Plate/Screen Modulation of Beam Power Tubes - Usual Suspects vs Sweep Tubes

[KK4YY]

Given the usual, easy method of modulating the screen along with the plate by using a simple dropping resistor, I'm wondering if anyone is aware of which tube characteristics provide for better, easier, more optimum, plate/screen modulation of beam power tubes?

I'm particularly interested in the difference of commonly used RF types (2E26, 6146, 4D32, 4-400A, etc.) versus TV sweep tubes (with their lower screen voltage limits) in class C plate/screen modulated service.

Any insight would be helpful.

tnx,
Don

[WD5JKO]

Don,

   You bring up some good points. Look at the Gonset G-76 method of plate and screen modulation of a 6DQ5. They have a VR tube in the G2 screen dropping resistor path, and then a G2 to ground resistor. The clamp tube ties in between the dropping resistor and the VR tube. This allows a simple 12AQ5 clamp tube when conducting (NO RF Drive) to extinguish the VR tube. The result is the G2 voltage goes to zero. This works very well, and no grid bias supply needed. As to how it modulates, I'd say dam good, but realize the screen grid (G2) voltage is derived at the junction of a resistive voltage divider meaning the screen modulation percentage is reduced....something necessary in most multi grid RF output tubes.

   Another method is to use a modulation transformer with a tertiary winding made to modulate the screen grid with the tertiary winding, and the plate with the main secondary winding. This has many advantages. The ART-13 as I recall used this technique. The turns ratio between the primary and the two secondaries will determine how much the plate and screen grid get modulated.

   The common multi grid RF tube plate modulated with a screen grid series dropping resistor is very simple, but has some disadvantages:

1.) The screen dropping resistor with the screen bypass capacitor form a low pass filter, and at the higher audio frequencies there can be appreciable phase shift between the plate modulation and the screen modulation. This can be compensated for with R-C compensation similar to what is done with a 10X scope probe.
2.) The percentage of screen modulation usually needs to be reduced from the percentage of plate modulation. Cutting that back to 50-70% might be a good place to start. This would require two screen dropping resistors; one from the modulated B+, and one from the unmodulated B+ such that the parallel equivalent of the two is equal to that of the single screen dropping resistor. Then play with the ratio of the two resistor values to determine the amount of screen modulation while keeping the parallel equivalent value constant.
3.) With big Tetrodes, a screen dropping resistor from the plate B+ supply poses a dangerous situation when the RF tube has NO RF drive. The screen voltage will rise to the same voltage as on the plate. NOT GOOD. A clamp tube will work here, but failures can occur.
4.) Look at the Globe King 500. Here a 4-250 or 4-400 uses a separate screen supply with high impedance. If the screen current is high, the screen voltage drops. Conversely if the screen current is low, the screen voltage rises. This helps keep the screen grid dissipation within the 35 watt rating. For AM use they switch in a large choke in series with the screen grid voltage path. When the RF PA tube is plate modulated, the screen grid will self modulate.
5.) With a separate screen supply, you MUST make precautions to insure that the screen grid is not at voltage when the Plate Voltage is OFF.

Maybe others can better say what I attempted to say.

Jim
Wd5JKO

[KK4YY]

Jim,

Thanks for your reply.

The Gonset G-76 method of modulating the 6DQ5 PA screen with a VR tube and voltage divider had escaped my attention. I had noticed that the 6DQ6B modulators were screen driven. I guess should have looked harder its overall clever design - and I will. It's interesting.

I briefly fooled around with using a mix of modulated and unmodulated screen voltage on my test bed DX-100. The lash-up didn't work well at first and I didn't pursue tweaking the ratio at the time. (I'm in the process of trying sweep tubes in its modulator using the "crazy drive" method of proportioned drive to the screen and control grids. If it works out well I'll publish the results in this forum.)

Getting the right amount of modulation to the screen appears to be the primary issue. Perhaps there isn't much difference between, say, a 6146 and a 6DQ5 in this regard. I was wondering if there were some inherent tube characteristic that made the job easier. I couldn't find anything about this in the literature - maybe it doesn't matter all that much.

Screen bypassing and screen protection seem to be two problems that just won't go away. They need to be addressed in any good design. Screen self-modulation requires a big chunk of iron which I prefer to avoid. Resistors are cheap and plentiful. Naturally, I'm trying to find a simple solution to a complex problem, but one that works well. Easy, right?

So, I'm going to file your report as a positive one for a sweep tube final given your thumbs-up for the G-76.

I'm looking at the application of sweep tubes as a plate modulated final but there aren't that many examples in the wild to draw insight from. Given their high emission vs low plate dissipation ratio they may not be the best choice for that service. But many of them can be had for small money, so it's worth looking into. Paralleling a few to achieve moderate power levels may produce its own set of problems. I'll need to give physical layout due thought and experimentation. I plan use sweep tubes as modulators but the RF final tubes are still up in the air. Maybe it will be two or three 4D32's. Maybe half a dozen 6LW6's. I could say that it's a blank sheet of paper right now, but the truth is it's many pieces of paper scattered about the shack. Either way, it's undecided at this time. This is all planned as a follow-on project after my DX-100 test bed experiments.

Don