I know it's code (refer to NEC 250.24(A)5 and 250.32), but I don't understand why you can't use a bonding jumper to connect the ground bus and neutral bus in the subpanel? I've been told it's because then you'd have current on the grounding conductor. In my attempt to understand this, I was thinking maybe it's a more of an issue of having parallel neutrals… but if they're grounded on both ends, how would that (not) work? Possible safety issue: if one of the grounds becomes disconnected, then you have a true parallel neutral… is that a viable or better explanation, or is current on the grounding conductor the real issue?
If so, then why isn't current on the grounding conductor(s) from the main panel a problem? I don't see the inherent difference between a subpanel and a main panel. Ultimately this question is, why bond neutral and ground in the main panel, but not the subpanel (besides NEC said so)? Why is the neutral from the main panel to the subpanel inherently different from the service neutral to the main?
edit- you may also refer to: How to properly ground a subpanel in detached building? for a relevant picture and because it is a related thread.
Best Answer
Ground and neutral are not parallel neutrals. I know it looks that way because they're bonded in the main panel. But shift into a different way of thinking about the purposes of the 2 wires. Think of the ground solely as a safety shield.
Let's try a few pairs of examples. The first is Code and the second bonds at the sub-panel also. The orange glow is on things which are "hot".
Seems awesome right? Poor old Code Man is in the dark. His power tried to return via neutral, and neutral is broke, so the power failed. Rogue Man is one happy guy and his life isn't disrupted. Ground is working great as a "backup neutral". He doesn't even know he has a problem!
Of course, ground is a thinner wire, so it might overheat, but so what? Or, what if both ground and neutral were cut?
Code Man is still in the dark and he's still gotta fix those wires. Rogue Man is dead.
In Code Man's installation, the hot went through the bulb, looking for neutral. It didn't find it, so it pulled the neutral up to 120V ( not enough power for useful work, but plenty to shock). It did the same for Rogue Man, but since he tied neutral to ground in the sub-panel, ground is now also 120V, including the service panel cover and the switch plate cover screws.
Suppose the sub-panel has its own ground rod. That doesn't help much. Earth tends to have high resistance, so the cover screws might be 103V instead of 120V.
I have the good fortune of working in EMT conduit in a steel building, which naturally forces the entire conduit system to ground. Ground is never part of the circuit in any way whatsoever. So I get to see it as intended, as a protective "shroud" around all things electrical.
Ground isn't quite yet a perfect envelope. It is in new work, but we still have a lot of old wiring out there that is not practical to outlaw entirely - such as NEMA 10 and switch-loop smart switches which poach ground as a neutral.
Why bond neutral at all?
That's a GREAT question. Not bonding ground would give you an isolated system. And that makes a lot of sense in some ways, like solving some of the problems you see above. But it has other disadvantages. I go into depth about that here.