transistor for open-drain / open-collector 25kHz PWM

[Harald Kapp]

Who says that in the 1000s of computers you quote the fan is controlled by an open drain transistor as you do? You can easily use a push-pull (totem-pole) transistor stage to drive the fan with a logic level pwm signal. Steve told you say in post #25.

I am absolutely right to think it's a problem with my circuit,

You're absolutely right here. But it is not the fault of the MOSFET being slow as you insist.
Lacking the driving transistor for the high signal (5V), you need to supply a pull-up resistor small enough to ensure a well defined high level. If the built-in pull-up is too high, add an external resistor to compensate. Your experiments have shown that this works.
Or follow Steve's advice and connect the pwm input of the fan directly to the controller's output. There is no need for the additional transistor. The controller's output has a push-pull stage and will happily supply a well defined logic signal to the fan.

 

[tehtehteh]

the intel specs sheet I put in the thread earlier says this, if you have something better please let me know

I refuse to believe they all just work without any standardisation

 

[tehtehteh]

sorry you changed your post as I was typing

 

[tehtehteh]

You're absolutely right here. But it is not the fault of the MOSFET being slow as you insist.
Lacking the driving transistor for the high signal (5V), you need to supply a pull-up resistor small enough to ensure a well defined high level. If the built-in pull-up is too high, add an external resistor to compensate. Your experiments have shown that this works.
Or follow Steve's advice and connect the pwm input of the fan directly to the controller's output. There is no need for the additional transistor. The controller's output has a push-pull stage and will happily supply a well defined logic signal to the fan.

some fans work on 3.3v logic, I don't want to put 5v into them without understanding the damage it could cause

 

[tehtehteh]

"This is leading nowhere. Threaad closed."

trying to shut me up for asking questions, not cool btw

 

[Harald Kapp]

except it goes completely against the specification so it's obviously not correct

This will not be the first part you'll come across in your life that's not up to spec.

some fans work on 3.3v logic, I don't want to put 5v into them without understanding the damage it could cause

Typically you can use a 3.3 V logic signal to control 5V logic as 3.3V is above the threshold of 5 V logic.

trying to shut me up for asking questions, not cool btw

Not for asking questions. For being insulting and not showing any tendency to take the good advice you've been given.
The last ime I re-opened the thread hoping you'd show some insight. This time I'll close this thread for good.