Signature failure, controller corruption & firmware corruption: three common terms for what is normally the same underlying event on an 860 EVO. The layer containing V-NAND dies is normally unaffected; it’s the layer that knows how to read from those dies which has failed.
The 860 EVO stores your data across multiple blocks of V-NAND dies, plus a translation layer describing where each block actually resides. Lose the translation layer and although the drive still holds all of your data, it no longer knows how to present it to you.
Although these three terms are used interchangeably in real-world applications, they generally refer to different phases of the same type of failure. The MJX controller located within the Samsung 860 EVO loads its operating firmware from a reserved area on the NAND flash memory when it first powers up. Prior to running the firmware, the controller checks whether it was loaded correctly. Signature failure is when the controller determines the firmware did not match what it expected, and it will then refuse to load. In this case, the reason for the signature failure would be due to some level of corruption of the firmware itself. Firmware corruption could occur if one or more modules in the service area were invalidated. Controller corruption is the symptom you see: a corrupted controller is essentially unusable as a drive, as there would be no valid firmware to run.
This will be typical behaviour. The drive may either completely disappear from the BIOS, have a blank model in the BIOS or show up with an absurd capacity (e.g., 0MB, 8MB etc.) as opposed to what it really is. A few drives will go into “safe mode” so they will identify themselves to you but refuse to allow you to read anything off them.
Loss of power in the middle of writing is the most frequent cause of the problem. Because the controller continuously updates its mapping tables, when you lose power unexpectedly, one or more service area modules may be written halfway through and have a bad checksum when they come up again. Ageing also contributes to this problem; as with all flash memory, the reserved block(s) containing firmware wear out. When the controller cannot find a good image of a service area module that it knows should exist, there is no place for it to go.
The third reason is “self-induced”: a firmware update that freezes or an update to a drive that was already unstable. Firmware updates by Samsung Magician will rewrite the service area, but if you interrupt an update on a drive that was malfunctioning prior to this point, it will leave the exact same condition that we have above. Therefore, attempting to run a firmware flash on a drive that is already malfunctioning can be considered the worst thing you could do to a drive.
The objective is to circumvent the failed firmware as opposed to repairing the drive to allow for use. The controller is placed in its vendor’s technical mode allowing for direct access to the service area and we extract the remaining functional firmware modules and translation tables from this location. If any of these modules are not readable then an equivalent from a donor (with the same capacity and version) can serve as an adequate substitute thus bringing the controller online enough to support addressing the flash memory.
With the drive now responsive, it is imaged in read-only mode and the file system reconstructed from this image versus using data from the original drive. In cases where the controller cannot be recovered at all then the next best option is to access the V-NAND directly; however accessing the V-NAND produces raw flash memory that will need to be de-interleaved, corrected for error correction code, and rearranged per the controller’s method of organisation prior to being usable. This process is both time consuming and requires specialised knowledge; therefore, this option may not be viable for each case.
First, there is encryption. The 860 EVO supports hardware-based encryption, as well as Class 0 or TCG Opal. If that is turned on (or used by BitLocker) then the data will be unreadable without the proper login information. No recovery service can bypass encryption, even with an exact copy of the disk. Without the key/password, a complete perfect image will yield nothing useful. Before you assume the worst, check for a BitLocker recovery key in your Microsoft account.
Second, there is TRIM. If you deleted files from a functional drive, TRIM has almost certainly already removed all of the underlying blocks so there would be nothing to recover. This is a completely different scenario than what is described on this page. On this page we are discussing failure of a drive containing your data (which is recoverable). A complimentary diagnostic lasting 48 hours determines which of these scenarios you are experiencing, prior to any quote being generated.
Prices are as follows: recovery of an SSD (such as Samsung’s 860 EVO) and NVMe starts at £300 +VAT. Each job includes a free 48-hour diagnostic and a detailed written quotation which outlines what you’ll need to do next. In most cases you only pay if we recover your data. Where controller level physical intervention is needed we require a 50% deposit before commencing the project and the balance isn’t due until successful recovery of your files has been completed. After receiving your quotation simply drop off your hard drive or SSD at John Eccles House, located within the Oxford Science Park or send it via insured postal service; please always send just the bare SSD. Avoid sending any further components. If the SSD is soldered to a motherboard within a laptop, then let us know the model number, so that we can immediately advise whether this is something we would be able to undertake.
This indicates that the controller has accessed the firmware stored in the drive’s reserved service area, and determined that it does not match the expected value; therefore, it will not read this firmware. The drive will appear to be either missing, blank or show a completely incorrect total capacity of the data. Your data has not failed; your read code has failed.
Yes. Most often. A fault with the controller or firmware causes the drive to stop presenting itself (to the OS), however, the V-NAND still contains the original data. Recovery involves accessing the flash memory through some other method than via the original route. First, vendor technical mode & repair of service-area, and if this fails, access the NAND directly.
No. Updating the firmware will rewrite exactly the same section of memory which was previously damaged. An update stalled upon an unstable drive can render it permanently unreadable. If you are interested in recovering your data, do not attempt to update the firmware prior to having it diagnosed.
Only with your login credentials. The 860 EVO offers Class 0 and TCG Opal encryption options and BitLocker possibly uses this. Without the decryption password (or recovery key) we cannot decrypt the information nor can anyone else. Check your Microsoft account (if applicable) or domain records (for domain accounts) for a recovery key as the first step.
For SSD / NVMe recovery our prices begin at £300 + VAT. We offer a free 48-hour diagnostic service and then provide a fixed written quote detailing what will be required to perform the recovery. In most cases you only pay if we recover your data. All controller level physical work requires a deposit of 50%, and the balance will only be invoiced when successful.