I stand corrected about feed/load impedance and measurement of filter response.
I should have said 'If the feed and load impedances are purely resistive and
reasonably close to the filter design impedance then the effect will be
negligible.' I will try to be more precise in future....
I am sure that the response of the filter in terms of phase, and to a lesser
extent amplitude is not optimum. However, it is a very simple filter and in
several cases it has effected a complete cure without affecting the speed of
the DSL circuit. (In my case ~5Mb.) In pure engineering terms I am sure that a
Bessel's function filter would be better, but it would need many more sections
to achieve a satisfactory cut-off. Additionally, I don't believe that the
perfect filter is possible as the DSL passband can include 160m, so that if
protection is required on that band there must potentially be a hit on DSL
Surely the art of good engineering (if that is not an oxymoron) is to design
equipment that achieves the desired effect with the least possible effort. I
think that the unknown OK station did just that.
RFI mailing list