Miller Effect - Adding WCF buffer overkill?

Schematic for the three triode preamp.

The 12AY7 mu follower is based on the RCA BC3C. The 12BH7 WCF is Jensen.

Question:

Are there any obvious mistakes in this design? Any changes that should be considered?

For example, a WCF would be attached typically to the previous tube's plate. Would it expect to see higher voltage?

BluegrassDan said:

It is a 200k series stepped attenuator with constant impedance.

Click to expand...

So, if the attenuation is dominated by the Miller effect, it should be independant on the attenuation setting...?

abbey road d enfer said:

So, if the attenuation is dominated by the Miller effect, it should be independant on the attenuation setting...?

Click to expand...

Yes. And it is.

BluegrassDan said:

Yes. And it is.

Click to expand...

That's not what you wrote in your OP.
Anyway I believe the choice of a constant-Z attenuator in this position is a mistake. You could use a 250k pot, which would result in a worst-case output impedance of 140kohms and 2.5dB attenuation at 20kHz. That would be easy to compensate by adding a small value cap across the cathode resistor of the 1st stage.
However I believe that the HF roll-off is also due to the secondary of the input xfmr meeting the Miller capacitance of the 1st stage.

New sweeps with the 3-triode arrangement. Output loaded with 1.3k resistor. 680ohms looked pretty good too.

And distortion sweeps...

34 dB Gain

And distortion at max gain. 60 dB.