Case for airflow

[ehume]

This is my rig, as of 2011-12-07:

 

 

What's all that stuff doing? Here's a diagram:

 

 

This is in a modded NZXT Beta Evo. It's not exactly the ideal case, y'know. So what would an ideal case be for air-cooled innards?

 

 

This gets you a nice airflow into the cpu heatsink. you also get the top air blocking air from the gpu, so your gpu's heat does not warm up the cpu.

 

Here's a standard setup, with a hint about why I do NOT like it:

 

 

See how the top exhaust fan draws air up to warm up the cpu? I do NOT like this setup.

[VaporX]

Does not work the way you think, while you might have dropped 1 to 2C for the CPU the overall case temp might have actually gone up. We have done a BUNCH of experiments on case air flow over the years and always find that the best OVERALL effect came from a chimney exhaust.

[Boinker]

you may experiment with it but I have found the best way for a side-panel to operate is in an exhaust mode. This varies from case to case and what Kind of Video card you have so do the tests and see what works best.

[vborets]

I think 3 fans can be removed without any problem

from top

from bottom

and middle 120 mm

[red1776]

I agree with Ed, I have extensively tested airflow configurations and with only rare exceptions the top fans should be pulling air out of the case. while pushing air down on top of the CPU heat-sink may give you a drop of a degree or two for the CPU, it is also redistributing the warmed air inside the rest of the case and usually disrupt the

 

linear flow of the shortest path of getting warmed air out of the case.

 

You might also try a negative pressure setup as well, right now it looks like you have it setup heavily positive pressure. (not knowing what your side panel mods might be that is)

 

 

[Waco]

Positive pressure is great for keeping your case clean

[red1776]

yep, positive is cleaner, negative moves more air...and requires more canned air...lots and lots of canned air

[VaporX]

negative moves more air

 

 

Not ture, negative air pressure is more restircted by case design on airmovement and thus tends to actually move less air. You are however both very correct about dust build up. One of the biggest reasons is with positive air pressure you control the intake points and thus can get filtering. Negative pressure will pull air from everywhere including the cracks and various other openings and thus bypass filters.

 

An additional caviate to positvie air pressure is that it helps reduce eddys of air the can form in the case between components and thus trap heat. In the OPs example the best setup is front in, bottom in, side in with top and back exhaust. You want the upper exhasut because it is uber effecient, the fans just enhance the natural air flow of heat rising and thus actually get a bit better exhuast performance than the fan is rated for.

[red1776]

not true to your not true , or we are talking about different things, thus my comment about how he has his side panel modded. having more CFM pulling air out of the case, moves more fresh air past components than a positive pressure. The problem is that when trying to set up a negative or positive setup, most will take it a bit too far and setup an 90/10 situation. cases vary greatly in their venting, panels and fan positioning, but i don't recall the last time i had a system setup that yielded lower temps with a positive setup. My current setup as an example is 2c cooler with about 75CFM + out, than the other way around. and this is (to the best extent possible, taking into account the effect on fan speed effects from case pressure)

 

again the caveats as you put it being the great variations in case designs and venting. and 2. if we are talking about your basic $40 case and only working with a single intake fan, and a rear exhaust, then yes i agree, posative in that case is the only option for the turbulence disrupting the eddy's, but that is clearly not the case here with the OP's setup, or desired setup.

 

 

 

 

 

I took this subject up a couple of years ago in practice after much reading and debate and and the problem with this is that what the two are designed to do, and how they are properly setup is not as easy as five fans out, and three fans in. i took this as far as introducing tracer into the different setups and observing the movement of air in the two setups. and with the exception of the 'very limited cheap case scenario' and cases that are more hole than case. the negative scenario has always been the more efficient cooling solution. (barring horrible wire management or extremely restricting filters) 

[VaporX]

Yeah I think we are discussing tow different ways to looking at this, that's cool. (Pun intended)

 

Looking at his case thought I will disagree. If we convert the top to a proper exhaust setup he needs to use the side as an intake, in fairness either way he should. The side as an exhaust will fight with the GPU fan for air movement and win every time. You want as much cool air to the fan for the GPU as you can get without disrupting the cases overall air flow. The key here is to keep a smooth air line. Hence the in from the front side and bottom and the out from the top and top/back. Think of this as water in a stream, fluid and air dynamics are very similar. A stream with a good flow will carry away dirt and other things. In this case the air flow will carry away heat. The flow of front to back and up will work with the natural flow created by the heat generated to allow the smoothest heat dissipation.

 

The problem with a negative flow case is the air intake areas. A negative pressure within the case means that the air is pulled from numerous areas, including those outside the normal air intake to compensate. While we know that dust is an issue due to this, so is actual air flow. The openings where a lot of the intake will pull from are not designed or placed for proper airflow so they will be restricted and work air into areas that will not be of much help. Additionally, as I pointed out they can result in eddys of air that will trap heat. A positive air flow means the air MUST escape, must find a way out. The good news is that since the case in question has a large chimney opening it has a ready made exit that not only allows for easy air flow out of the case but is position to take advantage of the air flow dynamics the heat from the system will generate.

 

BTW this will be a great section for the show and I am thinking we might even do it this week.

 

[ehume]

On my case, the side panel fan is a Kaze Maru 2, a 140mm fan set to intake. The side window is 120x120mm, with 120mm fan mounting holes. But the KM2 has 120mm mounting holes, and its 140mm fan takes up the whole 120x120mm square window, where a 120mm would take up less than 78% of the window. Better use of the space. In the diagram I left the choice of intake or exhaust open because some gpu's blow hot air in all directions. With lower front and bottom intakes the system could afford a side exhaust with such gpu(s).

 

In my own case I used a nibbler to eliminate the pillars from between the slots in the backplane. Now I have an open window there just as I have an open window up where the rear grill used to be.

 

I believe that removing the rear grill is key to developing what I call a positive flow case. With such an open area it cannot really build up a high pressure. Is it effective? Once I did an experiment. I removed both top fans and left those grill-free spaces open. In the other spaces I either removed fans or left them turned off. The only fans running were a pair of YL D12SH-12's acting as push/pull on a Megahalems. Even with the top open it was able to hold up a Kleenex over the lower front fanspace. And here is a picture of the power of a heatsink and an open rear grill space:

 

 

BTW -- this is a $40 case.

 

 

I'm toying with putting a plastic separator between the lower section -- the gpu space -- and the upper section -- the cpu space. Hobby shops sell sheets of stiff styrene for model railroaders to build houses. The stuff works great for all sorts of things.

 

The reason I am not concerned about older experiments where you say "sure, you cpu will get another 2c of cooling, but . . . " is that in the Intel world, since the LGA1366 there is nothing left on the motherboard that needs cooling. Lynnfield took the last of the hot chipset components and put it on-die. The mb components not only don't need the cooling anymore, they are in a position tp compromise the cpu cooling. And with an i7 8xx cpu, you really need to cool that. (And yes, I am aware that with each passing tick-tock, the need for even heavy-duty cpu cooling is passing.)

 

That leaves the gpu, which is still very hot on gamers' machines and is(are) in a position to really mess up the cpu temps with not very much gain for gpu temps. Hence the thoughts about partitioning the plenum into gpu space and cpu space. It also explains why I think slot pillars have to go.

Edited December 9, 2011 by ehume