Caution: this particular field is littered with _tons_ of pseudoscience claptrap. At the present time, it is not good evidence to really understand what (if any) physiological problems can be caused by 'EMFs'.
For the purpose of discussing things with gunther's acquaintance, EMF is 'Electro Magnetic Field', and it is the low frequency (60 Hz) near field (close to the wires) magnetic field produced by current flow in power conductors. This use of 'EMF' should be differentiated from 'Electro Motive Force' or voltage. It is also distinct from Electro Magnetic Waves (light), in that this is the near field effect of the currents in the wires, and is stronger close up, but falls of more rapidly with distance than the far field stuff.
As I said, there is no good evidence for effects from these fields...but do a web search for 'Transcranial Magnetic Stimulation', and understand that there are _well known_ effects on the brain from magnetic fields that are 1000-10000 times larger than the ones that the EMF people are worried about. Given that very strong fields have significant effects, it is plausible that very weak fields have some difficult to detect effects.
Because we are talking about near field magnetic fields, the things to look for is area contained within current loops. The lower the current, or the lower the area of the current loop, the less energy stored in the magnetic field. Ideally all current should return on a parallel conductor that is quite close to the source conductor. The net current should be zero, and the magnetic fields of the two conductors should cancel.
Check for circuits where conductors at the same potential have been joined; current can flow to the load on one circuit, then flow from the load on the other circuit, introducing a very large loop area. (Note: I understand that we are talking about AC circuits, and that the current flows back and forth...) The join can either be on a phase conductor, or on the neutral conductor.
Check for neutral to ground faults. These won't trip a breaker (though they will trip a GFI), and can introduce significant currents through the grounded structure.
Check for circuits wired with separate phase and neutral conductors, as is the case with old 'knob and tube' wiring.
I believe that a ring circuit _operating normally_ can introduce significant loop current. The impedance at every splice would need to be exactly balanced hot and neutral to eliminate loop current.
-Jon
