Does anyone know the real reason a sub-panel has the neutral isolated from the ground in the panel, but back at the main panel it is tied together?

Because a 'grounding' conductor differs from a 'grounded' conductor in that it does not carry current.

the exception being that service entrance conductors are allowed the dual usage of both.

See the 1999 NECH Commentary in Section 384-20:

384-20.
A separate equipment grounding conductor terminal bar is required to be installed and bonded to the panelboard for the termination of feeder and branch-circuit equipment grounding conductors. Where installed within service equipment, this terminal is bonded to the neutral terminal bar, as illustrated in Figure 384.5. Any other connection between the equipment grounding terminal bar and the neutral bar, other than allowed in Section 250-32 is not permitted. If this downstream connection occurs, current flow in the neutral or grounded conductor would take parallel paths through the equipment grounding conductors (the raceway, for example) back to the service equipment. Normal load currents flowing on the equipment grounding conductors could create a shock hazard. Exposed metal parts of equipment could have a potential difference of several volts created by the load current on the grounding conductors. Another safety hazard could be created by this effect where subpanels are used, since arcing or loose connections at connectors and raceway fittings, etc., could serve to create a potential fire hazard.

>Exposed metal parts of equipment could have a potential difference of several volts created by the load current on the grounding conductors.
And then the question becomes one of how many milliamps could flow.

And I insist that with a three wire electric clothes dryer, there are enough mA to be fatal.

If the neutral & ground were tied together somewhere on the interior of a facility, the grounding system would then carry a portion of the neutral current and a few problems would arise. For one, an unsuspecting person may open up a ground connection, become in series with the open connection, allowing the current to flow through him. This could be fatal. Another thing is that a loose connection anywhere on a raceway that serves an EGC would arc and potentially cause a fire. The safest place to bond is at the service equipment. Making this connection is necessary to allow ground fault current to flow back to the utility transformer with a low enough impedance to either operate the primary fuses or to allow the fault to burn clear with minimal arcing.

You guys already nailed this one, how 'bout an example. Sub-panel noodle and ground NOT isolated, neutral (aluminum, no penetrox) became high impedance, sub-panel fed insta-hot water heater (connection between water pipes and ground) now we had neutral (overcurrent) imposed on plumbing pipe. Cast iron DWV (That's Drain, Waste, Vent just in case) now a wet guy in shower touches cold or hot water spigot with big wet toe on drain and learns new dance in shower. Fixed this 20 years ago, and code became MUCH clearer after I drew the circuit for myself.

I know of an open neutral incident where the current flowing through the plumbing system generated enough heat to melt solder and cause major water damage. It sounds hard to believe that with water in the pipe the solder would melt, but that appears to be the case.

Redsy,
Boy I've heard of that, but never seen it, the only thing close is a welder who had clamped onto a pipe (with water in it) and the ground path was bad, did the same thing, melted the solder off the joints. What a mess, but a NEAT mess.

It's especially hard to believe when you have tried to solder a pipe that has a little water in it and the darn solder won't flow.

hmmm,
would anyone here consider the bond at the X-former, and isolation throughout the premisis ( all service equipment..) a cleaner method?

>would anyone here consider the bond at the X-former, and isolation throughout the premises
I'd say move it back to the meter. But that argument was that the bond has to be where we can inspect it.

If the bond is at the transformer, lightning will surely arc in the service panel. I would rather that it have a heavy gage jumper.

I don't see any problems with bonding at the service equipment. Does anyone have concerns about this?

N&G Isolation back to the X-former would have it's pro's and con's....
I'm just tryin' to think out of the box here...


were's Paul from UK ?

Sparky,

I am for separate Neutral and EGC downstream from the source, SERVICE or SDS.

The main "con" that I see is the fact that for services, the NEC would not have any say over the NESC.

I believe that separate Neutrals and EGC's would help reduce the amount of the electrical degrading of the water piping ( Plumbing ) industry. There should be considerably "less" current flow through such piping systems and 'earth returns' if the Neutral was completely an insulated conductor after the grounding at the source.
The EGC then would only carry minute 'leakage' ( primarily CAPACITANCE ) current from the ungrounded conductors except for "Fault Conditions".

Glenn

Right here Sparky..... Been so busy in the other areas that I hadn't looked in here for a few days!

With the three different systems used here that I described some time ago, there's always been discussion over which is best.
All in all, the old urban system has a lot to recommend it. Neutral is grounded only at the utilty xfmr star-point, and the house ground is via the armor of the distribution cable, which is also grounded at the utility's sub-station. In the event of a ground fault, there's bound to be some current flowing through parallel earth paths, but the bulk of the fault current is confined to the cable armor. The main disadvantage is cost, which is why such systems haven't been installed here for years.

I take the point about possible effects of lightning where the neutral isn't grounded at the service entrance, but all of these systems were/are completely underground. (There's also the point that storms are nothing like as frequent or severe as in most parts of the U.S.A.)

To recap our PME system: The main panel still has separate neutral & ground busbars, but the ground is bonded to the neutral at the service block, which is immediately before the meter. Originally used only in areas where it was difficult to get a low loop impedance with a local ground rod, this system has now become widespread.

Because of the possible hazards of a broken neutral, our rules specify stricter bonding requirements for houses using PME.

In non-PME systems, the minimum size conductor for bonding to water/gas pipes etc. used to be 2.5 sq. mm (slightly larger than #14 AWG), although 4 sq. mm (a little over #12) is now standard.

Where PME is employed, all bonding conductors have to be min. 10 sq. mm (a little larger than #8).

The third system is probably the least satisfactory, for reasons I've alreay explained.

> I don't see any problems with bonding at the service equipment. Does anyone have concerns about this?

Yes. Every metallic pathway from the bond point back to the xformer becomes a co-conductor with the neutral.

If you are using any conductive conduit between the main bond and the xformer and it is bonded at both ends (such bonding is practically unavoidable), that conduit will be a conductor of neutral current.

Sorry,
I was locked in to thinking O/H service.
Also I understand that even with O/H, the earth becomes a parallel path.
BTW,
Wouldn't service drops & laterals require an additional conductor to ground if bonding occurred at the Xfmr only? The utility would bear this additional cost?

>Also I understand that even with O/H, the earth becomes a parallel path.
Yes, it is. But it is a poor path compared to a metallic path. Nonetheless, EMF will be created by the imbalance.

>Wouldn't service drops & laterals require an additional conductor to ground if bonding occurred at the Xfmr only?
Absolutely! (I'm not the person who suggested this.)

>The utility would bear this additional cost?
Of course not! (They don't have to follow the NEC anyway.)

Step & touch voltage gradients due to poor soil conductivity?

Voltage gradient hazards are very real. There can be enough voltage between two places even a short distance apart to shock you.

For instance: I approved a trailer service for connection. The service was installed correctly. The power company went out and hooked it up. I got a call from the owner that he was getting shocked off the frame while lying on his back hanging pipe underneath it. I went out and found about 30 volts from the frame of the trailer to the earth. I shut off the disconnect and nothing changed.

To make a long troubleshooting story short, we got the voltage to go away by crossing the road and unscrewing the water pump fuse at his parent’s house. The service on the house was ungrounded, rusty pipe and corroded off clamp, NM-B buried to the water pump and rotted insulation. The current was traveling from the service out the pump wire, across the road past the trailer and up through the ground rods to the trailer pole service neutral then back to the transformer.

There was a big lossy voltage divider about a eighty feet long with 120 volts across it between the romex and the trailer ground rods. At the trailer frame it was 30 volts relative to the rods. I placed my meter between the ground rod clamp and a point 24” away from the rods and still got a couple of volts. And yes, they were having high power bills.

That’s why I always suggest driving a local set of ground rods nearby any distribution panel regardless of if the code requires it or not. Whether it is on a second structure, outdoor building, tool shed, picnic shelter, doghouse or whatever. If you are using power here and the circuit is grounded somewhere else, and even if you are not USING the power, there is a chance of shock when you get between the earth and a remotely grounded metallic object.