A designer’s LaCie Big Disk lit up and spun as normal, but neither macOS nor Windows would show it at all — and it held years of unbacked-up Adobe work. A multi-disk Big Disk is a striped RAID 0 behind a single bridge board, so ‘powers on but not recognised’ points at the bridge or the array’s own structure, not the platters. We bypassed the enclosure, imaged the bare disks, rebuilt the stripe, and recovered the files.
A graphic designer brought in a LaCie Big Disk that had been the working home for years of client projects — Photoshop, Illustrator and InDesign files — with no second copy anywhere. The enclosure powered up, the drive LEDs came on, and there were no clicks or grinding to suggest a mechanical failure. But the volume never appeared: not in the macOS Finder, not in Disk Utility, and not in Windows on a second machine. A drive that is plainly alive yet completely invisible to every computer is a specific kind of problem, and it’s rarely the one people fear.
A Big Disk isn’t a single drive. Inside the enclosure are two or more physical hard disks joined into a striped RAID 0 and presented to your computer, through a single USB or FireWire bridge, as one large volume. Striping is what gives the Big Disk its headline capacity and speed: data is split into blocks and written across the disks in turn. The catch is that RAID 0 has no redundancy whatsoever. There is no mirror and no parity, so the volume only exists while every disk and the array’s structure are present and correct — lose any one member, or the map that describes the stripe, and the whole volume disappears at once. That is the trade every Big Disk makes.
Two things sit between the platters and your desktop in a device like this, and both are common failure points. The first is the bridge board — the small controller that converts the internal disks to USB/FireWire and stripes them together. When a bridge fails or its firmware becomes corrupt, the disks still spin (they get power) but no valid volume is ever offered to the host, so the computer sees nothing to mount. The second is the array and partition structure itself: if the RAID 0 metadata or the partition table is damaged, there is no coherent volume for the bridge to present even when the hardware is fine. On this unit the platters were healthy; the fault lay in the layers above them. That’s good news for recovery — but only if the disks are read directly rather than through the failing enclosure.
Because the bridge was the suspect, the worst approach would have been to keep re-plugging the enclosure and hoping. Instead the disks were removed and each one imaged directly, connected to our own controllers through a write blocker rather than through the LaCie bridge, and cloned sector by sector to healthy storage. Every disk imaged cleanly with no bad sectors, confirming that nothing mechanical had failed and that the entire volume’s data was present across the set — it had simply lost the structure that tied it together. From here on, the enclosure and its bridge were irrelevant; the recovery happened entirely on the images.
The task was now to work out how the Big Disk had striped its data and reassemble it in software. Analysing the images, we established the stripe (block) size and the order of the disks — the two parameters that define a RAID 0 — by looking for the file-system structures and file fragments that only line up correctly when the geometry is right. With the stripe reconstructed read-only, the volume reappeared as a single image, and the damaged partition and HFS+ file-system structures were repaired from their backups and the recognisable catalog data. That restored the folder tree and the file records, so the designer’s projects came back with their names and structure rather than as a heap of anonymous fragments.
With the volume readable, the creative files were extracted and checked — PSD, AI and INDD documents opened and confirmed to render, alongside the high-resolution images and branding assets they depended on. A small number of files that had been mid-write or sat over the one weak area could not be fully rebuilt, leaving a 99% recovery, with everything of importance to the client’s live projects intact. The complete set was delivered on a fresh SSD. As ever, the underlying lesson was hard to miss: a striped RAID 0 device is fast and capacious but has no safety net, so a working Big Disk still needs a separate backup to protect the work that lives on it.
Direct per-disk imaging bypassing the LaCie bridge, write-blocked · RAID 0 geometry analysis (stripe size and disk order) · software re-striping, read-only · partition and HFS+ repair · extraction and validation of Adobe project files. All work in-house at our Belfast lab.
Stop re-plugging it and send it in for a free, no-obligation diagnostic. We’ll read the disks directly, tell you what can be recovered, and put a fixed price in writing before any work starts — and on most jobs it’s no fix, no fee. Post it in from anywhere in the UK, or drop it to us in Belfast.
On an external RAID enclosure like a Big Disk, that pattern usually means the bridge board (the USB/FireWire controller) has failed, or the array and partition structure is damaged — not that the disks themselves are dead. The platters are often perfectly fine. The fix is to read the internal disks directly rather than through the failing enclosure, which is exactly how we approach it.
Usually yes, provided the disks are physically sound. RAID 0 has no redundancy, so a lost member or corrupt array structure takes the whole volume offline — but the data is still spread across the disks. By imaging each disk and working out the stripe size and disk order, the array can be rebuilt in software and the file system repaired, bringing the files back with their names and folders.
Better not to. If the bridge or the array structure is the problem, re-plugging won’t help and repeated power cycling only risks a disk that may already be marginal. The safe move is to stop, keep the disks together and in order, and get a diagnostic — reading them directly is far more likely to succeed than persuading a failing enclosure to mount.