Sound Level :

The last time I took a hearing test was two years ago give or take a few months. At my age, the result was promising. I still could hear pretty well, although not as good as a 19-year old. I was in the hearing aid industry for 10 years and have some knowledge about sound, where people pay a lot of money to hear "sound". In recent years, I kind of involve with some PC stuffs, and notice an interesting trend about noise emitting from PC, where people are eager to pay money to get rid of sound.

When I say trend, I mean for sometimes PC enthusiasts did not seem to care about the loud noise generated by the "fame" delta and tornado fans; for awhile "the hall of fame" entitlement required ownership of such fans. It lasted for about a year or two before larger fan 90 and 120 mm began taking roots in favor of slower turning speed in order to reduce noise level. Users are attracted to water-cooling for the same reason - lower noise.
Fan controllers become hot seller in taming the delta and tornado allowing users the ability to slow down the screamers.

Since my hearing is considered normal (two years ago), I hope my hearing ability has not decelerated. I like to share with you what I have collected and gathered some information about sound level from FANS before I get to the age when a 6,000-RPM Delta fan is the only sound I can hear. A word from OSHA provides detail information on sound level requirements. An old Chinese saying "The tree sounds when the wind blows". It means more than the obvious, however taken the face value; it may just mean what it says. Now, I like to know "when the wind blows the Tree remains silent" is possible.

Objectives:

Does larger fan size reduce noise level?

Do dBA specifications provided by manufacturers correspond to real world situation?

Is water cooling the answer to quiet computing?

The Referee :
Not the best, with 94 dBA calibration built-in, enough to do what I intented to do here.




The Setup :



To build a certifiable sound room for this exercise would be out of my reach. Instead, I cleaned up my work bench and have a "clean" tablecloth over it to have a setup that is close to what we are likely to experience in our everyday environment. After all, we don't live in a sound test chamber where sound waves don't get bounced around. Here is the condition/environment of the test area.

Sound saturation ~37.4 dBA

Sound meter CEM DT-8850 35-100 dBA Low, 100-130 dBA High

Power Supply without fan connected (running fan less)

Air movement – calm with no wind speed

Fan tested is suspended by 2.5” foam

Microphone is ~ 4” from the test fan

All windows in the room are closed during the test


There is an air-vent located about 10 feet from the test bench; the air-vent is shut closed during the test. The sound meter fluctuates +/- 0.5 dBA; reading is taken after 30 seconds of continuous measurement; the lowest number is recorded.

Of course, A Sound Test Chamber will be something that meets all acoustic standards in the industry. A similar sound room and a bit smaller I used back in the old days, you could hear your own heartbeats after a few minutes sitting down.


Source: NASA


The Players :



A bunch of fans after my Pre-Spring clean up - all are used for a few hours at best for testing, some are relatively new and none are more than 1 year old.

Basis :

Most PC fans we purchase list the sound level of sort; label stating 18 dBA, 23 dBA etc. I have no idea how the ratings are obtained because I could never find out where those ratings came from. What I could find is something like this.



Sound properties and their perception to human ears is not linear; 40 dB is not twice as loud as 20 dB.

I consider the sound saturation in the test area at 37.4 dBA to be within the normal range as listed here about 40 dBA. This also establishes the accuracy level of the sound meter used in this review to be within an acceptable range for a non-clinical environment. The weighing factor of dBA and dBC is explained here.




Well, we now have some basis- (using dBA and not DB)

1) A very quiet room in an ordinary household is about 30 dBA at night. Normally, a quiet room has a range of 30- 40 dBA.
2) A small office has a sound saturation of 50-55 dBA.
3) 3 dBA increase is just perceptible
4) 5 dBA increase is twice as loud.
5) 10 dBA increase is four times as loud

Simply put, a PC that generates 50 dBA will be 10 times louder (in our perception) than one generates 40 dBA. Assuming you are in a quiet room of 35 dBA, you would barely hear the PC if it generates 38 dBA.

If the basis has any truth, I like to know the reason for marketing a fan that generates less than 30 dBA. Hence, I intend to find out. Please bear in mind, most of the fans you find claim less than 40 dBA. Please note that I am only trying to find out what is the real sound level generated by these commonly used fans in an ordinary household and small office other than in a sound chamber and not their air-movement capability.

Let's find out ->

Factor findings :

I am trying to establish the thought that "the tree sounds when the wind blows" - "cause & effect". I recorded the various fan rotation speeds using SF-609 controller and sound level of each setting. If the fan does not have speed sensor I indicated voltage used.



The chart answers a couple of questions in my mind:

1) The higher the fan rotation the higher the sound level.
2) Given the same fan speed larger fan does not mean lower noise level.



I am making an assumption that most PCs are about 18" from the user; second thought I believe the chart above shows high to low RPM readings corresponding to distance will better serve. Yes, the farther I am away the less the noise. At lowest setting 120 mm fans are not much quieter an 80mm at the same rotation.



I use one each 120, 92 and 80 mm size fans to emulate water cooling setup in a minimum number of fan scenario; 120 mm radiator fan, 92 mm exhaust fan and 80 mm front intake for hard drives. The same setup could be used for 100% air-cooling, say the 120 mm is on the Thermalright XP-120 for instance.

Granted, water cooling may well yield lower CPU temp only to be raised for more overclocking, hence noise is not reduced in some cases.

Thoughts :

I find manufacturer specifications on dBA rating are mostly out of line in real world. Based on what I’ve seen and experienced, a quiet room is about 35 to 40 dBA during day time and a very quiet room may be down to 30 dBA which is rare in today’s society with so much noise pollution. Ratings of most fans are below 35 dBA which means you would hardly hear them run. It is NOT the case here.

The size of the fan increase does not mean reduction of noise when rotation speed is the same as the smaller counterparts. However, large size fan could replace two smaller size fans and have the same effect on air movement, which in some cases do reduce noise level.

Water cooling does not provide any less noise pollution; in fact, it increases noise level because if a single 120 mm is not enough one could get a radiator that will house three of them.

The noise level generally will increase when the fans are mounted on heatsink and radiator from the effect of wind resistance and turbulence.


Conclusion:

"The Tree sounds when the Wind blows" is a valid statement. I find the options to reduce noise level from PC using active cooling water-cool or heatsink are:

1) Reduce the number fans used whenever possible.
2) Use sound insulation materials
3) Use larger fan only if it replaces two smaller ones.
4) Use low RPM fans whenever possible.
5) Stay away from buying high heat loss processors.

As I "mild" overclocker, I would have to cut my loss on the Prescott. As a consumer, When I read the specification stating an 80 mm fan with 3,000 RPM and 28 dBA or a 120 mm at 2,500 RPM with 32 dBA then I know the reading must be taken many feet or meters away. More about building a silent PC here.

This article is not to inflame anyone with my opinion based on the findings. It is merely an exercise in refreshing the little knowledge I have had learnt more than 15 years ago. Please feel free to make your comments forum thread