I ask this because I am busy working on my own design, and a circuit that Karel posted (below) made me have to think deeper. When reading up on these phase splitters, it is easy to understand them in isolation, and if capacitively coupled so DC isolated from preceding or succeeding stages. But some circuits (including the Mars EL84 amp) couple the output of the previous stage directly to the grid of the phase-splitter triode.
So my question is that doesn't this make biasing the phase splitter a bit tricky? If you want to bias the grid at -1,5v (say), then R and the current through the first stage AND R1 and the current through the phase splitter have to be set such that the voltage at the grid is 1.5V lower than at the cathode of the phase splitter. Would you do it by initially designing the first stage to give the voltage swing needed to drive the power stage (since gain of splitter ≈ 1) , and then choosing R1 to put the cathode at the right point? So you can't just make the two phase splitter resistors some convenient/commonly used value, but need to choose them so that the cathode voltage is the correct value relative to the grid.
Do you not then also need to use low tolerance resistors for R and Rk so that the first stage is close to where you want it? I know that the phase splitter resistors are typically 1% or 2%.
So my question is that doesn't this make biasing the phase splitter a bit tricky? If you want to bias the grid at -1,5v (say), then R and the current through the first stage AND R1 and the current through the phase splitter have to be set such that the voltage at the grid is 1.5V lower than at the cathode of the phase splitter. Would you do it by initially designing the first stage to give the voltage swing needed to drive the power stage (since gain of splitter ≈ 1) , and then choosing R1 to put the cathode at the right point? So you can't just make the two phase splitter resistors some convenient/commonly used value, but need to choose them so that the cathode voltage is the correct value relative to the grid.
Do you not then also need to use low tolerance resistors for R and Rk so that the first stage is close to where you want it? I know that the phase splitter resistors are typically 1% or 2%.
