It's custom, and you just set PWM fan temp source to specific sensor. 
 
@TomSawyear212, please share your results.
 

That's correct, the Noctua is hooked to a PWM fan headers on the motherboard which I control via Asus AI Suite 3.   With the current thermistor just sitting 'on top' of the backplate, my temperature variance and more importantly, response time is not great, so my fan curve is tight to accommodate;  Once I have the thin film unit I should see faster temperature response and should be able to tune my curve to get a better cooling/noise balance.
 
At idle, the Noctua runs around 650 RPM (virtually silent) and my GPU idles at 30-31c
 
Under load, Heaven/Valley, I peak out at around 48-51c and my T_Sensor will slowly get to around 40c at best (again, that will improve once I have the new Thermistor and 2nd Noctua in place).
 
FWIW, here is my current curve:
 
 
I'll report back tomorrow evening on how the new thermistor and 2nd F12 affect the numbers.
 
 
 
 
 
 

AdamInk
Hmmm. Then i dont know is it worth it to change to stock fan to noctua. My temps at idle about 25-27 Celsius, under load in gaming its max 52-53 C.
With Noctua will i have improvement in noise or temps what you think?

I think the goal here is to minimize idle noise/rpm as stock fan blowing full speed all the times.
@TomSawyer2112, did you try to link it to CPU, benching or gaming, it will be rough indication on the overall system load, including GPU. And this is watercooling, fan curve doesn't have to be very reactive or precise.
 
Another option, if you want absolute precision and your mobo supports it, is to use use NSSM
https://forums.evga.com/T...RPM-mode-m2678869.aspx

So the Noctua quiter then the stock EVGA fan?

Any fan is quiet at low rpm. Noctua are not, they all turn into whistling machines when approaching 1500rpm. This is not motor that is making noise, air moving through the radiator fins mostly, noise contribution from fan motor is relatively small.
 
As far as linking to CPU, keep in mind that you not only cool GPU but also help escape hot air from case, if set as exhaust, or pull in fresh in if intake. So from this perspective it's even better to account for overall system (cpu) load.

Well i think i try the Noctua, will order it. I dont know that if i plug it into the gpu , the gpu will adjust the fan curve based on temps?

AdamInk
Well i think i try the Noctua, will order it. I dont know that if i plug it into the gpu , the gpu will adjust the fan curve based on temps?

Can 100% confirm that the GPU/Precision software will not adjust the RPM of the Noctua (or any fan) if you plug it into the fan cable off of the GPU. If you are wanting to vary the RPM's of the Hybrid's radiator fan, putting it on some type of fan controller is a must (whether that be a motherboard header or dedicated fan controller).

TomSawyer2112

AdamInk
Well i think i try the Noctua, will order it. I dont know that if i plug it into the gpu , the gpu will adjust the fan curve based on temps?

Can 100% confirm that the GPU/Precision software will not adjust the RPM of the Noctua (or any fan) if you plug it into the fan cable off of the GPU. If you are wanting to vary the RPM's of the Hybrid's radiator fan, putting it on some type of fan controller is a must (whether that be a motherboard header or dedicated fan controller).

Curious to know how you confirmed this?  My BIOS is set to Slave and GPU is OC'd +100Mhz and Memory +400Mhz.  Idles at 35C with max temp around 55C during intense gaming.  Super quiet too!   I'm running a push/pull setup on the radiator with 2 x Corsair ML 120 Pro fans.  I have those connected to the GPU via a Phanteks 4 Pin PWM- Y Splitter Fan Cable (PH-CB-Y4P) connected to a Phobya 4 Pin PWM plug to 3 Pin (Socket) 30cm Adapter - Sleeved black.  Both can be purchased online (Amazon, Newegg, etc.).  I fooled around with some different fan profiles, but just decided to go back to the default setting.  The fans definitely change speeds depending on the GPU temp.  At idle they're running super slow.  While I may not be able to see the RPM speed of the GPU fans via Precision, again, they definitely fluctuate.  I'm very happy with my setup.
 
 

Andrew_WOT
I was doing some test with stock fan in stock configuration, measuring AF with paper strip. Same noise and same AF, idle or full load.
Is the trick in connecting aftermarket PWM fan via adapter? And we are talking radiator fan, just to make sure.
BTW, when you use Y-splitter doesn't it limit power draw to individual fan (I am most likely wrong on that, anyone knows)

Thanks for the info.  Yes, talking about the radiator fans.  I have both Corsair ML 120 fans connected to the Y-splitter, but not sure about the power draw for each individual fan? I know for a fact though that the RPM of the fans change with the cards temp.  The max I'm seeing is 55C.  In fact I just played a game of FIFA 17 (in 4K with max settings) and the GPU was @ 2088 MHz steady and Memory 5900 MHz and the max temp was 55C.  Just took a screen capture a minute ago of the current state.  As for the GPU fan speed control via Precision, there does not seem to be any custom control. 

SDPadre7
Thanks for the info.  Yes, talking about the radiator fans.  I have both Corsair ML 120 fans connected to the Y-splitter, but not sure about the power draw for each individual fan? I know for a fact though that the RPM of the fans change with the cards temp.  The max I'm seeing is 55C.  In fact I just played a game of FIFA 17 (in 4K with max settings) and the GPU was @ 2088 MHz steady and Memory 5900 MHz and the max temp was 55C.  Just took a screen capture a minute ago of the current state.  As for the GPU fan speed control via Precision, there does not seem to be any custom control. 

Same with me. I have two Cougar fans(CFD-120) connected with a Y-splitter and I'm seeing the same thing. The fans are definitely fluctuating with temperature. These fans are 3 pin fans and are rated at a max speed of 1200 rpm(+/- 10%). Very quiet and my temps max out at 48C on a 1440p monitor with settings on ultra.
 
I posted a video earlier in this thread showing the fans spinning at idle. In the video the top two front intake fans are being controlled by the motherboard and are spinning at 800 rpm. The bottom front intake fans in push/pull on the gpu radiator are clearly spinning slower than the top two. The radiator fans spin faster as temps increase. 
 
https://youtu.be/VQFp1cwDkIQ

I think I've dug to the bottom of it.
While it's 980 Hybrid, the design is the same, and the same confusion on whether it was controlled or not.
Turned out it was, voltage based on GPU temp, it's just stock fan is loud
https://forums.evga.com/FindPost/2468098
 
I will test with Corsair SP120 and 4pin to 3pin adapter to see how it works, might be a couple of days. May be all we need is just a quieter fan.

I'm miffed. Maybe it has something to do with multiple fans and a splitter being used on the radiator? Increased power draw?
 
But I know what I see and as I'm typing this I'm looking over at my gpu radiator fans and they are definitely not spinning at full speed.  

TomSawyer2112
By no means is this a truly scientific test, but in a quite room, I observed the following running each fan at 100% on an open desk:
 
dB reading in room with no fans running:  32 dB
eVGA stock fan:   64.2 dB total - 32.2 dB 'normalized'
Noctua NF-F12 Industrial:  65.5 dB - 33.5 dB 'normalized'
 
The eVGA had a very unpleasant buzz to it, guessing motor noise.
The Noctua was of course loud, but there was no underlying buzz or hum of motor, just the sound of a lot of air blowing.  
 
The Noctua moved more air to my <very> unscientific hand test.   It seems a bit more spread out, the eVGA more center concentrated.  No idea which translates to more effective cooling but assuming cooling capacity is identical, the Noctua certainly wins for sound quality and of course build, being pwm, etc.   Loud, sure, it's 2k rpm and moving a pretty large amount of air for a 120mm fan.
 
Noctua rates it at 29.7 dB pushing 71.69 CFM (121.8 M3/h) and 3.94 mm H20 static pressure (nice rating to say the least).
 
That's about as good as I can do guys...  I'll pop this guy on the rad along with the new thermistor when it arrives tomorrow.    - good times -  :-)
 
 

So what you think?
Noctua worth the money and get lower noise and temps? Or just the same with the stock fan?
Did you saw temp changes with Noctua?

Great question because the Noctua fans are certainly not cheap and even if the GPU is varying the voltage on its 2-pin fan port, the stock fan (to me) has a more annoying and therefore noticeable sound to it. Chances are, the same improvement could be had with an ML120 (or SP120), Cougars or other high quality fans.  
 
As far as temperature change, with a single Noctua at idle it is a bit warmer, 32c vs 29c give or take, but that is expected since I'm running the fan at only 640 rpm idle (dead silent).  But again at idle is when I want the PC to be as close to silent as possible, so a few degrees warmer with no load is acceptable for my goal.
 
My test was pretty simple as far as a standard fan connected to the gpu and then a 12v power source, but the fan was certainly spinning at a high speed on both sources. What really disappoints me was the fact that there was no control over the RPM through Precision, and certainly no way to run any GPU connected fan at very low rpm/noise level at idle.  I think it would still be interesting to take a 3 pin fan and split off the tach signal to see how much it varies when connected to the GPU fan plug, but my bet is probably not much.
 
 

petmic10

SDPadre7

petmic10
I'm miffed. Maybe it has something to do with multiple fans and a splitter being used on the radiator? Increased power draw?
 
But I know what I see and as I'm typing this I'm looking over at my gpu radiator fans and they are definitely not spinning at full speed.  

Same here.  I just recorded video with my system at idle and while playing NBA 2K17 (4K max settings) and there is a distinct difference in the fans speed.
 
@ Idle https://youtu.be/SbCoF9PDo7I
While gaming @ 4K https://youtu.be/uQ-NrmmpMU0

 
Definitely a difference there. I'm glad I'm not the only one who noticed. So the software has to be doing something. We just can't control it.

That's good to see, thanks for making the videos!  It is definitely putting some variance on the fan speed, not sure why the fan was running so high on my rig when I tested.  Are those ML120's?   The fan plug was not able to spin up a single NF12 (just kind of twitched) but a regular fan no problem.  How do the two compare to the single stock fan noise wise?