The most common sentence we hear when a dead laptop lands on the counter is “can you not just take the drive out?” For machines from Wester Hailes, Sighthill and the surrounding EH14 and EH11 streets, the honest answer these days is: sometimes. On a thin-and-light or entry-level laptop bought in the last five or six years, there is a good chance the storage is a chip soldered directly to the mainboard, with no connector and nothing to unplug. Data recovery on those machines follows a completely different route.
This guide explains how to work out which kind you own, what the warning signs of soldered-storage failure look like, and the one rule that decides whether your files come back or not. It is written from what we actually do on the bench in our EH11 workshop, a couple of minutes up the road from Westside Plaza.
Why “Just Take the Drive Out” Is No Longer a Given
For twenty years, laptop recovery had a reliable first move: undo the bottom panel, lift out the 2.5-inch drive or the M.2 stick, put it in a USB dock on a healthy machine, and copy the files off. The laptop could be completely dead — snapped hinge, liquid in the keyboard, blown charging circuit — and it made no difference, because the data lived on a part you could physically remove.
Manufacturers then started saving thickness and assembly steps by soldering the NAND memory straight onto the board. It shows up in three forms. Entry-level Windows laptops and 2-in-1s often use eMMC, a small chip with an integrated controller, usually 32GB, 64GB or 128GB. Mid-range machines increasingly use soldered UFS or soldered NVMe NAND. And every Apple silicon MacBook has its NAND soldered to the logic board and cryptographically tied to the Secure Enclave in that specific machine.
In all three cases the storage cannot be moved to another computer. If the board dies, the ordinary recovery playbook dies with it.
How to Check What Your Laptop Actually Has
You can find out in about a minute while the machine is still working, which is exactly when you should do it rather than after something goes wrong.
Open Device Manager and expand Disk drives. A conventional removable SSD reports a recognisable consumer model name — a Samsung, Crucial, WD, Kingston or SK hynix product string. An eMMC module instead reports a short, cryptic part code: on the bench we regularly see strings like BJTD4R, SDINBDG4 or HBG4a, which are the raw manufacturer part markings rather than a retail product name. If your disk has a name that looks like a chip stamp rather than a product, that is your answer.
A second check that takes ten seconds: open PowerShell and run Get-PhysicalDisk. Removable NVMe drives report a BusType of NVMe and a MediaType of SSD. Soldered eMMC very often reports a BusType of RAID or MMC and a MediaType of Unspecified, because Windows never gets a proper device profile from it.
The third clue is capacity. A machine sold with 64GB or 128GB of storage and no upgrade option is almost certainly eMMC. If your total capacity is a round 64GB and Windows Update has been complaining for a year, you own one of these.
eMMC Wear: The Failure Mode Nobody Warns You About
Here is the pattern that generic advice articles never mention, and it is the single most useful thing on this page.
eMMC is not built like a proper SSD. It has far fewer spare blocks, no meaningful write cache, and a much lower endurance rating. On a 64GB machine, Windows leaves so little spare capacity that every update cycle rewrites the same small pool of blocks over and over. Three or four years of that wears the chip out.
But it usually does not die outright. When an eMMC controller reaches the end of its life it commonly flips the whole device into a permanent read-only state — a protective mode designed to preserve what is already written. The symptoms are unmistakable once you know them: the machine boots but Windows cannot log you in properly, nothing saves, Word claims it cannot write the file, the browser forgets everything on restart, and updates fail with unhelpful errors.
People assume this is a virus or a corrupted Windows install and start reinstalling. It is not, and the good news is genuinely good: a read-only chip is still perfectly readable. Every document, photo and spreadsheet on it can be copied off intact — but only if nobody wipes the machine trying to “fix” it first.
The Rule We Follow: Copy First, Repair Second
On a machine with a removable drive, the order of operations is forgiving. Try a repair, and if it goes badly you can still pull the drive afterwards. On soldered storage there is no afterwards. A failed board repair, a reflash that does not take, or a Windows reset that runs to completion takes the only copy of the data with it.
So the rule on the bench is absolute: if a soldered-storage machine can still be made to boot anything at all, we copy the data off before we attempt any repair — even when the fault looks trivial. In practice that means booting a Linux live USB, mounting the internal storage read-only, and copying everything to an external drive over USB before touching the installed system.
Two practical notes on that. If BitLocker or Device Encryption is switched on — and it is on by default on a lot of these machines — you need the recovery key from the Microsoft account before a live session can read anything, so find it first. Our guide to tracking down a BitLocker recovery key walks through where it hides. And copying a full machine over USB from a live session is slow, so a couple of hundred gigabytes can take most of a working day. It is a booking-in job, not a wait-at-the-counter one.
A Wester Hailes Job That Shows Why the Order Matters
A student at the Napier campus at Sighthill brought us a four-year-old 2-in-1 in the spring. It had stopped saving files, then stopped logging in, and a relative had already tried “Reset this PC” twice — both attempts failed part-way through and rolled back, which as it turned out was the luckiest thing that happened all week.
Device Manager on a live boot showed a 64GB eMMC part, and the kernel log was full of write errors with every read succeeding cleanly. Classic end-of-life read-only mode. Because there was no drive to remove, we booted a live USB, mounted the internal storage read-only and pulled about 38GB off — a full dissertation draft, two years of coursework and a photo library that had never been backed up anywhere.
Had either of those reset attempts completed, the wear-levelling table would have been rewritten and most of that would have been gone. The machine itself was past saving as a Windows device, so the files went onto an external drive and the customer moved to a machine with a proper removable NVMe drive. That is the whole argument for copy-first in one job.
When the Board Itself Is Dead
If the laptop will not power on at all, live-booting is off the table and the job moves up a level. Where the fault is on the power delivery side — a blown charging circuit, a shorted rail, liquid damage around the connector — the practical route is to repair the board just far enough to boot once, then image the storage immediately. That is solder-level work and sits with our microsoldering service rather than a standard repair.
Where the board is genuinely beyond that, eMMC can sometimes be recovered chip-off: the chip is desoldered and read directly in a programmer. It is delicate, it is not guaranteed, and it only works because eMMC carries its own controller. Soldered NVMe NAND is much harder because the controller lives elsewhere on the board.
Apple silicon MacBooks are the honest exception. The NAND is paired to the Secure Enclave of that specific logic board, so if the board cannot be revived the data is not recoverable by anyone, in any lab, anywhere. We would rather tell you that up front than take a machine in on false hope. Our full data recovery service page sets out what is realistic for each type.
Disposal Is Different Too
One consequence people miss: when you retire a laptop with soldered storage, you cannot simply pull the drive out and keep or destroy it. The data goes wherever the machine goes. If the storage still writes, a full encrypted wipe before disposal is the answer; if it has already gone read-only, it cannot be wiped at all and the board needs physically destroying. That is part of what our IT recycling and e-waste disposal service handles, and it is worth thinking about before an old machine goes to a charity shop or a household waste site.
Related Reading
If your situation is a little different, a few of our other guides fit better. Our guide to getting files off a laptop that will not turn on covers the fault tree when the machine is dead rather than read-only. The SSD failure warning signs guide is the right read for conventional removable drives, where the symptoms and the timeline are quite different. And if the machine is still healthy, our backup guide for Windows PCs is the simplest insurance there is against ever needing this page.
What to Do If You Think This Is You
If your laptop has stopped saving files, keeps forgetting settings after a restart, or fails every Windows update with a write error, stop using it and stop trying to reset it. Those are read-only symptoms, the data is almost certainly still there, and every further attempt to repair it in place is a risk.
We collect across Wester Hailes, Sighthill, Broomhouse and Longstone from the workshop on Parkhead Drive, which is genuinely a few minutes away — often the same day. The data recovery service for Sighthill and Wester Hailes page covers how collection works locally, and customers a little further east are covered by our Dalry and Gorgie data recovery page. If the machine is still repairable afterwards, our laptop repair service picks up once the data is safely off.
Whether you are in Wester Hailes, Corstorphine, Livingston or Balerno, the advice is the same and it is the whole point of this article: on a machine with soldered storage, the copy comes first. Everything else can wait.
Last updated: 24 August 2026