Hi,
I wonder what would cause an outlet tester to read "bad ground" when all the grounds are installed?
I have a new out-building that has a 240 V-3 Wire, 60A service to it from a 200A main breaker panel. The building has 3-8 foot two tube fluorescents, a motion detector flood outside, 3-6 outlet surge suppressors designed for shop use, 4 reg recepts and all is fed from 2 gfi on two separate 15A ckts. The 70A panel has an isolated neutral. There is an 8' X 1/2" copper clad ground rod installed with a #8 gec. The neutrals and grounds are separated at the sub panel but not at the main panel. The sub panel is not bonded.
I get this reading from the first outlet on each circuit and then from the slave outlets.
The tester trips the gfi when the button is operated. Does this have anything to do with the ballast in the fluorescents?
Thanks for any replies.
-Regards
Greg


[This message has been edited by mustangelectric (edited 01-24-2005).]

Gregg, something I do not understand. You said the sub-panel is fed from MP via 3-wire, and the SP has isolated neutral bus. I assume by 3-wire you mean L1, L2, & N with no EGC. If that assumption is true, then your meter is correct, you do not have a ground.

If you run 3-wire, meet the conditions of 250-32 ( I think that is the code reference)of no metal paths, then you need a N-G bond in the SP. Pesrsonally I would always require 4-wire to be on the safe side.

Hi,
Thanks for the reply. Where is the NEC requirement to run a 4 wire service to an out-building? I established the ground at the separate service. Isn't this considered a separate service? Or it is treated as a separate service?

Are you saying that since it is a separate service then I need to BOND the panel and not treat it as a sub? I thought any panel after the first ocpd had to have an insulated neutral except for an out-building?

You must run all of the circuit conductors together for a sub panel in a single building but if you leave that building you can establish a ground at the separate service thus dropping the 4th wire.

So any three wire feeder needs to be bonded at the panel?

A four wire would not get a ground rod so I would not bond at the sub if in the same building.

I almost always run a 3 wire feeder to a outbuilding and install a ground rod there.

If this is treated as the first ocpd then it would have to be bonded.

-regards

Greg




[This message has been edited by mustangelectric (edited 01-24-2005).]

If you run 3 wire to an out building, you must bond at the sub panel. You also need to make sure no other metal path occurs from the main building to the out building. Either a 3 wire method or a 4 wire method can be correct if all the conditions are met. Right now you do not have a ground path back to the main service.

If you do not take an EGC to the separate building, you must bond the neutral in the panel. Otherwise, the only path back to the source in case of a ground fault, is a very high impedance path through the earth, which is likely not to trip the breaker.

Also, you must always establish a ground reference at a separate building (ground rod, ufer, etc.), even if you do pull an EGC with the circuit. The only exception there is if you are only feeding the building with a single circuit, then you don't have to drive a ground rod.

Laura Jenkins

Do not have a code book in front of me, but the reference is either 250.26, or 250.32 for your application.

It gives you two options for a sepearate structure.

A: is to supply the structure with 4-wire feeder (L1, L2, N, + EGS) just like any other sub-panel. The exception is you have to supply a ground electrode at the new service and bond it to the panel ground. But the neutral bus is not bonded, that bond exist at the main panel:

B: Is more restrictive. Absolutely no metal paths can exist between the two structures. If that condition you can run a 3-wire feeder (L1, L2, & N). With this option you set a ground electrode and GEC, and bond the neutral to the GEC just like setting a service.

As I stated earlier I never specify option B unelss I am 110% certain no metal paths exist like new construction. I use option A 95% of the time. By going Option A route it does not matter if a existing metal path exist or not, it is always safe.

You need to either add the N-G bond if the conditions allow, or add a EGC to your feeder. Either way a ground electrode is required.

Greg, the article you are looking for is 250.32(B)

Roger

Roger, THX. I can never seen to remember separate structures from common service. Don't ask me how I can remember the requirements and not the numbered reference.

Hi,
Thanks for the replies. I appreciate the code references and comments.

I am sure that the bonding will give me a good reading. I am glad I tested it and I am glad I asked!

sincerely

Greg



[This message has been edited by mustangelectric (edited 01-24-2005).]

You're welcome Dereck.

Roger

Hi,
Went back and tied in the bonding strap to the neutral bar. Tested the ckt again and got the green light.

There was no bond between the two systems.

That little tester just paid for itself!

Thanks for the help.

-regards

Greg

Just think back, you called me a hard-a$$.

Hi,
Thanks again for the replies. Derek I am sorry that something I may have said may have upset you. I do not remember using the term you used but if I did please accept my apology as it was not my intention to offend.

Sometimes e-mails and post can leave a lot to be desired as far as conveying a message. What may sound just fine in the spoken form may not come across the same in written form.

I vaguely remember the post but I also remember misinterpreting the comment.

The great thing about the site is that it allows a lot of different opinions.

Thanks for your feedback.

Once again, my apologies.

-regards

Greg

Greg, no need to apologize, I am poking some fun at you. It was a post about me firing an EC for some tower grounding work. There isn’t anything you can call this ole redneck that does not apply. I am just south of you in Ft. Smith and work for the major wireless telephone company in your area. You can figure out whom by “You Got That Right”. May look you up when building out up in that area if you can take directions from a SOB.

This thread reflects the confusion that there is, even in the NEC, regarding separate buildings on the same service. Heck, even the IAEI asked three of its' folks, and got three different answers! That's one reason the last two code cycles have had do much re-arranging in section 250.

KISS. Ground rods are for lightning, and only lightning. Each building gets one (or similar electrode).
The green wires are for bonding everything; that is, giving lost electricity a way to find its' way home again- "home" being the service point, where the ground and neutrals are connected.

Since a ground wire is a ground wire, and a neutral is a neutral, the two systems are kept separated everywhere else. You wouldn't want your conduit acting like a "neutral", would you? So, we don't make parallel paths possible.
Since pipe is inadequate for cleaing all but the smallest faults, we run a supplimental ground wire. While this wire is often quite a bit smaller than the feeder wires, it's there.

The effect of all this is to require that sub-panels have four wires run to them: two hots, a neutral, and a ground.

Just one quick one
You said # 8 ground wire
we use # 4 up here in upstate NY

Hi,
Thanks for the replies. Why would you want to use a #4 on a 70A panel? Take a look at table 250.66.

Another question I was thinking about..

Why would a ground rod at a pool pump panel not be required, especially if it is all about lightning?

-regards

Greg

Read 250.50. If the pool pump panel is in a separate building or structure than the main service an electrode is required at that building or structure. The pool pump panels I have installed were fed with a 4 wire feeder, neutrals and grounds separated and an electrode installed and connected to the equipment ground bar.

Mustang, I was trying to make a simple summary of an issue that, as I mentioned, even the NFPA gets confused over.
Just to play Devils's advocate, the rebar of a pool is required to be bonded, etc....making it identical to a Ufer ground, for all intents and purpoese. Remember, the "rod" is but one of several electrodes permitted.

For the GEC, the NEC gives a minimum size of #6, though many places require #4.
For the bond between panels, the water bond, etc., our local code mandates a minimum #8 size.

btw
as mentioned
If the pool pump panel is mounted in or on a structure the grounding electrode system mentioned in 250.50 is mandated.
and just to clarify

The structual steel of a pool and its use and interconnection to other non current carrying metal parts is addressed in 680.26 and that is bonding.

Yes here our inspectors require #4
and make us use solid #8 on pools and tubs
unless your talkin olimpic size pools then we use the tables and that good old cad-weld
pop fizzzzzzzzz

Hi,
Thanks for the replies. Then, according to what is being said here, if it is IN a separate building ALL sub panels require a ground rod. (does that include ON a separate building or structure? If not then I digress.) If it does include ON what will you support the panel with? When you put a screw through the hole of the panel you are attaching it to a structure. Even if you mount it on unistrut or treated post it is a STRUCTURE.

Another thing, I am not clear on what you mean by the following statement "For the GEC, the NEC gives a minimum size of #6". Looking at the 2002 NEC Table 250.66 you are allowed to use #8 as the minimum for the GEC. You are not required to use anything larger than a #6 copper for a driven rod. Is that what you were talking about or did I misunderstand your statement.

Looking at Stallcups Electrical Design Book, I have to disagree with the statement that the ground rod is used soley for lightning protection. The electrical system will not function properly without it as a whole.

250.32 covers two or more buildings or structures fed from a common service. 250.50 covers the requirements.

It is a myth that electricity always seeks the least path of resistance to ground.

Thanks again for the comments.

Regards

Greg



[This message has been edited by mustangelectric (edited 01-29-2005).]

Greg
Could you elaborate on why the system will not operate properly without groundrod(s)?

Pierre

Hi,
Thanks for the reply. I am no expert on grounding but art 250.4(A)(1),(2),(3),(4) and 250.4(B)(1) points out several other purposes.

Maybe someone else out there agrees with me. It doesn't really matter I am not trying to prove anybody wrong, I was just surprised to see the statement that the ground was for lightning only. That did not sound right to me.

I agree that it is a primary component.

Grounding is probably the least understood part of the NEC.

I just do not see why a ground rod would not be required at a sub panel for a pool or hot tub but would be for a tool shed with two branch circuits.

-regards

Greg

Mustang, thanks for your patience with me!

Each building needs a ground rod. It can have more. All rods, and all panels, have to connect to it in some manner.
If you had one panel into the building, that in turn fed the other panels, then only that first panel would need to be connected to the ground rod.
If you had two feeds coming into the building, both panels would have to tie into the ground rod, Ufer, etc.

You're right about one thing, though...I got the NEC backwards as to GEC size. For NEC, #6 is "max." while our local rule is #4 "min."

Far be it for me to quibble with Stallcup, but I point out that a number of countries...Norway comes to mind...do not use ground rods, etc. Last I saw, their industry was just as modern as ours. That is- and I stress this- as long as we are talking only about the ground rod and the wire to it from the panel.
In another sense, though, it IS essential to "ground" your system; that is, have a "grounded conductor" or 'neutral'. Simply put, really weird things, including large transient voltage surges, can occur as loads change on the "hot" circuits, unless there is some point where the different feeds (transformer windings) come together.
How can this happen? Well, imagine a 240 "delta" system. Should a short happen, there would be a 208 volt to ground potential. Yet, your transformers are isolated from ground. The result of this is a wildly varying load on the windings, heat, etc. Yet, if you deliberately ground part of the delta, you can control the system, the windings can 'stabilise,' because there is a nice, clean path back to the system.
Put another way, a completely ungrounded system is esesntially a system where all loads are in "series", rather than "parallel," so changes in one load greatly affects the others.