840 EVO demise

[Brado78]

Anyone check on Ebay lately?,, Looks like everybody is getting rid of their 840 Evo's , I guess the firmware fix wasn't a fix after all?, or are people just panicking? What do you guys think? :\

 

[VirtualLarry]

No, it wasn't a fix. There's yet ANOTHER fix coming out, this one, silently moves around old data on the drive, refreshing it, but using up write cycles.

 

[dave_the_nerd]

It's probably 2% legit problems and 98% panicky nerds.

 

[Magic Carpet]

No, it wasn't a fix. There's yet ANOTHER fix coming out, this one, silently moves around old data on the drive, refreshing it, but using up write cycles.

I remember installing one 6 months ago, thought it was done and dusted. Good thing, I stopped buying Samsung.

 

[Cerb]

No, it wasn't a fix. There's yet ANOTHER fix coming out, this one, silently moves around old data on the drive, refreshing it, but using up write cycles.

...which is what the release day firmware should have been doing in the first place. If it has enough problems reading data, it should mark it to be refreshed. I wouldn't worry myself over it if I had one, but Samsung is not exactly impressing me, here.

 

[coercitiv]

...which is what the release day firmware should have been doing in the first place. If it has enough problems reading data, it should mark it to be refreshed.

They probably thought they could improve parameters enough to keep the cells in good condition until they would get refreshed "naturally" due to wear leveling.

Nevertheless, they still should have went ahead and included cell refresh, for the simple reason it is a sure fix and it only generates P/E cycles when the drive has a low volume of daily written data to begin with.

 

[tweakboy]

There will be a resolution to this from the makers soon.

 

[Yongsta]

If I use Diskfresh on the 840 EVO and it becomes almost good as new again in performance... but that's a hassle. From now on I will avoid all EVO / TLC drives and stick to Pro / MLC drives. My 850 PRO is awesome.

 

[SSBrain]

They probably thought they could improve parameters enough to keep the cells in good condition until they would get refreshed "naturally" due to wear leveling.

Thing is, wear leveling doesn't seem to occur at all on static data under normal conditions. I've tested this on one of my 840. I had several tens GB of static data stored there for a couple months, then I disabled trim, and tried again using the drive normally for a few more weeks (as a sort of storage drive).

Static data eventually started degrading in speed, even though a few P/E cycles worth of writes in the available free space (mainly with large files, OS system images, etc) were performed. In theory wear leveling should have caused corresponding cells to get eventually refreshed.

So, my take is that for some reason Samsung made static wear leveling nonfunctional or limited it only to specific conditions.

 

[SeanFL]

Just checked on ebay, and most of the used ones are being offered at prices higher than a new 850evo on amazon.

 

[WhoBeDaPlaya]

No, it wasn't a fix. There's yet ANOTHER fix coming out, this one, silently moves around old data on the drive, refreshing it, but using up write cycles.

Hey, at least they're releasing something for the 840 EVO.
The 840 Vanilla users are shafted :|

It's probably 2% legit problems and 98% panicky nerds.

I don't know, seems to be a legitimate concern, like the silent data corruption in the Spinpoint F4 2TB :
http://hardforum.com/showthread.php?t=1566067

 

[BonzaiDuck]

I can regret that I bought three EVOs: a 120GB for my server boot-system disk, and two 256GB for the boot-disks of two household workstations. Two other systems use an 840 Pro and a Crucial MX100 -- both 500GB each.

I've kept an eye on the EVOs and haven't noticed any performance degradation yet. But it's a lesson learned about TLC, when you thought you were building systems with "T-L-C" . . . heh-heh.