Radiator Fans: Push or Pull???

[SaFrOuT]

dr_sharp ur way in talking and replying is not something that is acceptable on this forum

please be sure to remember that we are here to exchange knowladge not to exchange rude language without any proof while there are lots of tests done to prove that pull is better than push on the Rad made for w/cing PCs

@ gothchick

i really laughed a lot @ this

Who p*ssed in your wheaties?

[dr_sharp]

Originally Posted by SaFrOuT

dr_sharp ur way in talking and replying is not something that is acceptable on this forum

Yes, I agree. I've just read alot of posts recently by people (not this thread) that are complete BS and know nothing of the subject they speak of. They just attempt to repeat what they've heard from *slightly* more educated people. I am not claiming to be an expert otherwise I would explain it myself. It is frustrating to pose a question and recieve "answers" that are contradictory to what I DO know. I have taken fluid and thermal systems courses, along with heat transfer courses but this particular question is not your everyday question and I do believe that other than careful testing, the only way to properly answer it would be to use a CFD simulation. From a thermodynamics standpoint, there is absolutely no difference between push and pull. From a heat-transfer stand point, it is too complex of a problem to simply "explain".

My apologies, especially to gothchic, I appreciate the attempt to help but I'm looking for much more in depth answer.

Additional Comment:

Originally Posted by Go_Big

okay so pull i can have the fans ontop of the rad pulling the hot air from it right? not blowing air on it cause that would be push right?

Right, that is actually how my setup will be.

I am going for pull because as most know, heat "rises", so pulling air through and out the top of the rad is a good choice.

[SaFrOuT]

glad to read ur reply dr_sharp

and anyway a user tested the diff between psuh and pull @ XS using a PA120.3 ( as far as i remember or PA120.2) and using diff kind of fans including extreme and silent ones and the general conclusion that pull is Slightly better than push when using low pressure fans ( i.e. YateLoons on MCR or BIP or PA series )

so whenever anyone ask that question first i prefer telling him that it depends on teh place where he will place his rad, and tehn that it is knwon that pull is SLIGHTLY better than push on silent Rad that don't need a shroud

[jackdog]

Well I can't quote you any engineering studies off the top of my head, however every refrigeration/AC condenser and evaporator I have ever seen uses the pull method. Go outside your house and hold your hand over your AC's fan and I will bet you a six pack of your beverage of choice it is pulling the air across the condenser.

Also there is a dead spot at the center of the fan blades and that is why you use a shroud. I have a lot of tech books at work and I will see if I can dig up something on fan airfllow theory.

[doktaphex]

What about using both methods, fans front and back, would this not be the ideal situation.

Dok

[jackdog]

looks like there is no hard and fast this is the best way. I found this so if you have one of these Thermochill models just follow their recommendations

http://www.thermochill.com/faq.html

now if I just find the recommendations for a 77 Bonneville heater core

[crazyjv69]

does your rad look like one of these jack?

http://www.completeradiators.com/sho...t=Heater+Cores

[WiCKeD]

dr_s use the spatial side of your mind and think about it for half a second. One class in college isn't going to make someone an instant expert on a subject. Turbulence may help increase the amount of heat that is carried by the air molecules, but all of the bouncing around causes a loss of energy and you lose back pressure. End result is less air being pushed off the heatsink and some of it remaining nearer the rad. In a confined space, that air is more likely to be drawn back into the fan intake. With a less restricted stream, more of the energy imparted is directed at replacing the air with cooler air.

Fact, bad cable management increases turbulence... as well as CPU temps. Why? Because more energy is being wasted trying to cool the whole freaking case than providing a cool directed flow toward the CPU.

[dr_sharp]

Originally Posted by WiCKeD

Fact, bad cable management increases turbulence... as well as CPU temps. Why? Because more energy is being wasted trying to cool the whole freaking case than providing a cool directed flow toward the CPU.

You are mis-using the word "turbulence"

Bad cable management simply blocks the flow. You can call it what you want but at low velocities such as inside a computer case, it will not cause turbulent flow... how about... it "impedes" flow. I'm not going to get into an argument about terminology here but I think that needs to be clear.

Additional Comment:

Originally Posted by jackdog

Well I can't quote you any engineering studies off the top of my head, however every refrigeration/AC condenser and evaporator I have ever seen uses the pull method. Go outside your house and hold your hand over your AC's fan and I will bet you a six pack of your beverage of choice it is pulling the air across the condenser.

Though that is a good argument, there could be many explanations as to why most systems are built that way... namely safety (if an object goes into the fan, you don't want to chop something up and then spew it into your system, you'd rather chop something up and throw it back out). In a push system, the air will have a given velocity that needs to be dissipated and different components in the system may impede the air coming in (after it has passed through the radiator). With a pull system, the air has nothing impeding it other than atmospheric pressure. Maybe I just answered my own question? Who knows. This is all completely theoretical and being an undergraduate I don't trust myself. Maybe I'll ask some of the phds at work tomorrow

[jackdog]

Originally Posted by crazyjv69

does your rad look like one of these jack?

http://www.completeradiators.com/sho...t=Heater+Cores

Yeah the bottom one. $19.95 at the local auto parts store. I am using a $2.99 120v fan motor and 5 inch blade from a drug store in the pull position. Will post a picture this weekend.

By the way good point on the saftey issue Dr., I will do some more research as soon as I get a minute.

edit- read throughthe section dealing with fan selection in Stoecker's Industrial Refrigeration Handbook during lunch ~ yawn~. In section 6.18 he addresses push vs pull on fan selection.

Draw through has the advantage of greater air velocity across the coil, also in a push through situation the motor heat will also be absorbed by the coil. But fin spacing, thickness of the fins, diameter of the tubing, number of passes, fan blade diameter and type, distance from blade to coil, velocity of the water through the coil, and a number of factors come into play also.

Bottom line is follow the recomendations of the radiator manufacturer or if that is not available then try both and see which arrangement yields the better delta t across the coil. After all experimentation and fun is what it is all about anyway, for me at least.

[Rocky Mt. Hi]

Most of us who have reared children have heard worse from our kids than from Dr. Sharp. While Dr. Sharp’s presentation may be stepping on toes, I am pleased that he is asking for the why’s and how’s.

I’d like to make some educated guesses as to why I think a fan should “push” rather than “pull.” If you believe that my reasoning is pure BS (“Blue Sky”), I’d welcome your explaining why I am wrong.

Turbulent Flow – If “pushing” creates a more turbulent flow, then that is better for heat transfer than a smooth flow. When there is a smooth laminar flow through a radiator, a boundary layer makes contact with the radiator and absorbs heat. This layer is most effective when it initially enters the radiator because it is cool at that point. As it moves through the radiator, it warms up and becomes less effective. With a turbulent flow, the same thermal conduction occurs with cooler layers of flow, not with a warm boundary layer. You will see the same consensus about turbulence in discussion threads pertaining to water blocks.

Dense Air – When a fan “pushes” against a radiator, it creates a zone of dense air in the radiator. Dense air can absorb more heat than thin air. A fan that “pulls” will create a zone of thin air in the radiator that is less effective in absorbing heat.

I have not had the college courses that Dr. Sharp has taken. Please explain why you think my “physics” is wrong. Thanks very much!

[xiphmont]

Originally Posted by Rocky Mt. Hi

Turbulent Flow – If “pushing” creates a more turbulent flow, then that is better for heat transfer than a smooth flow. When there is a smooth laminar flow through a radiator, a boundary layer makes contact with the radiator and absorbs heat. This layer is most effective when it initially enters the radiator because it is cool at that point. As it moves through the radiator, it warms up and becomes less effective. With a turbulent flow, the same thermal conduction occurs with cooler layers of flow, not with a warm boundary layer. You will see the same consensus about turbulence in discussion threads pertaining to water blocks.

This is a sound point; turbulent flow is better at heat transfer. Weighing against the exhaust turbulence (a plus for push) is the problem that the exhaust flow from an axial fan is not uniform; it tends to be in a tight ring along the egde of the exhaust and expands outward in a cone. So there is a substantial amount of wasted dead space. This is a negative-- does it outweigh the positive of turbulence?

Originally Posted by Rocky Mt. Hi

Dense Air – When a fan “pushes” against a radiator, it creates a zone of dense air in the radiator. Dense air can absorb more heat than thin air. A fan that “pulls” will create a zone of thin air in the radiator that is less effective in absorbing heat.

True, but the pressure [density] difference is nearly negligable in the case of a computer fan. I doubt it makes a detectable difference.

Originally Posted by dr_sharp

Maybe I'll ask some of the phds at work tomorrow

They'll either tell you alot of fascinating things that won't actually answer your question, or [if they're the kind of guys that have to build things], they'll tell you to go test it. if you're lucky, maybe one of them has run the test already, but that's the only way a true engineer would solve the problem. Even complicated practical predictions are built primarily on empirical results. I don't know the answer either, but I am tempted to go test it out....

Monty

[dukes]

wont a fan push more cool are than pull warm air? Since cool air is more dence? /shrug what do i know tho i am just a carpet layer ((=

[xiphmont]

Originally Posted by dukes

wont a fan push more cool are than pull warm air? Since cool air is more dence? /shrug what do i know tho i am just a carpet layer ((=

heh, hadn't thought of that, but you're right about that too :-)

Monty

[SaFrOuT]

that si true if the diff in temp between the pulled and pushed air is big enough but in w/cing the air pushed or pulled through the rad the delta between its temp i don't think that it wil be more than 1C

maybe i am wrong for sure but i don't think so

[u4ikstilz]

I think it varies w/ rad and fan.
It's always best to try both out for yourself.
I think I had mine push when I had my BIX.
And I tried both ways, and found no difference.

[xiphmont]

Currently testing heat transfer of push and pull, with and without shroud. No reason to keep guessing ;-)

Monty

[jackdog]

Originally Posted by xiphmont

Currently testing heat transfer of push and pull, with and without shroud. No reason to keep guessing ;-)

Monty

Best idea I have heard yet.

I did a bit of research into this and in Stoeckers Industrial Refrigeration Handbook the only point he made was that in a refrigeration application the motor heat was absorbed in a push configuration and in a pull situation you would achieve more throw from the fan. Neither of which we would give a darn about in our application, in fact the heat transfer from the motor would be a negative in the equation although the wattage would be so minimal from our tiny fans it would be enough to matter.

I also asked one of our local tech college AC&R instructors and he cited the reason for most AC condenser fans being of the pull variety was the noise factor. As pointed out by others in this thread push will give more turbulence and turbulence = noise. Although I suspect the motor heat comes into play there because a one or one and a half HP motor can get a bit warm and generate a few BTU's which you would not want in a condensing application.

[xiphmont]

The y scale is 'temperature', so lower is better. As you can see, push and pull are so close together as to be about equal to the margin of error of the test. The curve fits reported the measurement error asymptote to be within about .75% of fit.

The rouded, normalized efficiencies (higher is better) with 1.0 == no fan were:

No fan at all: 1.0
Push with no shroud: 9.3
Push with shroud: 10.6
Pull with shroud:10.4
Push and pull fans, both shrouded: 13.1

That's right, running push + pull fans gets you about 25% improvement over running only push or pull. Not using the shroud (really, all it is is an airtight spacer that helps eliminate the dead spot directly behind the fan) loses you about 12%.

Monty

[dr_sharp]

Originally Posted by xiphmont

The y scale is 'temperature', so lower is better. As you can see, push and pull are so close together as to be about equal to the margin of error of the test. The curve fits reported the measurement error asymptote to be within about .75% of fit.

The rouded, normalized efficiencies (higher is better) with 1.0 == no fan were:

No fan at all: 1.0
Push with no shroud: 9.3
Push with shroud: 10.6
Pull with shroud:10.4
Push and pull fans, both shrouded: 13.1

That's right, running push + pull fans gets you about 25% improvement over running only push or pull. Not using the shroud (really, all it is is an airtight spacer that helps eliminate the dead spot directly behind the fan) loses you about 12%.

Monty

I'm confused as to how you were able to achieve the initial set temperature (120F) for each test. Did you just shut the pump/fan off and let the processor heat up? Aslo, you never stated what temperature we are looking at. I assume this is just cpu core temp?

In any case, nice work. I think we can definitly conclude that push+pull is superior. There is still no evidence, however, that only pulling is better than only pushing. In fact, if we were to attempt to draw conclusions, push actually looks better in these tests

Thank you for your efforts

[edit]

Originally Posted by Rocky Mt. Hi

Turbulent Flow – If “pushing” creates a more turbulent flow, then that is better for heat transfer than a smooth flow. When there is a smooth laminar flow through a radiator, a boundary layer makes contact with the radiator and absorbs heat. This layer is most effective when it initially enters the radiator because it is cool at that point. As it moves through the radiator, it warms up and becomes less effective. With a turbulent flow, the same thermal conduction occurs with cooler layers of flow, not with a warm boundary layer. You will see the same consensus about turbulence in discussion threads pertaining to water blocks.

I'm slightly confused about what you are trying to say but it sounds like you are confusing two different topics. Turbulent vs laminar flow and developing vs fully developed flow. At no point through the radiator will the flow be fully developed, it will always be developing (because the radiator is so shallow). Develeloping flow does have a higher temperature gradient than fully developed flow and in turn more heat transfer. I think this is what you're getting at?

Also, I think you mean convection. Moving fluid is convection... only solids or stationary fluids have conduction.

Originally Posted by Rocky Mt. Hi

Dense Air – When a fan “pushes” against a radiator, it creates a zone of dense air in the radiator. Dense air can absorb more heat than thin air. A fan that “pulls” will create a zone of thin air in the radiator that is less effective in absorbing heat.

The fan pushing air into the radiator will create a slightly higher pressure. This pressure could either compress the air and make it more dense or it could heat up the air. You will probably see a combination of this. As someone else stated though, this is virtually negligable with the fans we use on our pcs.