Fans, cfm, rpm and static pressure... anyone can undo the gordian knot?

[MrToad]

So... tirelessly I keep trying to fit the ubiquitous square peg in the round hole.

I think I'm starting to master the art of ice skating uphill :D

From what I've gathered so far, a fan's rpm rating is all good and dandy in open air, but is the static pressure what determines what's the volume of air that will actually go through a restrictive body (in our case, a radiator), correct?

Now (and this is where I want someone with the relevant knowledge to tell me that I'm completely wrong), after going through the data sheets of Sanyo Denki (which seems to be is the most reliable manufacturer when publishing data), a pattern seems to emerge.

It would look like there's a direct relationship between rotation speed and static pressure. In other words, (let's make it big so the numbers jump up) a 200mm fan which rotates @ 3600 rpm has an air flow of 11 m3/min with a static pressure of 360 Pa. The same fan, rotating @ 4800 rpm (33% faster) has an air flow of 14.7 m3/min (33% more) and... a static pressure of 640 Pa! (77% more).

This relationship seems to be constant throughout the board. As soon as you start dropping the rotation speed (searching for silence) the static pressure falls exponentially.

Therefore, if the above observation is correct, a large 220mm fan (like the YL D22BL-12H), which has a rotation speed of just 750 rpm, would not be able to pull any air whatsoever through a radiator, correct?

All this seemingly pointless observations are part of my (mad) idea of how, when and if a single 220mm fan could substitute 4 x 120mm fans.

*Edited to prevent further embarrassment...

[quiet]

What would 2500CFM give me?

http://www.homedepot.com/catalog/productImage/2e019933-0a90-4f74-b6ea-e4503cb4a96e_400.jpg (http://www.homedepot.com/webapp/wcs/stores/servlet/ProductDisplay?storeId=10051&langId=-1&catalogId=10053&productId=100012819&categoryID=502551)

:shrug:

[Grinch]

What would 2500CFM give me?

http://www.homedepot.com/catalog/productImage/2e019933-0a90-4f74-b6ea-e4503cb4a96e_400.jpg (http://www.homedepot.com/webapp/wcs/stores/servlet/ProductDisplay?storeId=10051&langId=-1&catalogId=10053&productId=100012819&categoryID=502551)

:shrug:

a BIG headache...:up:

[quiet]

You mean an Xtreme headache? :cool:

[MotF Bane]

What would 2500CFM give me?

http://www.homedepot.com/catalog/productImage/2e019933-0a90-4f74-b6ea-e4503cb4a96e_400.jpg (http://www.homedepot.com/webapp/wcs/stores/servlet/ProductDisplay?storeId=10051&langId=-1&catalogId=10053&productId=100012819&categoryID=502551)

:shrug:

It would give you Koolance testing conditions!

http://www.koolance.com/technical/myths/heat_exchanger_lab_results.html

[SNiiPE_DoGG]

It would give you Koolance testing conditions!

http://www.koolance.com/technical/myths/heat_exchanger_lab_results.html

ZOMG LOL!!!! that is the most bullshyte test i have ever seen

[Grinch]

ZOMG LOL!!!! that is the most bullshyte test i have ever seen

agree 100% :up:

[Muunsyr]

Static pressure, if I am not mistaken, is the air pressure the the fan will be able to maintain if the air is not moving (like if it is trying to blow into a box without any air being able to escape - or complete resistance). The other figure (volume of air per unit time) is normally the rating of the fan given no resistance whatsoever. A radiator will fall somewhere between the two - some resistance, but the air can still pass through.

[Vapor]

3600RPM to 4800RPM is a 33% increase...matches the CFM increase ;)

As for pressure....it's a mostly a crapshoot on ratings but generally the higher the RPM, the less loss there is to restriction (as a percentage). Also, for a fan design, the CFM through a radiator should be constant at any given RPM.

This means that if I have a Yate Loon SH at 1500RPM, it will push the same amount of air as an identical Yate Loon SM at 1500RPM.

For finding a good radiator fan, I'd trust user experience as well as the testing that Cathar and I are doing.

For getting a 220mm fan to do the work of a few 120mm fans....I honestly am not sure how that would work out. The 120mm fans would have less of a CFM loss due to restriction, but in the end, I don't know which would be quieter for a given amount of airflow.

[Waterlogged]

It would give you Koolance testing conditions!

http://www.koolance.com/technical/myths/heat_exchanger_lab_results.html

:rofl: :ROTF: :clap: AMEN!

[MrToad]

3600RPM to 4800RPM is a 33% increase...matches the CFM increase ;)

:embarrassment: :embarrassment: :embarrassment:

/me looks for a hole where to put the head in.

As for pressure....it's a mostly a crapshoot on ratings but generally the higher the RPM, the less loss there is to restriction (as a percentage). Also, for a fan design, the CFM through a radiator should be constant at any given RPM.

This means that if I have a Yate Loon SH at 1500RPM, it will push the same amount of air as an identical Yate Loon SM at 1500RPM.

For finding a good radiator fan, I'd trust user experience as well as the testing that Cathar and I are doing.

For getting a 220mm fan to do the work of a few 120mm fans....I honestly am not sure how that would work out. The 120mm fans would have less of a CFM loss due to restriction, but in the end, I don't know which would be quieter for a given amount of airflow.

So, in short, no conclusions extracted from data sheets are a substitute for real life experience, and given the fact that a 220mm fan would have a very limited application in our "field", I have a snowball's chance in hell of you or Cathar building specially designed shrouds and getting hold of 200-220mm fans to test them, so I will have to get my feet wet, and see if it works by myself...

If it doesn't, my bank account is never going to forgive me...

[Martinm210]

A fan in alot of ways behaves very similar to a water pump in that there is a curved relationship between static (0 flow) pressure and 0 pressure flow rate.

If you had calculated the pressure drop of air passing through a radiator, and had the curve for the fan, you could calculate the air flow you could expect for that fan just like you can for a water pump across your water cooling system.

Generally the deper 38mm type and the fans with larger fan blades will have a larger amount of pressure capability, and something with a very thin open finned design will have very little.

[MrToad]

Thanks for the replies.

I've borrowed one of those Antec 200mm fans from work, and I'm going to do some informal tests tonight.

We'll see...

Well, I've "informally" tested the fan with some home made tools (a cardboard shroud for the PA120.1 with some duct tape abuse, and a "sock" anemometer) and... well, great disappointment indeed. Somehow I was expecting a miracle that of course didn't happen.

Great in open air, but as soon as you impose some restriction, it just doesn't have the required "ooomph".

[bobo5195]

Yes there are some equations which relate rotational speed to a whole load of factors and over a fairly wide range they work pretty well. Combine with a few motor equations you can work out the curve of the fan graph pretty easily. Similar equations also hold for centrifugal pumps as well and most engineers should be able to derive by hand with sufficient arsedness.

An example link of some is below
http://www.airturbine.com/tools/fanlaws.html

There might be a textbook with some more somewhere.

I quite liked this article as well despite the fact it contains no fan laws.
http://www.acal-radiatron.com/fanselection.php

I’ve seen a link to noise characteristics somewhere, when a company used fluent and a sound evaluation code to work out how noise scaled with speed and other parameters.

Using a big ass fan is actually not a bad idea for radiators. Indeed it would be a cunning plan for a company manufacturing integrated products to do (hint hint) . I bet a PA160 performs quite well with a 160mm fan.

[iadstudio]

A fan in alot of ways behaves very similar to a water pump in that there is a curved relationship between static (0 flow) pressure and 0 pressure flow rate.

If you had calculated the pressure drop of air passing through a radiator, and had the curve for the fan, you could calculate the air flow you could expect for that fan just like you can for a water pump across your water cooling system.

Generally the deper 38mm type and the fans with larger fan blades will have a larger amount of pressure capability, and something with a very thin open finned design will have very little.

In theory this would apply for a radiator that had no fans mounted on it and only relied on the air presure in the case. So if one were to mount a passive radiator as a blowhole in an air tight case and have 2 fans blowing in the case with only the radiator to exhaust the air, I'd be able to get a decent "flow rate" of air through the radiator without mounting fans on it.

Does that make sense?
Where can one find figures for air flow presure drop through a radiator?

[Vapor]

specially designed shrouds and getting hold of 200-220mm fans to test themAnd there's the biggest issue.

1) I've tried testing a 172mm fan w/o a box frame (i.e., it had a roundish metal frame along the outside, no square frame that we're familiar with on 15-140mm fans) and it was just about impossible....but I'm sure one exists or could be made. but....

2) downscaling shrouds kill performance. Because with most shrouds airflow restriction is IMMEDIATELY at the edge of the blades (i.e., there is no 'reservoir' of air surrounding the fan) performance falls drastically. In my 140mm fan testing (on a testbed built for 120mm fans), the downscaling shrouds increased noise CONSIDERABLY and decreased airflow noticeably....performance of all my 140mm fans ended up below the average 120mm fan.

This is not to say that a 220mm fan will perform worse than a 120mm fan, but a shroud, unless properly designed, will considerably hamper performance. The ideal shroud would have a box shape for an inch or two and then begin shrinking down to the desired size, IMO. Maybe even a zone where it's bigger than the fan (as annoying as that would be to use).

3) Putting a shroud on each side of a fan (one for mounting to the rad, the other to the case, for instance) is even worse for performance. Again, properly designing the shroud will help, but it will make it take up more real estate.

4) Yes, I've been using a lot of numbered lists lately, they've become very convenient :p:

[MrToad]

Well... as I said, time to get my feet wet.

I've tried another fan, a 250mm one (the one you can find on the side panel of the TT armor/kandalf...). Useless.

So... time to step up. I've ordered four of these (http://store.thermalfx.com/merchant2/merchant.mvc?Screen=PROD&Store_Code=T&Product_Code=109E1712M503&Category_Code=172).

Knowing how US > UK deliveries work, it will take an age and a half to arrive, so I have plenty of time to order the rest of the parts and assemble them waiting for the fans.

We'll see if four of those can do the job of 9 120mm fans with a better air flow/static pressure/noise ratio.

More to come...

[jws2346]

And there's the biggest issue.
3) Putting a shroud on each side of a fan (one for mounting to the rad, the other to the case, for instance) is even worse for performance. Again, properly designing the shroud will help, but it will make it take up more real estate.
:p:

Now I'm completely K-fuked :shrug: . I have one E6600 OC'ed just 3GHz (at present) folding 24/7. I have a MCR220 rad with 4x D125SH fans (push/pull, my rad is external), MCP350 (unmodded), Apogee GT WB, 3/8" tubes runnin' some Pentrosin and distilled H2O. I have shrouds on both sides of the rad, the OP was talking about just putting shrouds on both sides of the fan, right? (I'm unfamiliar with that proceedure and have never seen it, I don't think) I sure hope so, it was just a little unclear what he was talking about. I'd hate to think I modded some Intel plastic packaging into fan shrouds for nothing :mad: . (I've also used the plastic containers for cat litter) Would this be like the "rad box" thing for hanging a rad off the back? :cool:
This fan thing is "killin'" me. I've read several posts, on several forums and they all say to use the Yate Loon fans
Edit: I've got a E6420 air cooled in a Antec 900 case (Zallman 9700 HSF). The case has a gimongus fan on the top (I think it's a 220 mm fan), is this the type of fan that's being disussed?
My other E6600 is down at the moment, no biggie. The WC'ing loop has been modded extensively, it has a sorry a$$ Via Agua 1300 pump with more foam under it than Sealy Posterpedic and it has a 120 mm fan blowin' on it to keep it cool, I do love those Yate Loon fans).
I'm collecting enough aluminum cans to get both some type of quad (either a Q6600, x3210 or a x3220) and a P35 mobo. I want some of that quad powa' for foldin'. (I wanted to wait and see if there's anything to AMD's new quad)
I'm also very impressed with the GTX WB in copper. Sure looks "puuurty"

[MrToad]

LMAO!

It will be some time before I order anything from the US again...

I've just been charged 50% tax on the fans I ordered from ThermalFX.

So, with delivery and tax that's going to be... hmmm... 133 quid ($266) for four fans.

Gosh. I hope they are the bee's knees. And hope that whoever was the customs officer responsible spends the weekend sitting in the toilet with his intestines hanging down the drain.

[Brian MP5T]

From what I've gathered so far, a fan's rpm rating is all good and dandy in open air, but is the static pressure what determines what's the volume of air that will actually go through a restrictive body (in our case, a radiator), correct?


Your point is very valid and true. There is always more than one thing to think about, to increase the effectivness of your cooling, moving more air volume will help, but much larger gains in heat transfer can be made by simply lowering the temperature of the air flowing through the core.

I have toyed with the idea of running a liquid cooled radiator into an old freezer. Basically, the fans would move the very cold air through the core.

Several benifits..

1. The fans would be completly quiet due to the construction of the RefriGGERATOR.

2. You would have a nice place to keep some drinks..

A Small unit can be easily found very cheap and I believe that under 3 Hours, you could have it completely set up.