, in effect, cannot go dead. And his "just the video files, not the full HDD" is honoured where it genuinely counts, exactly as this archive's selective cases establish: not in the figure (the fault sets that, and it's fixed in writing either way) but in the imaging order and the delivery — his video folders captured first, while the drive is strongest; verification by playback concentrated on them; and the handover scoped to what he actually asked for, on media sized to it, with the rest of the image standing behind it should he ever want more.

On the bench

The Atola Insight Forensic measured what the disconnections had been reporting: the drive's weak regions mapped as fact, the stall behaviour reproduced and characterised — the cable formally acquitted. Imaging ran on the PC-3000 Express under Data Extractor in the inverted discipline the fault demands: tight per-read timeouts, the video folders captured first in his stated priority, stalling ground deferred rather than fought, then swept up in dedicated end-stage passes. The videos came off the image complete — the expeditions that could never finish over the bus, finished in one pass on equipment that owns the clock — verified by playback, and delivered selectively as asked, with the old drive retired and the new SSD's era begun with everything it had been waiting for.

The outcome

The videos recovered first from a managed image, verified by playback, and delivered to the drive's successor — the disconnections diagnosed as the health report they were. Free assessment, one fixed written figure including VAT, no recovery, no fee. The decode, posted for every drive that keeps hanging up: "disconnecting" is the computer divorcing a stalling drive, not a cable coming loose — it's a read-health symptom wearing a connection costume; large files are the reads that trigger it, so the copies you most need are the ones that fail; every retry laps the same weak ground; and the fix is the standing inversion — stop serving files over a line the host keeps dropping, and image once, priorities first, on equipment that never hangs up.

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.