Blowers vrs. fans?

[I.M.O.G.]

I still have to insist;

Most of the blowers air will escape the easiest path out of the heatsink. (see pic at page 1 of topic, drawn as thick lines representing more air)

It's just like water, just like electricity, where the resistance is the smallest, thats where you'll find the highest flow.
For freakin' sake, take two straws of different diameters or lenght, blow into them, now tell me, where does it flow the most???

So, if you stick a blower(radial fan) ontop of a swiftech porcupine heatsink, you'll get most of the air escaping really close to the sides, just as "normal" axial fan would push the air.

The air doesn't want to force itself down to the bottom of the heatsink and then veer off 90 degrees and take an even longer route out of the heatsink.

Ofcoz, some miniscule amount of air will reach down there "in the venerable dead spot", but most of it will be trapped in vortices(as in vortex, I hope). My "?" in the picture posted above by me is where those vortices rule. We all think/know that vortices are good for heat exchange, but these ones are trapped, and not much air is rushing by to the rescue.

In my picture you can see thick lines representing where more air pass by... the thinner lines, the less air I predict will pass by...

Yes, the cooling of the dead spot IS BETTER with a blower, but so far as for practical solutions go, it has been a too high trade-off in CFM to make the idea work in practise.

Oh yes, radial fans has a potential to create better pressure, but that needs either high rpms(we know how that sounds) or bigger diameter paddle wheels(we all know already the devices are clumsy "as is").

Another factor speaking FOR AXIAL fans is the air coming of the blades is turbulent, and that turbulent air hits the fins/pins and then escapes the easy way to the sides.
The air from the radial fan is much less turbulent, of coz it might GET turbulent passing the fins/pins... but from the get-go the axial air is more turbulent.

Those appear as some mighty large assumptions, but knowing your posting I expect you have some reasons for thinking such besides just general theory.

Good quality blowers have been shown to perform quite well also... I don't know what you mean by "as for practical solutions go, it has been a too high trade-off in CFM to make the idea work in practise". Look at Hoot's article I linked earlier - and my current case and PSU has no fans but my centrifugal blower - there is no axial fan that can do that at any reasonable noise level.

Relative to Axial fans, even moderately sized wheels in centrifugal blowers produce much better pressure. They are clumsier than axial fans though, I'm right with you there - they definetly require some creativity to implement a real good blower elegantly.

Turbulence created by the fan has very little do with thermal transfer by air in these applications - it is different than water in the respect that we are dealing with volumes in the range of 70 cubic feet per minute easily with blowers. Turbulence helps with heat transfer with water in our applications because it gets more molecules involved given the same reasonable flow rates that are typical - but dealing with such large volumes of air, turbulence is much less necessary. Air flows freely enough that turbulence is neglible relative to total flow volume.

I'd like to know where your coming from with your view, and if you have links or anything you could refer me to so that I can understand it a little better.

 

[Paxmax]

Thank you Vonkaar for that intresting read.

While you said "Read around pages 15-25... air volume at various sensors..." Am I to understand that you mean "object temperatures at various sensors(points) during variable air volume" ?

From what I gathered from that document is that the hot-spot of a heatsink is really small in their simulation.
I mean, if the distance between sensors 1(above heatsource) & 3 where just 6mm and had a temperature difference of 8.5C. After that it didn't matter much. Distance between sensors 3 & 5 is roughly another 15mm and that temp difference between them was only 1.5C.

For me I interpret that as:
(First of all, this is not the situation I'm talking about, this simulation is about blowing air in another direction through a heatsink.)
For any fan to make a difference on covering the dead-spot it better deliver air pretty close to the heatsinks hotspot.
So for me, unless you move some serious air within the vicinity of 6mm from heatspot center you haven't covered the dead spot much...

If you move more air in the vicinity 6mm to 15mm from center thats great, but it won't make too much of a difference.

This document truly describes why WW and such are a success, it delivers coolness where it counts(ok ok, it benefits from a few more factors than just cold water close to the heat souce).

You(Vonkaar) mention "Air will go where it's placed (within at least a minor degree of error)".
Yes, I agree with that. An axial fan discharge it's are right off the blades, a radial fan discharges it's air from the paddlewheel.
But that "placement" you want would assume/need an infinite amount of nozzles right above heatsink. Yes, the flowpath is dictated by more than flowresistance alone, we have the momentum of the air as well.

So in the case of the radial fan, you have a lenght of tube where the air in the middle slows down slightly right after the discharge(before it even enters the heatsink) if the blowers output is put ontop of a standard pin heatsink. The closer the air stream(in the middle) gets to the bottom of the heatsink the slower will the air move, since it gots a much longer way to go, as compared to the air at the edges of the heatsink.

I can admit that I can't percieve how big the slowdown of air before it enters the heatsink is.

Ehmm... I'm sorry I can't follow your thinking regarding:
"If your model of physics were correct, passive heatsinks would do very little in a computer. If the airflow 'dodges' the paths of 'higher resistance', or becomes trapped in vortices, the cooling would be limited. The airflow (even in an open system) *will* explore all venues of escape."

I'm not sure/haven't thought about passive cooling or drawn an parallels.
-------------------------------

Hi IMOG!

Oh yes, these are mighty large assumptions!
Hopefully I'm not making a total *** of myself.

I couldn't agree more with you when you say "Good quality blowers have been shown to perform quite well"

And by good quality I would refer to bigger radial fans, since I feel that this world is short of good small radial fans.

I'm sorry for leaving that "practical unit" saying in the wide open, I'll explain it at the end of my post, my bad, sorry.

When you say "there is no axial fan that can do that at any reasonable noise level", I feel that that area has't been explored yet, most ppl never tried 150/180mm fans at reasonable rpms, neither have I so I can't really comment on that.

I don't even know how to compare radial fans with axial fans.
What should be the decisive point?

What matters the most??
Volume of air vs. noise
Volume of air vs. size of fan
Pressure vs. size of fan
Presure vs. noise

Not an easy task I say...

You can fit at least two axial fans of equal diameter(of a blower) in the same space that a radial fan occupies, since blowers are a bit clumsy.

Sure thing ducting will be a bigger hell with two fans but are you sure one blower will outpace two axial fans in terms of volume of air and noise? I know I'm not sure either.

Pressure, well, there's a race that surely will favor the radial, but you have to take to account that by using two axial fans they will shift posistion on the P/Q curve to maybe tie a radial.

--------------------
I thank you both for your constructive critisism to aid my thinking.
All of it is from my part PURELY my speculations I have yet no scientific data to cover anything I've said. All is derived from my common sense if you will, and toying around with fans of various sizes/types.

Maybe I should try to set up a heatsink with tiny flappery papers to simulate lack of flow in the middle above heatsource?
It would also be intresting to setup one or two 180mm fan with some ducting, just to see what it can do.
I'm not promising I'll do any of that though.

I too belive more air will circulate and/or touch the center of heatsink with a radial fan, but I honestly don't belive it will make THAT over-the-top everone should get it kind a difference.

Big radial fans has an advantage.
They can create a good pressure/volume with relatively low rpms.

When it comes to small practical units, where I consider 92mm axial fan a practical unit. I also consider the "aero blower" a practical unit, however it's performance is a practical joke.

It sucks 3 times the power, it delivers 1/2 of a flow at 3.5 times the size of an ordinary 80mm fan.

At some point it might best one 80mm fan on the P/Q curve.

So, if you got the space and need more air than a couple of 120mm fan can deliver than a radial fan might be the thing.

As for practical heatsink solutions I cannot percieve yet that a blower will excel. Radials are more a "total system" cooler.
With some smart thinking (like Hoot) you've got a killer total system cooler that replace alot of fans.

 

[madcow235]

Im not sure how the air wouldnt take the shortest distance, straight down, and why it would veer off to the sides, requires something to stop it or change its direction.

 

[cstarritt]

Pressure, well, there's a race that surely will favor the radial, but you have to take to account that by using two axial fans they will shift posistion on the P/Q curve to maybe tie a radial.

the only way using two axial fans could possibly raise pressure is if you stack them on top of each other. even then, i dont think you would get much of a rise in the pressure anyway. and even then, the only way to really get a good raise in pressure of it is if the two fans are counter-rotating.


my personal opinion is that radial fans win for cpu cooling when ducting outside air, and axial fans win for general case ventalation where there is neglegable static pressure. that is unless you are using one big *** blower to draw air in AND force it out.

 

[Paxmax]

the only way using two axial fans could possibly raise pressure is if you stack them on top of each other.

Yes, that is correct if you talk about MAXIMUM pressure when it moves zero CFM.

However, where a fan usually land is somewhere on the volume/pressure scale.

Which means, if you have one fan that moves 12 cfm with the 1 psi pressure it creates, you might shove 18 cfm with two fans cause together they can create 1.5 psi.
Their max psi might be at 1.9 psi with zero cfm, and it wouldn't matter much if there were 1 or 2 fans shoving against a pressure gauge.