Introduction

As an enthusiast it can be difficult to just "set it and forget it," to assemble a system and decree "this is as good as it's going to get." There's an inherent need to tweak and continue to tweak, to eke every last ounce of performance (within reason) out of our systems. Over the past few years, liquid cooling has become less the province of the extreme enthusiast and more accessible to the average user thanks to closed loop coolers manufactured by Asetek and CoolIT and brought to market by companies like Thermaltake, Antec, and Corsair.

The pump and radiator are only part of the equation, though; part of what makes even a decent closed loop cooling system tick is having a good fan configuration. Reading specs on the fan boxes helps a little, and visiting forums can certainly help, too, but we wanted something a little more definitive. After a couple of weeks of testing, we have results to share.

When dealing with fans being used for radiators, it's important to note that what makes a good case fan may not make a good radiator fan, and vice versa. We've gotten used to reading fan specifications that only list the maximum airflow of the fan, rated in CFM or "Cubic Feet per Minute." In the past couple of years, though, more and more manufacturers have been listing an additional specification, and this is the one we're interested in: static air pressure. Fans which produce high static air pressure are able to better focus and direct airflow, making them more ideal for forcing air through the densely packed fins of a liquid cooling radiator.

The propensity for manufacturers to list the air pressure specification over the past few years coincides with the increased popularity of closed loop liquid coolers, but during the same period of time we've also seen a gradual shift towards quieter computing. Builders place greater emphasis on having their systems run quietly, and why shouldn't they? If you can have good thermal performance with a minimal impact on ambient noise, why wouldn't you?

With these things in mind I've tested a collection of eight fans from Corsair, BitFenix, Nexus, NZXT, Cooler Master, and SilverStone to try and find the best balance between thermal performance and acoustics.

Testing Methodology
POST A COMMENT

112 Comments

View All Comments

  • Hummingbird - Thursday, August 23, 2012 - link

    I don´t think that this fan roundup really helps to decide which fan is the best out there as long as the Corsair H80´s radiator is far away in terms of it´s dimensions (depth and fin gaps/distance) from the standard radiators which are normally used for H2O setups. Reply
  • eric appla - Thursday, August 23, 2012 - link

    Thank you for efforts with the review Dustin. When I seen Anandtech doing fan review I was very excited that I'll see very detailed and scientific approach as I'm used to from AT articles. Unfortunately this review was not fully what I expected

    What I'd find useful in the Fan reviewes is also some CFM mesurements and noise measurements.

    Noise measurements are very tricky as even A weighted dBA noise level is not telling full story. I would suggest to provide noise spectral picture as it can show at what frequencies are the dominant components of the noise. Some fans do hum, some squeek.

    On the CFM and static pressure front it will be interesting to measure them in few scenarios.
    1) open air (case fan scenario)
    2) dust filter (case fan scenario)
    3) low restriction radiator for example XSPC RX series
    4) medium restriction radiator for example XSPC RS series
    5) high restriction radiator for example XSPC EX series

    Pretty much every radiator manufacturer makes few different radiators optimized for different fans based on the static pressure and airflow they can deliver.

    Full review like this will show people that there is no single best fan in the world, there are just best fans for particular usage scenarios.

    Challange is on :)

    For Fan reviews to use with radiators I usually go to websites specialized on this narrow subject such as martinsliquidlab or skineelab as they have the required test equipment. Have a look at what these guys do, I think you'll find it exciting and inspirational and as I know you based on your approach to reviewes you'll surely find a ways to improve their methodologies even further.

    Thanks again for all the efforts
    Reply
  • krumme - Thursday, August 23, 2012 - link

    A spectral plot wont bring us any closer, as there is dynamics involved to, and the charateristics of that dynamics. Besides that its expensive to do.

    Just add a colum with subjective impression/valuation and notes, fx. clicking, high picth tone..., even distributed...

    Its the subjective that matters in the end.
    Reply
  • Robalov - Thursday, August 23, 2012 - link

    Nice article and get's right to the point, however I'd like to see 20+ fans tested.

    They're are so many touting that their 'special' design knocks the socks off the others.

    Fans don't seem to get the air time (ba dum tiss) of other components are are left entirely up to personal recommendations on the forums, as the specs are just massaged or even outright lies.

    As an aside, CoolerMaster really are a horrible brand.
    Reply
  • BlueReason - Thursday, August 23, 2012 - link

    First off, it's nice to see Anandtech making an effort to evaluate fans, a component sorely under-prioritized in rig setups (not to mention extremely difficult to review).

    However, as stated by others, the NF-F12 really should have been tested. I don't even like any of Noctua's other fans, but the F12 is possibly the most specialized, single-purpose consumer grade fan on the market, engineered ground-up for heatsink performance. Whether it delivers on that performance is debatable (in my experience, it does), but its exclusion from the comparison is rather odd.

    Are any of the reviewed fans even PWM?

    On a side note: some F12 units seem to have acoustic issues, hence the occasional "loud" comments. My first F12 suffered from this and was replaced by Noctua, to good effect.
    Reply
  • khanov - Thursday, August 23, 2012 - link

    No no no. You can't say that those Corsair fans win, since they only manage to equal the performance of Noctua's "old gentleman" NF-P12. That is not a win from Corsair as those Noctua's have been available for 6+ years now (but still it is a good first effort from Corsair).

    None of the fans you tested can actually best the performance of a Noiseblocker M-12P at the same or lower sound pressure level on a low profile, high-fin-count rad. (and that is what these high pressure fans are for, right?). So the real winner is Noiseblocker, but you decided to exclude them along with Noctua. Shenanigans, Dustin.

    Noctua's NF-P12 are a good second best for many peeps into water cooling and are usually available at a reasonable price. If you are planning to push/pull a 3x120 or 4x120 rad for example, then the cost of 6x or 8x your chosen fan adds up very quickly. So to be reasonable about it the Noctua's should realistically win simply because the Corsairs are WAY overpriced right now (flavour of the month?).

    Noiseblocker's M-12P's are still the real performance winner (cooling vs perceptible noise) but may cost you several arms and legs to import depending where you live.

    I like that you have a WC setup and are now testing fans vs. rads. I look forward to future fan/rad. tests but just do it properly please. Excluding the real players is like saying the new Ford Fiesta is the fastest car on the Nürburgring because you deliberately excluded any real competition!
    Reply
  • kg4icg - Thursday, August 23, 2012 - link

    Nice review. I'm using the H100 instead of the H80 and yes I am using the stock fans. The thing is, I also have the Corsair Link hooked up inside my system which is controlling fan and pump speed instead of 1 button on the cpu block. Ironically I have more power hookup's for fans than I have room for fans in my case. Reply
  • Iketh - Thursday, August 23, 2012 - link

    Can't believe this fan was not included.

    http://www.newegg.com/Product/Product.aspx?Item=N8...

    It has winglets designed specifically for static pressure. The bearing it uses is also 100% silent. At max RPM, the fan will get loud though because of the winglets, but completely silent mid-low.
    Reply
  • Onus - Thursday, August 23, 2012 - link

    So, once again, Crappermaster is caught lying about their products. Amazing. Reply
  • thralloforcus - Thursday, August 23, 2012 - link

    I would have loved to see some Noctua, Noiseblocker, and Cougar fans tested. These seems to be pretty popular for radiators. I'm using the stock H100 fans right now in my H100, with two Noiseblocker M12-P fans for pull.

    Of course I'm always looking for better performance! I was using two Panaflo FBA12G12H1BX fans with the stock fans, but it seemed that the voltages given out by the H100 were too low, and the fans kept shutting off.
    Reply

Log in

Don't have an account? Sign up now