Why must ground & neutral be seperate in subpanel?

[krw]

Then perhaps you should correct those who are mistaken?

No one in this threadlet.

[tgdenning]

This is great but it doesn't mention wire size. If the ground is smaller than the current-carrying conductors, doesn't it become a fire hazard?

-tg

[w_tom]

[Doug Miller]

Code violation -- all conductors for a circuit must be in the same cable, raceway, conduit, etc. You need to run a new cable, 3 conductors plus ground as I described.

Another violation -- you *need* a grounding bus bar in the subpanel. Among other things, doing what you describe leaves the subpanel chassis ungrounded.

[Doug Miller]

He's talking about using that entire cable as the grounding conductor for the subpanel. Still not safe, still not Code-compliant... but not quite as bad an idea as you assumed.

[Doug Miller]

No. Under normal circumstances, the grounding conductor doesn't carry current anyway. In circuits 40A and above, the grounding conductor is permitted to be smaller than the circuit conductors; see NEC Table 250.122 for details.

[tgdenning]

But the ground is meant to protect against a short to the 'case', so if a short happens, the ground will not be protected by the circuit breaker---it will overheat.

I went through this exercise in running a circuit to an outbuilding. I never quite figured out what the code meant with respect to a ground for that building as well. There was something about livestock, but I don't remember it now.

-tg

[Doug Miller]

The breaker will trip long before the wire will melt.

[Charles Schuler]

I should have said "Not a good idea."

Problems are they are supposed to labeled as "ungrounded" and external testers won't work.

[Bob F]

Specifically, because it is a "short" condition, rather than an overload, which can last much longer.

Bob

[Mark Lloyd]

A lot of people read these things without making their presence known. Some will make up the wrong thing.

[Mark]

Will the TEST BUTTON on the GFI work if it does not have a ground?

If so, how does it work?

Mark

[joseph.hollyday]

Again, Thanks for the input. To clarify what I should have written, my plan was to tie my new cable into a ground bussbar which will be secured to the metal of the subpanel. The current neutral bar either floats or is grounded depending on one screw which makes the ground connection. I will remove that to make it float. Regarding the guage and being seperate from the supply, well sometimes something is better than nothing. 3 #12 conductors all tied to the main panel ground bus and to the sub panel ground bus is better than the current situation. Each #12 conductor is good for 20 Amps, so in theory my ground wire(s) can carry 60A combined, and it is only a 40A breaker. A lesser evil than the current situation which has not been problematic so far anyway. When I shop for the bussbar I will price a length of #4 bare copper and consider running that to the main instead. The issue then will be finding a lug in the main large enough to bond it to. Regards, Joe

[Doug Miller]

Or, I should have noted, long before the wire will get hot enough to cause any kind of problem.

That's better... but it's still a Code violation. All conductors for any circuit are required to be in the same conduit, cable, raceway, etc. Furthermore, Code also prohibits connecting conductors in parallel unless they're (I believe) 2ga or larger.

Correct.

Undoubtedly. But it's still a Code violation. If you're going to run a new cable anyway, why not run the right thing, connect it up properly, and be done with it?

Doesn't matter -- Code doesn't permit it.

Understood -- but you asked for the proper way of doing this. What you propose is not.

It's still a Code violation, because it's not in the same cable as the conductors supplying the subpanel. You need to replace the existing feed with

8-3 cable (given that you're using a 40A breaker) with ground, so that all of the conductors are in the same cable. And you won't have any trouble finding a lug in the main that you can connect #8 to.

[Don Fearn]

I think it was "Mark" who stated:

Yes.

Very well [*]

What it does is measure the current leaving the "hot" lead and compares it with the current returning in the "neutral" lead. If those are different by more than a few milliamperes, there is a Ground Fault, that is, current is finding its way to ground through some other path than "neutral". If that is the case, the GFCI Interrupts (the "I" part of "GFCI") the Circuit, very, very quickly, so that no damage is done to whatever caused the Fault (that could be a human body). The test button simulates a fault.

[*] I had it work once for me. I didn't know it had until the lamp I had dropped into water wouldn't come on again.

-Don

[Goedjn]

Right answer to the wrong question. The GFCI will work, but unless there's a capacitor or something in there that I don't know about, the test button won't.

[Mark Lloyd]

Yes.

The one I examined connects a resistor between neutral on the line side and hot on the load side. Ground is not involved.

[Mark Lloyd]

Yes, it will. The test button has nothing to do with ground. It creates an imbalance by routing some current from the hot side of the outlet around the GFCI to the neutral wire connected to it. This is the same sort of imbalance caused by a wet person touching hot.

[Don Fearn]

I think it was Goedjn who stated:

A resistor, I believe. It shunts a small amount of current around the balance circuit simulating an imbalance in the two sides, so if it's working right the GFCI will interrupt the current.

-D

[Mark]

OK that makes sense, and the test will work without a ground and without a load...

thanks

Mark