"Static Pressure Fans" Don't Exist

"Static Pressure Fans" Don't Exist - Part 4

Рет қаралды 1,183

6 ай бұрын

In this video I take a look at the same 120mm fans that I previously tested at 1,600rpm, but this time I've "down-clocked" them to 1,100rpm. What happens to the static pressure performance when the speed is lowered by 30%, and how does this translate to heatsink performance? Watch and find out!

Пікірлер: 35

@tim_lu 2 ай бұрын

Im very surprised that the Arctic P12 was so "Bad" in the Graph at 4:50. In different tests on youtube i have seen the Arctic perform on the upper places of the lists. First i tought that your fan is damaged or a faulty one but the Noctua performance is also bewildering me. And these Fans ar often head to head in other tests. (Btw. the Arctic and Noctua techservices are two of the bests in the industry and very obliging) I have also some Arctic coolers and fans myself and they work like a charm performance to noise wise. But maybe its just something other im not aware of or not seeing or as you said there are some variables in this tests that favor the other fans. Thank you for your good videos, i just found them randomly and i start to binging them! 👍

@FanPhys 2 ай бұрын

Thank you - I'm glad you found the videos useful. The P12 performs poorly I think because of the heatsink I'm using; with a less restrictive heatsink it will probably be ok. It's a very useable cheap fan.

@tim_lu 2 ай бұрын

@@FanPhys Its fun to watch enthusiastic creators like you. The deeper analysis helps to understand things in a deeper level. Im still surprised with the P12 and Noctua against your radiator because i have an AIO with a thick 35mm radiator and two P14 (they have lower static pressure then the P12s) and they work like a charm. And it corresponds with the testing of other KZfaqrs. So your findings are confusing for my knowledge and experience. I was wondering if the gap between your radiator from the curved design could make these difference but at the same time the Gentle Typhone has a simlilar design as the Noctua and he seems to work pretty good. In the end it isnt so important. Maybe just an anomaly. =)

@mattiasrask1575 5 ай бұрын

Very god work, nice to see such a detailed overview, only wish you measure fan noise to

@JJP1890 5 ай бұрын

I'm curious to know how a rad rather than a heatsink may change results as heatsink towers allow the air to flow out the sides while a rad doesn't. So I'm wondering if that extra resistance could show a greater performance difference between pressure optimized and flow optimized fans?

@FanPhys 5 ай бұрын

Interesting point - you're talking about ducting the flow, which is pretty standard with heatsinks in networking equipment because it improves the cooling efficiency for the reasons you mention. This is another thing that I will probably look into in a future video, and it will be interesting to see the results!

@JJP1890 5 ай бұрын

@@FanPhys Exactly, though the frame around a rad will introduce turbulence due to the frame wrapping around the front and back to provide the mounting points so may have higher airflow restrictions/back pressure than proper ducting.

@MtbWithBlue 5 ай бұрын

@FanPhys and @JJP1890 I too was curious about that especially with fans typically being mounted directly to the radiator. Arm chair science claims have said that fans pulling (mounted behind the radiator) was the worst, fans pushing (mounted in front of the radiator) was better, and push pull was the best, though marginally better. Though I feel like that claim is for coolant temp and not for case temp or overall airflow. The only push pull set up I have is on my test bench and this all gets thrown out the window for an open air setup. Keep up the good work!

@igors_lv 5 ай бұрын

For real static pressure fan, you would bolt 3 of them on intake and put 1-2 radiators on outlet of the case and that would work really good. The fact that you bolt fans straight to radiator or heatsink tells you that its juts airflow and you need to have it very close to have effect.

@infi84 5 ай бұрын

while certainly interesting I find measuring performance at fixed RPM quite futile, in the grand scheme of things nobody really cares how fast a fan actually spins, the only two metrics that really matter is performance and noise, so I think the question that needs to be asked is what is the max RPM a fan can run at while producing a similar loudness and noise profile, and that's where "fan optimization" comes in imho, if a fan has lower static pressure at 1100rpm as another fan that is "less optimized" but is just as silent at let's say 1500rpm where it manages to produce a higher static pressure than that other fan, then that would certainly be an "optimized fan", because it manages to create a higher static pressure at the same noise level, wether or not it needs to rotate at a faster rpm for that or not doesn't really matter in the end. Still, I apreciate all the insight and comparisons between the different fans, also I used to have 9000rpm delta fans in my PC in the early 2000s back when I didn't care about noise at all and heatsinks were a lot worse at dissipating heat ^^

@FanPhys 5 ай бұрын

Thank you! It's a fair point and I agree that most people want to know about noise, but noise is too subjective a metric which is why I made the series specifically about static pressure performance. Interestingly enough, I would agree that most fans in the PC market at the moment are primarily optimised for low noise, but the actual R&D behind such optimisation tends to be every man and his dog copying the design of the Gentle Typhoon.

@LogioTek 5 ай бұрын

Variables are also fan samples. It's more expensive and time-consuming but maybe do 2-3 different samples of each fan and show variation. With one sample you never know if it's an outlier or defective.

@FanPhys 5 ай бұрын

I would say if I was doing extremely high-stakes research then yes, that level of control would probably be appropriate, but for this sort of testing (mass produced electronic goods) it's pretty safe to assume every fan will function exactly as intended/designed. A defective fan would be one that didn't power up! x)

@emiljanduka51 5 ай бұрын

@@FanPhys and if a defective fan is shipped out thats the manufactures fault so even if the fan is defective its the brands fault and u shouldnt buy from them

@prostagma6055 5 ай бұрын

The reason i stumbled onto this channel was because I wanted to know what the best fan for MINISFORUM BD770i was but i feel like i been left with more questions than answers i am stingy so I'll probably go with artic p12. Do you find that there are any better budget options and thanks for blowing my mind i used to think static pressure used to mean a direct correlation for performance in heatsinks

@usmcp 5 ай бұрын

The P12 max is my go to, especially for mini it's builds. Even at 3k rpm, the noise isn't bad inside a case considering just how much air it pushes, for the price. There's a guy called STS who also also tests fans and while there are surely faster fans, it's diminishing returns.

@thetechchannel321 5 ай бұрын

Great vids keep up the great work,

@FanPhys 5 ай бұрын

Thank you! The P12 is a fine budget choice, with one added bonus being you can get it either in black or white (if that's your thang).

@tacticalcenter8658 3 ай бұрын

Why would you get a static pressure fan and not run it in the PQ range it was designed for? Seems like making the wrong choices.

@ChielScape 6 ай бұрын

The 4 variables you mention aren't that relevant. Your CPU and chosen stress test represent a particular heat load. Physics doesn't care about the brand or model. 100W is 100W. Your cooler is fixed in all your own testing, so it falls out of that equation and the ranking of fans is valid. If someone else does the same thing but with a different cooler, the heatload and ability of the heatsink may be different, but that only results in a linearly scaled different equilibrium temperature. The ranking of the same set of fans tested on another CPU/Stresstest/Heatsink combination should be roughly the same (with allowance for measurement inaccuracy) as yours. There are variables that matter, if they're not properly controlled for by either you, the other tester, or both, but none of the variables you mentioned fall into that category. If you would test on the ryzen, for example, and find that the fans' performance ranking is radically different, it would signify an error in your method somewhere, a variable from one fan being tested to another that you haven't properly controlled, not "oh its just a variable [from one tester to another], that just happens". An example would be disturbances in airflow from opening a door or window, or from CPU heat output not being consistent (have you verified CPU power draw?). On an unrelated note, I wanna ask you if you are aware of the channel Major Hardware with the Fan Showdown series. Ideally, I'd like to see you two exchange thoughts on fan designs and share your findings.

@FanPhys 5 ай бұрын

You've completely missed the point I made - these are not internal variables so of course they aren't relevant to my tests; these are external variables, thus they explain why my test results may not line up with those of other people doing similar tests with different gear and software. The internal variables have already been covered in previous videos in the series.

@ChielScape 5 ай бұрын

@@FanPhys It seems more like you're the one not getting my point. Internal variables are the only ones that can cause different results. As long as the methodology remains the same between two tests the external variables don't matter. (A different methodology isn't a "variable", its a completely different test). Lets say you rank 3 fans as A B and C, in that order, on your 100W heat load. You gets temps like A=70*C, B=75*C and C=80*C. Now I do the test on a 250W constant load I9 13900K, but I also watercool it with a 480mm radiator and I use 4 of each fan. You might say that's very different. I say it isn't. While my temps might be 55, 58 and 62*C respectively, they should still be in the order of A B C, because the ability of each fan to push a certain amount of air through a radiator or fin stack should not change relative to one another. If my results had been ranked C lowest and A highest temperature, then one of us has made some kind of internal error in their tests. Radiators and fin stacks aren't that different, especially if you are of the belief that static pressure is a meme and fans at the same RPM will never perform much differently aside from noise. So either one of the two tests is faulty, or the tests were entirely different tests. Like you can't compare a speed normalized test with a noise normalized test, they're not the same test. That difference is not a variable, certainly not one of those you mentioned.

@FanPhys 5 ай бұрын

@@ChielScape OK, I follow now - I think we are talking at cross purposes. I do not consider all other testing methodologies to be free from error and here’s why: many (possibly most) of them don’t accurately measure the ambient temperature, which needs to be measured close to the intake of the fan. Why? Because the testing environment will be filled with various regions and currents of warm and cool air whose temperatures do not influence the result of the test. If a temperature other than that at the intake of the fan is taken to be the ambient value, the test results may be similar, but only by coincidence. Respectfully, I still disagree with you on the subject of external variables. A different heatsink (different fin density, size, shape, pitch, area) will impart a different characteristic to the flow from each fan, which in turn means different flow velocities, a different turbulence characteristic, different viscosity / boundary layer behaviour, different thermal diffusivity, etc. etc., all of which directly affect the thermal performance of a heatsink. Those are just a few of many internal variables that have changed relative to each fan, all because of that one “irrelevant” heatsink swap. It is entirely possible that some of the fans will perform better on heatsinks that have, for example, a lower fin densitiy, resulting in a change to the rankings. As for testing radiators, that just introduces a whole ‘nother universe of physics (hydrodynamics) that is not worth getting into quite yet. :)

@ChielScape 5 ай бұрын

@@FanPhys If the fin density matters enough to change the ranking of the fans, your point about static pressure optimization is false. The two hypotheses cannot be true at the same time.

@FanPhys 5 ай бұрын

@@ChielScape I'm sorry but I don't follow. What point have I made about static pressure optimisation? My tests consistently and conclusively show that static pressure performance does not correlate to heatsink cooling performance with the chosen heatsink. This may not be true with a different heatsink, for the aforementioned reasons.