Quoth George Little:
What Winnie is saying is only true when there is fault current involving the grounding conductor. Normally you wouldn't have any current flowing in the grounding conductor hence no magnetic field.
For my points 1 and 2 I agree with George above. The issues that I was describing would only occur when a fault current is flowing, and would act to change the impedance of the fault current path, and would also act to dissipate power in the conduit system. I should have stated this more clearly.
But for point 3, I am specifically pointing out that there is _always_ some noise current flowing in the conduit system, even without an electrical fault. This may be caused by things like ground coupled TVSS devices, ground referenced signaling devices, magnetic coupling to the conduit system, etc. The whole reason for an 'isolated' ground is to provide a _single_ path right back to the building ground reference, with no loops. By running the 'isolated' ground via a different path from the circuit conductors, you permit currents in the conduit system to transformer couple and induce voltages in the isolated ground, IMHO eliminating the benefit of the isolated ground in the first place.
I should add that this is conjecture on my part. I do not know the magnitude of the noise voltages on conduit in industrial settings, and I don't know how much current is actually flowing, so I don't know the magnitude of the transformer induced voltages that I'm describing. But running the isolated ground in a way that would pick up electrical noise seems as 'wrong' to me as spending extra money for an IG receptacle in a home wired with NM-B.
-Jon
