Request RMA for worn OCZ Vector?

[Deders]

My Brother has an OCZ Vector 128GB that has been causing no end of trouble. He was getting regular crashes, corruption and diskcheck was starting nearly every time he loaded windows. Based on past experience he thought it was just something windows does but when we tried to install windows fresh it would blue screen.

OCZ themselves released an urgent firmware update 3.0 "Important: This is a mandatory update for system stability. OCZ urges all customers to update to 3.0 as soon as possible to ensure the continued reliability of their drives."

My brother, not being tech savy enough to know to check for these things had been using 1.3 for about 2 and a half years, half of the warranty.

So after the update we found that the drive was at 92% health, which by MLC standards means that over 2 million cells have gone completely if my maths is right.

He only uses it as a boot drive, and has only written 1759GB worth of writes to it, does this sound about right for an 8% drop in health?

I'm used to my Samsung 830 with its toggle nand, but I've written over 18TB, ten times as much as he has and I'm still at 100%.

He sent over this image of his smart details, and the response was if the smart data has any sudden drops then they will replace the drive.

TBH I don't think there is a lot you can tell from that picture, the details I would think you'd need are in the raw column which is obscured.

Here are non obscured images...

I'm not convinced that a lot of damage hasn't already been done, if it has lost 8% of its cells and Smart hasn't registered it then there could be more damage done to other cells.

What are people's thoughts on this?

 

[dave_the_nerd]

SSD health usually refers to the number of write cycles remaining, not the number of failed cells. So I don't think it means what you think it means. Your values for reallocated sectors and bad blocks would seem to indicate normal happy operation in that regard.

With that few writes, it should still be closer to 98% health remaining. But OCZ's utility may be doing other sorts of math. What does SSDLife report?

I'd have RMA'd or replaced it well before now, though, if it had been behaving as you describe.

 

[WilliamM2]

BTW, the 830 uses MLC nand, not the cheaper TLC nand.

 

[Deders]

With that few writes, it should still be closer to 98% health remaining. But OCZ's utility may be doing other sorts of math. What does SSDLife report?

SSD life says the same, 92%

BTW, the 830 uses MLC nand, not the cheaper TLC nand.

The vector also uses MLC, or at least it is advertised to.

 

[Deders]

So are people saying it is common to see the health figure drop from 100% in the first six months to a year?

 

[VirtualLarry]

So are people saying it is common to see the health figure drop from 100% in the first six months to a year?

Depends on usage, but, yes. I had an OCZ Agility 30GB drop to 72% health after like 3-4 months, doing BOINC on it. (BOINC, by default, checkpoints every WU every 60 seconds, so it was writing a bit excessively to the drive, and it was a small drive.)

SSD health usually refers to the number of write cycles remaining, not the number of failed cells. So I don't think it means what you think it means. Your values for reallocated sectors and bad blocks would seem to indicate normal happy operation in that regard.

This is true.

My Brother has an OCZ Vector 128GB that has been causing no end of trouble. He was getting regular crashes, corruption and diskcheck was starting nearly every time he loaded windows. Based on past experience he thought it was just something windows does but when we tried to install windows fresh it would blue screen.

OCZ themselves released an urgent firmware update 3.0 "Important: This is a mandatory update for system stability. OCZ urges all customers to update to 3.0 as soon as possible to ensure the continued reliability of their drives."

My brother, not being tech savy enough to know to check for these things had been using 1.3 for about 2 and a half years, half of the warranty.

What are people's thoughts on this?

Well, an SSD should not be causing crashes and blue-screens, if it were operating properly. Although, when my Mushkin 240GB was failing, it would cause sudden long pauses, and applications like web browsers would turn greyed-out for a while, in Win7. Not so many blue-screens, although I might have gotten one or two.

Has he done an A/B test, by installing a fresh HDD instead, and installing Windows on it (same key), and testing to see if it crashes? It could be the hardware. If it stops crashing, then I would slave the SSD to system (non-boot drive), and boot a Linux liveDVD and do a "secure erase" on the SSD, then, while it was still a secondary drive, boot to the HDD with Windows, and run the firmware update on it. Then I would disconnect the HDD, connect the SSD as the boot drive, and install Windows again on it, and see if it behaves. If it does, I would consider the problem fixed. If it was still crashing, and only with the SSD, I would RMA, given the evidence.

 

[Deders]

Depends on usage, but, yes. I had an OCZ Agility 30GB drop to 72% health after like 3-4 months, doing BOINC on it. (BOINC, by default, checkpoints every WU every 60 seconds, so it was writing a bit excessively to the drive, and it was a small drive.)



This is true.

I'm going to count boinc usage as extreme, I saw the effect it had on my SSD, writing gigabytes per day. Plus a 30GB drive has a quarter of the amount of cells compared to a 120GB model. As we've seen my brother's usage is at the opposite end of the scale.

Is it the case then that Samsung measures health differently? I'd assumed it was down to how many live cells were left because after 3 years of usage, and nearly 15GB written, it has yet to drop below 100%.

Can anyone shed any light onto how Samsung non-pro's behave?

 

[VirtualLarry]

Is it the case then that Samsung measures health differently?

Can anyone shed any light onto how Samsung non-pro's behave?

Different mfgs use different internal statistics. Or at least, they can. I'm not aware of any "standard" for determining drive health precisely.

 

[Deders]

Has he done an A/B test, by installing a fresh HDD instead, and installing Windows on it (same key), and testing to see if it crashes? It could be the hardware. If it stops crashing, then I would slave the SSD to system (non-boot drive), and boot a Linux liveDVD and do a "secure erase" on the SSD, then, while it was still a secondary drive, boot to the HDD with Windows, and run the firmware update on it. Then I would disconnect the HDD, connect the SSD as the boot drive, and install Windows again on it, and see if it behaves. If it does, I would consider the problem fixed. If it was still crashing, and only with the SSD, I would RMA, given the evidence.

We've most of this, except we updated the firmware first then secure erased it. His system works fine on the HDD and I've filled the drive 1.5 times and copied the data back off (with a restart in between to eliminate data cached in memory) without any issues.

It does seem looking online that many people have had issues with the drives, and OCZ tends to send out refurbs with only a years warranty for replacement.

 

[Elixer]

It does seem looking online that many people have had issues with the drives, and OCZ tends to send out refurbs with only a years warranty for replacement.

Hmm, they used to send out new units, haven't had to RMA their devices since the buy out by Toshiba, so, dunno if that policy has changed.

I know Sammy sends refurbs back.