Hi Pancho,

I am not sure if i understood your question correctly. However, I try to analyze switching mode PA in different way. First of all impedance looking in to the matching network do not care for large signal operation. And looking in to the PA drain has no constant bias as such. It is always supposed to be switching. So two conditions are there to analyze, when it is ON state and when it is in OFF state. In ON state there will be some RON of the device and in the OFF state parasitic capacitance.

Now if we take an example of say Class-E PA, then even though it might look like a harmonic generator (due to switching gate waveform), it is completely solved in time domain. Basically only fundamental current is allowed to flow even when gate waveform are squared one. And as i mentioned above the parasitc capacitance during off state is tuned out by adding an extra inductance in the series resonant matching circuit.

Bottom line is that i would be OK in defining the required otuput impedance (looking in to the matching network) for switch mode PA.

Regards,
--Vikas

Hi Pancho,

I am not familiar with PAs however here is what I would have done to measure the large signal impedance.

Put in the input of the PA a vsin source operating in the correct frequency and with the large signal amplitude

Run pss

Save current and voltage in the output node

calculate parallel R and C from the imaginary and the real part of the current and the voltage.

I hope I am not out of topic...

Regards

1. You don't need a lossy pad. That will just defeat the whole purpose of using a switchmode PA (efficiency). You have to start thinking in different terms. The desired signal to deliver to the load is the fundamental (typically) of the switching signal. The switching waveform itself does not need to be presented to the BPF. The trick is to come up with networks which insure proper switching at the drain/collector of the power device and delivers the fundamental to the load when the load is 50 ohms.

Class F amplifiers do this by using networks to pass the fundamental and present alternating open/short impedances at the first few harmonics of the switching waveform. If the fundamental network does not transform conviently to 50 ohm in the Class F output network, then an additional network can finish the transformation at the output so long as it does not disturb the impedances setting the harmonic terminations on the switching device.

Class D and, to some extent, Class E amps don't consider impedance matching. These work well for nearly resistive loads over the fundamental and first few harmonics.

2. There are two schools of thought on how to define the output impedance of a large signal nonlinear amplifier. Both agree that the impedance must be measured under large signal operation since the nonlinearity along with the amplitude of driving signal will or can cause a difference in the average impedance seen at each harmonic of the output switching waveform.
1.a Measure the small signal periodic impedance at the output (PSS and PAC) while the circuit is running at the desired output power.
1.b Drive the amplifier to the desired output and use a large signal source at the output to measure the large signal impedance at some offset to the carrier.

I have generally used 1.a, but for class AB circuits only.

3. Large signal stability of switching amplifiers is a tricky subject, but there have been a few researchers who have come up with some ways to analyze and compensate large signal stability of switching PAs.
Almudena Suarez from Spain has a book and a couple of good papers on the topic. I believe the book is on sale right now from the publisher. Her work even deals with analysis and compensation of chaotic instabilities.

Stability Analysis of Nonlinear Microwave Circuits
Almudena Suarez and Raymond Quere
Artech House
ISBN 978-1-58053-303-4
Copyright 2003
Pages: 356

Suarez, A.; Sanggeun Jeon; Rutledge, D., "Stability analysis and stabilization of power amplifiers," Microwave Magazine, IEEE , vol.7, no.5, pp.51-65, Oct. 2006

Feiyu Wang; Suarez, A.; Rutledge, D.B., "Bifurcation analysis of stabilization circuits in an L-band LDMOS 60-W power amplifier," Microwave and Wireless Components Letters, IEEE , vol.15, no.10, pp. 712-714, Oct. 2005

Sanggeun Jeon; Suarez, A.; Rutledge, D.B., "Analysis and elimination of hysteresis and noisy precursors in power amplifiers," Microwave Theory and Techniques, IEEE Transactions on , vol.54, no.3, pp. 1096-1106, March 2006

Hope this helps. It is a fairly deep subject.