Re: VTTC grid leak

["Tesla list" <tesla-at-pupman-dot-com>]

Original poster: "Bob (R.A.) Jones" <a1accounting-at-bellsouth-dot-net> 





 > Original poster: herwig.roscher-at-gmx.de


 >
 >  > The usually idea of the circuit in a class C  oscillator with a tuned
 >  > anode load is to automatically adjust the DC grid negative bias so that
 >  > a particular operating point or oscillation amplitude is archived.
 > - So the grid coil is responsible for driving the control grid with
 > correct phase and drive level and the grid leak combination sets the
 > desired negative bias. Are there interactions?

Yes there must be some interaction.
The pulses of grid current must combine with the drive and could produce
some phase shift or as the conduction angle for the valve varies that may
produce some phase shift.
But such effects are going to be small its going to be small.
Then there is the C looking in to the grid which produces phase shift with
the grid leak R.

 >
 >  > Because the tube is biased on the loop gain exceeds one. So any noise
 >  > or switch on transient grows in to an oscillation.
 > - So the control grid voltage is in phase with plate current?
 >
 >  > Assuming the time constant of the grid leak C and R is long compared
 >  > to the oscillation frequency
 > - 500 pF and 680 Ohms equals 340 ns
 >    half cycle of 6.5 MHz equals 75 ns
 >    Measured swing of control grid voltage from -280 V to + 280 V. Bias
 >    therefore 0 V.
 >
 > Increasing the value of the grid leak resistor to 5 kOhm changes the
 > swing to -330 V to + 200 V. The voltage still is sinusodial
 > (unexpected and a little bit strange). Bias calculates to -65 V (tube
 > is 4-125A).

Why would you expect it not to be sinusoidal ?  I assume you have tuned
circuit in the anode and the grid coil is coupled to it with sufficient
coupling that is low impedance compared to the input impedance of the tube.
So would it not just be a smaller version of the anode voltage with a very
small wiggle as the grid conducts?
 >
 >  > So the tube only conducts at part of the positive peaks of the
 >  > drive signal ie class C operation.
 > - I assume that the bias should be around -200 V. Correct?

I don't know what your correct bias should be.

Assuming no break out. I would have expected the total swing to decrease
because as you increased the negative bias the anode swing should decrease.
What is holding the amplitude constant? perhaps the anode swing is too
high(the load is too light) or is it the streamers.

 >
 > Are my comments correct?

I suggest the following design strategy.

Put say a 1 ohm resistor in the cathode circuit to monitor  the current.

For max efficiency you want a small conduction angle and you don't want to
exceed the peak valve current or what ever level you are willing to risk.
So you adjust the grid drive turns and grid leak R to get the required peak
current and power supply current.
But of cause you have to optimize it for pre and post break out so its a
compromise.

Bob Jones