a Pipe with a fan will it help ?

[H0nDA2000]

ok i saw this thread on a forum with a guy saying that what he did:

made a reall iprovment
i thought i try that .. i just whant to ask what if i put a fan blowing air in the pipe to the cpu fan ? so the air that the cpu fan uses is cold and new cz it comes from outside the case ?
here a quick draw i made ( am not that good at it )

thats the back of the case where i have 2 fans blowing out .. and i thought to put the one outside the case blowing IN tha case TO the pipe ...
your thoughts plz

 

[Blue]

Don't think that would work the same. It would prob. cause your CPU to overheat. The air would bring fresh air in but then would just circulate around the tube choking the CPU with hot air.

 

[JTurton]

That will work well, ONLY if you have more CFM pulling out than is being pushed in. If more is being pushed in you will have the problem blue speaks of which is gone over in the cooling 101.

Quote:
Positive Air Pressure Cooling: This is a type of cooling setup where you have more air (CFM) moving into your case than being pumped out. This creates heat pockets where the hot air can’t get out and just serves to heat your case up. Unless you are using a special cooling system, don’t use this.

You want this instead (from cooling 101)
Negative Air Pressure Cooling: This is the most common cooling setup. This is when more air (CFM) is pumped out of you case than pumped in. This creates a vacuum effect where hot air is sucked out and the inside temp stays pretty stable. A guide http://www.xoxide.com/computer-cooling.html

 

[dragon2309]

I think your intake and outtyake fans should swap places..... anyone agree??

 

[joelkyr]

this might help u
http://cache-www.intel.com/cd/00/00/06/20/62085_62085.pdf

ducting your CPU would definitely help(fresh air), Positive and Negative air pressure for me is a case to case basis, u have to consider the outside temp too.

 

[JTurton]

I think your intake and outtyake fans should swap places..... anyone agree??

Depends on which way the cpu fan is blowing. You don't want the 2 fans fighting each other.

 

[H0nDA2000]

That will work well, ONLY if you have more CFM pulling out than is being pushed in.

ok i think i got that ... my current setup.. i have 2 side fan blowing air in and 2 at the back blowing out...SO after reading the site u gave me ... if i make the one of the side fans blowing out will kind of solve this i think (although blowing out from the SIDE isnt so effective )

<dragon2309>.. well i was thinking that Cold "new" air from outside should be blown in the pipe

<joelkyr>.. thx for the link .. but i dont think it would be nice doing that coz i have a window on the side panel.. and i am not able to cut it

thx

 

[Yeti]

It will work better if there is an opening near the heat sink (or if the heat sink isn't covered) or else it'll basically be like Blue said.

That will work well, ONLY if you have more CFM pulling out than is being pushed in. If more is being pushed in you will have the problem blue speaks of which is gone over in the cooling 101.

Don't believe everything that you read What about the conservation of mass? (yes there will be a slight increase in density due to temperature rise, but think of it in the limit that there is no heat generation)

Positive vs negative pressure cooling depends on many many things... which is the reason companies like Fluent can charge exorbitant amounts of money for fluid modeling software. In this case it would be better to use "fresh" intake air to get a bigger temperature delta across the heat sink.

 

[dragon2309]

<dragon2309>.. well i was thinking that Cold "new" air from outside should be blown in the pipe

<joelkyr>.. thx for the link .. but i dont think it would be nice doing that coz i have a window on the side panel.. and i am not able to cut it

Whats all that about, i dont remeber saying that.

 

[H0nDA2000]

It will work better if there is an opening near the heat sink (or if the heat sink isn't covered)
In this case it would be better to use "fresh" intake air to get a bigger temperature delta across the heat sink.

ok i can understund about the opening.

[In this case it would be better to use "fresh" intake air to get a bigger temperature delta across the heat sink.] <== can u explain me this ? what do u recommend me to do ?

 

[Yeti]

[In this case it would be better to use "fresh" intake air to get a bigger temperature delta across the heat sink.] <== can u explain me this ? what do u recommend me to do ?

It's just my convoluted way of saying that it's better to use the ambient (as in outside the case) air as it is colder than the air inside the case which has been warmed by the various components (case temperature is generally a few degrees warmer than ambient).

What you are proposing to do is simply build an air duct. As long as you don't cover the heat sink (ie leave a place for the air to exit) and have both fans positioned to blow air onto the heat sink, it should work just fine the way you have drawn it .

 

[vortmax]

pulling fresh air across the heat sink is the best case senario. You need to think of this tho:

The heat sink fan is pulling air up from the board and out the top. For maximum cooling you want to go from turbulent flow to laminar flow. Air across the heat sink fins will most definatly be turbulent while coming out of the fan it will (quasi)laminar. Laminar = smooth. Going in reverse creates all kinds of eddies and vorticies which will trap air and not cool as efficiently.

You want to draw cool air across the fins as cooling efficiency is dependant on the temperature gradient between the air and heat sink. Remember tho, that the heat sink is significantly hotter then the outside air AND the air in the case. If you have proper chassis fans, then the case air temp will not differ from the outside air temp nearly as much as both varry from CPU temp, so it's not going to make a huge difference if you pull in air straight from the outside or just out of a well ventelated case. IF you are already having problems or are overclocking, then yes it might.

So idealy you would have a double pipe design. One pipe which is larger then the heat sink and another that is the exact size of the heat sink fan. The small pipe goes inside of the larger one. This completely thermally isolates the CPU from the case. Only problem is mounthing as the setup is quite large. A more efficient version would have shaped plumbing at the CPU interface to eliminate the 90* bend, as air doesn't like to flow like that, but running a negative pressure system like this should give some good cooling. You might also want to do some calcs to figure your outer pipe diameter so your CFM in = CFM out.

 

[Yeti]

The heat sink fan is pulling air up from the board and out the top

That's not necessarily true, it depends on the setup. In my previous post I was assuming the opposit (fan pushing air through heat sink)

For maximum cooling you want to go from turbulent flow to laminar flow.

...I understand laminar and turbulent and all that, but why will a transition from turbulent to laminar be advantageous?

Going in reverse creates all kinds of eddies and vorticies which will trap air and not cool as efficiently.

To be fair, it is eddies and such that make for the better heat transfer coefficient in turbulent flow. If everything is set up/designed correctly there shouldn't be much or any inefficiencies in reversed flow.

 

[Archangel]

well.. as far as i know, you have to push air towards the heatsink.. because if you pull air, it will only cool the outer edges of the heastink. ( so the cooling would be pretty crappy. )

this is about what i mean.

also.. if you make shure the air has an easy exit after it reached the heatsink.. a duct towards the CPU-Fan will help.

 

[vortmax]

when you are pushing air from laminar to turbulent, the frictional force causes convergence. This makes the heat sink a big limiting factor. Think about trying to fish a cable through something. Pushing the cable causes it to bunch up, while grabbing the end and pulling is much easier. It's about the same thing.

eddies do make for good cooling because they maximise the time air is in contact with the serface, but you have to be able to shed the vorticies, or you just have air trapped against the sink, storing heat. Vorticies don't like to shed unless they are embedded in a laminar flow. You will still create eddies and vortices with a sucking fan, but you will also ensure a more laminar flow profile after the heat sink.

i'm not sure about the outer edges of the heat sink seeing peak plow when pulling air. Maybe it would be better to draw air across the heat sink and not normal to it.

Anyway, I think we are debating some pretty fine points. I imagine any of these designs with a 120mm fan would cool a processor within anybodies tolerances short of water cooling.

anyone have access to a fluid dynamics model we could run this through?

 

[Yeti]

when you are pushing air from laminar to turbulent, the frictional force causes convergence. This makes the heat sink a big limiting factor. Think about trying to fish a cable through something. Pushing the cable causes it to bunch up, while grabbing the end and pulling is much easier. It's about the same thing.

But when you're talking about incompressible duct flow it doesn't really matter - like taking your fishing cable and putting it through a small tube (similar to bicycle brake cables). And what do you mean by friction causing convergence??

eddies do make for good cooling because they maximise the time air is in contact with the serface, but you have to be able to shed the vorticies, or you just have air trapped against the sink, storing heat.

I was under the impression that it was the better mixing in the boundary layer that was responsible for the better heat transfer.

anyone have access to a fluid dynamics model we could run this through?

I have access to Fluent and COSMOS Flow Works (Solidworks)... if I actually will have time to run a simulation is probably a better question

 

[vortmax]

but this isn't incompressble duct flow. You will have compression. It may be small but it's still there. Friction causes convergece. Friction serves to slow the air down. Fast moving air hitting a region of slow moving air creates velocity convergence

we are both saying the same thing in the second quote, just in different words. Mixing is caused be turbulent eddies, but they still have to shed. Think about a stream in the fall. You see larger rocks where leaves get trapped in the eddies behind them. In a non shedding eddie, those leaves stay trapped for a long time. If the eddie sheds, then they are free to move on. Imagine the leaves as heat leaving the rock. Thermal transfer by molecular diffusion is slow, inefficient and needs a high temp gradient. If you had a non shedding eddie, it's temp would rise and it would spread heat to the passing flow, but not as well as the heat sink passes heat to it. If that eddie sheds, then that pocket of hot air moves down the flow and it is replaced with cool air. That is boundary layer mixing on a finer scale. Without shedding you have low BL mixing.

 

[Yeti]

but this isn't incompressble duct flow.

Oh, but it is . Pretty much any fluids class/text will define incompressible flow as u<<c (where u is the fluid bulk velocity and c is the soundspeed). Taking a 12 cm duct with an air flow of 40 CFM you only get about 0.5% of the speed of sound (770 mph for air at 295K).

Okay, I get you with the boundary layer and vortex shedding stuff. But I still don't see why you would want to go from turbulent to laminar flow.

For maximum cooling you want to go from turbulent flow to laminar flow

 

[vortmax]

i take it you're an engineer. I'm at atmospheric scientist and yes, we treat air as compressable even though U is very much less then C. I would love it if you would call some of my old profs and tell them that the atmosphere is noncompressable.

 

[Yeti]

i take it you're an engineer.

Yep

I'm at atmospheric scientist and yes, we treat air as compressable even though U is very much less then C. I would love it if you would call some of my old profs and tell them that the atmosphere is noncompressable.

Why/under what conditions do you need to consider it compressible? As far as I know, and I'll admit that fluids isn't my strongest subject, engineers approximate it as incompressible to save on calculation time since it would make little or no difference in accuracy.

@H0nDA2000 - have you built or decided on how to build your duct system yet?