Odds are that your Audi uses pulse modulation on its headlight-sense wire.

The empeg player software handles this with some fancy code that watches the headlight-sense line and smooths out the pulses, then makes an educated guess as to the state of the headlight line.

My bet is that Mark's code doesn't do the same thing. A while back, he tried doing some debounce logic on the sensing code, and it broke the knob-illumination for me. So he took it back out.

I don't know how the code is handled, and I don't know if the empeg's dimmer-sense logic is in the kernel or if it's in the player. So I don't know if Mark can benefit from looking at how the empeg guys handle it. But I think that their implementation is the most robust as far as I can tell.

In the meantime, does anyone know of a simple circuit (possibly involving a capacitor) that can be soldered into the headlight-sense wire and will take care of the pulse modulation problem Lopan is having? It would need to be able to do this without screwing up the pulse on the "car" side of the connection because his dash lights would still need to function normally.