obcd wrote:It's not totally BS.
The protection diodes are indeed misplaced.
If the audio level exceeds 0.7V, the protection diode to ground will start to conduct and will cause a distortion and dc offset in the signal.
True, but you forget that with the current resistor division network (270R & 150R) divides the 3,3V PWM signal to a 1,18V PWM signal! So the maximum PWM output swing would be 0,59V DC with plus or minus 0,59V swing around that DC level, in other words, unless a negative DC level is already introduced on the output, (very unlikely) after the DC cap you will only have 0,59V going below GND, not enough to make the diode conduct, even at maximum audio level! So no exceeding of the -0,7V below GND, thus no distortion, and thus no DC level shift.
Also, as a side note, the diodes are not there to clip outgoing signals (to protect what is connected to the PI's audio output), but to protect the PI against any signals that somehow enter
the PI through the audio port!
Still placing the diodes on the other side of their DC blocking capacitors would not have harmed, and you could argue it would be "better", as the DC level would be well above ground, giving more "headroom", but then, the levels are designed such that that headroom is unnecessary.
Its still not a "big design error" though, as the wiki was making it out to be.