I explained why I thought the ammeter approach may not work. Trouble is, as you suggested, the ferroresonance of the circuit could produce dangerous voltages under light loads.
The way I always heard it done was a light capacitor phase shift circuit to start the first three phase motor and then larger and larger ones cut in / started from the resultant phantom phase stability created from each successive larger motor load.
--------- A simple AC voltmeter would probably do the same thing.
When L1-2 = L2-3 = L3-1 are the same voltage you have 60 or 120 degree displacements depending how you want to look at it.
The currents may be different in the unit unless the 3 phase motor is totally isolated and run on it's own.
-------------