I have a circuit that controls a fan linearly. I have a PWM output going through a low pass filter and into an NPN transistor for low side control.
It worked well with at first, however when I switched to a different fan there was a lot of pulsing a lower PWM. The fan starts spinning, and then a some point it turns off and then starts spinning right away. When I look at the collector voltage (Node A) with a scope it looks like the signal in image 2.
Because I don't want the fan at full speed I expect the voltage across the transistor to be high, and get to 0 when the fan is fully on. What is happening instead though, as the fan starts spinning the voltage across the transistor gradually increases (and so does fan speed) and then when it hits the peak it turns off and starts again. The higher the PWM the slower the pulsing interval. At some point though when the PWM is high enough the fans spins normally, but I would like it to spin at a lower PWM than that as well.
The solution that I have for this at the moment is kind of ugly. I added a capacitor across the fan and it seems to fix the problem. Not exactly, but it allows me to get to low enough PWM that I can use it.
It only masks the issue and it's also a massive capacitor. The capacitor stops the pulsing and keeps the voltage across the transistor constant. I have tried adding a diode instead of the capacitor but the issue still persists. This is what the signal looks with the capacitor added.
My primary goal is to understand what exactly is happening, and then obviously how to fix it.
The fan itself is a 12V, 3 wire, brushless fan, 0.33A rated current. It has auto restart and polarity protection. I tested with a power supply by ramping up the voltage to make sure it's not the fan and it works fine.