Grounding of residential subpanel

[pbunge]

An existing subpanel, probably over 40 yrs old, has a local ground #8 ground wire bonded to a copper water pipe and the neutral and grounds share the same buss bar. No ground wire in the #2-3 romex back to the service, about 100' away. Service has one 70 amp breaker dedicated to this condo. I believe the bank of about 10 service panels for the condo complex is grounded. No easy or esthetic way to run a ground to the service.

When replacing the subpanel is it permissible to install a ground rod under the subpanel, and bond a #6 copper wire to the ground rod and to a nearby 3/4" copper cold water pipe that enters the ground? Access openings to bonding clamps would be provided and the neutral buss would be isolated from the ground buss bar.

I have not been able to find code to support this and the bulding official says this is not acceptable but cannot give a reason. He says grounding is for lighting, a moot point for this area in So. California. I thought grounding was for safety, a path for current to flow should a short to a device occur, thus preventing the user of a device from being shocked.

Thank you. Should you want to reach me by phone:
(949) 433-8891.
Peter Bunge

[karl riley]

Peter, You have a dangerous situation here. At the subpanel, the path for a ground fault as you describe it would be to the copper water pipe to the assumed water pipe clamp at the service grounding point and then back to the transformer through the service neutral. This relies on the electrical integrity of the water pipe, which is not under the electrician's control.

You are under the impression that connecting to earth is for electrical safety (aside from lightning). This is a myth. Just think of how much current will flow from 120V through a grounding electrode of a nominal 25 ohms. 120/25 = 4.8A. Will 4.8 amps trip a breaker? Where is the safety?

I would suggest you go into the mikeholt website and order a book and/or video on grounding and bonding. It will clear up any misconceptions. (I don't work for Mike Holt!)

The neutral bus in the subpanel must be insulated and all equipment grounds must have their own bus. There must be a grounding conductor with the feed conductors (it could be a conduit) and it must run with the phase conductors, not in some other more convenient path.

Now, aside from basic safety, the situation you have now sets up a huge magnetic field loop from the normal use of power to the subpanel. The neutral return current for the subpanel is now splitting and returning on both the neutral conductor in the feed and the copper water pipe. Both paths will have a large magnetic field, depending on how much current. If there have been any malfunctions of computers, etc., you would be solving that situation as you solve the safety Code violation.

Karl Riley

I think you will haver some other comments by the end of the day.

[karl riley]

A quick correction of my post. As it is now, a fault current will return on both the water pipe and the feed neutral. The breaker will trip, but the parallel path of the current is not allowed.
Karl Riley

[dereckbc]

Peter, I agree with Karl, you have code violations. I am not real clear about the setup. Is the sub-panel in a detached building? It does not sound like it, but there a couple of options if it is.

I will assume it is not attached. Sounds like your neutral is multi-grounded at the sub-panel. That would cause load current to flow on the water pipe and violate 250.6 and 250.24(A)(5). From the main panel you will need to run phase, neutral, and an EGC to the sub-panel. You can bond the ground again to an electrode, but not the neutral.

If it is detached let us know, 250.32 offers a couple of options.

[ February 24, 2003, 03:12 PM: Message edited by: dereckbc ]

[jxofaltrds]

Peter

As stated above this is dangerous.

You may NOT drive a ground rod. You are in the same structure.

I see multiple problems here.

Stop do not do anything.

I am sure that the waterlines are "interconnected" between the condos.

You might have found a can of worms! :(

Please give a detailed explaination of what water, gas, electric is present.

Mike P.

[jxofaltrds]

I do not work for MH or KR.

Buy the books that these two "PROFESSIONALS" have. They will help you understand the dangers that others have made by their installations.

There are many "PROFESSIONALS" here one this site, I do not profess to be one of those. They will give you great advise. Whether by a "PRO" or a want-to-be like me, when someone says danger, please research this further.

Mike P. IMHO

[hurk27]

fault current will return on both the water pipe and the feed neutral From What I'm getting is this subpanel is fed with a 3 conductor H-H-N and the neutral is bonded at the subpanel to the case and the neutral bar. and the egc's are also connected to this bar. now you have a #8 ran to a water pipe from this subpanel as a water ground.It might not meet todays codes but this sounds like that it was code compliant for the "1960"'s yes you will have parallel fault currents on both the water pipes and the feeder neutral because of multible connections in the building. but where in the nec of the 60's is this not allowed? as we have neutral currents on our water pipes that have a commoned connection with a house down the road. I would install a ground rod at the main service bank and size the gec to the service. the problem is that to bring
this up to todays code would take running new SE 2/2/2/4 cable from the subpanel to the main service. and seperateing the grounds and the neutrals at the subpanel. and running the water ground back to the main service panel and bonding it there. and in no way do you ever depend on just the earth to take a fault to trip a breaker as the fault has to return back the path it came back to the neutral at the main service and earth will not do this. as was said before.

[karl riley]

Hurk27: Are you assuming that the prohibition of connecting neutral to ground on the load side of the main disconnhect was allowed in the 60s?
I doubt it, but I don't know the history of the Code on this point. Perhaps someone does know when this came in.

[pbunge]

Interesting how these condos, which are connected and share water pipes have been around for 40+ years. Of course that doesn't mean there isn't a risk involved in the existing state, but I wonder if alarm is warranted. Do the existing panels need to be changed for safety concerns?

I want to emphisize that the new panel neutral buss bar would be insulated from the panel thus there would not be a parallel flow of current back to the service unless it was a fault current. I do realize my mistake of assuming nill resistance posed by the ground rod/water pipe. I see the point of having the ground bond back to the neutral at the service in order to return current to the transformer and minimize resistance to fault current. I do not see why the path of the ground between the panel and service would make any difference, remembering that the only current flowing in the ground would be fault current. Thanks for the suggestion of the book on bonding and grounding, I'll order it.

[dereckbc]

Paul, if your work is going to be inspected, it is going to be tagged if you do not run an EGC with the feeder 215.6. To keep the impedance at a minimum, it is necessary to run the EGC within the same raceway or cable as the phase and neutral conductors. This allows the magnetic field developed by the circuit conductors to cancel, reducing thier impedance. The water pipe is not part of the path and may not have the low impedance neseccary to operate the breaakers.

Good Luck

Dereck

[pbunge]

Thanks, Derick
I was not aware of the effect of the magnetic field to reduce impedance. Is the affect of a magnetic field necessary to reduce the impedence for say 20+ amps fault current flowing through a #4 wire to to trip the breaker? I guess the concern is how much voltage is necessary to drive the current in order to trip the breaker. The higher the voltage the less safe the situation. A person grounding a faulty device would be subjected to whatever voltage is needed to drive the 20+ amps of fault current through the bonding wire to ground. Still seems that a #4 wire about 100' long would have very low impedance for 20 amps. Guess I need to do some math.
Peter

[dereckbc]

Peter; sorry for getting your name wrong, please accept my apology. 20-amps of fault current is not going to trip a 20-amp breaker in any reasonable amount of time if ever at all. It would appear as load.

A person making the fault would not likely exhibit a low enough impedance to trip a 20-amp breaker. That is why we have GFCI breakers in certain locations where it is likely. By design it takes at least 6 times the rated device current to make the device operate instantaneously, for a 20-amp breaker that would be 120 amps. In order to drive 120 amps by ohm’s law the impedance at 120 volts/120 amps is 1 ohm. The 1-ohm impedance is not the impedance of the EGC, but the impedance of the whole circuit, from the main panel where the N-G bond is located, through the breaker(s), phase conductors, contact resistances, and the EGC in series. So the EGC impedance would have to be less than .5 ohms.

Not sure where the # 4 AWG comes from, I assume you are talking about the feeder EGC with a 70-amp breaker. That is a different story. The # 4 AWG would protect the feeder circuit, and the down stream branch circuits. For a 70-amp feeder you are required to have a # 8 minimum EGC, so a 4 AWG is permitted. 20-amp breakers require a minimum 12 AWG EGC.

I am not trying to say what you have would not work, but it does not meet code, and could be dangerous. If there were property damage or someone was hurt, guess who is liable.

[pbunge]

Dereck, please accept my apology for mispelling your name! I appreciate your time to respond. What I'm needing is an answer to meet code and safety requirements preferably without replacing the existing H-H-N romex from the service with a H-H-N-G. Instead, adding a seperate #8 or larger ground wire in conduit to bond the subpanel with the service neutral and ground buss.

I am concerned about the issue Karl raised when he stated that the ground must run with the conductors because of the effect of the magnetic field. Would running a larger ground wire compensate for this effect of being routed independently?

[don_resqcapt19]

Peter,
There is no exception to 300.3(B) for this application. The required equipment grounding conductor must be in the same raceway or cable as the feeder circuit.
Don

[dereckbc]

Peter, I have been looking all morning trying to find an exception. The only thing I can find is for "nongrounding receptacle replacement" in 250.130(C). It does not apply to a feeder. 300.3(B) and 215.6 have got you between a rock and hard place. I concur with Don.

Good Luck

[pbunge]

I guess I should say thank you, Dereck and Don, again I appreciate your digging, yet it's like being between a hunk of concrete and the building (rock & hard place). How to do this without tearing up the building or concrete, not to mention dealing with the association. I really would like to know WHY the ground must be alongside the feeder cables. Is it primarily for organizing to avoid confusion about what wire is with what feed, or as Karl stated, has to do with the interacting magnetic fields?
Peter

[dereckbc]

It is to cancel out the magnetic fields, lower impedance, and reduce heating effects.

[karl riley]

Pb: it's also required by Code. No way around that. Karl

[rb]

Pbunge,

I can give you a very elementary reason (its the only one I know). When a conductor returning the fault current is located in close proximity to a conductor sourcing the fault current, the magnetic flux created in the two conductors cancel each other out. When these conductors are not located close to each other the magnetic flux created during the fault causes inductive reactance in the circuit and increases the impedence of the fault return path. Once this occurs the fault current may seek alternative paths back to the source. Paths that are not designed to conduct electricity.

Im sure there are many fellows on this site that can explain it a lot better than I can.

Ron Bethea

[pbunge]

Thanks Ron,
I appreciate the reasoning as it helps me to understand and accept the code better. I feel much more inspired to follow laws if I can understand the reasoning behind them. I did order Mike's book so I hope to be able to be less dependent on you guys, thanks for your input.

By the dialogue above it seems this is a very hazardous situation. Does this mean that the City should require all the condo's to have their grounding upgraded? I doubt that would go over well with the homeowners unless they could be convinced of impending danger; remember these have been around for 40+ years without burning down or killing people.
Peter

[don_resqcapt19]

pb,
I think that, if I were the AHJ, I'd permit a remote EGC for a feeder with a 70 amp OCPD. I don't think that the additional impedance caused by the fault clearing path being remote from the feeder will have any great effect on the fault clearing time of this circuit. If the circuit OCPD was larger, over 200 or 300 amps, I would not approve the remote EGC.
As far as the hazard level, of this installation there is a potential for harm. This is not much different from the 3 wire circuits that were permitted for many years on stove and dryer branch circuits. In these circuits a single conductor served as both the grounded and grounding conductor. There were proposals made over 3 or 4 code cycles before the rule was changed to require 4 wire circuits. All of the proposals cited the "potential" problems that could occur when using a single conductor for both functions. The CMP asked for any evidence of a problem that had really occurred. I don't recall anyone citing an actual real world problem that was caused because of the use of a 3 wire circuit.
Don

[pbunge]

Don
I appreciate your reasoning as being anchored in real life. Respectful of the intent of the code to ensure safety and yet not wasting time, effort and money to chase down potential risks that exist more in theory than real life.

Please confirm that I am hearing you right. Add an EGC and bond to the new panel, no bonding to a water pipe common to other units. Bond ground and neutrals together on the same buss bar as was done in the old panel.

Peter

[don_resqcapt19]

pb,
Run a new equipment grounding conductor from the subpanel to the main panel and connect directly or indirectly to the grounded conductor at the main panel. Connect the other end to an equipment grounding bus in the subpanel. Keep the grounding and grounded (neutrals) separate in the subpanel. A bond to the a water pipe from the subpanel equipment grounding bus would be fine. Please note that this is not a code compliant installation!! I feel that it is safer than the existing installation and I would permit it if I was the AHJ. The AHJ would be well within his rights to insist on a new feeder with an equipment grounding conductor from the main panel to the subpanel as this is what the code requires.
Don

[karl riley]

I do not think Don's advice can be improved on. It says it all.
Karl

[pbunge]

Don & Karl,
Thanks for you time and expetise.
Peter