touching tube grid starts oscillation?

hello I am making this https://dl3jin.de/tx_1x_1sh29b-data/tx_1x_1sh29b.gif the third schematic

Oscillation starts ONLY when I touch the grid with my finger.
What could be the problem?

- - - Updated - - -

neazoi said:

hello I am making this https://dl3jin.de/tx_1x_1sh29b-data/tx_1x_1sh29b.gif the third schematic

Oscillation starts ONLY when I touch the grid with my finger.
What could be the problem?

Click to expand...

Hm, that is strange, I connected the grid with the transformer case (shield) and oscillator now starts. Why does it happen?

I would suggest that you use "B" version of the schematics. Crystal-controlled oscillators are designed mostly with triodes which offer the Miller plate-grid capacitance feedback, needed to oscillate. If you use a pentode (versions A and C), there is no feedback and the oscillator can get some if you touch the grid with your finger.
You can try another schematic, with the quartz connected between plate and grid. In this case, make sure you apply a low plate voltage, I think the oscillator can run on 30-50V DC plate voltage. Tiny quartz resonators can be damaged by a high RF current.
Nowadays, vacuum tubes are no more used as good quartz-controlled oscillators. Transistors are better, and quartz crystals are more stable with them.

- - - Updated - - -

Also, not all quartz crystals are good for this oscillator. Some would not start at all. Try another one.

I would replace the 100K with a 1 M resistor, it'll damp the crystal less and make the interelectrode stray capacity more effective.
Frank