I do like a quiet PC. I'm always replacing fans in my retro PSU's and video cards, and looking for the cases and CPU heatsinks with the largest possible fan diameters.

Manufacturers' dB ratings don't really mean much. Their "measurement methods" are wildly incompatible with each other and far from what you'll actually get. And yes as you say, some noise signatures are less irritating than others at the same level. This is pretty tough with rare sizes that you can't find reviews for. You can always undervolt a noisy fan a bit with a resistor, though.

I nearly always end up doing the resistor trick, but was really hoping to avoid that one day. On occasion, I have ordered fans that were just how I liked them in terms of noise. As often happens, when I later go to order more, they are all out and have a replacement part, which is never the same.

firage wrote:

I do like a quiet PC. I'm always replacing fans in my retro PSU's and video cards, and looking for the cases and CPU heatsinks with the largest possible fan diameters.

Manufacturers' dB ratings don't really mean much ...

^ This.

It's all about subjective psychoacoustics.

Couple of illustrative trade tricks:

http://noctua.at/en/vortex-control-notches

http://noctua.at/en/stepped-inlet-design

My main rig is around 15-25 Db depending on fan speed (watercooling)
All my retro rigs are under 25 Db, the older the system the less noise it generates. I try to replace fans as much as possible with good ones as big as possible to get less Db out of it.
P1: only the speaker, can't hear the fans.
P2: 21 Db i's the loudest fan (all enermax)
Ss7: 18 Db from graphics card.
Athlon: same as P2 just more fan's, I replaced the loud 40Db cpu cooler with a 22 Db high cfm 120mm fan through a adapter.
I set my max at 30Db before I start modding and replacing fans with quiet ones.

Thanks for the responses. I am now stuck staring at these numbers on digikey. For example,

For a Cyrix MII at 300 MHz, do I go with 13 CFM of cooling at 29 dB(A) of noise, or 7.7 CFM at 17 dB(A)? Is 7.7 CFM enough? Room temperature never reaches above 24 deg. C, even in summer.

Another example, for a VIA C3 1.4 GHz, K6-III-500, and Tualatin 1.4 GHz, do I go with 13.8 CFM at 10.7 dB(A), 19.3 CFM at 22 dB(A), or 23.5 CFM at 27 dB(A). All fans are of the same brand/line of Sunon's DR MagLev w/Vapo-bearing.

Last example is for a standard 80mm x 80mm case fan. 33.0 CFM at 28 dB(A) or 23.7 CFM at 22 dB(A)?

And is 23 dB(A) for a tiny 30mm x 30mm fan at 9500 rpm going to have a high pitched buzz?

firage wrote:

I do like a quiet PC. I'm always replacing fans in my retro PSU's and video cards, and looking for the cases and CPU heatsinks with the largest possible fan diameters.

Manufacturers' dB ratings don't really mean much. Their "measurement methods" are wildly incompatible with each other and far from what you'll actually get. And yes as you say, some noise signatures are less irritating than others at the same level. This is pretty tough with rare sizes that you can't find reviews for. You can always undervolt a noisy fan a bit with a resistor, though.

^^This.

Also it's a matter of what the noises sound like. Some may be annoying to one person but not to another (some stock CPU HSFs are like this to me, eardrum bleeding to me but another may not even notice).

And there's loads of other noises, like resonance (which to me is usually actually a fun thing to take care of ^^) or read/write sounds which I don't typically mind at all but some other do.

Coil whine and the continuous rotation sound of some hard drives will make me go nuts however 😵

Noiseblocker has 80mm fans 18Db and 33 cfm. I prefer enermax for 80mm fans but both are good. 30mm is rare, take what you can get.

I like to use fan controllers, like the ones that came with the old zalman cpu coolers.

I did buy a few fan controllers, but they add quite a bit of cost and have extra messy and long cables hanging in the case, which is not so neat. You can mount them, but the cables are still there. I have 12 machines, case fans, CPU fans, graphic card fans - that's a lot of controllers. I'm hoping to not bother with the added complexity.

Seems like I will just have to try my luck. I think I'll skip on the 80mm x 80mm fans as these are so common for CPU case fans, I can get specialised quiet ones anywhere.

It would be neat if there was a switch on each fan which could set the RPM between 3 different speeds. I've noticed that for the same exact fan size, same manufacturer, same series, there will be about 3 different RPM versions with 3 different CFM's and noise levels. I assume they are trying to satisfy everyone's noise requirement. Why not make one fan with a 3-position switch or jumper?

I don't have any idea about how many dB, but fan noise is pretty ignorable for me as long as the computer is on the floor.
My main desktop is a fairly loud machine but I hardly notice because it's down by my feet, plus the desk is a bit of a barrier that I think further dampens or redirects the sound.
A horizontal desktop sitting underneath a monitor isn't bad, but I think the worst case is a tower sitting next to me on top of the desk with it's noise sources at ear level. That's when I might get picky about how loud things are.

shamino wrote:

... fan noise is pretty ignorable for me as long as the computer is on the floor.
My main desktop is a fairly loud machine but I hardly notice because it's down by my feet, plus the desk is a bit of a barrier that I think further dampens or redirects the sound ...

^ Yep, not having line of sight (direct noise path) from your working position to comp case greatly reduces subjective perception of noise.

Also, in a room with natural convection of air, by placing case on floor you automatically enjoy cooler temps (lower fan revs = less noise):

TELVM wrote:

shamino wrote:

... fan noise is pretty ignorable for me as long as the computer is on the floor.
My main desktop is a fairly loud machine but I hardly notice because it's down by my feet, plus the desk is a bit of a barrier that I think further dampens or redirects the sound ...

^ Yep, not having line of sight (direct noise path) from your working position to comp case greatly reduces subjective perception of noise.

Also, in a room with natural convection of air, by placing case on floor you automatically enjoy cooler temps (lower fan revs = less noise):

just remember to keep atleast 6 cm higher then the floor itself, this saves so much on the dust problems.

As close to 0dB as possible. ;] When I was building my Pentium 1 rig I pretty much eliminated all sources of noise by replacing the following:
1. Tiny CPU cooler with a 30mm fan --> replaced with oversized socket A cooler and a 60mm fan running at 5V.
2. Old PSU with 80mm fan --> replaced with a modern PSU with 120mm fan.
3. 3GB mechanical HDD --> replaced with a couple of CF cards.
4. CD-ROM drive --> slowed down to 2x by using a tiny dos utility.

Thanks to these few simple steps this system is now almost completely silent.

I remember how noisy my Pentium 1 and Pentium 2 rigs were back in the day and I just did not want to put up with this nonsense anymore.
Some more conservative vogoners won't aprove of that kind of heresy, but I do like mixing old and more modern parts. ;]

The only noise I like to hear is the sound of my floppy drive in action.

Not more than necessary, which means next to none when it comes to PIII systems and older. Those can be made virtually noiseless.

To update the original post, I ordered several fan sizes, RPM's, and dB(A) ratings and determined that around 21 dB(A) is my limit for what I consider quiet. More dB(A) and I feel like adding a series resistor to drop the voltage (and thus speed & noise) of the fan. I tested the CPUs on a heatsink to get an idea for their in-use noise, which is generally louder than a fan running not attached to the heatsink.