I am totally surprised, that even in 2013, one MAJOR source of noise in computers has not been properly addressed, yet is so elementary

FAN NOISE!

EVERY fan used in computers, regardless of blade shap, or rotation speed, size, bearings have one thing in common:

The blades are equally spaced around the circle! Choose any fan in your system, or on your bench, count the number of blades and divide 360 by the blade count. You will have the number of degrees between the same spot on every blade

Why is this important?

Air flowing thru a running fan has a little pressure wave in the air exiting the fan housing, little, tiny, bareley audible, whu, whu, whu, whu, in time with a blade passing any given spot in space. Like helicopter blades going Whup, Whup, Whup, you know the sound

now, fan blades are a tad smaller than choppa blades, but the pulses are of a higher frequency, and that frequency gets up into (human) audible range...

So, hold that thought...

Now theres another noise maker action at work, and that is where the arflow goes

Back in the 70's, EMD-GM locomotive company did some testing and research on locomotive radiator cooling fans, and they found that size, speed and HUB MOTOR MOUNTING were all crucial factors, so they came up with Q-Fans.



This photo is an 8-blade fan with the blades unevenly spaced.

A 9-blade fan has 3 sets of 3 blades, and in each set of blades, the middle blade is closer to the one (in the photo below) to the left.



The 48" fans generally run slower where possible.

This means that the pressure pulses are now no longer regular, and if you didnt already realise it, sound is a steady stream of pressure pulses

EMD also worked out that the bars that the motor is mounted on should be on the intake side of the fan, having the mounting bars on the exhaust side creates a siren effect. Your fan grilles can also contribute to the same kind of siren effect

This is all based on MUCH higher output fans in terms of speed of rotation, volume of air moved and static pressure. The noise created by slow turning fans is very different and therefore this research would not have any impact on quiet DC axial fans.

Fan grills are an accepted noise issue even in silent computing but their effect is far less given the lower speeds. Wire grills like this are used in preference to stamped or drilled grills in cases as they impeded the air less and in many circumstances grills can be removed entirely but in cases where grills are required, they have to go on one side of the fan in particular. There are also some grill designs like those from Silverstone which have a noticeable sound reducing effect and also act to focus air flow.

Zipang, do you know the comparative RPM figures?

The Loco fans operate at a maximum of 1500 to 1800 RPM
Faster fans will make higher-pitched sounds because the pressure-pulse frequency will be higher. Ditto for the siren effect

Yes, and this research is based on 1.2m diameter fans hence the speed of the fan tips would be 10 times higher for the same RPM as for a 120mm fan in a computer. 1500-1800rpm is also quite high for a 120mm fan which most people here would soncider far from silent. When you get down to 500-800rpm at 120mm, then this research becomes irrelevant given the far slower linear speed.

It's still worth considering rather than dismissing the concept outright...

Believe me I have considered it. I have a few friends in aeronautical engineering (we sit watching F1 debating wing angles and CFD) and one of them disagreed with fan designs in computers as being inefficient. This man has a PhD in aeronautical engineering so you would have thought he would have known! When I explained to him certain other influences such as noise, power draw, vibration, service life, positive/negative pressure impedance and a need for moving air to cool something rather than specifically to generate thrust he agreed that perhaps there was some merit in the designs they use!

I think I should point out that I have not dismissed it outright. If I had dismissed it outright I would not have given a thorough justifcation for why this might be appropriate for industrial cooling and not relevant for much lower airflow requirements. Yes, GE might have been able to reduce their noise but from what? 90dB to 85dB? It would hardly matter when you've got the noise of the rest of the train going on. Implement these changes on a low speed 120mm fan in a computer and what will you do? Make it worse in terms of some combination of noise, airflow or pressure as it is a totally different scale of size and speed. Aerodynamics do not scale well.

You also have to bear in mind that DC axial fans are not some amateurish effort purely made for computers. They have been in use throughout many decades cooling electrical equipment also. If this design really did offer such benefits for these applications, elements of it would have been implemented. Patents would have been taken and someone would have become very rich by making server and machine rooms slightly more tolerable places to be.

Another element of it that would not work on our scale is the number of blades. An even number of blades (8 they have chosen) causes resonance, this is bad for noise but may on their scale have counteracted other issues. All computer fans have odd numbers of blades - go on check! This is to reduce resonance, making them quieter for our application, even though this isn't a good thing when you're trying to cool a train.

So there is some more expansion of technical reasons why this is an inappropriate design for quiet computer fans, regardless of it perhaps being good at cooling massive train engines. If however you'd like to continue to yourself dismiss outright my arguments as dismissing your arguments, then by all means provide technical justifcation. Better still: patent it, bring it to market and prove me wrong.

8 blades in the upper photo was used for balancing, the 9 blade fan in the lower photo was the original concept, and yes i counted.

I have a slow-running 9-blader here on my system casing on the intake side, and a higher revving 7-blade at the exhaust end.

both run fairly quietly, altho i cant help wonder about air-movement effectiveness at higher speeds, and then what if we DID investigate the uneven spacing concept

computer fans have always been the same design tho since year dot, and i notice that the only real change has been in fan speeds, altho bearings and motors may have been revised, mainly the better stuff in top $ units

Dust admission into the motor/bearing area will contribute noise, as well as reducing efficiency, while dirt build-up on blades will also affect efficiency AND noise levels (I live in what can be a dusty area) Dust caking cannot be relied on to be even, and an out-of-balance situation can also arise

Maybe somebody could get busy with a 3D printer and do a PC fan or 2 that embodies the ideas and just try them out...

So I guess thres nowt more to consider on this line of thought, the ultimate expert has spoken and all is known that can be known about fan noise

That means I should also ignore the howl made by the electric radiator fans in my car when they kick in, it must be noise made by the front wheel bearings, or negative thermal expansion of the passenger cabin as the aircon cools it all down, although the 17°C inside feels (subjectively) a lot nicer than the 42°C outside. It might even be exhaust resonance...

Edh, just think of this: true scientific endeavour understands that there is no such thing as a totally definitive outcome on studies, and maybe you and friends need to get your pipes and bowls out, chew the fat a little more, do some more discussing, figure why the 12" radiator fans are noisy, noisier than 48" locomotive fans (and you might be surprised at just how quiet a modern 4,000hp EMD locomotive can be) or why they use a different design for the dynamic brake fans, you may just find some concepts that are worth looking into for computer cooling purposes. Some concepts are scaleable, some are not. Why do jet engines make so much noise, they are little more than an assemblage of fans. What if propeller engined planes had a ring around each propeller, making them ducted fans? What about the noise characteristics of a box fan compared to an open pedestal fan that you might use in your house on a hot day? Could we use drum/barrel fans, with scrolls, instead of noisy propeller blades?

You skimmed over my mention of regular pressure pulses being sound, and of the exhaust-side frame causing a siren effect

I realise that you are an absolute authority on the matter, but only religion presumes to be so authoratative that they close the book to further discussion and research.

Science continues to discuss...

Has anyone looked into the Dyson Air Multiplier concept?

Moving air is, itself, silent.

What we use to artificially move it, and what we move it through, is what makes the noise.

Sit outside enjoying the breeze on a warm day, it's not the breeze you hear, but the rustling of leaves as it blows through, moving them around...

I just watched "Richard Hammonds Miracles Of Nature" 2nd episode (around 41min) where they show a concept of a really quiet fan - made on a principle of an owl wing. The prototype has jagged back edge of a fan.

Here is one picture:




Well, I hope this really works and soon see some good products (in the show they mention PCs, aircons, airblowers, hairdryers... all being quiet ).

PS:
I wish I could replace the fan in my X61s with this one

bye

Cut off 3 of the blades of your 9 blade fan so that you have a 6 bladed fan with 3 pairs of 2 blades in an uneven pattern as you suggest. Then measure the noise and cooling performance at different RPM settings. To achieve the same airflow you will need to run a higher RPM and you can measure the SPL at both speeds and show that it is or is not better. If you do not have the equipment to do this measurement, perhaps donate a 9 bladed fan to MikeC at SPCR and suggest that he tries this experiment in his anechoic chamber. I would value Mike's input on whether this is viable or not.


There are many different designs of fan blade. Compare different units and you see that some are more sculpted than others, the materials differ, different blade counts etc. Look at Noctua to see different blade designs and there continue to be innovations in fan mounting from companies like Noiseblocker.


Quiet enough to be below the sound floor (~11dB) of an anechoic chamber? No, you're talking things that are an order of magnitude louder as a minimum. It's on a completely different scale of noisiness. Look at the SPCR lab reviews and you'll see many products that are so quiet they simply can't be measured with the equipment available. Train cooling just doesn't have relevance for silent computing. You're simply trolling by persisting with such an argument.


This has been tried. They tend to be less efficient at low speed and more efficient at high speed compared to a traditional propeller driven aircraft. However, they can not match a jet for operating altitude so can not replace turbofan aircraft and seeing how the advantage for small propeller driven aircraft over jet aircraft is doing short, slow flights they would lose this niche market if they were ducted.


Already in use and noisy for the same cooling effect. Take a look at things like the Cooler Master Jet series of CPU coolers. Silence has become of greater importance since that time and these designs have fallen by the wayside.


Do not make such accusations. I read it, it is not relevant when dealing with slow speed 120mm fans as I already stated, regardless of how important it might be with 1.2m diameter train cooling fans.


Strange you mention this, I have a friend who is a mechanical engineer and works with James Dyson! Interesting perhaps for a low static pressue application like a desk fan but consider this: if it was good for a higher static pressure (or vacuum) application then wouldn't Dyson have already fitted it to his vacuum cleaners? By your logic we would then have silent vacuum cleaners and that would sell like crazy! The trouble is it can't manage to generate much static pressure so it might be a little effective on an open test bench system (as would a desk fan) but in a case it would be pretty hopeless.


Yes, in a vacuum perhaps. However the turbelent airflow caused as that air interfaces with the surrounding air means that you simply can't blow air without making sound itself. By your logic, a nuclear bomb would also be silent! It is only the moving air caused by the nuclear explosion that makes the noise. Please don't start arguing with me about this as you have with everything else, I have worked as a nuclear physicist so I might just know...


If you ride a bike, take it up to 80km/h and you hear a lot of wind noise. Fall through the air in a pachachute jump and you hear a lot of wind noise. You simply can't move air past something without there being noise, even if you exclude the noise made by the air moving equipment. If you read any of the SPCR fan reviews you will see this as a common theme: the high quality fans only make the sound of air moving with no mechanical noises. If however you were to speed those same fan blade designs up 10 times faster on faster motors then yes, things like the siren effect would matter but it simply doesn't on silent computing fans.

I'm somewhat concerned with the direction your argument has taken. It verges on trolling that you consider a knowledgeable argument against your idea as being dismissive. The fact you also continued to post and make it personal when you do just because the target of your trolling did not respond is not helpful. Please do understand that I am giving you reasons why this wouldn't work not trying to shoot it down. I am not the enemy of quieter fans that you make out that I might be as if I could make a computer quieter I would do, hence why I'm here.

I would suggest you read and take on board some of the SPCR fan reviews to see what is relevant in a quiet fan for computing before you respond.

This was mentioned above, but bears repeating.

The Noctua fans (among others) at low speeds are incredibly quiet.

This is a similar concept to the Noctua 'vortex control notches':
http://noctua.at/main.php?show=nine_blade_design&lng=en