How can a sinusoidal transistor oscilator transmitter be created?

At your risk and at your responsibility.
Any voltage or current can be dangerous.


Is class A operation (electrons flowing all the time) possible on oscillators?


The mice book by Donald Stoner says:


"Class A operation is seldom found in a transmitter circuit other than for oscillators.
This mode is characterized by a pure sine-wave output independent of the flywheel action of a tank circuit.
Class-A operation results in a freedom from harmonic generation and improved frequency stability.
This condition is difficult to achieve in oscillator circuits because the transistor usually swings violently between saturation and cutoff and the output approaches the appearance of a square wave.
The flywheel action of the tank circuit helps restore the sine waveshape but it is seldom perfect if the original shape is badly distorted.
To obtain true class-A operation, careful attention must be paid to such details as biasing, impedance matching, and, most important, to drive level of the feedback energy."


So...


Does for example the nice, simple and famous Jostykit HF65 oscilator broadcasting to the FM band creates unwanted interfiering harmonics or overtones?


If likely yes, what can be done to improve it?


It is a nice grounded bass oscillator. A tank of LC circuit resonating at say 100MHz on the collector circuit.
A tap near Vb supply on said inductor for the low impedance antenna.
A 3pF capacitor for positive feedback between collector and emitter. (In grounded base configuration the emmiter signal in is in phase with amplified signal from collector.
But collector output is relative high impedance and emitter inout is low impedance.
Would this circuit operate better by impeadnce matching?
For example using the 75Ω or whatever tap to connect the 3 pF capacitor to feed energy to the low impedance emitter. (Instead of connecting 3pF from collector to emmiter.)
Would it work?
Would the use of a transistor not going very high be an aid too? For example a pnp one as used on the Jostykit HF365?
And also using a varicap as in HF365 for better modulation fidelity?


Do electron tube circuits benefit from the grid rectification effect of electron tubes effectively creating automatic gain control? [Armstrong]. Would a diode and then a similar to grid leak bias help a transistor in this direction?


References:


Transistor Transmitter for the Amateur - Donald Stoner - First edition - page 25


Jostykit HF65 schematic


Jostykit HF365 schematic


Operating Features of the Audion - E. H. Armstrong