Temperatures for lots of different case-fan settings tested

[HESmelaugh]

Introduction

Here is some testing I had been thinking of doing for a long time and now I finally got to it. I have posted some reviews and tests in the liquid cooling section and since they were always very welcome there, I thought I'd post something up in air cooling too. :-)

Basically, I tested temperatures with lots of different case-fan setups on a Cooler Master RC-690 case.


Testing method

Overview of thehardware used:

- Q6600 CPU @ 3.0 GHz
- Asus P5E WS Pro Mainboard
- Thermalright Ultra 120 eXtreme Heatsink
- Radeon HD 4850
- Arctic Cooling Accelero S1 GPU-Heatsink
- Corsair VX 550W PSU
- Cooler Master RC-690
- T-Balancer bigNG Fan controller

I used the following fans for this test:
- Scythe S-FLEX SFF21E on the CPU heatsink
- Scythe Kaze Jyu Slim 1000 on the GPU heatsink
- Nanoxia FX-1250 for all the 120-mm case fans
- Yate Loon D14SL-12 and Sharkoon System Fan for the 140-mm case fans

I realize that it would have been more elegant to use identical fans everywhere but since I don't have that many fans of one kind and this review was almost unsponsored (Nanoxia provided four fans - Thanks! ) I had to mix and match.

I ran all of the 120-mm fans at 900rpm and all of the 140-mm fans at 800rpm. I wanted to achieve a reasonable compromise between silence and performance.


Presentation

I'll present the test data in sets of pictures. I'll demonstrate how this works on the first and most basic of all fan setups, that I call "front in, rear out":


This first image shows an side view of the system with arrows symbolising what fans are in use. Intake fans are shown as blue arrows pointing inwards, exhaust fans are shown as red arrows pointing outwards.


The next image shows the temperatures measured after one hour of full load achieved with Prime95 and Furmark. Temperatures for the CPU, GPU, Northbridge, Harddisk and internal case temperature, measured with a digital sensor, are shown in the according locations.
All of the temperatures are shown as the difference between the measured temperature and ambient temperature, measured in Kelvin (1 Kelvin equals one degree Celcius of temperature difference).

For easier overview, in the following pictures, temperatures that are lower (better) than with this first, basic setup are colored blue, temperatures that are higher are colored red.


Part one: Silent

First, I try some settings using only two or three fans. This keeps the noise level down.

Front in, rear out, top out





Note that you would get similar results in cases where the PSU is located above the CPU heatsink and also draws air from the case outwards.


Bottom in, top out





Note that the CPU heatsink was rotated by 90 degrees to match the airflow going through the case.


Bottom in, top out with one less fan

Using only two fans, you get a very quiet setup. I tested this to see how much of an impact one less fan would have.






Passive intake

This is something I first encountered in an Antec P182 case. As shipped, there are only exhaust fans on the case. All the fresh air gets into the case via passive intake. This is also a very quiet setup.






More fans

Here I use the fans mountable on the side panel of the case.

Side exhaust






Side intake






AHELLUVALOTTAFANS!

Here, I use all or almost all the fans that can be mounted on an RC-690.

(Almost) all out






(Almost) all in






Balanced out/in






Very brief conclusions:

Some of the test-results were just as expected, but it's also clear that while more fans equals better temps, the placement of these fans does play an important part.

For silent systems, I would recommend either the good old "front in, rear out", which proved to be quite effective or the "passive intake" setup.

If you're using more than two or three fans, intake or exhaust fans on the side proved to be very good (whether they intake or exhaust not making that much of a difference). Finally, if you use tons of fans, having top and rear fans exhaust and all the others intake proved to be a well balanced solution for all components.

Hope you liked this test. Comments and feedback welcome! :-)

If you know German, you can also check out the original article.

Cheers,
Shane

[Okda]

honestly the results are what i expected but it is good to have them proved for you
Thanks for the effort done

[roofsniper]

yea i expected about the same results as well. the problem with testing case airflow is that every case is different and there is so many ways to do it. i have been trying to find new ways to get better airflow this review confirmed that the way i am doing now is the best for me. although i wish my case could have a top exhaust. also what are you using to measure temperature? what is the K? can't be Kelvin.

[jdmR]

lol if its kelvin, thats one hell of a cold system. its probably C. anyways, nice testing man, after seeing these numbers especially in this case, uve convinced me to go buy one, either this or the 590

[1_budz]

lol if its kelvin, thats one hell of a cold system. its probably C. anyways, nice testing man, after seeing these numbers especially in this case, uve convinced me to go buy one, either this or the 590

You won't be disappointed with either the CM 690 or CM 590. I own both of them and the airflow is good and the cable management is great as you can hide the cables behind the mobo.

[roofsniper]

lol if its kelvin, thats one hell of a cold system. its probably C. anyways, nice testing man, after seeing these numbers especially in this case, uve convinced me to go buy one, either this or the 590

yea exactly thats y it can't be kelvin but it can't be C either its saying that the case ambient is 5C at some points and the hdds are around 10 with the nb around 35. then the fan on the TRUE is only running at 900 rpm which wouldn't achieve those kinds of temperatures. im not sure what it is. but it doesn't really matter as the difference in numbers is all that matters.

[fatguy1992]

Nice work Very interesting

[HESmelaugh]

Thanks for the feedback.

To clarify, K is Kelvin but these aren't absolute temperatures. The numbers you see are always component temperature minus ambient temperature.

The temperatures were measured either by internal temp sensors or digital temp sensors attached to the t-balancer bigNG. I had some sensors placed outside the case to monitor ambient temperatures.

I agree that every case will be different. Also, it makes a big difference that I used an aftermarket cooler on the graphics card since this draws air vertically and stock coolers tend to have radial fans.

[bartx]

Very nice review In my case I use only one fan - rear exhaust. If I remember good, CPU temps decreased about 7-8C. For me it's enough

[roofsniper]

do you remember what your ambient was?

[HESmelaugh]

During all of the test runs, average ambient temperatures were between 21.5° C and 22.5° C.

[Ar3s]

Top In, Rear Out, Front In, Bottom Out, Side In. Positive Pressure so less dust in the long run too.

[emoners]

nice work!

[ThugsRook]

excellent work

ive run these same tests on my rig many times throughout the years, but my results arent quite the same as yours... intake fans always result as useless for CPU temps

[HESmelaugh]

... intake fans always result as useless for CPU temps

interesting. what heatsink are you using? I guess that different heatsinks would react very differently to the fan setups.

[ThugsRook]

interesting. what heatsink are you using? I guess that different heatsinks would react very differently to the fan setups.

using many different heatsinks on various platforms, currently a Noctua C12P on a e8200 775.

my results always ended with "2x side in/2x top out" as best, but my case isnt quite as big as yours.

[Lukeatluka]

Great, this is just what i needed.

I'm really glad that i found this thread.

Thanks!

[MomijiTMO]

Kudos on the test. I never really thought about case fans before so this was quite informative.

[KoHaN69]

somebody give this man a sticky!!!

[bettyfromhell]

can you try this configuration? : front intake , side intake , rear intake , bottom intake, top exaust and the heatsink rotated 90° right in order to make its fan exaust to the top.

thank you, very nice testing

[jdmR]

just wondering, for the side panel fan near the cpu, did u put it on the outside?

[HESmelaugh]

Aww, crap! I had totally forgotten about this thread!

Sorry for necromancing, but I really don't want to leave these questions unanswered...

bettyfromhell: Sorry, can't do additional testing since the system was disassembled shortly after the last test-run.

jdmR: Yes, I had to install the fans on the outside of the panel, because the TRUE heatsink is so tall it takes up almost the whole depth of the case.

Again, sorry for being so late...

Cheers,
Shane