The use of tubes with limited performances in RF, requires the use of more tubes for the same output power. The 6BG6G is a 'sweep' tube, used in old TV sets, and it's a straight derivation from the 807. The curves of both are very close each other, but the rated dissipated power is higher for the 807 and this one is better shielded internally to achieve proper performances at RF.
Despite the neutralization network, the text said that on 28 Mc/s the final amplifier works as doubler. The 6K6 is a relative to the 6V6 tube, and was meant for audio applications. Therefore, it might well be that its performance as frequency quadrupler is poor: from here, the need to double in the final stage.
In MOPAs for Radio Amateurs, usually the neutralization is applied to achieve, possibly, the same output power on all the HF bands, without sacrifying nothing. A neutralized PA should perform rather evenly on a broad frequency range. In reality, the capacitive neutralization network suffers a bit from a certain frequency dependance, and in fact all articles prescribe to set it on the higher bands.
This circuit has something strange, though, even in the neutralization: the capacitor from the plates to the bottom of the control grid tuned tank ... is 30 pF, which is a very high value for the purpose. Tests performed in my radio-bunker on my glorious MOPA, showed that the higher this capacitance, the higher is the neutralization dependance on the antenna tuning settings. A quick inspection on the PA stages of more recent transceivers, shows immediately that this capacitance very rarely reaches a rated maximum value of 10 pF, and usually all is made in such a way that the optimum point for neutralizing the final amplifier stays in the middle of its rate value.
Yet I haven't spent adequate time to investigate the hidden reasons for this behaviour, that is on the frequency dependence of the capacitive neutralization. There is, nevertheless, a piece of explanation in the fact that the 'C4' capacitor is exposed to the variable capacitors of the PA impedance matching network. At least, it is in my MOPA. Well: the stray capacitance between 'C4' and the other movable components of the antenna impedance network, changes upon the mechanical position of the rotor of a near, other, variable capacitor. And the tiny capacitance variations, result in an unbalance of the neutralization network. So the frequency dependance, is, at least partially, in my modest opinion, a secondary effect of an electrostatic variable, stray, coupling between close components!
In another famous MOPA, the MILLEN 90801, which I will describe soon, and which I'm extremely proud to own, the neutralization is inductive: there is a pick-up cois in the PA tank, which feeds back its signal to a single turn link put inside the control grid tank of the same PA tube. In this case, the pick-up coil generates the neutralization voltage from the magnetic field produced by the same PA tank coil. It's capacitive coupling to the PA variable capacitor, then, doesn't have, as a matter of fact, any influence. James Millen, is my opinion, understood this very well.
So, why this stubborn, insisting, discussion on MOPAs ? It' because if each of their technical details are investigated in deep, they can still teach something on Radio Engineering well beyond the year 2000 ! Thanks to the electron tubes technology, though ...
Radioman, Feb 2016
Posted by Radioman | Permanent Link