Grounding Vs Bonding Part (1-12) of 12

Grounding vs Bonding Part 1 of 12
Jan 1, 2005 12:00 PM, By Mike Holt, NEC Consultant

Safe installations start with a basic understanding of grounding and
bonding terminology
All Code references are based on the 2005 National Electrical Code. The
grounding and bonding requirements in this column apply to solidly grounded
systems that operate at not more than 600V, such as 120/240V, 120/208V,
and 277/480V.
If someone were to investigate which NEC Article suffers from the most
misapplications, violations, and misinterpretations, they'd find that Art. 250
easily claims that distinction. The situation is so dire that Art. 250 violations
are sometimes required. For example, many industrial equipment manuals
require violating 250.24(A)(5) as a condition of warranty. In particular, they
require installing an “isolated grounding electrode.” By this, they mean an
electrode without a low-impedance fault-current path back to the source
winding (other than through the earth itself). This creates a condition where
the ground-fault current return path to the source winding (utility transformer)
is on the order of several ohms rather than the fraction of an ohm that would
be provided by an NEC-compliant installation.
Fig. 1. Specific bonding requirements can be found in Part V of Art. 250.

If you apply basic physics and electrical theory, you can clearly see that Art.
250 is correct and equipment manuals that require “isolated grounding” are
wrong. Other standards agree: ANSI/IEEE Standard 142, Recommended
Practice for Grounding of Industrial and Commercial Power Systems (Green
Book) and Soares Book on Grounding use the same physics and electrical
theory as Art. 250. Also, IEEE-142 describes the correct way to provide an
isolated ground — and it isn't the method proposed by many equipment
manuals.
Art. 250 isn't a “preferred design specification,” as defined in Art. 90. It
provides the minimum requirements for a safe installation. These
requirements include providing paths to divert high voltage to the earth,
providing low-impedance fault-current paths for overcurrent protection
devices, and removing dangerous potentials between conductive building
components and electrical systems.
Fig. 2. Without an effective ground fault current path, you risk incurring
serious system damage.
Coming to terms with Art. 250. To correctly apply Art. 250, you must
understand how the NEC defines specific terms. This 12-part Code Basics
series will take that a step further and provide clarification of those terms that
can be especially confusing. Where the NEC uses the term “grounding” to
mean “connecting to the earth,” “earthing” will follow in parentheses. Where
the NEC uses the term “grounding” to mean “connecting to a conductive body
for the purpose of providing a low-impedance path to the source winding,” the
term “bonding” will follow in parentheses. A variation of this convention will be
used for “ground” and “grounded.”
Bonding (bonded). The permanent joining of metallic parts together to form an
electrically conductive path. This path must have the capacity to safely
conduct any fault current likely to be imposed on it (Fig. 1).
Bonding jumper. A reliable conductor sized per Art. 250 to ensure electrical
conductivity between metal parts of the electrical installation.
Fig. 3. This is one of three definitions found in 250.2.
Effective ground-fault current path. An intentionally constructed, permanent,
low-impedance conductive path designed to carry fault current from the point
of a ground fault on a wiring system to the electrical supply source winding
(Fig. 2).
Equipment grounding (bonding) conductor. The low-impedance fault-current
path used to connect the noncurrent-carrying metal parts of equipment,
raceways, and other enclosures to the grounded (neutral) conductor and
equipment grounding (bonding) conductor at service equipment or at the
source of a separately derived system.
Ground fault. An unintentional connection between an ungrounded conductor
and earth or metallic parts of enclosures, raceways, or equipment (Fig. 3).
Ground-fault current path. An electrically conductive path from a ground fault
to the source winding. The NEC uses the phrase “ground-fault current path,”
but fault current isn't traveling to the earth — it's traveling to the source
winding of the power supply.
Grounded (earthed). Connected to earth.
Fig. 4. The absence of an effective ground-fault current path creates a lethal

situation.
Grounding (earthing) conductor. A conductor used to connect equipment to a
grounding (earthing) electrode.
Grounding (earthing) electrode. A device that establishes an electrical
connection to the earth.
Grounding (earthing) electrode conductor. The conductor used to connect the
grounding (earthing) electrode(s) to the equipment grounding (bonding)
conductor, to the grounded (neutral) conductor, or to both in accordance with
250.142.
Main bonding jumper. A conductor, screw, or strap that bonds the equipment
grounding (bonding) conductor (service disconnecting means) to the
grounded (neutral) conductor in accordance with 250.24(B). For more details,
see 250.24(A)(4), 250.28, and 408.3(C).
Solidly grounded.The intentional electrical connection of one system terminal
to the equipment grounding (bonding) conductor per 250.30(A)(1).
System bonding jumper. The conductor, screw, or strap that bonds the
equipment grounding (bonding) conductor (metal parts of a separately derived
system) to one of the system conductors or terminal per 250.30(A)(1).
Why so much emphasis on wording? If you understand grounding- and
bonding-related terminology, you can then comply with Art. 250 requirements
and produce safe installations.
An illuminating example. Too often, metal parts are grounded (earthed)
instead of bonded. The accepted grounding practice for street lighting and
traffic signals in many parts of the United States, in which the ground rod is
used as the only fault-current return path, is a good example of how
commonly Art. 250 is misapplied. In this case, the metal pole of a light fixture
or traffic signal is grounded to a ground rod, but an effective ground-fault
current path isn't there.
For this scenario to make sense, you need to understand three concepts:
touch voltage, hazardous level, and surface voltage gradients.
 The IEEE defines “touch voltage” as “the potential (voltage) difference
between a bonded metallic structure and a point on the earth 3 feet from the
structure.”
 ANSI/IEEE Standard 142, [4.1.1], says the resistance of the soil outward
from a ground rod is equal to the sum of the series resistances of the earth
shells. The shell nearest the rod has the highest resistance and each
successive shell has progressively larger areas and progressively lower
resistances. This layering of shells results in “surface voltage gradients.”
A person could receive a lethal shock as little as one foot away from an
energized object.
The Table lists the percentage of total resistance and the touch voltage for
the light pole in Fig. 4 above, based on a 120V fault. As the Table shows, the
voltage gradient of the earth drops off so rapidly that a person in contact with
an energized object can receive a lethal shock one foot away from an
energized object if the metal parts aren't bonded to an effective ground-fault
current path to remove the voltage by clearing the fault.
Because the resistance of the earth is so high, very little current will return to
the power-supply winding if the earth is the only ground-fault return path. If a
metal lighting pole is only grounded (earthed) to a ground rod, then the earth
is the sole ground-fault current path, which is a violation of 250.4(A)(5).
Consequently, the circuit overcurrent protection device won't open and metal
parts will remain energized at a lethal level until someone makes contact with
them and the earth. Therefore, a ground rod doesn't lower touch voltage to a
safe value for metal parts that aren't bonded to an effective ground-fault
current path.
If the people involved in street lighting and traffic signaling in these locations
where the Code is misapplied understood the terminology of Art. 250, the
situation would be very different — and much safer. Sadly, this is only one
example of terminology-based misapplication; dozens of others exist.
So you can see the importance of understanding the terminology. But that
means taking time to learn definitions. The good news is that task isn't as
onerous as it might seem if you can remember the concepts of “earthing” and
“bonding.” In the first case, you're connecting to the earth. In the second,
you're connecting to a conductive body for the purpose of providing a low-
impedance path to the source winding. Many times when the NEC says
“grounding,” it's referring to bonding.
As this series progresses, you'll see these terms many times. More
importantly, you'll encounter them in your work — where a solid
understanding of grounding- and bonding-related terms will allow you to avoid
mistakes, such as those in the street lighting example. An old adage says,
“Words cannot hurt you,” but when it comes to grounding and bonding, not
understanding certain words can hurt everyone.
Stay tuned to Code Basics throughout the year for the rest of this series.
Feb 1, 2005 12:00 PM, By Mike Holt, NEC Consultant
General requirements for grounding and bonding differ
All Code references are based on the 2005 National Electrical Code. The
grounding and bonding requirements in this column apply to solidly grounded
systems that operate at not more than 600V, such as 120/240V, 120/208V,
and 277/480V.
Art. 250's requirements for grounding and bonding, which begin in 250.4, can
be broken down into two groups:
1. Grounded systems [250.4(A)]

2. Ungrounded systems [250.4(B)]
It may seem odd that the NEC has grounding and bonding requirements for
ungrounded systems, but 250.4's first requirement is for grounding in
grounded systems, so that's where we'll start.
Fig. 1. The earth shall not be considered an effective ground-fault current

path.
Grounding and bonding in grounded systems. The NEC requires you to
ground (earth) system windings to limit the voltage imposed on the system by
lightning, unintentional contact with higher-voltage lines, or line surges
(Sidebar below). Another function of this earthing is to “stabilize the voltage to
earth during normal operation” by providing a common reference point.
The NEC also requires you to ground (earth) metal parts of electrical
equipment in or on a building or structure. See 250.24(A) for services and
250.32(A) for separate buildings or structures. You accomplish this grounding
(earthing) by using a grounding (earthing) electrode conductor [250.64(A)] to
electrically connect the building or structure disconnecting means (225.31 or
230.70) to a grounding (earthing) electrode [250.52, 250.24(A), and
250.32(A)] (Fig. 1 above).
However, grounding (earthing) has its limitations. Grounding (earthing) of
electrical equipment doesn't provide a low-impedance fault-current path to
clear ground faults. In fact, the Code prohibits the use of the earth as the sole
return path because it's a poor conductor of current at voltage levels below
600V [250.4(A)(5) and 250.45(B)(4)].
In addition, grounding (earthing) the metal parts of electrical equipment
doesn't protect electrical or electronic equipment from lightning-induced
voltage transients (high-frequency voltage impulses) on the circuit conductors
inside the building or structure. Nor does it protect equipment within a
structure from transients generated from other equipment in that structure.
To provide protection from voltage surges, you must engineer a proper surge
protection system. The design should address surge protection devices (Art.
280 and 285) at service equipment, panelboards, and critical loads. Although
it isn't covered by the NEC, you could also consider point-of-use surge
protection in branch circuits.
Fig. 2. Bonding electrically conductive materials to each other and to the

supply source establishes an effective ground-fault current path.
Bonding in grounded systems has its own challenges. An “effective ground-
fault current path” is a permanent, low-impedance path for fault-current, and it
facilitates the operation of the circuit overcurrent protection device (OCPD)
(250.2) (Fig. 2). Because the earth is a poor conductor, it doesn’t permit
sufficient fault current to flow back to the system winding to open the OCPD.
Thus, a rod or concrete-encased electrode won’t assist in clearing a ground
fault [IEEE 142 Section 2.2.8] (Fig. 3).
A ground fault clears on a circuit through the automatic opening of the OCPD.
The time it takes for an OCPD to open is inversely proportional to the
magnitude of the fault current. Thus, the higher the ground-fault current value,
the sooner the OCPD will open and clear the fault.
Fig. 3. To quickly remove fault current from metal parts of the system, the
fault-current path must have sufficiently low impedance.
To quickly remove lethal touch voltage from metal parts after a ground fault,
the fault-current path must have sufficiently low impedance for the fault
current to quickly raise and facilitate opening the OCPD. For example, a 20A
circuit with an overload of 40A (two times the rating) would trip a breaker in 25
to 150 seconds. At 100A (five times the rating) the breaker would trip in 5 to
20 seconds.
Thus, the effective ground-fault current path is critical. And you establish it
through bonding. Start by connecting non-current carrying conductive
materials of electrical equipment together and to the electrical supply source
[250.4(A)(3)].
Do the same for electrical raceways, cables, enclosures, equipment, and
other electrically conductive materials “likely to become energized”
[250.4(A)(4)]. Whether something is “likely to become energized” is subject to
interpretation by the authority having jurisdiction (AHJ).
Dangerous voltages can also exist on metal water, sprinkler, and gas piping,
and exposed structural steel members, so if any such items are likely to
become energized, you must bond them to the effective ground-fault current
path.
Grounding and bonding in ungrounded systems. When used to describe a
system, the term “ungrounded” refers to its supply configuration and wiring
scheme. In a grounded system, the supply transformer secondary windings
may be wye-configured with the center tap grounded or it may be delta-
configured with a corner grounded.
Unlike a grounded system, an ungrounded system doesn't have a winding
grounded at the supply transformer. This allows for continued operation if you
have a ground fault on one phase. Presumably, qualified personnel will locate
and repair the fault before a ground fault on a second phase takes the system
down.
Ungrounded systems are common in factories in the southern United States.
These facilities typically have ground-fault monitors to alert maintenance
personnel to a ground fault on any phase. For example, the Daramic plant in
Owensboro, Ky., has a ground-fault monitor on the wall between the
maintenance offices and the machine shop — where maintenance personnel
pass by frequently.
But maintenance crews at such facilities must repair the fault promptly. If a
ground fault persists on an ungrounded system, the system could experience
overvoltage-to-ground as high as eight times the phase voltage. This
excessive voltage can puncture conductor insulation and create additional
ground faults.
Whether your system is grounded or ungrounded, you must bond enclosures
and equipment together. In ungrounded systems, bonding of electrical
equipment [250.4(B)(2)] serves a purpose similar to that set forth for bonding
electrical equipment in grounded systems [250.4(A)(3)]. The difference here is
you're bonding the ungrounded system equipment to each other, rather than
to each other and the source.
One consequence of this arrangement is the equipment-bonding path must be
capable of carrying the maximum fault current likely to be imposed on it.
Remember, the bonding system must be able to remove dangerous voltage
from a second ground fault.
The same difference and consequence applies to the bonding of electrically
conductive materials and other equipment in ungrounded systems
[250.4(B)(3)] vs. those in grounded systems [250.4(A)(4)].
Fault current paths. Requirements for establishing fault current are similar
for grounded and ungrounded systems. In neither case can you use the earth
as the sole equipment-grounding conductor or consider it a fault current path.
And in both cases you must install electrical equipment, wiring, and other
electrically conductive material likely to become energized in a manner that
creates a permanent, low-impedance circuit. Yet this requirement has one key
difference:
 For grounded systems, you must establish an effective ground-fault current
path. The circuit must be able to safely carry the maximum ground-fault
current likely to be imposed on it from any point in the wiring system where
a ground-fault to the supply source may occur.
 For ungrounded systems, you must establish a fault current path (not
ground-fault current path). The circuit facilitates the operation of overcurrent
devices if a second ground fault occurs.
A single ground fault can't be cleared on an ungrounded system because
there's no low-impedance fault-current path to the source. However, all metal
parts of an ungrounded system must be bonded together so the bonding path
will provide a low-impedance fault-current path in the event of a second
ground fault (line-to-line fault). This will allow the OCPD to clear the fault
[250.4(B)(4)].
You'll notice more of the differentiation between grounding and bonding as
you continue studying grounding and bonding in Art. 250.
Sidebar: Lightning and Surges

When lightning occurs, high voltages drive high current (as much as 40,000A)
into the earth for a fraction of a second. Typically, lightning strikes to wiring
occur outside. Therefore, grounding (earthing) the system windings will assist
the flow of lightning into the earth.
When a ground fault over 600V occurs, voltage on the other phases can rise
significantly during the fault (typically three to 12 cycles). This voltage surge
during the utility ground fault will be transformed into an elevated surge
voltage on the secondary, possibly destroying electrical and electronic
equipment. The lower the resistance of the utility grounding (earthing) system,
the lower the secondary voltage surge.
Grounding vs Bonding — Part 3 of 12
Mar 1, 2005 12:00 PM, By Mike Holt, NEC Consultant
Grounding and bonding at service equipment
All Code references are based on the 2005 National Electrical Code.
Because utilities provide grounded AC services and most facilities have at
least one utility service connection, a grounded AC service most likely
provides power to your premises wiring system. When you have one, your
premises wiring system must have a grounding electrode conductor
connected to the grounded service conductor [250.24(A)].
This brings up the question of how to comply with grounding electrode
conductor requirements. Because a grounding electrode conductor must
connect the grounded conductor to the grounding (earthing) electrode, the
question of how expands to include where. Can you make this connection
anywhere?
Fig. 1. Make your grounding electrode conductor connection at one of these

three points.
Location, location, location. Some inspectors require the grounding
electrode conductor to terminate to the grounded conductor terminal at the
meter enclosure. Other inspectors require the grounding electrode conductor
to terminate to the grounded terminal at the service disconnect.
The Code says you can make this connection at any accessible location, from
the load end of the service drop or service lateral up to and including the
service disconnecting means [250.24(A)(1)] (Fig. 1). The choice then
becomes an engineering decision that balances such factors as installation
costs, available space, and maintenance issues.
Connections. In this day of increased demand for uninterrupted power, many
facilities are dual-fed. This means separate lines run to the same services;
such services are referred to as “double-ended.” If the dual feeds are in a
common enclosure (or grouped together in separate enclosures) and they
employ a secondary tie, you can use a single grounding electrode connection
to the tie point of the grounded conductors from each power source
[250.24(A)(3)].
Whether your service is double-ended or not, you must install an unspliced
main bonding jumper between the grounded terminal and any metal on the
service disconnecting means enclosure. Ensure the bonding jumper and
installation comply with 250.28 and 250.24(C), respectively.
Your main bonding jumper is probably a wire or busbar. If you've installed this
jumper from the grounded conductor terminal (or bus) to the equipment-
grounding terminal (or bus) in the service equipment, the NEC allows you to
connect the grounding electrode to the same equipment-grounding terminal
(or bus or bar) to which you connected the main bonding jumper
[250.24(A)(4)].
Bonding of service equipment. A neutral-ground bond anywhere other than
at service equipment is a common cause of power quality problems. Such a
bond creates ground loops, which allow undesired current to circulate in the
system. Power quality problems often lead to the discovery and removal of
such a bond. But don't wait for power quality problems to reveal the bond.
Fig. 2. A grounded neutral conductor must be run to and bonded to each

service disconnect.
Another concern makes corrective action imperative. Load side neutral-
ground bonds allow objectionable current to flow on conductive metal parts of
electrical equipment — thereby violating 250.6(A). This objectionable current
can cause lethal electric shock. And it sets the stage for inadvertent
flashovers, overheating of equipment, and other problems stemming from the
presence of electricity in the wrong place.
So don't make (or allow) a neutral ground connection on the load side of the
service disconnect [250.24(A)(5)]. However, exceptions to this rule (250.142)
allow you to make such a connection for:
 Separately derived systems if you follow the requirements of 250.30(A)(1).
 Separate buildings if you follow the requirements of 250.32(B)(2).
Grounded conductor. Electric utilities don't typically provide an equipment-
grounding (bonding) conductor to service equipment, and they aren't required
to do so. Thus, you must run a grounded conductor from the electric utility
transformer to each service disconnecting means [250.24(B) and 250.130(A)]
(Fig. 2).
The grounded service conductor provides the effective ground-fault current
path to the power source winding. This path ensures that opening the circuit
protection device will quickly remove dangerous ground-fault voltage from the
circuit [250.4(A)(3) and 250.4(A)(5)].
Fig. 3. Very little fault current returns to the source if the earth is the only fault-
current return path.
The earth's resistance is too great for it to be an effective bonding jumper.
Very little fault current returns to the power source winding if the earth is the
only fault-current return path. But let's suppose the earth is your only fault-
current return path option. What would the consequences be? For one thing,
the circuit overcurrent protection device wouldn't open and clear the ground
fault. Consequently, metal parts like metal piping and structural building steel
would become — and remain — energized to circuit voltage (Fig. 3). The
system then poses a high risk of shock, arc flash, and fire.
You can calculate, for example, the voltage on a metal enclosure due to an
open service grounded conductor. Forensic engineers often crank out these
kinds of numbers when investigating a fatality case or damage to a facility. It's
easier just to comply with NEC requirements and eliminate such a voltage in
the first place.
So it's obvious you need a grounded conductor, but how big should it be?
Remember, this grounded service conductor serves as the effective ground-
fault current path. Thus, you must size it so it can safely carry the maximum
fault current likely to be imposed on it [110.10 and 250.4(A)(5)]. Size the
grounded conductor per Table 250.66 — based on the total area of the largest
ungrounded conductor. The grounded conductor must also have the capacity
to carry the maximum unbalanced current, per 220.61.
To test your understanding of the concept, consider the following scenario:
What's the minimum size grounded service conductor required for a 480V, 3-
phase service, where the ungrounded service conductors are 500 kcmil and
the maximum unbalanced load is 100A?
The unbalanced load requires a 3 AWG grounded service conductor — rated
for 100A at 75°C per Table 310.16 [220.61]. However, the grounded service
conductor can't be smaller than 1/0 AWG (Table 250.66). This minimum size
requirement ensures the conductor will accommodate the maximum fault
current likely to be imposed on it. Thus, the real answer is 1/0 AWG.
If you parallel your service conductors, do you use just the one conductor or
do you parallel your grounded conductor the way you parallel the current-
carrying conductors? The answer is neither.
First, you must install a grounded conductor in each raceway whenever you
parallel your service conductors.
Second, you can't simply divide your grounded conductor into two smaller
equal conductors. You would satisfy the requirement that the grounded
conductor must have the capacity to carry the maximum unbalanced current
per 220.61, but it could also result in a grounded conductor that's too small for
a given raceway.
To eliminate such a problem, size each grounded conductor per Table 250.66
— based on the total area of the largest ungrounded conductor in the
raceway. Note that regardless of the number you come up with, the grounded
conductor in each parallel service raceway can never be less than 1/0 AWG
(310.4).
Let's review with another quick quiz:
What's the minimum size grounded service conductor required for a 480V, 3-
phase service installed in two raceways, where the ungrounded service
conductors in each raceway are 350 kcmil and the maximum unbalanced load
is 100A?
The unbalanced load requires only a 3 AWG grounded service conductor, per
Table 310.16 (220.61). However, the grounded service conductor in each
raceway can't be smaller than 2 AWG (Table 250.66) [250.24(C)(2)]. This
ensures it will accommodate the maximum fault current likely to be imposed
on it. But ungrounded service conductors run in parallel can't be smaller than
1/0 AWG (310.4), so the answer is 1/0 AWG per raceway.
Properly grounding and bonding service equipment improves safety while
eliminating a common cause of power quality problems. You just have to
make the right connections in the right places. If you think in terms of
providing a low-impedance ground-fault path back to the source, you'll have
no problem.
Apr 1, 2005 12:00 PM, By Mike Holt, NEC Consultant
All Code references are based on the 2005 NEC. The grounding and bonding
requirements in this column apply to solidly grounded systems that operate at
not more than 600V, such as 120/240V, 120/208V, and 277/480V.
If you don't work with generators, you may think the requirements for
separately derived systems (SDSs) aren't relevant to you. However, that
thought may conflict with the reality of your situation.
People often forget that all transformers, except autotransformers, are
separately derived. This is because the primary circuit conductors have no
direct electrical connection to the secondary circuit conductors.
Fig. 1. A generator is considered to be a separately derived system if it has

no direct connection to other system conductors.
People also frequently forget that not all generators are separately derived. A
generator that supplies a transfer switch that opens the grounded neutral
conductor is separately derived [250.20(D) FPN 1] (Fig. 1 at right). This is also
the case if there's no grounded neutral conductor. But often generator setups
don't switch the grounded neutral conductor in the transfer switch.
To determine if you have an SDS, start with the definition given in Art. 100: “A
premises wiring system whose power is derived from a source of electric
energy or equipment other than a service. Such systems have no direct
electrical connection, including a solidly grounded circuit conductor, to supply
conductors originating in another system.” (Fig. 2).
Fig. 2. A separately derived system is defined as a wiring system whose
power has no direct electrical connection to the supply conductors originating
in another system.
System bonding jumper. Part 3 of this series explained that you can't make
neutral-to case (load side) bonds [250.24 (A)(5)]. It also discussed the
exception it provides for SDSs if you follow the requirements of 250.30(A)(1).
Bonding the metal parts of the SDS to the secondary grounded neutral
terminal — by installing a system bonding jumper — ensures quick removal of
dangerous voltage from a secondary ground fault. You can also accomplish
this by opening the overcurrent protection device of the secondary circuit
[250.2(A)(3)].
Overall, the system bonding jumper:
 Connects the equipment grounding conductors of the SDS (metal case) to
the grounded neutral conductor (typically the X0 terminal).
 Is sized based on the derived phase conductors. Connect the system
bonding jumper, per 250.28(A) through (D).
 Should be installed at only one location — either at the source or the first
system disconnecting means or overcurrent device.
 Shouldn't be installed at both the SDS and the secondary system
disconnecting means because it will allow dangerous objectionable current
to flow on conductive metal parts of electrical equipment, metal piping, and
structural steel in violation of 250.6(A) (Fig. 3).
Size the system bonding jumper per Table 250.66 based on the area of the
largest ungrounded secondary conductor [250.28(D)]. Where you run an
equipment-bonding jumper to the secondary system disconnecting means,
you must also size it per Table 250.66 based on the area of the largest
ungrounded secondary conductor.
Fig. 3. Objectionable current flowing on metal parts of a system creates a
personal hazard.
Where you run SDS conductors in parallel, install a secondary equipment
bonding jumper in each raceway (or cable) and once again size it per Table
250.66 based on the area of the largest ungrounded secondary conductor
[250.102(C)].
Grounding electrode conductor (GEC). For each SDS, ground the
grounded neutral terminal (typically X0) to a suitable grounding (earthing)
electrode of a type identified in 250.30(A)(7). Size the secondary system
grounding (earthing) electrode conductor per 250.66 based on the total area
of the largest ungrounded secondary conductor.
To prevent objectionable current flow in metal parts of electrical equipment,
metal piping, and structural steel, terminate the GEC to the same point on the
SDS where you installed the system bonding jumper. However, note the
following exceptions:
 Where the system bonding jumper [250.30(A)(1)] is a wire or busbar, you
can terminate the GEC to the equipment grounding terminal bar or bus on
the metal enclosure of the SDS (Fig. 4).
 Where the SDS originates in listed equipment suitable as service
equipment, you can use the GEC from the service or feeder equipment to
the same grounding electrode as the GEC for the SDS — if it meets the
size requirements.
 You don't need to ground (earth) SDSs rated 1kVA (1,000 VA) or less. But
you must install a system bonding jumper per 250.30(A)(1) to ensure the
clearing of ground faults.
Connect grounding (earthing) electrode tap conductors to the GEC without
splicing the common GEC. Install the GEC per 250.64. The basic
requirements mandate that the GEC is:
 Copper, where within 18 inches of earth [250.64(A)].
 Securely fastened to the surface on which it's carried [250.64(B)].
 Adequately protected if exposed to physical damage [250.64(B)].
Fig. 4. Where the system bonding jumper is a wire or bus bar, you can
terminate the grounding electrode conductor to the equipment grounding
terminal of the derived system.
If you run the GEC in metal enclosures, make them electrically continuous
from the point of attachment to cabinets (or equipment) to the GEC
[250.64(E)]. Bond building structural steel and all metal piping to an effective
ground-fault current path, per 250.104(D).
If you have multiple SDSs, you can ground the grounded neutral terminal (X0)
of each SDS to a common GEC. But the GEC and grounding (earthing)
electrode tap must comply with 250.30(A)(4)(a) through (c) (Fig. 5). Those
requirements can be summarized as follows:
 (a) Common GEC size. The common GEC must not be smaller than 3/0
AWG copper or 250 kcmil aluminum.
 (b) Tap conductor size. Size each grounding (earthing) electrode tap per
250.66 based on the largest separately derived ungrounded conductor of
the SDS.
 (c) Connections. Make grounding (earthing) electrode tap connections at an
accessible location with a listed connector or an exothermic weld. You can
also use listed connections to busbars not less than ¼ inch × 2 inch. If using
aluminum busbars, comply with 250.64(A).
Grounding (earthing) electrode. Terminate the GEC to a grounding
(earthing) electrode as close as possible to — and preferably in the same
area as — the system bonding jumper. The grounding (earthing) electrode
must be the nearest:
 Metal water pipe electrode, as specified in 250.52(A)(1), or
 Structural metal electrode, as specified in 250.52(A)(2).
If neither of these electrodes is available, use one of the following:
 A concrete-encased electrode encased in not less than 2 inches of
concrete, located within (and near the bottom of) a concrete foundation (or
footing) in direct contact with earth. This foundation must contain not less
than 20 feet of electrically conductive steel reinforcing bars (or rods) not
less than ½ inches in diameter [250.52(A)(3)].
 A ground ring encircling the building or structure, buried not less than 30
inches below grade. It must contain not less than 20 feet of bare copper
conductor not smaller than 2 AWG [250.52(A)(4) and 250.53(F)].
 A ground rod with not less than 8 feet of contact with the soil [250.52(A)(5)
and 250.53(G)].
 Other metal underground systems, piping systems, or underground tanks
[250.52(A)(7)].
Fig. 5. You can ground multiple SDSs to a common grounding electrode

conductor.
Observe the FPN in 250.30(7) to ensure quick removal of dangerous voltage
from a ground fault. Bond metal water piping and structural steel in the area
served by an SDS to the grounded neutral conductor at the SDS, per
250.104(D).
Grounded neutral conductor. If you install the system bonding jumper at the
secondary system disconnecting means instead of at the source of the SDS,
follow these requirements:
 Route the grounded neutral conductor with the secondary conductors, and
size it not smaller than specified in Table 250.66 based on the largest
ungrounded conductor for the SDS.
 If you install the secondary conductors in parallel, size the grounded neutral
secondary conductor in each raceway or cable based on the area of the
largest ungrounded secondary conductor in the raceway.
You can't size the grounded neutral secondary conductor smaller than 1/0
AWG (310.4). Remember, the grounded neutral conductor serves as the
effective ground-fault current path.
SDSs are common, so it's important to know — and correctly implement —
SDS requirements. A mistake in grounding and bonding can prevent a fault
from clearing, and it can allow dangerous potential to build on metal parts of
the electrical system. Your new knowledge of SDS grounding and bonding
requirements allows you to prevent both problems.
May 1, 2005 12:00 PM, By Mike Holt, NEC Consultant
Do you know the right way to ground and bond buildings or structures
supplied by feeders or branch circuits?
What little outbuilding was a badly-needed addition, but now you need to
supply power to it from the service of another building. You've sized the
disconnecting means (subpanel) and breakers correctly, but you're unsure
about the grounding. If you're smart, you'll ask yourself the following
questions:
 What are the grounding and bonding requirements for that subpanel?
 What are the grounding and bonding requirements for the building itself?
 Do you ground and bond the equipment in this building any differently
based on the fact that the building disconnect is powered from the service
of another building?
The benefits of proper bonding include clearing faults, preventing shock, and
reducing fires. Grounding provides a low-impedance path for lighting, and
properly grounded systems operate as intended. Improperly grounded
systems create fertile ground for power quality problems.
Fig. 1. Metal parts of electrical equipment in separate buildings or structures

supplied by the service of another building must be grounded to a suitable
electrode.
So how do you provide these benefits to a building or structure that doesn't
have its own service? The answer lies in complying with 250.32, and for that
purpose let's first look at the disconnecting means.
Disconnecting means. The purpose of grounding (earthing) the building or
structure disconnecting means to the earth is to limit elevated voltages on the
metal parts caused by lightning [250.4(A)(1)] (Fig. 1). Keep in mind that
grounding doesn't:
 Provide a low-impedance fault-current path to clear ground faults. In fact,
the Code prohibits the use of the earth as an effective ground-fault current
path since it's such a poor conductor of current [250.4(A)(5) and
250.45(B)(4)].
 Protect electrical or electronic equipment from lightning voltage transients.
You don't need a grounding (earthing) electrode where only one branch circuit
serves the building or structure [250.32(A)]. For the purposes of 250.32(A),
you can consider a multiwire branch circuit to be a single branch circuit.
Fig. 2. Proper bonding prevents shocks and reduces the likelihood of fires.
To quickly clear a ground fault and remove dangerous voltage from metal
parts, the building or structure disconnecting means must be bonded to an
effective ground-fault current path [250.4(A)(3)]. To establish this path when
your disconnecting means is supplied by the service of another building, your
installation must comply with either 250.32(B)(1) or 250.32(B)(2) (Fig. 2).
Equipment grounding (bonding) conductor. You can bond the building or
structure disconnecting means to an equipment-grounding (bonding)
conductor (as described in 250.118) that's installed with the circuit conductors.
Size this equipment-grounding (bonding) conductor per 250.122, based on
the rating of the feeder protection device [250.32(B)(1)].
Fig. 3. To prevent objectionable neutral current from flowing on metal parts, a
remote building disconnect must not be bonded to the grounded neutral
conductor.
Grounded conductor neutral. Be sure not to bond the grounded conductor
(neutral) to the disconnecting means or to the building because doing this
violates 250.6(A), and the resulting ground loop (parallel neutral current
paths) will allow dangerous objectionable current to flow onto metal parts of
the electrical installation (and onto metal piping and structural steel) (Fig. 3
and Fig. 4).
What if an equipment grounding (bonding) conductor isn't run to the building
or structure disconnecting means? In such a case, you can bond the
disconnecting means to the circuit grounded conductor (neutral). But this is
only permitted where there's no continuous metallic path between buildings
and structures and ground-fault protection of equipment isn't installed.
Fig. 4. Don’t create parallel neutral current paths during installation.

Neutral as effective ground path. Where the grounded conductor (neutral)
serves as the effective ground-fault current path, you must size it no smaller
than the larger of:
 The maximum unbalanced neutral load, per 220.61.
 The available fault current, per 250.122.
But maybe the size of this conductor should be your last concern because
using the grounded conductor (neutral) as the effective ground-fault current
path can have dangerous consequences.
Even if the initial installation doesn't result in dangerous objectionable current
on metal parts, the possibility remains for a future installation of metal piping
or cables between the buildings or structures that could create ground loops.
Thus, you should allow this only after careful consideration, and only as a last
resort.
The preferred practice is to not use the grounded conductor (neutral) as the
effective ground-fault current path. Instead, you should install an equipment
grounding (bonding) conductor with the feeder conductors to the building or
structure in accordance with 250.32(B)(1).
Grounding electrode conductor. The grounding (earthing) electrode
conductor (GEC) for a separate building or structure disconnecting means
must terminate to the grounding terminal of the disconnecting means. And
you must size it per Table 250.66, based on the largest ungrounded feeder
conductor.
Try the following practice question to reinforce these concepts. What size
grounding (earthing) electrode conductor is required for a building disconnect
that's supplied with 3/0 AWG?
The answer is 4 AWG. If you look at Table 250.66, you'll see this clearly. And
what if you supplied the disconnect with 500 kcmil cable? Then once again,
you'd look at Table 250.66. The answer, in this case, would be 1/0. However:
 Where the GEC is connected to a ground rod (or two, in a typical
installation), that portion of the conductor that's the sole connection to the
ground rod isn't required to be larger than 6 AWG copper [250.66(A)].
 Where the GEC is connected to a concrete-encased electrode, that portion
of the conductor that's the sole connection to the concrete-encased
electrode isn't required to be larger than 4 AWG copper [250.66(B)].
Lightning. With its vast power and potential for destruction, lightning
demands special attention. This is one reason for the existence of NFPA 780
“Standard for the Installation of Lightning Protection Systems.” You can sum
up Chapter 2 of the 2004 edition by simply saying, “Refer to the NEC.” Much
of what you need to know is in Art. 250. Another applicable standard is IEEE-
142 “Grounding of Industrial and Commercial Power Systems” — also known
as “the Green Book.”
Lightning is a high-frequency, multi-million volt electrical discharge that
obligates you to provide a path to the earth so it can dissipate there rather
than damage property or endanger people. To provide this path, you must
ground (earth) the building or structure disconnecting means to one of the
following electrodes [250.50 and 250.52(A)]:
 Metal underground water pipe [250.52(A)(1)]
 Metal frame of the building or structure electrode [250.52(A)(2)]
 Concrete-encased grounding (earthing) electrode [250.52(A)(3)]
 Ground ring [250.52(A)(4)]
Although such electrodes are usually available, that's not always the case. For
example, you might not have access to a metal underground water pipe
anywhere near your installation. Or you might need to do extensive damage
to finished surfaces to access one of these electrodes. So what are you
supposed to do? Fortunately, the NEC makes an allowance for this. Where
you don't have access to one of these other grounding (earthing) electrodes
— or they simply don't exist on your site — use a ground rod [250.52(A)(5)] or
a metal underground systems electrode [250.52(A)(7)]
Now you have answers to those questions you had when you were trying to
figure out the grounding and bonding for that little outbuilding. And you should
understand how to properly ground and bond any structure when the power
for that structure comes from the service of a separate building or structure.
The key to keeping a clear head about this concept is to think about the
purpose grounding and bonding serves, and what kind of a path you're
providing for the electricity. Remember that electrons that leave a power
supply are trying to get back to their source, not back to the earth. However,
lightning is trying to get to earth. Drawing the return path on paper will help
you see if your installation plan allows for what electricity is trying to do.
Ensuring your installation conforms to 250.32 will allow it to clear faults,
prevent shock, and provide a safe path for lightning.
Jun 1, 2005 12:00 PM, By Mike Holt, NEC Consultant
How solid is your understanding of grounding electrode systems?
NASCAR racers treat their tires with a care approaching reverence. Even the
most powerful engines are useless if the tires don't properly meet the track.
And large grounding conductors are useless if your grounding electrode
system doesn't properly meet the earth. In the first case, you start by selecting
the right tires. In the second, you start by selecting the right electrodes.
Electrodes you can use. A grounding electrode system consists of
grounding electrodes that are bonded together. When used in a grounding
electrode system, they're earthing electrodes and they make up an earthing
system. You can use the following electrodes for such an earthing system. If
more than one is present, they must be bonded together (Fig. 1.):
 Underground metal water pipe [250.52(A)(1)]
 Metal frame of the structure [250.52(A)(2)]
 Concrete-encased grounding electrode [250.52(A)(3)]
 Ground ring [250.52(A)(4)]
 Ground rod [250.52(A)(5)]
 Grounding plates [250.52(A)(6)]
If none of these electrodes is available, you must install one or more of the
following:
 Ground ring [250.52(A)(4)]
 Ground rod [250.52(A)(5)]
 Grounding plates [250.52(A)(6)]
 Metal underground systems [250.52(A)(7)]
Fig. 1. Connecting several electrodes improves the reliability of your
grounding electrode system.
Electrode requirements. Each type of electrode has a specific set of
requirements that you must follow.
Underground metal water pipe. When in direct contact with the earth for 10
feet or more, an underground metal water pipe can serve as a grounding
electrode (Fig. 2). If this pipe is interrupted by something, such as a water
meter, you must make it electrically continuous with a bonding jumper sized
per 250.66. If you use interior metal water piping to interconnect electrodes
that are part of the grounding electrode system, you can generally do so only
within 5 feet of where the piping enters the structure. See the exception in
250.52(A)(1).
Metal frame of the structure. If you use the structural frame as an electrode, it
must be of substantial cross-sectional area. Bond it to an electrode as defined
in 250.52(A)(1), (3), or (4) unless:
 10 feet or more of a single structural metal member is in direct contact with
the earth or encased in concrete that is in direct contact with the earth.
 The authority having jurisdiction (AHJ) approves some other means.
Concrete-encased grounding electrode (Ufer ground). Where the supply to a
structure consists of multiple services or feeders as permitted by 225.30 and
230.2, you must use the same electrode for grounding (earthing) enclosures
and equipment in or on that structure. The most practical method for meeting
this requirement is to ground (earth) each of the disconnecting means to a
common concrete-encased grounding electrode [250.52(A)(3)].
Fig. 2. Metal water pipe has been an accepted grounding electrode for more
than 50 years.
Electrically conductive steel reinforcing bars at least ½ inch in diameter or 4
AWG copper conductor can serve as a grounding electrode if the bar or
conductor is at least 20 feet long (conductive length), encased in at least 2
inches of concrete, and within and near the bottom of a foundation or footer
that is in direct contact with earth.
The steel rebar need not be one continuous length. You can conductively join
multiple sections with the usual steel tie wires to create a 20-foot, concrete-
encased grounding electrode (Fig. 3).
Ground ring. To qualify as a grounding electrode, a ground ring must contain
at least 20 feet of bare copper conductor that is at least 2 AWG. It must
encircle the structure and be in direct contact with earth at least 30 inches
below the surface [250.53(F)].
Fig. 3. The 2005 NEC requires a concrete-encased electrode in a grounding

electrode system for buildings with concrete footings or foundations in direct
contact with the earth for at least 20 feet.
Ground rod. A ground rod must have at least 8 feet of contact lengthwise with
the soil, which means it can't be less than 8 feet long [250.53(G)]. The upper
end of the ground rod must be at grade (or underground) unless you protect
the grounding electrode conductor attachment against physical damage per
250.10. See 250.52(A)(5) and 250.53(A) for details. If you encounter rock
bottom when driving the rod, you can drive it at an angle not to exceed 45°
from vertical, or bury it at least 30 inches deep.
Unlisted ground rods must be at least ⅝ inch in diameter. Listed ground rods
must be at least ½ inch in diameter. The diameter doesn't affect resistance to
ground (earth). See IEEE-142, Table 13 for reference. Larger diameters
increase mechanical strength (and service in the face of corrosion).
Fig. 4. It’s too dangerous to use metal underground piping of gas systems as
a grounding electrode. However, you must follow the bonding requirements of
250.104(B).
Ground plate. You can use a buried iron (or steel) plate at least ¼ inch thick
— or a nonferrous (copper) metal plate at least 0.06 inch thick — with an
exposed surface area at least 2 square feet. It must be at least 30 inches
below grade [250.52(A)(6)].
Metal underground systems. You can use such things as piping systems,
underground tanks, or underground metal well casings that aren't effectively
bonded to a metal water pipe.
All electrodes. Where practical, embed electrodes below permanent moisture
level. Make sure they're free from nonconductive coatings (e.g., paint or
enamel) [250.12 and 250.53(G)]. If more than one grounding electrode exists
at a structure, separate them by at least 6 feet. [250.53(B)]. Use fittings listed
for direct soil burial or concrete encasement if the grounding electrode
attachment fitting is underground or directly buried in concrete [250.68(A) Ex.
1 and 250.70].
Electrodes unsuitable for grounding. You can't use either of the following as
electrodes:
 Underground metal gas piping, which present a risk of explosion (Fig. 4 on
page 60).
 Aluminum electrodes, which are at risk of electrically-induced corrosion of
dissimilar metals [250.52(B)].
The jumpers that bond. If your grounding electrode system is within 18
inches of earth, use copper to bond the grounding electrodes together
[250.64(A)]. Size it per 250.66. Securely fasten the bonding jumper to the
surface on which it's carried, and protect it from physical damage [250.64(B)].
Fig. 5. The connection to metal water pipe must not depend on equipment
that may be removed.
Terminate grounding electrode bonding jumpers to the grounding electrode by
exothermic welding, listed lugs, listed pressure connectors, listed clamps, or
other listed means. When the termination is encased in concrete or buried,
use termination fittings that are listed and identified for this purpose (250.70).
The bonding connection to the interior metal water piping system (required by
250.104), must not depend on water meters, filtering devices, or similar
equipment that might be removed. Install a bonding jumper around removable
equipment to assist in clearing and removing dangerous voltage from a
ground fault to the metal water pipe [250.53(D)(1), 250.68(B) and 250.104]
(Fig. 5 above). You must supplement the underground metal water pipe
grounding electrode (if present) with one of the acceptable electrodes
discussed earlier.
Supplementary vs. supplemental. Don't confuse the requirements for
supplementary electrodes (250.54) with those for supplemental electrodes,
such as the underground metal water pipe [250.53(D)(2)].
A supplementary electrode establishes an additional electrical connection to
the earth but isn't part of the main electrical connection to the earth. An
example of a supplementary electrode is a ground rod installed next to a
machine tool. Technically, it serves no purpose, but some equipment
manufacturers who are unaware of grounding theory require them anyway.
You can't use a supplementary electrode as an effective ground-fault current
path for electrical equipment as required by 250.4(A)(5) and 250.4(B)(4). You
don't have to ground the supplementary electrode to the grounding electrode
system of the structure, and you don't have to size the grounding (earthing)
conductor for it per 250.66.
If the supplementary electrode is a ground rod, it doesn't have to comply with
the 25 ohm requirement of 250.56.
Lightning protection. If you use a grounding electrode for a lightning
protection system, you can't use it as a grounding electrode for the structure
grounding electrode (earthing) system required by 250.50.
The purpose of a lightning protection system installed per NFPA 780 is to
protect the structure from lightning damage. The concern that bonding the
lightning protection system to the building grounding system may create a
path for lightning into the electrical wiring or equipment within the structure
doesn't justify keeping the two systems separate; separation creates a
dangerous potential between them. You must bond your lightning protection
electrode system to the structure grounding electrode (earthing) system
(250.106).
Now you understand the types of grounding electrodes you can use — or in
some cases, must use — and some specific requirements that apply. You've
learned important requirements about the bonding jumpers used with these
electrodes. And you won't inadvertently confuse supplemental electrodes with
supplementary electrodes — you know an example of the former is a water
pipe and an example of the latter is a machine electrode and these have
different requirements.
With this newfound knowledge you can keep your grounding systems “on
track.”
Jul 1, 2005 12:00 PM, By Mike Holt, NEC Consultant
Errors in sizing and installing grounding electrode conductors can
cause your entire grounding system to fail
Part 6 of this series on grounding and bonding discussed grounding electrode
systems and compared them to the tires on race cars. This month's subject is
the grounding electrode conductor (GEC), which you can think of as the
suspension in a car. A properly tuned suspension allows the tires to do their
job and keep the car on the track. Similarly, the GEC connects the equipment
grounding conductor and the grounded (neutral) conductor to the grounding
electrode system.
The GEC performs this function at the service, at each building or structure
(where supplied by a feeder), or at the source of a separately derived system
(Art. 100). The GEC can be solid or stranded, insulated or bare. In most
cases, it must be copper, but you can use aluminum if the GEC isn't subjected
to corrosive conditions and isn't within 18 inches of the earth [250.64(A)].
The NEC doesn't require you to identify the GEC, but the generally accepted
practice is to apply green marking tape, strip off a few extra inches of
insulation (if used), or not identify the GEC at all.
Size matters. Using Table 250.66, size the GEC for the largest service-
entrance conductor (or equivalent area for parallel conductors), unless the
GEC connects to one of the three following electrode types:
 Ground rod — That portion of the GEC that's the sole connection to the
ground rod doesn't need to be larger than 6 AWG copper. Follow the
ground rod installation requirements contained in 250.52(A)(5).
 Concrete-encased grounding electrode (Ufer ground) — That portion of the
GEC that's the sole connection to the concrete-encased electrode doesn't
need to be larger than 4 AWG copper. Follow the concrete-encased
electrode installation requirements contained in 250.52(A)(3).
 Ground ring [250.52(A)(4)] — That portion of the GEC that's the sole
connection to the ground ring doesn't need to be larger than the conductor
used for the ground ring. A ground ring that encircles the building or
structure in direct contact with the earth must consist of no less than 20 feet
of bare copper conductor not smaller than 2 AWG.
So as you can see, sizing the GEC is simpler than it might seem. Prove this to
yourself by answering this question.
Fig. 1. The grounding electrode conductor is sized using Table 250.66, which
is based on the largest service-entrance conductor or equivalent area for
parallel conductors.
Q. What size GEC do you need for a 1,200A service supplied by three parallel
sets of 600 kcmil copper conductors per phase (Fig. 1)?
A. The equivalent area of three parallel 600 kcmil conductors is 1,800 kcmil.
So look in the far left column of Table 250.66 for the row that pertains to an
1,800 kcmil conductor. You can see that the last row, “Over 1100,” is the one
you want. The GEC size denoted by that row is 3/0 AWG.
Note that where a water pipe is used as the grounding electrode as shown in
Fig. 1, a supplemental electrode is required [250.53(D)(2)]. In this example,
the concrete-encased electrode is the supplemental electrode and 4 AWG is
the minimum bonding jumper required [250.66(B)].
Installing the GEC. If your GEC runs across a surface (or behind panels that
allow access), the NEC considers it “exposed” (Art. 100), and 250.64(B)
requires you to provide protection against physical damage. The first step in
providing protection is to securely fasten the GEC to the surface on which it
runs. Any additional steps will depend on its size. If the GEC is:
 4 AWG or larger, you must protect it from physical damage [250.64(B)].
 6 AWG and free from exposure to physical damage, you can run it along
the surface of the building construction without metal covering. If a 6 AWG
GEC is exposed to physical damage, you must run it in rigid metal conduit
(RMC), intermediate metal conduit (IMC), rigid nonmetallic conduit (RNC),
electrical metallic tubing (EMT), or cable armor.
 Smaller than 6 AWG, you must run it in RMC, IMC, RNC, EMT, or cable
armor.
Fig. 2. When terminating the grounding electrode conductor (GEC) to a
busbar, make connections exothermically or with listed connectors.
If you install the GEC in a ferrous metal raceway, you must make that
raceway electrically continuous by bonding each end of the raceway to the
GEC [250.64(E)].
Splices. You can run the GEC to any convenient grounding electrode
[250.64(F)]. But what if your GEC isn't long enough to reach? Can you splice
it? Yes, but only if you use one of the methods provided in 250.64(C):
 Use compression-type connectors listed for grounding, or exothermically
weld the splice.
 If needed, connect sections of busbars together to form a GEC.
 Terminate the GEC to a busbar sized not smaller than ¼ inch by 2 inches.
Fasten the GEC securely in place at an accessible location. Make
connections exothermically or with listed connectors (Fig. 2).
Taps. When a service consists of multiple disconnecting means as permitted
in 230.71(A), you can use a grounding electrode tap from each disconnect to
a common GEC [250.64(D)]. You must size the grounding electrode tap from
each disconnect to the largest ungrounded conductor serving that disconnect
per 250.66.
Fig. 3. Sizing the grounding electrode conductor for multiple connections can
be a little tricky. Just remember to size it based on the service conductors
feeding all disconnects.
Size the common GEC for the grounding electrode taps per 250.66. However,
take note that you must base its size on the service conductors that feed all
the service disconnects (Fig. 3).
Each grounding electrode tap must terminate to the common GEC in such a
manner that the common GEC will be free of splices and joints. In addition,
the grounding electrode tap can't be made within the service-disconnect
enclosure.
Enclosures. Each end of ferrous (iron/steel) raceways, boxes, and
enclosures that contain the GECs must be bonded to the GEC [250.92(A)(3)].
On the other hand, nonferrous metal raceways, such as aluminum RMC that
enclose the GEC, don't need to be bonded to the GEC.
Fig. 4. When bonding raceway to the grounding electrode conductor, you

must size it no smaller than the enclosed GEC.
If you use a bonding jumper with a bonding bushing or wedge [250.92(B)(4)]
to bond the raceway to the GEC, you must size the bonding jumper no
smaller than the enclosed GEC (Fig. 4).
The terminator. How should you terminate your GEC to grounding
electrodes? In all but the following two situations you must start by ensuring
the GEC attachment fitting remains accessible [250.68(A)]:
 The grounding electrode attachment fitting to an encased or buried
grounding electrode isn't required to be accessible.
 An exothermic or compression connection to fireproofed structural metal
isn't required to be accessible.
The second rule is to make the connection to the GEC in a manner that
ensures a permanent and effective grounding path. If you have an
underground metal water pipe electrode, you must install a bonding jumper
around insulated joints and equipment likely to be disconnected for repairs or
replacement [250.68(A)].
Continuity of the conductive bonding path for metal water piping as required
by 250.104(A) can't rely on water meters, filtering devices, or similar
equipment [250.53(D)(1)].
The GEC must terminate to the grounding electrode by exothermic welding,
listed lugs, listed pressure connectors, listed clamps, or other listed means. In
addition, termination fittings must be listed for the materials of the grounding
electrode. You can't terminate more than one conductor on a single clamp or
fitting unless the clamp or fitting is listed for multiple connections (250.70).
When the termination to a grounding electrode is encased in concrete or
buried, the termination fitting must be listed and identified for this purpose.
Ground fittings listed for concrete encasement or direct burial are typically
made from brass or bronze, not steel or aluminum alloy, so pay attention to
the color of the fittings to reduce errors in the field.
You're now well equipped to avoid mistakes in your GEC installations. You
should be able to ensure correct sizing, adequate protection, and proper
termination. Now your grounding system won't skid off the track because of
GEC errors.
Aug 1, 2005 12:00 PM, By Mike Holt, NEC Consultant
Don’t use the equipment grounding conductor for grounding
Would you be surprised to learn that the equipment grounding conductor
(EGC) doesn't connect anything directly to the earth (ground). Based on the
definition of EGC in Art. 100, the EGC connects the non-current carrying
metal parts of electrical systems to the system grounded conductor, the
grounding electrode conductor, or both. Thus, the EGC isn't really a grounding
conductor at all, but a bonding conductor.
According to Art. 100, bonding is the “permanent joining of metallic parts,” and
grounding is a connection “to earth.” It's an important distinction: If you're
joining metallic parts, you're bonding. If you're connecting to earth, you're
grounding.
This isn't just a matter of semantics. Understanding this difference will help
you avoid such mistakes as driving an “isolated ground” rod rather than
running a bonding jumper to the system grounded conductor, the grounding
electrode conductor, or both.
The EGC provides the low-impedance fault-current path that overcurrent
protection devices need to operate. It's those devices that remove dangerous
voltage potentials from between conductive parts of building components and
electrical systems [250.2 and 250.4(A)(3)].
14 flavors. Art. 250 provides a list of 14 types of components you can use
(singly or in combination) as an EGC (250.118). As you might expect, the first
item on the list is a conductor. This can be wire or busbar made of copper,
aluminum, or copper-clad aluminum. It can be bare or insulated.
Fig. 1. The combined length of conduit in the same fault return path can’t
exceed 6 feet.
The next three items are types of metallic nonflexible raceway: rigid metal
conduit (RMC), intermediate metal conduit (IMC), and electrical metallic
tubing (EMT). Chapter 3 in the NEC devotes an article to each of these: 342,
344, and 358, respectively. These raceways may outperform the conductors
within them as bonding paths, especially at high frequencies.
Varieties of metallic flexible raceway occupy the next four spots: flexible metal
conduit (FMC), liquidtight flexible metal conduit (LFMC), flexible metallic
tubing (FMT), and Type AC armored cable.
Listed flexible metal conduit — FMC installations must meet four requirements
[250.118(5)(a)-(d)]:
 The conduit terminates in fittings listed for grounding.
 The circuit conductors are protected by overcurrent devices rated 20A or
less.
 The combined length of the conduit in the same fault return path doesn't
exceed 6 feet (Fig. 1).
 The application requires flexibility.
Listed liquidtight flexible metal conduit — LFMC (Art. 350) installations must
meet five requirements [250.118(6)(a)-(e)]:
 The conduit terminates in fittings listed for grounding.
 For trade sizes ⅜ inch through ½ inch, the circuit conductors are protected
by overcurrent devices rated 20A or less.
 For trade sizes ¾ inch through 1¼ inch, the circuit conductors are protected
by overcurrent devices rated 60A or less.
 The combined length of the conduit in the same ground return path doesn't
exceed 6 feet.
 The application requires flexibility.
Flexible metallic tubing — To qualify as an EGC, FMT (Art. 360) must
terminate in fittings listed for grounding. The circuit conductors in the tubing
must be rated at 20A or less, and the FMC must be no more than 6 feet long
[250.118(7)(a)-(b)].
Type AC Cable — AC cable (Art. 320) is manufactured with an internal
bonding strip that's in direct contact with the interlocked metal armor. The
combination of the bonding strip and the interlocked metal armor makes the
cable suitable as an effective ground-fault current path.
MI cable — The ninth item on the list is the copper metal sheath of mineral
insulated (MI) cable (Art. 332). Take care with MI that you don't expose the
sheath to corrosion.
Type MC cable — Where listed and identified for grounding, MC cable is a
tenth option. But pay attention to whether it's interlocked, smooth, or
corrugated tube type:
 Interlocked Type MC must contain an EGC within the cable. The metal
armor of interlocked Type MC isn't suitable as an effective ground-fault
current path because — unlike Type AC — it doesn't have an internal
bonding strip.
 The sheath of smooth or corrugated tube Type MC is suitable as the
effective ground-fault current path. Thus, you can use it as an EGC without
a separate bonding conductor within the cable.
Metallic cable tray — You can use metallic cable tray as an EGC if continuous
maintenance and supervision ensure that qualified persons service it
[392.3(C)], and it meets two other requirements (392.7):
 Cable tray and fittings are identified for grounding.
 Cable tray, fittings, and raceways are bonded per 250.96 using bolted
mechanical connectors or bonding jumpers sized per 250.102.
Cablebus framework — You can use this if it meets the requirements of
370.3.
Other electrically continuous metal raceways — Items like metal wireways are
acceptable if they're listed for bonding.
Surface metal raceway — If it's listed for grounding, this is your fourteenth
option.
Fig. 2. Conductors with green insulation (or green with one or more yellow
stripes) can’t be used for an ungrounded or grounded neutral conductor.
Some ID, please. Individually covered or insulated EGCs must have a
continuous outer finish that's green (or green with one or more yellow stripes).
Conductors with green insulation (or green with one or more yellow stripes)
can't be used for an ungrounded conductor or grounded neutral conductor
(Fig. 2).
An insulated or covered EGC larger than 6 AWG may be permanently re-
identified at the time of installation at every point where the EGC is accessible
(250.119). But you don't have to identify EGCs larger than 6 AWG in conduit
bodies.
EGC identification must encircle the EGC. You can choose from three ways to
accomplish the identification:
 Stripping exposed insulation from the entire exposed length.
 Coloring the exposed insulation (or covering) green.
 Marking exposed insulation with green tape or green adhesive labels.
Sizing the EGC. Size wire EGCs per Table 250.122 based on the ampere
rating of the circuit-protection device. In no case does the EGC have to be
larger than the circuit conductors.
The EGC must be capable of safely conducting any ground-fault current likely
to be imposed on it (110.10 and Table 250.122). Otherwise, the EGC may
burn clear before the protective device responds.
Suppose you increase ungrounded conductor size to accommodate voltage
drop, excessive heating from harmonic currents, fault-current studies, or
future capacity. What about the EGC size? Any time you increase the size of
ungrounded circuit conductors — regardless of the reasons — you must
proportionately increase the size of the EGC [250.122(B)]. The following
problem illustrates this requirement.
Question: If you increase the size the ungrounded conductors for a 40A circuit
from 8 AWG to 6 AWG, you would also need to increase the size of your
EGC. If your EGC was 10 AWG, what size would it need to be now?
Answer: The circular mil area of 6 AWG is 59% greater than that of 8 AWG
(26,240 cmil/16,510 cmil) (Chapter 9, Table 8).
The EGC for a 40A protection device can be 10 AWG (10,380 cmil) (Table
250.122), but you must increase it in size by a multiplier of 1.59.
EGC size = 10,380 cmil × 1.59 = 16,504 cmil
Therefore, the new EGC must be 8 AWG, per Chapter 9, Table 8.
Fig. 3. Each raceway requires an equipment-grounding (bonding) conductor

based on the rating of the protection device.
When multiple circuits are installed in the same raceway or cable, you need
only one EGC in the raceway. Size it per Table 250.122 based on the largest
overcurrent device protecting the circuit conductors in the raceway.
When circuit conductors are run in parallel (310.4), you must install an EGC
with each parallel conductor set (310.4). Size it based on the ampere rating of
the circuit-protection device per Table 250.122 — unless ground-fault
protection is installed for cables systems (Fig. 3).
If ground-fault protection is installed, size the EGC based on the ampere
rating of the ground-fault protection per Table 250.122 where ground-fault
protection of equipment is installed. The installation must meet the three
following conditions, or you must size the EGC per the first method:
 Maintenance and supervision ensure only qualified persons will service the
installation.
 Ground-fault protection is set to trip at not more than the ampacity of a
single ungrounded conductor.
 The ground-fault protection is listed for protecting an EGC.
Fig. 4. When sizing feeder taps, the circuit protection ahead of the feeder will
provide guidance.
Size EGCs for feeders based on the ampere rating of the circuit-protection
device ahead of the feeder, per Table 250.122. In no case does the EGC
need to be larger than the circuit conductors (Fig. 4).
Don't let the “grounding” part of the EGC confuse you. It's really the
equipment bonding conductor. Don't think in terms of “grounding” equipment
— that is, connecting it to the earth (Art.100). Think in terms of connecting it to
the system grounded conductor, the grounding electrode conductor, or both).
Remember, the low-impedance fault-current path you create is necessary for
overcurrent protection devices to operate. Earth is not a low-impedance fault-
current path.
Sep 1, 2005 12:00 PM, By Mike Holt, NEC Consultant
Do you know when an isolated ground is appropriate and how to
properly install it?
Decisions on when and how to use an isolated ground (IG) involve design
issues you can't resolve based on the NEC alone [90.1(C)]. Before you try to
resolve those issues, you must understand what an IG is and isn't.
An IG isn't an arrangement whereby you drive a ground rod into the earth for
use as your sole grounding connection. In fact, doing so violates 250.4(A)(5),
which requires you to establish an effective ground-fault current path. The IG
concept probably arose from misunderstandings of neutrals and of the
differences between grounding and bonding — certainly, this “isolated ground
rod” idea did.
So what is an IG? To answer that, turn to IEEE-142, 5.5.54. The basic design
involves setting up a bonding system (for a given circuit or set of circuits) and
keeping it electrically isolated all the way to the service equipment or source
of a separately derived system.
Initially, designers also kept the grounding (earthing) system separate from
the source grounding (earthing) system. They would drive separate ground
rods for these systems and not bond them back to the source. They
mistakenly thought this would result in “separate” grounding systems — the
“isolated” one being “clean” compared to the “main grounding system.”
But such a practice would defy Ohm's Law, Kirchoff's Law, basic physics, and
several NEC references. Rather than solve a “dirty ground” problem, this
arrangement creates serious safety and operational problems. Thus, you
have no choice but to bond that “separate” ground rod system back to the
source after all. So an IG is actually a “separately insulated bonding conductor
to the power source” — not a separately isolated grounded one.
An IG system is sometimes designed with an independent counter-poise
ground (ground rods) that's bonded to the equipment grounding (bonding)
conductor. The NEC recognizes this independent grounding connection to the
earth as a “supplementary electrode.” According to Electric Power Research
Institute (EPRI) studies, a supplementary electrode is useless and its
presence could actually create a condition in which sensitive electronic
equipment could be damaged by lightning.
Because a supplementary electrode doesn't fulfill any of the NEC-required
functions, you don't have to bond it to the building grounding electrode
system, nor do you have to size it per 250.66 or make it comply with the 25-
ohm resistance requirement of 250.56 (250.54).
You can't use the supplementary electrode as the effective ground-fault
current path required by 250.4(A)(5) and 250.4(B)(4). By definition, this
supplements — but doesn't replace — NEC-required electrodes. Don't
confuse the requirements for supplementary electrode (250.54) with those for
the underground metal water pipe supplemental electrode [250.53(D)(2)].
An example of a supplementary electrode is a ground rod installed next to a
machine tool. Such an electrode serves no electrical purpose. Yet, some
equipment manufacturers require independent electrodes. They insist that
their equipment be electrically isolated from the structure's electrical system
[no equipment grounding (bonding) conductor].
Fig. 1. Despite evidence to the contrary, some equipment manufacturers still

believe a supplementary electrode is necessary to protect their equipment
from ground faults.
This practice violates 250.4(A)(5), which prohibits the use of the earth as an
effective ground-fault current path. If the metal enclosures of sensitive
electronic equipment were isolated or floated as required by some equipment
manufacturers, dangerous voltage on metal parts would remain from a ground
fault (Fig. 1).
You can isolate a metal raceway (containing circuit conductors for sensitive
electronic equipment) from the electrical equipment it supplies by using a
nonmetallic raceway fitting located at the equipment. However, the metal
raceway must contain an insulated equipment grounding (bonding) conductor
to provide the effective ground-fault current path to the power source (250.96)
(Fig. 2).
Fig. 2. If you use a nonmetallic fitting to isolate the metal raceway from a
piece of equipment, you must run an insulated grounding conductor between
it and the power source.
Some IG guidance. Many in the industry have their doubts about the IG's
effectiveness (“To IG or Not to IG?” on page 64). So what does all of this
mean for you if you think an IG is capable of solving your power quality
problems? For starters, there is no standard design you can adopt. However,
there are standards you can refer to for guidance on the basic principles and
requirements involved. Begin with the NEC and IEEE-142.
What about IG receptacles? By design, these devices have the grounding
terminal insulated from the metal mounting yoke. Therefore, you must connect
the grounding terminal of an IG receptacle to an insulated equipment
grounding (bonding) conductor that provides the effective ground-fault current
path to the power source winding (250.146).
IG receptacles must be identified by an orange triangle located on the face of
the receptacle [406.2(D)]. Sometimes the entire receptacle is orange, with the
triangle molded into the plastic face in a color other than orange. IG
receptacles installed in nonmetallic boxes must be covered with a nonmetallic
faceplate, because a metal faceplate can't be bonded to an effective ground-
fault current path [250.4(A)(3)].
Fig. 3. The use of spiral interlocked MC cable with a single equipment
grounding (bonding) conductor doesn’t meet the requirements of the Code.
IG receptacles require additional attention to wiring methods. For example,
the outer metal sheath of interlocked Type MC cable isn't listed as an
equipment grounding (bonding) conductor [250.118(10)]. Therefore, you can't
use this wiring method to supply an IG receptacle unless the cable contains
two equipment grounding (bonding) conductors. However, you can use
interlocked Type AC cable that contains a single insulated equipment
grounding (bonding) conductor, because the metal armor of the cable is listed
as an equipment grounding (bonding) conductor [250.118(8)] (Fig. 3 on page
63).
Now that you're familiar with what an IG is and what some of the requirements
are, you should be better prepared for taking steps to design and install one
should you decide you need one. Your first thought should be about whether
your installation conforms to the NEC and related standards. Before you
attempt to fix any problems by installing an IG, look very carefully at industry
standards, best practices, and conformance issues. If your installation is in
conformance, your second thought should be about how to correctly design,
install, and maintain that IG.
Sidebar: To IG or Not to IG?

The idea behind an IG system is that by bonding equipment with an insulated
equipment grounding (bonding) conductor directly to the power source, you
prevent contaminating equipment on one circuit with electrical noise from
another circuit. However, some research shows that may not always be the
case in practice.
IEEE 1100, Powering and Grounding Sensitive Electronic Equipment
(Emerald Book) states, “The results from the use of the IG method range from
no observable effects, the desired effects, or worse noise conditions than
when standard equipment bonding configurations are used to serve electronic
load equipment [8.5.3.2].”
Usually, you can prevent or solve noise problems simply by following best
practices and industry standards for electrical installations. However, if you
decide to go the IG route, first ensure your electrical infrastructure follows all
of the rules contained in Article 250 as well as Chapter 3 wiring methods. A
thorough review of your system against Chapter 2 of the NEC would probably
resolve any problems that remain. If not, various IEEE standards provide
more steps you should take before designing and installing an IG.
One reason for considering an IG is excess noise on a sensitive circuit. Such
noise is more likely to be of a higher amplitude on the current-carrying
conductors than on the ground circuit. So careful attention to wire separation
and routing will do far more for you than “isolating” bonding connections will
ever do.
Another reason for considering an IG is the idea you are going to “design out”
any chance of picking up noise through the ground (bonding) connection.
Because your IG must eventually tie into the grounding system, it's not really
isolated afterall. It's just “separately routed.” Is the idea that you're
accomplishing something by doing this just an illusion? Draw the circuits out,
and see what you think.
Except for some anecdotal accounts, scant evidence exists to suggest that an
IG cures any problems. In fact, as Chapter 5 of IEEE-142 points out, the IG
can make existing problems worse and create new ones.
But what about the various accounts of existing installations where problems
disappeared once an IG was installed? In many cases, these IG installations
are part of a larger bonding system repair project, so it's difficult to determine
exactly what fixed the problems. It's possible that IGs reduced symptoms at a
given facility, but as IEEE-1442 points out, IGs tend to mask problems rather
than fix them.
Of the IG systems that are properly designed, few are installed correctly and
even fewer are properly maintained. On top of everything else, engineering
opinions differ as to what constitutes a proper design.
Oct 1, 2005 12:00 PM, By Mike Holt, NEC Consultant
Does the NEC require you to ground piping systems and structural
steel? Be careful how you answer this question.
You've probably seen project requirements that call for the grounding of piping
systems and exposed structural steel. Those requirements, although well
intended, miss the mark. The stated intention of such requirements is nearly
always the removal of dangerous voltage on specific types of metal parts in
the event of a ground fault. These metal parts include exposed structural steel
members, electrically conductive metal water piping systems, metal sprinkler
piping, metal gas piping, and other metal piping systems. But these
requirements fail to make that intention a reality. That's because you remove
dangerous voltage on metal parts through bonding, not through grounding.
Fig. 1. You must size the metal water piping system bonding jumper in
accordance with Table 250.66.
Although you don't need to ground piping systems and structural steel, you do
need to bond them. And you need to bond them in a manner that establishes
an effective ground-fault current path [250.4(A)(4)].
Metal water piping systems. You don't need to bond isolated sections of
metal water piping connected to a nonmetallic water piping system. But where
you have metal water piping systems (rather than just isolated sections that
are metal), you must accomplish the bonding via one of three sets of
requirements. Which set you conform to depends on whether your installation
is a building supplied by a service, a building supplied by a feeder, or a
multiple occupancy building.
Building supplied by a service. You must bond the metal water piping system
to one of the following [250.104(A)(1)] (Fig. 1 above):
 Service equipment enclosure
 Grounded neutral service conductor
 Grounding electrode conductor, if sized per Table 250.66
 One of the electrodes of the grounding electrode system
Fig. 2. You must size the water piping system bonding jumper in accordance
with Table 250.66.
Size the metal water pipe bonding jumper per Table 250.66, based on the
largest ungrounded service conductor.
Here's an easy pop quiz for you. What size bonding jumper must you use for
the metal water piping system, if the service conductors are 4/0 AWG (Fig.
2)?
The correct answer is 2 AWG, as per Table 250.66.
Where hot and cold water pipes are electrically connected, you need only one
bonding jumper — to either the cold- or hot-water pipe. Otherwise, use a
single bonding jumper sized per 250.104(A)(1) to bond the hot- and cold-
water piping together.
Building or structure supplied by a feeder. You must bond the metal water
piping system of a building or structure supplied by a feeder to one of the
following:
 The equipment grounding terminal of the building disconnect enclosure
 The feeder equipment grounding (bonding) conductor
 One of the electrodes of the grounding electrode system
Size the metal water piping system-bonding jumper per Table 250.66, based
on the feeder circuit conductors that supply the building or structure. You don't
need to make this bonding jumper larger than the ungrounded feeder
conductors.
Multiple occupancy building. If the metal water piping systems in individual
occupancies do not mechanically connect to each other, you can bond the
metal water piping system (for each occupancy) to the equipment-grounding
terminal of the panelboard [250.104(A)(2)]. Size this bonding jumper per
Table 250.122, based on the ampere rating of the occupancy feeder
overcurrent protection device.
Other metal piping systems. You must bond metal piping systems (such as
gas or air) to an effective ground-fault current path, if they're likely to become
energized [250.104(B)]. According to the NFPA's National Fuel Gas Code, the
equipment-grounding (bonding) conductor for the circuit that may energize the
piping can serve as the bonding means.
Because the equipment-grounding (bonding) conductor for the circuit that may
energize the piping can serve as the bonding means, the NEC doesn't require
further bonding on the part of the electrical installer. Bonding of all metal
piping and metal ducts within the building provides an additional degree of
safety, but it isn't an NEC requirement [250.104(B) FPN].
Fig. 3. You must size the structural metal bonding jumper in accordance with
Table 250.66.
Structural metal. If exposed structural metal that forms a metal building
frame is likely to become energized, you must bond it to (Fig. 3) one of these
[250.104(C)]:
 Service equipment enclosure
 Grounded neutral service conductor
 Grounding electrode conductor, if sized per Table 250.66
 One or more of the electrodes of the grounding electrode system
This rule doesn't require you to bond sheet metal framing members (studs) or
the metal skin of a wood frame building, but doing so is a good practice. Size
the bonding jumper for the structural metal per Table 250.66, based on the
feeder or service conductors that supply the building (or structure). This
bonding jumper must be:
 Copper where within 18 inches of earth [250.64(A)].
 Securely fastened and adequately protected, if exposed to physical damage
[250.64(B)].
 Installed without a splice or joint, unless spliced by irreversible compression
connectors listed for the purpose or by the exothermic welding process
[250.64(C)].
Fig. 4. Pay careful attention to the bonding rules when working with a SDS.
Separately derived systems. In the area served by a separately derived
system (SDS), you must bond the nearest available point of the metal water
piping system to the grounded neutral terminal of the SDS. At the SDS end,
you must make this bond at the same location where the grounding electrode
conductor and system-bonding jumper terminate [250.104(D)(1)].
Size this metal water piping-bonding jumper per Table 250.66, based on the
largest ungrounded conductor of the SDS. You don't need a water pipe
bonding jumper, if you:
 Use the water pipe as the grounding electrode for the SDS, or
 Bond the metal water pipe to the structural metal building frame being used
as the grounding electrode for the SDS [250.104(D)(1) Ex 2] (Fig. 4).
Where exposed structural metal forms the building frame, you must bond it to
the grounded neutral conductor of each SDS. At the SDS end, you must make
this bond at the same location where the grounding electrode conductor and
system-bonding jumper terminate [250.32(A)]. Size each bonding jumper per
250.66, based on the largest ungrounded conductor of the SDS. You don't
need a structural metal bonding jumper, if you:
 Use the metal structural frame as the grounding electrode for the SDS, or
 Bond the structural metal frame to metal water piping being used as the
grounding electrode for the SDS.
Common grounding electrode conductor. Where you have installed a
common grounding electrode conductor for multiple SDSs — as permitted by
250.30(A)(4) — you must bond exposed structural metal and interior metal
piping (in the area served by the SDS) to the common grounding electrode
conductor [250.104(D)(3)]. But you don't have to install a separate bonding
jumper from each derived system to metal water piping (and to structural
metal members), if you have bonded the metal water piping and the structural
metal members (in the area served by the SDS) to the common grounding
electrode conductor.
Lightning protection system. Does the facility have a lightning protection
system? If yes, bond this system to the building (or structure) grounding
electrode system [250.106]. Do not use the grounding electrode for a lightning
protection system as the building (or structure) grounding electrode (250.60).
See NFPA 780, Standard for the Installation of Lightning Protection Systems,
for additional details on grounding and bonding requirements for lightning
protection.
Metal raceways, enclosures, frames, and other metal parts of electrical
equipment may require bonding or spacing from the lightning protection
conductors (per NFPA 780). Separation from lightning protection conductors
is typically 6 feet through air, or 3 feet through dense materials, such as
concrete, brick, or wood.
People often talk about grounding metal piping systems and structural steel.
The danger with such talk is it can mislead you into thinking you have met the
requirements of Art. 250 and several other standards, when you haven't.
Bonding deficiencies are notorious for permitting power-quality problems to
exist. But the worst part is they leave you with an unsafe facility. For safety
and performance, bond.
Nov 1, 2005 12:00 PM, By Mike Holt, NEC Consultant
Don’t let the bonding and grounding requirements of pools make you
feel like you’re in over your head
not more than 600V, including 120/240V, 120/208V, and 277/480V.
What comes to mind when you think of pools, spas, hot tubs, and similar
installations? Most people would say fun and relaxation. But the person trying
to properly ground and bond these installations is more likely to say confusion
and frustration. Why? Because of a phrase known as equipotential bonding,
which is a term some installers aren't familiar with.
To help us clear up the confusion, let's recap three basic concepts:
1. When you ground, you connect something to the earth. Think of this as
earthing.
2. When you bond, you create a conductive path between metallic objects.
3. Bonding is the method by which you create a low-impedance path for fault
current to flow.
The purpose of equipotential bonding is to bring metallic objects to the same
potential, thus reducing the shock hazard. This is obviously an important
consideration for people immersed in a container of water.
Fig. 1. Size the insulated copper equipment grounding conductor per Table
250.122.
The intention of equipotential bonding is to reduce earth voltage gradients in
the area around a permanently installed pool or similar installation by the use
of a common bonding grid per 680.26(B) and (C).
Equipotential bonding is not intended to provide a low-impedance ground-fault
current path to the source (which would assist in clearing a ground fault) as
required by 250.4(A)(3). It also has nothing to do with creating a path for fault
current. Therefore, the 8 AWG or larger solid copper equipotential bonding
conductor required by 680.26(C) isn't required to extend to (or attach to) any
panelboard, service equipment, or electrode.
Fig. 2. The junction box serving the luminaire must be listed as a swimming
pool junction box.
In addition to equipotential bonding, the Code requires any wiring method
used for equipment associated with a pool, spa, or hot tub to include an
insulated copper equipment grounding (bonding) conductor. You must size
this conductor per 250.122, but it can't be smaller than 12 AWG copper.
Various other requirements apply, depending on the type of raceway you use.
These requirements all work toward providing an effective ground-fault current
path.
Pool lights. When wiring luminaires in a pool, the goal is to light up the pool,
not its occupants. This is why branch-circuit conductors for an underwater
luminaire must contain an insulated copper equipment grounding (bonding)
conductor sized per Table 250.122. In no case can this conductor be smaller
than 12 AWG (Fig. 1).
You must not splice the equipment grounding (bonding) conductor for the
underwater luminaire [680.23(F)(2)], unless one of two conditions exists (Fig.
2 on page 46):
1. More than one underwater luminaire is supplied by the same branch circuit.
If so, the equipment grounding (bonding) conductor can terminate at a listed
pool junction box that meets the requirements of 680.24(A).
2. The equipment grounding (bonding) conductor terminates at the grounding
terminal of a listed pool transformer, GFCI, clock switch, or manual snap
switch that is located between the panelboard and a junction box connected
to the conduit that extends directly to the underwater luminaire.
The branch-circuit conductors for the underwater luminaire must not occupy
raceways, boxes, or enclosures containing other conductors on the load side
of a GFCI or transformer — unless one of the following conditions exists:
1. The other conductors are GFCI protected.
2. The other conductors are grounding (bonding) conductors.
3. The other conductors supply a feed-through type GFCI.
4. The other conductors are in a panelboard.
Junction boxes. The junction box (deck box) that connects directly to an
underwater permanently installed pool, outdoor spa, or outdoor hot tub
luminaire forming shell must be listed as a swimming pool junction box and
equipped with threaded entries or a nonmetallic hub [680.24]. It also must be
constructed of copper, brass, or corrosion-resistant material approved by the
authority having jurisdiction (AHJ). You must provide electrical continuity
between all metal conduit and the grounding (bonding) terminals within the
junction box.
Fig. 3. The steel tie-wires securing the rebar are suitable for bonding the
reinforcing steel.
You also must ensure the junction box has at least one more grounding
(bonding) terminal than the number of conduit entries [680.24(D)]. Typically,
there are four grounding (bonding) terminals in the junction box and three
conduit entries. Finally, you must provide a strain relief at the enclosure where
you terminate the flexible cord of an underwater luminaire [680.24(E)].
Bonding metal parts to the equipotential bonding grid. Ensure the
following five parts of a permanently installed pool, outdoor spa, or outdoor
hot tub are bonded together and to the equipotential bonding grid.
1. All metallic parts of reinforcing metal not encapsulated with a nonconductive
compound. The usual steel tie-wires that secure rebar together are
considered suitable for bonding the reinforcing steel together (Fig. 3 on
page 46). Where conductive reinforcing steel of the permanently installed
pool, outdoor spa, or outdoor hot tub shell and deck is not available, you
must provide an alternative means in accordance with 680.26(C) to
eliminate voltage gradients that would otherwise be provided by
unencapsulated, bonded reinforcing steel.
2. All metal forming shells for underwater wet-niche luminaires.
3. Metal fittings within or attached to the permanently installed pool, outdoor
spa, or outdoor hot tub structure, such as ladders and handrails.
4. Metal parts of electrical equipment associated with the circulating system
(water heaters and pump motors) and metal parts of equipment associated
with pool covers.
5. Metal cables, metal raceways, metal piping, and all fixed metal parts of
electrical equipment — except those separated from the pool by a
permanent barrier located within:
 5 feet horizontally of the inside walls of the permanently installed pool,
outdoor spa, or outdoor hot tub.
 12 feet above the maximum water level of the permanently installed pool,
outdoor spa, or outdoor hot tub; or any observation stands, towers,
platforms, or diving structures.
The equipotential bonding grid. All metal parts specified in 680.26(B) must
be bonded to an equipotential bonding grid with a solid copper conductor not
smaller than 8 AWG. The termination of the bonding conductor must be made
by exothermic welding, listed pressure connectors, or listed clamps that are
labeled as suitable for the purpose. An equipotential bonding grid must extend
under paved walking surfaces for 3 feet horizontally from the water
[680.26(C)].
The equipotential bonding grid must be formed from either or both of:
1. The structural reinforcing steel of a concrete permanently installed pool,
outdoor spa, or outdoor hot tub, tied together by the usual steel tie wires.
2. The metal walls of a permanently installed pool, outdoor spa, or outdoor hot
tub.
What if neither of these is available? Where structural reinforcing steel or the
walls of bolted or welded metal permanently installed pool, outdoor spa, or
outdoor hot tub structures are not available, you must construct an
equipotential bonding grid as follows:
(a) The equipotential grid can be constructed with 8 AWG bare solid copper
conductors that are bonded to each other at all points of crossing by
exothermic welding, listed pressure connectors of the set screw or
compression type, listed clamps, or other listed fittings [250.8].
(b) The equipotential bonding grid must cover the contour of the permanently
installed pool, outdoor spa, or outdoor hot tub, and deck extending 3 feet
horizontally from the water. The equipotential bonding grid must be arranged
in a 1-foot by 1-foot network of 8 AWG conductors, with a tolerance of 4
inches.
All equipotential bonding terminations must be by exothermic welding, listed
pressure connectors of the set screw or compression type, listed clamps, or
other listed fittings [250.8].
Dispelling common myths. Let's go back for a moment to a frequently
misapplied term: grounding. One of the myths about grounding (earthing) is
that it reduces shock hazards by bringing everything to ground potential.
Because the earth is not of uniform conductivity, that can hardly be the case.
Another myth is that grounding (earthing) provides a common reference point.
This also defies logic.
What about that common reference? If ground (earth) is not it, what is? You
can establish a common reference by constructing a bonding grid, not an
earthing connection. Connect your bonding path to the grounding grid, and all
of the objects are at an equal potential.
When you have equal potential, then by definition you do not have a voltage
difference. Without a voltage difference, there can be no current flow. Thus,
you use an equipotential bonding system — not a grounding system — to
reduce shock hazards.
Dec 1, 2005 12:00 PM, By Mike Holt, NEC Consultant
How do you make communications systems safe?
not more than 600V, including 120/240V, 120/208V, and 277/480V.
It's late Friday afternoon, and you're looking forward to the weekend,
imagining what you'll be doing in just about an hour. Then suddenly the
ringing of your cell phone jars you back to reality. After a brief conversation,
you know you can forget about your weekend. If smoke pouring out of the
server room wasn't bad enough, the news that Frank is on his way to the
hospital after simply touching a communication cable has your stomach in
knots.
Although no one wants to be stuck in this kind of situation, many facilities,
including office, home, and residential, contain NEC Chapter 8 violations that
could result in this exact scenario — or worse.
NEC Chapter 8. Some years ago, Sprint changed installation requirements to
mandate bonding its ground rod to the main bonding jumper. (Randy
Schmisny, former chair of the IEEE Kansas City Section, helped bring this
change about.) Sprint's previous stance was that its ground rod should be
“separate from” the power ground rod.
This new policy made its installations comply with Art. 250 and Chapter 8. The
impetus wasn't pressure to comply with the NEC, but rather the need to
prevent equipment failures and reduce service calls. Bonding its electrode to
the rest of the system provided cost savings through increased reliability,
performance, and safety. Comply with NEC Chapter 8, and you can enjoy
those same benefits.
NEC Chapter 8 contains the requirements for communications circuits (Art.
800), Radio and Television Equipment (Art. 810), Community Antenna
Television (CATV) and Radio Distribution Systems (Art. 820), and Network-
Powered Broadband (Art. 830). From a grounding and bonding standpoint,
these Articles have the same goals — the primary one being to eliminate
differences in potential. The grounding and bonding rules in these Articles
often differ in the details. For example, you'll find differences in the minimum
conductor size and whether or not you need an insulated grounding
conductor.
Fig. 1. Art. 800 covers the installation requirements for communications wiring
and equipment on voice, audio, video, data, and interactive services and
computer networks.
Where we get twisted. The telco typically provides the twisted-pair cable to a
terminal board at the structure. This terminal board is the network interface
device (NID). Art. 800 addresses twisted-pair wiring from the NID to the
premises (Fig. 1). We find this kind of wiring in such central station systems
as fire and burglar alarm, telephone, and telegraph.
A primary protector is a device that protects installers and occupants from
electric shock. You need a listed primary protector for each incoming
communications circuit [800.90(A)]. The point of entrance is where the cable
enters the structure or grounded raceway. Install the primary protector as
close as practicable to the point of entrance [800.90(B)] to reduce differences
in potential between communications circuits and other metallic systems and
objects. Those differences create personnel hazards and can lead to
catastrophic failures — especially when lightning is present.
If you're installing phone cable, ground the metallic sheath (or interrupt it by
an insulating joint) as close as practicable to the point of entrance [800.100].
This rule applies to other types of cable covered by Arts. 810, 820, and 830.
Four other rules also apply universally:
 Keep grounding wires as short as practicable, and run them in a straight
line. Why does this matter? Lightning tends not to travel through sharp
bends, corners, and loops. Instead, it tends to jump across them or flash
over to something nearby.
 As common sense suggests, you must protect the grounding conductor
where it is subject to physical damage — and that typically means you run it
in a raceway. If you use a metal raceway, bond each end of the raceway to
the grounding conductor.
 Ground cables and metallic raceways as close as practicable to the
entrance point.
 Use only grounding conductors, connectors, and fittings listed as suitable
for the purpose.
Grounding conductor. Art. 800 requirements for grounding the phone cable
and the primary protector are nearly identical to those of the other Chapter 8
Articles. The grounding conductor [800.100(A)] must be:
 Insulated and listed as suitable for the purpose (Art. 810 does not require
insulation).
 Copper or other corrosion-resistant conductive material (Art. 810 has more
stringent requirements).
 Not smaller than 14 AWG (this differs in the other Articles).
 As short as practicable. In 100(A), Arts. 800, 820, and 830 contain a length
limit of 20 feet for dwelling units. Art. 810 does not contain a length limit.
Sometimes, it isn't practicable to limit the grounding conductor to 20 feet Note
the exception in 100(A)(4) of the appropriate Article. The details differ slightly,
but essentially you can drive a separate ground rod that you must bond back
to the grounding conductor.
Fig. 2. The NEC requirement to bond all external systems to a single point
helps minimize the possibility of equipment damage and electric shock.
Though the details differ among the Articles, the requirement to bond all
external (entering a structure) systems (e.g., communications and power) to a
single point remains. This practice minimizes the possibility of equipment
damage — and electric shock — due to differences of potential between the
systems (Fig. 2). If you don't make this bond, your system will be at risk for
flashover, ground loops, power quality problems, and circulating currents. This
is why someone can die from shock by merely touching the shield of the
network cable on the back of a printer — even though all systems are properly
“grounded.” Maybe size doesn't matter, but bonding does.
Communications electrodes. If the structure has a grounding means, you
have several grounding conductor termination options. Terminate to the
nearest accessible point of the following locations:
 Grounding electrode system [250.50].
 Interior metal water piping system, within 5 feet from point of entrance
[250.52(A)(1)].
 Service bonding means [250.94].
 Metallic service raceway.
 Service equipment enclosure, or
 Grounding electrode conductor (or GEC metal enclosure).
In the rare case that the structure lacks a grounding means, install a ground
rod not less than 5 feet long and ½ in. in diameter [800.100(B)(2)(2),
830.100(B)(2)(2)]. For 810 and 820 installations, use a 10-foot rod per 250.52,
or bond to the grounded structure. Bond this to the grounding electrode
system with a minimum 6 AWG conductor.
Article 810. This Article differs markedly from the other Chapter 8 Articles;
however, it still requires you to reduce differences in potential. The difference
is in the details of how you do that.
For example, you must provide a listed antenna discharge unit for each lead-
in conductor from an outdoor antenna [810.20]. It doesn't matter if you locate
the discharge unit inside or outside, but you must locate it nearest the point of
entrance — and away from combustible material. If the antenna is indoors
(e.g., in an attic), you can skip the discharge unit.
Fig. 3. Grounding the antenna mast and discharge unit per 810.21 helps
prevent voltage surges caused by static discharge or nearby lightning strikes.
Ground the antenna mast and discharge unit per 810.21 (A) through (K) (Fig.
3). This grounding helps prevent voltage surges caused by static discharge or
nearby lightning strikes from reaching the center conductor of the lead-in
coaxial cable.
Satellite dishes provide additional challenges. Because the dish sits outdoors,
wind creates a static charge on the antenna and attached cable. This charge
can build up until it jumps across an air space — often passing through the
electronics inside the low noise block down converter feed horn (LNBF) or
receiver.
Manufacturers often mold copper-clad steel or bronze wire (17 AWG) into the
jacket of the coaxial cable to eliminate the need for a separate ground wire —
which simplifies grounding the satellite dish [810.21(F)(1)].
Avoiding confusion. To avoid Chapter 8 confusion, remember:
 Grounding and bonding requirements are in the same place in each Article
(except for Art. 810, which has its own structure).
 Make your grounding run short and straight.
 Focus on eliminating differences in potential.
 Use the Article for your specific installation.
Any time you work on a communications system, turn to Chapter 8. Bond to
equalize potential, but ground to protect from lightning.
This is the final article in this series. If you've read them all, you now
understand the difference between grounding and bonding — and the
purpose of one vs. the other. By not confusing the two, you can eliminate
some common causes of unsafe installations.

Grounding Vs Bonding Part (1-12) of 12

Enviado por

Dados do documento

Título original

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

Grounding Vs Bonding Part (1-12) of 12

Enviado por

Direitos autorais:

Formatos disponíveis

Grounding vs Bonding Part 1 of 12

Jan 1, 2005 12:00 PM, By Mike Holt, NEC Consultant

Fig. 1. Specific bonding requirements can be found in Part V of Art. 250.

Fig. 4. The absence of an effective ground-fault current path creates a lethal

1. Grounded systems [250.4(A)]

Fig. 1. The earth shall not be considered an effective ground-fault current

Fig. 2. Bonding electrically conductive materials to each other and to the

Sidebar: Lightning and Surges

Fig. 1. Make your grounding electrode conductor connection at one of these

Fig. 2. A grounded neutral conductor must be run to and bonded to each

Fig. 1. A generator is considered to be a separately derived system if it has

Fig. 5. You can ground multiple SDSs to a common grounding electrode

Fig. 1. Metal parts of electrical equipment in separate buildings or structures

Fig. 4. Don’t create parallel neutral current paths during installation.

Fig. 3. The 2005 NEC requires a concrete-encased electrode in a grounding

Fig. 4. When bonding raceway to the grounding electrode conductor, you

Fig. 3. Each raceway requires an equipment-grounding (bonding) conductor

Fig. 1. Despite evidence to the contrary, some equipment manufacturers still

Sidebar: To IG or Not to IG?

Você também pode gostar