Fans, Fans, Fans -- XS Fan Review!
Round 2 is here :)
As a follow-up to my first round-up, found here: http://www.xtremesystems.org/forums/...d.php?t=137832. I have tested many new fans and retested some of the best from the first roundup.
The environment and testbeds have changed slightly, so unfortunately results are not directly comparable, but this should still answer many questions about many of the most popular fans on the market. :)
Some test changes did occur!
1) CPU temps were ditched. It was an extraordinarily time consuming process that didn't really differ from the radiator test results.
2) Fans were tested from 5V to 12V in 1V increments. This is a change from Round 1's 4V-12V + minimum running voltage that was tested. Again, for the sake of reducing test time (min running voltage was a bit of a PITA and who runs their fans at <5V anyway?).
3) The dBA scale is, again, an artificial adjustment of raw data to 'consumerize' the data to numbers that we are more familiar with. My 30dBA is not exactly the same as Company A's 30dBA. Then again, Company A's 30dBA isn't the same as any other company's anyway ;) I have used data from a few fans with known-good dBAs and come up with a simple linear conversion for the data....neither the order nor magnitude of differences is changed! Because the environment has changed from Roudn 1, so has the adjustment method. Again, it's not at all exact science, but it does make the numbers appear friendlier without actually changing them :)
Other notes!
Again, an MCR120 was used as the radiator. All fans were tested in pull in relation to the anemometer and the fans were tested in pull in relation to the radiator as well. If you're wondering why, I believe the thread linked above explains.
Again, all fans have undergone a 24-hour burn-in prior to empirical testing. This is because most people run their fans for a long time. Long-term performance is a lot more important than out-of-the-box performance. I did do 0-hour CFM and dBA tests at 12V and 7V though don't have the results listed....where it matters, I'll tell you.
All fans were also properly warmed up immediately prior to testing.
And if the fan is not listed, I haven't tested it...
2 exceptions: FN122 and FM123.
The Silverstone 'Debacle'
I have tested both the aforementioned FN122 and FM123 for Round 2 and have not listed their performance. I hate them with a passion. Neither are anywhere near their CFM ratings and the FM123 is more annoying to listen to than the 220CFM Deltas in Round 1. I'd go as far as saying it's more annoying than a crying baby. Unlike the baby though, you can kick it. At this point, the FM121, FM122, FM123, and FN122 have ranked among the most annoying or worst performing fans I have ever tested/used. They may all be lemons, but if the lemon rate coming out of Silverstone is 80%, you should stay away. If they are good representations, you should stay away. So ultimately, based on my experience with them, you should stay away. And I really don't have a bias...I actually like the FN121 quite a bit. Shame it has retarded siblings.
I wanted to exclude all Silverstones from this test, but because I was using an FN121 as a carry-over all-star comparison from Round 1, that was not possible. It actually did fairly well in this round-up as well. Go figure.
The Sleeve Bearing Grind
It's common knowledge that sleeve bearing fans have positional tendencies....mostly poor longevity when stored or used blowing up/down. It has to do with the lubrication not effectively lubricating the entire bearing. It's annoying.
They also have low MTBFs....so they've been arbitrarily punished when it came down to selecting my favorite fans. Yes, they're often less expensive...but that's because they're cheaper to make. No major 'punishment' but if it was an on-the-bubble tie for two fans, nod went to the non-sleeve bearing fan. That said, there are some EXCELLENT sleeve bearing fans available and all for under $10 a pop.
Sanyo Denki Makes Some Really Nice Stuff
I'll spoil one result for all of you upfront: the Sanyo Denki San Ace 1011 is king of the hill. And by hill I mean universe. It's really, really good. In addition to its insane empirical results (see below), it makes NO extraneous noises and undervolts like a champ. I have five and they all perform like this and they all start at 4V or lower. They could kick Chuck Norris's ass....silently.
They're ball bearings as well, so you don't have to worry about the sleeve bearing inconvenience. They also perform past their listed specs--something that's exceedingly rare in these SilenX days. Good luck finding any though...and when you do, they're $25 a pop. I also really like Sanyo Denki's Eneloop batteries (hey, I said "stuff" in the title, that implies plurality).
Not All About the Blades
Something I plan to explain further later....but 'static pressure' (realistically: radiator performance) is not ALL about the blade structure. A motor design difference (or just a bad motor design) can hurt radiator performance just as much as a bad blade design. Just want to prepare everyone for when I go into this further....
Subjectivity Need Apply
Empirical data says one thing...but it must be combined with subjective data as well. Often times murmurs, clicks, even motor thudding goes unnoticed by the test equipment. I'll report my opinion on each fan and back it up with what I can.
Don't Give Me a Hallelujah!
Do not take this as gospel. There are other reviews out there from other people and sites that test fans very well. SPCR's recent reviews (after the testbed update) are very nice and a good resource for super-low-noise folk. Cathar has also tested some fans on a radiator in the WC section...unfortunately he no longer appears active on this forum. His testing was top-notch work and even though our data didn't corroborate on every result, that's important. It means there's variability between individual fans and exactly why this shouldn't be taken as gospel. That said, our data does match significantly :p:
Most importantly, trust your ears. If one fan sounds louder to you, that's what counts. Not what a dBA meter says--what your ears say. On the other hand, I wouldn't trust your hand for CFM measurements....at all. Really, don't do it...airflow patterns are more important than airflow quantity for what your hand will tell you.
A 250 Word Essay on Something You've Learned
Okay, I've learned a lot by testing some 50 fans with different designs. I really can look at a fan and its RPM spec and tell you how it will roughly perform. There are many key design features that lead to different fan characteristics. Blade design, motor design, bearings used, materials used, manufacturer, etc. Additionally, there are some 'rules' when it comes to approximating maximum fan performance (they're pretty damn solid ;)). I won't detail them here. Also remember I'm fairly reserved when it comes to passing judgement on a fan. So when I call out a fan without testing it, I'm doing it with extreme confidence because there is some glaring issue/problem/whatever with it.
One such example is the new Scythe Slipstream fans. If you have a moment, please compare CFM to RPM specs of the various fans in the series. Please remember that CFM to RPM is roughly linear, with a slight reduction in efficiency as RPM increases....now try the same thing for the Minebea series (very good CFM ratings ;)) and look at the numbers....tell me again those Slipstream specs aren't BS. I'm not saying they're bad fans. They may be very good in the open air, I haven't tested them....but I will say their specs (especially the 3 fastest) are in the league of SilenX and Silverstone. Their radiator performance will also be mediocre at best (probably why the Ultra Kazes were simultaneously introduced ;)). I don't know why manufacturers can get away with this...and worse yet, people believe them.
Bored Yet?
Enjoy the round up :)
