I would like you to cite that article.
What you're *really* asking, though, is "If I have multiple, redundant ground paths, do they ALL need to be large enough for the largest conductor?" And that's easily understood with a thought exercise. Suppose you run PVC conduit to a laundry room. You provision a 30A dryer recep and the mandatory 20A/120V laundry room circuit. You choose the following conductors:
- 10 AWG, H-H-N-G
- 12 AWG, H-N-G
However, the boxes in the laundry room are metal with EMT between them, so they are a valid grounding path, 250.118(4). Your first box has 2 ground screws so you ground both wires there. Is the 12 AWG ground a codevio because it can't support the dryer?
Of course not. Nothing wrong with it. The second ground is redundant, and the other ground is good enough for the dryer.
In this case, the EMT is the primary grounding path and the #14 is redundant.
It will be two 20 amp circuits. I plan on running a 12 gauge ground.
I was planning on grounding the metal boxes and each receptacle. Is this over doing it?
Grounding the boxes is required. Grounding the recep *may* be overdoing it, depending on how the recep mounts in the box.
I've read that grounding to the box is only necessary if the ears of the receptacle are sitting on drywall.
Thats a "web oversimplification" of a somewhat complex issue.
The deal is, here's what's not good enough for a receptacle: resting contact between mounting screw head and yoke. (that *is* good enough for a switch). So if your recep is proud of the metal box, and the only ground route is through the yoke and screw head, not good enough.
A "self-grounding recep" that has brushes which touch the mounting screw threads, that is good enough. NEC 250.146(B).
Otherwise, all these must be true on one attachment point:
- The yoke makes hard flush contact with the metal box.
- The metal box is bare, and is not spooged up with paint, rust, etc.
- Neither is the yoke.
- Those little paper squares that keep the screw captive, cannot be there.
Does the raceway need to be bonded at the panel, or is the nipple connection sufficient?
If the EMT conduit is properly assembled, then the EMT to the steel service panel case is a valid grounding path. Trust me, I've tested it lol.
Code can be funny about concentric knockouts (you knockout the 3/4" hole but around it is a 1" hole that could be knocked out) or stepdown adapters (you knockout a 1" hole by mistake and use a special washer to downsize it to 3/4"). But I can't find anything in Code that pertains to branch circuits or feeders.