OCZ using crappy flash chips?

D

danman

[H]ard|Gawd
Joined
Aug 17, 2005
Messages
1,668
http://blog.macsales.com/9438-not-all-ssd’s-are-created-equal-the-story-continues#more-9438

As you can see from the screen shot above, this OCZ SSD ordered Wednesday from a major retailer and received yesterday is still advertised as 120GB, yet is actually a 115GB version.

Then, when you watch the unboxing video below, you see the memory chips used are considered “OEM grade”—not the Tier 1 quality grade used in the OWC Mercury SSD family.

When we took the cover off of this third, direct from OCZ SSD, we found a ‘S’ stamped over Micron logo on all the flash devices (see the image to the left). This indicates the device is “off spec” product because it failed some parameter of Micron’s full performance and/or quality specification testing. “Off spec” memory is typically used in low-level applications such as toys, offering considerable cost savings over Tier 1 level to an SSD manufacturer.

“It is a very brave action to take, using these chips in a data storage device,” was the reply given.
 
I have an OCZ Vertex turbo 64Gb that was getting more fragmented since I didn't over-provision it when I installed Win7 Pro. So I'm going to do a reinstall with O-Ping, to help Win7 trim work better & I was thinking of getting a 100-120Gb OCZ ssd instead of using the 64Gb again but that's not happening now. Shot themselves in the foot I think, cause they just lost that sale.
 
I'm not surprised. To me, it has seemed that OCZ has a different approach to SSD manufacturing than I would prefer. This goes back to the indilinx drives and became especially apparent with the usage of pre-production Sandforce chips/firmware.
Intel and others have had their fair share of troubles (I've personally seen two X25M drives go, out of the three I have) but it is seems to me that OCZ is not doing things with 100% due care.
This new information only solidifies my opinion, even though it's coming from a competitor of OCZ. It would have been nice with a picture of the chips in question, the youtube video doesn't have enough resolution, even at 1080p, to read what they say. (Although the explanation in the comments state that OEM chips are lacking the manufacturer stamp, which they clearly do.)
 
Thank you for posting this information. OCZ and quality have never been synonyms in my mind, but I'm building a new system for a friend who was leaning towards this specific SSD after his own research. I suspect this will change his mind.
 
Yonzie said:
I It would have been nice with a picture of the chips in question,
Click to expand...

Huh? They did show a picture:

http://blog.macsales.com/9438-not-all-ssd’s-are-created-equal-the-story-continues#more-9438

You can clearly see the stylized "S" stamp on the flash chips:

OCZ-spectech.png
 
flegg said:
There goes my plans to buy Vertex 3?
Click to expand...

Not to worry. If you want an SSD with a Sandforce 2XXX series controller, there will be a number of other companies to buy from. G.Skill and Corsair, to name two. They will probably be released in the next few months.
 
john4200 said:
Okay, worry if you are so inclined.
Click to expand...

ASRRRRRRrRrrRRRrRRRRRGGGGgGGgGggGHHhhHHHhhhHHHHHh!!@!!!!!1


TYHte sky is faaaaling!!!!



Seriously, though, I've already decided to wait it out, until all major players are in the field, since my measly C300 is already giggles faster than my 250GB WD Scorpio Blue I was using as my desktop's primary HDD prior to buying the C300 :)
 
What's worse is OCZ completely locked out 2 threads involving users posting concerns on this article.
I think it's a valid concern.

Once again, WTF are the reviews on the C400's? I'd like to see the random 4K QD1 R/W performance...
 
I take that back - they just deleted any and all threads discussing the OWC article.

LOL. WOW.
 
Hopefully the OWC folks make something based on the Sandforce-2xxx. I like how they are so upfront about their products.
 
Sigh. Is there any way without cracking it open to know what chips you have inside?

It makes me wonder if this is why they are changing out HDD's without much question. OCZ has made a big point that its just a 32gb vs 64gb die issue but this makes me think it runs MUCH deeper.
:(
 
Thanks for bringing this to our attention. This remains to be independently verified though (I'm not surprised if this is indeed true).

Perhaps OCZ should issue a response statement.
 
Hasn't SF always advertised the ability to work with low-quality flash chips as one of the main benefits of their controller?
 
thefreeaccount said:
Hasn't SF always advertised the ability to work with low-quality flash chips as one of the main benefits of their controller?
Click to expand...

That not the reason why people were complaining, I think. OCZ did a switch-a-roo without making it clear in any way, except for degraded performance and slightly less storage space (3-4GB, which for a SSD running at 1.5-2usd per GB, is quite a bit).
 
Tbh, it's only a concern if they wouldn't honour their warranties, which I would presume they do. You don't know if their firmware contains adjustments to work with the chips in question.

It's like complaining that Intel sells bugged CPUs (all of them, especially the first ~1 year batches of any new cpu line) then masks those bugs by microcode updates shipped with M/B bioses.

If you are only complaining because the size does not match up (difficult to know, no one has posted any proof either way in this thread) then okay, complain about the lost gigabytes.
 
DJ_Datte said:
Tbh, it's only a concern if they wouldn't honour their warranties, which I would presume they do. You don't know if their firmware contains adjustments to work with the chips in question.

It's like complaining that Intel sells bugged CPUs (all of them, especially the first ~1 year batches of any new cpu line) then masks those bugs by microcode updates shipped with M/B bioses.

If you are only complaining because the size does not match up (difficult to know, no one has posted any proof either way in this thread) then okay, complain about the lost gigabytes.
Click to expand...

you haven't even read the first post in this thread. If it's just "lost gigabytes" to you, then too bad for YOU. I however, won't be buying again from a company that does false advertisement of their product (so far), change to slower parts without notifying anyone, has less capacity than advertised, did all this without using a different model number or any way to ID the "new" and "old" stuff without vioding warranty, and then try to suppress the threads and news posts that came up.

you've been around at least six months. All this bullshit happened within the last two.
 
Last edited:
http://www.ocztechnologyforum.com/f...nm-Spectek-Nand...see-here-for-specs-and-info

It isn't off spec nand......and OCZ didn't delete any threads, they consolidated new posts into one thread instead.


As far as the 64gb 25nm nand is concerned. I got two of those drives, they were exchanged completely free of charge for 32gb drives. I didn't even have to pay shipping. Sure it is a pain to wait a week or two for an RMA, but at least it was fixed.....and again, threads weren't suppressed, just consolidated. The original thread is still there, although it is a few pages back now.

I'm not saying that selling those drives was right in the first place, but at least it was fixed quickly and with no real cost to the consumer.
 
Last edited:
There are 2 separate issues issues at play here.

1) The 34nm to 25nm switch which OCZ also "combined" with the change from a 32gb die to a 64gb die. The 25nm/64gb die drives were the ones that totally blew. The newer 25nm/32gb die drives have better performance than the 25/64 but still do NOT perform at the levels the 34nm/32gb drives did (these were the ones that made the Vertex 2 so popular).

2) This 'low quality' NAND issue. What I'm about to say is all conjecture but I think this may be the cause of the 'time warp' and complete data loss issues that have popped up lately.

That said, I have seen and experienced that OCZ takes care of their customers. They are highly active on their boards and issue the RMAs with little or no question. I do think OWC may have made poor choice in their approach to this. To a point I do feel 'sorry' for OCZ because they are getting a little beat up as of late but it was their own fault for changing their drives to be lower performing without even changing a SKU or Model #. It will be interesting to see how this all plays out.

Edit, after reading the response thread to this issue on OCZ's forum I think OWC may find themselves in some seriously hot water.
 
Last edited:
ramrod126 said:
http://www.ocztechnologyforum.com/f...nm-Spectek-Nand...see-here-for-specs-and-info

It isn't off spec nand......and OCZ didn't delete any threads, they consolidated new posts into one thread instead.


As far as the 64gb 25nm nand is concerned. I got two of those drives, they were exchanged completely free of charge for 32gb drives. I didn't even have to pay shipping. Sure it is a pain to wait a week or two for an RMA, but at least it was fixed.....and again, threads weren't suppressed, just consolidated. The original thread is still there, although it is a few pages back now.

I'm not saying that selling those drives was right in the first place, but at least it was fixed quickly and with no real cost to the consumer.
Click to expand...

Lastly, do NOT post separate threads regarding this, mods will be instructed to trash all threads and PM those thread creators with the link to this thread...if you want to talk about this do it here
Click to expand...

he posted that thread TODAY, if you didn't care to notice. They've been "trashing" all other threads on their forums.
 
jeremyshaw said:
he posted that thread TODAY, if you didn't care to notice. They've been "trashing" all other threads on their forums.
Click to expand...

I am at the OCZ forums daily, sometimes several times a day. I see what goes on on a first hand basis. Do a little looking and you will see that the threads are still there, usually with a response as to why the thread is closed.

If YOU didn't care to notice:

1. http://www.ocztechnologyforum.com/f...6281-OCZ-Uses-poor-quality-NAND-in-their-SSDs

2. http://www.ocztechnologyforum.com/f...lity-grade-NAND-in-Vertex-2-(240GB-and-120GB)

3. http://www.ocztechnologyforum.com/f...OWC-Takes-a-shot-at-OCZ-Chip-Quality-Concerns

...and so what if they delete new threads and direct people to the thread that already exists for this discussion. I didn't realize it was wrong to want to keep the forum from being flooded with 100 threads about the same topic. It isn't censorship, it is housekeeping. If they were censoring people there wouldn't be some of the posts that are in the main thread.
 
They locked the thread again:

No more questions will be answered now.. until W Pacific time zone open of play tomorrow.

Thread locked.
Click to expand...

OCZ really tries to be the master of spin-doctoring, pretending to be honest when they are really deleting posts and locking threads. Who does OCZ think they are fooling? You would have to be awfully naive to believe OCZ about anything after their past behavior.
 
Is anyone familiar enough with the industry to explain how spectek reclaims the NAND, especially with the "AL higher than original tolerances"?

I would assume that when flash fails tests and is sent to spectek, spectek tests it in a more in depth manner and bins it appropriately. But how can flash modules which failed a test initially now pass at better than original requirements? Are they re-specced at a different voltage, etc?
 
ChrisBenn said:
Is anyone familiar enough with the industry to explain how spectek reclaims the NAND, especially with the "AL higher than original tolerances"?

I would assume that when flash fails tests and is sent to spectek, spectek tests it in a more in depth manner and bins it appropriately. But how can flash modules which failed a test initially now pass at better than original requirements? Are they re-specced at a different voltage, etc?
Click to expand...

I think that not just single chips fail originally - it's a large group of chips or part of a wafer. SpecTek then tests them individually, finds the failures, but also finds the passers? The passers are then rated AL or whatever.... I dono...
 
Brahmzy said:
I think that not just single chips fail originally - it's a large group of chips or part of a wafer. SpecTek then tests them individually, finds the failures, but also finds the passers? The passers are then rated AL or whatever.... I dono...
Click to expand...

That is what OCZ claims.

I'm skeptical, since if you look at SpecTek's website, the pictures of their chip packages show just the "S" logo and a part number on the chips (no "M" logo).

But the chips in question in the picture from OWC shows the "S" logo stamped over the top of some part numbers and the "M" logo. Obviously I cannot be sure, but it looks to me like Micron diced the chips from the wafer, packaged them up, printed the part numbers on the package, and then maybe the packaged parts failed final test in some way? Who knows. Then maybe they went to SpecTek, who stamped their "S" over the top of Micron's printing on the package, and it is not clear to me what criteria SpecTek used.

Apparently, the SpecTek grade given to these chips is the highest speed grade that SpecTek uses, but I don't see how that tells us anything about the overall quality and reliability of the chips, in relation to the chips that Micron sells without the SpecTek logo.

By the way, here is what (I think) is printed on the packages in question. A few of the characters are slightly obscured by the "S", but this should be close. If anyone can decode this from either Micron or SpecTek's spec sheets, that would be good to know:

9Z53 (at edge of package)
"S" and "M" logos
ODB18
NW181
PF458 AL
 
Last edited:
john4200 said:
That is what OCZ claims.

I'm skeptical, since if you look at SpecTek's website, the pictures of their chip packages show just the "S" logo and a part number on the chips (no "M" logo).

But the chips in question in the picture from OWC shows the "S" logo stamped over the top of some part numbers and the "M" logo. Obviously I cannot be sure, but it looks to me like Micron diced the chips from the wafer, packaged them up, printed the part numbers on the package, and then maybe the packaged parts failed final test in some way? Who knows. Then maybe they went to SpecTek, who stamped their "S" over the top of Micron's printing on the package, and it is not clear to me what criteria SpecTek used.

Apparently, the SpecTek grade given to these chips is the highest speed grade that SpecTek uses, but I don't see how that tells us anything about the overall quality and reliability of the chips, in relation to the chips that Micron sells without the SpecTek logo.

By the way, here is what (I think) is printed on the packages in question. A few of the characters are slightly obscured by the "S", but this should be close. If anyone can decode this from either Micron or SpecTek's spec sheets, that would be good to know:

9Z53 (at edge of package)
"S" and "M" logos
ODB18
NW181
PF458 AL
Click to expand...

Earlier today I almost made a post decoding it all but it doesn't really matter - if you decode the PF458 you'll get the part number which still doesn't matter because those last two letters (AL) mean its the highest tier.

OCZ has a 3 year warranty, if the chips are junk and your drive fails, they'll replace it. I'm willing to bet however that most people won't have a problem with their drives because of where these chips are sourced.
 
aphex187 said:
OCZ has a 3 year warranty, if the chips are junk and your drive fails, they'll replace it. I'm willing to bet however that most people won't have a problem with their drives because of where these chips are sourced.
Click to expand...

Yeah, but having a drive fail is a HUGE pita and potential data loss. Saying "it's okay if stuff might be crappy because they will just RMA it" doesn't really do it for me since I don't want to have to deal with swapping out bootdrives, system downtime, data loss, etc. I would rather buy a SSD that is least likely to crap out randomly.

I still can't say I know enough about the process to make a call one way or another, but I haven't seen any information that says this ram isn't lower spec than what comes off the standard micron line - just that it's the highest grade spectek offers.

I just don't see any reason to buy an OCZ product - what do they offer above and beyond the other SF controller manufacturers? If you want one of the first to market vertex3 drives is about the only reason I can see? Otherwise, why wouldn't I just buy from one of the other commodity sandforce vendors who don't have the problems surrounding them (if I want a sandforce drive?)
 
aphex187 said:
Earlier today I almost made a post decoding it all but it doesn't really matter - if you decode the PF458 you'll get the part number which still doesn't matter because those last two letters (AL) mean its the highest tier.

OCZ has a 3 year warranty, if the chips are junk and your drive fails, they'll replace it. I'm willing to bet however that most people won't have a problem with their drives because of where these chips are sourced.
Click to expand...

If you can decode all of that, please do.

AL is the highest speed tier, from what I can tell from SpecTek's documentation. But it does not tell me anything about the quality or reliability of the chips in relation to the chips that Micron sells without the SpecTek logo.

As for warranty replacement, you can't be serious. It is a huge hassle to replace an SSD. Even if my data is backed up, first I have to deal with the failure, then I have to waste my time (and fray my nerves) talking to OCZ to get an RMA, then remove the SSD and send it back, then wait and finally, hopefully receive the replacement, then install it and restore my data. That probably takes hours of my time. And that assumes OCZ is even still in business 3 years from now. At the rate that OCZ has been screwing up, conning their customers, getting big loans, and acquiring worn-out companies, I have my doubts if OCZ will still be around in 3 years.
 
Last edited:
ChrisBenn said:
I just don't see any reason to buy an OCZ product - what do they offer above and beyond the other SF controller manufacturers? If you want one of the first to market vertex3 drives is about the only reason I can see? Otherwise, why wouldn't I just buy from one of the other commodity sandforce vendors who don't have the problems surrounding them (if I want a sandforce drive?)
Click to expand...

+1 Completely agree.
 
ChrisBenn said:
Yeah, but having a drive fail is a HUGE pita and potential data loss. Saying "it's okay if stuff might be crappy because they will just RMA it" doesn't really do it for me since I don't want to have to deal with swapping out bootdrives, system downtime, data loss, etc. I would rather buy a SSD that is least likely to crap out randomly.
Click to expand...

I know it goes without saying, but I'd at least make sure that you have all data that you want to keep backed up in at least 2 other places. If the drives become faulty, i'm not even sure theres a recovery process for SSDs (is there?). A single drive failing should never result in data loss, especiallly important data.

Damnit I replaced a 3 times failing SSD with a vertex 2 (because it was between pay)...now I worry. It's only the OS, but still.
 
Anand at Anandtech is checking this out.

So in a nutshell (I posted this elsewhere):
How can a NAND chip be graded as full spec when it was initially discarded (by Micron)? An answer is that they were inadvertently discarded. Is this the case? If this is not the case, then they were purposely discarded because they did not meet Micron's specification, but meet Spectek's specification. Which may be a cause for concern.

This may be relevant:

http://www.behardware.com/articles/810-6/components-returns-rates.html
Return rates of various SSDs:
"For the first time, we are including SSDs in our returns rates. Here are the figures by manufacturer:
- Intel 0.59%
- Corsair 2.17%
- Crucial 2.25%
- Kingston 2.39%
- OCZ 2.93%

Just...an addition. :)
 
You know I had never really heard of this OWC company (since apparently they are geared towards Mac users), but after a bit of reading I may look at them next time I am looking for an SSD. Designed and built in the USA, and they appear to be very transparent about their process, neat!
 
You must log in or register to reply here.
Back
Top