120mm silent fan ducted to the heatsink

mycode · Post by **mycode** » Sun Jan 16, 2005 8:46 pm

I'm thinking of the following setup:

1. hang a 120mm silenX fan over the cpu heatsink with four rubber strings connected to the four cornners of the case;
2. pipe two fan adapters (120-to-80 and 80-to-60) together to connect the fan to the heatsink;
3. dremel the side panel to get a 120mm intake hole for the 120 fan;

My question is, with the distance created by the two adapters, is a fan with 58cfm output good enough to cool down an AMD Semptron 2200+ CPU with a cheap heatsink (Cool Master CP5-6j31c)?

Pjotor · Post by **Pjotor** » Sun Jan 16, 2005 9:49 pm

Hi mycode, and WELCOME TO SPCR!

Using just one fan adapter inhibits the airflow a great deal. "Piping" not just one, but two, adapters from 120mm to 60mm will do it even more so (only about 1/4 of the surface area left). SilenX fans are low-flow, and have low back pressure, which means cooling *will* suffer badly.

I really, really suggest getting a better heatsink -- one that can take at least an 80mm fan. Cooling-wise, your present heatsink *should* be enough to cool the CPU, but you won't be able to turn down the fan very much; it will create some noise.

Your idea does work, though; I have seen it in other places with good results. Let us know how and what you do.

Once again, welcome to the SPCR forums.

warlock110 · Post by **warlock110** » Mon Jan 17, 2005 5:00 pm

that's exactly what i'm trying to do, but i have a zalman fan bracket, so mounting is alot easier than your style, and i'm gonna use a yale loon 120mm at full speed, cooling a P3 sitting under the aero flow

, i hope it works, this is a very light rig, so i hope it's gonna work.

mycode · Post by **mycode** » Wed Jan 19, 2005 11:39 am

Thank you guys for your input.

Using a bracket is easier to mount the fan, but I guess hanging the fan by rubber strings or springs can reduce vibration. Using a rubber or silicon washer between the bracket and the fan should help as well.

Placing the fan over the heatsink directly will reduce the volumn of air flowing through the heatsink, that is the reason why some people use an off-centered setting. But here I'm trying to use adapters to capture every bit of airflow and "compress" (due to the fact that the air is passing through a narrow-down tunnel) it all to the heatsink. I guess this will keep the static back pressure of the bigger fan at equivalent level, if not increase it. If the inner skin of the adapters is very sleek, the air should be able to deliver smoothly. To deal with the pressure loss because of the turbulence generated at the central blind spot of the fan, I'm going to attache a parabolic cone to the center of the fan, if I can find one that match the size of the motor.

I'm not an aerodynamicist nor an acoustist, so all above are just my hypothesis. Every comments are welcome. If you can provide some information about similar trial on the web, you will definely save my time and effort. Thanks again!

Krispy · Post by **Krispy** » Wed Jan 19, 2005 5:07 pm

mycode wrote:

To deal with the pressure loss because of the turbulence generated at the central blind spot of the fan, I'm going to attache a parabolic cone to the center of the fan

I fear your chances of doing this without putting the fan blades out of balance are slim. You'll end up with more vibration, hence noise & reduce fan bearing life.

Wouldn't a home made duct offer a smoother reduction in size rather than joining together various adapters? It would be cheaper too.

I'm with Pjotor & a larger heatsink. Best of luck!

Tibors · Post by **Tibors** » Thu Jan 20, 2005 10:27 am

Krispy wrote:mycode wrote:

To deal with the pressure loss because of the turbulence generated at the central blind spot of the fan, I'm going to attache a parabolic cone to the center of the fan
I fear your chances of doing this without putting the fan blades out of balance are slim. You'll end up with more vibration, hence noise & reduce fan bearing life.

The parabolic cone needs to be on the side where the air leaves the fan. So it won't be attached to the rotor, but to the other (non moving) side of the fan hub. There is no reason why that would imbalance the fan blades.

I don't know if all this is worth the trouble, but it sure is an interesting experiment. Could you try to make measurements with and without the cone?

I agree that a homemade duct could be smoother than two adaptors. Besides a plastic funnel (from a household supplies shop or car repair supplies shop) or some rolled up sheet is probably cheaper than two adapters.