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

A Thought for calculating the Die temp

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

ILikeMy240sx

Member
Joined
Sep 28, 2003
Location
University of Michigan
Ok let's say that you know the temperature of water going into your CPU block and coming out of the CPU block. And the wattage that your CPU uses. Also the type of WB you use hence how effective your WB carries heat from the die to the water... Im sure you would have to do some testing to figure that out...

Are there any method to come up with an equation to calculate your die temp using variables above? I'm sure you'd need more variables so please feel free do add some on.

I basically want to get a close to accurate reading of the die temp without actually measuring it.

Am I dreaming or is this possible?

Edit: On the second thought you could get how much heat was transfered into the water with q=m*c*deltaT m being the flowrate. So you would have to know the flowrate. Anyway to get the die temp knowing how much heat was transfered?
 
Last edited:
That's how pHastus at procooling is doing it.

The problem doing it that way is you have to be able to measure temp with resolution and accuracy of .001 degree to get anywhere near reliable wattage figures. That = major $$$

Also, the lower in/out dT at high flow rates requires greater temperature measurement accuracy.

EDIT: Crap... I misread your post. I though you were talking about processor wattage. The most elegant solution I've seen is to tap into the on die diode and read it with an external diode reader. Calibrate the whole shebang and you have fairly good accuracy
 
Last edited:
Hmmm accruacy of .001 degree :eek: lolz Time for me to wake up now haha

But I would still be very interested in how pHastus... Any links or any info/tip from the man himself? :)
 
It is not really possible to accurately determine the CPU temperature with the variables identified. The problem is the interface between the CPU and the waterblock. That is a large variable, and if you are going through the effort to do this, and not just read the on-core thermal diode, you are going to need to accurately account for it. The effect of the thermal interface is determined by a number of factors you did not specify, like mounting force, and TIM thickness (which would be a function of mounting force). It would get very complicated, and therefore expensive, to get a more accurate reading than the built in thermal diode (it probably isn't even possible).
 
What he did @ procooling is very interesting and high-tech... I was aware of using diodes to get temp but didn't want to do all that to get an accurate die temp... Just wanted to use some variables and plug and chug into an equation heh..

Thank you for the link tho... Very interesting
 
Back