Data Recovery Case File · Electrical & Environmental Damage · The Casualty, Not the Cause
When an Ageing Power Supply Kills a Drive
His enquiry named the cause rather than the symptom, which is unusual and useful. A 250GB SATA SSD that "has failed due to an ageing power supply. The disk was far from being full — guessing from memory, about 25% used. The PC has been upgraded and now has a new SSD. Can you access the failed item with a view to restoring lost data?" The framing matters. This is not a drive that wore out or was mishandled: it is a drive that was on the receiving end of something else failing, which changes what is likely to be broken and where the data is likely to be. It also raises one practical warning he may not have considered, about which machine not to test it in.
| Media | 250GB SATA solid-state drive — failed following power supply deterioration in the host machine; approximately 25% of capacity in use; host since rebuilt with replacement storage |
| Reported situation | Failure attributed to an ageing power supply · drive lightly used relative to capacity · host machine upgraded with new storage · original drive retained for recovery |
| Fault class | Electrical damage from supply irregularity — controller or power-regulation components affected; memory typically unaffected |
| Equipment used | Not tested in the original host · board inspected under magnification with regulation components examined · vendor technological and safe-mode access (PC-3000 portfolio) · imaging on restored access; contents verified by opening |
The decode: what a failing supply does, and why storage suffers first
What "ageing power supply" actually means electrically: a power supply degrades in ways that are worse than simply switching off. Its filtering components dry out over years, so the output develops ripple — voltage that wavers rather than sitting steady. Under load it may sag, dipping below the tolerance the components downstream expect. And at switch-on or switch-off it can deliver spikes well above nominal. A supply doing all three still runs a computer, often for months, which is why these faults are noticed late.
Why storage is usually the casualty: drives take two supply rails and both matter — one feeding the electronics and, on mechanical drives, one feeding the motor. Drive boards contain small regulation and protection components sized for a stable input, and they are among the least tolerant things in a computer of a supply that wanders. Processors and memory sit behind their own substantial regulation on the mainboard; a drive has comparatively little. So when a supply misbehaves, storage frequently fails first and most visibly.
What is likely to be damaged on his drive: the electronics rather than the memory. Solid-state drives hit by supply irregularity commonly lose a power-regulation component or the controller itself, while the NAND holding the data is untouched — it is not powered directly and it is not what fails. That is the hopeful reading, and it means the realistic route is controller-level access through vendor technological modes once any board fault is addressed.
Why 25% used helps: two small ways. Less data means a shorter image and fewer regions that must read cleanly. And on solid-state storage, a lightly-used drive has done far less wear-levelling work, so the memory itself is in better condition than the failure suggests.
The warning about which machine: and this is the practical point. The obvious test is to put the drive back into the PC to see whether it works — and that PC contained the supply that damaged it. If the supply has been replaced during the upgrade, fine. If it has not, connecting the drive to it again risks repeating exactly the event that caused the fault, on a drive that has already been damaged once. Test on a known-good machine, or not at all.
On the bench
The drive was not connected to its original host, since the supply implicated in the failure may still be in it. The board was inspected under magnification with the power-regulation and protection components examined specifically, because supply damage concentrates there rather than in the memory. Access was then attempted through the PC-3000 portfolio's SSD support in the drive's vendor technological modes, with the firmware area assessed once the board could support operation. Imaging ran on restored access, and the contents were verified by opening before delivery on fresh media.
The outcome
The board damage identified and addressed, the controller accessed in vendor modes and the contents imaged and verified. 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, for anyone whose drive was killed by a power supply: an ageing supply does not simply stop — it develops ripple, sags under load and delivers spikes at switch-on, all while still running the computer, which is why it is noticed late; storage is usually the first casualty because drive boards carry comparatively little regulation of their own; the damage lands on the electronics rather than the memory, so your data is normally intact behind a damaged board; and do not test the drive in the machine that killed it unless the supply has been replaced.
Drive that failed when your power supply did
Don't test it in the same machine. That's the practical trap here: the obvious move is to plug the drive back in to see whether it still works, and the supply that damaged it may still be sitting in that computer waiting to do it again to a drive that's already been hurt once. Use a different machine, or leave it alone. Take some encouragement from what supply damage actually does — a failing power supply develops wavering output, sags under load and throws spikes at switch-on, and the components that suffer are the small regulation parts on the drive's own board rather than the memory holding your files. Your data is very probably intact behind damaged electronics. And if other drives were in that machine, treat them as suspect too: a supply that killed one has been feeding all of them.
Don't retest it in that machine — call Glasgow Data Recovery on 0141 404 0294; board inspected at component level, controller addressed in vendor modes, contents imaged and verified.
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.