26 120mm Fans Tested and Compared

[dqniel]

Two Yates for one Papst, my friend. Yates are about $6 stock and about $7 braided and sleeved. Papsts come sleeved and ready to go at $14 per fan. Of course, I don't pay retail... but that's a seperate issue.

I don't know where you're looking but I got my batch of 4 Yates for $12. One Papst is more than that.

[kuhla]

So I read over it in more detail today and so the final conclusion is essentially this:

Best Case Fan:
<27dB Sharkoon Golfball 2000
>27dB Papst 4412FGML

Best Radiator Fan:
<27dB Sharkoon Golfball 2000
>27dB Papst 4412FGML

Best CPU Heatsink Fan:
Sharkoon 2000 pretty much the whole way.

Am I right? Those results would be because those fans show the highest cfm per 1 dB, right?

If yes, I think I'm going to be getting some 4412FGML's soon

Links for the lazy:

Papst 4412FGML -->> http://www.performance-pcs.com/catal...oducts_id=4639

Sharkoon 2000 -->> http://www.performance-pcs.com/catal...ducts_id=20832

EDIT : One thing I forgot to mention is "best cfm to dB ratio within a certian reasonable limit" because when I calculated a few of the #s (from eyeballing it) I saw the 100+ cfm fans doing well over 2cfm per 1dB which is really good but those fans measured as being very loud. Am I right in assuming that too?

[Spawne32]

i just ordered 2 of the papst 4412FGML's and sold my 2 noctuas, ill prolly order 2 of the sharkoons also and compare them and use them in combination with a radiator for my watercooled setup

[Spawne32]

theres one thing i dont understand tho, out of the papst and the sharkoons, on the radiator tests, how come the papst did so much better then the sharkoon and how come the sharkoon did better on the cpu test over the papst, I had assumed that the sharkoon would have had the higher static air pressure because of the fin design over the papst, but it seems to be reversed?

[Vapor]

EDIT : One thing I forgot to mention is "best cfm to dB ratio within a certian reasonable limit" because when I calculated a few of the #s (from eyeballing it) I saw the 100+ cfm fans doing well over 2cfm per 1dB which is really good but those fans measured as being very loud. Am I right in assuming that too?

dB is logarithmic....ignore CFM/dB ratio. Yes, those fans are very loud. EDIT: also: dB is not loudness, it's intensity. Loudness is subjective.

theres one thing i dont understand tho, out of the papst and the sharkoons, on the radiator tests, how come the papst did so much better then the sharkoon and how come the sharkoon did better on the cpu test over the papst, I had assumed that the sharkoon would have had the higher static air pressure because of the fin design over the papst, but it seems to be reversed?

CFM through a radiator (NOT STATIC PRESSURE!....that's something a little different) is only part of the equation for cooling performance. Airflow dispersion patterns and hub sizes are huge factors. Think about it this way: if you have airflow through only part of a radiator, and you miss any critical parts, the CFM may be high, but the cooling just won't be there. However, dedicating airflow to useless parts of the radiator (on a TTBT, not all parts are equal), can alter the situation as well.

Hub sizes are equal...so airflow dispersion is probably the factor here. Also, look at the data points...CFM v dB is 1V increment (count the dots) and dT v dB is at 12V, 9V, 7V, 5V.....only at 10V and 11V did the ebmpapst noticably beat the Sharkoon.

[serialk11r]

Yea, the formula for db is like this. If something is x times louder than another something, then it is 10 log base 10 x more dba. (since its base 10 we can just write 10logx ) So thermaltakes are about 10 times louder than they claim, and the best cfm/db ratio fans are actually by far the noisiest (cfm/noise ratio). Hey vapor perhaps you could write that somewhere, so people don't get confused by the db thingy?

[FLMJIGGY]

Great thread!

I've used a few of these fans and I still go for Yates. CHEAP, bought about dozen or so and no probs, and nicely undervolt for silence. The Silvy FN is nice too but not worth the extra $ IMO.

CoolerMaster LEd is pretty decent too. With a 7v mod it is almost silent.
http://svc.com/tlf-s12-ep-19.html

I believe some of the Arctic Cooling' ue the S-Flex tech as well and are much cheaper.

Nexus + V - $ = Yate

The TRI-Cools from Antec are not worth i. Too loud on high- it does push some CFM but too loud for what it pushes. On low it's not too bad. Almost a 5v Yate

[Spawne32]

CFM through a radiator (NOT STATIC PRESSURE!....that's something a little different) is only part of the equation for cooling performance. Airflow dispersion patterns and hub sizes are huge factors. Think about it this way: if you have airflow through only part of a radiator, and you miss any critical parts, the CFM may be high, but the cooling just won't be there. However, dedicating airflow to useless parts of the radiator (on a TTBT, not all parts are equal), can alter the situation as well.

Hub sizes are equal...so airflow dispersion is probably the factor here. Also, look at the data points...CFM v dB is 1V increment (count the dots) and dT v dB is at 12V, 9V, 7V, 5V.....only at 10V and 11V did the ebmpapst noticably beat the Sharkoon.

Ok ok hold on *breaths deep* let me take all this in here... lol vapor can we measure the size of your brain by chance? Cus damn, i didnt even know there was that much crap to factor in.

[serialk11r]

Ok ok hold on *breaths deep* let me take all this in here... lol vapor can we measure the size of your brain by chance? Cus damn, i didnt even know there was that much crap to factor in.

lol...don't tell me you think he's like albert einstein or something.
BTW good to see you're not using that ultra anymore

[Vapor]

Cus damn, i didnt even know there was that much crap to factor in.

There's unfortunately tons to factor in....everything down to internal case airflow and HSF design (though the HSFs should be designed around typical fan characteristics). The lesser characteristics don't effect things AS MUCH, but they're still a factor. Overall, a better fan is still what will give you the biggest boost, rather than a bad fan with a well matched HSF.

Ultimately, this data should only be a reference point.

I am no Einstein....these are just fans and air movement....

[kuhla]

dB is logarithmic....ignore CFM/dB ratio. Yes, those fans are very loud. EDIT: also: dB is not loudness, it's intensity. Loudness is subjective.

I know all that but even if you look at it as if it were a linear graph it still holds some value to consider a straight cfm-per-dB "ratio".

So would you say this conclusion (quoted from myself, above) is accurate?

Best Case Fan:
<27dB Sharkoon Golfball 2000
>27dB Papst 4412FGML

Best Radiator Fan:
<27dB Sharkoon Golfball 2000
>27dB Papst 4412FGML

Best CPU Heatsink Fan:
Sharkoon 2000 pretty much the whole way.

[Vapor]

I know all that but even if you look at it as if it were a linear graph it still holds some value to consider a straight cfm-per-dB "ratio".

Comparing a Yate Loon SL to a Delta 220CFM behemoth with the same "ratio" just doesn't seem effective, IMO. The most valid considerations to make are: how much noise can you tolerate (or you desire) or how much cooling you need....and then finding the best options from that.

So would you say this conclusion (quoted from myself, above) is accurate?

Yes, that's what the data says...there are other factors that go into choosing a fan, like aesthetics, price, reliability.

[serialk11r]

I know all that but even if you look at it as if it were a linear graph it still holds some value to consider a straight cfm-per-dB "ratio"

So the 220CFM delta offers more air flow for the noise than a yate loon low speed?
Actually, the straight cfm/db ratio probably goes the other way, the noisier the fan is, the "better" the ratio (when you're comparing fans of different speeds).

[sacha35]

Nice thread, but i see you did not test the Xinruilian 120mm fan 0.4 amp, this fan kicks out some very good CFM but it is not very quite, but if it is cooling you are arfter then this has got to be the fan of all fans.

[Vapor]

That one?

[sacha35]

That one?

yep thats the one.

[Spawne32]

lol...don't tell me you think he's like albert einstein or something.
BTW good to see you're not using that ultra anymore

I think we should erect a statue in vapors honor at the front of XS and im not using the ultra because the zalman is more suited for my design, i was going for quiet and because of the zalmans cooling design, it makes virtually no noise at all. The ultra PSU's are actually good PSU's on the higher end models, if you would like to argue more about this i suggest you take it on over to jonny gurus forum where he can learn you a thing or two

BTW: i got my papst fans in today!! now to test them out

[kuhla]

Yes, that's what the data says...there are other factors that go into choosing a fan, like aesthetics, price, reliability.

Thanks, that decides it for me. I'll probably be putting in an order for 3 papsts sometime soon. The price is a bit high and I am unsure of the reliability but that doesn't bother me. Aesthetics is why I am not going for the Sharkoons. Black on white isn't quite my favorite color contrast

[Allsorts]

Thanks Vapor, I really enjoyed reading through your results.

Great work!

[bito]

How do the Panaflo Ultra quiet Low speed models do then.....68.9CFM @30dBA......also, IIRC correctly, these have an excellent pressure too, so good for radiators too !!
http://www.sidewindercomputers.com/pa12ulqu.html

[ziddey]

spendy, but looks good, if those specs are real. maybe i'll upgrade my yates with those someday. should have some decent 7v performance too hopefully

edit: newegg has it a little more expensive, but it looks like it's a 3pin rpm model

[Vapor]

How do the Panaflo Ultra quiet Low speed models do then.....68.9CFM @30dBA......also, IIRC correctly, these have an excellent pressure too, so good for radiators too !!
http://www.sidewindercomputers.com/pa12ulqu.html

That's in my test....Panaflo L1.

[Sparky]

Pretty sad.... Serialk11r is what, almost 6 years younger than me and sounds like he has a better grasp on the whole logarithmic stuff I'm in calc 2 right now and I still don't really understand logarithms, just how to survive through them

Last papst fan I had (actually I might still have somewhere) was one that I pulled from an old 8088 PC - 3 blades, Guessing about 150-160mm, solid steel, and 120V Once that thing got spinning you didn't want to stick your finger in it, it hurt pretty bad If I find it I'll snap some pics.

[SiGfever]

Very interesting review.

I definitely re-evaluated my fan choices. But then I have known for quite some time that Ian is one sharp guy.

[syne_24]

Wow I didnt realized you finished this already. Another vote to see the Sanyo Denki in there and possibly the Yates Loon SH in the high performance category.