I just experienced the end of a slow breakdown of an SSD.

This SSD has been installed in a media PC for 6 years. The first symptoms came early on with hangups that I thought were due to memory pressure. The OS would freeze for a time until I killed the browser so you can see how the problem looks memory related. On the 7th this month SMART started reporting the drive as bad. I noticed it a few days later on the 15th when the console reported "new mail". At first I thought it was a false alarm but then the hangups became progressively worse. Now the drive does not respond at all. No real data loss as this is a media PC.

I have included the smartctl report from the failed drive and a smartctl report from an identical drive that is still working.

Interestingly, smartctl reports no failed attributes found. But it also says the overall health assessment has failed.

I have also included a picture of the drive internals after I have opened it up. Nothing on the front or back of the board appears to be burnt. It is suprising how small the circuit board is and how little is inside these Sandisk drives.

One thing of note is that the retired block count, reserved block count and unknown sandforce attribute are higher on the failed drive than the working drive. Power on hours are much higher as well as the amount of data read and written to and from the drive. Clearly, the media usage of streaming videos took a heavy toll on the drive.

Notice how few power cycles are on both drives as I do not like equipment going from cold to hot and back again. I have had too many desktop PCs in the past not turn on due to a bad motherboard or power supply after running for a long time.

I use Samsung and Sandisk SSD drives. Sandisk is still an ok brand to me. On one hand the performance of the drive and the tiny circuit board inside compared to a Samsung drive is cause for concern. On the other hand I have reason to suspect that physical stress on the SSD when I was installing it played into the failure. The tablet-like laptop it was installed in was a very complicated fit and I was frustrated when installing it and likely damaged it.

I hope this is interesting to someone. I wonder why smartctl simultaneously shows "no failed attributes" but also shows the overall health as failed?

Attached Thumbnails

 

Attached Files

	DRIVE_FAILURE.txt (5.4 KB, 26 views)
	DRIVE_PASS.txt (4.7 KB, 17 views)

I have a bare disk just like that picture of your internals, it is called a DoM (Disk on Module), it came fitted in a thin client that I bought pre used.

I think they just put one inside a 2.5" case for convenience, perhaps all SSD are the same, I've never opened one up.

That is pretty interesting. It saves a lot of space like that. I suspect there is a little more inside the Samsung 2.5" SSD drives because they are noticeably heavier and to me perform a little bit better. I suppose this discussion is increasingly irrelevant now that the mass of people are moving toward m.2 drives. The replacement drive in the system is performing really well.

Many systems were not built with a SSD so there is also a lot of timing problems.

SSD's and usb's have an oddity that makes them slow down too.

Welcome to my world.

I fixed hardware for a living. Let me assure you that blown ICs or faulty ICs look exactly the same as good ones. For chips blown at the on the I/O legs (Most chips in industrial boards) you could use Analogue Signature Analysis to probe the legs, but with MSI/LSI chips the only viable option was functional testing.

It's all impossible now, because you can't get the parts.

The two parameters that might of triggered the fail are:
[code]
ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE
5 Retired_Block_Count 0x0032 100 100 000 Old_age Always - 341
232 Available_Reservd_Space 0x0033 100 100 004 Pre-fail Always - 14708775[/quote]

Available Reserved Space for the SanDisk from what I can tell starts at 100 and decreases.

Not relevant, but I have Dimension 4600 running a twenty year old drive, and that computer was on every day, and it still works, nearly twenty years on. I don't get it. Nothing has been touched except for RAM changed in 2008, upgraded graphics in 2010, and battery change couple of times in the past eleven years. Even the mouse, mouse mat, and screen are gone. All that is left is the keyboard with it. LOL

I only have it for Microsoft flight simulator 2004, Elite force, Men in Black the game.

Some hardware is built and lasts longer and others pack up after a few years or several.

I had an instructor in 1977 who told us of a test he started in 1961. They put NPN & PNP transistors on switching various loads regularly 24/7/365. 16 years later, it was substantially the same transistors doing the job, although a few had been replaced. At the time, this was a sharp contrast with thermionic valves, which would all would have weakened and nearly all have died. The weak link in your 20 year old box will be the hard disk, with it's moving, mechanical & magnetic parts.

I thought so. How ever that doesn't stop capacitors from going knackered, the optiplex GX270 small form factor, a 2001 dell system has a floppy disk, has a few bad capacitors, but it runs. I shouldn't of installed or attempted to install linux, as a result it became useless. I can't even run vice city, integrated graphics, I even tried the dell resource CD. But it isn't any good, but yes, erm, I guess the drive is more useful in that one for the 4600.

I can't play Mafia on the 4600, I probably should of done reinstalls of xp, due to the software being difficult to get on the web for certain software use. In the end it is pretty much old news.

But they all have their uses.