Need Help Understanding 250.24(C)

[George Stolz]

I ran into an interesting situation this week, and found my knowledge of theory lacking a bit.

The one-line is as follows:

-Service Disconnect A-
Large service, 3/? 120/208Y, say 600 or 800A, serving 10 apartment units. This service is connected to building steel, building steel has a bonding jumper to water pipe; service is also connected to ground rod.

-Service Disconnect B-
Same utility transformer supplies a CT can, nippled to a 400A 120/208Y disconnect serving house loads. It is a few inches away from the residential service. It has a GEC running to building steel.

-Service Disconnects C & D-
From Service B's CT can, there is an underground run into a gutter inside the Fire Riser room. There will be a disconnect for a 60 HP Fire Pump and a 2 HP Jockey Pump in the riser room.

Problem: The underground conduit run to C & D was accidentally undersized. The installer was going to compensate by running three phase conductors to the gutter, omitting the neutral, and then connecting service disconnects C & D to building steel.

When I tried to explain the problem (beyond simply saying it ain't code ), I found myself grasping at what would transpire if a ground fault were to occur in the Fire Pump. In previous discussions here, I know that the path away from the phase conductors, through building steel to the CT can would present a higher impedance to the fault, but had a hard time attaching numbers and explaining the concept to my satisfaction. Who can help me out here?

(The neutral will be present in the underground when this gets pulled in, I succeeded - I am just looking for a better explanation of the numbers and the theory.)

Thanks in advance.

 

[charlie b]

I am nowhere near a code book, but I vaguely recall there being something about pulling neutral wires with services. Is that the issue here? Was the installer proposing to use the GEC connection to building steel as a substitute for pulling a neutral? Even if that is the problem, the two pumps do not need a neutral, and the absence of one would not have an impact on a fault situation.

 

[George Stolz]

I am nowhere near a code book, but I vaguely recall there being something about pulling neutral wires with services. Is that the issue here? Was the installer proposing to use the GEC connection to building steel as a substitute for pulling a neutral?

Essentially, yes. More accurately, as a substitute for an EGC.

Even if that is the problem, the two pumps do not need a neutral, and the absence of one would not have an impact on a fault situation.

Yes, they require a neutral as a ground-fault current return path. I remember Don saying something about ground fault current not being inclined to follow a path that was not in close proximity to the supplying conductors.

 

[Dennis Alwon]

I think I understand what George is saying. He knows the code requires a neutral at the each service even if the loads do not require one. So know you have a fire pump that has no neutral at the service. The EC wants to run to the building steel to connect (indirectly) to the neutral of the other service. I am not sure whether there is enough impedance to be an issue. What is at issue, beside it being non compliant, is whether this would function during a ground fault- I think this is what George is getting at--. I would guess if the gec is properly sized it would work.

 

[George Stolz]

You got it. It has something to do with inductive reactance or something (theory being my weakest subject) where that GEC path would be a roadblock to fault current.

 

[don_resqcapt19]

Here are a couple of quotes from a thread on www.eng-tips.com.

As resqcapt19 said, the problem is high impedance for a ground fault making fault clearing slow. Even if the ground grid path to the LV ground source is shorter, the impedance will probably be higher because of the separation of the conductor from the ground return path. Inductance is proportional to log(D/r') where D is the distance between conductors and r' is the geometric mean radius of the conductor. The ground grid and EGC will be parallel paths for ground return current and the combination will have a lower impedance than either path.

There is a lot of history on this. Old IEEE papers by R. Kauffman and others showed that the impedance of grounding conductors outside of the raceway containing the phase conductors was so high as to render the grounding conductor ineffective. This is especially true for a steel conduit. As noted by others, this clearly not allowed by NEC and for good technical reasons.

I was aware of some tests done on single conductor cable in conduit a lot of years ago. The tests showed that 10 feet of steel conduit would choke the current to about 100 amps. A further effect of this would be a large voltage drop across the portion of the cable contained in the conduit. A second test was done with the conductor bonded to the conduit at each end. It was found that the current through the conductor was still limited to about 100 amps in the section of the cable contained in the conduit. The balance of the current was conducted through the steel conduit.
The tests were done by an Electrical Contractors Association to illustrate to the member contractors that grounding in accordance with the electrical code was (and is) important and not arbitrary.

 

[George Stolz]

Excellent, thanks Don. :thumbsup:

 

[T.M.Haja Sahib]

Then what is your proposed solution for the original problem?
QUOTE=George Stolz;1332237]Excellent, thanks Don. :thumbsup:[/QUOTE]

 

[Dennis Alwon]

Then what is your proposed solution for the original problem?
QUOTE=George Stolz;1332237]Excellent, thanks Don. :thumbsup:

I believe his solution is to fix it and bring a neutral in from the service.

 

[George Stolz]

Then what is your proposed solution for the original problem?

As I mentioned in the OP, I had swayed the installer before starting this thread - I was looking for more info to share with the installer while we're grunting over a pipe at 39.2% fill.