Re: LF: frequency doubler and divider

["James Moritz" <james.moritz@btopenworld.com>]

Dear Graham, Stefan, LF Group,

----- Original Message ----- 
From: "Graham" <g8fzk@g8fzk.fsnet.co.uk>

> Has anyone  looked  closely  at  Jim's  Class  D/E  EER   tx ?
Well, Jim has ;-) All that is used is a limiting amplifier to "square up" 
the input sine wave. For a push-pull PA, anti-phase square wave gate drive 
is obtained simply using an inverter. There is no requirement to multiply or 
divide the frequency. To obtain a 50% duty cycle square wave, the limiting 
thresholds need to be symmetrical around the average value of the input 
waveform. In the EER transverter, I used a transistor differential pair to 
provide a well-defined symmetrical limited waveform. The output of the 
differential amp is fed through some CMOS gates to increase the gain and 
make the output "square-er". You could use a voltage comparator instead, but 
it has to be a fairly fast one to avoid significant asymmetry in the output 
waveform due to differences in rise/rall times with positive and negative 
going signals, and other effects.
This works fine provided your signal source generates a symmetrical 
waveform, i.e. the wanted signal frequency with or without odd order 
harmonics, such as a sine wave or a square wave. If the waveform is 
asymmetrical (containing substantial levels of even-order harmonics), such 
as a rectangular waveform with non-50% duty cycle, or contains non-harmonic 
components (e.g. unwanted mixing products), then a bandpass filter is needed 
to remove the unwanted components. But most practical signal generation 
schemes do not contain very high levels of unwanted frequency components 
close to the signal frequency, so a simple single LC tuned circuit with low 
Q is an adequate bandpass filter.
Cheers, Jim Moritz
73 de M0BMU