SANTA CLARA, Calif. - Intel has found a design error in a support chip for the recently announced Sandy Bridge processor. The company says it has implemented a solution but that the error could cost the company $700 million.
The design error is in the Intel 6 Series support chip, code-named Cougar Point, that has been shipping since Jan. 9, Intel said.
In some cases the result is that the Serial-ATA ports within the chipsets may degrade over time, potentially impacting the performance or functionality of SATA-linked devices such as hard disk drives and DVD-drives, Intel said. The chipset is utilized in PCs with Intel's latest Core processors, code-named Sandy Bridge.
Intel said Monday (Jan. 31) it has stopped shipping the faulty support chip from its factories and is implementing a fix in silicon. The manufacture of a corrected version of the support chip has begun and Intel said it expects to begin delivering the updated version of the chipset to customers in late February and expects "full volume recovery" in April.
The Sandy Bridge microprocessor itself is unaffected by this issue and no other products are affected by this issue, Intel said, without disclosing the nature of the problem. Intel said it would accept the return of the affected chipsets, and plans to support modifications or replacements needed on motherboards or elsewhere within systems.
As the faulty chip has only been shipping since Jan. 9, relatively few consumers have been impacted, Intel said. The only systems sold on to customers are second generation Core i5 and Core i7 quad core computers. Intel said it believes that consumers can continue to use their systems with confidence, while working with their computer manufacturer for a permanent solution.
The halt in production of the current version of the support chip and pause while making a corrected version and ramping up production is expected to reduce revenue by about $300 million in the first quarter of 2011, Intel said. However, full year revenue is not expected to be affected as the company catches up sales later in the year.
But the total cost of the repair and replacement of affected materials and systems in the market is estimated to be $700 million. Intel said it will take a charge against cost of goods sold in both the fourth quarter of 2010 and in the first quarter of 2011.
Also Monday, Intel announced that it has completed the acquisition of the Infineon Technologies AG Wireless Solutions business. Intel said last August that it would buy the unit for $1.4 billion in cash.
Craig Berger, an analyst with FBR Capital markets, said the acquisition would add roughly $500 million in revenue to Intel's first quarter financials.
"Net, these issues are relatively immaterial for Intel and 2011 [earnings per share] estimates should not change too materially following the chipset recall issue and the inclusion of Infineon Wireless business," Berger said.
Thanks for the information.
However, it does seems strange that a circuit that "deteriorates" over time can be fixed by deleting a piece of late metal, as Intel executive Steve Smith said (see http://www.eetimes.com/electronics-news/4212726/Intel-has-late-metal-fix-for-design-error). I would like to know the nature of the deteroriation. Is it charge build up, electromigration? And how does the fix relate to the problem?
Can you provide the internet URL?
Details are on the net: it's a semiconductor process error that allows devices near the I/O pad for the SATA-III controller to deteriorate and stop performing over time of several months. Kind-of like cancer and smoking it's a statistical process, affecting 5-10% of the chips.
Intel pulled the chips, and apparently asked distributors to stop selling product; they'll probably reimburse the channel, and create a trade-in program.
I would have hate to have been the guy/gal that found this bug! It is a shame when a chip is released with problems and as hard as most designers work to prevent this sort of thing there is always the need to balance time to market and testing. This time they got caught, I hate when that happens :})
It's intriguing that it supposedly affects both hard drives and DVD drives. I can think of a number of ways that hard drives might degrade over time, but offhand I can't think of much of anything that could cause DVD drives to do so. Either that is a red herring or something interesting is going on. I'll be curious to get more details on this when it becomes available.
It will be great to know more details about timescales over which this could happen.
On more than one occasion I have seen such a post-silicon problem root-caused to a functional logic bug. For example, if under some corner-case scenario, if 1 flow control credit (of 1,024 or 64k credits) gets lost, performance can slowly de-grade over time. Specialized formal methods help in debugging.
I think Intel just about managed to get away with a major controversy. Had it been detected in the customer end, it could have been a disaster. I am curious, how early was Intel expectiing the interface to start mis behaving?
Join our online Radio Show on Friday 11th July starting at 2:00pm Eastern, when EETimes editor of all things fun and interesting, Max Maxfield, and embedded systems expert, Jack Ganssle, will debate as to just what is, and is not, and embedded system.