What is the viable option for high CFM low noise 120mm fans nowadays?

[blackjackel]

I was looking around and it looks like the top choices are:

the gentle typhoon AP-15, which isn't viable because its out of stock everywhere and likely not made anymore

The cougar vortex: but people complain about it whining at low RPM, or when mounted one way or another, which makes that not viable

The noctua NF-F12, which isn't viable cause it's 25 dollars...

So what's a viable 120mm fan solution for high cfm low noise these days?

 

[RnRollie]

Noiseblocker eg eLOOP B12-P
http://www.blacknoise.com/en/site/page/products.it

And the PWM versions of Delta, SanAce, Papst, Sunon,... : dial them back a notch & they are silent while still better performing as most "established/good/bling" fans

https://db.sanyodenki.co.jp/product_db_e/coolingfan/dcfan/cooling_dcfan.html
http://www.sunon.com/pro.php?c1=10
http://www.delta.com.tw/product/cp/dcfans/dcfans_main.asp


Golden Rule(s) for Fans:

CHEAP + SILENT = NO PERFORMANCE
CHEAP + PERFORMANCE = NOT SILENT
SILENT + PERFORMANCE = NOT CHEAP

+Most SILENT CaseFans are not well suited for CPU-HS-Towers or Radiators

.

 

[Blaylock]

Gentle Typhoon AP45's

Noiseblocker E-Loop B12-3

or

Corsair SP120

Biggest question is what color do you want, Back, White or Any?

 

[Bustos]

Golden Rule(s) for Fans:

CHEAP + SILENT = NO PERFORMANCE
CHEAP + PERFORMANCE = NOT SILENT
SILENT + PERFORMANCE = NOT CHEAP

+Most SILENT CaseFans are not well suited for CPU-HS-Towers or Radiators

.

+1

Looks like it may depend on what you're willing to spend on a fan and your dB tolerance for that CFM power. As stated above, if $25 is out of the question for a fan, then your options are rather limited.

 

[doyll]

High speed AP fans (AP-29, 30 &31) can be turned into PWM fans by adding the PWM lead. They are 7 blade instead of 9 blade and need 50-70rpm more speed for same airflow as AP-15
http://martinsliquidlab.org/2012/04/28/r9-fan-testing-gentle-typhoon-ap-29/

How to do it.
http://www.overclockers.com/Gentle-Typhoon-AP-29-to-PWM

en Rule(s) for Fans:[/u]

CHEAP + SILENT = NO PERFORMANCE
CHEAP + PERFORMANCE = NOT SILENT
SILENT + PERFORMANCE = NOT CHEAP

+Most SILENT CaseFans are not well suited for CPU-HS-Towers or Radiators

.

And to that I'll add
+Most SILENT Case Fans are not well suited for cases, especially if they have filters.

 

[blackjackel]

Just to update everyone on what I went with, Corsair SP120 and AF120 fans come in bundles of two, making them 10-15 bucks a pop, which makes them affordable AND good choices for performance and noise...

I now have SP120 on my radiator in push pull, and two AF120's on my case, I lowered the RPM's and I can barely hear them.

Sadly, now the loudest thing in my Case is my Thermaltake toughpower cable management 1200w, and I have to figure a way to slow down the fan on that thing, might have to open that sucker up to do it though.

 

[Blaylock]

That's a good choice. They aren't the quietest or the most powerful, but they are good a good middle of the road and probably the best value of the better fans. Those are what I'm considering for my future build too.

 

[Bustos]

Any improvement to your loop and case temps?

 

[BruceUSA]

CHEAP + PERFORMANCE = NOT SILENT


I wished this statement hold true to me but it is'nt. I am using 14 Delta fans and cost me $560 that is'nt so cheap after all at $40 each.

 

[NiHaoMike]

And the PWM versions of Delta, SanAce, Papst, Sunon,... : dial them back a notch & they are silent while still better performing as most "established/good/bling" fans

More to the point, many PWM (variable speed) fans will have some duty cycle point at which point the noise and power usage increase faster than the airflow. That's the point where the inverter is entering the "overmodulation" or "field weakening" (depends on implementation) region, instead of outputting pure sine waves like it normally does.

Also note that apart from fans designed for both modes of control, voltage controlling most PWM fans will not work properly. You won't get as much control range as with PWM control and cause the inverter to overmodulate, meaning it will operate noisier than using PWM for the same speed.

 

[RnRollie]

true, i should not have said "dial them back a notch", i should have said something like "set the duty cycle/PWM signal to 50% or something like that" ; since "dial them back" will most likely be associated with voltage control.

 

[ivanlabrie]

I have CM Blademasters and they are NOT quiet but definitely cheap and move some good cfm and air pressure.

I love 120x38mm fans, the good ole Panaflos of old were amazing

 

[NiHaoMike]

On most Delta fans without director vanes, the optimum point is actually more like 75% before it starts overmodulating.
Here's a rather extreme example of overmodulation:

Note that instead of being sinusoidal (after the inherent low pass filtering of the motor windings), the waveform begins to look more like a square wave. What is interesting is that the phase to neutral is still very close to a sine wave, but there's enough harmonics to make it start sounding "buzzy".

 

[Bobnova]

On most Delta fans without director vanes, the optimum point is actually more like 75% before it starts overmodulating.
Here's a rather extreme example of overmodulation:

Note that instead of being sinusoidal (after the inherent low pass filtering of the motor windings), the waveform begins to look more like a square wave. What is interesting is that the phase to neutral is still very close to a sine wave, but there's enough harmonics to make it start sounding "buzzy".

Yeah that's why I hear more non-air-noise over a given duty cycle on my Delta, Nidec, Joujye, Intel, AMD, and Noctua PWM fans.
Except wait, I don't. Weird. Almost like you're making things up to sound cool.

Every single PWM fan I have ever played with has a decently smooth noise curve when fed a "correct" (read: fits the datasheet) PWM frequency.
This is a non-trivial number of fans, over a wide variety of RPM and airflow specs.
So, please, quit making things up to sound cool, and stick to reality, if you're capable.

Re: scope shot, try running the "fan" (I'm far from convinced that's a fan at all) at the proper PWM frequency. <400Hz is not it, for any PWM fan I've run into. There are a few Nidecs that want 500-3000Hz, but even they aren't happy that low.


RnRollie, you nailed it, that's exactly the thing to do.

 

[NiHaoMike]

368Hz is the fundamental frequency going into the motor. It corresponds to about 5500RPM (full speed) for that particular motor. Note the "hashing" which is the PWM signal coming from the DSP in order to make the waveform more like a sine wave. But if you constrained the waveform to always approximate a sine wave, you'll lose performance since the output voltage would be less than for a square wave. Hence the inverter starts "clipping" when commanded to run at over about 75%. (One exception is Noctua which is designed to always output sine waves, but they're not exactly designed for high performance either.)

The PWM used for the inverter and the PWM used for the control signal are not the same. It was chosen because it's easy to generate with a digital circuit and easy to interpret with an analog circuit. Whether analog or digital (and resolver based or sensorless), the inverter measures the duty cycle and uses that to synthesize new PWM signals that are actually used to drive the motor.

 

[Bobnova]

It's a lovely theory, but it doesn't come close to explaining why the motor noise reflects the input PWM signal. Were what you said true, input PWM frequency would not show up in the motor noise.
Sadly, for the theory anyway, it does.

Care to expand that scope shot so the clipping is more visible? Or perhaps a video of PWM percentage varying things? Or a part number for the inverter IC? Or anything I couldn't make up?

I don't see anything "very close" to a sine wave there. What exactly are the traces measuring?

Sidenote: Real world testing continues to disagree with you from a practical perspective.

 

[NiHaoMike]

The incoming PWM frequency does matter in that if it's too low, the commanded speed would fluctuate and result in erratic behavior. If it's too high, analog controllers would often accept it just fine (turning PWM into an analog voltage is just a matter of low pass filtering it) but digital controllers often miscount and also behave erratically.

Top trace is one of the phases to ground, bottom trace is neutral to ground. The clipping is the part of the waveform where it "flattens". Note that the phase to neutral (which is what the motor cares about) looks more like a sine wave than a square wave (ignoring the high frequency carrier) despite the overmodulation. The current waveform couldn't be measured due to the way fan motors are constructed, but it would look even more like a sine wave since the inductance of the motor windings would attenuate the harmonics.

 

[wagex]

pwm signals are not sine they are on off analog. idk if clipping would be considered clipping if its intended to be that way. normal pwm _-_-_-_-_ sine ~~~~~~~~~ clipping occurs in a sine wave where the wave becomes so distorted that it creates flat peaks or plateaus instead of nice smooth round waves.

though i could be wrong im not an electrical engineer.

 

[NiHaoMike]

PWM is how inverters generate sine waves from a DC supply. The inductance of the motor windings averages it out. By changing the duty cycle, the resulting current can be a sine wave.

 

[wagex]

but the pwm signal from the motherboard doesnt go into the windings it doesnt power anything, its just a 5v signal for the board inside to figure how fast for the motor should be spinning.