Building input filter to power supply - confused

[jmw]

I'm trying to design a pi-type filter that will go on the input of a power supply that has discontinuous input current (such as buck, buck-boost, or DCM) to block switching harmonics from going back into the power source. I've been using TI's AN-2162 (http://www.ti.com/lit/an/snva489c/snva489c.pdf) as a starting point. I feel pretty good about sketching Bode charts and finding transfer functions and impedances once I have the circuit drawn in front of me. What I'm confused about from reading the note above is: what should (or shouldn't) be included in the filter circuit model?

Here's the model they start with:

Note that Vin is a small-signal source in the power converter from switching noise and Vout is where the DC power source is connected.

My question: shouldn't this include series resistances? The input impedance of the power converter Rin (= (DC value of V_in)^2 / (output power of the converter at the operating point)), and the series resistance Rser of the DC power supply.


Including these has pretty big effects on the transfer function, so I don't understand why they are not included in the analysis and examples from the application note. For example, without Rin, Cd effectively disappears from the transfer function, because from the first schematic, the impedance looking from Vout is (1/sC) || (sL + (1/sCd) || 0) = (1/sC) || sL. (I still get its importance in lowering the impedance when looking from the Vin side)

[HB9EVI]

Usually you use a common mode choke together with an X capacitor and 2 Y capacitor on the input of a SMPS.

see the schematic in https://de.wikipedia.org/wiki/Netzfilter

[jmw]

I would consider one of those but this is a non-isolated DC-DC converter project so there won't be common mode noise as any current that goes in the + terminal has to go out the - terminal.

[HB9EVI]

The usage of a common mode choke is not limited to AC; it has very well effect on the DC-side; I'm using plenty of high current SMPS with similar filters on the DC-side for noise reduction on the transceivers, actually for one I built a setup like in www.dg0sa.de/snt.pdf page 30; it eliminated all noise from the DC