Advice for Airflow Change

Been looking at my setup for a while and decided maybe I have the airflow a bit backwards...

Temps are good. Ambient for the room is ~24*c and internal case temps stay ~30-32*c
CPU @ idle is ~37-41*c
GPU @ idle is ~38-40*c

Anyways here is a pic of the system showing airflow currently.

Yeah, I would try and switch your airflow so the intakes are the front and side fans. It would exhaust out the back and up through the rad. I'm just a fan of heat flowing up.

I agree with BeHappy. You are trying to fight the fact that heat rises. I understand the reasoning for your setup, pulling cool exterior air from the outside through the radiator. I would say swap everything, and test temps to be sure.

Sorry to say, heat doesn't rise.

Heat actually doesn't rise. Air, metals, liquids become lighter with heat because the material expands. In a block of iron, the heat radiates equally outward in all directions. Since the metals are bound together, the hotter molecules can't rise.

Fluids are different because the molecules are lightly bound and as they heat up, the mass becomes lighter, more space between the molecules due to more energy. And air is a fluid. Thus the heavier air goes down. Not the warmer air, the heavier air. So the light air has to rise to make room for the heavier air.

In a forced air situation it doesn't matter. You reverse your fans, you'll see a slight increase in CPU/GPU temps because warmer air will go through the rad.

Hi there, I see you have your CPU and GPU water cooled, these are the hottest components in your system, your board will generate some heat but it's nothing compared to that of your GPU..

reversing the airflow will also allow cool outside air to blow directly to your GPU plate, this may help a tad..

You can try reversing your airflow, top and back exhaust whilst front and side intake, the air in your case should be a bit cooler as you are exhausting most of the hot air out directly from your radiator instead of pushing it back into your case.. Your current setup is just pushing all the hot air directly into your case which effect your mb temps... Its a double edge sword really, intake on the rads bring cool air through them directly, but pushes all that hot air into your mb, whilse exaust will push some warm air through your rads, but your mb temp should be lowered... give and take I guess..

In theory, you may see a bit of performance increase by about 1-2c.. however many factors comes into play, each rig is difference and it boils down to what brand fans you use, CFM, RPMs, radiator FPI.. a good example is I can have the same exact setup as you and use different fans, temp may be slightly different.. or have a different radiator where the air flow is not so great due to a high PFI...

25c is a nice sweetspot ambient temp, no AC required whilst your rig still can run at full bore without it feeling like a heater...

Try swapping it, it'll give you something to do

Mmm, I am not sure that blowing the hot air from inside the case through the rad is a good idea. External temp is usually much lower than internal.

Conumdrum said:

Sorry to say, heat doesn't rise.

Heat actually doesn't rise. Air, metals, liquids become lighter with heat because the material expands. In a block of iron, the heat radiates equally outward in all directions. Since the metals are bound together, the hotter molecules can't rise.

Fluids are different because the molecules are lightly bound and as they heat up, the mass becomes lighter, more space between the molecules due to more energy. And air is a fluid. Thus the heavier air goes down. Not the warmer air, the heavier air. So the light air has to rise to make room for the heavier air.

In a forced air situation it doesn't matter. You reverse your fans, you'll see a slight increase in CPU/GPU temps because warmer air will go through the rad.

Click to expand...

I was under a different impression of how thermal dynamics worked, thanks for clearing that up. Still seems that the airflow should be directed at the video card instead of just trying to pull air around it.

Yep, any experiment can help. GPUs need coolest air puled to the fan openeng if air cooled.

Knowing physics or the the basics helps. I didn't graduate HS, but it's just how it works I guess.

SLI or Xfire with cards so close needs help for sure. Fannage always helps.

manu2b said:

Mmm, I am not sure that blowing the hot air from inside the case through the rad is a good idea. External temp is usually much lower than internal.

Click to expand...

Yes, but imagine this, you are taking the hot air from your radiator and moving it around in your system, your motherboard and plates on your GPU will suffer..

vise versa, you will be blowing cool air into the plates of your GPU and the air circulating in your system is not as hot, thus your motherboard should have lower temps... but the air being pushed into your radiator is not as cool..

Like I said, it's sort of a double edged sword, I suggest you pop your stuff out, reverse the fans and give it a go... It's actually not a lot of work to reverse fans, I think I did my setup (incomplete) about 5-6 times already messing with fans which takes only 10 minutes to reverse... Trail and error is the best part and fun of custom water cooling

I'd switch them. In general the air is going to be warmer going into your radiator after passing through the case but the warmer air leaving the case will rise. As you have it now there is the potential for the warm air coming out the bottom to be sucked right back into the top as it rises. You could get a "thermal short circuit" the way they are now. Reversed there would be less chance as the hot air rises away.

The physics behind "heat rising" is that things in a system always go in the direction of equal or increasing entropy "The Second Law of Thermodynamics". But that is off topic.

Owenator said:

I'd switch them. In general the air is going to be warmer going into your radiator after passing through the case but the warmer air leaving the case will rise. As you have it now there is the potential for the warm air coming out the bottom to be sucked right back into the top as it rises. You could get a "thermal short circuit" the way they are now. Reversed there would be less chance as the hot air rises away.

The physics behind "heat rising" is that things in a system always go in the direction of equal or increasing entropy "The Second Law of Thermodynamics". But that is off topic.

Click to expand...

This is what I remember reading in my thermal dynamic book. So, I'll stick with my original post and suggest switching the airflow.

Conumdrum said:

Sorry to say, heat doesn't rise.

Heat actually doesn't rise. Air, metals, liquids become lighter with heat because the material expands. In a block of iron, the heat radiates equally outward in all directions. Since the metals are bound together, the hotter molecules can't rise.

Fluids are different because the molecules are lightly bound and as they heat up, the mass becomes lighter, more space between the molecules due to more energy. And air is a fluid. Thus the heavier air goes down. Not the warmer air, the heavier air. So the light air has to rise to make room for the heavier air.

In a forced air situation it doesn't matter. You reverse your fans, you'll see a slight increase in CPU/GPU temps because warmer air will go through the rad.

Click to expand...

You're using mass where you should be using density. Solids that heat don't become lighter with heat, perhaps they become less dense (marginally, might I add). Fluids also don't become lighter (or lose mass). They become less dense, which is why convection occurs in the first place ("heavier" air falls and "lighter" air rises).

I do agree that in forced air situations convection currents don't matter (at least, not really). However, why try and fight convection with forced airflow when you could adjust the forcing so that convection aids it? Either way, convection shouldn't really be bothered talking about. I'd say to keep the case cool you pull air from the case out through the radiator (instead of blow air through the rad into the case). That way case temperatures stay slightly cooler for all the passive air-cooled components in the case.

So if heat doesnt rise then how do hot air balloons work? And why is the middle floor always cooler than upstairs?
I think i proved my point...

Crapxxas119 said:

So if heat doesnt rise then how do hot air balloons work? And why is the middle floor always cooler than upstairs?
I think i proved my point...

Click to expand...

heat rising is massively inefficient, even me swooping my hand downwards in the air has un-done all the effects of heat convection currents, even a cheap generic 80mm fan can outdo the heat rising factor in a situation as small as a pc case, which is why the logic from hot air ballons cannot be cross applied, and +1 to everybody who was said this in this thread, swap every fan around, front/side intake roof/back exhaust.

Crapxxas119 said:

So if heat doesnt rise then how do hot air balloons work? And why is the middle floor always cooler than upstairs?
I think i proved my point...

Click to expand...

For my part I saw it as an exercise in precise language. Heat doesn't rise but hot air does. Heat goes from 'hot' to 'not so hot'. Hot air is less dense than cooler air so there is a bouyant effect which most people would phrase as hot air rises. Which is shortend into "heat rises". But this is all academic in my opinion. But still fun.

As far as the effectiveness of natural air convection, it is used in many many applications. Most flat panel TV's are designed with passive cooling that relies on it. The original Macintosh was fanless and used it. Even large industrial cooling towers use it. So it is a viable means of cooling but not anything near as usefull for our hobby of PC overclocking. We like fans and lots of em!

i personally believe this heat/hotair/whatever rising arugment needs its own thread.

Well maybe ill reverse all the fans tonight and see what temps look like.

First Ill do a couple of tests with the current setup.

onefstsnake said:

Well maybe ill reverse all the fans tonight and see what temps look like.

First Ill do a couple of tests with the current setup.

Click to expand...

Testing first - always a good idea!

I mentioned the short circuit because I did this myself on my case. It has two locations for fans on the side panel. I put a fan in the top location and left the bottom open. My video cards were ruuning hotter? When I opened the case to check the side panel was very hot in the area betweenthe fans/ As it turned out I was actually drawing in air that came out of the case from the lower opening and then putting it back in. This created a "loop" where the hot air just kept getting hotter and hotter.

Loving this info. It's great when everyone doesn't have to go through the whole process to find out little tricks like this.

Well I cant reverse the flow untill I get some hardware.
I have my fan screws going through the case then into the rads.
So by reversing these fans I will need some shorter screws to mount the rads to the case.

Unless I setup the fans in a pull config, which wouldnt be a good comparison.

OK well I got my 1st test done. 5min with P95 running Blend Test on 8 threads as well as FurMark running @ 1280x1024 16x

TEST 1: Ambient 24*c, Rad Fans running @ 1600rpm
Idle Temps: CPU 40-36-39-37
(after 5min) GPU 39
Load Temps: CPU 66-63-64-61
(after 5min) GPU 52
~600w load shown by my UPS