In message <[email protected]>, g3kev
<[email protected]> writes

If you read the literature published by one of the researchers/experimenters
of class D/E amplifiers, he states that if you push the amplifiers, which
most amateurs do, and do not pay attention to the LPF, you are then radiating
a distorted square wave, with all its odd order harmonics, he goes on to say
that for larger than normal antenna systems one would need possible more than
the normal 5 pole filter between the mosfets and the antenna.
I own a quality spectrum analyser and can verify what he says to be correct.
Very few signals generated by such amplifiers and used by radio amateurs have
a pure sine wave output after one 5 pole LPF, except at very low power
levels, also have a look at the ripple content super imposed on the signal
because the power supplies normally used by radio amateurs are not regulated
or well smoothed. .
With your modest amplifier and long wire a few feet above ground you have
nothing to worry about.
G3KEV

Quite true, but any RF filtering at the output of your transmitter will

have no effect on key clicks unless you manage to devise an RF filter
with only fraction of a Hz bandwidth. Filtering within the keying
circuit of the transmitter is the only way you can lengthen the rise and
fall time of the transmitted signal and as you rightly say, once the
keying characteristic has been defined, a linear amplifier should follow
in order to preserve the rise and fall.

The exception to this is in the G3YXM design, which I use and I believe
you use. Dave controls the mark/space ratio of the drive waveform during
keying, so even though followed by a non-linear 'switched mode' PA, the

73, Tom G3OLB.