TR Forums

As some of you may know, I came into possession of a new computer recently. Within 12 hours of getting it unpacked, it had two periods of minor instability, then a third that it did not recover from. A bluescreen lead to a windows system recovery attempt that failed. Subsequent boots didn't even get that far, failing to find any bootable hard drives in the system. I did some digging, and the BIOS could, in fact, see all of the drives, including the boot disk. Some digging on motherboard codes led me to it stopping on "Boot event" (thanks).

The only clue I had was that the BIOS flash utility would display the file systems for the other three drives, but not the boot drive. Fast forward another few hours, and I have a Windows 10 bootable stick, ready to test my theory about the harddrive. The motherboard finally gets around to telling me it is detecting a SMART failure on the drive, but it doesn't know what (the built-in SMART checks all pass with flying colors). I get the OS installed on the other SSD, and run CrystalDiskInfo on the disks, just to check. The three active drives are all fine, it can't find the fourth. I rummage around in disk manager thinking I just need to reinitialize the SSD, but it doesn't see it.

I'm out of time so I leave it for the night. In the morning, the drive has finally appeared and "Windows detected a hard disk problem." Judging by the power time listed in CrystalDiskInfo, the drive powered up 6 hours after my last reboot. The drive is pristine, except for a single code: 0xA4, which is vendor specific. Digging in http://media.kingston.com/support/downloads/MKP_521_Phison_SMART_attribute.pdf leads to:

170 AAh “Bad Block Count
(Early / Later)”
Counts the number of Bad blocks.
Raw Value Byte [1~0]: Early bad block count
Raw Value Byte [5~4]: Later bad block count
Formula
MABN: maximum acceptable bad block number
CBBN : Current bad block number
Spare unit percentage = ((MABN - CBBN)/( MABN)) *100
This formula calculates percentage of spare blocks. Value will range from 100
to 1.

CrystalDiskInfo is reporting this as a 0, which is literally out of range according to the documentation. Every other SMART attribute (except for temperature) is reporting a 100.



So what killed this drive? This isn't a HDD, and it obviously has been working fine for the previous owner, so what killed it via bad blocks?

Your expert knowledge is greatly appreciated!

Taking a guess here, but maybe the previous owner tried to update the firmware and didn't do it correctly and killed the drive. personally I don't bother updating the firmware/BIOS of hardware unless there is a need to do so. It is too easy to brick stuff.

The SSD has some reallocated sectors and read errors on it also. Anything that doesn't only have 0's in the raw values has something <1 in the attribute, but the formula for the attribute deems it low enough that it won't change it's SMART value from whatever it is by default.I'm not sure what F5 + F6 are but those attributes also have non-zero raw values

Edit: F5 + F6 are Max Erase Count and Total Erase Count respectively, so ignore that last sentence. Those two should be non-zero

whm1974 wrote:

Taking a guess here, but maybe the previous owner tried to update the firmware and didn't do it correctly and killed the drive. personally I don't bother updating the firmware/BIOS of hardware unless there is a need to do so. It is too easy to brick stuff.

Possible, but I don't think so. The other drive (the exact same mode) that is drive c:\ in that screen shot has the same firmware and is okay, though it does have 260 in the raw values for AA (not the original C:).

Sorry, left that detail out. There is another drive of the same model and size right next to it in the box.



These things happen, but the particular failure mode has left me scratching my head. (and did I cause it somehow)

APenguinNamedDire wrote:

whm1974 wrote:

Taking a guess here, but maybe the previous owner tried to update the firmware and didn't do it correctly and killed the drive. personally I don't bother updating the firmware/BIOS of hardware unless there is a need to do so. It is too easy to brick stuff.

Possible, but I don't think so. The other drive (the exact same mode) that is drive c:\ in that screen shot has the same firmware and is okay, though it does have 260 in the raw values for AA (not the original C:).

Sorry, left that detail out. There is another drive of the same model and size right next to it in the box.



These things happen, but the particular failure mode has left me scratching my head.

How do you know it has 260 in the raw values? I've never been able to change them into an actual numeral.

I assume they are in hex.

Easiest way is to go into your calculator (assuming windows), flip it into programmer mode, switch to hex, type your value, then switch it to decimal.
or use something like this:
http://www.binaryhexconverter.com/hex-t ... -converter

This is what I did to convert the AA to 170 when I read the SMART code from the documentation.

However, in this case, I should have written it 0x260, which is 608 in base10, I was just lazy.


Oh, and see the snippet of the document I posted to see what the raw value means!

Sounds like a crib death. It does happen.

Something is seriously messed up with that drive. The reallocated sector count is not used according to a kingston pdf, but displaying a non-zero raw value. It seems like it is a firmware issue.

Try flashing the latest firmware onto it, and if that doesn't work RMA it.

I'm not sure how you'd RMA promotional winnings but I'd imagine the drive is still covered under Kingston's warranty. As ludi says, it's possibly a crib death; I install hundred(s) of SSDs a year and when you deal with those sorts of quantities the 2% RMA rate becomes a tangible, real occurence.

Infant mortality. Not particularly common with SSDs, but it can happen.

So Damage didn't put the Skylake Damage Box through the testing wringer enough?? Where's that QA/QC?? j/k Damage

lol

SSD's should be under warranty.........I hope.

Well, when you buy them they have a 3 year warranty, but I don't know about promotional parts.

I'm trying to slip this in under the radars of most, I don't want to make any sort of big stink. These things happen, and I'm still incredibly happy with the machine. If anything had to go and there is no replacement, a hard drive was the best. I have a 128GB Samsung 830 (MLC) and an 850 (TLC) sitting about a foot away that I don't mind moving over.


I was curious about the failure mode, because I wanted to know:

a) did I cause it somehow
b) was it conceivably shipping (not because I thought that was at all likely, but it was insured)
c) is it a death that would be covered under a warranty if one exists for this (again, I don't know the specifics of how this part was handled)

I wasn't sure if infant drive death was the case here, but I do think it could explain the instability I was getting prior to the death. It has been rock solid since I moved the boot disk to the other SSD. (Of course, all the tweaking and overclocking got lost in the purge, as I did reset the BIOS at one point in my attempts to get it to see the drive).


Thanks for confirming, guys!

Didn't that system come with a pair of SSDs? It should definitely be covered under warranty. I would just pull and RMA the afflicted drive, and use the other as your main system disk.

Have you tried contacting Damage? He may have a spare. Or he may just tell you to deal with the RMA. But it's probably worth a shot.