Hey guys, let me reiterate:
I will not try to make an encyclopedic list of fans and everything you can do with each fan, but to establish a list of useful fans at useful voltages for silencing & cooling.
The 6 tasks I identified probably represent a few days work for each fan. Multiply by number of fans -- you're looking many weeks of work here even for a modest number of fans. I will not willingly add more complexity to the project, I've already done these tasks before, and it is painstaking work. So while I will consider some suggestions, I will not commit to anything more at this point.
Here's are some technical issues about recording or measuring fans at 5V: The noise floor
and dynamic range
The quietest fans at 5V are likely to be as low as 10-12dBA/1m, maybe less. There is simply no way to capture this sound or to measure it except maybe
by putting the mic closer. A lot closer. That immediately means it can't be compared directly against the other fans, which I hope to measure at 1m or at least 1/2 meter, no less. On top of that, unless the ambient noise is significantly lower than what's being measured or recorded, it will not be apparent. I have never seen lower than 13-14 dBA in my test lab, in my home or anywhere else except the anechoic chamber. (There is no way I can simply hang out at the anechoic chamber the way I would have to in order to get this testing done. It is far too busy a place.)
Also, even if I get the mic in close enough or get enough gain in the preamp to record the quietest fan above the noise floor of the room & recording system, there will be the question whether the loudest fan (say a 120mm) at 12V might overload the mic/recording system -- if I try to keep the recording level constant, which is one of my goals. I don't know yet, but I suspect this could easily happen.
So routine 5V testing is out.
This does not mean I will not comment on fans at 5V or less, just that I will not try to collect this data on a routine basis.
Now if a fan works well at 5V and can be measured / recorded, sure.
-- Again, this would basically double the testing work & I just don't think it is significant enough an issue to bother. PWM implementation seems to differ, and some PWM controllers seem to do ok with some fans while others don't do ok with the same fans. And vice versa. So the testing methodology here becomes ridicuously complex if you try to be fair to both the controllers and the fans. No, I am not touching this right now.
Another thing is that I am not interested in testing fans that will not be recommended. Maybe one or two common ones just for reference sake, but that just is not useful enough. Things like Vantec fans which most of us will use in a pinch if necessary, yeah sure, but not 46 dBA Deltas. What's the point? You can see from the spec alone that it would be useless to any silent PC builder.
The point is to make a short list of the quiet & useful fans & weed out any that clearly don't make the mark (and perhaops demonstrate why with a couple of examples) but to focus on the best. My assumptions are simple: Most fans I've heard are not quiet at 12V, and most fans do not ramp down graciously as you drop voltage. So in other words, most fans are bad, and we could spend till doomsday testing & measuring them. So let's not.
So recommendations for testing should include only fans that you KNOW or strongly suspect are pretty good.
And I'd like to reiterate one more time: Fans tend to vary a lot in quality, and a little difference to one in one cercimstance can be big to another in another circumstance when it comes to noise. So this recommended list will be very cautious.
For example, I've said before that the Panaflo 80L is my reference because it semed consistent and ramped down in speed really nicely. That was on the basis of several dozen samples. Maybe even a hundred. Well in the past 6 months, I've come across 3 more batches which were not nearly as consistent as those first; I've found some that were downright terrible. In most of these batches, there were enogh good ones that my basic opin of the Panaflo has not changed, but I am certainly much quicker to point out potential inconsistencies.