Data Recovery Case File · Solid State & Flash · Power Is Not Presence
Lit, and Invisible
His enquiry arrives with the elimination already done by people who know what they are doing. A personal USB stick: "it has an LED which lights up when inserted into a computer; however, no computer recognizes that it is inserted. I have had my company's IT department give it a try as well, and they experience the same — but feel the data is likely still there, just not accessible via conventional means." They are right on both counts. A lit LED and an absent device are not contradictory — they describe two different stages of the same start-up, and knowing which one failed is most of the diagnosis.
| Media | USB flash drive — power indicator illuminating on insertion; not enumerated by any host, confirmed across multiple machines including corporate IT testing |
| Reported situation | Stick inserting and illuminating · not recognised by any computer · behaviour reproduced independently by an IT department · data believed present · personal content held |
| Fault class | Power path functional with controller failing to complete initialisation — memory readable independently of the controller |
| Equipment used | No further insertion attempts · current draw measured on a controlled bench supply · memory packages read at chip level past the controller · translation layer reconstructed in software · files validated by opening |
The decode: what the light means, and what enumeration requires
What the LED actually reports: that the stick is receiving power from the port and that its power regulation is working well enough to light an indicator. That is genuinely useful — it eliminates a dead connector, a broken solder joint on the power pins, a short that would prevent the port supplying current at all, and any suggestion that the device is electrically dead. Something inside is alive.
What it does not report: anything about the conversation. On most sticks the indicator is wired to the power rail rather than driven by the controller, so it lights the moment current arrives — before any communication has been attempted, and regardless of whether the controller subsequently does anything at all. The light is a statement about electricity, not about the device working.
What being recognised requires: considerably more. The controller must start its own firmware, read its internal configuration, and then complete the introduction the interface demands — announcing what it is and what it can do. Only when that exchange finishes does the host list a device. His stick lights and never completes that, which places the failure in the controller: powered, and not starting.
Why the IT department's judgement is sound: they have separated the device from the environment by testing on their own machines, and they have reached the correct conclusion — that the data is likely intact but not reachable conventionally. Sticks keep two things apart: the controller, which conducts the introduction and manages the memory, and the memory packages, which hold the files. Controllers fail far more often than memory, because the controller sits on the power path and does the work. So a stick that is powered and silent has usually lost the part that talks, not the part that remembers.
What "not accessible via conventional means" actually means in practice: that no software can help. Recovery applications need the operating system to present a device they can scan, and where nothing enumerates there is nothing to hand over. The failure of every tool anybody tries says nothing whatever about whether the data survives.
How it is reached: the memory packages are read directly on a programmer, bypassing the controller entirely. What comes back is raw — scrambled, error-coded, and arranged by the controller's own internal scheme — so it is descrambled, error-corrected and reassembled in software through the translation layer until a filesystem appears.
What to stop: further insertion. The position is established across several machines and each attempt powers a controller that cannot start.
On the bench
No further insertion was attempted, since multiple hosts had established the position and each attempt powers a controller failing to initialise. Current draw was measured on a controlled bench supply rather than a host port, distinguishing a controller drawing nothing from one drawing abnormally — which separates a power-regulation fault from a controller that starts and stalls. The memory packages were then read at chip level past the controller, descrambled, error-corrected and reconstructed in software through the translation layer. Files were validated by opening.
The outcome
Current draw established on a controlled supply, the memory read past the failed controller and the files reconstructed and validated. Free assessment, one fixed written figure including VAT, 50% of parts and labour upfront with the balance only on successful recovery. The decode, for anyone whose stick lights up and appears nowhere: those are two different stages. The indicator is usually wired straight to the power rail, so it lights when current arrives and before any communication is attempted — it reports electricity, not function. Being recognised needs the controller to start its own firmware and complete an introduction, and yours does not. That points at the part that talks rather than the part that holds your files.
Stick that lights up but no computer sees
Stop inserting it — several machines have already settled the question. And don't read too much into the light: on most sticks the indicator is wired directly to the power rail, so it comes on the moment current arrives, before any communication has been attempted and regardless of whether the controller does anything at all. It tells you electricity is reaching the device, which is worth knowing, but not that the device works. Being recognised needs much more — the controller has to start its own firmware, read its configuration, and complete an introduction with the host. The encouraging part is what that implicates: sticks keep the controller and the memory separate, and controllers fail far more often, so the part that talks has failed rather than the part holding your files.
Stop trying it — call Newcastle Data Recovery on 0191 406 1051; current draw measured on a controlled bench supply, memory read at chip level past the controller, files checked by opening.
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.