Welcome to the Slashdot Beta site -- learn more here. Use the link in the footer or click here to return to the Classic version of Slashdot.

Thank you!

Before you choose to head back to the Classic look of the site, we'd appreciate it if you share your thoughts on the Beta; your feedback is what drives our ongoing development.

Beta is different and we value you taking the time to try it out. Please take a look at the changes we've made in Beta and  learn more about it. Thanks for reading, and for making the site better!

Asus, Gigabyte To Replace All Sandy Bridge Boards

samzenpus posted more than 3 years ago | from the return-to-sender dept.

Intel 180

J. Dzhugashvili writes "In the wake of Intel's announcement that all existing Sandy Bridge chipsets have a bug that causes degraded Serial ATA performance, top-tier motherboard makers Asus and Gigabyte have made public statements regarding their return policy for affected boards. Asus is promising 'hassle-free return and/or replacement', while Gigabyte says owners of affected boards are entitled to a full refund or replacement—and it recommends that users seek refunds. Both companies are advising users to contact the original place of purchase to proceed. On a related note, Gigabyte has announced that new Sandy Bridge motherboards with bug-free chipsets will be available in volume in April."

Sorry! There are no comments related to the filter you selected.

fp (0, Offtopic)

Anonymous Coward | more than 3 years ago | (#35083860)

props to all dead homiez

More tech detail (5, Informative)

FrankSchwab (675585) | more than 3 years ago | (#35083868)

For the chipheads, Anandtech has a good description of the underlying problem: []

Re:More tech detail (2)

icebike (68054) | more than 3 years ago | (#35083982)

Very good article.

I wonder if the wisest thing would be to just sit on one of those board till April, especially if your board is not yet experiencing the slow down. The article linked above suggests this problem gets worse with age.

But with replacement boards due in April I would opt for waiting. Of course some people can't/won't change out their own boards and warranty issues might not allow them to do so, but sending it back now gets you a refund, but you are still stuck without your computer.

Re:More tech detail (1)

TheGratefulNet (143330) | more than 3 years ago | (#35084334)

if its a phase lock loop drift, you WILL get errors. its not about 'if'.

just replace it. this is a 'bad clock' essentially.

That is what is being recommended (2)

Sycraft-fu (314770) | more than 3 years ago | (#35085402)

Newegg has a video out talking about it and they run down how you can still use the board, if you only have two devices just use the first couple ports and so on. Basically you can hold on to your board till the new ones come, then get a replacement.

The refund policy seems to just be to keep people happy. Some people look for any excuse to flip their lid and will do it over something like this and demand a refund. Intel headed all that off and told all suppliers to offer refunds, no questions asked, and Intel would reimburse them.

Personally if I were in the situation of owning a board, I'd just buy a SATA controller, since I have 4 harddrives, and then get a replacement board when they come out. The $30ish dollars would be worth it to just keep the system.

However Intel is offering people choices in what they want. Want a refund right now? No problem.

I think they learned their less on from the FDIV bug. They want to make sure nobody can accuse them of not dealing with the problem. Keep your board, get a replacement later, send it in get a refund now, they'll do whichever remedy makes you happy.

Re:More tech detail (1)

berwiki (989827) | more than 3 years ago | (#35084108)

I'm surprised these sorts of things don't happen more often.

There is so much going on with motherboards anymore, and there must be 1000+ motherboard variations in each generation. It's amazing that things like clock timings don't affect other portions of the board more frequently. Crazy.

Re:More tech detail (1)

Deflatamouse (37653) | more than 3 years ago | (#35084436)

Though there are a lot of motherboard variations, they are all built with mostly standard parts with well published design guidelines, e.g. an Intel reference design. I would say most boards out there are simply a small variation from the reference design. The chipsets are also likely designed with the tolerance for trace length, clock, voltage variations, etc.

I wonder if that's the whole story (1)

ron_ivi (607351) | more than 3 years ago | (#35084378)

As others point out, these motherboards don't need to be scraped, but could rather be resold as cheaper models with only the fast (6Gbps) SATA ports and none of the slow (3Gbps) ports by simply putting duct tape over the connectors connecting to the flaky circuits. I wonder if Intel's worried that similar problems may be scattered elsewhere across the chips, and therefor is budgeting for the cost of a full recall.

Re:More tech detail (0)

Anonymous Coward | more than 3 years ago | (#35084424)

I doubt it the leakage current increases over time. This would happen only if the gate oxide voltage was too high, causing some sort of soft breakdown. This requires a VERY high voltage.

N/PBTI is a reliability issue that causes the transistor threshold voltage (voltage at which it turns on) to *increase* over time, not decrease; the leakage current is dependent upon threshold voltage; higher Vth, higher I_leak. Thus, leakage current (I_leak) should be decreasing with time due to P/NBTI.

I think it is that somewhere somebody screwed up with the sizing of that transistor, and failed to simulate that leakage path properly over process and temperature.

In conclusion, Intel blames it on "over time" when it really is a process variation and temperature (PVT) problem, meaning they messed up from the get-go. Unfortunately, someone's gonna be in alot of trouble..

Re:More tech detail (0)

Anonymous Coward | more than 3 years ago | (#35084572)

OOps, i mean higher Vth, LOWER I_leak. Editing fail.

Re:More tech detail (1)

treeves (963993) | more than 3 years ago | (#35085254)

I didn't understand what the article said about just turning off the voltage to the transistor at issue to fix the problem.
How is letting the transistor fail different from disabling it by turning off the voltage to it?

Re:More tech detail (0)

Anonymous Coward | more than 3 years ago | (#35085352)

Just speculating here, but presumably the transistor should be off normally. Over time, it leaks more and more and the current is affecting the operation of the circuit downstream.

Re:More tech detail (1)

karnal (22275) | more than 3 years ago | (#35085898)

The transistor is embedded in the Northbridge somewhere - there's no physical way to turn it off without replacing the chip. They state in the article that it's as simple as turning it off - and I've seen others tripped up by this (in comments on the site) but replacement is still required.

I'll take one! (4, Insightful)

Dr. Spork (142693) | more than 3 years ago | (#35083886)

Wouldn't it make more sense to offer some sort of a substantial rebate and a correctly functioning SATA raid PCI-E card? Some of these motherboards - that are clearly getting scrapped - were very fancy. This seems like a terrible waste, since those boards basically worked.

Re:I'll take one! (0)

Anonymous Coward | more than 3 years ago | (#35083942)

I sincerely doubt they'll scrap them. All that is really needed is to desolder the main chipset chip and then resolder a fixed chip in. A fairly trivial fix that will probably cost a lot less than the cost of manufacturing a whole new board. Considering the volume of boards we're talking about (millions), it is probably worth setting up a whole manufacturing lines for replacing these chips.

Re:I'll take one! (1)

icebike (68054) | more than 3 years ago | (#35084054)

Fix the early returns as replacements for the later ones. Warranty repairs do not need to be new parts, just warranted parts.

But De-soldering is a messy task unless they set up a custom jig for each board type.

After you take out all the pluggable memory, CPUs, and video controllers, what is left is not that expensive. It might be less expensive to trash the board, file a claim with Intel and make new board.

Re:I'll take one! (1)

Anonymous Coward | more than 3 years ago | (#35084296)

It is ALWAYS cheaper to fix the board! Especially since this is just the chipset. You desolder the chipset, put in the new and and resolder it. There are tools to do this job efficiently, even for ballgrid chips. A few minutes to fix a board is not that long.

These boards will be produced for long period of time so a few dozen people just fixing old ones is all that is needed, and it saves the company millions (you still get maximum money from Intel).

Re:I'll take one! (2)

sexconker (1179573) | more than 3 years ago | (#35085002)

These boards will be produced for long period of time

So the 5 seconds between Intel's current socket and Intel's new, 100% incompatible, exactly-the-same socket is a "long period of time" now?

Re:I'll take one! (1)

X0563511 (793323) | more than 3 years ago | (#35085148)

What kind of chip is it?

You have fun desoldering a BGA in the middle of a bunch of other SMD/SMT components.

Re:I'll take one! (1)

Fallingwater (1465567) | more than 3 years ago | (#35085486)

Can't they just put the boards in the ovens and simply lift the damaged chip off?

Re:I'll take one! (1)

drinkypoo (153816) | more than 3 years ago | (#35083946)

It's not inconceivable that they will disable those ports, perhaps with software and epoxy, and sell them again to some other market...

Re:I'll take one! (1)

Dr. Spork (142693) | more than 3 years ago | (#35085238)

Interesting. If the price were right, I might be interested. But if they were going to do that, it would have saved them lots of trouble to offer this to existing owners along with a large refund. I mean, if you bought a Sandy Bridge chip, you don't want to wait until April to have a board to plug it into, right?

Re:I'll take one! (5, Insightful)

gstrickler (920733) | more than 3 years ago | (#35083968)

These motherboards will not be scrapped. The manufacturers have the tools and facilities to remove the defective chips and replace them. The defective chips may be scrapped, but the boards will be refurbished and used as replacement units.

Re:I'll take one! (1)

TheGratefulNet (143330) | more than 3 years ago | (#35084566)

are you sure?

it seems quite uneconomical to repair boards like this that cost in the $100 range.

a multi thousand dollar cisco router board, yes, you'd rework that.

a $100 commodity mobo? I doubt it. not only do you have to unsolder the chip, clean the board and resolder it but you have to fully test it again, too. does that seem worthwhile to you, for a $100 board? I don't know their success rate but I could see a number of the reworked boards failing for this or other reasons and so its a double loss.

when you start out with a fresh clean factory board, you can make them fast and know they are right. when you rework, unless its fully automated (is it?) - its going to be error prone and expensive.

boards are replaced whole. almost no one works at the board level anymore.

Re:I'll take one! (3, Insightful)

gstrickler (920733) | more than 3 years ago | (#35084976)

First, you don't scrap a $100 motherboard over the labor cost of removing and replacing one component (which Intel is providing for free). I'll cost at most $10 to replace the chipset, probably less than $5. Add another $5-$10 for testing and packaging, and for under $20 cost to the manufacturer you have motherboard that you can sell for a whole lot more than $20.

Second, these boards typically start upwards of $100 and go up to $300

Third, at the very least, they'll cut the traces going to the 3Gb SATA ports and/or remove those ports, re-label the board with different model number, put on an updated BIOS that disables the 3Gb SATA, and sell them to a secondary market with only the 6Gb SATA ports active. These could be sold in markets were prices are lower, sold to a small clone vendor building cheap systems, or sold retail as "refurbished" at a discounted price.

Anyway you look at it, most of these boards are not simply going to be scrapped.

Re:I'll take one! (1)

Archangel Michael (180766) | more than 3 years ago | (#35085278)

Labor. Obviously time is worth nothing to you. Stop wasting ours. Cost of the repairs will also include shipping, handling, accounting ..... you know ... overhead.

Grind the stuff up, and recycle.

Re:I'll take one! (1)

gstrickler (920733) | more than 3 years ago | (#35085644)

I'll take at most 10 mins with the right equipment. The people getting paid to do that aren't getting $60/hr, they might be getting $20-$30/hr IF they're in the US. So, how much labor cost is that? In all likelihood, these boards will be boxed up, shipped back to China repaired by people making a whole lot less than that.

The costs I cited include shipping, labor, packaging, etc. Even if I'm way low on the costs, and it costs $50 per board including shipping, packaging, labor, etc. that's still 1/3 to 1/6 what those boards sell for.

I know this is the internet, and slashdot isn't what it used to be, so this is asking a lot, but don't post about things you clearly know nothing about.

Re:I'll take one! (0)

Anonymous Coward | more than 3 years ago | (#35085386)

I'll cost at most $10 to replace the chipset, probably less than $5

Wow, you work cheap - Do you do housework?

Re:I'll take one! (1)

gstrickler (920733) | more than 3 years ago | (#35085650)

See my reply to the other commenter.

Re:I'll take one! (0)

Anonymous Coward | more than 3 years ago | (#35085582)

Everything sounds cheap if you're only counting cost of materials.

Re:I'll take one! (1)

gstrickler (920733) | more than 3 years ago | (#35085754)

Comprehension fail! There is labor and material costs cited in my post.

Re:I'll take one! (2)

multipartmixed (163409) | more than 3 years ago | (#35085008)

I don't know - if you've got, say, a million boards in the pipeline and you can fix them for $20 each then sell them for $60, it seems like there would be some money to be made.

"Fix for $20" might seem unreasonable, but I don't think so, based on the economies of scale available to Asus and Gigabyte. They already have 100% of the requisite QA/QC ability, and access to cheap labour. I'm sure Intel will give them the chips for free.

That said, the other poster's idea is more likely IMHO - re-badge the boards, rip off the defective ports, and sell them as-is.

Re:I'll take one! (3, Insightful)

Rudeboy777 (214749) | more than 3 years ago | (#35085062)

Perhaps, but many of these are not $100 motherboards. High-end P67 boards run in the $200-300 range.

This also would not be a small scale refurb operation -- thousands of identical boards could be processed in an assembly-line fashion making this much more cost-effective than a single worker refurbishing whatever came in the mail that day.

Re:I'll take one! (0)

Anonymous Coward | more than 3 years ago | (#35085326)

Its actually not that difficult. They just have to remove the BGA and put on a new one. The new one will be put on by a pick and place since no one hand solders BGAs of that size. Removing the BGA isn't that difficult and labor is cheap in Asia.

Re:I'll take one! (2)

Dr. Spork (142693) | more than 3 years ago | (#35085276)

Can they really desolder the chipset and solder on a new one, and do all this with acceptable reliability? These things are at least seven layer circuit boards. Do all the chipset pins even go through all the layers to make desoldering possible?

Re:I'll take one! (0)

Anonymous Coward | more than 3 years ago | (#35085488)

Modern ICs are surface mount, and have no pins that go through the PCB at all. This is most likely a BGA.

Re:I'll take one! (1)

gstrickler (920733) | more than 3 years ago | (#35085686)

Yes they can. That's how almost every PCB made is repaired. Most components today are surface mount, the pins don't go through any layers of the board.

Re:I'll take one! (1)

jimmydevice (699057) | more than 3 years ago | (#35085760)

Pin through hole died in the late 90's. Everything is surface mount.

Re:I'll take one! (1)

jimmydevice (699057) | more than 3 years ago | (#35085794)

Doh! except some discreets, like caps.

Intel is eating the cost (1, Informative)

Sycraft-fu (314770) | more than 3 years ago | (#35084046)

The mobo manufacturers don't really care. Intel has done this recall and is eating all the costs related to this problem. So if someone sends back a mobo, the company just sends the bill to Intel and gets their money.

For people that wish to keep their boards and just get a SATA controller (or for who use 1 or 2 ports since the 6gbps ports aren't affected) they can simply do so.

Re:I'll take one! (0)

Anonymous Coward | more than 3 years ago | (#35084082)

I think it could lead to warranty issues potentially or maybe Intel said recall them all, end of story. If I was affected I would want a new board.

Re:I'll take one! (3)

mewsenews (251487) | more than 3 years ago | (#35084092)

Handing out SATA cards would certainly be cheaper, but respectable companies repair or replace a product with defects.

Re:I'll take one! (0)

Anonymous Coward | more than 3 years ago | (#35084172)

Respectable companies would give the option. Would i rather use one of the PCI slots that are empty and get a 20 dollar bill or disassemble my entire computer and ship something off? Best case they cross ship and send me a box ready to go, but that still means tearing apart my computer and taking a box to be shipped. Id rather pop in a PCI card.

Re:I'll take one! (1)

Anonymous Coward | more than 3 years ago | (#35084342)

Unlike *Cough* Apple *Cough*

Re:I'll take one! (1)

francium de neobie (590783) | more than 3 years ago | (#35084158)

An add-on card means one more component to fail, one more component to find drivers for, more stuff inside the case, more power usage, and probably worse performance as well.

No thanks. I'd take the replacement in April.

Re:I'll take one! (2)

Lord Byron II (671689) | more than 3 years ago | (#35084420)

Considering that AMD's Bulldozer is coming out in May (most likely), it might make more sense, if you haven't already bought Sandy Bridge, to wait and see how AMD's chips stack up.

Re:I'll take one! (2)

Rockoon (1252108) | more than 3 years ago | (#35085366)

The 8 core bulldozer is rumored to benchmark 1.5x faster than an i7-950 (as well as 1.5x the similarly performing Phenom II x6 1100T) .. this info was supposedly leaked a few days ago.

Could be a fake, but I wouldn't be surprised given the PassMark numbers for the i7-950 (6346 @ $286), the Phenom II x6 1100T (6174 @ $260), and the i7-980X (10472 @ $1000)

AMD is moving to the same process size that Intel has been enjoying, and a smaller process size usually means large performance increments (just as it did with Intel), AMD has been working on this one for a very long time and they havent redesigned their architecture since the athlon 64's, so I think there is a good chance that the rumors or more or less true. AMD will likely be offering a single chip that can attain 9000..10000 on PassMark, territory that Intel had an exclusive to on single-CPU systems this past year.

The most important thing, tho, is price. If AMD drops these off at $300 like I suspect then Intel will have to seriously rethink its pricing strategy at the high end, but if AMD can't push them out for under $500 then Intel probably has nothing to worry about with its Sandy's pushing 9000+ PassMark's for less.

Re:I'll take one! (1)

LWATCDR (28044) | more than 3 years ago | (#35084174)

Some people will not want that solution and others will shout about Intel trying to cheap out. With this many products it is just simpler to deal with one solution for everyone. Hey if you don't want to send it in I doubt that they will send the motherboard police to your home and make you.

Re:I'll take one! (2)

Fallen Kell (165468) | more than 3 years ago | (#35084192)

Because there is a severe limit on PCI-e lanes on these chipsets. As it already stands, you are limited to two 8xPCI-e slots in use on the basic P67 chipset. On the H67 if you enable the internal CPU graphics, you are limited to one 8xPCI-e slot. That isn't much room left for a SATA controller card if someone plans on having a sound card or HD video capture card... This is also why on the higher end motherboards, they are including an additional bridge chip to expand the PCI-e lanes.

amd is much better at pci-e (2)

Joe The Dragon (967727) | more than 3 years ago | (#35084272)

Intels chips have limited pci-e lanes and you need to go to the high end i7 cpus just to get more then 20+DMI bus speed of 4 pcie lanes. For Sandy Bridge that may be a $400 cpu! + a high cost MB.

AMD lets you USE ANY CPU in a AM3 board with chip set choice with better pci-e lane setups. 890FX has 38 + 4 SB link. 890GX and other 800's 22 + 4 SB link.

790fx has 38 + 4 SB link lanes. 790X and 790GX has 22 + 4 SB link. 785G 20 + 4 SB link. 785E 22 + 4 SB link and most of the other 700 ones have the same.

Re:amd is much better at pci-e (-1)

Anonymous Coward | more than 3 years ago | (#35084950)

amd has shitty manufacturing tho. cpus burn out northbridges and southbridges, mobos cant handle cpu tdp under load etc.
for a quality product intel is the only game in town.

Re:amd is much better at pci-e (1)

h4rr4r (612664) | more than 3 years ago | (#35085244)

You will need to make some sort of citations for such an outrageous claim. As someone who tends to go back and forth between the two I have never seen said issues happen more to one than the other. With the exception of intel p4 boards, which burned up northbridges like crazy, but those boards also had bad caps and I imagine that was the real cause.

Re:amd is much better at pci-e (0)

Anonymous Coward | more than 3 years ago | (#35085782)

As someone who tends to go back and forth between the two I have never seen said issues happen more to one than the other.

Yeah, and you're a a whole fucking statistical universe, you are.

Re:I'll take one! (0)

Anonymous Coward | more than 3 years ago | (#35084874)

Not exactly. Those 16 PCIe lanes you are talking about have nothing to do with the P67 chipset. They are in the CPU itself, and this was first introduced in the 1156 socket CPUs. Before then all PCIe traffic went through the chipset.

In addition to the 16 PCIe lanes w/ direct CPU access (this is basically dedicated for graphics) there are an additional 8 PCIe lanes in the P67 chipset which is plenty to deal w/ any Sata cards, etc.

Some of the more expensive boards w/ several x16 slots do use bridge chips to expand the 16 CPU PCIe lanes to 32 lanes, but there is still only 16 lanes into the CPU, so I have little faith in these band-aid solutions. They add an additional layer, more complexity, and more latency. I didn't like the older nVidia chipsets (680i, 780i) as they were very buggy, and so I have little faith in their add-on NF200 chipset boards. Hydra hasn't been anything it was hyped to be either.

It has been shown that 16 PCIe2.0 lanes is plenty for 2 top-tier graphics cards in SLI/X-fire. If you want more than that then move onto a 1366 system or wait for the new 2011 socket later this year. Personally, I will always stick w/ 1 graphics card. SLI/X-fire in and of itself is buggy and a PITA.

Re:I'll take one! (1)

h4rr4r (612664) | more than 3 years ago | (#35085270)

Intel is coming out with yet another socket so soon?
Holy hell, no wonder I don't buy them that often.

Re:I'll take one! (1)

commodore64_love (1445365) | more than 3 years ago | (#35084276)

>>>This seems like a terrible waste

Not all will be scrapped. In the above article: "If a design only uses ports 0 & 1 off the chipset (the unaffected ports), then the end user would never encounter an issue and may not even be recalled."

Asus didn't handle the capacitor plague so well (2)

bzipitidoo (647217) | more than 3 years ago | (#35084870)

Nobody did, that I know of.

An Asus motherboard I bought many years ago worked fine for about 4 years, then over a period of 2 weeks started spontaneously rebooting or locking up at ever shorter intervals until the computer was unusable. (As I recall, it was a P4S333. I put a 1.7 MHz Pentium 4 in it.) I saw some capacitors had leaked on the motherboard.

Thought I'd see what Asus had to say about it, perhaps offer a replacement or upgrade at a prorated discount, or something. They told me to get lost. Motherboard was much too old to be in warranty. Businesses feel that they are excused the minute the magic warranty period expires, despite cases like this that are clearly the result of a flawed product. I didn't really expect any better, and to be fair the board was too obsolete to be worth messing with, but still, it was annoying.

Friend of mine who is good with a soldering pencil replaced the capacitors for me, and the board worked fine again. Still works fine today. Tried replacing capacitors on other afflicted boards. Sometimes that worked, and sometimes not.

Unexpected mishap (1)

korgitser (1809018) | more than 3 years ago | (#35083902)

I thought going for cougars was a safe bet.

Re:Unexpected mishap (1)

Deflatamouse (37653) | more than 3 years ago | (#35084368)

These cougar's ports wear out too easily

Intel is getting ahead of this one (1)

rickb928 (945187) | more than 3 years ago | (#35083924)

You will read of the details of this elsewhere, but I 'know a guy' at Intel, and this was slam-dunk gotta-fix-this for them, despite the cost. It was evident from the beginning that this had the makings of a legendary fail for them, and they bit off the $1B and just fixed it.

I'm hoping to get some tidbits on the actual cause, but for now it's pretty tight over there.

Not often that Intel makes these mistakes, and this is one they seem to be handling with integrity. Not like Nvidia [] .

Re:Intel is getting ahead of this one (1)

icebike (68054) | more than 3 years ago | (#35084122)

Quote from AnandTech Story []

Intel maintains that Sandy Bridge CPUs are not affected, and current users are highly unlikely to encounter the issue even under heavy loads. So far Intel has only been able to document the issue after running extended testing at high temperatures (in a thermal chamber) and voltages.

So, no, it wasn't evident from the beginning. They had to devise torture tests to even see it.
See Anandtech (above) for the desired tidbits you were hoping for.

Re:Intel is getting ahead of this one (1)

rickb928 (945187) | more than 3 years ago | (#35084740)

Actually, Intel got returns back from their customers (Dell, for one), and as soon as they analyzed the chipsets, they found the materials problem. The 'torture tests' were in response to the defective parts and the failure mode. That 'only' gets them to choosing the next step, is it hardware or firmware? Assuming,of course, that it is a defect, and not something else. My friend in driver development spends a fair amount of his time proving that the error is not his code. Less than a while ago, since he moved from Windows to Linux driver development. He's a lot less stressed for some reason.

From the beginning was intended to refer to 'from the beginning of Intel's analysis'. I don't know what beginning you were referring to, but at least Intel had to have a failed part to do any analysis at all. Unless you expect them to be clairvoiant. My friend assures me he is not, he's just a scientist.

I saw the Anandtech article, and others.

Re:Intel is getting ahead of this one (0)

icebike (68054) | more than 3 years ago | (#35084840)

"From the beginning" were your words, not mine. The fact that they had to bake it in an oven with higher than nominal voltages to induce it seems to suggest it was hard to reproduce in their testings.

And yes, since Intel made the chipset as well as the CPU, I kind of DID expect them to be clairvoyant about this since the problem was known. This particular part was slipped in AFTER the first rounds of tests, it wasn't in the prototypes according to the article.

Re:Intel is getting ahead of this one (1)

rickb928 (945187) | more than 3 years ago | (#35085316)

Intel so far has publically stated that this problem wasn't detected in their testing. I kinda believe that, and the root cause makes sense of that. The chip wasn't slipped in after the LAST round of tests, so the assumptions that Intel didn't test this design adequately will probably prove false. I haven't heard any credible report that Intel had any suspicion that this problem existed, or could, before shipment. Changed seppings may not have spotted it, for as you pint out, this needed extreme testing regimes to be revealed. Prototypes can't be relied upon as production examples for obvious reasons.

This would not be the first time Intel has had to retract a part because of materials problems. Just the first time in 2011.

And one item caught my attention. Intel seems to think that this would affect at least 25% of these parts over a three-year period, and lead to substantially shorter useful life for nearly all of them. In fact, I think they predicted that the vast majority of parts would show no failures for 1-2 years.

And another item. Anandtech reported that an Intel spokesman said that the fix was to remask and essentially diable the offending transistor. That doesn't square with what I know. But it is common in these cases for one team to blame another, so it coudl be a design team is being lamed by manufacturing for driving a device to failure, where design might be pointing out that the device wasn't manufactured to spec, and is failing because of that. Often, both teams get a trip to the woodshed, since resolving these problems has to be a team approach, and having teams covering their butts first can lead to delays and costs.

There are teams within Intel that can do their investigation without interference from the CYA gangs. Fortunately for us.

How do you tell the difference... (0)

Anonymous Coward | more than 3 years ago | (#35083936)

...between a normal Gigabyte board and a defective one? Thought they were all defective.

Re:How do you tell the difference... (1)

moeluv (1785142) | more than 3 years ago | (#35084310)

Gigabyte made the list of MB manufacturers I pass over, them and ASRock will never find their way into another build I do. I gave up on Gigabyte after trying 2 boards in succession and having the same boot issues with both.

Why the recall...? (1)

Wierdy1024 (902573) | more than 3 years ago | (#35084244)

It appears the problem only occurs with some chips, after some years, and where the issue occurs, it will only affect some of the SATA ports.

To me, it sounds like the best course of action from Intels point of view would to be to replace any failed chip when the user complains. The majority of users will never come across the issue, since most users won't have 3 or more SATA devices, and of those users that do, many will probably never get the problem, or if they do it'll be after the warranted period.

If I were Intel, I definitely wouldn't be recalling any chips that were already soldered onto a board without a direct user complaint. It might be fair enough to recall chips that aren't yet on a PCB, since then the cost is much lower, although even then I would just recall them, and then sell them again as slightly cheaper 2 port versions.

The refund procedure could be entirely Intel handled where all returned boards get dumped directly into "recycling", and the user sent a check for $100 (or more if they include a receipt for the board).

Since this ISN'T what Intel is doing, it makes me suspect there is something else more serious wrong with them...

Re:Why the recall...? (0)

Anonymous Coward | more than 3 years ago | (#35084370)

Remember the Pentium FP bug? Intel does, and they don't need a repeat of that on their record. It was impossible to keep that bug a secret. It is far more difficult to keep this type of problem today. Imagine a "wikileaks" type site getting ahold of internal documents about this chipset problem if Intel hadn't said anything...

Re:Why the recall...? (0)

Anonymous Coward | more than 3 years ago | (#35084558)

"...this type of problem SECRET today."

Re:Why the recall...? (2)

klashn (1323433) | more than 3 years ago | (#35084438)

This was the policy for the FDIV bug. People found out that Intel would only replace it if they were using the CPU that uses the floating point unit extensively. Those users that didn't use it extensively still claimed that they would use it, and thats why there was a big fiasco. This time around... Fix it, get it over with...

Re:Why the recall...? (0)

Anonymous Coward | more than 3 years ago | (#35084794)

So, because it's not what you'd do, there must be some bigger conspiracy ?
I take it you've never run a business?
If you have, please tell me who it is so I can make sure I'll never do business with you.
From your attitude, it's probably Dell ;)

Since Slashdot loves car analogies, what you're recommending is the same as Ford knowing that their steering racks sometimes lock up, but only replacing them out once the customer has crashed.
Do you think that's acceptable?

What Intel are doing is a thing called 'Being responsible'. Not just to their customers, but to their shareholders.
Sure, it's a big hit they'll take now, but it'll harm their reputation and business a damn sight less in the long run.

This is a good thing. (2)

Even on Slashdot FOE (1870208) | more than 3 years ago | (#35084288)

If more companies replaced their defective products at their own expense, this would be a better world. And people would be more focused on making things work before they ship the product.

Re:This is a good thing. (3, Insightful)

Shados (741919) | more than 3 years ago | (#35084452)

The problem is in the current world, a lot of companies go under if they do that.

So you have:

A) the companies that screw you over
B)The companies that don't...oh wait, those went under.

Ok, so you have A) the companies that screw you over. Thats it.

Take the news with a grain of sand (1)

JoeThoughtful (1945502) | more than 3 years ago | (#35084294)

Any engineer knows that there is going to be trouble when you make a bridge out of sand. It seems like another case of that mystical name/fate thing - where a name carries its own karma.

Re:Take the news with a grain of sand (1)

treeves (963993) | more than 3 years ago | (#35085338)

Interesting point, but I suspect it refers to a bridge on the Sandy River, a river near Portland, OR, which is (fairly) near Intel's R&D fabs.
They have a habit of naming stuff after rivers (and other places, often small towns), e.g. Nehalem, Tualatin, Clackamas, Willamette.

Now what about dell, hp, and others with bad chips (1)

Joe The Dragon (967727) | more than 3 years ago | (#35084298)

Now what about dell, hp, and others with bad chipsets? will they do a swap? make you pay to ship the system in? make you pay for a tech to swap a board? Let you swap the board on your own? Not even offer a free swap?

Re:Now what about dell, hp, and others with bad ch (1)

billcopc (196330) | more than 3 years ago | (#35084510)

I'm assuming Dell and HP will handle it through their usual warranty process, i.e. when customers call to complain their PC has died. Alternately, they might still be working on an official announcement of their own. A recall isn't something you launch on a whim.

all? (0)

Anonymous Coward | more than 3 years ago | (#35084358)

Is it actually "all" sandy bridge motherboards? Intel is only recalling certain boards for sandy bridge, at least. Some sandy bridge boards are unaffected.

I once dated a stripper named Sandy Bridge. (0)

Anonymous Coward | more than 3 years ago | (#35084444)

The way the arched her back was a sight to be seen. Many times I traversed her causeway to get to the other side of that gushing river.

What? (1)

immortalpob (847008) | more than 3 years ago | (#35084458)

The Gigabye press release is here [] and it says nothing like what this article says, there is no mention at all of returns. Q: What action should I take if I have already purchased a GIGABYTE 6 series motherboard? A: Firstly, please determine your computer setup. No action will be needed if you only use the SATA 3 ports. If you are using the SATA2 ports, then there are possibilities that the device’s performance will decrease after a period of usage. To ensure the highest standard of customer support & services, GIGABYTE recommends that all customers who purchased GIGABYTE 6 series motherboards contact their local dealer (retail store where you purchased the motherboard) at the end of April for a motherboard exchange. GIGABYTE will provide an equivalent new motherboard replacement.

Re:What? (1)

multipartmixed (163409) | more than 3 years ago | (#35085292)

Actually... this sounds pretty much like what I'd expect.

What do you think doesn't jive?

Thank you. (1)

unity100 (970058) | more than 3 years ago | (#35084642)

i'd rather not buy anything intel again. there has also been a time when they shipped defective batches in some q28345857something line of cpus too, for one of which one of my enthusiast friends paid approx $300 for a single cpu.

What About Intel's Own Motherboards? (1, Informative)

BaldingByMicrosoft (585534) | more than 3 years ago | (#35084644)

Here's my experience yesterday with Intel on their support chat. Not happy.


info: Please wait for a site operator to respond.
info: You are now chatting with 'Diego'
Diego: Hello. Thank you for using the Intel Customer Support chat service. We are glad to be of service. How can I help you today? Hi. I've read about your recent chipset issue. I just purchased and received a DP67BA motherboard. Is there a process to have it replaced with a corrected chipset?
Diego: In this situation, each place of purchase will be in charge of either replacing the motherboard.
Diego: Since the issue is very recent, there are no replacement units with the fix. Bear in mind this issue impacts all Intel® 6 Series Express Chipsets and Intel® Xeon® C200 Series chipsets on systems using SATA ports 2-5, if using ports 0-1, customers are not affected. Using ports 2-5 may impact functional issue rates over time, this is not impacted immediately. So, I need to contact the sales vendor regarding this issue, even though it's an Intel motherboard?
Diego: The issue is being under investigation by Intel and we are working as fast as possible concerning this
Diego: Please be aware that 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. Systems with only SATA ports 0 and 1 enabled are not susceptible to these functional issues.
Diego: It is not a problem that will be present in a short time Yes. I understand. This is a part I bought for a personal computer. I don't have a lot of money. I need the part to last properly for a long time. Are you telling me that I have to contact the sales vendor regarding this? I bought an Intel brand motherboard because I thought I would receive good support.
Diego: You may check with the place of purchase in case you would like to replace the motherboard in the future when a hardware fix is available on a new revision. We are working together with our chain or Authorized Distributors and resellers to cover this problem Okay then. I will forward this conversation to my sales vendor and see what they say about it. Thanks.
Diego: You are welcome
Diego: Is there something else I would be able to assist you with? No. Goodbye.

Re:What About Intel's Own Motherboards? (2)

base3 (539820) | more than 3 years ago | (#35085078)

AMD it is for the next build. Thanks for the info.

Re:What About Intel's Own Motherboards? (4, Insightful)

FrankSchwab (675585) | more than 3 years ago | (#35085274)

Frankly, I think that sounds like an excellent response from Intel. Did you expect them to send a car right over with a replacement Mobo?

They just announced the problem. They don't have 8 million replacement chips, or 8 million replacement motherboards in house. If they waited until they did have that many, somebody would bitch about them delaying the announcement.

The rep answered your questions truthfully, told you that the exchanges would be handled through the retailer, and suggested that your board will probably work just fine until a replacement is available. Again, what more did you expect?

Angry? Sure you can be angry. Someone sold you something that's defective, and that's causing you problems. They can't immediately make it right, and that's going to cause you more problems. Suck it up - it happens everywhere, all the time, and is part of life.

Re:What About Intel's Own Motherboards? (0)

Anonymous Coward | more than 3 years ago | (#35085422)

What's wrong with going through the vendors? They actually have the infrastructure to do RMAs for individual customers while Intel does not. From what other comments are saying, it seems Intel will be crediting the vendors, so they are still paying for the process. It looks like in your conversation, Diego gave you the answer after your first question, then you just got angry. How does this dialog show any wrongdoing on Intel's part?

Oh stop looking for reason to be angry (4, Insightful)

Sycraft-fu (314770) | more than 3 years ago | (#35085472)

I swear, some people are just looking for reasons to be pissed off.

So, Intel has offered you the ability to either get a refund now, or a replacement board later. What's more, to get your refund you don't have to deal with them, you deal with the people that sold you the board, Intel will reimburse them. Like if it is Newegg just contact them and they'll issue you a full refund RMA, no questions asked, for any product that contained a 67 chipset, mobo, laptop, etc.

You cannot be provided with a replacement right now, because there are none to be had, they are being made. However if you'd rather not get a refund and use the system as it is, that's fine you can do that and then get a replacement in a month or two when they are out on the market.

Seems to me as though they've done everything they can to rectify the situation. They are fixing the problem and everyone gets a free replacement when the fixed units are out. If you are unwilling or unable to wait, then you can send back the stuff for a full refund right now.

If this doesn't satisfy you I see only one of two situations:

1) You want a fixed chip right now. That means you are an idiot, expecting you can have something before it is physically possible.

2) You want Intel to issue you a refund directly, rather than the retailer because you feel that is a better "punishment" or something. In that case you are just being unreasonable. You can get your refund, just talk to the place that sold it. Hell you'll get it faster that way.

Seriously, I fail to see the big deal. There is a fuckup, they are doing what they can to fix it. Nothing else can be done that I can see.

Re:What About Intel's Own Motherboards? (2)

batkiwi (137781) | more than 3 years ago | (#35085554)

So is your vendor refusing to support you? If not, what's your problem? If your car has a recall, do you take it to the dealer (who is likely his own business), or call up corporate HQ and demand they ship you the new part?

They are using vendors as the channel of distribution. There's nothing wrong with that, and genrally ensures everyone gets served more quickly.

Newegg will refund, too (1)

klui (457783) | more than 3 years ago | (#35084698)

Newegg will also refund. []

Re:Newegg will refund, too (1)

Dr. Spork (142693) | more than 3 years ago | (#35085460)

Very informative video. If I understood it right, they are happy to let you keep your defective motherboard - which still has two SATA-3 ports that work perfectly well - and use it until April, at which point they will swap it out with a new one. If this is the deal, then that's actually very customer-friendly. From the blurb, I had the impression that if you want to participate in this rebate, you have to send back your mobo now, and your fancy Sandy Bridge CPU will just sit unused on your desk for almost three months, which in computer time is a good fraction of forever.

Verifying that you have a repaired board (1)

Anonymous Coward | more than 3 years ago | (#35084716)

How does Intel plan on representing chipsets that have been repaired/fixed? End-users are going to need a simple utility to verify this, as will us UNIX system administrators. :-)

I imagine that they could simply increment the PCI device ID revision attribute of the controller (from rev 0x00 to rev 0x01) or something similar, but they haven't disclosed how they're going to differentiate between "old/broken" and "new/working" chips.

The only thing I can find is a "stepping number" that's associated with the chipset (and that's based on [] ), but there isn't a way to obtain that information directly from the PCI bus; you have to already be speaking directly to the controller via PCI commands to get that. So, I sure hope Intel plans on disclosing how to verify that your board has a repaired chip.

Cougar Point, not Sandy Bridge (4, Informative)

pablodiazgutierrez (756813) | more than 3 years ago | (#35084742)

Cougar Point is a chipset (the set of circuits that normally come in a motherboard, separate from the CPU). Sandy Bridge is a family of processors. The announced problem is with the former, not the latter. A lot of tech news outlets are spreading the misinformation and causing quite a mess.

Re:Cougar Point, not Sandy Bridge (1)

PitaBred (632671) | more than 3 years ago | (#35085138)

Cougar Point only works with Sandy Bridge chips, and Sandy Bridge is a much more publicized name, so that's what the news outlets are running with. It may not be totally correct, but it's not totally incorrect either. Many people with Sandy Bridge CPUs don't realize they also have a Cougar Point chipset.

excellent! (1)

khraz (979373) | more than 3 years ago | (#35084846)

I'm hoping for a price drop on Sandy Bridge processors during this time, since there'll be no boards to put them in...

I've mostly bought AMD over the years but... (2)

Ironhandx (1762146) | more than 3 years ago | (#35085086)

I really have mostly supported AMD over the years. A lot times it comes down to the fact that they generally have the best price/performance in my/others budget range(most computers I build are in the $1000-1500 range) when I'm building computers and I also have a certain comfort level with them in that I've scrapped a lot fewer AMD cpus than Intel ones.

However, I have to say. I'm really impressed with how Intel is handling this. There must be a nice bit of extra support for board vendors as well, especially with the huge loss numbers they're predicting and how good the board manufacturers are being with this situation.

This could have been a really bad PR event. Instead I think I might be buying Intel when its time to upgrade again in 6-12 months.

Re:I've mostly bought AMD over the years but... (1)

h4rr4r (612664) | more than 3 years ago | (#35085304)

So that you too can experience broken parts?
To me this is them doing the least they could, the right thing would be a refund since you can't use any boards right now and a new board when they are available. This makes the CPU you bought from them worthless, so just offering a replacement in 2 months is nuts.

Re:I've mostly bought AMD over the years but... (1)

oracleguy01 (1381327) | more than 3 years ago | (#35085714)

If you read the article, it is an issue that isn't apparent right away (Intel had to use a extraordinary testing conditions to cause it to happen) and when it does, the 2 SATA3 ports are unaffected. From the Anandtech article:

So far Intel has only been able to document the issue after running extended testing at high temperatures (in a thermal chamber) and voltages. My recommendation is to try to only use ports 0 & 1 (the 6Gbps ports) on your 6-series motherboard until you get a replacement in place.

So it isn't like your computer is unusable, most people will never experience the issue between now and April when replacement boards are available. So the CPU you bought is hardly worthless.

Seems to me that they are handling it well enough, clearly they are working with their vendors to make the process easy. NewEgg will either give you a refund now or a replacement board in April, no questions asked.

glad i don't build or buy intel stuff (0)

Anonymous Coward | more than 3 years ago | (#35085088)

I'd have a lot of un happy gaming customers.

J. Dzhugashvili (0)

Anonymous Coward | more than 3 years ago | (#35085346)

Interesting, does J. Dzhugashvili stand for Joseph Dzhugashvili?

Yay for refunds! (0)

Anonymous Coward | more than 3 years ago | (#35085750)

I'm glad gigabyte is suggesting refunds. hopefully they make a new revision that will equal the competition's features. (i.e. the P67A-UD4 is rather lackluster compared to the P8P67 Pro).

but I wonder how many smaller stores will want to do refunds? especially if you wait until April, it would be way past the 30-day window. I bought mine at Microcenter, so I'm curious to see how they will handle this.

Load More Comments
Slashdot Login

Need an Account?

Forgot your password?