/sci/ - Science & Math » Thread #14209438

132KiB, 1178x1500, 81mSPKcv89L._AC_SL1500_.jpg

View Same Google iqdb SauceNAO

Air Flow Requirements for Cooling a Processor

Anonymous Thu 10 Feb 05:52:45 2022 No.14209438 View Reply Original Report

Quoted By: >>14209466

I'm curious if such a thing is even reasonably possible to estimate / calculate. I'm not here to ask "exactly what fans should I buy for my computer" or anything like that, mainly just curious about the thermodynamics behind this.

So far, looking it up, I've found a few different formulas:

Airflow[cfm] = watts / (20 * (1.7 / 60) * (Tin - Tout))

rate.of.cooling[W]=rate.of.heat.production=hA(Tsurface?Tair) Where, h is the convection coefficient, A is the surface area of the processor, Tsurface is the surface temperature of the processor and Tair is the average temperature of the inlet air.

CFM=1.79*watts/delta T

All of these seem highly suspect and also vary a suspect amount.

Is it calculable? Is there too many variables?

Anonymous

Anonymous Thu 10 Feb 2022 05:57:45 No.14209466 Report

Quoted By:

>>14209438
no, it isn't calculable, thermodynamics is all a bunch of fakery and nonsense which only has real world application in the context of passing exams which test how good you are at regurgitation of fakery and nonsense.

Anonymous

Anonymous Thu 10 Feb 2022 06:00:59 No.14209482 Report

Quoted By: >>14209659

>is it calculable
yes.
>are there too many variables
yes. What you want to do is simplify with assumptions. Work backwards. Look at the units of what you are looking for: Watts. Go from this desired output unit, rewriting it in as an equation using other units, until you arrive at a good group of inputs that you like. Consider your ideal: a processor which always consumes x Watts, with a nice round number for its size, in a room which is always cooled to y degrees in the unit you like.
>cfm
Not a great unit for this kind of thinking. Consider ft^3/s or m^3/s as a simplification.
Don't go into any thermodynamics rabbit holes, keep it simple. You aren't wanting to model air currents, you're wanting to make a few simple equations that give you your processor end temperature given fan specs and processor wattage.

Anonymous

Anonymous Thu 10 Feb 2022 06:24:41 No.14209573 Report

Quoted By:

What are you asking? To make a graph of a given heat source and heatsink purely from theoretical calculations? Not without computer modeling.

What you do in the real world is just build it and measure. Chips have it in the datasheet how many degrees Celsius per watt they rise over ambient for a given heatsink.

The takeaway that I've found from playing with them before is that any cfm of airflow is better than 0. Also don't forget about radiative heat loss.

Anonymous

Anonymous Thu 10 Feb 2022 06:54:43 No.14209659 Report

Quoted By: >>14209808

>>14209482
I don't know enough about thermodynamics to do that myself, I only took thermo 1 in my engineering classes. I know my room temperature is around 20 C, my processor always uses about 2-2.3 watts, it hits 64 C average under load, and the maximum fan size is 92mm diameter. Besides that I wouldn't really know where to go from there. I want CFM instead of cubic feet per second because fan airflow is measured in CFM.

Anonymous

Anonymous Thu 10 Feb 2022 08:04:09 No.14209808 Report

Quoted By: >>14211110

>>14209659
okay that's good info
in my experience and what I've seen from pc testing online, increasing airflow cfm barely helps
you need to decrease thermal resistance, remember it's analogous to a circuit

Anonymous

Anonymous Thu 10 Feb 2022 16:28:52 No.14211110 Report

Quoted By: >>14211436

>>14209808
I have heard this being relevant in PC building, but that's in a situation where most people already have an intake 120mm and an exhaust 120mm, which push a huge amount of air through the case. I can see how there might be no measurable difference between 30 CFM and 50CFM, but I feel like 4 vs 15 could be dramatic.

I don't know how to decrease thermal resistance.

Anonymous

Anonymous Thu 10 Feb 2022 17:41:12 No.14211436 Report

Quoted By:

>>14211110
>I don't know how to decrease thermal resistance.
increase airspeed
increase contact area
add fins

Capcode	All Only User Posts Only Moderator Posts Only Admin Posts Only Developer Posts
Show Posts	All Only With Images Only Without Images
Deleted Posts	All Only Deleted Posts Only Non-Deleted Posts
Ghost Posts	All Only Ghost Posts Only Non-Ghost Posts
Post Type	All Only Sticky Threads Only Opening Posts Only Reply Posts
Results	All Grouped By Threads
Order	Latest Posts First Oldest Posts First

Your latest searches