GregAC
commented
3 years ago I believe you're right. If you have silence in the audio if it produces a constant level in the middle of the range (0 if you have a signed range, translating to a level of 128 with the unsigned representation). So we'll end up with a constant voltage out of the RC circuit.
Analogue electronics is not my forte so I just based my circuit on other simple PWM audio examples you can find online with a little tinkering. As I mention in the article the 'Hardware Design with RP2040' document (https://datasheets.raspberrypi.com/rp2040/hardware-design-with-rp2040.pdf) provides a circuit that is likely better. It utilities a second capacitor, perhaps this helps remove the DC offset?
I'll experiment with a similar second capacitor in my design and see what happens when I get a chance.
prasadsangvikar
In your article (https://gregchadwick.co.uk/blog/playing-with-the-pico-pt3/), it seems that your circuit will leave a DC component at the speakers. It seems to work but the constant current flow and force on the membrane might stress the speakers (silence will be at 0.4V instead of 0V). Was that intended (or am I just wrong)?