With all that electricity passing through silicon at an amazing speed, there’s sure to be friction; and heat. The experts are continuously telling us how important it is to cool these tiny little pieces of silicon they call “CPUs” and “chipsets”. The problem is, parts of the cpu start to deform at high temperatures, also known as “burn out” temperatures. Although there are many many layers to the microprocessor, we focus our attention to the top of the chipset for the most effective overall cooling. So what if it is hot? If the silicon’s not deformed then the electrons can pass properly and we make the connection. So why cool it more than you need to?
Well, firstly lets look at silicon. Silicon comes in two types, P-Type and N-type, but effectively, the heat properties of the two behave the same. Silicon, as it heats up becomes more conductive, letting more electrons (put simply, electricity particles) through. Now I know what your thinking, “hold on, if it’s more conductive, doesn’t that mean it should be faster?”, but no. The speed of the computer isn’t on how fast the electricity moves through silicon, although, speeding that up would improve things. We’ve all heard the term, clock speed. Gigahertz. Megahertz. That is, the speed at which it can execute each instruction and synchronize each computer part. So to put it in lemans terms, the faster the clock speed, the quicker the computer (not taking architecture into account, for all you techies out there).
To put all of that simply, a CPUs heat is directly related to it’s clock speed; the higher the clock speed, the more power used and subsequently, more heat. Without adequate cooling, today’s CPUs can very easily reach well over 150 degrees Celsius, if there weren’t systems in place to prevent it. These systems include a clock speed limiter. When the CPU gets above temperature, the CPU management (BIOS – CMOS system) slows it down. This results in a complex system that gives performance increases with a cooler system, and performance slow down in a hot system.
While performance is a big aspect to look at in hot systems, so is the ongoing toll on the hardware. Consistently hot systems aren’t expected to last anywhere near as long a cooler systems. Thermal fatigue in silicon results in cracks in transistors, and transistor mounting failures, in particular. Of course, just one transistor failure on a complex system such as a motherboard or graphics card can result in the entire system failing. And a replacement board being an appropriate fix.
What’s a safe operating temperature?
Each computer system is different. Operating temperatures depend largely on the CPU type, the provided cooling, the age of the computer, and the motherboard heat management systems. But as a general rule, your CPU should be running between 40 – 60 degrees Celsius, and your system should be between 20-40 degrees. If it’s cooler, great. But generally it’s not good to have it running hotter. Laptop batteries are largely affected by heat so be sure to take out batteries if
What is appropriate cooling?
As much as some overclockers will say that you really do need a liquid nitrogen cooling system, a basic, air-cooled, computer fan system is usually fine. As long as your not not overclocking and running power intensive programs (like games), any processor will come included with appropriate cooling. If however, your in a particularly hot part of the world, be sure to look into better fan or even water systems.