Neon Filtering

Hi all,
        I spent half a day having a further look at all this. I in-
vestigated the properties of three circuits. Before describing them, 
I'd welcome input from others on any of these schemes.

    Characteristics : with R set so that damping was critical 
(Butterworth response), the response fell off gradually and definitely
towards a zero.
    Advantages : Simplicity, damping resistor will also help damp the
                 tuned circuit formed by the transformer inductance
                 and primary capacitor.
 Disadvantages : Transformer power is lost in the resistor as well as
                 the RF. Also, no damping between the filter cap and
                 transformer reactances.

         |                |        |
         |                |      -----
         -----\/\/\/\/-----      -----
   Chracteristics : At high frequencies, response falls off more
                    gradually than circuit one and tends to give
                    an R-C response (single pole) as the choke
                    becomes effectively left out of the picture
   Advantages : No transformer power is wasted in the resistor
 Disadvantages : Not as effective as (1) at suppressing RF having
                 a response more akin to an R-C filter. Also, no
                 damping applied to the Cp/transformer circuit.
 (This is a distinct disadvantage when used with a neon transformer.
  I think this circuit would be almost ideal for a pole-pig system
  where high primary powers are being used).
 Characteristics : Standard 2 pole response with resistors adjusted 
                   for critical damping.
 Advantages : Some damping for power components as well as RF
              Resistor in series with filter cap also helps damp any
              reaction between this component and the transformer
              secondary at high frequencies.
 Disadvantages : More components than the other two circuits.
                 Some 50/60 Hz power wasted in the series resistor.

There are other variations on these (e.g. resistor across the 
capacitor). I didn't test them all because some disadvantages were
immediately obvious. For example, a resistor in parallel with the
filter capacitor will have the transformer voltage impressed across
it continuously and will need suitable voltage ratings. Resistors
in series with chokes suffer the same thing. All circuits can roll-
off much faster but at the expense of peaking (ringing, Tchebyshev 
response) which is what we don't want. Personally I'd pick circuit 3 
as being most useful for a neon supply, but I'd like to hear of 
other's experiments/experiences with this.
    Finally, I found that the transformer secondary had little effect 
on the lower RF filter response in general.