What does Heat do to a CPU?

[Avid6eek]

This question may seem a little elementry, but what is it that heat actually does to the CPU core that causes it to malfunction? Why does increasing the Core Voltage help to alleviate this? Just wondering what actually causes the problems with overclocking....i'm not electrical/mechanical engineer. Thanks to anyone who can help answer this for me.

[zogmool]

The higher voltage forces the transisters to switch faster.<BR><BR>And, higher temps will raise the resistances of the circuitry slightly, though that shouldn't be a problem untill it gets VERY hot.

[auld clootie]

<img src="i/expressions/face-icon-small-smile.gif" border="0"> as zogmool says,increasing the core voltage speeds up the movement of electrons, - and this actually aggravates the problem,not alleviate it. You speed up your CPU,but generate more heat. Because heat itself speeds up electrons (generating even more heat),trouble comes when things get out of hand,resulting in "thermal runaway" - at some point the currents start to increase massively and blow up the chip's microscopic circuitry. When you "overclock",you need to compensate for the increased heat generated by improving the heatsink to maintain temperatures at a "reasonable" level- well below the "trigger" point.

[zogmool]

Actually, the electrons never speed up or slow down. They always move at the speed of light between atoms. The heat raises resistance because atoms get further apart, and the higher the resistance gets, the more heat is generated. At the point where the power becomes too much for the material to disipate, temperatures spike and the material is destroyed (maximum wattage).<BR><BR>It's not really that the processor is so hot that it stops functioning; rather, that the transistors can't switch fast enough to maintain a good signal strength. Operation becomes unstable because the signal becomes unclear.<BR><BR>(If you've not used an oscilloscope, this will probably not make any sense, but here you go) In a digital circuit, you need a uniform squarewave. Where there is the degradation of the signal from runing the cpu too fast, you can counter it by oversaturating the signal. So long as you can keep it cool, you can do so safely. (since upping the voltage will obviously raise the temp by creating more current flow)<BR><BR>Just incase: I wasn't trying to "flame" auld with that. I'm sorry if I came across that way... <img src="i/expressions/face-icon-small-sad.gif" border="0">

[auld clootie]

<img src="i/expressions/face-icon-small-smile.gif" border="0"> I stand corrected,zogmool. You're 100% right about electrons. I should have said that the atoms become more agitated in their molecules.That's an interesting point about boosting bandwidth by clipping.

[Avid6eek]

Thanks guys for answering this for me. It's not something you need to know, but it's nice to know what causes things sometimes. Thanks again

[CHE]

Localized parts of a semiconductor can reach temperatures high enough that a phase change occurs, not necessarilly solid to liquid, but a change in unit structure which is also a phase change. A short results and the failure is called thermal secondary breakdown failure. The chip basically fries and varying degrees and types of material failure occur when something fries.

[GNXtreme]

Feck, that is way tooo much for me!

[Toddblah]

Hot electron transfer is a primary killer of CMOS structures. One effect that occurs during the actual using of the transistors is the hot-electron-degradation of the gate oxide. <BR><BR>Hot electron degradation occurs when electrons are accelerated to energy levels which allow them to cross the barrier of the gate oxide. The electrons would then either cross the gate oxide completely or get stuck within the gateoxide. A stuck electron would incorporate a negative charge into the gate oxide.<BR><BR>Under normal circumstances, this isn't a real issue, as long as the CMOS structure(CPU) is operated within specified parameters. The problem primarily occurs when "overvolting" the CPU..which raises the operating temperature at roughly twice the additional voltage applied. Keep it cool, and only use the minimum voltage required to achieve your desired results.<BR><BR>The question, far from being elementary, is even still not fully understood.<BR><BR>If you lose a processor or two, it's okay. It happens to all serious overclockers. <img src="i/expressions/face-icon-small-smile.gif" border="0"><BR><BR><BR><BR>

[auld clootie]

<img src="i/expressions/face-icon-small-happy.gif" border="0"> LOL&nbsp;&nbsp; i'm with you,GNXtreme - aint edukashun wunnerful?<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; aint gonna be no overclocker - sounds too expensive!<BR><BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;(just joking,Toddblah - that is really interesting)&nbsp; <img src="i/expressions/face-icon-small-smile.gif" border="0">