Strange HSF performance w/ different FANS

[ku]

I thought I would try getting my CPU a bit cooler (have a Speeze Falconrock II), so I tried my 30cfm (really really old btw) case fan and put that on the heatsink first. It actually made it hotter by 2-3 degrees. Okay, I replace it with the original 28CFM fan and temps are back to the way they were. I figured I'd buy a 34CFM ball bearing fan, and pop it in. I got it, and guess what, 2-3 degrees hotter than the original fan! Don't really understand... why does fans that push more air perform worse than a fan that pushes less? Well, I compared the fans carefully and the only difference I can see is the shape of the blades. The original Speeze fan has straight blades that extend through the whole depth of the fan while the others don't. Could this be it? I don't know...

And btw, anyone wanna tell me what fans would cool better? Other than those crazy delta fans of course (nothing too loud)

 

[Zepper]

Back pressure from closed-bottom heat sink. The more CFM, the more back pressure. You ought to try a manually variable speed fan to see what the sweet spot is. Or even try reversing the flow direction with the faster fans - may need to seal the little space between the bottom of the fan and the top of the fins with duct tape or some such to make sure as much air is drawn thru the fins as possible.
. I thought about it some more and perhaps the air that is already between the fins gets trapped there and acts like a block causing more leakage thru the gap between the fins and the fan. Remember, air ( like water, electricity and students) takes the path of least resistence... That's why it's self-defeating to map out and build rectilinear (right-angled) walkways on campus. Best to just wait for the students to wear down paths and then put down paving on the most-used paths.
.bh.

It's Sunday, have some :wine: !

 

[Cerb]

Overall, axial fans suck at sucking.
So the fan comes in, sweeps air in and pushes it down and out.
But there isn't as much space to push down, so it doesn't push as much.
Various design factors (read: the way the fins are shaped, andf the motor used, but I'm not sure on specifics) can affect how well it works against restrictions. With a heatsink like that, you're now trying to push 100% of the air through a hole that now has 2/3 or less of the proper surface area.

Panaflo fans are well known to work very well in such situations, often besting the performance of fans with 20-50% more CFM ratings (even if accurate for free air) and power ratings.

CFM ratings can't be entirely trusted, like many fan dB ratings. The rated CFM is for free air. If you attach the fan to one side of some flat surface, and there are two separate volumes on each side, both open-ended, you'll get the rated CFM. Add ribbon cables, some drives, heatsinks, various cards in the way, and they won't quite stack up.

 

[Zap]

I've messed with fans also, using a Thermalright ALX-800 heatsink. I had four fans and the fastest RPM unit had the lowest temperature. Don't remember what brand/model, but it was some industrial looking 80x32mm unit off an old Swiftech Athlon heatsink. A close second was the slowest RPM fan. It was the Panaflo 12H. Third was a high RPM Sunon and fourth was an even higher RPM YSTech (though the industrial unit was highest RPM). Interesting to note that the rated RPM and CFM of the YSTech was highest, while it had the highest temperatures out of the three known fans. The blades of the YSTech were also the most curved.

My thought is that all else being equal, higher RPM = lower temperature, however fan blades can have a great effect on performance. I've noticed that some fans push air outwards more than others (imagine if airflow was like a cone, the farther from the fan the bigger dead spot in the middle). Those can have a very high CFM (perhaps being more efficient for that) while they are not good for use on a heatsink. That's why those TMD fans can work okay with lower RPMs and some of those "squirrel cage" blowers can do a decent job with really low CFM ratings.

 

[Zap]

Originally posted by: Zepper
Back pressure from closed-bottom heat sink.

That may be a very valid statement. Anyone remember the Alpha heatsink made especially for slotkets? The Alpha PEP66 series of HSF. I see that they've adapted it for other uses, but AFAIK the original market was for FCPGA slotket use. The heatsink used a bunch of really long thin fins and a shroud which kept airflow from exiting prematurely. A 60mm fan was used for the airflow with pretty much no back pressure because of the straight air flow path.