Mains Filter (Pi) Question

[weinstro]

Hello,

I'm in the process of building a DIY RFI filter for main power.

The filter contains both common mode and differential filter sections, with the differential section configured as a Pi filter (C-L-C).

When I model the equivalent impedance of this filter, as a function of frequency, I find that there are a number of poles and zeroes associated with this filter. At certain frequencies, the impedance is limited only by the DCR of the inductor. Large currents will flow if there is any significant noise present at these frequencies.

The phase fluctuates quite a bit as well. Unloaded, the model predicts the phase angle <-180 degrees at some frequencies. My recollection from control theory is that this indicates instability.

I can "stabilize" the impedance magnitude and phase curves by adding damping resistors in series with the capacitors. This increases the magnitude of the filter impedance by ~2X, however, but does eliminate the wild swings.

So, the question is, should I use these damping resistors or not?

Thanks and regards,

Rob