One of the nice or nasty things about 150Hz, is that is looks the same on all 3 phases, so if the 3f zero crossing is say 10 degrees after the 50Hz one, on one phase it will also be on the other 2 phases. So if the loads on all 3 phases were exactly equal, the 3 F, 9F components etc would all cancel. However this does not mean that you need a neutral to have distortion to the line waveform that gives a 150Hz repeating feature. In fact the 3 phase/ 6 diode bridge does it depressingly well because in reality the 6 diodes are never quite equal. So the answer is that the 3 F current in any one phase is imperfectly cancelled by the nearly corresponding 3F currents in the other two, and when the 150Hz current is   is 5-10 percent of the 50Hz RMS current then that 3F component is  is 30-20 dB down and is about the sort of thing you get with no special precautions taken.  I would expect any 3F currents flowing in the CPC to be a lot lower.

Mike.

thanks for your response Mike.

thanks for your response Mike.

forgive my ignorance, but does this depend on how things are connected, and specifically whether that PE is just a case earth? I imagine that this issue arises all the time with industrial motors and drives, which are generally 3-wire with a PE connection somewhere