Suppose for a moment you had a circular copper disc placed concentrically in the air gap between the poles of an upper and a lower circular magnet, the outer faces of the upper lower magnet being connected by a pole piece so as to complete the magnetic circuit. The edge of the copper disc is connected to a first wire and the centre of the copper disc is connected to a second wire that passes through a hole in the centre of the upper magnet and the pole piece so both wires are unaffected by any airgap flux. Basically a Faraday homopolar motor/generator setup. Nothing new so far. If now the magnets and pole piece are held stationary and the disc is rotated about the axis of the assembly we will get a small emf induced across the radius of the disc because there is relative motion between the flux and the disc. Conversely, if the disc is held stationary and the magnets and pole piece are rotated around the axis of the assembly, people more knowledgeable than I say that there will be no emf induced across the radius of the disc because the magnetic flux does not rotate axially with the magnet assembly so there is no relative motion between the flux and the disc. Let's now switch from generator mode to motor mode and where the disc is held firm and the entire magnet assembly is free to rotate (within the limits of the pole piece going from the outer faces of the upper to lower magnet colliding with the wire connected to the disc edge). If we now pass a current through the radius of the disc, the unmoveable disc will exhibit a torque in a certain direction and the magnet assembly may or may not exhibit a torque in the opposite direction. If the magnet assembly does exhibit a torque it will rotate until the pole piece hits the wire. Here is where things get interesting. If the magnet assembly does in fact exhibit a torque and rotate a certain distance then it is capable of doing work. However, established thinking says that the rotating magnet assembly will NOT induce a counter emf across the disc radius because magnetic flux does not rotate axially with an axially rotating magnet as it's source. This means that (ignoring dc resistance) there can be a supply current through the disc radius but no voltage across the disc radius because there is no counter emf generated by the rotating magnet. Watts input = ? amps x zero volts so there would be zero watts input to the system despite torque and motion = some watts output! The alternative to this of course is that the disc would exhibit a torque in one direction but the magnet assembly would not exhibit a counter-torque. Just what IS going on here?