Electronics vs Temperature

[cjdelphi]

No i got your attention by a vague title, someoneon fb asked an interesting question which brought up a 2nd question

1. Why do cpus die at around 140c when there's not much that melts, all i can think of is the trace lines melt or solder ? Why can't we run them at 300c and forget the fan, maybe the foward voltage drop / band gap drops when hot?

2. As people tried to answer why not,resistance came up and i quote

"Do resistors lose their resistance when super cooled" they should but

A Diodes fv goes down when hot, they go up when cooled... so if a diode is super cooled surely it does not behave like a semi conductor any more?

 

[(*steve*)]

"For every 10°C rise in temperature the lifetime of semiconductors is halved."

Whilst I'm not sure you'd want to bet your life on this statement being exactly correct, the correlation exists.

At high temperatures, ions are far more mobile and can migrate to places where they'll do bad things. A lot of care has been taken to implant ions in specific places and to place other materials in exact places. Increasing the temperature increases the speed at which these things move around (mostly by diffusion).

 

[JimW]

Who told you that they "die" at 140F? That is the temperature that the electrical timing specifications are no longer guaranteed. So a properly designed processor circuit may not function correctly if that temperature is exceeded. So the circuit may not work above that temperature, but the chip doesn't die.

JimW

 

[cjdelphi]

Who told you that they "die" at 140F? That is the temperature that the electrical timing specifications are no longer guaranteed. So a properly designed processor circuit may not function correctly if that temperature is exceeded. So the circuit may not work above that temperature, but the chip doesn't die.

JimW

I typed in a C for Celsius, not everyone lives in America

 

[(*steve*)]

not everyone lives in America

Be careful, them might be fighting words.