| I may soon be installing a generator panel. Fact is, I've never seen one, | much less installed | one. | | Rumor has it that during the ice storm of ???1998???, hydro workers were | shocked due | to potential on the neutrals which resulted from generators hooked up to | houses where the | main breaker was turned off and therefore the hot wires were not energized | but the neutral | was not disconnected and therefore had potential (wheww, run on | sentence....). | Problem is, I don't understand how this could occur. Since the neutral must | be grouded at | the service, it cannot have potential. With the neutral tied to ground this | becomes the | reference voltage against which all other ungrounded conductors are | measured.
There is an old saying that goes like this:
"Electricity follows the path of least resistance"
*ding* False!
Electricity follows all paths. It just follows each in inverse proportion to the impedance.
If the generator, or the service entranced from the generator, has neutral bonded to the grounding conductor or a ground electrode, then any contact between the neutral and ground forms a new path back to the generator. If the intended path is very low impedance, then the higher combination of people and ground impedance will have very little return current. However, if some faulty condition exists, such as an open neutral, then the people plus ground path could be the only path, or at least the most substantial path, back to the source.
This kind of problem happens very frequently with neutrals from utility sources. The same thing can happen with a generator. And it can be worse because people are expecting no power when the utility is off.
| I gather that generator panels disconnect the mains but not the neutral.
Many are like that. And doing so means the generator is *not* considered to be a separately derived system. And a system is supposed to have just one bond between neutral and ground. But often there are two (one is at the generator or its entrance) and that leads to neutral currents going back to the selected source going back over both neutral paths from the switch, and then crossing between them at the grounding wire.
| Two possible explanations come to mind: | 1) there's no truth to the rumor | 2) inadequate grounding leads to potential on the neutral
There is no such thing as grounding that is adequate enough to eliminate potential on the neutral. The best you can do is have a sufficiently low impedance to bring ground and neutral near equipotential. But when earth ground is part of that, all bets are off.
Here's some fun reading [PDF format]:
formatting link
| If 2) is correct,how would one best test for, and ensure, adequate neutral | grounding?
I would go with a switched neutral on the transfer switch. Each source would be separately grounded, and ONLY the grounding wires would be connected between systems (not the neutral). The code requires that such a switch ensure that with making contact, the ungrounded lines be connected last, and when breaking contact, the ungrounded lines be disconnected first (to ensure you don't have a momentary open neutral scenario).
If you are exceptionally paranoid, then don't source a neutral at all, and have a transformer after the transfer switch where you get your neutral from (and it gets bonded to the grounding conductor between the secondary winding and the secondary disconnect).