# CPU cooling - one fan vs. two?

## Main Question or Discussion Point

I have a question about fans and airflow, specifically related to processor cooling. I have a large heatpipe/radiator on my processor which came with one fan installed that blows air across the fins (a push fan). The assembly is set up to be able to add a second fan on the opposite side to pull air out of the radiator (push-pull configuration). They sell the cooler pre-installed with both fans for an extra \$20 or so. What I’m curious about is if this second fan is actually that useful, or is it mostly marketing hype?

My thought process goes like this:
The initial fan moves 120m3/h of air. If I put one of these on, it will push 120m3/h into and across the cooling fins. If I now put a duplicate fan pulling from the other side, it's going to pull that same 120m3/h that the first fan pushed into the fins. It would have more power (i.e., if the air were suddenly thicker and harder to move) but would not pull a greater amount of air or move it any faster. Am I right in my assumptions or totally off base here?

Last edited:

Related General Engineering News on Phys.org
QuantumPion
Gold Member
Your reasoning is incorrect. Fans work by creating a pressure differential. The stated flow rate for a particular fan is based on a single fan in isolation. Putting two fans together will increase the air pressure differential, causing a greater flow rate and thus greater cooling capacity.

If you want to do an actual calculation use actuator disk theory, which assumes that the pressure between the front and back of the fan is a discontinuous jump. Off the top of my head I believe using two fans will just make the flow "smoother" offering little cooling benefits. To increase the cooling rate you need to increase the velocity of the flow pass the fins.

Your reasoning is incorrect. Fans work by creating a pressure differential. The stated flow rate for a particular fan is based on a single fan in isolation. Putting two fans together will increase the air pressure differential, causing a greater flow rate and thus greater cooling capacity.
If each fan is the same, then they should have the same negative pressure on the back of each fan and positive pressure on the front. I know I'm simplifying, but shouldn't those be about the same value? If that's true, then the positive pressure in front of the push fan would be of equal magnitude to the negative pressure on the back of the pull fan, so it would still be the same speed (see the attached pic). I'm sure that it might increase slightly because the cancellations of pressure would decrease the amount of work each motor has to do, but I wouldn't think it that significant.

#### Attachments

• 3.8 KB Views: 1,189
QuantumPion
Gold Member
If each fan is the same, then they should have the same negative pressure on the back of each fan and positive pressure on the front.
This would only be true if the separation between the fans is so large that there is no net air flow from the first fan making it to the inlet of the 2nd fan, which is obviously not the case over the width of a CPU heatsink.

If that's true, then the positive pressure in front of the push fan would be of equal magnitude to the negative pressure on the back of the pull fan, so it would still be the same speed (see the attached pic).
No. The pressure difference would be +1 - (-1) = +2.

AlephZero
Homework Helper
Unless a lot of the area in the cooler is blocked off by the fins, the pressure increase across the fan will be small, and the airflow will depend mainly on the pitch of the fan blades and the RPM.

I don't think you will get twice as much flow from two fans unless they both run about twice as fast as one fan.

I suppose the company wants to sell you more fans to make more profit, but the bottom line is whether your CPU runs cool enough. If you don't have any problems with one fan, why get a second one?

AlephZero
Homework Helper
No. The pressure difference would be +1 - (-1) = +2.
The first fan is moving air that starts at zero velocity. The second fan is not. Therefore you will not get the same pressure difference across each unless they have different designs and/or run at different speeds.

If you could make an efficient multi-stage compressor just by stacking up a number of identical single stages, designing jet engines would be a lot easier than it actually is!

Instead of guessing why dont you look up actuator disk theory and take 5 minutes to get some results.

Instead of guessing why dont you look up actuator disk theory and take 5 minutes to get some results.
Not sure if this was directed at me or someone else, but I did try to look it up. I'm not a physicist, and I was unable to see how Wikipedia's explanation of it would relate to fluid momentum in a series of "disks" (their formula doesn't include anything for starting force or momentum). I would love to run an empirical test, but alas, I don't have two identical fans or the instrumental wherewithal to determine the actual speed/volume running across the fins.

If you could make an efficient multi-stage compressor just by stacking up a number of identical single stages, designing jet engines would be a lot easier than it actually is!
Good point! That lends at least a little credence to my feeling that just adding an identical fan wouldn't increase the flow by a significant amount.

russ_watters
Mentor
You don't need to become an aerospace engineer to figure this issue out.

Quantumpion is right that fans in series will increase the total flow by increasing the pressure output, but how much isn't a simple issue to figure out. Realistically, an axial fan is limited in its rpm and so stacking fans won't allow them to push much more air unless they really had a significant pressure restriction.

The general rule of thumb is that fans in series double the pressure at the same flow and fans in parallel double the flow at the same pressure. But this ignores the pressure requirements of the system they are blowing the air through. Applied to the same system, each will end up providing more flow and more pressure, but the amount depends on the performance curves of both the fans and the system.

I would look into a bigger or stronger fan before adding a second.

etudiant
Gold Member
There are a zillion computer enthousiast websites dedicated to examining exactly this issue, such as http://www.tomshardware.com/ or http://hardocp.com/ to name just two.
The impact of multiple fans has been extensively reviewed (some coolers have triple fans in series, with two radiators sandwiched between them) and always the benefit from the additional fan(s) is modest. There is no counter example afaik.

Mech_Engineer
Gold Member
It's been tested and is good for a couple of degrees (on overclocked processors) better cooling on computer hardware review websites.

I'm personally running a Cooler Master Hyper 212+ which comes with a 120mm fan on it already, and I added another fan which is run on a PWM circuit on the motherboard. I've found it to be a great setup, although I have to admit I never tested it without the second fan to see if I'm seeing a real performance increase.

Many times better intake/exhaust in your case will give you better overall performance, so take a look at that before adding a second fan to the CPU heatsink. Your case's interior needs to be as close to ambient as possible to maximize cooling capability of heatsinks, and reducing the temperature in there can help with all of your components' cooling systems.

Let me see if I can sum up the last few posts and maybe wrap the discussion up.

When taking physics, I remember many of the principles being discussed in “isolation” to focus on one part at a time (i.e., no gravity, no friction, etc.) If we apply that here and assume that the radiator is actually a straight, frictionless tube, then the rule of thumb that russ_watters mentioned holds true. Namely, the flow in would equal the flow out whether you had one or two fans; only the pressure available would differ. However, in the real world, the radiator is far from perfect. There are blockages on the edge of the fins, friction as it flows along the fins, and lots of places for the air to leak out along the way.

Enter the second fan.

What the second fan actually accomplishes is not increasing the overall airflow, but making sure that more of the air flows smoothly along the entire thickness of the radiator. The increased pressure would allow it to keep the speed up by overcoming some of the resistance of friction. That small volume of air that would have hit the first fin and blown out sideways will now be more likely to stay within the radiator, continuing to gather heat along the entire fin. This explains why hardware/overclocking sites (as mentioned by etudiant) and manufacturers report modest, but not phenomenal, gains by adding a second fan in push-pull configuration. Did that sum it up well?

Though not really part of the question, what Mech_Engineer says about case cooling is definitely important. Blowing hot air around inside a closed case won’t help no matter how many fans on the heatsink since there will not be a temperature differential to cool off the radiator. My case is pretty well ventilated, so that wasn’t an issue for me. The whole discussion actually got started with some friends because the fan on my Xigmatek Gaia was failing and I was debating about replacing it with two fans.

Two Fans

People overcomplicate the issue. A fan has two ratings: free air flow and maximum pressure with no flow (delta-P). Between those two extremes lies a curve of flow versus back pressure.

If you put two fans side by side (parallel), you effectively double the flow at all back pressure points but the maximum delta-P doesn't change.

Putting two fans in series doubles the available delta-P but keeps the maximum free air flow the same.

Plotting on an x/y graph with y=pressure and x=flow, parallel stretches the curve out on the x axis and in series stretches the curve up the y axis with the basic shape of the curve staying the same in each case.

The last thing to consider is that it requires 4 times the pressure to double the air flow through some restriction.

So - putting a fan on each side of a heat sink will give you about 40% more flow (square root of 2) because you've doubled the available pressure. Of course, other considerations such as heat flow off the processor into the heat sink and propagating through the heat sink also apply so you may not gain 40% more processor cooling.

You also gain redundancy.

See Chassis Plans White Paper - Cooling and Noise for a complete explanation.