Data Recovery Case File · Desktop Externals & Aging Drives · Two Observations, One Cause
Watching Throughput While Listening Turns Two Guesses Into One Finding
His enquiry contains a piece of diagnostic work most people never do. A drive giving read errors where recovery software initially accessed files and later crashed, after which "the drive began a soft, high-pitched whirring intermittently — monitoring shows an initial stable read speed which then drops off to zero when the whirring resumes." Either observation alone would be ambiguous; together, and correlated in time, they identify the fault without anything being opened.
| Media | External hard drive — read errors progressing to scan failure; intermittent high-pitched sound coinciding with throughput falling to zero |
| Reported situation | External drive returning read errors and failing to retrieve data · consumer recovery software initially accessing files · software subsequently crashing · intermittent soft high-pitched sound beginning thereafter · scans failing to complete on repetition · read throughput observed as initially stable then falling to zero as the sound resumes |
| Fault class | Mechanical fault confirmed by correlation of audible symptom with throughput collapse — sustained scanning implicated in progression; further scanning contraindicated |
| Equipment used | Sound and throughput correlated in time to distinguish mechanical from logical failure · no further scanning permitted · laminar flow bench inspection with head and platter assessment before any spin attempt · matched donor parts as indicated · imaging under capped timeouts in one supervised session |
The decode: why the correlation is worth more than either half
What a high-pitched sound alone would suggest: something mechanical, probably. It could equally be a supply component, an enclosure resonance, or a fan — a sound by itself invites a guess rather than a conclusion.
What throughput dropping to zero alone would suggest: the drive has stopped returning data. That is consistent with unreadable regions, with a controller problem, or with retry cycles — again several possibilities.
Why the two together are conclusive: they coincide. A sound that begins exactly when data stops flowing, and stops when it resumes, is the same event observed twice — mechanically and electronically — which rules out coincidence and rules out the explanations that would affect only one.
What that identifies: a mechanical action interrupting the reading. Heads unable to hold position, or a mechanism struggling, halts data return for as long as it lasts — and the drive returns to normal throughput when it succeeds again.
Why the initial stability matters too: the drive reads at a proper rate before the sound begins. That establishes the surfaces and the electronics can deliver data normally, so the fault is intermittent rather than general — which is more hopeful than a drive that never reaches speed.
Why the progression is the concerning part: the sound began during the recovery attempt and not before. The drive had read errors first and developed a mechanical symptom while being scanned, which suggests the sustained work contributed.
Why a full scan is such a demanding operation: it reads the entire surface without pause, including regions ordinary use never reaches. A drive with existing read difficulty spends that time retrying, and hours of continuous retrying is mechanical work.
Why repeating it is now the wrong response: each scan attempt reproduces the conditions that accompanied the deterioration. The scans are also failing to complete, so they are costing mechanism and returning nothing.
What is done instead: the drive is inspected before any further running, and where it can be read, imaged under strict per-sector timeouts so that no region absorbs the available time. Healthy areas are taken first while the drive is at its best, and difficult regions revisited rather than allowed to stall the capture.
What must not happen: no further scans and no further connection. He has already obtained the diagnosis; running the drive again adds nothing to it.
On the bench
Sound and throughput were correlated in time to distinguish mechanical from logical failure — an audible symptom alone admitting supply, enclosure or cooling explanations, and throughput collapse alone admitting unreadable regions, controller faults or retry cycles, whereas coincidence of the two identifies a single mechanical event interrupting data return. Initial stable throughput establishes that surfaces and electronics deliver normally, making the fault intermittent. No further scanning was permitted, sustained full-surface reading having accompanied the deterioration.
The outcome
Sound and throughput correlated to identify the fault, no further scanning permitted, and the drive inspected under filtered air before any rotation. Free assessment, one fixed written figure including VAT, 50% of parts and labour upfront with the balance only on successful recovery. The decode: the correlation is what settles it. A sound alone could be several things and a throughput collapse could be several others — but a sound that starts exactly when data stops flowing is one mechanical event seen twice.
Correlating what you hear with what you measure
If you can watch throughput while listening, do — it converts two ambiguous observations into one finding. A sound by itself could be a supply component, an enclosure resonance or a fan; a read rate falling to zero by itself could be unreadable regions, a controller fault or retry cycles. But a sound that begins exactly when data stops flowing, and stops when it resumes, is a single mechanical event observed twice. Then stop scanning: a full scan reads the entire surface without pause, and if a mechanical symptom appeared during one, repeating it reproduces the conditions that accompanied the deterioration.
Stop scanning — call Easy Data Recovery on 028 9002 0144; sound and throughput correlated to identify the fault, inspected under filtered air before any spin attempt, imaged under capped timeouts.
Request a quote online →
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.