Apologies for the on-topic post.
I have a class B amplifier with complementary Darlingtons for output, and an op-amp driving the tied-together bases with -ve feedback from the output. It's simple and works pretty well.
I was tinkering with LTspice to reduce the crossover distortion. I don't want class AB quiescent current, but putting three diodes between the bases to make the voltage 'gap' one junction rather than four and using pull up/down resistors improves things - less gap for the op-amp to cover, I guess.
However, using Zeners instead of ordinary diodes improves things still further as far as the simulation goes.
What are the fundamental differences between a PN diode and a Zener in forward conduction, or is this likely to be an artefact of the Zener models?
Cheers