5dBA 120mm Fan.

Control: management of fans, temp/rpm monitoring via soft/hardware

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

goatsandmonkeys
Posts: 91
Joined: Wed Oct 06, 2010 6:32 am
Location: new york

Re: 5dBA 120mm Fan.

Post by goatsandmonkeys » Mon Feb 21, 2011 10:32 am

This is fascinating thanks for everyone for posting.

could it be that this experiment explains mike's 6 year old results? Maybe the baseline noise of a room makes it seem that extremely quiet fans seem to not add 3db when doubled, but you can only detect the 3db bump when the fans are loud enough to make the baseline noise irrelevant.

PartEleven
Friend of SPCR
Posts: 279
Joined: Sun May 06, 2007 10:37 am

Re: 5dBA 120mm Fan.

Post by PartEleven » Mon Feb 21, 2011 12:16 pm

This is quite an interesting and lengthy analysis. Excuse me if this was answered somewhere in the previous posts, but where is your justification for the linear relationship between SPL and voltage? Other than your curve fitting, that is.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: 5dBA 120mm Fan.

Post by ces » Mon Feb 21, 2011 12:41 pm

This is all very interesting. And I am looking forward to where this thread ends up.

And all of us here are to to a greater or lesser extent intrigued by fans. But it occurs to me that of all the problems this website addresses itself to, fans are the simplest in actual execution.... that is until we get obsessive about it.

Here is a formula that will work for all but the most obsessive of us (and I am the first to admit that I have quite an excessive collection of fans, far more than I need, and it will likely grow soon after the next SPCR fan review):

Step 1. Noctua Heatsink.

Replace the stock CPU heatsink with any Noctua heat sink that will fit. There are better. There are cheaper. There may be even cheaper and better. But Noctua are always among the best. It is never a mistake to select a Noctua. Just make sure it is one that accepts 120mm fans. When Intel changes platforms, Noctua has always been there with a mounting kit they supply for for free.

You can argue whether or not there are better decisions in a given situation. But a Noctua decision is always a good decision.

Step 2. Forget Fan control.

Just use the stock Noctua fans that come with the heat sink. Use the Noctua power cables to set them at the lowest speed they will go. No need to muck around with fancy speed controls or anything like that.

Step 3. Replace Your Case Fans

Just replace all your case fans with 12 volt 800 RPM slipstreams (model SY1225SL12L). Don't muck around with speed controls or anything. Just plug them in at a stock 12 volts.

Don't use a case that uses fans smaller than 120mm. If a case requires fans greater than 120mm, don't use that case.

Step 4. Check your temps.

Your temps should be just fine. If they aren't fine, no problem. Go to the next optional step.

Step 5. Optional Nexus Fans

Get some Nexus Real Silent Case Fans (model D12SL-12) Again don't muck around with speed controls or anything. Just plug them in at a stock 12 volts.

Replace all the other fans with the Nexus fans. This might cost you 2 or 3 dBs, but it will still be reasonably quiet. And there should be an appreciable drop in your temps.

If this last optional step doesn't work, there is something wrong. Whatever the problem, it isn't a fan problem and it doesn't have a fan solution.

===========

Can you do better by futsing with all the stuff we all obsess about here? Sure. But the above will probably produce a quiet well cooled system that will work for all but the most obsessive of us. A quiet well cooled system that, if we had it at the time, would have preventing us bothering to come here in the first place.

Yes?
Last edited by ces on Mon Feb 21, 2011 12:50 pm, edited 2 times in total.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: 5dBA 120mm Fan.

Post by ces » Mon Feb 21, 2011 12:45 pm

PartEleven wrote:This is quite an interesting and lengthy analysis. Excuse me if this was answered somewhere in the previous posts, but where is your justification for the linear relationship between SPL and voltage? Other than your curve fitting, that is.
I don't think there is.

By the way, if you want to see the wide range of power utilization (watts) by fans with comparatively similar RPM, CFM, and even noise outputs, take a look at this:

http://www.xbitlabs.com/misc/picture/?s ... ig.png&1=1
http://www.xbitlabs.com/articles/cooler ... html#sect0

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: 5dBA 120mm Fan.

Post by ces » Mon Feb 21, 2011 1:01 pm

goatsandmonkeys wrote:could it be that this experiment explains mike's 6 year old results? Maybe the baseline noise of a room makes it seem that extremely quiet fans seem to not add 3db when doubled, but you can only detect the 3db bump when the fans are loud enough to make the baseline noise irrelevant.
That sounds pretty plausible to me.

goatsandmonkeys
Posts: 91
Joined: Wed Oct 06, 2010 6:32 am
Location: new york

Re: 5dBA 120mm Fan.

Post by goatsandmonkeys » Mon Feb 21, 2011 2:48 pm

ces wrote:This is all very interesting. And I am looking forward to where this thread ends up.
I'm enjoying where it already is :) I'm learning a lot. Also I am hoping that a big serious discussion about fans will positively influence the new fan test methodology.
ces wrote:
PartEleven wrote:This is quite an interesting and lengthy analysis. Excuse me if this was answered somewhere in the previous posts, but where is your justification for the linear relationship between SPL and voltage? Other than your curve fitting, that is.
I don't think there is.

By the way, if you want to see the wide range of power utilization (watts) by fans with comparatively similar RPM, CFM, and even noise outputs, take a look at this:

http://www.xbitlabs.com/misc/picture/?s ... ig.png&1=1
http://www.xbitlabs.com/articles/cooler ... html#sect0
Fascinating. the second link shows what you and others were saying about using more fans at slower speeds. most notably the slipstream 12H. at just over 30DB it pushes18.75 CFM and at just under 45 DB (about 30 times as loud) it pushes 36.67 CFM. two fans at the low speed woudl be 33DB and push 37.5 CFM

So what abut coolers with push/pull fans? do two fans in a series at the same speed produce less or more noise than two fans in parallel at the same speeds? do the fans in parallel produces more or less noise causing turbulence?

I enjoyed the recent "What I have Learned about Computer Performance from SPCR" thread with it's simplified rules. maybe we need a collection of general fan laws.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: 5dBA 120mm Fan.

Post by ces » Mon Feb 21, 2011 3:49 pm

goatsandmonkeys wrote:So what abut coolers with push/pull fans? do two fans in a series at the same speed produce less or more noise than two fans in parallel at the same speeds? do the fans in parallel produces more or less noise causing turbulence?
Good question. I saw somewhere someone doing an Einstein type of thought experiment. Starting out with two fans at opposite sides of the room, each pushing 80 cfm (a total of 160 cfm) and then asking what happens as you bring them closer to each other and put them on two sides of the tower heat sink. Are they still pushing a total of 160 cfm? Why not.

goatsandmonkeys wrote:I enjoyed the recent "What I have Learned about Computer Performance from SPCR" thread with it's simplified rules. maybe we need a collection of general fan laws.
I certainly would like to see a collection of general fan laws come out of this thread.

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Re: 5dBA 120mm Fan.

Post by MikeC » Mon Feb 21, 2011 4:05 pm

Great discussion, guys. Lots of food for thought, many interesting observations.

I think photonblaster's original post & comments have real merit.

1) The relationship between speed & noise is probably more linear (with some fans) to a lower level than the 10-11 dBA noise floor of the SPCR anechoic chamber. (That is the true noise floor, btw -- it is what we measure w/o any sound sources in the chamber and virtually none elsewhere in the house.) Conjectures by others on the noise floor limitation of bearing and electrical noise surely apply -- there has so be some minimum noise below which a fan simply cannot get go.

2) The extrapolation of "real fan noise" below the 11 dBA noise floor seems reasonable enough, and it does explain why doubling up on a fan that "measures" 11 dBA in the chamber has virtually no change on measured SPL. This has happened (tho I can't say with which fan right now).

3) The idea of using the extrapolated low level fan curves to get a better estimate of the SPL of multiple low speed fans seems sound. Obviously very useful if you don't want to mess with a whole bunch of fans to find out for yourself.

4) We could redo SPL readings for some fans at 0.5m or maybe even 0.33m to get a better sense of the noise curve at lower levels. This could be used to extrapolate "true" fan SPL at very low speeds -- in order to predict the noise of multiple slow/quiet fans. We could also just include 1m SPL measurements with 2 and 3 identical fans in tandem at various speeds. This really is relevant only for fans that are cream of the crop quiet. Absolutely no point with a typical buzzy ball bearing fan that doesn't start at less than 600rpm, for example.

5) I'm not sure what I can add to my post from 6 years ago. I've listened to and played with a lot more fans... but my basic approach has not changed:
-- multiple quiet fans are better than fewer higher speed fans for both noise and cooling
-- change the speed enough and the whole freq balance changes, along with bearing tics, etc
-- even when the noises are lower than ambient (by SPL), they can still be annoying/audible; it all depends...
-- speed "tuning" by ear against monitored component temps with fans that have a basically benign noise signature is still the best way to get exactly the PC noise level you want, along with acceptably safe temps. That "benign" noise signature is a somewhat subjective assessment; you may disagree with my judgment of what is benign and what is not. (Though there are lots of noises that are universally perceived as nasty, I suspect there are fewer sounds that are universally perceived as pleasant or benign.)

photonblaster
Posts: 59
Joined: Sun May 17, 2009 9:26 am
Location: North America

Re: 5dBA 120mm Fan.

Post by photonblaster » Mon Feb 21, 2011 4:21 pm

PartEleven wrote:This is quite an interesting and lengthy analysis. Excuse me if this was answered somewhere in the previous posts, but where is your justification for the linear relationship between SPL and voltage? Other than your curve fitting, that is.
PartEleven, I mentioned in my original post that I had to resort to using SPL vs voltage so I could predict what the two fans did together because that is the way SPCR did their measurments. IMO, the real relationship is between SPL and CFM.. CFM is close to linear with RPM which is close to linear with voltage, so SPL is close to linear with voltage. Good case to illustrate high correlation does not prove cause and effect.

photonblaster
Posts: 59
Joined: Sun May 17, 2009 9:26 am
Location: North America

Re: 5dBA 120mm Fan.

Post by photonblaster » Mon Feb 21, 2011 5:00 pm

ces wrote:===========

Can you do better by futsing with all the stuff we all obsess about here? Sure. But the above will probably produce a quiet well cooled system that will work for all but the most obsessive of us. A quiet well cooled system that, if we had it at the time, would have preventing us bothering to come here in the first place. Yes?
Great practical approach, ces. Maybe add start with a SPCR recommended case and power supply?
I do not consider myself obsessed about quiet systems, it is more of a personal challenge to produce a system whith relatively heavy OC but keep it quiet. But then some might say this is just a rationalization to camoflage my obsession :D

photonblaster
Posts: 59
Joined: Sun May 17, 2009 9:26 am
Location: North America

Re: 5dBA 120mm Fan.

Post by photonblaster » Wed Feb 23, 2011 4:42 pm

Here is an interesting question for those readers who suggested my original post has little merit. Suppose I did my original post in 2006? For instance I would have done the analysis below for the Noctua NF-S12-1200, data found at http://www.silentpcreview.com/article69 ... html#nexus At that time the noise floor level was about 18+ dBA or so.

Following the same analysis method, I would have estimated the noise from the fan alone to be about 11 dBA at 500rpm. This happens to be a relatively noisy fan. Since no one can have an office space with noise levels anywhere near this low, I guess this calculation to better estimate fan noise at low rpm would have no value as well? And you cannot have a fan alone anyway, right?

In hindsight, it probably would have been better to say...."noise from the fan in an environment where the dominant noise source is the fan". I just figured out where that might be for the really quite fans at low rpm. A little tough to get there, but consider having an anechoic chamber on the space station. Turn off all power except for the fan and the "no moving parts" PC, set the measurements on a time delay and go on a space walk. Of course you probably also would need a million dollar microphone to get the readings! Good thing there is no law against a scientist liking a little science fiction as well!
You do not have the required permissions to view the files attached to this post.

Modo
Posts: 486
Joined: Wed Apr 16, 2008 3:32 am
Location: Poland

Re: 5dBA 120mm Fan.

Post by Modo » Wed Feb 23, 2011 11:38 pm

Maybe it's just me, but the only important data is, when does a fan get inaudible in my system? This is what SPCR tests answer quite well. So what if fan noise levels can be extrapolated much lower? I, and probably everyone here, don't get such low background noise even on very quiet nights. And we all know that using too many fans means noise no matter what you do. So yeah, interesting theoretical distraction, but not much more.

Back to the real world, when's that new fan roundup coming? ;)

ame
Posts: 488
Joined: Tue Jul 10, 2007 10:35 pm
Location: Israel

Re: 5dBA 120mm Fan.

Post by ame » Thu Feb 24, 2011 2:02 am

photonblaster wrote:Here is an interesting question for those readers who suggested my original post has little merit. Suppose I did my original post in 2006? For instance I would have done the analysis below for the Noctua NF-S12-1200, data found at http://www.silentpcreview.com/article69 ... html#nexus At that time the noise floor level was about 18+ dBA or so.
To be accurate we seem to not be looking at the same review. The link you posted at the top of the page has 19db as noise floor level and not 18db as you say, certainly not 16-17 as your graph indicates.
There are 3 valid measurements you can use from that review (not 4) - 25db@12V, 22-21db@9V, and 20db@7V. the <19 cannot be represented as a point on a graph since you don't know it's value. The noise floor in your graph is 2-3 db lower than actual noise floor measured.

http://www.silentpcreview.com/article69 ... tml#noctua

Once again I think your math is solid. Its just you are not taking into account how sound really works. Its not just db levels that can be added.

This fan, if it was measured in todays SPCR low noise anechoic lab, using you'r formula would yield a score of 5-7 db for 'the fan alone' much like the other Noctua fan, indicating that the formula's result (11db vs 5-7db) is related to the room and does not produce an absolute level for the sound of 'the fan alone' as you call it, but rather a theorized number that does not apply in any way to real life.

To best see that db is relative and not absolute (thus good for comparisons rather than absolute conviction;)) look at nexus fan ratings from that review and compare them to current SPCR Chamber levels

Here are today's SPL @ 1m dbA levels
Reference Nexus 120mm fan measurements (noise floor 11dbA)
12V = 16 dBA@1m = 1100 RPM
9V = 13 dBA@1m = 890 RPM
7V = 12 dBA@1m = 720 RPM
5V = 11 dBA@1m = 530 RPM

and here are the numbers for the same fan (noise floor 19dbA) 2006/2008

12V = 22dBA@1m = 1080RPM
9V = 19dBA@1m = 850RPM
7V = <19dBA@1m = 680RPM
5V = <19 dBA@1m = 490RPM

It seems slight speed shifted due to possibly sample variance or slight voltage mismatch. But I am almost certain (and Mike C correct me if I'm wrong here) these fans are essentially the same.

The facts that you have showed a mathematically linear relationship between voltage and CFM and a near perfect logarithmic relationship for fan noise and noise floor (that only applies to fans and noise floors as measured is our current scenarios) are interesting, otherwise I would not bother posting. But there is no meaning to this as it does not apply to anything other than analysis of the test results we already have.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: 5dBA 120mm Fan.

Post by ces » Thu Feb 24, 2011 3:33 am

Modo wrote:Maybe it's just me, but the only important data is, when does a fan get inaudible in my system? This is what SPCR tests answer quite well. So what if fan noise levels can be extrapolated much lower? I, and probably everyone here, don't get such low background noise even on very quiet nights. And we all know that using too many fans means noise no matter what you do. So yeah, interesting theoretical distraction, but not much more
I have to admit that I am struggling to maintain an understanding of the discourse on this thread.

Still, it does seem relevant to the assumption of fan selection. Can someone interpret this discussion and apply it to the following decision:
(a) two slower fans moving 15 CFM each in free air
(b) one faster fan moving 30 CFM in free air

(for the sake of simplicity - let's say all three fans are slipstreams operating at different RPMs)

Which do you choose, under which circumstances?

And to what extent does impedance (back pressure) affect this decision? It does effect this decision. It has to. When you apply back pressure to these two scenarios, the CFM of scenario (a) will diminish more quickly than that of scenario (b). That means you will have to increase the speed of the two fans in scenario (a) to match the CFM performance of the single fan in scenario (b). Though I would propose the impedance issue differs depending on whether the two fans are being operated in series (such as with a CPU tower heat sink) or in parallel (such as in the situation of case fans).

To sort of summarize. There are only 3 variables of significance: 1. dB sound levels, 2. CFM, and 3. the impedance faced by the fans. (as MikeC pointed out all dBs are not equal, but to keep things from careening out of control, for the sake of argument, let's provisionally say all dBs are the same).

Can anyone take what we are learning here (or what they have learned elsewhere), and apply it in a manner to generate practical guidelines for dealing with the above decision... or heck even impractical theoretical guidelines... or even unfocused banter.... or anything.

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Re: 5dBA 120mm Fan.

Post by MikeC » Thu Feb 24, 2011 8:20 am

ces wrote:Can anyone take what we are learning here (or what they have learned elsewhere), and apply it in a manner to generate practical guidelines for dealing with the above decision... or heck even impractical theoretical guidelines... or even unfocused banter.... or anything.
OK, here's an attempt:

A. Objective: To achieve good cooling of core components with the absolute minimum audible noise in a PC. This ignores cost.
B. Assumption: At least 1 fan in the system & the ability to control the speed of each fan. Heatsinks are close to ideal for low airflow cooling performance.
C: Fan options: All 120mm fans, potentially up to 8. (Just to set a guideline) 1 on pack panel, 1 on top panel (like in P183), 2 on front panel, 2 on CPU HS, 2 on VGA HS.

1. Start with the minimum number of fans you believe is needed. Choose fans on the basis of best noise/rpm ratio at <16 dBA/1m by SPCR testing.
2. Tweak the speed of each fan by ear/temp while applying stress utility appropriate to your intended usage. If satisfactory noise/temps are reached, congratulate yourself, have a drink.
3. If not satisfied, add a fan somewhere -- or perhaps 2 -- and redo the process. Focus on getting the overall noise of the fans as quiet or quieter than before, while improving or maintaining temps (depending on whether temps or noise was the issue before). If satisfactory noise/temps are reached, congratulate yourself, have two drinks.
4. If not satisfied, repeat as necessary... or re-examine your fan selection and consider changing to another model of fan. Have drink to kill the pain.

photonblaster
Posts: 59
Joined: Sun May 17, 2009 9:26 am
Location: North America

Re: 5dBA 120mm Fan.

Post by photonblaster » Thu Feb 24, 2011 5:01 pm

ame wrote:There are 3 valid measurements you can use from that review (not 4) - 25db@12V, 22-21db@9V, and 20db@7V. the <19 cannot be represented as a point on a graph since you don't know it's value. The noise floor in your graph is 2-3 db lower than actual noise floor measured.
Thanks for taking the time to read my post carefully.
Good eye on the 18.5dBA value, but there is an explanation for that. SPCR rounds off their dBA values to an interger even though the sprectrum analyzer gives results to 2 decimal places. So the value when they say <19dBA is 18.5+/-0.5dBA. The exact value does not matter considering the intent of my last post. The least squares fit does not change that much regardless of where in the 18-19dBA.

ame
Posts: 488
Joined: Tue Jul 10, 2007 10:35 pm
Location: Israel

Re: 5dBA 120mm Fan.

Post by ame » Thu Feb 24, 2011 10:13 pm

photonblaster wrote:
ame wrote:There are 3 valid measurements you can use from that review (not 4) - 25db@12V, 22-21db@9V, and 20db@7V. the <19 cannot be represented as a point on a graph since you don't know it's value. The noise floor in your graph is 2-3 db lower than actual noise floor measured.
Thanks for taking the time to read my post carefully.
Good eye on the 18.5dBA value, but there is an explanation for that. SPCR rounds off their dBA values to an interger even though the sprectrum analyzer gives results to 2 decimal places. So the value when they say <19dBA is 18.5+/-0.5dBA. The exact value does not matter considering the intent of my last post. The least squares fit does not change that much regardless of where in the 18-19dBA.
How do you know it wasnt 14 db?

All you reall know is that is was lower than 19 or in other words not did not register above the amb noise.

Also how do you explain the Nexus and the diffrent outcome it would produce base on this equasion whene background noise is changed?

Modo
Posts: 486
Joined: Wed Apr 16, 2008 3:32 am
Location: Poland

Re: 5dBA 120mm Fan.

Post by Modo » Thu Feb 24, 2011 11:23 pm

Mike, what you have described works independent from the findings shown here. I think you are making my point.
ame wrote: All you reall know is that is was lower than 19 or in other words not did not register above the amb noise.
Thank you, too, for making my point. When the ambient noise of the lab is as low as a very quiet night in a very quiet location, then "does not register above the lab ambient noise" means "will be inaudible in my system at night". That's actually a better--practical--measurement than the theoretical extrapolations shown in this thread. Sorry to be Captain Obvious, but the ambient noise is a factor you should not ignore when planning a real system.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: 5dBA 120mm Fan.

Post by ces » Fri Feb 25, 2011 5:05 am

Modo wrote:Thank you, too, for making my point. When the ambient noise of the lab is as low as a very quiet night in a very quiet location, then "does not register above the lab ambient noise" means "will be inaudible in my system at night". That's actually a better--practical--measurement than the theoretical extrapolations shown in this thread. Sorry to be Captain Obvious, but the ambient noise is a factor you should not ignore when planning a real system
Does the analysis change if you treat the background noise, not as something special, but as the same as the noise of one more fan? This is an open question to anyone.

photonblaster
Posts: 59
Joined: Sun May 17, 2009 9:26 am
Location: North America

Re: 5dBA 120mm Fan.

Post by photonblaster » Sat Feb 26, 2011 5:25 am

ces wrote:Does the analysis change if you treat the background noise, not as something special, but as the same as the noise of one more fan? This is an open question to anyone.
The answer may depend on your point of view. I see two responses:
1) Purely mathematically, the noise value does not matter where it came from. It is a weighted average over all frequencies. It is a mesured number.
2) From a listening point of view, the backgound level probably has less "spikes" in the spectrum (you can see this in most reviews by SPCR, for instance look at frequency curve for the backgound noise in this article http://www.silentpcreview.com/article1150-page3.html ). So a fan with the same noise level as the background will be heard different.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: 5dBA 120mm Fan.

Post by ces » Sat Feb 26, 2011 7:48 am

photonblaster wrote:2) From a listening point of view, the backgound level probably has less "spikes" in the spectrum (you can see this in most reviews by SPCR, for instance look at frequency curve for the backgound noise in this article http://www.silentpcreview.com/article1150-page3.html ). So a fan with the same noise level as the background will be heard different.
So you are saying it is the same analysis, only that the spectrum of sound from of this "fan" is (a) more predictable than the spectrum of sound from random other "fans", and (b) closer to white noise than you would expect from a random selection of other "fans"?

That its sound spectrum is flatter and more predictable... what does that mean to the analysis? what does that mean in practical application?

Does that mean that, dB for dB, it's dBs are quieter than the dBs of real fans? Or the inverse, that, dB for dB, a fan's dB is louder than a background dB?

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Re: 5dBA 120mm Fan.

Post by MikeC » Sat Feb 26, 2011 8:00 am

ces wrote:
Modo wrote:Thank you, too, for making my point. When the ambient noise of the lab is as low as a very quiet night in a very quiet location, then "does not register above the lab ambient noise" means "will be inaudible in my system at night". That's actually a better--practical--measurement than the theoretical extrapolations shown in this thread. Sorry to be Captain Obvious, but the ambient noise is a factor you should not ignore when planning a real system
Does the analysis change if you treat the background noise, not as something special, but as the same as the noise of one more fan? This is an open question to anyone.
I don't get why you ask this.

Background noise in the anechoic chamber has virtually flat frequency response to below ~200Hz. No fan has such a spectrum. It's also quieter than most normal rooms, and the absence of reflections above 200hz makes it very unusual; most people have never heard/experienced anything quite like it.

Background noise obviously varies a lot in normal spaces where PCs are used, but I doubt it would resemble that of a fan in most homes or offices.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: 5dBA 120mm Fan.

Post by ces » Sat Feb 26, 2011 8:14 am

MikeC wrote:I don't get why you ask this.
Just trying to learn. I wouldn't overestimate the sophistication of the question or the questioner :)
MikeC wrote:Background noise in the anechoic chamber has virtually flat frequency response to below ~200Hz. No fan has such a spectrum. It's also quieter than most normal rooms, and the absence of reflections above 200hz makes it very unusual; most people have never heard/experienced anything quite like it.
So that means it is a pure white noise?

So it is a noise. It is a noise of a special character. So is a dB of this noise less or more intrusive than a dB of noise from a fan? I would think it would be less... though I have never experienced it. But it does count as noise if you are adding additional noise to it, doesn't it?

photonblaster
Posts: 59
Joined: Sun May 17, 2009 9:26 am
Location: North America

Re: 5dBA 120mm Fan.

Post by photonblaster » Sat Feb 26, 2011 9:25 am

ces wrote:
photonblaster wrote:2) From a listening point of view, the backgound level probably has less "spikes" in the spectrum (you can see this in most reviews by SPCR, for instance look at frequency curve for the backgound noise in this article http://www.silentpcreview.com/article1150-page3.html ). So a fan with the same noise level as the background will be heard different.
So you are saying it is the same analysis, only that the spectrum of sound from of this "fan" is (a) more predictable than the spectrum of sound from random other "fans", and (b) closer to white noise than you would expect from a random selection of other "fans"?

That its sound spectrum is flatter and more predictable... what does that mean to the analysis? what does that mean in practical application?

Does that mean that, dB for dB, it's dBs are quieter than the dBs of real fans? Or the inverse, that, dB for dB, a fan's dB is louder than a background dB?
This is getting out of my league...I am a numbers guy, and relating measured curves to how things sound is not my forte.
But, Yes, I think it is true that the room dBA curve is more like white noise just by looking at the curves...a little flater, less spikes.
The analysis does not care where the spikes are, if any. The dBA number totally removes any of these details. So adding dBA together logarithmically does not tell you if the result will sound strange or undesireable harmonics or whatever the listening experts can hear. I think you CAN say that adding two fans together (in a room where room sound is at least 3dBA lower than the fans at the distance you are at) gives you a much better chance of getting an uncomfortable sound than if one of the fans was replaced by a louder room at the same level. But again, this is getting out of my comfort zone, and I am making comments more based on logic than experience.

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Re: 5dBA 120mm Fan.

Post by MikeC » Sat Feb 26, 2011 10:20 am

yeah, you could say the ambient noise in the chamber is still a noise, and its overall profile (ignoring <200Hz) is like white noise -- ie random, pretty much the same level at all frequencies. It does add like any other noise, so if you had a true 11 dBA@1m noise source in the chamber, it should theoretically measure 14 dBA... tho I'm quite sure the chamber's lack of echoes keeps the rise lower than the theoretical +3 dB.

The ambient noise in the chamber is what is left after the ambient noise in the house is filtered by the chamber walls, ceiling, floor and acoustic treatment. The ambient noise in the house includes a couple of fridges many walls away, and the noise of Vancouver, again, filtered by the house. Most city noise is road traffic. There are no nearby highways to speak of, and the closest big street (4 lanes of traffic) is 1.5 blocks away. Maybe 200 yards -- and it's not a clear path, there are trees and houses in the way. In the morning, upstairs, if you get up really early and just lie in bed, you can hear vehicular traffic go from individually recognizable sounds (even it is faint) to a more general rumbling that gets louder through rush hour. Very little of this is heard downstairs (where the chamber is) -- the sound path is more blocked, and there are fewer windows (even tho they are all closed now in winter, more noise gets through upstairs). There's airplane noise, but we never record/measure if they are at all audible in the chamber, so that's not a factor, really.

Not sure what good all of that is for you.... but there you have it anyway: The source/character of ambient noise in the chamber.

photonblaster
Posts: 59
Joined: Sun May 17, 2009 9:26 am
Location: North America

Re: 5dBA 120mm Fan.

Post by photonblaster » Fri Mar 04, 2011 7:59 am

MikeC wrote:Great discussion, guys. Lots of food for thought, many interesting observations.

I think photonblaster's original post & comments have real merit.

1) The relationship between speed & noise is probably more linear (with some fans) to a lower level than the 10-11 dBA noise floor of the SPCR anechoic chamber. (That is the true noise floor, btw -- it is what we measure w/o any sound sources in the chamber and virtually none elsewhere in the house.) .......
4) We could redo SPL readings for some fans at 0.5m or maybe even 0.33m to get a better sense of the noise curve at lower levels. ..........
...etc..........................................
Thank you for the post Mike. It provided some information lacking to me in the first post (the underlined section)...I had assumed that the background noise level had a maximum noise value of 11dBA and fit the data to the model where the noise level was a variable. The low noise level fit for this fan has bothered me a little from the beginning and you have verified my "discomfort". If I now fix the noise level to 10dBA min (actually assume it is 10.5+/-0.5, the exact value in this range does not matter re this current post). This restriction has an interesting consequence..the calculate dBA now has to drop off FASTER than a straight line as you go to lower RPM. This may not be an unreasonable result..as the RPMs are decreased, the harmonic peaks move to lower frequencies where they are counted less be the A-weighting. So there are two pieces of physics acting to reduce the dBA, lower volume of air movement AND frequency reduction to lower the weighted noice value. I suspect that the dB would still fit a linear model with fixed background but I do not know the dB values. The smooth curve through the calculated points is a least squares logarithmic fit to the calculated data, but please consider this as just an aid to the eye to show the data trend even though it is interesting the best curve fit is a logarithmic one.

Your point number 4 would be great to do to verify the accuracy of these simplistic models, especially if both dB and dBA could be recorded. Maybe you could get a high school student to do a science fair project! :)
You do not have the required permissions to view the files attached to this post.

Post Reply