Improvement? Clearly. Arctic had been gradually refining its older “P” series fans, but it seems the limits were only truly surpassed with an all-new seven-blade aerodynamic design. Its first carriers are the P12 Pro (A-RGB) fans, which are likely to become popular again. Airflow per noise unit is higher than before—once more elite, even with otherwise inexpensive fans.
Results: Airflow through a plastic filter
Why is a value sometimes missing for a fan? There can be several reasons. Most commonly, it’s because the fan couldn’t be regulated to the target noise level. Some fans have higher minimum RPM (or they spin slowly but have an excessively noisy motor), or they are simply slower models that can’t reach higher decibel thresholds. Results are also omitted from graphs if the impeller scrapes against the mesh of a nylon filter. In such cases, we classify the combination as incompatible. And of course, a zero in the graphs naturally appears in situations where we record an actual 0.00. This often occurs at extremely low speeds when obstacles are present or during vibration measurements.
New to reading the sound profile graphs. Is it safe to say that the Momentum 12 has a better sound profile since the graph is smoother, and the dominant high frequency is lower? Although, the lower frequencies are much louder on the Momentums
Note: On Page 28, the 45 dbA table is missing (ID 3700)
The term “better” — or potentially “worse” — when referring to the acoustic profile will always lead to subjective conclusions. This is because people perceive the same frequencies differently. For this reason, there’s no simplified or universal answer.
That said, a higher (and more prominent) tonal peak is always the result of some technical imperfection, and ideally, these peaks should be as small and as evenly distributed across the sound spectrum as possible. When this is the case, we can say that fan A is acoustically more pleasant than fan B (which exhibits more pronounced tonal spikes).
However, again, it all depends on which frequencies you’re sensitive to. If the tonal spikes fall outside your hearing sensitivity range, you may not notice them at all — or you might be completely “deaf” to them. 🙂
And yes, thank you for pointing out the mistake — the table has been corrected. 🙂
Thanks for the test!
As always, much to praise, little to criticize.
SADLY, the critique part is key and BIG for me, and that is the much debated topic about limiting testing scenario to only 31 dBa being the lowest.
Same as with hwbusters.
At the latest Noctua NF-A12 test, someone said 31 dBa here at hwcooling is comparable to 39 dBa testing scenario at hwbusters.
That is correct?
https://hwbusters.com/best_picks/best-case-heatsink-radiator-fan-picks-hardware-busters/
“Most fans won’t even push any air at such low speeds/noise”
He commented at May 27, 2024.
It’s either too much testing needed to be done, or hwbusters arguing something something about “focusing on performance at adequate noise level.”
I conclude your peoples ears are bad. Simple as that.
Your definition of quiet =/= not mine.
I can still hear a 1st generation Noctua NF-A12 from 2 meters away spinning at 800 – 900 rpm (without obstruction); be it the air being pushed that is heard or the fan blades.
And I’m already 40 years old.
You people damaged your ears too much with techno and clubs in the 20s or what? 🙂
If I want “No noise” at evening or nights, than it means NO noise. Absolutely zero. it should be as quiet as the room noise itself.
I’m not paying 20 – 40 for a single fan to care about fan tests being done, with fans tested at RPM starting from 700 – 900 only, which is the case with all of the test here at hwcooling or hwbusters.
700 – 900 RPM is AUDIBLE and not quiet, especially if 3 fans are spinning simultaneously at the front.
Do you testers at both websites don’t understand your audience or what?!
Me (and I guess most others), don’t come here to care about fan performance from 700 – 2000 RPM.
I care about 300 – 1000 (maximum). In case of the new Noctuas G2 I might go up to 1200, because they are so quiet and the frequencies so low.
Why would I spent my time reading the tests from the last two websites available (offering tests done at amateur/professional level), and care about 31 dBa tests?!
Please reconsider your testing.
300 – 1000 RPM tests matters. I care about absolutely no-noise or whisper quiet fans, and how they perform at that level.
The comment on the previous thread said the opposite to be fair, “HWBusters (Cybenetics) 20 dBA level is roughly comparable to HWCooling 39 dBA level” where as you’re assuming “31 dBa here at hwcooling is comparable to 39 dBa testing scenario at hwbusters (Cybenetics)”. That comment also mentioned ” HWC 45 dBA level is comparable to HVB 25 dBA level (with HVB 25 being a bit quieter)”. Will have to wait for the G2 tests to be released over there but atleast for the G1, 20 DBA is 1454 RPM and while they don’t show noise levels at lower rpms, they do atleast show the PQ Charts for every 10% of the fan rpm range.
While I do agree with you that some tests at a lower noise level would be nice, in the case of the A12 G1, the 31 DBA test from HWC would already be ~30% RPM range (~750rpm 18 cfm) and using the scale above around ~8 DBA under HWB’s 20 DBA level so I’m not sure how much more they can get in terms of upgrading their test suite in scalability.
I for one, greatly appreciate their tests at the medium/higher ranges with my fan curves peaking out at 25 DBA equivalent from HWB for my 13700K (250W) + 4090 but I can see your point as well. I’m just part of the audience you assume isn’t around 🙂
Glad you guys finally got to release the P12 pros tests though, was wondering when arctic would release y’alls embargo on it!
I stand corrected, there is a speed percentage-noise chart over on cybenetics for the G1. https://www.cybenetics.com/evaluations/fans/4/
Will stand to see if there is one for the G2, but atleast for the P12 pros, this information does exist aswell.
https://www.cybenetics.com/evaluations/fans/175/
10% PWM @ 737 RPM is 1.9 DBA on their scale, and these fans do only start spinning at 600 RPM. Using the same math as before, the 31 DBA (HWC) tests would be around the slightly higher than 10 DBA on the HWB scale
1) HWC does not have the equipment to reliably and repeatably measure such noise levels.
2) …after installing a fan in a case, it is always noisier than if it were freely hanging in space (case ~ loudspeaker enclosure, mainly around 200-500Hz). 😛
3) …if one sound source has 30 dB, then 3 such close together will have a noise level of 34.8 dB (10*log10(1000)=30; 10*log10(3*1000)=34.8) 😉
4) I’m 46 years old and I can hear the blinking LED on my monitor or the ticking of a watch, so I sleep with earplugs… but if something can’t be measured with the given equipment, then it can’t be done, … 🙂
HWCooling’s 31dBA level corresponds to A12x25 G1 spinning at 787 RPM. Can you hear that too?
“At the latest Noctua NF-A12 test, someone said 31 dBa here at hwcooling is comparable to 39 dBa testing scenario at hwbusters.”
This is completely wrong. HWBuster’s 20 dBA level (which is their lowest normalized noise level) is essentially equal to HWCooling’s 42 dBA level (which is their second highest) – for example, for A12x25 G1, HWBuster 20 dBA = 1476 RPM, HWCooling 42 dBA = 1475 RPM. This means HWBuster do not test anywhere near a silent noise level, whereas HWCooling tests at much lower noise levels. In fact, I have yet to find a review outlet that tests fans at lower noise levels than HWCooling.
You can check page 13 – Results: Speed on this review to see which noise level you should be looking at. 31 dBA would be 787 RPM for A12 G1 and 903 RPM for G2, which should be close to what you want; likewise, if you’re looking for noise levels similar to A12 G2 @ 1200 RPM, you should be looking at 36 and 39 dBA results (which correspond to 1120 and 1289 RPM for A12 G2). Of course, you should also look at the frequency spectra (pages 28-31) which would indicate how the fan sounds, any potential irritating frequencies etc.
The difference in measured SPL level between HWBusters (Cybernetics) and HWCooling is mainly due to different distance between the microphone and the device under test.
One can check use a simple backside-of-a-napkin calculation by noting that in a free field the SPL is increased by 6 dB for every doubling of distance (assuming all else equal).
HWCooling uses a distance of 15 cm while HWBusters uses 116 cm (calculated from figure 1 in https://www.cybenetics.com/data/Cybenetics%20Chassis%20Test%20Protocol_en.pdf).
So a HWC SPL of 31 dBA is ca 13 dBA at 120 cm, so a difference of 18 dBA. As you noted a difference of only 39-31=8 dBA is not reasonable, but one of 42-20=22 dBA is much more reasonable and can be explained by different test setups.
Well, it’s clear that 31 and 39 are some kind of misunderstanding… there is no such level on HWB 😀 … well, English is not the native language of many on this forum, …mistakes happen 😉
It was not my intent to impugn this excellent site.
you didn’t do that in any way or form 😀 …
I was just trying to guess how user 300 RPM for life could have gotten some information wrong 😉
Since, there was mentioned the Sound propagation – the Inverse Square Law (by Plutten)…
I think it’s time for some mathematics 😛 (you are free to correct me; I am no expert in the field of acoustics or mathematics) 😀
1) distance difference
SLdiff = 10*log10(Power/(4*3.14*R1^2)/Power/(4*3.14*R2^2)=
=10*log10(1/R1^2/1/R2^2)=10*log10(R2^2/R1^2)=10*log10(R2/R1)^2=
=20*log10(R2/R1)
R1 – according to p. 6 of the Cybenetics Fan Test Protocol, Aris has 125 cm
R2 – HWC is 15 cm
SLdiff=20*log10(15/125)= -18.42dB
2) noise floor difference
Well, we don’t know the HWC noise floor, but we know it is below 30dB. Hence, we know that the HWC 45 dBA is far enough from the floor. Because 10*log10(31622.8)=45; 10*log10(1000)=30; 10*log10(31622.8+1000)=45.14 ~ 45.
So, the HWC 45dBA noise level should be equal to the HWB 26.5 dBA.
But for the lower levels we can just estimate the difference that would grow from 18.5 to 21.5 on the lowest HWC levels.
However, the HWB levels that are lower than 15 are also spoiled by proximity to the 5.3bB noise floor.
3) Same conditions
Well, they both have different chambers, but luckily Ľubo chose 15 cm, and his chamber seams to be at least 1m^3 big and has the appropriate dumping material and structure. If the fan is in the center, the echo must travel .5m to the wall and .5m back and it is dumped on the way, so the additional echo noise should be around .5dB, i.e. we can almost omit it. … so, we estimate the Bulgarian constat (no pun intended) in the favor of bigger chamber ~ 1db 😀
Therefore, the difference may vary from 19.5 to 22
The final qualified estimate:
HWC 33 ~ HWB 11 (but both levels are messed up by close noise floor proximity)
HWC 36 ~ HWB 15
HWC 39 ~ HWB 19
HWC 42 ~ 22
HWC 45 ~ 25
Here is BQ Silent Wigs Pro 4 120 for comparison:
https://www.cybenetics.com/photos/fans/71/Fan_Speed_Noise.png
HWC33 = 978, HWB10.1 = 1025
HWC36 = 1140, HWB15.4 = 1177
HWC39 = 1270, HWB20 = 1337
HWC42 = 1439, HWB22 = 1477
HWC45 = 1627, HWB25 = 1630
HWC60.1 = 2801, HWB40 = 2800
😉
The answers to most of your questions have already come up in our readers’ comments, but I’d like to clarify a few things. For example, the Akasa Alucia SC12 fan runs at 368 RPM in our “31 dBA” mode. Some fans aren’t even measurable in this mode because they simply stop spinning – they’re too loud to reach such a low noise level. This test mode is extremely quiet, right at the threshold of audibility, and I truly believe we understand what our readers are interested in. 🙂
Here are two simple demonstrations of the fallacy of hwbusters argumentation and giving certifications for fans airflow/performance, only as low as 20 dBA.
Many people including myself have asked going down to 15 and 10 dBA, because it matters.
https://www.cybenetics.com/evaluations/fans/103/
https://www.cybenetics.com/evaluations/fans/140/
https://cybenetics.com/index.php?option=fan-performance-database
For the “be quiet! Silent Wings 4 140mm” pro and non-pro version, they only gave a Gold certification for airflow, or no certification at all.
An uneducated reader in that matter, or someone not spending 1 minute time looking carefully at the PQ-charts or longwin, might now think “Whooo, Noctua, Diamond certification, so much better. Bad Gold certification”.
But spending time looking at the PQ reveals the be quiet fans in fact have slightly higher or equal airflow/noise ratio up to 60 %!
Only starting 70 % and above new Noctua fans start getting better, but only ~ 10 % at best.
Same fallacy comparing the Arctic P14 Max with the Noctuas.
https://www.cybenetics.com/evaluations/fans/106/
The fabled & praised P14 Max get a platinum airflow certification … but that’s only for 20 dBA again.
The obvious issue: That’s 20 dBA achieved at 1361 RPM and that is NOT quiet, neither no-noise.
The lower the RPM the worse the P14 max gets.
At 900 RPM they deserve a GOLD certification for airflow at best, and at 600 RPM it’s bascially SILVER.
Hwbusters didn’t even mentioned that the motor with those arctic is bad (noise etc.) and only starts being viable at > 900 RPM.
I was almost fooled buying these P14 Max due to their certification, only realizing last second how worse they are at 300 – 1000 RPM, were it matters being quiet.
But ye… let’s fool the readers with a platinum airflow certification.
So yes, they do test down to 300, 400, 500 … RPM if possible (if PWM or DC allows), but they do not rate anything below 20 dBA.
So same problem with here at hwcooling going down to 31 dBA only, in fact you testers here don’t even test at such low RPM.
Reconsider and start thinking about your core audience please.
People come here for high quality tests, not McDonalds fast-food like tests, which concludes they care about high quality tests, which in most cases means being whispers quiet, no-noise, all at good performance, or a good ration of them.
Think about it logically.
If I wouldn’t care about fans being quiet, no-noise, I wouldn’t visit this website, but would visit all the other 98 % of hobbyist-level tests at other websites, were they mostly tests at 1000 RPM and max-RPM only and not even remotely that good.
I can only conclude majoriy of people nowadays have damaged ears even in their 20s or 30s, and thus saying things like “I can’t hear the 40 € Noctua fans spinning at 600 RPM anymore at 50 cm”, which might explain why you testers don’t focus at lower RPM testing.
At home, don’t listen much with headphones, especially in-ears. They damage hearing the most!
All studies conclude this.
Speakers is the only way to go.
Have a good day and thanks for the testing.
–“…Many people including myself have asked going down to 15 and 10 dBA, because it matters…”
I personally also had email communication with Aris regarding adding 15 dBA testing (which would correspond to somewhere around 36 dBA on HWC) and he was against it.
As M writes above, HWC offers the lowest normalized level (31 dBA). …I think HWC is the top-notch portal for HW Cooling reviews 😉
–“…I can only conclude majoriy of people nowadays have damaged ears even in their 20s or 30s, and thus saying things like “I can’t hear the 40 € Noctua fans spinning at 600 RPM anymore at 50 cm”, which might explain why you testers don’t focus at lower RPM testing…”
1) Most people can filter out sounds around them, to the point where they say they can’t hear them. Some, like you or me, can’t filter them out that well… like for example, people with Asperger’s syndrome.
2) There is only one fan tester here. His name is Ľubo 😀 . And he tests not by his own hearing but by a quality sound level meter with a minimum level of 30 dBA
Have a good day and thanks for your comments and reactions 😉
I agree with so much with your statement regarding wanting quieter tests, but heavily contest with the statement “People come here for high quality tests, not McDonalds fast-food like tests, which concludes they care about high quality tests, which in most cases means being whispers quiet, no-noise, all at good performance, or a good ration of them.
Think about it logically.” The main result I come to this website for is “Airflow through a thicker radiator”. The fact that they test all these different normalized results in different scenarios is amazing quality in of itself already. When it comes to fast food tests, like you said, I instantly think of those sites that just list the results at max-rpm without any other situational information.
You’re conclusions seem to just be degrading in general, making light work of the results and seemingly everyone else that doesn’t mention the want of the methodologies you provided. I personally do hear my A14 G2’s spinning at 600 RPM inside of my case as front intakes (from my desk w/ my pc on the floor) but it’s just the matter of my noise floor being over that audible sound living in the middle of a city.
You should take a look at the test setups and, to use your words, “Think about it logically.” With a dose of knowledge you can make an informed guess from the measurements that can roughly apply to your own listening environment that most likely will greatly differ from the test setups/environments.
Not all 15 dBA measurements are created equal. It depends on the measurement method, which is influenced by the distance between the noise meter and the measured object, the environment in which the measurement is taken (such as reflections, ambient noise levels, and electromagnetic interference — especially for non-battery-powered equipment). Honestly, you could write a book on the topic. In short, it’s important to understand that higher noise levels under certain methodologies may correspond to lower ones in others. That’s the case here as well. I’d also like to highlight our interview with Noctua, where we explored how different microphone angles can yield different results.
PS: The main purpose of HWBusters/Cybenetics tests is to create a certification that reflects airflow and static pressure in controlled lab conditions. That’s how the author presents it too — it’s not meant to simulate real-world usage. Personally, even though their Longwin LW-9266 is an excellent instrument, I find it hard to draw practical conclusions from HWB’s measurements. But of course, others may see it differently.
Looking at the other arctic fans and tests done, I conclude now begins the usual arctic beta-testing.
2 – 4 iterations until the fan has all defects and flaws ironed out. Looking at the recent history, that would be 6 – 9 months, plus a small time-buffer to make sure all store/online retailers get the latest version 3/4 batch.
😉
@Ľubomír Samák
Is all your testers testing done with fans in the push configuration? No pull configuration with obstructions right in front of the fan? (as e.g. case fan intakes, blowing air inside)
Airflow, push without obstructions/through filters etc. only?
Static pressure, push without obstructions/through filters etc. only?
Please answer.
All case front fans (certain AIO configuration in the front excluded), is a pull configuration (air inside) with either small & solid metal bars right in front the fans (to screw the fan in place), or a nylon/metal filter 1 – 3 cm in front of the fans and may even additional cover in front of it.
I haven’t found a single case I like were there aren’t small & solid metal bars in order to screw the fans in place. It’s lenghty and narrow metal bars.
I don’t know the bigger picture, so I’m just telling you that for certain fans, no pull-testing might be not good and not tell the whole story.
Example: NoiseBlocker eLoop B12/14.
Got them 2016/2017 as case fans, expensive; they are still good nowadays.
Almost every tesers tested them and praised them, especially as case fans … and basically every tester failed, as they didn’t test the PULL-scenario properly, with small obstructions right in front of the fans and/or filters/covers.
Basically, the fans are useless in the pull configuration, even with the slightest solid, small metal bars in front of them; even a nylon filter 1 – 2 cm in front of them.
They start to produce a loud turbine-noise. The engineers didn’t engineer them right; the fan blades protrude slightly beyond the frame and come very close to any fan grilles, metal bars, or dust filters etc.
600 RPM is the maximum, after that they start a disproportionately loud wind-noise.
Had to swap them to the back/top, as push configuration (moving air outside), and since then they are very fine.
Considering buying x3 of those Noctua 140 mm G2 fans, which are (obviously) very expensive (though still somewhat justified), and would be unwise, not knowing how they behave in the pull configuration with obstructions.
I’m thankful for an anwers.
Have a nice day and thanks for all testing.
If I may answer, all the filters are tested in pull configuration (…and even the hexagonal grid if I remember correctly from discussion with Ľubo), and only the radiators are tested in push configuration. Ľubo tries to make it as practical as possible 😉
Hence, because of pull configuration, there are collisions on the unreinforced nylon dust filter with some fans. 😉
Thanks.
@Ľubomír Samák
Correct, yes?
Read the page “Everything changes with obstacles”, and it’s mentioned that filters are used, but not clear to me in what configuration (push or pull with filters in front or back).
–“…but not clear to me in what configuration (push or pull with filters in front or back)…”
yes I had the same problem that’s why I asked him (by email) 😉
Yes, Bufo is right, but please consider my earlier comment too. Determining whether a fan is more efficient in a push or pull configuration isn’t as simple as it might seem. 🙂
I’m sure your questions would be more to the point if you actually read the review. The test setup described in it. 😉
“All case front fans (certain AIO configuration in the front excluded), is a pull configuration (air inside) with either small & solid metal bars right in front the fans (to screw the fan in place), or a nylon/metal filter 1 – 3 cm in front of the fans and may even additional cover in front of it.
I haven’t found a single case I like were there aren’t small & solid metal bars in order to screw the fans in place. It’s lenghty and narrow metal bars.”
Yeah, that is really annoying but it’s not something that is part of the test setup for a fan.
So, either modify your case or add fan inlets to reduce the turbulence noise. For my Fractal Design Define 7 case I hollowed out the default Fractal 140 mm fans and put them in in front of my Noctua NF-A14x25 G2 used for front intake. That substantially reduced the noise.
it’s like magic…on the SK version, user epidot just showed the same solution for FD Torrent
https://postimg.cc/0bQLqLyJ
here is a nice video on the topic from ThermalLeft, who used to contribute comments here (now only occasionally) 😉
https://m.youtube.com/watch?v=WKs7_1gK4Zo
Thanks.
So Plutten mentions fan inlet; is that the same as a fan shroud?
You mention a spacer; doesn’t a spacer reduce fan performance (both air flow and static pressure) as the air isn’t “tunneled” efficiently anymore and leaks on all sides? The bigger the spacer the worse?
Or are these fans (all fans?) so well desined that 1 – 3 cm of spacer gap doesn’t make much a difference?
Watched some air flow-particle simulations videos/pictures, and they usually show that the air moves straight out, as a line, about 5 – 10 cm, snf only after that the air starts to swirl more and more.
@Ľubomír Samák
Would the tester say something please? You are knowledgeable.
The following Reddit thread, with pictures, describes it:
https://www.reddit.com/r/FractalDesign/comments/1kositm/fractal_design_define_7_with_fan_inlets/
As for leaking: These are used as case fans and barely have any gap at all, but one could use a sealant or some duct tape, of course.
— „… doesn’t a spacer reduce fan performance (both air flow and static pressure) as the air isn’t “tunneled” efficiently anymore and leaks on all sides? The bigger the spacer the worse?“
A well-designed spacer doesn’t reduce fan performance — on the contrary, it can enhance it.
Performance would only decrease if the spacer structure isn’t sealed properly and pressure escapes, such as when the spacer is leaky or doesn’t fit well between the fan and the obstacle.
Now to the second point — when a spacer increases fan performance or airflow. This typically happens when the spacer eliminates unwanted microturbulence at the fan’s intake, allowing for more laminar flow, which results in increased airflow at the outlet. In such cases, in addition to improved airflow/pressure, noise levels also drop because laminar flow is quieter than turbulent flow.
The tester has spoken. 🙂 If anything’s unclear, feel free to ask — I’ll gladly answer as best I can.
Thanks for that informative video.
I thank you too… thanks to your link I got to this maniacal approach…
https://www.reddit.com/r/Noctua/comments/1fafbt4/project_manifoldium_ducted_noctuafractal_build/
🙀👍 man oh man…
Very nice video. It’s just a pity that ThermalLeft has been busy with other projects lately, and their last public activity on their YouTube channel was quite a while ago — it’s been over a year since their last video. :/
— „Please answer.“
The push/pull configuration topic is particularly complex. Even if someone concludes that push gives higher airflow, it’s quite possible those measurements were taken at a point with less laminar flow than in the pull configuration. I’m not saying push can’t provide more airflow — but there’s definitely this risk present.
When measuring push/pull, it’s also important to distinguish whether the orientation is relative to the anemometer (or Pitot tube…) or an obstacle. What really matters is setting test conditions that lead to relevant, generally valid conclusions.
@Ľubomír Samák
Please consider adding another minus 3 – 5 dBA test, as I mentioned in my comments.
You do not do fully justice to all these new, good and/or expensive fans, who push enough air even at 500 RPM for certain scenarios, which results that the reader and potential buyer won’t see a big and clear differentiation in quality/performance of those fans.
See my 2nd comment about the arctic P14 Max fan.
I made a mistake with the “31 dBA equaling hwbusters 39 dbA testing”; it’s the other way around.
You people test at lower noise/RPM scenarios, so thanks for that.
Still I’d like to see going down to e.g. 500 RPM testing (example for those G2 Noctuas, were you testers mentioned you tested only down to ~ 750 RPM).
Here is another example to make the point clear:
Got a medium sized ATX case, good – very good airflow overall (not exceptional, but not far behind).
5 x 140 mm fans (x3 silent wings pro 4 in the front, x2 noise blocker eloop at back/top. Enough holes at the back of the case near the graphic card too).
Heavy rtx 4090 graphic cards (good cooler overall) limited to 360 watts, CPU + other components max. additional 100 watts; 450 watts total in a video gaming scenario. 360 AIO-cooler for CPU.
After enough testing with different fan speeds, even air filter removal or front/top cover removal, even setting the 3 graphic cards fans to very loud 2500 RPM, I concluded everything above 800 – 900 RPM for those 3 be quiet front fans, is basically a waste.
Even with graphic card fans to 2500 RPM (not being the limiting factor), setting double the RPM (800 – 1600), results in maybe ~ 2 C lower graphic card temperature. The rise in noise is disproportional to the lower temperature; that’s what I mean with “waste”.
The processor temperature is basically unaffected, even if the CPU consumes 140 watts (the CPUs themselves are the limiting factor nowadays, or their coolers). Setting the x3 120 mm fans on the AIO cooler to 600 – 800 RPM is more than enough.
400 – 500 RPM for those x3 front fans is also more than enough.
The graphic card limited down to 200 – 250 watts (300 – 350 watts total PC power consumption) during a gaming scenario, the front fans enough set to 600 RPM is enough; everything above that the same waste.
I even set the graphic card to 400 – 450 watts total, and even here 1100 – 1200 RPM was more than enough, and above a waste.
You testers understand the point I guess. > 98 % of all graphics card bought and used are way below the 360 – 450 watts; most are 200 – 320 watts.
PC cases have become so good regarding air flow/noise reduction, fans are so good regarding performance etc. AIO-coolers too, that 300 – 1000 RPM is very much enough.
Thanks for reading, please consider adding another minus 3 – 5 dBA test.
Have a nice day and thanks for all testing.
This is impossible, as the sound meter used is capable of measuring from 30 dB. 😉
https://www.reedinstruments.com/product/reed-r8080-sound-level-meterdatalogger
please… try to read other contributors’ comments (especially if they are shorter) 😉
Have a nice day
On second thought,… 🙀 … maybe it would work by measuring 31 dBA at a distance of 7.5 cm, that should be 25 dBA at the standard HWC distance (15 cm), but it would be unnecessarily laborious and would not yield much information entropy… so it would be relatively possible, but inefficient 😀
A shorter distance is possible, but 7.5 cm borders on the point where the fan’s airflow starts interfering with the measurements — even when measured from the side. That, of course, leads to distortion, which we want to avoid. 🙂
yes you are right, moreover even if the microphone was not directly affected, the parabolic collar would definitely interfere with the air flow (at a distance of 7.5 cm).
… so such a measurement is not possible 😛
You’re absolutely right—a shorter measurement distance introduces airflow distortion and can also affect perceived noise. And this varies from one fan to another. For example, the impact on the NF-S12A differs significantly from the NF-A12x25. I mention these two fans specifically because of their dramatically different airflow patterns. That pattern is something worth factoring into a methodology design—not just for noise measurement, but if you’re aiming for the most accurate comparison possible at the output stage. 🙂
“I made a mistake with the “31 dBA equaling hwbusters 39 dbA testing”; it’s the other way around.”
Nope, you’re still wrong, and you can see that if you read the posts here explaining why.
Arctic P12 Max: more air flow (better for cases)
Arctic P12 Pro: more pressure (better for radiators)
As for performance, the P12 Pro shows higher airflow (compared to the P12 Max) without obstacles, with a plastic filter, and even on a grille — especially at lower speeds, which are typically ideal for system cooling. You could argue there are scenarios where the standard P12 performs better, but in almost all tested cases, the P12 Pro has the upper hand over the P12 Max. 🙂
Looking forward to hear one of these myself as I have been very disappointed with a P14 pwm fan from Arctic that glowing reviews but a very very annoying mosquito noise from 1100rpm upwards. I replaced it with a Termalright fan that has a prrrrrrrrr noise but low freq which is more bearable but won’t keep it for long 🙂
Yes, the acoustic profile of the P14 PWM was often found unpleasant by many users, but with the P14 Pro, Arctic addressed this by shortening the blades and adding a ring. We’ll get to the results soon. Based on what we’ve seen with the P12 Pro, we can reasonably expect a significantly more pleasant acoustic profile. Let’s hope that’s true — and that the P14 Pro doesn’t suffer from any other shortcomings in exchange.
PS: Which Thermalright fan are you comparing it to? 🙂
Very impressive results, thanks so much for all of your hard work. Comparing by noise level is the most meaningful tests in my opinion so thank you for that.
Which testing would be most relevant to a single tower CPU cooler heatsink (eg Cooler Master Hyper 212)? Thicker radiator?
— „Which testing would be most relevant to a single tower CPU cooler heatsink (eg Cooler Master Hyper 212)? Thicker radiator?“
Out of the radiators we use for testing, the thicker one will be closer to the heatsink of the CM Hyper 212. It has the right characteristics—not just due to its thickness, but also because of its relatively sparse fin density. It definitely won’t be an identical scenario, but I’m confident that the results on the Hyper 212 Plus will be just as excellent as with our thicker radiator. That is, unless there’s an unusual case where a specific tonal peak or resonant frequency appears, caused by airflow interacting with the unique shape of the heatsink you mentioned. And yes, that can happen at certain speeds. We certainly can’t rule it out—just as we can’t guess how likely it is. Best to test it and see for yourself. 🙂
i have 3 of these in my system, 2 are used to replace the stock fans of my TR peerless assassin and one is used as an exhaust case fan.
they are very quiet at low rpm and they remain quiet up to roughly 1400 rpm, above that they become noticeably louder. But it’s fine since that’s more than enough for my usage (they keep my 5700X3D at about 82°C on full load an 60-65 in gaming).
Above said 1400 rpm threshold they start making an high pitch sound that’s a bit unpleasant. IDK why Arctic focuses on those high rpm when nobody is gonna use them at those insane speeds, i’d rather seeing them focusing more on the 500-1500 range.
Thanks for sharing your user experience. 🙂
At higher speeds, the priority will be the highest possible airflow, while noise is not really considered (as long as it holds together, haha…). With these fans, Arctic wants to replace not only the slower P12 fans but also the high-speed P12 Max. 🙂
What was the noise level at max speed in dba?
Thank you for your question. The noise-level measurements can be found, for example, in chapter 14. The graphs there represent measurements without an obstacle, meaning the noise is not influenced by the fan interacting with anything. In the following chapters you’ll also find data with obstacles. I assume you’re asking about maximum noise… in that case, look at the last graph in each respective chapter.
We should point out, however, that these values are not directly comparable with those you may find elsewhere. The numbers will differ because testing methodologies differ—some measure from a different (often much larger) distance, without an amplifier collar like the one we use, with a different meter, in a different environment… There are truly many variables in noise-level testing. The result is also affected by the angle at which the microphone captures the sound. We discussed this in detail, for instance, in our interview with Noctua. 🙂