Rob,
Just been making my way through the circuit - It is really nice! - All sorts of unbalanced signal paths and assymetries and frequency related anomolies.. the sort of circuit which one would giv a student to analyse if one was really sadistic! LOL..
I can see why your real circuit produces waveforms which are somewhat "nicer" than the simulations - Diodes (even "normal" ones) often introduce simulation innacuracies. I have seen the LED trick used for "tube" emulation before - but never seen a circuit employing LEDs in the feedback path and as shunts - I like the look of that, its a great idea!
What strikes me about the circuit is the obvious mix of hands-on experimentation combined with competence - there are no potentially destructive configurations, loadings are all fine - It actually reminds me a bit of some of Bob Moogs circuits - One can see that a lot was "designed" using a soldering iron (or these days, simulation alows one to "build" the circuit in much the same experimental way) - but unlike usual hobbyist circuits 'designed' in this way, it can be put into production and work reliably.
Good stuff!
Fred.
ps - Im a bit puzzleed about the input configuration (in and in1) and these being referenced to -V ?
Is R9 1k? and is the 150k RA to it R7? - Also, what is the value of the resistor on IN1 (R10 I think) ? I see 614.. oh - thats C14 sitting over a 1k ?