Data Recovery Case File · Solid State & Flash · Small Memory, Direct Access
Four Megabytes That Cannot Be Re-Recorded
This enquiry is not about photographs or documents. A data logger recording pressures from instruments below ground: "it was recently vandalised on a site and ended up flooded with water. It has 4MB of in-built memory which can be read like a USB stick to download the recorded data. After the flooding it appears as if the data is no longer there." Monitoring data is unusual in being genuinely unrepeatable — nobody can go back and re-measure last month — and the small size of that memory, which sounds like a limitation, is the most helpful fact in the message.
| Media | Industrial data logger with approximately 4MB of integral memory — device flooded following vandalism on site; recorded instrument data not presenting over its normal interface |
| Reported situation | Logger recording instrument pressures below ground · device vandalised on site and flooded · normally read over a mass-storage interface · recorded data no longer presenting after flooding · readings unrepeatable |
| Fault class | Liquid damage to interface and board with integral memory device physically separate — direct device read indicated |
| Equipment used | No power applied after receipt · board examined and residue removed before electrical assessment · memory device identified and read directly on a programmer where the interface was compromised · recorded structure interpreted against the logger's format · series validated for continuity and timestamp order |
The decode: why the size helps, and what has probably failed
Why four megabytes is good news: memory of that capacity in an embedded device is almost never the integrated controller-and-memory package found in a phone or a solid-state drive. It is a discrete memory chip on the board — a small serial device with a handful of legs, holding data in a simple arrangement. Such a device can be read directly, in place or removed and placed in a programmer, entirely independently of the logger's own electronics. That route does not exist on consumer storage of any size, and it is available here precisely because the design is simple.
Why "the data is no longer there" may not mean what it says: he is reading that from the logger's own interface, which is part of the board the water reached. A damaged interface can enumerate and present an empty or unreadable volume while the memory device behind it holds every reading intact. The report of no data comes from the damaged component, not from the memory — and separating those two is the whole assessment.
Why no power should be applied: corrosion is voltage-driven, and site water is worse than tap water — likely to be dirty, possibly saline, certainly conductive. A flooded board that has been powered since is corroding across its traces, and a memory device is only readable while its connections survive. This is the fault class where delay has a real cost.
What the residue means practically: a board that looks dry after site flooding is not clean. Dissolved solids remain, stay conductive, and draw moisture from the air — so residue is removed properly before any electrical assessment rather than the board simply being dried.
What makes this data different from most in this archive: almost everything here is irreplaceable because it is personal — photographs, records, work. Monitoring data is irreplaceable in a stricter sense: it is a time series of conditions that existed at moments now past. A gap cannot be reconstructed, re-derived or re-shot. That also means the recovery has an unusual acceptance criterion — the series must be checked for continuity and timestamp order, because a set of readings with an unnoticed gap is worse than one with a known gap.
The practical note for anyone running instrumentation: loggers on exposed sites are subject to weather, theft and vandalism, and a device that stores locally until collected holds an unbacked window of data the whole time. Shortening the collection interval costs a site visit; losing the window costs the record.
On the bench
No power was applied after receipt, corrosion being voltage-driven and site water conductive and likely contaminated. The board was examined and residue removed before electrical assessment, since dissolved solids remain conductive after a board appears dry. The memory device was identified and read directly on a programmer where the interface was compromised — memory of this capacity in an embedded design being a discrete chip rather than an integrated package, and therefore readable independently of the damaged electronics. The recorded structure was interpreted against the logger's format and the series validated for continuity and timestamp order.
The outcome
The board de-contaminated before assessment, the memory device read directly and the series validated for continuity rather than merely extracted. 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 with a damaged embedded device: small memory is a real advantage. A few megabytes in an industrial design is a discrete chip that can be read directly, independently of whatever else on the board has failed. And a device reporting no data is reporting through its own damaged interface — which is not the same as the memory being empty.
Instrumentation or logger damaged on site
Don't power it up, and get it looked at sooner rather than later — this is the fault class where delay genuinely costs you. Corrosion is driven by voltage, and site water is dirty and conductive, so a flooded board that gets switched on corrodes across its traces while you wait. Take encouragement from the small memory size, though: a few megabytes in an embedded design is a discrete chip on the board rather than the integrated package found in consumer storage, and it can usually be read directly whatever else has failed. That also means your device reporting no data isn't evidence the memory is empty — that report is coming through the same damaged interface. Ask for the recovered series to be checked for continuity, since an unnoticed gap is worse than a known one.
Don't power it — call Newcastle Data Recovery on 0191 406 1051; residue removed before electrical assessment, memory device read directly on a programmer, series validated for continuity and timestamp order.
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.