Data Recovery Case File · Solid State & Flash · Unplug It and the PC Works

Why a Data Drive Can Prevent Start-Up

His enquiry describes something that sounds contradictory and is entirely normal. A 2TB secondary solid-state drive: "I've had a secondary drive die on me, and I'm hoping you can help. Initially, it stopped the computer booting at all, however the system did make several attempts" before giving up. A drive that holds no operating system should not be able to stop a machine starting — except that it can, for a reason worth understanding. And the consequence is a free fix: disconnect it and the computer works normally again, which is also the right thing to do for the data.

The decode: why it blocks the boot, and what to do first

What happens before an operating system loads: the machine's firmware inventories every attached storage device. It asks each one to identify itself — what it is, how large, what its capabilities are — and waits for a reply, because devices normally give one. Only when that inventory completes does the machine hand over to whatever it is booting from.

Why a failing drive holds that up: a device that has been detected electrically but cannot answer leaves the firmware waiting. Rather than skipping it immediately, the firmware retries, waits out generous timeouts, and tries again — which is exactly the repeated attempts he observed. The process eventually gives up, but it can take minutes, and on some machines it stalls entirely. The drive does not need to hold the operating system to prevent the machine reaching it; it only needs to be attached and unresponsive.

What that means practically, and it should be done today: disconnect the failed drive. The machine will then start normally and be usable again, because nothing else is wrong with it. That single step solves the computer problem entirely and costs nothing — and it also stops the drive being powered and interrogated on every start-up attempt, which is the right thing for the data.

Why leaving it in is actively bad: every boot attempt is a failing device being powered, asked to initialise, and retried until timeout. If the controller is marginal rather than dead, repeated cycles of exactly that are the wrong way to spend whatever function remains.

What has probably failed: a solid-state drive that is detected and cannot answer, or is not detected at all, has usually had its controller fail to complete initialisation — reading its own internal structures on start-up and not getting through them. Manufacturers build technological and safe modes into their controllers for this, allowing the firmware area to be assessed, damaged modules repaired and addressing restored. That route is tried first.

The honest caveat: where the controller cannot be addressed at all, the fallback available on memory cards does not apply in the same way — on a solid-state drive the mapping between logical and physical locations, and often the encryption, live in the controller, so memory read past it cannot simply be reassembled.

On the bench

The drive was disconnected from the host before any further attempts, since every start-up powered it, asked it to initialise and retried until timeout — and disconnecting also restored the machine at no cost. It was assessed outside the machine on a direct adapter, with presence on the bus and initialisation state read directly. Vendor technological and safe modes were attempted through the PC-3000 portfolio's SSD support, with the firmware area and translation layer assessed and damaged modules repaired until the drive addressed its media.

The outcome

The drive removed from the host, assessed on a direct adapter and addressing restored in vendor modes. 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 machine stopped booting because of a data drive: it does not need to hold the operating system to prevent one loading. Firmware inventories every attached device before handing over, and a drive that is present but cannot answer leaves it waiting through retries and timeouts. Disconnect it — your machine will start normally, and the drive stops being powered and interrogated on every attempt.

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.