grid drive level vs output and modulation....

[K8DI]

So, question for those who know more than I..

In a given grid driven plate modulated class C amplifier, ignoring melting the grid/excess grid power dissipation for the moment, and also assuming some kind of protective bias.

What happens to amplifier output and level of modulation if the grid drive is much higher than normal (say double) datasheet values?  Does this cross a point where it starts to lower RF output at no modulation? does this get to a point where modulation is less linear, that is, more audio power is needed to push it to zero, or hit 100%? Besides cooking the grid, is there a limit to high grid drive?

And, what happens if grid drive level is under spec by a lot (say half of data sheet values)? Is unmodulated carrier level affected? Does it make it impossible to achieve 100% positive (or negative) modulation?  Besides high plate current due to low self-bias, is there a limit to low grid drive?

Ed

[Steve - K4HX]

At least one (big IMHO) consideration is modulation linearity.

From the "Radio Handbook", 1940 edition, page 160:

Class C Plate Modulation. The characteristic of a class C amplifier which makes it linear with respect to changes in plate voltage is that which allows such an amplifier to be plate modulated for radio telephony. Through the use of higher bias than is required for a c.w. class C amplifier and greater excitation, the linearity of such an amplifier may be extended from zero plate voltage to twice the normal value. The output power of a class C amplifier adjusted for plate modulation varies with the square of the plate voltage. Since this is the same condition that would take place if a resistor equal to the voltage on the amplifier divided by its plate current were substituted for the amplifier, it is said the stage presents a resistive load to the modulator.

[Opcom]

If you have enough grid drive for carrier level, it doesn't mean you have enough for +100% peaks during modulation. Sure the drive can be so low that you don't get enough RF to reach the carrier level you want.

I find I need 2x what the data sheet says for drive. Whether it's reserve, or because of grid circuit inefficiency I don't know. I just know that it's what the wattmeter says from the exciter to the grid circuit (50 ohm cable hookup).

I bias my final amp as far negative as permissible and apply drive as needed to furnish the drive to reach the peaks. I have way more drive than I need so I can get away with this - because I'm using a 100W ricebox as the VFO to drive a 4-1000. The conditions I'm runing call for 12W of drive but the meter says 25-30W is going in. The tube grid is rated 25W so it's OK.

I believe this causes the grid current pulse and therefore the plate current pulse to be as short/narrow as possible, and gives higher plate or output efficiency as long as there is enough emission for the peak plate current.

I don't really know a lot about it, just that this transmitter seems to work in this way.

[Steve - K4HX]

Some more goodness from the Handbook:

 Grid Excitation. A sufficient amount of grid excitation must be available for class B or class C service. The excitation for a platemodulated class C stage must be sufficient to drive a normal value of d.c. grid current through a grid bias supply of about 2-1/2 times cutoff. The bias voltage preferably should be obtained from a combination of grid leak and fixed C-bias supply. Cutoff bias can be calculated by dividing the amplification factor of the tube into the d.c. plate voltage. This is the value normally used for class B amplifiers (fixed bias, no grid leak). Class C amplifiers use from 1-1/2 to 5 times this value, depending upon the available grid drive, or excitation, and the desired plate efficiency. Less grid excitation is needed for c.w. operation, and the values of fixed bias (if greater than cutoff) may be reduced, or the value of the grid leak resistor can be lowered until normal d.c. grid current flows. This value should be between 75% and 100% of the value listed under tube characteristics.

[DMOD]

From some personal experience:

If there is a stated Maximum Grid current spec I look at the maximum grid current specification and initially divide it by 2.

For example, an 813 in Class C has a max grid current spec of 30 mA; so I set it up for a plate voltage of 2,250V with a 15 mA control grid current with an initial 10k resistor. That gives me a -150V grid bias. So multiply that by 1.17 for RF grid drive circuit losses, and we get about -176V. So we need as a miminum enough RF grid drive to give us -176V grid leak bias. Checking, Ig1 = 176V/10k = 18mA, still below 30 mA.

Now for sweep tubes such as the 6DQ5 in Class C, one will need about 2 mA of grid current max. Bias is usually developed across a 50k grid leak resistor for a grid bias of -100V. So I need enough RF drive voltage to develop a grid bias voltage of -100V across the 50k resistor. Looking at the 6DQ5 tube characteristics, this is 1/2 the "Peak Grid #1 Voltage" spec of 200V. This infers the maximum grid current must be 4 mA. So 2 mA is 1/2 the inferred maximum grid current.

So biasing is very much tube-type dependent.

Phil - AC0OB