When Thermal Throttling Goes Bad...

[Cricket]

When thermal throttling fails, expect smoke

Ouch.

Cricket

[blue60007]

Yikes!

[kram 2.0]

Ack! And the system still ran?

kram

[786ARS]

gets the sypathy bit from me though, (cringes at pic)

[freakitchen]

I'm guessing the system still ran because the pins (albeit charred!) were still making contact inside the socket. Wrenching it out was obviously the nail in the coffin.

Obviously watercooling is not without its risks, like anything else!

FK

[Panama Red]

My dual core 950 is running the same mobo and this makes me glad I'm using the stock hsf. I assume (bad word!) mine would shut down with a fan failure unlike a pump failure on water cooling. That poor sod is out a bunch of cash!

[rspassey]

Man, that really sucks. I feel sorry for the guy, even though he voided the warranty and can't expect it to be fixed for free.

[juppy]

Uh....ouch?

[catacon]

Bummer...

[HAL9000]

Realistically, I wonder exactly what happened... water cooled, air cooled, or cooling of any sort having fallen off of the CPU, it should have throttled back to avoid that kind of damage. This is the first time I have ever seen that kind of failure on an Intel chip. I don't think it would be fault of the motherboard as unlike the AMD chips, Intel chips don't require assistance from the BIOS for thermal protection, all protection is internal to the chip.

[rspassey]

Originally Posted by HAL9000 Realistically, I wonder exactly what happened... water cooled, air cooled, or cooling of any sort having fallen off of the CPU, it should have throttled back to avoid that kind of damage. This is the first time I have ever seen that kind of failure on an Intel chip. I don't think it would be fault of the motherboard as unlike the AMD chips, Intel chips don't require assistance from the BIOS for thermal protection, all protection is internal to the chip.

I agree, I have never seen something like this happen before. But, even if it was a faulty chip, he wasn't using the stock heatsink - therefore it isn't under warranty anymore.

[kosova]

1200 smackers out, poor fellow. I wonder if there is a reason why on the mainstream most computers are sold using air cooling? I'm not sure but I think it's because there can be a lot of hassle to water cooling, like a pipe/tube cracking and water spilling all over the motherboard, maybe this is a common belief/myth. At 96*C I thought that processor would pop like popcorn, guess I was wrong. That's Intel, it heats up but it can stand it--for the most part.

[David M]

Ethically, Intel owes him a new CPU. Their chip failed to shut down when it was designed to shut down. Technically, he invalidated his warranty and is not owed anything. Unfortunately he is out a new CPU but for reasons that are unrelated.

Thats just how it goes. If Intel wants to make themselves look good they would send him a replacement now that he has gone "public" and hundreds of people will read this thread. But Intel is such a massive company that he is no more than a bug on the windshield and the bottom line is that this incident will have no effect on Intel's bottom line...therefore, Intel has no financial reason to make good on their faulty CPU. Sometimes it's a cruel world. I feel bad about his circumstances as well.

Its too bad he cannot talk to someone within Intel. The best he can to is talk to someone at "Customer No-service" in Bhopal, India who has no decision making ability and can only go by a script. Getting help in his situation is like hitting a wall...I guess it is analogous to a bug hitting a windshield.

[kosova]

Good points David, one thing I don't understand is, the cpu overheated because it was not getting cooled and it melted, also damaging the motherboard right? Shouldn't Intel also have to deal with that? Their product was a hazardous that damaged something else. BTW I don't mean to offend anyone but I am curious about this legal situation, this could happen to anyone, like myself.

[HAL9000]

It comes down to liability of the owner. The CPU and heatsink are meant to be installed together as a processing/cooling solution, any modification voids warranty. If I buy a new car and toss nitros on it, blow the engine... is the car manufacturer at fault? I don't think so.

[rspassey]

Originally Posted by HAL9000 It comes down to liability of the owner. The CPU and heatsink are meant to be installed together as a processing/cooling solution, any modification voids warranty. If I buy a new car and toss nitros on it, blow the engine... is the car manufacturer at fault? I don't think so.

Good analogy, that basically sums it up.

[David M]

I don't see it that way...yes, nitrous invalidates the warranty on the engine if the engine blows up. But if the bumper falls off the car which has nothing to do with putting nitrous in the engine...should the warranty on the bumper be invalid?

My point is, the sensor that is supposed to shut down the CPU if it overheats failed. Which has nothing to do with why the CPU overheated.

The failsafe mechanism should have worked without regard to how the CPU overheated.

[HAL9000]

Sure it's related.... The CPU WITH the heat sink and fan are a package. To run the one, you must use an APPROVED method of cooling which means you use their cooling recommendations (Which is why OEM CPU's have a whopping 30 day warranty). Water cooling is NOT on that list. So in your analogy, if I take out the bolts holding the bumper on and insist on using bungee cords and it falls off, is the manufacturer at fault? The bumper should have worked without failure, just as the CPU should have worked even though in both scenarios, we have changed a critical, required part.

[blue60007]

Originally Posted by David M I don't see it that way...yes, nitrous invalidates the warranty on the engine if the engine blows up. But if the bumper falls off the car which has nothing to do with putting nitrous in the engine...should the warranty on the bumper be invalid?

You still voided the warranty on the *entire* car, not just one part of it.

[juppy]

Originally Posted by HAL9000 So in your analogy, if I take out the bolts holding the bumper on and insist on using bungee cords and it falls off, is the manufacturer at fault? The bumper should have worked without failure, just as the CPU should have worked even though in both scenarios, we have changed a critical, required part.

I have to agree with David on this, HAL. You're saying you'd take the bumper bolts out....that is changing the bumper. What David said was just if the bumper fell off without any changing, because that's what this chip did.....the shutdown mechanism failure was not due to having a different heatsink on there. The shutdown *should* have worked whether he had the stock hs, an aftermarket hs, water cooling, or NO heatsink whatsoever. That part of the chip is internally connected already from the factory and the heatsink doesn't affect it.....it should still shut the chip down regardless.

[HAL9000]

Wow.. going so far over heads gonna need the hubble to see it (j/k).... I'm NOT denying that the failsafe didn't do it's job, I am saying that by using improper installations, Intel is NOT liable for warranty.

BTW, changing the chip with the mandatory cooling solution, to me is equal to changing the chip as you have changed the design that includes heat dissapation which is a requirement.

[lil Jimmie]

In my biz any mods void the warranty!!!!!! Nut, washer, or blower... sorry your bumper hit the asphalt!

And yes I've had to eat a washer or two which cost more than the computer your sitting at cost.

[HAL9000]

I still remember the cylinder head incident... OUCH

[lil Jimmie]

Dave.. That was a $750.00 USD washer my cost.

but the same customer just brought in his '64 Lincoln to be modded, in which I'm in the middle of installing the airide at the moment.

[David M]

Yes, following Intels rules, they did invalidate the warranty. The fact remains that there is not a direct cause and effect relationship. Water cooling did not cause the thermal shutdown to fail. Any type of failed cooling device could have been used, whether it is factory air or water, and the thermal shutdown still would have failed.

[HAL9000]

I agree with that completely, never tried to argue it... just giving reason as to why Intel has just cause for voiding the warranty.

[juppy]

Originally Posted by lil Jimmie Dave.. That was a $750.00 USD washer my cost.

"Washer", as in the little round thing that goes on a bolt and under a nut, LJ? How the heck can one washer cost $750?! Handcrafted over a 3-year period by a monkey with a spoon and an ice pick, or what? How can anyone justify $750 for a single washer?

[David M]

I'm sure he meant it figuratively juppy.

So after all this..we agree? Intel has justification for invaladating the warranty..and there is no direct cause and effect.

Case closed!

Lock the thread! ....LOL

[Panama Red]

Originally Posted by juppy "Washer", as in the little round thing that goes on a bolt and under a nut, LJ? How the heck can one washer cost $750?! Handcrafted over a 3-year period by a monkey with a spoon and an ice pick, or what? How can anyone justify $750 for a single washer?

I'm willing to bet the "price" of the washer equates to the repairs/replacement parts necessary when the washer ended up in the wrong place - like down the intake manifold! I never did it personally but I've seen the results and it ain't pretty!

[HAL9000]

If I remember correctly, the "washer" was a cylinder head that slid off the forks of a forklift hitting a boat propeller before hitting the floor... I repeat.. OUCH.