So, thanks to many threads here, I'm starting to figure out this whole fan control thing. I'm in the middle of going full-Noctua on the cooling of my little Dell XPS mini-tower. Replacing the case fan and the power supply fans with Noctura 92mm fans and bolting a 92 mm intake fan to the mounts already in the case, etc.
This, of course, led to figuring out my options for connecting the fans. This Dell motherboard (and I suspect many Dell motherboards) has two fan connector headers. One is called the System Fan and is a 3-pin header. The other is called CPU Fan and has a 4-pin header. Both fan connectors are variable speed. The 4-pin header is PWM controlled. The 3-pin headers are controlled by variable voltage on the +12 volt wire. In the stock configuration, the rear chassis fan is connected to the 3-pin and the CPU cooler fan is connected to the 4-pin. Both vary based on temperature -- some dark Dell BIOS secret but probably CPU temp (the system reports both core temps from the CPU and temps from a sensor under the CPU heat sink).
Decision 1: Noctua Flex 3-pin fans or PWM 4-pin fans. The PWM 4-pins have two big advantages. First, they can be used with either the 4-pin header for PWM control or the 3-pin header for voltage control. Second, they have a higher top rpm and therefore a bit higher speed when the computer is at or near idle. IMO, this choice is easy. Go with the Flex 3-pin Noctuas for fan that will be connected to a Molex and left at a fixed speed. Go with the PWM Noctuas for any system with temperature based fan speed control. I don't want my case fans to be nearly stopped 90% of the time. I want them moving a little air!
Decision 2: Is there any real difference between the speed control on the 3-pin and 4-pin headers? To find out, I hooked up two identical 92mm Noctua PWM fans, one to the 3-pin header and one to the 4-pin header. I made sure that each of these was the ONLY fan sending tach info back to its header, so SpeedFan would give an accurate speed. It turns out that the voltage control of the 3-pin header and the PWM control of the 4-pin header are essentially identical on the Dell. They are clearly getting their underlying control from the same temp readings and algorithms.
Here's a 30 minute graph of the two fans at or near idle:
And, one with a couple bouts of driving the CPU hard with Prime95:
The green trace is the fan driven by the 4-pin PWM header. The red trace is the fan driven by the 3-pin variable voltage header. The traces are identical in shape, and just slightly offset in fan speed, with the PWM fan being a negligible bit faster (40 to 90 rpm).
My theory is that Dell (or AMD) drives the 3-pin header with exactly the same kind of circuit built into PWM fans that varies the voltage based on the PWM signal. I am certain that both headers are driven from the same underlying PWM control signal, with the fan version driving to a slightly higher voltage than the motherboard version on the 3-pin headers.
So, for me, it really doesn't make a darn bit of difference which headers I use. With the same fans, the results will be the same.
This, of course, led to figuring out my options for connecting the fans. This Dell motherboard (and I suspect many Dell motherboards) has two fan connector headers. One is called the System Fan and is a 3-pin header. The other is called CPU Fan and has a 4-pin header. Both fan connectors are variable speed. The 4-pin header is PWM controlled. The 3-pin headers are controlled by variable voltage on the +12 volt wire. In the stock configuration, the rear chassis fan is connected to the 3-pin and the CPU cooler fan is connected to the 4-pin. Both vary based on temperature -- some dark Dell BIOS secret but probably CPU temp (the system reports both core temps from the CPU and temps from a sensor under the CPU heat sink).
Decision 1: Noctua Flex 3-pin fans or PWM 4-pin fans. The PWM 4-pins have two big advantages. First, they can be used with either the 4-pin header for PWM control or the 3-pin header for voltage control. Second, they have a higher top rpm and therefore a bit higher speed when the computer is at or near idle. IMO, this choice is easy. Go with the Flex 3-pin Noctuas for fan that will be connected to a Molex and left at a fixed speed. Go with the PWM Noctuas for any system with temperature based fan speed control. I don't want my case fans to be nearly stopped 90% of the time. I want them moving a little air!
Decision 2: Is there any real difference between the speed control on the 3-pin and 4-pin headers? To find out, I hooked up two identical 92mm Noctua PWM fans, one to the 3-pin header and one to the 4-pin header. I made sure that each of these was the ONLY fan sending tach info back to its header, so SpeedFan would give an accurate speed. It turns out that the voltage control of the 3-pin header and the PWM control of the 4-pin header are essentially identical on the Dell. They are clearly getting their underlying control from the same temp readings and algorithms.
Here's a 30 minute graph of the two fans at or near idle:
And, one with a couple bouts of driving the CPU hard with Prime95:
The green trace is the fan driven by the 4-pin PWM header. The red trace is the fan driven by the 3-pin variable voltage header. The traces are identical in shape, and just slightly offset in fan speed, with the PWM fan being a negligible bit faster (40 to 90 rpm).
My theory is that Dell (or AMD) drives the 3-pin header with exactly the same kind of circuit built into PWM fans that varies the voltage based on the PWM signal. I am certain that both headers are driven from the same underlying PWM control signal, with the fan version driving to a slightly higher voltage than the motherboard version on the 3-pin headers.
So, for me, it really doesn't make a darn bit of difference which headers I use. With the same fans, the results will be the same.
Facebook
Twitter
Instagram