There is another guy that needs an efficient quiet fan I encountered, on Twitter, Edderic Ugaddan.
We can use this thread to discuss things and if anyone else is interested join us and draw from what we uncover.
I have thus far found:
a) Low cogging motors are important, I found only one low cogging motor that can be reliably sourced and can have the bearing replaced with quiet bearings (the bearings are a significant source of noise), the JDPower 3505C. The low cogging motors all use a ceramic ring for some reason. There are others for cheaper through ali express but they are always different.
B)anti cogging built into the drive electronics would be nice and we are making some progress here but there is still a lot to do to get to a first stage system and it will take even longer to do even better, probbaly.
C) we could use sensorless drive for sure as it’s a bit silly to use a sensor on a fan motor, however open loop appears to be reasonable for a fan, if you have current sensing to detect stall.
D)I also posted about current sensing and current optimization as a wacky drive method, which can achieve significantly better energy efficiency, I forget how much but I think like 50% better, than standard common drive approaches, probably due to improvements in de facto motor timing. It may be workable when combined with a stall detection method for a fan. Fans are a special case of relatively easy drive as the load doesn’t change much or fast
About the fan blades and rest of the fan:
A) the fan blades on most fans are really crummy and the pressure to flow ratio can be improved by a factor of two or more. This also implies being able to greatly reduce noise because lower blade velocity equals lower noise
B) centrifugal fans are by far the best way to go as far as I can tell, for unidirectional fans. The blades all go the same speed through the air and the rotational motion helps you instead of hurting you. Much quieter and more efficient as there is less air whipping around and things can be optimized. With an axial fan the area of the blades towards the middle have a lower airspeed etc. so while you can twist the blades and curve them it’s not as good.
D) It is practical to print fan blades with common FDM printers and get reasonable vibration characteristics, but the system needs to be mounted on something with a little spring to it, this can all be 3d printed easily. However plastics all exhibit creep and this is a problem for springs so the mounting springs may need to be metal, it’s not clear yet, some of the springs I made collapsed a little bit too much over time while others are ok after several months.
E) many fans are limited in their design parameters and this leads to noise production, for instance the NF-12 IPCC 3000 or whatever fan I have tested, it advertises 7 mm h2o stall pressure and 74 cfm (no restriction or impediment to flow, given typical flow curves you an assume about 25 cfm at 3.5 mm h2o which isn’t very good). It makes a buzzing noise because the spokes that mount the motor are too close to the blades, but they are limited in how thick the assembly can be for standard fans. If you make your own you can do better and it’s not hard.
It does take a lot of testing and experimenting, printing many many blades and testing them out.
RPM is a reasonable proxy for noise however it is influenced by other factors and unfortunately the noise meters on the market are not that sensitive or repeatable so I’m not sure they are any better than just trying to minimize rpm.
Torque to speed ratio of the motors is not a big deal as long as torque is high enough, optimizing it would optimize energy efficiency a little better perhaps but that’s probably too much work.
I have been able to source very quiet bearings, Z4, through alibaba. It’s time consuming and the shipping costs a lot but it’s doable.
I recommended Edderic explore the Mosquito but the Lepton 3.0 is wanted.