In article , Gary Coffman says...

To really know what's going on, you need to use a dual trace scope
setup to sum vector voltage and vector current in each circuit mesh.

A *floating* dual trace scope. Current can be measured with
a sampling resistor.

That'll tell you how much actual power is being supplied by each phase.
You want those to balance. It would be a surprising coincidence if that
occurs at the point where the amp clamp meter readings are equal.

Not *that* suprising. I suspect that once a fitch-type tuning
is done, this will be true. Ie, the phase-to-phase voltages
at the load motor will agree at the same point that the real
currents (measured not by just an amp-clamp) come into agreement.

I think I would have to invoke some kind of symmetry argument
as a first hand-waving approach.

The main thing that you want to worry about when tuning a rotary
with ordinary test equipment is to achieve rough voltage balance
phase to phase over the operating range of the converter. To do
that, you want the voltage deviations from equality to swing roughly
symmetrically about nominal as you go from no load to fully loaded,
while at the same time deviating no more than 10% of nominal in
either direction.

This becomes easier and easier to do when the idler motor impedance
is low. An argument in favor of oversizing the idler motor - a
lot of the finessing issues simply go away. My converter had no
tuning or pf correction, yet the phase to phase is well inside
ten percent over the entire operating range.

Jim

==================================================
please reply to:
JRR(zero) at yktvmv (dot) vnet (dot) ibm (dot) com
==================================================