Updated: 17 burning questions on Intel's Sandy Bridge chipset fiasco
Updated: 17 burning questions on Intel's Sandy Bridge chipset fiasco
Posted 01/31/2011 at 9:52am
| by Paul Lilly
64
Comments
Print
Santa Clara, we have a problem. That's the message Intel engineers had to deliver to company execs after discovering a "design issue" in the company's 6-Series chipsets. The issue is severe enough that Intel decided to halt shipments while it implements a fix.
"In some cases, the Serial-ATA (SATA) 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 in a statement. "The chipset is utilized in PCs with Intel's latest Second Generation Intel Core processors, code-named Sandy Bridge. Intel has stopped shipment of the affected support chip from its factories."
Intel said it has already corrected the issue and has started making a new version of the support chip that doesn't have the design flaw. In addition, the Santa Clara chip maker says Sandy Bridge processors and other other related products are unaffected.
Jump past the picture for all of the late breaking information on the situation.
6-series chipset fiasco FAQ from the Maximum PC Staff:
Q: What’s the problem?
A: Some of the SATA ports that come off of the Intel 6-series PCH or Peripheral Control Hub may begin producing errors over time. The problem isn’t immediate, but can happen over time. That means that a port that is working fine today, may start to generate data errors. This will affect performance and the port may simply stop working over time. Increased thermals and voltage on the ports may also contribute to the port failing sooner. The only good news is that ports 0/1 are unaffected by the bug.
Q: Is it the actual Sandy Bridge CPU that’s bad?
A: No. Intel said the problem is not the CPU, but the chipset. However, one thing that we don’t know yet is why Intel said it should only affect those with Core i5 and Core i7 processors. That implies that 6-series with Core i3 is fine. We are still waiting for more details.
Q: How did Intel find it?
A: The company said it came to light last week when customers, either OEMs or board vendors, brought it to Intel's attention. Intel labs was able to verify the problem and the decision was made yesterday to halt shipment of the parts.
Q: How is Intel going to fix this?
A: The company is correcting the problem in silicon and is in process of fabbing new chipsets without the problem. The bad news: boards and systems with the fixed chip will likely not be available for at least a month.
Q: Is the chipset in my motherboard or notebook impacted?
A: Intel didn’t get into specifics on Monday but the short answer is probably yes. However, one aspect of the problem is that it may only manifest itself on occasion. That is, some boards may exhibit the issue, while others will not. The problem seems to be unpredictable and only after three years, Intel projected 5 percent failures but said heavier use could accelerate the issue. The upshot is that while all of the chipsets have the bug, not all of them will fail over three years of use.
Q: So my notebook is bad?
The good news is that most notebooks only have two SATA ports which happen to run on ports 0/1. But as we said, that’s most notebooks. Some notebooks that use more than ports 0/1 or that used the others ports would be affected.
Q: Can this be fixed with a new BIOS or some other firmware update?
A: No. The problem is a hardware issue that cannot be fixed without replacing the affected chipset on the system. Since you can’t simply replace a chipset in a motherboard, the only way to fix it would be to complete replace the motherboard.
Q: So which ports are bad on my motherboard again?
A: Intel’s PCH supports six ports. Of those, ports 0/1 are unaffected. These are also the SATA 6Gb/s ports. The remaining four SATA 3Gb/s ports may or may not exhibit the breakdown over time. Many motherboards also support additional ports using third-party controllers such as Marvell’s SATA 6Gb/s. These would be unaffected. eSATA ports on the rear of the motherboard should also be unaffected as those mostly use third-party controllers from Jmicron.
Q: What is Intel going to do to make me whole?
A: That’s not completely clear at this point. Intel said this morning that: “Intel stands behind its products and is committed to product quality. For computer makers and other Intel customers that have bought potentially affected chipsets or systems, Intel will work with its OEM partners to accept the return of the affected chipsets, and plans to support modifications or replacements needed on motherboards or systems.” However, motherboard and PC vendors we spoke with Monday had no guidance on the topic yet. Stay tuned to this FAQ to find out what vendors will be doing.
Q: Will my motherboard vendor replace my motherboard?
A: That’s not known. As of this morning, board vendors had stopped shipping 6-series boards. They were being updated by Intel on details of the problem. Whether board vendors would replace the boards or not is not known yet so stay tuned. On pure speculation though, it’s quite possible that vendors may offer to replace the board with one with the newer spin of the chipset once they are available. They may also opt to let the consumer extend the warranty instead of replacing the board. Again, this is pure speculation, but we know from previous experience that you shouldn’t always expect vendors to simply replace it.
Q: My system has a 6-series chipset, will my OEM replace the motherboard?
A: We don't have answers from all OEMs at this point, but several have offered to outright replace boards once updated motherboards are available. Puget Systems, Maingear and Origin PC are offering to replace the bad boards or if the consumer wishes, send a SATA add-in card. For those who are willing to let it hang out, Origin PC has said that it will offer lifetime warranties on systems with the affected boards. Several vendors said they are also offering add-in SATA cards to replace the ports if the consumer does not want to deal with the hassle of downtime. Besides extending the warranty to three years and offering to replace the affected boards, Cyberpower PC is also offering to let those who have Sandy Bridge systems in hand upgrade to LGA1366-based motherboards with a Core i7-950 instead. We're hoping to have a roll call from OEMs this week as they sort out their policies.
Q: This sounds like a class-action lawsuit in the making.
A: How Intel will make people whole isn’t known yet, but yes, we suspect that someone unhappy with how this turns out will certainly file suit.
Q: Is it OK to practice schadenfreude now and prance about why I’m so lucky I built an AMD system or X58 box?
A: Yes. The full schadenfreude effect is in place. Please feel free to dance about forums and post about how it sucks to not be you. Just remember that schadenfreude (deriving pleasure from the misfortune of others) often leads to Karmic justice.
Q: Will Intel be creating a new chipset model with the fix so we can know when it's safe to buy a motherboard for Sandy Bridge?
A: Intel will not create a new chipset model, and a re-spin of the silicon will correct the issue, the company said. OEM's have privately told us that the B3 spin of the 6-series chipset is not impacted, but we have not been able to confirm that with Intel. As far as buying a system, notebook or board with the issue; the vast majority of stores and OEMs have already halted shipment of Sandy Bridge-based systems. Systems and boards will likely not start shipping until the fix is in.
Q: Is this the first time Intel has had a chipset problem of this proportion?
A: Actually no. Although the 6-series chipset screwup may be the worst in dollar amount, Intel actually blew it on the “Cape Cod” CC820 chipset back in 2000. That Pentium III chipset was designed to sidestep resistance to Direct RDRAM. The CC820 took the “Vancouver” 820 chipset and integrated a Memory Translator Hub so that the RDRAM-only chipset would work with SDRAM. Unfortunately, bugs in the chipset forced Intel to recall chipsets and boards based on the CC820. The final tune for that mistake was $253 million which adjusted for inflation would be roughly $312 million.
Q: I have a 6-series chipset in my system or notebook, is it safe to use?
A: Intel and numerous OEMs say that since the problem should occur over several years and only to a fairly small percentage of systems and notebooks, the risk of data loss is nil. If a port goes bad, you just move it to a port that works and the data will be there, Intel says.
Q: When will I be able to buy a Sandy Bridge motherboard?
A: This may be the biggest problem with the chipset: a major disruption of the supply chain. Intel says it will have the updated chipset in production this month but a full volume ramp would not return until April. As of Tuesday morning, major board vendors were still unclear on when they would have replacement boards in hand. However, some OEMs are saying that updated Sandy Bridge boards could be in their hands as soon as 4 to 6 weeks. Compliciating the matter is the approach of Chinese New Year when much of China and Taiwan are shut down for the holiday. More information will be available here as we get it.
More like this
64
Comments
Comments are closed on this article
Lar Burstein
March 02, 2011 at 5:07pm
Finally!
New Egg has 3 LGA 1155 boards on it's web site!!! Unfortunately, the ones that it has are Crossfire, not SLI... but the wait is finally nearing it's end!
zodi
March 01, 2011 at 6:18pm
I have asus P67pro, I just moved my stuff off the affect ports and on to the 4 SATA 6 ports.
I got a good deal with our computer store, they are going to swap the board and re-install the new one for free. Considering I built the system myself, it will be nice not have to waste a day putting the new one in.
aarcane
February 06, 2011 at 7:20pm
can we get a timestamp on this to indicate when it was last updated? or a changelog of somesort? I've been checking frequently, but I can never quite tell what's changed.
On an interresting note, I just got the March 2011 issue in my inbox, telling me how to build a sandy bridge system~
triptych
February 04, 2011 at 4:32am
After learning of this chipset bug, I contacted Newegg about my recently purchased ASUS P8P67-M Pro. I'm not sure how this affects anyone who may have already gone past the 30-day return period or if it is limited to ASUS motherboards (I'm assuming it's for all 6 series mobos). Here's the official response from Newegg (I've highlighted the most important part):
As always, Newegg remains 100% committed to our customers' total satisfaction. In keeping with our commitment to our customers, we are extending the return period for your motherboard by 90 days or until replacements become available from the manufacturer, whichever is greater. Intel expects to have a new revision of the P67 & H67 chipsets out around April, at which point first-run motherboards with this issue will need to be physically replaced in affected systems.
SpunkySpine
February 04, 2011 at 3:27am
Just change to ports 0/1 if you have errors?
Just a sec....
If I'm getting READ ERRORS from my HD...
Don't you suspect I also have WRITE ERRORS???
So much for just changing to the "Good" ports after I get frequent errors.
Just my thoughts.
ps: believe previous post failed... forgive my posting twice otherwise.
win7fanboi
February 02, 2011 at 8:47am
Ok Intel is making record profit and then this? I suppose its inevitable for a company to get arrogant/lazy when they have virtually no competition. AMD is getting a lot of breaks. They should jump in with CPUs that have USB 3 baked in while Intel is trying to keep itself from swallowing its tongue. To think I was planning to make my first build using the Sandy Bridge proc even though I detest the possibility of handing over money to Intel when they are dragging their feet to support USB 3. Well Intel this 'tock' just became a non starter, what else you got? They should use all that money they made to hire folks to insult management everyday to keep them humble :)
edit : <<insert Gordon's rant here>>
Caboose
February 01, 2011 at 3:49pm
Lets not forget that AMD had an entire CPU line with a TLB bug and they weren't recalled. Mind you there was a BIOS tweak to disable the bug but still. I had a Phenom x4. Really liked it, but I found that my system was much more stable when I moved over to a Phenom II x4.
logicmaster2003
February 01, 2011 at 10:38am
I guess I got lucky ! I got these 2 from microcenter this past saturday...
Asus P8P67-M PRO LGA 1155 mATX Intel Motherboard $149.99
Intel Core i5 2500K LGA 1155 Boxed Processor $179.99
zakn
February 01, 2011 at 6:30am
I dropped alot of MFers on my first read on this story, but moving 2 Sata (Since 0/1 aren't affected) cables (I run 3 drives and an optical) to the SATA6 channels isn't a big deal.
I'd take a Warrenty extention to say 5 years over joining the class action or sending the board back
JohnP
February 01, 2011 at 8:00am
Yeah, now that I have calmed down, I too am going to go with a "no big deal" attitude. If any of the SATA ports break, I will replace the board, not before. Moving my two hard drives onto the 6G ports is not exactly gene splicing and I already had my SSD drive on one to begin with. As for the optical drives, Who cares? Lose a port? Move it to another.
Oops, spoke too soon. My Samsung 1TB hard drive goes CLUNK, CLUNK, etc when attached to a 6G port. So only the SSD drive for now until a "6G enabled" hard drive comes around.
I also just thought about it and buying a replacement mobo for $200 once the issue is in production is what I will do. Then i can replace the board and send in the "broken" one for RMA. I have recycled a LOT more expensive mobos over the years for various reasons!
zakn
February 02, 2011 at 9:40am
The ports should be backward compatable. I moved everything off the affected SATA2 ports and I am now good to go.
I'm good to go unless I want to add drives.
mike323323
January 31, 2011 at 10:21pm
Max PC Staff! I think this needs a little editing! -- "The company has is correcting the problem in silicon and is in process of fabbing new chipsets without the problem."
(Also, "comments_top_anchor" doesn't exist in the source code!)
suntower
January 31, 2011 at 9:54pm
QUESTION: I just got a mobo Friday. I can RMA it, but it -seems- to be working OK now. So... if I wish to live 'dangerously', can I use it and rest assured that I am not frying my HDDs if I use ports 2-5?
IOW: Is the -worst- that can happen that I run into data errors or slow response? I'm not going to -damage- my HDDs right?
I can (temporarily) live with the possible errors, but I don't want to risk smoking my new HDs.
schneider1492
February 01, 2011 at 5:37am
I think that was covered when they said: "If a port goes bad, you just move it to a port that works and the data will be there, Intel says."
Brewskie
January 31, 2011 at 9:39pm
Ok. They found a problem and thankfully it was just before I actually bought one of the P67 mainboards.
Now, if I wait for three months to purchase the board how will I know it is the newer version?
Call it the P67"b"?
Conal_keaney
January 31, 2011 at 8:26pm
I was so close to just settling down on a lga 1155 system but decided just to stick it out and wait for the better LGA2011 CPUs to roll out. I might even wait some to see if Intel doesn't bust another "Intel".
Carlidan
January 31, 2011 at 5:44pm
What the hell. I just bought the 2600k at Fry's. If the cpu isn't affected why isn't Newegg selling them? Do I have to return the cpu? I just bought it Sunday.
itommyboy
February 06, 2011 at 2:27pm
No you don't have to return anything if you don't want to. If you just bought the 2600k processor only it's fine - the cpu's don't have the issues the sata ports on the new series motherboards do. The question for you however is what do you drop that 2600k onto at this point.
HeartBurnKid
January 31, 2011 at 5:06pm
"Q: Is it the actual Sandy Bridge CPU that’s bad?
A: No. Intel said the problem is not the CPU, but the chipset. However, one thing that we don’t know yet is why Intel said it should only affect those with Core i5 and Core i7 processors. That implies that 6-series with Core i3 is fine. We are still waiting for more details."
Have the Socket 1155 Core i3's even shipped yet? Newegg doesn't have them.
That's probably why Intel didn't specify them as affected; they're not in the wild.
gordonung
January 31, 2011 at 5:26pm
Hmm, that might explain it but I'm pretty sure that I've seen Core i3-2100 parts available on the Internet.
Vipsanius
January 31, 2011 at 4:53pm
schadenfreude...lmao
love the newest Q&A...and also hate those people right now...however, at least I still have a chance to send my build back for a refund, or keep it and not use the bad ports so I don't have to deal without a computer for a month or two
so does that mean people with a P67 board, who can get around the issue unaffected, be schadenfreuding those who will now have to wait months?
binkievanes
January 31, 2011 at 4:45pm
I noticed that Newegg and Amazon are still selling motherboards but took out the sandy bridge cpu's ??
To be honest I am really confused now and will postpone a new build for a few month at least untill the dust settles and we will now what is what.
binkievanes
January 31, 2011 at 6:53pm
amazon sells it all again
Newegg has erased all references to sandybridge and socket 1155 or chipset p69 from their site
Vipsanius
January 31, 2011 at 5:03pm
all I know is that the Asus Maxiumus IV I bought this morning and the Sabertooth I was going to buy for another build, are completely off the Newegg site (along with many others)...also, the 2 Asus boards left on the site are unable to be accessed
before, you could click on the motherboard and it would say "out of stock" and "deactivated"...I'm guessing they've restricted access to the pages to stop the flood of hate ratings I'm sure would have been coming
gordonung
January 31, 2011 at 4:39pm
If the optical/HDD runs off of ports 0/1, you should be fine. Since notebooks do not commonly have more than 2 ports, you would probably not see a problem since it is only the SATA 3Gb/s ports. If it runs off of the ports that are potentially bad, then you might see a problem manifest itself within three years. Again, it's tough to say until Intel releses more information on the issue and the OEMs/ODMs have time to digest it.
Gutter96
January 31, 2011 at 4:21pm
The "Caminogate" scandal as well! The Original Camino chipset (the 840 iirc) was recalled as well due to memory errors. In the end all it took was one resistor to repair, but the damage was already done. I guess it's not so bad to be rollin' an 890FX right now after all! :-)
scoop6274
January 31, 2011 at 4:10pm
Boy, do I feel lucky that I went with AMD instead of Intel. Cost less, works... hmmmmm.
Ceadderman
February 01, 2011 at 10:40am
Well actually Sandy Bridge costs about the same as 1090T. Maybe a bit more. I'm an AMD guy but I have to at least applaud Intel for bringing their price point down. Now if the other system manufacturers would set a reasonable price point I MIGHT entertain the idea of building a Sandy Bridge system.
Intel would have to have this issue completely resolved of course.
I don't mind issues. They come with the territory of system building. I did the NB fix on my Crosshair IV(41-43c @idle now) so it's not like I cannot work with minor deficiencies. This to me is not minor since I obviously cannot separate the layers of a MoBo to correct the issue. Nor would I want to even if I could.
Tech-Monkey
January 31, 2011 at 2:09pm
Here is a fix for Intel, how about dropping the 4 - 3 Gbs ports and just put in 6 - 6 Gbs ports? The ports are backwards compatible anyway.
JohnP
January 31, 2011 at 2:04pm
Cripes! Here I am, seriously fiddling with the damn setting on my ASUS P8P67 EVO board that is two weeks old. It really, really does NOT play nice with my hard drives, esp. My Vertex2 SSD drive! Drives just evaporate from the computer after sleeping or reboot or just stop responding. ASUS has been tearing out their hair trying to fix it via the UEFI bois updates (3 in the past week!)
Whew, least I know that it is NOT a setting issue AND I have my old i7 computer sitting on a chair right next to me!
patrickmaher
January 31, 2011 at 5:38pm
I've been doing some research on SSDs lately and I have seen warnings about problems with SSDs with SandForce controllers having problems after waking from sleep/hibernate. Not sure if that is your problem but it might be woth checking out.
http://www.ocztechnologyforum.com/forum/showthread.php?80715-Vertex-2-BSOD-after-sleep
JohnP
February 01, 2011 at 12:37am
Yeah, that sounds EXACTLY like what I am seeing so it probably is a SandForce/Intel chipset issue rather than this "degradation" problem. LOTS of teething issues with a new CPU architecture/ a new chipset/ and new mobo/ AND a new type of BIOS. ASUS is going nuts...
gordonung
January 31, 2011 at 4:35pm
Have you been running a lot of voltage to the PCH? Intel claims that it's almost impossible that anyone could see the issue from day one so your case (if it is indeed the problem) be quite interesting.
JohnP
February 01, 2011 at 12:40am
Gordon, you can rest easy. Just teething problems and a overwrought poster, not a SATA degradation issue...yet...
blkpanthr
January 31, 2011 at 6:19pm
this issue is unrelated.
If you go over to the Hard OC forums, youll see allot of people with problems with SSD and Cougar Point, but its been traced back to PLL overvoltage and Sleep modes on the intel SATA ports. Just disable PLL and it should clear itself up.
Asus and Intel are working together on a new version of the Intel RST drivers to fix the issue.
JohnP
February 01, 2011 at 12:32am
NO, PLL OVERVOLTAGE DISABLING DOES NOT FIX THE PROBLEM. I wish!
I have tried Disabling the PLL overvoltage settings, the C1/C3, resetting the Intel Rapid Storage Tech driver, different versions of the U-Fee BIOS, and a half a dozen other methods outlined in the 33 page tome of a forum post.
Don't be so glibe with "easy answers" if you have not READ ALL THE POSTS! You would clearly see that the issue remains unsolved.
blkpanthr
February 01, 2011 at 11:50am
please read my entire post, as in the last line. My repsonse was specifically directect towartds the sleep issue.
you can also remove the IRST driver and revert to the MS one in the meantimne, that has worked for many.
If YOU had read all the posts over there, you would see that disableing PLL overvolt has indeed fixed MANY peoples problems with sleep. Perhaps not yours, but many others. Id hardly classsify that as glib.
If you have the SSD randomly dissapearing problem, that is a SANDFORCE incompatability issue, youll have to wait for a new firmaware.
people are seriously over-reacting. As intel has said the issue with the Silcon has not been observed in the wild. Its purely on paper. 5%-15% will have problem in 3-5 years? big deal, i wont even have this mobo by the time it becomes an issue, and if i do, theres plenty of other ports.
Im usung the p8p67deluxe with no problems at all, and 4.8ghz stock settings, 4 sata drives, everything enabled, etc....
Vipsanius
January 31, 2011 at 1:56pm
wow, wth...I just ordered a brand new $4400 computer on newegg less than 2 hours ago...they've already pulled the items off their site, but when I called them to cancel, they said it had already been sent to the warehouse for shipping and can't be cancelled...seriously wth
newegg > FAIL
intel > FAIL
birds > falling from the sky
fish > turning belly up
this is the apocalypse isn't it???
Saigua
February 02, 2011 at 10:15am
G'wan, look at y'r board stats, benchmark y'r cold storage, log your BER, add alert triggers. Curate back!
Tech-Monkey
January 31, 2011 at 2:12pm
Just ask Newegg for the RMA number now, tell them copy of verbal request to stop shipment was denied and that you will be sending documentation to president of company. If he gets enough complaints of inability to cancel order and the cost of shipping boards back, he will figure out a way to modify order system to correct shipments.
Vipsanius
January 31, 2011 at 2:31pm
They told me to refuse shipment, and that when they got it back, I would get a full refund. I'll monitor the situation to see how bad it is until then. According to the Q&A just added to this announcement, I might be able to avoid using those ports to avoid problems. If that's the case, great!...I just got lucky as hell not having to wait a month plus. If not, I guess I'll be sitting on the front porch all day Thursday stalking the brown truck. Those guys are quick little boogers...usually gone before I even open the door.
Speaking of motherboards though, the whole reason I'm having to rebuild is because the DFI Lanboy I got 2.5 years ago decided to just burst into flames...literally...last Friday. Even my Samsung monitor takes 5 minutes to "warm up" because of leaky capicitors. What the hell happened to Quality Control?
Tech-Monkey
January 31, 2011 at 2:12pm
Just ask Newegg for the RMA number now, tell them copy of verbal request to stop shipment was denied and that you will be sending documentation to president of company. If he gets enough complaints of inability to cancel order and the cost of shipping boards back, he will figure out a way to modify order system to correct shipments.
calhydro
January 31, 2011 at 1:53pm
Does this mean the customer will have to identify the defect in their motherboard? Shouldn't Intel develope a ultility to detect this defect in the affected.motherboards? So vendors will continue to sell the defective boards to customers and we end up have to deal with Intel. Has chipset revision been determined to fix the problem and how will it be advertised so that someone can avoid buying the defective boards.
Future Sandy Bridge motherboard buyer.
gordonung
January 31, 2011 at 4:33pm
As I understand it, you don't have to identify it. If you have an existing Sandy Bridge board, you may or may not be have the issue manifest itself. The problem is that it may or may not happen and it may take three years or longer to even pop up. For the moment, board makers, system builders and Intel say use it and it will be fine as no one is seeing the bad SATA crop up from day one. Hopefully your board vendor will opt to replace your board with one with the newer spin of the chipset.
maxzorro
February 01, 2011 at 3:56am
I think the issue is that many vendors will unscrupulously sell the bad boards after the new ones have come out ... unless there is an easy way to identify the bad chip from the good one.
Adromeda
January 31, 2011 at 1:52pm
Wouldn't the easy fix now to just install Sata controllers in your PCI slot? Seems like that will fix the problem :).
Connect with MaximumPC
Featured Content
The gang discusses the GTX 690, Trinity, Ivy Bridge, Amazon, big-box stores, MMOs,...
This month's issue
Feature
Build a PC On Any Budget
Feature
Cloud Services Showdown
How-To
Encrypt Your External Drive
Build It
Upgrade an X58 Rig
Most Commented Articles
Latest Max PC Tweets
- maximumpc: Original RickRoll Video Returns to YouTube After 24-Hour Copyright Hiatus http://t.co/MaH9Sxyy2 days 4 hours ago
- maximumpc: Corsair has decided not to go forward with its $78 million IPO, citing weak equity market conditions http://t.co/A5a4DAzj2 days 4 hours ago
- maximumpc: Want to work at Maximum PC? Our Online Managing Editor, Alex, is moving away, so we need a new one. http://t.co/9Z7JCh0E4 days 2 hours ago
