There are audible signs of a failing mechanical drive that will occur over weeks. On an SSD there are none. The SMART report for the SSD will indicate flash wear reasonably well but very little about controller health (the controller is typically where these things fail), and therefore you may have an SSD that reports 96% life remaining on Monday and then, with no intervening event, by Tuesday may be completely undetectable. In short, because of this lack of forewarning, your response to the failure in question becomes much more important than with any other type of media we handle.
The behaviour of failing solid state drives (SSDs) is what surprises users most. It is the reason that your response to the post-failure situation is going to matter more than with any other type of media.
A mechanical drive lets everyone know when it's failing. The drive makes clicking noises and whining sounds, gets slower, and usually does so for days or weeks before it finally quits. This warning allows many jobs involving recovery from failed mechanical drives to arrive with the data still largely intact.
On an SSD, there is no similar warning system. SMART reports show reasonable information regarding flash wear, but minimal information regarding the health of the controller (which is usually the component that fails). Therefore, you could have a drive reporting 96% remaining life on Monday, with no intervening events, no noise and no warnings, and by Tuesday the SSD would be completely undetectable.
What comes next isn't theoretical; it's operational: with an SSD, your backup regime is the entire defence against data loss, as you don't have time to react. The hours immediately after a failure matter, because a controller that continues to enumerate sporadically is a much better prospect than one that simply went dead.
Failure of the flash itself is rare. Nearly every failure we've seen here was caused by something further up the chain.
A chip controlling how the flash works (i.e., its controller) fails to respond. The drive will be gone from the BIOS, report zero capacity, or appear as an “unknown” device. Usually the NAND behind it remains completely intact, and readable through other means.
The drive is still operating. However, it has lost track of the mapping between logical addresses and the physical locations of those addresses. It is possible, using vendor-mode access, to fix the tables loaded into memory upon startup, although this requires specific procedures unique to each manufacturer.
While this is the type of failure that most users fear, it rarely occurs in practice since consumer SSDs are usually replaced much sooner than their NAND endurance runs out. However, if an SSD does reach its end-of-life, it typically warns the user by going into read-only mode first (and therefore failing to write), and that warning is worth acting on.
Before you pay someone to attempt a deleted-file recovery from an SSD, keep in mind this one mechanism.
Deleted files remain on a mechanical disk (only the pointer is removed), where they may sit untouched for months. Because of this, deleted-file recovery from a hard drive is so often successful.
SSDs do not work like this. The operating system sends a TRIM command to the controller when a file is deleted. The controller then erases all blocks associated with the deleted file. The NAND memory used in SSDs is frequently cleared within minutes after deletion of the file, as part of routine housekeeping. Once a NAND block has been cleared, it cannot be recovered by any software, any lab, or any other method. All that remains is empty space.
Therefore, whether or not it will be possible to recover deleted files depends on several variables, including how long ago the file(s) were deleted and what type of configuration your SSD uses. If the TRIM process has completed, the answer is “No”. If you removed power to the drive immediately after deletion and/or if TRIM is disabled, there could still be a lot of information available. Another variable could be your RAID configuration, because some RAID controllers do not forward the TRIM command to each individual drive. In any case, we will determine which factors apply at no cost to you. We'd much rather let you know that attempting to recover your files will be fruitless than charge you to look.
Do not keep using the drive while you decide. Each minute an SSD remains plugged into a power source will give the controller another chance to garbage-collect the blocks containing the information you're looking to recover. If something you deleted is important enough, shut off the computer immediately.
A controller's function has more impact upon flash drives than does the brand (or “badge”) itself, since the controller determines whether or not we can gain vendor-mode access, and also determines how the data is scrambled when it is stored onto the flash NAND.
The cost of recovering a failed SSD or NVMe device is from £300 +VAT and is the same as for a typical mechanical hard disk. The price remains the same whether or not the controller needs to be bypassed in order to read the NAND, even though that work is more involved; the price is determined by class at diagnosis, rather than by how difficult the process proves to be.
Most times when we do an SSD/NVMe recovery, you only pay if we recover your data, as most of the time there isn't a fixed parts cost associated with the job (as is common on mechanical repairs).
Yes, frequently. Usually this is due to failure of the controller instead of failure of the flash (the NAND), which often leaves the NAND in full working order. The NAND can be accessed via vendor mode or by directly accessing the memory itself.
In most cases, no. We'll tell you that before taking any payment. Normally, TRIM has wiped out the deleted blocks in a matter of minutes after deletion. However, if the drive was disconnected straight away or TRIM had been disabled, then there's a real possibility. A diagnostic test will determine which scenario exists.
A controller lock-up, most often. The drive has lost or damaged the information that describes how the drive itself is laid out (geometry). This is an error within the firmware and is not due to actual mechanical or hardware damage. It is often one of the more recoverable errors as well.
No. Nearly always, the fault lies in either the controller or firmware of the device rather than the memory. In almost all instances, the data you are trying to recover is indeed located somewhere physically on the NAND (storage), and it is simply the ability to get to it that is impaired.
No, there isn't an inherent reason why this would be the case. However, these types of devices have tighter integration and vendors' methods for accessing them using their proprietary modes vary more from one vendor to another. One thing that does make recovery difficult is when the drive has enabled encryption at the level of the drive itself (several NVMe models enable this by default).
Unfortunately, we do not deal with those. Our services only cover drives that can be removed from your computer, and therefore we will not attempt to recover storage that has been soldered onto a laptop's motherboard or a Mac's board. On both Apple T2 and Apple Silicon computers, the “key” is also essentially baked into a security chip on the same board.
£300 +VAT flat rate, whatever the control unit turns out to have done with it. We first use vendor mode to read, and as a last resort only do chip-level reading where the control unit cannot be revived. You get an honest answer at the diagnostic as to whether TRIM has already removed the data you need.