Heat sink fan comparison tests

Inspired by the idea of two diagonal fans on a HS, I decided to test a few configurations to see which was better in my rig.

The original idea is here but I simplified it to just two fans attached together:

http://www.ocforums.com/showthread.php?t=298736

First a single panaflo 59cfm fan blowing on an SP-94 (p4 3.2 @ 3.52). I put my rig sideways for this experiment so it's a bit more efficient than normal. The side panel was open.
-----case/HS = 33/36c (sucking = 33/35.5 almost the same)




Next, two fans connected together in a V shape blowing into the HS. I put two pieces of cardboard on the sides of the V to direct the air downwards.
-------case/HS = 32.5/35.5c (about the same as a single fan, surprised)





Next, the same V configuration but with 80mm fans blowing into the HS.
-------case/HS = 32/38c (a lot hotter, smaller cfm did it)

Next, the two fans connected side-by-side and sucking from the HS rather than blowing on it.
-------case/HS = 28/34c (much cooler)



Finally, putting my rig back on its bottom with the side-by-side fans blowing into a duct and sucking off the SP-94.
-------case/HS = 29c/36c (a 2c diff for the sp-94 on its side.)

Interesting that all those configurations didn't make much difference to the over all temps. I suspect my CPU temp reading is 2c lower than it really is, so for comparisons you need to add 2c to my HS temps. My case temps seem accurate.

Hmmm, I'm looking at http://www.ocforums.com/attachment....tid=32764&stc=1 and I think it's not very efficient, because the fans are blowing into each other.... Maybe if they had a wider angle (now it looks like 80&deg) like 150&deg ish, they wouldn't blow into each other?

neonblingbling said:

Hmmm, I'm looking at http://www.ocforums.com/attachment....tid=32764&stc=1 and I think it's not very efficient, because the fans are blowing into each other.... Maybe if they had a wider angle (now it looks like 80&deg) like 150&deg ish, they wouldn't blow into each other?

Click to expand...

The angle is 60 and that was set because the width of the sp-94 is 92mm so it's a equalateral triangle with 2 fans. If you made the fan angle wider then at least part of the air would miss the HS and since the strongest air comes from the edge closest to the HS, it would also be your best air that is lost.

I fooled around changing the angle and allowing the air to blow on my face to see if it would change the force of the air. I found that the 60 deg angle is pretty good at forcing air and that wider angles only spread out the air and didn't seem to result in more air velocity. I was going to check it with some spread sheet calculations, but I haven't done that yet. My feeling was that from 60 to 100 deg is about the optimum air velocity, either way from that the air velocity definitely felt less on my face.

An angle of 150 deg would spread the effect out and you would need some way of converging the air to enter the HS. That's why I tried some 80mm fans because the angle becomes 70 deg, but the lower cfm of the fans made the cooling effect much less.

did you think about putting a vertical divider between the two fans?

not the full length to the heatsink tho, but about halfway. that would stop the fans blowing into each other and may decrease the backpressure.

Evil T C said:

did you think about putting a vertical divider between the two fans?

not the full length to the heatsink tho, but about halfway. that would stop the fans blowing into each other and may decrease the backpressure.

Click to expand...

I didn't think of that, but I'm not sure it would decrease the back pressure because the air would reflect off the wall and create a pressure of its own resisting the air coming behind it. Also the divider would flap around and probably vibrate with considerable noise. Still there is no better teacher than experiment so when I take my rig apart again I will try the 1/2 divider idea.

Both fans have RPM sensors and I didn't get any drop in RPMs with them hooked together. Normally as the back pressure builds the fan slows down. So from that I figured the effect of the other fan wasn't significant.

I thought the V fans would be pretty effective, but it didn't turn out that way. I think the main reason it didn't work better was the distance of 1/2 of the air flow from the HS surface. Air speed slows down considerably for each centimeter you are away and the top of the fans and where lots of air normally comes in at the top of the fans is about 9cm away. So I think what I gained in air volume from two fans, I then lost because most of the air beyond the hubs wasn't really contributing much. That seems to show up because when I used my dual parallel fans which sit within 2mm of the HS surface I got the best cooling by 2c (and that is with only 1/2 the air).

Still I'd like to see two 80mm tornadoes hooked in a V and tried on an sp-94. Probably not worth the noise.