Intel says flaw in Series 6 Sandy Bridge chipsets

[Modelworks]

I don't think that's the case, this is isolated to the controller that supplies voltage to those SATA ports from what I understand. It does not effect anything else.

Where it is located, is it a totally separate chip or internal to the controller ? If internal then they are all created on the same die at the fab and one defect can easily migrate to other areas.

 

[Anomaly1964]

Ok, forgive me if this is a Dumb question but... Intel designs the CPU obviously, do the in turn design the motherboards/chipsets as well and then motherboard companies actually manufacture them? I'm just trying to figure out how a bad SATA port is Intel's fault...?

 

[Modelworks]

Ok, forgive me if this is a Dumb question but... Intel designs the CPU obviously, do the in turn design the motherboards/chipsets as well and then motherboard companies actually manufacture them? I'm just trying to figure out how a bad SATA port is Intel's fault...?

Intel designs the cpu, chipsets, and support chips. Motherboard makers choose the layout of where the chips will be located on the board and what features they will support.

 

[solarissf]

I just spoke to ASUS, they told me even though they are aware of an intel problem it has not been verified that there is a problem with their board p67pro. That if there is a recall they will notify me directly. Is this me or are they trying not to do anything about the problem? Kept saying it was an intel chip problem, not a problem with their motherboard.

 

[Seven]

I just spoke to ASUS, they told me even though they are aware of an intel problem it has not been verified that there is a problem with their board p67pro. That if there is a recall they will notify me directly. Is this me or are they trying not to do anything about the problem? Kept saying it was an intel chip problem, not a problem with their motherboard.

Typical rep answer though.

 

[bcterps]

I'm guessing that if they fixed it this quickly then it was a manufacturing issue. If they had to change something in the design it would have taken much longer.

I think people would be amazed at how much of produced die is tossed out because of flaws, some fabs discard product at a rate as high as 80%.


Still this looks bad for quality control .

It's actually a design issue, not a mfg issue, according to Anand's latest update. Some engineer screwed up big time...

http://www.anandtech.com/show/4143/the-source-of-intels-cougar-point-sata-bug

 

[Diogenes2]

I just spoke to ASUS, they told me even though they are aware of an intel problem it has not been verified that there is a problem with their board p67pro. That if there is a recall they will notify me directly. Is this me or are they trying not to do anything about the problem? Kept saying it was an intel chip problem, not a problem with their motherboard.

Sounds like generic CYA at this time.. " Don't admit anything "...

Why talk about replacing boards before the new chips are available ?

 

[Hogan773]

I just spoke to ASUS, they told me even though they are aware of an intel problem it has not been verified that there is a problem with their board p67pro. That if there is a recall they will notify me directly. Is this me or are they trying not to do anything about the problem? Kept saying it was an intel chip problem, not a problem with their motherboard.

Dude it has been 24 HOURS since they (ASUS) found out about it. You expect them to have their full corporate policy on how a recall would work even when INTEL probably hasn't even designed the recall? You expect them to start making statements for the record like "yeah I know that board we manufactured has a defect"....

People need to just relax and wait a little bit. Corporate America doesn't move to the same speed of posters on Anandtech Forums.

 

[bcterps]

Yes. And keep in mind that if you have a Marvell SATA III (6 Gbps) controller it is not meant for optical drives, only data drives. So one thing to keep in mind when deciding what to do. That might be what I do since I can't think of anything else I plan on using the expansion slot for at the moment and I'm not using the 3Gbps ports right now anyway. It sticks in my craw a little bit that my new motherboard isn't exactly what I paid for but all things considered this is NOT the huge deal most people are making it out to be...at least not for me. I think the retailers should give people the option of either full free RMA for a fixed board when they come out, or get a $25-$50 cash rebate (covered by Intel) to make up for having paid for something you don't get or is hobbled, or a free add on card to restore the functionality.

My guess is people who bought less expensive boards without additional 6Gbps ports and need more ports will go for options 1 or 3 whereas a lot of people who bought more expensive boards and who still have enough available unaffected ports without using the 4 bad ones might just take the cash rebate and be happy.

That's how I'd handle the situation if I were Intel and their partners but what do I know? I'm just an idiot customer who paid full price for beta hardware.

This is a reasonably big deal for me. I built a mATX system with the best mATX board available (Asus P8P67-M Pro), and it only has 3 available SATA slots (2x6GB, 1xMarvell). With a SSD and a DVD drive, that really limits my ability to expand. Because it's an mATX board, I also have less PCI slots to work with. I was planning on doing Crossfire down the road which would have eaten up the last PCI slot. For the time being, I can manage, but for the longer term, I will need to RMA my board.

 

[Anomaly1964]

It's actually a design issue, not a mfg issue, according to Anand's latest update. Some engineer screwed up big time...

http://www.anandtech.com/show/4143/the-source-of-intels-cougar-point-sata-bug

Based on this...

The problem in the chipset was traced back to a transistor in the 3Gbps PLL clocking tree. The aforementioned transistor has a very thin gate oxide, which allows you to turn it on with a very low voltage. Unfortunately in this case Intel biased the transistor with too high of a voltage, resulting in higher than expected leakage current. Depending on the physical characteristics of the transistor the leakage current here can increase over time which can ultimately result in this failure on the 3Gbps ports. The fact that the 3Gbps and 6Gbps circuits have their own independent clocking trees is what ensures that this problem is limited to only ports 2 - 5 off the controller.

And then this on the Asus Site about the Sabertooth:

TUF Components (Alloy Choke, Cap. & MOSFET; Certified by Military-standard)

Certified for Tough Duty
Get rugged performance even in the most challenging conditions with robust TUF chokes, solid capacitors, and MOSFETs--certified through third-party, military-grade testing. TUF Chokes, also known as the “Alloy Choke”, is a made of a compound of various types of metal instead standard iron, enables the support of up to a massive 40A of rated current, 25% higher than conventional component. Furthermore, the single piece packaing also elimates the emission of virbation noise, delivering superb charactoristics as well as durability under extreme conditions.

Or does one really have nothing to do with the other?

 

[PingSpike]

I don't know about that not a full blown recall business. I'm sure the big OEMs are fine, but it kind of puts a bleh on all direct to consumer motherboards to me. I think I'm going to duck tape up my current rig and wait for z68. When is that coming out again? I just read Q2.

 

[solarissf]

yea, guess your right. just didn't like the tone on the phone ... just emanated "not our problem"

 

[Yables]

At least gigabyte have acknowledged the problem
http://www.gigabyte.com/press-center/news-page.aspx?nid=984


Understandably at this early stage no details on how the current P67 mobo owners get to swap to the revised units when they become available

 

[eddietandy]

I wonder if using all four SATA 2 ports makes the defect more likely to rear its ugly head than if you just used one port. My understanding is it's not a question of if, but when. So the more you use them, the quicker it will fail.

Another thing unclear is what Intel's 2 year estimate really means. Is that 2 years of 24/7 runtime, or based on a certain number of hours usage per day?

 

[Minjin]

Based on this...

The problem in the chipset was traced back to a transistor in the 3Gbps PLL clocking tree. The aforementioned transistor has a very thin gate oxide, which allows you to turn it on with a very low voltage. Unfortunately in this case Intel biased the transistor with too high of a voltage, resulting in higher than expected leakage current. Depending on the physical characteristics of the transistor the leakage current here can increase over time which can ultimately result in this failure on the 3Gbps ports. The fact that the 3Gbps and 6Gbps circuits have their own independent clocking trees is what ensures that this problem is limited to only ports 2 - 5 off the controller.

And then this on the Asus Site about the Sabertooth:

TUF Components (Alloy Choke, Cap. & MOSFET; Certified by Military-standard)

Certified for Tough Duty
Get rugged performance even in the most challenging conditions with robust TUF chokes, solid capacitors, and MOSFETs--certified through third-party, military-grade testing. TUF Chokes, also known as the “Alloy Choke”, is a made of a compound of various types of metal instead standard iron, enables the support of up to a massive 40A of rated current, 25% higher than conventional component. Furthermore, the single piece packaing also elimates the emission of virbation noise, delivering superb charactoristics as well as durability under extreme conditions.
/product_overview.jpg[/IMG]

Or does one really have nothing to do with the other?

They have nothing to do with each other.

 

[Hogan773]

http://www.asrock.com/news/events/201102ex/index.html

basically telling everyone to chill, use their Sata3 ports, and wait for further info.

Their English could use a little work though....perhaps they should hire someone who speaks English to proofread their press releases before they go out. There must be an out of work English major who is fluent in English who would do it via email for 5 bucks an hour! :awe:

 

[Anomaly1964]

They have nothing to do with each other.

Thanks for clearing that up for me...

 

[Minjin]

Thanks for clearing that up for me...

Each chipset IC probably contains somewhere around 50-100 million transistors (MOSFETs). In this situation, the gate oxide of a single transistor was made slightly too thin which caused the transistor to turn on too early. It isn't anything external or within control of the motherboard manufacturers. If it has the affected chip and uses the affected SATA controller (i.e. connects it to a SATA port), it will be affected by the bug.

To put it very simply, a MOSFET works by applying a voltage to the gate. This voltage via field interactions causes a metaphorical bridge to rise between the source and drain terminals, allowing conduction between the two. If you increase the gate oxide thickness, it takes more voltage at the gate to turn on the transistor. Conversely, if you decrease the thickness, the transistor turns on more easily. This can cause it to turn on when you don't want it to.

 

[Anomaly1964]

Each chipset IC probably contains somewhere around 50-100 million transistors (MOSFETs). In this situation, the gate oxide of a single transistor was made slightly too thin which caused the transistor to turn on too early. It isn't anything external or within control of the motherboard manufacturers. If it has the affected chip and uses the affected SATA controller (i.e. connects it to a SATA port), it will be affected by the bug.

To put it very simply, a MOSFET works by applying a voltage to the gate. This voltage via field interactions causes a metaphorical bridge to rise between the source and drain terminals, allowing conduction between the two. If you increase the gate oxide thickness, it takes more voltage at the gate to turn on the transistor. Conversely, if you decrease the thickness, the transistor turns on more easily. This can cause it to turn on when you don't want it to.

Great info...

So the reason Intel is on the hook here and not the MB manufacturers is because Intel gives them the specs to build with...?

 

[eddietandy]

Great info...

So the reason Intel is on the hook here and not the MB manufacturers is because Intel gives them the specs to build with...?

This would all happen within the chipset manufactured by Intel. The problem is already locked in by the time the mobo manufacturers get it.

 

[ahurtt]

Is this true re: Marvell controller? I was thinking of plugging my 2 HDDs into ports 0 and 1, and then moving my DVD to one of the Marvell Sata6.0 ports.

What motherboard do you have? I have the Asus P8P67 Pro and it says right in the manual that the Marvell 6Gbps SATA III ports are only intended for use with data drives, not ATAPI devices. This is most likely a limitation of the Marvell chipset and not the Asus motherboard itself. So if you have the same Marvell controller you probably have the same limitation. I was going to do the same thing as you initially until I noticed what it said in the manual about those Marvell ports not being for ATAPI devices. So I plugged my boot hard drive and DVD drive into Intel ports 0-1 and I plugged my secondary SATA drive into one of the Marvell 6Gbps ports.

 

[ahurtt]

I just spoke to ASUS, they told me even though they are aware of an intel problem it has not been verified that there is a problem with their board p67pro. That if there is a recall they will notify me directly. Is this me or are they trying not to do anything about the problem? Kept saying it was an intel chip problem, not a problem with their motherboard.

And it begins...I would have expected better from Asus...silly me. Technically I guess they are right...it is a chip problem. It's just that the chip with the problem happens to be soldered onto a motherboard they made which they themselves permanently soldered on there...Yeah Intel made the chip. But that doesn't absolve Asus of any responsibility for helping you fix the problem. What if this was the tack that Toyota took when it came to replacing all those faulty gas pedals in so many of its cars in that giant recall they did? "Well it's a problem with the gas pedal. Not our car. We didn't make the gas pedal...some supplier in Japan did." Sounds like this is the approach Asus is trying to take...I don't like the sound of this. Bad omen.

 

[Anomaly1964]

What motherboard do you have? I have the Asus P8P67 Pro and it says right in the manual that the Marvell 6Gbps SATA III ports are only intended for use with data drives, not ATAPI devices. This is most likely a limitation of the Marvell chipset and not the Asus motherboard itself. So if you have the same Marvell controller you probably have the same limitation. I was going to do the same thing as you initially until I noticed what it said in the manual about those Marvell ports not being for ATAPI devices. So I plugged my boot hard drive and DVD drive into Intel ports 0-1 and I plugged my secondary SATA drive into one of the Marvell 6Gbps ports.

I have my OPTICAL in the Marvel port and it is working fine...is that BAD, will it mess something up?

 

[Anomaly1964]

I have my OPTICAL in the Marvel port and it is working fine...is that BAD, will it mess something up?

Just checked, my Optical is SCSI...

 

[Dufus]

Hm, I wonder if Aida interprets this correctly at all

Please try downloading the latest hwinfo32 (v3.68) from http://www.hwinfo.com/download32.html

No installation required, no trial, just freeware and see what stepping you get under BUS-PCI Bus-Intel Cougar Point. Latest is B2.

Something kindly provided by mythlogic

ALL 6-series chipsets of B0, B1, B2 stepping (Which would be ALL of them no matter date of manufacture) have this issue. B3 stepping which corrects the issue will be out late this month.