• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Why do we use heatsinks and fans?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Beast Of Blight

Disabled
Joined
Oct 17, 2002
Location
Orange County
Why do we use heatsinks on CPU and GPUs?

I mean, Why don't we just have a fan blowing on the CPU itself? Wouldn't that cool it better than waiting for the heat to transfer into the heatsink? :confused:

This is just something that popped into my head lastnight after reseating a old proc
psycho.gif
 
Beast Of Blight said:
Why do we use heatsinks on CPU and GPUs?

Because the surfce area of a heatsink is much larger, and you can transfer more heat from die--> copper than die--> air.

What does leright have to do with this? :)
 
Cjwinnit is wise :)

Surface area rules when trying to dissipate heat: air cooling solutions and water cooling both provide much more surface area which more than compensates for the loss of efficiency you get from the multiple media transfers (ie.: core to sink and core to block->block to water-> water to rad.)

To directly cool a cpu core with air I am guessing that you would need an amount of airflow that would destroy the core anyway :)
Maybe somebody who are smart with math and formula can come up with a real cfm number ;)
 
In a half-sleep tired state of mind, where logic fails, I thought it would just be more effiecient to blow air right on the core to cool it rather than use a medium such as a Heatsink between the fan and the core.
psycho.gif
 
heh time and again it has been shown that no heatsink=DEAD AMD. If you want to try it be my guest but have a fire extinguisher handy...

The heat from the CPU is safely dissipated due to the high thermal condutivity of the aluminium and higher thermal conductivity of copper. Air has an absolutely lame conductivity and so we need lots of surface area to transfer heat to it, and a fan to blow hot air out of the way
 
This is the way i understand heatsinks and why we use them, i may be wrong but im new here and i am still learning. THe die of a CPU is very small this does not contact very much air, and therefore limits the amount of heat that can be transferred away from the core, to compensate for this the airflow must be increased to offset the limited surface area, however as heat output grows the amount of airflow that would be needed rises to an impossible level, a heatsink draws heat away from the processor die and to the metal, this allows there to be more surface area and more area from which heat can be transferred, it is analagous to this situation say you have a 100 watt heat output of an object, now if this were the size of a key on a keyboard that object would feel very hot, because there would not be sufficient area for the heat to be dispensed of effectively (if it were a computer processor it would surely be incinerated by now), however if this source was teh size of the side of your computer case the source would be cooler to the touch, because passive radiation would be taking away alot of the heat.
I hope this helps :)
 
its fun to see how long u can run p4's without heatsinks... a pretty long time. u take one of off a amd processor and u can swear u gotta go spend some mo money.
 
I'd guess you could use an equation like this, assuming you wanted to keep the core at say 40C:

(Total Heatsink surface area) / (Typical fan CFM) = (Core area) / (Target CFM)

Of course, there is bound to be more too it than that- you need real technical data, and there are many more factors involved.
 
One BIG factor in that is the ambient temp. 10C difference in ambient temp = 10C difference in core temp.
 
CPFitz14 said:
What i'd like to know, is why isnt the actual core a heatsink.... now that would be an idea.



-CPFitz-

The clock cycle of a modern chip is too small to do this, by the time the signal has covered the distance from one corner of the die to the other, more chip signals (due to the high clokc cycle) wil have been created.

Nice idea though.
 
my estimate on the CFM needed to cool the die on a socket a AMD would be atleast 5k or more and all of that would have to hit the die directly not just the cpu... im sure if u find a 5k+ cfm can and some type of 3x2cm attachment for it then hey more power to you
 
xb-70 said:
This is the way i understand heatsinks and why we use them, i may be wrong but im new here and i am still learning. THe die of a CPU is very small this does not contact very much air, and therefore limits the amount of heat that can be transferred away from the core, to compensate for this the airflow must be increased to offset the limited surface area, however as heat output grows the amount of airflow that would be needed rises to an impossible level, a heatsink draws heat away from the processor die and to the metal, this allows there to be more surface area and more area from which heat can be transferred, it is analagous to this situation say you have a 100 watt heat output of an object, now if this were the size of a key on a keyboard that object would feel very hot, because there would not be sufficient area for the heat to be dispensed of effectively (if it were a computer processor it would surely be incinerated by now), however if this source was teh size of the side of your computer case the source would be cooler to the touch, because passive radiation would be taking away alot of the heat.
I hope this helps :)

There's one more factor in this -- the temperature of the heat source. Hotter objects transfer heat to the air faster, so if you were able to run the CPU at a temperature of several hundred degrees, you wouldn't need a heatsink to cool things off.
 
Carnil said:
There's one more factor in this -- the temperature of the heat source. Hotter objects transfer heat to the air faster.

It also depends on the temperature of the air, and i'm led to believe the relationship is linear. A temperature difference of, say, 30*C (we will say 50* for the die and 20* for the air) will transfer air slower than a higher temperature difference.

This applies to all object that radiate heat.
 
DepthwD said:
my estimate on the CFM needed to cool the die on a socket a AMD would be atleast 5k or more and all of that would have to hit the die directly not just the cpu... im sure if u find a 5k+ cfm can and some type of 3x2cm attachment for it then hey more power to you

And if you had that much air flying around you'd probably be seeing sparks jump around over the core from the static buildup, no?
 
Because you increase the surface area by 50-100x. This allows the air to accept more heat in the amount of time that the air passes by.

Thank You,
Daniel
 
Back