If this is your first visit, be sure to check out the FAQ by clicking the link above. You may have to register before you can post: click the register link above to proceed. To start viewing messages, select the forum that you want to visit from the selection below. |
|
|
Thread Tools | Display Modes |
#31
|
|||
|
|||
Timothy Daniels wrote: "Johannes H Andersen" replied: Timothy Daniels wrote: Have you thought about just building a 4-sided plywood box around it that had cheap carpeting stapled to the inside surfaces? The box could even have a carpeted floor that didn't extend all the way to the rear wall so that cables could still drop down behind the desk. The front and rear walls could be doors that swung open for access to cables (rear) and to drives (front), and the front door would be open at the bottom (à la bar room door) to allow fresh air intake. Essentially, that would be giving up on quiet fans and relying instead on a sound absorbing partial cabinet around the case. The reason that companies haven't offered them is that they are so low tech (and bulky and heavy) that there would be no profit in them. But that makes them perfect for the homebuilder. A problem though with this solution is that it also keep the heat inside. Sound absorbing material is usually also heat insulating. Secondly, our sound perception follows a logarithmic scale; a small audible reduction of 3 dB require a 50% reduction in sound energy. Read the description again. There is no top to this cabinet, and there is free convection in the rear since there is no floor under the rear cables, and the front door does not go all the way down to the desk surface. There is no interference with air flow, and there is no dependence at all on conduction. All this cabinet does is force most of the sound to bounce at least once off a carpeted surface. A reduction of more than 50% of sound energy is easy. I did a crude experiment of just suspending a 2'x3' sheet of carpet against the wall behind my midi tower, and the reduction in sound was surprising. *TimDaniels* Then you have cut out reflections of noise from the wall. It's OK if you can remove noise by some simple means. It is unfortunate that the cabinet sometimes amplifies the noise, I've found that opening the cabinet on my PC actually reduces the noise, but then it diverts the airflow which you don't particularly want. Hence there are acoustic considerations as well. I have tried to shield my PC with a plywood box, but it gave little result and access was cumbersome if I wanted to do something. Instead I now have a good CPU heat sink with temperature control (the CPU fan rarely runs) plus a 50 Ohm 3 Watts resistor in series with the PSU fan. The PSU fan resistor was carefully chosen by feeling the temperature of the PSU components. |
#32
|
|||
|
|||
small audible reduction of 3 dB require a 50% reduction in sound
energy. o The letter after dB is the Weighting applied ---- (A) is used for humans - poor at 25dB(A) since it is an approximation -------- qualitative nature of the noise becomes important -------- for fans bearing choice, vibration, resonance, blade-to-housing gap ---- (A) is really aimed at industrial noise -------- hence there is a new ISO standard aimed at super-silent noise levels o A reduction of 10dB(A) is a 50% reduction in sound energy ---- 3dB(A) is the margin most people can detect ---- however this too depends on the noise level re 20s or 70s o Each 180-degree turn you make sound go thro reduces it 6dB(A) o The most cost effective sound dampening is treating the Cause ---- then treat the Effect - which is much more difficult re sound-proofing -------- you have lots of holes in a PC -------- you have case-fans outside the soundproofing inside the case -------- CPU-fan / drives / soundproofing / case-fan / ear :-) QuietPC 3-pk soundproofing is very good re price - good peelable stick, since it's very heavy & ok closed-cell foam. Of course the foam is a bit lame as it's just ~5mm thick (limited by PC space in present designs). You need mass for low frequency, and closed-cell foam for high frequency. Most PC sound emitters are few milli-watt, so it's a foam issue vs just mass. Resonance is an issue re fan & other motorised object mounting & vibration. Treat cause, treating effect is more difficult. The new BTX form-factor will initially solve a lot of problems: o Blow thro heatsink, blow thro PC design -- like the G5 ---- that will allow sub 30dB(A) PCs o However CPU & Graphics & HD power output are rising fast ---- 125W CPUs are the first, and frankly dual-CPU machines will hit 400W ---- IDE HDs are moving from 7200rpm to 10,000rpm & SCSI beyond 15300rpm So the silence may be short-lived - but at least better than the ATX alternative. Have fun :-) -- Dorothy Bradbury, backup email: www.stores.ebay.co.uk/panaflofan (Ebay) http://homepage.ntlworld.com/dorothy...ry/panaflo.htm (Free 1st-Class Shipping) |
#33
|
|||
|
|||
Gary D. wrote:
On Sat, 27 Sep 2003 12:35:43 -0700, Biffo wrote: On this day of our lord, Sat, 27 Sep 2003 10:42:28 +0100, Gary D. quilled: I replaced my PSU's fan with a Vantec Stealth 80mm fan (reviewed as being supposedly "silent") - it reduced the noise levels slightly but I still consider my PC to be as noisy as hell. Panaflow fans are apparently quieter than Vantec Stealth. Also you can get PSU's like the Zalman that run the fan in the PSU at very low rpm unless the temp get's inside the PSU gets over 50deg c. I didn't go for Panalflow fans because of an article I read on the web (can't remember where!) that said not all Panaflows were effective in noise reduction. Perhaps because they provide a higher airflow instead? There's always a trade-off between volume and volume, if you see what I mean :-) |
#34
|
|||
|
|||
"dorothy.bradbury" wrote: small audible reduction of 3 dB require a 50% reduction in sound energy. o The letter after dB is the Weighting applied ---- (A) is used for humans - poor at 25dB(A) since it is an approximation -------- qualitative nature of the noise becomes important -------- for fans bearing choice, vibration, resonance, blade-to-housing gap ---- (A) is really aimed at industrial noise -------- hence there is a new ISO standard aimed at super-silent noise levels o A reduction of 10dB(A) is a 50% reduction in sound energy ---- 3dB(A) is the margin most people can detect ---- however this too depends on the noise level re 20s or 70s I have to correct your science here. A doubling of absolute signal power represents an increase in dB of 3dB. Reference http://ccrma-www.stanford.edu/~oded/M151/decibels.html , but this is also generally known from any textbook. Applying a profile as in db(A) makes no difference since dB reduction is about relative level differences; putting in a profile will preserve ratios. ( Profile * signal_1 )/(Profile * signal_2) = signal_1/signal_2 "3dB(A) is the margin most people can detect" This is correct. |
#35
|
|||
|
|||
Human Sensitivity:
o 3dB is the difference in noise level a human can perceive o 1dB is the difference perceivable under laboratory conditions ---- vis., for general purchase use the figure of 3dB applies Human Perception: o 10dB increase in noise level is Perceived as a doubling in noise level o 10dB decrease in noise level is Perceived as a halving in noise level Energy Intensity: o 3dB corresponds to halving/doubling the noise sources o Intensity is important for exposure times as one exceeds ~85dB(A) The 10dB figure is rough, but standard practice - some prefer to use 6. Note this is a *Subjective* Human-Perception issue not an absolute. Building data is useful - since PCs are just another noise component: http://www.n-p.com/engineering/sound.asp gives OSHA data too. However, for PC use, the "A" weight is poor below ~25dB(A) levels: 1) it pooly represents the profile of human perception ---- it was designed to time & sensitity weight loud noises re H&Safety 2) it poorly represents the perception of noise components ---- frequency, harmonics, white-noise or not components & so on ---- eg, hum, bearing rumble, blade-tip-speed, diagonal v axial flow ---- for a consideration of fan noise components see Neise (1988) I'm well aware of the electronic usage, but when one is referring to human perception, human factors, the above figures are standard. Absolute silence is -20, anechoic chambers background often ~16. Which is one reason why fan design hits law of diminishing returns, and unfortunately many PSUs suffer coil-hum as much as fan-noise. Same with PWM solutions that use 1-2kHz PWM vs 10-20kHz. The solution is fan-&-system design, since correct use of baffles and moving fans off the case surface allows soundproofing to work. Work on sub & supersonic fan noise mgt is more advanced than PCs. That said, the PC even with BTX is somewhat behind the Dual-G5. -- Dorothy Bradbury, backup email: www.stores.ebay.co.uk/panaflofan (Ebay) http://homepage.ntlworld.com/dorothy...ry/panaflo.htm (Free 1st-Class Shipping) |
#36
|
|||
|
|||
"dorothy.bradbury" wrote: Human Sensitivity: o 3dB is the difference in noise level a human can perceive o 1dB is the difference perceivable under laboratory conditions ---- vis., for general purchase use the figure of 3dB applies Human Perception: o 10dB increase in noise level is Perceived as a doubling in noise level o 10dB decrease in noise level is Perceived as a halving in noise level This sounds correct, but you said in your earlier post: "o A reduction of 10dB(A) is a 50% reduction in sound energy" So you mixed up 'sound energy' with 'Percieved noise level' . Strictly, as 3dB reduction corresponds to 50% reduction in sound energy, it follows that 10dB corresponds to a residual factor of 0.5^(10/3) in other words 90% reduction in sound energy. |
#37
|
|||
|
|||
On Sun, 28 Sep 2003 09:31:37 +0100, Gary D.
wrote: I didn't go for Panalflow fans because of an article I read on the web (can't remember where!) that said not all Panaflows were effective in noise reduction. I was only a matter of having read that article that made me go for Vantec, simply a matter of one choice over the other, although I suspect you could be right! Panasonic (Panaflo), like Vantec, makes a variety of fan sizes and speeds. Their (Panaflo's) dual ball-bearing fans are rare in fans appropriate for PC applications, but their sleeve-bearing fans are much better than the typical sleeve-bearing fan. This allows the noise reduction of a sleeve bearing but without as much of a lifespan decrease. Certainly there are other good sleeve-bearing fans like Pabst, but in general most other good fans are ball-bearing. The fans generally regarded as very quiet are "L1A", as a portion of the model number where "L" is low, "1" means "with leads", and "A" simply means it has no sensor option, only 2 power wires. Certainly a sensor could be useful too, but this is the most common configuration. Their "M1A" series is also good for noise reduction if you need more ailrflow than a lower RPM fan can provide. Of course it's louder than L1A but good noise/airflow ratio. Another factor I think might still be contributing to the noise levels I'm experiencing is that my PSU unit is a cheap one that came with my PC and the PSU box fan aperture is not quite as large as the fan diameter, so some of the air flow is hitting the edges of the box aperture and this, I guess, must be causing turbulence and thus the annoying "hiss". To remedy this would involve removing the entire contents of the box and expanding the fan aperture; even then I reckon any resulting improved aperture might have rough edges (using just DIY tools) and it's unlikely I would achieve a professionally smooth job. Cheap power supplies often have small heatsinks and can produce more heat per power output, and also may have a more aggressive fan speed. Replacing the power supply would be the best alternative, and since your current power supply is still working, if you wanted to modify the new power supply you have ample time to do it, system downtime wouldn't be a factor. Also if the cheap power supply has a likewise cheap sleeve-bearing fan it may be subject to failure, is "usually" one of the first parts to fail in such a system. Dave |
#38
|
|||
|
|||
"Johannes H Andersen" wrote: Then you have cut out reflections of noise from the wall. Yes. The intention is to keep the noise from getting to one's ears, not from keeping within the tower. It's OK if you can remove noise by some simple means. It is unfortunate that the cabinet sometimes amplifies the noise, That is one purpose of the carpeting - to damp out vibrations of the cabinet walls. Cheap foamed-backed stick-on carpeting squares would be perfect for this. I've found that opening the cabinet on my PC actually reduces the noise, but then it diverts the airflow which you don't particularly want. If you leave a 6-inch clearance in the back and a 3-inch clearance in the front, the interference in air flow will be negligible. Remember that both the rear and the front walls of the cabinet do not contact the desk or the floor of the cabinet but leave large gaps for air flow. Hence there are acoustic considerations as well. I have tried to shield my PC with a plywood box, but it gave little result Try it again with carpeting lining the inner surfaces. and access was cumbersome if I wanted to do something. An access door in the front, and that 6-inch clearance at the back would help with that. Instead I now have a good CPU heat sink with temperature control (the CPU fan rarely runs) plus a 50 Ohm 3 Watts resistor in series with the PSU fan. The PSU fan resistor was carefully chosen by feeling the temperature of the PSU components. Yes, my fingers are the calibrated type, too. :-) But you must remember that the PSU fan contributes to the overall air flow in the tower, and if it has an intake on its bottom wall, it helps to ventillate the CPU. *TimDaniels* |
#39
|
|||
|
|||
"dorothy.bradbury" wrote: o A reduction of 10dB(A) is a 50% reduction in sound energy ---- 3dB(A) is the margin most people can detect ---- however this too depends on the noise level re 20s or 70s In the audio world, 1 dB is the accepted limit of human perception (if they're listening for it). 3 dB may be the limit for casual (or "unprepared") listening. As long as you're referring to human perception by adding the "(A)" to your dB figures, you might as well add "apparent" to "sound energy" terminology. And "db" is a measure of power, not energy; and instead of "energy", you ought to be saying "intensity" for perception effects. In short, you mix your terms from the physics basket and from the psychoacoustics basket. In the physics (and engineering) world, 10dB equals one Bell - a change in power by a factor of 10, not a factor of 2. A factor of 2 entails a change in power of 3dB. o Each 180-degree turn you make sound go thro reduces it 6dB(A) groan And if you spin around 10 times the sound will reduce by 120dB! If you merely mean "fluctuate" by 6dB(A), you have some explaining to do about how that sound does or doesn't reflects of walls, refracts around the edges of the ears, how it conducts through flesh and bone of the ear/head, etc., instead of that blanket one-size-fits-all statement. In college, conclusions stated without any comment on the methodology was called "hand waving". *TimDaniels* |
#40
|
|||
|
|||
Timothy Daniels wrote:
In college, conclusions stated without any comment on the methodology was called "hand waving". In engineering "rule of thumb" is adequate in many places, although I wouldn't expect a physicist to be happy with that :-) |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Computer Buying Advice | [email protected] | General | 5 | September 3rd 03 07:58 AM |
newbie - question about heat generated by computer | kony | General | 0 | August 31st 03 01:17 AM |
How to determine the age of a computer ? | Edward | General | 14 | July 10th 03 12:39 AM |
computer freezes when booting | Dan Chirica | General | 1 | July 6th 03 07:01 PM |
OEM drives? How silent are 2.5" hardisks? | Doru-Catalin Togea | General | 1 | July 3rd 03 11:49 PM |