Efforts to ensure continuity of power must not create potential hazards through improper grounding techniques.

Remote pumping stations, each with their own local disconnecting means and grounding electrodes, frequently cause difficulties in applying appropriate grounding rules. Often a property owner and the utility establish a remote service point where the power will be metered, and the owner takes over from there, providing the wiring to the remote stations. In some cases there may be from four to a dozen installations stretching out up to a mile or more in total distance. In the specific case involved here, the system is 480V corner grounded. Some utilities require this type of service in cases where there are no ground detectors in place.

At the point of supply to the customer wiring system, there will be a meter and a fused disconnect (or circuit breaker). This is the service disconnect for the installation, and it also allows the utility to troubleshoot and for overload protection for the transformer. Where the utility's primary distribution includes a grounded conductor (the usual case), there may be a 4-wire drop to the customer's pole. One option some utilities use is to connect the messenger wire to a driven ground at their distribution transformer only. The owner then usually distributes power from the meter location with three separate, insulated overhead conductors.

We received a report that when motors are connected to the load side of this distribution, there are instances of fault currents passing through local electrodes with magnitudes sufficiently high enough to literally bake dry the soil at the electrode. The result is an extremely high electrode resistance, and the frames of local motors and switchgear can be operating with a dangerous touch voltage.

When this happens, some owners treat the downstream equipment as though it were connected to an ungrounded system in order to increase reliability of service on these untended remote installations. This means that although there is a grounding electrode at each equipment location, there is no connection between the grounded phase conductor and this electrode.

According to the report we received, the owner, in cooperation with the utility, decided to lift the grounded phase off the center lug of the meter socket, and insulate it. In addition, he proposed to connect the load end of the messenger wire to a driven ground at his service pole. The meter enclosure and disconnect will be grounded through a short piece of solid copper, bonded to the pole ground. The proposed solution is shown in the drawing.

The EC&M Panel's response

We strongly disagree with the proposed solution. The system is a chimera, an incongruous combination of grounded and ungrounded distribution practices that fails to offer the minimum safety requirements of either.

Lifting the grounded phase from the service grounding terminals does not make the system ungrounded. A fault in the service equipment will still return to the transformer over the pole ground wire and the quadruplex messenger. Note, however, that the pole ground may not have the required size and would not be "routed with the phase conductors" as required in Sec. 250-23(b).

The owner's grounding system

However, our real concerns arise from what the description implies about the grounding system. Grounding electrodes are part of a ground reference arrangement that stabilizes the voltage to ground and dissipates surges. In performing this function, they are never intended to carry significant amounts of current for extended periods of time. At utilization voltages, they must never be relied upon for the return of fault current.

Fault current from an insulation failure can only bake an electrode if it is denied the properly constituted conductive path back to the system source, as required for these systems. In this case, that path would be over the grounded phase conductor by way of the main bonding jumper at each structure disconnecting means. According to the description, the customers are omitting these bonding jumpers in order to treat the system as though it were ungrounded. However, the system is grounded at the transformer. And where distribution systems are grounded at any point, Sec. 250-23(b) and Sec. 250-53(b) require that the grounded circuit conductor be brought to the service disconnecting means and bonded to the enclosure and any equipment grounding conductors.

Sec. 250-24(a) requires similar bonding for the downstream pumps, because the pumping stations would be considered "structures" in the application of that rule.

In both cases (service and remote station), you must make a bond to the grounding electrode conductor, but the primary fault current path is over the grounded circuit conductor, which should be identified according to the rules in Art. 200.

If you don't make this bonding connection, in the event of an insulation failure that energizes conductive materials that are connected to earth through an electrode, and with no other return path, current will flow through the electrode in accordance with Ohm's law. If the ground rod resistance is 25 ohms, then about 19A of current will begin flowing (480/25 = 19) through the electrode. This won't trip the feeder overcurrent protective device.

Since power is equal to [I.sup.2]R, this will produce about 9kW of heat. Most electrode resistance is concentrated near the electrode, and therefore this is where the heat will be concentrated. If continued, this will indeed bake the ground to the point that the electrode is worthless (and, incidentally, in violation of Sec. 250-84 at locations with a single electrode). As the electrode resistance deteriorates, the touch voltage on local conductive surfaces will approach 480V (depending on local voltage gradients), an extreme hazard.

A true, ungrounded, alternative

If the owner truly wants the continuity advantages of an ungrounded system, then by Code he or she may have it. But the owner must wire per the Code, without using identified phase conductors and with overcurrent protective devices in all three phase conductors. And the owner must make whatever arrangements the utility requires to connect it. On installations like this, where exposed overhead conductors run for great distances, lightning arresters should be provided at each pump location. These surges are often more destructive on ungrounded systems because there is no circuit path to ground.

With a ground fault on an ungrounded system, the system simply becomes corner grounded at the point of the fault. The touch voltages on the connected enclosures do not increase significantly above ground. There is time to arrange an orderly shutdown and correct the problem. Normally, if a second fault occurs on a different phase, the resulting short circuit through the intervening equipment grounding paths simply opens one or more overcurrent protective devices. In this case, however, ground detectors take on additional practical significance.

However unlikely, if such a ground fault is not corrected at one location, and another fault occurs on a different phase at a remote location, current will attempt to flow between the electrodes connected at each. This will produce a similar result as that described for the improperly connected grounded system. Overcurrent devices may well not operate. The phase-to-phase voltage between the faulted phases will drop by varying amounts depending on electrode resistance. This may decrease the life of the motors connected to the distribution, as well as increasing shock hazards over time. The enormous amounts of energy wasted while such a fault is in progress will be metered and charged to the customer.

The NEC does not require ground detection on ungrounded systems, but a fine print note in Sec. 250-5(b) refers to its usefulness. It might be possible to rig a radio operated monitoring device that would alert a central station when a phase ground occurs. We think that the NEC applies and should be enforced on all of the wiring downstream of the service point.

EDITOR'S NOTE:

These answers are given by our panel of experts. I am chairing this panel, and the other panel members include Bill Summers, James Stallcup, and Dan Leaf. The opinion expressed is that of the panel. If a panelist disagrees with the majority opinion, his explanation is printed following the answer. Although authoritative, the answers printed here are not, and cannot be relied on as formal interpretations of the National Electrical Code.