Use of NU2B "Class E" calculator

[Brian K1LI]

I have a question about the results from this program.

Using an ideal voltage-controlled switch in LTSpice, I simulated Bobby's example of a 75m KW http://classeradio.com/tools/nu2b/PushPull_KW.doc unit on the Design Tools page http://classeradio.com/tools/nu2b/nu2b_download.html of ClassERadio.com. Everything looks good, except that the full output power seems to be drawn twice - once from each DC supply - which cuts the efficiency roughly in half.

Am I looking at this the wrong way?

[Brian K1LI]

As usual, the fault lies with the operator. I didn't properly understand how to model a transformer in SPICE. For others who may want to try this, one "links" a pair of inductors - whose turns ratio is the desired impedance ratio - using the "K" SPICE directive.

For example, to achieve a factor of four (4) step up in impedance, draw two inductors with L1=1 and L2=4 (yes, that's not a misprint: no "nH" or "uH"!) and insert a SPICE directive stating "K L1 L2 1" for 100% mutual coupling. You'll notice on the schematic that phasing dots appear on each inductor to signify the "sense"; either inductor can be rotated to change the sense, as required.

This can be done with any number of inductors, not only with two.

I have not tried lower mutual coupling ratios to see how SPICE deals with leakage inductance. It's also worth noting that LTSpice adds a default 1milliohm resistance to all inductors to improve simulation efficiency.