Here is where the Split Load phase splitter shines. The lack of PSRR at the plate and the ultra high output impedance at the plate work in our favor. The Split Load phase splitter's plate resistor is connected to the top triode's cathode and the bottom is DC coupled to the grid of the top triode, while the cathode resistor is DC coupled to the grid of the bottom triode. Thus, as the cathode swings up and down, the plate resistor will follow without adding much additional loading to that cathode, as the Split Load phase splitter's plate impedance is hugely magnified by the large valued cathode resistor times the mu of the triode used. This means the output stage can work at driving the electrostatic headphones' capacitance without having to work hard at driving the phase splitter as well.
   This newly configured output stage works like the output stage of a conventional tube power amplifier, such as the Dynaco ST-70, in that the output tubes are required to deliver voltage gain, as well as power gain. This requirement is what forces the Split Load phase splitter's plate connection to the top triode's grid and its cathode connection to the bottom triode's grid. Reversing this arrangement results in the output tubes working as buffers, which would offer a very low output impedance, but  no gain.

   The alternative to both the transformer and the SRPP is more circuitry: more resistors, capacitors, and triodes. Basically, what needs to be done is to give each totem-poled pair of triodes its own phase splitter. In other words, the amplifier must contain cascading phase splitters. The first is the input stage, which will either be a Split load phase splitter or a Long Tail phase splitter. Then a phase splitter is needed to give the top and bottom triodes the correct signal drive voltage independent of the output signal swing.

DC feedback applied to the balanced push-pull output stage.

pg. 5

<PREVIOUS

NEXT >