@c75e3dac
Also, we can cheat the Nyquist-Shannon sampling theorem by rather than detecting and interrupting them, by triggering the crossing. Thus, we need hardware detection and not software detection.
For the same reason the #Klipper docs mentiones its PWM isn't the best.
@5756e48c #3dprinting
@c75e3dac
Good question and correct initial assumption!
No, I can't. The dimmer I use for my limmers are called TRIAC dimmers. AC signals have alternatinf current and voltage. The current is a sine wave and voltage a cosine wave. Thus, they are always out by 90°. The TRIAC detects when the current signal crosses the zero line and then only changes the PWM voltage. This removes flickering and is more energy efficient.
Detectong this switch must be done at 2x 60Hz due to the Nyquist-Shannon sampling theorem. This detection is done with a hardware interrupt pin to allow the software to run faster. The #RaspberryPi and #Klipper do not support this.
So I use a separate peocessor, which does allow this and communicate it back to Klipper and basically only gets the PWM signal and then does the rest.
@5756e48c #3dprinting
@c75e3dac
Good question and correct initial assumption!
No, I can't. The dimmer I use for my limmers are called TRIAC dimmers. AC signals have alternatinf current and voltage. The current is a sine wave and voltage a cosine wave. Thus, they are always out by 90°. The TRIAC detects when the current signal crosses the zero line and then only changes the PWM voltage. This removes flickering and is more energy efficient.
Detectong this switch must be done at 2x 60Hz due to the Nyquist-Shannon sampling theorem. This detection is done with a hardware interrupt pin to allow the software to run faster. The #RaspberryPi and #Klipper do not support this.
So I use a separate peocessor, which does allow this and communicate it back to Klipper and basically only gets the PWM signal and then does the rest.
@5756e48c #3dprinting
Oddbean new post about login | logout
Notes by Stark | export