Board-level faults · bench record · CDF-2025-0603
The Damage Arrived With the Power.
A digger on a neighbouring site caught a cable, and a garden centre on the A48 between Cardiff and Newport went without power for a morning. When it came back, the desktop in the office showed Default boot device missing
, while the BIOS was blunter still: no drive found on either channel
. Behind that message sat the plant database, every photograph on the website, and nine years of supplier accounts.
Does that match yours? Call us.
0800 6890668
Why that happens.
Data on a hard disk is not stored as electricity. It is stored as the direction tiny magnetic grains are facing in, and a direction offers a surge nothing to grab hold of. That single fact decides most cases like this one. The damage from an outage is almost never done while the power is off; it is done in the first moment it returns, which is far rougher than the cut itself. Drive makers expect it. A cheap component sits across the incoming supply for one reason only — to give up before anything costly does — and when it has, the board is finished. The disk behind it is not. So a drive that disappears the moment the lights come on is usually an intact archive trapped behind a burnt circuit.
What we used on this one.
Every step, in order →| The tools | What it did here | Why it helps |
|---|---|---|
| ACE Lab PC-3000 Express | Verified the replacement board — SMART log, defect tables and model string all answering | Opens the maker's service commands, which a plain ATA link will not pass |
| PC-3000 Data Extractor | Then imaged the surfaces, the drive’s own defect maps loaded into the tool first | Imaging driven from inside the drive's own firmware, by the PC-3000 |
| R-Studio Technician | Mounted the NTFS volume within the image and checked off every folder | Reads nearly every file system, and rebuilds arrays you can rely on |
How it went.
Find out where the energy actually stopped
Under a lens, the sacrificial part is obvious once it has blown — and spotting it answers very little. What matters is where the energy went next. A good deal of it often travels on past that point — into the interface chip, into the motor driver — so every stage beyond was tested before this board was called good. Skipping that is how you end up with a drive that dies halfway through the imaging run.
Carry the original ROM over to the donor board
A donor of the same revision gave us sound electronics. What it could not give us was this drive’s own memory, which had to be lifted off the burnt board. That ROM holds adaptive figures measured on the line for one particular head stack and one particular set of platters — these. Fit a board without it and the disk spins up, sounds entirely well, and returns noise.
Load the defect map before the first read
With the donor in place the Barracuda named itself correctly, full 2TB capacity and all, and its defect tables went into the imager before one sector was read. The copy ran through with no rereads whatsoever. That is how platters behave when the harm stopped at the circuit board.
Signing it off.
The lot was verified and went back on a new disk inside the same week. There is a UPS under the till machine and the office machine now, and the supplier accounts write themselves to a second drive overnight.
Pages people often read next.
More from the power cases.
Is that what yours is doing?
Power it down, get it to us, and hold off on any decision until the diagnosis says what can still be read.