As deduce from the sim, it appears the H2 is trading off for H3, that is when H2 level (as adjusted by diff feedback) gone up, H3 level gone down and vice versa, internal distortion cancellation details is unclear to me at this time.
Here is a preamp design I have been playing with. RIN1 and RIN2 are ganged 10k attenuators. They were originally parallel with 10K resistors to get the load I needed for a Jensen input transformer. This was to be a clean sounding 0 nfb amp but this thread made me think about being able to adjust the 2H. Adding R1 as a 2K5 pot lets one adjust 2H from null (R1 centered) to 3H + 25 dB. I’m still trying to grok the whole phase of H2 Nelson Pass described but I believe this scheme should allow one to adjust the phase of the 2H to either positive or negative.
Below is an LT spice file and jpegs of the schematic, H2 max fft and H2 min fft both at 2Vrms out.
Below is an LT spice file and jpegs of the schematic, H2 max fft and H2 min fft both at 2Vrms out.
