Hi, Elements
What you read about the M6E is also true of all the other Asus boards. They have the feature that either 3 or 4 pin fans can be connected to the fan headers. This is usually handy, but can cause trouble on at least one instance. When there are several fans to control together that exceed 1 amp current total, I tried to use a splitter that drew power from a PSU molex and used the MB header only for control and tachometer. That only worked on the true PWM header - the CPU fan header.
The CPU Fan header and CPU Optional header are truely PWM when 4-pin fans are plugged in to the CPU header. The board detects the difference and uses voltage control on both headers when a 3-pin fan is in the CPU fan header. The CPU and CPU_OPT headers respond to the CPU temperature sensor under the socket.
The three chassis fan headers also respond to the CPU temperature sensor. To get 3/4 pin fan interchangability, the chassis fan header PWM control is held in the ON state and voltage is varied. In this way, they are fully controllable. This is supposed to work okay for both types, but some have reported control voltage so low that some PWM fans won't start. That's only a matter of setting the minimum fan speed higher.
The three Optional fan headers each respond to the corresponding Optional temperature sensor, rather than to the CPU temperature. They are also 3/4-pin interchangeable, ie. fake PWM. I use a water temperature sensor and optional fan header to control my radiator fans. Those fans are PWM, but work fine with the voltage control. They total 0.8 amps current, so I can use one header for both fans.
Al the headers are controllable in BIOS. Fan Expert 2 covers only the CPU, which is mirrored by the CPU_OPT, and chassis fan headers. I rid myself of the OS burden of Fan Expert long ago. Needing BIOS to set Optional fan profile is no problem for me since I already use BIOS for the chassis fans.
The point is that, fan headers other than CPU may not be true PWM, but they control the fans.
Jeff
What you read about the M6E is also true of all the other Asus boards. They have the feature that either 3 or 4 pin fans can be connected to the fan headers. This is usually handy, but can cause trouble on at least one instance. When there are several fans to control together that exceed 1 amp current total, I tried to use a splitter that drew power from a PSU molex and used the MB header only for control and tachometer. That only worked on the true PWM header - the CPU fan header.
The CPU Fan header and CPU Optional header are truely PWM when 4-pin fans are plugged in to the CPU header. The board detects the difference and uses voltage control on both headers when a 3-pin fan is in the CPU fan header. The CPU and CPU_OPT headers respond to the CPU temperature sensor under the socket.
The three chassis fan headers also respond to the CPU temperature sensor. To get 3/4 pin fan interchangability, the chassis fan header PWM control is held in the ON state and voltage is varied. In this way, they are fully controllable. This is supposed to work okay for both types, but some have reported control voltage so low that some PWM fans won't start. That's only a matter of setting the minimum fan speed higher.
The three Optional fan headers each respond to the corresponding Optional temperature sensor, rather than to the CPU temperature. They are also 3/4-pin interchangeable, ie. fake PWM. I use a water temperature sensor and optional fan header to control my radiator fans. Those fans are PWM, but work fine with the voltage control. They total 0.8 amps current, so I can use one header for both fans.
Al the headers are controllable in BIOS. Fan Expert 2 covers only the CPU, which is mirrored by the CPU_OPT, and chassis fan headers. I rid myself of the OS burden of Fan Expert long ago. Needing BIOS to set Optional fan profile is no problem for me since I already use BIOS for the chassis fans.
The point is that, fan headers other than CPU may not be true PWM, but they control the fans.
Jeff
I favor water cooling, obviously. Form a Hyper 212 EVO air cooler to the water cooling, I got a 14C temperature drop at high load (same CPU and TIM, not delidded running the same stress test)). Looking at it the other way, I could crank up the overclock multiplier and voltage until it pulled 20 watts more power for the same temperature. I used 93C temperature just for this test, not 24/7.
I power the pump directly from a PSU molex. The D5 exceeds the 1 amp limit of the MB fan headers. The pump tachometer is connected to a fan header for monitoring. Two radiator fans run through a splitter to an optional fan header that responds to a water temp sensor. My case has three chassis fans. That's five headers. If I were to add a RAM fan, I'd use a temperature sensor stuck in one of the heat spreaders and an optional header controlled by that sensor. That's six headers, including two with added temp sensors. I think the optional temp and fan headers preclude the need for an external controller. If you have multiple radiators for the added load of GPUs, you may need more fans.
Jeff
I power the pump directly from a PSU molex. The D5 exceeds the 1 amp limit of the MB fan headers. The pump tachometer is connected to a fan header for monitoring. Two radiator fans run through a splitter to an optional fan header that responds to a water temp sensor. My case has three chassis fans. That's five headers. If I were to add a RAM fan, I'd use a temperature sensor stuck in one of the heat spreaders and an optional header controlled by that sensor. That's six headers, including two with added temp sensors. I think the optional temp and fan headers preclude the need for an external controller. If you have multiple radiators for the added load of GPUs, you may need more fans.
Jeff
1. Yep, that's what I'm saying
2. I can't tell any difference. Some have reported problems getting PWM fans to start on voltage control when set for really low rpm, but I haven't seen that. My minimum speed is 48%. I would say that the radiator fans are more quiet when driven by water temperature because they don't speed up with any short term CPU power increase. Water changes temperature slowly even when cooling at high power loads. RPM, and therefore noise, increase slowly and only as needed to cool the water. That has to do with what is sensed rather that any difference between voltage or PWM control.
3. The D5 pump came mounted in an EK disk-bay reservoir. It may be common for all D5's, but maybe not. It has two cables: power with a molex for connection to a PSU, and a single wire to a 3-pin fan plug. That single wire is the tach, is wired to the third pin and is picked up by monitor SW in BIOS and Windows. I've tried it on CPU_OPT, Chassis and Optional headers and all of them read the tach.
4. The Optional headers respond only to their corresponding sensor. The two-pin sensor headers are located right beside the corresponding fan headers. I suggest you download the M6F manual from the Asus support site. The layout diagram shows where these are on the board.
5. The water sensor I use is a fitting adapter.
http://www.frozencpu.com/products/21218/ex-tub-1992/Phobya_G14_Inline_Temp_Sensor_Coupler_-_2-Pin_-_...
The attached cable plugs right in to a sensor header. BIOS temp monitor and fan control tabs pick it right up.
That's not the only water temp sensor around. The M6F manual doesn't say explicitly, but the sensor has to be a 10K type.
6. There are pre-defined profiles and a 'manual' setting that will take a two-point curve about like Fan Expert.
7. When using a splitter, the power and PWM control are connected to all fans. The tachometer signal from one of the fans is returned to the header for monitoring. I have used 2- and 3-way splitters with BIOS control and all connected fans worked. Just be sure the total current from any one header is less than 1 amp.
I have also used a 3-way splitter with power directly from a PSU molex. In that case, only true PWM control will work. That splitter has a 4-pin fan plug with only the control and tach wires. Control from the header and power from the molex go to all fans. One fan's tach is fed back to the header.
Jeff
2. I can't tell any difference. Some have reported problems getting PWM fans to start on voltage control when set for really low rpm, but I haven't seen that. My minimum speed is 48%. I would say that the radiator fans are more quiet when driven by water temperature because they don't speed up with any short term CPU power increase. Water changes temperature slowly even when cooling at high power loads. RPM, and therefore noise, increase slowly and only as needed to cool the water. That has to do with what is sensed rather that any difference between voltage or PWM control.
3. The D5 pump came mounted in an EK disk-bay reservoir. It may be common for all D5's, but maybe not. It has two cables: power with a molex for connection to a PSU, and a single wire to a 3-pin fan plug. That single wire is the tach, is wired to the third pin and is picked up by monitor SW in BIOS and Windows. I've tried it on CPU_OPT, Chassis and Optional headers and all of them read the tach.
4. The Optional headers respond only to their corresponding sensor. The two-pin sensor headers are located right beside the corresponding fan headers. I suggest you download the M6F manual from the Asus support site. The layout diagram shows where these are on the board.
5. The water sensor I use is a fitting adapter.
http://www.frozencpu.com/products/21218/ex-tub-1992/Phobya_G14_Inline_Temp_Sensor_Coupler_-_2-Pin_-_...
The attached cable plugs right in to a sensor header. BIOS temp monitor and fan control tabs pick it right up.
That's not the only water temp sensor around. The M6F manual doesn't say explicitly, but the sensor has to be a 10K type.
6. There are pre-defined profiles and a 'manual' setting that will take a two-point curve about like Fan Expert.
7. When using a splitter, the power and PWM control are connected to all fans. The tachometer signal from one of the fans is returned to the header for monitoring. I have used 2- and 3-way splitters with BIOS control and all connected fans worked. Just be sure the total current from any one header is less than 1 amp.
I have also used a 3-way splitter with power directly from a PSU molex. In that case, only true PWM control will work. That splitter has a 4-pin fan plug with only the control and tach wires. Control from the header and power from the molex go to all fans. One fan's tach is fed back to the header.
Jeff
I haven't seen a way to change CPU/chassis header from responding to CPU temperature.
Sensors for the optional headers are separate. Asus e-store used to list them - a three-pack - but now appear to be sold out. Frozen CPU (the site I linked before), Newegg, Amazon etc. might have such. As I said, you want the 10K type.
Jeff
Sensors for the optional headers are separate. Asus e-store used to list them - a three-pack - but now appear to be sold out. Frozen CPU (the site I linked before), Newegg, Amazon etc. might have such. As I said, you want the 10K type.
Jeff
