Data Recovery Case File · Mac & Apple Ecosystem · Two Meanings of One Verdict
Told the Memory Chip Has Failed: What to Establish
His enquiry came from overseas and described a failure with no obvious cause. A MacBook Pro: "I shut it down as normal, no forced shutdown. Then the next day, it just wouldn't turn on. I took it to a repair centre here. They said there is a problem with the NAND chip. Have I lost all my data for good, or is there any chance of recovery? I have a photo of the chips I can send." He also offered a theory about insects getting into the machine overnight, which is less far-fetched than it sounds. The verdict he has been given is the important part, and it needs unpicking — because "a problem with the NAND chip" can mean two quite different things, only one of which is the end of the road.
| Media | MacBook Pro with soldered, hardware-encrypted storage — failed to power on following a normal shutdown; assessed overseas with a memory-chip fault reported |
| Reported situation | Normal shutdown followed by complete failure to power on · no impact, liquid or forced shutdown reported · repair centre attributing the fault to the storage memory · board photographs available · device to be brought to the UK |
| Fault class | Reported storage-memory fault requiring differentiation from a board-side addressing failure — outcome contingent on which applies |
| Equipment used | Board photographs reviewed in advance · board assessment distinguishing memory failure from supply or addressing faults · micro-level repair where the board is the fault · imaging on restored access with the owner's own credential |
The decode: two meanings, and why the distinction is everything
Meaning one — the memory itself has failed. On a Mac of this class the storage is soldered to the logic board and encrypted with keys held in a dedicated security chip. If the memory packages themselves have genuinely failed, that is the honest end of the road: there is no module to remove, no equivalent chip that holds his data, and no route around it. Said plainly because he asked plainly.
Meaning two — the board can no longer address the memory. Far more common, and it produces an identical symptom from outside. A failed power rail feeding the storage, a damaged trace, a fault in the controller path — any of these means the machine cannot talk to memory that is perfectly intact. From a repair centre's perspective the machine does not see its storage, and "a problem with the NAND" is a reasonable way to describe that. But it is a board fault, it is repairable at micro level, and the data behind it is fine.
How to establish which: by asking how the diagnosis was reached. Was the fault measured — the supply rails to the storage checked, the addressing path traced — or was it inferred from the machine not seeing its own storage? Those are different standards of evidence, and the second is what most repair centres can offer, entirely reasonably. That distinction is worth establishing before he concludes anything, and it is why a proper board assessment is the next step rather than a verdict.
Yes, send the photographs: genuinely useful. Clear board images show physical damage, corrosion, disturbed components and evidence of previous work, and they allow a realistic view to be offered before he ships anything internationally. They will not settle the question — that needs measurement — but they meaningfully narrow it.
On the insects: not a silly theory, and worth answering rather than brushing aside. Anything conductive bridging a live board can cause a short, and small insects are drawn to warmth and do get inside equipment. It would not change the route — the board is assessed the same way regardless of what caused the fault — but it does not need to be dismissed. A normal shutdown followed by a dead machine is equally consistent with an ordinary component failure, which needs no explanation at all.
The question to settle before travelling: the account password. On a machine of this class, even a perfect board repair yields nothing without it, and that check costs nothing.
On the bench
The board photographs were reviewed before the machine travelled, so that a realistic view could be offered in advance rather than after an international journey. On arrival the board was assessed to distinguish the two possibilities: the supply rails feeding the storage checked, the addressing path traced, and the memory packages and security chip inspected specifically — because that determination decides whether a recovery exists at all rather than merely how difficult it is. Where the fault proved to be board-side, repair proceeded at micro level until the machine would run and unlock its own storage against his password, with imaging beginning immediately on that access.
The outcome
The two possibilities separated by measurement, the board repaired where it was the fault, and the data imaged against the owner's own credential. Staged assessment with findings in writing, 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 given this verdict: "a problem with the memory chip" covers two situations — the memory itself having failed, which on a soldered encrypted Mac is genuinely the end, and the board being unable to address memory that is perfectly intact, which is repairable and far more common; ask how the diagnosis was reached, because measuring the supply rails and addressing path is a different standard from inferring it; send board photographs, which narrow the picture before you ship anything; and confirm you have the account password, because without it even a perfect repair returns nothing.
Told a soldered storage chip has failed
Ask how they know. On machines with storage soldered to the board, "the memory chip has failed" is used for two very different findings: the memory genuinely having failed, which on an encrypted Mac is the honest end of the road, and the board being unable to address memory that's completely intact — a failed supply rail, a damaged trace, a fault in the controller path. Both look identical from outside, because in each case the machine can't see its own storage. The question that separates them is whether the supply rails and addressing path were measured or whether the fault was inferred from the symptom. Take clear photographs of the board before you move the machine anywhere; they show corrosion, damage and previous work, and they let someone give you a realistic view before you ship it. And confirm you know the account password, because without it no repair helps.
Ask whether it was measured or inferred — send photos and call Glasgow Data Recovery on 0141 404 0294; board assessed to separate memory failure from an addressing fault before anything is concluded.
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.