Stable Fan Control: Grab a PWM Fan Controller?

Associate
Joined
2 Sep 2013
Posts
2,155
So, having finally been working on silencing my main rig for a while, I've finally managed to get it totally inaduible at a distance of around 75cm. And needs to be within 50cm range to even hear the slightest of fan whirrs (airflow largely) being run through the setup in the dead of night even when gaming and only when listening for it. Idealy, we just look to see if any of the LEDs are on to see if it's on now, so quite pleased with the outcome. :D

Quick reminder of setup I'm talking about:

Corsair 650D Case (5.25" Bay vacated for Noctua A14 PWM Fan as Intake set at 50% ~530RPM with LNA, 200mm Fan Grille removed and original fan replaced with Noctua A20 PWM Fan as Intake set at 60% ~480RPM, Rear has Noctua A12x25 as Exhaust at 50% ~960-1000RPM with LNA, PCI Slots has a cheap 80mm fan at 50% ~1050-1200RPM with LNA (this will be changed to a Noctua later).)
Motherboard is Asus Rampage IV (4) Extreme, fully populated with 64GB RAM and a 4930k.
4930k is at 4.3Ghz at ~1.29v, idles at 1.1Ghz at 0.85v
CPU Cooler is a Noctua U12A (Both fans at 50% ~720RPM-750RPM)
Powercolor RX580 8GB, with the fans and shroud removed (but kept the Heatsink) and strapped on two Noctua A12x25 and controlled by the card (Kept at 35% only). Undervolted and kept at default speeds.
Corsair Bronze HX650W PSU

Temps:
CPU (Under load) 77C, idle 25-27C
GPU 83C, idle 27C
GPU VRM 91C, idle 34C
Room Temp: 25-26C

But I've got a problem, one of the fans, that's being controlled by one of the Fan Headers on the Motherboard, strangely goes up and down by a small % on its own somewhat randomly. But this % is just enough to send the fan going from inaudible, to creating a type of harmonic resonance with the case (It's to do with the 5.25" cage that it partly sits on, it doesn't like activity on them, resulting in magnifying the sounds at certain times through the cage and onto the case).

So my question is: Will grabbing a PWM Fan Controller allow me a more steady (and stable) control of the fans in the system that the motherboard clearly will not permit me to control steadily with? Or will this (fan % and RPM going up and down) continue even with a Fan Controller? (I'm aware fans have a % where they sit under, but the fan going up and down seems off to me when other fans on other headers that aren't PWM on the motherboard don't do this)

Note, I've tested and found that ANY fan will cause the issue when sitting in the 5.25" bay when at 50% or above, basically over 550RPM for the 120mm or 140mm fans. But I do have in it right now a Noctua A14 PWM, but it can be swapped with a Noctua A12x25 PWM if needed, but I see no need for this since both experience the same issue as said already. At certain RPM's ~55% to 60%, resonance happens, any less and it'll be silent with the whole system, but any less than 50% and I have a lot less airflow coming into the case, which is a problem, as its current state (when it doesn't go above 50%) is effectively in harmony with the rest of the system to keep it from overheating and remaining inaudible. This is due to the ghetto mods I've done to the Intake and Exhaust areas of the case, resulting in restricted space arleady, so a further reduction of the fan speed will be problematic.

So if any PWM Fan Controllers are suitable, which internal ones are best? I'd ideally like to control the various Intakes and Exhaust fans, which number 4 (two Intake; A14 + A20, and two Exhaust; A12x25 + PCI Slot fan - that will become another A12x25 later on) to a % less than the 50% minimum on the R4E Motherboard, but have them ramp up to 50% when required. So any automatic/programmed or Windows controlled ability to control them to that level would be ideal as well. External (5.25" Bay) ones for example won't work in this case because the bay is already taken and I'd rather have it respond to pre-set changes rather than controlling them myself.

Thanks all.
 
It's not your question and I don't have any recommendation for best PWM fan controller, but thought I'd ask about the GPU. GPU temps look hot for RX 580 undervolted and 2 x NF-A12. Are the fans pressed against the heatsink firmly? Is their weight on the heatsink (pulling heatsink down from PCB and perhaps affecting GPU and VRM temps), or are you supporting the fans by other means?
 
Yep, the A12x25's are pressed onto the original Heatstink that came with the Powercolor RX580 (as much as physically possible that is).

I removed the original plastic shroud and the two original fans and attached two A12x25's onto the Heatsink via zipties, as the original fans still needed silly high RPMs to keep the GPU cool at an acceptable noise, but I wanted even less. In order to facilitate this with the older original fans I also allowed them to reach the 80C region (as they didn't throttle until 86-90C odd) and just slightly more to have the least amount of noise. But that needed like 50-65% of the fan to do so with the original fans and again, even then, the noise was very noticable, despite being low.

With the Noctua A12x25's, I am able to run them at 50% and they will never breach 75C with all things on etc, most of the time 70C odd with the reduced cool air from the Intakes (from ghetto trash modding the front of the case). 40% they would reach 77-78C odd, but they produced an audible amount of noise still. So I tried 35% and found I couldn't hear them then. But the amount of cooler air they're pushing onto the GPU now from the Intakes is far less than most (still enough, otherwise it wouldn't keep the GPU at 83C max whilst maintaining full frames). So they're hot because of the reduced ventilation at the Intakes and I've forced the ones on the GPU to run at 35% only (with LNA), so they're basically right at the limits of what you can take them to for the price of silence.

I'll grab some pics of the reduced ventilation and post them later, that'll give an idea on what I've done and why there's less cool air coming in, and subsequently, why the GPU is being allowed to hit the low 80s.
 
:)

My own GPU fan mod isn't pretty but also effective. Ziptied the first fan to the PCIe bracket on one end, and the other end is supported by some wires attached to the second fan and holes in the case. Both fans are linked tightly with zipties. So that instead of making the card sag more with the weight, the fans actually act as support since they are not directly attached to the heatsink, while being very snug against it and helping keep it level with the PCIe slot. Did this because I heard of black screen and/or temperature issues sometimes caused by the extra weight of regular fans attached directly to the heatsink.

Using Phanteks MP120s on an EVGA GTX 1070. Not as good as NF-A12s but still very good and quiet.
 
PWM control,whether mobo or external card etc will ALL behave like this
Ramping up and down.
Can you set the fans to manual with a fan curve in the Bios?That may be better for you noise wise :)
 
What Andarial said usually works very well. Worst case is higher idle/low load temps (30-35c instead of 25-29c), but still so low there are no long term problems.
 
So if any PWM Fan Controllers are suitable, which internal ones are best? I'd ideally like to control the various Intakes and Exhaust fans, which number 4 (two Intake; A14 + A20, and two Exhaust; A12x25 + PCI Slot fan - that will become another A12x25 later on) to a % less than the 50% minimum on the R4E Motherboard, but have them ramp up to 50% when required. So any automatic/programmed or Windows controlled ability to control them to that level would be ideal as well. External (5.25" Bay) ones for example won't work in this case because the bay is already taken and I'd rather have it respond to pre-set changes rather than controlling them myself.
AquaComputer's Aquaero/Quadro would give way best control ability inluding smoothing and low pass filtering of values to prevent fast pumping of speed from temperature changes.
https://forum.aquacomputer.de/weitere-foren/english-forum/109453-breaking-all-limits-aquasuite-x-6/

https://www.overclockers.co.uk/aqua-computer-aquaero-6-lt-usb-fan-controller-wc-31j-aq.html
https://www.overclockers.co.uk/aqua-computer-quadro-pwm-fan-controller-wc-338-aq.html
 
PWM control,whether mobo or external card etc will ALL behave like this
Ramping up and down.

Darn. :(

Can you set the fans to manual with a fan curve in the Bios?That may be better for you noise wise :)

Unfortunately, there is no "curve" on the RIVE (R4E) in the BIOS. The only fans that can even have any semblance of control is the CPU headers (and it's more a slope than a curve control), and I've already got that manually set to 50%, and have the A14 attached to it, but as you note, PWM will fluctuate, so that's part of the problem (since it occasionally exceeds the actual 50% to become enough rpms where it begins to resonate the case). And of course, the "other" Fan Headers on the RIVE are no better. They min at 50% for some, others at 60% and are all manually controlled in the BIOS or have their temps tied to something else that will never change them either, so kinda worthless. But, I also discovered that some of these Fan Headers are not true PWM headers, and so the cause the fans to run at some funny % and RPMs instead because they're trying to voltage control the PWM fan instead (or even possibly doing both; voltage and PWM, which might be why some of my tests of the fans kept stalling a while back now I think of it, hmm....). Which is what lead me here looking for alternatives. :)

What Andarial said usually works very well. Worst case is higher idle/low load temps (30-35c instead of 25-29c), but still so low there are no long term problems.

I agree with you and what ANDARIAL has said, I've allowed my setup to heat up without issues because of tollerances they can safely accept already. Unfortunately, the RIVE as said above already is terrible for fan control. And doing what I can with it already is what lead me to what I have so far, but the primary issue is that I can't set the fans any lower, because then the GPU will overheat and throttle at 90C. I've tried this already. So I just need to get the A14 at the intake under control, but none of the fan headers on the motherboard are suitable for this unfortunately. Whether they're not true PWM, Voltage isn't quite correct for Voltage control or plain can't reach the 50% I need (the Chasis Headers appear to be properly controllable but they min at 60% :mad:)

I think it's just bad luck, the RIVE appears to have been at the border between implementation changes at motherboard makers who began to implement more, some propietary, some half baked, and some proper fan controls, but the RIVE fell into the first two or three categories but not the last one. Speedfan for example doesn't work on RIVE at all, and coupled with the poor BIOS fan controls... doesn't make my life easy here.


Thanks EsaT, I'll have a look at those in a bit. That Quadro looks like what I might need and be after, if it can run the fans below 50% (since none of mine on the motherboard can bar the CPU header) so it can keep the fan at 45-50%, then that's likely what I'll end up with. Will look further into that one to make sure that does what I need. Thanks again. :)

:: edit :: Just read somewhere that it has PWM control between 0% and 100% control, so that's definately capable of what I want. Grabbing it now then. Thanks again. :)

Oh, screenies of this poor mans silent rig later on. :p
 
Thanks EsaT, I'll have a look at those in a bit. That Quadro looks like what I might need and be after, if it can run the fans below 50% (since none of mine on the motherboard can bar the CPU header) so it can keep the fan at 45-50%, then that's likely what I'll end up with. Will look further into that one to make sure that does what I need. Thanks again. :)

:: edit :: Just read somewhere that it has PWM control between 0% and 100% control, so that's definately capable of what I want. Grabbing it now then. Thanks again. :)
There should be similar to Aquaero maximum settings per channel setting limit to inside which controller functions:
https://youtu.be/NQcNl2ZE1Lk?t=4m24s
If that is set to half the max speed that's how fast fan will ever run even if control curve says 100%.
 
Darn. :(

Unfortunately, there is no "curve" on the RIVE (R4E) in the BIOS. The only fans that can even have any semblance of control is the CPU headers (and it's more a slope than a curve control), and I've already got that manually set to 50%, and have the A14 attached to it, but as you note, PWM will fluctuate, so that's part of the problem (since it occasionally exceeds the actual 50% to become enough rpms where it begins to resonate the case). And of course, the "other" Fan Headers on the RIVE are no better. They min at 50% for some, others at 60% and are all manually controlled in the BIOS or have their temps tied to something else that will never change them either, so kinda worthless. But, I also discovered that some of these Fan Headers are not true PWM headers, and so the cause the fans to run at some funny % and RPMs instead because they're trying to voltage control the PWM fan instead (or even possibly doing both; voltage and PWM, which might be why some of my tests of the fans kept stalling a while back now I think of it, hmm....). Which is what lead me here looking for alternatives. :)

I agree with you and what ANDARIAL has said, I've allowed my setup to heat up without issues because of tollerances they can safely accept already. Unfortunately, the RIVE as said above already is terrible for fan control. And doing what I can with it already is what lead me to what I have so far, but the primary issue is that I can't set the fans any lower, because then the GPU will overheat and throttle at 90C. I've tried this already. So I just need to get the A14 at the intake under control, but none of the fan headers on the motherboard are suitable for this unfortunately. Whether they're not true PWM, Voltage isn't quite correct for Voltage control or plain can't reach the 50% I need (the Chasis Headers appear to be properly controllable but they min at 60% :mad:)

I think it's just bad luck, the RIVE appears to have been at the border between implementation changes at motherboard makers who began to implement more, some propietary, some half baked, and some proper fan controls, but the RIVE fell into the first two or three categories but not the last one. Speedfan for example doesn't work on RIVE at all, and coupled with the poor BIOS fan controls... doesn't make my life easy here.

Thanks EsaT, I'll have a look at those in a bit. That Quadro looks like what I might need and be after, if it can run the fans below 50% (since none of mine on the motherboard can bar the CPU header) so it can keep the fan at 45-50%, then that's likely what I'll end up with. Will look further into that one to make sure that does what I need. Thanks again. :)

:: edit :: Just read somewhere that it has PWM control between 0% and 100% control, so that's definately capable of what I want. Grabbing it now then. Thanks again. :)

Oh, screenies of this poor mans silent rig later on. :p

Aquaero controllers allow you to setup your own fan curves based on component temps. They can do almost anything but wash windows. While I haven't tried it, I've read of aquaero could even be even programmed to start coffee maker before you wake up using relay output.
 
Back
Top Bottom