The service comes down to the pole into a box. That box has a 100A breaker that runs to the shop, and a 200A breaker that runs to the house. The pole to shop is about 200 yards and there are 3 large wires.
The meter and service are at the pole. The house and shop are subpanels.
The shop and house panels could have hot-hot-neutral-ground feeds and be wired with a separate ground bar and no N-G bond. Grounding electrode(s) required (also required in the other method).
Your description is hot-hot-neutral feed. The shop and house panels are wired the same as a service panel. Neutral-ground bond required (neutral and ground bars are effectively the same). [This method is not allowed in the 2008 NEC for new installations. Seems like what you have is common and safe.]
With a separate ground bar I suspect your panel doesn't have a neutral-ground bond. If missing it could be a mistake or a local wiring practice or inspection authority quirk. IMHO it is not safe. The NEC does not allow the earth to be used as a return. That is because the earth does not provide an effective return. The need for a metal (not earth) path for ground fault current back to the transformer is not understood as well as it should be.
Bud -- thanks for all the info! You a a bit over my head, but I'll try to figure this out.
I'm looking into my shop panel now:
There a 3 large wires - 4 ga. aluminum - coming up from the bottom (from the pole). One goes to the left and is fixed to the neutral bar, the other two go to the bars where the breakers attach (one on the left side, the other to the right).
The neutral bar has a place to afix white wires on both the left side and the right side (the panel hardware attaches the right bar to the left bar, so neutral whites could be brought in to either side -- in my case all are on the left side where the big wire from the poles is also.)
There is a ground bar attached to the grey metal box.
All circuits coming in connect bare ground to the ground bar, 110v whites to the neutral bar, 110v blacks to the breaker, and black and white wires to the 220v breakers.
From what I see, ground is not tied to the neutral in any way.
The ground bar has a larger wire running out the bottom of the box (but not with the 3 other large wires) and I see that wire ouside the shop going into the ground.
I don't quite understand your discussion about the ground and neutral being connected in the box.
Resorting to my old analog multitester (bec. the battery in the cheap digital one is dead...) breaker bar L to neu bar - 120v breaker bar R to neu bar - 120v breaker bar L to breaker bar R - 240v breaker bar L to grd - 120v breaker bar R to grd - 120v neu bar to grd - 1v (a trace of voltage, but not 0) resistance between grd and neu - 20 ohms (that's 2-0, and not K, etc, I was on the meter 1x setting)
Bud --
I did some reading tonight in some 'code' book at Lowes. I now see what you mean by neu-to-grn "bond". There is suppose to be this "bond" in the service panel, and then any sub panels should be fed by
4 wires - hot1, hot2, neutral, ground.But my shop panel is fed by 3 wires - hot1, hot2, neutral. There is only an earth ground in the panel.
I guess I should go pull the cover off the panel on the pole (which feeds the shop on a 100A breaker and the house on a 200A breaker).
What should the pole panel look like?
What should my shop panel look like? (given it has 3 wires from pole panel)?
What should my house panel look like? ( I need to look at it also)
. The voltage of interest is N-G. If there is voltage from N-G you can't measure the resistance. 20 ohms could be a reasonable measurement if the shop subpanel has no N-G bond and the path is through the earth. .
. That is true for a subpanel in the same building.
It is one way to wire subpanels in a *different* building. .
. That was a second way of powering a *different* building. But the NEC then requires a N-G bonding jumper. Without the jumper there is no metal path back to the transformer, and a short-to-ground will not reliably cause a breaker to trip (and can make the "grounds" hot). .
. Who cares. .
. You still don't know where the grounding problem is. From what I remember it could possibly be one circuit (but probably not).
Connect a light bulb, 100W or larger would be nice, from a hot to ground. Is the light fully bright? What is the voltage between the neutral and ground bus? Note all the grounds in the building, including the panel, may be "hot" when you do this. If you actually have 20 ohms N-G, a100W light bulb might give you about
15V from N-G. I would expect well under a volt if there is a N-G bond.As I said a couple times, the N-G bond is often a screw that looks like it could just be a mounting screw for the neutral bus. Recent ones are probably green. If not installed there will be an empty hole in the neutral bus. .
If the shop panel does not have the N-G bond I would certainly check the house (or check it anyway).
You may want to call an electrician.
(There is a minute possibility local practice is to not add the N-G bond. That would be very unwise.)
HomeOwnersHub website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.