LCL output networks - match anything with your transmitter

[IN3IEX]

RF output networks more complex than the usual pi are found on many military tube transmitters.
On the other hand it is not clear if these networks are the simplest/cheapest for the purpose.
The question is: can we make an output network that is:

1) simple
2) as versatile as a T tuner
3) low pass as a pi network

Possible answer:
Just make a LCL T tuner and use it in place of the pi network.

http://www.ing.unitn.it/~fontana/LCL%20Matching%20Networks.pdf

This is a theory paper !
Who will test it?
How to use it? What is the sequence of operations on the three knobs?
How about using a switched multi tap inductor (to anode) a variable capacitor and a roller inductor to save money?

[Opcom]

Considering that some transmitters have 3 knobs, 1 each for C, L, and C, this LCL network seems OK by me. As far as doing the design/build/test, I'm not the best qualified that's for certain. Better a BC engineer or other expert.

Where did that software come from? I could imagine something done to make the numbers with excel.

LCL.

It is not unlike half a 'balanced antenna tuner' where the variable C can be placed at either end of the variable L. Who is to say two L's can't be used, one before and one after the C. Then it is like the paper because in the first case he has almost no input L.

It is also like a PI-L (CLCL) network if the capacitance of the output tube or device is considered.

Most hams would not want to build all the stuff that's in a general coverage military set. Those sets have to cover almost every situation and every set of bedsprings and chicken wire someone might have to throw up in a pinch. In comparison the LCL presented is very simple, parts wise.

As for military stuff, some of it has an identifiable tuning and coupling section for the plate, and then a second loading and matching circuit for wire or other antennas.
The URT-12 is like that. There is a place in the middle of the tuning circuit where approximately a 50 Ohm coax can be attached. The set covers 2-30Mhz and expects a long wire.
In the GRC-106, there is a large turret of many coils with capacitors, really a set of many broad band transformers each covering some MHz range. This feeds a separate antenna matching circuit (all is built into the same casing) and the whole thing looks very complex even if the only plate circuits considered are in the single band used at the time. That set can tune a 50 Ohm load or a 15-21FT whip or a doublet. It can tune almost anything.

An inductor direct to the anode with no 'tuning' cap but a loading cap will work, at least in the Wawasee Black Cat linear amp having one 8417. The inductor was very long, maybe 1.25" compared to a usual 100W 10M tube style pi-network inductor which may be something like 4 turns 3/4" diameter. Almost anything can be made to work. That Black Cat arrangement was very low Q. Set the loading and forget it!


Before optimizing the new design with taps to save cost, the values all have to be worked out, one set of them for each frequency and Q. If the taps have to be moved to change the Q, then that's no good. If a tapped coil, better to pick a desired Q for each frequency and try to stick with it. One question is how many taps? who knows with this circuit, it could be more than one per band. 
With variable inductors, they can maybe be found by experiment but there would be more than one setting of the three knobs depending on Q.
Switching is OK until the power is high and a slew of vacuum relays has to be used. Then maybe a roller is more reasonable for cost.

[W7TFO]

The schematic looks just like the usual 'T' net used to match most BC am feedlines to vertical tower impedances.

Very handy to make the match and solve for +J0 reactance.  Most of the time the 'C' leg is a fixed value, but there you are dealing with a single frequency.

73DG