This TX uses one 6DQ6B and covers from about 3.0 up to 10.0 MHz, delivering more than 20W. Crystal controlled, with Crystal frequency correction, uses a Pi network with roller inductor for the widest impedance matching capability.
It has a solid state semi break-in circuit that drives two relais: one for control grid blocking, and the other for antenna changeover.
A panel meter is mounted to keep an eye on the control grid current and to make the antenna tuning procedure easy and reliable.
The 6DQ6B is fed by 300 V plate voltage and 150 V (OD3 stabilized) at screen grid. Relatively low voltages for a nice output performance.
The schematic is this:
It's a hybrid solution, in that it uses a solid state semi break-in circuit. At key down, both relais (control grid blocking + antenna change over) are activated. At key release, the control grid blocking section immediately changes its state stopping the tube to oscillate, while the antenna relais switches with a certain delay.
The capacitive divider (30 pF + 500 pF) at the Crystal section, plus the cathode chocke, are VERY TRICKY to set for nice CW tone quality. The values in the schematic are a first try and not optimized: they simply work acceptably with my FT-243 Crystals.
Any CW single tube transmitter works better when it operates as FREQUENCY MULTIPLIER. When it is used at the same Crystal frequency, the antenna impedance matching circuit might tend to resonate, completely 'blanking' the Crystal frequency determination. So, it is necessary to slightly de-tune the antenna circuit, with certain loss of output power, to have an acceptable tone quality and to have the transmitter working AT the frequency of the Crystal, and not at the one of the impedance matching network.
To easily and cheaply get on the air to make some chats with your colleagues here and there on the Shortwaves ... what more do you need ?
Radioman, Nov. 2014