Data Recovery Case File · Solid State & Flash · Agreement Is Not Corroboration
Programs That Work the Same Way Produce the Same Answer, Which Is One Test Repeated
His enquiry reports an unusually thorough effort and draws the natural conclusion from it. A stick unrecognised across three machines which nonetheless "will report that stick must be formatted", where he has "used 5 generic tools to locate data but 0 results found." Five programs agreeing is less informative than it appears: consumer recovery tools share a method, so running five of them is the same question asked five times.
| Media | USB flash drive — not enumerating usably across three hosts while prompting for formatting; five consumer recovery applications returning no results |
| Reported situation | Flash drive failing across three separate machines · hosts prompting for the device to be formatted · five consumer recovery applications applied · no content located by any of them · advice on next steps requested |
| Fault class | Device not presenting a readable volume — consumer tools operating above the host stack and therefore bounded by the same limitation; content presence unestablished |
| Equipment used | Agreement between similar tools assessed as a single result rather than five · device addressed through hardware with imposed timeouts rather than the host stack · controller state assessed in vendor technological modes · memory read past the controller directly from the package · translation layer reconstructed in software |
The decode: what those five programs share, and why it matters
How consumer recovery software reaches a device: through the operating system. It asks the host to read sectors, and the host asks the device — so every such program sits on top of the same chain.
Why that makes five results into one: they share the layer that is failing. If the host cannot obtain sectors from the device, no program built on the host can obtain them either, however sophisticated its analysis.
Why the programs report nothing found rather than an error: they interpret an empty read as an empty device. A tool that receives no data concludes there is no data, which is a statement about what it was given rather than about the stick.
Why his testing was nonetheless worth doing: it eliminates the software layer entirely. Five tools and three machines rule out the host, the operating system and the applications, which is a genuine narrowing.
What that leaves: the device. The failure sits below everything he has tested, in the controller or in its access to the memory.
Why the format prompt fits that: the device announced itself and offered nothing readable. A host that enumerates a device and cannot interpret a volume offers the only remedy it has, and that prompt is consistent with a controller that cannot reach its memory.
Why the general lesson is worth stating: repeating a method is not corroboration. Five instruments of the same kind agreeing tells you about the method, not about the subject — and it is the commonest way people conclude that content is gone.
What a genuinely different test looks like: addressing the device below the host stack. Equipment that communicates with the controller directly, with its own timeouts, asks a question the host cannot ask.
Why the memory is very likely intact regardless: nothing in his account implicates it. Stored content persists without power and is unaffected by a controller that cannot present it.
What must not happen: no accepting the format prompt, and no further tools. A sixth program of the same kind will return the same result, and the prompt is the only thing that could make this worse.
On the bench
Agreement between similar tools was assessed as a single result rather than five — consumer recovery software reaching devices through the host operating system, so every such program depends on the same chain and none can obtain sectors the host cannot. Empty reads are interpreted by such tools as empty devices, which describes what they were given. Testing across three machines eliminates host, system and application layers. The device was addressed through hardware with imposed timeouts.
The outcome
Agreement between similar tools read as one result, the device addressed below the host stack, and memory read past the controller. Free assessment, one fixed written figure including VAT; where a chip has to be removed, 50% of parts and labour is payable upfront with the balance only on success — otherwise no recovery, no fee. The decode: five programs that work the same way is one test repeated. They all reach the device through the operating system — so if the host cannot read it, none of them can either.
When several recovery programs all find nothing
Stop adding more, and refuse the format prompt. Consumer recovery tools all reach a device through the operating system, so they share the layer that's failing — if the host can't obtain sectors, none of them can, and they report that as an empty device rather than as an error. Your testing wasn't wasted: three machines and five programs rule out the host, the system and the software entirely, which places the fault in the device. But a sixth tool of the same kind will return the same answer.
Don't add a sixth — call Cambridge Data Recovery on 01223 655015; agreement between similar tools read as one result, addressed below the host stack, memory read past the controller.
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.