PWM fans stuck at 100%

jdcranke07

Honorable
Hey, Peeps. I'm attempting to run Noctua Redux S12B x8 off of a 4-pin fan head hub from Phobya. The header hub has nine 4-pin headers (1 for input, 8 for fans) and it came with a modular/split cable that allows power to be hooked up to a 4-pin Molex and the PWM hooked up to a header on the mobo. I, for some reason, can't get the PWM to work at all. It's either the fans are stuck at 100% or 0% and having the PWM plug plugged into a header or not won't turn them off (since power is coming from the molex). Anyone run into this before? BIOS and SpeedFan settings have no affect from my testing.

Equipment:
MSI X99S XPower Gaming Titanium
Intel i7-6850K
G.Skill Ripjaws4 3000mhz 32GB (4x8GB)
EVGA SuperNova 1200 P2
EVGA 980TI SC Classified (x2)
Asus Xonar Essence STX
 
Solution
OP, you suspect the real problem in your post earlier today.

Look at the mobo manual on p. 48 where the fan header pinouts are shown. For the CPU_FAN header, Pin #2 says "+12V", and Pin #4 says "Speed Control Signal". THIS header is using PWM Mode for control. Now see the labels on ALL of the other headers, Pin #2 says "Voltage Control", and Pin #4 says "NC". All of these headers are using the older 3-pin method called Voltage Co0ntrol Mode, and do NOT put out a PWM signal on Pin #4.

This latter type of header has become very common lately. It is the simplistic way to make a "universal" fan header. It CAN control 3-pin fans because this is the only method that works for them. It also CAN control 4-pin fans because they have a...


Sadly, when I go to Hardware Monitor in the BIOS, it gives me the option to run with SmartFan or Manual, not PWM or DC. In both SpeedFan and BIOS I have set all fans, including my pump to make sure, to manual and nothing has affected the 8 fans yet. I also have two rear 80mm Noctua fans that are affected the same way, but I'm not concerned with them. They are also on a completely different header. I have tried swapping fans around to different headers and I still have the same result in regards to the eight Noctuas I want PWM control over.
 


The fans work perfectly fine by themselves on the mobo headers. I'm wondering what's going on because I did have control of the fans in my last build that I used this hub in and SpeedFan worked very well with it. Now it won't.
 


I'm currently at a loss and I believe everything is plugged in correctly. Would it be possible to have a PWM header on the mobo that doesn't allow for PWM control? In the BIOS, I can turn reduce the RPM of System 2 & 3 and pump headers, but CPU, Rear, and OPTFan1 & 2 I can't so far. Once in WIndows, I can change the RPM of Pump, OptFan 1 & 2, System 2 & 3. If I reduce all fans to 0 rpm using SpeedFan in Windows, the rear fans, the 8 Noctuas on the hub, 1 out 2 fans on the System 3 and OptFan2 are still running at 100%.
 
I currently have two Cooler Master SickleFlow 2000rpm 120mm fans on OptFan2, System 2, and System 3 headers(total of 6). Then I have two 80mm Noctuas on the OptFan1 header, my pump on the CPU header, and my eight Noctuas hooked up to the rear header via pwm fan header hub and molex for power.
 
Hmm, not that I have heard of to be honest. Since why make a 4-pin PWM header with no PWM controll... makes no sense :)

There must be a setting somewhere that should fix it... But sadly I have no experience with MSI motherboards... but from what I read they have the program called MSI Command Center.
Have you tryed that or are you using it right now?
 
OP, you suspect the real problem in your post earlier today.

Look at the mobo manual on p. 48 where the fan header pinouts are shown. For the CPU_FAN header, Pin #2 says "+12V", and Pin #4 says "Speed Control Signal". THIS header is using PWM Mode for control. Now see the labels on ALL of the other headers, Pin #2 says "Voltage Control", and Pin #4 says "NC". All of these headers are using the older 3-pin method called Voltage Co0ntrol Mode, and do NOT put out a PWM signal on Pin #4.

This latter type of header has become very common lately. It is the simplistic way to make a "universal" fan header. It CAN control 3-pin fans because this is the only method that works for them. It also CAN control 4-pin fans because they have a backwards compatibility feature in their design. When a 4-pin fan receives no PWM signal from Pin #4 it cannot modify the voltage supplied on Pin #2 to alter the motor speed. BUT the voltage on Pin #2 is NOT a fixed +12 VDC, as a true PWM Mode system would provide. It is a VARYING DC supply, ranging from 12 to 5 VDC, just as any 3-pin fan header would do. The 4-pin fan motor gets that voltage, whatever it is, un-altered by a non-existent PWM signal, and the motor's speed IS controlled that way.

This design works for both 3-pin and 4-pin fans when only one is plugged into the header. It also works for either fan type if you use a SPLITTER to connect two or more fans to that header. A Splitter merely connects the fans in parallel to the header's voltage supplies. But when using a Splitter, you must consider that the header has a limit of 1.0 amps total current max to all fans connected to it. A HUB is a different device. It gets power for all of its fans from the PSU directly using a SATA or Molex output. It gets the PWM signal from the mobo header it is plugged into, and merely shares that out to all its fans. Then it is up to the fans to use the PWM signal internally to modify the voltage supplied (from the PSU, fixed at 12 VDC) to achieve speed control. Thus a HUB MUST have a PWM signal to share to its fans, and it can ONLY achieve control over 4-pin fans that can use that signal.

You say you have "Noctua Redux S12B" fans which are 3-pin models, but I bet yours are really the "NF-S12B Redux PWM" model with 4 pins. The problem you have is that the Hub (and hence the fans) is NOT receiving any PWM signal from the mobo header, so all those fans are not able to use it to modify the fixed +12 VDC supply coming from the PSU.

The solution is to NOT use a Hub, but change to using Splitters. In many cases this is not quite suitable, but for you it happens to work. Those fans (1200 rpm model) are spec'd at 0.075A each max current, so even 8 of them in parallel will consume 0.6 amps max, which is within the capability of one fan header. If you want to be cautious, get two Splitters and use them to connect only 4 of these fans to each of your SYS_FAN1 and SYS_FAN2 headers. To help, a SPLITTER has an input arm that connects to the mobo header and outputs for fans, but it has NO arm to connect to a PSU output. For your purpose, I suggest a Splitter that looks like a small printed circuit board with several male (with pins) 4-pin fan headers on it and one input cable. Here's an example

https://www.newegg.com/Product/Product.aspx?Item=N82E16882422016&cm_re=Coboc_fan_splitter-_-82-422-016-_-Product

Although the four ports on it are Labelled "CPU" and "Fan1 to 3", plug a fan into the "CPU" header on each one; do not plug your CPU cooler into these. Plug each Splitter into one of your SYS_FAN headers. The all these Ncotua fans will be fed the Voltage Control Mode signals from those mobo headers and their speeds WILL be controlled by those signals.
 
Solution


Thanks, I just learned a bit more about 4-pin fan headers. Yes, you are correct that I forgot to include the "PWM" part of the model number, so all of the Noctua fans are 4-pin PWM capable. Thank you for pointing out the amps that each of the fans consumes. I looked it up and somehow construed it to something else when I tried to remember it earlier. Sadly, each of the CM SickleFlow fans takes ~0.35A, so only two per header (three headers accounted for), the pump has it's own header, the Noctua fans have their own (totaling 5 headers), and then the rear 80mm fans can split with the other Noctua fans or be on a separate one themselves. Again thanks, I appreciate all the help.