Sub panel question

[concord]

Ok first post so be gentle,


The nuetral and ground need to be separate and isolated in a sub panel. I just can't see "why" this must be. Can someone slowly explain it to me without bagering me for not understanding;-)

JB

[roger]

Concord, could you please explain what type of inspector you are first. This is not meant to bager or berate you, but would help us understand where you are coming from.

Roger

[bennie]

Concord: Don't worry, I don't understand either, on a residental system with NM cable feeder.

[iwire]

I thought the reason was elevated voltage on the grounding conductors due to voltage drop from the load placed on the combination Grounding, Grounded Conductor?

Bob

[bphgravity]

It is simply the fact that objectionable current should not ever be on the grounding system considering it is bonded to every metallic and conductive component of the electrical system.

It is true that some installations would be of minor concern, but a metal raceway in metal building to metallic enclosures would be a serious issue. :)

[charlie b]

NOTE TO FORUM MEMBERS: I’m going to use conversational terms here, and make statements that are not precise. But it’s for illustration purposes only, so please don’t feel the need to correct me (unless I am way way off base). Here is an explanation of why the neutral and grounding conductors are connected to each other at the main panel only. The fundamental principal of interest here is that current takes all available paths (not just the one of least resistance) back to the source of power (not to planet Earth).

The “Normal” path of current is as follows: From the main panel (i.e., the source), via the “hot leg” of the feeder, to the sub-panel’s bus bars, out the breaker, along the “hot leg” of the branch circuit, through the load, along the “cold leg” of the branch circuit, to the neutral bus of the sub-panel, along the “cold leg” of the feeder, to the neutral bus of the main panel, thus completing the circuit.

Now let’s presume there is a short circuit between a hot wire inside the load to the external case of the load (or to any external metal part of the device). The “Fault” path of current is as follows: From the main panel (i.e., the source), via the “hot leg” of the feeder, to the sub-panel’s bus bars, out the breaker, along the “hot leg” of the branch circuit. Current will continue to flow back to the source as described above. But current will also follow two other paths. “New Path One” is from the case to the hand of the person who is touching the device, through that person’s body, out both feet, into the floor, along the dirt to the nearest ground rod, up the grounding electrode conductor to the main panel, into the ground bus, via the bonding jumper to the neutral bus, and thus returns to the source. Admittedly this is a high resistance path, but it does not take much current to ruin the person’s day. “New Path Two” is from the case to the ground wire (i.e., green wire or bare copper wire), along the ground wire of the branch circuit to the ground bus of the sub-panel, along the ground wire of the feeder to the ground bus of the main panel, via the bonding jumper to the neutral bus, and thus returns to the source. This is a very low resistance path, and the current that flows out the circuit breaker will be high enough to cause that breaker to trip, thus terminating the event and protecting the person who is touching the device. This is why there is a ground wire in the first place. But why do we separate the neutral and ground at the sub-panel?

The following would be the “Normal” path of current, if the neutral and ground were connected at the sub-panel: From the main panel (i.e., the source), via the “hot leg” of the feeder, to the sub-panel’s bus bars, out the breaker, along the “hot leg” of the branch circuit, through the load, along the “cold leg” of the branch circuit, to the neutral bus of the sub-panel. At this point, current separates into two or more paths. The first, as described in the “Normal” path, goes along the “cold leg” of the feeder, to the neutral bus of the main panel, thus completing the circuit. The other paths of current are from the neutral bus of the sub-panel, via the (illegal) bonding jumper to the ground bus of the sub-panel, back up the ground wire of every single device in the entire building, to the case (or any external metal parts) of every device in the building, into the hands of anyone who is touching any device in the building, through those persons’ bodies, out their feet, into the floor, along the dirt to the nearest ground rod, up the grounding electrode conductor to the main panel, into the ground bus, via the bonding jumper to the neutral bus, and thus returns to the source. See the comment above regarding ruining someone’s day.

[bennie]

Charlie: Very good. I am not referring to equipment ground wires with each branch circuit. I am addressing the number of feeder conductors, to a panel, from the service.

A three wire feeder will use the neutral for fault clearing, if the equipment ground wires are connected to the neutral bus in the panel.

I am only referring to a residential occupancy with NM cable as a feeder. I can understand when conductive paths are present, to separate the two grounds.

[roger willis]

Concord,
Welcome to the forum.
I won't say anymore, I think Charlie covered it all.

Roger

[ryan_618]

Charlie: Outstanding. :)

[brian john]

Very good Charlie, this would be an excellent 1st lesson for any course in electricity.

(not to planet Earth).
The number of electricians that thing the earth is a store house of used electrons scares me sometimes.

[el santo]

concord man i got a book here that puts it this way if the egc system were to touch the grounded conductor ( nuetral) system , which is carrying current all the time, there would be a parrallel path for this current to return to the service equiptment bonding point, such as the metallic raceway or the metal armor of an ac cable. this would create the possibility of a potential of voltage to exsist between all of the exposed dead metal parts of electrical equiptment and any other grounded metallic parts within that facility such as metallic pipe and appliances within the plumbing system. that is why u isolate the nuetral and the ground in a sub panel this would create a shock hazard if u touch the metal parts and it could melt the metal parts if the fault was strong enough

[eprice]

Charlie b,

Thanks for the explaination, but now I'm going to play the devil's advocate for a moment, because I think we need to dig deeper. :)

In your last paragraph, you describe what I'll call "illegal" paths because they utilize the illegal bonding jumper in the sub-panel through the ground conductors of the various branch circuits, through some unfortunate people etc, etc, until it gets back to the source. We can't know what the resistance of that portion of the path is after the sub-panel nuetral, but for discussion, let's say it is 25 ohms. I pick that number rather arbitrarily because it is a figure associated with ground rods :) the actual resistance may be less, but probably is more.

I suggest that in a legal installation without the illegal bonding jumper in the sub-panel there are some "legal" alternate paths that differ from the your "illegal" ones only slightly, as follows: from the neutral bus of the sub panel via the feeder neutral to the neutral bus of the main panel, then via legal main bonding jumper and the EGC of the feeder to the ground bus of the sub-panel and from there they are the same as the "illegal" paths you described.

Suppose the feeder is 60 ft. long, the resistance of the portion of the path through the feeder neutral and the feeder EGC would probably be about .01 ohms depending on the size of feeder and EGC. So, if the resistance of the portion of the "illegal" path from the sub-panel neutral to the source is 25 ohms, the resistance of the portion of the "legal" path from the sub-panel neutral to the source is 25.01 ohms. The current that would flow through the unfortunate person due to the illegal bonding jumper would be nearly the same as what could flow through them due to the "legal" paths that do exist in legal installations.

Am I wrong?

[hurk27]

Eprice
The only time the NEC allows the rebonding of the neutral-ground is when the subpanel is located in a remote building and there are no metal paths between the two buildings.
the reason this is allowed is because of the resistance of earth will allways be lower than the resistance of the neutral. so if there is a voltage drop across the neutral and the required GE is installed there will be no voltage potential between the grounding and earth. when there is a metal path then there is a possability of shock from this path and the grounding since this path will be at the potential of the main service ground-neutral and the voltage drop of the neutral feeding the sub panel will be between these two paths.

[bennie]

1940 Edition NEC, section 2523. No Additional Grounds.

No grounding connection shall be made to the grounded conductor on the load side of the service switch.

This probably originated around 1914, the words are the same today in the 2002 NEC.

This interpretation has also taken on a new spin from when it began. This has gone the same way as the separately derived system,and demonstrates my opposition to the use of "ed" and "ing"

The meaning is that the neutral/ground shall not be connected to earth after the main. Connecting to an equipment ground conductor is only grounding the neutral at both ends back to the service. This is not grounding (earthing) the neutral after the main.

[karl riley]

I think one reason it is difficult for some electricians to understand the effect of connecting neutral to grounding conductors in a subpanel is that they don't see anything different and they don't feel anything different. It becomes a theoretical point to discuss. Bennie evidently theorizes that the neutral current in a NM cable feeder will go obediently back to the main even when "grounded".

Since electricity is invisible we need an instrument to make it visible, to make it real. All you have to do is turn on a gaussmeter and watch the numbers, and then follow the current on all the various metallic pathways, HVAC ducts, wire mesh, pipes, steel girders, grounding conductors, sprinkler pipes, etc. You can't do most of this with a clamp-on ammeter, so if that's all you have, how would you know what's going on?

Even better, for a sense of the reality, is to go to Radio Shack and pick up a cheap telephone listening coil ($5?) and plug it into their inexpensive hobby amplifier and walk around listening to the huge hum as you get near a pipe or conductor carrying diverted neutral. I use earphones with the volume turned down. The sound can be so loud and annoying to the office personnel that I don't want to subject them to it. And the magnetic field on the feed, missing some of its neutral, is just as loud.

Everything Charlie described becomes a loud reality. And once you find and disconnect the N/G connection the noise disappears and peace returns.

If electricians would only carry a small $200 gaussmeter, or a $25 Radio Shack combo, they would know this is reality, not theory.

That's my sermon for today.

Karl

[bennie]

I stand corrected. I am referring to a residential occupancy, with wood framing.

I have a meter main on one side of my garage. The main panel is on the other side. I fed this with SER cable and separated the neutral and ground bus in the panel. There is no sense in doing this.

[karl riley]

Hi Bennie, you really keep a topic alive and keep the opinions coming until we get some kind of clarity. That's valuable.

Though steel buildings with conduit and various systems do give neutral current more paths, much of my work is done in residential wood frame houses. Even here, though, I have traced neutral all through the house. A panel in the house (actually a subpanel but with neutral bonded to the case) allows neutral to travel on the washer's equipment ground, to the copper water pipe supplying the washer, to a gas pipe which happens to be touching the water pipe, to a sheet metal vent pipe for the dryer, etc etc. All this comes together again at the bonding point at the meter outside the house. So wood frame doesn't mean you can't have neutral cavorting around the house in a way that results in areas that may have a high field where the resident doesn't want it.
Karl

[eprice]

Originally posted by hurk27:
Eprice
The only time the NEC allows the rebonding of the neutral-ground is when the subpanel is located in a remote building and there are no metal paths between the two buildings....
I am aware of this.

My previous post was a discussion of Charlie's example of the alternate current paths introduced by the illegal bond in the sub-panel. I was pointing out that in a legal system a path does exist from the feeder neutral through the main bonding jumper in the service panel, throught the feeder EGC, through the branch circuit EGC, through connected metal appliance cabinets, through human bodies, through the dirt, through the nearest grounding electrode, and back to the source. This path is nearly identical to the path created by an illegal bond in the sub-panel between neutral and grounding busses, and will differ in resistance only by .01 ohms or so because it must use the feeder EGC to get back to the sub-panel grounding bus.

I brought this up, because I think there needs to be another explaination for the code prohibition of the bond in the sub-panel.

[charlie b]

Eprice: Sorry I couldn’t get back to you earlier. It's been a busy couple of days.

The path you described will not carry current in the manner you describe. Let me return to my original statement about the fundamental principal at work here. Once current has left the source (i.e., via the hot leg), it seeks all available paths back to the source (i.e., normally via the neutral or cold leg). But once it has made it there, it does not go any further. I know it will continue to go around the same loop over and over; I don’t mean that. But if current goes from the sub-panel neutral bus to the main panel neutral bus, it’s next move is to the source. That's the end of the circuit. It will not proceed from the main panel neutral to the sub-panel (via the bonding jumper and the feeder EGC). The motive force for current flow (i.e., voltage) will drive current the other way (i.e., back to the source).

If you are taking a round trip flight from Salt Lake City to Chicago, when you arrive back in Salt Lake City you don’t hop a flight to Seattle. You’re home. For current, the source is home.

[brian john]

Karl:

With the advent of flexible CT"S readings are possible around most objects. One firm we work with have a 6-foot flexible CT great for measuring current on columns.

[brian john]

eprice:

In the "legal" system the only time there is current flow on the EGC (other that leakage current) is when there is a short/fault downstream from the grounded conductor/neutral. If the person or bodies you mentioned happen to inadvertently become this path the downstream voltage is typically too low to have a shocking impact on the inadvertent human conductor.

[bennie]

If you take an equipment ground wire and stick each end in the dirt, at different locations, you will get current flow, even with the service switch open.

[ July 11, 2003, 06:42 PM: Message edited by: bennie ]

[karl riley]

Brian,
yes, I have been happily using the flexible amp probe for years. First available only from a Scandanavian firm, now AEM sells a good one. I have 2 3' probes which can be connected in series and give a correct reading. Good for going around large bussways and snaking around a whole circuit of SE cables.

The gaussmeter will take care of everything you can't get the probes around, like steel beams in the wall, HVAC ducts, etc. It saves time because you don't have to snake anything around the pipe, etc. Just walk along beside it and watch the numbers. You can make rough estimates of amperage, but sooner or later you have to get an amp measurement to know exactly what is happening.

Karl

[brian john]

Karl:

I/we (the company) own several gauss meters, We utilize these for troubling shooting a variety of problems. Some are wiring errors and some are accidental neutral to ground faults. Both instruments make work easier.

There was a grounding seminar given recently in our area, one of the other techs went to this seminar and told me the following.

During the discussion on bonding at the main service an attendee inquired why the manufacture always seemed to ship a 1" to 1-1/4" green screw with every panel. Some attendees gasped others just looked quizzically. The presenter inquired what type of work this electrician did. The answer- Mostly residential, had been at it 20 plus years and was licensed.

[bennie]

Gentlemen: Analyze a service system with a separate ground wire from the transformer X-0 to the service panel ground bus.

Separate the neutral and ground bus. The neutral will be floating and earthed only at the pole.

Only fault and coupling current should flow on the ground wire. Load current will not.

The neutral will not be earthed after the main.

No need to earth the ground wire, except at the pole.

Check this out, I may be in the twilight zone again. :D

[karl riley]

Bennie, I am told this is how they do it in Germany and probably some other advanced European countries. Makes sense to me.
Karl

[karl riley]

Hey Bennie, good use of the word "earth" (using a noun as a verb, tsk tsk!) But it's a good usage. If we were to say "to earth" instead of "to ground" we could avoid confusion, because "to ground" can mean different things, but "to earth" has only one obvious meaning.

Then if we mean bond, we say bond. We could get rid of the word "ground" altogether.
Karl

[Ed MacLaren]

concord,
The answer to your question has been given by a couple of the guys. The sketch might help illustrate what they said.

To sum up, the illegal bonding jumper puts the (green) equipment grounding conductor (EGC) in parallel with the (grey) grounded neutral, and this permits some of the unbalanced load current to flow back to the transformer through the EGC which is not sized for such use.

Also, metallic conduits often serve as the EGC and can be exposed. A loose EMT coupling, for example, could have a considerable voltage drop across it.

An open neutral situation could result in the EGC carrying all of the unbalanced load current.

http://www.electric-ed.com/images/Service17.gif

Ed

[ July 13, 2003, 10:22 AM: Message edited by: Ed MacLaren ]

[bennie]

Karl: You caught me :o I used the "ed" incorrectly. I tacked it on to a noun.

I do agree with you, Don, and others, that using terms that are more technically correct, will take some confusion out of grounding. I am not sure the word "bond" is technically correct.

The word "fault" means failure. I would suggest "fault conductor" or something along that line.

[Ed MacLaren]

During my 35 years of teaching apprentices, I had often wondered why the "grounding/bonding" subject seemed to be so difficult for a majority of the students.

I became convinced that it was because we were trying to use the word "ground" for too many different purposes.

The word "ground" no longer has any meaning in a technical discussion, except to refer to the dirt itself, because any conductor that is associated with the word could be -
1. a system grounding conductor,
2. an equipment grounding conductor,
3. or the grounded (neutral) conductor.
Without specifying which one, it is impossible to communicate effectively.

Immediately after (about 15 years ago) we started using the term "bonding conductor" in place of "equipment grounding conductor", the level of confusion was greatly reduced.

Ed

[tom baker]

Exactly right Ed. The code requires a receptacle to be
406.3
To Be Grounded. Receptacles and cord connectors that have grounding contacts shall have those contacts effectively grounded.

The defintion of of effectively grounded is

Grounded, Effectively. Intentionally connected to earth through a ground connection or connections of sufficiently low impedance and having sufficient current-carrying capacity to prevent the buildup of voltages that may result in undue hazards to connected equipment or to persons.

So the code leads us to belive we connect the "ground" on a receptacle to earth, when what we are doing is making a low impedance path back to the source to clear a line to case fault or bonding.

No wonder we are confused.

[gwz2]

Let change Ed's diagram of July 13, 2003 10:16AM to:

At the left hand enclosure, remove the Switch and Fuse symbols and install a 1Ø 3W 120/240V 100A meter.
And remove the word "Main" from the Main Bonding Jumper text.

At the right hand enclosure remove the word "Illegal" from the Illegal Bonding Jumper text and name the panelboard as the Main Service panelboard.

In the 'Stick Man circle' add text that his hands are on each sice of a RMC (Rigid Metal Conduit) coupling.

Per 310.15(B)(6) the 100A panelboard could be supplied with 4 AWG conductors if a dwelling unit.

For a "Ball Park calculation using the DC ohms law and the following values:

120V at the meter location.
5 feet of 4 AWG for each of the conductors.
5 feet of 1" RMC
30 Amperes of Un-Balanced current on the circuit with;
R1 as the ungrounded conductor resistance to panelboard.
R2 as the resistance of the load to the grounded conductor.
R3 as the resistance of the Grounded conductor from the panelboard to the meter enclosure.
R4 as the resistance of 5 feet of the RMC.
R5 as the equivalent resistance of R3 and R4 in-parallel.

Chapter 9 Table 8, 1000' of 4 AWG is 0.308 ohms (R). [ R1 and R3 ].
Soares book on Grounding Table 3 shows 100' of 1" RMC to have 0.0154 ohms [ R4 ].

R1 = 0.308 ohms / 1000 feet = 0.000308 ohms times 5 feet = .00154 ohms = R1.

R3 = same as R1 above.

R4 = 0.0154 / 100 feet = 0.000154 ohms times 5 = 0.00077 ohms = R4.

R5 = [ 1/Rt = 1 / R3 + 1 / R4 ].
R5 = [ 1/Rt = 1 / 0.00154 + 1 / 0.00077 ]
R5 = [ 1/Rt = 649.35 + 1298.701 ]
R5 = [ 1/Rt = 1948.051 ]
R5 = 1 / 1948.051
R5 = 0.000513 ohms

E = I * R
R1vd = 30A * R1
R1vd = 30A * 0.00154 ohms = 0.0462 vd to the panelboard load.
R1vd = 0.0462.

R5vd = 30A * 0.000513 = 0.01539 vd from the load at the panelboard back to the meter enclosure.

120V ( at the meter ) - 0.0462 vd and the 0.01539 vd = 119.93841 vd at R2 ( the load).

Amps on R3 is: Evd / R3
Amps on R3 = 0.01539 / 0.00154 = 9.9935 amps.

Amps on R4 is: Evd / R4
Amps of R4 = 0.01539 / 0.00077 = 19.9870 amps.

If the Stick Man is holding a defective RMC coupling , He is subject to 19.9870 amps at 0.01539 volts.
R1 vd (voltage drop) = I * R1.

[don_resqcapt19]

gwz,
If the connection at the coupling is poor the current through the pipe will be limited by the resistance of the defective coupling. If the coupling is "open" the stick man will be sujected to a voltage equal to that of the grounded conductor voltage drop. In this example the voltage will be 0.00000256 volts and if you assume the stick man has a resistance of 500 ohms, then he will be sujected to 0.00000005 amps.
Don

[bennie]

Glenn: I agree with the figures. The stick man will only feel the potential difference between the two panels.

Remove the equipment ground wire, and with no other conductive paths, what difference does it make?

[gwz2]

Please note that the RMC, of this example, carries 2/3rd's of the Neutral current ( the un-balance current) between the meter and the Main Service panelboard.

When metal raceways are used between the meter and the Main disconnect of any typical service in witch the grounded circuit conductor goes thru the meter to the main diconnecting means.

Also, there will be a third parallel path ( thru the earth) with the grounded conductor if the Meter has a ground rod GEC and the metal water electrode GEC is from the panelboard.

Also, on every service with a Grounded conductor, there will be an 'earth' parallel path between the Main panelboard Grounding Electrode (GE)and the grounding of the utility companys' transformer grounding system.

Always, there will be parallel paths when there is a system with a grounded conductor circuit.

I think, more importantly, is the fact that the metal raceway will most likely carry more than 50% of the 'neutral current' between the meter and the Main disconnecting means.

[hurk27]

Glen the third path through Earth would only have the amoutn of current that the resistance drop of earth would allow. which would be the current applied from the voltage drop of the service neutral, if this voltage drop was a full 120 volts and the resistance of the electrod was 25 ohm's the current would be only 4.8 amps but since the actule voltage drop accross the service neutral would only be 2 or 3 volts the current on the gec would only be .12 of a amp of course the water ground would have more current on it but it to would be limited to the circuit resistance.

[websparky]

Gentlemen,

Very interesting and educational thread!

One question;

In the examples above with the stickman, as long as the other path for grounding exists, would he not receive zero voltage?

Thanks,
Dave

[gwz2]

I would think, that for the low voltage involved in the neutral parallel path of that 5 feet or so, that the current, even if very high, would not hurt most individuals because of the ' body ' resistance.

Really, some of the points I was trying to have comments on were:

1) The current flow on the RMC could reach welding status.

2) The current, hopefully, will not hurt a healthy person.

3) That an AC guru engineer would make some AC vs DC correction calculations - de-bunking the 'ball park' DC ohms law calculation. I do not know how!

4) On a 1Ø 120/240V system where one ungrounded conductor opens, it would be possible for the neutral current to approach the Amps of the service rating, ie. 100A, 200A, 400A, 600A, etc.
That short metal raceway could be carrying 50% or more of those AMPS!

5) On a Grounded Conductor 3Ø 3W system, It should act as the presented senerio.
On a 3Ø 4W system and losing either one or two phases will cause neutral current to 'run' high.
Seen many electricians always chose A & B for feeders rather than AB, BC, CA for feeders probably un-balancing the neutral current.

[ July 14, 2003, 07:17 AM: Message edited by: gwz2 ]

[don_resqcapt19]

Dave,
If our stickman is holding on to both sections of the conduit with an "open" connection at the coupling, he will be subjected to a voltage equal to that of the voltage drop on the other path(s). Under normal conditions this should not be enough voltage to cause any problems, but under fault conditions, until the OCPD clears the fault, this could be enough voltage to cause harm.
Don

[gwz2]

Don,

Quite a deference of my and yours Voltage Drop (vd)

across R4 when R5 is open. Mine would be 0.0462 +/- volts and yours is 0.00000256 volts.

Where did the calculations get mixed-up?

[gwz2]

For just 5' of 4 AWG (R1) from 120V meter to Load (R2) and 5' of 4 AWG (R3) from load back to meter, I get:

5' 4 AWG = 0.00154 ohms (R1)
5' 4 AWG = 0.00154 ohms (R3)

Voltage Drop (vd) across each of R1 and R3 is:

E = I * R

Evd = 30A * 0.00154 ohms = 0.0462 vd @ R1
Evd = 30A * 0.00154 ohms = 0.0462 vd @ R3
add the R1 and R3 together 0.0924 vd.

120V @ meter minus 0.0924 vd = 119.9076 vd @ R2

R = E / I

R = 119.9076 vd / 30A = 3.99692 ohms @ R2.

R1 @ .00154 + R2 @ 3.99692 + R3 @ 00154 = 4 ohms

E = I * R

120V = 30A * 4 Ohms

[don_resqcapt19]

Glenn,
I don't know what I did with my numbers. They are not correct. My point is that if the coupling is a high resistance or open connection the only voltage to drive current through the stickman is the voltage drop on the grounded conductor, so the amount of current will be very small. With the correct voltage drop of 0.0462 and a stickman resistance of 500 ohms, the current through him would only be 0.0462/500 or 0.0000924 amps.
don

[gwz2]

Would someone with the "AC savvy" make comments of the actual impedances etc. as opposed to the typical DC ohms Law of this 5 feet of 4 AWG and the RMC senerio ?

Glenn

[eprice]

Originally posted by charlie b:
The path you described will not carry current in the manner you describe. Let me return to my original statement about the fundamental principal at work here. Once current has left the source (i.e., via the hot leg), it seeks all available paths back to the source (i.e., normally via the neutral or cold leg). But once it has made it there, it does not go any further. I know it will continue to go around the same loop over and over; I don’t mean that. But if current goes from the sub-panel neutral bus to the main panel neutral bus, it’s next move is to the source. That's the end of the circuit. It will not proceed from the main panel neutral to the sub-panel (via the bonding jumper and the feeder EGC). The motive force for current flow (i.e., voltage) will drive current the other way (i.e., back to the source).
Thanks for the reply charlie.

I agree, the next move is back to the source which is the utility transformer. Yes, the service neutral is the best route back and most of the current will go that way, but some current will follow other paths. As has been mentioned before on this forum, if one of the grounding electrodes is a metal water line connected to the neighbor's metal water line for example, some of the current will follow the GEC, the metal water lines, the neighbor's GEC and the neighbor's service neutral. Another path would be through the GEC to any grounding electrode, then through the earth to the utility ground and back to the transformer that way. This would be a high resistance path but a few misguided electrons will take that path because it is there. And another path to the transformer is the one I am describing, through a feeder EGC, then a branch circuit EGC, through a metal appliance case, through an unfortunate person, through the earth, through the utility ground, and back to the transformer. The current would not follow the service ground back to the service panel in this case, because as you say that wouldn't get it any closer to home, but a small amount of current would use the path I describe through the utility ground because it is an available path from the service panel back to the utility transformer.

Now, the conclusion I am coming to is that this current I describe would be insignificantly small because of the resistance involved in that path, but I'm wondering if that wouldn't also be true of current in the path you describe in your original post through an unfortunate person if the path includes the earth.

I am thinking that the real danger created by the illegal bond in a sub-panel is the possibility it creates for a path back to the service panel as you describe via an unfortunate person that does not use the earth but some more conductive path such as a metal water line, or some other conductive element that is in contact with something that is bonded to the service panel.

[charlie b]

eprice:

I still think that you are visualizing paths that cannot carry current in the direction that you are describing. So let’s pin down a couple of basic principles, and see where we can go from there.

First, no current EVER flows in the GEC (i.e., the wire that connects the main panel’s ground bus to the grounding electrode – be that electrode a ground rod or a water pipe or whatever), and no current EVER flows through the grounding electrode itself, UNLESS some piece of equipment has a fault to ground (e.g., hot wire to case), or unless there is a lightning strike. Reason: There is not a complete path. The resistance between all current-carrying components (e.g., the transformers, panels, wires, and loads) and planet Earth is essentially infinite.

Secondly, if there is a fault within a component (e.g., hot wire to case), the current flowing in the fault path (i.e., source to panel to fault point to case to EGC to panel and back to source) should be high enough to trip the breaker and terminate the event.

Third, although it is true that current will take all available paths, it does not travel in all available directions. The direction of the push (i.e., from source hot to source neutral) is the only direction that the current can take. No matter what fault conditions exist, no matter if an unfortunate person is touching the wrong component at the wrong time, current cannot flow from the main panel, via the feeder EGC, in a direction that takes it to the branch circuit EGC. The voltage source will push any and all current within the EGC in the opposite direction – towards the main panel. This is not a matter of “how much resistance does one path have, in comparison to that of another path.” It is a matter of current goes in the direction in which it is pushed.

As to the potential danger of a fault path that includes planet Earth, there are (sad to say) a vast number of documented cases in which a person holding a device receives a shock, with the only possible path through their body necessarily including the dirt beneath their feet, and with fatal consequences. It only takes 0.1 amps to cause a fatality. On a 110 volt system, that means the total path resistance would have to be less than 1100 ohms. The resistance of the body itself can easily be under 300 ohms, and the resistance of the ground rod to earth should be under 25 ohms. Unless the person is wearing rubber boots and is standing on dry desert sands, the resistance of the dirt can easily be under 700 ohms. As I say, it has happened many times.

[bennie]

Charlie B: Please explain why you state the ground electrode conductor never carries any current, except under fault events?

[gwz2]

Is not the GEC and the grounding electrode a parallel path of the Grounded System Conductor of any Utilty supplied Service ie. Utilty transformer to the Service Equipment grounding system ?

I have withnessed the Water GEC path carry approximately 2/3rd's of what the grounded ( neutral ) conductor should be carrying.

Still no comments of my other question of the difference of an AC impedance law and a DC ohms law.
for the described 5 foot of 4 AWG , 30 Amps, 120 volt system.

[charlie b]

Originally posted by bennie: Charlie B: Please explain why you state the ground electrode conductor never carries any current, except under fault events? I mean that (1) Current must have a complete path from the source back to the source, and that (2) Under “normal conditions," the GEC is not part of a complete path.

If, throughout a distribution system, there is no “connection” (meaning either a direct short circuit or a high-resistance leakage path, possibly from degrading insulation) from a “current-carrying conductor” (meaning both the ungrounded and the grounded) to any “external metal parts that are not intended to carry current,” then there would be no complete path from the source back to the source, with the GEC being part of the path.

But let me now repeat something from my initial post on this thread: “I’m going to use conversational terms here, and make statements that are not precise.” So starting now, I’ll be a bit more precise.

(1) Any closed loop of conductive materials will experience an induced current, if the loop is in the presence of a AC magnetic field. Put your watch on the desk next to your computer, and current will flow around the band. In an electrical distribution system, there are many such loops. They consist of the ground rod, the GEC, the EGCs, water pipes, and other metal objects, with dirt being involved in some of the loops. The magnetic fields that surround the transformer and every current-carrying conductor will induce currents in each loop, with the result that some measurable currents may be present in the ground rod.

(2) Also, any two metal objects will have between them a capacitance, any single metal object will have a self-inductance, and any two metal objects will have between them a mutual inductance. In an electrical distribution system, the amount of capacitive and inductive reactances that exist among the various cables, conduits, and other metal objects can be high enough for there to be measurable currents present in the ground rod.

So I confess: I lied. But I told you at the beginning that I would.

[Ed MacLaren]

I mean that (1) Current must have a complete path from the source back to the source, and that (2) Under “normal conditions," the GEC is not part of a complete path.
I know it isn't "normal", but in an area that still has a metallic water main system, this scenario could easily exist for some time before being detected.

http://www.electric-ed.com/images/Neutral2.gif

Ed

[karl riley]

Ed, that's a great graphic. Were you aware that EPRI built such a dummy subdivision in MA to test how currents circulate on water pipes and neutrals?
Karl

[concord]

WOW!,

Ive been out of town. Thank you all so much for the responses!!! By the way I'm a residential home inspector.

Thanks again,
John

[sno]

Wow!
Thanks Karl...

That cleared everything up for me.