Photovoltaic Power Systems 2017

Photovoltaic Power Systems
For Inspectors, Plan Reviewers

& PV Professionals
BASED ON THE 2017 NATIONAL ELECTRICAL CODE

About the author
John Wiles is perhaps the most recognized name in the solar industry
for his numerous contributions to the development of codes and National
Electrical Code compliance for photovoltaic systems. He has written hundreds
of articles on Code-related photovoltaic system topics and is a regular con-
tributor to IAEI News.
Wiles retired from his full-time position as a research engineer at the
Southwest Technology Development Institute at New Mexico State Uni-
versity in 2013 after 24 years in the position. He continues to volunteer his
time to keep active in the codes- and standards-development processes; to
assist inspectors and plan reviewers with Code questions; to make “PV and
the National Electrical Code” presentations throughout the country; and to
consult with the PV industry on code-related issues.
Wiles is a member of several Standards Technical Panels for Underwriters
Laboratories and is active in formulating standards for PV equipment, such
as modules, inverters, charge controllers, combiners, cables, racks, and con-
nectors. He continues to write the “Perspectives on PV” articles in the IAEI
News and is active in the development of proposals for the 2020 NEC.
As an old solar pioneer, he lived for 16 years in a stand-alone, off-grid,
PV-powered home in suburbia. His new owner-designed and -built retire-
ment home has a 7.5 kW utility-interactive PV system with whole-house
battery backup, where he lives with his wife Patti, three dogs, and two cats.
& PV Professionals
THIRD EDITION
JOHN WILES
4
Published 2012, 2014, 2018 by

International Association of Electrical Inspectors
901 Waterfall Way, Suite 602
Richardson, TX 75080-7702
Copyright © 2012, 2014, 2018 by John Wiles
All rights reserved. First Edition published June 2012.

Printed in the United States of America
22 21 20 19 18 5 4 3 2 1
ISBN-10: 1-890659-81-3
ISBN-13: 978-1-890659-81-3
Photos used in this book were shot in situ or at tradeshows.

Use of the photos does not imply endorsement by IAEI of
the manufacturers or the products.
The material in this book has been extracted from and expanded upon the series of articles “Perspectives on PV”
found in IAEI News published by the International Association of Electrical Inspectors. The articles are based on the
author’s understanding of the 2005, 2008, 2011, 2014, and 2017 NFPA 70 National Electrical Code (NEC) 1; his
activities in developing that Code; his design reviews, inspections and testing of photovoltaic (PV) systems for more
than twenty years; and his interaction with electrical inspectors, PV systems designers, and PV installers throughout
the country. In all cases, the NEC is the requirement and local authorities having jurisdiction provide the interpreta-
tions of the Code.
DISCLAIMER
This book provides information on how the 2005, 2008, 2011, 2014, and 2017 National Electrical Codes apply to pho-
tovoltaic systems. The book is not intended to supplant or replace the NEC; it paraphrases the NEC where it pertains
to photovoltaic systems and should be used with the full text of the NEC. Users of this book should be thoroughly
familiar with the NEC and know the engineering principles and hazards associated with electrical and photovoltaic
power systems. The information in this book is the best available at the time of publication and is believed to be
technically accurate. Application of this information and results obtained are the responsibility of the user.
In most locations, all electrical wiring (including photovoltaic power systems) must be accomplished by, or under
the supervision of a licensed electrician and then inspected by a designated local authority. Some municipalities have
additional codes that supplement or replace the NEC. The local inspector has the final say on what is acceptable.
This book has not been processed in accordance with the National Fire Protection Association’s (NFPA) Reg-
ulations Governing Committee Projects. Therefore, the text and commentary in it shall not be considered the official
position of the NFPA or any of its committees and shall not be considered to be nor relied upon as a formal interpre-
tation of the meaning or intent of any specific provision or provisions of the 2005, 2008, 2011, 2014, or 2017 editions
of National Electrical Code.
Author and the publisher do not warrant or guarantee any of the products described herein nor have they per-
formed any independent analysis in connection with any of the product information contained herein. The publisher
does not assume, and expressly disclaims, any obligation to obtain and include information referenced in this work.
The reader is expressly warned to consider carefully and adopt all safety precautions that might be indicated by the
activities described herein and to avoid all potential hazards. By following the instructions contained herein, the reader
willingly assumes all risks in connection with such instructions.
THE AUTHOR AND THE PUBLISHER MAKE NO REPRESENTATIONS OR WARRANTIES OF
ANY KIND, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF FITNESS FOR
PARTICULAR PURPOSE, MERCHANTABILITY, OR NON-INFRINGEMENT, NOR ARE ANY SUCH
REPRESENTATIONS IMPLIED WITH RESPECT TO SUCH MATERIAL. THE AUTHOR AND THE
PUBLISHER SHALL NOT BE LIABLE FOR ANY SPECIAL, INCIDENTAL, CONSEQUENTIAL OR
EXEMPLARY DAMAGES RESULTING, IN WHOLE OR IN PART, FROM THE READER’S USE OF OR
RELIANCE UPON THIS MATERIAL.
1 National Electrical Code and NEC are registered trademarks of the National Fire Protection Association, Inc.,
Quincy, MA 02169
5
Table of Contents
1 An Overview of PV Systems and the 2017 National Electrical Code 8
2 PV Fundamentals and Calculations 42
3 PV Modules — Installation Considerations 65
4 The Inverter — Operation and Connections 81
5 Energy Storage Systems (ESS), Batteries in PV Systems 97
6 Grounding, Disconnects, and Overcurrent Protection 113
7 Utility Interconnections 133
8 Plan Checking and Inspecting 156
9 The Process of Inspecting PV Systems 168
10 The 15-Minute PV Inspection – Can You? Should You? 178
Appendix A 2014 and 2017 NEC Photovoltaic Electrical Power Systems

Inspector/Installer Checklist 184
Appendix B Bibliography 191

6
7
Preface
In a time where photovoltaic plan reviewers and inspectors are getting pressured to expedite the
inspection and review process, this book could not have come at a better time. If we are expected
to accomplish faster turnaround times with fewer inspections, we must be informed to ensure safe,
Code-compliant installations. Rubber-stamping plans and drive-by inspections may be what the
industry is pushing for, but are those actions what the customer deserves? The customer is relying on a
qualified inspector to verify that their PV system is safe and that it will continue to be safe for years of
operation.
What makes this book stand out is how it correlates to the National Electrical Code (NEC). When
citing corrections or comments, we need to be able to reference the Code to justify our calls. The last
thing that we should be doing is trying to enforce our opinion.
Stamping a set of PV drawings for approval or signing a permit card for an inspection does not
require skill or knowledge of the NEC requirements we need. The knowledge and skill before we sign
or stamp documents is where most of us need some help and guidance. Having a document such as
Photovoltaic Power Systems provides inspectors with a great tool for gathering information on what to
look for in plan review and during inspections.
Article 690 is a small section when compared to the entire NEC. The size of Article 690 does not
make it any less important than other articles found in the NEC. Due to its size, it is often not a focus
of the combination inspector. It is no wonder it gets overlooked when you stack up all the codes the
combination inspector must enforce.
When the NEC book and handbook are not enough to help you understand what or how you
should be enforcing the regulations, this document can provide clarity. The information provided in
Photovoltaic Power Systems has been compiled by someone who is known throughout the industry as a
PV expert. John Wiles has been a resource and has been providing training for more than 20 years to
inspectors and plan reviewers.
Not only is this book originally based on the 2011 Code and earlier editions, but also the 2014 and
2017 NEC. With knowledge comes credibility. This document will help plan reviewers and inspectors
know and understand what they are looking at and what to look for. If you want to understand what
is on the plans, this book will help. Even for those who have a solid understanding of PV systems, it is
helpful to have a book to refer to when questions arise. Whether questions come from us or from PV
designers or installers, this book can help answer these PV-related questions. You will find this book to
be an excellent resource.
Having ready access to this book can help us all be more informed about PV systems, where expertise
is often limited. John Wiles is known for having such expertise.
— Rhonda Parkhurst, Electrical Specialist

City of Palo Alto, California
8 Chapter 1 An Overview of PV Systems and the 2017 National Electrical Code
Chapter 1 An Overview of PV Systems and the 2017 National Electrical Code 9
01
An Overview of PV Systems and
the 2017 National Electrical Code
Photovoltaic (PV) power systems are being amps or more. These levels of voltage and current,
installed by the tens of thousands throughout the if not properly managed, pose shock, life safety,
United States. In states where financial incen- and fire hazards. These systems must be inspected
tives are available (like in California, New York, to ensure the safety of owners, operators, service
and New Jersey), the PV business is booming. personnel, and the public.
The first PV cells produced more than 50 years The Code requirements for a typical residential
ago are still producing power, and modern PV PV system are at least as complex as those for
modules are expected to produce energy for residential wiring, and the dc portions of the
the next 40 years or longer. The power output system coupled with the ac interconnection to the
from PV systems ranges from a few hundred utility grid make PV installations unique. Because
watts to many megawatts. Most of the systems the PV industry is growing rapidly, individuals,
are not operated or owned by any electric utility companies, and organizations (with varying
and therefore come under the requirements of degrees of knowledge, skill, and experience) are
NFPA 70 National Electrical Code (NEC). Unless installing these systems. Large, and some small,
otherwise noted, all references to the NEC will be PV-system integrators and vendors — working
to the 2017 NEC. with experienced electrical contractors who have
Systems as large as 700 megawatts have been jointly pursued additional PV-specific training
installed by third parties on private land in the and who work closely with the local permitting
United States, and are not under utility control and inspecting authorities — usually (but not
or ownership. Larger systems of up to 1500 always) perform the best, most Code-compliant
megawatts have been installed in other countries installations.
and will more than likely be installed in the On the other hand, individuals or organizations
United States in coming years. These systems with little or no experience or training installing
operate at 1000 volts to 1500 volts and, in the electrical systems of any type are installing many
larger commercial systems, direct current (dc) new PV systems. These systems may be unsafe
and alternating current (ac) can range up to 2000 (not Code-compliant) at initial installation;
develop hazardous conditions over the life of the training, experience, and skill requirements of PV
system; be hazardous to operate or service; and designers and installers obtaining this certification
fail to deliver the full performance of a well-de- will help ensure that safer, higher quality installa-
signed and -installed PV system. tions of PV systems take place.
The electrical inspector or plan reviewer, as the
authority having jurisdiction (AHJ), is the key PV System Types
player in ensuring that these less-than-ideal PV Two main types of PV systems are being in-
installations do not proliferate. Inspectors need stalled in the United States: utility-interactive
to demand additional training in the inspection (grid-connected) (see photo 1.1) and stand-alone
of PV systems and then inspect these systems (off grid) (photo 1.2). Both types use PV mod-
very closely. Yes, PV is a relatively unfamiliar ules connected in series and in parallel to form
technology, but 80% of the Code already familiar PV arrays that produce direct current energy at
to inspectors applies, and it is relatively easy to voltages ranging from approximately 12 volts to
learn the inspection requirements that are unique 1500 volts (photos 1.1, 1.2, 1.3 and 1.4). Refer
to PV systems. to Article 100 and Section 690.2 of the NEC
Several organizations in the PV industry pro- for definitions of the terms used to describe PV
vide training and certification for individuals. The equipment and systems. These systems will be
Photo 1.1 • Carport PV systems generate energy and keep Photo 1.3 • Commercial PV array mounted horizontally
cars cool. with some shading.
Photo 1.4 • Five mW utility-scale PV system with PV subar-

Photo 1.2 • 3.3 kW stand-alone, off-grid PV system. rays mounted on two-axis trackers.
defined and further described in the following Utility-Interactive Systems

chapters Utility-interactive (U-I) PV systems are by far
Generally, energy storage batteries are found in the most numerous type of PV system current-
stand-alone systems but are not normally found ly being installed. A typical residential system
in utility-interactive systems. Variations of each might have a PV array and an inverter (which
system are possible with some utility-interactive converts dc to ac) capable of delivering 3000
systems having battery banks to provide energy watts to 10,000 watts of ac power to either ac
when the utility power is not available. These mul- loads in the house or to the utility grid when
timode PV systems are becoming more common the PV power output is in excess of those local
as energy storage system prices decline, the utility loads. In residential PV systems, single and
power grid becomes less reliable, and natural di- multiple-inverter installations are common. The
sasters like hurricanes result in significant periods single inverter may have an ac output rating of
of utility power loss. Larger residential stand-alone 2000 watts to 7000 watts or more, and systems
systems will usually have a back-up generator, and are frequently seen with two to four inverters
these systems are known as hybrid stand-alone used to increase the system power output
systems (photo 1.5). (photo 1.6). A few residential PV systems have
had ac outputs up to 90 kW!
These residential-sized inverters interface with
the grid at 120 volts or 240 volts; are certified or
listed to Underwriters Laboratories Standard for
Safety 1741 (UL 1741 — Inverters, Converters,
Controllers and Interconnection System Equipment
for Use With Distributed Energy Resources); and
have all the necessary safety equipment built-in
and verified as part of the listing process.
In commercial systems, the three-phase
Photo 1.5 • 10 kW backup generator, powered by natural
inverters commonly used usually start at approx-
gas, used when utility is out during extended periods of imately 10 kW and go up to 250 kW and to 2.5
cloudy weather. MW (yes, 2.5 megawatt in a single inverter).
They interface with the grid at 208 to 480 volts
and higher (photo 1.7).
Stand-Alone Systems
Stand-alone systems are typically installed in
remote areas where the utility grid is not avail-
able or where the connection fees to the grid are
higher than the costs of an alternative energy
system. While stand-alone systems sales are far
lower than sales in the fast-growing utility-inter-
active PV system business, there is and has been a
Photo 1.6 • Utility-interactive inverter. External ac and dc steady market for off-grid systems.
disconnects not shown. The stand-alone inverter converts dc energy
Photo 1.7 • 2.2-MW Sunny Central Inverter. Courtesy of

SMA Solar Technology AG.
Photo 1.8 • Four 6-kW stand-alone inverters in an ac-cou- Photo 1.9 • Inverters and charge controllers for 10-kW off-
pled, battery-backed-up PV system grid, stand-alone PV system.
stored in batteries by the PV array to ac energy to and dc currents to the inverters can be hundreds
support the loads (photos 1.8 and 1.9). Inverter of amps at these higher voltages.
power ratings are from about 250 watts to 6500
watts for residential systems and, as before, PV System Component
multiple inverters may be connected together Descriptions
for greater power outputs. Battery banks usually PV Modules
operate at a nominal 12, 24, or 48 volts, so the The first thing inspectors see are PV modules.
current levels to the inverters can be hundreds of While most have glass fronts, aluminum frames
amps at full load. (colored mill-finish aluminum or anodized brown
Larger stand-alone systems can be found in na- or black), and plastic backs, some will be made
tional parks, at telecom sites, and at federal facil- with plastic frames or with no frames (photo
ities. These can be as small as residential systems 1.10). Others will be used as roofing materials
with ac outputs in the 2 kW to 10 kW range, but (photo 1.11) or laminated directly to standing
they can also have single inverters of 250 kW or seam metal roofs (photo 1.12). PV modules come
more. A few of these larger systems have multiple in many sizes and shapes.
large inverters with combined outputs approach- Inspectors need to determine the listing of the
ing 500 kW or more. Battery banks for the larger modules and the electrical ratings. These are print-
systems operate in the 200-volt to 600-volt range ed on the back of the module and may be available
in the instruction manual or module specification Photo 1.10 • (left) Framed PV modules (anodized and clear-
coated aluminum in natural aluminum color or brown/black).
sheet. Some unlisted, custom modules are being
installed in architect-designed projects. Unlisted Photo 1.11 • (top right) Building-integrated photovoltaics
modules are being sold through various channels (BIPV) PV modules as roofing material.
(including the internet), but unlisted modules no
Photo 1.12 • (bottom right) Thin-film PV modules lami-
longer meet Code requirements and should not nated to a metal standing seam roof.
be installed [690.4(D)]. Although appearances
may differ, these PV modules all produce elec- ing at dc nominal voltages of 12, 24, and 48 volts
tricity when illuminated and the normal cautions and are also used in higher voltage systems (up to
associated with any electrical power system should 1500 volts). They must be certified/listed by a na-
be followed. tionally recognized testing laboratory (NRTL) to
PV modules come in differing power and UL Standard 1741 [690.4(D)]. In these systems,
voltage ratings and the sizes and ratings are it is a normal practice to connect modules in se-
continually changing. The modules must be ries (called a PV source circuit [690.2]) to get the
connected in a manner that produces the needed proper voltage and then to connect each series
voltage, current, and power because the output source circuits of modules in parallel with other
of a single module is usually not sufficient to source circuits through a PV combiner to increase
operate the connected equipment or provide the the current to get the desired power level.
needed amount of energy. These combiners will usually contain the over-
current devices (fuses in the high voltage systems
PV Combiners or circuit breakers in the 12-, 24-, or 48-volt
PV combiners (PV j-boxes or PV combining systems) that are required to protect the module
enclosures) are common in PV systems operat- interconnecting conductors from fault currents
Photo 1.13 • PV combiner with fuses in the positive con- Photo 1.14 • PV combiner in white enclosure with manual
ductor only. Manual and contactor disconnects in positive switch opening both positive and negative conductors.
conductor only. Not 2017 NEC compliant [690.15(C)]. Plastic shield covers energized, exposed conductors.
and the individual modules from reverse currents. with multi-mode inverters (photo 1.17) and
Reverse currents may originate from paral- battery backup were popular for months follow-
lel-connected strings of modules; reverse currents ing the blackouts.
from the batteries in a system that has them; or Unfortunately, installation manuals for these
from backfeed currents from a utility-interactive complex inverters (particularly the stand-alone
inverter (unlikely in listed inverters). See chapter types) can be several hundred pages long. The
2 for additional details on the requirements for inspector should verify the proper dc and ac
combiners. conductor sizes and overcurrent protection.
Both are based on the rated ac power output of
Inverters the inverter. (See Sections 690.8 and 690.9 and
Inverters are found in both stand-alone systems Articles 705 and 706 in the NEC.)
and utility-interactive systems. They essentially Utility-interactive inverters have all the
convert dc)energy from the PV system (or the dc automatic ac utility disconnect devices built-in,
energy stored in batteries) to ac energy for use by which protects utility linemen who are working
local loads or for feeding into the utility system on supposedly de-energized utility feeders. The
(photos 1.15 and 1.16). Some utility-interactive utility-interactive PV inverter will not energize
inverters, known as multimode inverters, have the a dead line and, in fact, will disconnect from
capability to power selected load circuits from the line when the line voltage varies more than
batteries or the PV system when the utility is not -12% to +10% from nominal (typically 120, 208,
present. 240, 277, or 480 volts) or when the frequency
Many PV owners in California were surprised varies by more than -0.7 to +0.5 Hz from the
when their utility-interactive PV systems did normal 60 Hz.
not work during the rolling utility blackouts The inverter monitors the utility line voltage
created by the energy shortage and brownouts and frequency, and that voltage and frequency
a few years ago. Utility-interactive PV systems must remain stable and within tolerance for five
er will usually not include the Section 690.41

ground-fault device, so if it is not built into
the charge controller, an external, field-in-
stalled ground-fault protective device must be
used. The author developed a system of circuit
breakers in 1990 to meet the existing NEC
requirement, but like the use of a fuse GFDI,
it was not adequate to detect all low-level
faults in grounded conductors.
Photovoltaic systems mounted on buildings
will require a PV rapid shutdown system
(PVRSS) which is system manually activated
by first responders to reduce electric shock
hazards due to energized PV conductors on or
Photo 1.15 • Dual 4 kW utility-interactive inverters.
in a building (690.12). Many of these systems
will be built into the U-I inverter, but separate
versions are also available.
Photovoltaic systems with dc circuit
voltages over 80 volts will require a device
known as a dc PV arc-fault circuit interrupt-
er (DCPVAFCI) and, in most cases, these
devices will be built into the string inverter
(690.11).
PV and the 2017 National

Electrical Code
The 2017 NEC incorporated sweeping changes
Photo 1.16 • Transformerless (non-isolated) inverters, typ- in the way it addresses requirements for PV and
ically used on most recent residential and small commer-
related systems and that will resonate into the
cial PV systems. Courtesy of SMA Technologies AG.
future.
minutes before the inverter can resume power
transfer from the dc output of the PV system to Changes Are Here
the ac loads or to the utility. Two main areas stand out among the numer-
All PV systems, with a few exceptions, must ous changes that have been accomplished by
have a device known as a ground-fault detec- Code-Making Panel 4 and the NFPA Technical
tion/interruption device (GFID). See 690.41. Correlating Committee and its task groups. The
These GFIDs are normally built into all first major change is a redefinition of what a PV
utility-interactive inverters and some charge system is and the exclusion or removal of most
controllers. Early utility-interactive inverters non-PV system information and requirements
used a fuse that blew on ground faults, but from Article 690. The second major change is a
this method proved inadequate in detecting redefinition of what is meant by “grounding” and
all ground fault, particularly those in ground- “grounding requirements” as they pertain to PV
ed circuit conductors. The stand-alone invert- systems.
that does not involve energy storage or another

source of energy. Changes to Diagram 690.1 will
highlight these differences in the definition of a
PV system. Primarily, we are looking at the PV
system ending where the PV system disconnect
occurs. In the 2017 Code, the PV system discon-
nect may be either an ac or dc disconnect de-
pending on the configuration of the system. For
example, as shown in Figure 1.1 (part of Figure
690.1 from the 2017 NEC) the PV system dis-
connect on a pure utility-interactive PV system
occurs at the ac output of the utility-interactive
Photo 1.17 • Multimode Inverters. These inverters pro-
inverter. If there are ac PV modules involved, the
vide whole-house battery-backed-up energy and are ac
coupled to utility-interactive inverters. PV system disconnect will be at the output of the
combined outputs of all the ac PV modules in the
PV Systems Redefined system. These PV disconnects separate the PV
The first major change to Article 690 was to system from another energy source: the utility.
more narrowly define exactly what a PV system Even though it is acknowledged that the
is. As we know them today, PV systems may diagrams are not complete, the multimode dc-
include several interconnected sources of energy, coupled system diagram (Figure 1.2) and the
including battery storage systems, multimode stand-alone system shown (Figure 1.3) are possi-
inverters, generators, and similar devices. Photo- bly confusing in one aspect. Typically, the output
voltasic system disconnects are located in various of a PV array cannot be connected directly to a
locations throughout the “so-called” PV system in stand-alone or multimode inverter, to dc loads, or
a manner that is sometimes confusing. to an energy storage system as shown in these di-
In the 2017 NEC, the PV system is primarily agrams. If done as shown, the PV output voltages
concerned with and defined by the PV mod- to these circuits would be unregulated (changing
ules and any device connected directly to them throughout the day and as clouds appear) and
Figure 1.1 • Disconnects

for simple PV systems.
From Figure 690.1 in the
2017 NEC.
could possibly damage the connected equipment. to other parts of the Code.
Some sort of charge-management device (e.g., a
charge controller) is normally used, and under the Other Definitions Changed, Deleted,
new definitions, that charge management device or Added
or system should be grouped with the energy In Section 690.2, in addition to Functional
storage system. If that is the case, then the PV Grounded PV systems (described below), other
system disconnect would properly be connected definitions were changed. The following defini-
to an energy storage disconnect and connections tions were also added.
to the inverter and any dc loads would be made Generating Capacity. The sum of parallel-con-
directly to the voltage-regulated energy storage nected inverter maximum continuous output
system or battery through appropriate discon- power at 40°C in kilowatts.
nects. Also, in the two multimode diagrams Interactive Inverter Output Circuit. The
(Figure 1.3), the output from the multimode conductors between the interactive inverter and
inverter marked “Stand-alone system loads” in the service equipment or another electrical power
the dc-coupled multimode system diagram, and production and distribution network.
not identified at all in the ac-coupled multimode Photovoltaic System DC Circuit. Any dc con-
system, could probably be more properly iden- ductor supplied by a PV power source, including
tified as “Local ac loads” because they can be PV source circuits, PV output circuits, dc-to-dc
powered either by the utility or by the PV/energy converter source circuits, or dc-to-dc converter
storage system when there are utility outages. output circuits.
As the PV system proper became more narrow- The definition of Photovoltaic Systems Voltage was
ly identified, parts of Article 690 that deal with removed and now appears in Section 690.7.
other types of electrical systems have been moved Several definitions were modified, and they
include:
• Inverter Input Circuit. Conduc-
tors connected to the dc input of an
inverter.
• Inverter Output Circuit. Con-
ductors connected to the ac output
of an inverter
Grounding—New and Old

Grounding PV systems have
remained essentially unchanged
for the most part since PV came to
the Code in 1984. However, in the
2017 NEC, the terminology that is
used to describe the various ground-
ing functions has been changed.
Figure 1.2 • Multimode, utility-

interactive PV systems with energy storage
(FIG 690.1).
This was done to reduce confusion by simplifying past, but will exist for years and still be installed
terminology between the older “Grounded PV in smaller systems that are not on buildings.
Array/Isolated Inverter” PV systems and the The required Code requirements to address these
newer, more common “Ungrounded PV array/ varying “grounding” systems were cumbersome,
Non-isolated inverter” PV systems. difficult to understand, and hard to apply in
Grounding the PV System Circuit Conduc- many cases. The changes in the 2017 NEC ad-
tors. In the early years of PV systems, up to dress and simplify many of these grounding
about 2005, we primarily had grounded PV requirements.
arrays and grounded dc battery systems where
one of the circuit conductors was “grounded” or Grounding Definitions—2017 NEC.
“connected” to the grounding system, which is From the 2017 NEC, some definitions are the
composed of the equipment-grounding system, same as in the 2014 NEC.
the grounding electrode conductor, and the
grounding electrode system. From Article 100
In some cases, the grounding method was a “Ground. The earth.
solid conductor and, in others, the grounding “Grounded (Grounding). Connected (con-
method was by a fuse or circuit breaker like those necting) to ground or to a conductive body
used in Section 690.5, Ground Fault Protection that extends the ground connection.
(2014 NEC). In other cases, various resistances “Grounded, Solidly. Connected to ground
or solid-state devices were used to ground one of without inserting any resistor or impedance
the circuit conductors. device.”
Now we have increasing numbers of un- Author’s note: This definition could also include
grounded PV arrays and non-isolated (transfor- fuses, circuit breakers, and possibly relays and
merless) inverters that have no dc circuit con- contactors with the devices not allowed in a
ductors connected directly to ground. Grounded solidly grounded system.
PV systems, where one circuit conductor is “Grounded Conductor. A system or circuit
solidly grounded, are becoming a thing of the conductor that is intentionally grounded.”
Figure 1.3 • Stand-alone system with energy storage (FIG 690.1).

From Article 690

“Functional Grounded PV System. A PV It should be noted that there are no voltage
system that has an electrical reference to requirements associated with the requirements in
ground that is not solidly grounded. 690.41 and that the old 50-volt limit is gone. I sus-
Informational Note: A functional grounded pect it will take some time to figure out what we call
PV system is often connected to ground these various types of PV arrays. For example, the
through a fuse, circuit breaker, resistance PV array connected to a non-isolated inverter with
device, non-isolated grounded ac circuit or a 240 VAC output meets the requirements of (3)
electronic means that is part of a ground- because at some point there is a utility transformer
fault protection system. Conductors in these that has a grounded 240/120-volt center tap.
systems that are normally at ground potential
may have voltage to ground during fault Equipment Grounding
conditions.” Sections 690.43, 690.45, and 690.46 dealing with
Author’s notes: It is anticipated that revisions equipment-grounding conductors have remained
to the 2020 NEC will correct the grammatical essentially unchanged.
error in the phrase “functional grounded PV
system.” Also, it should be noted that a func- Grounding Electrode Systems
tionally grounded PV array using a non-isolated This section has again been revised in the 2017
(transformerless) inverter may have the dc array NEC and will be discussed in some detail in
circuit conductors at significant ac voltages with chapter 6.
respect to ground when the inverter is operating
normally due to the grounded neutral/center tap PV Circuit Overcurrent Protection and
on the utility 120/240-volt transformer. Disconnects
Section 690.9(C) now permits (note: not a
System Grounding mandatory “shall”) a single overcurrent protective
“690.41. PV System Grounding Configurations device (OCPD) to be used in dc PV-source and
One of the following system grounding dc PV-output circuits to protect the modules
configurations shall be employed: and circuit conductors even on ungrounded PV
(1) 2-wire PV arrays with one functional arrays where both circuit conductors—positive
grounded conductor. and negative—are ungrounded. No longer are
(2) Bipolar PV arrays according to 690.7(C) OCPDs required in both ungrounded conduc-
with a functional ground reference (center tors. However, the single conductor OCPD
tap) in source circuits, dc combiners, and in other
(3) PV arrays not isolated from the grounded locations in the dc system must be in the same
inverter output circuit polarity conductor. See photo 1.18.
(4) Ungrounded PV arrays Section 690.13(F)(1) requires that all types
(5) Solidly grounded PV arrays as permitted of PV systems have a disconnecting means that
in 690.41(B) Exception simultaneously opens all conductors (even those
(6) PV systems that use other methods that solidly or functionally grounded) of the PV sys-
accomplish equivalent system protection in tem from all conductors of other systems (photo
accordance with 250.4(A) with equipment 1.19). Solidly grounded conductors, in either
listed and identified for the use.” ac or dc circuits, should not be opened with PV
system disconnects. If a solidly grounded dc
Photo 1.18 • (top) DC combiners may look similar for many

systems. Although this is a combiner listed under the 2014
NEC, fuses are permitted to be in only one conductor. How-
ever, to meet 2017 NEC 690.15(15) requirements, as an
equipment disconnect/isolator, an output switch will be
required in both conductors unless one is solidly grounded,
and the now functionally grounded conductors should not
have white insulation.
Photo 1.19 • (left) The DC PV system disconnect or pos-

sibly an array isolator may have to be rewired on former-
ly grounded PV arrays when updating the inverter to a
non-isolated unit. A switch pole in each conductor will be
required under the 2017 NEC.
conductor is opened by a PV system disconnect

switch, some circuits could become ungrounded
and some conductors marked white may no lon-
ger be grounded. If a grounded ac neutral on an
inverter output is opened, internal circuits in the
inverter (such as filter circuits) may be damaged.
This issue will be clarified in the 2020 NEC.
Section 690.31(B)(1) allows a white conductor
marking on only those circuit conductors that are
solidly grounded.
Help for The Systems Designer from trolling controlled conductors from initiation has
The Professional Engineer been increased from 10 seconds to 30 seconds.
On larger PV systems that are 100 kW and larg- These modified requirements provide additional
er, some of the general requirements for voltage safety for first responders who must deal with an
and current calculations are too conservative for energized PV array and allow the PV industry,
the proper and safe installation of cost-effective in some cases, easier methods of meeting the
PV systems. requirements. (See photo 1.20.)
Section 690.7(A)(3) allows the professional There are now three new options that must
engineer (PE) to calculate the maximum system be used to meet the PV rapid shutdown system
voltages based on industry design practices. These (PVRSS) requirements for the conductors in-
might include such factors as microclimates, array side the PV array controlled boundary [690.12(B)
mounting devices, and array orientation. This (2)]. These requirements will become effective
new calculated system voltage may permit more January 1, 2019.
modules to be placed in series than the more
conservative 690.7(A)(1) and (2) would allow. 1. Use a PV array that has been entirely
Section 690.8(A)(2) allows the PE to calculate listed or field labeled as a rapid shutdown
a maximum current that can be no less than 70% PV array.
of maximum current calculated using the normal 2. Install a PV array with controlled con-
method of using 125% of the short-circuit cur- ductors inside the array boundary and
rent (Isc). In large systems, these methods using not more than 1 m (3 ft) from the point
simulated local irradiance may allow smaller, yet of penetration to 80 V or less within 30
still safe, conductor sizes. seconds of PVRSS initiation.
3. Install a PV array with no exposed conduc-
Other Changes in Article 690 tors or exposed conductive parts more than
Scope, 690.1 2.5 m (8 ft) from exposed grounded conduc-
The scope excludes large PV systems which, are tive parts or ground.
now covered in Article 691.
It should be noted that these three new options
DC-To-DC Converter Source and Output Circuits, will become effective on January 1, 2019. This re-
690.7(B) quirement implies that some sort of module-level
This new section establishes requirements for control will be required to meet the controlled
determining the maximum voltage of single or conductors inside the PV array boundary.
multiple series connected dc-to-dc converters. In some cases, module-level power electronics
(MLPE), such as microinverters, ac PV modules,
PV Rapid Shutdown Systems, 690.12 or dc-to-dc converters, may meet the PVRSS
This section received significant changes for requirement after they have been listed as rapid
2017. The previous boundary around the PV shutdown equipment. The 2014 NEC required
array that defined the point where conductors only that the equipment be listed and identi-
must be controlled to a lower voltage has been fied (meaning a listed relay could be used). The
decreased from 3 m (10 ft) to 305 mm (1 ft). The 2017 NEC, however, requires the equipment to be
length of allowable uncontrolled conductors listed and labeled as a rapid shutdown system or
inside a building has been reduced from 1.5 m (5 as RS equipment.
ft) to 1 m (3 ft). The time allowance for con- The new and revised PVRSS section of UL
Standard 1741 establishes complex and detailed plaque clearly showing a physical plan diagram of
testing requirements. One of these requirements is the PV system indicating which parts of the PV
that the ac output of inverters go to the controlled array are RSS controlled and under which edition
voltage limit within the required time frame after of the Code that part meets. Section 690.56(C)(1)
initiation (usually opening the ac utility discon- has detailed marking requirements for the various
nect) of the RSS system. This test was not in UL types of PVRSS.
1741 under the initial RSS requirements. Finally, the manual reset requirement has been
The initiation device for the PVRSS must deleted from the 2017 NEC.
plainly indicate the OFF position and the ON
position, where the OFF position indicates that Arc-Fault Circuit Protection (Direct-Current),
the PV array has been put in the controlled state 690.11
for all circuits controlled by that initiation device. This section of the Code was substantially
The initiation device shall consist of at least one reduced for 2017. An exception was added that
of the following: exempts PV systems from the arc-fault circuit
(1) The service disconnecting means. protection requirement that are not mounted on
(2) The PV system disconnecting means. buildings. Also exempted are PV systems with
(3) A readily accessible switch that indicates PV output and dc-to-dc converter output circuits
the ON and OFF positions. that are direct buried, installed in metallic race-
Because various PV systems are installed at dif- ways, or installed in closed metallic cable trays.
ferent times and may be expanded under different This exclusion was necessary because arc-fault
editions of the Code, Section 690.56(C) requires a equipment at the high current levels found in
large PV arrays does not exist.
Photo 1.20 • PV Rapid Shutdown Equipment.
Photo courtesy of SMA
The exclusion does not address larger PV sys- on inspecting large PV systems in recent editions
tems on the roofs of commercial buildings. These of IAEI News point out the continuing need for
systems are being addressed using multiple medi- close and detailed inspections of these installa-
um-size string inverters with arc-fault detection tions by an AHJ supervising a team of inspectors
capabilities mounted in subarrays on the roof. directly or carefully reviewing the certifications of
The manual restart requirement and the re- the independent professional engineer.
quirement for an annunciator have been removed,
but are included in the requirements in UL Article 705
Standard 1741. Definitions, 705.2
Definitions have been added or modified in
Article 691 this section.
For large-scale (5 MW or greater) PV systems, A microgrid system is “a premises wiring system
not under the ownership or operation of a utility, that has generation, energy storage, and load(s),
special design and installation requirements are or any combination thereof, that includes the
allowed. These requirements address numerous ability to disconnect from and parallel with the
areas, but in general require a PE-engineered primary source.” The requirements of this code
design; limited and controlled access; no local apply unless the microgrid system is under the
loads except as necessary to operate the gener- exclusive control of a utility.
ation system; and a full, utility-type substation The term utility interactive has been modified to
to make the connection to the local utility. Also, just interactive throughout section 705.
parts of the system that do not comply with the The definition of stand-alone system has been
requirements of Article 690 shall be fully justified moved to the new Article 710 – Stand Alone
and documented. Systems.
At the request of the AHJ, an independent
electrical professional engineer, retained by the Point of Connection, 705.12
system owner, may be required to evaluate the Sections (B) and (C) dealing with integrated
actual installation for compliance with the Code electrical systems and systems greater than 100
requirements and the engineered design. kW have been deleted from this Article. Article
Even though these large systems (in Articles 705.12 requirements found in 705.12(A) and the
690 and 691) are designed by a PE, and may renumbered 705.12(D) [now B)] apply to inter-
have independent review of the installation for connected electrical power systems of any size.
compliance with the design, it is still the respon-
sibility of the AHJ to ensure that the system is Point of Connection, Load Side 705.12(B)
safe and meets the various applicable codes. In This section has been modified to allow other
many cases, it is common for jurisdictions to interconnected (interactive) power sources to be
outsource inspections of large-scale PV systems connected on the load side of the system discon-
due to departmental manpower shortages. This necting means.
outsourcing could be to an independent electrical
inspection service or to a licensed professional Center-Fed Panelboards, 705.12(B)(2)(3)(d)
engineer. Keep in mind that all the large utili- Center-fed panelboards may now be treated
ty-scale systems being installed today use engi- like standard panelboards where the backfed PV
neered designs. breaker may be connected at the end (away from
Scott Humphrey’s excellent series of articles the main breaker) on only one of the two busbars
Photo 1.21 • (top left) Center-fed panelboard. PV connec- Photo 1.22 • (top right) Residential microgrid system—PV
tions now allowed by the 2017 NEC at one end only. Photo input.
by Dan Potkay.
Photo 1.23 • (bottom right) Microgrid system—generator
input.
and may be rated according to the 120% rule. This
rule states that 125% of the inverter output current system disconnect requirements in 690.13(F)(1)
rating plus the main breaker rating may not exceed require all conductors to be disconnected. This
120% of the busbar rating (see photo 1.21). will be addressed in the 2020 NEC, where only
ungrounded conductors shall have disconnect
Wire Harness and Exposed Cable Arc-Fault requirements.
Protection, 2014 – 705.12(D)(6) The equipment disconnecting requirements in
This section dealing with ac arc-fault protec- 705.21 are consistent with the equipment discon-
tion for 240-volt ac inverter output circuits up to necting/isolating requirements in 690.15 because
30 amps on flexible, exposed cables has been re- both require only the ungrounded conductors to
moved because no suitable equipment is available be disconnected.
to meet the requirement.
Disconnect Device, 705.22
Disconnecting Means, Sources This slightly revised section allows power-op-
Article 705.20 may require some harmo- erated disconnects to be used, but this allowance
nization with the revised system-disconnect for PV system disconnects in the 2014 NEC was
requirements in Article 690. These Article 705 removed when 690.17 was deleted. Again,
disconnect requirements require only ungrounded harmonization may be required in future editions
conductors to be disconnected, whereas the PV of the Code.
Article 710, Stand-Alone Systems

This new article gathers requirements previously
located in sections of Articles 690 and 705 relat-
ing to stand-alone systems. The Article primarily
addresses the requirements for a remote, off-grid
dwelling or other load and there is considerable
flexibility allowed in the characteristics of the
power source supplying that load.
Article 712, Direct Current Microgrids

Direct current distribution systems may be fed by
a PV system or another dc source. The ground-
ing terms defined in this Article differ somewhat
Photo 1.24 • Microgrid research project with utility-inter-
active PV (array on roof ), energy storage, and smart re-
from the new grounding terms found in Article
ceptacle outlets measuring energy and controlling smart 690. Section 712.2 defines a grounded two-wire
appliances. dc system and a grounded three-wire dc system as
having a solid or reference ground between one
Microgrid Systems, 705 Part IV circuit conductor and the equipment grounding
This part has four sections that outline the system. It also defines a reference-grounded dc
requirements for operation of a microgrid and system as one that is not solidly grounded, but
connection of that microgrid to a primary power has a low-resistance electrical connection to
source. These requirements are based primarily ground that maintains the circuit conductor
on the operation of utility-interactive inverters at equipment grounding system voltage under
found elsewhere in Article 705 and previously in normal operation. This sounds like the Section
Article 690 (photos 1.22, 1.23, and 1.24). 690.2 definition of functional ground, but could
use a circuit breaker or fuse as the low-resistance
Wiring Systems, 705.175 element. Section 712.2 also defines resistively
This section has been moved to Article 710, grounded as a system having a high-resistance
Stand-Alone Systems. connection between the circuit conductor and
the equipment grounding system. Of course, un-
Article 706, Energy Storage Systems grounded is defined and it (to a certain extent)
This new article was added to the 2017 NEC and agrees with that term in Section 690.2.
contains nearly all the references to batteries or On a dc PV system, we should certainly be well
other energy storage systems previously found “grounded” between these two Articles—I think.
in Article 690. Article 706 is lengthy and has re- (See photo 1.25.)
quirements that were not previously in the Code,
such as a requirement to determine and mark the UL Safety Standards
rated short-circuit current of the energy storage UL Safety Standards are developed to ensure
system. Article 480, Storage Batteries, remains that equipment certified (listed) to these stan-
in the Code and has some of the same new or dards can be installed according to the instal-
revised requirements that appear in Article 706. lation requirements found in the NEC, thereby
Chapter 5 will cover Article 706 in detail. yielding an essentially hazard-free electrical
Photo 1.25 • “We just know there is a ‘function ground’ here somewhere. Give us time; we’ll find it.”
installation. In the days preceding electronic careful adjustments of software- and hardware-
power equipment (yes, that was even before my driven internal parameters to operate safely.
time), application of the Code requirements was Each of these active devices, made by numerous
relatively straightforward: connect conductors to manufacturers, has different requirements for
passive devices like switches, overcurrent devices, safe installation. It is neither possible nor prac-
and transformers. Many electricians could do that tical to put detailed instructions in the Code for
with a little training, and even the inspections the installation and adjustment of each of these
were relatively easy. types of devices. The Code would become far too
However, AHJs are now faced with active ponderous to deal with in its present form.
electronic power-processing equipment scattered For this complex equipment, the best
throughout our electrical power systems, and PV the Code can do is to include Section 110.3(B),
systems are no exception. AHJs must now ensure which requires that the instructions and labels
that active inverters, active charge controllers, provided on certified (listed) equipment be used
rapid shutdown equipment, dc PV arc-fault and followed to properly install that equipment.
circuit interrupters, microinverters, and ac PV This requirement in the NEC places a signifi-
modules plus module-level electronics are prop- cant burden for ensuring the safe installation di-
erly installed and that they will operate safely. rectly on the quality and detail of the equipment
Many of these devices have multiple inputs and installation instructions that are required by the
outputs and require not only electrical connec- UL Standard for that equipment. These standards
tions to the rest of the electrical circuit, but also must be detailed and understandable because
several Nationally Recognized Testing Laborato- dards are so complex that the entire STP does
ries (NRTLs) other than UL such as TUV, ETL not, as a group, develop them. Various working
and CSA will be using them to test and evaluate groups are set up composed of STP members and
PV equipment. (See photo 1.26). other technically competent interested parties to
These requirements lead us to the effort draft sections of the Standard. After those drafts
required to develop these standards. Although are edited to comply with UL writing standards,
Underwriters Laboratories publishes the Stan- they are presented to the STPs for review, final
dards, the various UL Standards Technical Panels refinement, balloting, and integration into the
(STPs) write them. The STPs are composed of whole Standard. An example is the draft Stan-
balanced volunteer groups of manufacturers, dard for the Code requirement in 690.12 for a PV
AHJs, technical experts, users, government rapid shutdown system (PVRSS).
agencies, and general interest individuals. In some Underwriters Laboratories’ engineers devel-
cases, these groups may include more than fifty oped the basic outline and the initial contents for
people. There are standards and STPs for each of the draft PVRSS Standard and then published
the major categories of equipment in PV systems. them as a Certifications Requirements Deci-
These include UL 1741 for inverters (including sion (CRD) in March 2015. UL then formed
charge controllers, microinverters, rapid shut- a working group outside of the STP to further
down equipment, and ac PV modules); UL 1703 refine this draft. The working group, consisting
for PV modules; UL 1699B for dc PV arc-fault of more than fifty people, met on a weekly basis
circuit interrupters; UL 4703 for single conductor via webinar for more than twelve months to
PV cable; UL 6703 for PV connectors; UL 2703 investigate, evaluate, and develop the require-
for PV racks; UL 489B for dc PV overcurrent
Photo 1.26 • Rigorous application of standard require-
devices; and others. ments will minimize issues like this failed PV module sol-
In many cases, the requirements in the Stan- der bond.
ments for the PVRSS based on the very limited standard was developed to address systems that
requirement stated in Section 690.12 of the 2014 control hazards within the boundary of the PV
NEC. And as the PVRSS requirements in the array. This Standard, UL Standard for Safety for
2017 NEC became known, the draft Standard Photovoltaic Hazard Control (UL 3741), will be
was again revised. published in late 2018 or early 2019.
The working group had to consider all possible
ways in which various equipment manufacturers Harmonization with International Standards
could implement the Code requirement. Because Photovoltaic equipment is frequently designed
this PVRSS system provides a life safety func- and manufactured to be sold and installed in in-
tion, the testing and evaluation criteria that the ternational markets. The equipment must meet the
Standard required would have to be rigorous, codes and standards of each country. While there
extensive, detailed, and very complex. The PVRSS is some variation in the electrical codes between
system may have to work properly in an outdoor countries that have such codes, the connections of
PV system environment that includes extremes PV equipment to electrical systems is not too dif-
of temperature, humidity, UV irradiation, wind, ferent from country to country. Some editions of
snow, ice, and blowing dirt. The standard must the NEC have been translated into other languages
require testing to evaluate PVRSSs and PV rapid (photo 1.27). Unfortunately, the standards govern-
shutdown equipment (PVRSE) under these ing how equipment is built and tested can differ
conditions plus surges that might occur from significantly between countries. There has been a
nearby lightning strikes on the connected ac and long-term, on-going effort to harmonize safety
dc circuits. The PVRSS equipment must also and construction standards of PV equipment
operate reliably for the 40- to 50-year life of the between countries.
system. While there are UL Standards to evaluate A harmonized standard would reduce manu-
some of these conditions, and those standards facturers’ costs of certification and listing in mul-
were referenced in this new PVRSS Standard, tiple countries. Engineers from the United States
the unique nature of outdoor PV environments have participated in the development of interna-
required that new testing and evaluation proce- tional standards that are published by the Inter-
dures be developed. national Electrotechnical Commission (IEC),
Even simple products, such as power relays or based in Geneva, Switzerland. The International
contactors, that have been listed to various UL Standard, by its very nature, mainly addresses
Standards for the necessary voltage and current common requirements that apply in all countries
have frequently not been tested in wide tempera- to the design and testing of PV system electrical
ture extremes found in the PVRSS application. equipment. After an IEC Standard is published
Electronic devices must be evaluated for new formally, it is distributed to the various member
types of failure modes such as operation in high countries where they evaluate it and add sections
humidity environments and the response when to the standard that elaborate on the “Country
control circuits fail or are interrupted. Differences” that will be used along with the
The draft Standard is more than 20 pages common International Standard requirements to
long and, after review and modification by the evaluate products in a particular country.
UL 1741 STP, it will be added to the UL 1741 As United States safety standards become
Standard (already 144 pages long). harmonized with the international standards,
In addition to the changes made to include working groups in the United States spend two to
PVRSS in UL Standard 1741, an entirely new three times as long developing the international
of such vehicles and in locations throughout the

area where they’ll roam (photo 1.28). Similar
to cell phone coverage, charging stations will be
concentrated in metropolitan areas and then will
spread to less populated areas as the demand
for extended coverage grows. Owners of electric
vehicles will certainly have charging stations in
their homes and probably at their job sites. At
the very least, there will be a new type of recep-
tacle outlet to manage and probably relatively
high-current branch circuits. Article 750, Energy
Management Systems, has been expanded in the
2017 NEC. In addition, a new Article 706 applies
to charging systems for electric vehicles and ener-
gy storage systems for PV systems.
Photo 1.27 • Codes and standards are being harmonized Plans are also being made to have parked, fully
between countries
charged electric vehicles feed some of the energy
standards as they would a U.S.-only standard. The stored in the onboard battery bank back into the
working groups first develop the United States utility grid at peak demand times. Controlling
standard, and it is published as a UL Standard. this exchange of energy from grid to vehicle
They then work on the international standard and, and back, and ensuring that the car is ready and
finally, they develop the “Country Differences” charged when needed, will require communica-
when the international standard is published. This tion between the car, the owner, and the utility.
is a very long process involving possibly hundreds Such communication links may be wireless, over
of people and thousands of hours. All of this is the Internet, or through a hardwired connection
to ensure that PV equipment can be installed along with the power connections. Like utili-
according to the requirements in the NEC and will ty-interactive PV systems, these vehicle storage
operate safely when so installed. systems will require new Code changes and
The AHJ should now have a better idea why additional inspections to ensure public safety.
the instructions provided with a listed product
are critical and why those instructions must be Energy Storage Systems — Large and
reviewed during the permitting, plan review, Small
and inspection processes to ensure safe and Utilities will embrace the dispatchable energy
Code-compliant installations of PV systems. storage and generation systems. They will be
able to tap energy that has been stored or that is
Technology Advances available throughout the distribution network for
use to offset peak demand loads. This operation
Electric Vehicles will avoid having to increase the size of already
There are now both plug-in hybrid electric taxed power plants and transmission lines (photo
vehicles (PHEVs), having fueled engines plus 1.29). Backup generators at hospitals and other
electric motors and batteries, and pure electric locations are already being used in this mode of
cars (EVs), having electric motors and batteries, operation. These emergency power systems are
that require charging stations at the home base leased, operated, and maintained by third parties
Photo 1.28 • (top) Electric vehicles are back after a century of disuse in various forms, including EVs, hybrids, and fuel-
cell vehicles. They will dominate the transportation industry in the future.
Photos 1.29a and b • (middle and bottom left) The infrastructure of utility generation and distribution systems is fairly ro-
bust, but very old, and somewhat inflexible in dealing with increased use of energy sources and its associated issues. Smart
Grid programs are designed to modernize the entire system from the generation plant to end-use load.
Photos by John Watson.
who run them when not needed for emergencies cess that yields a long-lived battery that can be
and sell the power to the utilities during peak rapidly charged and deeply discharged virtually
load periods. New Article 706, Energy Storage an unlimited number of times. The batteries will
Systems, addresses these systems, and most of the be charged, and energy will be stored, during off
battery requirements in Article 690 have been peak demand periods and released back to the
moved to this location. grid during peak demand times. Of course, the
Flow batteries are coming to the market. These process will require utility-interactive systems to
batteries use stored liquid chemicals and a pro- interface with the utility grid and communications
systems to control the process. These systems, wind

turbines (NEC Article 694) (photo 1.30), and fuel
cells (NEC Article 692) operating from natural gas
will probably first appear in commercial buildings
that have the necessary space. These systems will
either be leased or owned, but in most cases these
new technology systems will require permitting
and inspections of the added mechanical systems,
the utility-interactive electrical connections, and
the communication circuits.
The Smart Grid

Energy demands throughout the country, and the
world, are increasing steadily and will necessitate
some combination of increasing the supply from
new generation plants (coal, gas, oil, or nuclear
and renewable), reducing the demand through
conservation, and restructuring of the existing
distribution and consumption system. Climate
change considerations and market conditions will
have significant impact on which energy sources
Photo 1.30 • Wind turbine farms are providing significant
will predominate in the mix. The infrastructure of
amounts of alternative energy today. Photo courtesy US DOE.
utility generation and distribution systems is fair-
ly robust, but very old, and is somewhat inflexible
in dealing with increased use of distributed Smart houses will soon become a reality.
energy sources and issues associated with moving Appliance manufacturers are already making
power from the sources to consumers in other dishwashers, clothes washers, and other applianc-
areas. The Smart Grid programs are designed es that communicate through either hardwired or
to modernize the entire system from generation wireless communication systems to the home-
plant to end-use load. owner, to the smart meter, and (in some cases)
Although many see the term Smart Grid and to the utility. When financially beneficial to the
think that it will not impact the premises wiring, consumer, or possibly when legislated, these
the NEC, or the inspection process, that would smart appliances will be remotely controlled by
be a misconception. Utilities are now installing the utility so they may be operated only during
smart meters as rapidly as they can find funds to times of low demand on the utility system. These
do so. These smart meters are computer (mi- appliances may have unique plugs and receptacles
croprocessor)-based and not only allow remote and possibly communication connections. All of
reading and power quality recording (real power, these new load connections must be inspected, of
reactive power, power factor, and more), but may course.
also serve as the interface between the Smart What will be the impact on the Code of a house
Grid and the smart house. Some smart meters that has load circuits and loads that may be remote-
can even allow power to be remotely disconnect- ly controlled or managed? How will service, feeder,
ed when bills are not paid. and branch circuit sizes be determined? Copper
conductor prices may rise so high that the electrical electrical systems in their homes were far ahead
equipment industry is forced to control power flow of their time! Of course, appliances requiring sig-
so that smaller conductors can be used. Eventu- nificant power for heat or mechanical motion like
ally, the use of electricity on the premises may be ranges, clothes washers, toasters, water heaters
scheduled so the maximum current ever drawn may and the like will usually need higher voltages to
be significantly less than that requiring a 100- or keep the current and hence the conductor sizes
200-amp service today. Smaller conductors and to reasonable sizes. But then there are heat pump
circuit sizes may reduce the ever-increasing costs of water heaters, induction ranges, and ultrasonic
electrical installations, but Code revisions would be washers that operate more efficiently than con-
needed. With the demise of the incandescent light ventional appliances.
bulb, do we really need three volt-amps per square
foot for general-purpose circuits? Oh yes, there will Renewable Energy Systems
be those 100+ inch flat panel displays on all four Large wind-power systems have been installed
walls to deal with. for many years, and many of those systems are
not owned and operated by utilities on utility
Is dc coming back? properties. Therefore, they come under the re-
There is a trend of going back to dc end-use quirements of the NEC and should be inspected
appliances. Most electronic appliances such as for safety using the requirements of Article 694,
cell phone chargers, radios, TVs, DVRs, DVD Wind Electric Systems. UL also has standards
players, cable boxes, satellite receivers, track for large and small wind turbines. Photovoltaic
lighting and the like—while being plugged into power systems for residential and commercial
a 120-volt ac receptacle outlet—run on low-volt- use have been around since the mid-1970s with
age dc. Fluorescent and LED lighting bulbs and substantial growth starting in the late 1990s.
fixtures also operate on direct current. Significant While ever-increasing numbers of residential
losses are incurred to transforming the 120-volt and small PV systems are being installed through-
ac line voltage into low-voltage dc. New Article out the country, real power production will come
712, Direct Current Microgrids, describes sys- from the numerous megawatt commercial systems
tems and circuits that may or may not be directly being installed and planned. Systems as large as
connected to the utility grid. There are also color 2000 megawatts are being planned and installed.
codes for dc branch circuits in Article 210. Some of these will be solar thermal systems along
At the present time, dc lighting fixtures are with the PV systems. In many cases, these large
being installed in commercial buildings and are systems are said to be “Behind the Fence” and not
powered during the day directly from PV power subject to the requirements of the NEC and in-
systems with no conversion to ac until the elec- spections, but they are mainly owned and operated
tronic ballasts are reached. Solar lighting power is by private companies under power purchase agree-
supplemented with utility power when necessary. ments (PPAs) and should be fully NEC compliant.
With the reduction in use of incandescent New Article 691, Large-Scale Photovoltaic (PV)
light bulbs over the next few years, the return Electric Power Production Facility, addresses these
of low-voltage dc power distribution systems larger (over 5 megawatts) systems.
for lighting and electronics is almost a certainty. True ac PV modules with microinverters
Shades of the 1970s and 1980s! Maybe those bonded to the back of the PV module with no
off-grid, long-haired solar hippies who insisted dc wiring subject to Code requirements. They are
on staying with the 12-volt dc PV systems and on the market in catalogs and in big box stores at
impulse-buying prices and are being installed by The Environment

PV professionals. Will these products be listed? With a salute to the U.S. Postal Service, neither
Will these be permitted? Will they be installed snow nor rain nor heat nor gloom of night stays
by qualified persons (those having the skills, these PV systems from producing power. Well,
training, and experience as per Article 100)? Will OK, the PV system will stop producing power
they be installed on dedicated circuits that will at night and will have reduced output when the
ensure public safety? Or will they be plugged into sun is not shining brightly, but when the sun next
the nearest GFCI outlet and violate Code regula-
comes up, it will be ready to produce the rated
tions and safety in many ways?
power (or more) day after day and year after year.
Of course, mainstream manufacturers are mak-
Environments in which PV arrays operate
ing and selling ac PV modules through normal
are harsh. The array, module, and string wiring
solar distributor networks and they are properly
are exposed to environmental conditions from
listed to UL Standard 1703 (module) and UL
the coldest temperatures in the winter to the
Standard 1741 (microinverter). They are being in-
hottest temperatures in the summer—plus the
stalled by professional, licensed PV installers and
added stress of high winds, sleet, hail, snow, and,
other qualified individuals and organizations.
nearly every day, additional heating due to solar
Dc-to-dc converters attached to or connect-
radiation.
ed to PV modules are well established in the
A PV system may be subjected to high levels of
market along with (in some cases) matching
sunshine (irradiance) on a continuous basis (for
inverters. Due to the nature of their operation,
three hours or more) when the sky is clear, the
many of the Code dc requirements do not apply
to these devices. The instruction manuals for sun is bright, and there is low atmospheric pol-
these listed products must be used to determine lution (humidity, dust, smog, etc.). Under those
compliance [110.3(B)]. The NEC does not conditions, a PV array may produce power and
specifically give guidance on how to deal with current that can reach 125% of the rated output.
them, and future editions of the Code may show A PV system is designed to accommodate these
a similar trend. situations.
All of these changing and emerging tech- A PV system will also be subjected to short peri-
nologies will create challenges for inspectors ods (minutes) of cloud-enhanced, snow-enhanced,
and plan reviewers and also opportunities to or water-enhanced irradiance that can be up to
excel. 150% of that on a bright, sunny day.
These environmental factors, plus the fact that
The Challenges this is an electrical power system that produces
Electrical inspectors, plan reviewers, and com- power in a daily cycle, means that all the conduc-
bination inspectors are being challenged today tors, any combiners, all overcurrent devices and
(and will be for the foreseeable future) with these disconnects, and each and every electrical termi-
evolving energy production and storage sourc- nal are subjected to a cyclic flow of current from
es that will be in use throughout the country. zero at night to a maximum at or above (for short
Many of these systems will appear connected to periods) the rated value of the device every day.
premises wiring and they will come under the Additionally, environmental heating and cooling
requirements of the NEC. Many multi-megawatt variations on daily and seasonal cycles impose
PV, wind, and solar electric farms will fall under additional stresses on PV systems.
the Code. This high-stress environment and operation
Photo 1.31 • Close-mounted PV arrays allow little room for inspecting the module and array wiring.
is unique compared with the typical residential systems; and where the inspector is also familiar
and small commercial electrical systems that with the technologies and equipment being in-
are familiar to the AHJ. When considering stalled. Experience with the contractors involved
that these systems may operate unattended and also speeds the inspection process.
unmaintained for years—if not decades—PV It is also acknowledged that many jurisdic-
systems demand the highest quality design, tions do not do plan reviews and, unfortunately,
material, equipment, and installation. In the plan this hampers the ability of the inspector to
review and final inspection of a PV system, the easily verify all Code requirements. Photovoltaic
AHJ has the final responsibility and authority modules on roofs are generally mounted close
to verify that the stringent requirements of the to the roof and it is generally not possible to
electrical and building codes are followed in the examine all the connections in the PV array, in-
installation. There are some details in various PV cluding conductor size, grounding and bonding,
subsystems that may need additional attention to and module wiring interconnections. (See photo
achieve a hazard-free, long-lived system. 1.31.) A related issue in a well-installed PV or
other electrical system is that it is sometimes
Yes, Time and Money Are in Short difficult to verify conductor types and conductor
Supply sizes where connections are made inside stan-
It is acknowledged that many residential elec- dard disconnecting devices and inverters. These
trical systems are inspected in less than an particular Code requirements are best reviewed
hour. This is probably acceptable for an electrical in a plan review stage even if that review is only
system that has remained essentially unchanged a half hour in the office looking at the detailed
for 100 years; where the inspector is quite famil- schematics and diagrams that were submitted
iar with Code requirements for those electrical with the permit application. If issues are found
during the office review, it is almost certain that Without going into the attic (and the flat-roofed
similar issues (or more critical ones) will be homes common in the southwest United States
found if time is taken to examine those areas that have no attic), it is difficult to verify that
during the field inspection. mechanical fasteners are indeed being inserted
While many larger PV systems are ground into the structural roofing components (photo
mounted, the great majority of residential PV 1.32). In many parts of the United States, stan-
systems, and smaller commercial PV systems, are dard spacing for trusses is 24 inches with older
mounted on the roofs of buildings. Unfortu- homes having structural members at 16 inches. A
nately, some jurisdictions have insurance policies quick check of the exterior mounting feet for the
that prohibit inspectors from going on the roof, array rack should show them to be consistently
making examinations of those critical PV system on some multiple of these dimensions. Racking
components difficult, if not impossible. If a ladder and mounting systems are continually evolving,
is available, and the installing contractor makes and any system must be installed according to the
all areas of the array easily accessible by providing manufacturer’s instructions.
ladders, the inspector might be able to view some The module interconnect wiring should be
of the array from the ladder without going onto firmly secured to the racking members or other-
the roof. The roof is no place to be in inclement wise supported. Those exposed, single-conductor
weather. cables should not be hanging loosely below the
modules where they are subjected to additional
The Array-Mounting and Mechanical mechanical stress from wind loading and the
Considerations possibility of abrading the insulation against
Combination inspectors who are also building in- the roof. Unfortunately, many of the plastic
spectors are probably more aware of the mechan- wire ties that are frequently used to attach these
ical requirements to fasten PV modules to roofs conductors to racking members, although listed
the building structures than typical electrical for outdoor exposure, do not have the necessary
inspectors might be. Most mounting systems, in- longevity for the years of service needed in a PV
cluding PV racking systems, require that the PV
Photo 1.32a and b • Roof trusses at 24 inches on center
array be connected to the structural elements of (more or less). But where are they after the sheathing and
the roof and not just to the thin roof sheathing. shingles are installed?
Photo 1.33 • Plastic cable ties are not durable in the PV ties or clips used to attach and secure the module
environment, particularly these white ones that are not wiring to the racking members could be discussed,
rated for outdoor use.
along with any listing or other information that
system. (See photo 1.33.) It is suggested that addresses the durability of those devices in the
metal clips with rounded edges be used to attach PV application. The installer might be queried on
these wires to provide the longevity needed for a the method of determining where the racking- or
PV system. module-attachment devices were located in rela-
Module connectors should be firmly seated tion to the structural members of the roof.
and, to some extent, able to be visually evaluated
at a distance. But where the conductors have been Torqueing Considerations
properly secured to the racks, the connectors may All electrical connections that require a
be less visible. threaded fastening have a torque requirement
The grounding and bonding methods used for for making that connection correctly. Thread-
the modules and the racks should be evaluated and ed screw or bolt connections are found on
compared with the instructions for those devices. disconnect switches, overcurrent devices, dc
combiners, and inverter inputs and outputs.
Can’t Get on The Roof? Because of the daily cycling of current in these
In jurisdictions where it is not possible for the circuits, it is critical that they be torqued to
AHJ to get to the roof to closely examine the the manufacturers’ specifications. The installer
aforementioned parts of the array closely, it is should be able to provide data on all torque
possible to query the installer about some of the specifications. The torque values that should
things mentioned above. For example, the meth- have been used should be readily available in
od of grounding and bonding the modules and the manuals and cut sheets provided with the
the racks can be discussed while looking at the permit application.
instructions for those pieces of equipment. The The installer should be able to show the AHJ
Photo 1.34 • All of this equipment has terminals, terminals, and more terminals. Each terminal must be properly torqued.
110.4]. Also, the torque tools can be used to

verify that the minimum value of torque has
been achieved. Obviously, no torque test should
be done on energized circuits.
The AC Utility Connection

Although Section 705.12 of the NEC is being
improved with each edition of the Code, the
requirements still confuse some installers and
AHJs. Following are a few simple checks that can
Photo 1.35 • Do not attempt to make a supply-side con- be done in the field to at least partially veri-
nection inside the meter main combo. The listing on the fy Code compliance.
load center would be violated. A supply-side connection, as allowed by
the torque wrenches and torque screwdrivers 705.12(A), will be made between the existing
used to make these connections. A visual revenue meter and the main disconnect on the
inspection should be made to see if all conduc- service. This connection is typically not allowed
tors are properly inserted into terminals and the to be inside a meter main load center combina-
AHJ should have the PV installer demonstrate tion where tapping either the internal conductors
on a few randomly selected terminals that they or the bus bar between the meter socket and the
have been properly torqued [NEC 110.14(D), main disconnect would violate the listing on the
Informative Annex I, NECH Exhibits 110.3, load center (photo 1.35).
device. If a meter adapter were used,

it would also have to be rated at 100
amps to use the full 100 amps from a
PV system.
Load-side connections, under
705.12(B), are made by backfeeding a
circuit breaker in the main load cen-
ter or making a connection elsewhere
in the electrical system at a subpanel
or possibly even by tapping a feeder.
Although 705.12(B) has been clari-
fied in the 2017 NEC, there are still a
number of considerations to be made
that are best checked in the office
during a plan review exercise. How-
ever, the most common connection
is a backfed breaker in the main load
center and several items are relatively
Photo 1.36 • PV supply-side meter adapter. easy to verify for Code compliance.
Courtesy of San Diego Gas and Electric. (See chapter 7 for the details on
utility connections.)
There are now several listed meter socket
adapters on the market designed for making a Continuing Education and Information
supply-side connection at the meter. The exist- The challenge for every electrical inspector and
ing meter is removed, the adapter is installed in plan reviewer is to keep abreast of new devel-
the existing meter base socket, and the meter opments as they start to appear in residential,
is reinstalled by plugging it into the adapter. commercial, and industrial electrical systems.
The adapter provides a set of terminals that are Inspectors and the plan reviewers need to know as
a supply-side connection. This adapter might much, or more, about these devices and systems as
conceivably even be used in a meter main load the people installing them. That has been true in
center combination to make a utility connection the past and it needs to be the standard of perfor-
without violating the listing on the load center. mance in the future if the inspection community is
While this is a listed device, utility approval is to ensure the safety of the public.
usually required for its use because many utilities Where the Code and the standards cannot keep
consider the meter and the meter base to be up with these new systems and devices, inspectors
under utility control. (See photo 1.36.) and plan reviewers must devote time to educate
The rating of a supply-side connection for a themselves on the systems that they are and
PV system is limited by the size of the service will be inspecting. Strong continuing education
entrance or any meter adapter or conductors programs for inspectors and plan reviewers must
involved with the service [705.12(A)]. be a part of the planning in every jurisdiction.
A 100-amp service entrance that has 100-amp Time and funding should be budgeted for classes,
meter and 100-amp service-entrance conduc- webinars, technical documents, and the equip-
tors could be fed by a 100-amp PV overcurrent ment needed to efficiently and proficiently review
and inspect these ever-changing electrical power requirements in some states to provide PV “ride
systems. through” of utility disturbances, including low
Inspectors and plan reviewers should have voltage (down to 50% of nominal) and frequency
electronic copies of all codes, handbooks for variations.
those codes, and technical data (including
manuals and specification sheets) for all types Installer Qualifications. While many PV sys-
of systems being inspected (and for equip- tems are being installed by competent individ-
ment that may be installed on those systems). uals and organizations, there are still numerous
Laptop computers, with screens that can be systems being installed by people who have less
read outdoors, with this information (updated than full competency in the required knowledge
as necessary) should accompany each inspector base and skills. Homeowners can install PV
as field inspections are conducted. Communi- systems in some jurisdictions with no permits
cation between inspectors and plan reviewers or inspections other than what the utility might
on a real-time basis via cell phone and wireless potentially require. Unqualified electricians, who
computer link will be required. Using a digital have had no specific PV training, are installing
camera and downloading and transmiting on- PV systems. New “PV installation companies”
site pictures will be necessary. are popping up left and right to take advantage
of the boom in PV installations without being
Facts of Life fully trained to do those installations properly.
More Dangerous. Photovoltaic systems are As an aside, as I write this, I received a call from
potentially more dangerous than typical residen- an electrician who had been hired to install a
tial or commercial electrical power systems. With PV system for a residential customer who had
voltages up to 600 volts dc on residential instal- bought the equipment himself. The electrician
lations and up to a 1000 volts dc on commercial had absolutely no knowledge of how to do the
installations, the voltages are higher than those job or that there were any requirements in the-
of typical electrical systems. If energy storage NEC for such installations. It also appears that
systems are involved, currents may also be higher the customer may have purchased an unlisted
than those found on non-PV systems. Energy inverter on the Internet where all things are
storage systems in factory-sealed containers may available.
be operating at voltages up to 600 volts dc in
residences. AHJ Qualifications. We don’t like to admit it,
but just as with some PV istallers, some author-
More Complexity. Photovoltaic systems are ities having jurisdiction and local inspection
more complex than typical residential or com- requirements are not always up to the task of
mercial electrical power systems and that com- ensuring that installed PV systems are safe and
plexity is continually increasing due to technol- meet the latest Code requirements and safety
ogy changes and changing Code requirements. standards. Proper training and education for local
There are now frequently changing requirements AHJs on the latest technology in PV systems is
for PV rapid shutdown systems to protect first critical for the safe operation of those systems for
responders (690.12). We have dc PV arc-fault cir- many years to come.
cuit interrupter requirements in 690.11. There are
also PV utility-interactive inverters with chang- The Electrical Specialist. There are AHJs in
ing requirements and designs due to new utility the inspection community who have come up
through the ranks as master electricians and have The Permitting and Inspection Process. The
additional training in PV systems. They have process for permitting and inspecting PV systems
been working with PV systems daily since the varies substantially throughout the United States.
beginning—some since as early as the mid-1980s. In some areas of the country, no inspections are
These people are the electrical specialist inspec- conducted on PV systems. In other areas, a per-
tors. These individuals are frequently involved in mit is nothing more than a form filled out noting
PV more than just at the inspection level. They the location of the electrical system and a paid
may be members of code-making panels for the fee. In some states, there are no inspections of
NEC; members of UL Standards Technical Pan- residential PV systems, and the utility is responsi-
els for PV equipment; active members in local ble for ensuring the quality of installations before
IAEI chapters and IAEI sections; and may even allowing the interconnection.
be writing articles in IAEI News. On the other hand, many jurisdictions wisely
require a full permitting package (with diagrams),
The Combination Inspector. At the other end a plan-review stage, and a detailed inspection be-
of the spectrum, the latest trend in the in- fore approving the system as NEC compliant. The
spection community is to hire a “combination” utility then performs their required inspections
inspector who will be responsible for enforcing prior to interconnection. After the system has
several disciplines, such as plumbing, electrical, been powered up, there may be a final inspection
mechanical, and building codes. The theory is to by the AHJ.
hire one inspector instead of two or more due to An expedited permitting process for PV
budget limitations and emphasis on other local systems less than 10 kW was developed by the
issues. This combination inspector may have Solar America Board of Codes and Standards
come from the plumbing industry and is now (now defunct) and a copy can be downloaded
inspecting not only plumbing systems, but also from the SolarABCS web site: http://solarabcs.
electrical power systems, including PV systems. org/about/publications/reports/expedited-permit/
This individual will typically not have a firm index.html.
background in the NEC, let alone the very Numerous other informative documents re-
specific safety requirements associated with PV lating to PV systems are available on the Solar-
systems. Combination inspectors should expect ABCs website at http://solarabcs.org/index.html
and have been provided with training in the Chapters 8, 9, and 10 cover the plan review
basic requirements of the NEC (its first four and inspection of PV system in greater detail.
chapters) and in basic electrical theory—and
where at all possible be mentored by a senior
electrical inspector through on-the-job training. Budget and Administrative Restrictions. In
Individuals who find themselves lacking many areas of the United States, funding is lim-
PV-system training or mentorship are typically ited for inspection organizations. Combination
not to be blamed. They are merely fulfilling an inspectors are being hired more frequently than
obligation placed on them by their employers. specialist electrical inspectors. Budgetary and
Consideration should be given by the jurisdic- manpower restrictions are forcing inspectors to
tion to hiring additional electrical specialist do more inspections per day and, in some cases,
inspectors or they could contract with profes- these restrictions allow only 15 minutes for an
sionals who specialize in PV systems where inspection. A thorough inspection of a PV system
possible. might require two or more hours, including plan
review. In many areas, this amount of time is just This prohibition is particularly troublesome
not available to the inspector. where many commercial PV systems are
Training is also an issue for AHJs who in- roof-mounted and 90% of the PV equipment
spect PV systems. Funding is frequently not that should be inspected is on the roof. In the
available, nor is time allocated to properly train southwestern United States, with common
the inspectors who will be working with the flat-roofed and low-sloped-roof buildings, PV
constantly changing requirements in the Code for systems (and sometimes inverters) are frequently
PV systems. To ensure the safety of the public, mounted on rooftops. Of course, with ac PV
inspectors should at least be as familiar with PV module system or microinverters, nearly all
systems as PV installers and PV system designers. systems that should be inspected are found on
This familiarity requires periodic and continuing rooftops.
training for inspectors.
It should be noted that PV systems and PV Resources
equipment are becoming complex to the extent There are many resources available to inspector
that instructions for installing these systems and plan reviewers. Most equipment manufac-
can no longer be fully included in the NEC. The turers have electronic downloadable PDF files
instructions for installation of that equipment of manuals that will be useful. Here are a few
will be found only in the manuals for the listed magazines (available in print and on the web)
equipment. These instructions can be, to say the that will enable inspectors and plan reviewers to
least, quite complex; far more complex than can keep abreast of the changing technologies.
be put in the NEC [110.3(B)].
Moreover, for unknown reasons, some juris- IAEI News. www.iaei.org/magazine
dictions have insurance policies that prohibit Solar Pro. www.solarprofessional.com
AHJs from going onto rooftops to inspect not Home Power. www.homepower.com
only PV systems but also HVAC equipment. Solar Today. www.ases.org
42 Chapter 2 PV Fundamentals and Calculations
Chapter 2 PV Fundamentals and Calculations 43
02
PV Fundamentals and Calculations
To understand how and why the Code is different meter (W/m2). This is an international constant
for PV systems compared with other electrical sys- and is near the average value of irradiance at sea
tems, it is necessary to first delve into the mysteries level on the surface of the earth. Modules are also
of what they are and how they operate. Inspectors rated at a standard module/cell temperature of 25
need to be familiar with the unique characteristics degrees Celsius (C) [77 degrees Fahrenheit (F)].
of PV modules because the ampacity calculations These two values of irradiance and temperature
and overcurrent device ratings don’t follow the are referred to as standard test conditions (STCs).
usual rules of thumb used in the electrical trades. When the module is exposed to these standard
Additionally, some PV installers may not be using test conditions (and connected to the correct
the proper calculations either. load), the module will produce the rated power at
a maximum power point voltage (Vmpp) and with
Current Sources a maximum power point current (Impp).
Photovoltaic modules are current sources of ener- During evaluation of the safety of the mod-
gy, as opposed to voltage sources such as 120-volt ule [performed by a nationally recognized testing
ac outlets in homes or offices or 12-volt dc batter- laboratory (NRTL), leading to the module being
ies in cars. The output of a PV module, in terms listed as required by NEC 690.4(D)], the ratings
of voltage and current, depends on the load placed of the module are also verified to be within some
on that module. The output is dc, and the load is percentage of the label values. The tolerance on
typically supplied by a utility-interactive inverter, the label values is usually 10 percent but may be
which changes the dc energy into ac energy. as low as 3 percent.
A PV module can be short-circuited indefi-
Standard Test Conditions nitely without damage. And, as will be shown in
Photovoltaic modules are rated for voltage and later chapters, the wiring, the switchgear and the
current output when exposed to a set of standard overcurrent protection are designed in a way that
test conditions. The standard solar intensity will allow entire PV arrays to be short-circuited
(called irradiance) is set at 1000 watts per square without damage.
Ambient Temperature output of a PV module or array that is not

Although Section 310.15(B)(3)(c) has an Infor- connected to any load, the voltage obtained will
mational Note on a source for ambient tempera- be the open-circuit (no load) voltage (Voc). A
tures, many people will just use the local weather current measurement would be zero for this
station data. open-circuit condition.
“Informational Note: One source for the If simultaneous voltage and current measure-
ambient temperatures in various locations is ments are taken on a PV module or a PV array
the ASHRAE Handbook—Fundamentals.” and these measurements are plotted for various
The temperature adjustments for raceways in loads, a graph that shows the electrical character-
sunlight were substantial in the 2014 NEC, but istics of a PV module could be shown. The graph
the requirements were reduced in the 2017 NEC. would have current (I) on the vertical axis and
When addressing PV circuits on the roofs of voltage (V) on the horizontal axis. This graph or
buildings, consideration should be given that the plot is shown in Figure 2.1 and is called an I-V
local weather station may be at an airport outside curve. Similar curves appear in data sheets for PV
the city and could be in an open area. The PV modules.
system may be installed in the city where the city The point to the right on the horizontal axis
heat bubble may increase ambient temperatures is the open-circuit voltage (VOC) and the current
over the weather station’s measurement. On a at this point is zero. On the vertical current axis,
flat-roofed building with parapets, limited airflow the curve intersects the axis at the short-circuit
may increase ambient temperatures on the roof current (ISC) where the voltage is zero.
even more. These increases are before the applica-
tion of any solar heating on conduits and should Power Equals Voltage
be considered when considering the 50-year Multiplied by Current
lifespan of a PV system in this environment. Each point on the I-V curve represents a value
What about the conductors and raceways be- of voltage and a value of current at a particular
hind those hot PV modules? If these conductors load. Multiplying the voltage (V) by the current
or raceways are within 3–4 inches of the backs of (I) will calculate the power (P) produced by a
the PV modules, an ambient temperature of 75°C module and delivered to the load.
has been used in many parts of the country. If the P=VxI
conductors or raceways are spaced further away, From this relationship, it can be seen that the
then a 65°C ambient temperature may apply. Lo- module delivers no power at either the open-cir-
cal situations may dictate different temperatures. cuit voltage point or at the short-circuit current
point because one of the factors of power is zero
at these points.
Current and Voltage However, if other points are examined on the
Measurements — The I-V Curve curve (producing power from different loads),
Measuring the module or array output under it will be noted that the power is not zero at
short circuit conditions will allow measurement these points. If the power output curve is added
of the short-circuit current (ISC), which will be to the I-V curve, the graph shown in Figure
used in PV system sizing and in many Code 2.2 is obtained, which includes the I-V curve in
calculations. A voltage measurement under blue and the power curve in red. The horizontal
short-circuit conditions will yield zero volts. axis for the combined graph is still volts, but
If a voltmeter is used to measure the voltage the vertical axis (on the right) for the power
Figure 2.1 • I-V curve for a PV module.
Figure 2.2 • I-V and power curve.

Figure 2.3 • I-V curve variations due to varying sunlight.
Figure 2.4 • I-V curve variations due to varying temperature.

curve is now marked watts. The curve for power rent is relatively insensitive to temperature, but
shows that it reaches a peak for some load both the VOC and Vmpp voltages will be affected. In
between ISC and VOC and this point is called the crystalline PV modules, VOC varies inversely with
maximum power point (Pmpp). Associated with temperature at approximately 0.5% per degree
the power at the maximum power point are Celsius. Peak-power voltage (Vmpp) varies in-
a maximum power point voltage (Vmpp) and a versely at approximately 0.4% per degree Celsius.
maximum power point current (Impp). Figure 2.4 shows this relationship; as temperature
It should be noted that the output voltage of goes down, voltage increases.
a PV module is not constant and varies with the
load. This output is changed by several external Turn Off the Light
environmental conditions in addition to the Photovoltaic modules in utility-interactive sys-
connected load. tems are connected in series and the open-circuit
voltage may approach 600 volts (dwelling), 1000
Sunlight Produces Current volts (commercial), and 1500 volts (utility-scale)
The current output of a PV module is directly in cold weather. The only way to effectively turn
proportional to the intensity of the sunlight off all electricity from a PV module or a PV array
falling on it. The rated currents (both ISC and is to cover it with an opaque material. Working
Impp) are output at the standard test-condition at night on array wiring is an option, but worker
irradiance of 1000 W/m2. However, PV modules safety would be a concern. Lightning in the
are exposed to irradiance values from 0 (at night) distant sky has been known to illuminate arrays
to 1500 W/m2 (if cloud, water, snow, or sand enough to produce electric shocks. High-inten-
enhanced) and the current follows changes in the sity lighting used by fire services may also create
intensity of the sunlight. A 10% reduction in the unsafe voltages in PV arrays.
irradiance value will result in a 10% reduction in
ISC and Impp. However, the open circuit voltage Determining the PV Size and
(VOC) is relatively unchanged with variations in Output to Match the Inverter
irradiance. Figure 2.3 shows the I-V curves for a In the PV design process, the array output must
PV module as the sunlight intensity varies from be matched to the utility-interactive inverter
1000 W/m2 down to 500 W/m2. As can be seen, input. As noted previously, the module voltage
the ISC changes in direct proportion to changes in and current outputs are not constant. The system
irradiance but VOC and Vmpp do not vary nearly as designer must first ensure that, in cold weather
much. conditions, the array output voltage will not
This is a significant fact. The voltage on a PV exceed the voltage rating of the inverter, the con-
module or PV array will generally be present at very ductor insulation, or other connected equipment.
low levels of light, such as at dawn or dusk. Pho- This limit will frequently be as high as 600–1500
tovoltaic arrays can have hundreds of volts on the volts. In a similar manner, all conductors, over-
wiring at dawn and dusk even before the sun direct- current devices, and switchgear must be able to
ly illuminates the front of the modules. Hazardous handle the array output current under worst-case
voltages on exposed terminals in dc combiners, conditions of high sunlight intensity. Exceed-
disconnects, and input terminals to inverters will be ing the voltage rating on inverters, conductors,
present. switchgear, or other equipment would be a Code
A second variation in module output and the violation and has been known to damage such
I-V curve is caused by temperature. Module cur- equipment. [See NEC 110.3(B).]
Early pioneers in PV, including PV mod- Adjustments—Short-Circuit

ule manufacturers, NASA’s Jet Propulsion Current
Laboratories (space programs), Underwriters The first adjustment is made to the module
Laboratories, and the National Fire Protec- short-circuit current. Until May 2012, UL Stan-
tion Association, realized that because of dard 1703 required that module manufacturers
these variations in the output of PV modules, have the following statement (or its equivalent)
special consideration would need to be made in instruction manuals that are shipped with
to handle these systems in the National Elec- every PV module.
trical Code (NEC). “Multiply the short-circuit current (ISC) marked
UL Standard 1703, Standard for Flat-Plate on the back of the module by 125% before apply-
Photovoltaic Modules and Panels, was written to ing any requirements established by the NEC.”
establish safety requirements (mechanical and This 125% factor is applied to the short-cir-
electrical) that PV modules would be required cuit current to ensure that the current used in
to meet. Until around 2012, requirements any Code ampacity or overcurrent device rating
in module instruction manuals were written calculations will never be exceeded by the actu-
that modified the values of the standard test al module output under high-irradiance condi-
tions. A 125% factor on ISC would be equivalent
condition parameters printed on the back of
to exposing the module to 1250 W/m2 of
modules.
irradiance, and that value of solar intensity has
These modifications dealt with the fact that
never been measured on a continuous basis.
sunlight intensity in many parts of the coun-
Transient values of sunlight may reach 1500
try could exceed the standard test condition
W/m2 due to cloud reflectivity enhancements
of 1000 W/m2 for three hours of more on
and reflections off sand, water, and snow. How-
numerous days throughout the year. Yes, PV
ever, these conditions are short-term because
modules can deliver more than rated current
the clouds move, as does the sun (through earth
and rated voltage as environmental conditions
rotation). These are not stable situations and last
vary.
for only a few minutes.
Three hours per day represent continuous
Instructions for most flat-plate PV modules pro-
duty in the Code. It was decided in the early hibit the installation of any permanent light-reflect-
days of PV systems that Code calculations ing or light-concentrating surface near the modules.
should be based on worst possible outputs and
that those outputs would be considered to be
Adjustments — Open Circuit
continuous 24 hours per day, 7 days per week,
Voltage
52 weeks per year and not vary on a daily As noted previously, open-circuit voltage (VOC)
cycle with the sun. Therefore, all voltages and varies inversely with temperature. Instruction
currents used in calculations for PV systems manuals for PV modules have the following
in the Code are adjusted from standard test statement or its equivalent:
condition measurements in a way that ensures “Multiply the open-circuit voltage (VOC)
electrical systems meet Code requirements for marked on the back of the module by 125%
safety and that all equipment will be operated before applying any requirements established by
within limits established by both the Code and the NEC.”
UL standards. For crystalline silicon, a factor of 125% on VOC
represents the open-circuit voltage of a module been modified to remove these two require-
at -40°C (-40°F). Temperatures lower than -40° ments and only the Code requirements will
are found in some parts of the country and remain for PV modules being installed where
further adjustments to VOC may have to be made the NEC is in force. Advocacy from electrical
where these temperatures are expected. inspectors throughout the country inspired
these changes.
NEC Leads UL Standard 1703
In 1996, during preparations for the 1999 NEC, The 125% Continuous Loads
the PV industry, UL, and the National Fire Pro- Factor
tection Association reached a joint agreement to The requirement to use 125% of the continuous
place all module output correction factors in the load (or 125% of the maximum current in PV
Code. Those proposals were agreed to by all, and circuits) has been with us for many years. It was
the 1999 NEC was modified accordingly. reportedly added to the Code to address nuisance
In Section 690.8, a requirement was estab- operation of overcurrent devices in enclosures
lished to multiply the module short-circuit when conductors were operated at 100% ampac-
current by 125%. This duplicated the module ity continuously. The temperature rise associated
manual instruction to adjust the Isc to account with I2R losses in the conductors, and in the
for normal and expected irradiance values above thermal overcurrent devices themselves, was
1000 W/m2. causing some of these devices to overheat and
Because many areas of the country never nuisance trip or blow. Using the 125% factor
have temperatures as low as -40°, a tempera- results in a conductor that is not loaded contin-
ture-dependent table was added to Section uously at more than 80% of rating. Note that
690.7 showing multiplication factors for Voc 1.25 and 0.8 are mathematical reciprocals of each
that vary from l.00 at an expected low tem- other (1/1.25 = 0.8 and 1/0.8 = 1.25), a fact that
perature of 25°C (77°F) to 1.25 at -40°. Again, will be used below.
instructions in the module manual duplicated
these Code requirements. A Second 125%
Because newer module technologies have Throughout the NEC, feeders and branch circuits
different temperature coefficients than crystalline (as well as overcurrent devices) are rated to oper-
silicon, the 2017 NEC permits module manu- ate on a continuous basis at no more than 80% of
facturers’ temperature coefficients be used where rating. Section 215.2(A)(1) has this requirement,
available rather than Table 690.7. which is also found in other sections of the Code:
UL modified Standard 1703 in 2012 to
remove the 125% multipliers on VOC and ISC “(1) General. Feeder conductors shall have
from PV module instruction manuals. This an ampacity not less than required to supply
solved a significant problem for inspectors the load as calculated in Parts III, IV, and V
and installers who were required to follow of Article 220. Conductors shall be sized to
NEC section 110.3(B), which states that all carry not less than the larger of 215.2(A)(1)
equipment be installed in accordance with (a) or (b).
included instructions. If followed, this section
of the Code (and 690.7 and 690.8) create a (a) Where a feeder supplies continuous loads
double calculation with far too conservative or any combination of continuous and non-
results. Module instruction manuals have continuous loads, the minimum feeder con-
ductor size shall have an allowable ampacity Energized and Safe for
not less than the noncontinuous load plus 125 Decades
percent of the continuous load. It has been shown how module output varies
with environmental conditions. The Code-
(b) The minimum feeder conductor size shall required corrections to the rated output are used
have an allowable ampacity not less than the to ensure that system electrical equipment never
maximum load to be served after the applica- has to handle more current or voltage than it
tion of any adjustment or correction factors.” was designed for. The rated output has been
adjusted for worst-case conditions under the as-
The intent of this code requirement, other than sumption that these worst-case conditions exist
the obvious rating factors, is that the 125-percent continually, even when it is known that output
factor (a) is not to be applied at the same time goes to zero every night. Photovoltaic modules
the “conditions of use” factors—such as tempera- will produce dangerous amounts of voltage and
ture and conduit fill corrections on ampacity—in current for 40 years or more. These adjustments,
(b) are applied. From a math point of view 1.25 and the other Code requirements, represent
(125%) and 0.8 are reciprocals, so they can be minimums to ensure that this sunlight-generat-
used two ways. The conductor size can be calcu- ed electricity is safely contained for that period
lated by taking 1.25 times the continuous load of time or longer.
(maximum current for a PV or other generator)
or by taking 0.8 times the conductor ampacity
(before adjustments for conditions of use) to find
PV Math—Some Sample
the maximum continuous currents.
Calculations
Looking at the PV array in a PV system, many
This same 125-percent factor (or the condi-
installers and inspectors are confused by new
tions-of-use factors if they result in requiring a
system voltage calculations that may be required
larger conductor) is applied to PV source and
by the Code specific to PV systems. Code Infor-
output circuits, so the dc PV source circuit or dc
mational Notes also address voltage drop that
output short-circuit current might be multiplied may be applied to the dc wiring from the array to
by 1.25 twice before selecting a conductor or the inverter. This section will cover both of those
overcurrent device. A factor of 1.56 (1.25 x 1.25 subjects.
= 1.56) is commonly used to determine ampac-
ity, but this is a shortcut and the slightly more PV Math—Module Open-Circuit
complex actual calculations will be addressed in Voltage
below. A PV module, or a string of series-connected
The VOC should be calculated following the modules, has a rated open-circuit voltage that
requirements of 690.7. The factors will be less is measured (and labeled on the module) at an
than 125 percent unless very low temperatures irradiance of 1000 W/m2 and a cell temperature of
(-40°) are expected at the installation location. 25°C (77°F). This voltage increases from the rated
Table 690.7(A) may be used, but it will yield voltage as the temperature drops below 25°C. It is
more conservative values for the cold weather necessary to calculate this voltage at the expected
maximum system voltage, which, in many cases, lowest temperature at the installation location to
may not result in the most cost-effective system ensure that it is less than the maximum input volt-
design. age of the inverter and less than the voltage rating
of any connected conductors, switchgear, and over- add to the confusion, PV module manufacturers
current devices (usually 600 volts). Because parallel present these temperature coefficients in two
connections of strings do not affect open-circuit different ways.
voltage, the number of strings connected in parallel
is not involved with this calculation. Percentage Coefficients
Where module temperature coefficients are One way of presenting these data is to specify
available, Section 690.7 of the NEC allows them as a percentage change, and they are ex-
the open-circuit voltage of a PV array to be pressed as a percentage change in VOC for a change
determined at the lowest expected temperature in temperature measured in degrees Celsius. Note
at the installation location. Alternatively, Table that the temperature used is a change in tempera-
690.7(A) can be used to determine a multiplier ture from the rated 25°C.
that was applied to either the module- or string- For example: The Voc temperature coefficient is
(a series connection of PV modules) rated VOC. given as -0.36% per deg Celsius, or -0.36% / °C.
The rated VOC is measured at 25°C (77°F) and The module has a VOC of 45 volts at 25°C
is printed on the back of the module and in (77°F) and is going to be installed where the
the module’s technical literature. To use Table expected lowest temperature is -10°C (14°F).
690.7(A), determine the lowest expected tem- Because the temperature coefficient is given in
perature, look up the factor from the Table for degrees Celsius, all numbers must be in degrees
that temperature (which ranges between 1.02 at Celsius. The change in temperature is from 25°C
24°C to 1.25 at -40°C), and multiply the factor to -10°C. This represents a change in tempera-
by the rated VOC. ture of 35 degrees. The minus sign in the coef-
For example, a module has a VOC of 35 volts ficient can be ignored if we remember that the
(at STC) and is going to be installed where the voltage increases as the temperature goes down
temperature dips to -17°C. The factor from Table and vice versa. Of course, if you are an engineer
690.7(A) is 1.18 and the cold temperature VOC or a mathematician, feel free to use the minus
for this module is 35 x 1.18 = 41.3 volts. sign in an algebraic equation.
If 12 modules were going to be connected in Applying the coefficient shows that the per-
series, the string VOC in cold weather would be 12 centage change in VOC resulting from this tem-
x 41.3 = 495.6 volts. perature change is 0.36% / °C x 35°C = 12.6%.
The string voltage at STC-rated conditions Note that where division and multiplication
could also be calculated first and then apply the are involved in a calculation, they are performed
temperature factor applied. In this case, the 12 from left to right. Where additions and sub-
modules in series would have a string open-cir- tractions are combined with multiplications
cuit voltage of 12 x 35 = 420 volts at 25°C. Then and divisions, the additions and subtractions are
the 1.18 factor is applied to get 1.18 x 420 = performed before any multiplications. Paren-
495.6 volts; the same answer as before. theses may be added to clarify the order of the
While Table 690.7(A) is still valid and was re- operations, and calculations inside the parenthe-
fined with 5°C increments in the 2011 NEC, new ses should be performed first.
modules may have different technologies than the This percentage change can now be applied to
silicon module technology used to develop the the rated VOC of 45 volts. And, at -10°C, the VOC
table. The use of the module temperature coef- will be 1.126 x 45 = 50.67 V.
ficients will provide a more accurate calculation For the mathematically oriented person, the
of the cold weather maximum system voltage. To equation looks like this:
VOC [cold] = VOC [STC] x (1 + ((-0.36)/100) x 550 / 78.2 = 7.03 modules and the
(-10-25))) = 45 x (1 + .0036 x 35) = 45 x 1.126 = correct answer would be seven modules.
50.67
7 x 78.2 V = 547.4 V
Eleven of these modules could be connected in
series and the cold-weather voltage would be 11 Eight modules could not be used because the
x 50.67 = 557.37 V, and that voltage is less than a open-circuit, cold-weather voltage would exceed
600-volt equipment limitation. 550 volts.
Millivolt Coefficients 8 x 78.2 V = 625.6 V

Other PV module manufacturers express the Voc
temperature coefficient as a millivolt coefficient. Expected Lowest Temperature?
A millivolt is one one-thousandth of a volt, or What is the source of the expected lowest tem-
0.001 V. perature? Normally, this temperature occurs in
A typical module with an open-circuit voltage the very early morning hours just before sun-
(at 25°C) of 65 volts might have a temperature rise on cold winter mornings. The PV modules
coefficient expressed as: are, in many cases, a few degrees colder than
the air temperature due to night-sky radiation
-240 mV per degree Celsius, or -240 mV/°C effects. The illumination at dawn and dusk are
sufficient to produce high VOC, even when the
If it is installed where the expected low tem- sun is not shining directly on the PV array and
perature is -30°C (-22°F), then there is a 55°C has not produced solar heating of the modules.
degree change in temperature from 25°C to And in many locations, cold temperatures are
-30°C. Again, the calculations must be accom- coupled with high winds and the winds can
plished in degrees Celsius, ecasue that is the way remove solar heating from the module, even in
the coefficient is presented. bright sun.
Millivolts are converted to volts by dividing the A conservative approach would be to get
millivolt number by 1000. weather data showing the record-low tempera-
ture of the area and use this as the expected low
240 mV / 1000 mV/V = 0.24 volts temperature. Other data show more moderate
low temperatures associated with the data used
The module VOC will increase 0.24 V/°C x 55°C to size heating systems. See the Informational
= 13.2 volts as the temperature changes from Note on 690.7(A).
25°C to -30°C. The National Renewable Energy Laboratory
The module VOC will increase from 65 volts maintains data that shows the record lows for
at 25°C to 65 + 13.2 = 78.2 volts at the -30°C many locations in the United States (http://
temperature. rredc.nrel.gov/solar/old_data/nsrdb/1961-1990/
In this PV system, the inverter maximum redbook/mon2/state.html).
input voltage was listed as 550 volts. How many Local airports and weather stations may have
modules could be connected in series and not historical data on low temperatures.
exceed this voltage? The maximum inverter volt- Also, weather.com has data on file accessble by
age of 550 volts is divided by the cold-weather ZIP codes (www.weather.com/weather/climatolo-
open-circuit voltage for the module of 78.2 volts. gy/monthly/zip code).
Photo 2.1 • DC-to-dc converter. Solar Edge Power Optimizer. Courtesy Solar Edge.
PV Math—Module Short-Circuit near the peak-power voltage (also called the

Current maximum power point). While this voltage can
In most silicon PV modules, the module vary with temperature—and temperatures vary
short-circuit current does increase very slightly as considerably—using the rated maximum power
temperature increases, but the increase is so small point voltage and maximum power point current
as to be negligible at normal module operating of the modules results in the easiest method of
temperatures. It is normally ignored. calculating voltage drops.
A typical PV array may have a single string of
Informational Notes—Voltage Drop ten modules in series connected to the inverter
In common, utility-interactive PV systems, PV 200 feet away with 10 AWG USE-2/RHW-2
arrays may operate from 50–60 volts up to near conductors. The maximum power point (mpp)num-
600 volts, depending on the system design. With bers for the module are:
nominal, peak-power, and open-circuit voltages
to deal with, installers and inspectors are some- Vmpp = 45V
times in a quandary as to how to calculate voltage Impp = 5.5 amps
drops from PV arrays to the inverters.
A utility-interactive inverter will normally For a single string of 10 modules, the string
operate in a manner that keeps the array voltage maximum power point numbers are:
Vmpp = 450V data when it is available. Normal cell operating

Impp = 5.5 amps temperatures are typically in the range of 42°C
to 52°C and is a measured temperature of the
Table 8 in Chapter 9 of the NEC gives conduc- module when the irradiance is 800 W/m2, ambi-
tor resistance per 1000 feet at 75°C. ent temperature is 20°C and the wind is blowing
For an uncoated, stranded 10 AWG conductor, across the module at 1m/sec. This temperature
the resistance is 1.24 ohms per 1000 feet. is used to adjust the Vmpp and Impp numbers
The total conductor length (both ways) must be presented at standard test conditions to values
used in the calculation and this is 400 feet. associated with the higher normal cell operating
The resistance for 400 feet of a 10 AWG con- temperature. The maximum power voltage will
ductor is 400/1000 x 1.24 = 0.496 ohms. be reduced and the maximum power current will
The current at the maximum power point is 5.5 be increased very little if at all. Voltage drop will
amps. Voltage drop is found by multiplying this be reduced.
current by the conductor resistance:
5.5 x 0.496 = 2.728 volts. Module-Level Power

Expressed as a percentage, 2.278/450 x 100
Electronics—When a PV
= 0.606% or about 0.6%, and that is much less Module Does Not Act Like a PV
than the Informational Note recommendation of Module
3% for most circuits. Of course, the losses in the Module-level power electronics (MLPEs) are
PV dc disconnect were not counted, but they are found within PV module junction boxes and in
typically less than 1% on these circuits. dc-to-dc converters connected to the module
output leads at the module (photo 2.1). Module-
level power electronics decouple the standard
Voltage Drop — Another test condition ratings of the PV module and
Perspective the module output characteristics from the
When dealing with the dc input circuits of a conductors leading to the rest of the string of
utility-interactive inverter, the inverter operates modules and the PV array and the inverter.
the array at the maximum power point with a The calculations shown above dealing with the
maximum power voltage (Vmpp) and a maximum PV module characteristics (primarily voltage,
power current (Impp). Because these parameters current, and power), are no longer applicable to
are affected by irradiance and temperature, it is the output of a module with module-level power
difficult to determine what voltage and current electronics.
should be used in the voltage drop equation. Module-level power electronics perform several
The procedure above uses module specifications functions, which vary from product to product.
at standard test conditions that are based on the Some will perform PV rapid shutdown system
rating conditions of an irradiance of 1000W/m2 actions (690.12) using a remote signal. Others
and a cell temperature of 25°C (which is not a will peak-power track the individual module and
very realistic temperature under actual operating provide an output that gets as much power from
conditions). One conservative approach is to individual modules as possible given the envi-
use the module manufacturer’s specified normal ronmental parameters. Still, other MPLEs will
cell operating temperature from the module attempt to maximize the output of the module
while also maximizing the output of the string temperature relationships of the output character-
containing that module. In nearly all cases, the istics of a normal dc PV module.
specifications relating to the output parameters of
the MLPE will not be comparable to the output
of the basic PV module. In nearly all cases, the Conductor Sizing and
manufacturer’s instructions provided with the Overcurrent Device Ratings
listed or certified module-level power electronics Inspectors or plan reviewers should be very
device must be followed (110.3 B). There is no familiar with the methods of calculating currents,
possible way that the NEC will be able to address conductor sizes, and overcurrent device protection
these numerous varying operating parameters and required in PV systems because, in many cases, PV
installation requirements for the numerous systems installers will not have performed these calculation
in a general manner. in a manner that follows NEC requirements.
For example, when a certain dc-to-dc converter Historically, most residential and light com-
is connected to an appropriately sized module, mercial electrical wiring and inspections of these
the maximum voltage output of this converter systems have involved indoor wiring at room
is rated at 60 volts. Moreover, on the surface, it temperatures [30°C (86°F) or less]. The am-
would appear that (at most) ten of these devices pacity tables in NEC Section 310.15 and Table
could be connected in series to a utility-interac- 310.15(B)(16) were developed with those condi-
tive inverter with a dc input rated at a maximum tions in mind. The commonly used molded-case
of 600 volts. However, the manufacturer’s litera- circuit breaker is rated for use with conductors
ture suggests that, optionally, 15 of these devices with 75°C insulation and they have a rated maxi-
could be connected to that 600-volt inverter— mum operating temperature of 40°C.
because the inverter from that manufacturer With these conditions and equipment charac-
communicates with each of the dc-to-dc convert- teristics in mind, some electricians have generally
ers and holds its output under actual operating used the 75°C insulated conductor ampacity
conditions to a total voltage per string of less tables in Table 310.15(B)(16) and not bothered
than 600 volts. too much with temperature corrections [Table
While this particular dc-to-dc converter is 310.15(B)(2)(a)] and terminal temperature limits
connected to a PV module that has the normal [110.14(C)] because they were not necessary or
PV module ratings on the label, they are to be were included in the tables being used.
essentially ignored and the ratings that apply to the However, dc PV conductors normally operate
dc-to-dc converter are to be used in determining in an environment that is too hot for conductors
whether it and the PV module have been installed with 75°C insulation. Conductors with 90°C
correctly. Some dc-to-dc converters have a very low insulation must be used, and appropriate tem-
output (one or two volts) when they are discon- perature and conduit fill corrections must be
nected from the inverter or from communications applied along with verification that connected
with the inverter. Of course, when the MLPE is equipment terminal temperatures (60°C or
fully contained within the module junction box, 75°C) are not exceeded. To do otherwise and
the certified or listed PV module with its internal use the short-cuts of the old days will result
electronics will have a label showing the output in conductors that may be larger than Code
characteristics of the combination. The output requirements (resulting in unnecessary costs) or
characteristics of the module-level power electron- that may have inadequate ampacities under the
ics may, in many cases, no longer follow the classic extreme conditions of use.
Throughout the Code, circuits are sized based Correction Factors. The maximum currents
on 125% of the continuous load plus the non- calculated in 690.8(A) after the application
continuous load. See 210.19(A)(1) and 215.2(A) of adjustment and correction factors.”
(1). This requirement establishes a situation
where conductors and overcurrent devices are Note for the Exception: In other sections of
not subjected to more than 80% of rating. (Note: the Code, such as 215.2(A)(1)(a), this exception is
1/1.25 = 0.80.) expanded to indicate that with this type of over-
Electricians typically use the 125% factor and current protection, the conductor ampacity can
then also apply the conditions of use factors be rated at 100% of the continuous currents (or
(temperature and conduit fill) sequentially. In maximum currents in the case of PV circuits). The
recent editions, the NEC has clarified the re- author is not aware of any dc overcurrent protective
quirements in 210.19(A)(1) and 215.2(A)(1), as device in an assembly rated for 100% operation.
well as in 690.8, so that both factors are not to be
applied at the same time to the same circuit. See 215.2(A)(1)(a) Exception No. 1: If the assembly,
the 125 percent requirement below. including the overcurrent devices protecting the
In the Code, we have at least two or three re- feeder(s), is listed for operation at 100 percent of
quirements that must be met in sizing conductors. its rating, the allowable ampacity of the feeder
First is the definition of ampacity found in Ar- conductors shall be permitted to be not less than
ticle 100. Ampacity is “The maximum current, in the sum of the continuous load plus the noncontin-
amperes, that a conductor can carry continuously uous load.
under the conditions of use without exceeding its
temperature rating.” Author’s Note: This book will not address the
Next is the 125 percent requirement, or condi- 2017 NEC requirements related to “adjustable
tions of use factors. From 690.8(B), Conductor electronic overcurrent protective devices” found in
Ampacity: (In part): 690.8(B), 690.8(B)(3), and 690.9(B)(3) because
(B) Conductor Ampacity. PV system currents these devices and their requirements are not well
shall be considered to be continuous. Circuit defined in either the 2017 NEC or in the relevant
conductors shall be sized to carry not less than UL Standards. At best, they are adjustable trip
the larger of 690.8(B)(1) or (B)(2)…. circuit breakers and are handled like other circuit
breakers. Revisions, clarifications, or deletions
“(1) Before Application of Adjustment and to these requirements are expected in the 2020
Correction Factors. One hundred twen- NEC.
ty-five percent of the maximum currents
calculated in 690.8(A) before the application Terminal Temperature
of adjustment and correction factors. Limitations and Operating
Temperature Limitations
Exception: Circuits containing an assembly, Section 110.14(C) requires that the temperature
together with its overcurrent device(s), that is of the conductor in actual operation not exceed
listed for continuous operation at 100 percent of the temperature rating of terminals on the con-
its rating shall be permitted to be used at 100 nected equipment.
percent of its rating. An added requirement for any listed equip-
ment, such as overcurrent devices, is that they not
(2) After Application of Adjustment and be used in a manner that deviates from the listing
or labeling on the product [110.3(B)]. Most PV be 24.3 x 1.25 = 30.375 amps [690.8(A)(2)].
combiners operating outdoors in the sunlight will An exception to these calculations is allowed
have internal temperatures that exceed the 40°C for PV systems of 100 kW or greater, where a
rated operating temperatures of commonly used professional electrical engineer may make al-
fuses and circuit breakers. Overcurrent devices ternate maximum current calculations using an
listed for (and required in) PV applications will industry accepted method.
have a 50°C-rated operating temperature.
B. AC inverter output circuits
Calculating Conductor Sizes and In the ac output circuits of a utility-interactive
Overcurrent Device Ratings inverter (or in the ac output circuit of a stand-
The following method for determining ampacity alone inverter), the continuous current is taken
meets the three Code requirements above and at the full rated output power of the inverter. It
finds the smallest conductor that can be used to is not measured at the actual operating current of
meet these requirements. The method also deter- the inverter (which may be a small fraction of the
mines the rating of the overcurrent device where rated current due to a small PV array connected
required. It is consistent with the requirements to a large inverter). Usually the rated current is at
found in Article 690 and elsewhere in the Code. the nominal output voltage (120, 208, 240, 277,
or 480 volts). The rated output current is usually
Step 1. Determine the maximum circuit specified in the manual, but may be calculated by
current [690.8(A)] dividing the rated power by the nominal voltage.
For stand-alone inverters, which can provide
Photovoltaic dc circuits and PV ac circuits are some degree of surge current, it is the rated power
not “load” circuits; the Code uses the term current that can be delivered continuously for three hours
instead of load. For Code calculations, all dc and or more [690.8(A)(3)].
ac PV currents are considered continuous and In some cases, the inverter specifications will
are based on worst-case outputs or are based on give a rated current that is higher than the rated
safety factors applied to rated outputs [690.8(B)].
power divided by the nominal voltage. In that
The term maximum is used in Article 690 instead
situation, the higher current should be used. This
of continuous current used elsewhere in the Code.
higher current usually has been determined at a
Daily variations in these currents is ignored.
lower-than-normal line voltage.
For a utility-interactive inverter operating at a
A. Photovoltaic dc circuits
nominal voltage of 240 volts and a rated power
In the dc PV source and dc PV output circuits,
of 2500 watts, the continuous current would be
maximum currents are defined as 1.25 times the
2500 W/240 V = 10.4 A.
rated short-circuit current (ISC) marked on the
A stand-alone inverter with a model number of
back of the module. If a module had an ISC of 7.5
3500XPLUS operates at 120 volts and can surge
amps, the maximum current would be 7.5 x 1.25
to 3500 watts for 60 minutes. However, it can
= 9.375 amps [690.8(A)(1)].
only deliver 3000 watts continuously for three
If three strings of modules (module ISC = 8.1
amps) were connected in parallel through a fused hours or more. The rated output current would be
dc combiner, the PV output circuit of the com- 3000 W / 120 V = 25 A
biner would have an ISC of 3 x 8.1 = 24.3 amps C. Stand-alone inverter battery currents
[690.8(A)(2)] and the maximum current would In either a stand-alone system or a battery-
backed-up utility-interactive system, the currents Step 1.

between a battery and an inverter must be based In our example, 1.25 x 50 = 62.5 amps, and this
on the rated output power of the inverter (con- would indicate a 6 AWG conductor.
tinuous for three hours or more) at the lowest
input battery voltage that can provide that output B. Conditions-of-use requirement
power [690.8(A)(4)]. The conductor, after corrections for conditions
This current is usually marked on the inverter of use, must have an ampacity equal to or greater
or found in the specifications, and that number than the continuous current found in Step 1.
should be used where available.
When not available, the input current can be Temperature correction factor = 0.91
calculated by taking the rated output power, Conduit fill correction factor = 1.0
dividing it by the lowest battery voltage that can
sustain that power, and also by dividing by the Required ampacity at 40°C is 50 / 0.91 / 1.0
inverter dc-to-ac conversion efficiency at that = 54.9 amps, and this would require an 8 AWG
battery voltage and power level. For example: conductor.
A 4000-watt inverter can operate at that power
with a 44-volt battery input voltage and has a The 6 AWG conductor is the larger of the two
dc-to-ac conversion efficiency (inverting mode) and is required.
of 85 percent. The dc continuous current will be
4000 W / 44 V / 0.85 = 107 A Example 2: Now there are six conductors in
the conduit and the temperature has increased to
On single-phase stand-alone inverters, the dc 50°C. The maximum current is still 50 amps.
input current is rarely smooth and will usually
have 120 Hz ripple current (root mean square Temperature correction factor = 0.82
[RMS]) that is larger than the calculated contin- Conduit fill factor = 0.8
uous current. The inverter’s technical specifica-
tions should list the greatest continuous current. Conditions-of-use requirement: 50 / 0.8 / 0.82
= 76.2 amps, and a 4 AWG conductor is needed.
Step 2. Calculate a conductor size [690.8(B)]. 125-percent rule: 1.25 x 50 = 62.5 amps calling
The conductor selected for any circuit must for a 6 AWG conductor.
meet the 125 percent requirement or the con-
ditions of use requirement. The correctly sized The 4 AWG conductor is the larger of the two
conductor is the larger of A or B below. and must be used.
Example 1: Three conductors are in a conduit Step 3. Terminal temperature limits[110.14(C)]

in a boiler room where the temperature is 40°C. The terminal temperature limits marked on the
The maximum current in all three conductors is equipment must be used. If no temperatures are
50 amps. A copper 90°C insulated conductor is marked, then a 60°C limit is used for circuits rat-
specified. ed at 100 amps or less or conductors 14-1 AWG.
For circuits rated greater than 100 amps and for
A. 125 percent requirement conductors greater than 1 AWG, a 75°C terminal
The conductor must have an ampacity of 125 temperature limit will be used [110.14(C)].
percent of the maximum current established in The following method is a terminal tempera-
ture estimation method and is not an ampacity Example 3. Take the 6 AWG conductor and
calculation method. It is used after the conduc- 50 amps of maximum current used in Example 1
tor size has been selected based on the ampaci- above. This conductor is connected to a terminal
ty calculation. with a 60°C marking.
Take the conductor size in Step 3 above. Find From Table 310.15(B)(16), a 6 AWG conduc-
the lowest terminal temperature limit for this tor in the 60°C column can carry a current of 55
conductor at any termination. Use that termi- amps.
nal temperature limit (either 60°C or 75°C) 1.25 x 50 = 62.5 amps. This is larger than the
to enter the ampacity per Table 310.15(B) 55 amps from the Table and this terminal will be
(16). For the conductor size selected, read out heated above 60°C.
the current in the correct column, either the When we increase the conductor size to 4
60°C column or the 75°C column. There are no AWG, the table gives us 70 amps, which is
temperature adjustments or conduit fill adjust- greater than 62.5 amps and the terminal will stay
ments in this estimation process. below 60°C.
The current from the table must be equal to
or greater than 125 percent of the maximum Example 4. Use the 4 AWG conductor select-
current. If the conductor meets this requirement, ed in Example 2 connected to a terminal with a
then the terminal temperatures are going to be 75°C temperature limit. The maximum current is
less than the 60°C or 75°C limit for that conduc- 50 amps.
tor and that maximum current. The 125-percent 1.25 x 50 = 62.5 A
factor accounts for many items not calculated in
this simplified temperature estimation process. A 4 AWG conductor in the 75°C column of
Photo 2.2 • Multiple small raceways routed across the building.

Table 310.15(B)(16) shows a current of 85 amps. and overcurrent devices listed for PV applications
Because this is greater than 62.5 amps, the con- are rated for use at 50°C. Photovoltaic combiner
ductor will operate cooler than the 75°C terminal boxes operating in outdoor environments may
temperature limit. No increase in conductor size experience ambient temperatures as high as
is necessary. 50°C. Exposed to sunlight, internal temperatures
An alternate, more conservative approach to may reach or exceed 55°C to 60°C. Anytime the
terminal temperature limitations would be to use an operating temperature of the overcurrent device
ampacity table of 60°C or 75°C for the conductor exceeds 50°C, it may be subject to nuisance trips
size calculations. However, this will result in a con- at current values lower than its rating. In these
ductor larger than necessary because the advantages situations, the manufacturer must be consulted
of the 90°C conductor insulation are not considered. to determine an appropriate derating. At high
operating temperatures, an overcurrent device with
Step 4. Calculate the rating of the overcurrent a higher rating will activate at the desired current;
device, where required. however, using a manufacturer’s certification of
a rating at a higher-than-listed temperature may
Because PV modules are current-limited, invalidate the listing on the device.
overcurrent devices are frequently not needed for
one or two strings of PV modules connected in Step 5. Verify that the overcurrent device pro-
parallel [690.9(A) Exception]. In systems with tects the conductor selected under the conditions
three or more strings of modules connected in of use
parallel, overcurrent devices are usually required. Where an overcurrent device is required, it
must protect the conductor under conditions
A. Rating determined from maximum currents of use. Conductors may be protected using the
The overcurrent device rating is determined round-up allowance found in 240.4(B).
by taking the maximum current for any of the
circuits listed in Step 1 and increasing that Example 5. A circuit has a maximum current of
maximum current by 125% (or by multiplying 70 amps. After conditions of use (4 conductors in
it by 1.25). Nonstandard values should be the conduit, 48°C) are applied, a 3 AWG, 90°C
rounded up in most cases [690.9B)]. conductor is selected to meet all ampacity and
In a few rare cases, an overcurrent device 75°C terminal-temperature requirements.
installed in an enclosure or in an assembly may The ampacity after conditions of use have been
be tested, certified, and listed as an assembly applied is 115 x 0.8 x 0.82 = 75.4 A.
for operation at 100% of rating. The con- The required minimum overcurrent device
ductor ampacity, before conditions of use are for this level of maximum current is 70 x 1.25 =
considered, can be 100% of the maximum 87.5 A.
current. The author knows of no dc overcur- A 90-amp overcurrent device would typically
rent devices installed in an enclosure for PV be used. It’s been suggested to use an 80-amp
systems that have such a rating. overcurrent device, but that would result in
running it at more than 80% of rating and, in dc
B. Operating temperature affects overcurrent PV circuits, could result in nuisance trips during
device rating short periods of cloud-enhanced irradiance.
Standard overcurrent devices are listed for a However, the largest overcurrent device that
maximum operating temperature of 40°C (104°F), could be used to protect the 3 AWG conductor
Photo 2.3 • Raceways spaced at varying distances from the roof.
with an ampacity of 75.4 amps is an 80-amp Table 310.15(B)(3)(c) has been removed from
overcurrent device. A 90-amp overcurrent device the 2017 NEC.
is the smallest allowed in this circuit. Photovoltaic professionals may elect to contin-
The conductor size would have to be increased ue to use Table 310.15(B)(3)(c) from the 2014
to 2 AWG for full compliance with NEC re- NEC (again, removed in the 2017 NEC) as a
quirements. safety factor to increase the durability of the PV
The ampacity of a 2 AWG, 90°C conductor source and output circuits over the 50-plus-year
under the conditions of use is 130 x 0.8 x 0.82 life of the PV system. This is based on the fact
= 85.28 A. With the allowed roundup, the next that conduits in sunlight that are 12 inches above
standard value of fuse is 90 amps. the roof, for example, have been observed to be
The 2 AWG conductor can be protected by the too hot to handle and are certainly significantly
required 90-amp overcurrent device. hotter than the surrounding ambient air tem-
peratures.
Raceways in Sunlight and Suppose, for example, there is a large PV
Varying Conditions of Use array being installed on a flat-roof commercial
Note that 310.15(B)(3)(c) no longer requires building. Each PV source circuit (aka string
all the temperature adders required by the 2014 of modules) requires two conductors (positive
NEC. Only conductors and raceways with and negative) plus an equipment grounding
bottoms less than 23 mm (7/8 in) from the roof conductor that will originate at the end of each
are subject to a temperature adder of 33°C (60°F). module string and be routed to a dc combiner
only a 90°C rating when it can be

exposed to higher temperatures
without damage.
Looking at the temperature adders
in Table 310.15(B)(3)(c) (2014
NEC) and the conduit fill factors in
Table 310.15(B)(3)(a) (2017 NEC)
will allow us to make the calcula-
tions.
In examining these two Tables, it
will be noted that the larger EMT
Table 2.1 • Table 310.15(B)(3)(a). raceway with an adjustment factor
of 0.45 for the 24 circuit conductors
inside the building (out of the high-tempera- in the shade is not as significantly affected as
ture, sunlit rooftop location). Each of these the smaller EMT conduit with no conduit fill
three-conductor circuits is installed in an EMT adjustment, where the smaller conduit is only 1/4
raceway and 12 such circuits are routed from inch from the roof and the adder is 33°C (2104
the 12 module locations to a centrally located NEC and 2017 NEC).
junction/pull box where all 12 circuits are routed The PV source circuit conductors are typically
into a single large EMT raceway with one 10 AWG USE-2 or PV wire. At an ambient
equipment-grounding conductor (photo 2.2). temperature of 51°C (estimated roof ambient
That large raceway, now with 24 circuit conduc- temperature), the temperature correction factor
tors and one equipment-grounding conductor, is 0.76 from Table 310.15(B)(2)(a). Adding
is routed across the roof and then down into the 33°C increases the temperature to 84°C and
building where it terminates in the dc combiner. yields a correction factor of 0.29. This reduction
As the smaller and larger EMT raceways run yields a corrected ampacity for 0.29 x 40 = 11.6
across the roof, they are all subjected to a number amps. This reduction factor, due to the conduit
of different environmental conditions. In some being close to the roof, is more significant than
locations, they are in the shade and are subject to the ampacity reduction due to conduit fill alone
the ambient temperature. At other locations, they in the larger EMT, which yields .45 x 40 = 18
are in the sun and may be 1/4 inch from the roof, amps.
or 5 inches from the roof on standoffs, or even 24 However, the large EMT may be in the sun
inches above the roof as they cross over HVAC and would be subject to both conduit fill and
ductwork (photo 2.3). Each of these environ- solar heating adjustments.
mental conditions must be addressed and will Looking at the larger EMT in the sun and
be coupled with the conduit fill factors that vary 5 inches from the roof gives us a conduit fill
with each of these circuits in their run from the correction of 0.45. And with a temperature
modules on the roof to the dc combiner. increase from 51 + 17 = 68°C, yields an additional
It is noted that XHHW-2 conductors are correction factor of 0.58. Multiplying the two
exempt from any temperature adders; this is a factors together gives us 0.45 x 0.58 = 0.261; this
superior type of conductor for PV installations is a more restrictive adjustment factor than the
[310.15(B)(3)(c) Exception]. However, engineers ½-in. conduit mounted against the roof and the
may wonder why the XHHW-2 conductor has ampacity is reduced (40 x .261 = 10.4 amps).
It would appear that while all different condi- the 2014 NEC Handbook uses a single hot section
tions of use should be examined, the physics of and then uses the entire circuit length to make
the situation indicates that attention should be the determination. In the PV environment, it
focused on the larger raceways in the sun con- appears reasonable to use just the length of the
taining more conductors. Plan reviewers could adjacent circuit with a higher ampacity, as this
easily do these calculations, but, unfortunately, section will be the heat sink for the shorter
this level of detail on conduit mounting is not conductor.
frequently found in plans submitted with permit If the most restrictive circuit section can be
applications. eliminated as the determining ampacity for the
Plan reviewers, inspectors, and installers take whole circuit, then the next most restrictive
note: In PV circuits, that electrician’s old rule of section should be examined in the same manner.
thumb of using a 30-amp fuse or breaker on a Eventually, considering all conditions of use, an
10 AWG conductor is frequently not valid. Also, ampacity of the circuit will be determined. At
the fusing in dc combiners should be carefully that point, other factors, such as terminal tem-
calculated along with the size of these source cir- perature limits, should be considered.
cuit conductors, as the PV module short-circuit
current continues to increase as larger and larger Sources of Fault Current
PV modules are being built. In most ac power circuits, the utility source of
energy becomes the source of the overload or
The 10-Percent, 10-Foot Rule fault currents, and the current in a circuit usually
Section 310.15(A)(2) may give some relief for the flows from the utility source to the load. The
high temperature conditions on the roof. requirements in Section 240.21 can be followed
where circuits are protected from overcurrents
“Selection of Ampacity. Where more than where they receive their supply: the utility end of
one ampacity applies for a given circuit length, the circuit.
the lowest value shall be used. However, in dc PV circuits, the PV modules
have limited current even when short-circuited
Exception: Where different ampacities apply under fault conditions. This limited current also
to portions of a circuit, the higher ampacity applies to the strings of series-connected mod-
shall be permitted to be used if the total por- ules and, to some extent, the subarray PV output
tion(s) of the circuit with the lower ampacity circuits where several strings of modules are con-
does not exceed the lesser 3.0 m (10 ft) or 10 nected in parallel. As we move from the modules,
percent of the total circuit.” to the strings and to the output of parallel con-
nected strings forming subarrays to the output of
On a roof with PV circuits, there are going to parallel connected subarrays, into the output of
be several sections of the circuit that will have the entire array, the available currents that could
different calculated ampacities. If the worst-case contribute to faults and overloads increase. As
(lowest-ampacity) section is less than 3.0 m (10 these currents increase, the circuit conductors are
feet) in length, or less than 10 percent of the increased in size appropriately and the ratings of
length of the adjacent section with a higher am- overcurrent devices are also increased.
pacity (whichever is less), then the higher ampac- The source of potentially high overload currents
ity may be used for both sections. The Code uses and fault currents is not the PV module or the
the term circuit portion and the example in string of PV modules. It is the combined output
of those strings in parallel and the combined inverter output circuits) and also connected
output of subarrays connected in parallel that to sources having higher current availability
pose the source of overload and fault currents. (e.g., parallel strings of modules, utility power)
Section 690.9(A) addresses this peculiarity of shall be protected at the higher current source
current-limited sources. connection.”
“Circuits and Equipment. PV system dc With current-limited sources, such as PV

circuit and inverter output conductors and modules and the ac output of utility interactive
equipment shall be protected against overcur- inverters, the location of the overcurrent device
rent. Overcurrent protective devices shall not for the circuits is going to be located at the source
be required for circuits with sufficient ampac- that has the highest available fault current: the
ity for the highest available current. Circuits parallel-connected strings of modules in the dc
connected to current limited supplies (e.g., circuits or the utility for the ac circuits.
PV modules, dc-to-dc converters, interactive
Chapter 3 PV Modules — Installation Considerations 65
66 Chapter 3 PV Modules — Installation Considerations
03
PV Modules — Installation
Considerations
Photovoltaic modules must be installed and con- distance (from zero to six inches or more), the
nected in a manner that meets Code requirements roof may be subjected to both uplift and down-
and that will ensure a safe, durable system with force wind loadings—again concentrated through
maximum output for possibly 50 years or more. the mounting feet of the rack. If the roof has
several layers of old shingles under the array, the
The PV Array — Mechanical structural limit of the roof may be approached.
Considerations Leaving up to two layers of old shingles in place
The PV array consists of individual PV modules is a common practice during re-roofing, so we
attached to a metal rack. That rack is usually can assume that the basic roofing structure has a
attached to the structural members of the roof in a safety factor allowing the extra load of old shin-
typical rooftop-mounted residential utility-inter- gles or the PV array, but possibly not fpr several
active PV system. Although not an electrical-code
issue, some attention must be given to the attach-
ment of the PV array to the building structure.
Various national, state, and local building codes
address the attachment of structures to build-
ings—especially in earthquake and high-wind
areas.
Most roofs in recent years have been built
using span tables in the building codes or using
trusses designed by professional engineers. Pho-
tovoltaic arrays may add up to 4 to 5 pounds per
square foot of dead weight to the roof structural
members, and that weight will be concentrated
through the rack mounting feet. Also, because Photo 3.1 • Array rack attachment point—used in dry cli-
mates. Without flashing, the roof may leak during heavy
the PV arrays are mounted above the roof some rain.
Photo 3.3 • Protective barrier “guarding” PV conductors.
safe” connectors (photo 3.2). The conductors

are typically USE-2 or PV wire as allowed by
NEC 690.31. Either conductor type is suitable
for exposed dc PV module wiring. Photovoltaic
wire is a “super” USE-2 that has a thicker jacket
(the conduit fill tables cannot be used); passes
a 720-hour accelerated UV test (and is marked
Sunlight Resistant); and has the flame and smoke
retardants of RHW-2. It can be used under and
within the PV array for the module intercon-
nections and in raceways in other locations. This
cable is used on all modules because it is required
Photo 3.2 • Module with attached cables and connectors.
on modules that can be installed both in the
layers of shingles and a PV array. United States and in other countries.
Array racks must be attached to the structural Although electrical connectors attached to the
elements of a roof (trusses or rafters). This will ends of module cables are “finger safe” when new,
require penetrating the roofing surface material if they are opened under load, the dc arc may
in a manner that is weatherproof for the life of damage the insulation and the connectors may
the PV array or for the life of the roof—whichev- then pose a shock hazard. Therefore, requirements
er is shorter (photo 3.1). in 690.33 establish the need for locking connec-
Stainless steel hardware is commonly used to tors. A tool will be required to open these locking
connect modules to racks. Galvanized hardware connectors. They will soon be appearing on most
is frequently used to bolt racks to the roofs. In (if not all) PV modules, although they are only
both cases, corrosion resistance is a must in most required when PV array wiring is operating above
climates. 30 volts and is readily accessible (690.33).
Another requirement that applies to readily
The PV Array—Cable Types and accessible PV source and output circuit con-
Installation ductors operating at over 30 volts is found in
Electrically, PV arrays consists of PV modules 690.31(A). These conductors must be installed
connected in series using permanently attached, in type MC cable or guarded raceways. Unfor-
exposed single-conductor cables with “finger tunately, most PV modules do not have junction
residential rooftop PV arrays are not

readily accessible.
The solution, as noted by the word
“guarded,” is to make this wiring not read-
ily accessible by placing a barrier behind
the modules that prevents the wiring
from being touched without removing
the barrier. Fences with locked gates may
not be a solution, because a basic mainte-
nance requirement for a readily accessible
ground-mounted PV array is keeping the
grass mowed—a task usually done by peo-
Photo 3.4 • Modules in landscape orientation.
ple not qualified to be near PV or other
electrical systems. Various barriers are on
the market that either attach to the back
of the PV module/rack assembly or to the
rack to prevent ready access to exposed
conductors (photo 3.3).
Conductor leads attached to modules
are 40 inches long (or longer) to allow
the series connection of modules when
they are mounted in a landscape orienta-
tion (photo 3.4). When the modules are
mounted in portrait orientation (photo
3.5), the excess lengths of conductors must
Photo 3.5 • Modules in portrait orientation.
be securely fastened against the module
racks to resist abrasive damage due to wind, sleet,
and ice. Many installers use plastic cable ties, but
unless they are of very high quality, they may not
last the required 40 years or more when exposed
to the extremes of heat and ultraviolet radiation
from sunlight. Some installers use a stainless-steel
pipe clamp (loop strap) with an EDPM insert.
There are numerous specially designed metal clips
for this requirement (photo 3.6).
Photo 3.6 • Pipe clamps (EDPM and stainless steel) used
to secure bundles of module conductors and special stain-
less steel clips to secure individual module conductors.
Cables and Connectors for PV
Modules
boxes with knockouts that would accept a Exposed cables and connectors used in PV source
raceway or MC cable. They come with perma- circuits are some of the most critical components
nently attached exposed, single-conductor cables of a PV system in terms of maintaining system
and connectors with no provision for attaching durability and safety. These components are
a conduit or other raceway. Fortunately, most exposed to the extremes of weather for the life
Photo 3.7 • PV module connectors (normally not load-break

rated). Photo 3.8 • Microinverter with ac and dc connectors list-
ed with the microinverter as isolating means. Courtesy of
of the system and may be required to be durable Enphase Energy.
and safe longer than the system will produce
power. Photovoltaic modules have a warranted and the 120-V ac receptacle in houses. Photovol-
lifespan for power output of 25 years, but they taic connectors, if they were to be listed, would
may produce dangerous amounts of voltage and have to be able to safely open a circuit at 600 volts
current for as long as 40 or 50 years. In many or even 1000 volts dc at the rated current of the
cases, there will be a functioning (or more likely, module, which could be (in some cases) as high as
nonfunctioning) inverter attached to a PV array 10 amps or 12 amps. Interrupting direct currents
output for that length of time. The quality of the is considerably more difficult than interrupting al-
components used, and the care taken in comply- ternating currents and, at this time, none of these
ing with the Code, determine the durability and PV connectors are rated for load break operation
ultimate safety of these cables and connectors. at full voltage.
When these connectors are installed on a PV
Connectors for PV Modules and module or other PV component, the connec-
Related Equipment tors are evaluated with that component and are
Recognized Components certified as suitable for that application with that
Photovoltaic modules come with connectors module or equipment. This certification still does
attached to the ends of the cables that have been not allow them to be operated under load, and
permanently attached to the PV module (photo nearly all of them will be marked “Not for Load
3.7). There are multiple connector manufacturers Break Operation” or similar wording.
supplying connectors to module manufacturers, The exception to this recognized compo-
and modules with a variety of connectors are nent rating relates to the connectors on some
being used. These connectors, as separate parts, microinverters that have been listed with the
are recognized components because they cannot inverter as a disconnecting means when used at
meet the more stringent requirements of a fully the lower voltages and less strenuous operating
listed connector, which would require that the conditions of these devices. These connectors
connector to be capable of being opened under meet the requirements of 690.15 as isolating
load, or opened at full current and full rated volt- devices (photo 3.8).
age, as well as meeting other requirements. An
example of a listed, load-break rated connector Compatibility Issues
assembly would be the ac plug used on appliances Some connector manufacturers are advertising
that some of their connectors are interchangeable ically strong and not separate under high wind
with connectors from other manufacturers in loadings and ice loadings (and possibly thermal
terms of an electrical and mechanical compati- stress) to which they may be subjected. There can
bility. On casual inspection by the AHJ and PV be no deterioration of the electrical contact due
installer, this mechanical and electrical compat- to a mismatch, however slight, of the metals used
ibility appears to be true. Unfortunately, Under- in each half of the mating connectors. When
writers Laboratories and the members of the viewing the connector as a splice in the cable, the
Standards Technical Panel for UL Standard 6703 Code requires the splice to be at least as robust
for connectors have established that none of these as the unspliced cable before the connector was
mixed pairs of connectors have been evaluated inserted at that point [110.14(B)].
for compatibility by an independent third-party When connectors are mixed and matched, the
Nationally Recognized Testing Laboratory like listing and certification of the PV module as a
UL, CSA, TUV and ETL—the four OSHA-rec- complete assembly—including the connectors—
ognized NRTLs authorized to certify PV equip- becomes invalid. If a mixed, unmatched, pair of
ment. Each connector manufacturer modifies the connectors fails either electrically or mechanically
materials and procedures used to manufacture and causes damage at some time in the future, the
their connectors in a proprietary manner. Even listing of the module, and compliance with NEC
though there may be an electrical and mechanical requirements [including Sections 690.4(D) and
compatibility at one point in time, there is no 110.3(B)] could be a significant issue.
continued evaluation to ensure that changes in At this time, connectors should only be mated
the production process of one manufacturer will in pairs from the same manufacturer and in the
result in their connectors remaining compatible same series. If equipment must be connected that
with connectors from another manufacturer. uses connectors from different manufacturers,
It is easy to understand how a slight modifica- then some sort of cable connector adapter must
tion of the metals used in one connector (while be field manufactured or purchased that will
being compatible with the metals in the mating allow the connectors to be mated in pairs from
connector by the same manufacturer) would not the same manufacturer. It is probably not a good
be compatible over long periods of time with the idea to cut connectors off of a given product
metals used in a connector by another manufac- and change those connectors to a connector set
turer. from a manufacturer that will mate with the
Similarly, a modification of the plastics used in connector set on the other product. Although
one connector (while meeting the requirements engineers at Underwriters Laboratories have said
to perform effectively with the plastics from the this procedure does not violate the listing on the
same manufacturer in the mating connector) may product, several PV module manufacturers have
not be compatible over the long term with the stated that cutting the connectors off their listed
plastics used by another manufacturer. Even min- module or other equipment will invalidate the
ute differences in the thermal expansion rating of warranty on that product. Therefore, it might be
different plastics could pose problems. more appropriate to make a short length of cable
These connector assemblies must remain with different connectors on each end to keep the
electrically and mechanically secure for the very mating pairs from the same manufacturer togeth-
long power production life of PV modules. They er, even though this will double the number of
must be watertight, not allowing water to reach connectors in the circuit and possibly reduce the
the electrical connections. They must be mechan- reliability of the system.
Photo 3.9 • Y cable adapter with MC-4 type connectors. Photo 3.10 • Inline fuse with MC-4 type cable connectors.
All connectors must mate with matching connectors from Connectors must be mated with cable connectors by the
the same manufacturer. Uncontrolled backfeed currents same manufacturer. Temperatures in installations exposed
to sunlight may result in fuses being operated outside
may pose problems.
their listed rating. Courtesy of Leader Group.
A Warning. It is becoming more common In situations where the Y adapter has fuses
to have to strings of modules combined with for both inputs internal to the adapter, or where
a Y-type cable adapter which may or may not inline cable fuses are used on each input, the
have fuses inside the adapter or connected to environment that the fuses operate in must be
the adapter for each input string (photo 3.9). considered (photo 3.10). Fuses listed for PV
Without fuses, the issue is: Are there any po- applications are listed with a maximum operating
tential external currents that can damage the temperature of 50°C. On a typical rooftop ap-
modules or string wiring? External currents plication, it is probable that the adapter and fuses
may originate from additional combinations of will be in an ambient temperature (around the
module strings elsewhere in this circuit or from cable/adapter) environment of 40°C to 50°C and
backfeed currents from the inverter. These are be subject to solar heating from direct exposure
the sources of external overcurrents that must be to sunlight. Although the 2017 NEC has deleted
addressed before this type of unfused paralleling any sunlight adders on temperature for cables
can take place (690.9 Exception). In many cases, that are more than 7/8 inch above the surface
it is difficult to obtain information on whether [310.15(B)(3)(c)], it seems evident that conduc-
the inverter can backfeed currents into faults tors anywhere within that 2014 NEC space of 36
in dc PV array wiring. These adapters are made inches will be subjected to additional solar heat-
by numerous manufacturers and, of course, the ing. In a recent PV installation, my tools on an
manufacturer of the Y adapter must be the same open stand about 12 inches above the roof in the
manufacturer who makes the connectors used on sun became far too hot to handle without gloves
the PV modules and the connector used on the in just a few minutes. It is likely that the fuses in
output cable from the Y adapter to ensure the these Y adapters will frequently be subjected to
listings on all connectors and modules remains more than 50°C, even when they are behind the
intact. All MC-4 type connectors are not equal. PV modules on a rooftop installation. This will
mean the fuses are operating outside of the listed

environment and the listing is invalid, which is a
violation of the Code. I urge caution any time Y
adapters, with or without fuses, are being used in
a PV application.
Factory Training and Factory Tools Required

When it comes to installing connectors on cables
in the field, caution is advised. In general, recog-
nized components may not be installed to other
equipment (or to a system by themselves) in the
field, but are supposed to be installed only in
the factory under factory-controlled conditions.
This is a gray area in the UL Standards and in
the Code. It is acknowledged, but not in writing
anywhere, that a recognized component PV
connector may be installed by a person having
manufacturer’s training on the installation of Photo 3.11 • Connector assembly tools. Courtesy of
the connector and using manufacturer-approved MultiContact/STAUBLI.
tools to assemble the connector to the cable
(photo 3.11). Because there are a significant caution when purchasing bulk spools of PV
number of different cables being used, the com- cable or wire in the United States that may
patibility of a connector with a particular cable have fine stranding. It is difficult to find
must be determined. terminals on equipment, such as dc combin-
ers and inverters, that can accept fine-strand-
Module and Exposed Cables ed cables in the United States. Note the
Construction warnings on this problem in 110.14(A) and
All PV module manufacturers are using “PV 690.31(H).
cable/PV wire” fastened to their modules. (See Although all PV cable or wire is marked
690.31.) Photovoltaic cable or PV wire is that sunlight resistant and has been tested for UV
cable meeting UL Standard 4703 for the use resistance with 720 hours of acceleration UV
on modules and in exposed, field-installed PV testing, some care must be exercised in selecting
source circuits. Of course, the old standby USE- conductors that will withstand decades of use
2 conductors can also be used for exposed, source and exposure in PV environments. It is acknowl-
circuit wiring. edged that conductors with black insulation will
UL Standard 4703 allows a wide variety of mate- last significantly longer than cables with colored
rials and construction configurations to be used for insulation, and this applies to PV cable and wire
the manufacture of PV cables so that they may be and to USE-2. Also, conductors with synthetic
used both in the United States and in Europe. In rubber insulation (thermoset), such as XHHW-
Europe, the requirement for PV cables means that 2, may last longer than conductors with PVC
the cables must be fine-stranded, flexible cables insulation (thermoplastic) such as THWN-2—
and have tinned conductors. Photovoltaic system although both can be used in the construction of
integrators and PV installers should exercise PV cable/wire.
Photo 3.12 • Plastic wire ties, especially white ones (not Photo 3.14 • Many potential issues: EMT clamp and tape
UV rated), are likely to fail in a few short years. used to secure bundle of cables, plastic wire ties, foam
sealant, and colored (non-black) PV wire insulation.
Photo 3.13 • Open conduit entry to dc combiner should Photo 3.15 • Proper termination fitting for exposed con-
not be used. ductors entering a dc combiner.
Mechanical Attachment insulated stainless-steel clamp shown in photo 3.6,

Even the highest quality cables and properly should be used.
mated connectors will not survive decades of In a related area, the entry of these exposed
outdoor abuse, unless they are properly secured to conductors into conduits or dc combiners should
the module and racking structure. Unfortunately, be made with an appropriate conduit or opening
nylon cable ties—even the black ones marked UV termination fitting. Open conduits or conduits
resistant—are not withstanding the test of time closed with nonelectrical foam products should
in the PV environment. Again, these cable ties not be used. (See photos 3.13, 3.14, and 3.15.)
have not been tested with more than 720 hours
of accelerated UV tests, which is equivalent to The Conductors
about two and a half years of outdoor exposure. As noted previously, single-conductor, exposed
In PV applications, particularly in the hotter cables (type USE-2 or PV cable/PV wire) will be
Southwest regions of the United States, some of used for the module interconnecting cables. Both
these are failing in just a few years (photo 3.12). cable types are available in black (the most durable)
Several manufacturers have brought to the market and in multiple colors, including white and green
stainless-steel cable clips that are suitable for this (for exposed grounding and equipment-grounding
application and will withstand the test of time. conductors). As noted in 200.6(A)(6), black cable,
These clips or something similar, like the rubber even when smaller than 4 AWG, may be marked
white as a grounded conductor, if necessary, at to require a “red is positive and black is negative”
the time of installation where used as an exposed color coding (at least for conductors in conduit)
outdoor conductor at the array. because there will be no grounded conductor.
Normally, the exposed single-conductor cables However, both ungrounded conductors can still
are transitioned to a conduit wiring method when meet Code if they are black. As mentioned, basic
the circuits leave the PV array. Conductors in black is preferred for all exposed outdoor conduc-
conduit, while they could be USE-2/RHW-2 tors due to an increased UV resistance.
(for flame and smoke retardant) or PV wire, are The newest bipolar PV arrays and bipolar
typically THHN/THWN-2 because they are less inverters should not be ignored. In these sys-
costly and the -2 rating is needed for the outdoor, tems, there will be red positive conductors,
wet environment and the high temperatures of black negative conductors, and possibly white
conduit in sunlight [310.15(B)(3)]. Unfortunately, grounded conductors (where these conductors
14-10 AWG conductors with THHN/THWN- are solidly grounded). Of course, some installers
2 insulation are not widely available due to low will use black for both ungrounded conductors
demand. Of course, THHN/THWN is available, (acceptable under the NEC), and this will pose
but it does not have a wet, 90°C rating. problems for the inspector and the troubleshooter
Demand will increase for the small-conduc- when the cables are misconnected. If the “center
tor THHN/THWN-2 conductors as inspec- tap” dc conductor is functionally grounded, then
tors start applying 310.15(B)(3) to rooftop it, too, may be black. There are numerous circuit
HVAC installations. Due to the limited marking, routing, and grouping requirements in
availability of 14-10 AWG THHN/THWN-2, 690.7(C), 690.31, and 690.41.
XHHW-2 would be a suitable, superior alter- As before, grounded conductors in PV dc
native and may not be subject to additional disconnects should not be switched. Bolted, iso-
deratings due to sunlight exposure or raceways lated, terminal-block connections are acceptable,
more than 7/8 inch above the roof. but not required. Section 690.31(F)(1) will be
Although most PV arrays installed during the clarified in the 2020 NEC and possibly by a TIA
recent past have had the dc-negative conductor for the 2017 NEC.
grounded (and colored white), newer arrays will
be installed to meet the functional grounding Wiring Methods—Continued
requirements of the 2017 NEC. Of course, there All circuits in a PV system, as in other electrical
are no designated color codes for ungrounded systems, must be wired using a Chapter 3 or a
conductors, but common sense would indicate 690.31 method that is suitable for the application
that ungrounded conductors would be clearly and the environment. The circuits between dc PV
marked. It would be understandable if they were disconnects and inverters are considered to be PV
colored red for positive conductors and black for output circuits (and also inverter input circuits).
negative conductors installed in conduit. They should be in a metal raceway or metallic cable
However, many installations use black conduc- assembly. Of course, local codes may dictate other
tors for both and that is still acceptable under the requirements, such as the need to use raceways
Code. In positively grounded systems where the inside commercial structures for all electrical wiring.
positive grounded conductor is colored white, the
ungrounded negative conductor would be most PV Module Grounding
clearly understood if it were black. Grounding PV modules to reduce or eliminate
Functionally grounded PV systems would seem shock and fire hazards is necessary but somewhat
difficult. Copper conductors are typically used must be used because this is the only point tested
for electrical connections, and the module frames and evaluated by the certification or listing agen-
are often aluminum. It is well known that copper cy for use as a long-term grounding point. UL
and aluminum do not mix (as was discovered from has established that using other points (such as
numerous fires in houses wired in the 1970s with the module structural mounting holes), coupled
aluminum wiring)—dissimilar metals are not to with typical field installation “techniques,” may
come into contact with each other (110.14). not result in low-resistance, durable connections
Photovoltaic modules have aluminum frames. to aluminum module frames. If each and every
Many are anodized for color or have mill finishes possible combination of nut, bolt, lock washer,
and many are clear-coated. The mill finish alumi- and star washer could be evaluated for electrical
num (and any aluminum surface that is scratched) properties and installation torque requirements,
quickly oxidizes. This oxidation, and any clear- and if the installers would all use these compo-
coat or anodizing, form an insulating surface that nents and install them according to the torque
makes for difficult long-lasting, low-resistance requirements, it might be possible to use the
electrical connections (in, for example, frame structural mounting holes for grounding.
grounding). The oxidation and anodizing is not a New grounding devices are coming to market
good enough insulator to prevent electrical shocks, that will eventually ease the problem of mod-
but it is enough to make good electrical connec- ule grounding. However, until they have been
tions difficult. evaluated with specific modules and the module
UL Standard 1703 (published by UL and instructions address these devices, they do not
developed and maintained by the Standards meet the requirements of UL Standard 1703 or
Technical Panel 1703) is used to certify and NEC Section 110.3(B).
list all PV modules sold in the United States. The latest PV module grounding methods use
It requires stringent mechanical and electrical engineered, manufactured mounting systems
connections between the various pieces of the or racking systems that have been listed to UL
module frame to ensure that these frame pieces Standard 2703 for PV module racks. This standard
remain mechanically and electrically connected evaluates and certifies racking systems for mechan-
over the life of the module. These low-resistance ical strength, for electrical grounding continuity, or
connections are required because a failure of for both. A module is typically mounted with clips
the module insulating materials could allow the to the rack and, in many cases, the clips mount and
frame to become energized at up to 600 volts to ground the module to the rack so that only a single
1500 volts (depending on the system design). copper equipment-grounding conductor attached
The NEC requires that any exposed metal surface to the rack is needed to ground all the modules
be grounded if it could be energized (250.4 and on that rack. The module instruction manual must
250.110). Installers of PV systems are required show where and when top clips for mounting or
to ground each module frame (690.43). NEC grounding can be used. The listed rack instructions
Section 110.3(B) and UL Standard 1703 require should indicate that the rack has been evaluated
that the module frame be grounded at the point for mounting and grounding with a particular
where a designated grounding provision has been PV module. Efforts are underway to categorize
made. The connection must be made with the PV module frames into categories related to size,
hardware provided (if any) using the instructions material, strength, and coating. Then, rack man-
supplied by the module manufacturer. ufacturers will have to test only a few categories
The designated point marked on the module to establish that racks will be able to mount and
Photo 3.16 • Tin-plated copper direct burial lay-in lug.
ground numerous PV modules that are in those specify the necessary hardware. These methods,
categories rather than testing each module. and the hardware, will be evaluated during the
Some PV module manufacturers are providing listing of the module. It is likely that thread-cut-
acceptable grounding hardware and instructions. ting or thread-forming screws of the past will no
Other manufacturers provide less-than-adequate longer be used.
(or no) hardware and unclear instructions. Revi- For modules that have been supplied with
sions of UL 1703 are addressing these issues. In inadequate or unusable hardware (or no hardware
every case, the module manufacturer’s hardware at all), there is one way to meet the intent of the
and instructions should be used (where possible) Code and UL Standard 1703.
to ground the module at the points marked on For situations requiring an equipment-
the frame. UL Standard 1703 requires that the grounding conductor larger than 10 AWG, a
module manufacturer outline the specific ground- lay-in, tin-plated copper, direct-burial lug with
ing methods to be used, and either provide or a stainless-steel #10 screw, nut, flat washers,
Photo 3.17 • Connecting tin-plated copper lay-in lug to Photo 3.18 • Grounding a metal roof—oops, outdoor rat-
aluminum. ed lug and wire needed.
Belleville spring, and lock washers can be used corrode. The same can be said for other screws,
to attach a direct burial lay-in lug to the module lugs, and terminals that are not suitable for
frame at the point marked for grounding (photo outdoor applications.
3.16). Before attaching the lug to the module, a The direct-burial lay-lugs are tin-plated lugs,
stainless-steel brush should be used to remove made of solid copper, with a stainless-steel screw.
any anodizing, oxidation, or clear-coating from They accept a 4 AWG to 14 AWG copper con-
the aluminum module frame. A thin coat of ductor. They are listed for direct-burial use and
antioxidant film should be placed on the clean outdoor use and can be attached to aluminum
aluminum surface. Flat washers are required to structures (the tin plate allows this). The much
prevent lock washers from digging into the soft cheaper tin-plated aluminum lay-in lugs look
copper of the tin-plated lug or the aluminum of identical, but have a plated screw. They are not
the module frame. The Belleville washer pro- listed for outdoor use. If module grounding is to
vides uniform tension, and a torque screwdriver be done with a 14 AWG to 10 AWG conductor,
should be used for all electrical connections then the lay-in lug may not be needed.
(photo 3.17). Some new grounding lugs have Other application-specific grounding lugs are
been listed for use without the antioxidant com- available that will penetrate the module-protect-
pound because the design of the lug penetrates ing and -insulating coatings without the need for
the oxidation. However, these should be evaluat- surface preparation of the module.
ed with a specific module because of the varying What size conductor should be used? The
thickness of the anodizing and clear-coat on minimum Code requirement is for the equipment
the modules. It is not acceptable to use the grounding conductor for PV source and output
hex-head, green grounding screws (even when circuits to be sized to carry 1.25 times the short-cir-
they a have 10-32 threads) because they are not cuit currents at that point, using the requirements
suitable for outdoor exposure and will eventually of 250.122 and based on the overcurrent device
Photo 3.20 • Improper conduit terminations will allow

Photo 3.19 • Rodent-damaged conductors. moisture and rodents to enter the enclosure.
protecting these circuits. For PV source circuits equipment-grounding conductors where there is
where there are no overcurrent devices, an assumed any possibility that these metal surfaces may be
overcurrent device per 690.9(B) shall be used per energized by conductor insulation failures (photo
250.122. While this may allow a 14 AWG conduc- 3.18). Rodent damage and abrasion of conductor
tor between modules, a conductor this small would insulation could cause module frames, racks, or
require physical protection between grounding metal roofs to be energized (photo 3.19).
points. An equipment-grounding conductor smaller Single-conductor exposed wiring (USE-2 or PV
than 6 AWG can be routed behind the modules, wire) is allowed only in the near vicinity of the PV
from grounding point to grounding point, if the array to interconnect the modules and to return
conductors are well protected from damage, as they the end of the string conductor to the origination
would be when run in the module or rack frames in point of the string wiring. At this point, exposed
a roof-mounted array [250.120(C)]. If needed, an wiring must transition to one of the more com-
8 AWG or 6 AWG conductor may be required (to mon wiring methods found in Chapter 3 of the
meet the Code or to satisfy an inspector) and then Code, unless the conductors have a dual marking,
lay-in lugs or a listed grounding device should be such as RHW-2 or XHHW-2. Typically, this will
used (690.45 and 690.46). be some form of raceway, such as electrical metal
The Code allows metal structures to be used tubing. If the array output conductors penetrate
for grounding and even allows the paint or other the surface of the structure before reaching the first
covering to be scraped away to ensure a good readily accessible dc PV disconnecting means, then
electrical contact. Numerous types of electrical they must be in a metal raceway where routed inside
equipment have parts that are grounded with the structure. Metal raceways include rigid metal
sheet metal screws and star washers. This works conduits, electrical metal tubing, intermediate metal
(particularly in the factory environment) on conduit, metal wireways, and flexible metal conduit,
common metals like steel, but not on aluminum and include Type MC metallic cable assemblies.
due to rapid oxidation. The transition fitting keeps water, dirt, rodents, and
The racks and any metal roofing panels under other material out of the conduit. A rain head or a
the PV array should be connected to earth with cord grip might be used (photo 3.20).
Photo 3.21 • Equipment-grounding conductor improperly connected to the module. (Incorrect hardware and incorrect
installation procedure.)
the frame pieces (so that any failure in module
UL Standard 1703—Grounding insulation or external conductor insulation will
versus Bonding result in all pieces of the frame receiving equal
The current edition (2012) of UL 1703, Standard voltage). Factory bonding also ensures that when
for Flat-Plate Photovoltaic Modules and Panels, the module frame is properly field-grounded at
delineates the differences between grounding one of the marked and tested points, the entire
and bonding. Bonding refers to the factory-made module frame is maintained at the ground (earth)
electrical connections between the four or potential under fault conditions.
more aluminum sections of the module frame. During the bonding process, all screw fasteners
Grounding refers to the field-installed electrical (when used) are precisely torqued to the specified
connection between the aluminum module frame value by automated equipment or by trained
and the equipment-grounding system (usually technicians using torque screwdrivers. Factory
consisting of copper conductors). bonding materials and methods are evaluated for
Bonding the frame pieces together at the low resistance and durability during the certifica-
factory, using very specific materials and methods, tion and listing processes. After the listing, if the
results in a durable electrical connection between manufacturer changes any of the bonding mate-
rials or methods, the changes must be reevaluated ing connections made by installers in the field.
by the listing agency. The materials (including any Instruction manuals and hardware (sometimes
screws or washers) are not specified generical- supplied) show techniques that are not always
ly—they are specified to the original equipment consistent with good electrical connections (pho-
manufacturers and must always be obtained and to 3.21). Field-made connections using threaded
used from those sources unless any change is fasteners are rarely torqued to the specified value,
reevaluated by the certification or listing agency. even when that value is given in a module’s
Compare this precisely controlled and evalu- instruction manual (because not all PV installers
ated factory bonding system with field-installed have or carry torque screwdrivers or wrenches).
grounding techniques used to connect copper The NEC requires that where threaded fittings
equipment-grounding conductors to aluminum or connections are provided with a torque value,
module frames. Grounding a PV module is a calibrated torque device shall be used to tight-
difficult at best, for a number of reasons. The first en that connection or terminal [110.14(D)].
is that module manufacturers may not realize Field-grounding connections may or may not be
the importance of this connection to the over- inspected by AHJs, and they are never tested for
all safety of the system. The revised UL 1703 overall continuity. Also, because PV systems can
distinguishes the differences between bonding operate without trouble for many years, there may
and grounding. Manufacturers may have the be little motivation to inspect these connections
impression that bonding techniques and materi- after the original installation.
als used in the factory may be applied to ground-
Chapter 4 The Inverter — Operation and Connections 81
82 Chapter 4 The Inverter — Operation and Connections
04
The Inverter—Operation
and Connections
Utility-interactive inverters range in size from connects. With these designs, an external dc
175 watts (photo 4.1) to 2.5 megawatts and come PV disconnect must be installed (690.15). Even
in all shapes, sizes, and colors (photos 4.2 and if the inverter has more than one set of input
4.3). New models are being introduced monthly. terminals for paralleling separate strings (source
These inverters will be listed by UL, CSA, ETL, circuits) of modules, external dc PV disconnects
and TUV Rheinland. All of these organizations must be used on each input (photo 4.4).
are designated as nationally recognized testing Other inverters have internal dc disconnects
laboratories by OSHA for testing and listing PV or disconnect housings that attach to the main
modules, inverters, combiners, and charge con- inverter section containing the electronics
trollers using standards published by UL. package. The method used to mount the internal
Some inverters have only a single set of dc disconnects, the ease and accessibility of the dis-
input terminals and no internal dc or ac disconnects, and the way they are separated from the
inverter proper vary from brand to brand
and from product to product. Installers
and AHJs must reach a mutual conclusion
on the suitability of these disconnects for
meeting the various isolation and discon-
nect requirements in the Code.
Because the inverters are listed with the
disconnects, it can be presumed that the
disconnects are properly rated for the dc
load break operation and that they would
meet the requirement for the isolator and
disconnect requirements of 690.15.
Meeting the requirement for equipment
Photo 4.1 • Westinghouse/Andalay AC PV modules using
Enphase microinverters. isolators and disconnects (690.15) will
Photo 4.2 • KACO utility-scale inverters.
Photo 4.3 • SMA Sunny Central 2.5 MW utility-scale inverter. Photo 4.4 • Inverter with internal dc disconnects
Courtesy of SMA-AG. plus additional external dc and ac disconnects.
require additional considerations. If an inverter to be removed to service the inverter, then some
required factory service, could the energized PV degree of safety is ensured. However, if ener-
source or output circuits be disconnected from gized conductors must be disconnected from
the inverter safely when there is no external internal switches and pulled through small con-
disconnect? If a disconnect housing is attached duit knockouts, the situation must be examined
to the inverter and that housing does not have carefully. Will qualified people, who know how
Peak Power Tracking

A PV array is a current source of energy and the
output power depends on the load that the invert-
er places on the array. No loading (zero current)
would operate the array at the open-circuit voltage
point (VOC). The heaviest loading (a short-cir-
cuit—hard to achieve) would operate the array at
the short-circuit current point (ISC). Neither of
these operating points would produce any power
output from the array, power being the product
of voltage and current and (at these points) one
of them is zero. However, for every condition of
sunlight intensity (irradiance) and array tempera-
ture, there is a load that will extract the maximum
power from the array that the array can produce
under those conditions. The utility-interactive
inverter will find that maximum or peak power
point—maximum power voltage (Vmpp) and maxi-
mum power current (Impp)—and track that point as
the sunlight and temperature vary throughout the
day. See chapter 2 for more information.
Automatic Operation
Photo 4.5 • Transformerless inverter with internal dc dis-
connect.
Today’s utility-interactive inverter is designed,
manufactured, tested, and certified or listed to
operate automatically in the PV system. There are
no transfer switches. The inverter seamlessly con-
verts dc power from the PV array into ac power
that is fed to the utility-supplied premises wiring
system. The output of the inverter is functionally
connected in parallel with the premises’wiring
(and loads) and the utility service.
One of the most important aspects of the in-
verter is the anti-islanding circuit. An anti-island-
ing circuit is designed to keep the utility electrical
system (both premises wiring and utility feeder)
safe in the event that the utility is being serviced
Photo 4.6 • Inverters with attached dc disconnects that or is disconnected at some point in the trans-
can be separated from inverters.
mission system, distribution system, or premises
to disable the array, be doing the removal? Or wiring system.
will an unqualified person try to pull energized Unlike an engine-driven generator, which can
conductors through the knockouts? (See photos feed power into a blacked out or disconnected
4.5 and 4.6.) local utility feeder system, an inverter anti-island-
ing system prevents an inverter from energizing Code requirement, between the array-rated output
the “dead” power system. current and the manufacturer’s specified maximum
This circuit prevents the inverter from deliver- dc inverter input current. Normally, a PV array is
ing ac power if the utility voltage and frequency rated in watts at standard test conditions of 1000
are not present, or if they are not within narrowly watts per square meter (W/m2) of irradiance and
defined limits. This circuit monitors the voltage a cell temperature of 25 °C. In most cases, an array
and frequency at the output terminals of the will operate, on average, at a lower power output
inverter. If the voltage varies more than plus because of normal and expected power lost due to
ten percent or minus twelve percent from the module heating. For this reason, inverter manufac-
nominal output voltage the inverter is designed turers typically suggest sizing PV arrays (standard
for (120, 240, 208, 277, or 480 volts), the inverter test condition dc watts) at ten to twenty percent
cannot send power to the output terminals. Nor greater than an inverter’s ac power output rating. It
is there any voltage on these terminals (from the does no short-term harm to connect an even larger
inverter) when the inverter shuts down if the PV array to an inverter because the inverter must
circuit from the utility is at zero voltage. In a limit its output to the rated value no matter how
similar manner, if the frequency varies from 60 much array power is applied. If this oversized array
Hz to more than 60.5 Hz, or less than 59.3 Hz, is used, the inverter will spend more operating
the inverter also cannot send power to the ac out- time each day at rated power output than it would
put. If the utility power is suddenly not present with a smaller array. The penalties for designing
at the output terminals for any reason (inverter a system in this manner will be increased mod-
ac output disconnect opened, service disconnect ule cost for the larger array; some lost power on
opened, meter removed from the socket, utility sunny, cool days; and possibly a slight reduction in
maintenance, or utility blackout), the inverter the inverter’s lifespan due to longer operation at
senses this and immediately ceases to send power full-power higher internal temperatures. However,
to the output terminals. systems installed where microclimates reduce
The anti-islanding circuit in the inverter irradiance during portions of the day, or where
continues to monitor the ac output terminals the local utility limits the inverter-rated power
and when the voltage and frequency from the output, this type of operation may result in more
utility return to specifications for a period of five energy being delivered from the output than
minutes, the inverter is again able to send PV would be possible with a conventionally sized
power to the ac output. When the inverter is not system.
processing dc PV power into ac output power,
it essentially stops taking power from the PV DC Input Fusing
array by moving the load on the PV array to a Some models of small (<10 kW) and large (>100
point where there is no power. Usually this is the kW) inverters have dc input fuses mounted
open-circuit voltage point for the PV array. inside the inverter or inside a combiner or
disconnect device attached to the inverter. The
Circuit Sizing smaller fuses (30 amps or less) are usually mount-
Direct Current ed in “finger-safe” fuse holders that allow the fuse
The dc input circuit and conductors to the invert- to be safely replaced in an unenergized (no load
er are sized based on the dc short-circuit current current) state.
in those conductors (that sizing is covered in pre- However, when fuse ratings go over 30 amps,
vious chapters). There is no direct relationship, or with values as high as 400 amps or more, these
source circuit fuse (those less than 30 amp “finger

safe” fuse holders). An optional disconnect at the
output of every combiner speeds this process and
makes servicing the combiner fuses safer, but all
disconnects must be located and opened (pho-
to 4.7). Combiners are available with internal
load-break-rated output disconnects that, if used
with a remote control, could speed the shutdown
and possibly meet some of the rapid shutdown
requirements (690.12).
When these fuses are present in the input of
larger inverters, the safest way to provide service
Photo 4.7 • PV combiners with internal disconnects on is to install a dc disconnect near the inverter on
the outputs of both polarity conductors.
each dc input to a fuse (photo 4.8). These grouped
disconnects can be opened easily, and with the
inverter turned off, the fuses can be safely removed
in a de-energized state. Section 690.15(C) and
240.40 require dc disconnects near these fuses
so that they can be safely de-energized from all
sources when being serviced.
Alternating Current (ac)

The ac output circuit of an inverter must be sized
at 125 percent of the rated output current of the
inverter (690.8). Some inverter manufacturers
specify the rated current or a range of values (due
Photo 4.8 • Disconnects, with fuses, for each input in- to varying line voltages from nominal). If this
stalled near the inverter
specification is not given, then the rated power
fuses are mounted in exposed fuse holders or may be divided by the nominal line voltage to
bolted directly to a dc busbar. One side of each determine a rated current. For example, a 3300-
fuse is tied together with the dc input of the watt inverter operating at a nominal voltage of
inverter. The other side of each fuse is hardwired 240 volts would have a rated current of:
to the output of a PV dc combiner and these
combiners will be scattered throughout the PV 3300 watts / 240 volts = 17.2 amps
array—sometimes over thousands of acres of real 1.25 x 17.2 = 21.5 amps
estate. Although the inverter can be turned off
and the dc input capacitors allowed to discharge This would be the circuit ampacity requirement
(up to five minutes), each fuse is still energized and it would usually be protected by a 25-amp
from its own input and the combined inputs of overcurrent device at the utility end of the circuit.
all the other fuses through the common busbar. These inverters are not capable of providing any
The only way to safely service these fuses is to sustained (more than a second) surge currents, so
go through the entire PV array, find all of the the rated output current is all that can be deliv-
combiners, and open or pull each and every ered. When faced with a short-circuit, the rated
output current is all that can be delivered, but— the inverter’s instruction manual and should not
more than likely—the reduced line voltage due be exceeded [110.3(B)].
to the fault will cause the inverter to shut down. Where fused or unfused disconnects are used for
With the available maximum current limited to the ac inverter output disconnect (or the required
the rated inverter output current and the overcur- utility disconnect), the circuit connected to the
rent device rated at 125 percent of that current, utility source should be connected to the supply
it is doubtful that the overcurrent device would side (top) of the disconnect with the inverter ac
trip or blow from currents from the inverter. output connected to the load side (bottom). A
That device would, of course, activate on the high warning label is required, because even when the
available fault currents from the utility. disconnect is opened, the inverter ceases to pro-
duce power within a fraction of a second and the
Connections exposed load-side terminals pose no shock hazard.
Dedicated Circuit However, the connected utility side terminals,
NEC 705.12(B)(1) requires that the inverter while protected, are still energized.
output be connected to the utility power source
at a dedicated disconnect and overcurrent pro- GFCIs and AFCIs
tective device (OCPD). In most systems, this is The ac output of a utility-interactive inverter
a backfed breaker in a load center/panel board should not be connected to a ground-fault circuit
[705.12(B)]. Inverters may not have their outputs interrupter (GFCI) or to an arc-fault circuit in-
connected directly to another inverter or directly terrupter (AFCI) circuit breaker, as these devices
to an ac utility-supplied circuit without first are not currently designed to be backfed and will
being connected to the dedicated disconnect or be damaged if backfed. These devices have termi-
OCPD. Utility-interactive microinverters and nals marked “line” and “load” and have not been
ac PV modules are exceptions to this rule be- identified, tested, or listed for backfeeding.
cause they are tested and listed to have multiple
inverters connected in parallel on a single circuit Ground-Fault Protection on Main
with only one OCPD or disconnect device for Breakers
an entire set of inverters. More details on the The NEC has had, for many editions, a require-
requirements found in 705.12(B) of the NEC and ment (230.95) that solidly grounded wye services
earlier editions will be found below. rated at 150–600 volts phase-to-phase and 1000
The OCPD must be sized at a minimum of amps or more have ground-fault protection
125 percent of the rated inverter output current (GFP). Connecting a PV inverter ac output
(or the total of the output-rated output current on the load side of these GFP-equipped main
from multiple microinverters or ac PV modules). breakers may pose safety issues.
It must also protect the circuit conductor from Circuit breakers are manufactured with numer-
overcurrents from the utility side of the connec- ous optional accessories, including (depending on
tion. It is usually not a good idea to install a larger the manufacturer and model) shunt trips, auxilia-
OCPD than the minimum required value (but ry switches, remote indicators, power operation,
allowing a round up to the next standard value adjustable trips, and GFP trip mechanisms.
is okay and needed) because the inverter may, as While UL Standard 489 requires tests for evalu-
part of the listing or instructions, be using the ating the backfeed suitability of the basic circuit
OCPD to protect internal circuits. The maximum breaker, most of the accessories are not evaluated
value of allowable overcurrent device should be in for backfeeding. In fact, backfeeding may have no
former to power a GFP device, and that

current transformer does not respond to
voltages on a tripped open main breaker.
The GFP device, powered by the current
transformer, should not be damaged by
backfeeding.
However, there is some confusion
and uncertainty about the older GFP/
breaker designs that are installed
widely and may still be on the market
(photo 4.9).
Although 705.32 and the Exception
no longer require that ground-fault
protection devices be identified and
Photo 4.9 • 1200-amp, 480-volt ground-fault-protected switchgear. listed for backfeeding, 110.3(B)
To backfeed or not? Only the manufacturer knows. requires that they be used according to
their listing and labeling. Ground-fault
effect on many of these accessories, and specific protection devices attached to circuit breakers
testing for backfeeding may be unnecessary. are not specifically tested and listed under UL
However, older and possibly some current Standard 489 requirements.
ground-fault trip mechanisms may be damaged Only the breaker manufacturer (the design
if the circuit breaker has voltages on both line engineer) can verify that the particular breaker
and load terminals after the breaker has been with GFP (by part number and model number)
opened by a ground-fault trip. UL 489 testing will not be damaged under PV inverter backfed
does not evaluate backfed GFP main breakers in conditions, where both line and load terminals
a manner that subjects the ground-fault device are energized at the same time during and after
to the conditions it would experience in a util- the GFP device trips the breaker. Bulletins and
ity-interactive PV system or possibly even in a information from sales departments are usually
parallel-connected generator installation where insufficient to make a determination.
line and load terminals are both energized After determining that the main breaker/
during and after a ground-fault trip. Photo- GFP is suitable, then the issue of protecting
voltaic inverters may have energized outputs the load circuits under ground-fault condi-
up to two seconds after the ac utility power is tions from all sources (utility and PV ) must be
removed from the output. addressed (705.32 EX). An engineering analysis
Circuit breaker manufacturers should be evalu- would be required that shows how and where
ating all accessories supplied with their breakers ground-fault currents are sourced. What are the
for operation under all possible application con- impedances involved in the utility source and
figurations. However, utility-interactive inverter the PV source? How much current can each
installations are relatively new applications and supply under varying types of ground faults?
inverters are being installed in electrical installa- Ground faults are not always “hard” low-resis-
tions that may be decades old. tance faults and may be arcing faults of varying
Informal discussions with manufacturers indi- impedances. Suppose the PV system sources
cate that most new designs use a current trans- enough ground-fault current to prevent the
main breaker GFP from tripping? How is the device. It should be tested a second time
ground fault contained or interrupted? to ensure that the device was not damaged
Some manufacturers are wary of putting during the first test.
some sort of GFP device on the inverter output
because this is a nonstandard connection and In many cases, it may be easier to implement a
any ground faults detected might only be those supply-side (of the main GFP breaker) PV con-
originating between the device and the inverter, nection as allowed by 705.12(A) and 230.2(A)(5)
not load ground faults. / 230.82(6). These connections will be discussed
Some manufacturers have a main breaker GFP in a later chapter.
that can take inputs from multiple ground-fault
sources like dual utility feeder systems. But these Functional Grounded Systems
would be found in limited, special instances It is likely that the type of PV system being
where there are multiple utility feeds. installed (with respect to grounding) will be de-
termined by the inverter topology. That topology
Summary will, in the future, most likely result in a func-
The requirements of the NEC are stringent, but tional grounded PV system. The internal methods
can be met. There are no one-size-fits-all solutions by which the inverter achieves the functional
to this issue. The following steps should be fol- grounding may, to some extent, determine the dc
lowed before connecting a PV system that could input and ac output circuit configurations. The
backfeed a GFP breaker. There may be others. inverter instruction manual will have the appro-
priate instructions.
1. Accurately determine that all ground-fault
protection devices are suitable for operation Newer Inverter Technologies
in a backfed manner with a utility-interac- Microinverters
tive PV inverter. The inverters that have been covered above are
2. Select an appropriate GFP device that can known as “string inverters” because they operate
be connected to the inverter’s outputs to with a string of series-connected PV modules.
control ground-fault currents from that These inverters range in power from more than
source. one megawatt down to approximately 700 watts.
3. Make an engineering assessment of the DC maximum system voltages range from 125
magnitudes of the potential and available volts to 1000 volts and higher. Inverters con-
fault currents from both utility and PV nected to one (or possibly two) PV modules are
sources to the load circuits being protected. classified as microinverters.
Circuit impedance calculations under fault The Enphase microinverter is a typical exam-
current levels for all sources and the load ple (photos 4.10 and 4.11). It is a small inverter
impedance should be made. (hence the name) that is designed to work with a
4. Determine the proper setting for all adjust- single PV module and to operate at a maximum
able-trip ground-fault protection devices that of approximately 70 volts on the dc input. The
will ensure that the load circuits are protected inverter is connected directly to the PV module
from all ground-fault current sources. using the existing conductors and connectors
5. With the GFP breaker being back fed with (now locking in most cases) attached to both the
current from the PV inverter, the GFP de- module and the inverter. Available units are rated
vice should be tested using the internal test in the 170-watt to 500-watt range, but as with
Photo 4.12 • Multiple microinverters connect to a trunk

Photo 4.10 • Enphase 400 W microinverter. Courtesy of cable. ANDALAY/WESTINGHOUSE ac PV modules with
ENPHASE. trunk cable integrated into module frames.
rated disconnects. With a power output in the

240-watt to 400-watt range (depending on the
model), the rated ac output current at 240 volts
will range from 1 amp to 1.2 amps. On the 14
AWG cable with a 15-amp overcurrent device,
the rated continuous current for that circuit is
limited to a maximum of 12 amps. This rating
will allow 1 to 12 inverters to be installed on the
same ac output cable. Trunk cables with larger
conductors will allow greater numbers of micro-
inverters to be connected.
Photo 4.11 • Westinghouse AC PV module with Enphase

AC Output Circuits
inverter showing ac and dc cables. The output circuit of any utility-interactive
inverter, including microinverters, up to the
other PV products, ratings and specifications first overcurrent protection device (OCPD)
change continually. is very much like an ac branch circuit. If the
The microinverter is a utility-interactive utility voltage is removed from this circuit (for
inverter. As part of a PV system (module and any reason), the circuit becomes de-energized
inverter), it must have dc ground-fault protection (dead)—just like a branch circuit. If there is
[690.41] per the requirements of UL 1741, just as a line-to-line or line-to-ground fault on this
is required in larger inverters. circuit, the OCPD responds in a normal manner
Microinverters have ac input and output to the fault currents generated by the utility. The
cables and connectors and are listed in a manner inverter can generate no more than its rated cur-
that allows multiple inverters to be connected rent per UL Standard 1741, and when the fault
with up to about 15 units on a single output occurs, the drop-in line voltage will normally
cable (called a trunk cable) (photo 4.12). The cause the inverter to shut down. When the
dc and ac connectors used with the Enphase branch circuit breaker opens in response to the
microinverter have been listed as load-break- fault, the inverter shuts down.
It would appear that these inverter output cir- Nevertheless, it is possible to mount an ac load
cuits could be wired using any Chapter 3 wiring center on a roof with proper solar shielding and
method suitable for the environment (indoors, or use it to combine the outputs of U-I inverters or
hot, wet, and UV outside, and hot in attics). sets of microinverters.
The metal raceway requirement of 690.31(G) The rating of any combining panel and the
applies only to the always-energized dc PV ampacity of conductor from that panel to the
source and dc PV output circuits on or in a backfed breaker in the main load center (as well
building. as the rating of the main load center and the
Flush-mounted inverters are starting to appear backfed breaker) must meet 705.12(B) require-
on the market, and they will have a ventilation ments that are covered in Chapter 7.
system to remove the internal heat generated The ac output conductor for a set of inverters
during the inverting process. must have an ampacity of 125 percent of the
An ac GFCI device should not be used to continuous currents for all the inverters on that
protect the dedicated circuit to the microinverter circuit. The backfed circuit breaker in the panel
or ac PV module, even though it is an outside must be rated the same, and if an odd current rating
circuit. None of the small GFCI devices (5 mA to is determined, the breaker rating should be the next
30 mA) are designed for back feeding and will be larger size. The breaker must protect the conductor
damaged if backfed. In a similar manner, most ac under the conditions of use, and the conductor am-
AFCIs have not been evaluated for backfeeding pacity must be derated for those conditions of use.
and may be damaged if backfed with the output of The ac output circuit from each set of inverters
a PV inverter. must have an equipment-grounding conductor
to facilitate OCPD operation during ac ground
Combing Multiple Sets of faults. Some microinverters have a three-wire
Microinverters or AC PV output through a four-contact connector. The
Modules unused terminal in the connector is reserved for
In multiple strings of inverters, there is no NEC future use. The three active pins in the connector
requirement that an ac combining panel (load are 240 V L1 and L2 and a neutral. There is no ac
center) be located on the roof. In fact, most equipment-grounding conductor. This lack of an
NEMA 3R load centers must be mounted equipment-grounding conductor in the cable re-
against a surface to keep water from penetrating quires that the equipment-grounding conductor
holes in the back panel. Such a surface may have for the microinverter or ac PV module be an ex-
to be added to properly mount a 3R load center ternal connection to the inverter case, where the
on a roof. There might be problems meeting case is metal. This external equipment-grounding
110.26 clearance requirements. conductor must be connected to the fixed wiring
A further issue with an OCPD on a roof is system (usually, but not always conduit) where
heating of the device over its rated 40°C oper- that wiring system originates. The ac module or
ating temperature. Gray load centers in the sun microinverter instructions will cover grounding
will normally operate 10°C to 20°C hotter than requirements and should be followed [110.3(B)].
the local ambient temperature (which, in some Unless the microinverter bracket has been de-
cases could be 40–50°C). This may be difficult signed and evaluated as a grounding or bonding
to compensate for when considering available jumper, grounding the microinverters does not
equipment, the size of ac conductors attached to ground the rack or the modules (and vice versa).
inverters, and listing restrictions on the inverters. There is only one ac neutral-to-ground bond in
nation, the microinverter dc connec-

tion to the PV module may have to
be disconnected to replace the micro-
inverter should it or the module fail.
While the voltage will be a maximum
of approximately 70 volts with current
inverter designs, the current may be
in the 3-amp to 7-amp range and the
connectors could possibly be damaged
at this voltage and current, posing a
possible safety hazard. While very few
inspectors may request a costly and
impractical load-break-rated discon-
Photo 4.13 • SUNPOWER AC PV Module. Courtesy of Sunpower. nect, the Code-compliant solution
is quite simple. The back of the PV
an ac electrical system. That bond is made in the module must be accessed to reach
existing service-entrance equipment. No addi- these dc connections, and this generally requires
tional neutral-to-ground bonds should be made that the module be unfastened from the mount-
when installing a PV system unless a supply-side ing system. Because the module is accessible
service-entrance connection is made or trans- and is being accessed, just by putting a blanket
former isolation establishes a separately derived or other opaque material over it per 690.18
system. (2104 NEC) will reduce the dc output voltage
and current (and the ac current) to near zero,
AC PV Modules allowing the module/inverter dc connectors to
Take a normal dc PV module in the factory and be safely opened. Opening this connection with
connect a microinverter to it, fasten the microin- the module blacked out will likely be safer than
verter to the back of the module, and you have an opening the same connectors on a module in a
ac PV module (photo 4.13). AC PV modules are high-voltage string of modules.
becoming more common in the market. One is Each microinverter or ac PV module will have
the ac PV module by Westinghouse Solar/Anda- an ac output trunk cable to allow the multiple
lay Solar and it has a unique frame that is also the inverter parallel connections. This cable may
module mounting rack (photo 4.12). Photo 4.14 carry currents in bright sunlight of 1 amp or
shows an LG/Enphase ac PV module. Because more at 240 volts (in current designs) from the
the dc wiring between the module and inverter is first module/inverter in the set to as much as 12
listed with the entire ac PV module and becomes amps at 240 volts through the last connector of
an integral part of the product, dc requirements the set that has multiple series-connected devices.
in the Code no longer apply to the ac PV module. Servicing the single ac PV module or utility-in-
The ac PV module is a utility-interactive device teractive microinverter could be accomplished by
and has a similar ac output cabling system to the covering the module to reduce the dc and, hence,
microinverter addressed above. the ac current to zero.
However, not covering all modules in the set
DC Connections would allow current from other, non-covered,
In a standard PV module/microinverter combi- modules/inverter to flow through the cable and
the ac PV modules or microinverters will

enhance safety, but is not a Code require-
ment. A common 60-amp unfused, pullout,
air-conditioning disconnect costs less than
$10 at building supply centers. It provides
the disconnect; a place to terminate the
ac trunk output cable from a single set of
microinverters or ac PV modules; a place to
originate the field-installed wiring system
to the ac load center in the house; and is
usually cheaper than a separate junction
box and cover.
Some microinverters and ac PV mod-
ules may have dc and ac connectors fully
listed as load-break-rated disconnects.
The Enphase microinverter has both the
ac and dc connectors listed as load-break-
rated disconnects. Such a rating would
meet 690.15 disconnecting/isolating
means requirements and the separate
rooftop disconnect would not be required
(although the device mentioned above
would still be an inexpensive transition
point between the inverter/module ac cord
Photo 4.14 • AC PV module, LG NeON 2 AC, as defined in
and the conduit conductors leading from
690.2 and 690.6. Courtesy of Enphase.
the roof ).
at 240 volts, could damage the connector and
possibly pose a shock hazard when opening these AC PV Module Maintenance
ac connections under load. To some extent, the —A Gray Area
hazard is minimized because the inverter anti-is- Combinations of a microinverter and a PV
landing circuits shut down very rapidly, reducing module with exposed dc connectors and dc
any arcing when the ac connector is opened. conductors between the PV module and the mi-
Opening the ac circuit at the PV backfed croinverter are being certified and listed as ac PV
breaker in the building service-entrance panel modules. Some of these products have instruc-
would be a safe solution if that breaker could be tion manuals that say the microinverter may not
locked open. However, breaker locks are few and be removed or disconnected for service from the
far between and lock-out/tag-out procedures are PV module to affect a repair of either component.
not generally used in residential and commercial In fact, such a separation may be difficult in the
electrical systems. field and would violate the listing on the ac PV
Where disconnect/isolation requirements for module and violate the warranty.
each microinverter can be met by the use of con- Other manuals give specific instructions for
nectors on the dc input and ac output, the instal- removing the microinverter from a PV module
lation of a separate ac disconnect on the roof near for repair. At issue is the definition of an ac PV
module as a factory assembled unit and the 690.6

exemption from all dc Code requirements for
these products with exposed dc connectors and
dc conductors. However, exposed connectors are
subject to loosening or being opened in the field.
Connectors and conductors are exposed to envi-
ronmental degradation, ground faults, arc faults,
and animal damage.
Also at issue is the microinverter-to-PV
module frame bonding when the mechanical
or electrical connection is broken in the field.
When the microinverter is replaced, how is the
bonding connection quality verified and how is
the certification or listing maintained without
NRTL evaluation?
At some point, these issues will be addressed
in UL Standard 1741 and possibly in the NEC.
Advantages Will Boost the Market

Photo 4.15 • Numerous parts are required for a string-in- The use of microinverters and ac PV modules will
verter system.
proliferate due to several advantages they offer
over conventional string inverters.
The first is a simplified set of installation require-
ments and a reduced number of separate parts. See
photos 4.15 and 4.16 for quantitative differences
in the amount and types of equipment involved in
installing an ac PV module system versus a conven-
tional string inverter system.
In a dc series-connected string of PV modules,
module mismatch is sometimes an issue that
affects the string performance. Modules come out
of the factory with slight (up to 10%) variations in
performance. The string of modules in a dc system
cannot deliver current above the current delivered
by the weakest module in the string. The mismatch
between module currents results in some lost
power compared with a dc string of modules that
is equal in every specification.
Photovoltaic modules near the top of an array
on a sloped roof may operate hotter than modules
lower down on the roof due to hot air rising be-
Photo 4.16 • Far fewer parts are needed to install the An-
dalay AC PV module system, which includes the rack built hind the modules. Depending on how each string
into the heavy module frame. of modules is connected, some loss of power may
inverter. If a short circuit or ground fault were to

occur in these ac output circuits, the dedicated
branch-circuit breaker would open and the circuit
would go dead. Opening the main service discon-
nect or the backfed PV breaker will de-energize
those PV ac output circuits and is one method of
achieving compliance with 690.12 rapid shutdown
requirements—a boon to fire fighters.
Disadvantages
There may be some cost impact of using ac PV
modules or microinverters on each module when
compared with using single-string inverters.
However, two factors must be considered. The
cost of dc switchgear and the required conduit (or
other appropriate wiring method) for dc conduc-
tors inside a building, plus the cost of the single
Photo 4.17 • DC-to-dc power converter attached to the inverter, must be compared with the added cost
back of each PV module. Courtesy of SoarEdge. of multiple small inverters or ac PV modules with
an inverter on each module.
occur if hot modules are connected in series with Then there are the life cycle costs. Modules are
cooler modules. guaranteed for power production for 25 years,
Shading is also a problem in a conventional but can be expected to produce power for as long
string-inverter configuration. The shading of a as 50 years. Large inverters generally have an
single module will result in a power loss from average longevity of about 15 years. Microinvert-
that module, but may also reduce power from the er manufacturers, using different construction
other, non-shaded modules in the string. methods and topologies, are predicting signifi-
The microinverter and the ac PV module work cantly longer lives for their products. However,
at the individual module level. Each inverter ex- the failure of a microinverter in a system or on
tracts the maximum power from that module no an ac PV module may necessitate disassembling
matter what the other modules in the PV array some portion of the PV array to gain access to
are doing. The output of each is independent the failed device. Time will reveal all.
of the other modules/inverters in the set. The
outputs of the microinverters or ac PV modules DC-to-DC Converters
are connected in parallel, rather than in series, DC-to-dc converters are on the market. They are
and this isolates one from another. small boxes with leads that attach to the output
The outputs are at 120, 208, and 240 volts ac and of each module. These boxes act somewhat like
these ac output circuits act much like ac branch a microinverter by decoupling the actions of one
circuits. They go dead when the ac utility power is module from others in the string so that shading
removed at any disconnect in the circuit, so they is less of a problem. Of course, the output of
do not pose the safety hazards associated with the these devices is dc, not ac like the microinverter.
daytime “always-energized” dc circuits operating Some of these devices can be used on just a few
at hundreds of volts between the modules and the modules in a PV array, but others are required for
every module. Some devices work with a stan- or topology. However, because all of these dc-
dard string inverter, but others require a special to-dc converters should be listed per 690.4(D),
inverter made by the same manufacturer. Most installers and inspectors are generally going to
are connected in series like a normal string used have to rely on instruction manuals and label-
with a string inverter. ing supplied with these new devices for proper
In some cases, the specialized inverter re- system design and installation [110.3(B)]. The
quired with some of these devices communicates concepts of short-circuit current and open-cir-
and controls the performance of the individual cuit voltages used in the design and installation
dc-to-dc converters connected to each module. of standard string module systems will not
Removing the ac power from the inverter or apply to these products.
turning it off can, in some cases, shut down the
individual converters, removing dc voltages from Summary
the wiring. Numerous microinverters and ac PV modules
These devices have a wide variation in system are being installed, and while they are directed at
design and connection requirements. There the smaller residential systems, systems as large
are several sections of the NEC that address as several megawatts have been installed using
general requirements for dc-to-dc convert- microinverters. They are being sold in home
ers and they are found in Figure 690.1(a), improvement centers, building supply houses, and
690.2, 690.4, 690.7(B), 690.8(A)(5) and (6), local electrical supply houses and the public is
690.9(A), 690.11, 690.15, and 690.53(3). The buying them. Inspectors should become familiar
requirements are self-explanatory and are not with these devices and the Code requirements
associated with any particular product design that apply to them.
Chapter 5 Energy Storage Systems (ESS), Batteries in PV Systems 97
98 Chapter 5 Energy Storage Systems (ESS), Batteries in PV Systems
05
Energy Storage Systems;
Batteries in PV Systems
Article 706, Energy Storage battery systems operating at below 60 volts dc, and
Systems this would include PV systems using 12-volt and
Article 706, Energy Storage Systems (ESSs), was 24-volt nominal battery systems and a few 48-volt
added to the 2017 NEC, and all references to bat- nominal systems. However, some of the require-
teries or other energy storage systems previously ments, in Article 480, like for battery disconnects,
found in Article 690 are now in this section. This apply only to systems operating over 60 volts dc.
new section is quite lengthy and has requirements Installers of these low-voltage, but high-current,
that were not previously in the Code, such as a battery systems will have to exercise common
requirement to determine and mark the rated sense in designing these systems, as will AHJs
short-circuit current of the ESS. Article 480, inspecting them. Article 720, Circuits Operating
Storage Batteries, remains in the Code and has at Less Than 50 Volts (direct current or alternating
some of the same new or revised requirements current), is not going to help because it excludes
that appear in Article 706. PV systems in Article 690, Parts I and VIII.
Note that a PV system, as defined in Article
690, is separate and distinct from any energy “706.1 Scope. This article applies to all
storage system that it may be connected to. This permanently installed energy storage systems
separation, and the new grounding definitions (ESS) operating at over 50 volts ac or 60 volts
and requirements for a PV system, may compli- dc that may be stand-alone or interactive with
cate the grounding requirements for both systems other electric power production sources.”
where they are combined. It will take time, and
possibly the 2020 NEC will clarify this. Energy Storage Systems
Section 690.71 states that an energy storage Defined in Detail in the Code
system connected to a PV system must be installed “Energy Storage System (ESS). One or more
in accordance with Article 706, which applies to components assembled together capable of stor-
ESSs operating over 50 volts ac or 60 volts dc. It ing energy for use at a future time. ESS(s) can
would appear that now only Article 480 applies to include but is not limited to batteries, capacitors,
and kinetic energy devices (e.g., flywheels and nents but, instead, are composed of individual
compressed air). These systems can have ac or dc components assembled as a system…
output for utilization and can include inverters
and converters to change stored energy into Informational Note: Other systems
electrical energy. will generally be comprised of dif-
ferent components combined on site
“Energy Storage System, Self-Contained. to create an ESS. Those components
Energy storage systems where the compo- would generally be tested and listed
nents such as cells, batteries, or modules and to safety standards relevant to the
any necessary controls, ventilation, illumina- application.”
tion, fire suppression, or alarm systems are
assembled, installed, and packaged into a This chapter will examine only chemical energy
singular energy storage container or unit. storage systems (batteries) that might be found
in an off-grid, stand-alone PV system or in a
Informational Note: Self-contained residential or small commercial PV system with
systems will generally be manufac- battery backup for part or all the loads. In many
tured by a single entity, tested and cases, these battery systems will be classified as an
listed to safety standards relevant to Energy Storage System, Other, because they will be
the system, and readily connected on field-assembled and installed at the final location.
site to the electrical system and in the
case of multiple systems to each other. Batteries in PV Systems
Electrical power outages are becoming more
“Energy Storage System, Pre-Engineered common in recent times with man-made and
of Matched Components. Energy storage natural disasters and aging utility infrastructure.
systems that are not self-contained systems With natural disasters, such as Hurricane Sandy,
but, instead, are pre-engineered and field-as- tornadoes, and other severe weather conditions,
sembled using separate components supplied many people who are already using PV systems
as a system by a singular entity that are (and many that do not have PV systems) are
matched and intended to be assembled as an going to be interested in using PV systems in
energy storage system at the system installa- the event of electrical power outages. Electrical
tion site. inspectors can expect to see increasing numbers
of battery-backed-up, utility-interactive PV
Informational Note: Pre-engineered power systems.
systems of matched components for
field assembly as a system will gen- PV Plus Batteries Means Power When
erally be designed by a single entity the Utility Goes Out
and comprised of components that These backup systems allow owners to operate
are tested and listed separately or as some or all the loads in a building using a special-
an assembly. ly designed and configured PV system with bat-
teries in the absence of the utility service (photo
“Energy Storage System, Other. Energy 5.1). These systems can be as small as a system
storage systems that are not self-contained or that can power a radio or cell phone charger.
pre-engineered systems of matched compo- They can also be as large as necessary to run all
appliances and loads in a residence or commercial

building. The size and number of electrical loads
that can be operated, and the period of time they
can be operated, depends on the size of the PV
power system, the size of the battery bank, and
the size of the specialized inverter.
There are characteristics of PV systems with
batteries that are different from those relating to
standard utility-interactive PV systems. Obvi-
ously, batteries pose some unique problems that
an inspector must review, and the connection of
inverters to not only the electrical system in the
house but also to the utility requires looking at
some different Code sections than are normally
used.
The multimode inverter that is used has charac-
teristics of both the utility-interactive inverter
and the standalone, off-grid inverter with features
that are unique to the multimodal inverter. These
inverters will be listed to UL Standard 1741.
These inverters will have two sets of ac input/
output terminals and a connection for the battery
Photo 5.1 • Battery-backed-up, utility-interactive PV sys- bank. Photo 5.1 shows the batteries and the
tem. DC coupled with PV array on the roof of a house. multimode inverters in a system being installed.
Figure 5.1 • Components in a battery-backed-up, utility-interactive PV system. The PV array may feed the batteries by
two different means. See text and the following Figures.
Figure 5.2 • DC-coupled system interconnections and power flows.
Figure 5.1 shows the basic elements of a The designated protected (backed up) loads
battery-backed-up, utility-interactive PV system. may be supplied by either the utility (when pres-
Green arrows represent dc power/energy flow ent) or by the PV inverter output (supplied from
and red arrows represent ac power/energy flow. the batteries when the utility is absent). Where
Double-headed arrows represent bidirectional the PV system power output exceeds the building
power/energy flow. loads, the excess energy is fed into the utility and
renewable energy credits or net-metering bene-
DC-Coupled Battery Charging fits may be accrued. At night, or at other times
There are two main types of battery-backed-up, when the PV production is low, power for the
utility-interactive PV systems. The first (and old- loads is purchased from the utility and fed to the
est) is what is called a dc-coupled charging system. main loads through the main panel or through
As shown in Figure 5.2, the PV array has a nom- the multimode inverter to the protected loads.
inal voltage of 24 volts or 48 volts and normally In general, the battery stays fully charged at all
operates through a charge controller to charge a times, but there are some systems in which the
battery bank. The battery bank is connected to a stored energy in the battery can be sent (“sold”)
multimode, utility-interactive inverter and that to the utility with proper programming of the
multimode inverter is connected to the house loads equipment.
and to the utility using two separate and distinct ac When the utility is not present, the PV array
input/output circuits. When the utility is present, and battery combination and the multimode
the PV system charges the batteries through the inverter continue to operate the loads connected
charge controller. Power is taken from the batteries to the protected loads subpanel to the extent
(or directly from the PV system when the batteries that the size of the PV system and the capac-
are fully charged) through the multimode inverter ity of the battery bank can supply the energy
where it is converted to ac power. required by those protected loads. The multi-
Photo 5.2 • DC-coupled system. Three black charge con-

trollers on the right and four black inverters in the center.
Silver enclosures contain ac and dc distribution equip-
ment.
Photo 5.4 • 60-A charge controller.
utility interactive once again. Photo 5.2 shows a

dc-coupled battery charging system. The three
charge controllers are on the right and the four
inverters are in the center between the ac and dc
distribution panels.
Charge Controllers
Standalone systems and dc-coupled utility-in-
Photo 5.3 • 30-A charge controller with display. Courtesy
teractive systems with battery banks will also
of Blue Sky Energy. have charge controllers that regulate the state-
of-charge of the battery bank. Charge controllers
mode inverter will not send power to the main come in many sizes, shapes, and colors (photos
(unprotected) loads or to the utility connection, 5.3, 5.4, and 5.5). When properly adjusted, they
but continues to monitor that utility connection protect the batteries from being overcharged.
for voltage and frequency. The main panel gets Installers are responsible for adjusting these
no power from any source. When the utility devices properly. Inspectors should verify good
comes back online with the proper voltage field terminations, proper conductor sizes, and
and frequency characteristics, the multimode appropriate overcurrent devices protecting those
inverter will reconnect and the system becomes conductors.
AC-Coupled Battery Charging

Figure 5.3 shows a more recent type of system,
known as an ac-coupled charging system, where
the PV modules are usually configured in a
high-voltage string configuration (200 volts to
600 volts) and provide dc voltage to a standard
utility-interactive inverter. The output of the
utility-interactive inverter is connected to the
protected load subpanel with a backfed breaker
[705.12(B)] and that subpanel is connected to
the load ac input/output terminals of the mul-
timode inverter. The battery again is connected
to the multimode inverter dc input/output. The
utility is connected to its unique ac input/output
on the multimode inverter, and when the utility is
present, it feeds through the multimode inverter,
generally keeping the batteries charged at all
times and providing energy to the protected load
subpanel. The utility-interactive inverter sees the
proper voltage and frequency supplied by the
utility and continues to convert dc PV energy
into ac energy that can be used by the loads (both
Photo 5.5 • 96-A charge controller. Courtesy of MidNite. protected and main) and be fed to the utility.
Solar When the utility goes down or has a brownout
Figure 5.3 • AC-coupled system interconnections and power flows.

Photo 5.6 • AC-coupled system. The gray utility-interactive string inverters are at the top and yellow hybrid inverters be-
low. The storage batteries are in green cases. The rotective plastic shield between batteries and electronic equipment is
not shown.
(voltage or frequency variation), the multimode can operate 24 hours a day, so the total amount
inverter senses this and stops sending power to the of PV array energy that can be stored in the
now unenergized utility lines (and the main load battery—and the capacity of the battery and
panel). However, it continues to monitor them for size of the inverter—determine how long the
proper voltage and frequency, which would indi- loads can be operated and how many loads can
cate that the utility is back online. At this time, on be connected at any one time.
the load ac input/output of the multimode invert- Photo 5.6 shows an ac-coupled, battery-
er, the battery supplies energy to the inverter and backed-up, utility-interactive system. The gray
it will become the correct frequency and voltage utility-interactive inverters are above the yellow
reference source to supply not only the protected multimode inverters and the batteries are in the
loads, but also to keep the utility-interactive in- rear of this very compact installation. There is
verter connected to the PV system, operating and normally a clear insulating service panel in front
producing energy (in the daytime). of the batteries; the panel was removed when the
Again, the size and number of loads that can photo was taken.
be connected and operated for any short or In either case, with dc charging or ac-coupled
long period of time depends on the size of the charging of the batteries, the certified or listed
PV array and the capacity of the battery bank. multimode inverter ensures safety for the power line
Typically, the PV array may only supply energy and utility personnel anytime the utility is shut-
for four to six hours per day. Loads obviously down or operates abnormally.
Battery Considerations battery are sized based on the rated output of the
Batteries, although not considered a source of charge controller, irrespective of the size of the
energy, can store considerable amounts of energy. PV system feeding it. These conductors should be
They should not be considered current-limited sized at 125 percent of the rated output current
sources as are PV modules, but, rather, they have of the charge controller [706.23(C)]. There
the characteristics of a constant-voltage output, should be an overcurrent device and a disconnect
like an ac feeder with large amounts of available at the battery end of the circuit to protect these
short-circuit current. Batteries must have over- conductors from high short-circuit currents orig-
current protection and disconnects on the output inating at the battery. Depending on the location
conductors. Batteries operating over 100 volts of the charge controller with respect to other
(conductor to conductor or conductor to ground) components, there may be disconnects required
shall also have a maintenance disconnecting on the input and output of the charge controller.
means for both ungrounded and grounded con- A main PV system disconnect located between
ductors that is accessible only to qualified persons. the PV array and the charge controller will be
Where battery circuits are subject to field servic- required, complying with 690.13.
ing and operating at over 240 volts (conductor to Available short-circuit currents. The battery
conductor or conductor to ground), they shall have banks used in these types of systems will typical-
provisions to disconnect the battery series-con- ly have an available short-circuit current at the
nected strings into segments that are less than 240 output conductors from the battery bank less than
volts (706.21, 706.30). The current between the 15,000 amps. Conductor lengths, connections, and
battery and the multimode inverter is bidirection- conductor resistances limit the available short-cir-
al. It flows to the batteries when the batteries are cuit current. Any overcurrent devices or discon-
being charged by the multimode inverter or the nects must have ratings that can handle currents
charge controller, and it flows from the batteries of this magnitude. Current-limiting fuses and dc
when the multimode inverter is in the inverting rated circuit breakers are generally available with
mode supplying the protected loads with ac power. sufficient ratings and should be used.
In the dc-coupled charging system, the con- Conductors. The conductors between the bat-
ductors between the charge controller and the tery bank and the multimode inverter must carry
bidirectional currents. The multimode
inverter will use utility power or power
from the utility-interactive inverter in
ac coupled systems to keep the battery
charged and currents will flow from
the inverter to the battery. When the
multimode inverter is operating in the
inverting mode and supplying protect-
ed loads with energy, the currents will
flow from the battery to the multimode
inverter.
In general, the discharging currents
flowing from the battery to the inverter
Photo 5.7 • Class B stranded conductor, top; fine stranded will be larger than the charging currents
conductors, bottom. flowing from the inverter to the battery.
This is because the typical multimode inverter

will be able to draw more current from the
battery than it can provide to charge the battery.
Therefore, the conductors between the batteries
and the inverter must be sized based on the
maximum rated output of the multimode inverter
in the inverting mode of operation. However,
if the multimode or stand-alone inverter in-
struction manual indicates a higher current in
another mode of operation, then that current
should be used. This continuous current (called
the maximum current) should be specified in the
inverter specification or installation manual and
the conductor sized at 125 percent of this max-
imum current (706.20). Of course, the battery
conductors should be in a raceway along with an
equipment-grounding conductor, which would be
used to ground any metallic battery rack and bat-
tery disconnect or overcurrent device enclosure.
The size of the equipment-grounding conductor
would be based on the rating of the overcurrent
device protecting the circuit (250.122).
Many premanufactured battery cables are
Photo 5.8 • Battery disconnect and overcurrent protection
made with fine-stranded cables consisting of located near the batteries.
type appliance wire material (AWM) conductors.
These cables are not suitable for use in battery PV Battery Circuit Overcurrent Protection and Dis-
systems because they are not mentioned directly connects. An overcurrent device should be located
in the NEC as one of the Chapter 3 wiring ma- at the battery end of the circuit to protect this
terials suitable for field-installed wiring. In many conductor from high available fault currents from
cases, automotive battery cables and welding the battery. This overcurrent device will be sized
cables have been used for their flexibility, but they at 125 percent of the standalone or multimode
are not allowed in this application by the Code. inverter rated dc current in the inverting mode,
To address flexibility issues, see the fine-stranded which is the same number used to size the con-
cable requirements in Section 110.14. ductors. An overcurrent device at the inverter end
THWN, RHW, THW and other commonly of the circuit is normally not required because the
available conductors listed in Chapter 3 of the inverter typically cannot source the same high fault
Code are acceptable for battery installations. currents that the battery can. A battery disconnect
Most commonly available plastic (thermoplastic) should be installed at the battery end of the circuit.
or rubber-insulated (thermoset) conductors are Normally, if the inverter is within four to five feet
considered acid- and moisture-resistant. of the battery bank, it may not be practical or pos-
However, the use of fine stranded flexible cables sible to put a disconnect any nearer to the battery
may pose termination issues. See NEC 110.14 than this distance. Therefore, the disconnect for
and photo 5.7. this circuit can be near or at the inverter—usually
in a power center. However, if the distance between rent protection and disconnects must be used in
the battery and the multimode inverter is more both ungrounded conductors; therefore, increasing
than four to five feet, or if the inverter is in a dif- the cost of an ungrounded battery system.
ferent room than the battery bank, then there must
be a disconnect at the battery end of the circuit in AC Circuit Considerations
addition to the overcurrent protection required at Multi-wire branch circuits. Many houses today
that location. Photo 5.8 shows a battery discon- have several multi-wire branch circuits that have
nect/overcurrent protection enclosure using circuit two branch circuits with a shared neutral conduc-
breakers mounted just above a valve-regulated tor and are wired with a 14–3 AWG/with ground
(sealed) lead-acid (VRLA) battery bank. These type NM cable. Multimode inverters come with
batteries release no hydrogen gas or acid fumes either 120-volt ac outputs or 120/240-volt ac
during normal operation. outputs. Neither of these multimode inverters
Grounding. The nominal battery voltage in should be connected to load circuits in a building
these systems is 48 volts dc. The operating voltage that are part of a multi-wire branch circuit. [See
may be as high as 62 volts to 65 volts. Normally, NEC 690.10(C).] The inverters in the inverting
multimode inverters do not ground one of the mode (standalone mode), in some cases, may
battery circuit conductors and the grounding not be in synchronization with the utility power
system used by the multi-mode or standalone frequency waveform. This could cause over-
inverter will have to be coordinated with the loading of the shared neutral that is associated
grounding system used by the connected PV sys- with multi-wire branch circuits. If any of the
tem (690.41). Equipment grounding is required circuits needing battery backup power protection
throughout the combined system and the require- are multi-wire branch circuits, they should be
ments for this grounding have remained relatively segregated in their entirety (both circuits) in the
unchanged in recent editions of the NEC. special protected loads load center that is con-
If the system uses dc-coupled battery charging, nected to the multimode inverter ac output.
the connection to earth for the PV system will Utility connections. One of the characteristics
usually be done through a distinct and sepa- of most multimode inverters is that they can pass
rate ground-fault detecton interruption system power from the utility through to the protected
(GFDI) as required by NEC Section 690.41. load circuits at a greater power level then they can
In some cases, the charge controller may have a supply power to the utility in the utility-
GFDI built in. interactive mode. This indicates that the circuit
On an ac-coupled system, the utility-interactive and overcurrent device (typically a breaker)
inverters will have their normal GFDI internal between the utility connection and the multimode
circuitry and will also be required to meet the inverter must be rated at the full pass-through
requirements for PV system grounding in 690.41. current capability of the inverter. A common value
However, in ac-coupled systems, the dc battery of this circuit breaker would be 60 amps or 70
circuit may still have to be solidly grounded to amps. However, in the utility-interactive mode, the
keep costs down and to be compatible with avail- inverter may only be able to source 33 amps from
able equipment that has been designed for use in the PV system into the utility. In previous editions
grounded systems. Where an ungrounded battery of the Code, the 60-amp or 70-amp breaker would
system is used, there is no provision for having an be used in the 705.12(D) calculations to determine
overcurrent device in just one of the ungrounded panelboard/load center busbar ratings and conduc-
conductors as there is for PV dc circuits. Overcur- tor sizes. The danger to the circuit from overload-
Photo 5.9 • Flooded lead-acid batteries with acid containment system.
ing, however, is related to the 33-amp output of manner as any normal utility-interactive system.
the inverter when feeding the utility. The dc PV circuits are connected in the same
Section 705.12(B)(2) now requires the calcu- manner as those circuits in a standard utility-
lations for this requirement to be based on 125 interactive PV system for the ac coupled system.
percent of the rated utility-interactive inverter The dc-coupled systems require additional con-
output in the utility-interactive mode. In this siderations for the low-voltage battery charging
example, 41.25 amps (1.25 x 33) could be used in circuits.
the calculations. The circuit breaker connecting the
inverter to the load center can still be rated at the A Closer Look at Batteries
higher 60 amps or 70 amps required to allow the Energy storage systems (in the form of batteries),
protected loads to be operated in the pass-through when included in photovoltaic power systems,
mode of operation. Of course, the circuit conduc- are critical and important items that need close
tors must be sized to carry the higher currents in scrutiny during the plan review and inspection
the pass-through mode, including any ac currents process.
from the utility used to charge the battery.
Aside from the battery circuits and the unique Types of Batteries
characteristics of the utility interconnection There are two main types of battery systems that
covered above, the multimode inverter in the have been used (and are still being used) in PV
battery-backed-up, utility-interactive PV system systems at this time. Both are based on a lead-
is connected to the utility in much the same acid battery technology (older than the NEC).
The oldest technology uses a flooded lead-acid

battery that has removable vent caps and re-
quires the addition of water on a regular basis
(photo 5.9). A related technology known as the
valve-regulated, lead-acid (VRLA) battery uses
an internal chemical design to allow the battery
case to be effectively sealed (photo 5.10). No
water can be added to VRLA batteries.
The newer battery types of nickel cadmium and
lithium ion have not found their way into PV
systems in any substantial numbers yet, probably
Photo 5.10 • Valve regulated lead-Acid (VRLA) batteries.
due to the higher cost of these technologies.
Increasing use of these battery types in electric hydrogen and sulfuric acid fumes. Both battery
vehicles is lowering the costs and they are being types should be installed in a well-ventilated area
used more frequently in PV applications. (See and should normally not be installed in a living
the AC Battery Section below.) Most lithium space.
ion battery systems require a charge management In normal operation, the amount of hydrogen
system for each battery cell and that system gas and sulfuric acid fumes are very limited and
increases the cost substantially. Other advanced are quickly and easily dispersed into the sur-
technology battery types, such as chemical rounding room without problems. However, if
“flow” batteries, are generally found only in large, these gases are restricted from being diluted with
experimental installations in utility and industrial the surrounding air, an explosive combination of
applications. gases and air is possible. Garages and outbuild-
ings are ideal locations for these batteries if they
Mechanical Installation, Venting, and Acid are well ventilated, not sealed, not used as living
Containment spaces, and do not have living spaces over them.
Batteries contain substantial amounts of lead, No attempt should be made to construct a
which makes them very heavy. The floor that they manifold or venting system for these batteries.
rest on must be sufficiently strong, particularly for Power venting systems have a tendency to fail
larger battery banks that may weigh several tons. and the high-vent, low-vent system used for
Where battery banks are installed in racks that combustion appliances, such as gas furnaces and
are more than a foot or so high, those racks must water heaters, is not applicable to batteries beca-
be secured to a substantial wall. Building codes, suse there is no driving energy source equivalent
especially in earthquake zones, require substantial to the heat energy source of a heating system.
mechanical securing of batteries. Power-vented manifolds on the batteries
During the normal charging process, flood- should not be used, as they have been responsible
ed lead-acid batteries will emit hydrogen gas for explosions and fires in the past. Batteries
and sulfuric acid fumes into the surrounding installed in well-vented locations are acceptable.
environment. The type and adjustment of the Catalytic battery vent caps, employing a platinum
charge controller will determine the amount of catalyst, are available for use on flooded lead-acid
outgassing. Although VRLA batteries are sealed, batteries to reduce hydrogen outgassing and re-
if the charge controller is misadjusted, becomes duce the requirement to add water to the battery.
defective, or fails, VRLA batteries will also outgas Flooded lead-acid batteries contain liquid
tion, disconnects, and charge-control functions.

Even in the best of circumstances, it is generally
not possible to mount and connect an overcurrent
device or a disconnect any closer to the battery
terminals than three or four feet.
Conductors must be sized based on the contin-
uous current requirements, which will be based
on charging currents from either the PV array or
an inverter/charger or on discharging currents,
usually to the inverter. Depending on the system
design, there may be more than one circuit con-
nected to the battery—one circuit for charging,
Photo 5.11 • Flooded lead-acid batteries in lockable poly-
and one circuit for discharging. Where there is a
ethylene containers.
single circuit used for both charging and discharg-
sulfuric acid, and VRLA batteries include a similar ing, the conductor must be sized on the largest
jelled electrolyte (photo 5.9). These batteries continuous current in either mode. Although
should be installed in a manner that prevents the Code typically requires that conductors be sized
them from being mechanically abused. Many local and protected based on continuous loads (3 hours
building codes require some sort of acid contain- or more), battery-based systems may require a
ment for these batteries should the cases become slightly different approach.
cracked. Installing batteries in an acid-resistant With standalone, off-grid inverters, and multi-
outer container may be sufficient. Note the acid mode inverters in the standalone mode, inverters
containment system under the batteries in photo have a substantial capability to surge currents
5.8. There are no generally available, certified or above their steady-state rating for periods of
listed battery containers or battery boxes. For minutes to an hour. This may create voltage drop
smaller systems, it has been found that heavy-duty in the conductors that poses operational problems.
polyethylene toolboxes provide sufficient acid In this case, voltage-drop calculations must also
containment and also (with lockable tops) meet be made to ensure that the inverter will operate
Code requirements to prevent unqualified people properly under all conditions of continuous use
from accessing the electrically energized battery and surges.
terminals [110.27, 706.30]. (See photo 5.11.) Battery banks connected with sets of series-con-
nected cells or batteries and parallel sets of cells
Location, Location, Location or batteries should be wired so that the length of
In small- and medium-sized PV systems, bat- conductors (and the resulting resistance) would be
teries operate at a nominal 12, 24, or 48 volts. the same for each series connection to minimize
To minimize voltage drop in the conductors charging and discharging current and voltage
at typical high operating currents, batteries are imbalances and premature cell or battery failure.
generally installed as close as possible to the loads (See figure 5.4.)
that they serve. In PV systems, these loads are
usually a multimode utility-interactive inverter or Overcurrent Protection
a standalone inverter. In many cases, particularly Batteries can source high fault currents. It is
in standalone systems, a power center is used that somewhat difficult to obtain specific short-circuit
provides a central location for overcurrent protec- current data on batteries and even more difficult
are commonly available and will provide some

protection for downstream components. Di-
rect-current-rated circuit breakers, on the other
hand, have little or no current-limiting capability.
Any disconnecting equipment in the dc battery
conductor path should have sufficient short-cir-
cuit interrupt capabilities where circuit breakers
are used.
Normally, in small residential and small com-
mercial battery banks, the individual strings of
batteries (series-connected cells or batteries)
are not fused. Series and parallel connections
are made within the battery bank and then
the overall output is provided with overcurrent
protection. This would generally indicate that the
conductors within the battery bank be able to
handle the entire current of the battery bank (and
not be sized to handle just a proportional part of
the total battery current).
In parallel-connected sets of batteries or cells,
the individual cells and batteries may not age at
Figure 5.4 • Series-parallel battery connections. Note: the same rate and the currents may tend to be
Current travels through equal lengths of conductors in higher in some of the series connections than in
each of the series paths. Battery conductors should be the
others. For that reason, it is somewhat difficult to
same size and length in each path. All terminations should
be identical. Source: Trojan Battery Co. fuse individual strings of batteries because they
may be required to carry more current as other
to determine what those short-circuit currents strings of batteries age and carry less current.
may be at the output of the battery bank, which Also, installing an overcurrent device inside the
may consist of numerous series and parallel battery bank on a single string or several strings
connected individual batteries. poses both maintenance and potentially explosion
Tests and estimates made several decades hazards.
ago indicate that for typical residential and In larger battery installations (where room-size
small commercial battery banks, the inter-cell/ battery banks are involved), individual strings
inter-battery connectors, the battery conductors, of batteries or cells may have a fused disconnect
the contact resistances, and other factors will typ- on each string. These fused disconnects must be
ically limit the available short-circuit current at located in areas that allow easy access and are not
the output of the battery bank to less than 15,000 subject to hydrogen gas or sulfuric acid fumes.
amps. Therefore, any overcurrent device protect- The size of the conductors for each battery string
ing these conductors should have an interrupt and the overcurrent device rating require careful
rating of at least 15,000 A. Both dc rated fuses consideration because of the potential for un-
and dc rated circuit breakers with that rating are equal currents in the battery series circuits as the
available. battery bank ages.
Direct-current-rated, current-limiting fuses Proper maintenance of the battery banks is the
key to maintaining battery longevity and keeping

the various currents balanced. However, aside
from access requirements (706.33), the NEC does
not address the maintenance of electrical systems.
The batteries are able to generate more
short-circuit current than the typical charging
source and there should be an overcurrent pro-
tection device near the batteries. The overcurrent
device must be sized at least 125 percent of the
continuous current (either charge or discharge,
whichever is greater, including any long-term
surge currents). If the charging currents are great-
er than the discharging currents, an overcurrent
device may be required at the charging source.
A disconnect may be required at the battery
to not only service the fuse (a circuit breaker
includes the disconnect), but also to provide a Photo 5.12 • AC battery system. Courtesy of Enphase
method of disconnecting the battery circuit from
other connected devices. Sections 706.30 and from backfeeding the utility. These devices are
706.31 establish disconnect requirements and cell quite complex (installation, set up and operation
interconnections. manuals may run hundreds of pages), but are
Fuses that bolt directly to the battery terminals usually a single piece of equipment with one or
are generally not acceptable in Code-compliant more external current sensors. The device includes
installations because they are very difficult for a multimode inverter and batteries—usually some
unqualified people to service on the always-en- form of lithium-ion cells. The NEC has little
ergized battery terminals. Also, the fuses may be information on the installation of these listed
in a hydrogen gas/acid fume environment under units and they must be installed and operated
certain circumstances, resulting in OCPD failure per the instructions provided with the listed unit
or an explosion should the fuse open in that [110.3(B)]. These AC batteries are made by at
environment. least three companies with more on the way.
The AC Battery Summary

The latest addition to energy storage systems Plan reviewers and inspectors should check
is the ac battery (photo 5.12). No, this is not a battery installations for substantial mechanical
contradiction in terms. This system was designed installation; correct conductor types and sizes;
to provide a degree of energy backup during proper placement and rating of overcurrent de-
utility outages; to store PV energy for use when vices and disconnects; good conduit installation
utility rates are high; and to prevent PV energy techniques; and proper ventilation.
Chapter 6 Grounding, Disconnects, and Overcurrent Protection 113
114 Chapter 6 Grounding, Disconnects, and Overcurrent Protection
06
Grounding, Disconnects, and
Overcurrent Protection
Grounding Fault Protection Device. In other cases, various

Grounding—New and Old resistances or solid-state devices were used to
Grounding PV systems has remained essentially “ground” one of the circuit conductors.
unchanged for the most part since PV came In recent years, there are increasing numbers
to the Code in 1984. But with the 2017 NEC, of “ungrounded” PV arrays and non-isolated
the terminology that is used to describe various (transformerless) inverters which have no dc
grounding functions has been changed. This was circuit conductors connected directly to ground.
done to reduce confusion and simplify terminol- Grounded PV systems in any but the very
ogy between older “Grounded PV Array/Isolated smallest PV systems are becoming a thing of
Inverter” PV systems and newer, more common the past, but will exist for years [690.41(A)(5)].
“Ungrounded PV Array/Non-isolated Inverter” The Code requirements to address these varying
PV systems. “grounding” systems were cumbersome, difficult
to understand, and hard to apply in many cases.
Grounding the PV System Circuit Con- The changes in the 2017 NEC address and
ductors. In the early years of PV (up to about simplify many of these grounding requirements.
2005), we primarily had grounded PV arrays and However, it is noted that proposals for the 2020
grounded dc battery systems where one of the NEC, if adopted, may further clarify these new
circuit conductors was “grounded” or “connected” requirements.
to the grounding system. This system is com- Some of the grounding definitions in the 2017
posed of the equipment-grounding system, the NEC are the same as in the 2014 NEC.
grounding electrode conductor, and the ground-
ing electrode system. From Article 100
In some cases, the “grounding” method was a “Ground. The earth.
solid conductor, and in other cases, the grounding “Grounded (Grounding). Connected (con-
method was by a fuse or circuit breaker like those necting) to ground or to a conductive body
used in Section 690.5 of the 2014 NEC, Ground that extends the ground connection.
“Grounded, Solidly. Connected to ground It should be noted that there are no voltage
without inserting any resistor or impedance requirements associated with the requirements in
device. 690.41 and that old 50-volt limit is gone. It will
“Grounded Conductor. A system or circuit take some time to figure out what we call these
conductor that is intentionally grounded.” various types of PV arrays. For example, a PV
array connected to a non-isolated inverter with a
From Article 690 240 VAC output will meet the requirements of (3)
“Functional Grounded PV System. A PV because at some point there is a utility transformer
system that has an electrical reference to ground that has a grounded 240/120-volt center tap.
that is not solidly grounded.
Informational Note: A functional grounded Equipment Grounding
PV system is often connected to ground Sections 690.43, 690.45, and 690.46 dealing with
through a fuse, circuit breaker, resistance equipment-grounding conductors have remained
device, non-isolated grounded ac circuit or essentially unchanged.
electronic means that is part of a ground-
fault protection system. Conductors in these Grounding Electrode Systems
systems that are normally at ground potential Section 690.47 on grounding electrode systems
may have voltage to ground during fault has been substantially revised and clarified. The
conditions.” most significant change is for functional ground-
ed pv systems (not solidly grounded). In these
As an aside, my high school English teacher systems, the normal ac equipment-grounding
would have objected to the term functional ground- conductors from the inverter to associated distri-
ed and preferred functionally grounded, but then bution system are allowed to serve as the ground-
she has not seen the rest of the NEC. Functional ing connection for any ground-fault protection
ground would probably have been acceptable— equipment and for the equipment grounding
another potential update for the 2020 NEC. of the PV array. This is a change from previous
690.47(C)(3) in the 2014 NEC that required
“690.41. PV System Grounding Configurations this conductor to meet the requirements of what
One of the following system grounding was called a “grounding conductor” (bonding,
configurations shall be employed: size, and continuity), but was actually a bonding
(1) 2-wire PV arrays with one functional jumper, as well as the requirements for an equip-
grounded conductor. ment-grounding conductor.
(2) Bipolar PV arrays according to 690.7(C)
with a functional ground reference (center tap). Grounding the Array—690.47(D) Grounding
(3) PV arrays not isolated from the ground- Electrode System
ed inverter output circuit. Section 690.47(D) has been revised in every
(4) Ungrounded PV arrays. edition of the Code since it was originally written.
(5) Solidly grounded PV arrays as permitted Here is what is says, in part, in the 2017 NEC:
in 690.41(B) Exception. “(A) Buildings or Structures Supporting a
(6) PV systems that use other methods that PV Array. A building or structure supporting
accomplish equivalent system protection in a PV array shall have a grounding electrode
accordance with 250.4(A) with equipment system installed in accordance with Part III
listed and identified for the use.” of Article 250.”
This grounding electrode system could be the

existing grounding electrode system or a newly
installed one if no grounding system exists.
“PV array equipment grounding conductors

shall be connected to the grounding electrode
system of the building or structure support-
ing the PV array in accordance with Part VII
of Article 250. This connection shall be in
addition to any other equipment grounding
conductor requirements in 690.43(C). The
PV array equipment grounding conductors
shall be sized in accordance with 690.45.”
Section 690.43(C) requires that these PV array

and mounting structure equipment-grounding con-
ductors be routed with the circuit conductors where
Photo 6.1 • Array grounding electrode required. A 2
those conductors leave the vicinity of the PV array. AWG copper conductor attached to a rebar to makes a
690.47(B) permits, but does not require, concrete-encased electrode (sometimes referred to as
additional auxiliary grounding electrodes for UFER)—prior to concrete pour.
ground-mounted and roof-mounted arrays. They
are to be installed in accordance with 250.52 and Disconnects
250.54, which do not require them to be bonded The various sections of the Code dealing with
to other grounding electrodes in the system. Of disconnects in PV systems have been signifi-
course, they would be connected together with cantly revised in the 2017 NEC. As noted in
the other grounding electrodes through the chapter 1; NEC Figure 690.1(B); and as shown
equipment grounding conductors. in Figures 1.22, 1.23, and 1.24, the PV system
Also, any building or structure that supports disconnect can be either ac or dc and is the
a PV array must have a grounding electrode demarcation between the PV system and any
system (photo 6.1)—either an existing one or an other source of energy including the utility
added one. The PV array equipment-grounding grid, energy storage systems, and other energy
conductors will be connected to that grounding sources like fuel cells or wind systems. Section
electrode system, and that grounding electrode 690.13 deals primarily with the PV system
system may be the existing site grounding disconnect, and Section 690.15 establishes
electrode system if one exists. These PV array requirements for the disconnection of PV
equipment-grounding conductors are routed with equipment (photo 6.2).
the dc circuit conductors. Evidently through the use of master keys and
Solidly grounded systems will require a fire axes, fire services consider the interior of a
grounding electrode conductor per Section locked home to meet the readily accessible re-
250.166 connected to the grounding electrode quirement, because about half of the homes in the
system. The grounding electrode conductor or United States have the main ac service disconnect
main bonding jumper will be connected at any located inside the structure. In the future, some
point on the PV output circuit (690.42). consideration should be given to the differences
Photo 6.3 • There is no longer a specific requirement to

have a dc PV disconnect located at or near the point of en-
try of the conductors to a building. Marking requirements
have changed from this 2011 NEC commercial installation.
there is no longer an exception in 690.13 related

Photo 6.2 • Disco Madness. How many disconnects can
you find? At least nine are visible in this 2014 NEC installa-
to that installation requirement that allows nearly
tion, but how many are inside the enclosures? And what any location for the PV system disconnect.
would it look like under the 2017 NEC? An Informational Note to 690.13(A) alludes
to the 690.12 rapid shutdown system as dealing
between an ac service and a dc PV circuit entering with energized conductors entering the build-
a building with respect to disconnect requirements. ing, but the author feels that this is a separate
system/function that is not directly related to
No Longer at the Point of Entry. Section the widespread requirement of having all power
690.13(A) requires that the PV system discon- sources equipped with a disconnect as they enter
nect be in a readily accessible area, and this may a building [230.70(A)(1)]. This issue may or may
be a dc disconnect in some systems or an ac not be clarified by the 2020 NEC.
disconnect in others [See NEC Figure 690.1(B)]. While some AHJs and jurisdictions may con-
Note that there is no longer a specific require- tinue to follow the 2014 NEC guidance in this
ment for a dc PV disconnect to be located inside area, the 2017 NEC apparently allows or accepts
or outside nearest the point of entry of the dc the required dc isolation device/disconnect on the
conductors into the building as required by the dc input of the inverter or the input to a charge
2014 and previous editions of the NEC, such as controller as a suitable dc PV disconnect, no
690.13(A) in the 2014 NEC (photo 6.3). matter where it is located. As long as that loca-
While there is a requirement in 690.15 in the tion is readily accessible, the disconnect can also
2017 NEC for an isolation device within 10 feet serve as the PV system disconnect where it meets
of the inverter, that device may not be located the definition and requirements of that device. In
at the point of entry of the dc conductors if the some cases, the disconnect may be the PV system
inverter is more than 10 feet from that point of disconnect as defined in the 2017 NEC. In other
entry. Of course, the dc conductors inside the cases, it may just be a dc disconnect at the input
building are still required to be in a metal raceway to a utility-interactive inverter.
or installed in Type MC cable [690.31(G)], but Various parts of 690.13 require that the PV
Photo 6.4 • Disconnect labels based on 2014 NEC. 2017 NEC requirements are slightly different.
system disconnect open all conductors [690.13(F)]; ductors and it is noted that, in many cases where a
be rated for the available short-circuit current grounded conductor is involved (usually a neutral)
and voltage in the circuit [690.13(E)]; be rated as in that circuit, at least a three-pole disconnecting
service equipment for ac supply-side PV connec- means would be required. It will be difficult to
tions [705.12(A), 690.13(C)]; and may be a single achieve the ac PV system disconnect using the
disconnect for multiple inverter systems and have backfed circuit breaker in a load center due to that
up to six switches for more complex PV systems neutral switching requirement. On the other hand,
[690.13(D)]. As mentioned previously, the color Section 705.21 requires that only the ungrounded
coding of a solidly grounded conductor that has a conductors be disconnected at this point.
disconnect will need to be clarified in the Code. Proposals have been made to correct the
requirement in 690.13 in the 2020 NEC, but in
To Open the Grounded Conductor or Not? the meantime, it is suggested that the require-
Because the PV system disconnect may be the ments of 705.21 be followed rather than those
connection of the output of a utility-interactive of 690.13(F). A Tentative Interim Amendment
inverter to the utility power source, the require- (TIA) for the 2017 NEC may rectify this issue
ments of disconnects and Article 690 should be before the 2020 NEC is published and adopted.
consistent with those requirements in Article 705.
The requirements in 690.13(F) require that the Isolation Devices—a Subset of
PV system disconnect simultaneously open all Disconnecting Means
conductors of the PV circuit connected to other Section 690.15, Disconnection of Photovoltaic
sources. There is no exception for grounded con- Equipment, requires that isolating devices for
all non-solidly grounded conductors be located than 30 amps, an equipment disconnecting

within 3 m (10 ft) of PV modules, ac PV mod- means must be used to isolate the components.
ules, fuses, dc-to-dc converters, inverters, and The equipment disconnecting means must
charge controllers. open all conductors of a circuit simultaneously
Note the change from previous editions of the and have an interrupt rating sufficient for the
Code. The 2014 NEC in 690.13(F) specifically available short circuit current and voltage at its
excluded the requirement for a PV dc disconnect terminals [690.15(B)]. Section 690.15(D) lists
at either the module or the array. That exclusion the types of devices that are suitable for use as
does not appear in the 2017 NEC, so an isolation equipment-disconnecting means. These devices
device (circuits of 30 amps or less) or a discon- include:
necting device (circuits more than 30 amps) will “(1) A manually operable switch or circuit
be required at the PV modules (and ac PV mod- breaker
ules, etc.) and possibly the PV array. While the (2) A connector meeting the requirements of
PV module (ac PV modules, dc-to-dc converter) 690.33(E)(1)
connectors could be used as module/converter (3) A load break fused pull out switch
isolating devices, in many installations these (4) A remote-controlled circuit breaker that
connectors are neither visible or accessible after is operable locally and opens automatically
the PV array is installed on the roof. But visibility when control power is interrupted.”
or accessibility is not a Code requirement for
these isolation devices. Although a disconnecting Equipment disconnecting means that may have
means on the output of a combiner (where used) both line and load terminals energized when
is no longer specifically required [see 2014 NEC, open shall be marked with the electric shock
690.15(C)], such a disconnect/isolation device as warning statement in 690.13(B).
required by 690.15 might meet the PV module Photo 6.4 shows the front of a PV dc discon-
isolation device requirement at the PV array. nect with the labels required by 690.13(B) and
A disconnecting means can always be used as 690.53 (2014 NEC). The 690.13(B) warning
an isolation device, but for circuits with a maxi- is required because the load terminals of this
mum of 30 amps or less, the isolating device may disconnect are connected to the inverter dc input,
be a connector (690.33) if listed and identified which may be energized for up to five minutes
with a specific piece of equipment; a finger-safe after the disconnect has been opened. The filter
fuse holder; an isolation switch that requires and energy storage capacitors in the inverter will
a tool to open; or an isolating device listed for be discharged after this time. The 690.53 label
the intended operation. These types of isolating with the system dc voltages and currents will
devices shall not be required to have an interrupt allow the AHJ to determine if the correct con-
rating, nor shall they be required to open all ductors, disconnects, and overcurrent devices have
conductors of a circuit simultaneously. They will been installed. Section 690.53 in the 2017 NEC
be able to open the maximum circuit current or requires a revised set of information:
be marked “Do Not Disconnect Under Load” or
“Not for Current Interrupting.” (1) Maximum voltage
It is not clear how the disconnecting means Informational Note to (1): See 690.7 for voltage.
marking requirements of 690.53 will be met
when those disconnecting means are connectors. (2) Maximum circuit current
For circuits with a maximum current greater Informational Note to (2): See 690.8(A) for
tally touched when the door of the disconnect

is open. In the PV dc disconnect/isolator, the
PV source or output circuits should always be
connected to the line-side terminals where a
switched disconnect is used.
The dc input to the inverter is connected to the
load-side terminals and the 690.13(B) warning
label is required as shown in photo 6.4. The dc
input terminals are connected to energy storage
capacitors inside the inverter and may be ener-
gized for up to five minutes. On the ac discon-
nect, the utility connection goes to the protected
line-side terminals at the top.
Marked “Line” and “Load.” To Backfeed or
Not? AHJs have noted that many of the utili-
ty-required ac PV disconnect switches (visible
blade, lockable open) have terminals marked
“Line” and “Load,” but are being backfed because
the utility connection should be made to the
touch-shielded, protected upper “Line” terminals
(photo 6.5). Section 690.13(F)(2) states:
Photo 6.5 • Line terminals at top and protected with
shield; unprotected load terminals at bottom.
“(2) Devices marked “line” and “load.” De-
calculation of maximum circuit current. vices marked with “line” and “load” shall not
be permitted for backfeed or reverse current.”
(3) Maximum rated output current of the charge
controller or dc-to-dc converter (if installed). This section appears to confirm the require-
ment that switches (as devices) so marked should
Power flows in a PV system from the PV array not be backfed (photo 6.6). However, Section
through the dc PV disconnect, the inverter, the ac 705.12(B)(4) provides slightly different guidance:
disconnect, various isolation devices, and finally
to the grid. This power flow sometimes confuses “(4) Suitable for Backfeed. Circuit breakers, if
installers on how to properly connect the dc backfed, shall be suitable for such operation.
and ac disconnects. Note the upper line-side
terminals on the disconnect shown in photo 6.5 Informational Note: Fused disconnects,
are covered by an insulated cover. Also note the unless otherwise marked, are suitable for
switchblades, the fuse holder terminals (if any), backfeeding.”
and the load-side lower terminals are exposed
and easily touched. Discussions with personnel at UL add additional
A general safety rule is that the most dangerous weight to the Informational Note and confirm
circuit should be connected to the protected, line- that there are no issues with backfeeding fused or
side terminals. If this is done, it is less likely that unfused switched disconnects unless they are specif-
energized terminal connections will be acciden- ically marked “Not suitable for back feeding” or the
Photo 6.6 • Note the factory embossed “LINE” marking on

this unfused ac disconnect (safety switch) just below the
word “UTILITY” on the field-installed label. Okay for back-
feeding with the utility connection to the line terminals.
equivalent. The “Line” or “Line and Load” markings

on these switches are used to identify the shielded,
protected terminals where the utility (or other most
likely energized circuit when the switch is opened)
is to be connected. Proposals for the 2020 NEC
should clarify this area.
Photo 6.7. AC and dc connectors on this microinverter are
Utility-Interactive Inverters on the Roof certified/listed as load-break-rated disconnects.
In installations using microinverters on the to microinverters and ac PV modules. To meet

roof, or installations where the string inverters the disconnect/isolation requirements at the
are located on the roof of the building, NEC inverter location, most microinverter companies
705.70 has specific disconnecting requirements have had their ac and dc connectors listed as
related to those types of installations where the load-break-rated disconnects when used in this
inverters are in not in readily accessible areas. application. This information should appear in
While this section is self-explanatory, some the instruction manual for the particular prod-
additional information is needed when it comes uct and is covered in the Code by 690.15(C)(1).
Photo 6.8a • An 8-kW string inverter can be paralleled

with two others on a special combining cable and using a
custom splicing device with no OCPD for each inverter.
Courtesy of HiQ Solar.
(See photo 6.7.)
AC PV Modules and Microinverters

In recent times, ac PV modules and microinvert-
ers are the only types of inverters that are paral-
Photo 6.8 b • Custom splicing device used to par-
leled at their outputs on a common circuit (called allel three 8-kW utility-interactive inverters with no
the trunk cable) without a dedicated overcurrent OCPD for each inverter output. Courtesy of HiQ Solar.
protection device for each inverter. As the power
rating of these devices increases (currently up to method) from sets of microinverters or ac PV
about 700 watts), some more precise definition of modules are run from the roof to an ac inverter
these inverters will be needed in the Code because combining subpanel in a readily accessible area
the power levels of these smaller microinverters where each of the backfed breakers for each of
are approaching the “string” inverter power levels. the sets of microinverter outputs serves as the
There are now three-phase, 480-volt ac string second ac disconnect required by 705.70.
inverters rated at 8 kW that have been listed
for parallel operation of three inverters with no Standalone PV Systems
overcurrent protection for each inverter when In the larger standalone systems that are used
used with a listed combining “cable” and where to power dwellings or other structures, the ac
specially listed ac connectors are used as the output of the inverter is normally fed to a power
output disconnect for each inverter (photos 6.8a distribution panel. This power distribution panel
and 6.8b). should have a main breaker, and that breaker is
Because the ac connectors are listed as load- equivalent to the main ac service disconnect on
break connectors at each microinverter and these utility-connected electrical systems. NEC Article
newer string inverters, there is no requirement for 710 provides limited guidance on the require-
an additional ac disconnect for the trunk cable ments for disconnects and their location. If the
on the roof. In many cases, several trunk cables inverter is in the interior of the house, there
(or their extensions using an appropriate wiring appears to be no specific location for the main
Photo 6.9 • (left) PV rapid shutdown initiator located near

incoming service entrance conductors and readily identifi-
able by the fire service.
Photo 6.10 • (top right) Inverters with attached dc

disconnects. AC disconnects are in separate nearby
panelboards.
ac disconnect. However, AHJs who inspect such fied for clarity and are addressing new equipment
systems sometimes consider requiring an external in each new edition of the NEC. To paraphrase
ac disconnect on the dwelling or at least a direc- the Bard again: We know what disconnects are,
tory on the outside of the building indicating the but know not what or where they may be.
location of the ac disconnect inside the building.
When the inverter is located in a separate struc- PV Circuit Overcurrent
ture or is exterior to the dwelling, then the nor- Protection
mal Code rules for installing an ac disconnect for Section 690.9 establishes the requirements for
the conductors entering the building or structure overcurrent protection associated with the now
would apply [230.70(A)]. In some jurisdictions, redefined PV system circuits, both dc and ac.
the fire service may have established requirements Overcurrent protection requirements for batteries
for disconnects on standalone PV systems. (energy storage systems), stand-alone PV systems,
and dc and ac microgrids are covered in other arti-
The currently evolving changes in the Code cles in the Code and in other chapters in this book.
and in UL standards. NEC 690.12, Rapid Shut- Sometimes overcurrent protection is not need-
down System for PV Systems, will have some ed. In any situation where circuit conductors have
effect on how the PV system can be controlled sufficient ampacity for the sum of all available
from a single point in an emergency situation currents in that circuit under normal or fault
(photo 6.9). conditions, overcurrent protection is not required
Many utility-interactive inverters have an [690.9(A)].
internal dc disconnect or one in an enclosure
that is attached to the inverter. That disconnect Special Consideration and Location. Al-
will serve, in most cases, as the dc isolator for the though circuits are normally protected from
inverter (690.15) (photo 6.10). overcurrents at their source, PV source and
Disconnect requirements with respect to lo- output circuits and interactive inverters differ
cation vary significantly with each particular PV from that general rule. Because several pieces of
installation. Those requirements are being modi- PV equipment, such as PV modules, dc-to-dc
listed for the PV application (photo 6.11). The

overcurrent devices listed for the PV application
will usually have a greater operating temperature
range, from -40°C to 50°C (versus 40°C for a
standard OCPD), and have different time history
curves.
In very rare cases, an overcurrent device in an
assembly may be listed for continuous oper-
Photo 6.11 • Fuse listed for PV applications. ation at 100 percent of its rating [690.9(B)].
converters, and interactive inverters, have cur- These assemblies are vanishingly rare in PV
rent-limited outputs, circuits connected to those systems. This OCPD/assembly 100% com-
devices require special consideration with respect bination rating should not be confused with
to overcurrent protection. These circuits, when the fact that UL Standards require that all
connected to a source that can deliver higher fault overcurrent devices hold at 100 percent rating
currents (e.g., utility source and parallel-connect- and blow/trip within one hour at 135 percent
ed strings of modules) into that circuit than the of rating. These tests are conducted on the
current-limited source, shall be protected with bare (well-ventilated) OCPD device without
an overcurrent protective device at the higher an enclosure. A load center, for example, is an
current source end of the circuit [690.9(A)]. assembly/enclosure and the circuit breakers in-
An Exception to 690.9(A) also exempts the stalled in it are derated by the 0.8 factor (rating
overcurrent device requirement on PV source equals 125 percent maximum current) due to
circuits or dc-to-dc converter outputs when the internal heating provided by the circuit break-
available short-circuit currents from all sources ers themselves and the internal conductors.
do not exceed the maximum overcurrent device
current rating for a specific PV module or dc-to- Undefined Device. The adjustable electronic
dc converter. overcurrent devices addressed in 690.9(B)(3),
Another Exception to 690.9(A) also exempts 690.8(B), and 690.8(B)(3) appear to be adjustable
PV source circuits from overcurrent protection trip circuit breakers, and after setting per 240.6,
where there are no external sources of overcur- they should be applied as any other circuit break-
rents from parallel connected source circuits or er. The author strongly feels that because they
backfeed from inverters (and probably charge have no current-reducing function, other than
controllers). The backfeed capability of inverters the normal trip setting, they should not be used
under array fault conditions should be a speci- to establish conductor ampacity as apparently
fication in the inverter instruction manual. This allowed by 690.8(B) and 690.8(B)(3).
specification may not be clearly stated in the
instruction manual and a change to UL Standard “Circuit conductors shall be sized to carry
1741 may be required to clarify the requirement not less than the larger of 690.8(B)(1) or (B)
and define an appropriate test. (2) or where protected by a listed adjustable
electronic overcurrent protective device in
Ratings. Most overcurrent devices used in accordance 690.9(B)(3), not less than the
PV systems will be rated at 125 percent of the current in 690.8(B)(3).”
maximum circuit current determined in Section
690.8 and where used in dc PV circuits, shall be Section 690.8(B) appears to imply that these
devices can reduce a high current to a lower cur- The NEC assumes that each ungrounded con-
rent and then a smaller conductor associated with ductor is connected to some source of overcur-
the lower current can be used when protected with rents that might potentially damage that conduc-
this type of OCPD. As far as can be determined, tor under fault conditions. This source could be
there are no devices that can provide power circuit a power supply, a utility service, or a battery that
(branch circuit) overcurrent limitation/protection supplies currents in excess of the ampacity rating
in this manner outside of some very sophisticated of the conductor. The NEC, in 240.21, requires
electronic devices. Hopefully, the 2020 NEC will that the conductors be protected at their source
clarify this issue. The 2020 NEC may also define of supply. Photovoltaic modules are current-lim-
how actual current-limited inputs on some devices ited devices, and their worst-case, continuous
and energy management systems will deal with outputs for Code calculations are 1.25 times the
limiting the current from multiple sources con- rated short-circuit current. Therefore, the module
nected to a single device or circuit. Look to Article cannot generate sufficient current to damage the
705 and 750 for changes in this area. conductor attached to it in a short-circuit con-
Optional Overcurrent Protection. Under the dition. An exception to Section 690.9(A) allows
2017 NEC, other than very small solidly ground- conductors and PV modules to be used without
ed PV arrays (690.41), most of the dc circuits OCPDs where there are no sources of external
in PV source and PV output circuits will not be currents that might damage that conductor or PV
connected directly to earth (solidly grounded). In module.
previous editions of the NEC, these ungrounded Additionally, UL Standard 1703 requires that
PV arrays would require an overcurrent device modules must have an external series OCPD
in both (positive and negative) of the unground- if external sources of current can damage the
ed circuit conductors [2014 NEC, 690.9(E)]. internal module conductors. The module can be
Section 690.9(C) in the 2017 NEC, now permits damaged if reverse currents are forced through
the use of an overcurrent device (where overcur- the module (due to an external or internal fault)
rent protection is required) in only one of these that are in excess of the values of the maximum
conductors, not both. If that option is selected, series fuse marked on the label on the back of the
all dc overcurrent devices in other PV source module. Again, if there are no sources of external
and output circuits must be in the same polarity currents that exceed this marked value, then no
conductor in those circuits. OCPD is needed to protect the internal module
wiring.
Fusing of DC PV Module External sources of current (apart from the
Circuits in Utility-Interactive module or series-connected strings of modules)
PV Systems vary from system to system. These currents can
In most electrical systems, the NEC requires originate from modules or series-connected
every ungrounded circuit conductor to be pro- strings of modules that are connected in parallel
tected from overcurrents that might damage to the module of interest, from batteries in the
that conductor. Overcurrent protective devices system, or from utility currents backfeeding
(OCPDs), either fuses or circuit breakers, provide through utility-interactive inverters.
that function. However, some of the smaller utili- In systems with batteries and charge control-
ty-interactive PV systems may not need OCPDs lers, the batteries are a predominate source of
in the dc circuits that are connected to the PV currents and OCPDs will be required on each
modules. module or series-connected string of modules.
Photo 6.13 • DC source circuit combiner with output dis-

connect. Note protective “dead front” shield over exposed
energized surfaces/terminals.
6.12 and 6.13). Inspectors should check for

screw-cover enclosures and warning labels if the
combiner contains circuit breakers or fuses and
has internal, exposed, energized terminals. UL
Standard 1741 may be changed to reflect the re-
quirement for such warning labels and, in the fu-
Photo 6.12 • DC source circuit (string) combiner at top.
ture, may require the combiners to be dead front
Usually, only one OCPD will be required to (no exposed energized surfaces) when opened.
protect all modules connected in a single series Some combiners are presently dead front.
string. A properly sized and located OCPD will In utility-interactive systems, a few inverter
protect not only the conductors, but also the designs may be capable of allowing current
modules from external overcurrents [690.9(E)]. from the utility to flow backwards through the
The ratings of the overcurrent devices in the inverter into faults in the PV array. Systems
combiners must be consistent with the ampacity using these types of inverters would typically
of the conductors connecting the modules in require an OCPD at the inverter dc inputs
the source circuit and not more than the rating or an OCPD on each string of modules or
of maximum series fuse marked on the back of OCPDs in both locations (see Figure 6.1).
the module. The combiner might be viewed as a Many of the smaller utility-interactive
branch-circuit load center connected in reverse inverters (below about 6 kW ) are designed so
acting like a PV source-circuit combining panel. that they cannot backfeed currents from the
The overcurrent devices in this enclosure are utility into array faults. However, there are cur-
located in the proper place in the PV circuits to rently no specific tests in UL 1741 to validate
meet NEC and UL requirements. These combin- the lack of backfeeding from the utility, so a
ers should not be mounted in direct sunlight to manufacturer’s certification should be obtained
minimize the internal temperature rise (photos that the inverter cannot backfeed from the
Forward current
Module Isc =8A 1.25 Isc =10A maximum
Module ...
Isc = 8A
Fuse
Fuses rated at
1.56 Isc=12.5
(use 15A)
... Fuse
Fuse
4 x 1.56
Isc=49.9
(use 50A)
... Fuse Fuse +
To inverter
... Fuse
Line to line –
fault
Reverse currents into fault up to 3 x 1.25 Isc + 50 A=80 amps

if string fuses are not used
Figure 6.1 • Part of a large system. Possible reverse currents from in-
verter and parallel strings into fault in one string. Note: all fuses are in
the positive conductor.
utility into an array dc wiring fault. PV modules or strings of modules can be con-
String fuses are required, and an inverter fuse is nected in parallel and still meet the NEC and
required if the inverter can backfeed into fault in UL requirements (marked on the back of each
dc system. The 2017 NEC permits the use of only module) before an OCPD is needed on each
one fuse in each circuit, but any dc fuses in other module/string of modules? UL marks the mod-
circuits must be in the same polarity conductor. ules based on reverse-current tests. The NEC
requires that the manufacturer’s instructions and
The General Case—For Most Larger labels be followed [110.3(B)]. The intent of the
PV Systems module marking is to protect the conductors
The most common situation occurs in systems internal to the module at the marked level from
where there are multiple strings of modules reverse currents. This is a maximum value for the
connected in parallel. The non-faulted strings OCPD. Lesser values can be used if they meet
may be able to supply sufficient overcurrents the NEC requirement of 1.56 x ISC to protect
(through the parallel connection) to damage the conductor associated with the module or
either the conductors or the modules in the string of modules. In some cases, the value of
faulted strings. A basic question is: How many the module protective overcurrent device is
Module Isc = 8A
Maximum Module Series Fuse 15A
Load/inverter cannot source fault currents
... +
LOAD/
INVERTER
Conductor ampacity = 1.56 Isc = 1.56 x 8 = 12.5 A minimum
Conditions of use may require higher rating. NEC 690.8
Module can generate 1.25 Isc = 1.25 x 8 = 10A worst case
Internal module components can handle up to 15 amps of reverse currents
Figure 6.2 • Single string of PV modules. Where there are no external source of overcurrents that can damage the mod-
ules or the conductors, no overcurrent protection is required. NEC 690.9(A) Exception.
less than 1.56 x ISC. This poses a Code conflict this case, we are assuming that the inverter or the
[110.3(B) vs. 690.8/9] and is an issue for UL or batteries are a potential source of overcurrents.
the listing agency to rectify. The OCPD will have a minimum rating of 1.56
Many installers of 12-, 24-, and 48-volt PV x n x ISC amps. It is sized at this value to allow
systems ignore the module OCPD requirement maximum forward currents from the array to pass
and connect modules/strings in parallel. Can it through without interruption and to keep the
be done and how? David King, when he worked overcurrent device from operating at more than
at Sandia National Laboratories, and the author 80 percent of rating.
have smoked (destroyed) a few modules and Examine a circuit where there are n modules/
determined that the module OCPD requirement strings connected in parallel. Place a ground-fault
is valid, even at low voltages. in one module/string. Examine the sources of
It is easy to see that in a one-string system, an fault current that would affect that module string.
OCPD is needed only when the inverter or bat- Let us ignore current from the faulted module/
tery is a source of overcurrents. No fusing would string itself since the wiring in that string is
be required in a one-string system if there were already sized to carry all forward currents gener-
no battery or inverter that could source overcur- ated in the string.
rents (see Figure 6.2). First, there is the potential back feed cur-
Consider n modules or strings of modules rent from the battery or the inverter in those
connected in parallel. The NEC requires that an systems with these components. It is limited
OCPD be installed in the combined paralleled to the NEC-required OCPD of 1.56 x n x ISC
output of all strings (modules) to protect the between its input and the combined array output.
cable from reverse currents from batteries and This current is added to the current from the
backfeed of ac currents through an inverter. In remaining modules connected in parallel. In this
case, the current is (n-1) x 1.25 x ISC. The 1.25 However, if we try to parallel three of these
is required [representing the maximum current modules/strings, the fault current equation yields
defined in (690.8)] because of daily-expect- a fault current of 29+ amps that exceeds the 20-
ed irradiance values that are greater than the amp limit on the module. The single OCPD is 3
STC-rated ISC. x 1.56 x 3.8 = 17.8 amps (because OCPDs at this
rating are not common, a 20-amp OCPD must
I-fault = 1.56 x n x ISC + (n-1) x 1.25 x ISC be used). The two parallel-connected modules
contribute 2 x 1.25 x 3.8 = 9.5 amps for a total
With a little algebra, the resulting fault current is: potential fault current of 29.5 amps. This is
significantly above the maximum series protective
I-fault = (2.81 x n-1.25) x ISC amps. (Fault fuse of 20 amps.
Current Equation.) In most cases, it is not possible to parallel more
than two modules/strings with a single OCPD
Note that this equation does not account for unless the marked maximum series OCPD is very
rating roundup of the OCPD, so each system large in relation to ISC for the module. Some of the
must be checked with the actual OCPD values. thin-film technologies may be able to do this and
If the module can pass the UL reverse current that will be an installation benefit for them.
test at this I-fault value or greater and be so Questions about driving voltages to produce
marked (the maximum protective series fuse on these currents? The faults can occur anywhere in
the label), then it is possible to parallel n mod- the module/string, so a fault involving a single
ules/strings (pick your n) without a series OCPD cell could be the trouble spot, and driving voltages
for each module/string. over one volt could produce the reverse currents.
For example, a PV module is rated at 60 watts What about currents generated within the
and has a maximum series OCPD requirement faulted module string? In the portion of the
of 20 amps, which is marked on the back of the module/string below the fault (toward the
module. The ISC for this module is 3.8 amps. Here grounded end of the module/string), the currents
are the required calculations and checks for two flow in the forward direction toward the fault
strings in parallel. and may or may not cause problems. As far as the
The paralleled circuit OCPD installed at the contribution to the fault current is concerned, the
output of the two paralleled strings will be 2 x 1.56 contribution only appears in the fault path/arc
x 3.8 = 11.8 amps. Assume a 12-amp OCPD is and does not affect the ampacity of the conduc-
used because the NEC now requires module/string tor. Above the fault (toward the ungrounded
OCPDs in one-amp increments up to 15 amps. end), the currents in that portion of the module/
Fuses are available in these values except there is a string appear to oppose the external fault currents
jump from 10 to 12 and then to 15. This OCPD that are trying to reverse the flow of current, but
could potentially allow 12 amps of fault current to the string is reversed biased, and the external
reach the faulted module/string from backfeed from driving currents are flowing. Because the location
a charge controller/battery or from the utility grid of the fault cannot be controlled ahead of time,
through a utility-interactive inverter. worst-case currents must be assumed.
Another 1.25 x 3.8 = 4.75 amps could come from The increased marking value of 20 amps on the
the parallel-connected module/string for a total of example module allows for two modules/strings
16.75 amps. This is acceptable because this module to be connected in parallel and it does make it
is marked for 20 amps. easier for the installer to use a single OCPD with
1.25 Isc =12A

max fault current
... +
...
Module Isc = 8A
Maximum Module Series Fuse 15A Line to line
Load/Inverter cannot source fault currents fault
Conductors rated at 1.56 Isc = 13A
Figure 6.3 • Two strings in parallel. No source of overcurrents that can damage modules or conductors in faulted string.
No backfeed from inverter. No overcurrent devices required. NEC 690.9(A) EX.
a larger conductor to meet both the NEC-re- currents from being backfed through the inverter
quired conductor protection and the UL-required from the utility to faults in the PV array. This
module protection with one large OCPD instead removes one source of currents in the above
of a two smaller OCPDs plus a larger OCPD. equation. With these products, it is possible to
Conductor ampacity must also be addressed have two (and sometimes more) strings of mod-
if modules are going to be paralleled on a single ules in parallel with no OCPDs in the dc circuits.
OCPD. The conductors for each string must be The inverter manufacturers should be contacted
able, under fault conditions, to carry the current for information in this area. The above equations
from the other parallel strings (modules) plus the can be modified by deleting the combined-cir-
current that may be backfed from the inverter cuit OCPD and then solved to determine both
or battery. In the case with n strings in parallel the requirements for OCPDs and the necessary
and a single OCPD in the combined output, the ampacity of the conductors.
conductor ampacities would be as follows: In this case, the reverse current flowing through
Each of the string conductors would have to the forward fuse (n x 1.56 x ISC) is set equal to
have an ampacity of 1.25 (n-1) x ISC + 1.56 x n zero or removed from the equation. In a system
x ISC. If the equation is factored, the required with n strings of modules connected in parallel,
ampacity becomes A = (2.81 x n-1.25) x ISC. As if one of the n strings develops a fault, the fault
before, OCPD roundup is not considered and the current is now reduced to:
values should be recalculated with actual OCPD
values. The combined output-circuit conductors I fault= (n-1) x 1.25 x ISC. For two strings in
would require an ampacity of 1.56 x n x ISC. parallel, n = 2 and the fault current becomes:
Modern, utility-interactive inverters I fault = 1.25 ISC.

Many utility-interactive inverters on the market
have redundant internal circuitry that prevents The NEC requires that all PV wiring generally be
Photo 6.14 • Retaining kit for a backfed circuit breaker. Required where a circuit breaker is backfed from a standalone in-
verter output.
sized at 1.56 x ISC. The required module series protec- Overcurrent Protection in Related Articles.
tive fuse is nearly always greater than 1.56 x ISC. Article 705 establishes additional requirements for
Therefore, in a system with two strings of mod- overcurrent protection in the ac inverter output
ules connected in parallel, there are no sources circuits. Article 710 does the same for ac circuits in
of fault current that exceed the ampacity of the stand-alone systems. In general, conductors shall
conductors or the requirements for a module be protected from all possible sources of overcur-
protective fuse. No dc string or array fuses would rent (705.30, 705.65, and Article 240).
be needed. NEC Section 690.9(A) Exception If circuit breakers are backfed, they shall be
applies. (See Figure 6.3.) suitable for such operation [705.12(B)(4)]. This
If there are more than two strings of mod- usually indicates that they will not be marked
ules connected in parallel, then the calcula- “Line” and “Load.” Also, if a backfed breaker is
tions outlined above will have to be made to fed by a listed interactive source (PV utility-in-
ensure that (n-1) x 1.25 x ISC is less than the teractive inverter), the breaker does not have
module series protective fuse value. If not, to be mechanically fastened to the panelboard
fuses should be used in each string. Again, [705.12(B)(5)]. However, where a backfed break-
the actual value of available fuses should be er is connected to the output of a standalone
used in the calculations. inverter (a voltage source and not interactive),
breaker, do not require that breaker to be secured.

The interrupt and short-circuit ratings of the
equipment in an interconnected system shall con-
sider the contributions from all sources (705.16).
Section 705.31 requires that an overcurrent
device be located within 3 m (10 ft) of the
connection point of the service conductors in a
supply-side connected PV system [705.12(A)].
This overcurrent device would usually be
associated with the start of the feeder leading
to the output of a utility-interactive inverter.
An exception allows this overcurrent protection
to be located more than 3 m (10 ft) from the
connection point if either a current-limiting
breaker or a cable limiter (photo 6.15) are
installed at the connection point to the service
conductors.
The author notes,that current limiting breakers
have different mechanical and electrical char-
Photo 6.15 • Cable limiters. Attached directly to cables acteristics than cable limiters. Where current
or bolted to busbars. No overload protection is provided.
limiting breakers include a disconnect function
Last-ditch cable protection from bolted line-to-line and
ground faults. and will initiate a feeder to the inverter output
at the breaker location, cable limiters have no
that breaker shall be mechanically fastened to the overcurrent rating (they are sized based on the
panelboard per 408.36(D) [photo 6.14]. conductor size) and cannot provide that discon-
Section 710.15(E) on standalone systems is nect function.
not clear in this area, because it mentions that While both devices may be current limiting,
backfed circuit breakers connected to an inter- the cable limited provides only a last-ditch
connected supply (interactive inverter?) require protection before the cable melts under severe
fastening, but the section does not address the overcurrent conditions. The current-limiting
output of a standalone (not interconnected) circuit breaker provides that protection, possibly
inverter. In Article 705, interconnected sources at a lower current limit, and provides the overload
require or imply the use of interactive sources protection necessary to establish the start of a
and those sources, when connected to a backfed feeder.
Chapter 7 Utility Interconnections 133
134 Chapter 7 Utility Interconnections
07
Utility Interconnections
Connecting the utility-interactive inverter to should be used. In a few cases, the inverter manu-
the utility grid properly is critical to the safe, facturer may specify a higher rated current based
long-term, and reliable operation of the entire on a lower than nominal operating voltage (the
system. The ac output circuit requirements and low end of the anti-islanding range) and this val-
the circuits that carry the inverter current in the ue of rated current should be used. Voltage drop
premises wiring are somewhat complex. However, and conditions of use considerations may require
meeting Code requirements can and should be a larger conductor. Higher rated overcurrent
accomplished to ensure a safe and durable system. devices should not normally be used; even though
Inspectors need to know this material and how they may be appropriate for the conductor being
to apply it because many PV installers are not used, they may not provide required protection
familiar with the details of the requirements. The for internal inverter circuits. The inverter manual
first step is to get the power out of the inverter. will generally specify the maximum allowable ac
output overcurrent device rating and this value
Inverter Output Circuit should be followed [110.3(B)].
In general, the output circuit on an inverter to Even though power and current flow from the
the first overcurrent device should be sized at inverter to the utility, it should be noted that the
125 percent of the rated output current (the utility-end of this circuit is where the currents
maximum current) of the inverter that, in turn, is originate that can harm the conductors when
determined from the specifications or by dividing faults occur. Utility-interactive inverters cannot
the rated power by the nominal ac output voltage generate surge currents and usually shut down
[690.8(A)(3)]. For example, a 2500-watt, 240- under conductor fault conditions. Any overcurrent
volt inverter will have a rated output current of: protection should be located at the utility end
2500 / 240 = 10.4 amps of the inverter ac output circuit and not at the
1.25 x 10.4 = 13.02 inverter end of this circuit [690.9(A)].
A 15-amp circuit breaker is the next larger In load-side connected systems, it is good
standard size and at least 14 AWG conductors practice to install the inverter near the backfed
load center so that the backfed breaker commonly ty- interactive PV inverter to the supply side
used to interconnect the inverter with the utility of a service disconnect is similar to connecting
can also be used as the ac inverter equipment a second service-entrance disconnect to the
disconnect/isolator required by 690.15. This places existing service. Because these PV conductors
the overcurrent device at the utility-supply end of are unprotected like service-entrance conductors,
the circuit and groups the ac disconnect for the many of the rules for service-entrance equipment
inverter with any internal or nearby dc disconnect/ should be considered. Sections 230.2(A)(5) and
isolator. A disconnect/isolator must be located 230.82(6) permit these parallel power production
within 3 m (10 ft) of the inverter if the inverter is systems as additional services. However, they
located more than that distance from the backfed do not meet the definition of a service found in
circuit breaker [690.15(A)]. Article 100 of the NEC.
In general, the “Tap Rules” of Section 240 do
Load-Side and Supply-Side not apply unless specifically mentioned because
Utility Connections they were not developed to address two sources
There are two types of connections allowed by of power in a tap circuit, nor were they developed
the Code for interfacing the output of the utility- to ensure safe operation when one source is an
interactive inverter to the utility power. They are unprotected utility power source.
made on either the supply side [705.12(A)] or Although the PV output is not considered a
the load side [705.12(B)] of the main service dis- service, the unprotected conductors are exposed
connect of a facility or structure. The load side of to the same potential fault currents as the ser-
the main service disconnect is the most common vice-entrance conductors. It is suggested that
connection used for the residential system and this PV connection be as robust as any service
the smaller commercial system under about 10 entrance. Section 230.91 requires that the service
kW. Section 705.12(B) covers the requirements overcurrent device be integral with the service
and it is heavy reading at best. disconnect or located adjacent to it. A circuit
breaker or a fused disconnect would meet these
Supply-Side Utility requirements (photo 7.1). A utility-accessible,
Connections visible break, lockable (open) fused disconnect
Many larger PV systems cannot meet the require- (aka safety switch) used as the new PV service
ments/restrictions for a load-side (of the service dis- disconnect may also meet utility requirements
connect) connection to the premises wiring system for an external PV ac disconnect in areas where
and a supply-side connection must be considered. utilities require such an additional disconnect
(photo 7.2). Section 690.13(C) requires that the
Code Considerations PV system disconnect connected to the supply
The supply-side connection (also known as a ser- side of the service disconnect be listed as suitable
vice-entrance connection) is allowed by the NEC for use as service equipment.
and is addressed in several sections in the Code. Section 230.71 specifies that the service
Section 705.12(A) allows a supply- (utility) disconnecting means for each set of service-en-
side connection as permitted in 230.82(6). Sec- trance conductors shall be a combination of no
tion 230.82(6) indicates that solar photovoltaic more than six switches and sets of circuit breakers
equipment is permitted to be connected to the mounted in a single enclosure or in a group of
supply side of the service disconnect. enclosures. Section 690.13(D) allows up to six
It is evident that the connection of a utili- switches or circuit breakers for each PV system.
Photo 7.1 • Added lugs provide a supply-side connection. Photo 7.2 • Utility-required ac disconnects. They could have
But is it Code legal? The added lugs probably violated the been combined with a combiner into one disconnect.
listing on the meter/main combo.
Common industry practice indicates that an ac cent of the rated output current from the PV in-
disconnect associated with a PV system is not verter(s). But in small systems, a question arises:
counted against the six allowable disconnects How small can it be? Section 230.79 addresses
for the existing utility service. This is a gray area the rating issue for service disconnects. It must
subject to AHJ interpretation. be pointed out that these conductors are not
service conductors, but for all practical purposes,
Location and Directory they are treated as service conductors. Although
Section 230.70(A) establishes the location re- not a Code requirement, many AHJs and PV
quirements for the service disconnect. Sections installers choose to treat these conductors as
705.10 / 690.4 / 690.54 require that a directory service conductors. Some inspectors have looked
be placed at each inverter and service equip- at 230.79(A) and say that it can be as low as 15
ment location showing the location of all power amps if that value is at or above the rating of
sources for a building and the locations of all PV the inverter output circuit. The connection of
system disconnects (photo 7.3). Locating the other specific loads is allowed at this level and is
PV ac disconnect adjacent to or near the existing common.
service disconnect may facilitate the installation, I would suggest caution here, because the
inspection, and operation of the system (photo connection is to service-entrance conductors
7.4). Many utilities require the service disconnect rated at 100 amps and above. The typical 15-amp
and the PV ac disconnect to be co-located. circuit breaker with 10,000 amps of interrupt
capability, in this application, may not be able to
Size Matters withstand the available fault current because it
Obviously, the size of the new PV system is not protected and coordinated with any main
disconnect on the ac inverter output circuit is breaker. Of course, Section 110.9 should be fol-
important. It will normally be sized at 125 per- lowed and available fault current calculated. Also,
Photo 7.3 • Directory and labels required by NEC and Photo 7.4 • PV ac disconnect above closed and locked-out/
utility. tagged-out service disconnect
a service-entrance-rated 30-amp fused discon- requirement for the smallest PV ac disconnect

nect with 15-amp fuses could be used; listed fuse for PV inverter supply-side connections. Section
adapters may be required [see UL Product Spec 230.79(D) requires that the disconnect have a
(formally UL White Book) Category IZZR]. minimum rating of 60 amps. This would apply to
Other considerations are the size of the a service-entrance rated circuit breaker or fused
service-entrance conductors, the new PV connec- disconnect. Again, this is not a Code requirement,
tion conductors, and the size of the terminals on but the suggestion of this text and the author.
available switchgear rated at 30 amps or 60 amps. Section 230.42 requires that the service-entrance
The added conductors between the existing ser- conductors be sized for the sum of the non-con-
vice-entrance conductors and the new PV system tinuous loads plus 125 percent of the continuous
ac disconnect will be subjected to available fault loads. All currents in a PV system are worst-case
currents and will have no overcurrent protection and are considered to be continuous. The actual
except that provided by the fuse on the primary rating should be based on 125 percent of the
of the utility distribution transformer. Making rated output current for the utility-interactive PV
them as large as possible, with an upper limit of inverter as required by 690.8. The service connec-
the size of the existing service-entrance conduc- tion conductors, in this case, must have a 60-amp
tors, would seem prudent, but small disconnects minimum rating from 230.42(B). Temperature and
will not accept very large conductors. Cable conduit fill factors must be applied.
adapters may be required to make the connection For a small PV system, say a 2500-watt,
between large service-entrance conductors and 240-volt inverter requiring a 15-amp circuit and
smaller terminals on lower current overcurrent overcurrent protection, these requirements would
devices or disconnects (photo 7.8). appear to require a minimum 60-amp rated
Although these circuits do not meet the disconnect with 15-amp fuses; listed fuse adapt-
definition of service-entrance circuits, the author ers would be required. Fifteen-amp conductors
suggests that Section 230.79(D) be used as the could be used between the inverter and the 15-
conductors be as short as possible with

the new PV service disconnect mounted
adjacent to the connection point. Section
705.31 allows up to 3 m (10 ft) of un-
protected conductors between the point
of connection to the service conductors
and the first overcurrent protective
device.
Making these PV feeder connection
conductors as large as the service-en-
trance conductors, while not a Code
requirement, would also add a degree of
Photo 7.5 • Meter-main combo. Supply-side connections made by safety. Of course, the added disconnect
connecting to internal conductors are not allowed. Listing on the
combo would be violated.
must be able to accept the larger conduc-
tors. Conductors installed in rigid metal
amp fuses in the disconnect. Section 230.42(B) conduit would provide the highest level of
requires that the conductors between the service fault protection.
tap and the disconnect be rated not less than the Neutral-to-ground bonding in the ac PV system
rating of the disconnect; in this case, 60 amps. disconnect is a gray area. Such a bonding may
How we would deal with the 60-amp disconnect, create parallel paths for neutral currents with
15-amp overcurrent requirements using circuit two adjacent neutral-to-ground bonds and two
breakers is not as straightforward. A circuit breaker grounding-electrode conductors to a common
rated at 60 amps would serve as a disconnect, and grounding electrode, but illustrations in the NEC
it could be connected in series with a 15-amp Handbook in Article 250 indicate that these paral-
circuit breaker to meet the inverter overcurrent lel paths do not cause currents that are objection-
device requirements. In this case, the requirements able. Utility requirements and the location of the
of 705.12(B)(2) should be applied for the series PV production or renewable energy credit (REC)
connection between the two circuit breakers. meter should be considered.
Sections 110.9, 690.13(F), and 705.16 require The actual location of the supply-side connec-
that the interrupt capability of the equipment be tion will depend on the configuration and location
equal to the available fault current. The interrupt of the existing service-entrance equipment. The
rating of the new disconnect/overcurrent device following connection locations have been used on
should at least equal the interrupt rating of the various systems throughout the country.
existing service equipment. The utility service On smaller residential and commercial systems,
should be investigated to ensure that the avail- there is sometimes room in the main load center
able fault currents have not been increased above to connect to the service conductors just before
the rating of the existing equipment. Fused they are connected to the existing service discon-
disconnects with RK-5 fuses are available with nect. In other installations, the meter socket has
interrupt ratings up to 200,000 amps. lugs that are listed for two conductors per lug. Of
Section 230.43 allows a number of different course, adding a new pull box between the meter
service-entrance wiring systems. However, con- socket and the service disconnect is always an
sidering that the PV connection conductors are option.
unprotected from faults, it is suggested that the Combined meter/service disconnects/load
Photo 7.6 • General Electric, 125-amp main-lug-only service-entrance-rated load center. With appropriate breakers, the
panel has a 22,000-amp rating. Two stabs can be used to convert the panel to a backfed main breaker panel—but then
supply-side connections cannot be made with remaining breaker positions.
centers frequently have significant amounts of socket, and the old socket covered.
interior space where the connection appears to In larger commercial installations, the main
be possible between the meter socket and the service-entrance equipment will frequently have
service disconnect. However, tapping this inter- busbars that have provisions for tap conductors.
nal conductor or busbar in a listed device, such Any holes intended for connections must be
as a meter-main combination (aka meter-main marked “Tap Locations” or similar. The PV con-
combo), would violate the listing on the device nection to this tap point can only be made with
and should not be done (photo 7.5). the approval and instructions from the manu-
Where the service-entrance conductors are facturer of the equipment or by the organization
accessible, a new meter base (socket) could supplying the service equipment (usually a UL
be added ahead of the combination device. A 508 Industrial Controls Shop). These organizations
connection box would then be added between the can tap the equipment and maintain the listing on
new socket and the combination device. The me- the equipment.
ter would then be moved from the combo device In all cases, safe working practices dictate that
to the new socket, jumper bars added to the old the utility service be de-energized before any
Photo 7.8 • Cable adapters allow larger cables to be con-

nected to terminals designed for smaller cable sizes (re-
ducer). They might also be used to terminate fine-strand-
ed conductors where appropriately rated (shoo-pin).
Courtesy of Greaves.
circuit serving as an equipment disconnect/isola-

tor (690.15) will force the conductor and busbar
calculations into 705.12(B) territory.
Photo 7.7 • Supply-side PV connection to output terminals Section 705.12(A) remains unchanged in the
of meter socket using conductors that are too small. Double 2017 NEC; however, a few changes were made
lugging is also a Code violation.
in other sections in Article 705 that relate to
connections are made. Additional requirements supply-side connections. A basic requirement in
in Article 230 and other articles of the NEC may 705.12(A) is:
apply to this PV feeder connection.
In some locations, a six-circuit breaker main-lug- “The sum of the ratings of all overcurrent de-
only panel is used as a service-entrance panel. If vices connected to power production sources
one of those six circuit breaker positions is unused, shall not exceed the rating of the service.”
it may serve as a supply-side connection for a PV
system (photo 7.6). The size of the PV inverter Note that there is no specific, clearly worded
output routed through this circuit breaker could be prohibition from having both supply-side and
as large as the service-entrance rating or the busbar load-side connections on the same PV system
rating in the load center, whichever is smaller. if the requirements for each are met. However,
caution should be exercised where the sum of the
Other Considerations supply-side PV output OCPD and any load-
Although 705.12(A) governs the supply-side side PV output OCPD exceed the rating of the
connection, as soon as the circuit leaves this new service. In the absence of loads on the system, PV
PV service-entrance disconnect, the requirements currents could exceed the service rating.
of 705.12(B) apply if there are any overcurrent As an example: A 400-amp service has a
devices between the service-entrance overcurrent 400-amp main panel with a 400-amp main
device and any inverter in the system. Photo- breaker. An 80-amp load-side PV connection is
voltaic inverter ac combining panels, or load made, and a 350-amp supply-side PV connec-
centers needed for multiple inverter installations, tion is made. Normally loads on the 400-amp
and even a circuit breaker in the inverter output panel would absorb some part of the 80-amp
PV output. But, if the loads were not present

or turned off, then the sum of the PV currents
flowing into the service conductors from both
PV systems could total 430 amps, exceeding the
service rating.
Sections 705.16, 705.20, and 705.21 establish that
all sources and all equipment shall have disconnect-
ing means and that fault currents from all sources
shall be considered when determining the required
interrupt ratings of those disconnecting means.
Ground Fault Protected Circuits. Section
705.32 requires a PV system supply-side con-
nection where ground fault protection is required
for the loads. This would typically occur where a
ground fault equipped main circuit breaker was
installed as required on solidly grounded wye
services at 1000 amps or greater and 150-1000
volts (230.95). An exception allows a load-side
connection where the loads are protected from
ground fault currents from all sources and that
would include ground fault currents sourced by Photo 7.9 • Splicing blocks can be used to tap services
with multiple conductors per phase.
the PV system.
Technical considerations relating to the suit- current device size to the more robust disconnect
ability of GFP main breakers for backfeeding, the fuse holders. This action would probably ensure
selection of an ac ground fault device for a PV that the disconnect was service entrance rated
inverter output, and the settings of two interact- and would also minimize (to nearly zero) any
ing ground fault devices generally indicate that additional unprotected conductors. It would also
a PV system should be connected only to the meet 705.31 requirements.
supply side of a GFP main breaker. Available Fault Currents. Utilities in various
One Possible Solution. To avoid the issues parts of the country are making efficiency im-
associated with those shown in photo 7.7, where provements in their distribution systems. They are
space permits, it could be possible to open the installing more efficient distribution transformers
service conduit or cable, insert the ac PV fused in newer subdivisions, and those transformers
disconnect in the conduit run and then make the have lower impedances and lower losses. The
appropriate connections inside the disconnect result is that the available short-circuit current at
provided there is sufficient room for such connec- the transformer secondary will exceed the com-
tions and the splicing devices. Using a disconnect monly used 10,000-amp number, and the service
with the same rating as the service conductors entrance equipment at the building will need to
might facilitate the connections (proper terminal be rated for the higher available fault currents.
size) as would special multi-conductor terminals The AHJ should require that the PV installer
(where available) on the line side of the discon- get the specification for the available fault current
nect. Fuse size adapters could be used to match at the transformer in writing from the utility and
the possibly smaller required PV ac output over- provide the calculations for the available fault
connection, all conductors of each phase and the

neutral conductor must be “tapped” to ensure
that all currents remain balanced [300.3(B)(1),
310.10(H)]. Splicing blocks and other devices
are available to facilitate these connections, but
they do require space in the enclosures (photo
7.9).
Supply-side service-entrance connections are
useful for larger PV systems where the condi-
tions of the load-side connection cannot be met.
These supply-side connections normally require
that the power be removed from the service to
ensure a safe installation. After the first over-
current device at the PV service disconnect is
passed, load-side requirements may be imposed,
depending on the complexity of the system.
Load-Side Utility Connections

Code-making panels since 1984 have maintained
that 705.12(B) and the prior 690.64(B)(2) will
Photo 7.10 • There must be an open space somewhere for be rigorously applied to any circuits supplied
a backfed breaker for a load-side PV connection. from multiple sources where the circuit is pro-
tected by overcurrent protective devices (OCPD)
current at the service entrance equipment. This from each source (photo 7.10). Such sources
calculation will also apply to the first disconnect/ would include the output of PV inverter(s)
overcurrent protection used on the supply-side and the utility supply, and those circuits would
PV connection circuit. In these cases, equipment include all existing and added premises wiring
rated at 22,000 amps or higher may be required. that could be subjected to currents from either or
It should be noted that there have been a few both sources.
cases where the distribution system has been
upgraded and the existing service entrance equip- Introductory Paragraph 705.12(B)
ment is now underrated. “705.12(B) Load Side. The output of an
To provide high levels of protection during interconnected electric power source shall be
faults or line surges (where surge protectors are permitted to be connected to the load side of
used) for the PV inverter, it is suggested that the service disconnecting means of the other
fast-acting current-limiting fuses (e.g., type source(s) at any distribution equipment on
RK-1) be used as the first overcurrent protection the premises. Where distribution equipment,
device on the PV ac circuit nearest the connec- including switchgear, switchboards, or pan-
tion point. elboards, is fed simultaneously by a primary
Multiconductor Connections. In some cases, source(s) of electricity and one or more
each phase of the service-entrance conductors other power sources, and where this distri-
and the neutral will consist of two or more bution equipment is capable of supplying
paralleled conductors. When making the PV multiple branch circuits or feeders, or both,
the interconnecting provisions for the other 705.12(B)(2)(1) Feeder Ampacity

power sources) shall comply with 705.12(B) “(1) Feeders. Where the power source output
(1) through (B)(5).” connection is made to a feeder at a location
other than the opposite end of the feeder
As in past editions of the Code, “distribution from the primary source overcurrent device,
equipment” is not specifically defined. Of course, that portion of the feeder on the load side of
it is usually understood what the term encom- the power source output connection shall be
passes. But, it also must be understood that protected by one of the following:
distribution equipment in the form of junction (a) The feeder ampacity shall be not less than
boxes for taps or new panelboards can be added the sum of the primary source overcurrent
to the premises wiring at almost any location that device and 125 percent of the power source
is allowed by the code. So essentially connections output circuit current.
for utility-interactive inverters can be made at (b) An overcurrent device on the load side of
many points on the load-side circuits that are not the power source connection shall be rated
in existing distribution equipment. not greater than the ampacity of the feeder.”
705.12(B)(1) Dedicated Overcurrent and In 705.12(B)(2)(1)(a), a PV connection is made

Disconnect to the feeder somewhere along the feeder, but not
“Each source interconnection of one or more at the end opposite the utility connection. The
power sources installed in one system shall be portion of the feeder, from the connection point
made at a dedicated circuit breaker or fusible to the load end of the feeder, can be subjected to
disconnecting means.” currents that are additive and can be as high as
the rating of the existing utility end overcurrent
This section has been clarified. Multiple PV device protecting the feeder plus the output of
systems may be connected to an existing utili- the PV inverter. Hence, the conductor from the
ty-supplied electrical system, but each connection connection point to the load end of the feeder
must be through a dedicated overcurrent protec- must have an increased ampacity equal to the
tive device and a disconnect. “Dedicated” means sum of the existing overcurrent device protecting
that no loads are to be placed on this circuit be- the feeder and 125% of the inverter output rating
tween the breaker and the PV system ac output. as noted in this section. (See Figure 7.1.)
In 705.12(B)(2)(1)(b), an allowance is made
705.12 (B)(2) Bus or Conductor Ampere Rating to protect the existing feeder by installing an
“One hundred twenty-five percent of the inverter overcurrent device on the feeder on the load
output circuit current shall be used in ampacity side of the connection point at the connection
calculations for the following:” point. This allowance, with the added overcurrent
Note that the title of the section remains bus protection rated the same as the existing feeder,
or conductor ratings and will apply to both as will allow the existing feeder to be retained and
they are defined in the introductory paragraph. not be replaced as may be required in 705.12(B)
This 125% of rated output allowance may slightly (2)(1)(a). The addition of this overcurrent device
reduce the required busbar and conductor ratings will prevent excess load currents or faults from
required by the following calculations compared exceeding the ampacity of the feeder. The added
to the previous requirement to use the inverter overcurrent device should be placed adjacent to
output circuit breaker rating. the power source connection on the load side
Figure 7.1 • 705.12(B)(2)(1)(a) Increased feeder (on load side of PV connection) ampacity required.
of that connection to protect the remainder of in the new requirements, we can assume (some-
the existing feeder from overcurrents and faults times not a good thing to do) that there is no
(240.21). Placing the added overcurrent device ampacity correction required on the feeder under
at the load end of the existing feeder does not that situation. The size of the existing feeder was
provide that protection. See figure 7.2. determined by the existing overcurrent device
This section represents a significant change from protecting that circuit from utility currents.
past code requirements. It presents requirements Additionally, while locating the power source
for feeder size and overcurrent protection when output connection at the opposite end of the
the utility interactive inverter connection is not feeder from the utility source will prevent the
at the opposite end of the feeder from the utility feeder from being overloaded by additive cur-
connection. It should be noted that the ampacity rents, it is obvious that 125% of the rated power
of the feeder between the main OCPD from the source output current must not exceed either the
utility and the inverter point of connection and rating of the utility-end overcurrent device or the
the conductor from the point of connection to ampacity of the existing feeder.
the inverter output OCPD are not addressed. The The general requirement should be (proposed
following will discuss that omission. for the 2020 NEC): The ampacity of a circuit that
has sources protected by overcurrent devices at
PV Opposite Utility on the Feeder (Not each end shall be no less than the rating of the
addressed by Code—Update Coming in the largest overcurrent device. This requirement can
2020 NEC) be used to size feeders or size the rating of the
Since the situation where the PV connection is inverter ac output OCPD.
at the opposite end of the feeder is not addressed In all cases, the largest OCPD in this situation
Figure 7.2 • 705.12(B)(2)(1)(b) Additional breaker (on the load side of the PV connection at that connection) required to
protect existing feeder from excess currents from PV plus utility.
should be the main OCPD protecting the feeder. power source output current, in general any tap
Also, where the feeder is not “tapped” at the end conductors may have to be increased on exist-
point, the conductor from the inverter connection ing taps and calculated to be a larger size than
point to the inverter output OCPD should also meet normal on new taps.
this rule and be sized the same as the existing feeder.
Busbars
Inverter Output Circuit (the tap conductor) Size “(3) Busbars. One of the methods that follows
“(2) Taps. In systems where power sources shall be used to determine the ratings of busbars
output connections are made at feeders, any in panelboards.
taps shall be sized based on the sum of 125
percent of the power source(s) output circuit Busbar Rule (a)
current and the rating of the overcurrent (a) The sum of 125 percent of the power
device protecting the feeder conductors as source(s) output circuit current and the rating
calculated in 240.21(B).” of the overcurrent device protecting the
busbar shall not exceed the ampacity of the
This section applies to both existing load taps busbar.
on a feeder and any taps added after a power Informational Note: This general rule assumes
source connection has been made to that feeder. no limitation in the number of the loads or
All taps, both old and new, will be subjected to sources applied to busbars or their locations.”
this requirement and because the currents used in
the various tap rule calculations will be increased This worst-case requirement presented in
due to adding the main OCPD to 125% of the 705.12(B)(2)(3)(a) assumes that the utility cur-
rent through the existing main breaker and the contains loads, the sum of 125 percent of the
current from the output of the utility-interactive power sources(s) output circuit current and
inverter may add and that current may create an the rating of the overcurrent device protecting
overload on the busbar. There are no restrictions the busbar shall not exceed 120 percent of the
on the location of the main utility breaker or the ampacity of the busbar. The busbar shall be
PV backfed breaker. If the busbar has a rating sized for the loads connected in accordance
equal to the sum of these two values, then no with Article 220. A permanent warning label
overload would be possible. shall be applied to the distribution equipment
It should be noted that reductions in the size adjacent to the back-fed breaker from the
of the utility breaker are not prohibited in this power source that displays the following or
section and could be accomplished if allowed by equivalent wording:
other sections of the Code, load calculations and
equipment limitations. WARNING:
POWER SOURCE OUTPUT CONNECTION–
Busbar Rule (b) DO NOT RELOCATE THIS OVERCURRENT DEVICE.
“(b) Where two sources, one a primary power
source and the other another power source, The warning sign(s) or label (s) shall comply
are located at opposite ends of a busbar that with 110.21(B).”
705.12(B)(2)(3)(b) is similar to the require-

ment found in previous editions of the Code. If
the two sources (utility and PV) are at opposite
ends of the busbar, then the sum of those two
sources can be as high as 120% of the busbar
rating (photo 7.11). With this location of sources,
it is not possible to overload the busbar. Note
that the busbar must be sized for the loads that
are connected. The reason for the value of 120%
is lost in history, but may be related to potential
thermal overloading of the panelboard. The
warning label is self-explanatory and the new
requirement referring to section 110.21(B) gives
additional information on the specifics of the
appearance and durability of the warning label.
Busbar Rule (c)

“(c) The sum of the ampere ratings of all
overcurrent devices on panelboards, both load
and supply devices, excluding the rating of
the overcurrent device protecting the busbar,
Photo 7.11 • Oops, the electrician got it wrong. The three
backfed PV breakers should have been mounted at the
shall not exceed the ampacity of the busbar.
top of the panel as far as possible from the main breaker The rating of the overcurrent device protect-
located at the bottom. ing the busbar shall not exceed the rating of
the busbar. Permanent warning labels shall the busbar could not be drawn through the
be applied to distribution equipment that load breakers. And again, under this condition,
displays the following or equivalent wording: no PV backfed breaker could be added. How-
ever, as the sum of the load breakers is reduced,
WARNING: there becomes an allowance for adding a back-
THIS EQUIPMENT FED BY MULTIPLE SOURCES. fed PV breaker with increasing ratings. In the
TOTAL RATING OF ALL OVERCURRENT DEVICES, extreme case, there could be a situation where
EXCLUDING MAIN SUPPLY OVERCURRENT DEVICE, there are no load breakers and only a single or
SHALL NOT EXCEED AMPACITY OF BUSBAR. multiple backfed PV breakers rated, in total the
same as the busbar. In any of these cases, no
The warning sign(s) or label(s) shall comply with matter where the PV breaker is installed on the
110.21(B).” busbar, the supply and/or load currents cannot
exceed the rating of the busbar.
705.12(B)(2)(3)(c) provides an alternate But, it should be noted that in existing load
method of sizing the PV backfed breaker, or centers, with the sum of the load breakers
determining the size of the required busbar if the totaling more than the busbar rating, it is
PV backfed breaker rating is known. This section unlikely that load and load breakers can or will
will most likely be used when connecting a PV be removed.
inverter output to a lightly loaded subpanel or It would not be wise to install a backfed PV
when sizing inverter ac combining panelboards. breaker that was larger than the main breaker
After excluding the main breaker from the in those instances where the busbar rating is
utility, the sum of all remaining breakers, both larger than the main breaker. If this were done,
load breakers and the PV supply breaker may the main breaker could trip from overcurrents
not exceed the rating of the busbar. There are through the larger PV breaker.
several aspects to this requirement that need close However, this section needs to be used with
inspection and consideration. caution because there is no restriction on the
First, the main breaker before the addition of position of the backfed PV breaker. Suppose a
any PV has been sized to protect the busbar from 50-amp PV breaker were installed near the top
possible overload from utility currents. The main of the 100-amp busbar in the load center near a
breaker will always be equal to or smaller than 100-amp main breaker and there were 50 amps
the busbar rating. For example, many load centers of load breakers. The Code requirement is met
have a 125-amp busbar, but only a 100-amp main with this configuration. However, what happens
breaker. In a normal panelboard or load center, if someone disregards the warning label or the la-
the ratings of the load breakers will total more bel simply falls off over time? Added load break-
than the rating of the main breaker or the busbar ers in the empty positions could pose an overload
in nearly all circumstances. If this situation exists, situation on the busbar. I suspect that many
then no PV can be added because the require- jurisdictions are going to have to emphasize the
ment cannot be met because the sum of the load permanent nature of that warning label to cover
breakers already exceeds the rating of the busbar. the materials that it is made of and the manner
However, if the sum of the load breakers were in which it is fixed to the panel board. Also, some
equal to the rating of the busbar, that busbar consideration might be made to permanently
would still be protected both by the main covering unused panelboard breaker positions.
breaker and by the fact that excess current over It might be wise to adopt a local jurisdiction
requirement that the backfed PV breaker always and the possibility of installing additional break-
be installed as far as possible from the main ers and loads in unused spaces in the panelboard.
utility breaker and an additional warning label Fault studies may involve looking at the electrical
as required in (b) be placed adjacent to this PV time versus current profiles for each of the circuit
breaker, or other PV overcurrent device. breakers involved to ensure that all portions of
the busbars will be protected under various fault
Center-fed Panelboards and Multiple-ampac- scenarios from currents sourced both from the
ity Busbars utility through the main breaker and from the PV
“(d) A connection at either end, but not both system through the backfed PV breaker.
ends, of a centerfed panelboard in dwellings
shall be permitted where the sum of 125 Marking (3), Suitable for Backfeed (4), and
percent of the power source(s) output circuit Fastening (5)
current and the rating of the overcurrent These sections are largely unchanged from the 2014
device protecting the busbar does not exceed NEC and are self-explanatory.
120 percent of the current rating of the
busbar.” Examples
“(e) Connections shall be permitted on 1. A dwelling has a 125-amp rated service
multiple-ampacity busbars where designed panel (busbar rating) with a 100-amp
under engineering supervision that includes main breaker at the top. How large can the
available fault current and busbar load calcu- backfed 125% of inverter output current
lations.” be assuming that it can be located at the
bottom of the panel? Circuit breaker round
There was no provision in earlier codes to up is allowed.
address center-fed panelboards or multiple-am-
pacity busbars and it was not possible to install PV OCPD + main OCPD <= 120% of
the PV breaker at the opposite end of the busbar panel rating
from the main breaker because there were two
or more busbars connected to the main breaker. 120% of panel rating = 1.2 x 125 = 150 amps
705.12(B)(2)(3)(d) was specifically added to
the 2017 Code to address the common situation PV + 100 <= 150, therefore the 125% of
where PV needs to be connected to a center-fed inverter output current can be up to 50
panelboard. Although not clearly stated, there amps
was no intent to allow center-fed panelboards
to be installed under sections (a) through (c) 2. Suppose it was 100-amp panel with a
of 705.12(B)(2)(3). PV connections are now 100-amp main breaker, how much 125% of
allowed on center-fed panelboards and mul- inverter output current could be added?
tiple-ampacity busbars under the conditions
noted in these sections. Engineering supervision PV + 100 <= 1.2 x 100 = 120
typically indicates that the analysis of the PV
connection will be made and stamped by a pro- The maximum PV backfed 125% of invert-
fessional engineer. The load calculations will look er output current would be 20 amps.
not only at the breakers installed on the busbars,
but also the loads connected to those breakers, 3. A 200-amp main panel with a 200-amp
main breaker would allow up to 40 amps OCPD cannot be located at the bottom of the
of 125% of inverter output current, which panel or at the end of the circuit, it is not possible
could be any combination of 125% of to install the backfed breaker without changing
inverter output currents that added up to something. That 120% allowance drops to only
40 amps on either line 1 or line 2 of the 100%. Any panel that has a main breaker rated
120/240-V panel. the same as the panel rating in the above equa-
tions would not allow any OCPD to be added.
PV + 200 <= 1.2 x 200 = 240 The 100%-of-the-panel-rating factor (instead of
PV <= 240-200 = 40 amps 120%) would equal the rating of the main break-
er, and the equation would force the PV breaker
4. When working the problem from the rating to be zero. Of course, in some unusual
inverter end, we start with the continuous situations, the other allowances in 705.12(B)(2)
rated inverter output current. This is usually might be used.
the rated power divided by the nominal line In a few cases, an NEC Chapter 2 load
voltage, unless the inverter specifications list a analysis might reveal that the service for the
higher continuous output current (sometimes dwelling needed to be only 150 amps, but a
given at a low-line voltage). 200-amp panel was installed with a 200-amp
main breaker just to provide extra circuit
A 3500-watt, 240-volt inverter has a rated breaker positions. In this case, a 150-amp main
ac output current of 3500/240 = 14.58 breaker could be substituted for the 200-amp
amps. breaker if the panel is listed for interchange-
able main breakers. Even without the bottom
The output circuit must be sized a 125% of position being open, 50 amps of PV breaker
14.58 = 18.2 amps [690.8(A)(3) and (B)(1)]. The could be installed.
next larger overcurrent device would be a 20-amp
OCPD and this would be consistent with the use Systems with multiple inverters
of 12 AWG conductors if there were not any very Many residential and small commercial systems
large corrections applied for conditions of use or use more than one inverter (photo 7.12). If the
voltage drop. This system could be connected to local utility requires an accessible, visible-blade,
a 200-amp panel or a 100-amp panel providing lockable disconnect on the ac output of the PV
the backfed 20-amp breaker could be located at system, then more than one inverter could not be
the bottom of the panel (utility input/breaker at connected directly to the main panel. The two, or
the top). more, inverters would have to have their outputs
There is sometimes a tendency to use that 30- combined in a PV ac inverter combining subpan-
amp breaker and those 10 AWG conductors that el (PV ac subpanel) before being routed through
happen to be on the truck. While this would pose the utility disconnect, where required, and then
no problems for conductor ampacity or circuit to the main panel. The utility disconnect (which
protection, the inverter specifications may limit may now be the PV System disconnect) is not
the maximum size of the output OCPD and normally fused, but some are, depending on the
larger values may not be used [110.3(B)]. system configuration (sometimes fused in a sup-
ply-side connection). The PV ac subpanel rating,
No bottom breaker position? the rating of the disconnect, and the ampacity of
From the above equations, if the backfed PV the conductor to the main panel are also con-
Photo 7.12 • Three inverters with PV ac combining panel to the left. The combining panel is required where the utility
will allow only one ac PV disconnect or one REC meter (not shown).
trolled by 705.12(B) requirements and 705.12(B) 4200/240 = 17.5 amps

(2)(3)(c) might be used. 1.25 x 17.5 = 21.875 amps; use a 25-amp
breaker and 10 AWG conductors.
Here is another example.
The dwelling has a 200-amp main service panel The 20- and 25-amp breakers are mounted
with a 200-amp main breaker and there is an in the bottom of a main-lug-only panel used
empty breaker position (2-poles) at the bottom as a PV ac subpanel. Normally, no loads will be
of the panel. The utility requires an external connected to this subpanel. It will be dedicated to
disconnect switch and it is desired to install a the PV system. If no loads are connected to this
PV system that has a 3500-watt and a 4200-watt ac inverter combining panel, the 705.12(B)(2)
inverter. A PV ac panel will be used to combine (3)(c) would allow the breakers to be positioned
the outputs of the two inverters and the output anywhere on the busbar. The panel/busbar would
of that PV ac panel will be routed through the be rated at least at 40 amps (18.2 + 21.875 =
utility disconnect and then to a single backfed 40.075) but normally, a 60–100-amp subpanel
breaker in the main service panel. would be used.
The ratings of the output circuits of each The next step is to calculate the backfed breaker
inverter are: that must be placed in the main service panel to
handle the combined output of both inverters
3500/240 = 14.58 amps from the PV ac subpanel and to protect the con-
1.25 x 14.58 = 18.2 amps; use a 20-amp break- ductor carrying those combined outputs under
er and 12 AWG conductors. fault conditions from high utility currents.
The combined currents from both inverters are: “strings” of microinverters or ac PV modules may
be connected to the added combining panel (green
14.58 + 17.5 = 32.08 blocks). Or, a single inverter could be connected to
an existing load center (red blocks). In some cases,
The overcurrent device should be 40 amps (1.25 multiple inverters might be connected through an
x 32.08= 40.1 ), and this meets Code requirements ac combining panel and then backfeed an existing
for a 200-amp panel with a 200-amp main load center. Let’s start our examination of the
breaker. Note that the allowance of using 125% requirements at the inverter end of the circuit.
of the inverter currents allows a 40-amp backfed
breaker, whereas if the inverter output breakers Inverter Output Circuit
had been used, the resulting 45-amp backfed All utility-interactive inverters have a rated out-
breaker would have not met Code requirements. put current that cannot be exceeded. There are no
With overcurrent devices on each end of the surge currents in these output circuits and NEC
circuit between the inverter output ac combining 690.8 requires that the circuit and the overcurrent
panel and the 40-amp backfed breaker in the protective device (OCPD) be rated at 125% of
main panel, all conductors and equipment in this that rated output current. When the calculated
circuit should be rated for at least 45 amps which OCPD value is a nonstandard value, the next
is the total of the two inverter output breakers standard higher value should be used, but not
(20 + 25 = 45). This exceeds the rating of the to exceed the maximum overcurrent value given
40-amp backfed breaker in the main panel and in the technical specifications for the inverter.
is discussed above where there are overcurrent Conductor size should be selected so that it can
devices on the ends of a circuit. carry the inverter current and is protected by the
OCPD rating.
One Diagram Is Worth a Thousand The asterisk (*) by the 690.8 in the diagram
Words indicates that if there is an overcurrent device
Many people do better with diagrams than they mounted at the inverter, then the requirements
do with words, so figure 7.3 should be just up their of 705.12(B), and not 690.8, will apply. Some
alley. This big picture diagram can be used with installers and manufacturers use a circuit break-
many types of utility interactive PV systems. These er or fused disconnect at the inverter to meet
systems all start with a meter connected to the the requirements of 690.15 to have an equip-
utility as shown on the left (blue). After that, we ment disconnect at the inverter. The inclusion
may be dealing with an existing service disconnect of an overcurrent device at this location may
and the connected existing load center or with a require the output conductors from the inverter
PV supply-side connection, which is similar a sec- to be larger [705.12(B)] than would otherwise
ond service entrance on the existing premises wir- be required by 690.8.
ing system. In either case, the NEC requirements
of Article 230 apply as noted at the bottom of the After the First Inverter Overcurrent
diagram. In most jurisdictions, the local utility will Device
require a PV disconnect on the ac output of the Any conductor or busbar that can have power
PV system, and many areas will use a Renewable flowing from more than one source of supply
Energy Credit (REC) meter to measure the PV (under normal or fault conditions) such as the
system output. As shown, one or more single utility and a PV inverter, and where the conduc-
inverters may be connected or even one or more tor is protected by an overcurrent device on each
Figure 7.3 • The Big Picture Diagram.
supply source must meet 705.12(B) requirements. through the overcurrent device/disconnect on
This is the longstanding 120% allowance [when either an existing service disconnect or through
705.12(B)(2) conditions can be met]. Section the overcurrent device/disconnect on an added
705.12(B) is going to apply to all conductors and PV supply-side connection, the requirements of
busbars from the first overcurrent device connect- 705.12(B) apply all the way to the first overcur-
ed to the inverter output all the way to the service rent device connected to the inverter output.
disconnect.
These busbars and conductors would include Making the PV Circuit to Service Conductor
the busbars of any backfed main panel boards Connection
connected to one or two inverters or sets of Most, but not all, supply-side PV connections
microinverters, and any busbars in PV ac invert- are made to the service conductors between the
er combiner panels. The conductors or feeders meter location and the main service disconnect.
between the panelboards or load centers and the This assumes that the meter is the closest device
main service disconnects are also subjected to toward the utility supply and that a net-metering
the requirements of 705.12(B) as noted on the connection is desired, allowed and/or required
diagram. by the local utility. In some jurisdictions, a zero
sequence or cold meter socket is required by the
The Main Disconnect and on to the local utility and the main disconnect may be
Meter ahead of the meter socket. In either case, the al-
Any circuit between the meter and the service lowance of NEC 705.12(A) is for the connection
disconnect would be considered a service-en- to be made on the utility side of the main service
trance circuit and be governed by the require- disconnect.
ments of Article 230. However, after passing Local utility requirements will determine where
the PV connection can actually be made, and, of

course, the utility will have to be notified to turn
off power to the service before the connection is
made. It is usually acceptable to open the raceway
between the disconnect and the meter, insert a
j-box and make the connection using appropriate
splicing devices.
As mentioned previously, the PV inverter
output connection cannot be made to the internal
bus bars or conductors between the meter socket
and the main breaker on a meter-main combo
panel or in large switch gear unless those conduc-
tors or busbars are factory marked as tap points,
the manufacturer’s instructions and hardware are
available, and the utility approves the connection.
Otherwise the listing on the combo panel would
be violated.
Special, listed PV connection meter socket
adapters are available from several sources, but
may or may not be approved for use by a local
utility. These automatically provide a set of “PV
Photo 7.13 • Meter adapter used to provide sup-
ready” terminals that are connected to the load ply-side PV connection. Courtsey EATON.
side terminals of the meter (photo 7.13).
Several load center manufacturers make “ranch become more common and as battery systems
panels”, which have a 400-amp input busbar for this application proliferate in the industry
and two, 200-amp breakers connected to that [Article 706]. See photo 7.14.
busbar. One of the breakers feeds a normal load A utility-interactive multimode inverter is
center busbar and the second 200-amp breaker is used in these applications and that inverter will
available for a supply-side PV connection. have one ac input/output that is connected to the
In all cases, the connection will be to ser- utility, a second ac input/output that is connected
vice-entrance conductors and any devices, to the load circuits and a dc input/output that is
methods or equipment used must comply with connected to a battery bank. The PV array may
the requirements for such connections in this be connected to the multimode inverter as an “ac
environment. For example, a short-circuit current coupled” system where a separate utility-interac-
rating (SCCR) might apply to any meter socket, tive inverter converts dc PV energy to ac energy
busbars, fuse holder, disconnect or terminal block and feeds it to the ac load circuit that is connect-
which might carry fault currents. ed to the multimode inverter. The multimode
inverter charges the batteries from either the ac
Battery Backed up Systems. PV systems utility or the ac output from the PV array.
containing energy storage systems (ESS), usually Or, the system may be “dc coupled” where the
a battery bank, are becoming more common PV array is connected to a charge controller that
as utility outages due to natural disasters (hur- charges the batteries directly. The multimode
ricanes, snow storms, flooding and tornados) inverter will discharge the batteries into the load
Photo 7.14 • Typical 10 KW ac output whole-house battery-backed-up PV system. System is ac coupled with additional
utility-interactive inverters located outside at the PV array.
circuits or the grid when there is excess energy be some subset of the maximum building load
available or charge the batteries from the grid where only a few critical circuits are connected
when necessary. See Chapter 5 for additional to the battery backed up PV system, but the
details. maximum battery charging load must still be
In either type of system, it is important to note added to this load.
that the ac utility input on the multimode invert- Each type of multimode inverter from the
er may, at times, be a load on the utility feeder or several manufacturers will have software con-
service since it can pass through utility currents trols available to limit and otherwise control
to supply loads under normal operating condi- the various pass through ac currents to the
tions. That input can also supply utility currents connected loads, the ac currents that can be
to charge the batteries when excess PV energy is made available for battery charging and the
not available. maximum ac currents that can be supplied to
The amount of potential maximum load on or supplied from the utility source. The in-
the utility supply must be calculated and used struction manual for these relatively complex
in determining conductor sizes, overcurrent and devices must be carefully reviewed to ensure
disconnect device ratings and service entrance that the appropriate software settings have
ratings. The load may be the entire building load been made and that the external circuitry is
when a “whole house” PV battery backed up compatible with these adjustments. In general,
electrical system is installed plus the load due these software adjustments must not be accessi-
to maximum battery charging. Or, the load may ble to unqualified people.
Summary of NEC 705.12(B) quirements might also be applied in jurisdictions

It is obvious that the new 705.12(B) require- using earlier editions of the Code by accepting
ments are significantly different from those in alternate methods and materials waivers based on
past years. While many of them make sense the 2017 NEC clarifications.
from an engineering point of view, the real world The load-side connection for the utility-inter-
faced by inspectors and plan reviewers may be active PV inverter is not the easiest subject to
somewhat different where people typically ignore understand, but the correct application of these
instructions, ignore the code, and ignore warning requirements will yield a safer, more durable
labels. On the other hand, PV systems have not system. When the requirements of load-side
changed significantly from 2011 to 2017 and the connections become complex and expensive, a
electrical systems they are being connected to supply-side connection may be used.
have not changed significantly, so these new re-
156 Chapter 8 Plan Checking and Inspecting
Chapter 8 Plan Checking and Inspecting 157
08
Plan Checking and Inspecting
A formal plan review process and plan reviewers not made in future installations, hopefully, by the
are not found in numerous smaller jurisdictions same contractor.
due to budget constraints and the perceived lack However, we now have the ubiquitous pho-
of need for such activities. This is probably an tovoltaic (PV) power system springing up in
acceptable situation where routine inspections in- numerous neighborhoods on residences and
volve residential electrical systems and the small- commercial buildings throughout the country.
er commercial systems. Those electrical systems The photovoltaic (PV) industry is dynamic
have remained essentially unchanged for decades and fast changing. New PV modules are being
with the exception, of course, of equipment such released monthly and each new module has a
as arc-fault circuit interrupters and ground-fault different size and different ratings. New inverters
circuit interrupters and the like. and other equipment are also coming to market
The load center that I installed in 1960 in the with significant frequency. Although the equip-
home I grew up in is nearly identical to the load ment is connected with the conductors and uses
center I installed in my retirement home in 2010. overcurrent devices that we are familiar with, the
Yes, each edition of the National Electrical Code outdoor environment and the rapidly changing
(NEC) provides changes in the requirements for equipment generally point to the need of some
residential and commercial electrical installations, sort of plan review process to ensure that the
but the equipment is very similar from year-to- somewhat unique PV Code requirements are
year and the methods of installing that equip- being met.
ment are very familiar to the electricians and Most inspectors would agree that an unfamiliar
inspectors. The electrical contractors are familiar installation, a new contractor, and unfamiliar
with the equipment and installation procedures equipment as well as new and slightly different
and although there are differences from instal- Code requirements dictate that a longer-than-usual
lation to installation, there are many similarities. inspection may be required. In fact, it is difficult
Mistakes in the installation occasionally get made when inspecting the typical PV system as installed
and are found by the inspectors, corrected and are in the field to fully ascertain that all the Code
Photo 8.1 • (top) Module labels and wiring not

readily accessible
Photo 8.2 • (left) Conductor markings not easily

accessible in roof-mounted dc source-circuit com-
biner. Pre-2017 NEC “grounded” PV array
Even without a formal plan review process or

dedicated plan reviewers, it would behoove the
typical inspector to spend a few minutes in the
office with the permit application and the plans,
which will be described shortly, and verify that,
at least, on paper the proposed PV system meets
requirements have been met. The modules are Code requirements.
fastened to the roof and the rating labels cannot be
seen so the ratings for open-circuit voltage (Voc), The Plan Review
short-circuit current (Isc), and maximum protec- Of course, to do a plan review one must have a
tive overcurrent device cannot be verified (photo plan. Many jurisdictions do residential electrical
8.1). The wiring is typically behind the modules or permitting across the counter with no substantial,
in conduit and the markings on the conductors in detailed plans required for the electrical system.
many cases cannot be verified and, like the module A form is filled out, the fee is paid, and the
labels, some of the equipment such as combiners is permit is issued. At some point, an inspection is
not readily accessible (photo 8.2). requested, and the inspector makes the inspection
Figure 8.1 • Three-line diagram of a 2.5 KW PV system. PV system disconnect is the 2-pole backfed breaker. Extra 6 AWG
GEC from inverter is not required by 2017 NEC.
and either approves or disapproves the installa- ing a plan review to determine if the basic design
tion. While this works for normal residential and of the PV system meets minimum code require-
for some small commercial electrical installations ments. Figure 8.1 shows a three-line diagram of a
where the equipment and Code requirements are small PV system with a single string of modules
fairly well standardized, it is probably not the best as an example and such a diagram/schematic
option for PV installations, as noted above, where should be required in any permit application.
the equipment and the installation techniques are
changing frequently. Equipment Lists and
Some jurisdictions are adopting an expedited Specifications
PV permitting process for PV systems under 10 A list of the equipment used and the specifica-
kW. This expedited permitting process usually tions for that equipment should be included with
includes a diagram and other information and is, the permit. This list would include the PV-spe-
in some cases, based on the process described on cific equipment such as the PV modules, the
the SolarABCs’ website. That expedited per- inverter, the fuses, and circuit breakers. Listing/
mitting report and substantial amounts of other certification and rating information must be
PV-related information can be found at: http:// included. The specifications of this equipment
www.solarabcs.org/about/publications/reports/ are necessary to determine if the conductors have
expedited-permit/index.html. been properly sized and that the fuses and circuit
The following paragraphs contain some sugges- breakers used in the dc parts and ac parts of the
tions that have been found useful when perform- system are properly rated. Factory cut sheets or
Figure 8.2 • One-line PV system diagram
pages from instruction manuals are the preferred several additional pieces of information. An in-
way to present this information. stallation instruction manual for the PV module
Since the PV modules and inverters are con- should be included along with a specification
stantly changing, the preferred option would be sheet for the module since frequently the spec-
to include copies of the installation manuals for ifications are not contained in the instruction
the inverter, module, and dc PV combiner with manual. An installation instruction manual for
each submission. Since not all module installation the utility-interactive inverter should also be
manuals include the technical specification, the included along with an instruction manual for
technical specifications should also be included. any dc combiner, where such a combiner is used
to combine source circuits. Cut sheets should be
The Diagram provided to show the ratings of all disconnects
A one-line diagram such as shown in figure 8.2 and overcurrent protective devices used in the
should accompany the permit application. Since system. Information should be provided on
the details of disconnects and grounding are the ratings of any existing electrical equipment
not familiar to all involved, a three-line diagram such as load centers (main breaker and busbar
would be even better as shown in figure 8.3. rating) that the PV system will be connected to
While a formal CAD-generated diagram on or that PV ac currents may pass through all the
24” x 36” paper is not required, something better way from the PV inverter output to the service
than a “back of the envelope” sketch should be entrance conductors.
presented. The circled letters in the figures will
be referenced below to indicate information that The System
should appear on or be attached to the plan. On the one- and three-line diagrams, the fol-
This diagram should be accompanied by lowing information should be indicated, or that
Figure 8.3 • Three-line PV system diagram. Either the ac disconnect or the utility disconnect can serve as the PV system
disconnect.
information should be attached. The following Due to the exposed, outdoor location and high
numbers reference items in figures 8.2 and 8.3. operating temperatures, all conductors should
have insulation rated for 90°C and wet conditions
A. PV Array (in conduit, THHN/THWN-2, XHHW-2 or
A.1. The type and number of PV modules in each RHW-2). Exposed conductors ( USE-2 or PV
series string should be indicated. The open-circuit Cable/PV Wire) must be suitable for the hot,
voltage (Voc) of each module, times the number of wet, outdoor environment.
modules connected in series, times a cold tempera-
ture factor (690.7) equals the maximum systems B. Conduits
voltage and must be less than the maximum direct B.1. Conduits will typically be used throughout
current (dc) input voltage of the inverter and less the system and specifically after the dc wiring
than the voltage rating of connected direct-cur- leaves the PV array and runs through the struc-
rent (dc) equipment (wires, overcurrent devices, ture. They will be installed in various locations,
disconnects). some of which may be in sunlight. See 310.15(B)
The coldest expected ambient temperature (2) and 310.15(B)(3) for temperature calculations
should be noted on the diagram. A label on the and adders. The average high and the height of
back of each module as shown in photo 8.3 will the conduit above the roof should be specified.
give the electrical parameters needed for the Conduit fill and conductor ampacity calculations
code-required calculations. These values should for conduit fill and temperature calculations
appear on the diagram. A photo of the label should be included or attached.
should be included since this label will not be One source of temperature correction
visible once the modules have been installed. factors for conduits in sunlight based on the
A.2. The ampacity of module interconnection 2014 NEC (still useful, if not specifically
cables, not less than the larger of: 1.56 times the required in the 2017 NEC) is Copper Devel-
module Isc or conditions of use applied to 1.25 Isc. opment’s web site:
C. Module and String Overcurrent

Protection and PV DC Disconnect
C.1. Overcurrent protective devices (OCPD) in
dc circuits may not be required when there are
only one or two strings of modules connected in
parallel to the inverter. Three or more strings of
modules typically require an OCPD in each string.
The current rating of the OCPD, when required,
should be 1.56 Isc for that circuit (690.8, 690.9).
The voltage rating of the OCPD should be not
less than the maximum PV systems voltage found
in A.1. The strings may be combined in parallel in
a combiner box ahead of an unfused dc PV dis-
connect or combined at the output of the dc PV
disconnect (figure 8.2 and photo 8.4). See chapter
3 for a more detailed look at the calculations and
Photo 8.3 • PV module label showing electrical ratings the requirements for OCPD in the dc PV array
circuits. Any OCPD connected in series with a
module or string of modules should not have a
value greater than the maximum series fuse value
marked on the back of the module (photo 8.3).
C. 2. The PV array output should be connected
to the top or line side of the main dc PV dis-
connect. The circuit from the dc PV disconnect
to the inverter dc input should be connected to
the bottom or load side of the disconnect. Both
ac PV output conductors must be switched by
Photo 8.4 • DC PV disconnect with required markings. the disconnect except in those very small solidly
2017 NEC requirements are slightly different. See 690.13 grounded PV systems.
and 690.53.
C. 3. PV output conductors, after any com-
http://www.copper.org/applications/electri- bining of series strings, should have an ampacity of
cal/building/pdf/rooftop.pdf not less than the larger of: 1.56 times the module
Isc times the number of strings in parallel or con-
B. 2. The dc PV source circuit and dc PV out- ditions of use applied to 1.25 Isc times the number
put conductors must remain outside the struc- of strings in parallel. The overcurrent device must
ture until they reach the readily accessible PV always protect the conductors. Additional details on
dc disconnect unless the conductors are installed sizing conductors and overcurrent devices will be
in a metallic raceway [690.13, 690.31(G)]. The found in chapter 6.
exception for the use of metal conduits inside
the building does not apply to the ac output of Module Grounding
utility-interactive inverters since these circuits Starting with the PV module, a grounding detail
are similar to ac branch circuits. that is going to be used to ground the module
Photo 8.5 • Top clip mounting system must be men- Module Mounting
tioned and installation locations identified in the module
In a similar manner, a mounting detail for the
instruction manual to keep the module listing valid. Note
that drilling two extra holes in this clip has violated the PV module should be included with the diagram.
listing on the clip. This mounting detail should be consistent with
the mounting instructions for the PV module.
frames should be provided on or with the dia- Again, if the module is not mounted according to
gram, and this grounding detail should be consis- the module instructions, the listing is invalid, and
tent with the grounding instructions provided in the module installation does not comply with the
the module instruction manual. If the module is requirements of the NEC. The use of a mounting
not grounded in a manner consistent with the device or racking system that has top clips to
module instruction manual, the listing on the mount the modules has become common recent-
module will be invalid. Since the NEC requires ly (photo 8.5). The module instruction manual
that the module be certified/listed [690.4(D)], must show the use and location of top clips for
then the installation of an unlisted module no mounting the module if this mounting system
longer meets NEC requirements. A frequent is to be acceptable and the installation to remain
problem in this area is the use of some sort of a code-compliant. This alternate mounting system
grounding device that does not use the marked will be in addition to the normal mounting holes
grounding holes shown in the module instruction that are drilled in the back frame of the module
manual and the instruction manual makes no that are normally used for mounting. Although
mention of grounding the module at positions not required by the NEC, the racking system
other than the marked grounding holes. may be listed to UL Standard 2703 that governs
Conductor Conditions of
Use
The conductors from the PV mod-
ules to the inverter will pass through
several different areas on or in
the building; each area will have a
different set of conditions of use. At
some places, they may be in free air
behind the modules in the shade or
in sunlight. (See photo 8.6.) At other
locations, they may be in conduit.
In some locations, the conduit will
Photo 8.6 • Conditions of use may vary greatly along each source be in sunlight close to the roof and
circuit from the module to the inverter. And, some circuits may need at other locations the conduit may
additional protection from Mother Nature.
be significantly above the roof but
still in sunlight and frequently the
the mechanical and electrical grounding require-
conduits will be in shade on the roof. Usually,
ments for the racking system.
at some point, the conduit will be inside condi-
tioned space as it is routed to the inverter. The
Racking Systems diagram should be annotated to show these
Underwriters Laboratories (UL) Standard 2703 different conditions of use and the temperatures
details a list of performance requirements for and spacing from the roof expected in each
PV racking and mounting systems. Although segment of the conductor run. Of course, the
the Code does not require racking or mounting worst-case conditions of use will be applied to
systems to be certified/listed, a racking system determine the conductor size. It should be noted
that complies with this Standard will ensure that that, in some installations, the conductor size for
the racking system is structurally robust, and that the field-installed conductors may exceed the
the racking system can be used as an equipment size of the conductors attached to the modules
grounding conductor. themselves because the conditions of use for the
The certification/listing of the racking system field-installed conductors and conduit in sunlight
to UL 2703 does not, however, automatically may be more severe than those associated with
guarantee that the racking system and its module the conductors attached to the modules which
mounting, and module grounding provisions are typically in free air. An ampacity calculation
meet code requirements. At this time, both the showing the conditions of use and the required
racking system instructions and the module conductor size should be shown for each separate
instructions must indicate that these two devices condition of use or at least the worst-case con-
have been evaluated together as a system in order ductor size and that calculation should be shown
to ensure that the mounting of the module to the (310.15).
racking system and the grounding of the module
to the racking system comply with all provisions D. The Inverter
of both standards and, therefore, will meet the D.1. The inverter must be listed for utility-inter-
requirements of the NEC. active (U-I); use 705.6.
D.2. The inverter maximum input voltage must

not be exceeded in cold weather [110.3(B)].
D.3. Most PV systems must have a dc ground-
fault protection device (GFPD) [690.41(B)].
When a GFPD is built in to the inverter (nearly
all U-I inverters), there should be no external (to
the inverter) bond between the grounded circuit
conductor and the grounding system. Addition-
ally, an equipment grounding conductor must
be routed from the PV array through the dc PV
disconnect all the way to the inverter.
D.4. In addition to ac and dc equipment
grounding conductors, the inverter providing a
PV system functional ground may have a require-
ment for connecting the ac equipment-grounding
conductor [690.47(A)]. The dotted lines in figure
8-2 show alternate routing and bonding for the dc
Photo 8.7 • Inverter with internal ac and dc disconnects.
These may meet the requirements for inverter isolating
grounding electrode conductors (required by the
devices (690.15) 2014 NEC).
Photo 8.8 • Inverters installed with external ac and dc disconnects

E.2. The inverter ac disconnect

should be “grouped” with the
dc inverter disconnect and both
should be “near” the inverter. The
AHJ determines the definitions of
“grouped” and “near.” Most systems
use the PV disconnect as the dc
inverter disconnect, but if the PV dc
disconnect is on the outside of the
building (no longer a requirement
in the 2017 NEC) and the inverter
is on the inside, a second dc inverter
disconnect may be required inside
Photo 8.9 • Utility-required ac disconnect. Inverters are located at the the building at the inverter loca-
ground-mounted PV array. tion. The same thing would apply
if the backfed circuit breaker in
D.5. DC and/or ac disconnects internal to the the building load center were on the outside
inverter are acceptable if the inverter is readily wall and the inverter were on the inside. A
accessible and the AHJ judges that only quali- disconnect (usually a circuit breaker) would be
fied people will service the inverter (photo 8.7). required inside the building near the inverter.
Internal dc disconnects are common in the smaller E.3. From the above, it becomes obvious that
inverters. Otherwise, external disconnects will be the system diagram should show the approximate
needed (photo 8.8). Internal disconnects, if circuit physical location of all components.
breakers, will not protect the ac output circuits
from utility-sourced fault currents and an exter- At the DC PV Disconnect and The Inverter
nal OCPD will be needed at the utility point of The inverter instruction manual should be con-
connection. sulted to determine any instructions for imple-
menting the PV system functional ground.
E. Inverter AC Output Overcurrent Device
and Disconnect Inverter AC Output Circuit
E.1. Any OCPD located in the inverter ac output The conductor from the inverter output to the
should be rated at 1.25 times the rated continuous first overcurrent device must meet the require-
output current of the inverter. The rated continu- ments of NEC Section 690.8 for size. And, the
ous current is specified in the inverter manual or overcurrent device itself should be rated accord-
is calculated by dividing the inverter rated output ing to the requirements of Sections 690.8 and
power by the nominal ac line voltage. This OCPD 690.9 of the NEC. The conductor sizes for the
may be a backfed breaker located in the dwelling ac output circuit of the inverter including any
load center, the place where any possible fault cur- conductor downstream of the inverter after the
rents for the inverter ac output conductor would first overcurrent device that may carry both
originate. A backfed breaker in the dwelling load inverter output current and utility current must
center could also be the PV system disconnect and comply with NEC Section 705.12(B). Someplace
the inverter ac disconnect (690.13, 690.15) if the in the permitting package, the calculation should
inverter were located near the load center. be shown for each circuit involved all the way
back to the service-entrance equipment. Backfed Point of Connection-Load Center

circuit breakers in each panelboard or load center G.1. Most of the smaller residential and commer-
that will handle PV current should be marked cial PV systems will make the point of connec-
on the diagram showing their correct location, tion with the utility through a backfed breaker in
which in most cases will be at the opposite end the building. NEC Section 705.12(B) establishes
of the busbar from the main breaker or main lugs the requirements. See chapter 7 for the details of
for that panelboard. Chapter 7 provides addi- these connections.
tional details on the various PV system-to-utility
connections. The Inspection
With the some of the Code requirements verified
F. Utility-Required AC Disconnect during plan review, the field inspection will usual-
F.1. Many utilities require a visible-blade, lock- ly go much smoother. Code requirements that are
able-open disconnect in the ac output circuit of found not to be met on paper may be flagged for
the inverter. This disconnect is usually located the installer, and they should be corrected before
within sight of the service-entrance meter so that taking valuable time going to the site and at-
emergency response people can easily find it. The tempting to ensure that everything meets Code by
top terminals (line side) of this disconnect should a visual inspection of the installation. I have
be connected to the circuit that comes from the reviewed several hundred plans for PV systems
utility (supply-side connections, [705.12(A)]) over the last 30 years (5 kW to 4 MW) and note
or the ac load center (load-side connection, that many systems fail to meet all NEC require-
[705.12(B)]) because they will usually be ener- ments on paper. The systems that do comply with
gized by utility voltage. The bottom terminals the NEC requirements are usually plans from PV
(load side) should be connected to the circuit installers/systems integrators that have been in
from the inverter. This disconnect may be fused business a number of years and have been install-
or unfused depending on the specific require- ing similar systems over that period.
ments of the utility and its location in the circuit.
Photo 8.9 shows an ac disconnect between the
Summary
REC meter and the revenue meter that serves
All the above information should be included in
as both PV system disconnect and the utili-
plans submitted for obtaining a permit for the in-
ty-required ac PV system disconnect. The utility
stallation of a PV system. The more information
point of connection is inside the house through a
submitted, the easier it will be for the PV system
backfed circuit breaker in the load center.
designer/installer to communicate to the inspec-
F.2. The utility disconnect must have a mini-
tor/permitting official that the system design and
mum current rating of 1.25 times the maximum
component selection meet the requirements of
continuous output current of the inverter (690.8)
the NEC. It is far more cost effective to change
and be protected by the downstream (toward the
the design on paper before any hardware is
utility) OCPD. For a supply side connected PV
purchased and installed than it would be after the
system [705.12(A)], the disconnect must also be
system has been installed. Ready for the inspec-
rated as suitable as service entrance equipment
tion? Read on.
[690.13(C)].
168 Chapter 9 The Process of Inspecting PV Systems
Chapter 9 The Process of Inspecting PV Systems 169
09
The Process of Inspecting
PV Systems
While large, utility-scale photovoltaic (PV) The Way It Should Be Done—

power systems account for more of the installed The Author’s Opinion
megawatts and gigawatts of PV power in the The Overall Process. The inspection process for a
United States, the typical inspector will be PV system should begin with the permitting pro-
inspecting far more residential and small com- cess that should include a full-system description
mercial PV systems than utility-scale PV systems. which shows a three-line diagram and manuals
While failures in those utility-scale PV systems for all PV-unique equipment being installed.
can be spectacular, these systems are usually Chapter 8 addresses the details. Hopefully, per-
ground-mounted in remote areas, and damage is mits will be issued only to individuals or organi-
restricted to the equipment itself, and of course zations who are qualified to install PV systems
the pocketbooks of organizations involved direct- and have the necessary licenses. In the best of
ly with the system. circumstances, the permit applicant will be well-
On the other hand, when residential and small
commercial PV systems fail, the resulting damage
can affect homes, commercial buildings, and
people (including first responders) in a very real
manner. To ensure the safety of the public, the
inspection community including administrative
personnel and managers, as well as the people
who are responsible for funding must work
together to provide an exceptional final PV in-
spection before the system is turned on. That final
inspection will ensure the safety of an electrical
power system that may be generating energy
Photo 9.1 • Some PV systems may experience trouble,
for the next 40 or 50 years, possibly without any now or in the future, when installed by less than compe-
further attention. tent organizations.
to get adequate knowledge and training so that he

or she can effectively inspect PV systems to ensure
the safety of the public. Acquiring this background
is not as easy as it sounds (photo 9.2).
For example, the author is involved in confer-
ence calls a couple of times a week dealing with
the standards associated with PV systems. New
codes and standards are reviewed on a regular ba-
sis and assistance is provided in writing the same.
Even the author has trouble keeping up with the
changing technology, the changing standards, and
the changing codes associated with PV.
As inspectors, we must take it upon ourselves
Photo 9.2 • Hands-on field inspection training for AHJs
to get as much training in this ever-changing
versed in installing residential and commercial technology as possible. There are online training
electrical systems, with additional training in PV programs, and there are hardcopy materials that
code requirements and installations. can be used. The first example that comes to
A plan review step in the comfortable confines mind of hardcopy material (or electronic) is the
of an office environment should be performed on National Electrical Code Handbook (NECH). In
the permit application material. Unfortunately, many cases, the informative material found in the
many aspects of the PV installation cannot be NECH is based upon the background and tech-
verified after the system is installed. Modules will nical information that went into developing the
be mounted close to a roof, and the labels cannot code requirement. Every inspector should have a
be seen. Conductors may be installed in not copy of the Handbook and review this document
readily accessible areas, and sizing and types will almost continually. It takes nearly three years to
be difficult to verify. gain familiarity with the current version of the
The initial inspection should verify that all Code and then the new edition is released.
requirements in the NEC have been met and Another source is these “Perspectives on PV”
that the equipment that has been installed agrees articles written by the author that are archived
with the information provided in the permit on the IAEI website. We should also review the
application. Typically, there is no utility interface IAEI Publications Catalog (http://www.iaei.org/
connection at this time, and several parts of the web/store) for all the significant documents that
system operation cannot be evaluated for safety. summarize the new code requirements. These
In most cases, the initial PV inspection will be documents are updated every code cycle, and the
followed by a utility inspection where ac elec- IAEI Publications Catalog is probably the best
trical power is connected to the system by the source for a broad spectrum of documents that
utility. After this is done, a subsequent inspec- can be used in educating the inspector on the
tion should be made to verify that equipment nuances and changes in the Code.
like PV Rapid Shutdown systems and the like Another source of wide-ranging technical
are functioning as intended. articles on PV systems, PV equipment, and PV
installations can be found on the Solar Amer-
First Comes Training and Education ica Board of Codes and Standards (solarabcs)
It is imperative that the inspector of PV systems website (http://www.solarabcs.org/).
One source of online training that is widely d. PV module mounting and grounding details
advertised in IAEI Magazine and other publica- showing that these installation details are com-
tions is the “PV Online Training Course for Code patible and comply with the instructions in the
Officials” developed by the Interstate Renewable PV module and the mounting rack manuals.
Energy Council (IREC), IAEI and others. It Requiring this material in the permit applica-
can be accessed here: www.pvonlinetraining.org. tion will help to ensure that qualified persons are
Be advised; a game is included in this training making that application. The material also will
program. See the excellent article in IAEI Mag- be useful in the plan review process (described
azine (Nov-Dec 2016) on this program (https:// below) that will facilitate the actual on-site
iaeimagazine.org/magazine/2016/11/04/become- inspection.
solar-smarter-with-pv-training-online/).
Expedited Permitting. For those jurisdictions
Building Officials and Administrators. Inspec- not having a formal permitting process, an
tors must be provided with adequate funding expedited permitting process developed by the
and time allowances for the necessary study Solar American Board of Codes and Standards
and training. Workshops and presentations may (SOLARABCS) and written by Bill Brooks and
be brought to a central location so that many others can be found here: http://www.solarabcs.
inspectors can participate, or video conferencing/ org/about/publications/reports/expedited-permit/
webinars can be set up to ensure that all inspec- This expedited permitting process is very useful
tors maintain currency in their respective disci- for PV systems of 10 kW and less and includes
plines, in this case, PV systems. diagrams, fill-in forms, and code-based calcula-
tions. It has been used by numerous jurisdictions
The Permit throughout the country and modified for use
The permit for a PV system should include with additional systems by other jurisdictions.
more than a simple form and the payment of a
fee. That permit application should contain the The Plan Review
following: Advantages. One advantage of having the de-
a. A three-line diagram of the system showing, tails described above submitted with the permit
as a minimum, conductor sizes, conduit sizes, application is that a quick review of this material
equipment grounding circuits, disconnects, over- will give some indication of the capabilities and
current protection, and the location and method competencies of the organization making the ap-
of making the utility interconnection. plication and, hopefully, the person installing the
b. The code calculations for the maximum system system. If the basic system design in the permit
voltage including the expected lowest tempera- package does not meet code requirements, it is al-
ture at the site, the ampacity calculations for each most certain that the system will not be installed
circuit including the ambient temperatures used, in a safe and code-compliant manner. Another
and the NEC 705.12 calculations for the inter- advantage of performing a plan review is the fact
connection point. that the review can verify partial code-compli-
c. A copy of the installation manuals for the PV ance in the warm, quiet, office environment and
module, the dc combiner, the inverter, the PV not out in the cold, blustery or hot, sunbaked
rapid shutdown system/equipment and any other environment of a rooftop PV system.
PV unique equipment, as well as, cut sheets for As mentioned above, it is physically not pos-
the various disconnects and any load centers. sible to fully access all the labels and markings
on the PV modules and many of the cables than rated maximum equipment voltage
after the PV array has been installed. The PV [690.7]
arrays are, in many cases, mounted within a • Cable types suitable for the environments,
few inches of the roof and it would require suitable for the circuit, and properly rated
unmounting a PV module to view the data [690.8]
label on the back. Depending on the module • Conduit and raceway selections appropri-
grounding method being used, grounding ate for the environments and code-com-
hardware might have to be replaced when a pliant [Chapter 3]
module is removed from the racking system. • Conduit fill calculations correct [Annex C]
Conductors installed under the modules and • Disconnects suitable for the type sys-
the connectors used between the PV modules tem-grounded or ungrounded
and the field-installed wiring may not be easily • Disconnects properly rated and properly
accessible. located [690 Part III]
Additionally, reviewing the system on paper • Overcurrent devices properly rated and
allows the AHJ to see the overall system and properly located [690.9]
how the various components are interconnected. • Utility point of connection properly rated
It will also show the flow of power through the and located [705.12]
system which will assist in determining if dis- • Any utility-required disconnects and
connects and overcurrent protection devices have meter properly rated and located
been located properly. • Equipment grounding and system
To a certain extent, some of this information grounding circuits correct
is provided by a full permit package and can be • PV Rapid Shutdown System equipment
used to verify code compliance. Of course, this located in circuits where required by
assumes that the equipment and materials in instructions [690.12], [110.3(B)]
the permit package match the equipment and • DC PV Arc Fault system incorporated
materials that will be installed; and that may not into the PV system [690.11]
always be the case. This item should appear on • DC combiners properly rated
the field inspection checklist. • Inverter and inverter circuits properly
rated for dc inputs and ac output
Items to Look For. Here is a brief checklist
of items that should be verified during a plan The Onsite Inspection
check. Significant numbers of items that are not First impressions are important. Have you
code-compliant may signify that the basic PV worked with this PV installer or systems integra-
system design is inadequate and that the permit tor before? Is the site uncluttered and the system
package be returned to the submitter for correc- ready for inspection? Did they make the roof ac-
tions before any field inspection can be under- cessible in a safe manner? From a distance, does it
taken. A more detailed checklist will be found in look like good workmanship has been employed?
Appendix A. As with any electrical power system, atten-
• All equipment listed for the specific tion to details during the installation process
application [690.4(B)] is critical, and the overall first impressions
• Modules and racks compatible for mount- give some indication as to whether the install-
ing and grounding [110.3(B)] er has exercised good work procedures and
• Calculated maximum system voltage less habits.
Equipment installed to match the permit?

The equipment that is installed should match the
equipment that was listed in the permit applica-
tion package. Variations in any components such
as the PV modules or the inverter may necessi-
tate recalculation of various parameters and the
replacement of conductors, overcurrent devices,
and other equipment.
Let the installer operate the system. From
a liability and safety point of view, it is usually
a good idea to let the installer of the system
perform all handling of the system components
Photo 9.3 • Inspecting PV systems can result in unpleas-
and operation of those components during the
ant surprises.
inspection (photo 9.3).
The installer should have the necessary tools the main service panel and the grounding elec-
and test equipment available to open equipment, trode(s) should be verified by visual inspection.
make measurements, and exercise the equipment The installer might be asked to show that various
in various modes to demonstrate compliance bolted grounding connections have been properly
with Code requirements. Unfortunately, the PV torqued, and that conduit nuts and grounding
systems have reached a level of complexity that bushings are tight. Grounding/bonding bush-
it is no longer possible to verify full code-com- ings should be used on all metal conduit fittings
pliance by just looking at a static assembly of operating over 250 volts ac or dc (250.97). (See
wires, conduits, circuit breakers, PV modules, photo 9.6.)
inverters and other equipment. In most cases, On to the roof—Or not. What are some AHJ
the equipment must be operated to demonstrate options when the local jurisdiction does not per-
the proper safety features and compliance with mit personnel to climb on the roof? The installer
NEC requirements. In many cases, the utility will should be asked to provide close-up pictures of
require an inspection by the local jurisdiction the PV module mounting and grounding meth-
before they make the interconnection and they ods and devices used. The pictures should show
may necessitate a second inspection by the AHJ that the instructions for the module and rack
of the operational system. mounting system were followed.
The installer should be able to show the torque Although not required by Code, some jurisdic-
screwdrivers and torque wrenches that were used tions are requiring the use of UL 2703 certified
in making the electrical connections and to show rack/mounting systems on the roof of residences.
where the torque values for the various pieces of Since these racking systems cannot directly guar-
equipment can be found. All electrical contacts antee the quality of the actual attachment to the
that are bolted or screwed must be torqued to the structure of the house, the author feels that any
proper value to ensure a safe and durable connec- well-assembled module mounting and grounding
tion. [See 110.3(B) and photos 9.4 and 9.5.] system should suffice.
Proper grounding is critical for the long- The same considerations might be given to
term safety of the system. Before touching ground-mounted installations where Unistrut em-
any parts of a new PV system, the equipment bedded in concrete-filled holes might be judged an
grounding systems from the PV modules to adequate mounting system. For ground-mounted
systems that are readily accessible, the exposed

single-conductor module wiring must be made
“not readily accessible” by fencing or guarding
[690.31(A) and photo 9.7].
Instructions supplied with new modules will
have the following phrase or equivalent: “This
module remains certified to UL Standard 1703
only when mounted and grounded according to
the included instructions.”
A picture of the back or underside of the
array will indicate if the loose, single-conductor
Photo 9.4 • Various sized torque tools are necessary for module wiring has been properly secured. Any
making code-compliant electrical connections. conductors not well-secured for decades of use
will be subject to wind-driven movement and
possible abrasion of the insulation, creating a
potential shock or ground-fault hazard in the
future (photo 9.8).
Another relatively recent issue in PV installa-
tions is the fact that the mating pairs of connec-
tors used on modules and field-installed USE-2
and PV wire must be from the same manufactur-
er and of the same series. Even though connec-
tors from two different manufacturers are said to
be “MC-4 compatible,” they have not been tested
and certified/listed with each other and therefore
it would be a violation of the connector and
Photo 9.5 • Improperly torqued splicing devices have
module listing to use them together (photo 9.9).
melted.
See chapter 3 for additional details.
Back on the Ground

At the utility point of connection. In a
705.12(B) utility-interactive inverter output
connection, the rating and location (position on
the load center busbar) of the backfed PV breaker
should be verified. Note that the 2017 NEC uses
the term interactive power source or just power source
when referring to the utility-interactive inverter.
The 120% or 100% rule should be met, and for the
120% rule, the backfed PV breaker(s) should be
Photo 9.6 • Metallic conduit operating at 400 volts dc
at the opposite end of the busbar from the main
connected to a concentric knockout with no grounding/
bonding bushing. Grounding lug (not listed for this enclo- breaker (photo 9.8). Other sections of 705.12(B)
sure) attached with sheet metal screw violates Code and (2) may be used and these should be substantiated
disconnect listing. with the appropriate calculations.
Many residences and commercial buildings

have subpanels. Where subpanels are involved,
attention must be directed to ensuring that any
feeders or main-lug-only subpanels have appro-
priate ampacity or ratings to deal with backfed
PV currents in the main panels adding to the
currents from the main service breaker. In some
cases, the feeder may be protected by a break-
er at its source end, and the subpanel may be
protected by a main breaker in that subpanel. If
these protections are not afforded, the ampacity
of the feeder and the rating of the subpanel will Photo 9.7 • Exposed, readily-accessible module wiring
“guarded” by grounded, galvanized hardware cloth at-
have to be increased. Having the backfed PV
tached to back of module racks.
breaker mounted in the last subpanel and in the
last position in that subpanel will avoid some
complications, but feeders and the main panel
must be checked for Code compliance in Article
705.
For a supply 705.12(A) supply-side connec-
tion, the rating of the overcurrent protective
device should be checked, and the location and
wiring (utility connection to the top of any
safety switch) of any utility required or used ac
PV disconnect verified. The overcurrent device
connected to the output of each inverter may
not exceed the maximum overcurrent protective
Photo 9.8 • Improperly secured module wiring will be
device as specified in the inverter installation subject to future wind-driven damage.
manual; yet another reason to have the manual
available during the inspection. The presence of
the required overcurrent device within 10 feet of
the service entrance connection point (705.31)
should be verified.
It is the opinion of the author, since those
(unnamed) conductors between the actual
connection to the service-entrance cables
and the first overcurrent protection device
are subject to the full available fault currents
from the utility and because those conductors
may be smaller than the service-entrance
conductors, they should be treated as ser- Photo 9.9 • Bottom: Connectors from the same manufac-
turer/series properly paired for a code-compliant installa-
vice-entrance conductors per the requirements
tion. Top: Connectors from two different manufacturers im-
of Article 230 with respect to routing and properly paired, resulting in a certification/listing violation,
mechanical protection. a Code violation and a potential safety hazard.
After the System is Powered PVRSS with the well-marked initiator, and the
With the complexity of modern PV equipment, AJH should verify that the measured voltage
full verification of the installation for safety and drops to the required value of 30 volts within 30
compliance with the NEC requirements (includ- seconds. Note: at some future date, the ac voltage
ing Section 110.3(B) that requires compliance level may be reduced to 15 volts due to proposed
with equipment installation instructions) will changes in UL Standard 1741. Both line-to-line
usually require the AHJ to inspect the system and line-to-ground measurements will be re-
during operation. quired.
Verify equipment adjustments and settings. String inverter systems. The string inverter
Equipment like string inverters, multimode system will usually require that the dc conduc-
inverters, charge controllers and PV Rapid Shut- tors from the array and the ac conductors from
down Equipment (PVRSE) may have software/ the inverter be controlled by the PVRSS. Some
firmware adjustments that must be properly inverters may be listed as PV Rapid Shutdown
made to ensure the safe and correct operation of Equipment (PVRSE) and will meet the voltage
the product [110.3(B)]. The AHJ should have the requirements on dc inputs and/or ac outputs
installer demonstrate that all adjustments have without external equipment. Appropriate meters
been completed in accordance with the product should be connected to the dc input conductors
instruction manuals. that go to the inverter (normally, at the closed
dc disconnect) and to the ac output conductors
PV Rapid Shutdown System (PVRSS). of the inverter. The installer should activate the
The inspector should have the installer demon- PVRSS initiator, and the AHJ should verify that
strate the proper operation of the PVRSS since all voltages go to the required 30-volt levels with-
this system is directly involved with life safety in 30 seconds. AC voltage requirements may be
issues for first responders. The ac power to the reduced to 15 volts due to future changes in the
building that the PV is installed on should UL standard. Line-to-line and line-to-ground
be turned off, as it would be during any first measurements will be required.
responder activity. In fact, the main service
disconnect or the ac PV disconnect may have to Arcs Will Happen
be marked as a second PVRSS initiator in some As the PV system ages, it is expected that there
systems to ensure the ac inverter output circuits will be greater opportunities for series arcs to
are controlled by the PVRSS in addition to the form in the dc array wiring and the modules.
dc circuits. The main ac service disconnect for Connectors, even those properly mated in pairs,
the building may or may not be the PVRSS ini- may loosen over time from wind driven vibra-
tiator. In any event, the ac output circuits of the tions where conductors have not been securely
PV system should not remain energized after the fastened. There will be some connector pairs that
PVRSS system has been initiated and the facility are mated from different manufacturers that
main service disconnect has been opened. get past the inspections and may create failures
AC PV module or microinverter system. In sooner rather than later. Module solder bonds
the case of an ac PV module system or a micro- have failed in the past, and with numerous new
inverter system, the installer should connect a manufacturers of PV modules on the market, we
digital voltmeter (DVM) to the ac output of the may expect to see such failures in the future.
operating ac PV module array with line voltage The requirements for PV system to have dc
indicated. The installer should then activate the PV arc-fault circuit interrupter (DCPVAFCI)
are similar, but not identical in 2011, 2014, and islanding software/firmware. It will be up to the
2017 Codes. Verification that the inverter, dc PV AHJ to determine that the correct inverter has
combiner or charge controller has the DCPVAF- been installed based on the local requirements. It
CI will be accomplished by the markings on will not be possible on the typical PV inverter to
the inverter, dc combiner or possibly the charge easily test these characteristics in the field, so the
controller. At this point, it is unclear how the markings and the certification/listing mark will
UL Standard 1699B will deal with products that have to suffice.
must meet the differing requirements of various
Code editions. It is possible that an installer Summary
controlled software function will allow a single Photovoltaic power systems are proliferating
inverter to meet the varying requirements of and are going to be a major player in our nation’s
several editions of the NEC. energy mix. They are going to become more com-
plex. The inspector community must be able and
Ride-Through PV Inverters willing to inspect these systems as the last line
Various states and utilities are starting to require of defense in ensuring the safety of the public.
the use of utility-interactive inverters that will The plan review stage is becomisng increasingly
stay on line and produce full available array important due to the increasing complexity of the
power during utility voltage and frequency systems and the number of not-fully-qualified
variations that exceed the normal anti-islanding individuals and organizations installing these
limits of +10% to -12% on voltage and +0.5 Hz systems.
to -0.7 Hz on frequency. The voltage variations Those of us in the inspection community and
may be as low as 50% of the normal line voltage. those associated with the inspection community
These inverters will also be required to come including inspectors at all levels, chief inspectors,
back on line at full array power within one (1) building officials, and administrators responsible
second after the utility voltage and frequency for funding need to work diligently to increase
have returned to either the normal or widened the competency and quality of our inspection
limits. This is considerably different from the process and our inspection force. Additional
current five (5) minute delay in powering up time must be made available to inspect these
after a utility outage. There will be certified/listed systems. Additional time and funding must be
utility interactive inverters on the market that will made available to educate and train the inspector
meet these new requirements and they will be workforce, and above all, those inspectors must be
well marked as such. However, the great majority fully supported when they make the tough calls
of the available inverters will have the current anti on code violations.
178 Chapter 10 The 15-Minute PV Inspection – Can You? Should You?
Chapter 10 The 15-Minute PV Inspection — Can You? Should You? 179
10
The 15-Minute PV Inspection–Can
You? Should You
In some jurisdictions, inspectors have as little verters, disconnects and overcurrent devices) that
as 15 minutes to make a residential electrical has been installed is the same as previous systems
inspection. A common question is, “Can I inspect and those previous systems have been inspected
a residential PV system in 15 minutes?” This and no significant issues were found.
chapter will examine that question and take up Here are some of the items that an inspector
the question, “Should only 15 minutes be allocat- should verify during the site visit. They are listed
ed for inspecting a residential PV system?” in order of importance and in order of safety for
Let’s start with an ideal situation. The inspec- the inspector. For a more complete checklist,
tor is familiar with PV systems in general and see the Inspector/Installer Checklist found in
has inspected quite a few. He or she receives an Appendix A.
application for a permit for a PV system, and that
application is accompanied by all the material Grounding
outlined in the preceding chapters in this book. Proper grounding of the PV system is extremely
A plan review of the supplied material shows important because the PV modules will be gen-
no major problems in code-compliance, and the erating hazardous amounts of energy for the next
installer quickly rectifies the few minor problem fifty years or more. Proper grounding is the first,
areas found. A team consisting of a PV vendor/ the last, and the most important area (in the au-
installer with a history of good PV installations thor’s experience) that requires code-compliance
employing or working with an electrical contrac- in a PV system. Proper grounding of all exposed
tor/electrician who has a commercial electrical metal surfaces that may become energized as the
license and some PV experience has done the system ages or as accidents happen will provide
design of the system and the installation. The the highest levels of protection against shock
system is nearly identical to others that this or- and fires. Proper grounding will also facilitate
ganization has previously installed and have been the action of the ground-fault protection and
inspected with no significant issues. The system arc-fault protection equipment that these systems
configuration and the equipment (modules, in- will have. As the inspector moves through the
As PV systems mature and UL

standards and the Code evolve, it
is hoped that the grounding of PV
systems will become more robust.
AC Point of
Connection to the
Utility
While the premises ac load center
is open to check the grounding
connection, the location and value
of the backfed PV circuit breaker
can be noted. It should match the
Photo 10.1 • Improperly wired 240-volt inverter; neutral wired to PE value on the permit application
terminal and shall not be greater than 20%
of the load center rating. It should
PV system, grounding will be a critical inspection be located at the opposite end of the busbar
item in several locations. from the utility input. This assumes that the
Most smaller PV systems (below 10–20 kW) main breaker and the load center have the same
may have all the PV equipment, both ac and rating. See NEC 705.12(B)(2). This requirement
dc, grounded by a single “grounding” conductor limits the backfed PV breaker to a maximum
connected from the modules to the ground- of 20-amps on a 100-amp load center and to
ing bus bar in the existing ac load center. The a maximum of 40 amps on a 200-amp panel.
module frames, the PV array mounting rack, Breakers larger than this indicate that the utility
and the dc disconnect are connected with a dc connection should have been made on the supply
equipment grounding conductor that connects side of the service disconnect. See Chapter 7 for
to the inverter. From the inverter through one or more details.
more ac disconnects, an ac equipment ground-
ing conductor (which may also serve as any PV Inverters
system functional ground) continues the connec- The inverter should be opened to check the
tion to the ground bus bar in the ac load center field-installed connections. Some inverters will re-
[690.47(A)]. quire metric hex socket drivers (or Allen wrenches)
The first item an inspector should verify is that to open. One manufacturer makes a sealed inverter
the ac equipment grounding conductor from the with permanently attached cables for connections
PV system inverter has been connected properly to the adjacent ac and dc disconnects.
in the ac load center grounding bus bar and that Inverters with a 120-volt output should have
the ac load center has a proper connection to line (ungrounded), neutral (grounded), and equip-
ground (earthed). If this equipment grounding ment grounding conductors between the load
has not been done properly, a ground fault in the center and the inverter. Inverters made outside the
PV array or elsewhere in the system may put U.S. may have the equipment grounding terminals
several hundred volts (with respect to the ground marked PE for “protective earth.” Some 240-volt
where the AHJ is standing) on the ungrounded inverters have only line 1, line 2, and equipment
exposed metal surfaces of any PV equipment. grounding conductors with no neutral (grounded)
terminal, because such a connection would make

a double bond between ground and the neutral
conductor that is prohibited by NEC Section
250.6 (photo 10.1).
The dc input connections to the inverter may
include one or more sets of positive and negative
conductors as well as at least one dc equipment
grounding conductor routed to either an external
dc disconnect or to the PV array (photo 10.2).
AC and DC Disconnects
Each disconnect should be properly grounded
with the equipment grounding conductors or
metal raceways. Following and verifying the
equipment grounding conductors backwards
Photo 10.2 • Inverter with two dc inputs (lower left). EGC from the ac load center through the system to
on lower right. Improper double lugging for small surge the PV modules is important to ensure that each
suppressor conductors. Color coding based on 2014 NEC.
exposed metal surface that may be energized is
grounded. Grounding using sheet metal screws
is prohibited by the Code and the use of thread
cutting screws and aluminum lugs is questionable
(photo 10.3). Most listed fused disconnects and
circuit breaker enclosures have ground-bar kits
with specific mounting instructions and locations
that should be used to maintain the listings of
the devices and to provide the highest quality
grounding connection (photo 10.4). Metal
conduits with dc circuits operating over 250 volts
will usually require grounding/bonding bushings
(250.97).
While any of the disconnect or isolation device
enclosures are opened, the color coding of the
conductors should be checked. PV systems in-
Photo 10.3 • Improperly grounded dc disconnect; violates stalled under the 2017 NEC will typically not have
250.8, 110.3(3), 250.96(A), and 250.4(A)(5) any of the dc PV array conductors grounded and
no white conductors should be seen. The exception
conductor, while others will have line 1, line 2, would be the small solidly grounded PV array
neutral, and equipment grounding conductors. outside a building as allowed by 690.41.
The inverter manual (submitted with the permit There is no specified color code for the un-
request) will show the proper connections. Invert- grounded conductors, and any color is permitted
ers requiring no neutral connection must not have as long as gray, white, green, and green and yellow
the neutral terminal in the utility circuit attached are not used. In conduit, conductors with colored
to anything, particularly an equipment grounding insulation can be used for polarity and circuit
Photo 10.4 • Properly installed, listed ground-bar kit Photo 10.5 • Readily accessible module wiring properly
secured and guarded.
identification. However, colored insulations on
exposed wiring may not prove durable over the life most residential PV systems will more closely
of the system due to a lack of carbon black in the resemble the equipment and workmanship on
insulation resulting in a reduction in the ultraviolet a commercial electrical installation than those
(UV ) radiation resistance. Black insulation has items in a residential electrical system. There
proven to be the most durable. will usually be surface-mounted disconnects and
Both circuit conductors (positive and negative) much of the wiring will be in exposed, sur-
should be routed through the disconnect enclo- face-mounted conduit.
sure even when only the ungrounded conductor is The installer should have a ladder on-site the
switched. Avoiding a “switch loop” configuration day of the inspection to facilitate examining
ensures that both circuit conductors are always the installed PV array. A quick look at the PV
in close proximity (lowering circuit inductance) array on the roof should verify that any exposed
for best functioning of overcurrent devices and to wiring is firmly secured to the PV modules or
allow a bolted connection point for the grounded the mounting structure and is not dangling down
conductor on an isolated “neutral bus” in the where it would be subject to physical damage
enclosure, if required. (photo 10.5).
In the “switch-type” dc PV disconnect, the al- If the backs of the PV modules can be closely
ways “hot” conductors from the PV array wiring observed, proper grounding of the modules
source or output circuits should be connected should be checked. The hardware supplied by
to the top (protected) “Line,” terminals on the the module manufacturer should have been used
switch while the lower, exposed, “Load” terminals as shown in the instruction manual delivered
should be connected to the inverter. On any ac with the permit application. Each PV module
switched disconnect, the upper “Line” terminals must be grounded, and if exposed, single-con-
should be connected to the utility power conduc- ductor cables touch the mounting racks or a
tors that come from the backfed ac load center or metal roof, those objects should also be ground-
a supply-side utility connection. The lower “Load” ed. See Chapter 3 for more details on module
terminals should be connected to the inverter. grounding.
The conductors used for module interconnec-
Workmanship and the Roof tions should be as specified in the permit appli-
The equipment used and the workmanship on cation with respect to size (AWG), insulation
type, and temperature rating. Any PV combiners equipment being installed, or the installer would
containing overcurrent devices exposed to sun- normally dictate that the inspection takes more
light should be noted and the plans and technical time. How much? Some residential PV inspections
data reviewed to determine if adequate tempera- for new inspectors are somewhat of a training ses-
ture deratings were applied. Conduits in sunlight sion and with a knowledgeable installer, examining
will also be exposed to higher-than-ambient and discussing all the details relating to a durable,
temperatures. safe (for 50-years) installation might take two or
more hours.
Inspect in 15 Minutes?
Yes, it might be possible to perform the above Should We Do 15-Minute
inspections in 15 minutes if the inspector has Inspections?
spent some time at the plan-check stage and is See the little girl in photo 10.6? That PV system
experienced in PV systems employing this inverter she is touching will still be producing power
and the installer is there to answer questions, open when her grandchildren are her age. It will take
the inverter and other equipment as necessary and more than a 15-minute inspection to ensure that
to provide a ladder for roof access. However, any the PV system will be as safe then as it is now.
problems found in the above areas should warrant Fifteen minutes is probably insufficient time to
a closer look at the entire system and when more ensure the public safety of a system that may
details are examined, the inspection time can grow. operate, with possibly little attention, over a
A lack of familiarity with either PV in general, the 40–50-year period.
Photo 10.6 • A 40 kW PV system. Safe now, but will it still be safe in 40 years” Courtesy James Worden
184 Appendix A 2014 and 2017 NEC Photovoltaic Electrical Power Systems Inspector/Installer Checklist
Appendix A 2014 and 2017 NEC Photovoltaic Electrical Power Systems Inspector/Installer Checklist 185
A
2014 and 2017 NEC Photovoltaic
Electrical Power Systems
Inspector/Installer Checklist
The following checklist is an outline of the CHECKLIST FOR
general requirements found in the 2014-2017 PHOTOVOLTAIC POWER
National Electrical Code (NEC) in Articles SYSTEM INSTALLATIONS
690 and 705 that deal with Photovoltaic (PV)
Power Systems installations. 1. PV ARRAYS
The checklist is only a guide and applies to □ PV modules listed to UL Standard 1703?
any component used or installed in a PV system [110.3, 690.4(B)]
other than devices inside a listed, factory-as-
sembled component. a. Mechanical Attachment
The local authority having jurisdiction
(AHJ) or inspector has the final say on what
□ Modules attached to the mounting struc-
is or is not acceptable. Local codes may mod-
ture per the manufacturer’s instructions?
ify the requirements of the NEC.
[110.3(B)]
This list should be used in conjunction with
Article 690, Article 705 and other applicable
□ Roof penetrations secure and weather
articles of the NEC and includes inspection
tight? (110.12, 110.13)
requirements for both stand-alone PV systems
(with and without batteries) and utility-inter-
b. Grounding
active PV systems. Where Article 690 or 705
differ from other articles of the NEC, Article □ Each module grounded using the supplied
690 or 705 takes precedence. (690.3, 705.3) hardware, the grounding point identified
NOTE: The 2017 NEC has many detailed on the module and the manufacturer’s
changes in Art 690 and 705 with older mate- instructions? Note: Bolting the module
rial being placed in new articles and numerous to a “grounded” structure usually will not
small revisions of the remaining material. Code meet NEC requirements [110.3(B)] and
references will generally be to the 2017 NEC. may not comply with the instructions
for grounding the PV module. Array PV Strain reliefs/cable clamps or conduit used
mounting racks are usually not identified on all cables and cords? (300.4, 400.10)
as equipment-grounding conductors, unless
certified/listed to UL Standard 2703. □ Listed for the application and the environ-
(690.43) Module instruction manual must ment? Fine stranded, flexible conductor
specifically show/indicate grounding and conductors properly terminated with ter-
mounting method. minals listed for such conductors? (110.14)
□ Properly sized equipment-grounding con- □ Cables and flexible conduits installed and
ductors routed with the circuit conduc- properly marked? (690.31)
tors? (690.45)
□ Exposed conductors in readily acces-
c. Conductors sible areas in a raceway or guarded
if over 30 volts? [690.31(A)] Note:
□ Conductor type?—If exposed: USE-2 or Raceways cannot be connected to
PV wire for grounded PV arrays and PV most modules. Conductors should be
wire for ungrounded PV arrays. All PV installed so that they are not readily
modules will use PV wire. 2017 NEC al- accessible (i.e., guarded).
lows USE-2 or PV wire for both grounded
and ungrounded systems. 2. OVERCURRENT
PROTECTION
□ Conductor insulation rated at 90°C (UL-
1703) to allow for operation at 70°C+ near
modules and in conduit or cables exposed
□ Overcurrent devices in the dc circuits list-
ed for dc operation? If device not marked
to sunlight? [Table 310.15(B)(3)(c)]
dc, verify dc listing with manufacturer.
Auto, marine, and telecom devices are
□ Temperature-corrected ampacity calculations not acceptable.
based on 125% of short-circuit current (Isc)
or the 156% Isc without conditions of use
(take the worst case)? □ In PV circuits, OCPD must be listed as
PV device [690.9(B)].
Note: Suggest temperature derating factors
of 65°C for conductors behind modules in □ Rated at 1.25 x 1.25 = 1.56 times
installations where the backs of the module short-circuit current from modules?
receive cooling air (4" or more from roof ) (UL-1703, 690.8, module instructions).
and 75°C where no cooling air can get to Overcurrent devices listed for PV appli-
the backs of the modules. Ambient tem- cations are required. [690.9(B)].
peratures (near and at the array location)
more than 40°C may require different □ Each module or series string of modules
derating factors. have an overcurrent device protecting the
module? [UL-1703 / NEC 110.3(B)]. Note:
□ Portable power cords allowed only for Frequently, installers ignore this require-
tracker connections? [690.31(C), 400.3,7,8] ment marked on the back of modules.
Listed combiner PV combiner boxes □ Twist-on wire connectors listed for the
meeting this requirement are available. environment (i.e., dry, damp, wet, or direct
One or two strings of modules generally burial) and installed per the manufacturer’s
do not require overcurrent devices, but instructions?
three strings or more in parallel will
usually require an overcurrent device. The □ Pressure lugs or other terminals listed for
module maximum series fuse must be at the environment? (i.e., inside, outside, wet,
least 1.56 ISC. [690.9(A)] direct burial)
2017 NEC: Only one conductor of an un-

□ Power distribution blocks listed and not just
grounded PV source or PV output circuit
UL Recognized?
is permitted to have an OCPD. If used,
other OCPD in the dc circuits must be in
the same polarity conductor.
□ Terminals containing more than one con-
ductor listed for multiple conductors?
□ Located in a position in the circuit to pro- □ Connectors or terminals using flexible,

tect the module conductors from backfed
fine-stranded conductors listed for use with
currents from parallel module circuits
such conductors? [690.31(H), 690.74(A),
or from the charge controller or battery?
110.14]
[690.9(A)]
□ Locking (tool-required) connectors on read-
□ Smallest conductor used to wire modules ily accessible PV conductors operating over
protected? Sources of overcurrent are 30 volts? [690.33(C)]
parallel connected modules, batteries, and ac
backfeed through inverters. [690.9(A)] 4. CHARGE CONTROLLERS
2017: See Article 706,
□ User-accessible fuses in “touch-safe” holders Energy Storage Systems
or fuses capable of being changed without
touching live contacts? Disconnects from □ Charge controller listed to UL Standard
all sources of voltage in dc combiners at 1741? [690.4(B)] Exposed energized termi-
the inverter? (690.16) nals not readily accessible?
3. ELECTRICAL □ Does a diversion controller have an

CONNECTIONS independent backup control method?
[706.23(B)(1)]
□ Pressure terminals tightened to the recom-
mended torque specification? [110.3(B), 5. DISCONNECTS 2017 NEC:
110.14] Substantial changes
□ Crimp-on terminals listed and installed with □ Rapid shutdown system installed? (690.12)
listed crimping tools by the same manu- Applied to inverter inputs, module out-
facturer? [110.3(B)] puts, batteries (where used), and combiner
outputs? Listed equipment is available and protection for entire dc system with possible
the UL Standards addressing the require- exception of source circuit or module protective
ments are published. Operation verified? fuses.
□ Disconnects listed for dc operation in dc 6. INVERTERS (Stand-‐alone

circuits? Automotive, marine, and telecom Systems)
devices are not acceptable. □ Inverter listed to UL Standard 1741?
[110.3, 690.4(B)] Note: Inverters listed to
□ PV Disconnect readily accessible and located telecommunications, automotive or other
at first point of penetration of PV conductors? standards do not meet NEC requirements.
Location not specified in the 2017 NEC.
□ DC input currents from the battery cal-
□ PV conductors outside structure until culated for conductor and fuse require-
reaching first readily accessible disconnect ments? Input current = rated ac output
unless in metallic raceway? [690.13(A), in watts divided by lowest battery volt-
690.31(F)] Metallic raceway now required age divided by inverter efficiency at that
all the way to the inverter dc input. power level. [690.8(A)(4)]
□ Disconnects for all current-carrying conduc- □ Cables to batteries sized 125% of calculated
tors for the PV system? (690.13) inverter input currents? [690.8(A), 706.20]
2017 NEC: PV System Disconnect must
disconnect all circuit conductors—even □ Overcurrent/Disconnects mounted near
on solidly grounded systems -but solidly batteries and external to PV load centers
grounded conductors should not be opened. if conductors are longer than 4–5 feet to
(Look for a possible change to this require- batteries or inverter?
ment in the 2020 NEC.)
□ High interrupt, listed, dc-rated fuses or
□ Disconnects for equipment? (690.15/690.17) circuit breakers used in battery circuits?
2017 NEC: Equipment isolation discon- AIR/AIC at least 20,000 amps?
nects may disconnect only the ungrounded [706.21 110.9, 110.10]
conductor.
□ No multi-wire branch circuits where single
□ DC combiner has output circuit discon- 120-volt inverters are connected to 120/240-
nect/isolator internal or within 3 m (10 ft)? volt load centers? [Art 100–Branch Circuit,
[690.15(A)] Multi-wire), 710.15]
□ Grounded conductors not fused or switched 7. BATTERIES

(except PV system disconnect in 2017 For 2017 NEC: See Article 706, Energy
NEC)? Storage Systems
Note: Listed PV Power Centers are available □ No separate batteries cells are listed.
for 12, 24, and 48-volt systems and they contain
charge controllers, disconnects, and overcurrent
□ AC Battery systems are generally self-con- 8. INVERTERS (Utility —

tained and will be listed as an assembly. interactive Systems)
□ Building-wire type cables used? (Chapter 3) □ Inverter listed to UL Standard 1741 and
Note: Welding cables, marine, locomotive identified for use in interactive photovol-
(DLO), appliance wire material (AWM) taic power systems? [690.4(B), 705.4].
and auto battery cables don’t meet NEC Note: Inverters listed to telecommuni-
requirements. Flexible, listed RHW, or cations and other standards do not meet
THW cables are available. Article 400 NEC requirements.
flexible cables larger than 2/0 AWG are
OK for battery cell connections, but not in □ Backup charge controller to regulate the
conduit or through walls [690.74, 400.8]. batteries in systems with multimode
Flexible, fine-stranded cables require inverters when the grid fails? [706.23(B)]
limited-availability, specially-listed terminals
[110.14, 690.74]. See stand-alone inverters □ Connected to dedicated branch circuit
for ampacity calculations. with back-fed overcurrent protection?
[705.12(B)] or connected as a supply-side
□ Access limited? [706.30] connection with overcurrent protection
within 10 feet? [705.12(A), 705.31]
□ Installed in well-vented areas (garages, base- □ Listed dc and ac disconnects and overcur-
ments, outbuildings, and not living areas)?
rent protection? (690.15, 17)
Note: Manifolds, power venting, and single
exterior vents to the outside are not re-
□ All requirements of 705.12(B) or
quired and should be avoided [706.10(A)].
705.12(A) met?
□ Have the conductor routing and protection NOTE: Square wave or modified sine wave
requirements of 706.20 and 706.32 been inverters may be listed to UL 1741 but are
met? Cables to inverters, dc load centers, not compatible with many power tool battery
and/or charge controllers in conduit? chargers, smoke alarms, and other listed elec-
tronic devices and should not be used with
□ Conduit enters the battery enclosure these devices. The manufacturer’s instruction
below the tops of the flooded batteries? manual will usually have the warning state-
(300.4) ment [110.3(B)].
Note: There are few listed battery boxes. 9. GROUNDING

Lockable heavy-duty plastic polyethylene tool NOTE: Grounding and Disconnects exten-
boxes are usually acceptable. sively revised in 2017 NEC.
□ Only one bonding conductor (grounded

conductor to ground-only on solidly
grounded PV systems) for dc circuits on
grounded PV arrays and one bonding □ Conductor insulations other than black
conductor for ac circuits (neutral to ground) in color will not be as durable as black
for ac system grounding? in the outdoor UV-rich environment.
Note: The utility-interactive inverter will □ DC color codes correct? They are the same
generally provide the functional ground for the as ac color codes—grounded conductors
system. Instructions for that functional ground are white or gray and equipment-grounding
will be in the inverter installation manual. conductors are green, green/yellow, or bare.
[200.6(A)]
□ System/inverter grounding meets require-
ments of 690.47? □ Ungrounded PV array conductors on
ungrounded PV arrays will not be white
□ Equipment grounding conductors properly in color. Note that functionally grounded
sized (even on ungrounded, low-voltage PV systems under the 2017 NEC will not
systems)? (690.43, 45, 46) have any dc PV source or dc PV output
conductors with white insulation. The
□ Disconnects and overcurrent in both color white will only be used in solidly
ungrounded conductors in each circuit on grounded PV systems (690.41).
12- volt, ungrounded systems or on un-
grounded systems at any voltage? (690.9, 11. Markings
690.13, 690.15, 690.31) Revised for 2017
NEC with functional grounding. □ All field-applied markings correct?
Note that functionally grounded PV [690.13(B), 690.31(B), 690.51, 690.53,
systems under the 2017 NEC will not 690.54, 690.55, 705.10, 705.12]
have any dc PV source or dc PV output
conductors with white insulation. □ Meet color and letter size requirements?
(690.56)
□ Bonding-grounding fittings or bushings
used with metal conduits when dc system 12. DC PV Arc Fault Circuit
voltage is more than 250V dc? (250.97). Protection?
Grounding bonding bushings used where
grounding electrode conductors are in □ Usually installed in the inverter or on larger
metallic raceways and /or enclosures? systems in the array field. May be multiple
devices (690.11). UL Standard 1699B is
10. CONDUCTORS (General) the applicable standard.
□ Standard building-wire cables and wiring 13. PV Rapid Shutdown

methods used? [300.1(A)] Systems?
□ Wet-rated conductors used in conduits in □ Installed per 690.12 and local require-
exposed locations? (100 Definition of ments? Operational?
Location, Wet)
191
B
Bibliography
Molded-Case Circuit Breakers, Molded-Case Standard for Connectors for Use in Photovoltaic
Switches, and Circuit-Breaker Enclosures for Use Systems. UL Standard 6703, 2014.
with Photovoltaic (PV) Systems. UL Standard
489B, 2016. Standard for Flat-Plate Photovoltaic Modules and
Panels. UL Standard 1703, 2012.
National Fire Protection Association. National
Electrical Code® (NEC®) Handbook. Quincy, Standard for Flat-Plate Photovoltaic Modules and
Massachusetts: NFPA, 2017. Panels. UL Standard 1703, 2002.
National Fire Protection Association. National Standard for Industrial Control Equipment. UL
Electrical Code® (NEC®) Handbook. Quincy, Standard 508, 1999.
Massachusetts: NFPA, 2014.
Standard for Inverters, Converters, Controllers and
National Fire Protection Association. NFPA 70: Interconnection System Equipment for Use With
National Electrical Code. Quincy, Massachusetts: Distributed Energy Resources. UL Standard 1741,
NFPA, 2017. 2010.
National Fire Protection Association. NFPA 70: Standard for Mounting Systems, Mounting Devices,
National Electrical Code. Quincy, Massachusetts: Clamping/Retention Devices, and Ground Lugs
NFPA, 2014. for Use with Flat-Plate Photovoltaic Modules and
Panels. UL Standard 2703, 2015.
Photovoltaic (PV) DC Arc-Fault Circuit Protection.
UL Standard 1699B, 2011. Standard for Photovoltaic Wire. UL Standard
4703, 2011.
Solar American Board of Codes and Standards.
Expedited Permit Process Report - Revision 2, last Wiles, John, “Perspectives on PV archives,” IAEI
accessed on 12/13/2017. http://solarabcs.org/ Magazine, 2015-2018.
about/publications/reports/expedited-permit/
index.html.
192

& PV Professionals
Third Edition
Author John Wiles
CEO David Clements
Director of Education L. Keith Lofland
Technical Advisor, Education, Codes, and Standards Joseph Wages, Jr.
Director of Publishing Kathryn P. Ingley
Publications Manager Bryan Nyary
Creative Director John Watson
Technical Reviewer James Rogers

Photovoltaic
Power Systems
& PV Professionals
In a time of rapid advancement in photovoltaic power systems, this third edition of
John Wiles’ acclaimed book serves as a comprehensive manual for inspectors, plan
reviewers, and installers to ensure National Electrical Code-compliant PV system
installations. Updated for the 2017 NEC, this extensive guide covers everything plan
reviewers, installers, and inspectors need to know about these systems.
John Wiles is perhaps the most recognized and influential name in the solar indus-
try. He’s worked extensively in the development of the NEC and UL Standards and
is an active trainer on Code-compliant PV systems. Wiles has written hundreds of
articles on Code-related photovoltaic system topics and continues to write Perspec-
tives on PV articles for IAEI News.
Chapters include:
· PV fundamentals and calculations · Disconnects

· Inverters
PV module installation considerations · Overcurrent protection
· Energy storage systems · Utility interconnections
· Grounding · PV system inspections
Plan checking
· ·
International Association
of Electrical Inspectors
901 Waterfall Way, Suite 602,
Richardson, TX 75080-7702
v.2

Photovoltaic Power Systems 2017

Uploaded by

Copyright:

Available Formats

Photovoltaic Power Systems 2017

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Photovoltaic Power Systems 2017

Uploaded by

Copyright:

Available Formats

What is the book about?

What is the book about?

Who is the author of the book?

Who is the author of the book?

Photovoltaic Power Systems

For Inspectors, Plan Reviewers

BASED ON THE 2017 NATIONAL ELECTRICAL CODE

BASED ON THE 2017 NATIONAL ELECTRICAL CODE

Published 2012, 2014, 2018 by

Copyright © 2012, 2014, 2018 by John Wiles

All rights reserved. First Edition published June 2012.

Photos used in this book were shot in situ or at tradeshows.

1 An Overview of PV Systems and the 2017 National Electrical Code 8

2 PV Fundamentals and Calculations 42

3 PV Modules — Installation Considerations 65

4 The Inverter — Operation and Connections 81

5 Energy Storage Systems (ESS), Batteries in PV Systems 97

6 Grounding, Disconnects, and Overcurrent Protection 113

7 Utility Interconnections 133

8 Plan Checking and Inspecting 156

9 The Process of Inspecting PV Systems 168

10 The 15-Minute PV Inspection – Can You? Should You? 178

Appendix A 2014 and 2017 NEC Photovoltaic Electrical Power Systems

Appendix B Bibliography 191

— Rhonda Parkhurst, Electrical Specialist

Photo 1.4 • Five mW utility-scale PV system with PV subar-

defined and further described in the following Utility-Interactive Systems

Photo 1.7 • 2.2-MW Sunny Central Inverter. Courtesy of

er will usually not include the Section 690.41

PV and the 2017 National

that does not involve energy storage or another

Figure 1.1 • Disconnects

Grounding—New and Old

Figure 1.2 • Multimode, utility-

Figure 1.3 • Stand-alone system with energy storage (FIG 690.1).

From Article 690

Photo 1.18 • (top) DC combiners may look similar for many

Photo 1.19 • (left) The DC PV system disconnect or pos-

conductor is opened by a PV system disconnect

Article 710, Stand-Alone Systems

Article 712, Direct Current Microgrids

of such vehicles and in locations throughout the

systems to control the process. These systems, wind

The Smart Grid

impulse-buying prices and are being installed by The Environment

110.4]. Also, the torque tools can be used to

The AC Utility Connection

device. If a meter adapter were used,

Ambient Temperature output of a PV module or array that is not

Figure 2.1 • I-V curve for a PV module.

Figure 2.2 • I-V and power curve.

Figure 2.3 • I-V curve variations due to varying sunlight.

Figure 2.4 • I-V curve variations due to varying temperature.

Early pioneers in PV, including PV mod- Adjustments—Short-Circuit

Millivolt Coefficients 8 x 78.2 V = 625.6 V

PV Math—Module Short-Circuit near the peak-power voltage (also called the

Vmpp = 450V data when it is available. Normal cell operating

5.5 x 0.496 = 2.728 volts. Module-Level Power

backed-up utility-interactive system, the currents Step 1.

Example 1: Three conductors are in a conduit Step 3. Terminal temperature limits[110.14(C)]