In my opinion the Gentle Typhoon is the winner.... posthumously anyways.

After some years (about 7) I mean it's a relatively overrated worry: in my experience the only fan which actually did delude me was the 80mm Scythe S-Flex (which sports a fluid dynamic bearing!).

Fanboy is a term used to describe a male who is highly devoted and biased in opinion towards a single subject... ...and might aggressively defend their supposed superiority compared to the other brand(s), be skeptical or in denial about negative reviews of the product, and exert a high level of brand loyalty..

You're a Noctua fanboy, ces, period: I guess probably I'm not as "politically correct", but in "my english" it is not an offense, and I don't use it as an insult.
At any rate, if you did take offense, or if it's an insult in "real english", I beg your pardon.


But this opinion (even before that questionable) is rather OT here: there's a couple of official, dedicated threads about the L12.

Here they come the fans: and if we can say anything in topic, then it's that Noctua fans of any size are often no more effective or even ess effective than the very best ones (that is not such a plus for such expensive fans). They're not bad but overrated, cooling & noise wise.

In case, that "L12 case" (pun not intended) just confirms this fact: it seem that it take two fans (and an about four times spending) to "beat" a single common, cheap, sleeve bearing Nexus Real Silent Case Fan.

I don't think anyone here proposes that NH-L12 is a anything but a good heatsink. (Notice how I'm trying to be diplomatic here to avoid further offtopic discussion)

This thread, however, is about fans =)

Fantastic to see a new fan roundup and great to see these brands in there that have been talked about a lot but never had serious testing. Thanks SPCR! The more fans the merrier by the way Everyone wants to see how their fan compares.

A little touchy aren't we?

I personally use Prolimatech, Scythe and Noctua heatsinks... and Scythe, Noctua and Nexus fans. I like good stuff that tests well. I am a fan of the quality of the testing done by SPCR.

And there I thought you were the fanboi.

If I am overlooking any facts, please correct them. The adhominem attacks on me or any other person, just gets in the way of the facts. I welcome any attempt to shut me down or up with facts. The name calling... what do you hope to accomplish with that?

I propose you just stick to the facts. What facts am I in error on?

I would submit that in this instance (the anomalous results of different fan configurations and brands on that particular heat sink) you really can't separate the two subjects. You have to admit, the fan performance results are non-intuitive.... don't you... two inferior fans out performing a single superior fan where you would not have expected them to do so?

Now THAT is a pertinent question. I however don't see the results as being non-intuitive myself.

We know from experience that noise drops VERY rapidly as RPMs (or perhaps rather "air speed change", which will not be 1:1 with RPM when comparing different fan blades) decrease. So 2 slightly inferior fans running on lower "air push ratio per fan" cooling slightly better than 1 better fan at the same SPL makes sense to me.

... but I'm merely a generalist and have no specialist knowledge in fans and air turbulence, so if there's a reason as to why this shouldn't be making sense - please do enlighten me!


As an aside, I can see how 2 identical fans with identical noise signatures could end up being 1 + 1 = 2 perceived noise. But in this situation the fans are also not identical, and not operating under the same conditions (1 is pushing, 1 is pulling), so the resulting noise likely ends up being 1 + 1 = "1.5" perceived noise with some waveforms partially cancelling eachother or simply peaking at different frequencies. (Please ignore the fictitious ratios, they are there to illustrate a point).


Edit: Case in point: why do chassis makers now put multiple slow fans in their cases, rather than 1 fast?

The numbers you use are difficult to correlate to either decibels or "times" perceived level. In general, a 10 dB increase = 2x perceived loudness. The smaller increments are difficult translate to a times perceived level.

But here's an empirical observation: 2 identical noise sources (ie fans) should add up to +3 dB over one. But at the low levels we most often deal with at SPCR, the measured SPL never seems to add up to 3 dB. Most often, it's not much more than 1 dB. This applies to both quiet fans and quiet HDDs, typically under 20 dBA. This applies in the anechoic chamber as well as in live rooms... though I have to admit it's been a while since we did any extensive measurements in live rooms.

Look at our measurements of system noise in the Silent PC Build Guides from a couple of years ago. There were several home server guides with multiple drives (one with 6 or 9 HDDs, iirc). The general rule is 2 identical noise sources leads to +3 dB, 4 sources leads to +6 dB, and 8 sources leads to +9 dB... but this was not borne out with the HDD noise. Here's one of those builds: http://www.silentpcreview.com/Nine_Susp ... nd_Testing

Anyway, it's true that 2 very quiet fans tend to outperform one. But if the difference is a decibel at <15 dBA and under say 2C, I doubt it matters to anyone except uber obsessives.

That be me

On the two Noctuas vs the Nexus... is that really a fair comparison? All the fans were tested using straight DC correct? In real life the Noctuas will be running off of PWN current. Will that not add some noise to them that might change the relative performance visa vis the Nexus?

Will the PWM impact on noise be greatest at the low RPM range?

Maybe my pulled-out-of-my-ass hypothesis on (partial) waveform cancellation wasn't so off (except I didn't think it applied to this case)? Thanks for relaying the observation!

Well, you can obsess about it all you want, but unless you sit within a foot of the fans in a really quiet room, you won't hear a 1 dB difference unless it is qualitative -- ie, a buzzy nasty 13 dBA vs a smooth broadband 12 dBA. Actually, that kind of qualitative difference would swamp the 1 dBA SPL difference. Even 15 dBA vs 16 dBA is very subtle.

As for whether it's a fair comparison, yes it was.

All the non-PWM fans were tested at different speeds by varying the voltage. That's the only way they can be run/tested, PWM does not work on them.

All the PWM fans were tested with PWM control -- except in the cases when 550rpm could not be reached using the Asus Fan Xpert 2 because of software control limitations.

In other words, all the fans were driven optimally.

Actually... except for poorly implemented PWM controllers a few years back (when they made both PWM and straight voltage fans sound worse), we don't really hear any qualitative differences today between slowing a PWM fan with PWM or voltage reduction. From a practical view point, it's just that in many cases a PWM fan will go from max speed at 12V to min at 8 or 9V -- very difficult to make fine speed adjustments.

nice review.

I made my own 120mm fan a few years ago...but at 44193 hours, may as well get on a waiting list for the scythe gentle typhoon.

my own fan is still doing very good, sleep next to my pc.

Mind, ces: there ain't no such thing as a PWM current.

PWM has all to do with duty cycle, not current (rectius, voltage): it's just a switching signal.
Definitely it's a more efficient (so not a better one tout court) way to drive a fan, that is power loss in the switching device is low, but it's always the same DC (or AC) to run the fan.
So, if you obtain the reference average voltage (and current) through a resistor, while not switching it on and off at a suitable high frequency, then you're wasting power over it, but that's all.

Mind ces, these things may be interesting, but the awful fact is that Noctua NF-F12 PWM is still clearly inferior, price wise, cooling wise, noise wise.

So, as the original debate was something like: "why don't you test some Noctuas", well the answer is still: 'cause they "sucks".
Ok, now I'm doing some "fanboyism": they don't "sucks". They just aren't the real best (and they are not priced accordingly).

Do you think I confused anyone with the term "PWM current"

Luca for future reference.... how would you have me refer to what I called PWM current. Is "PWM electricity" acceptable?

PWM "Electricity ', "current", "electron stream" "pulsed electron stream", "electron pulses" or what?

Personally I have always thought that the term "pulse" is somewhat misleading when used in the context of PWM c***ent. Don't you?

First off, fantastic review as always.

I'm left wondering about my current application on a radiator though (the stock Corsair fans are not exactly impressive IMO). The Scythe posted best results, and someone has already mentioned good static pressure, but the big hub blocking so much of the airflow through the radiator would seem to be a bad thing.

I'm tempted to wait for the test of the new Corsair fans for their targeted application design, but the claimed SPL on their site doesn't sound especially encouraging for a "Quiet Edition" fan.

About the L12 testing, there's nothing to actually debate but data are probably incomplete: IMVHO top mounting data might have been useful to somehow better assess the comparison.

I kinda feel like being a bit pedantic on the subject, so... Here we go.
PWM refers to a signal. Also, those PWM fans with a 4-pins connector don't actually run off the PWM signal: they run off 12V DC, the PWM signal being only a control signal. So one correct way to put it could be that "those Noctua fans will be controlled by a PWM signal" - or, more simply, "those Noctua fans will be PWM controlled".

In short, there were two things wrong with your statement:
- the word "current": in practice, what matters on the PWM line is the voltage (either zero or a set value, 3.3V being the recommended one), not the current. Besides, since it's a control signal, the current drawn from the PWM line is supposed to stay very small (absolute maximum is 5 mA).
- the idea that the fan runs off the PWM line: 4-pin PWM fans run off a 12 V DC line, the PWM line being only used to control them, not to power them. This is an important distinction, because PWM control can also be used with regular, 3-pin fans - and in that case, PWM is applied to the line that actually power the fan (the one which is nominally a 12 V DC line).

At least, that's what I got from the spec sheet for PWM controlled fans (it's on formfactors.org). I'm no specialist.

There's an integrated circuit chip in the fan itself, which has 12V DC as input and also receives the control signal. It outputs the PWM pulses based on what the control signal says. At 50% duty cycle this means that the pulse and the gap between pulses are the same width. This fan control chip also monitors the fan speed through the tacho (yellow) wire so that the reported fan speed is used as part of a feedback loop. If for example a a PWM fan is stopped manually the chip will detect this and continue attempting to restart the fan until the tacho wire reports that it is running again. A PWM fan can restart even at low duty cycle settings such as 20%. I have confirmed that this is the case in practice by manually stopping a Noctua NF-P12 PWM fan running at 350 rpm. The time to restart was around 2 to 3 seconds. There was no sign of an emergency 100% duty cycle starting pulse or anything like that, the fan just seemed to resume running at the same speed as before.

Since this is a pretty decent thread, I'm going to delete the series of baby-ish posts with no on-topic content. Hopefully the two posters will continue their little pissy-fest via PM.

Linus @ NCIX thinks Mike is crazy rich: http://youtu.be/9uUXt7mE6Qg?t=9m
Also, has access to a real CFM & pressure meter.
Linus & Mike need to get together to produce epic reviews.

Thats a shame, i guess this review came a little to late. Maybe its a matter of too many samples coming back to Scythe as RMA considering people talk alot about sample variance when it comes to bearing chatter.

I decided to buy 2 of these today as soon as i realized they wont be around anymore. Hope i understand this correctly in that the Scythe Gentle Typhoon fan is the best high pressure fan for horizontal mounting because its a ball bearing fan? Meaning they would be a good choice on a VGA heatsink like the Arctic S1?

I also bough the new Noiseblocker fans, not the multiframe, the ones called e-Loop or something. They were pricey but at least i will be able to report back on performance.

Edit: The two GT 800rpm samples i bought were dead quiet at 12v. The NB eLoop 800rpm only spin at 600rpm at 12v så i RMA them to NB.

">" means "greater than"
and
"<" means "less than"

so "slowed to >1200 RPM" means "slowed to greater than 1200 RPM" when I think you meant to say "slowed to less than 1200 RPM" ??

In which case it should be "slowed to <1200 RPM".

Perhaps avoid using the < > symbols?

Great review. Unfortunately, as has been observed, 800 and 1150rpm GT's are unavailable anywhere pretty much (though EU outlets still seem to have stock). Noiseblockers are the same way - unobtainable.

I wonder how noisey my case would be with 5 1450rpmGT's?

Just for the record...

I have had some five NB multiframes S2 for more than one year one, and have to say that they have been a pleasure and the best ones I have had with regard to noise -I tried Scythe GTs and BeQuiet SilentWings too, which are very quiet.

That said, *all three* 120 m NBs I tried to run horizontally have developed a ticking noise almost immediately, no exception. I consider them a no-no for horizontal usage. Hope this info is useful for you guys coming to SPCR looking for hard data.

Just to add some extra data, I have to say that I have a pair of 80 mm NBs, and they did not develop that noise when used horizontally.