Safely disconnecting a utility ground fault vs. NEC

Phil, I'm not sure I understand the problem that concerns you. On a utility service with a grounded neutral, the neutral is grounded once at the utility transformer and again at the building service entrance. Are you concerned that, simultaneously, (a) the neutral-ground connections at the utility transformer and at the building service become disconnected, (b) the utility neutral is energized, and (c) a person contacts the utility neutral? If so, do you have personal knowledge of this happening?

Regards, Chris Johnston

You want a 3 conductor transfer switch. That will allow you to connect another independant power source to the other leg or just disconnect completely. That is the simplest, code compliant, way to cut the cord.

| Phil, I'm not sure I understand the problem that concerns you. On a | utility service with a grounded neutral, the neutral is grounded once | at the utility transformer and again at the building service entrance. | Are you concerned that, simultaneously, (a) the neutral-ground | connections at the utility transformer and at the building service | become disconnected, (b) the utility neutral is energized, and (c) a | person contacts the utility neutral? If so, do you have personal | knowledge of this happening?

Whan you say "at the building service entrance" you aren't specific enough to specify whether it is before, or after, the point where the main disconnect takes place. But I can understand the ambiguity since in most cases it is moot due to the neutral not being switched. But if the neutral is switched with the other conductors, then it matters. If the neutral is connected to the ground before the switch, then the switch will not protect against voltage coming across from the neutral to the ground wire.

Scenario 1: A broken neutral between the point where the transformer is grounded, and the point where the service splits between 2 service drops. Unbalance current in the neutral wire at another service drop will be divided among the various ground connections the service drop side of the neutral trying to find a way back to the transformer, since the proper path is broken. I don't remember which forum, but there was a case where someone actually experienced this. They had 5 amps of current across the bond between the neutral and ground, even with the main switched off (which did not switch the neutral). The power company eventually fixed "something" and it went away.

Scenario 2: (this might be what you described) The pole transformer becomes either ungrounded or has an internal fault. Primary voltage can be introduced on the secondary, either by capacitive coupling, or direct fault. I saw a video of a case where the fault actually happened and a lineman was hit. He survived (stunned by was able to walk away) because he was wearing gloves, but was injured because they were low voltage gloves (which otherwise make sense since he was attempting to cut the low voltage service drop).

My objective is to be able to disconnect entirely from the service drop, either by manual action, or by ground fault detection. That means all conductors, with a grounding wire system that is not connected around that switch ti the service drop. Connecting neutral to ground _after_ the main disconnect accomplishes that, but it does not comply with NEC.

| You want a 3 conductor transfer switch. That will allow you to connect another | independant power source to the other leg or just disconnect completely. | That is the simplest, code compliant, way to cut the cord.

It may be a transfer switch, or it may be just a plain switch or breaker. But definitely a 3 pole unit would be needed to disconnect the neutral with the other conductors.

But where do you attach the neutral to ground?

If before the switch, then the ground is still connected to the utilty even though the switch opens or the breaker trips.

If after the switch, then it runs into NEC 250.24(A)(1) and (5).

Your objective can be achieved by installing a service disconnecting means that opens the neutral. If the switch opens all conductors of the service simultaneously then that is permissible under the US NEC. The code permits the main disconnecting means to be adjacent to the service Over Current Protective Device (OCPD). They need not be the same device. So you could install a three pole switch ahead of a two pole main breaker or fuse pullout that would open all three conductors of the incoming three wire service. If the service is three phase then a four pole switch can be used ahead of a three pole breaker or fused pull out. The bonding screw or strap can be tied-in on the load side of the switch if it is located in the same enclosure. For protection from lightning induced surges you would install a secondary surge protector on the line side of the switch.

-- Tom H

| Your objective can be achieved by installing a service disconnecting | means that opens the neutral. If the switch opens all conductors of the | service simultaneously then that is permissible under the US NEC. The | code permits the main disconnecting means to be adjacent to the service | Over Current Protective Device (OCPD). They need not be the same | device. So you could install a three pole switch ahead of a two pole | main breaker or fuse pullout that would open all three conductors of the | incoming three wire service. If the service is three phase then a four | pole switch can be used ahead of a three pole breaker or fused pull out.

This much was part of my plan.

| The bonding screw or strap can be tied-in on the load side of the | switch if it is located in the same enclosure. For protection from | lightning induced surges you would install a secondary surge protector | on the line side of the switch.

Can you cite NEC code that overrides 250.24(A)(1) and 250.24(A)(5) with respect to bonding on the load side of the switch?

What would the secondary surge protecter be connected to for ground?

Would that surge protecter deal with primary voltage?

| You want a 3 conductor transfer switch. That will allow you to connect another | independant power source to the other leg or just disconnect completely. | That is the simplest, code compliant, way to cut the cord.

But what about the grounding? Where will you attach the ground in a code compliant way while not providing a path between the ground inside and the neutral outside?

You can't really do it. If you drive a separate ground rod, it will still provide a path to the ground rod on the other side of the switch. The earth is a decent condutor.

I think you are worrying to much about a non-issue.

Charles Perry P.E.

250.24(A)(1), the grounded wire is to be connected to the grounding wire between the serice drop/lateral, and the terminals of the disconnecting means. Then 250.24(A)(5) prohibits connecting at any points on the load side of the disconnecting means.

The problem I have with this, is that by connecting the grounded conductor before the disconnection, it really isn't a 100% disconnection since the grounding system (e.g. the grounding wire, not the neutral wire) in the building is still attached to the utility neutral wire. Faults or breaks in the utility wiring, or even in neighbor wiring sharing the same transformer, could introduce voltage relative to ground on the neutral wire. While the grounding connection would make for a path for that voltage to go to ground, no ground is ever perfect, so if there is ever any other point in the grounding system (e.g. any equipment which has a chassis attached to the grounding wire) which might get grounded to earth, it also forms a path where current could flow. If that path includes a human being, then what we have is two paths in parallel, one near the service entrance (the normal grounding point), and one including a human. Since ground is a resistive component (relative to ground return paths) this will mean some part of the return current flows through the human. Since the grounding system has become part of the circuit under a supply fault, it is not doing its job in protecting people.

A ground fault detector at the main disconnect won't detect this since it is current flowing on the grounding wire itself once it has reached that wire from the connection with the neutral. Even if a disconnect breaker included disconnecting the neutral (allowed by NEC only if the neutral is disconnected with or after the other conductors), it won't stop this hazardous condition.

What I propose is that where a main service disconnect also disconnects the neutral conductors (simultaneously with all other conductors), that the neutral be connected to the grounding wire on the load side of the disconnect. This way, when the conductors are disconnected, there is no possible path for a supply/line side fault to energize the grounding wires in the building. This will also allow a ground fault current detection to work properly since it will be detecting all current coming in that does not have an equivalent return current. Such a detector would then be able to interrupt such faults via a shunt trip or such on the main disconnect.

I generally find when the NEC has a rule, there's a good reason for it. But I cannot figure out the reason for having the grounding point of the grounded conductor (neutral) being on the line side of the main breaker, as opposed to the load side. Obviously, in cases (and this is most of them) where the neutral is not switched at all, it's moot, electrically. But it is allowed to disconnect the neutral when the other conductors are as well, and in such a case, it now matters which side things are on. But I just do not see the reason why the NEC has it this way, and as far as I can see, it is safer as I proposed.

Reference from the NEC 2005 draft:

• 250.24 Grounding Service-Supplied Alternating-Current Systems.

*

• (A) System Grounding Connections. A premises wiring system supplied by a

• grounded ac service shall have a grounding electrode conductor connected

• to the grounded service conductor, at each service, in accordance with

• 250.24(A)(1) through (A)(5).

*

• (1) General. The connection shall be made at any accessible point from

• the load end of the service drop or service lateral to and including the

• terminal or bus to which the grounded service conductor is connected at

• the service disconnecting means.

[snip]

• (5) Load-Side Grounding Connections. A grounding connection shall not be

• made to any grounded conductor on the load side of the service disconnecting

• means except as otherwise permitted in this article. [ROP 5 65]

[snip]

--

----------------------------------------------------------------------------- | Phil Howard KA9WGN |

| | (first name) at ipal.net | |

-----------------------------------------------------------------------------

A solution would be to use a 4-pole main circuit breaker at the service, with the 4th pole breaking the neutral. The service neutral would be grounded on the line side of the main breaker, thus complying with the NEC. When the breaker opens,the load side neutral is disconnected from the line side neutral and grounding connection.

The part that puzzles me is how the breaker is to be automatically tripped under the two scenarios you mention. In scenario 1, no one with an interest in continuous service is going to set a GF relay at 5 amps. In scenario 2 you could use an overvoltage relay to trip the breaker.

A personal comment, please. I sometimes wonder whether the folks who write the NEC ever consider situations such as you describe. I know that they continue to give the utility companies a pass on using multi-grounded neutrals on ballfield and parking lot lighting. Every so often you read of someone being electrocuted by this situation. I find it disturbing that the utility companies maintain a prominent presence on some of the NEC committees, yet maintain that, in most circumstances, they are not subject to the NEC.

Cheers, Chris Johnston

- snipped-for-privacy@ipal.net wrote:

| A solution would be to use a 4-pole main circuit breaker at the | service, with the 4th pole breaking the neutral. The service neutral | would be grounded on the line side of the main breaker, thus complying | with the NEC. When the breaker opens,the load side neutral is | disconnected from the line side neutral and grounding connection.

Now to find a 4-pole breaker.

| The part that puzzles me is how the breaker is to be automatically | tripped under the two scenarios you mention. In scenario 1, no one | with an interest in continuous service is going to set a GF relay at 5 | amps. In scenario 2 you could use an overvoltage relay to trip the | breaker.

If there is current coming in on the supply neutral seeking ground, that tells me there is something wrong. I'd want my GF detection set way down there for these things, though I'm not sure where. I don't want to have to switch neutrals on all the branch circuits, yet I do want to open the circuit on 5 milliamps. The traditional ground fault which can be interrupted by opening hot wires should catch that in the branch circuit where GFCI is installed. But coming in on the neutral remains the issue. And if, for now, that means opening the main breaker to clear it, so be it. If that happens too much I'll address that when it happens.

I have more of an interest in safety over continuous operation for my home. In a data center I might install, it could be different; it might have a high impedance ground on multiple separately derived systems. But I'm the networking person on those; I just work with electricians sometimes (like providing them with load needs for the cabinets).

| A personal comment, please. I sometimes wonder whether the folks who | write the NEC ever consider situations such as you describe. I know | that they continue to give the utility companies a pass on using | multi-grounded neutrals on ballfield and parking lot lighting. Every | so often you read of someone being electrocuted by this situation. I | find it disturbing that the utility companies maintain a prominent | presence on some of the NEC committees, yet maintain that, in most | circumstances, they are not subject to the NEC.

I would wonder why they want multiple grounded neutrals.

|> | You want a 3 conductor transfer switch. That will allow you to connect | another |> | independant power source to the other leg or just disconnect completely. |> | That is the simplest, code compliant, way to cut the cord. |>

|> But what about the grounding? Where will you attach the ground in a code |> compliant way while not providing a path between the ground inside and the |> neutral outside? |>

| You can't really do it. If you drive a separate ground rod, it will still | provide a path to the ground rod on the other side of the switch. The earth | is a decent condutor.

Setting NEC aside for the moment, why do I need to have a ground rod on the line side of the main disconnect if it is disconnecting all current carrying conductors (including neutral)? Why can't I have the grounding attachment only on the load side, and not at all on the line side, thus leaving the service drop/lateral open ended and not connected to anything when the disconnect is opened? If there is any hazard from a wire there not connected to ground, why would that hazard be of concern for the neutral wire but not the hot wires (which obviously would not be grounded)?

Now back to NEC; it isn't allowed. But what is the technical reason?

I read in sci.engr.electrical.compliance that snipped-for-privacy@ipal.net wrote (in ) about 'Safely disconnecting a utility ground fault vs. NEC', on Fri, 16 Jul 2004:

Draw a schematic including the cables as resistors. With long cable runs, there may be too much loop resistance for a line to ground short to cause enough fault current to clear the protective device (not a GFCI but an ordinary fuse or thermal-magnetic circuit breaker) in a short enough time. Grounding the neutral at several places reduces the loop resistance.

I'm sorry for typing too fast and not proof reading carefully enough. For this purpose the Service Disconnecting means continues to be the main breaker or fused pull out. The three or four pole switch is not the service disconnecting means it is an isolating switch. Your not overriding anything. The bonding point is in the service equipment enclosure. It is one of the points "from the load end of the service drop or service lateral to and including the terminal or bus to which the grounded service conductor is connected at the service disconnecting means." Isolating switches are used to isolate service equipment from its source of supply. This is often done with service rated transfer switches so that the transfer equipment can be serviced without utility company intervention.

-- Tom H

|> | The bonding screw or strap can be tied-in on the load side of the |> | switch if it is located in the same enclosure. For protection from |> | lightning induced surges you would install a secondary surge protector |> | on the line side of the switch. |> |> Can you cite NEC code that overrides 250.24(A)(1) and 250.24(A)(5) with |> respect to bonding on the load side of the switch? |> |> What would the secondary surge protecter be connected to for ground? |> |> Would that surge protecter deal with primary voltage? | | I'm sorry for typing too fast and not proof reading carefully enough. | For this purpose the Service Disconnecting means continues to be the | main breaker or fused pull out. The three or four pole switch is not | the service disconnecting means it is an isolating switch. Your not | overriding anything. The bonding point is in the service equipment | enclosure. It is one of the points "from the load end of the service | drop or service lateral to and including the terminal or bus to which | the grounded service conductor is connected at the service disconnecting | means." Isolating switches are used to isolate service equipment from | its source of supply. This is often done with service rated transfer | switches so that the transfer equipment can be serviced without utility | company intervention.

If the isolating switch is not considered to be the service disconnecting means, then it should be able to be grounded _between_ the real service disconnecting means and that isolating switch, right? If that is true, then the isolating switch would only need to be 3 pole for 3 wire single phase (4 pole for 4 wire three phase for comparison), right? If right so far, then why not let the isolating switch actually be a circuit breaker itself (which adds automatic overload and fault interruption capability)? If that can be done, what we have is two main breakers in sequence, with the neutral grounded between them. Now the 2nd breaker is redundant, it would seem, and pointless (aside from a code issue) as all it does more is disconnect the neutral of the unenergized load side from ground.

I read in sci.engr.electrical.compliance that snipped-for-privacy@ipal.net wrote (in ) about 'Safely disconnecting a utility ground fault vs. NEC', on Wed, 14 Jul 2004:

I think there has been a lot of misunderstanding and diversion in this thread, so I'm going back to the beginning to try to clarify Phil's question. A simple ASCII schematic would have helped, I think, also fewer words. I think the above sentence is clear enough. [snip]

Yes, if you don't change something..

Yes, this is a hazard, but of low probability, I would say. Still, it only need to happen once...

I think some people didn't understand this point, though. If a high- current fault, or an open neutral, occurs outside the allegedly-isolated premises, a voltage will appear on the utility's local grounded conductor, and thus on the grounded conductor inside the premises, relative to other grounds more distant from the fault. IEC 60364 calls for equipotential bonding to minimise this possibility. But introducing a remote ground into a bonded environment is still hazardous.

You can prevent this, as far as I can see, only by disconnecting the grounded conductor of the premises from the utility's ground. It may well be against Code to do this with a switch, but that would presumably not rule out physically disconnecting a wire. I'd recommend temporarily grounding the ground conductor network to a ground rod distant from the utility's ground in that case.

What you need to do, of course, is not to close the three-pole switch (L1, N, L2) with the ground still disconnected. I doubt that there is a simple way to ensure this automatically, so some form of 'lock-out' safeguard need to be used.

Isolating switches are often installed with a rating that would not be adequate as a service disconnecting means. Switches used as isolating switches are often not listed as suitable for use as service equipment. The reason that circuit breakers are seldom used as isolating switches is that since they open automatically they would have to be rated with a high enough withstand to open the circuit under load. That is a far more demanding function then just breaking an idle but energized circuit.

-- Tom H

| Isolating switches are often installed with a rating that would not be | adequate as a service disconnecting means. Switches used as isolating | switches are often not listed as suitable for use as service equipment. | The reason that circuit breakers are seldom used as isolating switches | is that since they open automatically they would have to be rated with a | high enough withstand to open the circuit under load. That is a far | more demanding function then just breaking an idle but energized circuit.

I guess you meant that the breaker is rated to open the circuit under fault? The switch would have to be able to open under at least some portion of the load, and have the stay-closed rating at full load.

But if a circuit breaker were used anyway ... what then?

So if that is correct, you cannot (per NEC) disconnect the neutral from the premises electrode grounding system under any circumstances. Renders all the other words/issues moot.