@owend I understand this fan is "not" PWMable but in practice it works, it is just super noisy. I don't think it changes anything in this case however.
what is your M308 setting for this fan?
Since you asked: M308 S1 A"Hotend" P"temp2" Y"pt1000", but I can reproduce this on a spare Duet2 Wifi board with a more typical M308 S1 P"e0temp" Y"thermistor" T100000 B4138.
In order to prove my point and eliminate the fan and frequency from the equation, I hooked up my logic analyzer to measure the PWM duty cycle (the "fan" here is a simple resistor), and this is what I got:
921a666b-8456-4ad9-8b1e-1fc2d1ed116b-image.png
The config here was M106 P4 S0.1 H1 T20 so it should have been running at 10%, however the firmware did indeed send 50% instead of what I configured. You can also see I reverted back to the default 250Hz frequency.
If I disable thermostatic mode, then the fan control works as expected with M106 P4 S0.1:
aaa52522-cd69-470f-8139-cda4aa594264-image.png
(the duty cycle is inverted so 90% here means 10% in RRF.)
So I think this proves there is an undocumented limitation (or hopefully just a bug) with thermostatic mode and PWM cannot go lower than 50% in that mode.
@o_lampe that's not a bad idea, assuming they can start turning at a low voltage. But I suppose getting better fans is the real solution. 🙂
Cheers,
Ben.