Sign in with
Sign up | Sign in
Your question

Does Push/Pull System in CPU heatsink work?

Last response: in Overclocking
Share
September 14, 2009 5:54:35 AM

Lots of people seem to think that push/pull system, meaning one fan sucking the air into the cooler's fins and the other fan sucking the air out of the cooler's fins, results in lower temperature than single fan system. Is this true and can it be proven? IMO if both fans spin at the same RPM there should be virtually no difference on the temperature of the heatsink. The only difference would be the strain on the fan itself. :o 
a b à CPUs
a b K Overclocking
September 14, 2009 6:01:32 AM

I believe the pressure of the two would make the temps better. It does help the temps because ive tested it, i just dont know why.
a b à CPUs
a b K Overclocking
September 14, 2009 7:40:05 AM

The second fan may help the first overcome resistance and turbulance in the radiator fins, or it may suck in additional air from the sides of the radiator. Depends on the design.
Related resources
a b à CPUs
September 14, 2009 7:52:14 AM

air simply does not want to go through the fins so easily - the second fan pulls air from between the fins etc helping the fan push air into the fins etc and extracts the air flowing through the fins

simple test - blow air at your fingers while leaving small gaps in between - the air flow behind your fingers is perhaps 1/10th the original air speed etc
a b à CPUs
a b K Overclocking
September 14, 2009 7:52:49 AM

When I checked fan rpm on my system the dual fan configuration operated at slightly higher rpm than a single fan.

September 14, 2009 8:40:37 AM

Using an axial flow fan it is optimal to use a fan with eight or more blades with bevelled edges and curved tips. This is the most efficient style PC cooling fan. It's a lower speed fan (1800 rpm or less) that produces less noise while moving the same volume of air compared to higher speed fan (2400rpm or higher).
As well a fan with a sleeve bearing is preferred because it accommodates the thrust effect when the fan is under load. (The fan is under load when it pulls air across the heat exchange surface.)
If this is for your cpu heatsink this style of fan is found on the more expensive ($50 or more) cpu coolers.

On a side note: Sleeve bearings have a lower MTBF (Mean Time Between Failure) rate compared to roller/ball bearings. In other words, ball bearing have a longer working life compared to sleeve bearings. However there are hybrid sleeve bearings called fluid bearings or fluid magnetic bearings that have longer working life than roller/ball bearings. Fans with this style sleeve bearing are found on more expensive cpu coolers.
a b à CPUs
September 14, 2009 9:48:29 AM

With a single fan, wouldn't the air at some distance from the fan slow down or scatter? I think with a dual fan setup you get a sort of tunnel of wind, unlike a single fan wherein you get a sort of inverted funnel. At some point with a single fan, the air would slow down or scatter.

Though I don't have wind tunnel or CFD data to prove this.
a c 172 à CPUs
a c 197 K Overclocking
September 14, 2009 2:12:17 PM

I tried it once (Q9550, TRUE, two S-Flex SFF21F fans, Antec 900 case, HSF oriented vertically). No significant improvement in my case. May have been because the top fan was pretty close to the big case fan.

Of course, YMMV.
a b à CPUs
a b K Overclocking
September 14, 2009 5:03:21 PM

jsc - About 4 or 5 years ago there was an article about case fans done by aeronautical engineers who work with wind tunnels. The distance between the second cpu heatsink fan and an exhaust fan on a case panel needs to be at least 1 to 1.5 inches to be effective. It's darn near impossible to get the second cpu fan and the panel fan in synch. When they are too close to each other and are out of sync the extra turbulence is counter productive..

Wish I had bookmarked the article. I stumbled into it accidently and now I can't remember where it is located or how I found it.
a b à CPUs
a b K Overclocking
September 14, 2009 5:08:06 PM

Makes sense. Rotating the heat sink 90 degrees in a normal case is not an answer either. So maybe its only effective with a weak fan on the exhaust fan side of the heat sink. Or with the cooler rotated 90 degrees in a case with a top exhaust and some headroom.
September 14, 2009 8:43:24 PM

Lets make it more specific, instead of thinking of the effect of other fans on the system, lets consider a radiator outside of the case with 2 fans pushing air through the fins or 4 fans in push pull configuration. Clearly in this case no external influences would apply (except ambient temps of course). I know from turbine design fans placed concurrently next to each other at specific degrees will yield greater air pressure, but since I have no control over the fans turbine-like set up is not achievable.
After running the push /pull set up I would say I've experienced maybe 2-3C (negligible) improvement over the single fan set up. If anyone has a push pull system running please submit your single and push pull temps.
a b à CPUs
a b K Overclocking
September 14, 2009 8:59:33 PM

This review of one such multi-fan-capable cooler shows that the gain may depend on fan speed . . . and if one fan is fast enough, the second may not matter (ex/noise).

http://www.legitreviews.com/article/1053/1/

I'm sure you can find other examples and comparisons by googling "Scythe Mugen 2 review", and any other cooler that's 2-fan capable.
a b à CPUs
September 15, 2009 1:22:28 PM

I tried it recently with my Core Contact 92, the most difference I got is 2C. The system seems to work, but it probably doesn't warrant the power consumption of a second fan.
a b à CPUs
a b K Overclocking
September 15, 2009 2:44:46 PM

Yes, the use of mutiple cpu heatsink fans will not produce spectacular results. A difference of about 2C or 3C is all that one should expect. For the typical user it probably won't make any difference. For a hardcore gamer it might.
August 18, 2011 5:53:02 PM

I have also tried dual fan config and i got the same results, 2-3c lower temp. But due to the power consumtion of the second fan, didnt make a blind bit of difference. :??: 
a b à CPUs
a b K Overclocking
August 18, 2011 6:04:39 PM

Necro thread FAIL
December 22, 2011 10:11:06 AM

Adding a second fan creating a push/pull system could yield as high as 5-6C decrease in temperature depending on the original size of the heatsink and the cooling air flow or air currents (directional flow patterns) inside your PC along with the rooms ambient temperature at the time. If you set up the airflow diagram with enough input fans to match the output fans and also you should have some sort of graphics card cooler or fan blowing on that card, you will get the temps down. Obviously larger cases with more fans are better for cooling. At an Idle, you should gain 2-3C, but when heavy gaming or 100% load using Ortho Stress Test, or some other program, that's when you will see you have better cooling capability using the second push/pull fan. That's when you need it. The other advantage is that if you get lazy as far as cleaning up the dust inside occasionally from the heatsink, fan filters, etc...and your air flow along with cooling has been diminished, your extra fan then is providing you just a little bit of extra CPU temperature insurance. IMHO Niki ***AIRFLOW DIRECTION & WIND TUNNELS ARE MY LIFE***!!!
February 1, 2012 9:37:50 PM

As you are an expert,

if by example you have a rectangular heatsink with the fins alongside the longest part of the block. you place a fan on one end to be able to push or pull some air in the heatsink... if the heatsink and the fan are in a tight enclosure with holes at both ends... is it better to push or pull the air in the heatsink?

Regards

nikimarie said:
Adding a second fan creating a push/pull system could yield as high as 5-6C decrease in temperature depending on the original size of the heatsink and the cooling air flow or air currents (directional flow patterns) inside your PC along with the rooms ambient temperature at the time. If you set up the airflow diagram with enough input fans to match the output fans and also you should have some sort of graphics card cooler or fan blowing on that card, you will get the temps down. Obviously larger cases with more fans are better for cooling. At an Idle, you should gain 2-3C, but when heavy gaming or 100% load using Ortho Stress Test, or some other program, that's when you will see you have better cooling capability using the second push/pull fan. That's when you need it. The other advantage is that if you get lazy as far as cleaning up the dust inside occasionally from the heatsink, fan filters, etc...and your air flow along with cooling has been diminished, your extra fan then is providing you just a little bit of extra CPU temperature insurance. IMHO Niki ***AIRFLOW DIRECTION & WIND TUNNELS ARE MY LIFE***!!!

November 3, 2012 8:18:25 PM

I just installed a second fan on my hyper 212 and it dropped temps 5C, which I think is pretty good
!