I find this post very interesting and amusing. For the IGC to be even remotely effective, the IG EGC should originate at the service N-G bond point. That would mean from the service disconnect device you would need two EGC circuits and two ground bars in every panel down stream to accommodate the IG EGC. I agree with Bob (iwire), the IG EGC circuit could be done with 12 AWG.
Here is the funny part. There is no NEC requirement to do this, NONE... If you wanted to do this cheaply, you could simply originate the IG EGC at the panel that serves the receptacle by attaching the green/yellow IG EGC right to the EGC bus in the sub-panel. IMO according to the NEC you could go even cheaper by simply bonding the IG EGC pin of the receptacle to the box the yoke is mounted in.
Enough nonsense, here is my recommendation as an engineer in the know (sorry Bob, couldn’t help myself): Instead of running IGR, run dedicated 20-amp circuits to the locations. If you really wanted to make a PQ circuit, oversize the grounded circuit conductor or even both the N & L conductors. Some electronic equipment is sensitive to N-G voltages. By over sizing the neutral conductor you minimize the N-G voltage to less than 2-volts peak to peak (IEEE spec).
[This message has been edited by dereckbc (edited 02-05-2006).]
