[Mondial]

Morzh, what you are in effect describing is a resonant circuit. For the inductive current to equal the capacitive current and cancel, their reactances have to be equal but out of phase.

This is exactly the situation in a parallel resonant circuit where the L and C reactances cancel and the impedance at the resonant freq ideally approaches infinity. Since this parallel resonant circuit is in series with the DC supply, a trap is formed and the 120 Hz component is not passed.

It really works pretty well. The 118 power supply uses a separate choke and resonating cap to trap the ripple freq. I actually connected a scope to both sides to see the ripple reduction. On the input side there was about 60 V P-P of ripple while at the output it was reduced to a fraction of a volt. Changing the cap value affected the null, so with the exact specified value what was left was only 60 Hz and higher harmonics, the 120 Hz component was totally gone!