Noctua NF-P12 ... vs ten other fans. Done.

Bluefront · Post by **Bluefront** » Mon Jan 21, 2008 5:24 am

Hummm......similar testing methods, but judging the results differently. This test apparently judged the differences of the water temperature. I judged the differences in the fan RPM reduction, relative to the intake restriction and voltage. (second test).

In his test the Noctua NF-P12 and the S-Flex 1600 ran almost the same water temperature, at the different voltages. In first test in the computer, my CPU temperatures were slightly in favor of the S-Flex. Like I said.....do the tests again, and the results would differ. And I don't see him using a filter on the intake side, which might explain the results.

But the Noctua certainly doesn't match the hype.

DG · Post by DG » Mon Jan 21, 2008 6:07 am

How's the noise of the noctua and scythe slipstream, compared to a nexus fan, at the same rpm?

Bluefront · Post by **Bluefront** » Mon Jan 21, 2008 6:22 am

I've never heard a nexus 120mm.....but the YL 120mm which is the same as a Nexus but turns faster at 12V, sounds similar to the two you mentioned, when all three are running the same RPMs.

That's using my own ears.....no electronic measurement tools.

Again....I'm willing to preform another test comparing the Noctua, the S-Flex, and the Slip Stream, if someone can propose a test method where the Noctua could conceivably perform much better.

WEKS · Post by **WEKS** » Fri Jan 25, 2008 6:06 am

I always figured the NF-P12 were the "best" because, while not cooling as well as others (but still within just a few degrees), they did it a much lower noise level. And this noise level is what we people here at SILENTpcreview should prioritize, right?

Bluefront · Post by **Bluefront** » Fri Jan 25, 2008 2:01 pm

There certainly is a limit to how quiet you can get when using a fan. The limit depends on the parts you are trying to cool, and how hot you want the thing to run. The idea is to find a quiet fan that pushes enough air through your system, to meet your temperature limit.

In order to get enough airflow, I would have to run a Noctua S12 at a high voltage in the setup I'm using.....making it noisy.

In theory, the P12 could push more air than an S12, at a lower voltage, and possibly quieter. That's what this sort of fan test is all about. It's not about finding the quietest fan alone....that's easy.

cltang · Post by **cltang** » Sun Jan 27, 2008 12:43 am

I know this may be a little paranoid (and silly :P) concern, but Noctua's SSO-bearing seems to contain an electro-magnet.
I was wondering whether this magnet, as well as other fans claiming "magnetic bearings" (are they the same type?) might possibly cause
any sort of electro-magnetic influence on other components' signal integrity?
(Say, if they're placed near a HDD, etc.)

Bluefront · Post by **Bluefront** » Sun Jan 27, 2008 1:58 am

I suspect any magnetic effects of this sort of bearing are so minute as to be a non-issue with a fan. Any evidence to the contrary is lacking....

WEKS · Post by **WEKS** » Sun Jan 27, 2008 8:47 am

I happened upon this review of the P12 just now.

http://www.hardwarecanucks.com/forum/ha ... eview.html

Where they test it against, among others, a Yate Loon D12SL-12, on a Ultra-120 Extreme and a Black Ice Extreme single fan radiator.

kike_1974 · Post by **kike_1974** » Mon Jan 28, 2008 7:10 am

WEKS wrote:I happened upon this review of the P12 just now.

http://www.hardwarecanucks.com/forum/ha ... eview.html

Where they test it against, among others, a Yate Loon D12SL-12, on a Ultra-120 Extreme and a Black Ice Extreme single fan radiator.

In that review they compare only CPU temperatures at default fan speeds.

There is no mention to noise... (well... there is, but only subjective) so I can't make any conclusions from that. The only fan in the comparison with a similar speed to the P12 is the Yate Loon, the others are slower, so worse temperatures were already expected with them...

I think it would be nice a review comparing fan performance at the same noise level (go SPCR, go

).

WEKS · Post by **WEKS** » Mon Jan 28, 2008 7:50 am

kike_1974 wrote:
WEKS wrote:I happened upon this review of the P12 just now.

http://www.hardwarecanucks.com/forum/ha ... eview.html

Where they test it against, among others, a Yate Loon D12SL-12, on a Ultra-120 Extreme and a Black Ice Extreme single fan radiator.
In that review they compare only CPU temperatures at default fan speeds.

There is no mention to noise... (well... there is, but only subjective) so I can't make any conclusions from that. The only fan in the comparison with a similar speed to the P12 is the Yate Loon, the others are slower, so worse temperatures were already expected with them...

I think it would be nice a review comparing fan performance at the same noise level (go SPCR, go ).

During the listening tests I was continually amazed that this fan was operating at 1300RPM while maintaining a smaller acoustical footprint than the Ultra Kaze which operates at 1000RPM

Just because it's subjective doesn't mean it is any less true. Sometimes, if not most often, pure decibel readings don't tell the whole story. But, non the less, I'm waiting for a updated fan guide from SPCR with the P12 included.

Bluefront · Post by **Bluefront** » Mon Jan 28, 2008 8:14 am

WEKS.....I think you're wasting your time. AFAIC....the best you can hope for is an airflow against resistance about equal to other much less expensive fans, that are just as quiet. Nothing like what you might suspect after reading the hype.....

WEKS · Post by **WEKS** » Mon Jan 28, 2008 11:26 am

Well, I like discussing/debating hardware, regardless if I'm "wasting" my time or not.

Also, a hell of a lot reviewers seem to agree with the "hype". In fact I've yet to read a review that has anything bad to say about them, expect of the high price and color choice.

Now I'm considering of buying one or more these, so the question is: Trust 10+ reviewers with their reviews on various sites, or a forum poster. Your thread here made me second guess my decision, which I welcome, because that usually leads to a better decision (so please if you do more test or something, do post it).

Now here where I live the S-Flex series fans are just 1â‚¬ cheaper than a P12, so I'm leaning towards a P12 (if it wasn't for the P12 I'd go for the S-flex, no doubt).

Felger Carbon · Post by **Felger Carbon** » Mon Jan 28, 2008 11:42 am

WEKS wrote:Well, I like discussing/debating hardware, regardless if I'm "wasting" my time or not.

If the discussion is not based on empirical data, it's so much hot air. We got lotsa people wanting answers about backpressure, a goodly number (including yourself) debating, and a single-digit number of experimenters.

One of those experimenters - me - is batting .000 right now on backpressure. Bluefront is fighting the good fight, but by basing his experiments on filters his results are irrelevant to 90 to 99% of SPCR readers. I also wonder about RPM change as the appropriate metric, but remember my batting average on this subject - Bluefront is much more likely to be right. I understand MikeC has taken a shot or two at this, but not on a high-priority basis, it would seem.

Anybody know anybody else trying to gather actual empirical data?

WEKS · Post by **WEKS** » Mon Jan 28, 2008 11:55 am

At the end of the day though, I think the only thing I will really trust, is a SPCR article.

Bluefront · Post by **Bluefront** » Mon Jan 28, 2008 12:26 pm

FC....Over the years I have tried several different methods to judge restriction effects on fan output. And I don't think it should matter if the restriction is on the intake or the exhaust side of a fan. The first method I used involved CPU temperature change when a restriction was placed in front of a CPU fan intake. I kept the voltage constant, and watched the temp go up....which of course points to a reduction in CFM, caused by the restriction.

That works, but the CPU temp number results are too close to judge a whole lot. I needed something that would return a wider number difference. I can read RPM changes down to one, down to maybe 400rpm. And I have already used rpm drop to judge various filters with respect to restriction. That works very well. This last test is the same....except instead of changing filters, I just changed fans. The results are the same....an rpm drop results from the restriction. And I can judge the difference.

IMHO....you can use this test result to also judge how a fan would perform when blowing through a tight radiator or set of fins. I am open to suggestions how to judge resistance effects any better than my test. The first set of tests was preformed in a real computer. I saw what happened....and came to a conclusion accordingly.

KenAF · Post by **KenAF** » Thu Jan 31, 2008 3:47 pm

Do I understand correctly that you've only tested resistance on the intake side?

How about testing with resistance on the opposite side, as you would get blowing through a large heatsink with closely spaced fins?

mattthemuppet · Post by **mattthemuppet** » Thu Jan 31, 2008 6:30 pm

just to add a note about the DS12SL-12 I have - the lowest I can get it idling using Speedfan is ~570rpm, which is inaudible at my ambient noise level (pretty quiet after 10pm). A DS12SM-12 set at it's lowest level using Speedfan was perhaps a smidgen noisier, but not a reliably detectable difference. no rpm data I'm afraid as it was a 2 wire fan.

The DS12SL-12 becomes audible over background at ~750rpm, but seems to cope okay as the exhaust on a intake filtered case (idle ~83W, gaming ~95W, stress test 106W), though I had to increase it from 570 to ~900rpm when ambient temps reach 35C.

Interesting tests, though the differences in fan order between the PC test and free air test suggests either that the PC adds extra/ different impedences or that the difference in filter area between the 2 tests changes how well the fans cope with the impedence. Is it possible to make a free air filter set up that has the same area as the filter used in that case? That would remove one potential source of variation at least.

Bluefront · Post by **Bluefront** » Fri Feb 01, 2008 2:16 am

Humm....I'm not at all convinced it makes much difference what side of the fan the airflow resistance is placed, as long as the resistance is a sufficient distance from the fan. If you put the resistance very close to the fan's exhaust side, you get back-pressure effects which are different from resistance effects.....IMHO. Now if I had a good/accurate way to tell the two apart, a test might be worth it......except. I lately try to avoid hanging a fan directly on a heatsink, blowing into it. It always sounds noisier to me. I'll hang a fan on a heatsink blowing outward.....if possible.

If you place a filter directly on the exhaust side of a fan, the airflow is terrible, compared to the same filter on the intake side. I can testify to this. That's back-pressure's effect on airflow. The other thing.....I am mostly concerned about a fan's ability to handle resistance on the intake side (filters). That's why I ran this series of tests.

In the very first tests, I had the fan installed in a case, sucking through a filter.......this is the normal way I would run the thing, and is a good test of a fan's ability to handle resistance on both sides. The Noctua failed this test, compared to three other normal fans, with respect to airflow.

The P-12 is a disappointment in other ways also. I cannot measure noise....other than by ear. The Noctua also fails the "ear" test when confronted with resistance, compared to other fans. IMHO.

KenAF · Post by **KenAF** » Sun Feb 10, 2008 2:23 pm

Bluefront wrote:Humm....I'm not at all convinced it makes much difference what side of the fan the airflow resistance is placed, as long as the resistance is a sufficient distance from the fan. If you put the resistance very close to the fan's exhaust side, you get back-pressure effects which are different from resistance effects.....IMHO. Now if I had a good/accurate way to tell the two apart, a test might be worth it......except. I lately try to avoid hanging a fan directly on a heatsink, blowing into it. It always sounds noisier to me. I'll hang a fan on a heatsink blowing outward.....if possible.

If you place a filter directly on the exhaust side of a fan, the airflow is terrible, compared to the same filter on the intake side. I can testify to this. That's back-pressure's effect on airflow. The other thing.....I am mostly concerned about a fan's ability to handle resistance on the intake side (filters). That's why I ran this series of tests.

Thanks.

It is my understanding that Thermalright recommends that you blow through their heatsinks, so I figured that back-pressure effects would be worth testing. Most people also have a second exhaust fan [relatively close to their heatsink] exhausting air out of their case.

It seems to me the logical test would be to set your fan controller to adjust your fans to maintain a certain temperature on the CPU. Then you would compare the RPMs and perceived sound of each fan at that temperature.

Kuroimaho · Post by **Kuroimaho** » Mon Feb 18, 2008 6:51 am

Thank you very much for this test Bluefront.
Couldn't decide whether to go for the Kama flow or the Noctua which costs twice as much this settled it.

Farinorco · Post by **Farinorco** » Thu Feb 21, 2008 9:33 am

Wow, I've seen this excellent test while looking for Scythe SlipStream tests here in the forum (thank you very much for your awesome work here, Bluefront!), and as I see it, the second test is perflectly coherent and confirms the first one results, even when it could seem otherwise at a first look. Of course in my interpretation I'm making some asumptions that could but don't have to be right, so I hope any of you more expert silencers can back up or refute what I say...

The main thing that, in my opinion, helps both tests to make sense together is the fact that the spinning speed of a fan is a huge factor in the effective pressure of the fan, so it's in its ability of handling resistance (I've read by there than the pressure of a fan changes with the square of the speed, but I don't know exactly where and hence its reliability...). Also blade design, motor and so, but a fan at a certain rpm will handle much better the resistance than the same fan at half the speed, which will be much more prone to suffer variations in its performance with an increasing/decreasing in resistance. Neither one or the other test makes constant that speed variable, so it has to be considered when interpreting the results.

So, what I see in the results of the first test by Bluefront, is the S-Flex doing the best here. Yate Loon is showing so little variation with the resistance increasing as the S-Flex, but it is spinning much faster. Ignoring the higher minimum speed of the Yate Loon fact, even without the filter, it has to spin much faster than the S-Flex to dissipate the same amount of heat. Then, even when spinning faster (1175rpm VS 1020rpm), both are reacting similarly to the increase of resistance (so are putting similar pressure), so I'd conclude that at equal speeds (and similar noise) the Yates have much worse pressure than S-Flex. Then, Noctua is getting worse results than S-Flex, but probably better ones than the more speedy Yate Loons. And surprisingly, the SlipStream seems to do it more or less so good than the S-Flex at load (the idle results are tricky to read because of the slightly improved temps and specially the apparent higher minimum speed at same voltage, Slipstreams seems to have less than usual reactivity to undervolting).

In the second test, the fans are running at full speed and 5v, so the advantage of the higher speed of the Yate Loons, obtained due to its worse cooling performance, is decreased (at 11.5v, Slipstreams are increasing its speed in ~22%, Noctuas ~26% and Yate Loons only ~14%, comparing with the load first test) enough to put it in the worst place amongst the four. The S-Flex tested here is the 800rpm one, and even so, it manages to stand firm at more or less the same pressure level than the 1200/1300rpm Slipstream/Noctua, so I suppose that if the 1200rpm was tested instead the 800rpm one, it would win again.

So, what are my opinions on these fans based on Bluefront's test results?

Yate Loon: the worst of the four. I've seen (mainly at XtremeSystems) that there are different models of Yates for each label/id. The more common ones are mediocre (not bad) performers, and the more rare are top notch performers, near the S-Flex. There are physical and design differences (that I don't know) so it seems not to be a quality control problem. Here at Spain I can't see any shop selling Yates, so I haven't seen anyone of them aside internet images. It seems a little risky to buy Yates if you are looking for a top notch performer fan... (except if you know exactly where to buy the good ones, people by there says that Petra's Shop usually -not always- have the good ones for any reason, while most of the other shops don't)

S-Flex: no surprise here, this fan has a reputation of being one of the best high pressure fans in 25mm thickness...

Slipstream: the greatest surprise (to me) in this tests. If I believe the "mediocre" (not "bad") Yate Loon performance theory, this fan do have good pressure, even if it's worse than S-Flex one. Better than Yate Loons and the new Noctuas, or it seems so. It's a known fact that in free flow conditions, it's moving more air than S-Flex (and any other 120x25mm fan by there) at the same rpm, so if it has worse pressure than S-Flex, I suppose that SS it's gonna be a better performer until a certain level of resistance for a given speed, and then S-Flex it's going to turn the tables on that and start being a better performer. By the test results, it seems that by chance Bluefront's first test system is around this point, at least at ~1000rpm, cause both are performing similarly well on this test. I'm also noticing (not this thread alone) that Slipstreams seems to lower speed much less than usual when lowering its voltage. Looking at its design and low power consumption I suspect that the Slipstreams need little strength to spin their blades, and rely on inertia/air movement more than usual (but this is pure speculation), which can be one reason to its low mechanical noise, and a problem if you want a wide range of options by speed controlling the fans...

Noctua: following the previous line, this fan would have a good but not top notch pressure performance. If it's true that these Noctuas sound around the same as other top performer fans when spinning slightly faster, as some reviews say, it could end up being a better performer in high resistance conditions in terms of noise:performance, due to its increased speed at same noise level, even if it's not in terms of rpm:performance (in the end, a merely statistical concern, as I'm more interested in the noise it makes while cooling my rig, than the rpms shown in a control software). I can't give an opinion on this though, as I've not heard/seen one of this in action, and there are few probabilities that I pay 20â‚¬ for a single fan to know it...

After all this I don't know what fans I'd must buy to cool my future watercooling radiator yet...

Wow, I think I've already bored you enough with this...

Felger Carbon · Post by **Felger Carbon** » Thu Feb 21, 2008 1:35 pm

Farinorco wrote:(I've read by there than the pressure of a fan changes with the square of the speed, but I don't know exactly where and hence its reliability

I'd conclude that at equal speeds (and similar noise) the Yates have much worse pressure than S-Flex.

S-Flex: no surprise here, this fan has a reputation of being one of the best high pressure fans in 25mm thickness...

I suspect that the Slipstreams need little strength to spin their blades, and rely on inertia/air movement more than usual (but this is pure speculation), which can be one reason to its low mechanical noise...

Some facts: the energy of a moving column of air (as produced by a fan) is proportional to the cube of the air velocity. Mechanical energy is pressure times distance (check out the definition of "horsepower" sometime). Distance is directly proportional to air velocity, so that means (to make the math come out even) that pressure must be proportional to the square of the air velocity, and the air velocity produced by a fan is, of course, proportional to its RPM.

It's also a fact (please take my word for this) that the air resistance to flow (when passing through an impedance such as an air filter) is proportional to the square of the air velocity. So both pressure and resistance are proportional to the square of the RPM, meaning that the pressure squared term provides no advantage; that's just how the math works.

Because the Yates are among the lowest-pitch fans and the SFlex among the highest pitch, I'd strongly suspect the Yates provide higher pressure at a given RPM. Don't have any numbers for that, though. Anybody?

Fans produce noise in two ways: the "whoosh" of the air, which is proportional to the cube of the air velocity, and the blade/bearing noise, proportional to the cube of the RPM. At a given airflow, the "whoosh" of the SFlex and Slipstream are the same, but the RPM of the Slipstream is lower, which is why the Slipstream is quieter than the SFlex in an unimpeded airflow situation. Which is better in an impeded situation is what Bluefront's experiment was about.

Now that everything is perfectly clear, I'll go back to sleep. It's snowing here! Drat.

Bluefront · Post by **Bluefront** » Thu Feb 21, 2008 1:54 pm

Frankly I'm lost here with all the test results, and trying to make anything out of it. I've been using the SlipStream 1200 in my latest setup (which was what the tests were all about). The idea of course was to use the quietest fan that kept my setup at a temperature point (completely arbitrary) that satisfies me. The poor Noctua is in a box somewhere.

When I actually ran the Noctua in my mostly completed setup, it responded badly to the LIS2 PWM control....nasty noises. The Scythe fans were not so bad, but I still had to use M-Cubed attenuators to quiet them down. Amazingly.....the Enermax fan from the test, was quiet as a pin using the PWM controller. But it does not blow enough CFM for the setup.

So there's other things to consider when picking a fan.....I knew that of course. But the Noctua was an even bigger disappointment than I first imagined. You just never know.....

Koolpc · Post by **Koolpc** » Thu Feb 21, 2008 3:20 pm

I've tried lots of different fans and to be honest the Noctua is the best one so far. Quiet and cool.

Farinorco · Post by **Farinorco** » Fri Feb 22, 2008 3:03 am

Felger Carbon wrote:Some facts: the energy of a moving column of air (as produced by a fan) is proportional to the cube of the air velocity. Mechanical energy is pressure times distance (check out the definition of "horsepower" sometime). Distance is directly proportional to air velocity, so that means (to make the math come out even) that pressure must be proportional to the square of the air velocity, and the air velocity produced by a fan is, of course, proportional to its RPM.

It's also a fact (please take my word for this) that the air resistance to flow (when passing through an impedance such as an air filter) is proportional to the square of the air velocity. So both pressure and resistance are proportional to the square of the RPM, meaning that the pressure squared term provides no advantage; that's just how the math works.

Wait! Are you saying that rpm have nothing to do with tolerance to impeded situations?

Don't understand the following as a try to refute you, but a try to understand well all this matter, as is important for me and my next pc build:

I understand "pressure" as the ability to overcome resistance without reducing much the airflow created, so if a fan has more pressure, its airflow will show more tolerance to impedances, so if you put a filter in its way, its airflow will be less reduced than if it would have less pressure. Maybe that's my former mistake, so you know, if it's so, correct me please.

As far as I understand you (correct me when I'm wrong please) the pressure of a fan grows with the square of the velocity (and rpm), but resistance to airflow created by the impedances grows in the same proportion, making the capacity of a fan to overcome resistance completely independant from its rotational speed.

Then I simply can't understand why when using a highly restrictive watercooling radiator and a little restrictive one, with slow (low rpm) fans, the non-restrictive gets a much superior performance, but as you increase fan rpm the difference narrows, and there is a point (a certain fan rpm) when the more restrictive radiator begins to get a better performance, and from this point, the difference begins to widen in favor of the more restrictive radiator. It seems to me that as you increase rpm (hence velocity) the capacity of the fans to overcome the impedance of the radiators IS growing and then becoming the resistance made by the radiator less and less important to the end results... What am I doing wrong? I need to understand all this.

What I like more about this kind of forums with serious and informed posters is that I can learn a lot about true facts and no urban legends and so

. So thank you for your clarification, Felger...

EDIT: Ehm, another little question which answer could be yes or no with a little little explanation if no, thinking about you've said, if more distance means less velocity, and less velocity less resistance for a given impedance, does it means that if a put a given impedance, lets say a filter, far away from the fan, it's going to be less restrictive than if its near it? Even if the way from the filter to the fan is a closed space (like a duct)? Would be interesting to know it for my mdf tower design... thanks again...

Bluefront · Post by **Bluefront** » Fri Feb 22, 2008 3:41 am

Farinorco.....realize that SPCR people judge things with an emphasis on noise. So my tests involved trying to find the fan that pushed the most air at the least RPM, when sucking through a filter resistance.

To come to a conclusion, I judged temperature differences in the first tests, and RPM reductions in the final tests. These two sets of results are comparable, but not completely so.

Then there's the actual resistance thing....and if it makes any difference if it's an intake resistance (the filters I use), or an exhaust-side resistance, or a combination of the two, when trying to pick out the best fan from all the testing.

The best fan would be completely silent in it's operation, and cool everything better than any other fan. Obviously just not possible.... but doing the tests certainly narrowed down the field of candidates. At very low rpms, these fans sounded pretty much the same, but there were differences that I tried to measure. At the very least I determined that the Noctua was not much better (if at all), than it's close competitors.

Farinorco · Post by **Farinorco** » Fri Feb 22, 2008 3:57 am

Bluefront wrote:Farinorco.....realize that SPCR people judge things with an emphasis on noise. So my tests involved trying to find the fan that pushed the most air at the least RPM, when sucking through a filter resistance.

Yeah, me too. And that's why I find your tests so interesting to me, because I'm trying to build a system with so many or even more resistance than yours (I don't think that sucking through a filter would be much harder than sucking through a radiator... specially if previously have had to pass a filter too

), and that's why I need to understand the effects of having low rpm fans combined with high resistances...

And over all, why I need to understand the results you obtained in your tests, because those tests can be extremely useful to me

One thing I know for sure is that I'm not gonna buy a Noctua

. At around 20â‚¬ each one, probably I wouldn't done it even if they were the best fans around...

jhhoffma · Post by **jhhoffma** » Fri Feb 22, 2008 7:52 am

Farinorco wrote:I don't think that sucking through a filter would be much harder than sucking through a radiator... specially if previously have had to pass a filter too

What radiator are you using, Farinorco? If you haven't installed one yet, maybe you should look at one with the greatest fin-spacing you can get (the so-called Ninja approach) where the impedance on a fan will be the least.

Personally, I don't think you'll notice much of a difference (in noise or cooling performance) between S-Flex, SlipStream, or Y/L (at equivalent speeds) in your application.

tutu · Post by **tutu** » Sun Mar 02, 2008 11:17 am

How does a NF-P12 compare to the noise levels of a NF-S12 800 or 1200? I'm planning on getting three in a P182 (CPU, Exhaust, Intake) running on a low setting.

I might use 2 if necessary (either use passive cooling or negate a intake fan).

Post by **NeilBlanchard** » Sun Mar 02, 2008 11:35 am

Welcome to SPCR!

I prefer the Scythe Slipstream 800RPM fans -- they are the quietest fans I know of (except for the 500RPM).