Class E at double the RF frequency

[AB2EZ]

I have a Class E transmitter with a pair of RF modules operating in push pull... i.e. one of the standard/popular configurations.

I was thinking about the following... and I am wondering if anyone has tried this:


The transmitter is currently set up to operate on 75 meters.

Suppose I drive it with a 3.65 MHz (approximately) input signal... but I add the two RF outputs, rather than subtracting them. I.e. I reverse one of the output transformer windings.

The new output will be a single-ended series of Class E pulses at 7.3 MHz (approximately).

I could easily modify the output tank circuit to be resonant at 7.3 MHz. [E.g. use half of the output coil and half the tuning capacitance]

To obtain the 3.65MHz drive signal, I could (as an option) run the 7.3MHz (40m) output of my transceiver through a 2:1 frequency divider at the input of the Class E transmitter. That would keep the class E transmitter and the receiver in the transceiver on the same 40m frequency.

As far as even harmonics of 7.3MHz are concerned... since I would no longer have a push-pull configuration... I would have to take steps to keep the transmitted level of even harmonics  under control. In addition, since the two out-of-phase rf modules would not be putting out identical signal levels (there would always be some amount of difference), there would be some amount 3.65 MHz power in the combined signal... and that, too, would have to be managed. The selectivity of my antenna tuner, in combination with my 40m dipole antenna, is what I would use to manage the harmonics and the 3.65MHz sub-harmonic.  

Stu

[WBear2GCR]

obviously, if the input signal is driving both "sides" in phase, you will get canceling at the output side, and a lot of heat... I take it you are thinking of a waveform like 120Hz full wave rectification from 60 hz.?  otoh, if you are thinking like square waves, as in digital, it seem to me that you'd have to shorten up the duty cycle to fit the out of phase component inbetween the inphase component in order to make the frequency double... or so it seems to me...

Interesting idea... seems to me that IF you can make this "alternating doubling" work, you make two halfs, each doing the doubling, then run those two halfs (each one then an altered P-P stage) into a PP output to kill the second harmonic as before. More or less a sort of bridged amplifier I suppose...

Might work out to be "band switchable" between any two bands?



                   _-_-bear

[WA1GFZ]

push push doubler, all you need to do is reduce the drive duty cycle and it should work

[W3RSW]

with reduced efficiency I think?

[kb3ouk]

Ok, refresh my memory: A push-pull doubler works by having the input circuit tuned to the drive signal's frequency, and the output is tuned for the doubled frequency, right?

[AB2EZ]

After thinking about this for a few more days... I decided that it was worth trying.

Operating normally (push pull), each cycle of the drain voltage waveform in each of the two, 180 degree out-of-phase, 2-FET modules (i.e. viewed individually) has a textbook class E pulse shape that is narrow enough to suggest that the waveform contains a significant amount of 2nd harmonic. [Remember, even harmonics get cancelled out, in the output combining transformer, when you subtract one of the two 180 degree out-of-phase class E waveforms from the other]. Using the FFT computation/display feature of my scope... each of the two drain voltage waveforms (i.e. viewed individually) has a 2nd harmonic component that is around 6dB down from the fundamental component.

I reversed the output transformer winding of one of the two modules, so that the two outputs would add (rather than substract), and I adjusted the input (drive) signal frequency to 3.650 MHz.

The resultant output was at 7.3 MHz; but only about 40 watts (after adjusting the series tuned circuit and the loading capacitor for maximum output). The output in the normal (push pull) configuration is around 170 watts.

The drain voltage waveform had only about half the peak amplitude that it has in the normal configuration; and the drain current was only about half as much as it is in the normal configuration.

It appears that there may be too much drain-to-source capacitance to allow the drain voltage to build up. I didn't try removing or reducing the values of the shunt capacitors (which are currently 500 pF for each of the 2-FET modules)

Stu

[kb3ouk]

500 pf would be too much for 40 meters, try again with only 250 pf for the shunt capacitance.

[AB2EZ]

Based on the results of my experiment (described above)... and all things considered... it is probably easier (or just as easy) to build a separate 40m RF deck (-:

Stu