Guide to fan noise and case thermodynamics

[cloudkat]

Hello all,
I was directed here by one of your regulars. I look forward to becomming part of this very active community.

The reason I am posting here is because I have written the following article and posted it on a couple of other forums, and was asked to post it here. I hope you enjoy it.

This thread is (I hope) going to help clear up some misinformation and (also hopefully) give the community a little more knowledge of one of the most overlooked components in a computer case.
A case fan is rapidly achieving the status of a ?major component? when dealing with high-end systems especially when OC?ing.
Fans are used not only in the case itself, but added to passive heatsinks, and more than ever before we are seeing modders/OC?ers replace existing fans on HSFs, PSUs and GPUs to achieve either higher air flow, lower noise levels, or a mixture of both.

I started doing some serious research into fans when I realized that my newest case mod will have a total of 13 fans in it when it is complete. That?s right, 13 FANS !! (The case will be actually divided into sections, so fans are needed to direct the air flow from section to section. I will be posting some work logs in the near future in the Case modding/worklogs section of this forum. The case mod I am speaking of will be called ?From Hell?)
Anyways, with that many fans my PC was going to sound like a Harrier Jump-Jet taking off in my room. So I decided to do some research. This thread is the fruits of my labor. I hope you can bring something away from this post.

Let?s start with a little technical info:

Fans are commonly listed with many numbers and letters giving you detailed specs on it.
Such as this: 120x25mm, 3pin, 12V DC, 1600RPM, 63.57CFM, 27dBA, 2 ball bearing

The 120x25mm is just the size of the fan. The 25mm is the thickness of the fan. This is of important consideration when working in small areas or in places where you don?t want the fan so big it takes away from the looks of the overall effect. This is also very important when replacing the stock PSU fan because some PSUs do not have room in it to mount a thicker fan. I ran into that problem with my old OCZ Modstream PSU when I switched out its stock fan for a quieter one that moved more air. Some fans are as big as 46mm thick. The two most common thicknesses are 25mm and 38mm, with 25mm being the most commonly found by far.
The 3pin is simply how it connects to power. If you are looking to plug into your MB fan headers, make sure you get the right fan or buy the 4pin-to-3pin adapter (if it doesn?t come with it).
I?m sure that the ?12V DC, 1600RPM, 63.57CFM, 27dBA? stuff you are all familiar with, so I will not go over them except to say that CFM stands for Cubic Feet per Minute. And dBA stands for A-Weighted Decibels.

So in a nutshell, CFM is how much air the fan moves per minute and dBA is how loud the fan is at that given CFM. These are the two most important stats when comparing fans.

Lastly, there will normally be a stat that says one of the following:
Sleeve bearing, Ball bearing or S-DFB (Sony-Dynamic Fluid Bearing)
Instead of getting into the mechanics if each one, here is just a quick breakdown of the 3 types and the estimated lifespan of each type. (these are not my estimates. These estimates are the industry wide MTBF (Mean Time Between Failures) and all companies use them)
*Sleeve bearing ? this is by far the most common type. Easy to make, cheap to produce. These fans are ?normally? the loudest and have a MTBF (or estimated lifespan) of 30,000 hours.
*Ball bearing ? becoming fairly common to find. Usually quieter and lasts 6x longer, but generally costs more than a sleeve bearing fan with equal specs. Has a MTBF of 180,000 hours
*S-DFB (or just DFB) ? this is the newest technology in fans. Harder to find, costs the most. Doesn?t last quite as long as Ball bearing types, but capable of being the quietest of the 3 types. Has a MTBF of 150,000 hours.

<Continued below>

 

[cloudkat]

Well, now that that is over, let?s get to the ?meat? of this thread.

When looking around and comparing fans, there is a very simple formula to come up with its CFM/dBA ratio. This will tell you which fan has the best ?air flow-to-noise? ratio. Simply divide the dBA into the CFM. Thus a fan rated at 63.57CFM, 27dBA would have a ratio of 2.35 (or this fan moves 2.35CFM for every dBA). Obviously you want to have the highest possible ratio while still maintaining a comfortable noise level.

Now, this is all fine and good, for just ONE fan. But how many PC cases have you ever seen that had only one fan?? So how do you come up with the total sound (or loudness) of a full PC??
This gets a little tricky if you want to do it exactly. But don?t worry. This doesn?t need to be an exact science. We can do this in just generalized ways that will give us a good idea of our TOTAL sound.
*NOTE* For those of you who are more ?scientifically minded?, you may want to check out THIS link for a much more detailed, and exact way to figure out total sound levels.

For us that are not rocket scientists, the simple formula for adding multiple fans that have the SAME dBA RATING is simply shown by the following chart:
*NOTE* This chart is for adding multiple sound sources with the EXACT SAME dBA rating ONLY !!
This chart shows the effect of something called ?harmonization?.

Number of fans - Added dBA
2 - 3
3 - 4.8
4 - 6
5 - 7
6 - 7.8
7 - 8.4
8 - 9
9 - 9.5
10 - 10

As an example; we have multiple fans that are ALL rated at 30dBA. If we put two of those in a case, the total noise level would be 33dBA. If we put 7 of them in a case, the total noise level would be 38.4 (30+8.4). Get it? Good.

Now some of you may be asking, ?why did I even bother to add this part??? Well, having a case that all the fans are the same is very possible. Here is just one example of how:
Say you are running a water-cooled system. You may have 2 or 3 120mm fans on your res, and you have a 120mm fan in your PSU, and maybe a couple of 120mm case fans. The other parts are all cooled by your WC setup.

<Continued below>

 

[cloudkat]

OK, only in specific set-ups would you have ALL your case fans rated the same. To figure out your total sound of many fans that have different ratings all you need to do is use the chart below and start adding them up.
To use the chart, first find your highest rated (or loudest) fan. (If you do not know what your fan is rated at, you can usually find out by just a little searching on the net.) Next write down the dBA ratings of ALL the fans in your case. (Don?t forget to add in the PSU and GPU fans, as well as any NB or SB chipset fans you may have)
Now that you have all your ratings on paper, all you need to do is subtract each fan?s rating from your highest rated fan to come up with a difference. Then use the chart to find how much sound you add to your total based on that difference.
Do that for each fan, writing down the difference and added sound on your piece of paper until you are done.
Now that you have all your numbers, all you have to do is add all of them up, then add it to your highest rating.

Difference between dBA - amount to add to highest dBA
0- .9 - 3
1-1.9 - 2.5
2-3.9 - 2
4-4.9 - 1.5
5-7.9 - 1
8-10 - 0.5
Anything over 10 is 0 because the loudest fan is too overpoweringly loud compared to the other

As an example; let?s say we have a Zalman 9700 (CNPS9700 LED), it has a rating of 35dBA. We also have OCZ GameXStream 700W PSU with a rating of 34.3dBA, an Antec Nine Hundred PC case and an 8800GTS with a Thermaltake TMG ND5 aftermarket cooler that is rated at 16dBA.
The Antec case comes stock with three 120x25 fans rated at 79CFM @ 36dBA (high) and a single 200x30mm fan rated at 134CFM @ 32dBA (high).
The loudest part are the three 120mm case fans @ 36dBA each (damn that is loud, but it IS stock).
*NOTE* since there are 3 of them all rated the same, remember that you will be using the 1st chart to add up to a single number for all fans rated the same. This holds true even if you have multiple sets of fans that are rated the same. (ie: 4fans @ 36dBA, 2 fans @ 32dBA, and 2 fans @ 24dBA)
Anyways, 3 fans @ 36dBA gives us a total of 40.8dBA (now you can see why 3 fans only rated at 36dBA each can be so damn loud in your case)

The 200mm case fan @ 32dBA gives us a difference of 8.8dBA giving a +0.5 modifier.
The PSU @ 34.3dBA leaves us with a 6.5dBA difference which, by looking at the chart above, gives us a +1dBA modifier.
The CPU HSF @ 35dBA gives us a modifier of +1dBA
The GPU is so silent that the CPU HSF is greater than a 10dBA difference, so there is no modifier for that.
So our case total would be 43.3dBA. (40.8dBA for our ?loudest part? + 0.5 for the 200mm fan + 1 for PSU + 1 for the CPU HSF)
This is actually WAYYY too loud in my opinion. So what can we do about this? Let?s find out.

First let?s handle those loud a$$ case fans. Lets replace those 79CFM/36dBA (2.19 ratio) fans with some Scythe ?Minebea NMB Silent IC Series? (4710KL-04W-B29-V52) fans rated at 76CFM/31dBA (2.45 ratio) (nice move there). We will also replace the 200mm fan with a new one. This becomes a problem because of the 200mm size. You can either replace it with a smaller 120mm fan or a bigger 220mm fan by just drilling a few holes. For the purpose of this example, we will replace it with a monster 220x30mm fan by Yate Loon. The D22SL-12H is rated at an incredible 135CFM/25.5dBA (an awe-inspiring 5.29 ratio)

Now let?s replace the PSU fan with a Scythe ?Slip Stream? SY1225SL12M fan rated at 68.54CFM/24dBA (2.86 ratio)

So our loudest part is still the case fans (remember ?harmonization?) @ 35.8dBA. Subtracting the other parts one at a time, we come up with:
PSU @ 24dBA gives a modifier of +0
CPU HSF @ 35dBA gives us a modifier of +3
220mm case fan @ 25.5dBA gives us a +0 modifier
The GPU?s aftermarket TMG ND5 is still more than 10dBA so no modifier added for it.

So our new total is 38.8dBA. (35.8dBA for the loudest part + 3 for the combined modifiers)
Holy crap?we just shaved off almost 5dBA simply by switching some fans out. And 5dBA is a very noticeable difference in sound volume.
We could cut it down even more by replacing the CPU HSF with a different HS that we could buy our own fan for. By doing that and using another Scythe ?Minebea NMB Silent IC Series? (4710KL-04W-B29-V52) fan rated at 76CFM/31dBA (2.45 ratio) as the fan for the new HS we would end up with a system that has a total sound level of only 37dBA (I?ll let you do the figuring on that one as a test ). That gives us a total of 6.3dBA difference between before and after. Not bad at all.

*NOTE* I am very aware that there are many different variables in determining a ?total case? noise level. Things such as HDD noise, the differences in exhaust and intake sides of fans, the total area of the case, acoustical resonances of different case materials, ect., ect..
This is not meant to be an all inclusive guide. This is just simply a guide to try and help get the best possible air flow with the least possible noise.

<Continued below>

 

[cloudkat]

Now that you know how to pick good fans (using the CFM/dBA ratio), let?s look at what to do, and not do, with them.

Understanding case thermodynamics has become almost a necessity in the modern modding community. With PC components running hotter, more modders putting stuff in their PC cases (ie: res, rad, stuffed animals, and all manner of other stuff), and of course the noobie (who, after all, is who this is written for), understanding a few simple little ?rules? will mean the difference between overheating and the best possible case/component temps.

*NOTE* If you are an experienced PC person, you can probably just stop reading right here. The following is very basic stuff on how to achieve the best PC case cooling.

First, and probably the most important, thing to understand is air flow. By keeping as much stuff out of the way of your case?s air flow as possible is a great first step. Try to keep all your cables, wires, ect out of the way (hidden) as possible. Use wire ties, zip ties, ect to move and hold everything along the edges and/or behind other parts. If you are still using the old ?ribbon? type flat cables for your IDE components, you may want to change them out for newer ?rounded? IDE cables. Utilize your case?s ?dead space?. (ie: try running MB header wires and your case?s front USB/audio wires under the MB. Drill holes to run zip ties through to hold stuff along the sides, top and bottom of the case.) This not only helps promote good air flow by having as little air restriction as possible, but it just plain looks better.

The next thing to understand is that a case will be cooler when it has a negative air pressure.
Now I am not going to get into the physics behind this. Suffice it to say that you want more air being pulled OUT of the case than you have being pushed INTO the case. In other words, add up the total CFM of all your intake fans, then add up the total CFM of your exhaust fans (don?t forget to add in your PSU fans into your exhaust fan total). All you really need to have is roughly 10% more air minimum going OUT than you have going in. If you can hit right around the 10% to 30% mark minimum your doing good. If you want more info on the physics behind this, just do a little searching on the net.

It is common knowledge that heat rises, correct? We all know this. But if that is the case, why do so many people who mod their case?s neglect to put an exhaust fan in the top of their case? If you are going to mod your case, think about adding a top exhaust fan. You will be glad that you did. I know that for some case mods, due to the theme, or other reasons, this is not practical. That?s cool. But if you can, do it.
Also, speaking of ?heat rises?, proper air flow direction is very important as well. I have seen more than a few cases where the owner put the exhaust fan in the front of the case and had the intake fans in the rear. (No, I am not kidding, I have seen this many times with customer PCs)The idea behind doing this is simply that the CPU HSF is much closer to the rear fans, so blowing air right on it should help keep it cooler (Hey, you have got to give them credit for trying). Doing this takes the cool air and brings it in high in your case and tries to blow it out low in the case. The problem with that should be obvious once you just simply remember that ?heat RISES?. Not to mention that in doing it this way the PSU is immediately pulling a lot of this cool air right back out again.
Another thing to keep in mind when designing your case?s air flow is that we are trying for a smooth flow of air through the case. Not a turbulent one.
I say that to say this; I have seen many people add fans in their case that blow right on the GPU(s) or chipsets, or whatever. This disrupts the smooth flow of air from the bottom of your case to the top. You may get a couple of degrees lower temps on that single component, but the trade-off is that by disrupting the flow and creating a turbulence you now actually GAIN those few degrees back in your overall case temp.
To effectively do this, those cases, or mods, that have a side panel fan are great. But only if it is set up as an exhaust fan, NOT an intake fan, and ONLY if the fan is positioned high enough on the side panel to exhaust hot air from the CPU. Those cases with a low mounted side panel fan should be avoided if possible as using that as an intake will disrupt your air flow, and using it as an exhaust fan will be almost useless because all you really will be doing is exhausting some of the cool air coming in from the front of your case. If you have to use a case that has a low mounted side panel fan, just don?t put a fan there. Leaving it blank is a better option than using it as either an intake or exhaust fan.

Another overlooked thing is the direction of air flow of the CPU?s HSF. It is more efficient (cools better) to pull hot air AWAY from the CPU as opposed to blowing air AT it. If you buy an aftermarket HSF, be sure that you mount the fan blowing AWAY from the CPU. With some coolers this is not possible. But if you can, make sure of the direction of air flow.
Also, while on the subject of HSFs, some coolers will allow you to mount them where the fan can blow straight at the back of the case. This obviously would be the best solution considering that the case?s rear exhaust fan(s) can immediately take the hot air and expel it from the case. So when adding a new HSF, first see if it can mount in an ?East-West? orientation, as opposed to the traditional ?North-South? orientation that you may automatically want to use because it has become habit.

One last thing, your optical drive; you know that heat rises, and you also know that your optical drive produces heat. With those two things in mind, try not to mount your optical drive in the top most spot in your case. Instead use the second spot, or even lower. This will allow two things to happen. One, it opens up a ?dead air? space that the heat of your case can go to until it can be exhausted from the case. Secondly it helps your optical stay a little cooler running because it is no longer in the ?hot air zone?, while also allowing the heat that it produces to rise up and away from it.
Anyone who questions the amount of heat an optical drive can produce need do nothing more than play a music CD for a few songs or put in a movie DVD and watch it for about 15 minutes, then immediately eject the CD or DVD and feel how hot it is.

<Continued below>

 

[cloudkat]

I know that a lot of this may seem very small and/or petty. But how many of us have done crazy amounts of work just to get that extra 10-20 points on our benches, or drop and extra degree or two in heat. How many of you have voided the warranty on your high dollar CPU by lapping it? Voided the warranty on your high priced GPU by putting on an aftermarket cooler? Voided the warranty on your PSU by changing out a fan? Cut up you HDD to add a window to it just for looks?
See what I mean? So while some of this may seem small or petty to you, some of us will do anything and everything to gain just that little extra performance, a little lower temps, a little less noise.
So all I ask is that if you didn?t find this post helpful, please do not post negative stuff. Someone else may actually learn something (I hope).

One final thought for those of you who want to get the best case temp/noise ratio. If you will simply pull off BOTH side panels of your case, you can then disconnect ALL of your case fans. The only fans you need running in that case is the CPU, PSU and GPU (if it has one). This will give you ambient room temps inside your case, which is the absolute best temp you can get on air cooling alone.
Now this obviously isn?t something that everyone can, or will do. Some people have put A LOT of time, effort and money into custom modding and/or paint to NOT show it off. That piece of advice is for those who do not care that their PC is sitting there wide open.

Anyways, that?s it. I hope that at least someone can get something out of this. If just ONE person can bring something good away from this post, then it has all been worthwhile.

 

[oz0]

Excelent post, thank u very much 4 this. Found very interesting what u said bout the optical drive. But talking about noise and fans, dont u use rubber pads and rubber bolts 4 the fans? what do u think about them?

 

[cloudkat]

Originally posted by: oz0
Excelent post, thank u very much 4 this. Found very interesting what u said bout the optical drive. But talking about noise and fans, dont u use rubber pads and rubber bolts 4 the fans? what do u think about them?

I have not seen any definite testing on this to say for 100% certainty, but the concept would seem valid only in specific cases. Such as if you have a small, high RPM fan, it would vibrate, thus leading to the belief that rubber grommets and/or bolts would quieten it down. But on bigger, low RPM fans, they do not vibrate much (if any), thus the grommets and/or bolts would do no good, or at least offer no noticeable difference.

The idea behind using such things is to cut down on the fan's vibration against the case. If you use a good quality fan (lowest possible RPMs with highest possible CFM) you should have very, VERY little vibration, if any at all. Especially if you screw the fan to the case tightly, and not just half-a$$ed.

Again, to my knowledge there have been no real tests on this, so this is just all a guess using a little logic.

 

[cloudkat]

Let me clarify my last reply a little. Even the best quality, low RPM fan will vibrate. Even if it is such a small amount you would not ever be able to know it without professional testing equipment.
What I am trying to say in my last reply is that a good fan will vibrate so little, especially if screwed down tight, that adding rubber pads, grommets, or bolts would not give you any noticeable difference.

Think of those crappy little 80mm and 92mm Antec lighted fans that were so popular a few years ago. Those things looked cool, but CHRIST they were loud and vibrated like mad. Those are the kind of fans that would DEFINITELY gain a benefit from rubber pads and stuff.

But, I guess just to be on the safe side, it couldn't hurt to add them to any fan anyways, just for $hits-and-giggles, as the saying goes.

 

[esjrobles]

Great post men, I've been trying to look for this kind of guide. Now I know that I have to revert back my upper Fan into exhaust fan not an intake. LOL.

 

[DerwenArtos12]

Good guide for the basics. I have to say though that there is a LOT more to fans than just this. Blade design and RPM have very big factors in the tone of the sound the fans emit. dbA alone doesn't tell the whole story for how a fan will sound. I don't know that I would bet on your harmonics numbers either. No offense but, that is assuming all the fans have not only the same DBA ratio but, that they're the exact same fans. Also, regaurdless of how quiet or loud a fan is, having more than one fan will be louder, regaurdless of how loud the first fan is and how quiet each subsequent fan is. Also, your scale for dbA modifiers, what is that actually supposed to acomplish, I don't think that anything obove 10dbA is too overpoweringly loud that you won't hear other fans. Average room ambient noise is generally above 20dbA. I realize that those are supposed to be differentials but, I don't think it's accurate. What refrence materials did you use to come up with that?

 

[zorrt]

Louder fan doesn't overpower the quieter one...

Flat IDE cable actually can be better than the rounded ones. Sure rounded ones allows air to flow around them easier, but flat ones allow you to bend and fold the cable to position it out of the way and often with todays motherboard layouts it is more beneficial I believe.

Not too sure what you mean by fan blowing away from the cpu heatsink opposed to blowing at it. Most heatsinks are designed to have air blowing through the fins. By having a fan blowing away from the fins wouldn't it just mean you've restricted airflow from the fan? I still think its a better idea to have air flowin through the fins, just like they were designed too.

Good effort on the guide though even though I don't agree with most of the things on it at least youve taken the time to write one up.

 

[DerwenArtos12]

Originally posted by: zorrt
Louder fan doesn't overpower the quieter one...

Flat IDE cable actually can be better than the rounded ones. Sure rounded ones allows air to flow around them easier, but flat ones allow you to bend and fold the cable to position it out of the way and often with todays motherboard layouts it is more beneficial I believe.

Not too sure what you mean by fan blowing away from the cpu heatsink opposed to blowing at it. Most heatsinks are designed to have air blowing through the fins. By having a fan blowing away from the fins wouldn't it just mean you've restricted airflow from the fan? I still think its a better idea to have air flowin through the fins, just like they were designed too.

Good effort on the guide though even though I don't agree with most of the things on it at least youve taken the time to write one up.

QFT. I realize now I may have sounded a little harsh in my original post. I make custom rounded IDE cables out of flat because with the big plastic covers over the rounded they're not nearly flexible enough to route nicely. As for the blowing away from the heatsink; most if not all fans are designed as pushers, not pullers, the overall efficiency of the fan will drop pretty notably if used as a puller instead of a pusher. To see what I mean take a fan with rpm monitoring and hook it up, preferably a low speed one as the difference is easier to see with them, and put a very restrictive filter on the side the fan is drawing from and note the RPM drop, then put it on the side the fan is pushing towards and note the RPM drop there, it will in almost every occasion be higher on the pulling side than on the pushing side.

 

[bunnyfubbles]

I'll keep my optical drive in the top most slot. Have been doing it since forever and have yet to have a drive fail before I've replaced it solely as a performance and/or technology upgrade.

I can sympathize with some of what you're trying to share, but come on...'if it ain't broke, don't fix it' - otherwise you're essentially fear mongering.

 

[Navid]

Originally posted by: cloudkat
When looking around and comparing fans, there is a very simple formula to come up with its CFM/dBA ratio. This will tell you which fan has the best ?air flow-to-noise? ratio. Simply divide the dBA into the CFM. Thus a fan rated at 63.57CFM, 27dBA would have a ratio of 2.35 (or this fan moves 2.35CFM for every dBA). Obviously you want to have the highest possible ratio while still maintaining a comfortable noise level.

Let's say we have a 60CFM 30dBa fan and a 66CFM 33dBa fan. They both have a ratio of 2 and according to you, they are as good.
The 66CFM fan moves only 10% more air but it is twice as loud.
If you use a fan controller and drop the 66CFM to 60CFM, it's noise will never go down to 30dBa (3dB reduction is like halving the noise). So, you are much better off with the 60CFM 30dBa fan. The two fans are not as good.

That ratio is meaningless!
The reason is that dBa is on a log scale while CFM is on a linear scale.

To compare apples and apples, you want to know what the noise of each fan is at the exact same CFM. Or, you want to know what the CFM of each fan is at the exact same dBa. The useless ratio does not tell you either.

 

[cloudkat]

Originally posted by: DerwenArtos12
Good guide for the basics.

Thank you. And remember that this is ALL it was ment to be.....the basics.

I have to say though that there is a LOT more to fans than just this. Blade design and RPM have very big factors in the tone of the sound the fans emit. dbA alone doesn't tell the whole story for how a fan will sound.

Agreed. There are many other factors involved here. But again, this was not meant for people like you who already have a strong understanding of fans and acoustics. This is just a (hopefully) good starting point for those either new to the game, or have never really thought much about this part of a PC.

I don't know that I would bet on your harmonics numbers either. No offense but, that is assuming all the fans have not only the same DBA ratio but, that they're the exact same fans. Also, regaurdless of how quiet or loud a fan is, having more than one fan will be louder, regaurdless of how loud the first fan is and how quiet each subsequent fan is. Also, your scale for dbA modifiers, what is that actually supposed to acomplish, I don't think that anything obove 10dbA is too overpoweringly loud that you won't hear other fans. Average room ambient noise is generally above 20dbA. I realize that those are supposed to be differentials but, I don't think it's accurate.

Now here is where I will have to disagree with you. The harmonics numbers are accurate, and do not have to be the exact same fan, only the exact same dBA level. This is because all of those "variable" that we know exist are already in play when the fan is tested to come up with a total dBA rating that the manufacturer can print or advertise. The only other variables that will effect the harmonics scale are those that are specific to the individual. Such as how close you sit to the PC, how good your case dampens acoustical sound, the positioning of the components in your case, ect., ect.. But when just looking at the pure effect of adding multiple dBAs, this harmonics scale is very accurate. Here are just two sites to look at:
http://www.engineeringtoolbox....ding-decibel-d_63.html
http://www.acousticalsurfaces....coustic_IOI/101_12.htm
If you do some searching you will find many more sites that all say this same thing. I had 5-6 sites that I found before I quit looking because they all said the same thing.

What refrence materials did you use to come up with that?

Well, here are just some of them:
http://www.comairrotron.com/engineering_notes.asp
http://www.cpemma.co.uk/
http://www.silentpcreview.com/article690-page1.html
http://www.silentpcreview.com/article687-page1.html
http://www.engineeringtoolbox....-power-level-d_58.html
http://www.engineeringtoolbox....nd-pressure-d_711.html
http://www.ebmpapst.us/techarticles.asp

There is more, but this should give you the info you need.
As you start going through these and reading them, hopefully you will see why I chose to write this article. There is allot of info in just those few sites. So trying to compile it all into a single article, AND try to keep it at a decent size instead of asking you guys to read a novel, was one hell of a task.
I may not have done the best job at it, and i'm sorry for that. But I tried. All I was doing was trying to help those who have not done the massive amount of research on their own.

 

[aussiestilgar]

Good beginner's guide for people with no clue where to start. Still, I think SilentPCReview.com is the best resource for fans (also XS forums) and worth a visit for anyone new or experienced. However I felt that if you have so many fans, it may be more cost effective to just get a fan controller and slow them down a little rather than replace them all. I also wanted to say that despite the basic ratios and specs of the fans (CFM, dbA etc.) the majority of specs listed by manufacturers cannot be trusted (silenx?). It is far better to read a review site when deciding on which fans to get, than to choose a fan based on the specs on the box.

 

[cloudkat]

Originally posted by: zorrt
Louder fan doesn't overpower the quieter one...

If there is more than a 10dBA difference, yes it does. Please see the links that I referenced in the above post and read for yourself. Just keep im mind that there are many different variables. But let's just assume that we have a 12'x12'room that is totally soundproof. In the middle of that room you place a PC with just one fan in it that has a rating of 45dBA at a distance of 3'. You know this because you used a meter and measured it yourself. Now we go in and add another fan to that same case that is rated at only 33dBA. If you go the same 3' distance away and pull another measurement, it will still only be 45dBA.
Again, I am not trying to argue with you, just go read some of the links I posted above and see for yourself.

Flat IDE cable actually can be better than the rounded ones. Sure rounded ones allows air to flow around them easier, but flat ones allow you to bend and fold the cable to position it out of the way and often with todays motherboard layouts it is more beneficial I believe.

I agree with you here, but only to a point. Just as you can bend and fold a flat cable, you can also bend and tie/tape/glue/whatever the rounded one out of the way.

In another post you state that you make your own IDE cables. So do I, for the exact same reasons that you do. We can get the benefit of both flat and rounded cables. But once again my article is being mistaken for something that was written for an experienced user. This guide was written for people who have not advanced as far as you have in PC knowledge.

I can't speak for you, or anyone else for that matter, but speaking for me, I can remember when I first got started with BIY PC building. I just always used the flat IDE cables that came with the MB. It was not until later that I found out about rounded cables and their benefits, so I started using them. Then even later, I found a tutorial written by someone on a forum somewhere on building your own rounded IDE cables and the benefits of doing so.
In other words, I was not born with that knowledge. I had to do allot of reading and searching to end up where I am at today.

With all the forums and websites available to people nowdays, there is more info than ever available to the those new to BIY PCing. And having people that have already "been there, done that" write up guides is a tremendous help.
Just look at these forums. There are many "How-To" and "Noobie's" guides that are stickied. How many times have you seen a noob come here on these forums and ask stuff like, "How do I OC my CPU?", or "What is the best way to test my OC?". It is because of these type questions that people with the knowledge write guides.
I would have KILLED to have the amount of info available to me when I first started in PC building as people have available to them nowdays. (Yes, I know, I'm showing my age there) But, depending on how long YOU have been doing this, don't you wish you would have had this kind of info and support available to you when you started?
Just ask some of the "old-timers", and I bet they will agree with me on this.

Not too sure what you mean by fan blowing away from the cpu heatsink opposed to blowing at it. Most heatsinks are designed to have air blowing through the fins. By having a fan blowing away from the fins wouldn't it just mean you've restricted airflow from the fan? I still think its a better idea to have air flowin through the fins, just like they were designed too.

It is a proven Law of physics that it is more efficient and cools better to pull hot air AWAY from a heat source than to blow cool air AT it.
As an example, Take a top fan mounted HS and measure not only the core temp, but measure the temp of the MB area surrounding the CPU socket. Now take that same exact fan and turn it over to where the fan is blowing AWAY from the CPU. Look at all your temps again. You SHOULD see a difference in your core temp, and you will DEFINITELY see a difference in you MB temps because you no longer have the fan blowing hot air from the HS at it. Now again, keep in mind, this is talking in a "generalized" sense. Some HS may NOT perform better this way for one reason or another. But just looking at the laws of physics, give it a shot. Allot of people have had great results just doing this simple thing.

Good effort on the guide though even though I don't agree with most of the things on it at least youve taken the time to write one up.

Thank you once again for the compliment. And don't worry about disagreeing with me.
It is through health, mature debates like this (not flame wars) that we learn. If you can prove me wrong, I have learned something. If I can prove you wrong, you have learned something. And isn't learning something one of biggest (if not THE biggest) reason that we come to forums like these?

 

[Krazy4Real]

Definitely a good guide for someone to start out with. Thanks for posting.

 

[cloudkat]

Originally posted by: bunnyfubbles
I'll keep my optical drive in the top most slot. Have been doing it since forever and have yet to have a drive fail before I've replaced it solely as a performance and/or technology upgrade.

I can sympathize with some of what you're trying to share, but come on...'if it ain't broke, don't fix it' - otherwise you're essentially fear mongering.

That is the cool thing about it being YOUR PC. You can do whatever you wish with it.
I never said that by having your optical at the top it would cause it to fail. All I was pointing out was that it gets hotter in that position and it blocks up the "hot air zone". So by moving it you can not only open up a larger, higher-in-your-PC-case "hot air zone" to be exhausted, but you can also help lower the operating temp of your optical. Producing less heat is a good thing, right?

I disagree with your statement, "if it ain't broke, don't fix it' - otherwise you're essentially fear mongering." And I am sure that most of the other modders and OCers out there will as well.
If your statement is true, then why do people OC? Lap their HSs? Spend all that money on aftermarket HSF? And a million other things like those?
It is simple, to get every last friggin' OUNCE of performance we can squeeze out of our systems. Each and every little thing we can do to gain better temps, or better performance we are willing to try. So, it is with that in mind that I do not put my opticals up top. I use that as the "hot zone" and have exhaust fans there.
You, as I said, may do as you wish.

 

[cloudkat]

Originally posted by: Navid
Let's say we have a 60CFM 30dBa fan and a 66CFM 33dBa fan. They both have a ratio of 2 and according to you, they are as good.
The 66CFM fan moves only 10% more air but it is twice as loud.
If you use a fan controller and drop the 66CFM to 60CFM, it's noise will never go down to 30dBa (3dB reduction is like halving the noise). So, you are much better off with the 60CFM 30dBa fan. The two fans are not as good.

That ratio is meaningless!
The reason is that dBa is on a log scale while CFM is on a linear scale.

To compare apples and apples, you want to know what the noise of each fan is at the exact same CFM. Or, you want to know what the CFM of each fan is at the exact same dBa. The useless ratio does not tell you either.

You obviously missed something in the article. If you have the two fans you talk of, and they both have the same ratio but one has a lower dBA, then the lower dBA one would be the better of the two as long as its CFM was sufficient to do the job you need it to do.
Here is what I said in the article, "you want to have the highest possible ratio while still maintaining a comfortable noise level". (ie: lower dBA is better)
So how you came up with this: "They both have a ratio of 2 and according to you, they are as good", is beyond me. I never said that equal ratios made them equal fans. I specifically stated, "This will tell you which fan has the best ?air flow-to-noise? ratio". That is all. Nothing more. No hidden meaning. All the ratio is designed to do is to allow you to compare different fans on an equal basis.
Kind of like if you can buy 3 candy bars for $1, or a case of 20 for $6, which would be the best deal. You break them down to equal ground. In this case, how much are you paying for each individual candy bar in each option.
So in the example fans you used, if the prices were equal, and they both would do what you wanted them to do, which one would you buy? The one with the lowest dBA or course.
See, that is all the ratio is for. It is up to you at that point to deside if you want the extra 6CFM at increased noise, or do you want the quieter fan.

Now all this is only generally speaking. If you have a HS that seems to work best when you are using over 60CFM, then obviously the louder 66CFM fan will be your best choice. But if you have a HS that cools its max ability with only 50CFM, then anything OVER that 50CFM is waisted, so finding the quietest fan at (or around, it doesn't have to be exact) that 50CFM mark is the best fan for you. And by using that ratio, you can determine which fan would be the quietest while still achieving your 50CFM goal.
THAT is what the ratio is for.

It amazes me why people put these crazy high CFM (and thus LOUD) fans on their HSs. A HS can only remove so much heat, so fast. If you can achieve that max performance of you HS with a quiet 70CFM, why in the name of the Gods would you put a monster 100+CFM fan on it. All you are doing is creating more noise. There is no gain to be had in cooling. The HS can only do so much.

 

[Sheninat0r]

Nice guide, but pretty basic... I felt it was missing a couple of things.

First off, since you did your research at SPCR I would hope to see a warning about manufacturer fan specs in this... Most fan manufacturers way inflate CFM and dBA levels, and their specs are never a proper way to judge a fan. For beginners, that would be a very important thing to know, lest they run out and buy cases of SilenX crap based on the 74CFM/14dBA specs. An outright lie, in case anyone doesn't know already.

A quick discussion of motor/blade geometry and how it relates to CFM and noise would have been nice to see too, and a section about backpressure - it's up there with CFM and dBA as important fan specs. Noctua's NF-S12 and NF-R8 have amazing CFM/dBA ratios, but literally ANY restriction, on either side of the fan, will destroy that ratio because of its blade design. Speaking of important fan specs, RPM is probably the biggest factor in determining airflow and noise.

Another thing to include, or to link to [I know XS has a two nice fan roundups by Viper], would be a list of recommended fans in each CFM range to help people [beginners, who you seem to be aiming this article at] pick. Definitely not necessary, but nice to have either way.

Also, negative pressure is not always the best case setup - it may or may not get you that extra degree, but it will lead to dust getting sucked in through every crack in your case. Properly filtered positive pressure systems will be much cleaner and will probably perform about the same. A discussion of the pros and cons of each would be better than saying "negative pressure is better", because sometimes it's not. That being said, I have a negative pressure system with one Tri-Cool on low exhausting and no intakes, and I don't have a dust problem [well, there's a little but nothing to be worried about] - mostly because I keep my house clean.

I skimmed the article quickly, so pardon me if you did cover some of this without me seeing. All in all, it's a good guide that covers the bare minimum to know before buying fans, but it isn't as complete as it could be.

 

[Navid]

Originally posted by: cloudkat
Simply divide the dBA into the CFM. Thus a fan rated at 63.57CFM, 27dBA would have a ratio of 2.35 (or this fan moves 2.35CFM for every dBA).

My point is that when you say "this fan moves 2.35CFM for every dBa", one can expect that for 4dBa, it would move 4 times the amount of air. This is basically what a ratio means.

As an example, speed is the ratio of distance to time. If you say your speed is 10 Miles per hour, that means in 2 hours, you will move 20 Miles. Or, in 3 hours, you will move 30 Miles. The ratio always holds as long as the speed is constant.
Note that both Mile and Hour are linear units not logarithmic.

But, if your fan moves 2.35CFM for every dBa, it will not move 4 times the air for 4dBa. Because 4dBa is only twice as loud as 1dBa, not 4 times as loud.

The ratio will have a meaning if you use a linear unit for both the numerator and the denominator.

CFM/dBa is a meaningless ratio.

 

[toslat]

Originally posted by: Navid

Originally posted by: cloudkat
Simply divide the dBA into the CFM. Thus a fan rated at 63.57CFM, 27dBA would have a ratio of 2.35 (or this fan moves 2.35CFM for every dBA).

My point is that when you say "this fan moves 2.35CFM for every dBa", one can expect that for 4dBa, it would move 4 times the amount of air. This is basically what a ratio means.

As an example, speed is the ratio of distance to time. If you say your speed is 10 Miles per hour, that means in 2 hours, you will move 20 Miles. Or, in 3 hours, you will move 30 Miles. The ratio always holds as long as the speed is constant.
Note that both Mile and Hour are linear units not logarithmic.

But, if your fan moves 2.35CFM for every dBa, it will not move 4 times the air for 4dBa. Because 4dBa is only twice as loud as 1dBa, not 4 times as loud.

The ratio will have a meaning if you use a linear unit for both the numerator and the denominator.

CFM/dBa is a meaningless ratio.

There is a difference between 'perceived' loudness (frequency dependent) and intensity.

The response of the human ear to sound is pseudo logarithmic. This is the reason for dB in measuring loudness, while the A in dBA refers to a frequency dependent weighting standard.

Hence 73dBA is not twice as loud as 70dBA (80 dBA is) but it has twice the intensity. Doubling the 'perceived' loudness usually requires an increase of 10dB

 

[cloudkat]

Originally posted by: Navid

My point is that when you say "this fan moves 2.35CFM for every dBa", one can expect that for 4dBa, it would move 4 times the amount of air. This is basically what a ratio means.

As an example, speed is the ratio of distance to time. If you say your speed is 10 Miles per hour, that means in 2 hours, you will move 20 Miles. Or, in 3 hours, you will move 30 Miles. The ratio always holds as long as the speed is constant.
Note that both Mile and Hour are linear units not logarithmic.

But, if your fan moves 2.35CFM for every dBa, it will not move 4 times the air for 4dBa. Because 4dBa is only twice as loud as 1dBa, not 4 times as loud.

The ratio will have a meaning if you use a linear unit for both the numerator and the denominator.

CFM/dBa is a meaningless ratio.

Why would you break it down to dBA per CFM, then try adding it up again??
The only reason I can see for this is if you are under volting your fan by use of either a fan controller or some other method. But if that is the case, as you know, or should know, that brings in a whole different set of rules and problems. This ratio is not meant to do anything more than give you basis for comparing the noise level to CFM of different fans when you are looking to buy one. I do not know how many ways I can say this.

While, as I said, I love a healthy debate, I will have to concede this one to you. If you have not grasped the meaning, or use, of the CFM/dBA ratio by what I have already explained, then I can not help you any further.


What I do suggest though, is if you are so passionate about this "CFM/dBa is a meaningless ratio" thing, then please go email the many manufacturers of fans that use this simple ratio as part of their website, part statistics and marketing. Maybe one of their design engineers can explain it to you better than I can. If it is such a useless ratio, I doubt very seriously that so many review sites, distributors and manufacturers would use it.

But, as I said, I can not find any more ways to explain the same thing. I honestly think that you are "over thinking" this whole thing. You are obviously a very intelligent person. But did you ever stop to think that maybe, just MAYBE, this ratio is as simple as I am saying it is.

*EDIT* = I guess I should have mentioned some places to start if I am going to ask you to email manufacturers and such. While I was doing my research, I bookmarked allot of pages, but I don't remember if I kept any of the pages that didn't contain any info that I used in my article. The only one I can remember off the top of my head is Scythe. I think I remember that one only because it is probably my favorite fan manufacturer. Anyways, here is a link to their product info page where you can clearly see that they list a CFM/dBA ratio on each fan. That would seem to me that it wouldn't be such a useless stat. But anyways, keep us informed on how your emails go.
http://www.scythe-usa.com/supp...temp/002/fan_spec.html

 

[zorrt]

Originally posted by: cloudkat
In another post you state that you make your own IDE cables.

Umm no I didn't. That was Derwen

Btw, I too joined the round IDE bandwagon. Waste of money. Not sure if they're all the same but the one I got is impossible to tidy it up properly. I had to tuck it into a spare slot under my dvdrom.

What you can do with the flat one with modern motherboard is fold the cable near the connection on the motherboard, run it FLAT along the back of the case, then fold it again to bring it towards the drive. Much better than trying to tuck round cables here and there and would work better too in terms of airflow as theres NOTHING to restrict air. But of course this only applies if you have a modern PC, lets say all SATA hdd and an IDE cd/dvdrom. I know I for one wont bother upgrading to a SATA dvdrom until my IDE dies.