Push-Pull design with Rads. Why?

[kohlersc]

There are a number of things I don't understand well and this is one of them.

What is the full purpose of a push-pull fan design around a rad? If anyone can even just point me to a post about this that would be great.

Also if using a push-pull design do you need to use a shroud? What are the benefits.

I guess I am just looking at a run down of how it compares to just running fans on one side of a radiator and how it necessarily helps and/or how much of a benefit you actually see from running this kind of set up.

[silverphoenix]

Only thing I see is if you have fans that put out less pressure it helps overcome that deficiency, therefore indirectly you achieve closer to the fan's rated CFM with the same restriction.

[cadaveca]

single-side fans airpressure is not constant within radiator. the fans must work more to push the rated CFM due to constriction of radiator fins. This slows the air...and we all know slower air yields less cooling.(assuming fans are sucking air through rad).

Add in the second set of fans, the fan blowing air in is not restricted as much at inlet, but is slightly restricted at outlet by radiator fins.

If the fans is blowing into rad, the fan must work against teh restriction of the rad to continue to push the air through the rad. The air slows as soon as it hits the rad, and by the time it exits, it's moving far slower than when it entered.

so, either way, air slowed down through rad.

It LOST energy to slow. It lost energy by friction...which creates a bit of heat. Although this AMT of heat given off is very small, it is detrimental to the air absorbing heat from the rad.


Add two sets of fans...the air should enter and exit at the same speed. you don't get twice the airflow from adding a second set of fans, but those fans will WORK LESS as the load from moving the air is spread out between them.


In the end, the air then needs to absorb energy to maintain the same speed...and lost some due to friction...where does that energy come from?

The heat from the water?

[Bun-Bun]

PUSH/PULL allows you to use quiter fans for the same objective or to turn louder fans down.

I have 114CFM panaflo's on my GTX240 rad and by adding a couple 70 CFM fans to the top I am able to turn the panaflo's down to 7v making the machine much quiter and keeping the same performance as just the panaflo's at 12v.

[The_Beast]

single-side fans airpressure is not constant within radiator. the fans must work more to push the rated CFM due to constriction of radiator fins. This slows the air...and we all know slower air yields less cooling.(assuming fans are sucking air through rad).

Add in the second set of fans, the fan blowing air in is not restricted as much at inlet, but is slightly restricted at outlet by radiator fins.

If the fans is blowing into rad, the fan must work against teh restriction of the rad to continue to push the air through the rad. The air slows as soon as it hits the rad, and by the time it exits, it's moving far slower than when it entered.

so, either way, air slowed down through rad.

It LOST energy to slow. It lost energy by friction...which creates a bit of heat. Although this AMT of heat given off is very small, it is detrimental to the air absorbing heat from the rad.


Add two sets of fans...the air should enter and exit at the same speed. you don't get twice the airflow from adding a second set of fans, but those fans will WORK LESS as the load from moving the air is spread out between them.


In the end, the air then needs to absorb energy to maintain the same speed...and lost some due to friction...where does that energy come from?

The heat from the water?

wow that is a great explanation

[cadaveca]

LoL i just hope it's right

[Movieman]

LoL i just hope it's right

I like a guy with a good sense of humor!
Your comments made sense and generally logic is correct.

[Vapor]

Stacking fans does increase CFM as well as decrease CFM-loss due to restriction.

e.g., a 100CFM fan on a radiator may only push (or pull) 75CFM through the radiator. That's a 25% loss.

Stacking two ideal (ignoring rotational direction for now) 100CFM fans may increase throughput to 125CFM. When this is put on a radiator, the fannage 'unit' can usually push nearly 105CFM through the radiator. This is only a 16% loss AND a higher 0-point. It does have a slight disadvantage when it comes to air dispersion patterns however.

Where this is useful is both in increasing performance and/or decreasing noise. For instance, if your goal is to have 75CFM pushed through your radiator, you can use a single loud 100CFM fan or two medium-noise 72-ish CFM fans. Two medium-noise fans are usually quieter and emit a lower-pitch sound as well.


DISCLAIMER: don't quote me on those figures, it GREATLY depends on the fans used and what I used is the best-case scenario I've seen. The way most people do push-pull doesn't give 105% efficiency (closer to 98% or so I'd estimate) since almost nobody does counter-rotation and even a radiator separating two fans doesn't reduce all rotational airflow patterns that counter-rotating fans do.

[kohlersc]

Thanks much. Those were some excellent explanations. As long as you guys are right it sounds good to me.

[MrToad]

Stacking fans does increase CFM as well as decrease CFM-loss due to restriction.

e.g., a 100CFM fan on a radiator may only push (or pull) 75CFM through the radiator. That's a 25% loss.

Stacking two ideal (ignoring rotational direction for now) 100CFM fans may increase throughput to 125CFM. When this is put on a radiator, the fannage 'unit' can usually push nearly 105CFM through the radiator. This is only a 16% loss AND a higher 0-point. It does have a slight disadvantage when it comes to air dispersion patterns however.

Where this is useful is both in increasing performance and/or decreasing noise. For instance, if your goal is to have 75CFM pushed through your radiator, you can use a single loud 100CFM fan or two medium-noise 72-ish CFM fans. Two medium-noise fans are usually quieter and emit a lower-pitch sound as well.


DISCLAIMER: don't quote me on those figures, it GREATLY depends on the fans used and what I used is the best-case scenario I've seen. The way most people do push-pull doesn't give 105% efficiency (closer to 98% or so I'd estimate) since almost nobody does counter-rotation and even a radiator separating two fans doesn't reduce all rotational airflow patterns that counter-rotating fans do.

Hmmm... informative, as usual.

Now, wouldn't making a counter-rotating fan involve changing the orientation of the fan blades?

How can this be done with an off-the-shelf product?

[CyberDruid]

Hmmm... informative, as usual.

Now, wouldn't making a counter-rotating fan involve changing the orientation of the fan blades?

How can this be done with an off-the-shelf product?

I have the same question. How does one make or find counter-rotating fans?

[Bun-Bun]

Hmmm... informative, as usual.

Now, wouldn't making a counter-rotating fan involve changing the orientation of the fan blades?

How can this be done with an off-the-shelf product?

There are counter rotating fans availible... they are just not popular so you don't have every make and model availible in counter rotating...

[CrazyJoe]

Check out this thread here - New Equipment for Cathar

He mentions in his first post about the differences in the push and pull effects of fans. To sum up, push allows for more flow, yet, pull has much more even flow across the rad. Push and pull gives you the best of both worlds.
It looks like Cathar is planning to show the differences for each in relation to differing shroud distances. This would be great information to have for future builds.

[Vapor]

I have the same question. How does one make or find counter-rotating fans?

Yes it does...unfortunately there is no way to do this manually. It means buying a fan that spins the 'wrong way,' of which there are only a few (Delta Tri-blade and some sub-mediocre Arctic Cooling models come to mind).

Using a Tri-Blade LE and a Sharkoon 2000 (very evenly matched at ~73.5CFM each) I got tremendous results. However, the Tri-Blade isn't the quietest offering for fans at that speed (though, it's not bad). I haven't done anything more than anecdotal testing of two non-ideal fans, but getting two of the best, even though they spin in the same direction, may end up have comparable air movement at a given dBA level with a radiator separating them.

[Wolfsburg18]

Ok, so it is clearly established that a two fan set up of push pull is best but then I would follow up with the question... Is it best to have the fans matched.

Ex: I have 3 Yate Loon med fans pushing through a PA120.3

Should I just buy 3 more identical Yate Loon fans for the top or should the fans on top have a higher CFM rating to create a suction effect through the rad?

~Wolfsburg
***Update to sig coming soon

[CrazyJoe]

I'd go with the same Yate Loons on both sides. No point in adding the extra noise and the PA's historically run great with the low cfm fans.

[Marci]

It means buying a fan that spins the 'wrong way,' of which there are only a few (Delta Tri-blade and some sub-mediocre Arctic Cooling models come to mind).

Panaflo and many other manufacturers do make counter-rotating fans, but, very few do so in a DC variant. I have a pile of metal-framed 120mm papst and panaflo 240v AC fans here from my phasechange days that are counter rotating.

[Bun-Bun]

Yes it does...unfortunately there is no way to do this manually. It means buying a fan that spins the 'wrong way,' of which there are only a few (Delta Tri-blade and some sub-mediocre Arctic Cooling models come to mind).

Using a Tri-Blade LE and a Sharkoon 2000 (very evenly matched at ~73.5CFM each) I got tremendous results. However, the Tri-Blade isn't the quietest offering for fans at that speed (though, it's not bad). I haven't done anything more than anecdotal testing of two non-ideal fans, but getting two of the best, even though they spin in the same direction, may end up have comparable air movement at a given dBA level with a radiator separating them.

I wonder, how would fin count per inch affect the rotational air patterns? Like a PA120.3 vs GTS360 (extremes on either end of the fin count)

[bobo5195]

single-side fans airpressure is not constant within radiator. the fans must work more to push the rated CFM due to constriction of radiator fins. This slows the air...and we all know slower air yields less cooling.(assuming fans are sucking air through rad).

Add in the second set of fans, the fan blowing air in is not restricted as much at inlet, but is slightly restricted at outlet by radiator fins.

If the fans is blowing into rad, the fan must work against teh restriction of the rad to continue to push the air through the rad. The air slows as soon as it hits the rad, and by the time it exits, it's moving far slower than when it entered.

so, either way, air slowed down through rad.

It LOST energy to slow. It lost energy by friction...which creates a bit of heat. Although this AMT of heat given off is very small, it is detrimental to the air absorbing heat from the rad.


Add two sets of fans...the air should enter and exit at the same speed. you don't get twice the airflow from adding a second set of fans, but those fans will WORK LESS as the load from moving the air is spread out between them.


In the end, the air then needs to absorb energy to maintain the same speed...and lost some due to friction...where does that energy come from?

The heat from the water?

Friction inside the rad is not important. Insignificant.

The best way to even pressure accross the rad is to place the fan further away the air will naturally find an equal pressure point. lower restriction or more precisely makes airflow more normal to the radiator plane reducing turbulence which lowers sound level. which is much much more important than the heat in relative magnitude. DIfferent air flow levels will also play havoc with my luvly radiator approximations.

A counter rotating fan is a fine idea as it creates higher efficentcy by tiurning the air one way than screwing it the other you gain relative air speed. Some helicopter

Push pull does allow you to cheat (in a good engineering sense, like when the y put vacuum cleaners under F1 cars to stick them to the track) and add two fans to the same rad area. Which lowers overall noise.

Fin density is a design variable you trade fan back pressure for increased thermal density.

[cadaveca]

Uh, friction is what allows the air to absorb heat from teh rad. The heat from the friction is insignifigant, for sure, but without friction, there'd be NO COOLING!!!!

[tiro_uspsss]

Uh, friction is what allows the air to absorb heat from teh rad. The heat from the friction is insignifigant, for sure, but without friction, there'd be NO COOLING!!!!

bulldust - according to ur statement, if u could magically take away that *very very very* small amount of friction that is occurring when the air passes, there would b no cooling - rubbish - there would still be cooling, as there is still contact between air & rad fins.
just like there is still heat transfer happening between ur cpu's ihs & hs/wb (& there is even less friction there - i dare say ) due to contact

friction isnt what allows it to absorb heat from rad, its contact.

[bobo5195]

Uh, friction is what allows the air to absorb heat from teh rad. The heat from the friction is insignifigant, for sure, but without friction, there'd be NO COOLING!!!!

I mean clearly you know more than the textbooks or the people who have spent not only their lives but also their pHds on the exceptionally hard core and misunderstood area that is boundary layer physics. If you are correct you surely up for a nobel prize. ( this is a joke as what you are saying contradicts a lot of current fluid mechanics as practiced by some very intelligent profs who could continually slap everyone here around till they admit they know nothing about fluid mech and are in fact idoits. If your insight is correct it would earn a Nobel and billions of dollars but I feel that some intelligent guys at Stanford/MIT/Imperial/Cambridge might just say no)

Friction is important but more of it in this context doesn’t mean there is more heat absorbsion. Continuum mechanics would say that more air flow velocity is,

Agreed without viscosity that there would be no heat transfer but beyond on that to understand what is going on you need a pHd as its exceptionally complex.

[cadaveca]

You can post your comments without the insults, thanks. I'm quite aware of reality. Did you not read my second post?

You just felt the need to repeat what dude above you posted, and add an insult...does this make you feel good?

I hope it did.

[Bun-Bun]

I mean clearly you know more than the textbooks or the people who have spent not only their lives but also their pHds on the exceptionally hard core and misunderstood area that is boundary layer physics. If you are correct you surely up for a nobel prize. ( this is a joke as what you are saying contradicts a lot of current fluid mechanics as practiced by some very intelligent profs who could continually slap everyone here around till they admit they know nothing about fluid mech and are in fact idoits. If your insight is correct it would earn a Nobel and billions of dollars but I feel that some intelligent guys at Stanford/MIT/Imperial/Cambridge might just say no)

Friction is important but more of it in this context doesn’t mean there is more heat absorbsion. Continuum mechanics would say that more air flow velocity is,

Agreed without viscosity that there would be no heat transfer but beyond on that to understand what is going on you need a pHd as its exceptionally complex.

Need a pHd? Hardly... I know people out in the feild who know more the guys with pHd's...

[Xeon th MG Pony]

Same, met a guy with multiple degrees in A/C yet I could run circles around him in both repair and design. Need more then just paper you need to actualy do it to understand it, you learn more by siting infront of some thing and playing with it then learning the theory then you will just learning either or.

Like a grade 12 Sceince teacher, he all ways insisted the thermometer was wrong, becuase he had a murcury one on a peice of aluminium, a metals surface will all ways be cooler then the surounding air do to convection and radient cooling Took him 5 years to admit he was wrong!

Paper/theory tries its best to immitate/model life, not the other way around!