Thank you W9IQ. I always appreciate your review. I did assume a voltage drop for the input power sense diode but I don't recall actually measuring it.
This is a look at how the 2KL works in the time domain. It departs from the classic behavior of class AB. The classic AB amplifier behaves as shown in this University of Pennsylvania article.
https://www.seas.upenn.edu/~ese319/Lecture_Notes/Lec_22_ClassAB_Amplifier_12.pdf
Below is the input to the input transformer of PA1 with the 2KL set for 500 watts output on 160 M. There is some ripple but as we shall see, that ripple is likely created by the very non-linear junction behavior of the power transistors.
This is what is found at the transformer outputs. There is obvious clipping as the base emitter junctions of the power transistors clamp the input. The purple trace uses the math function of the oscilloscope to do A-B of both channels with the sum bearing a strong resemblance to the input signal.
This is the output measured on both collectors of each device on PA1. Notice that both device’s collectors are driven into hard saturation long before 500 watts output is reached.
The power supply voltage for the 2KL power devices is +40 VDC. The reason the collector voltages rise above +40 VDC is during the time that the devices are turned off, the output transformer acts similarly to a heavily loaded flyback transformer. What is interesting is that it is the flyback voltage created by the output transformer and the possibly the following low pass filter that tends to negate the distortion caused by collector saturation. The purple trace is the A-B sum which shows substantially less distortion than one might expect from just looking at the collector voltages.
It is also easy to see in the time domain that even with matched devices, there is clearly a gain difference at RF frequencies for the device with the blue trace having greater gain than the device with the yellow trace.
