another poor man's car engine conversion

You might like this.http://ibis.experimentals.de/downloa...lvibration.pdf

I've read this several times in the past. What strikes me is the
absence of test equipment like wireless load cell torque sensors on
the shafts. If used, any torsion oscillations could be seen on an
oscilloscope long before they became destructive.

High speed shafting is as old as the industrial revolution. Solutions
to torsion oscillations are just as old. Yes, there are potential
problems but there are also well tested solutions. The trick is
knowing what they are and how to use test instrumentation.

The other thing that jumps out about the Bede saga is that fact that
they were trying to put an experimental engine/prop drive into an
experimental airplane before it was de-bugged and they were doing it
under a deadline.

That's snakebite country. Even things that work everywhere else are
almost certainly going to bite you under those conditions.

It's vital to work one problem at a time. If it's an experimental
engine and prop drive, work on that until it's been running sweetly
for a long time. Then, maybe, think about designing an airframe
around it.