Super User is a question and answer site for computer enthusiasts and power users. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I've seen a lot of posts about how the maximum temperature of a CPU is somewhere around 100°C, and then the computer will shut down to prevent damage. But what would happen if all of the fans for a computer were turned off, and all hardware and software temperature interlocks were defeated? How would the computer's performance change as the temperature rose? I know people joke about being wary of reaching the melting point of aluminum (for some Macs), but would the temperature ever get that high?

To clarify, I'm looking for a degree-by-degree timeline of what the temperature effects would be on a running computer.

Specifically, what would happen to, say, a new MacBook Pro?

share|improve this question
up vote 4 down vote accepted

Through some early overclocking (PII's and Athlon Tbirds) you soon learn that a PC with faulty/inadequate cooling doesn't last long.

A degree-by-degree timeline with effects will vary by processor to processor and all the other hardware. The usual result on most older standard hardware (as mentioned by Ignacio) is what's known as 'blown caps' (ie failed capacitors). This would happen VERY quickly on an overclocked CPU (ie a couple of seconds on an old Athlon, you sometimes wouldn't even see POST). Non-overclocked Pentiums would generally last a few minutes, depending on the passive cooling (ie good/bad/non heatsink) you may even see ~20-30 minutes out of some.

Another symptom would be the 'Magic Smoke' (tm) escaping from the CPU. I've even seen flickers of flame!

I mention older hardware partly through experience, but also because there were very few 'failsaves' or inbuilt shut-offs into CPU's and motherboards. These days hardware is made with these in by default, most you cannot even 'turn off'.

Interesting info about chips overheating.

share|improve this answer

The temperature increase is caused by the transistors in the electronics switching on and off. Once a certain temperature is reached the transistors would depolarize, and the temperature would stop increasing.

share|improve this answer

That is really hard to say because of all the different effects and parts in a computer. But as a guideline, a computer turns all electrical energy into heat. So to know how much it consumes, you can simply add a meter to the power socket and calculate how much electricity it takes in to see how much heat it produces (we can assume a simple 1:1 relation).

So if your computer needs 100 Watt electricity, it produces 100 Watt heat.

If that heat can't go anywhere, different things will happen:

  1. Capacitors will blow, breaking power distribution to the CPU and graphics card.
  2. Transistors will burn because the heat isn't distributed evenly.
  3. Thin wires will melt and break
  4. The resistance for electrons in your computer will rise.

As the damage builds, there are two possible outcomes:

  1. The damage builds too quickly and the power supply fails.
  2. The damage builds slowly enough that some part of the PC can start to burn. Afterwards, you have roughly two minutes to leave the room alive unless there is nothing else (no paper, wood or plastic) which can catch fire. The fumes will be quite toxic.
share|improve this answer
Actually, silicon has a negative thermal resistive coefficient, which means that the resistance will actually drop. – Ignacio Vazquez-Abrams Aug 22 '12 at 8:14
@IgnacioVazquez-Abrams: That's true for the (tiny) transistors but what about the (long) wires which connect them? – Aaron Digulla Aug 22 '12 at 8:18
The coefficient for conductors is linear while for semiconductors it's exponential, so the increase in the wires can't compensate for the decrease in the semiconductor. Especially since the wire can eventually wick the heat away while it tends to stick around a bit longer in the semiconductor. – Ignacio Vazquez-Abrams Aug 22 '12 at 8:21
@IgnacioVazquez-Abrams: Thanks, fixed my answer. Anything else? – Aaron Digulla Aug 22 '12 at 8:23
I wouldn't say that it makes the computer faster so much as it exacerbates the heat problem; a decrease in resistance causes an increase in current consumption leading to more heat being produced leading to... – Ignacio Vazquez-Abrams Aug 22 '12 at 8:26

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .