At risk of boring some.....  this may be of interest to others as well.

Ko, although you CAN pull the crystal some way with the trimmer and it IS
suggested in the paperwork from David Tonge I believe that is NOT the actual
purpose of the tuned circuit in the oscillator. The circuit of the
oscillator is a bit of a brute force cross-coupled affair using the bias, or
tail, transistors in the mixer. The problem is that the feedback is
enormous, and all modes of the crystal can be excited. The circuit should be
tuned so that the fundamental of the crystal cannot be heard in a receiver.
The fundamental will be about 25kHz above one third of the overtone
frequency marked on the crystal. With heavy feedback the fundamental
resonance may be excited as well. This will be you a severe problem on 73kHz
(about 3 times the 25kHz) as it will appear in-band as a strong signal.
Roger G2AJV had problems with this. My version does not have a problem. The
circuit should be adjusted, I believe to make the maintaining amplifier
degenerative at about 9MHz. In most Colpits overtone oscillators the LC in
the emitter is normally tuned to geometric mean (not the mean) of the
fundamental and the overtone frequencies. This means that the fundmamental
feedback is attenuated whilst the overtone frequencies are accentuated. I
suspect if you can pull the rock onto frequency you will not achieve this,
meaning you may have some big spurious signals around. You can only try it
but do look for the crystal fundamental on the Rx before you screww the lid
back on! The other alternative might be to take the crystal out and build a
Colpits osc with it, which you can trim only frequency......it will probably
be a lot more stable too. Because of crystal heating due to the high
dissipation the converter is not really stable enough for very narrow band
modes.

Cheers de Alan G3NYK
[email protected]