Valve frequency multipliers

I am trying to use a 15 Mc/s crystal oscillator to generate a 150 Mc/s
signal, the obvious multiplication ratios are x5 and x2.   The 150 Mc/s
has to be distributed to two other units by a 120-ohm screened cable.

The whole thing must be done with the minimum number of valves and no
semiconductors.  The Colpitts-derived xtal oscillator is an EF91 and the
multiplier stage(s) can be either another EF91 or an ECC91. 

I have tried picking the x5 signal (75 Mc/s) off the anode of the
oscillator with a tuned circuit but can only get a couple of volts
pk/pk.  This isn't enough to drive the ECC91, which I  had hoped could
be used as a 'push-push' doubler, it also won't drive an EF91 over
enough of the curved portion of its characteristic to give sufficient
frequency-doubled signal.

Alternatively, I have tried using a parallel-tuned circuit at 15 Mc/s in
the anode of the xtal oscillator to drive one of the triodes of the
ECC91 which can then act as the multiplier.  There is a whopping great
15 Mc/s signal going into the grid of the triode (about 25v pk/pk) and,
with the cathode earthed, this develops enough grid-leak bias that the
valve is conducting anode pulses of over 20 mA about 10% of the time.

I would have thought that under those conditions the triode would have
given a large signal at 75 Mc/s in an anode circuit tuned to that
frequency - but it doesn't appear to.  I can't use the triodes as
straight earthed-cathode amplifiers at those frequencies because of the
Miller capacitance effect, but they should be perfectly satisfactory as
multipliers where the grid and anode circuits are  tuned to different
frequencies.

Does anyone know how to determine the optimum conditions for generating
the 5th and 2nd harmionics in valves?

--
~ Liz Tuddenham ~
(Remove the ".invalid"s and add ".co.uk" to reply)
www.poppyrecords.co.uk