Sorry for the hiatus, I had an exam to submit over the weekend and
work is being a time hog at the moment.

Re something with a few watts of drive, I do have a couple of ham
friends, who might be able to provide this if I brought the project to
them. Only problem is that the power supplies et al are spread over
the workshop at the moment, but this might be a good route if I cant
get any further. It has occured to me that I might have broken the
tube, melted the grid or something, so this could prove the tube
works.

I have studied up on the tube capacitances and the interaction in the
circuit, and decided to rework the project into a different design -
either hartley or colpitts. Have come across a patent also that
describes exactly what I want to do (tube amp for plasma excitation),
and they use a modified hartley. But looking at the colpitts, I like
the way that the tube capacitances are in parallel with the tank caps.
They should swamp the tube values I think.

Re the 12V supply, the idea is that the resistor biases the grid (at
DC) to ground, the cap smooths the current through the resistor, and
as electrons flow into the grid and make it more negative, at 12V the
diode starts to conduct and clamps the DC grid bias at -12V. Of
course, the tube end of the grid circuit is being driven high and low
by the coil, so this shouldent clamp the grid voltages - just the grid
DC level (bias). Thats the idea anyway - might be missing something. I
have checked grid current and it peaks at about 50mA with the voltage
cranked up to about 2kV IIRC.

Re choosing 25MHz, I originally was shooting for 50MHz but had no joy
getting the circuit to work. It worked great at 5 MHz or so, so I felt
a compromise would be 25MHz. So thats where I am currently aiming for.
The higher the frequency the less reactance I will get from the 'load'
if I drive it capacitvley which is what I want to do (because of
striking the discharge, which is difficult to do inductivley). There
is a need to seal the plasma tube for certain experiments to raise the
pressure in it, and some of the substances I wish to excite are not
compatible with metals (an example is sulphur, which is a major part
of the experiments, and its performance is described in various
patents - its usually excited with microwaves in the current art, but
I want to use very small discharge chambers which make this hard) so I
cannot have any metal in the discharge envelope (I also have the kit
onhand to make these quartz glass capsules).

As you say I have heard of 13.56MHz also as a standard frequency for
this kind of work, but I am aiming higher for the above loading
reasons.

Now, as for grounding the grid, I have considered it, but have not
come across any designs of oscillators where the grid is grounded, and
I dont have enough confidence in my circuit redesign skills to work
out the details myself. I could of course work out the circuit, but
what Z will be where, and how is this related to that phasewise - an
example is the increased capacitance from the anode to the case (which
is now the grid). Just dont know.

I will get some photos of the rebuilt thing once its a colpitts, I am
also replacing the caps with some homemade ones made from heavy copper
foil, since I am a bit unsure of the quality of the russian doorknobs
I am currently using.

Finally, is this a uni project - well, not yet, but I am a physics
undergrad and am studying some related fields, so you could say its
related. It also relates to some harder core physics experiments I
want to start soon, with a cyclotron being one project, which I need
powerful RF drive for. Its all great fun - unless you cant get your
damn RF power house to work!

Current interest for this project relates to finding a discharge lamp
(high intensity) whos spectrum compliments the zooanthellae
photosynthesis absorbtion of marine corals. I could write a ream about
this but its off topic so I will leave it at that for the moment. But
its a means to an end currently - although I would like to have a
power oscillator capable of spanning 1 to 100 MHz at several hundred
watts for plenty of other experiments.

Thanks for the contributions so far. Hope to nail this one soon.