Could some of you recommend some 92mm fans for CPU heatsink that are quiet
but move are anyways. Links would be helpful. Thanks a lot for anyone
with input. I don't want to just go by Db ratings because they can be
deceptive.
In terms of design concept, this fan has the potential to be
a quieter solution:
http://www.verax.de/verax.php?SID=2274042a9a0a02187a&template=cms_entry_technik&groupid=1&entryid=7
The fin shape provides smoother acceleration for the air. The fans
they make are super-expensive, and also are not available in all
popular sizes.
One contributors to fan noise is bearing type (sleeve is 3dB quieter
than ball bearing, but sleeve has potentially shorter life, unless
perhaps it is something like the Panaflo hydrowave design). The fan
body can make a difference, in terms of placement of struts and whether
the fan body is perforated or not.
The material the fan is constructed of, could also make a difference.
A fan with a metal body might offer better rigidity, and avoid
resonances better.
One brand of fans has a magnetic bearing in the hub, but I've never
read any user reports of how effective that is at reducing noise.
I don't know if fan motor design makes much of a difference.
The only innovation I know of, was the "tip magnetic" motor, which
puts the magnets on the periphery of the fan, rather than in the hub.
I think that ended up being more noisy than the traditional design.
I'm not aware of much in the way of attempts to smooth the pulsations
of the brushless DC motor, at least for use in cheap fans.
Another source of noise, is the nature of the power source. If
you use PWM (a.k.a Q-fan), the switching frequency of the PWM
power source can beat against the rotational frequency, and
"cogging noise" is generated by the fan motor. What that means is,
as the fan speed is reduced by Q-fan, the fan begins to produce
a whine due to cogging noise. A simple rheobus or a Zalman fanmate
is a cleaner source of power for the fan, and can avoid cogging noise.
Not all Q-fan implementations use PWM - some use linear control and
don't suffer from that problem. You can tell which type your CPU
header uses, by downloading the datasheet for the SuperI/O chip that
controls the fan.
In terms of noise, there are fans that have wideband noise (the
"rush of air"), versus the ones that tend to make a dominant tone.
When a fan with a dominant tone is measured by a weighted sound
measuring instrument, it tends to have good looking spec numbers.
But would be damn annoying to the user. I worked on a minicomputer
once, that had four fans on it with dominant tone type noise.
If I closed the door on the room I was working in, then walked
100 ft down the hallway, I could still clearly hear the fans. That
is the kind of fan you want to avoid.
Most of the "quiet fans" being offered for sale, simply trade off
reduced air flow for reduced noise. I picked up a fan like that
locally, and was unimpressed by how much air it moved. A simple
comparison of CFM for the various products, will show you which
fans are kinda worthless to own.
Some of the "quiet PC" sites have articles on fans. Here is a
sample:
http://www.silentpcreview.com/article25-page1.html
The data collected by the author of that article is here:
http://www.silentpcreview.com/files/fanspecs/fanspecs.pdf
http://www.silentpcreview.com/files/fanspecs/fanspecs.xls
If you scatterplot the data in the XLS file, and then
select a fan along the top diagonal edge of the plot,
those should be good fans. You may need to edit the file
a bit (for example, just put 92mm fans in the table etc).
Or, if you sort the data in the columns by dB instead of
by CFM, you may be able to find a winner that way (because
with the columns of data, you can see the fan size at the
same time as the other stats).
HTH,
Paul
