Roelof,
yes, 20-35 Ohms is a (very rough) approximation for the source impedance of the PA (12V), as seen by the filter. But as a nonlinear system (class D), the usual way of (conjugate) impedance matching is not applicable.
Anyway, I gave up trying to simulate PA+filter: the results deviated too much from reality.
Next time I will try 6xBS170...
73, Gottfried
