Iateronmly

Well it's 2007, it's -30 outside, and I have cabin fever somethin' fierce. So what is a person like me to do? Well I bought 10 square feet of transparent green cell cast acrylic, 3'x2' of ultra thin cosmetic grade diamond plate, and 12' of 3/4" 90 degree channel. I was working on a mATX tower I bought, became disgusted with it, and decided it wasn't worthy of 250mm fan goodness so....

I'm going to build a new custom cube with TWO 250mm case fans and keep it under one cubic foot of area. Each 250mm fan is rated at 105cfm. Fun with numbers - that means it's total area would be exchanged once every 1/3rd of a second/210 times per minute.... BS, but well you know still funny thought. Now you're probably wondering why I'm writing this in the Daily Disturbance and not just starting a worklog. Well the reason for this line of thought and the title is this: understanding CFM ratings, and why you'll never get that type of airflow through a case. I'll be adding cliff notes at the end.

The first thing is - those pesky ratings on fans. First off CFM ratings - I have no solid proof, but I don't think anyone tests them, or if they do that memo get's lost on the way to marketing so they wing it(please don't get me started on my views on sound ratings). But let's pretend in the fantasy land of hypotheticals - all those fans actually were tested, and they were tested under the same conditions. They'd be tested at equal pressure/open air with no restrictions in front of the intake or exhaust side. In short, they wouldn't do us any more benefit - let me explain why.

No two situations are the same, and that starts throwing variables into the situation. No case will give you a true free flowing situation, and pressure or vacuum will develop, often small amounts, that even if you saw in a chart you'd scoff - but they make big changes. Since different fans have different pitches to the blades, they may infact have as little as three blades and more than 11. A fan's flow change when presented with these inevitable things - and will perform differently even if their free flowing cfm ratings were identical on the test bench.

Returning back to fantasy land of hypotheticals - a fan with 11 blades will move more air at lower rpm than a fan with less blades at the same speed. Reality, there is of course a limit to all good things - and more blades/less blades lends itself to certain qualities, but can also hamper. I have a silverstone 120mm fan with it's 11 blades, but rpm for rpm it actually pushes less air in my case than an Enermax with less blades I have at the same rpm.

Axial fans produce a lot of turbulence compared to cross flows while exiting. This turbulence means that for the airflow to stabilize and flow uniformly there must be distance between the tip of the fan blade and the item offering resistance. Proximity of another surface dictates a lot in regards to amount of pressure that must be developed in that distance to over come. Reason why many watercooling folks use shrouds - it allows more air to flow through with the same fan, so better performance to noise is achieved instead of throwing Delta screamers to achieve the same ends.

Probably should talk about the three different types of fan designs: axial, radial, and cross flow. Radial you don't see because it isn't suited to a PC. But now cross-flow you might've seen... but not commonly. If you think about it, and read the marketing about some of the cross flow fans on the market, you'd almost think cross flows are the way to go. Sadly even though cross-flow fans offer better air flow to noise level in theory - they're horrible at dealing with restriction as their design doesn't allow it to build pressure - air will just recirculate in the cage and that is true in both theory and reality. I wouldn't mind seeing cross-flow fans mature, and better bearing designs used so they could be quieter so don't think I'm hating on them. But their cfm to noise ratio isn't the sole deal breaker for me - due to a cross flow fans problems with restriction, being able to use a filter is a major problem and if they get dirty they're even more of a pain to clean.

With all this said - remember this very important aspect. Turbulence may mean reduced CFM traveling through your case, but it's vital as air must travel over your heat generating parts to cool them! The trick is to reduce unneeded restriction so all the potential flow of air will be used where it helps. Also don't overlook fan placement to help reduce stagnant areas of air.

__________________________________________________ _________________________________

Cliff Note Section:

Fan placement is arguably just as if not more important than the amount of air moved. If the path air travels doesn't cause turbulence over a heat generating area - then it's the same as not having any air flow at all.

Fan CFM ratings are all across the board in terms of accuracy to published numbers - use them as a rule of thumb as they're mostly feel good numbers. A 100cfm rated fan will always be better than a 40cfm rated model, but when two are within 10cfm you might find results to be different than what you'd expect. Also when placed in chassis - performance characteristics may change marginally or greatly.

To keep airflow optimal, as we all know, use less restrictive fan grills.

Shrouds are worth investing in if you use a radiator and have the clearance space for them if you're trying to optimize your cooling to noise ratio.

If you want to use a side fan to blow directly on specific areas such as the CPU/GPU/n.bridge areas keep in mind these (minimal) distances that afford decent airflow in regards to distance from a flat surface:**

If you want to place a....
60mm fan - it should be a 1/2" (12mm) from flat surface.
80mm fan - it should be 3/5" (15mm) from flat surface.
92mm fan - it should be 1/2 (17mm) from flat surface.
120mm fan - it should be 1" (25mm) from flat surface.
250mm fan - it should be 2 1/4" (58mm) from flat surface.

__________________________________________________ _________________________________


Iateronmly - Keep It Chilly

** This information was posted by Seyeklopz but didn't offer supporting information, but I feel it's reasonable figures so I felt obliged to share them as they're as accurate as any numbers I'd have guessed. The figure for 250mm is my personal guess based upon what was offered.

__________________

Master Yunk

Different kind of DD post than I expected.

Still appreciated though

__________________

Iateronmly

Quote: Originally Posted by Master Yunk Different kind of DD post than I expected. Still appreciated though

Okay fine - next time I'll write a DD "10 things not to insert into your cases blowhole's hole"

__________________

$SOLID$ Necro

Eh...just use a 300MM 500CFM fan and buy some ear plugs like I did, that will "Get er' done"...

LOL, good write up...Some very intresting info!

Wait til ya see what I was given at work today
(But..dammit, I forgot to bring it home!)

__________________

Iateronmly

Quote: Originally Posted by $SOLID$ Necro Eh...just use a 300MM 500CFM fan and buy some ear plugs like I did, that will "Get er' done"...

I'm shooting for 25dbA max.

__________________

ex421

nice, I need to replace my fans, I upgraded my system but theres really only intake fans.

$SOLID$ Necro

Quote: Originally Posted by Iateronmly I'm shooting for 25dbA max.

Yeah..was just teasin ya, one of the main reasons I even delved into that project was to see how much flow i needed to cool a PC effectively and be realy quiet...obviuosly a large fan spinning at low rpm was the best way.

My rev 2 of it acomplished that goal, but I ran into some issues with the fan motor I was using not being wired properly (Needed a capacitor) so it generated alot of heat.

By the time I had it sorted out, somone had beet me too it and went into production, so the project was abandoned.

I have been toying with a new version of it over the holidays with Apha, we are looking at putting a new "Spin" on the fan design that should prove intresting (Both in terms of flow and DB levels) but this time I am keeping the details "Quiet" so they don't get out before I can find someone to put it into production.

__________________

lobey

Quote: Originally Posted by Iateronmly Okay fine - next time I'll write a DD "10 things not to insert into your cases blowhole's hole"

I know the answer to that...kittens.

Good write up though, I found it to be quite informative.

Iateronmly

Quote: Originally Posted by $SOLID$ Necro I have been toying with a new version of it over the holidays with Apha, we are looking at putting a new "Spin" on the fan design that should prove intresting (Both in terms of flow and DB levels) but this time I am keeping the details "Quiet" so they don't get out before I can find someone to put it into production.

Variable pitch with tip drive? Or more exotic?

__________________

Im_gumby

Quote: Originally Posted by lobey I know the answer to that...kittens.

So sorry, the correct answer is gerbils.


Gerbils.

__________________

Gzim

Quote: Originally Posted by Iateronmly Okay fine - next time I'll write a DD "10 things not to insert into your cases blowhole's hole"

Nah, I liked it! It was like reading a post by alpha, except I didn't have trouble maintaining interest. Although I kind of got confused a few times....

Would there be increased turbulence when using a shrould to put an oversized fan mounted on either a case or radiator?

__________________

Iateronmly

Quote: Originally Posted by Gzim Would there be increased turbulence when using a shrould to put an oversized fan mounted on either a case or radiator?

Turbulence is just a trait of any air movement really - it's just on the entering and exit of a fan air will move anything but straight in and out. I really need to add diagrams.... anyways. If you look at most fan adapters, most of them aren't just thicker for the sake mounting - but give it distance that helps the airflow "straighten out" and lets it pass through a radiator at a higher velocity/volume.

But I don't really apply a lot of stuff that could be to this discussion for these reasons:

1) Even though some may like to think they'd love to understand all there is to know - it's not practicle for any of us to undertake over PC air cooling.

2) My goal is just to help people get the most out of their fan placement - that's it. I just got tired of people complaining about temps when they have very nice looking, but restrictive grills and proposing fan placements on sketchs with MS paint. Not that there is anything wrong with MS Paint - but I've seen some funky ideas.

3) Do number2 without needing explain the Venturi effect, Bernoulli's principle, incompressable flow and related subjects. There's many fine universities that'll be able to instruct far better than an internet forum and instructors that know far better than I.

__________________