A summary, a starting point, and directions to other info sources on the many aspects of fans. This is NOT comprehensive, but may help answer common questions to avoid the repetition common to forums like this one. You're welcome to suggest additions to this list.
Read this stopgap article
as well as this sticky.
= cubic feet / minute. Airflow measurement.
= Decibels, A-weighted (to approximate human hearing). Usually measured at 1 meter in an anechoic chamber. Often unreliable due to improper technique by manufacturer and lack of info about the quality of the noise: a high pitched whine is much more annoying than a smoother broadband noise of the same dBA rating. Check our web links
for sources of info on noise and fan noise. (Comair-Rotron is the best.)
Both of the above are usually measured with the fan suspended and without impedance to airflow. They represent a no-load condition. Fans on HS and in PC cases will always be louder and produce less airflow than specified. Aside: The highest CFM/Noise fans seem to be the high speed, fairly noisy ones. Slow them down and they're no different than the ones that are already slow.
C. Common code in fan part numbers:
L, M, and H usually refer to low, medium and high speed (rpm). B=ball bearing, S=sleeve bearing. The diameter of the fan in mm is usually in the part #: 40, 60, 70, 80, 92, 120...
D. A = amps
or current. Ranges from under 0.1A to over 0.5A for most 12VDC fans used in PCs. In the absence of other data, most fans under 0.1A are quiet; anything over 0.15A will likely be too loud at 12VDC.
Rules of Thumb:
1. Most quiet fans are low airflow.
Air turbulence noise is the #1 noise source in fans, followed by vibrations mechanically transmitted to the case, and then bearings. Fan blade design and dymanic balancing, therefore, may be more critical than the type of bearing used. Blade design is not easy to assess. Easy enough to check on balance -- if you already have the fan on hand: just feel how much it vibrates in your hand, compare to another. There will be variations within the same model due to manufacturing toleranes. (Avoid fans made on Mondays & Fridays.
2. Any impedance
within 1/2 inch of either side of the fan results in increased turbulence noise - the amount varies with airflow. An impedance on the intake side causes more noise than the same impedance on the exhaust side.
2.A. Constriction of airflow
through any kind of channel is a kind of impedance and will restrict airflow and increase noise -- from very little to much more, depending on particulars. For such channels, aerodynamics do matter; internal surfaces should be smooth and in/out vents rounded. Fan adaptors usually impede airflow somewhat.
3. Mechanical decoupling
works best with very soft grommets that eliminate direct contact between fan and mounting plate. Ordinary rubber grommets don't do much. At low speeds, decoupling has little effect.
4. Small fans (60mm and smaller) are too noisy
if they are going to be effective at all because they have to spin that much faster to move any air. The high speed combined with the small blades causes high pitched noise, which is more annoying than any other, for most of us.
5. Good 80~92mm fans represent the best compromise
, with good 120mm fans being useful if you have the space to make them fit. Note that there is a dead zone directly beneath the central hub -- the air flows out from the fan in a spiral or cylindrical pattern. As you get farther away, the effect dissipates, but airflwo is also more diffused. Bigger fans usually have a bigger dead zone and less concentration of airflow (which explains why larger higher CFM fans sometimes don't work as well as smaller ones for HS cooling).
Check out the YSTech TMD fan, which is driven from the rim and thus has no central dead zone. Too buzzy and whiny for me even at reduced voltage, but quieter and more effective for HS than conventional fans with similar CFM and noise rating.
6. Our reference fan is the Panaflo FBA08A12L1A
80mm "Hydrowave" bearing fan. Why? Cheap (in the US, anyway) consistent, virtually inaudible at 7V or less (in the typical PC case and room acoustic). More reasons in our HS Test Methodology
article. I often call it Panaflo 80L - lazy shorthand. In fact, I may start using P80L.
There are some as good or better, but it's a good value even at US$10-12 when you listen to all the others, including many sold as quiet. (I have not heard a Papst that does not click -- at least not yet. Send me some that don't click!)
with noisy fans can be replaced with various fanless HS from Zalman. But fanless doesn't always work, and the HS replacement is not always necessary. If there's room above the PCI/AGP slots, a low noise 80-92mm fan on a Zalman 123
fan bracket (or facsimile) will bathe both sides of the VGA card with enough airflow to replace tiny fans, usually w/o any mod of the HS. Often good with a 5-7V Panaflo (80mm, L). This is the simplest, cheapest effective quiet VGA cooling - if you have room. The 123 is an excellent solution for CPU fan mounting, and allows a second fan to be mounted over the VGA card, as John Coyle showed us here
Screaming little fans on motherboard chips
can usually be removed if you have some airflow from the CPU HS fan and a VGA overhead fan (mentioned above).
that use <80mm fans are to be avoided.
Handle fans with care
. 2 reasons -
1) blade edges are sharp and fingers are softer; even at slow RPM, blades win.
2) Dropping a fan can cause bearing wear & imbalance. Even a small amount of impbalance will cause more vibration.