Unfortunately, the question is something like how to convert a VW to a
Mercedes -- you pull the radiator cap from the VW, drive it off, drive
the Mercedes under, and put the radiator cap onto the Mercedes.

Harmonics are caused by distortion of the generated wave. This
distortion is caused by overdriving the oscillator, which is necessary
unless some other method of limiting is employed. An oscillator is an
amplifier with gain greater than one whose output signal is fed back to
the input. When it starts, the signal amplitude increases until
something limits it. In a simple oscillator like this one, the limiting
factor is the smashing of the waveform as the transistor gets overdriven
each cycle. Distortion is inevitable and necessary. So, you can either
reduce the distortion or you can try to filter it.

Filtering presents a couple of problems. One is that you're apparently
radiating enough signal (and harmonics) without an antenna. That means
that even if you had a perfect filter someplace, you still have enough
energy being radiated from the portion of the circuit that's ahead of
the filter. If you simply try to trap or reduce the harmonics at the
collector, you might stifle the oscillation by cleaning up the waveform
and eliminating the smashing.

In general, less feedback will reduce the waveform distortion. The
tradeoff is that the oscillator will start more slowly, and if the
feedback becomes too little at some temperature or battery voltage,
it'll stop altogether. But it you'd like to try this, fiddle with the
values of C1, C2, and C4. I believe that decreasing C1 or C2, or
increasing C4, will reduce the feedback.

If you want try improving the collector waveform, increase the Q of the
tank by decreasing L1 and increasing C6 to compensate. Or you can try
putting a series LC network (with fairly high L/C ratio) from the
collector to ground, tuned to the second or third harmonic, to act as a
trap for one particular harmonic. Any of these schemes could make the
oscillator stop, though.

A professional approach would be to add a buffer stage, and build the
oscillator and buffer in a shielded box, with good filtering on any
lines going in or out except the output signal conductor. Then, a filter
at the output would be effective, so one could be designed to take care
of the harmonics to any extent desired.

Roy Lewallen, W7EL

Diego Stutzer wrote:
Hi,

I built a small FM Radio as shown on
http://members.aol.com/torkraemer/uk...ukwsender.html.
As it is it works pretty well, allthough I operate it without any
adition antenna.

Of course there are pretty high (unwanted) peaks in the frequency-band
at multiples of the operating frequency.
Because I don't want to use an additional antenna, I can't filter the
signal on its way from L1 to the antenna with a common "2-port
Bandpass".
Putting such a Bandpass (or Lowpass) between the DC-Source and the
LC-Circuit did not seem to work well.

So how can I damp those harmonics on L1 Coil? I thought of some kind
of frequency dependent impedance in parallel with the Coil.

I tried this with another LC-prallel Circuit which has an impedance of
approx. 0 Ohms at 200 Mhz. This works as expected but seems a bit
unsophisticated and the Impedance falls too slow as the frequency
approaches the 200Mhz.

Unfortunately I have no Idea how to design a "higher order
1-port-Bandpass", or even more likeably a Highpass.

What would you recommend?
What shoud I read?
Any suggestions are welcome.

Thanks to all, and sorry for the lenght of the message.

Diego S.