Where, exactly, does the sound come from?

[scalar]

When talking about fan noise, the interesting question is exactly where the noise is coming from. Let's assume that you have very quiet bearings and the motor drive is silent, and the rotor is properly balanced.

Where does the noise come from, if all these factors are eliminated? Are the blades vibrating and producing the sound? Or is it something else?

I have a theory that the real source of fan noise is from small spinning tornadoes of air, known in scientific terms as a "vortex", that is like slicing a peeling off the surface of cheese, and the peeling curls up into a tight roll.

The rotational forces inside this tightly spinning curl of air keep it coherent and in motion as the fan blade slices through the air, like a tiny smoke ring that spirals through air for an inch or so before dissipating.

I am of the opinion that the noise we hear is a sort of miniature shockwave created as these tightly compressed rolls of air collide with objects in their path, such as the support beams thar support the hub of the fan rotor, and the air grille over the back of the fan.


Now, this does not account for air entering the fan through a grille, but I am of the opinion that something similar is occurring when a large amount of air is trying to rapidly sqeeze through a limited space, such as a sheet of metal with perforations in it.

As the air tries to enter these holes, it is in effect creating a small vortex of air that spins through the holes at a high rate of speed, creating a large amount of contact friction with the surface area of the grille holes. A coarsely-punched set of holes may further contribute to this, since the knife-edge of the punching may help in the formation of these vortexes.


I do believe there is a way to test these theories, and to find ways to minimize the formation of the vortexes. There are scientists researching how insect wings function, who are doing so by using large-scale models of insect wings suspended in a tank of oil, which simulates the viscosity of air at such a small scale. The slow moment of the wing through the oil can be tracked with air bubbles in the oil that are mapped with a scanning laser assembly:

Little wings, big flap
http://www.nature.com/nsu/010823/010823-10.html

Such a system could be similarly applied to the design and manufacture of air grilles and fan blade design, to understand exactly where the noise sources are, and how they could be reduced or eliminated.

(I'm considering contacting these scientists and finding out what would be the cost of duplicating their efforts, in an attempt to better understand fan/grille airflows..)

-Scalar

[wgragg]

I've often wondered (if this is even possible) if constructing a case where almost all avenues for air entrance (besides the intended air intake at the front) were somehow closed would help or hurt either cooling or noise. On the one hand I can see a more true wind tunnel type of effect and possibly less turbulence, but on the other, I could see it possibly having a detrimental impact on say cooling the power supply or the drives in the upper bays.

You pose an interesting theory there.

Wendell

[scalar]

Ah, here we go. The flow mapping system is known as:
Digital Particle Image Velocimetry (DPIV)


Flows of water around swimming fish (method and implementation)
http://darwin.bio.uci.edu/~edrucker/home/dpiv.htm

How DPIV works:
http://www.holomap.com/dpiv.htm

Particle Image Velocimetry at NASA
http://www.grc.nasa.gov/WWW/OptInstr/piv/index.htm

STEREOSCOPIC DIGITAL PARTICLE IMAGE VELOCIMETRY
http://www.vki.ac.be/research/themes/mt/stereo.html


A direct application for what I'm interested in doing: observing airflows over fan/compressor blades

Tip-Clearance Vortex Characterized With Three-Dimensional Digital Particle Image Velocimetry
http://www.grc.nasa.gov/WWW/RT2001/5000/5520wernet.html

Surge Flow in a Centrifugal Compressor Measured by Digital Particle Image Velocimetry
http://www.grc.nasa.gov/WWW/RT1999/5000/5520wernet.html

[scalar]

Whoah, the fish locomotion researcher has some awesome articles on DPIV. Take a look at page 14 of this PDF, with the frames animation of the vortex moving over the fish's body:


Drucker, E. G. and G. V. Lauder. 1999. Locomotor forces on a swimming fish: three-dimensional vortex wake dynamics quantified using digital particle image velocimetry. Journal of Experimental Biology 202: 2393-2412. [Cover illustration / (PDF)]
http://darwin.bio.uci.edu/~edrucker/hom ... eb2130.pdf


Someone else I should write to about a hobbyist DPIV test bench.

-Scalar

[pdf27]

Hi,
One of my fourth year modules this year is on flow induced sound and vibration, so I've picked up a few things that might be relevant.

Firstly, any vibrations at all will make noise. Either in the form of a pressure wave passing over a surface at less than the speed of sound, or a vibrating surface, they make noise. IIRC, where the movement is subsonic the decay is exponential, while it only declines as X^3 for supersonic motion.

Secondly, the vortices you are talking about are themselves noise. What you hear is after all only pressure variations on your eardrum, and any turbulent flow (i.e. anything that has been through a fan, grill or similar) will have a large number of tiny vortices in it. These have a low pressure at the centre because they are spinning very fast, meaning there is a pressure differential across the core of each vortex.
This means that the disturbed air will sound louder than the surroundings, and that it will cause anything it flows over to vibrate (possibly resonate). This means (as you will probably have already spotted) that you want to cut down on turbulence (e.g. free flow fan grills) and use low speed fans wherever possible (off the top of my head, turbulence intensity varies with something like air speed cubed).

I've got a friend who is doing a similar sort of thing for his fourth year project (zapping water flows over a delta wing with lasers to see what the air particles are doing), and the kit involved isn't cheap.

[Elliot]

I have a theory that the real source of fan noise is from small spinning tornadoes of air, known in scientific terms as a "vortex", that is like slicing a peeling off the surface of cheese, and the peeling curls up into a tight roll.

From ARX FAQ:

Airflow noise is mostly associated with wide band noise. There are two major contributing factors. First source is vortex shedding noise, which is generated by fan blades and trailing edges cutting through air, and therefore can be better managed through improved fan blade profile, pitch angle and trailing edge design . The second source is turbulence noise, which is created by airflow streams. Inlet and outlet disturbance, sharp edges and angles tend to intensify air turbulence and result in increased noise.

[scalar]

Yes, okay, so the general concept is covered in the FAQ. But how many fan and/or case manufacturers are actually applying these principles?

When designing for silence, do they consider the implications of airflow interactions with other components, or are they just looking at individual components, not interacting with anything else? I note that many fan noise ratings are from highly unrealistic tests, like suspended midair in free space with no nearby obstructions or finger guards, etc.

I am of the opinion that the testing and designing for silence can and should be more sophisticated. The main question currently is the cost of such testing, which I hope to eventually determine here..

-Scalar