Posted by Steviegtr on 20/02/2021 16:55:22:
…
Surely you would not expect the average operator to switch over while the inverter was running on the other machine.
If you thought it could happen then simply use the aux contacts to break the stop circuit. So then it would be idiot proof. …
As human error is the main cause of accidents, I certainly wouldn't trust Mr Average let alone myself! Even highly trained operators like Airline Pilots make amazing mistakes, not because they're stupid, but because something out of the ordinary destroys their situational awareness for just long enough to cause a cock-up.
Though it would reduce the risk, I don't think auxiliary contacts would save the day. 3-phase from a generator is three smooth 50Hz sine waves, changing rather slowly, and well-behaved. A VFD is different: it simulates 3-phase by pulse-width modulating three DC outputs. On an oscilloscope, it's obvious a VFD delivers a hideously spiky wave-form compared with a generator. Normally, the spikes don't matter because the motor's inductive windings choke out the pulses. But, if a winding is disconnected before the others, as when one of three contacts opens a millisecond before the others, then strange things happen. In particular, the floating winding is likely to react to pulse inputs as if it were a car ignition coil, generating spectacularly high voltages.
As the magnitude of the unwanted voltage is related to pulse rise time it can be hundreds of times bigger than when ordinary 3-phase is switched. Easily capable of rising high enough to puncture the motor's insulation or to jump switch contacts. In the worst case the spark creates a low resistance ionised path through the air allowing ordinary low voltage to form a power arc inside the switch or motor.
Although there's potential for spectacular failure it depends on the timing. How high the voltage goes depends on when the contact opens relative to the VFD's pulse waveform. There's an element of luck in it ranging from 'no problem' to 'electrical fire'. More likely than a catastrophe, the volts would spike back the other way into the VFD, and wreck the electronics. It's really not a good idea to switch a live VFD.
It is possible to buy VFDs specifically designed to allow hot switching. Last time I looked, they were all big beasts well beyond anything Joe Public might buy. But who knows, there might be a small version available now – there's certainly demand for them. Off hand, I can't think of a way of making one cheaply: cleaning up spiky power waveforms needs a decent filter and hefty inductors aren't cheap!
Dave