# IS this the correct way to control the speed of a 4 wire CPU fan?

The circuit is used outside of a PC.

simulate this circuit – Schematic created using CircuitLab The fan hums (but still spins) on the MED and LO settings. Does this mean the circuit is the wrong way to control the fan? If this circuit is the wrong way to control the fan, please edit the schematic to make it right.

• Use a variable 1kHz PWM input to control speed – Tony Stewart EE75 Apr 19 at 12:47
• If input is pulled up to V+ then probably an open collector , or a push pull, lots of 555 solutions for PWM , also 4000 series Schmitt trigger relaxation oscillators with offset bias to control PWM duty cycle – Tony Stewart EE75 Apr 19 at 12:53
• THen I would guess hi input is 12V from a pullup R and the % low reduces speed – Tony Stewart EE75 Apr 19 at 13:05
• Does grounding PWM input stop fan? Use a resistor to measure V/R then use an open collector or CE switch to vary duty cycle AT 1kHz. report V and I in your question with RMP response to PWM input. On/off – Tony Stewart EE75 Apr 19 at 15:36
• There’s a hundred ways to do this.. with 4000 series logic on 12V using a 4093, this is one way. Variable speed and sleep input switch electronics.stackexchange.com/questions/330184/…. Other ways use an Op Amp for a positive feedback oscillator – Tony Stewart EE75 Apr 19 at 15:51

There were two basic types of PWM made. One type is an always off type that has a pull down resistor that kept it off when the main voltage is applied to the power wire. The pwm logic high of 5V controlled the on state of the fan. The other type is an always on, or free running which has a voltage divider network internally attached to the pin, and a logic low or grounding this pwm pin turns the fan off. The pwm logic low controls the off time state of the fan. The pwm fan can be controlled several ways. The most inefficient is the current limiting method the OP listed above, because the fan coils will oscillate from low frequency harmonics caused by the reduction of current flow in the fan coils. The basic pwm control is what I call the self idle circuit, that uses its own RPM signal to introduce the logic low for speed control:

simulate this circuit – Schematic created using CircuitLab

The fan's RPM sensor receives about 5V to operate the sensor. When the sensor pulls down the voltage when it conducts, the pwm fan interprets this as a logic low and temporarily tun off the fan until the voltage is above the normal TTL logic high threshold. The way a microcomputer controls pwm is with adjusting an oscillator signal to control this off state. The oscillator based solutions is just simulating what these micro controllers do, but at a less sophisticated methods by using 555 based timers and positive feedback oscillators.

• Depends on the fan because if it is one that uses a transistor for the PWM, it would bias dynamically like that, but if it is a MOSFET in the PWM circuit of the fan, it will be a switch on-off operation and varying a voltage on the PWM doesn't do anything until it switches past the threshold voltage.. – David Mikeska Apr 19 at 20:01