A contractor runs PCV to a detached building. Proper sizing, three wires and has a ground rod at the building.
I give him a red tag for Article 250.52 (NEC2005) in which a lively debate starts.
When you run a sub-panel 3 wire feed or 4 wire feed. I always run 4-wires plus a ground rod. What do you guys think?

Joe, although I would run an Equipment Grounding Conductor because of preference, it is not required if 250.32(B)(2) applicable.


Roger

Thank you Roger, have a great day

Roger states that it is not required.. true, but running a seperate EGC would be alot better. It would avoid the whole paralell path thing down the road, when your not there.


Dnk....

[This message has been edited by Dnkldorf (edited 12-19-2005).]

[This message has been edited by Dnkldorf (edited 12-19-2005).]

What does 250.52 have to do with 3-wire -vs- 4-wire feeders? Why would you give him a violation of 250.52 when you said he had a rod?

Question: As Roger points out, if there is no grounding conductor run with the feeder, the grounded conductor is used to bond metal frames, etc....does this mean that the neutral bar is not insulated/isolated from the ground bar as it would be if this subpanel was located in the main building where the service is located?

HC ... Yes, it looks just like a service entrance.

"I give him a red tag.... in which a lively debate starts."

The color red gets most really lively.

So did he get a green one yet, or an orange?
(That is, if there are no bonded metalic paths between buildings or not.)

Even the IAEI, when they presented this application to three of it's own experts, got four answers!

Codewise, the guy can probably get away wit only a ground rod, treating the feeder as if it were a separate service.

I am something of a simpleton, and like to consider the "service" to be where the PoCo drops in, with everything from that point to be part of the same system. But I can see where the code might not limit me to that approach.

In my construction days I always ran four wires. I learned fast that just because the contractor does not do it like I do means it is incorrect.
Yes red-tagged always give lively debates but I always call the contractor when it is going to cause extra work and money.
I gave him a white/purple sticker, that is what New Jersey gives out when the final passes.
Thnaks for all your feed back

"I learned fast that just because the contractor does not do it like I do means it is incorrect."

Is that a typo? If not, I am glad you don't work here... Is there a failure of coffee and donuts being provided?

Not trying to be disrespectful, but the code is not always cut and dry. There are options to design, and for the most part conditions for them. Just because one way is more common, does not automatically mean its wrong. I keep a code book in my truck, (As we all should) but often will have it hand for an inspection if I have a vaguary, or something un-orthodoxed. Not that this particular situation is. (You can also be guaranteed I checked before I installed it.) The code book also makes a good backing when they sign my card, but also if there is any question, its right there. Most Inspectors don't have any problem with me handing to them when they question something. That said, I have had only one red tag, (Not my fault) in 15 years, other than prefferential corrections requested by Inspectors. Which, I technically did not have to do. Most Inspectors here will often sleep on it, and not give any tag or a temp, unless it it is totally obvious, and un-disputable. Even willing to bring the card, and pick up the tag from you. Anyway, not saying I'm never wrong, (And I have been at times) but if it's right, expect a debate of some kind. "Do you take suger and cream in your coffee?"

Now there is a statement that can be interpreted as meaning two different things, depending on how you 'pause' when you read it. Joe has definitely been reading codes for a while as he has taken on code writing style.

No offense meant Joe.

Just curious, I'm actually on a job where I am running 4 wires because the only place where the ungrounded conductor and the ground can be bonded is in the main panel (service entrance). Any other location to panels downstream the neutrals and grounds have to be separated to avoid loosing a neutral and the ground picking it up and carrying it back to the main panel or creating another path. How could you treat this as a separate service and only run three wires?

Ky the "open neutral" scenario is not the reason for the rule about regrounding the neutral. They just don't want voltage gradients within the building. When you get to the next building, with a new ground electrode system, you can start a new "pseudo service" regrounding the neutral ... as long as you don't have a bonded metalic path back to the previous building.
The ramifications of an open neutral in building #2 are exactly the same as one in building number one.

That is not entirely true. There are a few instances where you may be allowed or required to bond the grounded and grounding conductors.

Separately derived systems for one.

Separate buildings or structures for another.

Check out the option allowed by 250.32(B)(2)



The 'other path' problem is covered by this part of 250.32(B)(2)


I agree with Greg this has nothing to do with the possibility of loosing the neutral.

Why is this parallel path only a problem on the load side of the service disconnect??? There are often multiple parallel paths on the line side. What makes it safe on the line side, and under the same conditions, unsafe on the load side?
As far as using an EGC for the subpanel, in one respect it is not better. The grounded conductor will be larger than the EGC and make a better fault clearing path than using the fourth wire for a remote building subpanel.
Don

The thinking is that you are establishing a grounding system bonded at the service (or 2d building disconnect)and any ground shift would make the ground reference rise and fall with that shift. The reality is dependant on the quality of your ground electrode. That is the reason why I like the Ufer, that also catches the steel in the floor (and walls in Florida). Add the bonding to plumbing and equipment.
If you do get a ground shift from the utility grounded conductor, it will be reflected in anything you are likely to touch and you are literally the bird on the wire.

"Ground" is really a theoretical condition anyway. We measured up to 35v between buildings in some surveys we did. As soon as you start stringing other metallic paths between buildings you have to really start thinking about bonding everything to a single point. That will be the 4th wire.
In our data applications we did use a separate EGC and we supplimented it with big bonding wires run with the data lines. There ended up being a lot more copper on the ground side than what we had on the phases.

gfretwell,
Grounding really doesn't do anything to provide a stable voltage reference. If there is voltage on the system or conductor that is being grounded, it doesn't go away when you ground it. In reality, the grounded conductor that the utility supplies is not at "earth" potential if you measure to "remote earth" (earth not close enough to the grounding electrode system to be influenced by the voltage drop in the earth around the grounding electrode). The utility grounded conductor has a voltage to remote earth equal to the voltage drop on the primary and secondary grounded conductor from the source to the point where you take the measurement. There is no way to eliminate this voltage. This is the main source of stray voltage problems in pools, and the code actually requires that we energize the pool bonding grid!!!

The code rule for single point bonding has only to do with parallel paths for grounded conductor current. You are still the "bird on the wire" when there are parallel paths. Parallel paths are required by the code on the line side of the service and prohibited on the load side. The electrons really don't know what side of the service that they are on, so why are the parallel paths safe and required on one side of the serivce and unsafe and prohibited on the other side of the service?
There are no studies or data to show that useful to solve data noise problems.
Don

Don, it is not "noise" problems we fix in data when we use fat redundant bonding conductors, it is to give lightning a more attractive path than the data line.
As for the rest, we pretty much agree. The function of service grounding/bonding is to provide the equipotential reference in the building.

I also agree the NEC is silent on the utility side of the disconnect.

gfretwell,
If the lightning path is anywhere near the data lines, I don't think that additional conductor size for the grounding conductors will make any difference. You need to keep the lightning outside of the building and away from the data systems. If you have to run data between buildings you need to use fiber so lightning isn't a problem.
Don

Don we woulds have loved to be able to talk customers into fiber but it ain't happening in the point of sale or PC world. It is not lightning itself that we were protecting against. It is just the ground shift and transients induced by lightning. I guarantee your problems between buildings will drop to close to zero if you pull a 8ga copper (it could probably be as small as 16ga), along with the LAN cable and bond the machine frames together. Also add ferrite beads to the data cables and the normal point of use protectors.

We were frying PCs and servers almost every afternoon here until we started doing this.

It became a standard for one major hotel chain after we proved it worked.
At the time we were ready to pee on the proverbial spark plug if we could stop this problem. Later the same principle got applied to a big insurance company's remote printer (RS422) and a major retailer's cash registers (IBM "S" loop).

I hear and understand you, I was just laying out the NEC rule as it applies to a separate building.

I never said it made sense.

I don't understand what "doesn't make sense".

You are trying to establish an equipotential grounding grid within a building. If you parallel neutral current in there it will see the voltage drop (actually "rise") on the EGC.
An example. Lets say you compute a VD (rise) on the neutral conductor in a sub-panel 100' from the service panel at 3 volts. Now if I used the neutral at that distant sub as the EGC the case of my equipment would be 3v above the potential of building steel. If I pulled an equal sized EGC and bonded it the rise would only be 1.5v but that is still not a good thing.
If this sub panel was bonded to building steel the steel becomes part of the neutral path, still not a good idea.
Once you get to another building, with no parallel paths you can start over with a new "ground". It will probably be a different level between buildings but if you don't have any other paths, who cares?

In the data world we had a parallel path and it was going to be the thing that resolved "ground shift" differences. Buy a few routers, LAN cards or system boards and a fat copper wire starts looking pretty cheap.

IBM was very careful to make the wire black and call it a "drain" so we didn't kick the article 250 tar baby.

What does not make sense is the invisible line drawn at the service disconnect switch.

On one side parallel paths are plentiful and required on the other side they are a violation.

Can you explain how this makes sense?

gfretwell,
With the price of media converters about $250 a pair and fiber about the same as plemun or wet location CAT-5, I bet what you are doing is very close in cost to using a fiber link, assuming that you are using Ethernet communications. Also I have been seeing a lot of wireless ethernet for point of sales terminals this year.
Don

Because the Service Disconnect tells the Electrons flowing through it, not to wander away from each other, and if they do (such as in parallel conductive pathways), the Service Disconnect gets mad and pouts all day.

The key problem here, is that the Service Disconnect has some minor control issues, and really should see a Counselor about it before things turn very ugly.

(quoted from an article that was produced by the newly founded "Co-Op"- between the IEEE, and the American Board of Professional Psychologists - the "AKA ABPPIEEE - EI,EI,O")

Man, I really need to get out more

Please do not read this message as an insult, just wanted to make everyone laugh!

Scott35

Don I have been out of it for a while but most of our interfaces were not ethernet. Fiber starts becoming a lot more expensive when you need a "S loop" (twinax, RG62AU or whatever) to fiber converter and then converting back. In fact I have never even heard of one. Fiber, overall, was a lot more expensive 10 years ago.
Stringing an 8ga was not really that expensive. In fact we got the first bunch for free from FPL. They were involved in the finger pointing along with IBM, the building owner, the tennent customer, the server vendor and an electrical contractor.

Bob, would it make more sense if we say the magic line is drawn at the grounding electrode conductor connection?
Everything downstream of that is a pure EGC with no current on it, referenced back to the ground electrode, while the neutral is free to fluctuate at the whim of unbalanced current, harmonics and voltage drop.

No it would not.

I think you may be approaching my question from the wrong side.

I am certainly not saying that parallel paths on the load side of the service disconnect are acceptable.

The question is as Don laid out.

Why are parallel paths acceptable on the supply side of the service disconnect?

That is what makes no sense.

That is true, right now I can buy a 4 fiber indoor cable for about the same price as #8 copper and an outdoor rated one for about twice the cost of the copper. You do have to add the fiber ends at about $12 each. We use a connector that does not require field polishing and you can make the fiber termination is just a few minutes.
Don

We work at quite a few campus style facility's and all of them are presently or are in the process of switching to fiber for all connections between buildings.

They will use the fiber for voice, data, video security etc.

Many are also using fiber for these systems in the buildings as well, with Cat 5 or 6 filling in the gap to the last device.

Bob, parallel paths on the supply side of the disconnect are _known_ to cause problems. I suspect that the reason that they are permitted on one side and prohibited on the other side is a combination of 'balance of risk/benefit', jurisdiction, and economics.

At the level of the distribution network, multipoint grounding is the norm. This makes the earth itself a big parallel path, and we have had stories in this forum of problems caused by this. I can't find the thread, but I recall an EC who was trying to figure out how a customer was getting shocks from a hose tap outside his house. The piping was copper, the electrical system properly grounded. But there was apparently an Earth voltage gradient between the house grounding electrode and the patch of soil next to the hose tap.

But on balance, I expect that multipoint earthing substantially improves overall safety because of things such as lightning discharge.

At the level of the secondary side of the transformers, we have stories of plumbers getting shocked cutting pipes, because the metal water distribution network was causing a very good parallel path between separate structures sharing the same transformer. My guess is that if the NEC could eliminate this risk, it would...but multiple buildings sharing the same transformer secondary are the domain of the POCO, not the NEC.

At the level of multiple services on the same building, this is a parallel path that is explicitly permitted, but we are talking a very short parallel path (the separate ground/neutral bonds in each disconnect) with everything grouped together, which should eliminate most of the risk. At some point you must always have parallel paths; the combined neutral/ground bar in a residential panel is _not_ a single point, and if you look closely, you will see parallel paths there...but for the purpose of the NEC it is considered a single point.

-Jon

Jon I know I presented the question like I did not have a clue but I do have some idea.

The fact is that these parallel paths on the utility side do cause problems and have been the subject of much litigation.

Try a www search using "stray current cow".

Or

"stray current pool"

Here is some info
http://www.mikeholt.com/documents/strayvoltage/pdf/ElectricalShockShower1-25-99-1-27-00Rev1.pdf

We may get to a point that the cost / benefit balance may change.

Bob

I have been catching up on the reading since my last post on this subject. I was wondering since there are people that disagree which is better, or is it preference, running 4 wires (one being the EGC) and using the grounding system from the original main panel or 3 wires and starting a separte grounding system at the new service panel?

I would say, if there is ever a chance that you will have a parallel path, go with the 4 wire.
If this is just a remote shed out in the yard that will never get connected in any way to the house, 3 is fine, perhaps even better since "ground" is a local thing.

edit to add;

If you are running 6ga and driving a rod you are actually making your GES better.

[This message has been edited by gfretwell (edited 12-27-2005).]

I have been told by a couple of different electricians that began their careers in Europe that the neutral is never bonded at the customer service unit / Service Disconnecting Means. I don't know if I understood them correctly; one spoke French as his first language and I speak only English; and I never got to ask if there is a bonding, grounding, earthing conductor run with the service conductors. Does any one here have accurate intelligence on that issue.
--
Tom Horne

"This alternating current stuff is just a fad. It is much too dangerous for general use." Thomas Alva Edison

That is fascinating. I hadn't even considered thinking of that. I don't have much to add to this thread other than "thanks" for making this much easier to understand than reading it in a book.
--Ron

okay so im still confused...? IS there any diagrams that show this .

Which this out of this involved thread are you interested in.
--
Tom Horne

We have many temporary metal frame (Tents) structures on the project. We are driving a ground rod bonding the structure and then going into the panel and connecting to the ground bar with a #6 GEC. This is in addition to the EGC run with the phase conductors.