Anyone have a dead C2D?

[ishould]

I've been trying to figure out the long term affects of high temperature/volts on a C2D and I'm having a hard time finding any situation where one has died because of it. Does anyone have a dead core 2 duo from running high volts for too long? Something over 1.6 perhaps

 

[FatFunkey]

im not 100% sure if the C2D suffers from electromigration....it probably does (almost all electrical circuits suffer from it..hints why they break down over time) but not on the level of the Northwood (S.N.D.S)

 

[nvalhalla]

I haven't heard of a voltage death of C2D. Of course, I haven't heard much about CPU deaths, other than SNDS. I have 10 year old CPUs that still work fine.

 

[nvalhalla]

Damn, my first double post! I don't know how that happened... I clicked submit, switched to another tab for a different website, came back to this tab a few minutes later and it was still on the reply page, so I clicked it again. BLAM, double post.. weird. Look, they're 7 minutes apart!

 

[MU_Engineer]

im not 100% sure if the C2D suffers from electromigration....it probably does (almost all electrical circuits suffer from it..hints why they break down over time) but not on the level of the Northwood (S.N.D.S)

You're right. All transistors can suffer from electron migration if given too many volts for too long. High voltage can also be bad because the heat production scales quadratically with a voltage increase and only linearly with anything else. Heat will kill ICs, too as it leads to thermal breakdown. There is also self-heating that occurs when a chip is running warm, especially with SOI chips. The substrate's electrical resistivity changes when it gets hot and allows a much greater power draw. That makes more heat, which changes the resistivity more, and allows yet more power... It's a vicious cycle.

I personally don't overclock and in fact run my CPU undervolted to reduce the heat output and allow for slower and quieter fan speeds. But if you want to overclock, you should be safe if you keep the temps in check and don't go more than 7.5% over the highest Vcc_HFM figures as quoted in the Intel Core 2 Duo Reference Guide. Intel sets the Vcore for the top speed (Vcc_HFM) and the lowest SpeedStep speed (Vcc_LFM) on a chip-by-chip basis. There is a range of values and if your chip is on the lower end, you can be guaranteed to have your chip work fine with it set at the maximum Vcore because Intel certifies the chip will work fine at that Vcore. The +7.5% figure is the maximum deviation from the default Vcore that Intel will allow the motherboard to supply to the CPU. So if your PSU and mobo are rock-solid in their voltages, you can use the wiggle room to add more volts. This is an extremely conservative way to come up with a max Vcore for a chip, but you will not suffer from any voltage-related ills by doing this.

 

[ishould]

How prone are C2D's to this cycle of breakdown? Has it happened (or started happening) to anyone's?

 

[jackluo923]

There's a guy who was using a C2D and his WC pump broke down. So in the end he was running the C2D at 100+C and it was throttling down to 1.0Ghz. Now his processor is overheating even at stock speed. I believe that is what you call a vicious cycle.

http://forumz.tomshardware.com/hardware/E6600-slowly-dieing-ftopict242728.html

 

[ishould]

I was wondering more about the long term affects of something running in the 60's/70's. I don't think anyone in their right mind would think they could run in the 100's for very long lol.

 

[ethernalite]

You'd be surprised at how resilient processors are. Everyone freaks out about temperatures in the 60°C range, but it isn't all that ridiculous. Intel's mobile processors are rated to 100°C. Mind you, they don't have IHS like desktop chips do. Certain compounds in the IHS that glue it and the processor die together may experience thermal breakdown. Plus, the mobile processors run at lower voltages, reducing the risk of damage from high voltages compounded exponentially by the high temperatures.

I'm not saying that you should run your desktop chips at 100°C (Intel says the maximum for desktop chips is 60°C, but other than the lack of an IHS, there isn't any real architectural or manufacturing difference between a Core 2 Duo mobile and desktop processor), but I wouldn't worry about temperatures in the 60° range. 70's are getting up there, and I wouldn't recommend it, but you probably will still get a long functional life out of a chip (3+ years) with anything below 80°.

Disclaimers: YMMV. IANACPUE (I Am Not A CPU Engineer)

 

[MU_Engineer]

You'd be surprised at how resilient processors are. Everyone freaks out about temperatures in the 60°C range, but it isn't all that ridiculous. Intel's mobile processors are rated to 100°C. Mind you, they don't have IHS like desktop chips do. Certain compounds in the IHS that glue it and the processor die together may experience thermal breakdown. Plus, the mobile processors run at lower voltages, reducing the risk of damage from high voltages compounded exponentially by the high temperatures.

They also have lower amperage demands as well. A Core 2 Duo E6600 operating at 1.30 volts draws 50 amps at its rated maximum draw of 65 W. A Core 2 Duo T7700, also at 2.40 GHz, has a Vcore of about 1.20 volts but will only draw 30 amps as the wattage dissipation is 35 watts, not 65 watts.

I'm not saying that you should run your desktop chips at 100°C (Intel says the maximum for desktop chips is 60°C, but other than the lack of an IHS, there isn't any real architectural or manufacturing difference between a Core 2 Duo mobile and desktop processor), but I wouldn't worry about temperatures in the 60° range. 70's are getting up there, and I wouldn't recommend it, but you probably will still get a long functional life out of a chip (3+ years) with anything below 80°.

There are slight architectural differences in the transistors IIRC. The ones in the Merom CPUs are tuned more for efficiency and low power dissipation rather than for speed. This lets them draw less power and dissipate less heat. If you look at the Processor Reference Guides from Intel, the Tjunction_max is a line, depending on wattage draw. Note that CPUs with lower wattage demands (and thus lower voltage demands) have a higher Tjunction_max than ones that draw more volts/amps/watts and put off more heat. Just extrapolate those thermal resistance curves out for the draw values for mobile chips and you'll see where the 95-100 C temps can be reached safely.

Disclaimers: YMMV. IANACPUE (I Am Not A CPU Engineer)[/quote]

Neither am I; I'm trained as a biological engineer. I've taken a few circuits-type classes as part of the biomedical instrumentation portion of my degree, but I don't hold a candle to what a real double-E knows, particularly one that specialized in CPU design. Computers are a hobby for me, well, actually they are my job this summer It always pays to be "that guy" when somebody needs some technical help. And in a university, there are often quite a few people that need some coding or scripting work done, or just need somebody to walk them through things.

 

[ethernalite]

Yea, that's what I was trying to say with the voltages (which means higher wattage/amperage) combined with the higher thermals. You clearly explained it better than I.

I was just trying to make the point that the processor will not go up into smoke and instantly die because of (relatively) high temperatures, but rather simply reduce the lifespan (albeit exponentially) due to the regular causes of CPU death.


edit: and just to reiterate, I do not recommend temps above 70°, for the reason MU_Engineer said. The only ancedotal experience I've had with very high temps is a Prescott I used in a HTPC (I had a spare one lying around, don't laugh). It was stored in a small cabinet with no circulation, and ran at about 75° and would throttle itself during high loads. It lasted for a year and a half, and I only swtiched to a Core Duo CPU with a passive heatsink a couple of months ago. As far as I know, it still works.