Voltage Stressing A CPU?

[Monkfish]

What I wish to understand is the short-term or long-term dangers posed to a modern CPU by increasing its supply voltage while maintaining normal operating temperatures (or even lower than normal using exotic cooing methods).

As I see it, a transistor or CPU can be destroyed by supplying a higher than rated voltage in one of two ways: the voltage itself (e.g. static discharge), or heat spikes caused by higher currents (e.g. thermal runaway).

If the chip is being cooled then how does a higher voltage actually stress the chip? Does it produce heat spikes that cannot be conducted away sufficiently quickly?

How exactly does a static discharge destroy a transistor (i.e. is it too simply heat-related at the molecular level)? How does a higher then spec supply voltage actually stress a chip?

Perhaps someone frequents this forum with a deep knowledge of transistors or CPUs?

Thanks you

 

[IRstuff]

Higher voltages coupled with higher speeds from the cooling of the chip will aggravate a number of failure modes.

This was first tried with the 286, over 20 years ago, overclocked at 50% higher supply voltages. The CPUs lasted about 1 month. Cooling will not mitigate the failure modes.


TTFN

 

[Monkfish]

Are you aware of any online sources that discuss these failure modes in detail?

 

[IRstuff]

Go to the semiconductor manufacturers' websites. Their reliability reports list the types of failures.

TTFN