Re: PA phase shift changing with amplitude
If you lowpass filter that distorted waveform....you will have a phase shift, as I indicated. Think of the Fourier analysis of the waveforms that make up the distorted one, and you will see that the fundamental has shifted.
As I said, the matched coil (antenna if you want to call it that) connects to the cathode of D1. At my operating point, the lumped element pi network there approximates a quarter wave line. It works fine as a narrowband TR switch, but thought I'd include it in the schematic anyways.
In the next version I build I'm planning on having the two stages separated but on the same board, so I'll match both their inputs and outputs to 50 individually. Should make things more straightforward.
Yeah I'll probably use the chokes for that purpose just to save space.
Could you elaborate a bit more on this please?
Yes that makes sense, noted.
You mean something like using a 1:4 impedance transformer so my 50ohm load looks like 12.5ohm (or whatever)? Yeah I'd have to increase Id, but then I could probably lower my drain voltage as well.
The only simulation software I'm competent with is LTspice, which doesn't have nonlinear models for any RF FETs, and I've never been able to find a nonlinear spice model for any of the FETs I use. I tried ADS for a bit but was pretty much overwhelmed by it (though they at least had a model for the MRF136 which I could use, and possibly there are more libraries out there).
I need 6db difference between each pulse within one pulse sequence, but I will need a decent overall range of output power. So I might need to change between 33dB and 39dB, or 28dB and 36dB. Sorry for the confusion. Overall I'd be happy with a working dynamic range of 20dB.
See reply above.
Regarding modelling and simulation.
When you visit Polyfet Home Page try to locate a device the looks like yours. The provide Pspice model files for virtually all of their devices. The files for eagle are text files, so you copy them into your favorite simulator (and modify parameters / syntax).
Start with some simulation to get the datasheet specs (DC transfer curve, capacitance versus voltage, etc). If there are significant differences, you need to modify some parameters (can be time consuming). Last year I used a Polyfet SK702 device model for a feasibility problem.
Regarding impedance transformation, you can do that with a L network (series L, parallel C to load) or CLC Pi network, so you aren't fixed to standard impedance steps as known for transmission line transformers.