why is heat bad for hardware

bum · Post by **bum** » Fri Jun 18, 2004 11:14 pm

ok i know this question may sound stupid but im gona ask anyway.

why exactly is heat bad for hardware ? i mean, everyone know's its bad, everyone knows that fans, heatsinks and liquid cooling solutions make things colder, but why exactly is heat so bad. what does it do to freeze or screw up components ? and as the old age question goes, if chicken can be cooked at 180C and still be good, then why cant my silicon ?

madmag9999 · Post by **madmag9999** » Fri Jun 18, 2004 11:20 pm

becouse heat creates friction and slows down the reaction of the transistors. the colder u get those transistors the faster they move.

TaranT · Post by **TaranT** » Sat Jun 19, 2004 2:33 am

(Calling dwchang!! Or mneideng if he ever comes here...))

Try this, bum. Heat causes almost all materials to expand in physical size. That includes the metal wires - if you can call them that, being as small as they are - and junctions in the chips. If they expand enough to cross the space between each other, electrical shorts can occur.

I'm not sure how true that is, but it's the first thing that comes to mind. Physical expansion at nano-scales can be deadly.

This article is much more detailed, but probably uses language that only a few of us can understand. (Not to mention that the English is imperfect.)

The two major problems he describes are:

(1) Electromigration...which I suppose means the metal "wires" will essentially start evaporating. Thinner wires means higher resistance means lower performance.

(2) Oxidation...which happens when heat promotes "dirt" - for lack of a simpler word - to accumulate inside a transisitor. A transistor's performance, and there are millions inside a typical CPU, will degrade and possibly fail outright.

bum · Post by **bum** » Sat Jun 19, 2004 7:12 am

thats for that taran and madman. so based on what ya told me if a chip runs at dangerously high temps for a long period of time then the cpu could actualy start eroding away. hmmm this is interesting.

but on the other hand, what if a cpu gets too cold ? freon coolers can take cpu's to -30C (-22F) degrees and lower. could a cpu get so cold that it will slow down the movement of electrons , or possibly something worse ?

Corran · Post by **Corran** » Sat Jun 19, 2004 8:26 am

When approaching Absolute Zero (-273 degrees C) it would become a superconductor.

http://www.ornl.gov/info/reports/m/ornlm3063r1/pt2.html

Heike Kammerlingh Onnes recognized the importance of his discovery to the scientific community as well as its commercial potential. An electrical conductor with no resistance could carry current any distance with no losses. In one of Onnes experiments he started a current flowing through a loop of lead wire cooled to 4 K. A year later the current was still flowing without significant current loss. Onnes found that the superconductor exhibited what he called persistent currents, electric currents that continued to flow without an electric potential driving them. Onnes had discovered superconductivity, and was awarded the Nobel Prize in 1913.

bum · Post by **bum** » Sat Jun 19, 2004 12:28 pm

WTMOTHERF ? ok wait, so theoreticaly, if a cpu we're cooled to -274C (persumed absolute zero) then it could have limitless speed, or to be more precise, it could run at the speed of light (300,000 km per second or 186420 miles per second, which is the presumed speed limit of the universe) and thierfor be the most powerful single processor possible. i know this is all very theoretical and i could be missing alot of info as i ramble about on this, but, disregarding the dificulty of the task and conflicts with other physical forces, would what i just sujested be possible ?

TaranT · Post by **TaranT** » Tue Jun 22, 2004 12:00 am

Maybe at absolute zero, nothing moves at all. On the other hand, if the processor is doing anything, it'll make some heat, so you can never really get to the lowest temp.

People are trying to develop optical computers, either all-optics or optical hybrids. Something like that would run at nearly c - which electronic computers can't do. A Google search on "optical computer" will bring up some links for you.

bum · Post by **bum** » Tue Jun 22, 2004 1:52 am

bum wrote:thats for that taran and madman. so based on what ya told me if a chip runs at dangerously high temps for a long period of time then the cpu could actualy start eroding away. hmmm this is interesting.

but on the other hand, what if a cpu gets too cold ? freon coolers can take cpu's to -30C (-22F) degrees and lower. could a cpu get so cold that it will slow down the movement of electrons , or possibly something worse ?

madmag laughed at me when he read that

madmag9999 · Post by **madmag9999** » Tue Jun 22, 2004 2:32 am

actualy, i laughted at this

bum wrote:WTMOTHERF ? ok wait, so theoreticaly, if a cpu we're cooled to -274C (persumed absolute zero) then it could have limitless speed, or to be more precise, it could run at the speed of light (300,000 km per second or 186420 miles per second, which is the presumed speed limit of the universe) and thierfor be the most powerful single processor possible...

but if it could really happen that would be very sweet.

TaranT · Post by **TaranT** » Tue Jun 22, 2004 2:53 am

I have to laugh at myself for writing that a processor would run at c. That makes no sense at all.