college project: case pressure

NoNameNoNumber

NoNameNoNumber

NoNameNoNumber

Soldato
Joined
25 Nov 2009
Posts
5,392
getting my college project underway now, decided im going to play around with the relationship between case pressures and temperatures.
ive spoken to a few companies that make fans and one or two have them have been really helpful :)
got 3 freebies off noctua and 3 xilence redwings too
there should be some nexus fans on the way some time next week aswell.
basically what i intend to do is keep prime 95 running at all times and take temp readings off all of the probes firstly with just the fan on the cpu cooler and the two antec case fans on the medium speed exhausting air.
next i will add 1 then 2 (inlet) fans to the front of the case and then a third on the side (over the gpu) gradually increasing the case pressure from sub 1 bar to just over and log the results.
i will also try all (5) fans exhausting air then reverse them and see what happens too.
anybody got any other suggestions to help improve this project?
any help much appreciated :)
 
I'm unsure about this. What're your plans for measuring pressure at various points, or are you just going to swap fans around and write down the load temperatures?

I can offer that static pressure scales with the square of rpm, and suggest undervolting as a means of reducing rpm to increase the data that can be gathered. Moving the psu out of the case will be helpful for maintaining a controlled environment.

Do you hope the results will be applicable to other cases, or only your own? If only your own, which case is it, and will you be doing any ducting/blocking off of vents.

Mainly though I can't see there being enough data for a report. There's an awful lot of variables that you can't reasonably control, predicting airflow patterns is very complicated. Are you considering finite element work to compare and contrast with results achieved with the fans?

Engineering student might mean you're familiar with ansys and the like, in which case there is ample material for a study.
 
Does your cpu cooler really care what the case pressure is? Surely the biggest difference in switching more fans on will just increase flowrate of air, which is what ultimately cools the cpu, the case pressure is a byproduct of flow demand. I guess it might be a worthwhile comparison if you compare the effects of a fixed number of fans through the various combinations of in/out. I.e. with 4 fans you either have zero in 4 out, 1 in 3 out, 2 in 2 out, 3 in 1 out or 4 in zero out. But it is still going to be difficult to arrange them such that one scenario doesn't just cool the cpu more because the flow path of air is directed more at the cooler in one setup than another and thus could be completely unrelated to pressure or even total flow.

Edit. Also, if you are going to do this, you maybe want to make a custom perpex box rather than a standard pc case. just so you can position the fans equally to test combinations.
 
Last edited:
I done a bit of very crude experimenting with different fans on my lapped TRUE black a few weeks ago. Fans used were sharkoon silent eagle 2000's, sharkoon silent eagle modular 400-1200 pwms and arctic cooling 500-1350 pwm's. The sharkoon pwms gave a max temp of 76c on highest core, got a 3c drop using the full blown silent eagle 2000's, more noise for a very small drop in temps, the current arctics perform pretty much the same as the sharky pwms did, im at the limit as far as conventional fans go, dont want to use deltas.
 
AiiR said:
freebies you say? :D
Click to expand...
Yup :D they wanted proof of what im doing though
Bubo said:
Does your cpu cooler really care what the case pressure is? Surely the biggest difference in switching more fans on will just increase flowrate of air, which is what ultimately cools the cpu, the case pressure is a byproduct of flow demand. I guess it might be a worthwhile comparison if you compare the effects of a fixed number of fans through the various combinations of in/out. I.e. with 4 fans you either have zero in 4 out, 1 in 3 out, 2 in 2 out, 3 in 1 out or 4 in zero out. But it is still going to be difficult to arrange them such that one scenario doesn't just cool the cpu more because the flow path of air is directed more at the cooler in one setup than another and thus could be completely unrelated to pressure or even total flow.

Edit. Also, if you are going to do this, you maybe want to make a custom perpex box rather than a standard pc case. just so you can position the fans equally to test combinations.
Click to expand...

A perspex box had crossed my mind- i would just mount the mobo and hard drive in there and take temperature readings.
and yes pressure does affect heat dispersion

setter said:
I done a bit of very crude experimenting with different fans on my lapped TRUE black a few weeks ago. Fans used were sharkoon silent eagle 2000's, sharkoon silent eagle modular 400-1200 pwms and arctic cooling 500-1350 pwm's. The sharkoon pwms gave a max temp of 76c on highest core, got a 3c drop using the full blown silent eagle 2000's, more noise for a very small drop in temps, the current arctics perform pretty much the same as the sharky pwms did, im at the limit as far as conventional fans go, dont want to use deltas.
Click to expand...


im using an antec 300 at the minute.
i have 5 temp probes (that i am aware of) 2 on the cpu, 1 gpu, 1 nb and 1 hdd.
i will be using many combinations of the fans i have to hand to create different scenarios hopefully.
a perspex case will probably give me more reliable readings as i can eliminate any leakages of air.
 
JonJ678 said:
I'm unsure about this. What're your plans for measuring pressure at various points, or are you just going to swap fans around and write down the load temperatures?
Click to expand...
Pretty much

JonJ678 said:
I can offer that static pressure scales with the square of rpm,
Click to expand...
ooo thats useful :)


JonJ678 said:
and suggest undervolting as a means of reducing rpm to increase the data that can be gathered.
Click to expand...
the noctuas are pwm and the xilence i intend to run off a fan controller

thanks for you input :) when i get it up and running i will keep you ll informed :cool:
 
what thermal paste are you using on your cpu and cooler? this could play a big role in your results, make sure you record the room temps too! you dont want to get some odd results because your house is warmer / cooler then it was before. Consider all possiblitys.
 
chaosquall said:
what thermal paste are you using on your cpu and cooler? this could play a big role in your results, make sure you record the room temps too! you dont want to get some odd results because your house is warmer / cooler then it was before. Consider all possiblitys.
Click to expand...

Its just the stuff that came with the chip and heatsink. it was already applied.

yes ive got a thermometer to keep an eye on that too :)
 
I can't see how you are going to get any meaningful results as you're effectively increasing/decreasing airflow through the case which will have a significantly larger impact on temperature than just increasing pressure.
 
Pressure does not affect temperatures in the conditions you will be experimenting in.

I design bespoke heatpipe cooling devices for electronic equipment (CPUs included) - if you need some serious math help then please feel free to ask here or PM me; there is far too much to speak about in a single reply. I am guessing you're in college or similar as if you're in a degree you've just hit a world of pain if you expect to get a good mark (unfortunately!!)
 
Dessimat0r said:
Use a smoke device and see how the smoke is circulated around the case.
Click to expand...

This would be irrelevent as pressure is being measures not airflow, therefore the smoke would not be giving any indication of pressure as well as this not being measureable


Im trying to get my head around this and cannot see how pressure will affect things as said above airflow is airflow and surely moving more air moves more heat.

Unless the thermal aborbsion of air increase with pressure due to increased volume.

Also IMO using fans of different manufacturers will cause bad results as they wont be moving equal amounts of air at equal pressures. Am I correct in thinking static pressure will be relative to case pressure.

As far as measuring the pressure I remember it being used on the TV show Mythbusters that a U-shaped plexi or glass tube can be used to measure pressure providing half is in the atmosphere and the other half is in a sealed enclosure. On this basis an air tight acyrlic case could be used allowing you to measure pressure.
 
Last edited:
Ovr said:
if you need some serious math help then please feel free to ask here or PM me;
Click to expand...
cheers:)

Ovr said:
there is far too much to speak about in a single reply. I am guessing you're in college or similar as if you're in a degree you've just hit a world of pain if you expect to get a good mark (unfortunately!!)
Click to expand...
yes its college, its more about how i go about managing the project i think than actual results. so long on my report covers everything.

frozennova said:
Unless the thermal aborbsion of air increase with pressure due to increased volume.
Click to expand...
bang on.

frozennova said:
Also IMO using fans of different manufacturers will cause bad results as they wont be moving equal amounts of air at equal pressures. Am I correct in thinking static pressure will be relative to case pressure.
Click to expand...
think im gonna go down the pespex box route where i could have (eg)
3 xilence redwings on the front and 3 noctuas on the back.

frozennova said:
As far as measuring the pressure I remember it being used on the TV show Mythbusters that a U-shaped plexi or glass tube can be used to measure pressure providing half is in the atmosphere and the other half is in a sealed enclosure. On this basis an air tight acyrlic case could be used allowing you to measure pressure.
Click to expand...

one for youtube i think :)
 
However heat transfer from heatsink to air is not limited by the heat capacity of the air. It's limited by the various convection relationships and by the temperature gradient. The convection relationships are very much affected by airflow past the heatsinks.

This is before you consider how much above atmospheric pressure the air in a computer case can be. While air may move around it swiftly, the difference in density between various reasons is going to be slight, and heat capacity related to the absolute pressure not the gauge pressure.

So, I think you have a negligible pressure increase, which would be exerting a negligible effect on heat transfer even were the pressure difference considerable, and which is completely disguised by the convection relationships.

Unless you don't actually mean pressure, but as an engineering student I'd hope you know what pressure is.
 
It seems to me you are mixing up the oft used term "case pressure" with the more relevant issue of instantaneous pressure a fan generates in order to go about its business of accelerating a volume of air to create flow. If you think varying case pressure *alone* has an impact on heat transfer then see if you can blag a session in a decompression chamber or something and forget about trying different fan combinations, just use the same PC at different absolute pressures. I predict there will be no measureable difference. Tip, take the hdds out or they might explode.
 
You must log in or register to reply here.
Back
Top Bottom