17 comments

[ 2.7 ms ] story [ 44.1 ms ] thread
It's weird that OCZ tries to somehow salvage itself from the reputation of a company that knowingly produces unreliable hardware and knowingly lies to customers.
(comment deleted)
I think they were eventually acquired by Toshiba.
I love how Kristian opens up with a flow chart of OCZ's design process. Every OCZ product fails in step 4 -- or seems to jump from 2 to 5.

I read the title of the article as "How SSDs are Made (Poorly)". This was a bad acquisition for Toshiba.

They forgot to include the "Testing unfinished products on client PCs" phase :-)
And the history of black/grey market meandering. I remember watching a discussion on their official support forum from users requesting help on broken drives purchased new in box off of ebay. OCZ didn't even bother to check serial numbers for legitimacy they just made a blanket rule against anything purchased on ebay. I'd have chalked this up to typical support laziness but after (respectfully) posting suggestions for various methods that can be used to distinguish clones from legit they deleted the whole thread. Just a horribly shoddy company.
I'm surprised that they're looking to replace some of the more mundane manual tasks with automation. Labor is a lot cheaper than machines in China.

I somehow doubt that putting stickers on a SSD is the bottleneck in their production facility.

I had three OCZ SSDs - a Vertex 2 and two Vertex 3s. One lasted a month and a half; the others lasted a month. It's surprising that Toshiba would want to be associated with this brand.
Yep, those Vertexes turned OCZ into a joke (for me, at least).

They fail as fast as they write :-).

Intel and Samsung drives from the same generation just keep going...

OCZ basically did the beta testing for SandForce, and paid the price with their eventual bankruptcy. But by then they had acquired Indilinx to provide an in-house alternative to SandForce controllers, and that's a significant chunk of what Toshiba was interested in buying. Winning in the SSD market seems to require making your own controller and/or flash, and there's only one good source for buying a good high-performance controller on the open market (Marvell).
Pretty empty article TBH. This is not only standard, but very simple pick and place assembly that any CM can do.

Most of what goes into 'Making an SSD' is the ASIC and firmware design on the controller. The actual assembly is incredibly mundane from there.

Will never, ever buy anything branded with OCZ. This company nearly poisoned the SSD market.
I completely missed this - what did they do?
They rushed several products to market, got tons of market share because they were setting performance records, and then they got basically all of the backlash when the bugs in the controllers and firmware started causing lots of drive failures. They also got in a bit of trouble once or twice for swapping out components with cheaper stuff that didn't perform as well. On the other hand, they were doing a great job of pushing the price/performance frontier so all the more reliable and more expensive drives that eventually made it to market still had to at least match OCZ's performance. That led to the very surprising situation where Intel basically withdrew from the consumer SSD market for a few years because they couldn't keep pace without sacrificing their QA standards.
Few companies have poisoned the well as thoroughly as OCZ has, by treating their customers and investors as utterly disposable. There are a lot of people out there who won't forget what they did. Toshiba would do well to retire the brand, dissolve the org and integrate the assets fully.
I think they will. They shipped "Hitachi" drives for awhile, but now are branding them Toshiba. I think the same thing will happen with OCZ-- they introduce a Toshiba branded drive, that's exactly the same as the OCZ and over time make these more available than the OCZ branded ones until the market has transitioned.