Intel identifies major flaw in Sandy Bridge chipset; begins recall

Intel has announced that it has identified a flaw in the design of Serial ATA ports on its newest 6 Series "Cougar Point" chipset, according to a press release. In some cases, this flaw causes ports to degrade over time, causing users to suffer from poor performance on any Serial ATA devices such as hard drives, DVD drives or BluRay drives. Only the newest second generation Core i5 and i7 series chips use the the 6 series chipsets.

Unfortunately, this flaw cannot be corrected in software. Intel has halted all current shipments of the flawed chipsets and is implementing what it is calling a "silicon fix". Intel is currently in the midst of manufacturing new chipsets containing the fix. It expects to start shipping the updated chipsets in late February. Recalls of already shipped models are largely expected, and as a result the company is expecting to take a $300 million hit to quarterly revenues.

Intel says they've only been shipping second-generation Core i-series chips and the related chipsets since January 9th, and doesn't think many people will be affected by this issue. For computer manufacturers and other Intel customers that have bought already affected chipsets or systems, Intel will work with partners to accept the return of the chipsets, and has plans to support the modifications and/or replacements required to correct the aforementioned issue.

Users of computers containing the 6 Series chipset can continue to use their computer with confidence whilst Intel works on a more permanent fix.

Report a problem with article
Previous Story

Windows Mobile outsells Windows Phone 7 in Q4 2010

Next Story

Wikipedia is a man's game

62 Comments

Commenting is disabled on this article.

It really shouldn't have happened in the first place. But it is nice to see that they are stepping up and fixing it before it went to much further. This whole issues makes u wonder just why did Intel buyback 10billion in stock the week before? makes since now.

Glad I waited to upgrade. Figure there would be some minor bugs but this sounds pretty serious.

Hope Bulldozer doesn't have any major bugs.

The question now is, will the new motherboards have a different revision number or a model number for us to easily identify the new revision?

Why it's usually always better to wait for 2nd gen products of any new product. Gets rid of most of the flaws and kinks along the way.

Morisato said,
Why it's usually always better to wait for 2nd gen products of any new product. Gets rid of most of the flaws and kinks along the way.

Funny thing isn't Sandybridge the second gen of the Core i series? Glad intel dealt with this quickly, it could have been the nvidia fiasco all over again.

Since i doubt degradation events like thermal runaway can make it to the consumer-level, my bet on the degradation comes from electromigration.
I guess the circuit design did not have any compensation for that phenomenon.

Was going to be ordering a new system this coming weekend. Glad this news came out this week instead of next week. I'd be ****ed if my new system had to be returned for a recall right after I bought it! Isn't this the kind of thing that Intel should be testing BEFORE releasing a new chipset?

roadwarrior said,
Was going to be ordering a new system this coming weekend. Glad this news came out this week instead of next week. I'd be ****ed if my new system had to be returned for a recall right after I bought it! Isn't this the kind of thing that Intel should be testing BEFORE releasing a new chipset?

According to the article on Anandtech:
Anandtech said,

Intel has been testing its 6-series chipset for months now. The chipset passed all of its internal qualification tests as well as all of the OEM qualification tests. These are the same tests that all Intel chipsets must go through, testing things like functionality, reliability and behavior at various conditions (high temps, load temps, high voltage, low voltage, etc...). The chipset made it through all of these tests just fine.

Intel ran their standard testing/validation on these chipsets and they passed. The randonmess of this issue meant it wouldn't necessarily have shown up during that and in this case, it didn't manifest itself during that period. Perhaps Intel needs to beef up their testing/vaildation procedures?

Same here! The new revised motherboards better have obvious new product codes to differentiate between the dud and the new one.

McoreD said,
Same here! The new revised motherboards better have obvious new product codes to differentiate between the dud and the new one.

What really worries me is that there are still systems with Sandy Bridge CPUs available for sale on many sites. Having previously worked in retail, I know that it can sometimes take weeks for recalled items to be fully purged from the distribution system. I'm not taking a chance on ordering one of the Sandy Bridge systems right now. I'm just going to take the mild performance hit and go with a previous-gen system. Hell, it is still going to be a major upgrade from my current system.

thatguyandrew1992 said,
This sucks, but Intel is doing the right thing. I'm glad that they are replacing the mobos.

I seriously doubt they are doing it out of the goodness of their hearts. I can only imagine the lawsuits that would be heading their way if they didn't recall the boards.

jafoman said,

Of course.

On engadget: Jimmy sent us a chat log with an Intel customer service representative indicating that this recall only affects "some desktop boards based on Intel P67 chipset,"...

though some others might be affected.


djnv2010 said,
someone call in Will-I-AM, LMAO!! glad i bought my I7-870 when i did!!

Why? "I gota feeling" that Intel will fix this pretty fast

djnv2010 said,
someone call in Will-I-AM, LMAO!! glad i bought my I7-870 when i did!!

I'm quite sure this was the result of a Dirty Bit.

Julius Caro said,

which is oh so conventional


Perhaps you should take a deep breath and let the Neowin staff handle their website the way they wish? Spewing bitter posts just for the sake of your battered ego is not going to do much good.

"In some cases, this flaw causes ports to degrade over time, causing users to suffer from poor performance on any Serial ATA devices such as hard drives, DVD drives or BluRay drives."

Permanently? Is it just the mobo or do even the drives themselves degrade?

GS:mac

Glassed Silver said,
"In some cases, this flaw causes ports to degrade over time, causing users to suffer from poor performance on any Serial ATA devices such as hard drives, DVD drives or BluRay drives."

Permanently? Is it just the mobo or do even the drives themselves degrade?

GS:mac

Just the ports, not the disks

/- Razorfold said,
More information: http://www.anandtech.com/show/...eries-chipset-begins-recall

It only affects the SATA 3gb/s ports..so if you're using the 6gb/s ones (the first 2 ports) you will be fine.

I would still get a replacement. Having only two functioning SATA ports would be a great hindrance to most people, maybe not right away when the machine is new, but if they add more storage, they will see this flaw. I know that I am not an outliner when it comes to desktop towers, and I have 4 SATA ports used. My dad has at least 6 used. A minimal setup will use two, that is to say one HDD and one Optical drive

Sraf said,

I would still get a replacement. Having only two functioning SATA ports would be a great hindrance to most people, maybe not right away when the machine is new, but if they add more storage, they will see this flaw. I know that I am not an outliner when it comes to desktop towers, and I have 4 SATA ports used. My dad has at least 6 used. A minimal setup will use two, that is to say one HDD and one Optical drive


True. Heck; it would be a major problem for *me* (three SATA devices; one optical and two hard drives), and I'm still pushing a Celeron DC in mATX.

PGHammer said,


True. Heck; it would be a major problem for *me* (three SATA devices; one optical and two hard drives), and I'm still pushing a Celeron DC in mATX.

Oh that's sad. I feel for you. I don't think I could get out of bed if I were still stuck in the world of Celeron's.

Well that sucks for any early adopters. But, then again as an early adopter you go in to it knowing issues like this can occur. Intel seems to be handling the situation well though.

GrandTheftAuto said,
I bet you got this from Engadget and you haven't sourced back to it.

While this particular news item will be ALL over the place, I have found neowin to echo on stuff only published by engadget without giving any props to engadget. Like when the internet caught up on that android SMS bug, Cnet published it first, THEN engadget, so it became big (and the bug had been there for months). Neowin conveniently linked to the bug reports at google is if they had caught wind of it out of nowhere.

Julius Caro said,

While this particular news item will be ALL over the place, I have found neowin to echo on stuff only published by engadget without giving any props to engadget. Like when the internet caught up on that android SMS bug, Cnet published it first, THEN engadget, so it became big (and the bug had been there for months). Neowin conveniently linked to the bug reports at google is if they had caught wind of it out of nowhere.

Yeah, and how do you know, like, for a fact that they didn't?

Julius Caro said,

While this particular news item will be ALL over the place, I have found neowin to echo on stuff only published by engadget without giving any props to engadget. Like when the internet caught up on that android SMS bug, Cnet published it first, THEN engadget, so it became big (and the bug had been there for months). Neowin conveniently linked to the bug reports at google is if they had caught wind of it out of nowhere.

Who's to say that news staff isn't watching bug reports? I know for a fact some people are. Even so, you seeing it on Engadget doesn't make them the only viable source.

Tzvi Friedman said,

Who's to say that news staff isn't watching bug reports? I know for a fact some people are. Even so, you seeing it on Engadget doesn't make them the only viable source.

The bug report was there for a few months. Cnet published a story based on some e-mails from some dude. Cnet thanked the dude for his contribution and linked to the bug reports. Engadget published the story and linked to Cnet as the source. Neowin, no longer than 2 days later, published the same story, but they wrote it as if they had magically learned about the bug at the same time as engadget as cnet, without linking to any.

Now come on. The bug was there for months. At least have the decency to link to whoever uncovered it.

http://www.engadget.com/2010/1...ging-bugs-thatll-get-you-f/

http://www.zdnet.com/blog/hard...es-to-random-contacts/10796

http://code.google.com/p/android/issues/detail?id=9392


The bug was first reported in June. JUNE. at least 5 months before it became 'big' (prompting google to actually fix it real quick).

It was first reported by Cnet on the 31st (they thanked some dude for the heads up). A day later, this was posted on neowin. http://www.neowin.net/news/and...roulette-with-text-messages

There's no way in hell you could play the "coincidence" card here.