Re: Running a generator on a boat?
I do a lot of things in my back yard that I wouldn't do at a corporate event......
It's my understanding that without a ground/neutral bond, GFIs don't work unless there is a ground. There needs to be a "zero" reference, there's no way for a GFI to differentiate a current loss without it. Somebody correct me if I'm wrong.
I can't find a definitive answer on boating forums, most of the references to this situation are from "bowfishers", and there are a lot of opinions but nothing concrete enough for me.
Matt,
The obligatory Code/Licensed Electrician/Authority having Jurisdiction disclaimer about dealing with electrical stuff goes right here. That said:
A GFCI/RCD looks for a differential in hot to neutral current. It will trip if this exceeds a certain small threshold, and is agnostic about the alternative path for that current.
In a classic three wire situation, with a standard breaker, if there is a fault to ground the current flows in the ground wire back to the ground-neutral bond point, and it is there that the neutral return path that acts as the current sink to trip the breaker,
not the ground. This is because the ground rod to soil connection typically has a fairly high impedance (Z), certainly higher than the neutral return path.
Indeed, if neutral and ground were not bonded, and the third/green wire ran straight to a typical grounding rod alone, it is conceivable that the conventional breaker would never trip as there would not be enough fault current flowing in the third/green wire into the ground rod and finally into the earth to trip the breaker.
For your
non-inverter genset,
unlike the Honda, the ground and neutral sink points (i.e. the engine block) are one and the same on the generator itself. The third/green wire merely provides an alternative (i.e. fault) return path to the engine block in the event a piece of gear shorts something internally to its case. The Honda is more complicated because of its putative connection to the mains panel of a residence in an emergency outage situation. See TJ's comment below.
The grounding rods we drive have several nominal purposes that they perform with various degrees of success:
- Path to earth for lightning. This works much better in a building full of steel in the earth than with a typical ground rod tied to a genset or stage. Indeed the impedance of the ground rod at the much higher frequency components found in lightning could be very high due, rendering the rod effectively an insulator.
- Path for fault current. As we discussed above, the majority of the fault current instead follows the neutral return path at the ground-neutral bond point in the typical case with a high Z path to earth through the ground rod.
- Genset potential reference to earth. Holds the voltage potential of the generator near the local ground/earth voltage potential, so a conducting path between the genset and (physical) ground has as few volts of potential as possible. This works well as long as the amount of current in the grounding rod is essentially zero. If substantial current flows in the ground rod, its potential is raised above the earth potential via an Ohm's law calculation ( V_raised = (I_ground rod) * (R_ground rod to soil) ).
Ideally the gensets we use would be tied to ground through an essentially zero impedance connection, but that is rarely the case. As long as there is not a substantial fault/loop/local galvanic current in the ground, a standard grounding rod will hold the generator's potential near the earth potential. Once current is flowing into the soil from the grounding rod, though, the potential of the genset will rise above the local earth potential.
In the (not always) hypothetical circumstance where the generator's potential has had an opportunity to float substantially above the earth potential, the genset should be kept physically out of contact by those whose reference would be local earth potential, and who have a conducting pathway to the earth. It is always prudent to de-energize the genset through an insulated path, in the event it has floated above the local earth potential by whatever circumstances.
PS The measurement of the ground rod to earth impedance is similar to what is known as a four probe measurement, and can be conducted using a meter like
this B&K.
