DC Power-System-Design-for-Telecommunications PDF

DC POWER SYSTEM
DESIGN FOR
TELECOMMUNICATIONS
Whitham D. Reeve
IEEE Press Telecommunications Handbook Series

Whitham D. Reeve, Series Editor
IEEE PRESS
A JOHN WILEY & SONS, INC., PUBLICATION

DC POWER SYSTEM
DESIGN FOR
TELECOMMUNICATIONS
IEEE Press
445 Hoes Lane
Piscataway, NJ 08854
IEEE Press Editorial Board

Mohamed E. El-Hawary, Editor in Chief
J. B. Anderson R. J. Herrick F. M. B. Pereira

R. J. Baker S. V. Kartalopoulos N. Schulz
J. F. Brewer M. S. Newman C. Singh
T. G. Croda G. Zobrist
Kenneth Moore, Director of IEEE Book and Information Services (BIS)

Catherine Faduska, Senior Acquisitions Editor
Jeanne Audino, Project Editor
Technical Reviewers
Thomas G. Croda
Marco Migliaro
Michael S. Newman
Percy Pool
Roy E. Thompson
DC POWER SYSTEM
DESIGN FOR
TELECOMMUNICATIONS
Whitham D. Reeve
IEEE Press Telecommunications Handbook Series

Whitham D. Reeve, Series Editor
IEEE PRESS
A JOHN WILEY & SONS, INC., PUBLICATION

Unless indicated otherwise, all illustrations are owned and copyrighted by the author.
Copyright © 2007 by the Institute of Electrical and Electronics Engineers, Inc. All rights reserved.
Published by John Wiley & Sons, Inc. Published simultaneously in Canada.
No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by
any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under
Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the
Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center,
Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at
www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions
Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-
6008, or online at http://www.wiley.com/go/permission.
Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in
preparing this book, they make no representations or warranties with respect to the accuracy or completeness of
the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a
particular purpose. No warranty may be created or extended by sales representatives or written sales materials.
The advice and strategies contained herein may not be suitable for your situation. You should consult with a
professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any
other commercial damages, including but not limited to special, incidental, consequential, or other damages.
For general information on our other products and services or for technical support, please contact our
Customer Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-
3993 or fax (317) 572-4002.
Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be
available in electronic format. For information about Wiley products, visit our web site at www.wiley.com.
Library of Congress Cataloging-in-Publication Data is available.
ISBN-13 978-0-471-68161-8
ISBN-10 0-471-68161-X
Printed in the United States of America.
10 9 8 7 6 5 4 3 2 1
CONTENTS
Preface ix
1 Introduction 1
1.1 Basic Requirements for Telecommunications Power Systems 1
1.2 Applications Review 1
1.3 Direct Current Power System Elements 5
1.4 Power Sources and Loads 12
1.5 General Design Considerations 18
1.6 Standards, Practices, and Codes 24
2 Electricity Review 31
2.1 Electromotive Force 31
2.2 Current Flow 31
2.3 Voltage Sources 32
2.4 Resistance 33
2.5 Wire and Rigid Busbar Conductors 39
2.6 Ohm’s Law and Electrical Circuits 44
2.7 Electrical Power 49
2.8 Current Density 52
2.9 Kirchhoff’s Current and Voltage Laws 55
2.10 Alternating Current 66
References 83
v
vi CONTENTS
3 dc Power System Components 85

3.1 Nominal System and Operating Voltages 85
3.2 Noise and Transients in dc Power Systems 86
3.3 Rectifier Systems 91
3.4 Powerboard 109
3.5 Monitoring and Control 117
3.6 Primary and Secondary Distribution Systems 122
3.7 Voltage Conversion 135
References 148
4 Telecommunications Batteries 151

4.1 Batteries and Cells 151
4.2 Battery Voltage and Discharge Curves 165
4.3 Temperature Performance 174
4.4 Battery Life 176
4.5 Charge Acceptance and Thermal Runaway 179
4.6 Cell Impedance and Conductance 183
4.7 Short-Circuit Currents 184
4.8 Cell Gassing and Ventilation 187
References 187
Bibliography 188
5 System Design 189

Part I Direct Current Power Systems Component Design 189
5.1 System Design Considerations 189
5.2 Battery System Design 207
5.3 Rectifier System Design 223
5.4 Powerboard Design 233
5.5 Secondary Distribution Design 243
5.6 Voltage Conversion Systems Design 243
5.7 Other Design Considerations 249
References (Part I) 262
Part II Direct Current Power Systems Circuit Design 263
5.8 Circuit Design 263
5.9 Direct Current Power System Bonding and Grounding 367
References (Part II) 378
6 System Installation and Maintenance 381

Part I Direct Current Power System Installation 381
6.1 Installation Safety Guidelines 381
6.2 Material Compatibility 382
6.3 Circuit Wiring 382
6.4 Terminal Connector Lugs 385
6.5 Fasteners 391
6.6 Power Wiring Installation 397
CONTENTS vii
6.7 Power Equipment Frames 411

6.8 Auxiliary Framing 412
6.9 Cable Racks 415
6.10 Sleeves, Wall Openings, and Floor Openings 417
6.11 Basic Installation Sequence 418
6.12 Battery System Installation 419
References (Part I) 425
Part II System Maintenance 426
6.13 Overall Maintenance Requirements 426
6.14 Annual System Maintenance Check 428
6.15 Rectifier System Maintenance 429
6.16 Inverter System Maintenance 432
Part III Battery System Maintenance 433
6.17 General Requirements 433
6.18 Vented Lead–Acid Batteries 449
6.19 Valve-Regulated Lead–Acid Batteries 469
6.20 Battery Failure Case Studies 476
References (Part III) 481
Appendix A Comments on Historic dc Power System Configurations 487
Appendix B Mathcad Worksheet Examples 491
Index 495
PREFACE
This is a book on the design of dc power systems that operate at nominal voltages of 24
and 48 V direct current (dc) and use lead–acid batteries and are used in public network
telecommunications systems and other exclusive-use environments. Other voltages have
been used over the years, but modern telecommunications systems in the United States
use only these two voltages.
This is not a book on the theory of electrical power production, transmission, and dis-
tribution. Further, this book does not cover power-limited dc systems or components used
with line-powered digital transmission technologies, such as the high bit-rate digital sub-
scriber line (HDSL), T1-Carrier, and some proprietary line-powered access network
equipment, nor does it cover older dc power technologies, such as counter emf (electro-
motive force) cells and end-cells (although these are briefly discussed in Appendix A).
Further, this book does not cover audio tone and ringing systems (a traditional component
in older telecommunications power systems), 130-Vdc systems, uninterruptible power
systems (UPS), battery technologies other than lead–acid, or alternating current generat-
ing systems.
This volume was written for practicing telecommunications engineers and technicians
because of their stated need for a book that concisely and systematically describes how to
design telecommunications dc power systems. No work of this type exists in any lan-
guage, so far as I am aware.
I started writing this book in the fall of 1973 when I first started working in the
telecommunications industry; it took more than 30 years to write. The company I worked
for at the time had neglected all of its dc power systems, and they all were badly over-
loaded, some to a frightening extent (busbars too hot to touch and warped so badly that
wooden blocks had been strapped between opposite polarities to prevent a short). As a
ix
x PREFACE
central office engineering manager, I suggested that “we need to do something,” and I
was told to “take care of it.” And so I started my education, which continues to this day.
Most of it has been the hard way.
It is important for the reader to know that dc power system design, as with all types of
design, does not so much involve right ways or wrong ways but involves systematic
analyses to make the most correct choices based on a number of economical, operational,
electrical, and physical constraints. Just about all statements of practice and rules of
thumb in this book have exceptions, and the reader is strongly cautioned to use common
sense and good engineering judgment.
In preparing this book, I have kept two points in mind: first, to make statements as
clear as possible, considering the technical nature of the subject; and, second, to present
the facts accurately and as fully as necessary. I have sought to use plain language so that
readers can obtain a clear understanding, but it is impossible to discuss any technical mat-
ter without using terms peculiar to it. Definitions of terms used in this book are traceable
to industry standards and dictionaries and industry practice.
I have made some assumptions as to the reader’s previous knowledge and experience.
First, although I have tried to minimize the amount of mathematics, readers must be com-
fortable using algebra, which is the use of symbols and letters to state a generalized solu-
tion and then the substitution of specific numbers in place of the letters and symbols to
solve a particular problem. Second, I have avoided the development of most concepts
from so-called first principles. I do not believe such development is necessary to the un-
derstanding and design of low-voltage dc power systems. However, after introducing and
briefly describing the basic components of dc power systems in Chapter 1, I have laid
down the basic principles of electricity in Chapter 2. More experienced readers may find
Chapter 2 unnecessary, although those same readers may find the information interesting
and worth reviewing. The remaining chapters describe the power system components
(Chapter 3), battery systems (Chapter 4), system design (Chapter 5), and system installa-
tion and maintenance (Chapter 6).
Several reviewers have provided very helpful suggestions and criticism. I am especial-
ly grateful to Marco Migliaro, Percy Pool, and Roy Thompson, who carefully went over
the manuscript and made many important suggestions that greatly improved the general
presentation.
This book, as it is presented in this first edition, is far from being perfect. However, it is
a good start, and I believe that the demand is sufficient to warrant its immediate publication.
But it is my purpose to work and ask others to work at its improvement and elaboration.
I ask for private communications from readers stating where they have found the writ-
ing unclear, or suggesting information not contained in it (I already have a list for a future
edition). Such suggestions will be a valuable aid to future enlargement and revision. You
can contact me at w.reeve@ieee.org.
Finally, although the focus of this book is telecommunications systems, much of the
material also applies to low-voltage dc power systems used in other industries.
WHITHAM D. REEVE
Anchorage, Alaska
September 2006
CHAPTER 1
INTRODUCTION
This chapter provides a brief description of requirements and the basic elements and com-
ponents that make up telecommunications direct current (dc) power systems, including
their associated alternating current (ac) power sources. Also included in this chapter are
brief descriptions of system design considerations. Design considerations are described in
greater detail in subsequent chapters.
1.1 BASIC REQUIREMENTS FOR TELECOMMUNICATIONS

POWER SYSTEMS
The basic requirements for a telecommunications power system are listed in Table 1.1.
Designing to achieve these requirements entails the analysis of many details and consider-
ations (Table 1.2).
1.2 APPLICATIONS REVIEW
Telecommunications dc power systems and modern network equipment in the United

States most often operate at nominal voltages of 24 or 48 volts dc (Vdc). Other voltages
Table 1.1 Basic Requirements for Telecommunications Power Systems

Requirement Achieved by
앫 Provide uninterruptible power to 앫 Using batteries to bridge interruptions
critical loads
앫 Be safe to workers and the public 앫 Using relatively low voltages, current limiting and
grounding
앫 Have long life (20–30 years, or more) 앫 Using conservative design and planned preventive
maintenance
dc Power System Design for Telecommunications. By Whitham D. Reeve 1

Copyright © 2007 the Institute of Electrical and Electronics Engineers, Inc.
2 INTRODUCTION
Table 1.2 Detailed Requirements and Considerations

Detailed Requirement Design Consideration
Voltage tolerance range of the load 앫 Requirements for minimum and maximum
equipment equipment operating voltages
앫 Circuit conductor voltage drops
앫 Battery float, equalize, and final discharge voltages
Control of noise on the supplied voltage 앫 Voltage and current ripple limits on rectifier output
circuits
앫 Conducted emissions to and from load equipmenta
앫 Proper grounding and bonding
앫 Power filters near load equipment
Monitoring and signaling 앫 Alarm thresholds
앫 Signal alarms when performance thresholds have
been exceeded
앫 System control and metering
Protecting and limiting of electrical 앫 Equipment malfunctions and circuit faults
circuits 앫 Overcurrent protective device ratings
앫 Conductor current rating (ampacity)
System operation and maintenance 앫 Skill and training levels for daily operation and
routine repair
앫 Ease of repair and routine maintenance
Expansion and extension 앫 Growth and changes during the life of the system
앫 Minimizing or eliminating service outages
Capital investment and operating and 앫 Engineering economy
maintenance costs 앫 Budget constraints and requirements
Floor space and structural requirements 앫 Building modifications and strengthening before and
after initial system installation
앫 Floor and wall penetrations for cabling
Specifications and regulations 앫 Compliance with industry standards, codes, and
company practices
앫 Listing with independent testing laboratories
Prime and standby power sources 앫 Long-term requirements
앫 Reliability and availability
앫 Operating costs
앫 Electrical distribution
앫 Voltage and current ratings
앫 Standby system fuel type and prime mover
technology
a
Conducted emissions are noise currents that follow a conductive path into or out of the equipment, for example,
via power or signal leads.
have been used over the years to meet specific system requirements, but they are not used
in modern systems and are beyond the scope of this book. Also beyond the scope of this
book are power-limited and current-limited higher voltage systems used to power access
network equipment and subscriber terminals such as optical network nodes, high bit-rate
digital subscriber line (HDSL) technologies and other line-powered transmission technolo-
gies. Telecommunications power systems are used in many types of facilities and locations,
1.2 APPLICATIONS REVIEW 3
Fig. 1.1 Power systems applications in the telecommunications network.
including core network central offices, access nodes, user premises, and where network op-
erators are collocated (Fig. 1.1).
Central offices, also called wire centers, contain core telecommunication facilities in-
cluding end office, access tandem, and transit switching systems, mobile switching cen-
ters, packet switches and routers, fiber-optic and other digital landline transmission sys-
tem terminals, terrestrial microwave radio terminals, and satellite earth stations (Fig. 1.2).
Core network facilities provide user (subscriber) features and services.
Access networks deliver the services and features from the core to the user and include
landline facilities, such as metallic twisted pairs and optical fibers, fixed and mobile wire-
less base stations, and, in some cases, satellite earth terminals dedicated to individual
users. Access nodes are important in any modern access network. They include transmis-
sion and multiplexing facilities, such as digital loop carrier (DLC) systems, optical net-
work units (ONU), optical network terminations (ONT), and other electronic systems de-
ployed close to the user.
All but the simplest telecommunications terminals require on-site power systems for
the facilities that are located on user premises. These power systems can be low-current
systems (fractions or a few amperes at 24 or 48 Vdc) or complex systems equivalent to
those used in the core network. The so-called lifeline POTS (plain old telephone service)
lines generally are line powered over the metallic twisted pair plant from the central office
or serving access node. In some modern technologies, such as voice-over-packet [e.g.,
voice-over-asynchronous transfer mode (VoATM) and voice-over-Internet protocol
(VoIP)], the network equipment located on the user premises requires more power than
can be delivered over the access network. In those cases, a local power supply including
battery is required to ensure that regular telephone service is available during commercial
power outages at the premises.
Collocation of telecommunications core and access network equipment owned by dif-
ferent network operators is another application for dc power systems. In some cases, such
as a “carrier hotel,”1 space may be provided for a number of different network operators,
each with its own power system. In other cases, a single large centralized power system
provides power to numerous network operators.
1
A “carrier hotel” is a euphemism used to describe a building that is occupied by a number of different and
sometimes competing network operators. These network operators, including competitive local exchange carri-
ers (CLEC) generally differ from the incumbent local exchange carrier (ILEC) serving the area in that the
CLECs do not have extensive access network infrastructure.
4 INTRODUCTION
Fig. 1.2 Applications—central office buildings (wire centers) (1 and 2), switching systems (3 and 4),
mobile wireless (5), satellite earth stations (6 and 7), terrestrial microwave radio (8), access mode (9).
1.3 DIRECT CURRENT POWER SYSTEM ELEMENTS 5
1.3 DIRECT CURRENT POWER SYSTEM ELEMENTS
A telecommunications dc power system consists of seven basic elements, shown schemat-

ically in Figure 1.3:
앫 Rectifier (charger) system

앫 Battery system
앫 Charge bus
앫 Discharge bus
앫 Primary distribution system
앫 Secondary distribution system
앫 Voltage conversion
The following brief descriptions are keyed to the more detailed block diagram in Fig-
ure 1.4. All components are described in greater detail in Chapter 3 (dc Power System
Components). The design and selection of the various system components are described
in Chapter 5 (System Design).
①1 Rectifiers (also called chargers) convert the prime power source ac voltage to di-
rect current (dc). Rectifiers serve three main purposes:
a. Power the loads when commercial ac power is available.

b. Supply float charge to the battery to overcome battery internal losses.
c. Recharge the battery upon restoration of commercial power after failure while si-
multaneously supplying the normal equipment load.
②
2 Battery, which is an energy storage device, powers the loads during prime power
ac outages (Fig. 1.5). The battery always is connected directly to the discharge bus (de-
Commercial ac
Power Input
Fig. 1.3 Simplified schematic of a telecommunications dc power system.

6
Inverted ac
Distribution
Primary dc
Distribution
Converted
dc
Distribution
ac Power
Supply
Switchboards
and
Distribution
Inverted ac
Distribution
Primary dc
Distribution
Converted
dc
Distribution
Fig. 1.4 Basic elements of a telecommunications dc power system.

Fig. 1.5 Battery system (battery and battery disconnect switches are in the left half and the recti-
fiers and powerboards are in the right half of the picture). (Photo courtesy of Power-One Inc.)
fined below) so there is no switchover time or interruption when the prime power source
(and standby source, if equipped) fails or if the rectifier system fails.
③
3 Charge bus provides a centralized location for connecting rectifiers to the battery.
The charge bus carries equipment load current and float charge current for the batteries
during normal operation. When the ac power source fails, the charge bus carries battery
discharge current to the discharge bus. Upon restoration of the ac power source, the
charge bus carries equipment load current and battery recharge current. A separate busbar
is provided for the positive (+) and negative (–) outputs of the charge bus.
④
4 Discharge bus provides a centralized location for connecting the battery and recti-
fiers to the primary dc power distribution system. In most power systems, the charge and
discharge buses are rigid copper busbars separated by a current shunt and, in some instal-
lations, a low-voltage disconnect device (Fig. 1.6). A current shunt is a low-resistance,
high-power resistor for measuring load current. The load current causes a small voltage
drop across the accurate resistance, which is measured by a voltmeter in the monitoring
and control system. The low-voltage disconnect device disconnects the loads from the
battery to prevent battery overdischarge or to prevent undervoltage damage to equipment.
As with the charge bus, a separate discharge busbar is provided for the positive (+) and
negative (–) outputs. In most installations the charge and discharge return buses are one
continuous busbar.
⑤④
5 6 Distribution systems provide central locations for feeding loads and for protect-
ing circuit wiring. The distribution systems also provide a convenient way to isolate indi-
8 INTRODUCTION
Fig. 1.6 Typical charge and discharge bus arrangements.
vidual loads from each other during fault conditions. The primary distribution system ( 5 )
has the first overcurrent protective device (either fuse or circuit breaker) between the dis-
charge bus and the load. Sometimes the primary distribution system is given the unwieldy
name battery distribution fuse bay (BDFB) or battery distribution circuit breaker bay
(BDCBB), depending on the type of overcurrent device. Although the primary distribu-
tion system can directly feed loads if desired, a secondary distribution system ( 6 ) may be
used. The secondary distribution system is an intermediate protected distribution network
between the primary distribution network and the load equipment. There can be one or
more secondary distribution systems, which serve individual loads or groups of loads
downstream from the primary distribution system.
⑦7 Voltage converters consist of two basic types: DC–DC converters and inverters.
DC–DC converters change the battery–rectifier system voltage and polarity to other uti-
lization voltages and polarities. Inverters change the dc back into ac, typically 120 Vac,
60 Hz, for ac-powered load equipment that requires an uninterruptible, or protected, ac
supply. Each such voltage conversion system usually has its own distribution system for
connecting and protecting load circuits.
In addition to the main components just described, all dc power systems include ancil-
lary systems such as
⑧8 Prime and standby ac power systems. Commercial electric utility ac power sys-
tems almost always are the first choice for providing power to telecommunications facili-
ties so, when available, they are called prime power sources. At many sites, but not all of
them, standby power is provided by an engine–generator set fueled by diesel, propane
(liquefied petroleum gas, or LPG) or liquefied natural gas (LNG)2 (Fig. 1.7).
Many remote sites are not served by commercial power supply systems. In this case,
on-site generation is required in which case the local generation system is the prime pow-
er source. A number of technologies may be used including internal combustion en-
gine–generator sets, wind generator sets, thermoelectric generators, solar (photovoltaic)
generators, fuel cells, and, occasionally, hydroelectric stations. Hybrid power systems,
consisting of two or more technologies, are used at many sites (Fig. 1.8). The battery re-
2
LPG is either propane or a mixture of propane and butane, but domestic bottled gas normally is plain propane.
LPG is produced in the extraction of heavier liquids from natural gas and as a by-product in petroleum refining.
LNG is created by refrigerating subterranean natural gas. This reduces the volume by approximately 600 times
for transport. LNG is then converted back to a gas by warming it.
Fig. 1.7 Electric utility (prime) and engine–generator (standby) ac power systems.
serve time requirements generally are relaxed at sites with multiple primary sources or
standby generation.
⑨9 Alternating current power distribution system. Commercial ac systems consist of
an electric service entrance and metering and distribution switchboards or panelboards
(Fig. 1.9). Panelboards include circuit breaker panels for overcurrent protection of ac cir-
cuits and a transfer switch for transferring from the prime ac supply source to the standby
source upon failure of the prime source. Not all sites have a standby source, in which case
the transfer switch and standby generator system components are not equipped and the
service entrance is connected directly to the distribution panelboard. Alternating current
distribution systems are used not only for the rectifiers but also for lighting, heating, ven-
tilating, and air conditioning (HVAC), convenience receptacles, and other ac-powered
equipment at the site. Depending on the prime service voltage, the ac distribution system
may include step-up or step-down transformers to match supply voltages to load equip-
ment utilization voltages.
Fig. 1.8 Block diagram of a hybrid power supply system.

10 INTRODUCTION
ac Load
Equipment
Fig. 1.9 Simplified on-site ac power distribution system with standby generator.
⑩
10 Load equipment includes all the network equipment and associated operational
support systems at the site that are required to be powered by the dc system. Depending
on the site purpose, this can include core switching, multiplexing and transmission equip-
ment, wireless base stations, and related dc and ac-powered equipment.
①
11 Monitoring and control system includes alarm collection, processing and sending
systems, and metering. Control systems include rectifier float/equalize control and tim-
ing, parametric recorders, and local area network (LAN), serial port, and modem inter-
faces.
Other important infrastructures required for power system operation are:
앫 Cable and rigid bus supporting structures are cable and ladder racks and raceway
systems used to physically protect and support wire and cable.
앫 Equipment frames are assemblies of equipment or components mounted on a com-
mon support structure. Other names are relay rack, bay, and cabinet; however, re-
lay racks are open on all sides while bays and cabinets usually are enclosed frames
with doors for access to the front and rear.
앫 Powerboard is an equipment frame or frames containing dc power equipment such
as the charge and discharge busbars, rectifiers, primary distribution system, and
power system controllers, meters, and alarm panels (Fig. 1.10). Power system
frames usually are dead-front, which means no live parts are exposed to persons on
the operating side of the equipment. Live-front, on the other hand, means live (con-
ductive) parts with potential difference with respect to ground are exposed to per-
sons on the operating side of the equipment. All modern power system frames are
dead-front.
앫 Special floor structural components or physical support systems may be required
for batteries due to their high weight and small footprint.
앫 HVAC and lighting are required in all installations.
앫 Grounding system is a system consisting of earth grounding electrodes and interior
grounding busbars and components. All dc power systems used in telecommunica-
tions are referenced to earth ground. Telecommunications facilities grounding elec-
trode systems may consist of buried copper grids or plates, ground rods, water
Fig. 1.10 Powerboard frame assembly.
wells, concrete encased electrodes (footings and slabs), buried copper radials at an
antenna supporting structure and other buried metallic components, and building
structural steel. Grounding systems also include lightning and surge suppression
systems.
앫 Older end office switching systems required central ringing voltage and precision
audible tone supplies (e.g., ringback tone and dial tone), which were provided as
part of the dc power system. However, almost all modern switching systems have
integrated power supplies for these purposes and their design is beyond the scope of
this book.
A common method of improving the reliability and performance of dc power systems

is through redundancy of components, such as rectifiers, circuits, and batteries. Typically,
component groups, such as rectifiers, are sized and equipped such that the failure of any
one of them does not result in any service degradation and the remaining components in
that group carry the load. Circuit redundancy is provided by running two independent cir-
cuits (A and B) from the primary distribution system to all secondary distribution systems
and to all load equipment. It also is possible to provide two independent power systems
that are interconnected only by common reference to the earth grounding electrode sys-
tem.
12 INTRODUCTION
1.4 POWER SOURCES AND LOADS
A system may include a number of different power sources, including prime and standby
power sources and energy storage systems, such as batteries.
1.4.1 Prime Power Sources

Prime power sources by their nature are very reliable. The most common prime power
source is the commercial electric utility that serves the telecommunications facility. Com-
mercial electric service normally is less expensive than alternate prime power systems.
Commercial electric utility systems have quite high reliability, but they are neverthe-
less subject to interruption or outage, and the reliability is not 100% when measured over
months, a year, or several years. Any interpretation of what is termed an interruption in-
cludes the magnitude and duration of the voltage dip. For purposes of telecommunica-
tions dc power system design, interruption is defined in terms of a minimum duration and
assumes the voltage is zero during that period. Interruptions may last from fractions of a
millisecond (microinterruptions) to weeks. Interruption statistics vary widely by region
and area of the country and only gross generalizations can be made about them.
It is important that commercial electrical service interruptions do not affect network
equipment. All modern network equipment is based on processor-controlled digital tech-
nology, and these systems cannot withstand even a short loss of voltage. Some equipment
or systems that support network equipment can withstand varying degrees of interruption.
For example, the dry air pressurization systems for waveguide and coaxial cable transmis-
sion lines can loose power for significant time periods—hours or even days—depending
on how leakproof they are. Other equipment may operate correctly if the interruption does
not exceed a few seconds but may fail or need to be restarted on longer interruptions.
Short-time interruptions in the millisecond range may be bridged by the capacitors in
rectifier and chassis power supply output filtering circuits. Longer duration interruptions
are bridged first by batteries, which are online all the time, and ultimately by standby pow-
er systems such as engine–generator sets if the site is so equipped. If not equipped, the net-
work equipment stops working or falls out of specification when the batteries discharge to
the minimum equipment operating voltage or when the battery voltage reaches the thresh-
old of the low-voltage disconnect (LVD) and the LVD disconnects the load.
Electric utility circuit switching and automatic circuit reclosing during faults (e.g., line
slap during wind storms, lightning, or trees falling into lines) typically cause interruptions
lasting less than 0.5 s. Most utility interruptions last less than 30 min, but some may last
much longer; the interruption frequency and duration can be higher in some rural areas.
Longer interruptions usually are due to
앫 Electrical equipment failures (e.g., damage to service transformers, distribution

lines, or substations)
앫 Natural disasters (earthquakes, floods)
앫 Weather (hurricanes, tornados, freezing rain, and severe storms)
앫 Cascading faults in large interconnected electric power grids (when large power
systems collapse, the resulting outages may last for several days or even weeks in
some areas)
앫 Human error
1.4 POWER SOURCES AND LOADS 13
Where commercial electric service is not available, an alternate prime power source
must be used to power the telecommunications power system. A variety of prime movers
(means to mechanically power an electrical generator) and other energy sources are avail-
able, both nonrenewable and renewable (Table 1.3). The most common prime mover falls
in the broad category of reciprocating internal combustion (RIC) engine. A number of dif-
ferent RIC engine technologies are available including air and liquid-cooled spark igni-
tion (e.g., gasoline) and air- and liquid-cooled compression ignition (e.g., diesel). Both
spark and compression ignition engines can be set up with multifuel capability. Gas-tur-
bine (GT) engines also have been used successfully as prime movers, particularly in larg-
er installations. So-called microturbines are seeing greater use as power sources in small
systems. Many full-size GT and microturbines have been adapted from aircraft engines or
aircraft auxiliary power units (APU) and can have multifuel capability.
1.4.2 Standby Power Sources

On a unit power or energy basis, standby power sources usually are considerably more
expensive than a comparably rated prime power source and are run only when the prime
power source fails. Although they may use the same basic technologies as prime power
sources, standby power sources are designed to run for much shorter time periods such as
hours, days, or a few weeks at a time. Standby sources run, on average, a few hours a
month or year, although there are many sites that vary considerably from the average. At
locations where the prime power source is reasonably reliable, standby sources may run
only a few hundred hours over their lifetime, and most of that time is during regularly
scheduled exercising tests.
The most common standby power sources are diesel, natural gas, or propane-fueled RIC
engine–generator sets. Gasoline seldom is used because it is too difficult to safely handle
and store. Where engine–generator sets are used in standby service, they usually have high-
er power ratings than comparable sets used in prime power service (this can be done with-
out compromising the engine because of the relatively low duty cycle in standby service).
Table 1.3 Prime Movers and Other Energy Sources

Nonrenewable Energy Source Prime Mover or Generator
a
Biomass Reciprocating internal combustion (RIC) engine
Diesel RIC or gas-turbine (GT) engine
Gasoline RIC engine
Heavy oil RIC engine
Jet fuel RIC or GT engine
Kerosene RIC or GT engine
LPG and LNG RIC or GT engine
Wood gas RIC engine
Renewable Energy Source Prime Mover or Generator
Hydrogen Fuel cell (no prime mover)
Sunlight (solar) Photovoltaic (no primer mover)
Water Water turbine
Wind Wind turbine
a
Timber and agricultural residues, nonhazardous solid waste, wastewater, and animal manures.
14 INTRODUCTION
For convenience, the standby system generation voltage and frequency generally is the
same as the prime power source. In remote unattended sites there may be advantages to di-
rectly generating direct current at the telecommunications system operating voltage.
Standby systems can be designed to automatically start and stop and to switch to and
from online operation, or they may be manually started and switched. In some installa-
tions, there is no permanent standby system installation except for a generator electrical
receptacle and transfer switch. In this case, a portable generator set is hauled to the site,
plugged in and run for the duration of the outage. The battery must be large enough to
bridge the interruption until the generator arrives and is placed online, while accounting
for inevitable problems in dispatching and setting up portable equipment.
1.4.3 Alternating Current Power System Loads

The ac power system at a telecommunications facility serves a variety of ac loads. The
loads can be categorized according to their importance as nonessential, essential, and pro-
tected (Fig. 1.11).
Nonessential Load An ac load that does not need to operate during prime ac power
outages of any length.
Essential Load An ac load that must operate during extended prime ac power out-
ages but can tolerate power interruptions lasting a few or several seconds, or possi-
bly longer. During longer prime power source outages, essential ac loads are pow-
ered by a standby engine–generator set if the site is so equipped.
Protected Load An ac load that cannot tolerate any interruption of its input power.
Protected ac loads (other than the dc power system rectifiers) are powered by an in-
verter that, in turn, is powered by the dc power system battery during outages.
Some protected loads only need to operate long enough for an orderly shutdown. In
these cases, an Uninterruptible Power System (UPS) is used to power the load.3
1.4.4 Energy Storage

A battery is an electrochemical storage device that takes energy from the prime power
source through the rectifier system and stores it as chemical energy. Upon failure or inter-
ruption of the prime power source or rectifier system, the battery reconverts the chemical
energy to electrical energy and powers the network equipment loads. Both lead–acid and
nickel–cadmium (NiCd) batteries are used in telecommunications with lead–acid types
being the most common and the only type considered in this book.
Two basic lead–acid technologies are used—vented lead–acid (VLA) and valve-regu-
lated lead–acid (VRLA). A VLA battery consists of cells that allow gases evolved during
operation to escape from the cell containers. A VRLA battery consists of cells with a
pressure relief valve that prevents the relatively low amounts of evolved gas from escap-
ing during normal operation, allowing the gases to recombine and eliminating the need to
periodically add water to the cells.4 However, during abnormal operation the gases are al-
lowed to escape through the pressure relief valve.
3
In general, most stand-alone uninterruptible power systems have a relatively short battery reserve time—typi-
cally 15 min to 1 h—compared to the telecommunications dc power system.
4
The recombination process is not 100% efficient so a small amount of gases will eventually escape.
Nonessential ac Loads
Essential ac Loads
dc Load
Equipment
Protected ac Loads
Fig. 1.11 Alternating current (ac) load categories—nonessential, essential, and protected.
Batteries used in telecommunications are classified as secondary (rechargeable) and

for stationary (fixed) service.5 Because of their different electrical characteristics, individ-
ual battery designs can be categorized according to their reserve times:
앫 Batteries for short-term loading (< 1 h)

앫 Batteries for long-term loading (> 1 h)
Telecommunications applications fall into the long-term loading category, and that cate-
gory is the focus of this book. All telecommunications batteries of concern here use lead-
active materials in the plates and a dilute electrolyte of sulfuric acid in water. In VRLA
batteries the electrolyte is immobilized as a gel or absorbed in a glass fiber mat separator
between the lead plates. In VLA batteries the electrolyte is a free liquid.
A battery, or battery string, consists of a number of cells connected in series. Each cell
has a nominal voltage of 2 V. Modern telecommunications dc power systems in the Unit-
ed States use 24 or 48 V (or both) and the battery strings consist of 12 cells and 24 cells,
respectively. Other cell counts are sometimes seen in older systems. Battery plant capaci-
ty can be increased by connecting individual strings in parallel (Figs. 1.12 and 1.13). Bat-
teries and battery systems are described in detail in Chapter 4 (Telecommunications Bat-
teries).
There are numerous other energy storage devices that are promising for telecommu-
nications power system applications, particularly fuel cells (fuel cells convert the chem-
5
Primary batteries are not rechargeable and are disposable. A familiar example is a simple flashlight battery.
16 INTRODUCTION
Fig. 1.12 Twelve-cell (24-V) battery strings (upper, series arrangement; lower, series–parallel
arrangement).
ical energy of hydrogen to electricity without combustion and with little pollution).
Also, there are battery active materials other than lead that are becoming more impor-
tant. However, at the time of this writing (2005), none of these alternate technologies
adequately meet the cost requirements of the typical installation supplying several
tens to thousands of amperes of load current over 3-h or longer reserve times. These al-
ternate technologies presently are suitable in small applications where at most only a
few amperes are required, and their size and cost do not present undue economic penal-
ties.
1.4.5 Voltage Magnitudes

1.4.5.1 Alternating Current The ac power systems in the United States operate at
a frequency of 60 Hz, and the service voltages at buildings and facilities are standardized.
For many facilities, the service voltages are nominally 120 V line-neutral and 240 or 208
V line-line, depending on whether the service is one-phase or three-phase, respectively.
Fig. 1.13 Battery systems (24-cell VRLA on the left and 24-cell VLA on the right).
Another common three-phase service voltage is 277 V line-neutral and 480 V line-line.
Larger facilities may be served at primary voltages in the range of 7 kV line-neutral and
13 kV line-line, or higher.
1.4.5.2 Direct Current The nominal dc system voltages used by modern telecom-
munications facilities in the United States are either +24 or –48 V. A +24 Vdc system has
its negative terminal bonded to the facility earth grounding system. A –48 Vdc system has
its positive terminal bonded to the earth grounding system. The positive terminal on 48-
Vdc power systems originally was connected to earth to reduce galvanic corrosion due to
leakage currents on buried lead-sheathed outside plant cables and other metallic pipes
(water, gas) and components buried in the ground. Although lead is no longer used to
sheath outside plant cables, the practice continues to minimize corrosion on existing lead-
sheath cables and other buried metallic components. Systems that use 24 Vdc normally do
not have any buried components and thus are not subject to corrosion. For example, 24-V
private branch exchanges (PBX) normally do not have outside plant twisted-pair cables,
and 24-V radio frequency (RF) base stations do not have buried outside plant coaxial ca-
bles.
Both 24- and 48-V systems fall in the 60-Vdc class and are considered central supply
systems that directly feed the loads. Other voltages have been used in the past including
±130 Vdc for powering T1-carrier span lines6 and coin telephone collect and return cir-
cuits. These higher voltage systems fall in the 160-Vdc class and are not covered in this
book because they are not used in modern telecommunications systems.
1.4.6 Direct Current Power System Loads

Most load equipment uses distributed, or decentralized, power supplies to convert the 24-
or 48-V direct fed voltages to component or electronic voltages, such as 5 and 12 V (Fig.
1.14). These voltages are provided by on-board or chassis-mounted DC–DC converters
(e.g., 48 to 5 Vdc).
All load equipment has a voltage tolerance specified by a minimum and maximum op-
erating voltage. The minimum voltage determines when the equipment stops operating or
no longer meets operating specifications as batteries discharge during a prime power
source or rectifier system failure. The maximum voltage determines the upper limit at
which the equipment can safely operate without damage either from electrical stress or
heat damage. The type of battery technology and the voltage at which it is charged affects
the maximum operating voltage.
Load equipment spends most of its time at a voltage somewhat higher than the nominal
system voltage, typically 9 to 14% higher. For example, –48-Vdc systems with 24-cell
VRLA batteries typically operate at 54.0 to 54.5 V. The actual magnitude of the operating
voltage varies with the type of battery technology and other factors, but the overall
change from the minimum to maximum is around 25 to 30% (e.g., 44 to 56 Vdc). Some
modern load equipment has a much wider range, on the order of 100% (e.g., 36 to 72
Vdc). However, in a given central office or telecommunications facility, the equipment
6
T1-carrier is a type of twisted-pair transmission system originally deployed in the early 1960s that uses line-
powered repeaters to extend its operating distance to as far as 200 miles (320 km). Although T1-carrier still is
used today, line powering is by individual line interface units called central office repeaters and not by a central-
ized system.
18 INTRODUCTION
Facility ac Power System
dc Power System
Fig. 1.14 Decentralized power for network equipment shelves.
with the highest minimum operating voltage and the lowest maximum operating voltage
determines the actual design voltages for the power system.
Network equipment loads fall into two basic categories, resistive and constant power,
with most modern loads in the constant power category. Resistive loads draw current in
proportion to the applied voltage. Therefore, when a discharging battery powers a resis-
tive load, the load current decreases as the battery voltage decreases. On the other hand,
the current to constant power loads is inversely proportional to the voltage, and the load
current increases as the discharging battery voltage decreases.
1.5 GENERAL DESIGN CONSIDERATIONS
A telecommunications dc power system is required to provide the specified voltage range

at the load equipment under all operating conditions. When the commercial ac power sys-
tem fails, the dc power system must operate the load equipment for a specified battery re-
serve time. Power system design involves many variables including battery reserve time,
circuit current ratings, and voltage drop in the battery circuits and dc distribution circuits.
Other design considerations include the choice between rigid bus and cable bus in feeder
and distribution circuits, rectifier sizing, and voltage conversion device sizing.
1.5.1 Design Loads

For convenience, dc load currents are separated into two basic categories called normal
and peak. Normal current specifies the average load current during normal operation.
Normal load current is used to size rectifiers. Rectifiers are relatively easy to add to an op-
erating system and their ratings and quantities are based on the initial and ultimate normal
1.5 GENERAL DESIGN CONSIDERATIONS 19
equipment loads. In some telecommunications switching systems and wireless base sta-
tions, the load current depends on traffic levels, so the normal load current takes into ac-
count peak traffic conditions associated with these equipment types.
Peak current specifies load current during worst-case conditions of low battery dis-
charge voltage. Peak currents include inverter loads if the inverter is offline during nor-
mal conditions (if the inverter in online during normal conditions, its load is included in
the normal category). Peak load currents are used to size equipment that cannot be ex-
panded without incurring considerable operational risk, including primary distribution
circuits, primary overcurrent protective devices, and system charge and discharge buses.
Peak load currents also are used to determine initial and ultimate battery capacities; how-
ever, unlike charge/discharge buses and the primary distribution system, batteries may be
added relatively easily in an operating system.
1.5.2 Voltage Drop

A telecommunications dc power system consists of four basic types of circuits—battery
circuit, rectifier circuit, primary distribution circuit, and secondary distribution circuit
(Fig. 1.15). A system may have cascaded secondary distribution circuits (Fig. 1.16). A
circuit always consists of two conductors, feed and return, so the circuit forms a loop, and
voltage drop calculations must include the drop in both conductors.
During normal operation, the rectifiers power the load equipment and the dc electrical
circuit consists of the rectifier, primary distribution and secondary distribution circuits.
Since the rectifier output voltage is well regulated and considerably higher than the mini-
mum equipment operating voltage, the voltage drop from the rectifiers to the load equip-
ment is not an important design issue. When the ac input, ac source, or the rectifiers fail,
the battery powers the load. In this case, the electrical circuit consists of the battery, pri-
mary distribution, and secondary distribution circuits. The voltage drop from the battery
to the load when the rectifiers are not operating plays an important role in keeping the
equipment in operation for the longest possible time as the battery discharges and its volt-
age decreases.
Fig. 1.15 dc power system circuits.

20 INTRODUCTION
Fig. 1.16 Cascaded secondary distribution circuits.
The maximum allowed total voltage drop from the battery to the load equipment dur-
ing battery discharge is around 4%. Since the important factor is the total voltage drop,
the voltage drop in each segment (battery circuit, primary distribution circuit, and sec-
ondary distribution circuits) can be designed for any value as long as the sum of the volt-
age drops does not exceed the specified maximum. More specific requirements are dis-
cussed in Chapter 5, System Design.
1.5.3 Conductors
Copper conductors consisting of rigid bus or cable bus (Fig. 1.17) are used in telecommu-
nications dc power systems, although rigid aluminum busbar has been used occasionally.7
Current-carrying capacity (ampacity) and voltage drop are the two criteria used to select
the conductor size in any given circuit. Voltage drop is determined from the conductor re-
sistance at the desired design temperature and load current. Since conductor resistance de-
pends on length, the circuit path distance must be known. Each circuit consists of two
conductors, feed and return, so the total conductor length is always twice the path dis-
tance. Feed and return conductors always are paired and never shared with other circuits.
The feed conductor is the circuit conductor operating at the system voltage and the return
conductor is the circuit conductor operating nearest ground potential.
Ampacity is the term used to describe the current the conductor may safely carry with-
out overheating.8 In low-voltage dc systems, currents are relatively high. On short con-
ductor runs, ampacity usually determines conductor size, while on long runs voltage drop
most often determines conductor size. On any given circuit, the conductors must be large
enough to meet both criteria (ampacity and voltage drop) simultaneously.
In the United States, conductor sizes are stated in American Wire Gauge (AWG, or
just gauge) or circular mils (CM). Conductor sizes 4/0 (0000) and smaller are described in
7
Aluminum cable conductors are not used in telecommunications dc power systems because of the relative diffi-
culty making aluminum wire connections due to aluminum’s tendency to cold-flow and corrode.
8
The term ampacity is peculiar to the National Electrical Code (NEC). Engineering books about electrical pow-
er cables use the terms current rating and current-carrying capacity.
Fig. 1.17 Rigid bus (upper) and cable bus connected to rigid bus (lower).
22 INTRODUCTION
AWG (e.g., 1/0 AWG) while conductors larger than 4/0 are described in circular mils
(e.g., 250,000 CM). The circular mil is a measure of the cross-sectional area of a conduc-
tor and is equal to the square of its diameter, where the diameter is measured in mils (1
mil = 0.001 inch). A common variation of CM is kilocircular mil (kcmil), which is 1000
CM (kcmil sometimes is seen in older documents as MCM). Conductor properties are de-
scribed in greater detail in Chapter 2, Electricity Review.
1.5.4 Overcurrent Protection

Basic electrical circuit design requires conductors to be protected at their power source
from overloads and short circuits by an overcurrent protection device, such as a fuse or
circuit breaker. The source for all circuits except battery circuits is the primary and sec-
ondary distribution panels (Figs. 1.18 and 1.19). In many instances it is impossible to pro-
vide reliable battery circuit operation at the same time as reliable overcurrent protection.
Therefore, other means are used to reduce the probability that a fault occurs on battery
circuit conductors.
A short circuit is an unintentional contact of an ungrounded conductor to a live,
grounded, or dead metal part, whereas an overload is a higher-than-normal current due to
an equipment fault or operational problem.
An overcurrent protection device must protect each load branch circuit. During fault
conditions, if the overcurrent protection devices are properly coordinated, only the over-
current device closest to the fault trips open and upstream devices are unaffected. The
choice between a fuse and circuit breaker is dictated by a number of factors, including
cost (the initial cost of fuses usually is less), convenience (resetting a circuit breaker is
easier, safer, and faster than replacing a fuse) and fault current-carrying capabilities (some
fuses can carry very high fault currents and may be used to protect a circuit breaker under
high fault conditions).
1.5.5 Rectifiers
Rectifiers must be able to serve load currents and simultaneously recharge a discharged
battery. Once the battery is recharged, the rectifiers closely regulate the float voltage to
ensure a long battery life. Battery recharge times typically range between 12 and 24 h.
The rectifier system is based on N + 1 redundancy, where N is the number of rectifiers re-
quired to meet load and recharge requirements and one additional rectifier is added for re-
Fig. 1.18 Overcurrent protection.

1.6 STANDARDS, PRACTICES, AND CODES 23
Fig. 1.19 Primary distribution (left) and secondary distribution (right).
dundancy. This means there always are at least two rectifiers in a system. Rectifiers al-
ways are configured in the hot-standby mode so they can pick up the load immediately
upon failure of one of the other rectifiers. Rectifiers may be modular plug-in types or
fixed mounted (Fig. 1.20).
1.5.6 Battery Capacity

Battery capacity is determined from the peak load current, the required reserve time, an-
ticipated growth, and the need to provide for expansion as the power system grows. The
reserve time depends on whether a standby engine–generator set is installed at the site and
can vary from as little as 3 h to upwards of 12 h or more. Remote sites may require a re-
serve time measured in days or weeks. Other considerations, such as cost, weight, floor
space requirements, and system reliability, affect the battery system design. The growth
in battery capacity is related to the growth in rectifier capacity because both increase with
load current.
Where service reliability is critical, two battery strings may be installed, each with at
least 50% of the total required capacity. If full N + 1 redundancy is required, each of the
two strings must have 100% of the required capacity (Fig. 1.21). In either case, multiple
strings allow easier battery maintenance without jeopardizing service to the load equip-
ment. The number of battery strings in any given installation is the minimum required to
achieve reserve time and maintenance requirements.9
9
The author has been told of systems consisting of 64, 72, and 100 battery strings in parallel.
24 INTRODUCTION
Fig. 1.20 Modular switch-mode rectifiers (left) and fixed mounted ferroresonant rectifiers (right).
(Left photo courtesy of Power-One, Inc.)
1.6 STANDARDS, PRACTICES, AND CODES
Standards, practices, and codes are written documents that provide interoperability, effi-
ciency, and uniformity in components and systems and uniformity in design, construc-
tion, operation, and maintenance methods. This section briefly describes such documents
related to dc power systems.
Standards in the United States are voluntary consensus standards. The American
National Standards Institute (ANSI), headquartered in Washington, D.C., administers, co-
ordinates, and safeguards the integrity of the U.S. voluntary standardization and conformi-
ty assessment system. While many standards specify detailed requirements, some specify
only general requirements and details are provided through company or industry practices.
Two important organizations that develop national standards related to telecommuni-
cations dc power systems are the Institute of Electrical and Electronics Engineers (IEEE)
and the Alliance for Telecommunications Industry Solutions (ATIS) Network Interface,
Power and Protection (NIPP) Committee (formerly T1E1).
Standards may be obtained from ANSI at http://webstore.ansi.org or Global Engineer-
ing Documents at http://global.ihs.com. IEEE and NIPP Committee standards also may
be obtained as described in the following paragraphs. All standards are updated or reaf-
firmed at least every 5 years (the dates are not shown in the following listing because the
most current should be used in design work; where standards are cited in later chapters,
the date in effect at the time this book was written is given).
Practices reflect common design, installation, and operation methods, but a given
practice may not be universally accepted throughout the industry. There may be numer-
ous acceptable practices for accomplishing the same task or meeting the same end. Prac-
tices can be formal and written or informal and verbal and in many cases technically un-
justified (“We’ve always done it that way”). Standards and practices are not legally
binding unless adopted by reference by federal, state, county, or municipal jurisdictions.
Codes generally specify safety requirements and are adopted by governmental juris-
dictions and, as such, are legally binding. Examples of codes that may apply to dc power
systems are the National Electrical Code® and International Fire Code.10
10
The NEC® does not apply in spaces used exclusively for telecommunications equipment. Additional detail on
the application of various codes is covered in Chapter 5 (System Design).
Fig. 1.21 Redundant VRLA battery installation (upper, B-Blue; lower, A-Red). (Photos courtesy of
Nick Nichols.)
26 INTRODUCTION
1.6.1 ATIS Network, Interface, Power and Protection Committee

NIPP Committee (NIPP— http://www.atis.org) develops standards for interfaces, power,
and network protection. Important NIPP Committee standards are:
앫 ANSI/T1.304, Ambient Temperature and Humidity Requirements for Network

Equipment in Controlled Environments
앫 ANSI/T1.307, Fire Resistance Criteria—Ignitability Requirements for Equipment
Assemblies, and Fire Spread Requirements for Wire and Cable
앫 ANSI/T1.311 DC Power Systems—Telecommunications Environment Protection
앫 ANSI/T1.313, Electrical Protection of Telecommunications Central Offices and
Similar Type Facilities
앫 ANSI/T1.315, Voltage Levels for DC Powered Equipment Used in the Telecommu-
nications Environment
앫 ANSI/T1.319, Equipment Assemblies—Fire Propagation Risk Assessment Criteria
앫 ANSI/T1.328, Protection of Telecommunications Links from Physical Stress and
Radiation Effects and Associated Requirements for DC Power Systems (a baseline
standard)
앫 ANSI/T1.329, Network Equipment Earthquake Resistance Standard
앫 ANSI/T1.330, Valve Regulated Lead-Acid Batteries Used in the Telecommunica-
tions Environment
앫 ANSI/T1.333, Grounding and Bonding of Telecommunications Equipment
1.6.2 Institute of Electrical and Electronics Engineers (IEEE—

http://standards.ieee.org)
앫 ANSI/IEEE Std 446, IEEE Recommended Practice for Emergency and Standby
Power Systems for Industrial and Commercial Applications (IEEE Orange Book)
앫 ANSI/IEEE Std 450, IEEE Recommended Practice for Maintenance, Testing, and
Replacement of Vented Lead-Acid Batteries for Stationary Applications
앫 ANSI/IEEE Std 484, IEEE Recommended Practice for Installation Design and In-
stallation of Vented Lead-Acid Batteries for Stationary Applications
앫 ANSI/IEEE Std 485, IEEE Recommended Practice for Sizing Lead-Acid Batteries
for Stationary Applications
앫 ANSI/IEEE Std 946, IEEE Recommended Practice for the Design of DC Auxiliary
Power Systems for Generating Stations
앫 ANSI/IEEE Std 1187, IEEE Recommended Practice for Installation Design and In-
stallation of Valve-Regulated Lead-Acid Storage Batteries for Stationary Applica-
tions
앫 ANSI/IEEE Std 1188, IEEE Recommended Practice for Maintenance, Testing, and
Replacement of Valve-Regulated Lead Acid (VRLA) Batteries for Stationary Ap-
plications
앫 ANSI/IEEE Std 1189, IEEE Guide for Selection of Valve-Regulated Lead Acid
(VRLA) Batteries for Stationary Applications
앫 ANSI/IEEE Std 1375, IEEE Guide for the Protection of Stationary Battery Systems
1.6.3 Telecommunications Industry Association

The Telecommunications Industry Association (TIA—http://www.tiaonline.org/) is a
trade association of telecommunications equipment manufacturers.
앫 ANSI/J-STD-607-A, Commercial Building Grounding (Earthing) and Bonding Re-
quirements for Telecommunications
1.6.4 Telcordia
Telcordia (http://www.telcordia.com) develops numerous technical requirements for its
clients, the Regional Bell Operating Companies (RBOC) and others. Because of the
RBOC’s purchasing power, most telecommunications equipment manufacturers build
equipment to Telcordia requirements; however, these are not national standards although
some have become ad hoc or de facto standards. Of particular interest are the widely quot-
ed NEBS, or Network Equipment Building Systems, documents:
앫 GR-1089-CORE, Electromagnetic Compatibility and Electrical Safety Generic Cri-

teria for Network Telecommunications Equipment
앫 GR-63-CORE Network Equipment Building Systems Requirements: Physical Pro-
tection
1.6.5 Rural Utilities Service

The Rural Utilities Service (RUS—www.usda.gov/rus/telecom/publications/bulletins) is
an arm of the U.S. Department of Agriculture—Rural Development Agency that loans
money to rural telecommunications service providers and has developed technical re-
quirements and approved equipment lists for use by its borrowers. Three documents par-
ticularly relevant to this book are
앫 Bulletin 1751E-302—Power requirements for digital central office equipment

앫 Bulletin 1751E-320—Emergency generating and charging equipment
앫 Bulletin 1751F-810—Electrical protection of digital and lightwave telecommunica-
tions equipment
1.6.6 Electrical Generating Systems Association

The Electrical Generating Systems Association (EGSA—http://www.egsa.org) only re-
cently became a nationally approved standards developer but not all of its requirements are
national standards as of this writing; however, EGSA has been developing performance re-
quirements for electrical generating systems since 1965 and its requirements are widely
recognized. Unlike NIPP and IEEE standards previously listed, EGSA standards are free.
앫 EGSA 100B: Performance Standard for Engine Cranking Batteries Used with En-
gine Generator Sets—Requirements for rating, classifying, applying, installing, and
maintaining engine cranking batteries
앫 EGSA 100C: Performance Standard for Battery Chargers for Engine Starting Bat-
teries and Control Batteries (Constant Potential Static Type)—Requirements for
28 INTRODUCTION
voltage and temperature limits, application, and accessories for charging engine
cranking batteries
앫 EGSA 100D: Performance Standard for Generator Overcurrent Protection, 600
Volts and Below—Performance specifications for circuit breakers, field breakers,
thermostats, thermistors, and other temperature detectors
앫 EGSA 100E: Performance Standard for Governors on Engine Generator Sets—
Classifications, performance requirements, and optional accessories for generator
set engine governors
앫 EGSA 100F: Performance Standard for Engine Protection Systems—Performance
specifications for engine control systems including temperature, level, pressure, and
speed sensing
앫 EGSA 100G: Performance Standard for Generator Set Instrumentation, Control and
Auxiliary Equipment—Requirements for generator set engine starting controls, in-
strumentation, and auxiliary equipment
앫 EGSA 100M: Performance Standard for Multiple Engine Generator Set Control
Systems—Performance requirements for manual, automatic fixed sequence, and
random access generator set paralleling systems
앫 EGSA 100P: Performance Standard for Peak Shaving Controls—Requirements for
parallel operation and load transfer peak load reduction controls
앫 EGSA 100R: Performance Standard for Voltage Regulators Used on Electric Gen-
erators—Application and performance requirements for generator voltage regula-
tors
앫 EGSA 100S: Performance Standard for Transfer Switches for Use with Engine
Generator Sets—Classifications, applications, and performance requirements for
transfer switches for emergency and standby transfer switches
앫 EGSA 100T: Performance Standard for Diesel Fuel Systems for Engine Generator
Sets with Above Ground Steel Tanks—Application and performance requirements
for diesel fuel supply systems with above-ground steel tanks for diesel engine driv-
en generator sets
앫 EGSA 101G: Glossary of Electrical and Mechanical Terminology and Defini-
tions—Definitions of terms specific to the on-site power industry
앫 EGSA 101P-1995: Performance Standard for Engine Driven Generator Sets—Clas-
sifications of use, prime mover configuration and ratings, and performance require-
ments for complete generator sets
앫 EGSA 101S: Guideline Specification for Engine Driven Generator Sets, Emergency
or Standby
앫 EGSA 107T: Performance Standard for Generator Test Methods
앫 EGSA 109C: Code Listing: Safety Codes Required by States and Major Cities—
Listing of national and international codes and standards adopted by U.S. states and
selected major cities
1.6.7 Practices
Many large network operators have developed their own power system practices. Perhaps
the most well known, but now obsolete, practices are the old AT&T Bell System Prac-
tices (BSP) and the General Telephone & Electronics Practices (GTEP). The develop-
ment of these types of practices changed course when AT&T was divested of the Region-
al Bell Operating Companies in 1984 and when Bell Atlantic and GTE merged into Veri-
zon in 2000.
Smaller companies to some extent use the practices of larger companies. However, in
many cases the larger company’s practices are out of date or simply do not specify the
best methods for smaller companies.
1.6.8 Codes
앫 National Electrical Code (NEC), NFPA 70-2005 (http://www.nfpa.com)—Applies
to electrical installations in public and private buildings and industrial electrical
substations. It does not apply to “Installations of communications equipment under
the exclusive control of communications utilities located outdoors or in building
spaces used exclusively for such installations” [NEC par. 90.2(B)].
앫 National Electrical Safety Code (NESC), ANSI C2-2002 (http://standards.ieee.
org)—Although the NESC has little to do with telecommunications power sys-
tems, its grounding requirements apply to outside plant enclosures such as access
nodes.
앫 International Fire Code (IFC)—International Code Council (http://www.iccsafe.
org).
앫 International Building Code (IBC)—International Code Council (ICC, http://www.
iccsafe.org).
앫 U.S. Dept. of Labor, Occupational Safety & Health Administration (OSHA,
http://www.osha.gov)—Title 29 Code of Federal Regulations.
1.6.9 Other Applicable Standards and Recommendations

National Fire Protection Association (NFPA—http://www.nfpa.com)
앫 NFPA 75-2003, Standard for Protection of Information Technology Equipment
앫 NFPA 76-2003, Recommended Practice for Fire Protection of Telecommunications
Facilities
앫 NFPA 110-2002, Standard for Emergency and Standby Power Systems
앫 NFPA 111-2001, Standard on Stored Electrical Energy Emergency and Standby
Power Systems
International Telecommunications Union—Telecommunications Standardization Sec-
tor, ITU-T) (http://www.itu.int)
앫 K.27 Protection Against Interference—Bonding Configurations and Earthing Inside
a Telecommunications Building
앫 K.31 Bonding Configurations and Earthing of Telecommunication Installations In-
side a Subscriber’s Building
앫 K.35 Bonding Configurations and Earthing at Remote Electronic Sites
30 INTRODUCTION
European Telecommunications Standards Institute (ETSI—http://www.etsi.org)

앫 EG 201 147—Equipment Engineering (EE); Interworking Between Direct Cur-
rent/Isolated (DC/I) and Direct Current/Common (DC/C) Electrical Power Systems
앫 ETS 300 253—Equipment Engineering (EE); Earthing and Bonding of Telecom-
munication Equipment in Telecommunication Centers
앫 ETSI EN 302 099—Environmental Engineering (EE); Powering of Equipment in
Access Network
CHAPTER 2
ELECTRICITY REVIEW
This chapter emphasizes telecommunications direct current (dc) power systems but also
introduces relevant alternating current (ac) topics. The concept of current flow is intro-
duced followed by a discussion of voltage sources and resistances. DC voltages, currents,
and power and a number of dc circuits are solved using Ohm’s law and Kirchhoff’s cur-
rent and voltage laws. Finally, certain topics associated with ac voltage, current, and pow-
er as they apply to rectifier input and inverter output circuits are discussed.
2.1 ELECTROMOTIVE FORCE
All matter is made from atoms. An atom consists of a positively charged central nucleus,
composed of protons and neutrons, around which one or more negatively charged elec-
trons rotate. In conductors, some of the electrons can freely move or drift from atom to
atom when a difference in potential is applied to the two ends of the conductor. The po-
tential difference can be due to an excess of electrons at the negative terminal of a battery
and a deficiency of electrons at the positive terminal due to chemical action in the battery,
or it can be from other sources such as an electrical generator, thermocouple, photoelec-
tric cell, or piezoelectric generator.
The potential difference is commonly called electromotive force, or emf, and is ex-
pressed in volts. In this book, emf is called by its most familiar name, voltage. It is the al-
gebraic difference between the voltages at two points in a circuit that determines the force
with which charges move. Generally, the Earth (“ground”) is taken as the zero reference
point or zero reference plane (Fig. 2.1).
2.2 CURRENT FLOW
The movement of the free electrons, or charges, constitutes electric current (or current
flow). The direction of current flow through a load in electrical engineering is taken to be
from a region of higher potential (such as the positive terminal of a battery) to a region of
lower potential (such as the battery negative terminal) as shown in Figure 2.2. Within a

32 ELECTRICITY REVIEW
Fig. 2.1 Voltage source and reference plane. In this illustration, the negative terminal of the voltage
source is connected to the reference plane, or ground, and the other terminal is positive with respect
to ground. A voltage source may be connected the other way, that is, with a positive ground.
Fig. 2.2 Current flow.
voltage source, current flows from the negative terminal to the positive terminal. A bat-
tery is a voltage source, and current flows within the battery cells via ions.1 The opposite
direction of current flow in a circuit is sometimes seen in textbooks (the so-called electron
theory of current flow). It makes no difference which direction is used as long as it is used
consistently throughout the circuit.
2.3 VOLTAGE SOURCES
There are four types of voltage sources used in telecommunications dc power systems:
앫 Alternating current power sources including commercial electric utility (or prime
power source) and standby power system
앫 Batteries
1
Ions are atoms or groups of atoms in which the number of electrons is different from the number of protons,
giving a net positive or negative charge.
2.4 RESISTANCE 33
앫 Rectifiers
앫 Voltage conversion devices (dc–dc converters and inverters)
Rectifiers convert ac currents and voltages from the commercial electric utility or prime
power source to dc currents and voltages for use by network load equipment and for
charging the batteries. As power storage devices, batteries allow network equipment to
operate during commercial electricity outages. DC–DC converters and inverters provide
voltages for loads that do not operate at the nominal system dc voltage. DC–DC convert-
ers convert the nominal system dc voltage to another dc voltage (e.g., from –48 Vdc to
+24 Vdc or from +24 Vdc to –48 Vdc) while inverters convert the nominal system dc
voltage to ac voltage. Since inverters are powered by the dc system, including the batter-
ies, inverters allow ac powered equipment to operate during commercial electricity out-
ages. Batteries, rectifiers, voltage conversion devices, and other dc power system compo-
nents are covered in Chapter 3.
Direct current voltage sources always have some internal resistance, which causes the
output voltage to decrease under load. Rectifiers and voltage conversion devices have in-
ternal regulating mechanisms that maintain a fairly constant output voltage regardless of
load.
2.4 RESISTANCE
2.4.1 Conductors
The current in an electrical circuit depends not only on the applied voltage but on the cir-
cuit properties as well. For example, if a copper conductor is connected across the termi-
nals of a battery or other voltage source, current will flow. If the conductor makes a poor
contact at one of the terminals or at some other point in the circuit, the current will de-
crease even though the voltage remains unchanged. Also, heat will be dissipated at the
point of poor contact. If the conductor is cut in the middle and some type of load equip-
ment is connected in the circuit, the current will further decrease and more heat will be
dissipated. In each case, a current decrease accompanies the insertion or addition of the
poorer conducting medium or of the load.
The property of an electric circuit that prevents or decreases current flow and at the
same time causes electric energy to be converted to heat energy is called resistance. A re-
sistor is a component that has intentional resistance. Resistance is measured in ohms.2
Load equipment has equivalent resistance, which in some cases may depend on the volt-
age across it or the current through it. Conductance is the inverse of resistance (1/R,
where R is the resistance), and its unit of measure is the siemens3 (formerly mho). The
higher the conductance, the lower the resistance. The term conductance is useful in some
contexts, for example, when the ability to conduct is more relevant than the ability to re-
sist current flow.
Resistance may be accounted for by considering that electrons moving through a con-
ductor must pass through the molecules or the atoms. In doing so, they collide with other
2
The resistance unit, ohm (), was named after Georg Simon Ohm, a Bavarian mathematician and physicist,
who developed Ohm’s law.
3
The conductance unit, siemens (S), was named after Werner von Siemens, a German inventor and industrialist.
In 1866 he discovered the dynamo-electric principles that helped establish electricity as a power source.
electrons and with the atoms and in the process dissipate heat. The number of collisions,
and the heat dissipated, in a given time varies with the square of the current. The velocity
of the electrons (and their electric energy) decreases and thus the current is reduced.
When this happens, the voltage must be increased to maintain a given current.
Electrons move relatively easily from atom to atom in a conductor, but they do not
move easily in an insulator. Oils, glass, dry paper, cotton, rubber, porcelain, and plastics
are nonconductors or insulators. Wood, either dry or impregnated with oil, is a good in-
sulator, while wet wood and wood treated with some preservatives is a partial conduc-
tor.
Conductors may be categorized as metallic, electrolytic, or gaseous. With metallic
conductors, such as copper wire, conduction is due to interatomic movement of the elec-
trons within the conductor. This movement is not accompanied by any movement of ma-
terial through the conductor or by any chemical action. With electrolytic conductors, con-
duction is accompanied by movement of material (ions) through the conductor and
usually by chemical action. The diluted sulfuric acid used in lead–acid batteries is an ex-
ample of an electrolytic conductor. With gaseous conductors, conduction is due to move-
ment of free positive ions and free negative ions, or electrons, into which the gas atoms
become divided when the gas is ionized. The ionized gas in a neon sign is an example of a
gaseous conductor.
The best metallic conductors are silver and copper and alloys of these elements. Table
2.1 shows the conductance of some metallic conductors of the same cross-sectional area
relative to silver. Copper is the only metal used in telecommunications dc power systems
wiring, but aluminum is common in commercial and industrial ac wiring.
A 1- resistor will allow 1 A of current flow if 1 V is applied across it.4 Also, if 1
A flows for 1 s through a 1- resistor, the heat energy dissipated by it is 1 W-s (1 W =
1 J/s).5 The resistivity of a material is the ohmic resistance of a unit cross section of the
material per unit length. The resistivity, or resistance between opposite faces of a cen-
timeter cube (face dimension of 1 cm), of soft or annealed copper6 at 20°C (68°F) is
1.724 × 10–6 (1.7241 -cm or 0.67879 -in.), or approximately 98% of pure cop-
per, while the resistance between opposite faces of a centimeter cube of hard rubber is
about 1 × 1016 . The ratio of the two resistances is about 6 × 1021, or 6 followed by
21 zeros. Resistivity can be specified in other units, such as ohms-circular mil/ft (-
CM/ft).
Since the resistances of insulators are ordinarily very large numbers and the resistances
of other electrical components may be very small numbers, it is convenient to use prefix-
es to scale resistance values. Table 2.2 shows the common values encountered in dc de-
sign work. For example, if an insulator has 10,000,000 resistance, the equivalent resis-
tance is 10 M, and if a conductor has 0.001 resistance, the equivalent resistance is 1
m.
4
The current unit, ampere (A), was named after André Marie Ampère, a French research mathematician, who
published a theory of electrodynamics phenomena in 1826. Ampere frequently is shortened to amp. The voltage
unit, volt, was named after Alessandro Giuseppe Antonio Anastasio Volta, who invented the voltaic pile, or bat-
tery, in 1800.
5
The power unit, watt (W), was named after James Watt, a Scottish instrument maker, who produced the first ro-
tary-motion steam engine in 1781. The energy unit, joule (J), was named after James Joule, an English physicist,
who in the mid-1800s conducted experiments to demonstrate the unity of heat and motion forces.
6
Annealed copper has been heated and held at a certain temperature to allow full recrystalization and then
cooled to make it less brittle.
2.4 RESISTANCE 35
Table 2.1 Relative Conductance

Conductor Relative Conductance (%)
Silver 100
Copper 94
Gold 66
Aluminum 62
Tin 14
If a conductor has a constant cross section, its resistance will vary directly with the
conductor length. If a conductor has a constant length, its resistance will vary inversely
with its cross-sectional area. These two properties are summarized in
L
R = (2.1)
A
where R = resistance ()

= resistivity of a cube of material with a face dimension of 1 cm or 1 m (-cm or -m,
respectively)
L = length of the conductor in the direction of current flow (cm or m)
A = cross-sectional area at right angles to the direction of current flow (cm2 or m2).
In the above expression, the same dimensional units must be used throughout (such as all
centimeters or all meters).
Example 2.1 Determine the resistance of a 100-cm length of copper wire with a cross-
sectional area of 0.133 cm2 and a temperature of 20°C (68°F).
L 100
R = · = (1.7241 × 10–6) = 0.00130 , or 1.3 m
A 0.133
2.4.2 Temperature Coefficient of Resistance

The resistance of copper (and other nonalloyed metals) increases appreciably with tem-
perature. The temperature at which electrical conductors operate varies with the current
and depends on the ambient temperature and installation conditions. For example, for a
Table 2.2 Common Resistance Units

Name Multiplication Factor
microhm () 1 /1,000,000
milliohm (m) 1 /1,000
ohm () 1
kilohm (k) 1,000
Megohm (M) 1,000,000
given current, a wire in conduit will operate at a higher temperature than the same wire in
free air because the wire in free air is better able to dissipate the heat developed during
operation.
Over a limited temperature range, the resistance of copper wire is a linear function of
the temperature. If the resistance at a temperature is known, the resistance at any other
temperature can be determined from
RT = R0[1 + 0(TT – T0)] (2.2)
where RT = resistance at temperature TT ()

R0 = resistance of copper of 100% conductivity at the initial or standard temperature ()
TT = temperature at which resistance is desired (°C)
T0 = initial or standard temperature (°C)
0 = temperature coefficient of resistance corresponding to T0 (/°C)
At 20°C, the temperature coefficient for copper is 0.00393/°C.7 The coefficient at other
temperatures over the temperature range of 0 to 75°C may be determined from Figure 2.3.
The curve is extrapolated using a trend function above 50°C.
Example 2.2 Determine the resistance of No. 14 AWG copper wire in ohms/foot at
75°C. The resistance at 20°C is 0.00252 /ft.
RT = R0[1 + (TT – T0)] = 0.00252[1 + 0.00393(75 – 20)]

= 0.00306 /ft (over 20% increase)
Table 2.3 gives copper conductor resistivity at various temperatures in ohm-circular

mil/ft.
Copper melts at around 1100°C (2000°F) but wire is never operated even close to this
temperature. Wire operating temperatures are limited by insulation and the terminal tem-
perature ratings. High heat thermoplastic and cross-linked polyethylene are the most com-
mon insulations and their maximum operating temperatures are usually 60, 75, or 90°C
(140, 167, or 194°F, respectively) depending on the insulation material and working envi-
ronment (dry or wet). Operating insulated wire above their maximum operating tempera-
tures can damage the insulation.
2.4.3 Resistance in Series

A series circuit is one in which the resistor (or other electrical component) terminals are
connected end-to-end (Fig. 2.4). In such a circuit, the same current flows through each
component but the applied voltage divides among the components. If a number of resis-
tors having resistances R1, R2, R3, . . . , RN are connected in series, the total (or equivalent)
resistance of the combination is
RT = R1 + R2 + R3 + . . . + RN (2.3)
The temperature coefficient of resistance for most nonalloyed metals is around 0.004 /°C.
7
2.4 RESISTANCE 37
Fig. 2.3 Temperature coefficient of resistance for copper.
Example 2.3 Determine the equivalent resistance of three resistors, with values R1 = 1.0
, R2 = 1.5 , and R3 = 2.0 , connected in series. From Eq. (2.3), the resistance is
RT = R1 + R2 + R3 = 1.0 + 1.5 + 2.0 = 4.5
2.4.4 Resistance in Parallel

A parallel circuit is one in which one terminal of each resistor (or other electrical compo-
nent) is connected to a common point and the other terminal is connected to a second
common point (Fig. 2.5). Under these conditions, each component of the parallel system
Table 2.3 Resistivity of Copper Building Wire

Resistivity (-CM/ft) Temperature (°C) Temperature (°F)
10.7 20 68
11.1 30 86
11.6 40 104
11.8 50 122
2.3 60 140
12.7 70 158
12.9 75 167
Fig. 2.4 Resistors in series.
is across the same applied voltage, but the total current divides among them. If a number
of resistors are connected in parallel, the equivalent resistance of the combination is
1
RT = (2.4)
1 1 1 1
+ + + · · · +
R1 R2 R3 RN
For the case of two resistors in parallel
R1R2
RT = (2.5)
R1 + R2
For three resistors in parallel
R1R2R3
RT = (2.6)
R1R2 + R2R3 + R1R3
Example 2.4 Determine the equivalent resistance of three resistors, with values R1 = 1.0
, R2 = 1.5 , and R3 = 2.0 , connected in parallel. From Eq. (2.6), the resistance is
R1R2R3 (1.0)(1.5)(2.0)
RT = = = 0.462
R1R2 + R2R3 + R1R3 (1.0 × 1.5) + (1.5 × 2.0) + (1.0 × 2.0)
In a parallel resistance circuit, the equivalent resistance always is less than the smallest re-
sistance in the circuit. Remembering this helps to quickly gauge whether a calculation
was done correctly. In the above example, the equivalent resistance is about 0.5 (round-
ed), which is less than the smallest resistance (1.0 ) in the circuit.
The concept of series and parallel resistors tracks perfectly with the connection of
wires in series and parallel circuits. Wires have resistance, and an example of a series cir-
cuit has feed and return conductors from a battery to the primary distribution bus, from
the primary distribution bus to the secondary distribution bus, and from the secondary dis-
tribution bus to the load (Fig. 2.6).
Fig. 2.5 Resistors in parallel.

2.5 WIRE AND RIGID BUSBAR CONDUCTORS 39
(a)
(b)
Fig. 2.6 Series circuit using conductors: (a) physical circuit and (b) equivalent electrical circuit.
Similarly, an example of a parallel circuit has feed and return conductors connected in
parallel to reduce the overall resistance and voltage drop in the circuit (Fig. 2.7).
2.5 WIRE AND RIGID BUSBAR CONDUCTORS
In the United States, the American Wire Gauge (AWG) and circular mil are used to de-
scribe wire size.8 AWG is based on a constant ratio between diameters of successive
gauge numbers; that is, the diameters taken in order form a geometrical progression. The
diameter of No. 4/0 is defined as 0.4600 in. (11.68 mm) and the diameter of No. 36 as
0.0050 in. (0.127 mm). There is a span of 39 gauge numbers between No. 4/0 and No. 36.
Hence, the ratio of any diameter to the diameter of the next greater gauge number is
冪莦
39 0.4600 39
= 兹9 苶2
苶 = 1.1229
0.0050
Since the cross section of a circular conductor varies as the square of its diameter, the ra-
tio of any cross section to the cross section of the next greater gauge number is (1.1229)2
= 1.2610. The sixth power of 1.1229 is 2.0050, which for practical purposes is 2. There-
fore, the ratio of any diameter to the diameter of a gauge number that is 6 gauges greater
is 2. It follows that the cross section either doubles or halves for every three gauge num-
bers.
The smallest wire size normally used in dc power wiring is 16 or 18 AWG. Wires larg-
er than 4/0 AWG are measured in circular mils or kilocircular mils.
8
In most places outside of the United States, metric sizes in square millimeters (cross-sectional area) or millime-
ters (diameter) are used.
(a) (b)
Fig. 2.7 Parallel circuit using conductors: (a) physical circuit and (b) equivalent electrical circuit.
The circular mil is a measure of wire cross section. A mil is one-thousandth of

an inch, or 0.001 in. A circular mil is the area of a circle whose diameter is 1 mil (Fig.
2.8). The area in square inches of 1 circular mil is (/4) · (0.001)2 = 0.7854 × 10–6 in.2.
The advantage of the circular mil is that circular areas measured in terms of this unit
have a very simple relation to their diameters. Also, with the circular mil as the unit, the
factor does not enter computations of cross sections. The general relation may be
written
D2in.
CM = 2 = 1,000,000(Din.)2 = D2 (2.7)
(0.001)
where CM = circular mils (upper case is used to avoid confusion with the abbreviation
for centimeters)
Din. = diameter of wire (in.)
D = diameter of wire (mils)
The concept of circular mils may be summed up in two rules:
앫 To obtain the number of circular mils in a circular conductor of given diameter, ex-
press the diameter in mils, then square it.
앫 To obtain the diameter of a conductor having a given number of circular mils, take
the square root of the circular mils, and the result will be the diameter in mils. Di-
vide the diameter in mils by 1000 to obtain the diameter in inches.
Fig. 2.8 Circular mil.

Example 2.5 A 6-AWG wire has a diameter of 0.162 in. Determine its cross-sectional
area in circular mils:
0.162 in. = 162 mils

(162)2 = 26,244 CM
Example 2.6 Determine the diameter of a wire with a cross-sectional area of 250,000 CM:
苶5
Diameter = 兹2苶0
苶,0
苶0苶0
苶 = 500.0 mils = 0.500 in.
The next standard wire size larger than 4/0 AWG is 250,000 CM. Wire sizes generally in-
crease in 50,000-CM increments above 250,000 CM in building wire applications. Com-
mon sizes are 350,000, 400,000, 500,000, and 750,000 CM. It is more convenient to state
these wire sizes in terms of the kilocircular mil, or kcmil, so that 250,000 CM is 250
kcmil and 750,000 CM is 750 kcmil.9
The unit circular mil–foot can be used to make quick approximations of wire resis-
tance. The resistance of a circular mil–foot of standard copper at 20°C is 10.371 . For
approximations, this is rounded to 10 with an error of about –3.6%. Knowing this resis-
tivity value, the resistance of any length and size of conductor may be determined from
Eq. (2.1), where is the resistance of a circular mil–foot of copper (10 , approximately),
L is in feet and A is in circular mils.
Example 2.7 Determine the resistance of a 750,000-CM copper wire that is 250 ft long.
This example can be solved by direct application of Eq. (2.1):
250
冢
R = 10 = 0.00333
750,000 冣
or by thought: A 250-ft wire that has a cross section of 1 CM has a resistance of about 10
× 250 = 2500 . However, the cross section of the wire in question actually is 750,000
CM; therefore
2500
R = = 0.00333
750,000
A wire can consist of one or more strands. If it consists of one strand, it is called a sol-
id wire, and if more than one strand, it is called a stranded wire. Solid wires typically are
not used in telecommunications dc power systems because of their stiffness. Stranded
wires, which are more flexible, are used exclusively. Two types of stranding generally are
used—coarse strand (also called concentric strand) and fine strand. Coarse strand is built
in concentric layers (or lays) according to the simple geometric properties of circular
wires. Six strands will just fit around a center strand of the same diameter. Each addition-
al strand layer has 6 more wires than the previous layer. Thus, the first layer has 6 wires,
the second has 12 wires, the third layer has 18 wires, and so on. Following this basic con-
struction rule, concentric strand wires will contain 1, 7, 19, 37, 61, 91, 127, . . . strands.
Figure 2.9 shows the cross section of a 19-strand wire and Table 2.4 shows the number
of strands used in common concentric strand building wire. Concentric strand conductors
9
Some older conductor data shows the unit MCM, which is the same as kcmil.
Fig. 2.9 Cross section of 19-strand building wire.
Table 2.4 Typical Coarse-Strand (Class B) Building Wire Strand Counts

Wire Size Number of Strands Strand Diameter (mils)
14–2 AWG 7 24.2–97.4
1–4/0 AWG 19 66.4–105.5
250–500 kcmil 37 82.2–116.2
600–1000 kcmil 61 95.0–128.0
1250–1500 kcmil 91 109.9–128.4
1,750–2,000 kcmil 127 112.2–125.5
used in telecommunications dc power systems typically (but not always) are built accord-
ing to American Society for Testing and Materials (ASTM) B8 Class B requirements
[1].10
The diameter of a concentric strand conductor is greater than a solid conductor of the
same gauge due to the stacking of the strands. For example, a 4/0 AWG solid wire is
0.460 in. in diameter while a 4/0 AWG 19-strand wire is 0.528 in. in diameter. However,
the effective cross-sectional area (the total area of the individual strands) of solid and
stranded wires is the same.
Fine-strand wire is used to obtain greater flexibility. With fine-strand wire the unit
strands themselves are also stranded with very fine wire. Fine-strand wire is built much
like a rope (and is called a rope-lay construction). Fine-strand wire typically is used in
sizes larger than 4/0 AWG where installation may be difficult or damage to the insulation
is possible due to many bends. Because large fine-strand wire is easier to handle and sup-
port, there is less chance of insulation or conductor damage during installation. Fine-
strand wire also is used for equipment connections in seismic areas where the stiffness of
coarse strand wire may damage terminations during a seismic event.
Fine-strand wire used in telecommunications dc power systems typically (but not al-
ways) is built according to ASTM B172 Class I requirements [2]. Class I stranded con-
ductors consist of individual 24 AWG wires. Other B172 classifications use 30 or 34
AWG wires. Figure. 2.10 shows the cross section of a 105-strand wire made up of 7 mem-
bers of 15 bunch-stranded wires (the strands are bunched into a circular cross section).
Table 2.5 shows typical strand counts for fine-strand wire used in telecommunications dc
power applications (ASTM B172 allows other constructions and strand counts).
It is important to understand how the different stranding configurations affect the con-
ductor diameter and the selection of a connector lug for terminating the conductor. The
10
Numbers in brackets indicate reference numbers at end of chapter.
Fig. 2.10 Cross section of 105 (7 × 15) strand, rope-lay, bunch-stranded wire.
following chart shows the number of strands and diameter of a 4/0 AWG conductor in
different ASTM classifications:
ASTM Class B C D G H I K M
No. Strands 19 37 61 133 259 532 2,107 5,320
Diameter (in.) 0.528 0.529 0.530 0.599 0.601 0.613 0.627 0.645
The diameter change for the 4/0 AWG class B, C, and D concentric strand conductors
is small and the same connector lug will fit them all. However, the class G through M
rope-lay, bunch stranded conductors have significantly larger diameters and most likely
will require different lugs. Some connector lug manufacturers flare the barrel to make it
easier to insert the fine-strand wire.
In many applications, wire strands are coated with a thin layer of tin or lead to reduce
tarnishing from insulation materials or the air (tarnish is a form of surface oxidation or
corrosion and acts as an insulator). The physical and electrical characteristics of tinned
wire are slightly different than uncoated wire. Some network operators use only tinned
wire in telecommunications dc power systems.
Table 2.5 Typical Fine-Strand (ASTM B172, Class I) Wire Strand Countsa
Wire Size Number of Strands Strand Layout Strand Diameter (mils)
6 AWG 63 1×7×9 20.1
4 AWG 105 1 × 7 × 15 20.1
2 AWG 161 1 × 7 × 23 20.1
1/0 AWG 266 1 × 19 × 14 20.1
2/0 AWG 342 1 × 19 × 18 20.1
4/0 AWG 532 1 × 19 × 28 20.1
250 kcmil 637 7 × 7 × 13 20.1
350 kcmil 882 7 × 7 × 18 20.1
500 kcmil 1225 7 × 7 × 25 20.1
750 kcmil 1862 19 × 7 × 14 20.1
1000 kcmil 2527 19 × 7 × 19 20.1
1250 kcmil 3059 19 × 7 × 23 20.1
1500 kcmil 3724 19 × 7 × 28 20.1
2000 kcmil 4921 19 × 7 × 37 20.1
a
The stand layout uses a simple nomenclature, A × B × C, where A = number of rope-stranded conductors, B =
number of bunch-stranded members that make up each rope-stranded member, and C = number of wire strands
in each bunch-stranded member.
Rigid, rectangular copper busbars are used in many installations in place of round wire
conductors. The busbars may be bare or tinned, as with stranded wire. Typical dimensions
are 2, 4, 6, and 8 in. wide by 1–4, 3–8, and 1–2 in. thick. Busbars normally are mounted with the
long dimension vertical for convection cooling (Fig. 2.11).
The cross-sectional area of rigid busbars is given in circular mils, but the calculation of
the circular mil area from bar dimensions is not as simple as for round wire:
Tin.Win. 4 4
(0.001) 冢冣
CMBusbar = 2 = 1,000,000(Tin.Win.)
冢冣 (2.8)
where Tin. = bar thickness (in.)

Win. = bar width (in.)
Example 2.8 Determine the cross-sectional area in circular mils of a 4-in. × 1–4-in. busbar
and determine the diameter of a round conductor with equivalent area. From Eq. (2.8)
0.25 in. × 4.0 in. 4

CMBusbar =
(0.001)2 冢冣
= 1,273,239.5 CM = 1273 kcmil

and solving Eq. (2.7) for the diameter
苶M
D = 兹C苶 = 兹1
苶,2
苶7苶3苶,2
苶3苶9苶.5
苶 = 1,128.4 mils = 1.128 in.
2.6 OHM’S LAW AND ELECTRICAL CIRCUITS
Ohm’s law states that, for a steady current, the current in a circuit is directly proportional
to the total voltage in the circuit and inversely proportional to the total resistance in the
circuit. This is summarized mathematically in
V
I= (2.9)
R
where I = current (A)

V = voltage (V)
R = resistance ()
Ohm’s law may be rewritten as
V = IR (2.10)
Example 2.9 Figure 2.12 shows a simple electrical circuit containing a resistor and a
battery. Determine the current.
The values from the illustration are plugged into Eq. (2.9) as follows:
V 50
I = = = 33.3 A
R 1.5
Example 2.10 Figure 2.13 shows an electrical circuit containing two conductors, an in-
dicating lamp, and a battery. The two conductors have equal resistances and their value is
2.6 OHM’S LAW AND ELECTRICAL CIRCUITS 45
Fig. 2.11 Normal busbar orientation.
shown along with the resistance of the indicating lamp. Determine the current and the
voltage across each of the conductors and the lamp.
The components are essentially three resistors connected in series, so using Eq. (2.3),
the total resistance is the sum of the individual resistance values, R1 + R2 + R3, or 1.5 + 1.5
+ 0.8 = 3.8 . Using Eq. (2.9) the current is
V 50
I = = = 13.2 A
R 3.8
The same current flows through each resistance. Equation (2.10) is used to find the volt-
age across the components. The subscripts in the following three expressions indicate the
voltage across each component due to the current though it:
V1 = IR1
V2 = IR2
V3 = IR3
The current was found previously as 13.2 A. Therefore, the voltages are
V1 = IR1 = 13.2 × 1.5 = 19.7 V
Fig. 2.12 Simple electrical circuit.

Fig. 2.13 Electrical circuit consisting of two resistors and a lamp.
V2 = IR2 = 13.2 × 0.8 = 10.6 V

V3 = IR3 = 13.2 × 1.5 = 19.7 V
A quick cross-check can be made by adding the individual voltage drops to give (19.7 +
10.6 + 19.7), or 50 V, which is the voltage of the battery.
Example 2.11 Three conductors with equal resistances are connected in parallel and to a
battery as shown in Figure 2.14. Determine the resistance of each conductor and the cur-
rent through it.
Equation (2.9) can be rearranged to solve for the resistance. Note that the voltage
across each conductor is the same. The resistances are equal, so the total current will di-
vide equally through each conductor. Since the total current is 15 A, the current through
each conductor is 5 A, and the resistance of each conductor is
V 50
R1 = = = 10
I1 5
A quick cross-check can be made by determining the equivalent resistance of the three
conductors in parallel and then solving for the current. From Eq. (2.4), the equivalent par-
allel resistance RP is
Fig. 2.14 Three equal resistances in parallel.

2.6 OHM’S LAW AND ELECTRICAL CIRCUITS 47
1 1 1
RP =
1 1 1
= = = 3.33
— +–
– — +–
— 0.1 + 0.1 + 0.1 0.3
10 10 10
The total current through the equivalent resistance is
V 50
IT = = = 15 A
RP 3.33
which is the original current that was given.

The above example illustrates the division of current in a parallel circuit, but it is sim-
plified by using equal resistances. If the resistances are different, the current will not di-
vide equally but, instead, will divide inversely as to the resistances.
Example 2.12 Two unequal resistances are connected in parallel as shown in Figure
2.15. Determine the current through each resistance.
The voltage across each resistance is the same, and the current through each resistance
is given by
V 50
I1 = = = 3.33 A
R1 15
V 50
I2 = = = 2.50 A
R2 20
The total current in the circuit is the sum of the two branch currents, or 5.83 A. This can
be checked by first finding the equivalent parallel resistance and then using Ohm’s law as
follows:
1 1 1 1
RP = 1
= = = 8.57
1 = 1 1
+ + 0.0667 + 0.0500 0.1167
R1 R1 15 20
Using Ohm’s law
V 50
IT = = = 5.83 A as expected
RP 8.57
Fig. 2.15 Two unequal resistances in parallel.

Fig. 2.16 Series–parallel circuit.
A circuit may consist of groups of parallel resistances in series with other resistances as
shown in Figure 2.16. In this case, each group of parallel resistances is replaced by its
equivalent parallel resistance and the entire circuit is then treated as a series circuit.
Example 2.13 In the circuit of Figure 2.16, determine: (a) equivalent resistance of each
parallel group of resistances; (b) total equivalent series resistance; (c) total current; (d)
voltage across each series component; and (e) current in each resistance. Use the follow-
ing values: R1 = 5 ; R2 = 10 ; R3 = 12 ; R4 = 15 ; R5 = 20 ; R6 = 25 ; and VT = 50
V. The circuit is redrawn in Figure 2.17 with all values shown.
(a) The equivalent parallel resistances of the 10- and 12- resistors and of the 15-, 20-,
and 25- resistors are
1 1
R10||12 = 1 1
= = 5.45
10 + 12 0.100 + 0.083
—
– —
–
1 1
R15||20||25 =
1 1 1
= = 6.38
15 + 20 + 25 0.0667 + 0.0500 + 0.0400
—
– —
– —
–
(b) The total series resistance is (5 + 5.45 + 6.38), or 16.83 .

(c) The total current is found from Ohm’s law using the total series resistance
Fig. 2.17 Example series–parallel circuit. Resistances are in ohms.

2.7 ELECTRICAL POWER 49
V 50
IT = = = 2.97 A
RS 16.83
(d) The voltage across each series component is found by solving Ohm’s law for volt-
age and using the equivalent resistance of each series component:
V1 = ITR1 = 2.97 × 5 = 14.85 V

VR10||12 = ITR10||12 = 2.97 × 5.45 = 16.19 V
VR15||20||25 = ITR15||20||25 = 2.97 × 6.38 = 18.95 V
A quick check shows the total voltage to be (14.85 + 16.19 + 18.95), or 50 V, as expected.
(e) Since the voltages across the resistances are now known, it is a simple matter to
calculate the current through them using Ohm’s law. The current through the 5 resis-
tance is the same as the total current:
I5 = 2.97 A
The currents through the other resistances are
16.19
I10 = = 1.62 A
10
16.19
I12 = = 1.35 A
12
18.95
I15 = = 1.26 A
15
18.95
I20 = = 0.95 A
20
18.95
I25 = = 0.76 A
25
Note that I10 + I12 is 2.97 A and I15 + I20 + I25 is 2.97 A, as expected.
Voltage drop is an important consideration in low-voltage dc power systems, and wire
sizes frequently have to be larger than the sizes required for current-carrying capacity to
compensate for it. Some older telecommunications equipment has a fairly narrow operat-
ing range and landline end-office switching systems cannot meet line supervision and sig-
naling requirements unless the voltage is above a certain value, so proper wire sizing is
one of the most important jobs in dc power system design.
2.7 ELECTRICAL POWER
In dc circuits, the power dissipated, P, is
P = VI (2.11)
A simple substitution from Eq. (2.10) gives
P = I 2R (2.12)
and substituting Eq. (2.9) in Eq. (2.11) gives
V2
P= (2.13)
R
If any two of the three quantities (voltage, current, and resistance) are known, the preced-
ing three equations can be used to solve for the third quantity. It is important to remember
that the above equations apply only to dc circuits and need to be modified slightly, as ex-
plained later, for ac circuits.
When 1 A flows between two points with a potential difference of 1 V, the power dis-
sipated is 1 W. The watt frequently is too small a unit, particularly when large amounts of
power are being considered. The kilowatt (kW), which is 1000 W, is frequently used. For
example, the modular rectifiers used in telecommunications commonly are rated 1500 or
3000 W, which are equivalent to 1.5 and 3.0 kW, respectively.
Power is the rate of doing work, or the rate of energy use. Therefore, electrical energy
is the product of electrical power and time:
E = Pt (2.14)
where E = electrical energy (Wh)

P = power (W)
t = time (h)
If power P in Eq. (2.14) is stated in kilowatts, the energy will be in kilowatt-hours (kWh).
Example 2.14 Determine the amount of energy consumed in 30 days by a steady load of
3000 W. There are a total of 720 h in a 30-day period, so the electrical energy consumed
is
E = Pt = 3000 × 720 = 2,160,000 Wh, or 2160 kWh
The amount of energy consumption is relevant when calculating the electric bill or
amount of fuel used by a standby engine–generator set. The rate of energy consumption,
or power, is relevant when calculating the size of the standby engine–generator set or
calculating circuit conductor sizes or the size of a transformer serving ac loads in a
building.
The components used in dc power systems are not perfect. Their efficiency is not
100%, and some of the energy used by the components is dissipated as heat. For example,
the efficiency of rectifiers at full load is around 90% (i.e., 10% of the power input is con-
verted in the rectifier to heat), some heat is dissipated in wiring between the rectifiers and
battery and the loads, and inverters and DC–DC converters may have low efficiency
when lightly loaded. Therefore, the power required by a dc system is always greater than
the power delivered by it to the loads.
2.7 ELECTRICAL POWER 51
Fig. 2.18 Example feeder circuit for power and voltage calculations.
Example 2.15 Figure 2.18 shows a feeder circuit consisting of positive and negative
conductors supplying a load. The feeder is connected to busbars operating at a constant
voltage of 27 V. The feeder is 100 ft long and consists of 4/0 AWG conductors operat-
ing at 30°C. The maximum load is 150 A. Determine: (a) Voltage drop from the bus-
bars to the load and the voltage at the load; (b) power dissipated in the feeder wires; (c)
power dissipated in the load; and (d) total power supplied and efficiency of power trans-
mission.
(a) From Chapter 5, the resistance of 4/0 AWG wire at 30°C is 0.05197 /1000 ft or
0.00005197 /ft. Since the current flows through both wires, there is a voltage drop in
both wires as determined by their resistance. Each wire is 100 ft long, so its resistance is
100 × 0.00005197, or 0.005197 . The total resistance of both wires is 2 × 0.005197, or
0.010394 . Using the wire resistance and the 150 A load in Eq. (2.10), the voltage drop
in each wire is 150 × 0.005197, or 0.780 V, and the total voltage drop in both wires is 150
× 0.010394, or 1.559 V. Therefore, the voltage at the load is the busbar feeder voltage less
the voltage drop in the feeder wires, or 25.44 V.
(b) From Eq. (2.12), the power dissipated in each feeder wire is
PW = I 2RW = (150)2 × 0.005197 = 116.9 W (233.8 W for both wires)
(c) The power dissipated in the load is determined from Eq. (2.11) using the load volt-
age [calculated in (a) above] and the load current
PL = VLI = 25.44 × 150 = 3816 W (or 3.816 kW)

(d) The total power supplied to the circuit has to equal the power consumed by it;
that is, 233.8 + 3816, or 4049.8 W. The power supplied can be calculated using Eq.
(2.11):
PT = VI = 27 × 150 = 4050.0 W
which is only slightly different because of rounding. The efficiency can be found by tak-
ing the ratio of the power delivered to the load to the power supplied by the source, or
PL 3816
Eff = = = 0.942 = 94.2%
PS 4050
2.8 CURRENT DENSITY
As mentioned previously in Section 2.5, a circular mil–foot of copper at 20°C has a resis-
tance of 10.371 (or, approximately, 10 ). A typical (or “normal”) current density for
wiring is 1 A per 1000 CM, or 0.001 A/CM. The approximate voltage drop through a cir-
cular mil–foot carrying 0.001 A is
V = IR = 0.001 × 10 = 0.01 V
Another circular mil–foot conductor, carrying 0.001 A, also will have a voltage drop of
0.01 V between its ends. If these conductors are placed side by side, the voltage drop
across the two still will be 0.01 V. With any number of conductors, each having 1 CM
cross section, a length of 1 ft, and a current of 0.001 A, the drop between the ends of each
conductor will be 0.01 V. The wires may be separated or they may be made into a cable.
Four separate conductors are shown on the upper part of Figure 2.19, each with a di-
mension of 1 CM-ft and each carrying 0.001 A. The voltage drop across each must be
0.01 V. On the lower part of Figure 2.19, the same four conductors are bundled together,
and since each is carrying 0.001 A, the total current is 0.004 A. The voltage drop across
the wire bundle is still 0.01 V. If any number of circular mil–foot conductors, each carry-
Fig. 2.19 1 Circular mil conductors carrying 0.001 A.

2.8 CURRENT DENSITY 53
ing 0.001 A, are added in parallel to this bundle the drop remains 0.01 V. The following
rule may be deduced from this discussion:
The voltage drop per foot of copper conductor whose resistivity is 10 (circular mil–foot) is
0.01 V if the current density is 0.001 A per circular mil (1 A per 1000 CM). Further, if the
current density is other than 0.001 A per circular mil, the voltage drop will be in direct pro-
portion to the current density.
Example 2.16 An equipment load of 110 A is located 75 ft (conductor path distance)

from the power bus. The bus is operating at 54 V. Determine the wire size required so the
voltage drop does not exceed 2.0 V.
For the wire to operate at the normal current density (1 A per 1000 CM), it must have
110 × 1000, or 110,000 CM. The total voltage drop is 0.01 × 75 × 2 = 1.5 V (note the to-
tal conductor length is twice the conductor path distance). The allowable voltage drop is
2.0 V, and so a proportionally smaller wire may be used:
1.5
冢冣
110,000 = 82,500 CM
2.0
The actual current density in the smaller wire is
110,000
= 1.33 A/1000 CM
82,500
which is one-third higher than the normal current density.
Another way of looking at this problem is as follows: The voltage drop per unit length
of conductor is known as the voltage gradient in the conductor and is the slope of the volt-
age graph similar to Figure 2.18. The relation between the voltage gradient and current
density is shown by combining Eq. (2.10) with Eq. (2.1), that is,
IL
V = IR =
A
from which
V I
=
L A 冢冣 (2.15)
Therefore, the voltage gradient (voltage drop/unit length) equals the product of the cur-
rent density (amperes per unit area) and the resistivity (ohms–unit length). As has been
shown for copper conductors, the resistivity is approximately 10 (circular mil–foot),
and the voltage drop per foot is 10 times the current density in amperes per circular mil.
The previous example may be solved from the voltage gradient point of view. By sub-
stitution in Eq. (2.15), the voltage gradient in the wire is
2.0 110
= 10
2 × 75 A 冢冣
Solving for area A gives
75
冢冣
A = 10 × 110 × 2 = 82,500 Cm
2.0
and the actual current density is
110
= 0.001333 A/CM (1.333 A/1,000 CM)
82,500
As before, this density is one-third higher than the “normal” density of 0.001 A per circu-
lar mil. The implication of this higher current density is that conductor heating will be
higher than normal and conduit or cable rack fill may have to be reduced to ensure the
heat is dissipated.
The above methods may be used to determine the power loss in a copper wire. At the
normal density, the approximate power loss is
P = I 2R = (0.001)2 × 10 = 0.00001 = 1 × 10–5 W per circular mil-foot
where P is the approximate power loss (it is approximate because the resistance is round-
ed to 10 ).
In general, the total power loss at the normal density is 0.00001 × CM-L where CM is
the conductor circular mil cross section and L is its length in feet. The actual power loss is
proportional to the square of the ratio of the actual to the normal current density; that is,
D2
P = P0 2 (2.16)
(D 0)
where P = actual power loss (W)

P0 = power loss at normal current density (W)
D = actual current density in amperes per 1000 CM
D0 = normal current density in amperes per 1000 CM
Example 2.17 Determine the power loss in the previous example.
P0 = 0.00001 × 82,500 × 75 × 2 = 123.75 W at the normal current density
The actual power loss is
1.333
冢冣 = 219.9 W
2
P = 123.75
1.000
As a check, the power loss is readily determined from the voltage drop and current. In this
example, the voltage drop is 2.0 V and the current is 110 A. Hence, the power loss in the
wire is
P = 2.0 × 110 = 220 W (slightly different due to rounding)
A current density of 1 A/1000 CM is equivalent to 1273 A/in.2 (a conductor with a cross

section of 1000 CM has a diameter of 31.6 mils = 0.0316 in., and its cross-sectional area
in square inches is
2.9 KIRCHHOFF’S CURRENT AND VOLTAGE LAWS 55
d2 (0.0316)2 1
= = in.2
4 4 1273
One way to control the temperature rise in the wire is to limit the current density or,
looked at another way, the lower the current density, the lower the temperature rise. For a
single copper wire, safe current densities are 1200 to 2000 A/in.2 (~1060 to 640 CM/A).
Where wire is cabled or bundled, it cannot dissipate heat as effectively, and the safe cur-
rent densities decrease to 1000 to 1600 A/in.2 (~1270 to 800 CM/A) or less depending on
the configuration. Temperature rise is related to how effectively the heat generated by re-
sistive losses can be dissipated to ambient air, and a low current density does not always
ensure a low temperature rise.
Where two conductors are joined, there is a certain amount of contact resistance and
heat is developed at the point of contact. Generally, the greater the clamping pressure and
the greater the contact area, the smaller the contact resistance will be. However, the nature
of the contact surfaces affects the contact resistance; smooth surfaces have less contact re-
sistance than rough surfaces. Contacts are designed such that for a given current, the con-
tact area is large enough to keep the contact resistance from being too high and causing
excess heating.
The current densities for conductor contacts are much less than for conductors them-
selves. For example, copper-to-copper spring contacts in switches are limited to 60 to 80
A/in.2 (21,200 to 15,900 CM/A), copper-to-copper clamped contacts are limited to 100 to
125 A/in.2 (12,730 to 10,180 CM/A), and fitted and screwed copper-to-copper contacts
are limited to 200 to 250 A/in.2 (6370 to 5090 CM/A) [3].
2.9 KIRCHHOFF’S CURRENT AND VOLTAGE LAWS
Kirchhoff’s laws provide a means for solving electrical networks that may not be easily
solved by simple application of Ohm’s law.11
1. Kirchhoff’s current law (or first law): In any electrical network, the algebraic sum
of the currents that meet at a point (or junction or node) is zero.
2. Kirchhoff’s voltage law (or second law): In any closed electrical circuit, the sum of
all the voltages and all the resistance drops, taken with their proper sign, is zero.
The first law may be expressed as follows:
冱 I = 0 at any junction (2.17)
The second law may be similarly expressed:
冱 V + 冱 IR = 0 in any closed electrical circuit (2.18)
Equation (2.18) is frequently written
冱 V = –冱 IR (2.19)
11
Gustav Robert Kirchhoff, a Prussian, developed the laws named after him in 1845.
Both Eq. (2.18) and Eq. (2.19) have the same meaning—the sum of voltage drops around
a closed loop is zero.
Kirchhoff’s current law is obvious:
The total current leaving a junction must equal the total current entering the junction.
In Figure 2.20, four currents, I1, I2, I3, and I4 meet at the junction. The directions of the
first three currents are toward the junction, so the currents have plus signs as they add to
the quantity of electricity at the junction. The direction of current I4 is away from the
junction, so it has a minus sign since it subtracts from the quantity at the junction. Ac-
cording to Kirchhoff’s current law,
I1 + I2 + I3 – I4 = 0
Example 2.18 Assume I1 = 5 A, I2 = 8 A, and I4 = 17 A. Determine I3.
I1 + I2 + I3 – I4 = 5 + 8 + I3 – 17 = 0
or
I3 = +17 – 5 – 8 = +4 A
where the plus sign indicates that the current direction is toward the junction as shown.
The second law is really another application of Ohm’s law, Eq. (2.9), and its basis is
straightforward:
Starting at a certain point in a circuit with a given voltage, and following the paths around
the circuit until reaching the starting point, the voltage must be the same. Therefore, the volt-
age sources encountered in the path must equal the voltage drop in the resistances, with
every voltage being given the proper sign.
Example 2.19 Two voltage sources, V1 = 24 V and V2 = 18 V, are connected in series

opposing (their positive terminals are connected together) as shown in Figure 2.21. The
voltage sources have internal resistances R1 = 2 and R2 = 1 , respectively, and the ex-
ternal resistance RExternal = 5 . Determine the current and voltage at each part of the
circuit.
Fig. 2.20 Kirchhoff’s current law at a junction.

Fig. 2.21 Voltage variation in a circuit.
Since the two voltage sources act in opposition, the net voltage of the two is 24 – 18 =
6 V. The current is
V1 – V2 24 – 18 6
I = = = = 0.75 A
R1 + R2 + RExternal 2+1+5 8
First, consider point A as the zero volt reference point. This is an arbitrary reference point
for the purposes of this example, but in many real circuits, the point that is Earth ground-
ed is the zero reference point. In moving from A to B there is a 24-V rise in voltage due to
source V1. However, there is a voltage drop of 0.75 × 2, or 1.5 V, due to the internal resis-
tance of V1. Therefore, the net voltage at B is 24 – 1.5, or 22.5 V as shown in Figure 2.21.
From B to C there is a drop of 18 V due to going from the positive to the negative termi-
nal of V2 plus the 0.75-V drop due to the 1 internal resistance of V2. This makes the net
voltage at C = 22.5 – 18 – 0.75 = +3.75 V. From C to A there is a voltage drop of 3.75 V
due to the current of 0.75 A in the 5- resistance. When A is reached, the voltage has
dropped to zero. Therefore, as specified by Kirchhoff’s second law, the sum of all volt-
ages in the circuit, taken with their proper sign, is equal to the sum of the IR voltage drops
as follows:
Voltages IR Voltage Drops

V1 = +24 V V1 IR drop = –0.75 × 2 = –1.5 V
V2 = –18 V V2 IR drop = –0.75 × 1 = –0.75 V
Total = +6 V External resistance drop = –0.75 × 5 = –3.75
Total = –6 V
冱 Voltages = –冱 IR Voltage Drops = +6 = – (–6)
The current also may be found by applying Kirchhoff’s second law directly. Starting at
point A, which is assumed to be zero (reference voltage),
+24 – I × 2 – 18 – I × 1 – I × 5 = 0
I × 8 = +6 V
I = +0.75 A
A problem frequently encountered and a source of error with Kirchhoff’s second law is
the assignment of algebraic signs. Following these rules can minimize problems:
앫 A voltage rise should be preceded by a + sign

앫 A voltage drop should be preceded by a – sign
For example, while moving through a battery from the negative terminal to the positive
terminal, the voltage rises, and this voltage should be preceded by a + sign. On the other
hand, in going from the positive terminal to the negative terminal, the voltage drops, and
so a – sign should precede this voltage. These voltages are due to voltage sources. There-
fore, the sign preceding them is independent of the direction of the current.
In going through a resistance in the same direction as the current, the voltage drops.
Therefore, the voltage drop through a resistance in the direction of the current, whether
the internal resistance of a voltage source (such as a battery internal resistance) or an ex-
ternal resistance, should be preceded by a – sign. In going through a resistance in the op-
posite direction of the current, the voltage rises (negative voltage drop) and the voltage
should be preceded by a + sign. It should be noted that the algebraic signs preceding volt-
age drops or rises across resistances depend only on the direction of the current and are in-
dependent of the polarity of any voltage sources in the circuit.
In applying Kirchhoff’s laws to any circuit, the following conditions are necessary for
the solution:
앫 Kirchhoff’s current law must be applied to a sufficient number of junctions to in-

clude every unknown current at least once.
앫 Kirchhoff’s voltage law must be applied a sufficient number of times to include
every element in the circuit at least once.
Example 2.20 Three voltage sources (batteries) are connected as shown in Figure 2.22.
Assume the battery internal resistances are negligible compared to the external resis-
tances. Use the following circuit values: V1 = 22 V, V2 = 20 V, V3 = 24 V, R1 = 0.5 ,
Fig. 2.22 Application of Kirchhoff’s laws to solve for voltages and currents.
R2 = 3 , R3 = 1 , R4 = 1 . Determine the voltages and currents at all points in the

circuit.
Starting at point A and applying Kirchhoff’s voltage law to the path ABCDA, the fol-
lowing equation may be written:
+V1 – I1 × R1 – I2 × R2 + V2 – I1 × R4 = 0 (a)
Starting at point F and going along the path FEBCDF
–V3 + I3 × R3 – I2 × R2 + V2 = 0 (b)
Equations (a) and (b) provide two equations for determining three unknown quantities (I1,
I2, and I3). Three independent equations are necessary to solve for three unknowns. A
third equation may be obtained by applying Kirchhoff’s voltage law but (a) and (b) al-
ready meet the requirement that every element be included at least once. For example, a
third path FEBADF may be followed; however, when the third equation is combined with
either (a) or (b), the result is either (b) or (a), showing that another condition must be
placed on the circuit if a solution is to be found. This third condition must be obtained by
applying Kirchhoff’s current law to one of the junctions, say B:
+I1 – I2 – I3 = 0 (c)
The signs in (c) are based on the assumed current directions: I1 is toward the junction (+
sign) and I2 and I3 are away from the junction (– sign). This single equation meets the re-
quirement that every unknown current is included at least once.
With the three equations, (a), (b), and (c), it is possible to solve for the three currents.
The circuit is redrawn in Figure 2.23 with the numerical voltages and resistances shown.
Rewriting (a) with numerical values
+22 –0.5I1 – 3I2 + 20 – 1I1 = 0 (d)
and simplifying
1.5I1 + 3I2 = 42 (e)
Fig. 2.23 Example circuit with numerical values.

Similarly, for (b)
–24 + 1I3 – 3I2 + 20 = 0 (f)
or
1I3 – 3I2 = +4 (g)
And for (c)
+I1 – I2 – I3 = 0 (h)
or
I1 = I2 + I3 (i)
Substituting (i) in (e) gives
1.5(I2 + I3) + 3I2 = 4.5I2 + 1.5I3 = 42 (j)
Multiplying (g) by 3 and (j) by 2 gives
3I3 – 9I2 = 12 (g)
3I3 + 9I2 = 84 (j)
and then adding (g) and (j) gives
6I3 = 96 (k)
or
I3 = +16 A
Substituting this value in (g) gives
16 × 1 – 3I2 = +4
Reducing and solving for I2 gives
I2 = 4 A
From (i)
I1 = I2 + I3 = 4 + 16 = +20 A
Check by substituting the currents in (d) and (f)
+22 – 20 × 0.5 – 4 × 3 + 20 – 20 × 1 = 0
–24 + 16 × 1 – 4 × 3 + 20 = 0
As a further check, Kirchhoff’s second law may be applied to path ABEFDA, which was
not originally used in the above solutions, and the appropriate substitutions made
+V1 – (I1R1) – (I3R3) + V3 – (I1R4) = 0
+22 – (20 × 0.5) – (16 × 1) + 24 – (20 × 1) = 0
The application of Kirchhoff’s laws to solving circuit problems is greatly simplified,

and the chance of error is greatly reduced if systematic procedures are followed:
1. Make a clear diagram of the circuit. Letter the diagram and label every given quan-
tity.
2. Indicate the assumed direction of current in each element of the circuit. Designate
the currents in the different elements by I1, I2, I3, . . . IN.
3. Use Kirchhoff’s current law at each junction to write an equation involving the un-
known current. If the direction of the current is toward the junction, its value enters
the equation with a positive sign; if away from the junction, with a minus sign. Not
all junctions will yield a new relation between the unknown currents. The number
of unknown currents can be reduced by applying Kirchhoff’s current law to certain
junctions.
4. Use Kirchhoff’s voltage law to write an equation for each closed portion of the cir-
cuit (loop). Proceed as follows: Begin at any convenient point in a circuit loop and
pass completely around it in either direction. When a voltage source is passed over,
write down its value preceded with a sign that is opposite of the sign attached to the
polarity at the end of the source first encountered. When a resistance is passed over,
write the value of its voltage drop (IR drop) preceded by a minus sign if the direction
is the same as the current arrow through the resistance assigned in step 2 or a plus
sign if the direction is opposite. Set the algebraic sum of the values written down to
zero upon completing the loop. This process will yield more equations than are nec-
essary to find the solution to the problem. Verify the sign of each quantity.
5. Select from the equations written down in steps 3 and 4 as many independent ones as
there are unknown values.
6. Solve the equations algebraically for the unknowns by canceling terms or simplify-
ing.
Fig. 2.24 Example parallel battery circuit.

7. Verify the solutions by substituting in the equation of a circuit branch or path not
already used.
Example 2.21 Three batteries are connected in parallel as shown in Figure 2.24. The cir-
cuit includes battery internal resistances as well as external resistances. The arrows indi-
cate the assumed current directions. Determine the currents.
Since there are four unknowns, four independent equations are necessary. Applying
Kirchoff’s voltage law in circuit path ABDA, starting at A,
+30 – 3I1 + 2I2 – 30 = 0 (a)
In circuit DBCD, starting at D,
+30 – 2I2 + 1I3 – 30 = 0 (b)
and in circuit DBED, starting at D
+30 – 2I2 – 9I4 = 0 (c)
The equations given in (a), (b), and (c) each include a current that does not appear in any
other equation. A fourth equation, (d), can be obtained by applying Kirchoff’s current law
to junction B, giving
I1 + I2 + I3 = I4 (d)
or
I1 + I2 + I3 – I4 = 0 (e)
A number of different steps may be used to simplify and eliminate terms. One such proce-
dure follows:
1. Combine (a) and (b) to obtain (f). This expresses current I1 in terms of I3:
3I1 = I3 (f)
2. Simplify (a) to obtain I1 in terms of I2:
3I1 = 2I2 (g)
or
I1 = –32 I2 (h)
3. Simplify (b) to obtain I2 in terms of I3:

2I2 = I3 (i)
4. Substitute (f) in (e) and rearrange to obtain (j):
I1 + I2 + 3I1 – I4 = 4I1 + I2 – I4 = 0 (j)
5. Multiply (j) by +9 and combine with (c) to obtain (k):
–36I1 – 9I2 + 9I4 = 0 (j)
–2I2 – 9I4 = –30 (c)
–36I1 – 11I2 = –30 (k)
6. Substitute (h) in (k) and solve for I2:
–36(2–3 ) I2 – 11I2 = –30
or
–24I2 – 11I2 = –30
and
–35I2 = –30
or
30
I2 = + = +0.857 A (l)
35
7. Substitute (l) in (i) and solve for I3:
I3 = 2I2 = 2 × 0.857 = +1.714 A (m)
8. Substitute (l) in (h) to and solve for I1:
I1 = –32 × I2 = –32 × 0.857 = +0.571 A (n)
9. Solve (d) for I4:
I4 = I1 + I2 + I3 = 0.857 + 1.714 + 0.571 = 3.142 A
Since the currents all have positive signs, the assumed directions are correct. As a quick
check of the solutions, use circuit path ABCDA, starting at A,
+30 – 3I1 + 1I3 – 30 = 0 (m)

Substituting I1 and I3 in (m) gives
+30 – 3 × 0.571 + 1 × 1.714 – 30 = 0
thus verifying the solutions.

When using Kirchhoff’s laws, the direction of the currents must be assumed, but it
does not matter which direction is used as long as the same direction is used throughout
the analysis. If the assumed direction is not the actual direction, the current will have a
minus sign when the equations are solved, thus indicating the current actually is in the
other direction.
Example 2.22 The circuit of Figure 2.24 is redrawn in Figure 2.25 to show the assumed
directions of I1 and I2 to be opposite of the previous example. Determine the currents I1
through I4.
The circuit is solved using the same procedures as the previous example. Applying
Kirchoff’s voltage law in circuit path ABDA, starting at A,
+30 + 3I1 – 2I2 – 30 = 0 (a)
In circuit DBCD, starting at D,
+30 + 2I2 + 1I3 – 30 = 0 (b)
and in circuit DBED, starting at D,
+30 + 2I2 – 9I4 = 0 (c)
Applying Kirchoff’s current law to junction B gives
I1 + I2 + I4 = I3 (d)
or
+I1 + I2 – I3 + I4 = 0 (e)
Fig. 2.25 Example battery circuit with I1 and I2 current directions reversed.
Using the same basic procedures as before:
1. Combine (a) and (b) to obtain (f):
3I1 = –I3 (f)
2. Simplify (a) to obtain I1 in terms of I2:
3I1 = 2I2 (g)
or
I1 = 2–3 × I2 (h)
3. Simplify (b) to obtain I2 in terms of I3:
2I2 = –I3 (i)
4. Substitute (f) in (e) and rearrange to obtain ( j):
+I1 + I2 + 3I1 + I4 = +4I1 + I2 + I4 = 0 ( j)
5. Multiply (j) by +9 and combine with (c) to obtain (k):
+36I1 + 9I2 +9I4 = 0 (j)
+2I2 – 9I4 = –30 (c)
+36I1 + 11I2 = –30 (k)
6. Substitute (h) in (k) and solve for I2:
+36(2–3 ) I2 + 11I2 = –30
or
+24I2 + 11I2 = –30
and
+35I2 = –30
or
I2 = – 30
35 = –0.857 A
–– (l)
7. Substitute (l) in (i) and solve for I3:
I3 = –2I2 = –2(–0.857) = +1.714 A (m)

8. Substitute (l) in (h) to and solve for I1:
I1 = 2–3 I2 = 2–3 (–0.857) = –0.571 A (n)
9. Solve (d) for I4:
I4 = –I1 – I2 + I3 = –(–0.571) – (–0.857) + 1.714 = +3.142 A
In this example, the currents I1 and I2 have minus signs indicating the assumed
current directions are opposite to the actual directions, as is apparent from the
original example. The signs of I3 and I4 are positive indicating their assumed di-
rections are correct. As a quick check of the solutions, use circuit path ABCDA
starting at A:
+30 – 3I1 – 1I3 – 30 = 0 (c)
Substituting I1 and I2 in (c) gives
+30 – 3(–0.571) – 1(+1.714) – 30 = 0
thus verifying the solutions.
Kirchhoff’s laws provide a set of procedures for determining voltages and currents in a
circuit. Other methods, such as Maxwell’s mesh equations and network analysis tech-
niques, are available to solve the same types of problems and, in many cases, are easier to
use when the circuits are complex. References 4 and 5 provide additional detail. However,
telecommunications dc power systems and the associated circuits usually are quite simple,
and Kirchhoff’s current and voltage laws and Ohm’s law are sufficient for almost all pur-
poses.
2.10 ALTERNATING CURRENT
2.10.1 Applications
The basic circuit analysis methods described in the previous sections apply to ac circuits
but several additional considerations are necessary in order to use them properly. This
section provides a brief review of ac electricity and how it is applied and used in telecom-
munications power systems.
Two basic applications of ac circuits are covered in this section—powering the recti-
fiers from the facility ac distribution system and powering inverter loads. The ac distribu-
tion system itself, including the ac service and metering equipment, ac load centers, and
ac utilization equipment (such as air conditioners, lights, convenience outlets), are not
covered. These ac systems are not unique to telecommunications dc power systems and
are covered in References 6 and 7.
Except in remote sites where the prime power system generates direct current,
telecommunications buildings and enclosures are powered by alternating current. Alter-
nating current is most common because it can be generated economically and its voltage
can be readily raised and lowered so that energy can be transmitted economically over
considerable distances.
2.10 ALTERNATING CURRENT 67
Fig. 2.26 Sine wave.
2.10.2 Sine Wave Voltage and Frequency

An ac generator converts mechanical energy into electrical energy by the principle of
electromagnetic induction. Unlike dc, where the voltages and currents have constant or al-
most constant values at all times, the values of the alternating voltage or current vary with
time and can be represented by a smooth curve called a sine wave (Fig. 2.26).
The voltage sine wave curve is a graph of the equation
v = Vm sin(2ft) (2.20)
where v = instantaneous voltage

Vm = maximum voltage
= pi (3.14159 . . .)
f = frequency (Hz)
t = time (s)
sin = trigonometric sine function
The current sine wave is identical in shape, but the values of i and Im are substituted for v
and Vm in Eq. (2.20). When the voltage or current wave goes through one complete set of
positive and one complete set of negative values, it has completed one cycle. A wave of
frequency f completes one cycle in 1/f seconds. The unit of frequency is the hertz (Hz),
which indicates the number of complete cycles occurring in one second.12
12
The frequency unit, hertz, named after Heinrich Hertz, a German physicist in the late 1800s, replaced cycles
per second (cps) in the 1960s.
In North America and some other parts of the world, the frequency of ac power sys-
tems is standardized at 60 Hz; that is, the voltage wave completes 60 cycles in one sec-
ond. Other frequencies, such as 50 Hz, are used in other parts of the world.
Example 2.23 A 60-Hz voltage sine wave has a maximum value of 100 V. Determine
the instantaneous voltage 2.083 ms after its zero crossing.
The quantity 2ft in Eq. (2.20) is in radians so, if a calculator is used to solve this
problem, it must be set to calculate trigonometric functions with angle inputs in radians.
To convert radians to degrees, multiply by the quantity 360/(2) ⬇ 57.3° per radian.
Assuming the voltage is rising after the zero crossing, its value after 2.083 ms will be
v = Vm sin(2ft) = 100 sin(2 × 3.14159 × 60 × 0.002083) = 70.7 V
There are three important values associated with voltage or current sine waves:
앫 Instantaneous value—designated as v or i—may be any value between zero and

maximum depending on the instant chosen.
앫 Maximum value—designated as Vm or Im—the peak value of voltage or current
reached twice in each cycle (one positive and one negative) when the instantaneous
voltage or current is at its maximum.
앫 Effective value—designated as Vrms or Irms (or just V or I)—the ac voltage or current
value that produces as much power or heating effect as a dc voltage or current of the
same value. The effective value also is known as the root mean square, or rms, val-
ue. A voltage of 1 V rms across a resistance will produce the same average power as
1 Vdc. When discussing ac voltages and currents, the subscript rms usually is
dropped.
The effective value is the most common value used in ac circuit analysis. It is the value
measured with common multimeters and the value used to indicate service and utilization
voltages.
The relationship between the effective and peak voltage values for sine waves is (sub-
stitute Im and I when the relationship between effective and peak currents is desired)
Vm = 兹2
苶V = 1.414V (2.21)
Similarly,
1
V = Vm = 0.707Vm (2.22)
兹2苶
Example 2.24 Determine the peak voltage when V = 120 V and V = 208 V.
Vm = 1.414V = 1.414 × 120 = 169.7 V for rms voltage of 120 V

= 1.414V = 1.414 × 208 = 294.1 V for rms voltage of 208 V
When the current or voltage waveform is distorted, it is not a pure sine wave but is com-
posed of the fundamental frequency and harmonic frequency components. In this case, the
rms value of the distorted wave includes the added effects of the harmonics and for current
is calculated from (the rms value for a distorted voltage wave may be similarly expressed)
冪莦莦
n
I= 冱 (I )
h=1
h
2
(2.23)
where Ih = rms current value of harmonic h

n = number of harmonics in calculation
Equation (2.23) may be expressed as:
I = 兹(I
苶苶1)苶+
2
苶苶I
(苶)
2苶苶
2
+苶·苶·苶·苶
+苶I
(苶)
n苶
2
(2.24)
The total rms current is calculated as follows:
앫 Square each current component including the fundamental and harmonics.

앫 Add the squared values together.
앫 Take square root of the result.
Example 2.25 Find the total rms current if the first (fundamental) through seventh har-
monics are 10, 1, 0, 3, 3, 0, and 1 A, respectively.
Squares of each component = 100, 1, 0, 9, 9, 0, 1

Sum of the squares = 100 + 1 + 0 + 9 + 9 + 0 + 1 = 120 A2
Square root = 10.95 A
Other terms are of interest when discussing ac voltage and current amplitudes are peak-to-
peak and crest factor. Peak-to-peak amplitude is the total amplitude of an ac waveform
measured from maximum positive to maximum negative peaks. The crest factor of an ac
waveform is the ratio of its peak (crest) to its rms value. Crest factor affects the measure-
ments of distorted waveforms with an average value responding meter. Analog meter
movements and inexpensive digital meters respond to the average value of an ac voltage
or current but typically are calibrated to display the rms value. While this calibration is
accurate for an undistorted sine wave, the meter will incorrectly indicate the rms value of
a distorted waveform. True rms meters are available for measuring distorted waveforms
but most only are able to accurately display the rms amplitude for waveforms with a crest
factor less than approximately 1.1 to 2.0 for handheld multimeters and 5.0 for laboratory
instruments. The crest factor of a pure sine wave is 1.414 (or 兹苶 2) and of a triangular
wave is 1.732 (or 兹苶3).
2.10.3 Phase
Voltages and currents in an ac circuit have the same fundamental frequency (60 Hz), but
they may not necessarily pass through their corresponding zero values at the same instant
in time. Figure 2.27(a) shows 60 Hz voltage and current sine waves with rms amplitudes of
120 V and 75 A (peak values of 169.7 V and 106.1 A, respectively). Both the voltage and
(a)
(b)
Fig. 2.27 Phase relationship of 60-Hz ac voltage and current: (a) in-phase and (b) out-of-phase by
30° or 0.00139 s.
current pass through zero amplitude at the same instant, increasing positively, and, there-
fore, are said to be in phase with each other. Figure 2.27(b) shows the same voltage and cur-
rent combination but not passing through zero at the same instant. The current waveform
passes through zero, increasing positively, later than the voltage waveform. Therefore, the
voltage and current are not in phase and, further, the current lags the voltage.
The time lag between the voltage and current shown in Figure 2.27(b) is 0.00139 s,
which corresponds to 30° of phase. Phase is represented by the angle . The relationships
between the various parameters are
1
360 f冢冣
t = seconds (2.25)
or
= 360ft degrees (2.26)
For example, at a frequency of 60 Hz, the time corresponding to 25° is
25 1
冢
360 冣冢 60 冣
= 0.00116 s
2.10.4 Power in Alternating Current Circuits

2.10.4.1 Voltage and Current In-Phase Calculating the power in a dc circuit is
simple—take the product of the dc voltage and dc current. In an ac circuit, the situation is
slightly more complicated. In an ac circuit, the product of the instantaneous voltage and
current determines the power at that instant (the instantaneous power). The peak power
occurs when the peak voltage and current occur at the same time, that is, when the voltage
and current are in-phase. The voltage and current in a circuit are in-phase when there is
only resistance in the circuit and, under this condition, the average power is given by
P = VI watts (2.27)
2.10.4.2 Voltage and Current Out-of-Phase by Angle If the voltage and

current differ in phase by an angle that lies between –90° and +90°, the average power
is given by
P = VI cos watts (2.28)
The quantity cos (trigonometric cosine function of angle ) is called the power factor
(PF) of the circuit. The power factor can never be greater than 1.0 and can be leading or
lagging based on the phase difference between the voltage and current. A lagging power
factor indicates the current lags the voltage wave and the circuit has more inductive reac-
tance than capacitive reactance. A leading power factor indicates the current leads the
voltage wave and the circuit has more capacitive reactance than inductive reactance.
Solving Eq. (2.28) for the power factor gives
P
PF = cos = (2.29)
VI
Example 2.26 The phase angle between the voltage and current in a circuit is 30°. De-
termine the power factor.
PF = cos = cos(30°) = 0.866
Inductive reactance can be thought of as the ac resistance of an inductor. An inductor can

be a coil of wire (with or without a magnetic core) or regular circuit wiring. When current
flows through the inductor it stores electrical energy in the resulting electromagnetic
field. With a steady dc current, the inductance has no effect, but with an alternating or
changing current, it opposes the change. An inductor acts like a short circuit (has low im-
pedance) at low frequencies and an open circuit (has high impedance) at high frequencies.
Inductance is measured in henrys13 (millihenry, microhenry). Inductive reactance, XL, is
proportional to the frequency f and inductance L and is given by
XL = (2 fL) ohms (2.30)
Similarly, capacitive reactance can be thought of as the ac resistance of a capacitor. A ca-

pacitor acts on the principle that when there is a voltage difference between two electrical
conductors, there is some charge storage on them. This sets up an electric field in the
space between the conductors. With a steady dc voltage applied to the plates, the capaci-
tance has no effect, but with an alternating voltage, the capacitance opposes the change. A
capacitor acts like an open circuit (has high impedance) at low frequencies and a short cir-
cuit (has low impedance) at high frequencies. Capacitance is measured in farads14 (micro-
farads, nanofarads, picofarads). Capacitive reactance, XC, is inversely proportional to the
frequency f and capacitance C and is given by
1
Capacitive reactance XC = ohms (2.31)
2fC
Example 2.27 A 10-F capacitor and 10-mH inductor are used in a circuit operating at
60 Hz. Determine the reactance of each component.
XL = 2fL = 2 × 60 × 10 × 10–3 = 3.77

1 1
XC = = = 265.2
2fC 2 × 60 × 10 × 10–6
2.10.5 Alternating Current Circuits

2.10.5.1 Circuit with Resistance Only Figure 2.28 shows an ac circuit contain-
ing only resistance. The instantaneous current and voltage are
i = Im sin(2ft) amperes (2.32)
v = Ri = RIm sin(2ft) volts (2.33)
13
The unit of inductance, henry (H), was named after Joseph Henry, an American scientist, in 1893.
14
The capacitance unit, farad (F), is named after Michael Faraday, an English electrical experimentalist, who de-
veloped the concept of electric and magnetic forces in 1845.
Fig. 2.28 Circuit with resistance only.
The relationship between the effective (rms) voltage and current is
V = IR volts (2.34)
When the circuit has resistance only, = 0° and cos = 1.0 (power factor is 1.0). For the
in-phase case, Eq. (2.28) reduces to Eq. (2.27) and
P = VI = I2R watts
Note that with resistance only, the ac circuit follows the same laws as a dc circuit in re-
gard to the relationships between the voltage, current, resistance, and power.
2.10.5.2 Circuit with Inductance Figure 2.29(a) shows a circuit with inductance
only. The current waveform in this circuit lags the voltage waveform by exactly 90°. Dur-
ing those periods when the current is increasing from zero to maximum, the energy re-
ceived from the power source is stored in the magnetic field of the inductance; during
those periods when the current is decreasing from its maximum value to zero, all the
stored energy is returned to the source. Therefore, over a complete cycle, the net energy
taken by a pure inductance is zero.
Practical inductors have some amount of resistance, as shown in Figure 2.29(b), so the
phase difference between the current and voltage is not quite 90°. All power expended in
such a circuit is accounted for in the resistance. The total impedance of this circuit is
苶2苶+
Z = 兹R 苶苶X 苶L2 ohms (2.35)
(a) (b)
Fig. 2.29 (a) Circuit with inductance only and (b) circuit with inductance and resistance.
Note that in the case of a pure inductance (in which case R = 0), Eq. (2.35) reduces to
Z = XL ohms
The voltage across a circuit with inductance and resistance is
苶2苶+
V = I兹R 苶苶X 苶L2 = IZ volts (2.36)
and
V V
I = = amperes (2.37)
兹R
苶苶+
2
苶苶
X L苶
2
Z
Equation (2.37) corresponds to Ohm’s law for the ac circuit—the current in an ac circuit
is directly proportional to the voltage across the circuit and inversely proportional to the
impedance of the circuit. Note also that the angle by which the current lags the voltage,
and thus the power factor of the circuit, is given by
R R
PF = cos = = (2.38)
兹R
苶2苶+
苶苶X 苶L2 Z
2.10.5.3 Circuit with Capacitance Figure 2.30 shows a circuit with capacitance
only. The current waveform in this circuit leads the voltage waveform by exactly 90°.
During those periods when the current is increasing from zero to maximum, the energy
received from the power source is stored in the electric field of the capacitance; during
those periods when the current is decreasing from its maximum value to zero, all the
stored energy is returned to the source. Therefore, over a complete cycle, the net energy
taken by a pure capacitance is zero.
2.10.5.4 Circuit with Resistance, Inductance, and Capacitance Figure 2.31

shows a circuit with resistance, inductance, and capacitance. The current waveform in this
circuit leads the voltage waveform by between +90° and –90° as determined by the rela-
tive amounts of inductive and capacitive reactance.
The total impedance of this circuit is
苶2苶+
Z = 兹R 苶苶X
(苶苶–
L 苶苶
X苶C)苶 ohms
2
(2.39)
Fig. 2.30 Circuit with capacitance only.

Fig. 2.31 Circuit with resistance, inductance, and capacitance.
The voltage across the circuit is
V = I 兹R
苶2苶+
苶苶X
(苶L苶–
苶苶X苶
C)苶 = IZ volts
2
(2.40)
and the current is
V V
I = = amperes (2.41)
兹R
苶苶+
2
苶苶X
(苶苶–
L 苶苶
X苶C)苶
2 Z
The power factor of the circuit is given by
R R
PF = cos = = (2.42)
兹R苶苶+
2
苶苶X
(苶L苶–
苶苶XC苶)苶
2 Z
Note that when XL = XC, the power factor is 1.0 because the net of the inductive and ca-
pacitive reactances is zero.
Example 2.28 Determine the 60-Hz power factor of a series circuit consisting of a 20-
mH inductor, 15-F capacitor, and a 100- resistor.
First, calculate the inductive and capacitive reactances:
XL = 2 × 60 × 20 × 10–3 = 7.54
1
XC = = 176.8
2 × 60 × 15 × 10–6
The total impedance is
Z = 兹R
苶2苶+
苶苶X
(苶L苶–苶苶
X苶C)苶 = 兹1
2
苶0苶0苶2苶+
苶苶7
(苶.5
苶4苶苶
–苶7
1苶6
苶.8
苶)苶2 = 兹1
苶0
苶,0
苶0苶0
苶苶+苶8
2苶,6
苶4苶9
苶 = 196.6
and
R 100
PF = = = 0.509
Z 196.6
2.10.6 Other Forms of Alternating Current Power

There are three types of power in ac circuits:
PApparent = VI volt-amperes (VA) (2.43)
PReal = VI cos watts (W) (2.44)
PReactive = VI sin volt-amperes-reactive (VAR) (2.45)
Equation (2.43), which is the product of rms voltage and current, has a measurement unit
of volt-amperes, or VA. It is called the apparent power because it may not indicate the ac-
tual power used or delivered if the circuit contains capacitance or inductance. Eq. (2.44) is
the real or active power in watts, or W. It is called real power because it is the power that
actually does work (causes an incandescent lamp to glow, charges a battery, or powers
load equipment). Finally, Eq. (2.45) is called reactive power. It indicates the power that is
stored in the circuit’s capacitance and inductance and is given in volt-amperes-reactive,
or VAR. The three types of power are related by
PApparent = 兹(P
苶苶Re苶)
al苶苶
2
+苶P
(苶苶
Re苶
ac苶
tiv苶
e)苶
2
(2.46)
Equation (2.46) may be written using a more formal nomenclature:
S = 兹P
苶2苶+
苶苶Q2苶 (2.47)
where S = apparent power

P = real power
Q = reactive power
Another way of defining power factor is the ratio of real to apparent power:
PReal
PF = (2.48)
PApparent
While Eq. (2.48) is equivalent to the definitions already given in terms of the phase angle
between the voltage and current (displacement power factor), the power ratio provides
clearer insight into the power factor correction used in all modern ferroresonant and
switch-mode rectifiers. In these rectifiers, the ac current drawn by the rectifiers normally
would be highly distorted and, as a result, would introduce relatively high harmonic cur-
rents into the ac power system. These harmonics distort the shape of the current wave-
form, which leads to what is called distortion power factor. Distortion power factor is the
ratio of 60-Hz current magnitude to the effective (rms) current magnitude. The total rms
current includes harmonics and will be larger than the current at the 60-Hz powerline fun-
damental frequency.
An ac load with high harmonic currents causes numerous problems in electric utility
transformers and building wiring, so all modern rectifiers have power factor correction
circuits that reduce the harmonics and thus improve the distortion power factor.
Example 2.29 The rms voltage across a circuit is 120 V and the rms current through it is
10.5 A. The current lags the voltage by 29°. Determine the apparent, real, and reactive
power and the power factor.
PApparent = VI = 120 × 10.5 = 1260 VA
PReal = VI cos = 120 × 10.5 cos(29°) = 1102 W
PReactive = VI sin = 120 × 10.5 sin(29°) = 611 VAR
PReal 1102
PF = = = 0.875 lagging
PApparent 1260
2.10.7 Single-Phase and Three-Phase Voltages

The phase angle concept applies to voltages in a multiphase system, such as a three-phase
electric service, and in a single-phase service as well. In the three-phase system, the volt-
ages have exactly 120° phase difference, and the voltages on the two legs of a single-
phase electric utility service have 180° phase difference.
Electric power is delivered through a service transformer or bank of transformers lo-
cated on or near the facility property. A simple transformer has a primary winding and a
secondary winding (Fig. 2.32, upper). The voltage transformation depends on the primary
and secondary turns ratio. For example, if the primary has 30 times more turns than the
secondary, or N/M = 30, the turns ratio is 30 : 1. If a voltage of 7200 V is applied to the
primary (a common distribution voltage), the secondary voltage will be 7200/30 = 240 V
(a common service voltage). In this type of service transformer, the secondary winding
has a center tap, so the ratio on each half of the secondary effectively is 60 : 1 and the
voltage from the center tap to each outer leg is 120 V (Fig. 2.32, lower).
The service transformers are step-down types that reduce the electric utility distribu-
tion voltage to a much lower utilization voltage. The voltages used in modern distribution
systems vary from 12,700 to 34,500 V line-line. These voltage levels are classified as
medium voltages. The typical line-neutral voltages of 7200 and 14,400 V have corre-
sponding line-line voltages of 12,470 and 24,940 V, respectively. In older areas, many
Fig. 2.32 Transformer schematics: upper, Simple transformer with N : M turns ratio; lower, trans-
former with center-tapped secondary.
distribution lines still use line-line voltages of 2400 and 4800 V. Many other voltages
have been used over the years. Single-phase services use one step-down transformer and
three-phase services may use one transformer with three integral windings (i.e., a three-
phase transformer) or two or three individual transformers (i.e., single-phase transform-
ers) depending on the configuration.
The service voltages at the service transformer secondary windings may be measured
from line-to-neutral or from line-to-line. With most single-phase services, there are two
lines (L1 and L2) and a neutral (total of three wires) and the voltages have a 180° phase
relationship (Fig. 2.33). The most common voltages are 240 V line-line and 120 V line-
neutral. The nomenclature “120/240 V” indicates the service voltage for a typical single-
phase service.
The general relationship between line-line (L-L) and line-neutral (L-N) voltages for a
single-phase service is
VL-L = 2VL-N or VL-N = 1–2 VL-L (2.49)
The advantage of a three-phase electrical service is in the efficiency of generation and dis-
tribution of large amounts of power compared to single phase. However, most ac loads in
telecommunications facilities are single phase and are connected line-neutral or line-line.
To maintain load balance, the individual loads are spread as evenly as possible across the
three phases.
Three-phase services may be connected in a delta or wye configuration, which refers
to the way the windings in the service transformers are presented schematically. Several
variations of the delta configuration have been used over the years, including un-
grounded and corner grounded delta. These normally are found only in industrial instal-
lations. One variation of the delta configuration that uses a center-tapped transformer to
serve line-neutral loads may be found in many older telecommunications facilities.
Either two or three transformers may be used (Fig. 2.34). The nomenclature for a three-
phase, four-wire delta service is “240/120 V” to differentiate it from a 120/240 V sin-
gle-phase service.
The three-phase, four-wire delta combines the characteristics of a single-phase service
for serving line-neutral loads such as lighting with the characteristics of a three-phase ser-
vice for serving three-phase loads such as motors and large rectifier systems. Line-neutral
Fig. 2.33 120/240 V, one-phase, three-wire service. The small schematic to the right of the trans-
former shows the phase relationship between the two legs L1 and L2.
Fig. 2.34 Three-phase, four-wire delta services.
loads are connected L1-N and L2-N and line-line loads are connected L1–L2, L2–L3, and
L1–L3.
The advantages of the three-phase, four-wire delta service voltage configuration is the
ease with which an existing single-phase service may be upgraded to three-phase by
adding one or two single-phase transformers and the fact that the added transformers al-
most always are smaller than the original single-phase transformer, thus providing some
cost savings to the electric utility. The disadvantages are the inherently unbalanced load
on the primary distribution lines and the potential problems and damage due to certain
types of electrical faults. Also, the voltages from line-neutral depend on the phase leg in
question. For L1-N and L2-N, the voltage is 120 V but for L3-N, the voltage is 208 V.
The latter is called the “wild” leg or high leg. All line-line voltages (L1–L2, L2–L3, and
L1–L3) are 240 V.
With a wye-connected three-phase service (the most common type of three-phase ser-
vice), there are three lines (L1, L2, and L3) and a neutral (a total of four wires), and the
voltages have a 120° phase relationship. There are two common service voltages with 208
V line-line and 120 V line-neutral and 480 V line-line and 277 V line-neutral (the latter
typically found in larger installations) (Fig. 2.35). The nomenclatures used to indicate
these service voltages are “208Y/120 V” and “480Y/277 V,” respectively. The general re-
lationship between line-line (L-L) and line-neutral (L-N) voltages for a three-phase wye-
connected services is
Fig. 2.35 Three-phase, four-wire, wye service configurations.
1
VL-L = 兹3 苶VL-N or VL-N = VL-L (2.50)
兹3 苶
A variation of the three-phase, four-wire service,
called a network service, is used to derive one-phase, three-wire services from a three-
phase transformer installation (Fig 2.36). In this case, the one-phase services are connect-
ed to any two phases and neutral, and the line-neutral voltages on both legs are 120 V and
the line-line voltage is 208 V. Such an installation is convenient and inexpensive to the
electric utility that has an existing three-phase service and needs to connect one-phase ser-
vices nearby.
The load power on a single-phase service is the sum of the power on the individual
legs, or
P1–
= (VA-N IAPFA) + (VB-N IBPFB) (2.51)
For a balanced single-phase load (equal line currents, line voltages, and power factors)
P1–
= 2VL-N ILPF = VL-LILPF (2.52)
The calculations of power in three-phase circuits are straightforward if the loads on each
phase are known. Assuming the loads are perfectly balanced, and in terms of line-neutral
Fig. 2.36 Network service derived from a three-phase transformer installation.
voltages, the power is
P3–
= 3VL-N ILPF (2.53)
or, in terms of line-line voltages,
P3–
= 兹3
苶VL-N ILPF (2.54)
Power can be calculated based on a single-phase representation of each of the three phas-
es. If the loads are not balanced
P3–
= (VL1-N IL1PFL1) + (VL2-N IL2PFL2) + (VL3-N IL3PFL3) (2.55)
2.10.8 Alternating Current Wiring

Wiring in ac circuits has slightly different electrical characteristics than the same wiring
in dc circuits because of skin effect and proximity effect and the inductive and capacitive
effects on ac current. Skin effect is the tendency for alternating currents to flow closer to
the surface of a wire with higher current density than in the center. Proximity effect is the
interaction of the alternating magnetic fields of conductors in close proximity. Skin effect
and proximity effect increase the resistance of the wire at power frequencies (60 Hz) and
is most noticeable in large wire sizes (> 250 kcmil).
The type of conduit, whether magnetic such as steel or nonmagnetic such as alu-
minum, used to enclose the wiring also affects the electrical characteristics. The power
factor of the load itself will affect the voltage drop in the wiring. However, except in mul-
tifloor or large sprawling buildings, ac voltage drops are small and detailed calculations
seldom are needed.
In those cases where it is desirable to perform ac circuit analysis of the type applica-
ble to rectifier and inverter wiring, the process can be simplified by assuming certain
conditions and using approximations. However, it is helpful to first look at a vector di-
agram of the voltage and current relationships in an ac circuit (Fig. 2.37). A vector dia-
Fig. 2.37 Vector diagram of ac circuit voltages and currents with lagging power factor.
gram shows the magnitude and phase angle of the source and load voltages and the line
current (the lengths are exaggerated in the illustration for clarity). The vector arrow
lengths are proportional to the magnitudes and the angles are shown directly. The exact
formula for the voltage at the load end of the circuit, which can be derived from the vec-
tor diagram, is
VLoad = 兹V
苶2苶
So苶
ur苶
ce苶
–苶IX c苶s苶苶苶
(苶苶o –苶R
I苶苶in
s苶苶) 苶2 – (IR cos + IX sin ) (2.56)
Usually, the parameter of interest is the actual voltage drop itself, which is approximated
by
VDrop = IR cos + IX sin (2.57)
where VDrop = line-neutral voltage drop in the circuit (V)

I = line current (A)
R = line ac resistance in one conductor ()
X = line reactance in one conductor ()
= angle between the load voltage and load current
cos = load power factor
sin = load reactive factor
The voltage drop calculated above is the drop in one conductor. For single-phase circuits,
VDrop must be multiplied by a factor of 2, and for three-phase circuits, must be multiplied
by 兹苶 3 = 1.732 as in
V1–
Drop = 2(IR cos + IX sin ) (2.58)
苶(IR cos + IX sin )

V3–
Drop = 兹3 (2.59)
REFERENCES 83
From the above formulas, the resistive voltage drop is in-phase with the line current and is
given by the term IR cos . The reactive voltage drop component, or IX sin , is 90° out of
phase with the current. The sign of sin depends on whether the load power factor is lag-
ging, in which case it is positive, or leading, in which case it is negative. Lagging power
factor is more common.
Example 2.30 The resistance in a single-phase circuit is 0.01 and the reactance is 0.02
. The line current is 9.1 A and the load power factor is 0.9. Determine the voltage drop.
First, determine the phase angle between the load voltage and current and then use that
to calculate the voltage drop. Since the power factor is 0.9, then cos = 0.9. Therefore,
angle is the inverse cosine of 0.9 and
= cos–1(0.9) = 25.8°
and
sin(25.8°) = 0.436
The voltage drop is
V1–
Drop = 2(IR cos + IX sin ) = 2[(9.1 × 0.01 × 0.9) + (9.1 × 0.02 × 0.436)] = 0.322 V
To simplify voltage drop calculations, wiring characteristics, including the effects of con-
duit, are summarized in tables in terms of voltage drop per 10,000 A-ft of wire with mul-
tiplication factors to adjust between single-phase and three-phase systems. The table val-
ues are then scaled to the actual circuit conditions being analyzed. Chapter 5 (System
Design) provides a table for this purpose and its use is described there.
REFERENCES
1. ASTM B8-99, Standard Specification for Concentric-Lay-Stranded Copper Conductors, Hard,

Medium-Hard, or Soft, 1999 (www.astm.org).
2. ASTM B172-01a, Standard Specification for Rope-Lay-Stranded Copper Conductors Having
Bunch-Stranded Members, for Electrical Conductors, 2001 (www.astm.org).
3. T. Croft and W. Summers, Eds., American Electricians’ Handbook, 14th ed., McGraw-Hill,
New York, 2002 (http://books.mcgraw-hill.com).
4. H. H. Skilling, Electrical Engineering Circuits, Wiley, New York, 1965.
5. M. E. Van Valkenberg, Network Analysis, 2nd ed., Prentice-Hall, Englewood Cliffs, 1965.
6. ANSI/IEEE Std 241™-1990, IEEE Recommended Practice on Electric Power Systems in Com-
mercial Buildings, IEEE, 1990 (www.ieee.org).
7. D. Beeman, Industrial Power Systems Handbook, McGraw-Hill, New York, 1955.
CHAPTER 3
dc POWER SYSTEM COMPONENTS
Telecommunications dc power systems are made from numerous components and equip-
ment groups, including rectifiers, powerboards with charge and discharge busbars, prima-
ry and secondary distribution systems (including overcurrent protection), batteries, and
voltage conversion devices. This chapter describes all of these components in detail ex-
cept batteries, which are described in Chapter 4, Telecommunications Batteries.
3.1 NOMINAL SYSTEM AND OPERATING VOLTAGES
Modern telecommunications dc power systems in the United States are 60 Vdc class and
operate at nominal 24 or 48 V, although some equipment still exists in networks that op-
erates at a nominal voltage of 12 or 130 Vdc. Nominal system voltages and operating
voltage ranges are shown in Table 3.1.
A number of factors affect the dc power system operating voltage ranges including
battery technology (VLA or VRLA), voltage drops in the circuit segments, rectifier out-
put regulation, and equipment voltage limits. Specific details and rationale behind the
ranges shown are provided in Chapter 5 (System Design).
Because the thinner air at higher altitudes reduces the convection and forced-air cool-
ing capability of equipment, the power output of most dc power system components must
be derated when operated at altitudes exceeding 3300 ft (1000 m) above sea level. There
also is additional derating at 6600 ft (2000 m) for circuit interrupting devices (such as cir-
cuit breakers). For interrupting devices, the thinner air is a poorer dielectric, thus reducing
their interrupting capability. Derating applies to mountaintop sites as well as high-altitude
regions in the United States. Power producing equipment such as rectifiers, converters,
and inverters generally have a lower maximum current output capability or a lower maxi-
mum ambient operating temperature at altitude than at sea level. A typical temperature
derating requires a linear decrease in ambient operating temperature from +50°C (122°F)
at sea level to +40°C at 10,000 ft (3000 m).

86 dc POWER SYSTEM COMPONENTS
Table 3.1 Nominal System Voltages and Voltage Rangesa

Nominal System Operating Voltage
Voltage (Vdc) Range (Vdc) Remarks
12 10.7–15.0
24 20.0–28.3
48 42.75–56.7 Applies to load equipment that provides network-to-
user interfaces for dc signalingb
48 40.0–56.7 Applies to load equipment that does not provide
network-to-user interfaces for dc signalingc
a
Source: ANSI T1.315 [1]. See Chapter 5 for explanation of how ranges are derived.
b
Network-to-user interfaces that use dc signaling include loop-start and ground-start analog line circuits (the for-
mer often referred to as POTS, or plain old telephone service), loop-reverse battery (LRB) analog trunk circuits,
and E&M analog trunk circuits. The design of some switching systems that use these circuits is such that the
lower end of the operating voltage range has to be increased to achieve the desired supervision distances on
metallic twisted-pair access network facilities [2–5].
c
Not all equipment can meet the lower and upper limits.
3.2 NOISE AND TRANSIENTS IN dc POWER SYSTEMS
3.2.1 Noise Characteristics

Noise is any undesired signal and can take many forms. Of particular importance in dc
power systems are noise voltages and currents that interfere with load equipment opera-
tion or even the operation of the dc power system itself. The noise can be coupled into
equipment via power and signal circuit conductors (conductive coupling) and via electro-
magnetic (inductive and capacitive) coupling caused by improper or inadequate shielding,
bonding, and grounding.
It is impractical to build rectifiers, converters, and rotating electrical generating ma-
chinery and transformers that are entirely free of harmonic voltages and currents in their
input and output waveforms. The dc output from rectifiers contains many ac components
called ripple, and their input ac current waveforms can be distorted. The term ripple indi-
cates what the dc voltage looks like on an oscilloscope (Fig. 3.1). The distorted voltages
and currents (noise) are electromagnetically coupled into telecommunications systems
through outside plant cables and signal and power cables in central offices.
The coupled noise can be measured with or without a filter. When measured with a fil-
ter, it commonly is called weighted noise, and, when measured without a filter, it is called
unweighted noise. It should be noted that all noise measuring test sets have limited fre-
quency response even when in the unweighted, or “wideband,” mode, but the response
usually is flat throughout the desired band rather than specially shaped as in the weighted
mode. In either case, the measured noise power is determined from the root mean square
(rms) voltage of the ac components within the frequency band or bands of interest.
Weighted measurements indicate the interfering or annoying effects of specific fre-
quencies on human telephone conversations. Certain frequencies are more or less filtered
by the combination of the telecommunications channel, telephone set, and human ear and
have different annoying effects. Frequencies that are heavily filtered by the system or are
not annoying have less weighting while unfiltered and annoying frequencies have more
weighting. Frequencies outside of the voice band (below approximately 200 and above
approximately 3400 Hz) are already heavily filtered by analog circuit transmission inter-
3.2 NOISE AND TRANSIENTS IN dc POWER SYSTEMS 87
Fig. 3.1 Ripple voltages in a dc power system. The oscilloscope screen capture was taken with a
ground reference at a –48-Vdc secondary distribution bus about 100 ft from the primary distribution.
The waveform shows noise and small spikes superimposed on 120-Hz ripple. The peak–peak ampli-
tude of the ripple is around 100 mV.
faces and have less interfering effect so are completely excluded from voice-band mea-
surements.1
The C-Message filter is the standard filter for voice-band measurements (Fig. 3.2). It
is based on the 500-type telephone set manufactured by Western Electric in the 1950s.
Although this type of telephone set has not been manufactured for many decades, the C-
Message filter still is widely used for noise measurements in the telecommunications in-
dustry. Noise measured with this filter is given in units of dBrnC (decibels with respect
to reference noise using the C-Message filter). Reference noise is defined as 1 picowatt,
or 0 dBrnC, which is equivalent to a power level of –90 dBm. Reference noise voltage
in a 600- circuit impedance is 0.0245 mV (0.0000245 V). Noise measured with a C-
Message filter is lower than noise measured over the voice band with no filter (un-
weighted).
The major noise sources in dc power systems are rectifiers and load equipment, but
noise may come from external sources as well. Many types of external equipment, such as
electric motors, and other external noise sources, such as lightning and ac power distribu-
tion and transmission systems, influence overall noise performance especially if the
1
Frequencies in the band from approximately 200 to 3400 Hz are called voice frequencies and the band is called
the voice band. Analog telecommunications channels that operate in this band are called voice-grade circuits.
The noise requirements for voice-grade circuits are specified in [6]. Voice-grade circuits can carry not only hu-
man speech but also analog voice-band data from fax machines and dialup modems. For a more complete dis-
cussion of analog transmission, see [2,7].
10
20
Loss (dB)
30
40
50
100 1,000 10,000
Frequency (Hz)
Fig. 3.2 C-message filter curve, voltage–frequency, or current–frequency response.
telecommunications facility is inadequately bonded and grounded. Rectifiers have fairly

large capacitors and inductors for output noise filtering, but this has little effect on noise
induced into power and signal cabling by other sources. To reduce noise from these other
sources, it is common to use individual powerline filters located as close to the loads as
possible (e.g., in the fuse or circuit breaker panel located in an equipment frame).
The typical noise output specification for rectifiers is based on industry experience and
normally is 32 dBrnC (1.6 nW, or 0.975 mV across 600 ). Reduction below this value at
the rectifier output through additional internal filtering has little effect on the noise at the
loads, and additional filtering adversely affects the rectifier’s response to step loads and
other transients caused by load equipment.
So-called wideband (unfiltered) noise is specified in volts or millivolts with and with-
out a battery (as a filter) in the measuring circuit. Fully charged batteries look electrically
like very large capacitors so they have a significant filtering effect. Wideband noise mea-
surements usually specify a battery ampere-hour (Ah) capacity that is four times (4×) the
load current. For example, the wideband noise from a 50-A rectifier would be measured
with a 200-Ah battery connected to its output. Typical values for modern modular recti-
fiers are 100 to 500 mV peak to peak in any 3-kHz interval over a 20-MHz bandwidth.
Table 3.2 shows various noise parameters that apply at the system level. Individual com-
ponents may operate at different noise levels and dc powerline filters located close to load
equipment may be required to meet operating requirements.
3.2.2 Transients
Transients are temporary deviations of the steady-state voltage to higher and lower values
(Fig. 3.3). Transients that decrease the voltage can cause equipment dropouts; transients
3.2 NOISE AND TRANSIENTS IN dc POWER SYSTEMS 89
Table 3.2 Maximum System Noise Levels

Nominal System Voice-Band Noise Wideband Noise Peak–Peak Noise
Voltage (Vdc) (dBrnC at 600 ) (mVrms) (mVpp)
12 — — —
24 56 100 240
48 56 100 480
Source: ANSI T1.315 [1].
that increase the voltage can damage power supplies and electronic components. The
opening of overcurrent protection devices, component failures, and online or offline
switching of older rectifiers can cause transients. Transients in a circuit can cause nui-
sance tripping of overcurrent devices. All modern rectifiers have a current walk-in feature
that eliminates transients during online switching.
Transients in telecommunications dc power systems are not well studied or document-
ed, but values for testing 48-V equipment have been specified (Table 3.3).
3.2.3 Direct Current Powerline Filters

Powerline filters are used to reduce ripple voltages and currents on the power conduc-
tors. Although they frequently are called battery filters, powerline filters do not filter
battery noise (batteries do not generate noise). Powerline filters are located close to
loads or groups of loads and are used to reduce noise that is electromagnetically coupled
into the dc power circuits. Such filters usually are an L- or -filter topology (Fig.
3.4) and are connected in series with primary or secondary distribution circuits (Fig.
3.5).
Powerline filters usually have input overcurrent protection (component CB in Fig.
3.4). Inductor L1 carries the full load current and is made from heavy magnet wire
wrapped on an iron core. Typical values are 1 mH. The L filter has only one high-value
electrolytic capacitor (C1), whereas the filter has two capacitors (C1 and C3). Typical
values for C1 and C3 are 6300 F. Since electrolytic capacitors are ineffective at higher
frequencies, a small value capacitor (C2, typically 0.01 F) suitable for high frequencies
is used in the output circuits. All components have a voltage rating of 150 to 200%, the
Steady-State
dc Voltage
Fig. 3.3 Transient on a dc power system.

highest operating voltage. For example, a nominal 48-V filter has components rated 75 to
100 V. Resistor R1 drains the charge from the capacitors when power is removed from the
filter. It usually is sized to discharge the capacitors in 1 min or less and typically is 30 k
to 1 M.
Since electrolytic filter capacitors have limited life (5 to 10 years) and almost always
fail shorted, the capacitors may be connected in series with a fuse (Fig. 3.6). When the ca-
pacitor fails, the fuse opens and isolates the short from the rest of the power system, thus
preventing a capacitor explosion or other catastrophic circuit failure. When the capacitor
fails and the fuse opens, the filter loses its effectiveness. The capacitor fuse usually is
Table 3.3 Transients in dc Power Systems

Maximum Maximum
Nominal System Amplitude Duration
Voltage (Vdc) (Vpk) (ms) Remarks
12 — — Not specified
24 — — Not specified
48 75 10 Maximum rate of rise and fall 10 V/ms
Source: ANSI T1.315 [2].
Fig. 3.4 dc powerline filter topologies (upper, L-filter; lower, filter) .
dc
Powerline Filter
Fig. 3.5 Powerline filter connections.

3.3 RECTIFIER SYSTEMS 91
Fig. 3.6 Capacitor fuse. If the capacitor fails shorted, the protection fuse opens. The protection
fuse is large enough to withstand capacitor charging current when voltage is first applied. The alarm
indicating fuse is a small value fuse that opens immediately after the protection fuse opens.
wired in parallel with an alarm indicating fuse so that operating personnel are alerted to
the problem.
3.3 RECTIFIER SYSTEMS
Rectifiers convert the source ac voltage to a direct voltage and supply direct current to the
load equipment. During normal operation, the rectifiers determine the voltage accuracy
and regulation and can heavily influence the noise on dc power circuits. Rectifiers always
are provided in an N + 1 redundant configuration giving a minimum of two rectifiers in
any given system (Fig. 3.7).
Fig. 3.7 Rectifiers are equipped for N + 1 redundancy: N rectifiers are required to serve load and
battery recharge requirements and one rectifier is added as a spare.
3.3.1 Technology
Three basic technologies have been used in telecommunications rectifiers since the
1970s:
앫 Silicon-controlled Rectifier (SCR)

앫 Controlled ferroresonant transformer
앫 Switch-mode rectifiers
A typical rectifier consists of a power section and a control section (Fig. 3.8). The power
section can use various types of semiconductor switches. The SCR is a unidirectional
semiconductor device that, when used in rectifiers, acts as a phase-controlled electronic
switch. Another name for an SCR is thyristor. When the SCR is turned on, it conducts
current in one direction. SCRs usually are arranged in a full-wave or bridge circuit and are
turned on and off as required to regulate the dc output voltage. The power section in-
Fig. 3.8 SCR rectifier simplified block diagram.

cludes an input transformer, SCRs, and output filters. The control section includes the
SCR trigger circuits and current and voltage regulator functions.
A conceptual schematic of a simple SCR rectifier is shown in Figure 3.9(a) to illustrate
its operation. Two SCRs (SCR1 and SCR2) and two power diodes (D1 and D2) form a
bridge circuit. The trigger control circuit delivers two trigger pulses per period of the ac
input voltage waveform, one for each half-wave. The pulses are mutually phase shifted by
ac Power Source
ac Input
Voltage
ac Power Source ac Power Source
Fig. 3.9 SCR rectifier conceptual schematic: (a) conceptual schematic, (b) circuit with SCR1 turned
on and SCR2 off, (c) circuit with SCR2 turned on and SCR1 off, and (d) waveforms at various circuit
locations.
180° and have a duration of 180° – a, where a is the amount of phase shift. Both half-
waves are rectified and phase controlled by the trigger pulses. For example, if a = 67.5°,
then the half-waves are on for 180° – 67.5° = 112.5° of their total period and off for the
rest of the time.
When a positive trigger pulse is applied to the gate of SCR1, it turns on and the result-
ing circuit is shown in Figure 3.9(b). Current flows through both SCR1 and D2. SCR1
turns off after the zero crossing of the input voltage waveform because, with a negative-
going ac input voltage at its cathode, the SCR’s anode has a negative voltage compared to
its cathode. At this point, SCR2 can be turned on. When SCR2 receives a trigger pulse
and conducts, the resulting circuit is shown in Figure 3.9(c). Current flows through both
SCR2 and D1. The voltages at various points in the circuit are shown in Figure 3.9(d).
The output of the SCR/diode bridge consists of a series of truncated half-waveforms, the
duration and rms voltage of which depend on the trigger time and duration. The output
voltage is then filtered to remove the ac components and leave only the dc component. By
delaying or advancing the trigger time with respect to the phase of the input waveform,
the output voltage can be varied from 0 to 100% or regulated within tight boundaries.
Controlled ferroresonant transformer rectifiers use a specially designed transformer
(sometimes called a controlled-output or constant-voltage transformer) that operates in a
saturated mode and uses a resonant circuit as the regulation stage (Fig. 3.10). Energy cir-
culates between the capacitor and the inductance in the transformer secondary winding to
maintain the transformer core in saturation. With the core saturated, the secondary (out-
put) winding develops an almost constant voltage even though there may be relatively
large changes in the voltage across the primary (input) winding.
SCRs or triacs may be used to control the resonant winding. Triacs are front-to-back
SCRs and thus are bidirectional (e.g., CR1 in Fig. 3.11). When the triac is switched on, it
allows current to flow in either direction. The triac is self-commutating in that it turns it-
self off at the waveform zero crossing (end of each half-cycle). When the triac is off, the
resonant winding circuit is capacitive (C1). The capacitance resonates with the winding
inductance and tends to raise the secondary output voltage. If the triac is continuously
switched on at the beginning of each ac cycle, the inductor (L1) is effectively connected
across the capacitor and the net reactance is designed to be inductive, which effectively
lowers the secondary output voltage. By varying the point in the waveform at which the
triac is turned on, the resonant winding circuit can be made to vary from capacitive to in-
ductive, thus varying the secondary output voltage up or down. The phase control circuit
determines when the triac is turned on and maintains the desired output voltage and cur-
rent. This control overcomes some of the sensitivity to input frequency that older ferrores-
onant transformer rectifiers have.
Rectification of the ac waveform in SCR and controlled ferroresonant transformer rec-
tifiers produces large ripple components on the dc output, and filtering the ripple out re-
quires large and heavy filter inductors and capacitors. Also, the switching action generates
spikes and electromagnetic interference (particularly radio frequency interference, or RFI)
that requires additional filtering. SCR rectifiers are particularly noisy. Since phase-con-
trolled rectifiers discontinuously draw power from the ac supply, rather than continuously,
the input current waveform is highly distorted. This results in an inherently low distortion
power factor and the need for additional electronics in the rectifier to improve it.
Switch-mode rectifiers are the most common technology used in new systems. When
switch-mode rectifiers were introduced, they were relatively expensive but offered size
and weight advantages not possible with other technologies. Several generations of
Fig. 3.10 Controlled ferroresonant transformer rectifier conceptual schematic.
switch-mode rectifiers have been deployed in telecommunication networks, each genera-

tion smaller and lighter than the previous generation.
While SCR and controlled ferroresonant rectifiers operate at the fundamental power-
line frequency (60 Hz), switch-mode rectifiers convert the powerline frequency to much
higher frequencies—tens of kilohertz. At higher frequencies, transformers and reactive
components (inductors and capacitors) are much smaller than at lower frequencies, thus
allowing the size and weight of the rectifier to be significantly reduced. As each genera-
tion was deployed, the frequencies got higher and the rectifiers got smaller.
Another advancement that led to smaller size and weight was the development of cir-
cuit integration and later very large scale integration (VLSI). Such circuits allowed the re-
placement of many discrete components (capacitors, resistors, and transistors) with a sin-
gle chip that usually does the same job better and more reliably.
The switch-mode rectifier consists of three stages (Fig. 3.12). The first stage rectifies
the input voltage (converts it to dc). The dc powers the high-frequency dc–ac inverter in
ac Input
dc Output
Fig. 3.11 Triac-controlled ferroresonant transformer rectifier simplified schematic.
the second stage. The output of the high-frequency inverter feeds the third stage, which
rectifies the high-frequency ac and provides the regulated and filtered dc output. Since the
inverter operates at a high frequency, relatively small filter components can be used to
eliminate high-frequency ripple on the output. The inverter block may use field-effect
transistors (FETs) in a push–pull arrangement (Fig. 3.13).
One of the design challenges associated with switch-mode rectifiers is the reduction of
high-frequency emissions to acceptable levels. All modern rectifiers must meet FCC
emission requirements.2 In addition, most rectifiers marketed for telecommunications ap-
plications are designed to meet Telcordia Network Equipment Building Systems (NEBS)
requirements as well as the applicable NEMA standard, which includes performance re-
quirements, and ANSI T1 standards, which include, among other things, vibration, tem-
perature, and fire spread and ignition.3
There are three most noticeable differences between current generation rectifier prod-
ucts and previous generations:
Modularity All current generation rectifiers are modular and are installed as plug-in
units in equipment shelves. Modular telecommunications rectifiers must be hot
swappable; that is, the rectifiers can be removed from the shelf and replaced with-
out affecting the dc power system. These rectifiers are easily replaceable, whereas
previous generations were fixed mounted and fixed wired in equipment frames. Re-
placing one of these older units requires that all wiring be disconnected.
Compact Size Previous generation rectifiers were large and heavy, providing a cur-
rent output capacity of a few to several hundred amperes at 48 V in a 23-in. × 7-ft
2
Federal Communications Commission, Part 15, Subpart B, Class A (commercial) [8].
3
For example, see [9,10,11,12,13,14].
ac Power
Source
Fig. 3.12 Basic switch-mode rectifier block diagram.
equipment frame. Current generation rectifiers are very compact and provide as
much as 1200 to 1400 A at 48 V in the same frame size.
Cooling All current generation modular rectifiers, because of their compact size and
limited capability to dissipate heat through convection, use fan cooling. Most previ-
ous rectifier generations use convection cooling. The addition of rotating machinery
(fans) to rectifiers has reduced their overall reliability and increased their mainte-
nance. In some rectifiers, fan failure shuts down the rectifier, but in others it acti-
vates an alarm and reduces the rectifier output current to prevent overheating. The
fairly loud buzzing noise made by older rectifier technologies (usually caused by
transformer vibration) has been replaced by a fairly loud fan noise from switch-
mode rectifiers.
Table 3.4 compares the characteristics of various rectifier technologies.

ac Input
dc Output
Fig. 3.13 Switch-mode rectifier simplified schematic. FET1 and FET2 are very efficient electronic
switches that are alternately turned on and off to chop the dc input and convert it to a high-frequen-
cy ac. The chopped dc is coupled to the secondary of transformer T1 where it is subsequently recti-
fied by D1 and D2 and then filtered by the output filter circuit.
Table 3.4 Rectifier Technology Comparison

Controlled
Parameter SCR Ferroresonant Switch Mode
Output regulation ± 0.5–1% ± 0.5% ± 0.5%
Current limit 125% 125% 100–105%
Input surge immunity Poor Excellent Excellent
Output ripple, rms 30 mv 30 mV 10 mV
Output ripple, peak–peak 500 mv 100 mV 100 mV
Output noise 32 dBrnC 32 dBrnC 32 dBrnC
Efficiency, 40% load 78% 81% 80–85%
Efficiency, 100% load 80% 87% 88%
Power factor, 40% load 0.70 0.94 0.97
Power factor, 100% load 0.70 0.99 0.97
EMI emission Poor Excellent Good
EMI susceptibility Poor Good Good
AC system noise Bad Excellent Excellent
Performance, 57–63 Hz Okay Excellent Excellent
Audible noise Buzz Buzz Fan noise
3.3.2 Rectifier Features

Telecommunications rectifiers have features that are characteristic of their application in
high-reliability telecommunications systems. Some features in older rectifiers were op-
tional and were only provided if specified at the time of purchase. Newer rectifier prod-
ucts generally are fully featured at no extra cost.
Operation Mode Rectifiers may operate in two modes—float and equalize. With
VLA batteries, the float mode operates load equipment and float charges the batteries dur-
ing normal operation, and the equalize mode operates load equipment and recharges the
batteries after an ac power failure or equalizes the cells in a battery when maintenance
data indicates that such a charge should be applied. The equalize mode normally is not
used with VRLA batteries because the float mode normally recharges the battery. The ac-
tual float and equalize voltages depend on the battery technology. Most rectifiers have a
switch on their front panel to control the mode and an associated indicator lamp; other
rectifiers just have a control input from an external controller or switch.
Load Sharing and Parallel Operation The minimum number of rectifiers in a dc

power system is two (one active and one backup). Multiple rectifiers must coexist, that is,
operate in parallel. Unfortunately, field experience has shown that such operation is not
guaranteed with all rectifier brands, although it should be. For example, rectifiers from
one manufacturer may not properly operate in parallel with rectifiers from another manu-
facturer.
Load sharing is different than parallel operation, although in order for rectifiers to load
share they must also operate in parallel. With load sharing, each rectifier supplies current
proportional to its rating. For example, three 50-A rectifiers serving a 100-A load each
would supply 33 A, and two 50-A rectifiers and one 100-A rectifier serving the same load
would supply 25 A from each 50-A rectifier and 50-A from the 100-A rectifier. Load
sharing is not particularly accurate, being around 10%. Load sharing requires that the rec-
tifier load sharing control circuits be connected together, usually in a controller or termi-
nal block in the powerboard.
Remote Control In some installations, it is desirable to turn rectifiers on and off from a
remote location. The remote control can be from a fire or smoke detection system, an alarm
control system, or a network operations center (NOC). The control may be by closing a dry
(i.e., no voltage) relay contact or by applying a ground to the rectifier control input.
Voltage Sensing Rectifiers regulate their output voltage by automatically adjusting it

up or down to maintain the desired value. The sensing can be done locally at or near the rec-
tifiers or rectifier frame or remotely at the battery terminals (Fig. 3.14). If remote voltage
sensing is used, the sensing is done at or close to the battery terminals. Remote sensing pro-
vides the most accurate voltage control at the point where it is required—the batteries.
However, under float conditions, the battery current is very small, and the voltage drop in
a properly designed circuit from the rectifiers to the battery terminals is very small.
Therefore, local sensing is sufficiently accurate in most installations, and the added ex-
pense of running sense leads to the batteries is not justified. However, where many battery
strings are operated in parallel or where the float current causes a significant voltage drop
(more than approximately 100 to 200 mV), then remote sensing is recommended. With par-
allel strings, a logical sensing point is a collector bar near the batteries.
Fig. 3.14 Remote (upper) and local (lower) voltage sensing. The rectifiers regulate the voltage at the
sensing point.
Temperature-Compensated Charging Related to voltage sensing and of partic-

ular importance to VRLA batteries is temperature-compensated charging. With temper-
ature-compensated charging, a temperature sensor on the battery controls the rectifier
output voltage through the remote voltage sense leads (Fig. 3.15). The compensation
reduces the chances of thermal runaway in VRLA batteries and is not required with
VLA batteries. Many existing rectifier systems can be retrofitted with an external tem-
perature compensation module (Fig. 3.16); new systems have temperature compensation
built-in.
With no temperature compensation in Figure 3.15(a), the rectifier output voltage is ad-
justed to the battery float voltage (54.5 V in this example) and is regulated by the sense
voltage input; there is no temperature sensor and no compensation. With temperature
compensation in Figure 3.15(b), the rectifiers initially are adjusted to a lower voltage be-
cause of the voltage drop across the temperature compensation module. In this example,
the voltage drop of the temperature compensation module at the normal operating temper-
ature of 25°C (77°F) is 6.0 V, so the rectifiers initially are adjusted to 48.5 V without the
module in the circuit. When the module is inserted in the circuit, the rectifier sense leads
are 6.0 V lower than the output voltage, and the rectifier output automatically increases
by 6.0 V to 54.5 V to compensate for it. If the battery temperature increases, the voltage
drop across the temperature compensation module in Figure 3.15(c) decreases by, say,
0.6 V to 5.4 V at 30°C (86°F). The rectifier sense voltage input sees this as a voltage in-
crease from 48.5 to 51.1 V. However, because the rectifier is adjusted for 48.5 V at the
dc Output –
dc Output +
dc Output –
dc Output +
dc Output –
dc Output +
Fig. 3.15 Conceptual diagrams of 48-V temperature-compensated charging setup: (a) no tempera-
ture compensation, (b) with temperature compensation at normal operating temperature, and (c) with
temperature compensation at elevated operating temperature.
sense voltage input, it automatically decreases the output voltage by 0.6 V to 53.9 V in or-
der to maintain the sense voltage input at 48.5 V.
The relationship between battery temperature and battery voltage usually is a linearly
decreasing voltage with increasing temperature (Fig. 3.17). In the other direction, with de-
creasing temperature, the rectifier output voltage is increased with falling temperature but
not indefinitely. Typical temperature coefficients are –2.5 to –4.5 mV/cell/°C. This range
works out to be 30 to 54 mV/°C for a 12-cell battery (24-V system) and 60 to 108 mV/°C
for a 24-cell battery (48-V system).
Fig. 3.16 Temperature compensation module suitable for retrofit applications.
Rectifier Current Limitation Modern switch-mode rectifiers are constant-power de-

vices, and their output current capability is inversely related to their output voltage. For
example, a 3000-W rectifier can deliver 65 A at 46 V and 55 A at 54.5 V. This means the
rectifiers can provide more recharge current to a discharged battery, when system voltage
is low, and potentially recharge it faster. However, the higher recharge current can in-
crease the risk of thermal runaway in VRLA batteries.
Rectifier current limitation can be used to minimize the chance of thermal runaway
and also to increase system efficiency. This usually is a programmable function that con-
trols rectifier current to some predetermined value. It also can be used to turn off unneed-
2.4
2.35
Charging Voltage (V/cell)
2.3
2.25 -4.5 mV/cell/ C
2.2
2.15
2.1
-20 -10 0 10 20 30 40 50
Temperature (°C)
Fig. 3.17 Battery cell voltage as a function of battery temperature for temperature-compensated
charging (typical).
ed rectifiers (or place them in a hibernate state) and to limit the number of rectifiers (and
thus the charging current) that are available to recharge the battery. This would be used
when a rectifier system is initially oversized for the initial load or battery capacity. The
unneeded rectifiers can be left plugged in. If there are any rectifier module failures, the
controller automatically brings a hibernating rectifier back online. The added benefit of
this scheme is the overall rectifier system operates much more efficiently, potentially re-
ducing the electricity consumption at the facility.
Voltage Regulation Voltage regulation is defined as

VNL – VFL
Regulation (%) = × 100 (3.1)
VFL
where VNL = no-load voltage (V)

VFL = full-load voltage (V)
In telecommunications rectifiers regulation is ±0.5% under all load and input conditions.
The relatively tight regulation is required for long battery life. The regulation at various
operating voltages is shown in Table 3.5.
Efficiency Efficiency is defined as the ratio of the input power to the output power, ex-
pressed as a decimal fraction or in percent and is around 0.85 to 0.9 (85 to 90%) for mod-
ern rectifiers at full load. As a percentage
POutput
Efficiency (%) = × 100 (3.2)
PInput
where POutput = Output Power (W)

PInput = Input Power (W)
Switch-mode rectifier efficiency is fairly constant at around 80 to 85% at loads above

25% of rated current. The efficiency of controlled ferroresonant rectifiers peaks at around
90% at full-load current. Below full load the efficiency drops to around 80% at 25% load
(Fig. 3.18). For example, a fully loaded rectifier that delivers 3000 W to network equip-
ment at 90% efficiency actually draws 3333 W from the ac power system. The rectifier
dissipates the difference (333 W) as heat. Rectifiers are seldom intentionally operated at
100% rated current except during battery recharge. In a properly designed dc power sys-
tem, rectifiers spend most of their time at 25 to 50% of rated current.
Table 3.5 Rectifier Voltages at ±0.5% Regulation

Nominal System Cell Voltage Operating Regulation at Regulation
Voltage (V) (V/cell) Voltage (V) 0.5% (V) Range (V)
24 2.17a 26.04 ±0.130 25.91–26.17
24 2.27b 27.24 ±0.136 27.10–27.38
48 2.17a 52.08 ±0.260 51.82–52.34
48 2.27b 54.48 ±0.272 54.21–54.75
a
Typical VLA cells.
b
Typical for VRLA cells.
100
Power Factor
Controlled Ferro-Resonant
90
Switch-Mode
% Efficiency or % Power Factor
Efficiency
80
Controlled Ferro-Resonant
70
60
50
40
0 10 20 30 40 50 60 70 80 90 100
% of Full Load
Fig. 3.18 Typical rectifier efficiency and power factor versus load.
Rectifier operating costs are directly related to their efficiencies and are especially im-
portant in areas where electricity costs are high.
Example 3.1 A 24-V system is equipped with six 50-A rectifiers and the total load is
steady at 75 A. Determine (1) the yearly electrical energy costs for ferroresonant and
switch-mode rectifiers when the rate is 2.5 cents/kWh, 5.0 cents/kWh, 10.0 cents/kWh,
20.0 cents/kWh, 40.0 cents/kWh, and 80.0 cents/kWh, and (2) the yearly costs of losses at
the same energy rates. Assume the system operates at 27.0 V.
In this example, the load on each rectifier is 12.5 A (50 A × 75 A/300 A) so they are
operating at 25% of rated load. The output power of each rectifier is 338 W (12.5 A × 27
V). Assuming the efficiency is 80% for ferroresonant rectifiers and 85% for switch-mode
rectifiers, the input powers are 422 W (338 W ÷ 0.8) and 397 W (338 W ÷ 0.85), respec-
tively. The energy consumed by each rectifier in one year is 3697 kWh (0.422 kW × 8760
h) and 3478 kWh (0.397 kW × 8,760 h). Operating costs are tabulated below.
Unit Energy Cost Yearly Operating Cost per Yearly Operating Cost per Difference
($/kWh) Ferroresonant Rectifier ($) Switch-Mode Rectifier ($) ($)
0.025 92 87 5
0.05 185 174 11
0.10 370 348 22
0.20 739 696 43
0.40 1479 1391 88
0.80 2958 2782 176
The power losses for the ferroresonant and switch-mode rectifiers are 84 W (422 W – 338
W) and 59 W (397 W – 338 W), and the yearly energy losses are 736 and 517 kWh, re-
spectively. The costs of the energy losses are tabulated below.
Unit Energy Cost Yearly Cost of Losses per Yearly Cost of Losses per
($/kWh) Ferroresonant Rectifier ($) Switch-Mode Rectifier ($)
0.025 18.40 12.92
0.05 36.80 25.85
0.10 73.60 51.70
0.20 147.20 103.40
0.40 294.40 206.80
0.80 588.80 413.60
The foregoing analysis can be taken one additional step by calculating the present
worth of the energy savings. Assuming 5% annual cost of money over a 10-year study
period (7.857 present worth factor assuming monthly payouts), the present worth of
the energy savings of the switch-mode rectifier is shown below. Where energy costs
are high and the study period is fairly long (10 to 20 years), the savings per recti-
fier approximately equal the purchase price of one or two switch-mode rectifiers. A sim-
ilar analysis can be made at other costs of money, time periods, and rectifier efficien-
cies.
Per Rectifier Per Rectifier System

____________________________________ _________________
Unit Energy Cost Switch-Mode Advantage Present Worth Present Worth
($/kWh) ($) (5%, 10 Years) (5%, 10 Years)
0.025 5.48 43 258
0.05 10.95 86 516
0.10 21.90 172 1032
0.20 43.80 344 2065
0.40 87.60 688 4130
0.80 175.20 1377 8259
Power Factor Correction The rectifier load in telecommunications facilities can

represent the majority of the total ac load, so it is important to use high power factor
rectifiers. Current or voltage distortion caused by low-load power factor leads to nu-
merous problems in electric utility service transformers, power and energy metering
equipment, and building electrical circuits. Most electric utilities have a penalty clause
in their tariffs for low power factor, so it is in the end-user’s interest to maximize pow-
er factor to minimize this penalty. Also, overheating from low distortion power factor
can damage building electrical wiring and step-up and step-down transformers. Power
factor correction in rectifiers reduces the distortion on the ac input voltage and current
waveforms and thus improves the performance of ac electrical systems. The power fac-
tor of ferroresonant and switch-mode rectifiers is approximately 0.9 and higher for most
load levels (Fig. 3.18).
Input Voltage Range Modern switch-mode rectifiers have a wide input voltage
range, typically 176 to 264 Vac or 96 to 264 Vac at 47 to 63 Hz.4 A wide input voltage
range allows operation with 120/240 Vac single-phase systems and 208Y/120 Vac three-
phase systems without adjustment. Older technologies require taps on the input trans-
former to be changed for operation at different service voltages. When the taps are prop-
erly set, the rectifiers will meet the specified output voltage regulation with as much as
10% variation in the ac input voltage. Large ferroresonant rectifiers (400 A and larger) re-
quire 480Y/277 Vac service voltage or a step-up transformer where lower service volt-
ages are used.
High-Voltage Shutdown All modern rectifiers have internal selective high-voltage

shutdown circuits. These circuits sense a high-voltage condition in the rectifier, possibly
due to internal failure, and shut it down by tripping the input circuit breaker to prevent
the rectifier from damaging load equipment and other rectifiers. Many rectifiers also
have remote high-voltage shutdown capability so that a high-voltage condition may
be sensed at a different location and the rectifiers shutdown through an external control
circuit. The high-voltage shutdown controls, both local and remote, are nonrevertive
(i.e., reset is not automatic) and require manual intervention to bring the rectifier back
online.
Visual Indicators and Controls Some manufactures eliminate visual indicators

from rectifiers and put them in a controller (Fig. 3.19), while others provide a full range of
operational indicators on the rectifier front panel itself. A full range includes
앫 Light-emitting diode (LED) bar graph, liquid-crystal display (LCD), or meter to in-
dicate output current
앫 LED bar graph, LCD, or meter to indicate output voltage (this meter may also mea-
sure current as determined by a selector switch; bar graphs have low resolution and
are only used to indicate voltage or current within a range)
앫 Power on/off
앫 Equalize/float mode
앫 Test
앫 Rectifier fail alarm (RFA)
앫 Low-current alarm (LCA)
앫 High-voltage alarm (HVA)
Many types of rectifiers are stand-alone devices that have user-adjustable controls on
the rectifier front panel but others require a cover or door to be opened. Other types re-
quire a separate controller module installed in the shelf assembly with the rectifier mod-
ules (Fig. 3.20). Still other types of modular rectifiers use a separate controller chassis
that connects to the rectifier shelf assembly by a control cable. Most modern types do not
have potentiometer-type adjustments but are set through a craft user interface. User-
adjustable controls typically include
4
Rectifiers used in the United States only need to operate over the range from 57 to 63 Hz, but many rectifier
products also are sold in countries where the fundamental powerline frequency is 50 Hz, which requires a range
from 47 to 53 Hz.
앫 Float voltage adjustment Used to adjust rectifier float voltage output

앫 Equalize voltage adjustment Used to adjust rectifier equalize voltage output. In
most rectifiers, the float and equalize voltage settings are interdependent. Usually,
the float voltage is set first and then the equalize voltage may be set.
앫 Current limit adjustment Rectifiers will go into current limiting under several
conditions, including short circuit on the discharge or charge bus or associated cir-
cuits, overload caused by load equipment failure that is not cleared by an overcur-
rent protection device, and battery recharge after restoration of an ac power failure
or after a load test. To the rectifier, overload and battery recharge conditions are in-
distinguishable. Many switch-mode rectifiers do not have an adjustable current lim-
it control. Some rectifiers can operate indefinitely under overload conditions but
rectifier life generally will be reduced. Ferroresonant rectifiers have inherent cur-
rent limiting, which may be supplemented by an adjustable control.
앫 Load simulate adjustment This is very useful in testing the rectifier over its full
current range; otherwise, external power supplies or special procedures are required
to fully test and adjust rectifier load sharing and overload controls. Load simulation
is missing from older rectifiers.
앫 Load sharing adjustment This control calibrates the rectifier for proper load shar-
ing.
앫 High voltage alarm and shutdown adjustment Typical high-voltage alarm settings
are 56.0 V for a 48-V system and 28.0 V for a 24-V system, but this varies with the
installation and load equipment tolerance.
앫 Low-current alarm adjustment The low-current alarm (LCA) usually is activated
if the rectifier carries less than 10% of rated load current. There can be many rea-
sons for an LCA and not all are actual alarm conditions. For example, the load cur-
rent may be very low compared to rectifier capacity and under normal float condi-
tions the rectifiers are very lightly loaded. This situation is encountered where long
battery reserve times are needed and the resulting ampere-hour capacity is large rel-
Fig. 3.19 Typical rectifier front panel. The dimensions of this switch-mode rectifier module are 3.94
in. wide × 3.30 in. high × 9.25 in. deep and it is rated 1.6 kW at 24 V. The LED indicators on the upper
right side of the front panel are ac, dc, and alarm. (Photo courtesy of Argus Technologies, Inc.)
Fig. 3.20 Rectifier shelf assembly with controller and rectifier modules (same type as in Fig. 3.19
but 48 V in this case). The shelf assembly mounts in a standard 19-in. equipment frame. The con-
troller (bottom photo) plugs into the left-hand slot of the rectifier shelf assembly and is required for
rectifier operation. LED indicators on the upper-left are OK, minor alarm, and major alarm, and there
is a reset button just below the LEDs. At the bottom left of the controller is a 10BaseT Ethernet jack
for connecting the controller to the central office local area network. The dimensions of the shelf as-
sembly are 17.4 in. wide × 3.4 in. high × 12.0 in. deep. (Photos courtesy of Argus Technologies, Inc.)
ative to the normal equipment load. Another situation is where rectifier load sharing
or float voltage is slightly out of adjustment and the rectifiers do not properly share
at low load levels. In many installations, this is not cause for alarm because the low-
er rectifier will pick up the load upon failure of another rectifier in the system.
앫 Filter capacitor charge Rectifiers are equipped with relatively large output filter
capacitors, which draw considerable charging current when voltage is applied to
them in an uncharged condition, such as when a rectifier is first turned on. The large
current draw can trip rectifier overload circuits or open overcurrent devices in the ca-
pacitor circuit. Well-designed rectifiers have a precharging control switch that allows
the capacitor to be safely charged before the output circuit breaker is turned on.
Internal Protection and Control Functions Internal protection functions are in-
herent to most modern rectifier designs and operate without user intervention. Typical in-
ternal controls are
앫 Walk-in circuit When a rectifier is turned on, the output current increases gradual-
ly. This reduces transients on the dc bus and prevents overshoot of the rectifier out-
put that could lead to a high-voltage shutdown. A typical walk-in circuit requires 5
to 10 s to start picking up the load and another 5 to 10 s to reach the required value.
앫 Low line voltage shutdown If the ac input voltage to the rectifier falls below the
3.4 POWERBOARD 109
preset threshold, the rectifier will automatically shut down to prevent rectifier dam-
age or to prevent the rectifier from operating outside its specified limits. Normally,
this is a revertive control, and the rectifier automatically restarts when the input
voltage returns to normal.
앫 Line phase loss or unbalance shutdown This applies to three-phase rectifiers and
shuts them down to prevent damage from unbalanced input conditions.
앫 Voltage sense lead protection Remote voltage sense leads are much smaller than
battery circuit conductors and usually are run on the same cable racks with them.
The leads may be damaged during the installation or removal of conductors or acci-
dentally disconnected. To prevent runaway rectifier output circuits, internal rectifi-
er controls will either force the rectifier output to a low voltage or shut it down
completely. This type of control is nonrevertive.
3.4 POWERBOARD
A powerboard consists of one or more equipment frames that support the charge and dis-
charge busbars, rectifier system, monitoring and control system, and primary distribution
system. Small single-bus powerboards with ratings up to approximately 1600 A are con-
veniently packaged with charge/discharge busbars, monitoring and control, and primary
distribution in one frame. In sizes 400 A and smaller, the rectifier system may be pack-
aged in the same frame (Fig. 3.21). In sizes larger than 400 A, a second frame, sometimes
called a supplemental frame but more often just a rectifier frame, holds the rectifier sys-
tem.
Common bus ratings for small powerboards are 50, 100, 200, 400, 600, 800, 1200, and
1600 A (actual values vary with the manufacturers). Above 1600 A, powerboard bus rat-
ings increase in approximately 800- to 1000-A increments. In some systems, the charge
and discharge busbars are hung from the ceiling or attached to the tops of frames (Fig.
3.22), and only the primary distribution system is located in equipment frames. Rigid cop-
per busbars are used in telecommunications powerboards for convenience of construc-
tion, standardization of mounting arrangements and economy. However, small power-
boards (< 50 A) may use cables to connect components. Very small power systems, up to
approximately 15 A, may have all components and limited primary distribution in one
equipment shelf (Fig. 3.23).
Busbar mounting arrangements must be designed to withstand mechanical forces due
to shipping and installation, seismic forces during earthquakes, forces from contraction
and expansion of components due to temperature changes, and magnetic reaction forces
during electrical faults. Short circuits between the current-carrying conductors (feed and
return) or between the ungrounded current-carrying conductor (feed) and ground can
cause large magnetic reaction forces. Two current-carrying conductors in close proximity
will repel or attract each other depending on the relative current directions. For thin bus-
bars, the force is approximately [15]5
2 × 10–7 × i1i2L d d2 + s2
F =
d2
2d 冤
tan–1

s
– s 冢冣
ln
s2 冢冣冥 newtons (3.3)
5
This expression for force between two conductors assumes that the bars are thin with respect to their width and
is a simplified version of a more detailed expression by H.B. Dwight in Chapter 37 of Electrical Coils and Con-
ductors, Their Electrical Characteristics and Theory, McGraw-Hill, 1945.
Fig. 3.21 Powerboard: front and rear view of a powerboard in one frame with termination busbars,
shunts, controls, meters, and rectifiers, and corresponding schematic.
where F = force (N)

i1 = current in one conductor (A)
i2 = current in other conductor (A)
L = length (m)
s = spacing between conductors (m)
d = busbar width, long side (m)
The force is small for normal operating currents but can be quite large for fault currents.
For example, assume 4-in. (102-mm) busbars are spaced 1 in. (25 mm) apart, are 10 ft (3
m) long, and carry 100 A. For a paired circuit in which the two currents have the same
magnitude and flow in opposite directions, the force tends to push the two conductors
apart. Using Eq. (3.3) the force is about 0.12 N (0.0266 lb).
Short-circuit currents from batteries in dc power systems can be quite high and can last
for several minutes. The resulting magnetic reaction forces also can be quite high. For ex-
ample, if the paired conductors of the previous example are shorted and the fault currents
are 10,000 A in each conductor, the repulsive force is about 1150 N (258 lb) using Eq.
3.4 POWERBOARD 111
Fig. 3.22 Busbars mounted at the top of equipment frames, ready for connection to batteries. This
bus is rated 10,000 A (note 10,000 A shunt in middle left of picture). (Photo courtesy of Power-One,
Inc.)
(3.3). This force is applied almost instantly (rise time limited by the ratio of circuit resis-
tance to inductance) and forces the conductors apart.
3.4.1 Charge and Discharge Busbars

The charge and discharge buses consist of one set of busbars that operates at the system
voltage and another (return bus) that is bonded to the earth electrode or grounding system.
The charge and discharge buses at system voltage may be one and the same bus but, in
Fig. 3.23 Small 48-V power system with rectifiers and primary distribution in a single equipment
shelf mounted in a 23-in. channel-type equipment frame. Up to four rectifier modules may be in-
stalled. Shelf dimensions are 17.5 in. wide × 7 in. high × 12 in. deep.
many systems, they are separated by a low-voltage disconnect (LVD) contactor, current
shunt, or both (Fig. 3.24). The charge and discharge segments of the return bus usually
are combined.
In small, low-current systems, the busbars are relatively small and do not have enough
physical space for terminating all the cables from the rectifiers and batteries, in which
case separate cable termination busbars are used (Fig. 3.25).
Copper busbars are used in powerboards because of their rigidity, strength, and high cur-
rent-carrying capability. For example, a single 2-in. × ¼-in. busbar mounted vertically in
free air can safely carry slightly more than 700 A when the ambient temperature is 30°C and
the bar is allowed 40°C temperature rise, and a 4-in. × ½-in. busbar mounted vertically in
air can safely carry almost 1900 A under the same conditions. When busbars are paralleled,
they are spaced equal to their thickness (e.g., parallel ¼ in. busbars are spaced ¼ in. apart).
The maximum current density normally is in the range of 1 A per 1000 CM to 1 A per 1600
CM. Table 3.6 shows the busbar sizes used in typical powerboards.
Current shunts used in dc powerboards are low-resistance, high-power resistors con-
nected in parallel with a high-impedance voltmeter. They are designed to carry almost all
of the current so that only a negligible portion flows through the voltmeter (Fig. 3.26).
Although the indicator actually is a voltmeter, it is calibrated in amperes with full-scale
equal to the shunt’s current rating. The most common shunt voltage drop is 50 mV at rated
load current, but other values such as 25, 40, and 100 mV have been used in some systems.
Fig. 3.24 Charge and discharge buses with low-voltage disconnect and current shunt.
3.4 POWERBOARD 113
Fig. 3.25 Rectifier and battery cable termination busbar (return bus not shown).
Example 3.2 The main shunt in an 800-A powerboard has a voltage drop of 50 mV at
rated load. Determine the shunt resistance and power dissipated at rated load. The resis-
tance of the shunt is
V 0.050
R = = = 62.5 × 10–6 = 62.5
I 800
The power dissipated at full load is
P = VI = 0.050 × 800 = 40 W
Most shunts consist of parallel resistance components to provide more surface area for
heat dissipation (Fig. 3.27). This minimizes heating of the resistive element and improves
its accuracy over the full operating current range.
When a shunt is connected between the charge and discharge buses and measures the
total system load, it is called a main shunt. Shunts also may be used on individual pri-
mary or secondary distribution circuits or to measure the load on circuit breaker and
fuse panels (Fig. 3.28). In some powerboards, there is no main shunt, in which case the
Fig. 3.26 dc current shunt application.
Table 3.6 Typical Powerboard Busbar Sizes

Powerboard
Rating Busbar Size Area Current Density
(A) (in.) (kcmil) (CM/A)
1–
50, 100, 200 ×1
4 318.3 1592
1–
400, 600 ×2
4 636.6 1061
1–
800 × 2 1–2
4 795.8 995
1–
1200 ×4
4 1273.2 1061
1–
1600 (2) × 4
4 2,546.5 1592
Fig. 3.27 400-A current shunt consisting of four parallel resistance elements in center of photo
(note small leads to current indicator).
3.4 POWERBOARD 115
Fig. 3.28 Shunt applications: an analog ammeter would have two or more scales or the selector
switch would select different ranges in a digital ammeter to accommodate different current ratings in
the main and distribution shunts.
controller automatically adds the individual distribution loads to arrive at the total cur-
rent.
Low-voltage disconnects are high-current relays (contactors) that may be connected
between the discharge bus and the primary distribution system (Fig. 3.29). Their voltage
and current ratings are the same as the busbars on which they are mounted. In high-
current applications or where redundancy is needed, two LVDs may be wired in paral-
lel. Where this is not practical, the loads are split across more than one discharge bus
to lower the individual LVD current rating. A low-voltage disconnect may be installed
to
앫 Prevent overdischarge of the battery

앫 Provide a means to shed loads to prolong battery reserve time
앫 Prevent damage to load equipment from low voltage
The arguments against using low-voltage disconnects are
앫 They are an electromechanical element that is subject to failure.

앫 They are subject to accidental operation.
앫 The low-voltage disconnect may open due to an initial dip in the battery voltage
Fig. 3.29 Low-voltage disconnect on the main bus.
(so-called coupe de fouet) when a fully charged battery first starts to discharge. Al-
though the battery voltage probably recovers, the LVD already has interrupted the
discharge bus. This phenomenon is more fully described in Chapter 6, Part III, Bat-
tery Maintenance.
Low-voltage disconnect contactors are electrically held; that is, when the system bus
voltage is greater than the disconnect threshold, the LVD contactor is in the operated
state (closed) and is electrically held in the closed condition by the contactor solenoid.
As the battery discharges and the system bus voltage drops, the disconnect threshold is
reached where the contactor releases and disconnects the circuit. Battery discharge is
stopped, and the battery is protected from overdischarge when the load is disconnected
from it by the LVD. The LVD does not disconnect the rectifier circuits. When the rec-
tifier power source is restored, all available rectifier current initially is available to
recharge the battery via the charge bus. As the battery voltage rises, the LVD reconnect
threshold eventually is reached, which causes the LVD contactor to close, thus restoring
current to the load. At this time, the rectifier system provides current to both the battery
and load equipment.
Low-voltage disconnects normally are set with hysteresis such that they disconnect
(drop-out) at a lower voltage than they reconnect. For example, the disconnect voltage
threshold may be set to 44.0 V and the reconnect voltage to 48.0 V. As the battery
discharges, it will be disconnected when the voltage at the LVD sensing point reaches
44.0 V. As the battery recharges the LVD will reconnect the load when the voltage
3.5 MONITORING AND CONTROL 117
at the sensing point reaches 48.0 V. The actual set-point voltages used in any given sys-
tem depend on a number of factors. More details can be found in Chapter 5, System
Design.
Low-voltage disconnects are used primarily to protect battery systems from overdis-
charge, but they also may be used in load-shedding schemes and to prevent load equip-
ment from operating at low voltages that may damage them.
A load-shedding scheme allows less important or lower priority loads to be discon-
nected first, which reduces the discharge current and extends the discharge time. For ex-
ample, in a three-step load-shedding scheme (Fig. 3.30), the first LVD (LVD1) may be
set to disconnect at 46.0 V, the second (LVD2) to disconnect at 45.0 V, and the third
(LVD3) at 44.0 V. When the discharge bus voltage reaches the first threshold, LVD1 will
disconnect the low-priority loads thus reducing the battery current. The battery will con-
tinue to discharge but at a slower rate. When the second threshold is reached, LVD2 will
disconnect the second load group, and so on.
3.5 MONITORING AND CONTROL
Monitoring and control systems can include simple analog metering for voltage and cur-
rent and dry relay contact closures and indicating lamps for alarms, or they can be more
Fig. 3.30 Low-voltage disconnect in a three-circuit load shedding application.

complex digital metering, data logging, and alarm collecting systems with various types
of local and remote connectivity (e.g., EIA-232 or 10/100BaseT Ethernet using Internet
Protocol, IP). Figure 3.31 shows typical functions.
3.5.1 Alarm Circuits

Typical monitoring, alarm sending, and alarm indicating functions include
앫 High-voltage alarm
앫 Low-voltage alarm
앫 High or overcurrent alarm
앫 Rectifier system minor alarm (operates upon failure of one rectifier in an N + 1 sys-
tem)
앫 Rectifier system major alarm (operates upon failure of more than one rectifier in an
N + 1 system)
앫 Fuse/circuit breaker alarm
앫 Low-voltage disconnect (LVD) alarm
앫 Emergency or fire alarm disconnect control
앫 Automatic and manual rectifier equalize control
The high or overcurrent alarm setting is based on the powerboard bus (or main shunt) rat-
ing and the rating of the largest rectifier in the system as follows
IAlarm (%)(ICapacity – IRectifier)

IOCA (%) = (3.4)
IShunt
where IOCA (%) = setting of overcurrent alarm (%)

IAlarm (%) = percent of full load desired for overcurrent alarm (A)
ICapacity = total capacity of the power system (A)
IRectifier = current rating of largest rectifier in the system (A)
IShunt = main shunt current rating (A)
Alarm circuits may be “dry” or “wet” and may be based on solid-state electronic
switches or electromechanical relays. A dry circuit has no voltage or ground on the alarm
contacts, whereas a wet circuit has either battery or ground. Sensing circuits associated
with dry contacts supply battery and ground for indicating lamps or LEDs (light-emitting
diodes) and alarm sending systems. A typical alarm output circuit consists of one or more
form-C relay contact sets, each of which includes a normally open and normally closed
contact (Fig. 3.32). Depending on the function, separate sets of contacts are provided for
minor, major, and catastrophic alarms.
Alarm circuits can indicate minor (MIN), major (MAJ), or catastrophic (CAT) condi-
tions. A minor alarm indicates a fault that is not service affecting or affects only a small
percentage of the total service. An example of a minor alarm is the failure of one rectifier
in an N + 1 redundancy rectifier system. A major alarm is more severe and affects a larg-
er percentage of the total service. The failure of two or more rectifiers in a rectifier system
normally is a major alarm. A catastrophic alarm indicates total service failure. The failure
of all rectifiers and battery discharge to where the low-voltage disconnect opens is a cata-
Fig. 3.31 Monitoring and control.
Fig. 3.32 Form-C relay contact set. The symbols for normally closed and normally open relay con-
tacts on the left are unique to the telecommunications industry. The contact position shown in
schematics usually is the position of the contacts when the device is “on-the-shelf” and not pow-
ered.
strophic alarm condition. In modern programmable equipment, many types of alarms can
be classified in software according to their perceived severity.
Fuse and circuit breaker alarms usually are operated in parallel and, on a system basis,
are detected as a group. Small circuits (less than 15 A) use alarm indicating fuses [Fig.
3.33(a)] that have integral means of closing an alarm contact when the fuse burns open.
Alarm indicating fuses also are connected in parallel with larger cartridge fuses to provide
the indicating function [Fig. 3.33(b)].
Circuit breakers used in telecommunications power systems include an auxiliary set of
form-C contacts to indicate circuit breaker handle position—opened or closed (Fig. 3.34).
In this case, the normally closed contacts are connected in parallel to the alarm detecting
system.
3.5.2 Battery Equalize Control

The equalize control may be manual or automatic. Older rectifier systems have an equal-
ize control switch on each rectifier. The rectifiers also may be manually controlled as a
group from the powerboard with a single toggle switch. More modern systems use a cen-
tral controller in the powerboard to control all rectifier functions, including equalize.
Automatic equalize controllers monitor the system voltage (Fig. 3.35). The rectifier
equalize control circuit is armed if the voltage falls below a programmable threshold,
which indicates the rectifiers are not supplying current and the battery is discharging.
Thresholds typically are 24.0 or 48.0 V depending on the system voltage. When ac power
(a) (b)
Fig. 3.33 Fuse alarm circuits. (a) Alarm indicating fuses are connected in parallel with a larger car-
tridge fuse. When the larger fuse burns open, so does (b) the alarm indicating fuse.
Fig. 3.34 Circuit breaker alarm circuits (alarm contacts are shown with the circuit breaker in
off/tripped position). When the circuit breaker is in the on position, the normally closed (NC) contacts
are mechanically held open, and the bus voltage is blocked from the alarm circuit. When a circuit
breaker trips or is turned off, the normally closed contacts connect the bus voltage to the alarm cir-
cuit.
Fig. 3.35 Automatic equalize controller.

returns and the rectifiers are able to deliver output current, they switch to the equalize
mode for a selectable time period, typically from 0 to 99 h. When the timer expires, the
rectifiers automatically are reset to the float mode. Equalize controllers normally are not
used with VRLA batteries but are found in many powerboards used with VLA batteries.
3.5.3 Capacitor Precharge Circuit

Some load circuits have large capacitances associated with them. When voltage is applied
to a discharged capacitor, it acts like a short circuit and may trip upstream overcurrent de-
vices. Some powerboards contain circuits that precharge the capacitors prior to the distri-
bution circuit breaker closing. Precharging prevents the circuit breaker from tripping
when it is initially closed due to the current transient from the capacitors in the circuit
load equipment. Each large circuit breaker position (typically larger than 125 A) is
equipped with a precharge push-button and associated circuit (Fig. 3.36).
3.6 PRIMARY AND SECONDARY DISTRIBUTION SYSTEMS
The primary distribution system includes the first overcurrent protection device after the
discharge bus and normally is located with or close to the discharge bus. Primary overcur-
rent devices usually feed equipment groups such as switching systems or secondary distri-
bution systems. The secondary distribution system includes intermediate overcurrent pro-
tection devices after the primary and may be located adjacent to the primary or in any
other location.6
Overcurrent protection devices such as circuit breakers and fuses in the primary and
secondary distribution systems protect all branch circuits on the load side of the discharge
bus. The overcurrent protection is located at the point closest to where conductors are
powered (Fig. 3.37) and also at any point in a circuit where the conductor size is reduced
unless an upstream overcurrent device also is suitable for protecting the smaller conduc-
tor.
3.6.1 Overcurrent Protection

System protection is “the detection and prompt isolation of the affected portion of the sys-
tem whenever a short circuit or other abnormality occurs that might cause damage to, or
adversely affect the operation of any portion of the system or the load that it supplies”
[16 (page 39)]. Overcurrent caused by an overload or a short circuit results in overheating,
burning, or other damage to circuit wiring and equipment if it is allowed to persist.
The primary objective of electrical system protection is to prevent human injury.
Secondary objectives are to minimize load equipment and circuit component damage,
and to limit service interruptions caused by equipment failure, human error, or adverse
natural events. Achieving the primary objective in many cases means achieving the sec-
ondary objectives. Equipment damage and service continuity are traded off according to
the operating requirements and importance of the equipment being protected. The pri-
mary and secondary objectives can be achieved only by the correct combination of cir-
6
Primary distribution centers and secondary distribution centers are sometimes called battery distribution fuse
bays (BDFB) or battery distribution circuit breaker bays (BDCBB).
3.6 PRIMARY AND SECONDARY DISTRIBUTION SYSTEMS 123
Fig. 3.36 Capacitor precharge circuit. Prior to closing the circuit breaker the push-button is
pressed and current flows through the charging resistor, charging fuse, and indicating lamp to the
load capacitors. When the lamp extinguishes, the capacitors in the circuit are charged. At this time,
the circuit breaker can be closed and the button is released. The charge resistor is sized to limit the
current through the charge fuse.
cuit overcurrent protection, wire size and insulation, and physical protection. The nom-
inal current rating of a device refers to the maximum current it can carry during normal
operation.
Properly sized overcurrent protection devices open the circuit before any damage oc-
curs. The total of the individual primary or secondary distribution overcurrent device rat-
ings almost always is greater than the powerboard bus rating. This is because the individ-
ual distribution circuits usually operate at a 50% load factor (the actual circuit operating
current is no more than 50% of the overcurrent protective device rating) although they are
designed to continuously carry 80% of circuit current-carrying capacity. Telecommunica-
tions dc circuits normally are not designed to carry more than 80% of the conductor cur-
rent rating during normal operation. As long as the actual total load does not exceed the
bus rating there will be no problems.
Overloads are overcurrents between 1× and approximately 6× the nominal current rat-
ing while short circuits are overcurrents above approximately 10× nominal current. Cur-
rents between approximately 6× and 10× are considered transition currents. Overload cur-
rents do not leave the normal current-carrying circuit path from the source through the
load, whereas short circuits bypass the load. Short circuits usually result from line-to-line
or line-to-ground faults. Overcurrent protection devices react differently to overload and
Fig. 3.37 Overcurrent protection device for a circuit is located where the circuit receives its power.
short-circuit currents because of their current–time (IT) or delay characteristic curves. It

is possible the delay characteristic curve for a given device is different for dc than ac cur-
rents, thus it is necessary to confirm their basis before use.
Overloads are caused by connecting larger or additional load equipment to the circuit
or failure in the load equipment that causes higher than normal current. Short circuits in
low-voltage dc circuits usually are caused by mechanical damage to wire insulation dur-
ing installation, removal, or replacement, accidental contact with tools, or catastrophic
failure of load equipment.
Besides protecting the conductors and equipment, the overcurrent protective device
isolates the fault from the power source so that nonfaulted circuits can continue working
in a normal manner. The entire electrical system, including the power source, wire, elec-
trical circuit components, and load equipment must be protected from faults.
Protection devices must be coordinated. The ratings of all devices connected between
the current source and load are selectively coordinated when the device nearest to the
fault is the only one that opens and only the affected equipment is disconnected from the
electrical system. For example, a fault in the primary or secondary distribution circuits
should not affect the battery circuit (Fig. 3.38).
The selective coordination process compares the delay characteristic curves of the pro-
tection devices connected in series. Circuit breakers and fuses are available with various
delay characteristic curves. Figure 3.39 shows a typical circuit breaker delay characteris-
tic curve, and Table 3.7 provides tabular information corresponding to this curve. These
characteristics apply to any circuit breaker with the delay 51 curve shown. Figure 3.40
and Table 3.8 provide similar information for a typical fuse. Note that a separate curve is
provided for each fuse value because different fuse link masses and shapes affect the
melting characteristics. Fuse curves such as these normally represent average conditions
and do not show minimum melt or total clearing conditions.
As an example of a circuit breaker delay curve application, consider a 50-A circuit
breaker and a fault that draws 200 A (4× or 400% of circuit breaker rating). The opening
mechanism in a circuit breaker with a delay 51 trip characteristic will unlatch in as little
Fig. 3.38 Coordination of overcurrent devices. A fault at D should open the overcurrent protection
device at C only and not the devices at A or B.
Fig. 3.39 Example of delay 51 circuit breaker trip points. (underlying graph with permission of Air-
pax Corporation, Power Protection Products.)
as 0.03 s (point A in Fig. 3.39), but the circuit will not completely open until as much as
0.2 s has elapsed (point B). If the load on the delay 51 circuit breaker in this example is
62.5 A (125% of rating), it may trip anywhere between 0.3 s (point C) and 3 h (point D) or
more (it may never open).
Coordination of circuit breaker interrupt ratings (as opposed to overload ratings) can
be complicated. A short circuit can occur at the load terminals of the distribution protec-
tion device or almost anywhere in the battery or distribution circuits. Some locations are
higher risk than others (e.g., exposed terminals), and locations closer to the battery will
experience higher fault currents. Coordination should ensure that the first device to inter-
rupt the fault has an interrupt rating greater than the available short-circuit current.
In modern telecommunications dc power systems, two conductors carry current from
the power source to the load (feed and return conductors). The return conductor is con-
nected to a return bus, which may be in a powerboard or other convenient location. In
modern systems, the return bus is intentionally bonded to the earth ground system at only
one location. A third conductor (equipment grounding conductor, EGC, or frame ground)
is used to bond equipment frames and structures to the earth ground system. In modern
systems, the return conductors and EGCs are separate conductors, and the EGC is meant
Table 3.7 Delay Characteristics for Delay 51 Circuit Breaker

Delay 100% 125% 150% 200% 300% 400% 600% 800% 1000%
51 & 51F No trip 0.5– 0.25– 0.15– 0.05– 0.03– 0.015– 0.01– 0.01–
10 s 3s 0.9 s 0.3 s 0.2 s 0.15 s 0.09 s 0.04 s
Data provided by Airpax Corporation, Power Protection Products.
Fig. 3.40 Example of TPS fuse delay characteristic curve. (Underlying graph with permission of
Bussmann.)
to carry only fault currents and does not carry any operating current. Older systems may
have different bonding configurations (Table 3.9).
Short circuits can occur between the feed conductor and return conductor or between
the feed conductor and the EGC or other grounded metallic part. In low-voltage systems,
short circuits usually are low impedance (terms used are solid, bolted, or welded) rather
than high impedance from arcing or flashover. However, high-impedance faults can occur
in battery systems where a cell container fails and electrolyte leaks to a grounded metallic
component.
The type of overcurrent protection device—fuse or circuit breaker—used in any partic-
ular application is determined by cost, preference, available space, bus configurations,
and maintenance, environmental, and circuit performance requirements. Proper selection
results in a device with the lowest rating that will not inadvertently open (false trip or nui-
sance trip).
Table 3.8 Delay Characteristics for 30-A TPS Fuse

Delay 100% 125% 150% 200% 300% 400% 600% 800% 1000%
Typical No trip > 1000 s ~ 130 s ~ 80 s ~ 0.7 s ~ 0.09 s ~ 0.03 s ~ 0.012 s < 0.01 s
Data from the corresponding delay characteristic curve.
Table 3.9 Comparison of Return Conductor and EGC Configurations for Modern and Older
Systems
Component Modern Older
Return bus Intentionally bonded to the earth Intentionally and incidentally bonded
ground system at only one to ground at many points or
location locations
Equipment grounding Separate from the return Not always separate from the return
conductor (EGC) conductor and meant to only conductor—may be shared with
carry fault currents the return conductor
The overcurrent device must sense the fault and disconnect the circuit from the power
source before injury or damage occurs. Only the feed conductor is disconnected by the
overcurrent device. Fuses and circuit breakers provide both sensing and disconnecting
functions in one device.
A fuse depends entirely on thermal effects for its operation—the fuse element melts
and opens the circuit. The time it takes to open depends on the length of time the overcur-
rent has been flowing and the magnitude of the overcurrent. Fuses are single-use devices
(not resettable) that are discarded and replaced after they have opened.
Circuit breakers are electromechanical devices that may be reset and reused after an
overload or low-value short circuit, but most manufacturers recommend replacement if
the device ever interrupts a short circuit with a fault current close to its interrupt rating.
Circuit breakers depend on thermal effects or magnetic effects, or both, for their opera-
tion. Thermal operation is similar to fuses in that the current magnitude required to open
the device depends on how long it has been flowing. Circuit breakers that are designed to
protect only against short circuits use an electromagnet to sense and quickly actuate the
opening mechanism. Thermal-magnetic circuit breakers combine the advantages of sens-
ing and protecting against overloads as well as sensing and quickly opening on short cir-
cuits. Circuit breakers used in telecommunications are trip-free, which means they cannot
be held on with an overload or short circuit present on the circuit.
Circuit components, such as rectifiers, filters, regulators, and electronic circuits, have
significantly different overload withstand characteristics compared to wire and cable.
Many electronic circuits and components require extremely fast fault clearing to ade-
quately protect them from thermal damage.
3.6.2 Fuses
Fuses are the most common overcurrent protective devices used in telecommunications
systems, especially older systems—they are low cost and reliable. A fuse operates by
melting an internal, shaped-metal link, typically a lead alloy (fast-acting fuses may use a
silver wire link). Since a fuse depends on its self-destruction to protect a circuit, the fuse
must be replaced after it opens [replaceable element fuses (type RL) have been used in
telecommunications but all new applications since the early 1980s use one-time fuses].
When a fuse burns opens, the melting current and the circuit interrupting current may be
different. Fuse-clip size and condition and the size of the conductor attached to it can in-
fluence fuse performance (large conductors act as heat sinks). In addition, fuse and clip
corrosion can cause heating due to the higher resistance of the corroded surface and can
lead to nuisance interruption. Also, if a fuse holder is overheated, the fuse contract springs
may relax, leading to poor contact and increased resistance.
Fuse elements may deteriorate over time due to chemical and physical stresses in the
element during repeated short duration overloads. For example, some power supplies pro-
duce a short high-current inrush (capacitor charging current) followed by a relatively low
running current. The short duration inrush current warms and thermally stresses the ele-
ment but does not last long enough to melt it. Cyclical expansions and contractions of the
fuse element can lead to eventual mechanical fatigue and premature failure.
Fuses must not be used where the circuit operating voltage is higher than the fuse volt-
age rating. On the other hand, a fuse may be operated at a lower than rated voltage but the
arcing and clearing times may be different than at rated voltage. Fuses should not be used
in direct current circuits unless they are rated for dc.
3.6.2.1 Fuse Operating Parameters The operation of all fuse types is defined by
several parameters:
앫 Melting time
앫 Arcing time
앫 Total clearing time (sum of melting and arcing time)
앫 Peak let-through current
앫 Voltage rating
The time required to melt the fuse element on a specified overcurrent is the melting
time, whereas arcing time is the elapsed time from the melting of the fuse element to the
final interruption of the current. The peak let-through current is the maximum instanta-
neous current through the fuse during the total clearing time (Fig. 3.41).
Fuses are available in two basic configurations—small value, alarm indicating fuses,
Type 70 and type GMT, and cartridge fuses (Fig. 3.42). Several fuse panel types are
available for telecommunications applications depending on the fuse type and required
current rating (the following are modern and not historical)7:
앫 Type 70 fuse panels with individual fuse position ratings up to 20 A

앫 Type GMT fuse panels with individual fuse position ratings up to 15 A
앫 Type TPL and TGL, cartridge fuse panels with individual fuse position ratings from
70 to 800 A
앫 Type TPN, TLN, and TGN, cartridge fuse panels with individual fuse position rat-
ings from 1 to 600 A
7
TPx fuses are manufactured by Bussmann, TLx fuses by Littelfuse, and TGx fuses by Ferraz-Shawmut.
Fig. 3.41 Typical fuse current limitation curve.
앫 Type TPS, TLS, and TGS, cartridge fuse panels with individual fuse position rat-
ings from 1 to 70 A
Fuse positions rated more than 600 or 800 A are seldom used in telecommunications dc
power systems. Loads are split in such a way to reduce the rating of individual circuits to
less than 600 or 800 A, typically 100 to 200 A.
3.6.2.2 Time Delay Characteristics The thermal energy required to melt the fuse
element is specified by its melting integral, which is represented in fuse catalogs by
time–current (current–time) or delay characteristic curves (Fig. 3.43). The delay charac-
teristic curve is determined by the element construction and material. To minimize false
opening, fuses are selected such that their minimum opening energy is greater than the
maximum inrush energy. Some load equipment requires a fast-blow fuse while others re-
quire a slow-blow fuse. Slow-blow fuses have an inherent thermal delay that allows them
to withstand large inrush currents for short durations without burning open.
Because there is no overlap between fuse curves shown, these fuses appear to provide
selective coordination; however, the curves represent average values and do not show the
possible ranges due to manufacturing tolerances. Curves generally are considered to be
accurate to ±10%, but for selective coordination purposes an accuracy of ±20% is a more
conservative assumption.
As a rule of thumb, for a given fuse family, selective coordination can be achieved
when the fuse ratings have a 2:1 ratio. For example, a 100-A fuse will coordinate with a
downstream 50-A fuse. The 2:1 ratio may not apply when different fuse families are in-
volved. Also, when fuses and circuit breakers are connected in series, selective coordina-
tion generally cannot be ensured by rules of thumb but only by more detailed analysis
(Chapter 5, System Design).
3.6.2.3 Short-Circuit Capability Cartridge fuses have good short-circuit interrupt

capability. The TxN (TPN, TGN, TLN depending on manufacturer), TxS, and TPL series
Fig. 3.42 Telecommunications fuses. Type TPN, TPL, and TPS are cartridge fuses and type GMT
and 70 are alarm indicating fuses.
all have 100,000-A interrupt rating. Type NON fuses used in older dc power systems
have 50,000-A interrupt rating up to 60 A and 10,000 A at higher values. Smaller alarm-
indicating fuses have much lower ratings because they are used on relatively low-current
circuits whose resistance limits the available fault currents to low values. The type 70 fus-
es have 1000-A interrupt rating and type GMT have 450-A interrupt rating when used in
60-Vdc class systems.
3.6.2.4 Direct Current Resistance Small value alarm-indicating fuses have sig-
nificant dc resistance and can introduce significant voltage drop (Table 3.10). On the oth-
er hand, cartridge fuses have very low resistance and do not introduce significant voltage
drop.
3.6.3 Circuit Breakers

A circuit breaker provides manual on–off switching and automatic operation during over-
current conditions. Overcurrent trip is indicated by the handle moving to the off or
midtrip position.
3.6.3.1 Thermal Circuit Breakers Thermal circuit breakers depend on tempera-

ture rise in a thermal sensing element for actuation. When an overcurrent condition exists,
the thermal sensing element deflects and trips a spring-loaded switch to open the circuit.
The most common thermal elements consist of a laminate of two or three different metals.
Fig. 3.43 Time delay characteristic curves, type TPL fuse. (Underlying curve provided by Buss-
mann.)
The low expansion side may be the metal alloy Invar, the center may be copper for low
resistivity or nickel for high resistivity. Metals used in the high expansion side vary con-
siderably.
Temperature rise in the sensing element is caused principally from internal resistance
heating due to the load current through the circuit breaker (I2R heating). However, the
heating or cooling effects from external sources also affects the thermal element by alter-
ing its room temperature calibration. The physical size, shape, configuration, and electri-
cal resistance of the thermal element determine the circuit breaker current rating. In some
cases, a heater coil is placed adjacent to, and electrically in series with, the thermal ele-
ment to augment self-heating of the thermal trip element. This is especially true in circuit
breakers rated below 5 A.
3.6.3.2 Thermal-Magnetic Circuit Breaker To protect wiring, upstream compo-

nents and the breaker itself from unnecessarily long thermal and mechanical stress during
Table 3.10 Type GMT Fuse dc Resistance and Voltage Drop

Ampere Nominal Cold
Rating Resistance Voltage Drop @ 50% Voltage Drop @ 100%
(A) () Load Factor (V) Load Factor (V)
0.18 6.25 0.56 1.13
0.20 5.70 0.57 1.14
0.25 4.20 0.53 1.05
0.375 2.00 0.38 0.75
0.5 1.52 0.38 0.76
0.65 1.25 0.41 0.81
0.75 0.980 0.37 0.74
1 0.665 0.33 0.67
1.33 0.480 0.32 0.64
1.5 0.385 0.29 0.58
2 0.120 0.12 0.24
2.5 0.0904 0.11 0.23
3 0.0670 0.10 0.20
3.5 0.0415 0.07 0.15
4 0.0350 0.07 0.14
5 0.0285 0.07 0.14
7.5 0.0113 0.04 0.08
10 0.00840 0.04 0.08
12 0.00660 0.04 0.08
15 0.00580 0.04 0.09
high fault currents, an electromagnet may be added to decrease the tripping time of the
thermal breaker (Fig. 3.44). This magnetic circuit usually consists of a few turns of a large
cross-section magnet wire in series with the thermal element.
The magnetic-assist mechanism usually has a trip point well above the normal over-
load trip range. This has little effect on the normal thermal trip response time, but under
high overcurrent conditions the current produces sufficient magnetic force to trip the cir-
cuit breaker without waiting for the bimetal thermal element to heat up and deflect.
Fig. 3.44 Thermal-magnetic circuit breaker.

3.6.3.3 Magnetic Circuit Breakers A magnetic circuit breaker (also called in-
stantaneous trip circuit breaker) is similar to a thermal-magnetic circuit breaker except it
does not have the thermal element. It uses a solenoid in which a movable core, held with a
spring in a tube and damped with a fluid, is moved by the magnetic field of a series coil
that is carrying fault current. The movable core trips the spring-loaded contacts.
Magnetic circuit breakers are designed to open quickly (“instantaneously”) under
short-circuit conditions. The instantaneous trip current usually is on the order of 6 to 10
times (6×, or 600%, to 10×, or 1000%) the current rating of the circuit breaker. A typical
thermal circuit breaker will open a 10,000-A fault in about 40 or 50 ms, but a magnetic
circuit breaker will open the same fault in about 10 ms. Under higher short-circuit current
conditions, the operating speed of magnetic circuit breakers may be as fast as 3 to 4 ms.
The time-to-trip of a typical magnetic breaker is illustrated in Figure 3.45.
A magnetic breaker can be reset immediately after tripping, although the delay mecha-
nism itself does not immediately reset. If the fault is still present, the circuit breaker will
Fig. 3.45 Time–current curve for a magnetic circuit breaker. The circuit breaker shown will not re-
spond to currents less than 5 times (5×) its current rating. The minimum opening time for very large
faults (100×) is limited by the mechanical inertia of the circuit breaker components. The dotted lines
indicate the time ranges due to mechanical and electrical tolerances.
trip again but with reduced time delay. Thermal breakers cannot be reset until their heat-
ing elements cool down.
3.6.3.4 Short-Circuit Capability It is important that the fault current available at

the overcurrent device not exceed the current that the device can safely interrupt. Circuit
conductor resistances and fault resistance affect the available fault current at any location
in a circuit. In a properly designed circuit, short-circuit currents do not destroy the circuit
breaker.
The primary factors that determine the available fault current are the battery capacity
and its internal resistance, rectifier capacity, resistance of the circuit conductors, and con-
nection resistances. These factors, in addition to the fault resistance, determine the actual
fault current.
The effective current capacity of a circuit can be computed roughly by a simple differ-
ential measurement of the circuit output voltage from no load to full load. For example, if
a 50-V circuit supplying 30 A has a 2-V drop, the total resistance back to the power
source is R = 2/30 = 0.067 , and the short-circuit current at the load end of this circuit is
limited to the power source voltage divided by the total circuit resistance, or 50 V/0.067
= 750 A. The interrupting capacity for typical circuit breakers is shown in Table 3.11.
3.6.3.5 Transient (False or Nuisance) Tripping The fast operating speeds of

magnetic circuit breakers can cause nuisance tripping on high-amplitude transients. When
a transient of sufficient energy flows through a circuit breaker, it responds in an instanta-
neous trip mode and trips. If the transient is nondamaging to load and circuit equipment,
as it is in many cases, the trip is classified as a nuisance trip or false trip.
Inrush transients are most severe when the power source has low impedance and the
circuit voltage is high. Transient-tolerant protectors must accept the first surge of current
without tripping, while still providing maximum circuit protection. This is accomplished
either by shunting high magnetic flux peaks away from the circuit breaker internal arma-
ture or using an inertial device to dampen the armature. Both methods affect the delay
characteristic curves.
3.6.3.6 Time Delay Characteristics Trip delays must be long enough to avoid
nuisance tripping caused by harmless transients, yet fast enough to open the circuit when
a real fault exists. Basic trip delay characteristics are:
앫 Instantaneous Usually trips in under 100 ms but more typically within approxi-
mately 15 ms, for very sensitive circuits where low overloads of short duration may
be harmful, or where high fault currents should not be allowed to persist.
Table 3.11 Circuit Breaker Interrupting Capacity

Current Maximum Interrupting Third-Party
Range (A) Voltage (Vdc) Capacity (A) Approval
100–250 65 65,000 UL 489
100–250 65 65,000 UL 489A
275–1200a 65 65,000 UL 489A
a
Parallel poles.
앫 Fast Delay Trips in less than 10 s, for circuits and electronic applications where
temporary overloads of 200% cannot be tolerated for more than a few seconds.
앫 Slow Delay Trips in 10 to 100 s, for most large transformer-coupled loads where
brief overloads can be tolerated without damage. Slow delays allow turn-on surges
to pass without tripping.
앫 Very Slow Delay Trips in more than 100 s, for protection of wiring where a limit-
ed overload will not cause damage.
As the fault current increases, the time delay decreases to a point where the circuit
breaker’s mechanical inertia and arc quenching capability determine the circuit interrupt-
ing time. At very high fault currents (10 times the rated current), the practical limit for
magnetic circuit breaker is about 4 to 10 ms.
Circuit breakers used in telecommunications usually are available in three delay varia-
tions: 51, 52, and 53, with 52 being the most common (Fig. 3.46 and Table 3.12). Delay
51 is a relatively short delay device for general-purpose applications in distribution cir-
cuits. Delay 52 and 53 have much longer delays and are useful in circuits where inrush
currents and false tripping are potential problems. Using a long delay circuit breaker in-
curs some risk because conductor insulation may melt or the conductor itself may be
damaged before the circuit breaker opens.
3.6.3.7 Direct Circuit Resistance Circuit breakers typically have very low resis-
tances and, therefore, low-voltage drops at rated load. However, the voltage drop may be
a significant factor if the load voltage is critical, particularly for low-amperage circuit
breakers (5 A and less). Typical resistance values and corresponding voltage drops at 50
and 100% load factors are shown in Table 3.13.
3.6.3.8 Trip and Switching Configurations Molded case disconnect switches

and circuit breakers can be equipped with a shunt trip mechanism that allows the device
to be opened from a remote location (Fig. 3.47). When used on a battery disconnect
switch or circuit breaker, a push-button switch usually is located near the door to the bat-
tery room so that firefighters or others can operate it to disconnect the battery during an
emergency. The shunt trip usually is powered directly by the battery with fuses providing
overcurrent protection for the push-button circuit wiring. Light-emitting diodes (LED) or
lamps can be provided to indicate switch or circuit breaker states. Figure 3.48 and Table
3.14 show some possible combinations.
3.7 VOLTAGE CONVERSION
Two basic voltage conversion devices are used in telecommunications power systems:
DC–DC converters and inverters.
3.7.1 DC–DC Converters

Most telecommunications switching and transmission equipment operates from –48 V but
the radio frequency (RF) equipment used in many wireless base stations operates from
+24 V. Some older RF equipment used in paging and land mobile radio (LMR) installa-
Fig. 3.46 A dc circuit breaker IT characteristics. (Graph with permission of Airpax Corporation,
Power Protection Products.)
136
3.7 VOLTAGE CONVERSION 137
Table 3.12 dc Circuit Breaker Current–Time Characteristics Comparison

Delay 100% 125% 150% 200% 300% 400% 600% 800% 1000%
51 & 51F No trip 0.5–10 s 0.25–3 s 0.15– 0.05– 0.03– 0.015– 0.01– 0.01–
0.9 s 0.3 s 0.2 s 0.15 s 0.09 s 0.04 s
52 & 52F No trip 9–90 s 5–40 s 2–15 s 0.9–5 s 0.5–3 s 0.03–1 s 0.01– 0.01–
0.28 s 0.08 s
53 & 53F No trip 100– 50– 22– 8–45 s 4–25 s 0.5–5 s 0.01–2 s 0.01–
1000 s 400 s 150 s 0.1 s
Data provided by Airpax Corporation, Power Protection Products.
tions operates from +12 V. Equipment requiring other voltages and polarities has been de-
ployed over the years, for example, +130 Vdc, –130 Vdc, and +48 Vdc for network-con-
trolled payphones (coin phones).
If a site has a mixture of equipment, a separate battery plant may be provided for each
voltage or polarity but, depending on the load currents, it may be more economical to use
dc–dc converters instead. While voltage conversion is the most common application for
dc–dc converters, it is not the only one. Converters may be used to
앫 Change voltage (e.g., from +24 Vdc to –48 Vdc)

앫 Change polarity (e.g., from +24 Vdc to –24 Vdc)
앫 Isolate ground systems or reference planes (note: DC–DC converters have been
marketed in the telecommunications industry that do not have isolated inputs and
outputs and obviously cannot be used for this purpose.)
Table 3.13 Typical Circuit Breaker Resistance and Voltage Dropa

Rated Current DC Resistance Voltage Drop @ 50% Voltage Drop @ 100%
(A) () Load Factor (V) Load Factor (V)
1 1.38 0.690 1.380
2 0.371 0.371 0.742
5 0.055 0.138 0.275
10 0.017 0.085 0.017
20 0.006 0.060 0.120
30 0.003 0.045 0.090
50 0.0019 0.048 0.095
100 0.000375 0.019 0.038
150 0.000375 0.028 0.056
200 0.000225 0.023 0.045
225 0.000225 0.025 0.051
250 0.000225 0.028 0.056
400 0.000125 0.025 0.050
600 0.000083 0.025 0.050
800 0.000063 0.025 0.050
1000 0.000050 0.025 0.050
1200 0.000042 0.025 0.050
a
DC resistances generally have a ±50% tolerance.
Fig. 3.47 Trip control (red push-button on lower left is shown for illustration only as a trip control
serves no useful purpose located so close to the disconnect switch or circuit breaker). (Photo cour-
tesy of Electric Equipment & Engineering Company).
Fig. 3.48 Switch and circuit breaker position indication and remote trip. The circuit breaker and
switch contacts are shown with the circuit breaker off or tripped.
Table 3.14 Switch and Circuit Breaker State Indicators (Typical)

Configuration Red Green Yellow
No indication
Disconnected Off
Connected—Green On
Off–On Off On
Electrically tripped Tripped
Full status Off On Tripped
Connected—Red On
On–Off On Off
앫 Provide voltage stability

앫 Compensate for voltage drop
앫 Relocate power equipment to reduce floor space requirements in the power room
앫 Optimize voltage
The dc–dc converters in telecommunications applications are similar to switch-mode

rectifiers and operate on the same basic principles except that the input rectification stage
is not needed (Fig. 3.49). The DC–DC converters use a transformer to isolate the input
section from the output section (the input dc is converted to ac, which is coupled through
the transformer, and then converted back to dc at the output). This allows a load with one
type of grounding system to be isolated from the power system that has another type of
grounding system. It also helps to prevent disturbances and electromagnetic interference
(EMI) on one side of the converter from being coupled to the other side.
Although the conceptual schematic shows field-effect transistors in the inverter sec-
tion, other types of semiconductors may be used. One type of dc–dc converter uses a
20-kHz switching rate and power transistors in the inverter section (Fig. 3.50). The trig-
ger circuits control transistors Q1 through Q4. Q1 and Q3 operate as a pair and are
turned on simultaneously. When Q1 and Q3 are on, Q2 and Q4, which operate as an-
other pair, are turned off. In the next half-cycle of the switching frequency, Q1 and Q3
are turned off and Q2 and Q4 are turned on. The alternating switching action develops
a chopped dc voltage in the primary of transformers T1 and T2, which is coupled to
the transformer secondary circuits as a square wave where it is rectified by diodes D1
and D2. Subsequent filter circuits smooth the output and remove ripple and switching
noise.
Converters provide a well-regulated output voltage over a wide range of input volt-
ages. For example, a typical 48-V converter used for polarity conversion will hold the
output voltage within ±1% (+47.5 to +48.5 V) with a 30% change in input voltage (–42 to
–60 V). Some converters operate over an input voltage range of 18 to 72 V.
A converter plant has better voltage regulation than a battery plant during discharge
conditions. With a battery plant, the load voltage steadily decreases as the battery dis-
charges, but converters provide fairly good regulation throughout the battery discharge
period.
Voltage drop between the battery plant and the load may be reduced by using larger
conductors or by using converters to provide voltage compensation. In this situation, the
converter regulates the output voltage at the desired value when the voltage drop from the
dc Power
Source
Fig. 3.49 The DC–DC converter conceptual schematic.
battery to the load would otherwise be unacceptable. The converters may be located clos-
er to the load, effectively reducing equipment floor space requirements in the power room
and also reducing the conductor size from the converters to the load equipment (assuming
that voltage drop and not ampacity is the limiting factor in conductor sizing).
With voltage optimization, the converter output voltage is set to the most desirable op-
erating point of the equipment served. Certain equipment may operate more efficiently at
a certain set point voltage or may have increased life or reliability when operated at re-
duced voltage. Converters with nominal output of 48 V are usually set at 48-V output,
whereas 48-V equipment connected to a battery plant normally operates at 52.1 or 54.5 V
depending on the battery technology.
Where the load current exceeds some threshold value, usually around 100 to 200 A,
the costs of a new converter plant generally will be higher than the costs of a dedicated
powerboard and battery plant. Within and above this range, it usually is cheaper to install
a dedicated battery plant. The decision to use a battery plant or a converter plant must take
into account not only capital recovery costs but also operating costs, including disposal,
over the life of the plant. There are many other variables to consider including the avail-
ability of floor space, structural requirements for battery plants, and availability and costs
of ac distribution for rectifiers.
dc Input
dc Output
Fig. 3.50 The DC–DC converter schematic showing basic circuit components.
It may be worthwhile to replace an existing battery plant with a converter plant. In this
case, the converter plant is connected directly to the existing distribution buses (Fig.
3.51). The charge bus, rectifiers, and battery are eliminated.
The DC–DC converters typically are used in a converter plant assembly with distribu-
tion equipment, meters, controls, and alarms dedicated to the converter plant. The stand-
alone application, where a single converter is used that has built-in metering, alarms, and
overcurrent protection, is the limiting case of the converter plant. However, single con-
verters seldom are used; instead, converter plants are arranged in a common-bus or split-
bus configuration.
With the common-bus configuration the converters and overcurrent device distribution
panels are connected to a single common bus (Fig. 3.52). Since the output of converters is
limited to 100% of rating, they may not provide enough fault current to open overcurrent
devices in the converter distribution bus if those devices are too large. Therefore, a fault
in a converter, circuit, or load can disrupt other equipment on the same bus.
The split-bus configuration consists of two common-bus configurations (Fig. 3.53). It
reduces the problem with the common bus in that a fault on one of the buses normally will
not affect the other bus. The split-bus converter plant lends itself to modern applications
in which the load equipment has dual-bus power inputs (A and B). The split-bus configu-
ration is slightly more expensive than the common bus because of the extra bus work,
overcurrent protection, and converters. Each bus is shown with N converters, which effec-
tively provides 2N redundancy. In an alternate configuration, each bus could be equipped
with N + 1 converters.
The two configurations can be varied according to existing system configuration or re-
(a)
(b)
Fig. 3.51 Replacing a battery plant power system with a DC–DC converter system: (a) before and
(b) after.
liability requirements. For example, the converters with their output connected in a split-
bus configuration could be powered by a common-bus configuration on their input. Simi-
larly, the common-bus output configuration could be powered by a split bus on their in-
put.
Modern converters use switch-mode power supply technology and have followed the
same evolutionary path as rectifiers. Modern converters are modular and have relatively
small output ratings (amperes to tens of amperes ranges) compared to older products. A
Fig. 3.52 The DC–DC converter common-bus configuration.
Fig. 3.53 The DC–DC converter split-bus configuration.
143
large converter plant requires numerous converters, and a dedicated battery plant with
rectifiers may be a better alternative.
The DC–DC converters limit their output current, usually to 100% of output rating,
by reducing their output voltage when an overcurrent occurs. This also limits their abil-
ity to open an overcurrent protective device when a fault occurs. To ensure that over-
current devices trip under fault conditions, the individual overcurrent devices on the out-
put bus or buses normally are rated no more than one third the total converter output
current capacity. For example, if the total converter plant has a 100-A rating, the largest
individual overcurrent device would be rated no more than 30 A. The exception is
where a single overcurrent device feeds the load, in which case this kind of coordination
is not possible.
As with the main dc power system, the output from DC–DC converters and associated
systems are referenced to earth ground either through isolated bonding networks or com-
mon bonding networks.
3.7.2 Inverters
Some equipment deployed in telecommunications networks operates only on alternating
current. If this equipment must operate without any interruption whatsoever or cannot with-
stand an interruption longer than approximately 50 ms when the primary power source
fails, it may be connected to an inverter. Inverters are powered by the main system voltage
such as 24 and 48 Vdc. The most common output ac voltage is 120 Vac, one-phase, 60 Hz,
although larger loads may require 208 or 240 Vac, one-phase or three-phase, 60 Hz.
The inverter must regulate the ac output voltage within specified tolerances, typically
±5%, over a relatively large range of input voltage. A typical 48-V inverter will meet out-
put specifications from 42- to 56-Vdc input, while 24-V inverters typically meet output
specifications from 21- to 28-Vdc input. Another important requirement of inverters used
in the telecommunications environment, one that differentiates them from inexpensive
consumer inverters, is the requirement that they provide a relatively distortion-free sine
wave output voltage waveform.
Inverters add load to the main dc power system and may affect battery reserve times
and rectifier system rating. Bus rating also may be affected. It may be more economical to
install an uninterruptible power system (UPS) for ac loads (UPS design is beyond the
scope of this book).
Three basic inverter configurations have been used (Fig. 3.54):
앫 Commercial ac primary with inverter standby—also called passive standby

앫 Inverter primary with commercial ac standby—also called active standby
앫 Inverter primary with no standby—also called continuous operation
In addition to the basic configurations just described, inverters can be provided in re-
dundant arrangements to guard against individual unit failure (Fig. 3.55). When redun-
dancy is required, the inverters can be configured for parallel or standby operation. With
parallel operation, the inverters must have compatible phase synchronization controls so
that all units operate in synchronization and in parallel. There is no break or interruption
on the ac bus when a unit fails. With standby operation, the redundant inverter is powered
but disconnected from the ac output bus by a switch or relay. There typically is a short in-
ac
Distribution
dc
Distribution
To ac Load
Equipment
ac
Distribution
dc
Distribution
To ac Load
Equipment
ac
Distribution
dc
Distribution
To ac Load
Equipment
Fig. 3.54 Inverter configurations.
terruption (5 to 50 ms) between the time the failed inverter goes offline and the standby
inverter is switched into the circuit.
Figure 3.56 shows a conceptual schematic of an inverter. Several methods are used for
inverter output voltage control, including phase angle control, pulse-width control and
pulse-frequency control. In the phase-angle switching control method, the outputs of two
independent inverter circuits are connected together in one package so that their output
voltages add. One of the inverters is controlled by varying the phase of the thyristor trig-
ger pulses so that its output voltage is shifted in phase relative to the output voltage of the
Inverter
ac Bus
dc Input
ac Output
dc Input
ac Output
dc Input
ac Output
Fig. 3.55 Redundant inverter configuration.
dc Power Source
ac Load Equipment
Fig. 3.56 Simplified inverter schematic to illustrate operation.

other inverter. The larger the phase difference between the two inverters, the smaller is
the sum of the two voltages. With this type of voltage regulation, the output voltage can
be held constant regardless of input voltage or load current variations.
The pulse-width control methods trigger and then turn off the SCRs repeatedly during
each cycle of the ac output voltage. In this way, the individual half-waveforms of the out-
put voltage are formed from a number of pulses. By varying the pulse width or pulse rep-
etition frequency, it is possible to regulate the output voltage.
A center-tap circuit is used to achieve the inversion, or periodic reversal, of the di-
rect voltage input. The trigger control alternately triggers the two thyristors SCR1 and
SCR2 into conduction, which allows the input current to flow alternately through the
transformers two primary windings P1A and P1B. The alternating current in the prima-
ry is coupled to the secondary winding S1 and appears as a square wave voltage at the
required fundamental frequency (60 Hz). The output is filtered by L2, L3, and C2 so
that the inverter output is a relatively clean (undistorted) sine wave. One of the advan-
tages of having a transformer separating the input and output is the isolation it provides,
and the output circuit may be referenced to a different grounding system than the input
if desired.
When SCRs are used in ac circuits, they turn off when the alternating voltage across
them goes through a zero crossing. In dc circuits, however, a quenching capacitor must be
used in the circuit to turn off the thyristors (C1).
Pulse-width and pulse-frequency control are used in the inverter shown conceptually
in Figure 3.57.
Circuit Description The dc input is supplied through circuit breaker CB1 to the in-
verter circuits. Capacitor C1 filters the input to prevent coupling of inverter noise back to
the input circuit. The trigger circuit generates pulses to turn on thyristors SCR1 and
SCR2. Two separate pulse trains are generated at a basic frequency of 60 Hz and 180° out
of phase with each other. When SCR1 is triggered and starts to conduct, full input voltage
ac
Output
dc Input
Fig. 3.57 Inverter simplified schematic. This inverter uses pulse-width and pulse-frequency control.
is applied across ferroresonant transformer T1 primary windings P1A. Primary windings

P1A and P1B act as a step-up autotransformer and apply twice the input voltage across
capacitor C2.
At the proper time, the trigger circuit applies pulses of opposite phase to SCR2 causing
it to conduct. At this time, both thyristors are turned on and C2 is shorted out. C2 dis-
charges through SCR2 and current attempts to flow in reverse through SCR1 causing it to
turn off. Since SCR2 still is conducting, full input voltage is applied across primary wind-
ing P1B. Through autotransformer action, twice the input voltage is applied across C2 but
the voltage has opposite polarity to the previous cycle. At the proper time, the trigger cir-
cuits turn on SCR1 causing it to conduct and discharge C2. When C2 discharges, the re-
verse current turns off SCR2. This cycle repeats itself at the frequency determined by the
control circuits (60 Hz). The changing voltage polarity appears as a square wave across
the primary windings of T1.
Inductor L1 and capacitor C2 form a resonant circuit that provides a commutating
pulse to turn off the conducting thyristor. Capacitor C4 prevents high-frequency oscilla-
tion, while diodes D3 and D4 prevent the voltage across T1 primary from exceeding the
input voltage. These diodes also compensate for leading and lagging power factor in the
inverter load circuits by providing a return path to the dc input circuit for reactive cur-
rents.
The square wave input to the transformer is coupled to the secondary windings S1 and
S2 and filtered by the transformer’s inherent inductance giving a distorted sine wave at
the output. The transformer secondary windings inductances and capacitor C3 form a sec-
ondary resonant circuit that helps to maintain transformer core saturation and output volt-
age regulation even though the load or input voltage may change.
If there is an output overload or short circuit, the field in the core will collapse because
the energy drained from the ferroresonant circuit exceeds what is required to maintain
core saturation. The transformer output voltage decreases and limits the output current to
a safe value.
Inductor L2 limits harmonic currents at the output of the transformer and thus limits
output voltage waveform distortion. Capacitors C5 and C6 reduce electromagnetic inter-
ference by suppressing high-frequency currents at the inverter output.
REFERENCES
1. ANSI/T1.315-2001, Voltage Levels for DC-Powered Equipment Used in the Telecommunica-

tions Environment, Alliance for Telecommunications Industry Solutions (ATIS), 2001.
2. W. D. Reeve, Subscriber Loop Signaling and Transmission Handbook: Analog, IEEE Press,
Piscataway, NJ, 1992.
3. ANSI/T1.401-2000, Network to Customer Installation Interfaces—Analog Voicegrade
Switched Access Lines Using Loop-Start and Ground Start Signaling, ATIS, 2000.
4. ANSI/T1.405-2002, Network-to-Customer Installation Interfaces—Direct Inward Dialing Ana-
log Voicegrade Switched Access Using Loop Reverse-Battery Signaling, ATIS, 2002.
5. ANSI/T1.409-2002, Network-to-Customer Installation Interfaces—Analog Voicegrade Special
Access Lines Using E&M Signaling, ATIS, 2002.
6. ANSI/IEEE Std 820-1984, IEEE Standard Telephone Loop Performance Characteristics, IEEE,
1984.
7. W. D. Reeve, Analog Telephone Channels and the Subscriber Loop, Communications Hand-
book, J. Gibson, Ed, CRC Press, Boca Raton, FL, 2002.
REFERENCES 149
8. Code of Federal Regulations (CFR) Title 47, Part 15, Radio Frequency Devices, Federal Com-
munications Commission, 2005.
9. GR-63, NEBS™ Requirements: Physical Protection, Telcordia, 2002.
10. GR-1089, Electromagnetic Compatibility and Electrical Safety—Generic Criteria for Network
Telecommunications Equipment, Telcordia, 2002.
11. ANSI/PE-7-1997 (R2003), Communications Type Battery Chargers, National Electrical Manu-
facturers Association, 2003.
12. ANSI/T1.304-1997, Ambient Temperature and Humidity Requirements for Network Equip-
ment in Controlled Environments, ATIS, 1997.
13. ANSI/T1.307-2003, Fire Resistance Criteria—Ignitability Requirements for Equipment As-
semblies, Ancillary Non-metallic Apparatus, and Fire Spread Requirements for Wire and Ca-
ble, ATIS, 2003.
14. ANSI/T1.319-2002, Equipment Assemblies—Fire Propagation Risk Assessment Criteria,
ATIS, 2002.
15. D. Fink, and J. Carroll, Eds., Standard Handbook for Electrical Engineers, 10th ed., McGraw-
Hill, New York, 1969.
16. ANSI/IEEE Std 242™-2001, IEEE Recommended Practices for Protection and Coordination of
Industrial and Commercial Power Systems, IEEE, 1986.
CHAPTER 4
TELECOMMUNICATIONS BATTERIES
This chapter provides a detailed description of the components that make up a battery and
what may be expected from them in telecommunications applications. Battery sizing and
design are covered in Chapter 5 and battery maintenance and installation considerations
are covered in Chapter 6.
4.1 BATTERIES AND CELLS
4.1.1 Types
Batteries are the energy storage component in a telecommunications power system. The
basic battery technology most often used is lead–acid.1 Two types are used: vented
lead–acid (VLA) and valve-regulated lead–acid (VRLA). These terms indicate the
method used to handle the gases that are produced during operation. The VLA cell (also
called flooded or wet cell) allows the gases to freely escape from the cell container while
the VRLA limits the amount of gases that escape except under abnormal conditions. Both
types are categorized as secondary batteries, which indicates they are rechargeable (as op-
posed to primary batteries, which are not rechargeable and are not used in telecommuni-
cations applications) and stationary by nature of their fixed, nonvehicular installation.
A battery stores electric energy and delivers it to the load on demand. Rectifiers
(chargers) recharge the battery after a discharge and provide energy to loads during nor-
mal operation. Also, during normal operation, the rectifiers provide float current to the
battery to make up for losses due to the self-discharge of the cells in the battery. A
telecommunications battery spends most of its life in the fully charged state. Depending
on the length of time that rectifiers are not available due to ac power source failure, the
discharge portion of the cycle could last from a few minutes to 12 h or more. The battery
discharges at a rate determined by the reserve time and discharge current. Upon restora-
tion of the rectifier’s ac power source, a fully discharged battery recharges in (typically) 8
to 24 h and then floats at a constant voltage until the next power source failure.
1
Lithium–metal polymer (LMP) and nickel–cadmium (NiCd) batteries are used in some small-capacity applica-
tions.

152 TELECOMMUNICATIONS BATTERIES
4.1.2 Cell Configurations

A cell is the basic electrochemical unit in a battery. At a minimum a cell consists of two
electrodes immersed in an electrolyte, in a container (or jar). The cell converts chemical
energy into electric energy by an electrochemical reduction–oxidation (redox) or decom-
position–recomposition process, which is more fully described in later paragraphs. Re-
versing the redox process recharges the cell.
Batteries used in telecommunications usually consist of 12 cells (24 V nominal system
voltage) or 24 cells (48 V) connected in series.2 Each lead–acid cell has a nominal voltage
of 2 V. The series cell combination is called a battery string (Fig. 4.1). Battery strings are
connected in parallel to provide higher capacity (Fig. 4.2). In smaller cells sizes, more than
one cell may be combined in a package called a monobloc (also called monoblock). The
cells in a monobloc may be connected in parallel to increase capacity or in series to increase
voltage. A typical monobloc has three cells in series and a nominal voltage of 6 V, although
six cells in series for a nominal voltage of 12 V also are common. A battery system consists
of one or more battery strings in parallel and associated mounting frames or racks.
Battery capacity is measured in units of ampere-hours (Ah), which sometimes is de-
noted as “C” in manufacturer’s literature. Ampere-hours indicate the load current mea-
sured in amperes that may be drawn from the battery over a time period measured in
hours.3 For example, if a 12-cell battery is discharged to an end voltage of 21.0 V (1.75
V/cell) by a load of 12 A in 8 h, the battery capacity supplied to the load under those con-
ditions is 12 A × 8 h = 96 Ah. The stated capacity has to be qualified with a particular end
voltage, battery temperature, nominal full-charge electrolyte specific gravity, and dis-
charge rate because battery capacity varies with each of these factors (this relationship is
described in more detail later).
Vented lead–acid cells range in capacity from about 50 to 8000 Ah. Figure 4.3 shows
line drawings of two sizes: 200 and 8000 Ah. VRLA cells range in capacity from much
less than 50 Ah to around 2000 Ah. Examples of two VRLA cells are shown in Figure
4.4. VLA cells always are mounted with the vent and terminal posts at the top while most
VRLA cells may be mounted in any position but usually vertically with posts at the top or
horizontally with posts at the side. Some larger VRLA absorbed glass mat (AGM) cells
must be mounted horizontally for proper operation.
4.1.3 Active Components

Cells have three active components (Fig. 4.5)—negative electrode (or plate, which is the
terminology used in this book) of almost pure sponge lead (Pb), positive plate of pasted
lead dioxide (PbO2, also called lead peroxide), and electrolyte of water-diluted sulfuric
acid (H2SO4). The plates are built from lead alloy in the form of a grid frame, which not
only conducts the electrical current but also holds the active materials. The active materi-
als are made from lead oxide mixed into a paste with electrolyte and water, pasted into the
grids, and allowed to dry. Cells for telecommunications batteries generally are assembled
and placed in the container or jar and filled with electrolyte. Current is then forced
2
Some 48-V installations may use 23 cells because of load or power equipment limitations, and other cell counts
may be found in older installations. Also, higher voltages, such as 130 V, exist in older telecommunication sys-
tems. These systems consist of 60 to 65 cells in series, although other cell counts have been used.
3
A charge of 1 Ah is equivalent to 3600 coulombs (C).
4.1 BATTERIES AND CELLS 153
Fig. 4.1 Series connection of 24 cells to form a 48-V battery string. If the capacity of each cell is
100 Ah, the capacity of the series string is 100 Ah. In this illustration, there are eight modules con-
sisting of three cells each.
Fig. 4.2 Parallel connection of three 24-V battery strings to increase capacity. If each string is rated
100 Ah, the capacity of the parallel combination is 300 Ah.
Fig. 4.3 Examples of VLA cells. The flame arrester vent is in the top-middle of the cells and each
cell has four terminal posts.
through the plates to form them electrochemically into the negative and positive plates.
The final step is to test the cells prior to shipment and service. The negative plate is a gray
metallic lead color while the lead dioxide positive plate is a chocolate-brown color.
In some situations, it is desirable to assemble, ship, and then store VLA cells in a dry
condition (no electrolyte) for some time prior to service. Manufacturers refer to this as a
“dry-charged” battery. In this case, the cells are built as already described except that the
plates are formed in tanks and then dried before being assembled. Once assembled the
plates are placed in the container (jar) without any electrolyte. The cells are sealed to keep
out air and moisture and shipped to the storage site. The cells then may be stored in a
cool, dry location for as long as 2 years. Some manufacturers also supply moist-charged
cells. Just prior to being placed in service, the cells are filled with electrolyte and fully
charged using elevated (activation) voltage. Although the manufacturer will provide the
correct specific gravity electrolyte, it can be mixed on-site if desired. If mixed on-site, the
specific gravity normally is mixed to about 10 points (0.010) lower than the fully charged
specific gravity. For example, if the final specific gravity should be 1.215, the electrolyte
would be mixed to 1.205 before it is added to the cells.
Cells have more than one positive plate and more than one negative plate. The plates
of a given polarity are connected in parallel (Fig. 4.6). There may be an even number of
plates, giving an equal number of positive and negative plates or, as is most often the
Fig. 4.4 Examples of VRLA cells. Depending on the cell dimensions, one to six cells may be mount-
ed in a steel frame to form a module. Compare the size and weight of the VRLA cells to the VLA cells
in the previous figure.
Fig. 4.5 Cell active components.
Fig. 4.6 Multiplate cell.

case, odd number, giving one more negative plate than positive. If odd, the number of
positive plates can be calculated from
Total – 1
Positives = (4.1)
2
Example 4.1 An 8000-Ah cell has a total of 41 plates. Determine the number of positive
plates.
Solution The number of positive plates is (41 – 1)/2 = 20.
4.1.4 Electrochemistry
The overall chemical reaction from the charged state to discharged state and from the dis-
charged state to the charged state is shown in (4.2). These reactions apply to both VLA
and VRLA cells.
Discharge 씮
Pb + PbO2 + 2H2SO4 2PbSO4 + 2H2O (4.2)
씯 Charge
where H2O = water

H2SO4 = sulfuric acid
Pb = lead (negative plate)
PbO2 = lead dioxide (positive plate)
PbSO4 = lead sulfate
During discharge, the lead in the negative plate partially dissolves in the sulfuric acid and
goes into a solution state. When load current flows through the battery on discharge, ion-
ization occurs in which each lead atom loses two electrons as shown in (4.3):4
Pb 씮 Pb2+ + 2e– (4.3)
where Pb2+ indicates ionized lead and e– indicates a free electron.

The electrons flow through the external load circuit from the negative plate to the pos-
itive plate (Fig. 4.7). The positive plate accepts the electrons from the load during dis-
charge as the plate is reduced. The electrolyte completes the internal battery circuit by
furnishing ions and, thus, a conductive path between the negative and positive plates. A
microporous plate separator prevents the positive and negative plates from shorting out
while allowing ions to flow from plate to plate through the electrolyte.
The lead atom in the positive plate absorbs the two electrons as shown in (4.4):
Pb4+ + 2e– 씮 Pb2+ (4.4)
where Pb4+ indicates ionized lead (short four electrons) in the lead dioxide positive plate
and e– indicates a free electron flowing from the negative plate through the load.
4
An ion is an atom or group of atoms in which the number of electrons is different from the number of protons
and thus having a net positive or negative charge.
Fig. 4.7 Cell discharge circuit.
The reduction (decomposition) at the positive plate breaks the lead dioxide molecule
into Pb2+ and O2. While these reactions are taking place, the sulfuric acid molecule
(H2SO4) breaks into hydrogen ions (H+) and sulfate ions (SO42–). The H+ ions from the
sulfuric acid and the O2 from the lead dioxide unite and form water (H2O) as shown in
(4.5):
4H+ + PbO2 + 2e– 씮 Pb2+ + 2H2O (4.5)
The lead ions at the negative and positive plates from (4.3) and (4.4) unite with the sulfate
ions to form lead sulfate as shown in (4.6)
Pb2+ + SO42– 씮 PbSO4 (4.6)
The chemical reactions (4.3) and (4.6) at the negative plate combine as shown in (4.7):
Pb + SO42– 씮 2e– + PbSO4 (4.7)
Similarly, the reactions (4.5) and (4.6) at the positive plate combine as shown in (4.8):
PbO2 + 2e– 씮 PbSO4 + 2H2O (4.8)
As a final check, (4.7) and (4.8) may be combined to give the discharge direction shown in
(4.2). The electrolyte acid concentration is reduced as the battery discharges. The elec-
trolyte in a fully discharged lead–acid battery has a specific gravity of about 1.065 to 1.100.
When the cell is recharged, the chemical processes just described are reversed. Recti-
fiers force a current into the cell in the opposite direction by providing a voltage source
greater than the cell. At the positive plate, the water and lead sulfate break into lead diox-
ide, hydrogen ions, sulfate ions, and two free electrons as shown in (4.9):
2H2O + PbSO4 씮 PbO2 + 4H+ + SO42– + 2e– (4.9)

The hydrogen ions and sulfate ions combine (the action of recomposition) to form sulfu-
ric acid, and the negative plate gains the two electrons lost by the positive plate. There-
fore, the lead atoms in the lead sulfate at the negative plate get back their two electrons
and change back into plain lead as shown in (4.10):
PbSO4 + 2e– 씮 Pb + SO42– (4.10)
Reactions (4.9) and (4.10) combine as shown in (4.11), which is identical to the charge di-
rection shown in (4.2):
2H2O + 2PbSO4 씮 Pb + PbO2 + 2H2SO4 (4.11)
As a cell approaches full charge, it cannot absorb all the energy from the charging cur-
rent, and the excess energy dissociates water by electrolysis into its components hydrogen
and oxygen. The water ionizes slightly into hydrogen ions (H+) and hydroxyl ions (OH–).
The positive hydrogen ions are attracted to the negative plate, where they receive an elec-
tron and become a hydrogen atom. Each hydrogen atom unites with another hydrogen
atom to form a molecule of hydrogen gas as shown in (4.12):
2H+ + 2e– 씮 H2 (4.12)
The hydroxyl ions at the positive plate lose four electrons and break into water molecules
and oxygen gas molecules as shown in (4.13):
4OH– – 4e– 씮 2H2O + O2 (4.13)
Reactions (4.12) and (4.13) combine to provide water electrolysis as shown in (4.14):
2H2O 씮 2H2 + O2 (4.14)
Once the cell reaches full charge, the charge voltage remains constant. The charge voltage
is the recommended float voltage for the battery, and the current decreases to a float val-
ue. The energy from the float current is consumed by continued electrolysis and by inter-
nal resistance losses.
The gassing reactions above indicate the evolution of hydrogen (H) and oxygen (O) in
a ratio of 2 : 1. For every 1 Ah of overcharge, about 0.73 liters (0.026 ft3) of gas are pro-
duced of which 0.49 liters (0.017 ft3) are hydrogen and 0.24 liters (0.0085 ft3) are oxygen.
The open vent on a VLA cell allows the gases to escape. The electrolysis reduces the
amount of water in the electrolyte, which effectively raises the specific gravity and drops
the electrolyte level. The water loss must be replaced regularly to prevent the electrolyte
level in the cells from dropping to the point that the plates will become exposed to air and
damaged. Gas bubbles may be seen clinging to the plates and bubbling to the surface in
cells with transparent plastic containers. The negative plates will have twice as many (hy-
drogen) bubbles as the (oxygen) bubbles at the positive plates.5
5
It is the incorrect and poor practice of some technicians to rap the cell containers with their knuckles to break
the bubbles loose. However, the bubbles are part of the natural chemical reactions and should be left to break
loose on their own.
A VRLA cell operates on the same electrochemical basis as the VLA cell but depends
on the recombination of oxygen and hydrogen during overcharge. In a recombinant cell
the same electrolysis takes place as shown in (4.14). The oxygen that is evolved at the
positive plate diffuses through the plate separators and reacts with the negative plate to
produce lead oxide (4.15). This causes a slight depolarization (reduction in the overvolt-
age required to drive the electrochemical process) and inhibits the further release of hy-
drogen, (4.16) and (4.17). The lead oxide reduces to lead sulfate (4.18), which on further
charge goes to lead and sulfuric acid (4.19). The result is no net loss of oxygen or hydro-
gen (or water), although some loss does occur since the recombination reactions are not
100% efficient:
2Pb + O2 Overcharge씮 2PbO (initial reaction) (4.15)
4H+ + 4e– Overcharge씮 2H2 (hydrogen evolution at (4.16)

negative plate)
O2 + 4H+ + 4e– Overcharge씮 2H2O [oxygen reduction (water (4.17)

recapture)]
2PbO + 2H2SO4 Overcharge씮 2PbSO4 + 2H2O (intermediate reaction) (4.18)
2PbSO4 + 4H+ + 4e– Overcharge씮 2Pb + 2H2SO4 (final reaction) (4.19)
Under conditions of gross overcharge, such as when a battery string has one or more
shorted cells or if rectifiers are misadjusted, more than normal float current flows into the
cell. In this case, not all oxygen and hydrogen is recombined and the gas pressure within
the cell builds up. To prevent an explosion due to overpressure, all VRLA cells are
equipped with a self-resealing pressure relief valve to vent the gases and reduce the inter-
nal pressure. While the risk of dry-out of a properly installed and operated VRLA cell is
small, electrolyte still can be lost through a failed vent, post seal, cover seal, or a crack in
the container or by slow migration of water vapor through the cell container itself.
Figure 4.8 is a more detailed illustration of a battery throughout a full cycle—fully
charged (floating), discharging, fully discharged, and charging. When in the fully charged
state, the cells contain maximum lead in the negative plate, maximum lead dioxide in the
positive plate, and minimum lead sulfate (upper left). The acid concentration in the elec-
trolyte is at its maximum level (maximum sulfuric acid and minimum water). During dis-
charge, electron current flows from the negative plate through the load to the positive
plate. As the battery discharges, the amount of metallic lead in the positive and negative
plates decreases and lead sulfate increases (upper right). At the same time, the electrolyte
acid concentration decreases (amount of sulfuric acid decreases and the amount of water
increases). When the battery is fully discharged, the lead sulfate is maximum and lead in
the plates in minimum (lower right). The acid concentration is at its minimum level (min-
imum sulfuric acid and maximum water). When the battery recharges, current flows from
the rectifiers into the negative plate, through the battery, and out the positive plate. The
lead sulfate is converted to lead at the negative plate and lead dioxide at the positive plate
(lower left). During recharge the acid concentration increases (amount of water decreases
and amount of sulfuric acid increases).
Fig. 4.8 Battery cycle.
Cells also consist of a separator and container (including lid or cover) and other packag-
ing components. The separator physically separates and electrically insulates the positive
and negative plates from each other but allows transport of ions between plates during
charge and discharge. Separator failure is one cell failure mechanism—in this case, the two
plates short out and discharge the cell. In lead–acid batteries, the separator is a microporous
rubber or fiber material (older cells used glass or even wood). The separator is not required
for electrochemical operation, but it absorbs part or the entire electrolyte, keeping it close
to the plate and increasing cell efficiency. The separator used in VRLA cells is a highly
porous, absorbent, microfiberglass mat and plays a doubly important role by soaking up the
electrolyte and immobilizing it. Electrolyte normally cannot leak out of the VRLA cell un-
less it is forced out or oozes out under pressure during abnormal operation.
Regularly discharging and charging a cell works active material loose from the posi-
tive plates. This material settles to the bottom of the cell (Fig. 4.9), so containers must
have extra space at the bottom for collection; otherwise, too much sediment will pile up
and short the plates. Separators have a smooth side next to the negative plate and a
grooved side (backweb) next to the positive plate. The grooved side is next to the positive
plate to allow gas that is formed to escape and any loosened active material to fall to the
bottom of the container. Most VLA cells used in telecommunications applications have a
fiberglass mat attached to the separator backweb to prevent active material from shedding
from the positive plate. The mat also keeps a ready supply of electrolyte near the positive
plates.
Fig. 4.9 Sediment at the bottom of a cell. (Photo courtesy of M.W. Migliaro.)
Unlike VLA cells and gelled electrolyte VRLA cells, which have excess electrolyte to
compensate for water loss during operation, absorbed electrolyte VRLA cells contain
only a minimum amount of electrolyte. For example, a typical 1400-Ah VLA cell has ap-
proximately 13 gal of electrolyte, but a typical 1400-Ah VRLA cell has only 4.1 gal. Ab-
sorbed electrolyte VRLA cells often are referred to as “electrolyte starved.” As a result,
the container for absorbed electrolyte VRLA cells is much more compact than VLA for a
given ampere-hour capacity, and they can be mounted in any position. VLA cells and
gelled electrolyte VRLA cells are mounted with their vent at the top to prevent spilling.
4.1.5 Cell Containers

Various materials are used in the cell containers. Most of the original lead–antimony
telecommunications batteries used a hard rubber and were flexible enough that plate
growth or mishandling seldom cracked the container and allowed it to leak; glass jars
were used in some smaller batteries. Except in the largest sizes, lead–calcium and modern
lead–antimony cells generally use a transparent plastic container made from styrene,
polycarbonate, or flame-retardant polyvinyl chloride (PVC). Early lead–calcium cell con-
tainers were sensitive to stress caused by excessive plate growth and subsequently
cracked and leaked, leading to a large number of battery fires. These packaging problems
largely were solved by controlling the calcium to better control plate growth.
The limiting oxygen index, or LOI, is used to measure the relative flame retardant
characteristic of a material. LOI is the minimum concentration of oxygen in an
oxygen–nitrogen mixture that will support continued combustion of a particular material
once the flame source is removed. Air is 21% oxygen so any material with an LOI of 21%
or less will burn readily in air. A higher value indicates less flammability. Typically, a
polycarbonate or flame-retardant PVC material with an LOI > 28% is used in cells where
flame-retardant properties are required, such as in all central office applications. It should
be noted that many standard catalog cells are not flame-retardant (e.g., styrene containers
do not have LOI > 28%) and, if a flame-retardant container is needed, it must be specified
in the battery purchase specification and costs extra.
Large hard rubber cells (4000 Ah and larger) sit directly on the battery room floor and
have a steel angle frame built around them in seismic applications. Small to medium cells
with plastic cases (above 100 to 3900 Ah) are mounted on battery racks, usually two tiers
high (Fig. 4.10), while the smallest cells (100 Ah and smaller) may be rack mounted or in-
stalled on shelves in a frame or relay rack or cabinet (Fig. 4.11).
Except in some very small installations, the frames, racks, and shelves are conductive
steel and, for safety purposes, are bonded to ground. Also, concrete floors on which the
racks or cells sit are conductive. Any electrolyte leakage usually leads to an electrical path
to ground and possibly arcing, which can ignite the cell containers. Therefore, much ef-
fort has gone into perfecting container and cover materials and posts and seals to prevent
leakage. The potential for acid leaks is one of the reasons VLA batteries seldom are in-
stalled in network equipment rooms and, instead, almost always are installed in separate
rooms.
Vented lead–acid batteries have been used in their present form for about 100 years.
VRLA batteries did not find widespread use in telecommunications until the mid-1980s.
The majority of installations today use VRLA technology, and almost all outside plant
equipment and remote telecommunications enclosures such as access nodes and digital
loop carrier remote terminals use VRLA technology (NiCd batteries and a limited amount
of alternate technologies, such as lithium–metal polymer, are used in some applications).
VRLA batteries have been given a number of different names, mainly by manufacturers
eager to differentiate their products. Some names are “maintenance free” (a definite mis-
nomer), SLA (sealed lead–acid), AGM (absorbent glass mat), “gel-cell,” suspended or ab-
sorbed electrolyte, and others.
Fig. 4.10 Medium VLA clear plastic 24-cell battery in seismic zone 4 rack.
Fig. 4.11 Small 6-cell VRLA monoblocs on a frame-mounted tray (four monoblocs visible in front
and four in back in middle-bottom of picture). (Photo courtesy of M.W. Migliaro.)
Roughly 60% of the cell weight is lead plates and lead components. Although pure lead
is the ideal plate material, it is too soft except in certain battery types such as the round cell
designed by the AT&T Bell System (Fig. 4.12).6 For modern stationary battery applica-
tions, adding antimony (approximately 4% by weight, chemical element Sb) hardens the
lead. Other variations also have been used, including 1 to 2% antimony with some seleni-
um to act as a grain refiner. These are called “lead–selenium” or “low-antimony” cells.
Another lead-hardening method uses calcium (approximately 0.03 to 0.05% by weight,
chemical element Ca). Tin also is used (approximately 0.8% by weight, chemical element
Sn). Most modern VLA installations use what are called lead–calcium cells, although these
cells actually use a tertiary alloy consisting of lead, calcium, and tin. The tin provides addi-
tional corrosion resistance for the plate grid. Also, some manufacturers add aluminum to
the lead–calcium–tin alloys to better control calcium content. Very little of the aluminum
winds up in the final grid, but it is used to provide a layer on top of the molten alloy during
casting to minimize oxidation and subsequent loss of calcium. VRLA cells use lead–calci-
um or lead–tin, although at least one manufacturer uses a patented alloy with antimony.
The first lead–calcium cells were deployed in the telecommunications network around
1950, but it took about 20 years to solve all the problems associated with them. Lead–cal-
cium is viewed as a superior material combination mainly because the water consumed by
the lead–calcium VLA battery is much smaller (by a factor or 1/10 to 1/20), as is the float
current when compared to lead–antimony. The reduction in water consumption leads di-
6
The round cell uses pure lead plates, which minimize plate corrosion and growth, and a positive plate paste that
provides additional mechanical stability. The round shape minimizes the stress effects of plate growth that does
occur. The plates are supported in a unique structure that relies less on the plates themselves. The original design
life of the round cell was 30 years, although current marketing materials claim 70-year life. Because the round
cell uses different manufacturing processes and materials than conventional stationary cells, the costs are much
higher (approximately 2 to 2.5 times for a given ampere-hour capacity).
Fig. 4.12 Battery consisting of round cells and seismic zone 4 rack installed in southeastern Alaska.
Round cells are not very space efficient, but they have a relatively long life when properly operated
and maintained.
rectly to reduced maintenance costs because the electrolyte level in the battery does not
need to be checked and water replenished as often. There is very little water loss from a
VRLA cell because of the recombination reactions that take place in the cell while it is
under positive pressure (i.e., regulated by its valve).
4.1.6 VLA and VRLA Comparison

VLA Advantages
앫 Container usually is transparent so visual examination is possible.
앫 Longer life.
앫 Not subject to thermal runaway.
VLA Disadvantages
앫 :Requires regular electrolyte maintenance (water addition).
앫 Releases hydrogen gas during charging.
앫 Can be installed only in one position (vent at top).
앫 Racks must be specially designed for seismic applications.
앫 Requires flame arrester vent.
앫 Requires spill containment system (except in small sizes).
VRLA Advantages
앫 Normally low hydrogen gas venting (except during overcharge).
앫 Higher energy density.
4.2 BATTERY VOLTAGE AND DISCHARGE CURVES 165
앫 Can be installed in any position (absorbed electrolyte cells).

앫 Smaller footprint (can be stacked vertically).
앫 Acid containment system normally not required (unless by local codes).
앫 Most battery racks inherently seismic rated.
앫 Built-in flame arrester vent (if specified).
VRLA Disadvantages
앫 Opaque container or steel frame around modules so visual examination is impossi-
ble.
앫 Shorter life.
앫 Failure mechanisms not seen in VLA cells.
앫 Subject to thermal runaway.
앫 Very heavy floor loading in stacked installations.
4.2 BATTERY VOLTAGE AND DISCHARGE CURVES
4.2.1 Cell Voltages and Specific Gravity

The open-circuit voltage of a fully charged lead–acid cell depends directly on the elec-
trolyte specific gravity. Specific gravity is the ratio of the electrolyte density to water den-
sity at the same temperature. Pure water has a specific gravity of 1.000 and pure sulfuric
acid has a specific gravity of 1.835.7 The specific gravity used in telecommunications
VLA cells usually is 1.215, for absorbed electrolyte VRLA cells is 1.300, and for gelled
electrolyte VRLA cells is 1.260 to 1.280 (however, variations exist among different man-
ufacturers of VRLA cells) at 25°C (77°F). The approximate relationship between cell
voltage and specific gravity is
OCV ⬇ 0.845 + SG (4.20)
where OCV = cell open circuit voltage (V)

SG = electrolyte full-charge specific gravity at 25°C (77°F)
For the typical specific gravities previously mentioned, the open-circuit voltage of a VLA
cell is 2.06 V and of an absorbed electrolyte VRLA cell is 2.15 V. The open-circuit volt-
age and specific gravity depend on temperature. The temperature coefficient for open-cir-
cuit voltage is variously reported as 0.2 to 5.5 mV/°C (0.33 to 9.1 mV/°F) at 25°C (77°F)
[1–3], and the temperature coefficient for specific gravity is 0.0006 to 0.00075/°C (0.001
per 3°F) at 25°C (77°F) for SG in the range of 1.215 to 1.300 [1]:
VPC(T) = VPC(25°C) + V(25 – T) (4.21)
where VPC(T) = cell voltage at temperature T (V)

VPC(25°C) = cell voltage at 25°C (V)
T = temperature (°C)
V = temperature coefficient for open-circuit voltage (0.2 to 5.5 mV/°C)
7
One cubic centimeter of water at +4°C weighs 1 g and the same amount of sulfuric acid at the same temperature
weighs 1.835 g.
SG(T) = SG(25°C) + SG(25 – T) (4.22)
where SG(T) = specific gravity at temperature T

SG(25°C) = specific gravity at 25°C
T = temperature (°C)
SG = temperature coefficient for specific gravity (0.0006 to 0.00075/°C)
Example 4.2 The measured cell voltage is 2.170 V at 33°C (91°F). Determine the volt-
age corrected to 25°C (77°F).
Solution First, rearrange Eq. (4.21) to solve for VPC (25°C):
VPC(25°C) = VPC(T) – V(25 – T)
The difference between the base temperature and the actual temperature is 25 – 33 =
–8°C. Assuming V = 5.5 mV/°C, the correction is –8°C × 0.0055 mV/°C = –0.044 V so
the cell voltage corrected to the base temperature is 2.170 V – (–0.044 V) = 2.214 V.
Example 4.3 The measured cell voltage is 2.200 V at 15°C (59°F). What is the voltage
corrected to 25°C (77°F)?
Solution The difference between the actual temperature and the base temperature is 25 –
15 = 10°C. Assuming V = 5.5 mV/°C, the correction is 10°C × 0.0055 mV/°C = 0.055 V
giving a cell voltage corrected to the base temperature is 2.200 V – (+0.055 V) = 2.145 V.
Note: This indicates that, although the measured voltage appears proper, the cell really is
being undercharged based on the operating temperature.
Example 4.4 The measured specific gravity in a VLA cell is 1.204 at 33°C (91°F). De-
termine the specific gravity corrected to 25°C (77°F).
Solution The difference between the base temperature and the actual temperature is 25 –
33 = –8°C, and, assuming SG = 0.0006/°C, the correction is –8°C × 0.0006 = –0.005 giv-
ing a corrected specific gravity at 25°C of 1.204 – (–0.005) = 1.209 SG.
Example 4.5 The measured specific gravity in a VLA cell is 1.215 at 15°C (59°F). De-
termine the specific gravity corrected to 25°C (77°F).
Solution The difference between the base temperature and the actual temperature is 25 –
15 = +10°C, and, assuming SG = 0.0006/°C, the correction is +10°C × 0.00006 = +0.006
giving a specific gravity corrected to the base temperature of 1.215 – (+0.006) = 1.209
SG.
4.2.2 Float and Equalize Charge Voltages

Battery design life will not be achieved unless the float voltage is well regulated. The reg-
ulation normally is ±0.5%, which for most 48-V systems is ±0.260 to 0.273 V and for
most 24-V systems is ±0.130 to 0.136 V (depending on whether VLA or VRLA batteries
are used). The rectifiers largely determine the regulation, although poor design of the rec-
tifier dc circuits can affect it. Rectifier dc circuit design is covered in Chapter 5, System
Design.
In addition to the overall battery float voltage regulation, the individual cell voltages
and corresponding specific gravities must be maintained within certain limits, typically
±0.02 V and ±0.04 V of the average cell voltage for VLA lead–antimony and VLA
lead–calcium cells, respectively, and within 10 points of the specific gravity (for VLA,
specific gravity is 1.215 ±0.010). The voltage tolerance for VRLA cells typically is
greater, and some manufacturers specify three different tolerances depending on cell age;
for example, ±0.13 V for < 6 months, ±0.08 V for 6 to 12 months, and ±0.06 V for > 12
months.
Since the specific gravity of VRLA cells cannot be measured, only cell voltage has op-
erational relevance. For example, say a 24-cell VRLA battery is floated at 54.48 V, which
is equivalent to 2.27 V/cell. If, in this example, the cell voltage tolerance is ±0.06 V, the
acceptable voltage measured at any individual cell could range from 2.21 to 2.33 V. Cells
that operate at a higher than normal voltage are in a relative overcharge condition and
cells that operate lower than normal are in an undercharge condition. VLA and VRLA
batteries require different float voltages (and have different voltage tolerances).
Another difference worth noting is the need for equalizing in VLA batteries. The volt-
age across individual VLA cells or specific gravity in float operation will drift apart over
time, particularly if subject to cycling, and exceed the tolerance mentioned above. Raising
the battery charging voltage for some length of time (typically 24 to as much as 200 h)
until all cells meet the voltage tolerance is known as equalize charging or equalizing.8
During the equalize period, the cell voltage for higher than normal cells will decrease and
for lower than normal cells will increase. Equalizing too often or for too long will cause
excess gassing, excess loss of water from the electrolyte, and shorter battery life.
Lead–calcium VLA batteries usually do not need equalizing as often as lead–antimony if
operated at higher float voltages (e.g., 2.20 to 2.25 V/cell for a nominal 1.215 specific
gravity).
Equalizing normally is not used with VRLA batteries because their design and higher
float voltages prevent cell voltage drift, although, if the VRLA cells do drift due to inap-
propriate float operation, the manufacturer may recommend an equalize charge.
The actual float and equalize voltages vary with the cell type and, particularly, with
electrolyte specific gravity. Table 4.1 shows typical values that apply to most battery
types. Manufacturer’s instructions always should be consulted to confirm the values to be
used in an operating system. Typical equalize time periods for VLA batteries are shown
in Table 4.2. Generally, a battery is equalized until successive measurements of specific
gravity or cell voltage show no change over a period of several hours. Load equipment
may limit the equalization voltages used. For example, some older equipment has a maxi-
mum operating voltage of 56.0 V so the maximum equalize voltage would be 2.33 V/cell,
assuming no voltage drop between the battery and the load equipment.
4.2.3 Battery Discharge

The discharge curve for lead–acid cells is fairly flat (Fig. 4.13). The discharge period is
the time required for the cell voltage to reach its final value (final cell voltage or end-of-
discharge voltage). When the cell transitions from float to discharge, there is an immedi-
8
Other terms sometimes used in place of equalize charge are freshening charge and check charge.
Table 4.1 Float and Equalize Voltages

Cell Voltage 12-Cell System 24-Cell System
Type (V/cell) (V/cell) (V/cell)
Float
VLA (lead–antimony)—1.215 SG 2.15–2.17 25.80–26.04 51.60–52.08
VLA (lead–calcium)—1.215 SG 2.17–2.25 26.04–27.00 52.08–54.00
VRLA—1.300 SG 2.25–2.27 27.00–27.24 54.00–54.48
Equalize
VRLAa—1.300 SG 2.30–2.35 27.60–28.20 55.20–56.40
a
VRLA batteries should not be equalized or operated at elevated voltages unless recommended by the manufac-
turer.
ate voltage drop due mostly to the cell internal resistance. As the cell discharges, the volt-
age continues to decrease due mostly to polarization. Hydrogen bubbles that form on the
positive plate cause polarization. This has two effects—the bubbles cause a substantial in-
crease in the resistance at the contact surface between the plate and the electrolyte, and
the hydrogen reacts chemically with the plate in such a way that it sets up a voltage that
opposes the cell voltage.
The cell voltage on discharge (under load) is lower than the open circuit voltage at the
same charge levels. Similarly, the voltage required to charge a cell is higher than the
open-circuit voltage. Figure 4.14 shows how cell voltage and specific gravity varies over
typical discharge–charge cycle.
Five parameters describe the discharge characteristics of stationary batteries: (1) capac-
ity in ampere-hours at (2) a specified discharge rate in hours to (3) a final cell voltage in
volts/cell (also called end-of-discharge voltage or cut-off voltage), (4) battery temperature,
and (5) the nominal full-charge electrolyte specific gravity. For telecommunications bat-
teries in North America, the nameplate discharge characteristics are specified at the 8-h rate
Table 4.2 Equalizing Voltages and Time Periods

Cell Voltage Time
Type (V/cell) (h)
VLA (lead–antimony)—1.215 SG 2.24 80
2.27 60
2.30 48
2.33 36
2.36 30
2.39 24
VLA (lead–calcium)—1.215 SG 2.24 222

2.27 166
2.30 105
2.33 74
2.36 50
2.39 ^34
Fig. 4.13 Lead–acid cell discharge curve.
to 1.75 V/cell [some VRLA battery manufacturers use a 10-h rate and some small-capaci-
ty (< 100 Ah) VRLA batteries are specified at a 20-h rate]. A fully charged cell with C8 Ah
capacity will discharge to 1.75 V/cell in 8 h at 25°C (77°F) if the discharge current I in am-
peres is C/8. Looked at another way, if a battery is discharged to 1.75 V/cell in 8 h and the
current required to do that is I amperes, the cell capacity, C8, in Ah is I A × 8 h.
Manufacturers typically produce a basic cell design and then vary the number of plates
to provide different capacities in a product series. One set of discharge characteristic
curves applies to the series and the curves are then scaled according to the number of pos-
itive plates. Table 4.3 shows the ampere-hour capacity and number of plates for a series
of VLA cells, and Figure 4.15 shows the corresponding discharge curves.
Example 4.6 The nameplate capacity of a battery is 680 Ah. Determine the current re-
quired to discharge this battery to 1.75 V/cell in 8 h at 25°C.
Solution The current is 680 Ah ÷ 8 h = 85 A.
Example 4.7 Determine the capacity of a battery that is discharged in 8 h to 1.75 V/cell
at 25°C when the discharge current is 150 A.
Solution The ampere-hour capacity (at the 8-h rate) is 150 A × 8 h = 1200 Ah.
As shown above, if the cell is discharged at its nameplate rate at 25°C, it will reach 1.75 V
in 8 h (middle curve in Fig. 4.13). The depth of discharge is 100%, which corresponds to
ac Power outage
ac Power returns
Fig. 4.14 Typical cell voltage and specific gravity variations over discharge–charge cycle. In this il-
lustration, a low-voltage disconnect (LVD) relay automatically disconnects the load at 1.75 V/cell. Also
shown (lower graph) is the rectifier current that powers the equipment and recharges the battery. Not
obvious from the upper graph is a slight increase in cell voltage when the load is disconnected.
42.0 V for a 24-cell battery and 21.0 V for a 12-cell battery. If the cell is discharged at a
higher current, it will discharge in less than 8 h to 1.75 V (lower curve in Fig. 4.13) and if
discharged at a lower current, it will discharge in more than 8 h (upper curve in Fig. 4.13).
Under these conditions, the cell actually has less-than-nameplate capacity at the higher
discharge current and more-than-nameplate capacity at the lower discharge current. This
is illustrated in Figure 4.16, which shows the percentage of nameplate capacity for typical
Table 4.3 Example VLA Battery Series—Capacity and Ah/Positive Plate are at 25°C to 1.75 V/cell
and 8-h Discharge Rate
Capacity Number Number Total Ah/Positive
(Ah) Positive Plates Negative Plates Plates Plate
840 5 6 11 168
1008 6 7 13 168
1176 7 8 15 168
1344 8 9 17 168
1680 10 11 21 168
1848 11 12 23 168
2016 12 13 25 168
lead–acid batteries (at 8 h rate to 1.75 V/cell) versus actual discharge rate. Note that the 8-
h discharge rate yields 100% nameplate capacity and that faster discharge rates (< 8 h)
yield less than 100% nameplate capacity.
The actual discharge capacity of a stationary battery also depends on the final cell volt-
age. If the final cell voltage is higher than nameplate (> 1.75 V/cell), the battery is dis-
charged to less than nameplate capacity. In other words, if the battery is not fully dis-
charged to 1.75 V/cell, it still has some unused capacity, and it has not reached 100%
depth of discharge. Similarly, if the final cell voltage is lower than nameplate voltage
Fig. 4.15 Discharge curves corresponding to the example battery series with different end-of-
discharge voltages.
Fig. 4.16 Percentage nameplate capacity vs. actual discharge rate for typical VLA cells.
(< 1.75 V/cell), the battery will have more than nameplate ampere-hour capacity. Howev-
er, to prevent damage from overdischarge, telecommunications batteries should not be
discharged to less than 1.75 V/cell unless allowed by the manufacturer.
Discharging a battery to 1.75 V/cell releases the most capacity, but the battery system
may power equipment that is unable to operate at such a low final voltage. For example,
several end office and transit switching systems used in the current network do not oper-
ate below 44.0 or 44.5 V. Also, there can be an up to 2.0-V drop from the battery termi-
nals to load equipment. To support 44.0 V at the load with a 2.0-V distribution drop, the
final battery voltage must be 46.0 V. This corresponds to 1.92 V/cell. There is no practi-
cal reason to discharge a battery below this voltage if the equipment it powers will not
work. Therefore, for design purposes, the battery final voltage would be 46.0 V and the
average cell final voltage would be 1.92 V.
4.2.4 Discharge Factor

A discharge factor (DF), also called capacity factor, is used to summarize the variations in
cell capacity with varying discharge rate and final cell voltage. Figure 4.17(a) and 4.17(b)
show three-dimensional (3D) plots of the discharge factor versus battery reserve time and
cell final voltage for VLA and VRLA cells. The 3D plot shows that as discharge rate, or
reserve time, decreases below nameplate (8 h), the discharge factor increases above 1.0.
Also, as the final cell voltage increases above nameplate (1.75 V/cell), the discharge fac-
tor increases above 1.0. While the 3D plot shows the overall effects, two-dimensional
plots are easier to use in battery sizing calculations (see Chapter 5, System Design).
4.2.5 Self-Discharge
A cell will discharge even if it is not connected to a load. Self-discharge, also known as
local action, takes place continuously whether the cell is charging, discharging, or stand-
(a)
(b)
Fig. 4.17 (a) VLA discharge factor showing overall effects of reserve time and cell final voltage. (b)
VRLA discharge factor showing overall effects of reserve time and cell final voltage.
ing in an open-circuit condition. Floating the battery at a constant voltage overcomes the
effects of self-discharge. The plate materials, lead and lead dioxide, are thermodynami-
cally unstable in electrolyte and react with it. Oxygen evolves from the reaction of the
electrolyte with the positive plate (lead dioxide), and hydrogen evolves from the reaction
of electrolyte with the negative plate (lead). Self-discharge increases rapidly with increase
in temperature and specific gravity.
If one part of a battery runs warmer than the rest, the warmer cells will have a higher
rate of self-discharge, and their capacity gradually falls below the others. Therefore, bat-
teries must be located so that sunshine or space heaters and air conditioners do not affect a
portion of the battery by introducing differential temperatures between cells in the same
string. Differential temperature is limited to 3°C (5°F).
The rate of self-discharge when on open circuit is quite high for VLA and compara-
tively low for VRLA cells. Self-discharge rate increases with temperature and age and
can be minimized by storing batteries at cooler temperatures between 5 and 15°C (40 and
59°F). All lead–acid batteries in storage require periodic recharging. Typical self-dis-
charge values at 25°C (77°F) are
앫 6 to 7% per month for lead–antimony VLA

앫 3% per month for lead–calcium VLA
앫 1 to 2% per month for VRLA
4.2.6 Overdischarge
A cell will become overdischarged if it drops below its rated final cell voltage. During bat-
tery discharge, weak cells will be exhausted well ahead of normal cells and then will be-
come overdischarged or oversulfated. The latter may cause the plates to buckle and the
grids to crack. Continued overdischarge may reverse the polarity, making positive plates
out of the negatives and negative plates out of positives. Reversal usually destroys the cells.
4.3 TEMPERATURE PERFORMANCE
Telecommunications batteries are optimized for operation at 25°C (77°F). Operation

above 27°C (80°F) is not normally recommended except in special designs because of
greatly shortened life. Operation at lower temperatures generally will increase battery life
but it lowers battery capacity (Fig. 4.18). The temperature factor (TF) in Figure 4.18 is in-
dicative of lead–acid batteries, but there may be considerable variation among manufac-
turers and cell types. It may be used as a first-order estimate to adjust battery nameplate
capacity for operation at lower than optimum temperatures. The temperature factor equals
1.0 for operation at 25°C (77°F). Since batteries should not be operated at higher than
normal temperatures, the temperature factor used in battery sizing calculations never will
be less than 1.0.
Low temperatures may freeze the electrolyte and the resulting expansion may rupture
the battery container or warp the plates. This danger applies not only to operating batter-
ies but batteries in storage. As batteries self-discharge in storage, their specific gravity de-
creases, thus changing the freezing point. Lead–acid batteries in storage require periodic
recharging. The freezing point of electrolyte versus its specific gravity is shown in Figure
4.19.
Fig. 4.18 Percentage battery capacity (temperature factor) vs. temperature. This curve is meant to
show the general relationship; there may be significant differences among manufacturers and cell
types.
Fig. 4.19 Electrolyte freezing point vs. temperature.
175
4.4 BATTERY LIFE
4.4.1 End-of-Life Definition

The positive plates generally are the life-limiting components in modern stationary
lead–acid cells used in telecommunications—as a battery operates, the positive plates cor-
rode.
Telecommunications batteries are considered at the end of their useful life when their
capacity decreases to 80% of nameplate. For example, when the capacity of a 1000-Ah
battery decreases to 800 Ah, the battery is, by definition, at end of life. Conversely, if 800
Ah are required at the end of life, a 1000-Ah battery must be installed initially.9 The 80%
factor is taken into account in battery sizing calculations by the end-of-life (EL) factor,
which is 1.25. Other end-of-life factors may be used if justified.
The initial capacity of new cells may be slightly lower than the nameplate capacity.
The capacity will increase to 100% in a short time after the battery is connected to the
power system and charged and under ideal conditions will stabilize at 100% for most of
the battery’s life. As the battery approaches end of life, the capacity will decrease. Tests
made on a large number of operational VRLA battery systems show that a large number
of cells start to lose significant capacity within a few years of installation.
Cycling (discharging and charging) will decrease the battery life, as determined by
the cycling regimen. The aging mechanisms primarily are changes in the performance of
the positive plate active material and positive plate grid oxidation. As a battery is dis-
charged and then recharged, the lead dioxide active material changes from a highly ac-
tive crystalline form to a less active amorphous structure. Also, the lead metal used in
the positive plate grid oxidizes to lead dioxide, which has lower electrical conductivity.
As the grid conductivity decreases, the cell internal resistance increases and diminishes
the current delivery on discharge and charge acceptance on recharge. Another conse-
quence of grid oxidation is plate growth due to intergranular corrosion of the positive
plate grid. Lead dioxide has about 20% greater volume than lead metal. An obvious sign
of plate growth in VRLA cells is bloating and bulging and in VLA cells is plate warp-
ing.
A battery that is seldom or lightly cycled will last longer than a battery that is continu-
ously and deeply cycled. Most telecommunications applications are subject to long float
periods with occasional cycling. The amount and depth of cycling depends on several fac-
tors:
앫 Length and depth of discharge—the lower a battery is discharged (greater depth of

discharge), the shorter its life.
앫 Amount of recharge before the next discharge cycle—if a battery is not fully
recharged before the next outage that causes a discharge, the shorter its life.
앫 Length of time in the discharged state—the longer a battery remains in the dis-
charged state before it is recharged, the shorter its life.
The life of a modern lead–calcium VLA cell in a central office environment can be
20 years under noncycling float service. A daily discharge of 10% reduces the life by a
9
Over the years, the definition of end of life has changed from 50% to 75% and presently is 80% according to
[4, 5].
4.4 BATTERY LIFE 177
factor of 1–2. A depth of discharge higher than 10% on a daily basis is not recommended
for lead–calcium VLA cells. The life of a lead–antimony VLA cell is about the same as
lead–calcium; however, lead–antimony cells may be discharged to 50% on a daily basis
with some life reduction. VRLA cells generally have a shorter useful life than VLA,
typically 5 to 10 years in a controlled environment and 2 to 5 years in an uncontrolled
environment such as outside plant access nodes where high temperatures are common.
In a cycling environment, VRLA cells have better performance than VLA cells (Fig.
4.20). Tests on the cycle performance of AGM batteries show that there is a long time
period where the capacity is relatively stable before the capacity sharply decreases (Fig.
4.21).
An interesting difference exists between manufacturer’s claims and warranty cover-
age. Marketing materials typically claim the batteries have a 20-year life, but the warran-
ty will provide full replacement value within only 1 or 2 years after new and a linear de-
crease in value over time.
4.4.2 Life-Limiting and Failure Mechanisms

In addition to thermal runaway, which is unique to VRLA batteries, there are many fail-
ure mechanisms in both VLA and VRLA lead–acid batteries (Fig. 4.22).
Fig. 4.20 Cycle-life dependence on depth-depth of discharge for VLA and VRLA batteries. This
chart is based on batteries used in automotive and electric vehicle applications (data from Chapter
11 of [6]), but tests on stationary VRLA AGM batteries show results similar to the VRLA curve shown
(Chapter 13 of [6]).
Fig. 4.21 Capacity loss vs. discharge cycles to 80% depth of discharge for typical VRLA batteries.
Capacity is relatively constant before steadily decreasing.
Fig. 4.22 Cell failure mechanisms.

4.4 BATTERY LIFE 179
All failure mechanisms lead to capacity decrease, which is manifested as reduced life.
Some failure mechanisms in VRLA batteries may lead to catastrophic failures. The most
common VRLA failures are:
앫 Grid corrosion leading to capacity loss

앫 Plate shorting
앫 Plate growth leading to bulged or bloated cell, damaged post seals, and leaky seals
앫 Dry-out due to water loss through vent, water vapor transmission through container
walls, and consumption of water by the positive grid corrosion reactions
The causes of failure in any battery usually are:
앫 Excessive cycling
앫 Post seal leakage (mainly a problem with VRLA but can also cause problems with
VLA)
앫 Loose intercell connections
앫 High or low float voltage
앫 High or low temperature (freezing)
앫 Discharge without recharge
앫 Overdischarge
앫 Cracked and leaking cell container (Fig. 4.23)
4.5 CHARGE ACCEPTANCE AND THERMAL RUNAWAY
Batteries are not 100% efficient and not all of the energy put into a battery by charging is
available during discharging. The charge acceptance, or efficiency, is around 85 to 90%;
for every 100 Ah removed from a battery during discharge, 110 to 115 Ah have to be re-
turned to it during recharge. This efficiency must be taken into account when calculating
the recharge time (see Chapter 5, System Design).
The charge acceptance of a completely discharged battery is initially low, but the bat-
tery accepts current more readily after it is slightly charged. The charge acceptance stays
at a high level until the battery is about 80% recharged. As the battery becomes fully
charged, some of the charge energy is used to generate hydrogen and oxygen and some is
used to make up for internal losses that otherwise would be manifested as self-discharge.
Batteries accept more charge current at higher temperatures than at lower tempera-
tures. This can lead to thermal runaway in VRLA batteries because their compact packag-
ing limits their ability to dissipate the generated internal heat. Figure 4.24 shows that
some cells in a typical VRLA module stacking scheme are surrounded on four sides by
other cells. These “internal” cells generally will run hotter than other cells. As the cells
are charged, internal losses heat the cells increasing their charge current, and the heat pro-
duced by the reaction exceeds the heat removed. Higher charge current leads to higher
temperatures until the cells are damaged, possibly catastrophically (Fig. 4.25). Polypropy-
lene containers used for VRLA cells start to soften at 95°C (203°F) and melt at 155°C
(311°F).
While VRLA cells are more susceptible to thermal runaway when being charged after
a discharge, similar conditions can be brought on by the simple failure of one or more
(a) (b)
Fig. 4.23 Leaking cell at customer premises—the cell cracked due to shipping or installation han-
dling, allowing electrolyte to leak onto the grounded metal battery enclosure. Since electrolyte allows
current flow, the resulting short circuit generated enough heat to burn a hole through the battery jar
(a) about 1 in. across and enclosure (b) about 3–8 in. across like a blowtorch.
Fig. 4.24 Typical VRLA modular stacking scheme. The upper-middle cell cannot dissipate heat as
readily as cells on the edges.
4.5 CHARGE ACCEPTANCE AND THERMAL RUNAWAY 181
Fig. 4.25 Heat damage to a 2-cell VRLA monobloc from thermal runaway—note discoloration in the
area of the letters “Lib.” (Photo courtesy of M.W. Migliaro.)
cells during normal float operation. For example, if 1 cell of a 12-cell string shorts out
while under float, the voltage in the remaining 11 cells must rise to the float voltage. If the
float voltage is 27.0 V (2.25 V/cell for 12 cells), the voltage across the remaining cells is
2.45 V, which clearly is an overcharge condition. If 2 cells fail shorted, the voltage across
the remaining 10 cells is 2.70 V.
The production of hydrogen sulfide (H2S) has been observed during VRLA thermal
runaway. Hydrogen sulfide is a poisonous, colorless gas with an offensive smell similar
to rotten eggs. It can be very damaging to equipment and a serious health hazard to people
because it can paralyze the respiratory system, leading to serious injury or death. Thermal
runaway is not observed in VLA batteries because they are better able to dissipate internal
heat.
Using temperature-compensated charging can reduce the chances of thermal runaway
in VRLA batteries. The compensation is based on a negative temperature coefficient
(NTC) in the range of –2.5 to –4.5 mV/°C/cell (–1 to –3 mV/°F/cell); that is, the battery
voltage is reduced as the temperature increases. If the temperature falls, the battery volt-
age is then increased; however, it is not increased indefinitely as it would eventually dam-
age the battery. Figure 4.26 shows a typical compensation curve. Note the plateau at 10°C
(50°F) in this example where further decreases in battery temperature do not affect the
voltage. Figure 4.27 shows the susceptibility for thermal runaway under float and charg-
ing conditions.
Ideally, temperature probes for temperature-compensated charging would be mounted
inside each cell. However, practical installations use external temperature sensors mount-
ed on cell terminal posts. This provides satisfactory results since lead plate grids are good
thermal conductors (Fig. 4.28). In older installations, a standalone temperature compensa-
tion module may be used, which is mounted on the battery frame. Although the stand-
alone module does not sense post temperature, it still senses battery temperature rise (or,
more accurately, the temperature of the battery frame or module structure). Further details
on temperature-compensated charging, including application information, can be found in
Chapter 5, System Design.
Fig. 4.26 Typical voltage vs. battery temperature for temperature-compensated charging.
Fig. 4.27 Susceptibility to thermal runaway. VRLA batteries are more susceptible to thermal run-
away during charging after a discharge because more heat is generated than during float operation,
as shown. Thermal runaway is unlikely when the battery is operated below the respective line for
each condition.
4.6 CELL IMPEDANCE AND CONDUCTANCE 183
Fig. 4.28 Temperature compensation probe mounted on cell terminal post.
The float current required by a fully charged VLA battery is related to its capacity. A
rule of thumb for lead–calcium is 4 and 10 mA per 100 Ah of capacity at 25°C when float-
ed at 2.17 V/cell and 2.25 V/cell, respectively. For example, the float current for a 2000-Ah
VLA lead–calcium battery at 25°C that is floated at 2.25 V/cell will be approximately 20 ×
10 mA = 0.2 ±0.1 A. The float current for lead–antimony cells is about 5 times greater than
that for a VLA lead–calcium cell when the cells are new and 20 times greater when the cells
are at end of life. Float current increases with temperature and float voltage.
4.6 CELL IMPEDANCE AND CONDUCTANCE
The equivalent circuit of a cell is shown in Figure 4.29(a). This circuit is useful when con-
sidering the various failure mechanisms and to show how the components affect the cell
resistance. For example, the terminals, plate straps (paralleling bars), and internal inter-
connecting welds account for 44% of the total resistance. A somewhat simplified and
more common representation of a cell is shown in Figure 4.29(b). This circuit is useful in
battery and cell testing.
Battery and cell testing generally has two objectives: (1) determining capacity and
(2) predicting cell failure. Much effort has gone into determining capacity from internal
resistance, impedance, or conductance measurements but the technology to do so does
not yet exist. The only reliable and accurate capacity measurement is based on dis-
charging the battery. However, to be done accurately, the battery must be disconnected
from the dc power system and discharged into a load bank. This is difficult and risky in
any system because during testing, which can require 5 to 12 h or more (or 2 to 3 h with
current technology), the power system reserve is compromised. The risk is higher espe-
cially in systems with only one battery string, in which no reserve would be available
during the test period. Capacity testing is not done too often (usually at 5-year intervals,
if ever) because the tests can be very costly and difficult, if not impossible, especially in
remote areas.
Internal resistance, impedance, or conductance measurements can be used to predict
cell failure. Also, cell voltage trends and specific gravity measurements (the latter is ap-
plicable only to VLA cells, although there is a movement in the industry away from spe-
(a)
(b)
Fig. 4.29 (a) Cell equivalent circuit showing the resistance breakdown. (b) Simplified cell equivalent
circuit.
cific gravity measurements) may be used to predict cell failure. For example, a cell may
be headed for failure if its voltage or specific gravity decreases over time or drops quickly
and does not recover by proper charging.
A cell with an internal short usually will float at its open-circuit voltage rather than the
float voltage. For example, the open-circuit voltage for a typical VRLA cell is 2.15 V. If
the system float voltage is 2.27 V/cell (54.48-V system voltage), a cell measurement of
2.15 V probably indicates an internal short. Cell voltage generally, but not always, is a re-
liable indicator for VRLA batteries.
True impedance measurements include both a resistive and reactive component, where-
as conductance is the real part of the admittance (reciprocal of impedance). However, both
types of measurements give the same indication. Additional information on battery main-
tenance and testing is provided in Chapter 6, System Installation and Maintenance.
4.7 SHORT-CIRCUIT CURRENTS
Batteries can supply a large amount of energy to a fault but not indefinitely. The length of
time depends on the discharge rate. For bolted faults close to the battery, the battery will
discharge quickly, but the short-circuit currents may last long enough to damage conduc-
4.7 SHORT-CIRCUIT CURRENTS 185
tors and components or start fires. Telecommunication batteries are designed to not rup-
ture when shorted. It is necessary to consider fault current in battery circuit design, which
is covered in more detail in Chapter 5, System Design.
There are several ways to calculate battery fault current depending on what informa-
tion is available. The fault currents should be obtained directly from the manufacturer if
possible. Alternately, the fault currents can be calculated from the cell open-circuit volt-
age and cell resistance using Ohm’s law:
VOC
ISCC = (4.23)
RCell
where ISCC = short-circuit (fault) current at cell post

VOC = open-circuit voltage, from Eq. (4.20)
RCell = cell resistance including post and plate connection resistances
If the manufacturer provides detailed discharge curves for individual cells or for a gen-
eral design covering various cell capacities, the cell internal resistance may be determined
graphically. The internal cell resistance is the slope of the “initial volts” line in the dis-
charge curve (this is not the same resistance that would be measured with an impedance
tester):
V
RCell = (4.24)
I
where RCell = cell internal resistance ()

V = voltage change (V1 – V2) corresponding to I on initial volts curve (V)
I = current change (I2 – I1) corresponding to V on initial volts curve (A)
If the current scale on the initial volts curve is based on amperes, then the values may be
used directly in Eq. (4.24). However, if the current scale is based on amperes per positive
plate, then the curve ampere values first must be multiplied by the number of positive
plates in the cell. Figure 4.30 shows initial volts versus amperes per positive plate for an
example cell design (this figure should not be assumed to represent all cells).
The total number of positive and negative plates may be even or odd. If even, the num-
ber of positive plates, NP, is exactly one-half the total. For example, if the cell has a total
of 16 plates, as determined from the data sheet, then NP = 8. If the total is odd, the number
of positive plates is determined by first subtracting 1 from the total number of plates and
then dividing by 2 per Eq. (4.1). For example, if the total number of plates is 19, then NP
= 9.
Example 4.8 Determine the fault current available from a 17-plate cell represented by
Figure 4.30.
Solution The number of positive plates in this cell are (17 – 1) ÷ 2 = 8 positive plates.
First, find the slope of the initial volts curve. Choose convenient points on the curve such
as 40 and 80 A per positive. The corresponding initial voltages are 1.90 and 1.78 V, giv-
ing a cell internal resistance per Eq. (4.24) of (1.90 – 1.78 V) ÷ [(80 – 40 A/positive) × 8
positive plates] = 0.000375 . Next, assume the cell has a specific gravity of 1.215 and
calculate the cell open-circuit voltage. From Eq. (4.20) the cell open-circuit voltage =
Fig. 4.30 Example initial volts vs. amperes per positive plate.
2.060 V. Finally, determine the short-circuit current. From Eq. (4.23), the short-circuit
current is (2.060 V/cell) ÷ 0.000375 = 5493 A.
The foregoing calculations considered an individual cell. The data provided by manu-
facturers generally include the resistance effects of one cell interconnect strap. If not, the
fault current available from a battery string could be less than from an individual cell due
the resistance of the cell interconnection straps. The total series resistance of a battery is
RB = RCNC + RI (4.25)
where RB = total battery equivalent resistance including cell interconnection resistance ()
RC = cell internal resistance ()
NC = number of cells in battery (24 cells for 48-V battery and 12 cells for 24-V bat-
tery)
RI = total resistance of interconnection straps () if not included in the battery
manufacturer’s discharge data for the cell
The short-circuit current available from a battery is
VB
ISCB = (4.26)
RB
where ISCB = short-circuit current at battery terminals (A)

VB = battery open-circuit voltage (VOC × Nc) (V)
RB = total battery equivalence resistance including cell interconnection resis-
tances ()
Figure 4.31 shows typical short-circuit current magnitudes available at the terminals of
telecommunications batteries. Variations can be expected for different brands and for
models within a brand.
REFERENCES 187
Fig. 4.31 Battery short-circuit current vs. battery capacity.
4.8 CELL GASSING AND VENTILATION
As previously mentioned, VLA cells emit hydrogen during normal operation as well as
during abnormal operation. Flame arrestor vents are available and should be used on all
VLA cells to prevent a spark outside the cell from igniting the hydrogen in the cell. Bat-
tery manufacturers do not always provide flame arrestor vents unless they are specified
with the battery order. VRLA cells can emit hydrogen during normal and abnormal oper-
ation. When a valve in a VRLA cell fails open, the cell turns into a VLA cell.
Hydrogen concentration in air above 4% by volume is the lower explosive limit. To
prevent such a concentration, battery rooms for both VLA and VRLA batteries always
should be ventilated. Ventilation does not always mean forced-air ventilation, as convec-
tive or natural air movement through inlet and outlet air vents may be sufficient in many
situations, particularly with small batteries. Hydrogen diffuses easily through vents and
door cracks without the help of fans.
REFERENCES
1. D. Linden, and T. Reddy, Handbook of Batteries, 3rd ed., McGraw-Hill, New York, 2002.
2. P. Milner, Float Behavior of the Lead-Acid Battery System, Bell System Technical Journal, Vol.
49, September, 1970.
3. Stationary Battery Installation and Operating Instructions, Gould Electronics & Electrical Prod-
ucts, Document No. GB-3384B, 1991.
4. ANSI/IEEE Std 450™-2002, IEEE Recommended Practice for Maintenance, Testing, and Re-
placement of Vented Lead-Acid Batteries for Stationary Applications, IEEE Standards Associa-
tion, 2002.
5. ANSI/IEEE Std 1188™-1996, IEEE Recommended Practice for Maintenance, Testing, and Re-
placement of Valve-Regulated Lead-Acid (VRLA) Batteries for Stationary Applications, IEEE
Standards Association, 1996.
6. D. Rand, P. Mosely, J. Garche, C. Parker, Eds., Valve-Regulated Lead-Acid Batteries, Elsevier,
San Diego, 2004.
BIBLIOGRAPHY
1. Bark, M., Ed., Electrochemical Power Sources, Institute of Electrical Engineers, New York,
1980.
2. Bode, H., Lead-Acid Batteries, Wiley, New York, 1977.
3. Crompton, T., Battery Reference Book, Butterworth-Heinemann, Jordan Hill, Oxford, 1996.
4. ANSI/T1.330-1997 (R2002), Valve-Regulated Lead-Ccid Batteries Used in Telecommunica-
tions Environment, ATIS, 2002.
CHAPTER 5
SYSTEM DESIGN
PART I DIRECT CURRENT POWER

SYSTEMS COMPONENT DESIGN
This part describes guidelines for collecting and preparing the necessary information and
the procedures for systematically designing each of the basic components in the dc power
system. Although some of the guidelines are very specific, the problem of dc power sys-
tem design is like any other.
앫 There is no single correct method.

앫 All practices and procedures and rules of thumb have exceptions and equally valid
alternatives.
앫 There are no substitutes for common sense and good engineering judgment.
5.1 SYSTEM DESIGN CONSIDERATIONS
A telecommunications dc power system must provide the specified voltage range at

the load equipment and battery when the prime power source is available and for the
specified reserve time when the prime power source or rectifier system fails. The design
of a dc power system to meet this and other requirements includes, among other things,
considerations of battery capacity, component and wire current ratings, voltage drops,
as well as the need to simultaneously recharge the battery and operate the load equip-
ment.
The design process involves preparation and execution. Preparation includes planning,
gathering data, and developing system requirements. Execution includes developing de-
signs based on those requirements.
New systems are the easiest to design because there are no constraints caused by ex-
isting and many times inadequate infrastructure. The new infrastructure is built to ac-
commodate the new systems, and the designer can choose from a wide variety of new

190 SYSTEM DESIGN
equipment. Expansion or retrofit of existing systems, particularly older systems, usually

is more difficult because of the physical and electrical limitations inherent to the equip-
ment already in place. Major parts of dc power systems used in telecommunications,
particularly main buses, are difficult to expand because they cannot be taken out of ser-
vice.
5.1.1 Basic Design Steps and Guidelines

The design of dc power system components follows the basic sequence shown in Figure
5.1.
5.1.2 Planning Interval

Direct current power systems have very long life—many systems are still in service that
are 50 years old or older. Designing systems to last this long requires planning to not only
meet initial requirements but to allow flexibility and expansion over time. Some compo-
nents such as main buses are very difficult to expand while in service and therefore are
designed to meet ultimate load requirements. The planning interval to be used depends on
the application.
Fig. 5.1 Basic design sequence.

5.1 SYSTEM DESIGN CONSIDERATIONS 191
There is little question that long-term planning has never been more difficult than in
the telecommunications industry since 1996.1 This is particularly true of incumbent local
exchange carriers (ILEC)2 subject to competition, which up to that time enjoyed a monop-
oly that allowed them to control and plan the use of spaces in their central offices. While
these ILECs may still control and plan their own technology deployment, they are oblig-
ated under certain conditions to provide space in their central offices for equipment be-
longing to competitive local exchange carriers (CLEC). In doing so, ILECs may not gen-
erally reserve unused central office space for their own exclusive use, making it very
difficult to determine what, if any, space is available for either ILEC or CLEC equipment
on even a short-term basis. The equipment placed in central offices directly affects dc
power systems and, if it is impossible to forecast equipment placement and load growth, it
is equally impossible to forecast dc power system requirements over the long term. How-
ever, there are indirect methods that may be used to estimate the equipment load and these
are described later.
When planning the dc power system, it is necessary to determine the initial and ulti-
mate planning intervals. In a central office environment, the initial interval generally is
the present capacity requirements plus 2 or 3 years’ growth. The ultimate interval normal-
ly accounts for 10 to 20 years’ growth, although certain components, such as the power-
board main bus, may require a prediction well beyond 20 years.
The guideline for central office powerboard planning is to deploy new or additional
units or increased bus capacity when the actual load on the existing powerboard has
reached 50 to 80% of its bus rating. For example, an 800-A powerboard would be ex-
panded when the load exceeds 400 A (50%) but before it reaches 640 A (80%). As a
matter of good practice a powerboard is not loaded more than 80% of its rated capaci-
ty.
Direct current power system components for small telecommunications enclosures,
such as access node remote terminals (less than approximately 700 access lines), usually
are designed and equipped for ultimate capacity. The dc power components in larger re-
mote terminals and equipment enclosures can be designed and equipped on an incremen-
tal growth basis. For example, rectifiers and battery strings may be initially equipped to
handle 2 or 3 years’ incremental growth and then expanded as the load grows beyond the
initial capacities.
Data gathering can be time consuming and tedious, but the activity nevertheless is nec-
essary (Fig. 5.2). Many larger companies have technical groups dedicated to compiling
and processing planning data. From a power system perspective, the results of such ef-
forts are an estimate of the initial load, information on the growth rates throughout the
planning intervals, and an estimate of the ultimate load. Different power system compo-
nents may have different initial and ultimate requirements and planning intervals.
Floor Space Requirements and Availability The floor space required by a dc

power system depends on the types of battery and rectifier technologies, charge/discharge
bus current ratings and physical configurations, primary distribution requirements, and
1
The Telecommunications Act of 1996 (TA96), enacted by the U.S. Congress and signed into law in February
1996, fundamentally affected the way telecommunications systems are planned, operated, and built.
2
ILEC is a contemporary name defined by the Federal Communications Commission (FCC) as a network opera-
tor that provided telecommunications exchange services on February 8, 1996, the day TA96 was signed into
law.
192 SYSTEM DESIGN
Fig. 5.2 Information planning for dc power systems.
growth requirements. Power system components must have adequate space around them
for safe and efficient operation and maintenance. Some considerations are
앫 Complete batteries or individual cells or modules will require eventual replacement

and there must be space for lifting devices to remove the components and place the
new ones.
앫 Electrical codes (where applicable) and safety practices require a safe amount of
workspace around energized equipment so the planner must take into account main-
tenance and installation aisle space.
앫 Existing building spaces may be inadequate in terms of structural strength or avail-
ability and remodeling may be required.
앫 Ventilation and air conditioning systems may be inadequate for batteries and addition-
al power system and network equipment and may require modification or expansion.
Current Drain Studies Historical load growth is very useful in predicting future load
growth in many situations. Modern power systems include monitors that log many para-
meters associated with operation, and external data loggers can be permanently or tem-
porarily installed in power systems that have accessible current shunts or data logging in-
terfaces. Most modern dc power system controllers have integrated data loggers, which
are accessible through a craft interface.
Wire Center Forecasts A typical wire center provides regular services such as
POTS (plain old telephone service) and DSL.3 Historic growth data normally is readily
3
DSL, or digital subscriber line, is a generic name for transmission technologies that provide digital services
over metallic twisted cable pairs or optical fibers.
available in company records. Forecasts are developed by projecting past growth and
comparing it with estimates of future requirements.
New Technologies Deployment New technology equipment generally is more

compact than the equipment it replaces, but the power it consumes per unit volume is
higher. Examples of new technologies are fiber-to-the-home (FTTH) and fiber-to-the-
curb (FTTC) and the associated infrastructures for television and broadband services de-
livery.
Maintenance Requirements At some point in their lives, old systems and equip-
ment become obsolete because they do not meet current requirements, replacement parts
may no longer be available, or software systems are no longer supported and kept up to
date. Examples are rectifiers, voltage conversion devices, and power system controllers
that no longer meet system needs.
Collocation Requirements Most ILECs view collocation as a major impediment to

their financial success; other ILECs and network operators view collocation as a source of
revenue and business collaboration. Collocated equipment requires power, but the ques-
tion of how much power can be very difficult to answer without detailed knowledge of fu-
ture requirements or at least some method to estimate them.
Particular types and brands of equipment will be used in most central offices. For ex-
ample, the major load in central offices owned by local telephone service network opera-
tors will be an end office switching system. Equipment in the access network will be re-
mote switching terminals (RST) or access nodes (digital loop carrier and optical fiber
nodes) served by the end office switch. If the node is for transmission purposes, the major
load may consist of terrestrial microwave radio terminals or fiber-optic terminals. Many
sites will consist of a combination of switching and transmission equipment. In any case,
the loads may be determined from the manufacturer’s installation and engineering guide-
lines.
5.1.3 System Operating Voltages and Ranges

The operating voltages described in this section are specified by industry [1]. The speci-
fied ranges are determined by assuming that in a typical facility some loads may be close
to the power system (low voltage drop in the distribution loop) and some far away (high
voltage drop in the distribution loop). Equipment exists in the network and will be avail-
able in the future that safely operates outside of the ranges given here; however, the dc
power system must be designed to accommodate all equipment in a given facility. The
equipment with the highest minimum voltage and lowest maximum voltage generally will
set the operating range of the power system in a given facility.
The operating range for nominal 24-V systems is specified as 20.0 to 28.3 Vdc. Most
24-V systems have a negative ground (positive operating voltage with respect to ground).
The minimum voltage, 20.0 V, corresponds to an end-of-discharge condition where the fi-
nal cell voltage is 1.75 V/cell (21.0 V final battery voltage) and a maximum of 1.0 V drop
in the distribution loop (from the battery terminals to the load equipment terminals). The
maximum voltage, 28.3 V, corresponds to a 12-cell battery being equalized at 2.35 V/cell
(28.2 V battery voltage), a minimum of 0.0 V drop in the distribution loop, and rectifiers
operating at the high side of +0.5% regulation (0.14 V).
194 SYSTEM DESIGN
Two operating ranges are specified for nominal 48-V systems:
앫 Range 1: 42.75 to 56.7 V

앫 Range 2: 40.0 to 56.7 V
The 48-V systems used in telecommunications have a positive ground (negative op-
erating voltage with respect to ground).4 The minimum voltage in range 1—42.75 V—
is based on design limitations in some existing end office switching systems that use dc
signaling on network-to-user interfaces such as loop-start line circuits (POTS) and
ground-start analog line circuits, loop-reverse battery (LRB), analog trunk circuits, and
E&M analog trunk circuits. This includes Lucent 5ESS and Nortel DMS-10 and DMS-
100, and others. To accommodate 42.75-V minimum equipment voltage and 2.0-V max-
imum distribution loop voltage drop, the final battery voltage must be 44.75 V (1.86
V/cell).
The minimum voltage in range 2—40.0 V—is based on the same thinking as for 24-V
systems but with a maximum distribution loop voltage drop of 2.0 V. Therefore, the min-
imum voltage corresponds to an end-of-discharge condition where the final cell voltage is
1.75 V/cell (42.0 V final battery voltage) and a maximum of 2.0-V drop in the distribution
loop (from the battery terminals to the load equipment terminals). This range does not
consider any limitations, if applicable, in the end office switching system.
The maximum voltage in both ranges 1 and 2 is 56.7 V, and this corresponds to a 24-
cell battery being equalized at 2.35 V/cell (56.4-V battery voltage), a minimum of 0.0-V
drop in the distribution loop, and rectifiers operating at the high side of +0.5% regulation
(0.28 V).
Equipment and systems that operate satisfactorily at lower voltages of 40.0 to 42.0 V
are available, but for any group of interdependent network equipment, the system mini-
mum voltage is determined by the loads requiring the highest minimum voltage. A large
amount of equipment in use today requires 44.0 V minimum. Therefore, if network equip-
ment that operates down to 42.0 V is interdependent on equipment that operates down to
44.0 V, the minimum operating voltage for the group is 44.0 V.
Similar considerations are required at the high end of the range. For example, if
equipment is used in a facility that cannot withstand prolonged operation at a voltage
higher than 56.0 V, then the high limit for the power system must be set to 56.0 V even
though there may be equipment in the same facility that operates satisfactorily at, say,
60.0 V. It should be noted that prolonged operation at higher voltages almost always re-
duces equipment reliability even when the equipment is rated for the higher voltage. It
also should be noted that a considerable amount of equipment used in the network to-
day is designed to operate at voltages outside of the ranges just discussed, either higher
or lower or both.
5.1.4 Load Characteristics

One consideration in dc power system design is how the load current changes as the
voltage decreases on battery discharge. The load characteristics of individual network
components depend on their technology. There are three basic types of load characteris-
tics:
4
See Chapter 1, Introduction, for a brief discussion of why this polarity is bonded to earth ground.
앫 Resistive—current (and power) decreases as voltage decreases [Fig. 5.3(a)]

앫 Constant current—current stays constant (and power decreases) as voltage decreas-
es [Fig. 5.3(b)]
앫 Constant power—current increases (and power stays constant) as voltage decreases
[Fig. 5.3(c)]
Most modern network equipment can be classified as constant power, in which case the
current increases as the battery voltage decreases.
5.1.4.1 Estimating Equipment Loads All telecommunications facilities will

have one or more of the equipment types shown in Table 5.1. This table can serve as a
checklist for new facilities to ensure that the power system design takes into account all
equipment that may be used initially or in the future.
Network equipment load characteristics depend on its basic function and technology.
Digital end office circuit switches have a large fixed load component plus a relatively
small variable component that depends on traffic levels. If the central office facility has a
significant amount of pair gain equipment, such as digital added main line (DAML) and
universal digital loop carrier (UDLC), the load is boosted by the pair gain equipment,
which also is traffic dependent on a one-for-one basis with circuit-switch load (each off-
hook or ringing DAML line corresponds to an off-hook or ringing circuit-switch line).5
Some rules of thumb are
앫 Each on-hook (idle) line circuit draws ~ 0.8 W.

앫 Each off-hook (busy) line circuit draws ~ 2.5 W.
앫 Each ringing line circuit draws ~ 1 to 2 W.
앫 Each off-hook DAML line draws a few watts.
Most central offices include transmission and multiplex equipment and packet switching
and routing equipment. The dc load caused by this type of equipment is constant regard-
less of traffic. The load currents in two typical central offices that consist of digital end
office switching systems are shown in Figures 5.4(a), 5.4(b), and 5.5.
The load characteristics of wireless radio frequency (RF) equipment are quite different
from switching systems. With trunked channelized RF equipment, there is a relatively
small fixed (static) load component and a large variable load component. The variable
component depends on traffic (each active channel contributes to the load), and daily
variations are fairly predictable—midmorning, midafternoon, and evening (Figs. 5.6 and
5.7).
For example, a typical 30-channel AMPS/TDMA cellular site has ~70 A fixed load,
and the load per active RF channel is ~ 5 to 7 A at 27 Vdc.6 Since most modern RF equip-
ment uses constant power switch-mode power supplies, the current increases as the volt-
age decreases when the system battery discharges.
Wireless systems, like all telecommunications systems, experience peak traffic during
disasters such as airline crashes, tornados, hurricanes, earthquakes, and floods. The peak
5
Pair gain equipment is access network equipment that is used to temporarily or permanently provide more than
one POTS line on one metallic twisted cable pair.
6
AMPS stands for Advanced Mobile Phone service and TDMA stands for Time Division Multiple Access. Both
are modulation technologies used in mobile wireless telecommunications services.
196 SYSTEM DESIGN
(a)
(b)
Fig. 5.3 (a) Resistive load characteristics. (b) Constant current load characteristics.
(c)
Fig. 5.3 (c) Constant power load characteristics.
Table 5.1 Network Equipment Types Connected to the dc Power System

Analog loop transmission End office and transit circuit- Mobile switching centers
equipment switching and packet- (MSC), either circuit or
switching systems packet switching
dc–dc converters Fiber-optic terminals Modem servers for dialup
Internet service
Digital circuit multiplication High bit-rate DSL (HDSL) Primary multiplexers (channel
equipment (DCME) span termination equipment banks) and M13 multiplexers
Digital cross-connect Host digital terminals (HDT) Operation support systems
systems (DCS)
Digital subscriber line access Internet and data network Optical line termination (OLT)
multiplexers (DSLAM) equipment such as ATM equipment
switches, routers, and
related servers
Digital subscriber line Inverters Pair gain systems such as
(DSL) loop termination digital added main line
equipment (DAML) and universal
digital loop carrier (UDLC)
Digital microwave radio Local area network (LAN) Remote office test line
terminals routers and related equipment equipment and testboards
such as hubs, switches, and
data controllers
Echo cancellers Mobile and fixed wireless radio T1-carrier span termination
frequency (RF) equipment
(a)
(b)
Fig. 5.4 (a) Central office—local exchange network operator. Measurement of main bus load over a
4-day period (Friday afternoon to Tuesday morning) during July 2001. The primary load is a circuit
switch with approximately 12,000 access lines; other loads include digital microwave radio and fiber-
optic termination equipment, and almost all other items listed in Table 5.1. The switching system pri-
marily serves residential subscribers. (b) Same location as (a) except the scale has been expanded to
show only one day, Saturday. This graph clearly shows that the central office serves mostly residential
subscribers—low load during early morning hours and high load during late evening hours. A central
office that primarily serves business subscribers will show a midmorning and midafternoon peak.
198
dc Load Current vs. Time

dc Load Current (A)
Fig. 5.5 Central office—local exchange network operator. Measurement of main bus load over a 7-
day period in October 2001. The primary load in this central office is a circuit switch with 2000 ac-
cess lines. The switching system serves a mix of business and residential subscribers. Note the daily
load variations of about 10% due to line traffic.
traffic effects may be localized to a particular community, commuting corridor, or central

office or may be spread over a wide geographic region. The design basis for public
telecommunications systems is not peak traffic but instead to achieve acceptable blocking
levels during high-average traffic conditions (e.g., the average of the busy hours during
the 10 highest traffic days during the busy season, 10HDBSBH—10 high day busy season
busy hour). Nevertheless, wireless cell-based base transceiver stations (BTS) and their as-
sociated power systems need to be designed to handle simultaneous calling on all chan-
nels.
Equipment loads can be estimated from manufacturers’ data sheets if they are avail-
able. Unfortunately, in many cases accurate and detailed information is difficult to obtain
without actually buying the equipment. This makes designing a new facility for a new
switching system somewhat difficult unless that same system is used elsewhere by the
network operator and detailed engineering information is available. One public source of
useful load information for a variety of network equipment that may be found in rural
central offices is [2].
Each generation of network equipment uses less power than its predecessor, which
implies that the overall load in a given facility is likely to decrease over time. However,
what really happens is the load increases as the amount of equipment increases with
system growth (this is true of both end office and transit switching systems, which
represent the majority of the load in most central offices). By the time the load equip-
(a)
(b)
Fig. 5.6 (a) RF cell site—wireless network operator. Measurement of main bus load over a 7-day pe-
riod in July 2001 at a site in southcentral Alaska. At that time the site included 29 AMPS and TDMA RF
channels in a ratio of 15 : 6. Note that the fixed load component is around 70 A with a peak around 230
A (at 24 V). Compare with part (c). (b) Same location as (a) except the scale has been expanded to show
only one day during midweek, Thursday.
200
(c)
Fig. 5.6 (c) Same location as (a) except the data were taken about 4 years later in May 2005. At that
time the site included three CDMA RF channels and 29 AMPS and TDMA RF channels.7 Compare with
(a) and note that the base load has significantly increased while the peak load has significantly de-
creased. This probably is due to a larger percentage of users having CDMA handsets than TDMA or
AMPS (CDMA RF carriers are online all the time).
ment reaches maximum capacity, it is obsolete and requires either total or partial re-
placement. This occurs at approximately 8- to 12-year intervals. When the major group
of equipment is replaced, the load decreases significantly, and another load growth cy-
cle starts.
One indirect method that may be used to estimate equipment loads in a facility is to as-
sume that all power delivered to network equipment is dissipated as heat in the equipment
room. While this is completely true of modern transmission systems, a small fraction of
the power delivered to landline end office switching systems is dissipated in the outside
plant and customer premises installation and for wireless systems is dissipated in the an-
tenna and through propagation outside the facility. In all cases, this is ignored in the esti-
mates.
Assuming that the network equipment complies with Telcordia’s Network Equipment
Buildings Systems (NEBS) requirements [3]8, then the heat dissipation can be calculated
7
CDMA stands for Code Division Multiple Access and is a modulation technology for mobile wireless services.
8
NEBS requirements are technical requirements originally prepared by Bell Communications Research (Bell-
core) for its Bell Operating Company (BOC) clients. Bellcore is now known as Telcordia. Because the BOCs
have large market power, many equipment manufacturers build according to NEBS requirements so they can
sell to the BOCs. NEBS requirements are not national standards but are considered de facto standards by a large
segment of the telecommunications industry in the United States.
202 SYSTEM DESIGN
dc Load Current (A)
Fig. 5.7 800 MHz trunked radio (specialized mobile radio—SMR) RF site used for fixed wireless lo-
cal loop service. Measurements show the main bus load over a 7-day period in midwinter 2001. This
site consists of 4 analog FM voice channels operating at 24 V and is located in southwest Alaska.
from those requirements. NEBS requirements specify objective maximum heat dissipa-
tion in terms of area or volume for three configurations:
앫 Entire central office equipment space

앫 Individual equipment frames
앫 Individual equipment shelf or chassis in a frame
For all usable area in the central office equipment space:
앫 Objective maximum for usable central office space—79.9 W/ft2
Example 5.1 The floor plan for a small central office is shown in Figure 5.8. Room 104
is the equipment room. Determine (1) the objective maximum heat dissipation for this
space and (2) the estimated current from a 48-Vdc power system to produce this heat.
Solution The dimensions of room 104 are 24 ft 0 in. long by 19 ft 8 in. wide, and the
area is 472 ft2. Using NEBS objectives, the maximum objective heat dissipation is 79.9
W/ft2 × 472 ft2 = 37,713 W (37.7 kW). At a nominal load voltage of 48 Vdc, the current
required to generate this amount of heat would be approximately 786 A.
It is well known that not all equipment and not all central office spaces are designed to
meet NEBS requirements, but this amperage calculation can be used as a starting point for
system design.
Fig. 5.8 Central office floor plan for Example 5-1.
The NEBS objectives for the maximum heat dissipated by an individual equipment
frame are:
앫 Convection cooled equipment—134.7 W/ft2 maximum heat dissipation

앫 Forced-air fan-cooled equipment—181.2 W/ft2 maximum heat dissipation
Applying these objectives requires knowledge of the effective floor area covered by the
frame. The area includes one-half aisle in front and back (Fig. 5.9). Equipment frames
with higher heat dissipation, whether or not the equipment is fan cooled, may require
wider aisles in order to meet the overall objective of 79.9 W/ft2 for the entire space.
Equipment frames in a telecommunications facility are aligned in rows, or lineups
(Fig. 5.8). The spaces between lineups are aisles. Various names are given to the aisles
depending on whether they are in front or back of the lineup. Since the maintenance on
most equipment is done from the front, the front aisle sometimes is called a maintenance
aisle but more often just front aisle. The rear aisle in many types of installations is nar-
rower because routine maintenance access to it is unnecessary and extra space for workers
to move around is not critical. This aisle sometimes is called the installation aisle but
more often just rear aisle. If equipment cabinets require access only from their front, they
may be mounted back-to-back or close to walls to conserve floor space. In seismic areas,
equipment frames and cabinets that are not fastened together usually are separated by 6 to
12 in. In other cases, both front and rear access is required for installation and routine
maintenance and front and rear aisle widths are the same. Most central offices have a mix-
ture with different aisle widths.
The overall design requirements for equipment frames, cabinets, and panels are de-
scribed in [4], and a Universal Telecom Framework, which encompasses dimensional pa-
204 SYSTEM DESIGN
Fig. 5.9 Equipment frame area for NEBS heat dissipation (frame in center of drawing). This illustra-
tion shows the plan view of equipment frames and aisle spacing.
rameters, performance (such as loading) and application criteria, is described in [5]. For
purposes of this discussion, equipment frames can be open rack structures (relay racks) or
enclosed cabinets. Network equipment frames usually are 12 to 15 in. deep, although the
trend is toward deeper frames. Traditional frame widths are either 20 –83 in. (19-in. frame)
or 24 –83 in. (23-in. frame), but the trend is toward wider frames. Frame heights of 7, 8, 9,
and 111–2 ft have been used, but in many installations 7- or 8-ft frames have replaced the 9-
and 111–2-ft heights.
Transmission and other equipment installed since the late 1980s in many central of-
fices use frames that are 26 in. wide and 7 ft high. Traditional circuit switching system
frames usually are much wider (30 to 32 in., or more) and deeper (15 to 18 in.). As a mat-
ter of engineering policy, many companies use only 23 in. (nominal) frames and do not
use 19 in. frames in new installations, traditional switching systems being the exception.
The width of nominal 23 in. frames for heat calculation purposes usually is assumed to be
26 in. (rather than the actual width).
Some of the latest generation switching systems fit in “standard” 19- or 23-in. frames,
but the equipment shelves are up to 24 in. deep (30 in. of depth with cabling). Data
switching, routing, and server equipment originally designed for enterprise applications
has been installed in central offices even though it was never intended for that environ-
ment. This equipment usually is 20 to 24 in. deep and overall has unwieldy dimensions
and ill-fitting mounting arrangements, which affect how they are mounted and how many
can fit in a frame.
The equipment depths used in NEBS objectives are 12, 18, or 24 in. (if the actual depth
is between sizes it is rounded up; e.g., if the actual depth is 14.5 in., it is rounded up to 18
in.). Recommended aisle widths in the NEBS objectives for typical frame depths are
shown in Table 5.2, but considerable variations exist in practice. The main point is to de-
sign aisle widths that meet both heat dissipation and physical maintenance requirements.
The total floor area of an equipment frame for calculation purposes is
WFront WRear
AFloor = ft2

冢
WFrame DFrame + +
2 2 冣 (5.1)
144
Table 5.2 NEBS Objective Aisle Widths [3]

Frame Depth (in.) Front Aisle (in.) Rear Aisle (in.)
12 30 24
18 54 30
24 42 30
where AFloor = floor Area (ft2)

WFrame = frame width (in.)
DFrame = frame depth (in.)
Wfront = front aisle width (in.)
Wrear = rear aisle width (in.)
To convert dimensions in inches to millimeters, multiply by 25.4 and to convert dimen-

sions in square feet to square meters, multiply by 0.0929.
Example 5.2 Determine the footprint including aisles of a 23-in. (nominal) equipment
frame that is 18 in. deep where the aisles are 36 in. (front) and 30 in. (rear).
Solution For calculation purposes, the frame width is assumed to be 26 in. Using Eq.
(5.1)
26(18 + 36
—
– 30
2 + –
—2)
AFloor = = 9.21 ft2
144
The NEBS objectives for the maximum heat dissipated by an individual equipment
shelf or chassis (Fig. 5.10) are based on a volumetric value determined by the shelf foot-
print (equivalent floor area including one-half aisles) and height.
앫 Convection cooled—20.9 W/ft2/ft of vertical frame space

앫 Forced fan cooled—27.9 W/ft2/ft of vertical frame space
The height includes any required heat deflectors (air ramps), spacers, and fan cooling
trays even if separate from the chassis. For example, if a shelf is 10.5 in. high but requires
Fig. 5.10 Equipment shelf dimensions; the overall footprint includes shelf depth plus one-half aisle
widths (not shown).
206 SYSTEM DESIGN
an open space or deflector of 1.75 in. above it, then the height dimension used in the cal-
culations is 12.25 in. Most network equipment has a vertical dimension in increments of
1.75 in., which often is referred to as one “rack space unit” (RSU) or “rack mounting
unit” (RMU). Therefore, an equipment height of 5.25 in. is equal to 3 rack mounting
units. A 7-ft frame can hold at most 42 rack mounting units (42 × 1.75 in. = 73.5 in. =
6.125 ft) because of structural components at the top and bottom.
The dc load current for frames and shelves may be determined by using the same con-
cepts as for the total central office space.
Example 5.3 Determine the objective maximum heat dissipation for a 7-in. high (4 rack
space units) by 15-in. deep convection cooled equipment shelf mounted in a 23-in. frame
with 36-in. aisle in front and 30-in. aisle in back.
Solution According to general practice, the 15-in. depth is rounded up to 18 in. The total
depth is 18 in. (one-half front aisle) + 18 in. (equipment) + 15 in. (one-half back aisle), or
51 in. = 4.3 ft. The area is the frame width, 26 in. = 2.167 ft, times the total depth of 4.3 ft,
or 9.318 ft2. The equipment height is 7/12 = 0.583 ft, and the maximum objective heat
dissipation is 20.9 W/ft2/ft × 9.318 ft2 × 0.583 ft = 114 W. This wattage corresponds to a
load current of approximately 2.4 A at 48 V input voltage.
Example 5.4 A fan-cooled next-generation switching equipment shelf is 14 rack mount-

ing units (24.5 in. = 2.042 ft) high, 12 in. deep and fits in a 23-in. frame. Three of these
will fit in a 7-ft frame. Each shelf requires two 50-A power feeds, but the actual load nev-
er exceeds 12.5 A (total of both power input buses). (1) Determine if this equipment
meets NEBS objectives when the front and rear aisle widths are 30 in. and 24 in., respec-
tively. (2) Determine how many of these equipment shelves may be mounted in a 7-ft
equipment frame if NEBS objectives for equipment frames are to be met.
Solution (1) The total depth is 15 in. (one-half front aisle) + 12 in. (equipment frame) +
12 in. (one-half rear aisle) = 39 in. = 3.25 ft, and the width is 26 in. = 2.167 ft. The floor
area is 3.25 ft × 2.167 ft = 7.042 ft2. At a nominal input voltage of 48 V, the heat dissipa-
tion is 48 V × 12.5 A = 600 W. The heat per volume is 600 W ÷ (7.042 ft2 × 2.042 ft) =
41.7 W/ft2/ft. This equipment does not meet NEBS equipment shelf objectives of 27.9
W/ft2 (by a factor of about 1.5).
(2) The NEBS objective for an equipment frame with fan-cooled equipment is 181.2
W/ft2. The floor area is 7.042 ft2. Therefore, the maximum allowable heat dissipation is
181.2 W/ft2 × 7.042 ft2 = 1276 W. Since one shelf dissipates 600 W and two shelves
would dissipate 1200 W, two equipment shelves may be mounted in the frame (even
though three physically fit in the frame).
5.1.4.2 Load Conditions Two load conditions are used to size dc power system
components (except the battery system):
Normal Average load current during normal “busy hour” operation. Normal load
current is one of the parameters used to size rectifiers.
Peak Peak load current during worst-case conditions (lowest discharge voltage and
peak load) and includes prediction of future growth estimates. Peak load currents
are used to size the primary distribution circuits and conductors, primary overcur-
5.2 BATTERY SYSTEM DESIGN 207
rent protective devices, and system discharge bus capacity. These components are
difficult to add or expand once the system is operational.
Where most loads are constant power, neither condition may be suitable for battery
size calculations; however, a conservative design may use peak currents for battery sys-
tem design. This may be too conservative in some applications because it assumes the
load current drawn by constant power equipment to be at a maximum level from start to
finish of battery discharge, which is not the case. A less conservative but more accurate
battery system design uses an average voltage at the load during battery discharge.
The average voltage depends on the battery voltage and the voltage drop between the
battery terminals and equipment terminals, which in turn depends on the average load
current and circuit and battery internal resistance. It would be a waste of time to try to de-
termine precise values for the variables because of the many required assumptions. In-
stead, it is sufficient to use typical average values for the load operating voltage: 44 to 48
V for 48-V systems and 22 to 24 V for 24-V systems. The lower values would apply
where most loads are on circuits with maximum voltage drop and the higher values would
apply where most loads are close to the battery system and connected by circuits with low
voltage drop. For example, the average current to a 1300 W load would be 29.6 A for an
average load voltage of 44 V and 27.1 A for an average load voltage of 48 V.
5.1.4.3 Collocation Collocation agreements specify the obligations and conditions

under which the equipment space owner allows others to occupy and use it. The occupa-
tion may be voluntary and desirable as a matter of business or it may be involuntary and
(probably) undesirable as a matter of regulation. An example of the former is a space
provider for many network operators and an example of the latter is an ILEC subject to
competition.
The ILEC collocation agreements normally do not allow batteries in equipment rooms;
however, if the ILEC puts its own batteries in its equipment rooms, it normally cannot pre-
vent CLECs from doing the same thing (ILECs cannot hold CLECs to different standards
than they themselves use). Equipment installed under a virtual collocation agreement nor-
mally is powered from the same powerboards and distribution systems as the ILEC’s own
equipment. Equipment installed under a physical collocation agreement normally is pow-
ered from distribution circuit breaker or fuse panels dedicated to collocation equipment.
The amount of power provided to collocated equipment is in amperage increments
specified in collocation tariffs or agreements. Typical increments are 40 and 15 A. The
feed to collocation equipment normally is redundant. For example, if a collocater requires
a 40-A service, a redundant feed consisting of a 40-A feed from the A-bus and a 40-A
feed from the B-bus is provided. The collocater may be billed on a bulk basis (e.g., for the
40-A feed just described) or in billing units or increments. If the billing unit or increment
is 15 A, then a 40-A service would be billed as three units of 15 A each.
5.2 BATTERY SYSTEM DESIGN
The results of the battery system design are
앫 Choice of technology—vented lead–acid (VLA) or valve-regulated lead–acid

(VRLA)
208 SYSTEM DESIGN
앫 Total battery capacity in ampere-hours (at the nameplate 8-h rate at 25°C to 1.75
V/cell)
앫 Number of battery strings required to provide the total capacity
5.2.1 Choice of Technology

The choice of battery technology includes a number of considerations, such as initial and
operational costs, weight, floor space requirements, anticipated growth, battery life, and
reliability. The choice between the VLA and VRLA batteries can be made based on the
information provided in Chapter 4, Telecommunications Batteries; however, there is one
additional important consideration not described there:
앫 When analyzing the choice between VLA and VRLA, engineering economic analy-
sis should play a part in the decision.
Engineering economic analysis is one of the many analysis tools seldom used anymore in
the telecommunications industry, although it is an important engineering process.9 Nu-
merous books are available that describe engineering economic analysis methods, some
targeted toward telecommunications [6–10]. An example will illustrate a specific applica-
tion to battery technology selection.
Example 5.5 Determine the annual charges, including capital recovery, over a 20-year
period for a 24-cell 1000-Ah VLA battery compared to a 1000-Ah VRLA battery. The
material cost, including delivery to the site, for the VLA battery is $15,000 and for the
VRLA battery is $18,000. Installation costs are estimated to be $4500 for the VLA and
$3000 for the VRLA battery. Do the analysis using a 5% annual cost of money and make
the following assumptions: Labor cost is $80 per hour (including overhead); there are no
operating costs other than maintenance; and there is no inflation. For this example, ignore
all costs associated with floor space (the VLA battery generally will require more floor
space) or remodeling that may be required.
Solution First consider battery life: A reasonable assumption is that VRLA battery life
will be 10 years (i.e., it must be replaced at 10-year intervals) and that one cell will re-
quire replacement 5 years and another 7 years after the new installation. Each cell re-
placement is estimated to cost $2030, which includes 16 labor-hours and replacement cell
cost of $750 (this assumes a warranty fight between the manufacturer and the network op-
erator is not worth the trouble; here, it is assumed the costs of an army of lawyers is much
greater than the cost of a replacement cell). The replacement cost of the VRLA battery is
estimated to be the same as the initial battery. The VLA battery is assumed to last 20
years with no cell replacements before end-of-life. The costs of cell and battery replace-
ments include the costs of disposal.10
Now consider maintenance: Routine maintenance costs for the two battery technolo-
9
Engineering economic analysis is not particularly difficult; however, to do the analysis correctly and as accu-
rately as possible, numerous assumptions about the future are required and these predictions are difficult to
make.
10
Battery manufacturers generally do not charge for disposal, but shipping from the site to the disposal depot is
the user’s responsibility.
gies will be different. The VLA battery requires routine inspections, cell resistance, volt-
age and specific gravity measurements, and water replenishment. The VRLA battery re-
quires routine inspections and cell resistance and voltage measurements. Chapter 6 has a
more complete list of maintenance tasks, but for this example it is assumed that, including
travel time to the site, annual VLA battery maintenance requires 40 labor-hours ($3200
per year) and annual VRLA battery maintenance requires 20 h ($1600 per year).
The above assumptions are summarized in terms of investments and maintenance ex-
penses for each year of the study period (Table 5.3). This table also shows the present val-
ue factor, component present value dollar amount, and total present value for each battery
technology at the 5% interest rate. Under the conditions stated and with the assumptions
made, the sum of the present values of investments and expenses are essentially equiva-
lent over a 20-year study period and 5% interest rate (a coincidence).
Table 5.3 Capital Investment, Expense, and Present Value (PV) Summary for Example 5-5
VLA VLA
VLA Investment Present VRLA Investment Present
Year and Expense Values and Expense Values P/F
0 $15,000+ installation $19,500 $18,000+ installation $21,000 1.000
@ $4,500 @ $3,000
1 $3,200 $3,046 $1,600 $1,523 0.952
2 $3,200 $2,902 $1,600 $1,451 0.907
3 $3,200 $2,765 $1,600 $1,383 0.864
4 $3,200 $2,634 $1,600 $1,317 0.823
5 $3,200 $2,509 $1,600 + cell replacement $2,846 0.784
@ $2,030
6 $3,200 $2,387 $1,600 $1,194 0.746
@ $2,030
8 $3,200 $2,166 $1,600 $1,083 0.677
9 $3,200 $2,064 $1,600 $1,032 0.645
10 $3,200 $1,965 $1,600 + battery replacement $13,876 0.614
@ $18,000 + $3,000
11 $3,200 $1,872 $1,600 $936 0.585
12 $3,200 $1,782 $1,600 $891 0.557
13 $3,200 $1,696 $1,600 $848 0.530
14 $3,200 $1,616 $1,600 $808 0.505
@ $2,030
16 $3,200 $1,466 $1,600 $733 0.458
@ $2,030
18 $3,200 $1,331 $1,600 $666 0.416
19 $3,200 $1,267 $1,600 $634 0.396
20 $3,200 + battery replacement $8,558 $1,600 + battery replacement $8,520 0.377
@ $15,000 + $4,500 @ $18,000 + $3,000
Total present value $66,735 $66,651

210 SYSTEM DESIGN
Each cost component is converted to an equivalent present value using the present val-
ue factor (also called single payment–present value):
P 1
= i n (5.2)
F
冢
1 +
100 冣
where P/F = present value factor
i = annual interest rate (%)
n = time period (year)
Discussion As expected, future payouts are deemphasized by the time value of money.
Even when the time value of money is not taken into account, both scenarios cost about
the same. Of course, any change in the assumptions, particularly those associated with
maintenance expenses throughout the battery system life, may cause one battery technol-
ogy to be considerably less expensive than another. For example, at 5% cost of money, a
20% decrease in the expected VLA annual maintenance expenses (from $3200 per year to
$2560 per year) over the 20-year period will yield an $8000 (rounded) reduction in the to-
tal present value, and a 20% increase in the expected VRLA annual maintenance expens-
es (from $1600 per year to $1920) will yield a $3700 (rounded) increase in the total pre-
sent value. The differences are more dramatic at lower costs of money and less dramatic
at higher costs of money.
5.2.2 Battery Capacity

The factors used to calculate battery capacity are
앫 Equipment load current

앫 Battery reserve time
앫 Discharge factor (capacity factor)
앫 Temperature factor
앫 End-of-life factor (aging factor)
앫 Design margin (uncertainty factor)
These are summarized in11
AH8hour = IEQtBRFDischarge FT FEL FMargin (5.3)
where AH8hour = capacity referred to the nameplate discharge rate of 8 h at 25°C to a final
cell voltage of 1.75 V/cell (Ah)
IEQ = equipment load current (A)
tBR = battery reserve time (h)
FDischarge = discharge factor (no units)
FT = temperature factor (no units)
11
The letters AH are used in this book rather than C, which is found in some literature, to avoid confusion with
the C used in capacitance equations.
FEL = end-of-life factor (or aging factor, no units)

FMargin = design margin (or uncertainty factor, no units)
5.2.2.1 Battery Reserve Time There are two general guidelines for battery re-
serve time depending on whether or not the facility is equipped with a permanent standby
generator:12
앫 Standby generator available: 3 to 5 h reserve time

앫 Standby generator not available: 8 to 12 h reserve time
The telecommunications industry’s baseline reserve time requirement for a facility that
has a permanent standby generator is 3 h [11]. Operational experience may indicate a
longer reserve time is needed. Also, in rural areas or where fuel or repair resources are not
readily available, a longer time such as 5 hours is more appropriate for a baseline.
Where a standby generator is not permanently installed, the baseline reserve time is
8 h [11], but a longer reserve time may be needed based on operational experience, the
required driving time to the site with a portable generator and other operational condi-
tions. For example, power outages may last for weeks after severe storms. Reserve
times measured in weeks are seldom provided at any site; however, a reserve time
longer than 12 h may be appropriate if portable generators have to be rotated through an
area. The selection of reserve time should be based on economic analysis and opera-
tional conditions.
The above guidelines are very general in nature. Whatever the specific situation, the bat-
tery system is to be designed to bridge prime power source interruptions. An “order of mag-
nitude” value for electric utility service failures is two per year and for average duration
(time per failure) is one to two hours for facilities with a single electrical service entrance.
Some telecommunications facilities have two separate service entrances (and separate dis-
tribution circuit feeds from the electric utility) to increase overall service reliability. Where
multiple circuit supplies have all circuit breakers closed (no manual or automatic switching
required on failure of one of the circuits), the failure rate is about 17% and the average du-
ration is about 40% of single circuit services. Interruption rates for multiple utility circuit
supplies with manual or automatic switching schemes are comparable to single circuit sup-
plies, but switching schemes provide smaller average duration, on the order of 15 min.
Failure rates are highest for utility supplies operated at distribution voltages and lowest for
supplies operated at subtransmission and transmission voltages (> 35 kV).13 Only very
large facilities are fed at subtransmission and transmission voltages.
The frequency and duration of commercial and industrial electric supply interruptions
depend on numerous factors. For example, electrical services fed by overhead lines in
heavily forested areas will experience more and longer faults from trees falling onto the
lines—either deadfalls or windfalls or due to snow load. During the 1970s and early
1980s, electric utilities installed a large amount of underground residential distribution
(URD) cable that had high failure rates due to premature insulation breakdown. If such a
line feeds the telecommunication facility, there will be numerous extended interruptions
until the cables are replaced. Some areas of the country have a high incidence of tornados,
and some coastal areas are subject to severe seasonal storms. Some parts of the country
12
State public service commission regulations may specify reserve times outside of the ranges given.
13
These values are adapted from [12], Table 40, for electric utility supplies to industrial plants.
212 SYSTEM DESIGN
may have inadequate electricity grids with a high risk of extended service outages, while
other areas may have just the opposite.
Although the average interruption lasts 1 to 2 h, battery system design has to encom-
pass more than an average interruption. If the site has a standby power source, such as an
engine–generator set, the battery must supply power only between the time the prime
power source is interrupted and the time the standby power system switches online.
Standby systems normally are setup to ignore momentary interruptions and to start and
switch online after some time delay—typically 2 or 3 min or possibly longer (30 to 60
min, or more). A site with frequent short interruptions, say, < 10 min, would be set to
wait, say, 20 min. A typical preheated diesel engine–generator set requires 15 to 45 s to
start and switch online, so where there is a functional standby system the battery may
have to operate only for time durations of a few minutes to an hour. However, standby
power systems do not always start and run reliably and they eventually must be refu-
eled.14 In some installations, the standby generator set is the least reliable part of the sys-
tem (usually because it is not maintained to the same high level as the telecommunica-
tions equipment).
The following example will illustrate some of the factors to be considered when decid-
ing battery reserve time.
Example 5.6 Recommend the battery reserve time for the following conditions: A
telecommunications facility is located in a suburban area approximately a 4-h drive from
the network operator’s dispatch center. It is the network operator’s policy to not dispatch
a technician to a site unless a commercial power outage exceeds 2 h (all sites have alarm
systems that detect power outages). Power reliability is considered “okay” because “there
have not been many outages” over the years. The outages that have occurred lasted be-
tween 30 min and 12 h, although one outage about 4 years ago lasted 3 days. The building
does not have a standby generator, but it is equipped with a manual transfer switch and an
ac receptacle for temporary connection of a portable engine–generator set.
Solution Since this site does not have a standby generator, the minimum battery reserve
time is 8 h. However, if an outage exceeds 2 h, it will take at least 4 h for a technician to
reach the site from the dispatch center. To that time (a total of 6 h) needs to be added the
time it takes to locate the technician on call-out duty and for the technician to get to the
dispatch center, find a functioning portable generator, and hitch it to the truck. Once at the
site, additional time is required to unhitch the generator, set it up, and put it online. This
can easily add 2 h to the time required. In this case, the total elapsed time is 8 h and the
entire baseline reserve time is used up. It is possible the generator needs to be fueled be-
cause the last time it was used the dispatch center forgot to refuel it. It also is possible the
generator has a history of hard starting and difficulty being placed online. Any number of
other things could go wrong and they probably will. With these additional considerations,
it would be wise to add at least another hour or two, giving a total recommended reserve
time of at least 9 or 10 h for this site. Only a fool-hardy engineer assumes everything will
go right every time.
5.2.2.2 Discharge Factor The next parameter to be determined is discharge factor,

FDischarge. This factor takes into account the actual discharge rate if different than the
nameplate rate of 8 h and cell final voltage if different than the nameplate final voltage of
14
See [12] for data.
1.75 V/cell. The first part requires determination of the battery final voltage (VBF), which
is the minimum voltage to which the battery can discharge and still maintain equipment
operation. The battery final voltage must be greater than the minimum equipment operat-
ing voltage (VME) by an amount equal to the voltage drop from the battery terminals to the
load equipment terminals and includes all circuit conductors and connections in between
(Fig. 5.11). The maximum voltage drop, suggested in [1], is
앫 1.0 V for 24-V systems

앫 2.0 V for 48-V systems
The circuit from the battery terminals to the load equipment terminals includes the battery
circuit, primary distribution circuit (including discharge bus), and one or more secondary
distribution circuits. The total voltage drop is
VTotal = VBattery + VPrimary + VSecondary (5.4)
where VTotal 1.0 V for 24-V systems and 2.0 V for 48-V systems
VBattery = voltage drop in battery circuit (V)
VPrimary = voltage drop in primary distribution circuit, including discharge bus (V)
VSecondary = voltage drop in secondary distribution circuits (V)
and the battery final voltage is
VBF = VTotal + VME = VBattery + VPrimary + VSecondary + VME (5.5)
where VBF = battery final voltage (V)

VME = minimum operating voltage of load equipment (V)
The final cell voltage (VCF) equals the battery final voltage (VBF) divided by the number
of cells (NCell) in the battery (12 cells for 24-V systems and 24 cells for 48-V systems), or
VBF
VCF = (5.6)
NCell
Example 5.7 Determine the battery final voltage and cell final voltage for the circuit
shown in Fig. 5.12.
Fig. 5.11 One-line diagram showing the total voltage drop from battery terminals to load equipment
terminals.
214 SYSTEM DESIGN
Fig. 5.12 Example 5-7 circuit diagram.
Solution The total voltage drop in this example is
VTotal = VBattery + VPrimary + VSecondary = 0.5 + 0.5 + 0.5 = 1.5 V
The battery final voltage is
VBF = VTotal + VME = 1.5 + 44.0 = 45.5 V
and the cell final voltage is
VBF 44.5
VCF = = = 1.90 V/cell
NCell 24
Once the final cell voltage has been determined, the discharge factor can be read from
charts depending on the technology [Figs. 5.13(a) and 5.13(b)]. These charts are based on
typical VLA and VRLA battery systems and can be used for most system designs. Where
higher precision is needed, discharge tables or curves should be obtained from the battery
system manufacturer and used to develop the discharge factors.
Example 5.8 Determine the discharge factors for VLA and VRLA batteries when the
cell final voltage is 1.90 V/cell and the discharge time is 6 and 12 h.
Solution Enter the horizontal axis at 6 and 12 h and move vertically to intersect the de-
sired cell final voltage curve and then left to the vertical axis to read the discharge factor.
The discharge factors for VLA technology from Fig. 5.13(a) are approximately 1.45 for 6
h and 1.25 for 12 h. For VRLA, the discharge factors from Fig. 5.13(b) are approximately
1.25 for 6 h and 1.05 for 12 h.
Example 5.9 The required reserve time for a VRLA battery is 5 h to a final cell voltage
of 1.92 V/cell. Determine the discharge factor.
(a)
(b)
Fig. 5.13 (a) Discharge factor for VLA batteries (typical); the solid horizontal and vertical lines repre-
sent the 8-h nameplate discharge rate at 25°C. (b) Discharge factor for VRLA batteries (typical); the
solid horizontal and vertical lines represent the 8-h nameplate discharge rate at 25°C.
216 SYSTEM DESIGN
Solution Entering Figure 5.13(b) at 5 h on the horizontal axis and move up vertically. It
is necessary to interpolate between the 1.90- and 1.94-V final cell voltage curves, then
move left across to the vertical axis where the discharge factor of 1.35 is read. This can be
interpreted to mean the battery will require 35% more capacity because of the higher than
nameplate discharge rate and final cell voltage.
The two discharge factor curves can be approximated in terms of two independent
variables (battery reserve time, tBR, and cell final voltage, VCF) by the following expres-
sion. This expression is based on regression curve fitting of data from several manufactur-
ers and is sufficiently accurate for initial designs; expect variations when comparing to
specific battery models.15
2
b d VCF g V CF VCF
FDischarge = a + + cVCF + 2
2
+ eV CF +f + 3
3
+ hV CF + i + j (5.7)
tBR t BR tBR t BR tBR t 2BR
where the coefficients are
Coefficient VLA VRLA

a –1105.7643816 –295.4606616
b 201.7986178 79.2665474
c 1822.6791610 491.4691826
d –9.1965015 –3.0649118
e –1000.9629849 –271.8618864
f –224.5513525 –88.2716728
g 1.3124170 0.7333677
h 183.2697637 50.1408711
i 63.3158840 25.2446729
j 3.8187552 0.7855183
5.2.2.3 End-of-Life Factor (Aging Factor) The end-of-life factor for telecommu-
nications batteries is, by definition, 1.25, since a battery is considered at end of life when
its capacity has decreased to 80% of its nameplate ampere-hour rating. For example, if
calculations reveal that the required battery capacity is 100 Ah at end of life, then a bat-
tery capacity of (100 Ah × 1.25 =) 125 Ah is required to be installed today.
5.2.2.4 Temperature Factor Under normal circumstances, the temperature factor

always is 1.0, where a factor of 1.0 indicates battery operation at the nameplate temper-
ature of 25°C (77°F). Telecommunications batteries are not intentionally operated above
nameplate temperature because their life is reduced even though they may have slightly
more capacity (a temperature factor < 1.0). On the other hand, there are numerous in-
stances where batteries may be intentionally operated at lower temperatures. For example,
in remote mountaintop sites, the costs of heating the equipment enclosure to normal room
temperatures may be prohibitively high so the equipment and batteries are operated at a
temperature less than 25°C. In this case, for a given capacity requirement, a larger battery
would be required.
The performance of telecommunications batteries at lower-than-normal temperatures
depends on the brand, model, and technology, and only a general relationship between
temperature and capacity is provided here (Fig. 5.14).
15
Curve-fitting program used to develop this expression is DataFit (V8.1) by Oakdale Engineering of Oakdale,
PA (http://www.oakdaleengr.com/).
Fig. 5.14 Temperature factor. This chart shows the general relationship for typical batteries; there
may be considerable variations across brands and models. To use this graph, enter it along the hori-
zontal scale at the desired operating temperature, move up vertically to the intersection with the
graph, move left horizontally to the vertical scale, and read the temperature factor. For example, the
temperature factor at a temperature of 15°C (59°F) is approximately 1.12.
5.2.2.5 Design Margin (Uncertainty Factor) All engineering problems involve a

design margin that reflects the engineer’s imperfect knowledge of the way real systems
operate. In the case of battery system design, the design margin is 1.0 and typically
1.10 to 1.15.
5.2.2.6 Capacity Calculations Battery capacity calculations are illustrated in the

following example.
Example 5.10 Determine the required capacity for VLA and VRLA batteries based on
the following parameters and a minimum equipment voltage of 40.0 V. Repeat for a min-
imum equipment voltage of 43.0 V and explain why there is a difference, if any:
IEQ = 100 A
tBR = 12 h
FT = 1.0
FEL = 1.25
FMargin = 1.1
VTotal = 2.0 V
218 SYSTEM DESIGN
Solution With a minimum equipment voltage of 40.0 V and total voltage drop of 2.0 V,
the final battery voltage is 42.0 V or 1.75 V/cell. From the discharge factor curves for
1.75 V/cell and 12 h reserve time, the discharge factors are approximately 0.90 for both
VLA and VRLA batteries (by coincidence). Therefore, from Eq. (5.3)
AH8hour = IEQtBRFDischarge FT FEL FMargin
AH8hour = 100 A × 12 h × 0.90 × 1.0 × 1.25 × 1.1 = 1485 Ah (VLA)
AH8hour = 100 A × 12 h × 0.90 × 1.0 × 1.25 × 1.1 = 1485 Ah (VRLA)
If the minimum equipment voltage is increased from 40.0 to 43.0 V, the final battery volt-
age also is increased from 42.0 to 45.0 V, or 1.88 V/cell. From the discharge factor
curves, the discharge factors for 1.88 V/cell and 12 h reserve time are 1.15 for VLA and
1.05 for VRLA batteries.
AH8hour = 100 A × 12 h × 1.15 × 1.0 × 1.25 × 1.1 = 1898 Ah (VLA)
AH8hour = 100 A × 12 h × 1.05 × 1.0 × 1.25 × 1.1 = 1733 Ah (VRLA)
In the first part of this example, the battery is allowed to fully discharge to 1.75 V/cell, but
in the second part it is allowed to only partially discharge before the required battery final
voltage is reached. The effect of this partial discharge is to increase the required battery
capacity by about 28% (VLA) and 17% (VRLA) to meet the same reserve time require-
ments. In this example, the choice of battery technology has a nominal effect on the cal-
culated capacity.
5.2.3 Number of Strings

After calculating the total battery capacity, it is necessary to determine the number of bat-
tery strings. As a general rule, the number of battery strings should be the minimum nec-
essary. Accurate voltage support of each cell during float operation is more difficult when
more than one string is installed and operated in parallel. Also, achieving equal charge
and discharge characteristics from each parallel string is notoriously difficult.
The battery capacity may be dictated by the anticipated growth pattern and the need to
simplify expansion. It usually is easier to design and install additional strings of the same
battery type and capacity as those already in place. It is not necessary that all battery
strings have the same capacity, but the ratio of the capacity (in ampere-hours) of one
string to another should not be greater than 2 : 1. Because of different float voltage, and
charge and discharge characteristics, different battery technologies (VLA and VRLA)
should not be connected in parallel except in temporary special circumstances. For exam-
ple, when converting a facility from lead–calcium VLA to VRLA, the voltage of the VLA
may be raised to equal the required VRLA float voltage, the VRLA string connected in
parallel with the VLA string, and then the VLA disconnected.
The growth in battery capacity is related to the growth in rectifier capacity because
both increase as load current increases. The battery capacity increment is approximately
IRMtBR
AH = (5.8)
FR
where AH = incremental battery capacity (Ah)

IRM = rectifier module rating (A)
tBR = battery reserve time (h)
FR = recharge factor
The rectifier module rating and recharge factor is explained in Section 5.3.
If the total required battery system capacity is to be divided equally across two or more
strings, the capacity of each string is determined from
AH8hour
AHString = (5.9)
NString
where AHString = individual battery string capacity (Ah)

NString = number of battery strings in parallel
Where service availability is critical, and some level of battery redundancy is needed, it is
good practice to install two or possibly three battery strings. Multiple strings allow easier
battery maintenance without jeopardizing service to the load equipment. The battery re-
dundancy can be calculated from
FRedundancyAH8hour
AHString = (5.10)
NString
At least three scenarios may be considered when redundancy is to be provided. In the first
scenario, two battery strings are used, each with 50% (1–2) the required total capacity. In
this case, FRedundancy = 1.00 and NString = 2. This method costs a little more and requires
more floor space than one string of 100% capacity, but it ensures service continuity be-
tween the times a cell fails and it is replaced. It does not provide full reserve time when
one string is out of service (it provides approximately one half the reserve time during
that period).
The second scenario provides two battery strings, each with 67% (2–3) the total required
capacity. In this case, FRedundancy = 1.33 and NString = 2. This scenario has higher cost than
the first scenario but provides a longer reserve time (approximately two thirds the reserve
time) while one string is out of service. The third scenario provides two battery strings,
each with 100% of the total required capacity. In this case, FRedundancy = 2.00 and NString =
2. This scenario is twice the cost (in terms of battery material and floor space) of the first
scenario but provides full reserve time while one string is out of service. Of course, other
scenarios with other FRedundancy and NString are possible, such as providing three strings,
each with 50% of the total required capacity (FRedundancy = 1.50 and NString = 3). In this
case, the loss of any one string does not affect the design reserve time. The cost is approx-
imately 50% higher than the first scenario.
There is no technical limit on the number of battery strings that may be connected in
parallel but there are practical limits. As a general rule, the battery manufacturer should
be contacted if more than five strings are to be paralleled. Some manufacturers do not rec-
ommend more than four or five parallel strings.
Short circuit-currents and voltage drop usually dictate practical parallel string limits.
Battery short-circuit currents from parallel strings may exceed the interrupt capacity of
overcurrent devices or the mechanical strength of rigid busbar or cable bus mounting struc-
tures.
(a)
(b)
Fig. 5.15 Battery rack structures: (a) VRLA batteries—integral frame and rack components. Two of
the same model batteries with different style insulating covers. (b) VRLA—24 modules in trays
mounted in a box frame or cabinet assembly. The leads between tiers are flexible interconnect straps
that connect the cells in each tray.
220
(c)
(d)
(e)
Fig. 5.15 (c) VLA—seismic battery rack. The cross braces and rails surrounding the cells indicate
this battery rack is for seismic applications. (Photo courtesy of Schultz Brothers Electric Company.)
(d) VLA—nonseismic battery rack. Note there are no rails around the cells. (Photo courtesy of M.W.
Migliaro.) (e)—Stepped VLA rack. This photo shows a two-step installation that is difficult to main-
tain. The rack has no maintenance aisle in the back, and the cells on the upper step at the back are
difficult to access for routine maintenance and for removal and replacement. (Photo courtesy of
M.W. Migliaro.)
222 SYSTEM DESIGN
5.2.4 Battery Racks

Above approximately 150 Ah, VRLA batteries have an integral rack structure and can be
stacked and bolted together [Fig. 5.15(a)] or placed in a specially made cabinet [Fig.
5.15(b)]; however, stacking heights may be limited and manufacturers should be consulted
for high-risk seismic installations.16 Below 150 Ah, the battery or cells are placed in a tray
mounted in a regular equipment frame. However, in high-risk seismic areas, special equip-
ment frames normally are required because of battery weight. In high-risk seismic areas,
VLA racks are designed for specific battery and cell dimensions [Fig. 5.15(c)]. VLA racks
for low-risk seismic areas have much less structure [Fig. 5.15(d)]. Racks must be designed
and installed so the battery can be maintained, including removal and replacement of indi-
vidual cells. This generally precludes VLA racks with more than two tiers and stepped
racks [Fig. 5.15(e)]. Battery rack bonding and grounding is covered in Section 5.9.3.
5.2.5 Battery System Checklist

The following checklist can be used as a basic specification. Suggested wording is pro-
vided in parentheses, and variables are shown in brackets where:
a = number of module stacks (VRLA only)

b = number of cells in a module (VRLA only)
n = number of cells in a string
x = desired total ampere-hour rating
x = desired individual string ampere-hour rating (for multiple string installations)
y = number of parallel strings
z = battery disconnect switch rating in amperes
1. Application: (Specify system application such as “Float application in

telecommunications central office.”)
2. Battery type: (VLA or VRLA with plate growth provision.)
3. Nominal system voltage: (–48 V or +24 V.)
4. System float voltage: (Specify voltage or voltage range.)
5. Total capacity: No less than [x] Ah total at 8 h rate to 1.75 V/cell at 25°C
6. Battery arrangement: [y] parallel [x] Ah, [n]-cell battery strings, each string
consisting of model (specify model if desired) with [a] stacks
of [b] cells per stack.
6. Cell container and cover: Flame retardant 28% LOI or better.
7. Design life: 20 years under float at 25°C.
8. Safety vent: Flame arrester [if VLA] or self-resealing with flame arrester [if
VRLA].
9. Cell interconnection straps: Lead-plated solid copper with stainless steel hardware.
10. Battery terminals: Lead-plated solid copper drilled for at least two 2-hole lugs, side
mounted with mounting insulators and brackets.
11. Mounting and structure: (Horizontal or vertical) with provisions for individual cell
replacement.
16
Seismic risk areas, or zones, have been defined by four zones, where low-risk seismic zones were 0, 1, and 2
and high-risk seismic zones were 3 and 4. However, such zoning has been replaced in building codes [13] and
national standards [14] by maps showing contours of design ground motion and are more indicative of what ac-
tually happens during an earthquake.
5.3 RECTIFIER SYSTEM DESIGN 223
12. Seismic: Zone (zone 1, 2, 3, or 4) or seismic description.

13. Temperature: (Specify abnormal temperature requirements.)
14. Battery disconnect: Each string with battery-mounted disconnect switch (circuit
breaker) rated at least [aa] A, short-circuit withstand rating [bb]
A, and including all mounting components and hardware.
15. Required accessories: 앫 Insulated interconnection strap covers
앫 Insulated terminal covers
앫 Cell number set (1-24 or 1-12) and polarity markings (+ and –)
앫 Oxidation inhibitor compound
앫 Installation, operation, and maintenance manual
앫 Module lifting strap or lifting device
5.3 RECTIFIER SYSTEM DESIGN
Rectifier system design requires the selection of
앫 Technology
앫 Quantity
앫 Current rating
앫 Input voltage rating
앫 Features and options
Rectifiers used in telecommunications applications are equipped for N + 1 redundancy,

operated in parallel, and almost always setup for load sharing. The rectifier system ca-
pacity must be large enough to not only operate the load equipment but simultaneously
recharge the batteries in the desired time. If maximum efficiency is desired, only enough
rectifier capacity is equipped to achieve this end and no more; otherwise, all rectifiers
will operate in a low-current, inefficient condition. With modular rectifiers, it is rela-
tively simple to adjust the capacity by unplugging unneeded rectifier modules or using
modern controllers that can be programmed to activate only the necessary rectifiers. The
controllers on some modern rectifier systems can be set up to automatically shutdown
unneeded rectifiers during float operation and to bring them back online when needed
for battery recharge.
The rectifiers used for VLA batteries also may be used for VRLA batteries except that
temperature compensation should be added or retrofitted when VRLA batteries replace
VLA batteries. In some cases, the rectifiers may need to be replaced in order to provide
temperature compensation. The rectifiers also must have adequate output filtering to pro-
tect the VRLA cells from ac ripple, which may lead to thermal run-away.
5.3.1 Technology
New systems almost always will use hot-swappable, modular switch-mode rectifier tech-
nology. If frame or floor space is limited, the logical choice is switch-mode technology.
Existing systems that use ferroresonant rectifier technology may remain with that technol-
ogy to ensure compatibility and to minimize spares and maintenance requirements. How-
ever, it may not be possible to retrofit some ferroresonant rectifiers for temperature-com-
pensated charging.
224 SYSTEM DESIGN
5.3.2 Output Configurations

The different dc output circuit configurations from standalone rectifiers and modular rec-
tifiers must be accounted for in the circuit design. Standalone rectifiers have individual
output circuits that are connected by cabling or busbar to the charge bus or a rectifier ter-
mination bar that is part of the charge bus [Fig. 5.16(a)]. Modular rectifiers are installed in
a shelf assembly that buses the outputs to one common termination point on the assembly
[Fig. 5.16(b)]. This termination point is then connected to the charge bus by cabling or
busbar.
(a)
(b)
Fig. 5.16 Rectifier output circuit configurations: (a) standalone rectifier chassis output circuit and (b)
modular rectifier shelf assembly output circuit.
5.3.3 Rectifier Quantity

Even the smallest system will have at least two rectifiers—one to carry the load and
recharge the battery and another configured for hot-standby operation. There is no techni-
cal limit to the number of rectifiers, but a simple economic analysis may reveal that a
small number of large rectifiers is cheaper than a large number of small rectifiers, or the
other way around. The capacity of a 7-ft × 23-in. bay with modular rectifiers is at least
900 A and can be much higher. It should be noted that a single bay with 1400-A capacity
and operating at full load will dissipate around 7.5 kW due to the rectifiers not being
100% efficient. This is a large amount of heat in a small space and must be handled by the
facility cooling system.
5.3.4 Rectifier System Capacity

Industry baseline requirements specify that the rectifier system should be designed to
recharge a fully discharged battery to 95% of its nameplate capacity rating within at least
24 h while simultaneously operating all load equipment [11]. This requirement should be
met when the equipment is operating at its busy hour load (normal condition).
If the rectifier system output current is equal to AH8hour/20 A (where AH8hour is the bat-
tery nameplate capacity in ampere-hour at the 8-h rate and 25°C), the battery will
recharge to about 95% of its rated capacity in 24 h. For example, a 450-Ah battery will
recharge in 24 h if the available charging current is (450 ÷ 20 =) 23 A. This charging cur-
rent is over and above that required to power the load. As battery cells age, they require
longer recharge time for a given charging current.
The time required to recharge a battery to 100% of its nameplate capacity can be cal-
culated from
AH8hour
tRecharge = FBatLoss hours (5.11)
IRS – IEQ
where tRecharge = battery recharge time (h)

AH8hour = battery nameplate capacity at 25°C and 8-h rate to 1.75 V/cell (Ah)
IRS = rectifier system capacity with N rectifiers (A)
IEQ = normal condition equipment load current (A)
FBatLoss = battery loss factor (typically 1.10 to 1.15 for lead–acid batteries)
Equation (5.11) assumes 25°C battery temperature. Battery temperatures below 16°C re-
quire that the calculated recharge time be doubled and battery temperatures below 5°C re-
quire that the recharge times be quadrupled.
Example 5.11 Determine the recharge time for a 660-Ah battery if the rectifier system
capacity is 100 A and the load is 50 A.
Solution From Eq. (5.11)
AH8hour 660 Ah
tRecharge = FBatLoss = × 1.15 = 15 h
IRS – IEQ 100 A – 50 A
226 SYSTEM DESIGN
Typical recharge times are 12, 18, and 24 h. If the desired recharge time is 8 h or less, the
manufacturer should be contacted for their recommended charging regimen.
The total normal condition load current is one parameter used to calculate rectifier sys-
tem capacity. Note that normal condition load current is used here instead of peak (peak
load currents are encountered only on battery discharge when the rectifiers are not operat-
ing and system voltage is decreasing).
Rectifier system capacity can be calculated two ways—(1) using the current required to
recharge the battery in a certain time period or (2) using a recharge factor. The first method
is advocated in [15] and allows more analytical control over battery recharge time.
With the first method, the rectifier system capacity is calculated by solving Eq. (5.11)
for IRS, or
AH8hour
IRS = IEQ + FBatLoss amperes (5.12)
tRecharge
With the second method, the rectifier system capacity is calculated from
IRS = IEQFRecharge amperes (5.13)
where FRecharge = recharge factor

= 1.2–1.4 (typical for VLA)
= 1.15–1.2 (typical for VRLA)
Higher recharge factors allow faster recharge times but increase the possibility of thermal
runaway in VRLA batteries (unless temperature-compensated charging is used).
Where the rectifier system is used at a site that is 3300 ft (1000 m) or more above sea lev-
el without a corresponding decrease in ambient temperature, rectifier output currents usu-
ally are derated to ensure proper cooling. However, if the ambient temperature is decreased
by 2°C for every 1000-ft (300-m) altitude increase, derating may not be required. Rectifiers
may also require derating or adjustment for high ambient temperatures. Typically, at ambi-
ent temperatures above 40 or 50°C the rectifier maximum output current needs to be de-
creased for proper cooling of the electronic components. Derating of rectifier output cur-
rent for ambient temperatures and altitudes may be found in the manufacturer’s data sheets.
Where modular rectifiers are to be used, the total number of modular rectifiers (NRM) is
determined from the rectifier module current rating (IRM) from
IRS
NRM = + 1 (5.14)
IRM
rounded up to the next higher integer value. This assumes that all rectifier modules have
the same rating, which would be true in new installations and replacements. Where addi-
tional rectifier capacity is added to an existing installation, the redundant rectifier (or rec-
tifiers) must be at least as large as the largest rectifier in the system. For example, if an in-
stallation has two 50-A and one 100-A rectifiers for the load and battery recharge,
redundancy would be provided by adding a second 100-A rectifier or two additional 50-A
rectifiers. This protects the system from the failure of any rectifier in the system.
Example 5.12 An 800-Ah battery is used in a system where the equipment load current
is 100 A. Determine the total rectifier system capacity and number of rectifier modules if
each module is rated 50 A.
Solution Assuming tRecharge = 24 h, then
AH8hour 800 Ah
IRS = IEQ + FBatteryLoss = 100 A + × 1.15 = 138 A
tRecharge 24
and
IRS 138 A
NRM = + 1 = + 1 = 3.76 or rounded to 4
IRM 50 A
In this example, the total rectifier capacity is (4 × 50 A =) 200 A, 150 A working plus 50
A standby.
5.3.5 Battery Voltage Sensing

Tight regulation of the battery terminal voltage is important for long battery life. The rec-
tifiers control the voltage accuracy at the battery terminals and use voltage sensing con-
nections for regulation. The sensing leads can be connected at the powerboard bus [Fig.
5.17(a)], directly to the battery terminals [Fig. 5.17(b)] or to a battery circuit collector bar
[Fig. 5.17(c)]. The latter two have the advantage of sensing the voltage at the point where
it is most important to regulate—at or near the battery terminals. However, these methods
have the disadvantage that the small sense lead wires (typically 12 or 14 AWG) usually
are installed in the same cable rack as the much larger battery wires and potentially may
be damaged by the larger wires during installation or removal. One way to work around
this problem is to put the voltage sense leads on L-brackets fastened to the cable rack car-
rying the battery leads.
When the battery voltage sense leads are connected directly to the battery terminals,
they should have overcurrent protection as close as possible to the battery terminals; how-
ever, it is the practice of some companies to omit this protective device. The sense current
is very small, and a small value fuse or circuit breaker (5 to 10 A) may be used but, in any
case, the fuse or circuit breaker rating should not exceed the sense lead conductor current
rating. In an installation with more than one battery string in parallel, the voltage sensing
can be done at the battery collector bar. Another, simpler, way is to sense the battery volt-
age at the powerboard where the rectifiers are installed. During float operation, when tight
voltage regulation is required, the current and voltage drop in the battery circuit is small,
so there is little sacrifice in regulation accuracy.
5.3.6 Rectifier ac Input Circuit Requirements

The rectifier system must be matched to the ac service voltage. If the service voltage is
lower or higher than the rectifier ac input rating, a step-up or step-down transformer is
used to change the service voltage to the proper value. Power is delivered by electric
utilities through either single-phase or three-phase connections. Common service volt-
ages are discussed in Chapter 2, Electricity Review. The preferred voltages for most rec-
tifier installations are 208 and 240 V with line–line connections. Large (400, 800, and
1600 A) ferroresonant rectifiers require 480-V line–line voltage. In large facilities, the
service may be delivered at primary voltages, such as 12,470Y/7200 V, in which case
an on-site step-down transformer converts the primary voltage to normal secondary uti-
lization voltages.
(a)
(b)
(c)
Fig. 5.17 (a) Battery voltage sensing at the powerboard bus. (b) Battery voltage sensing at the bat-
tery terminals. (c) Battery voltage sensing at the battery collector busbar.
228
Each individual rectifier chassis or module requires a dedicated ac input circuit. The
importance of a dedicated circuit cannot be overemphasized. Feeding more than one recti-
fier from only one circuit breaker jeopardizes the system reliability. If that one circuit
breaker opens for any reason, the remaining rectifier capacity may not be sufficient to op-
erate the system properly. The ac input circuits are unlike the dc output circuits from
modular rectifier systems, in which the outputs are bused across all rectifier modules in
the chassis assembly. This arrangement ensures that the failure of any rectifier does not
affect the dc bus (except for loss of rectifier capacity).
Rectifiers may be fed from a single dedicated panelboard or two panelboards for redun-
dancy (Fig. 5.18). Installations with a large quantity of modular rectifiers may require
more than one panelboard because lighting and appliance panelboards are limited to 42
poles and have space only for 21 two-pole circuit breakers (thus a maximum of 21 recti-
fiers with line–line connections may be connected to such a panelboard). The rectifier load
on the ac panelboard should be balanced. If the rectifiers are connected to a single-phase
panelboard, they are connected in such a way that each ac bus is loaded as evenly as possi-
ble. Since there are two buses in a single-phase panelboard, the load can be completely
balanced if there is an even number of rectifiers. In a three-phase panelboard, perfect bal-
ance can be achieved only if the number of rectifiers is evenly divisible by 3. Small recti-
fiers (~500 W) usually are connected line-neutral [Figs. 5.19(a) and 5.19(b)], and larger
rectifiers usually are connected line–line [Figs. 5.19(c) and 5.19(d)]. Rectifier ac circuit
design is discussed in Section 5.8.
Fig. 5.18 Single and redundant ac panelboard configurations for rectifier input circuits. Modular
rectifier shelf assemblies are shown, but the same concept applies to standalone rectifiers. Other ac
wiring configurations are possible.
230 SYSTEM DESIGN
(a)
(b)
Fig. 5.19 The ac panelboard connections for rectifier ac circuits: (a) line-neutral connections in a
single-phase panelboard using one-pole circuit breakers, (b) line-neutral connections in a three-
phase panelboard using one-pole circuit breakers.
120/240 Vac, 1-Phase, 3-Wire
(c)
(d)
Fig. 5.19 (c) line–line connections in a single-phase panelboard using two-pole circuit breakers,
and (d) line–line connections in a three-phase panelboard using two-pole circuit breakers.
232 SYSTEM DESIGN
Rectifiers are not 100% efficient and dissipate heat whenever powered, even when un-
loaded. Cooled air and ventilation are required in rooms with rectifiers to carry away this
heat.
5.3.7 Rectifier dc Output Circuit Requirements

The voltage drop in the rectifier output circuits is not normally a serious design issue.
Therefore, considerable leeway is available in the voltage drop of rectifier output circuits.
They are typically designed for a maximum drop of 1.0 V (48-V systems) or 0.5 V (24-V
systems) at full load from the rectifier output terminals to the charge bus. Low voltage
drop is easy to achieve because of the normally short run.
In addition to voltage drop, the rectifier output circuits must adequately carry the full
output current (including overload current). The output of many modern rectifiers is rated
in watts (or kilowatts) rather than amperes, and the output current increases with decreas-
ing voltage. However, controllers for these rectifiers can be set to operate the rectifier in a
current limit mode. For example, a 1.5-kW rectifier can deliver 55.5 A at 27 V and 75 A
at 20 V but could be set to current limit at 50 A regardless of its output voltage. Rectifier
output circuits always are sized for the maximum output current. For power-limited recti-
fiers (and not current-limited rectifiers), the maximum output occurs at the lowest output
voltage. For example, a rectifier rated 3.6 kW with minimum output voltage of 42.0 V,
the maximum output current will be almost 86 A. Rectifier output circuit conductor sizing
is covered in Section 5.8.
5.3.8 Rectifier System Checklist

The following checklist can be used as a basic specification when only rectifiers are being
purchased. Suggested wording is provided in parentheses. When the rectifier system is
being purchased with a powerboard, the requirements for temperature compensation and
controls should be coordinated with the powerboard design; see Section 5.4.8.
1. Application: (Specify system application such as “Telecommunications

central office” or “Remote microwave repeater site.”)
2. Configuration: (Specify standalone or modular.)
3. Ambient temperature and (Specify expected ambient temperature range and altitude
altitude: of site.)
4. Nominal system voltage: (–48 V or +24 V)
5. dc output voltage: (Specify voltage or voltage range.)
6. Rectifier system capacity: No less than [x] A total at [y] V output at [z]°C.
7. Compatibility: (If applicable, specify paralleling and load share requirements
and type and model of existing rectifier system.)
8. Output terminations: (Specify type of terminations for dc output circuits, such as
rigid busbar or cable bus; provide details if specific hole
spacings and diameters are required.)
9. ac input voltage: (Specify ac input voltage and phase configuration to be used.)
10. Temperature-compensated (If VRLA, specify number of probes and distance from
charging: rectifier system or compensation controller to battery
system.)
11. Alarm system: (Specify type of alarms required and interface requirements
for alarm system connections.)
5.4 POWERBOARD DESIGN 233
12. Controls and craft interfaces: (Specify if rectifier system will be used to equalize the battery
and if an automatic equalize timer is required; also specify
the craft interface requirements, local area network
interfaces, serial port interfaces, and if data logging is
required and the format.)
13. Spares: (Specify fuse and other consumable spares, such as “For every
fuse rating equipped, provide minimum of [xxx] spare fuses;
provide factory recommended spares for all consumables
(e.g., fan filters), if any.” Sparing of modular rectifiers
usually is handled by ordering an extra module.)
14. Documentation: (Specify documents, such as “Two complete sets of
documentation for installation, operation, and maintenance
of all components provided under this specification.”)
5.4 POWERBOARD DESIGN
Many powerboards are packaged such that rectifiers, controllers, meters and alarms, and
primary distribution occupy one frame (sometimes called initial bay). Additional recti-
fiers and distribution can be configured in additional frames (sometimes called supple-
mentary bays). In new installations, it may be desirable to order and install all supplemen-
tary bays for anticipated growth as part of the initial installation and equip them with
busbars, low-voltage disconnects, and shunts for easy future connection of additional rec-
tifiers or distribution (Fig. 5.20).
Powerboard bus design requires the selection of
앫 Voltage rating
앫 Bus type and current rating
앫 Conductor termination bars
앫 Current shunts
앫 Control options, including low-voltage disconnects
앫 Distribution options
5.4.1 Voltage Rating

Powerboards used in 24 and 48 telecommunications applications are 60-Vdc class unless
they also serve higher voltage loads, such as 130 Vdc, in which case the powerboard must
be 160-Vdc class. Although both 24- and 48-V powerboards have the same voltage class,
individual components usually are set or adjusted for the actual operating voltages. For
example, the alarm set points for 24-V systems are different than 48-V systems.
5.4.2 Bus Type and Amperage Rating

The powerboard bus usually is composed of a rigid copper busbar assembly, although
small powerboards (50 and 100 A) may use a combination rigid and cable bus arrange-
ment. The bus usually is designed for ultimate peak load currents. Ultimate current is
used because it is very difficult if not impossible to increase the current rating of a power-
board without taking it out of service. Even then, physical limitations may preclude ex-
234 SYSTEM DESIGN
Fig. 5.20 10,000-A charge and discharge busbars with current shunt. Battery termination and dis-
tribution busbars are highlighted. (Photo Courtesy of Power-One, Inc.)
pansion, in which case the entire powerboard must be replaced. Packaged powerboards up
to 1200 A rating are available in discrete sizes (typically 50, 100, 200, 400, 800, and 1200
A), and any design rating calculations should be rounded up to the next larger size. Larg-
er powerboards usually are built in increments of 1000 to 2000 A (2000, 4000, 6000 A,
etc.).
Powerboard buses should not be continuously loaded more than 80% of their rated ca-
pacity unless they are specifically designed to be loaded at 100%. The powerboard bus
rating is
IBus = 1.25IRSU amperes (5.15)
where IBus = bus rating (A)
IRSU = ultimate rectifier capacity (A)
1.25 = inverse of 80% derating factor
The ultimate rectifier system capacity is used because that is the maximum continuous
operating current the powerboard bus will ever carry. The ultimate rectifier capacity is
calculated using the methods described in Section 5.3 except the ultimate equipment load
current is used. Depending on the rectifier capacity calculation method, an estimate also
may be required of the ultimate battery system capacity. If the central office space is to
comply with NEBS requirements, the calculations discussed in Section 5.1 can be used as
a cross check.
An alternate method for calculating the powerboard bus rating is based on the recharge
factor discussed in Section 5.3. In this case
IBus = 1.25IEqU FRecharge amperes (5.16)
where IBus = bus rating (A)
IEqU = ultimate equipment load current (A)
FRecharge = recharge factor
The ultimate equipment load current may be based on Section 5.1 estimates or some other
estimating method.
5.4.3 Conductor Terminations

Sufficient space must be provided in the powerboard for terminating both feed and return
conductors from
앫 Battery systems
앫 Rectifier systems
앫 Primary distribution systems
The busbars usually are predrilled for connector lugs with standard bolt-hole diameter and
spacing. Lug sizes will depend on the conductor sizes. In telecommunications dc power
systems, the voltage drop constraints usually require large or parallel sets of conductors.
To minimize field installation problems, any requirements for larger than normal and for
parallel conductor terminations should be given to the powerboard manufacturer before
the powerboard is ordered.
Two-hole lugs are used wherever possible. In telecommunications applications, the
hole diameters and hole spacings are smaller than in most industrial and commercial ac
electrical equipment. Typical values are shown in Table 5.4. Also, to ensure low-resis-
tance crimps, long barrel connector lugs should be used where space permits (crimping
methods are discussed in Chapter 6, System Installation and Maintenance).
All power conductors are paired—for every feed conductor to be terminated on the
charge bus, discharge bus, or overcurrent device, a return conductor must be terminated
on a return bus. For example, if the powerboard is equipped with twenty primary distribu-
tion circuit breakers, the return bus must be able to terminate at least twenty return con-
ductors for those circuits alone. If parallel conductors are used, the termination require-
ments increase by a factor of 2, 3 or more depending on the number of parallel conductors
(Fig. 5.21). Connector lugs normally should not be stacked unless stacking is unavoidable
because of space constraints. The actual stacking configuration and the use of special lugs
for stacking or stacking spacers will depend on the situation (Fig. 5.22).
In many powerboards, separate termination bars are provided for the battery system
and rectifier system feeds but a common termination bar is provided for all return conduc-
Table 5.4 Telecommunications Connector Lug Hole Diameter and Spacing (not all combinations
are available)
Conductor Size Nominal Bolt Size Hole Spacing
8 AWG 0.25, 0.375 in. 0.625, 0.75, 0.8125, 1.0 in.
6—4/0 AWG 0.25, 0.375, 0.5 in. 0.625, 0.75, 0.8125, 1.0, 1.75 in.
250 kcmil 0.25, 0.375, 0.5 in. 0.75, 0.8125, 1.0, 1.75 in.
300–500 kcmil 0.25, 0.375, 0.5 in. 0.75, 1.0, 1.75 in.
600–750 kcmil 0.375, 0.5 in. 1.0, 1.75 in.
800 kcmil 0.5 in. 1.75 in.
1000 kcmil 0.375, 0.5 in. 1.0, 1.75 in.
1500–2000 kcmil 0.5 in. 1.75 in.
236 SYSTEM DESIGN
Fig. 5.21 Conductor terminations on termination bars (lower) in primary distribution bay. (Photo
Courtesy of Schultz Brothers Electric Company.)
Fig. 5.22 Stacking configurations using special stacking lugs, stacking spacers, and back-to-back
arrangements.
tors. This results in a large return bar compared to the individual feed bars to accommo-
date all of the conductors.
5.4.4 Current Shunts

Shunts must be matched to the bus rating, whether on the main bus or distribution buses. In
some installations, the ultimate bus capacity is much larger than the initial load, and a small-
er shunt is installed to accurately measure the lower initial currents. As the load grows, the
shunt is replaced with a larger unit. Systems that have been properly planned will have pro-
visions for mounting the new shunt before removing the old one, thus maintaining continuity
of service. If provisions have not been made for changing out the shunt, a temporary jumper
must be placed across the old shunt while it is removed and replaced. The jumper must have
a current rating at least equal to the load during the change-out. Consideration should be giv-
en to changing the shunt when the load reaches 50 to 80% of the existing shunt rating.
5.4.6 Control Options

The powerboard controller handles alarm collecting and sending, low-voltage disconnect,
rectifier control, craft interface, and data logging and monitoring. The controller may in-
clude an automatic equalize timer if a VLA battery system is used.
5.4.5.1 Equalizer Control Equalize timers normally are set for some fixed time pe-
riod. The timer usually does not start until the batteries have been on discharge for a set
period, such as 10 or 15 min (Fig. 5.23).
End of ac Start of ac
Power Outage Power Outage
Start of ac End of ac
Power Outage Power Outage
Fig. 5.23 Equalize timer operation. If the ac power is restored before the equalize activation timer
expires, the rectifiers remain in the float state. If the ac power is restored after the equalize activation
time expires, the rectifiers are switched to the equalize state. The rectifiers will remain in the equalize
state until the equalize length timer expires and then switch back to float.
238 SYSTEM DESIGN
5.4.5.2 Low-Voltage Disconnects Low-voltage disconnects (LVD) can serve a

number of purposes, as described in Chapter 3, DC Power System Components. When an
LVD is used to protect the battery from overdischarge, the set points can be determined as
follows:
The LVD has two settings—one for disconnect on discharge of the battery and another
for reconnect on recharge of the battery. Hysteresis is provided to allow the battery to be
recharged to some extent before the load is reconnected to the rectifiers. The hysteresis nor-
mally is 7.0 V for 48-V systems and 3.5 V for 24-V systems but other values are used in
practice.
The disconnect setting used in a given installation is based on the voltage drop between
the battery terminals and the LVD voltage sensing point. The LVD and its voltage sensing
point usually are in the powerboard on the battery side of the discharge bus (Fig. 5.24), so
the voltage drop from the battery terminals to the powerboard is the important factor.
The battery should not be allowed to discharge below 1.75 V/cell, or 42.0 V for a 24-cell
48-V battery and 21.0 V for a 12-cell 24-V battery. The voltage drop between the battery ter-
minals and the LVD must be added to the minimum allowed battery discharge voltage to ar-
rive at the set point for the LVD. For example, if the voltage drop during discharge is 0.25 V
for a 24-V system, the LVD would be set to drop out at 20.75 V. With 3.5-V hysteresis, the
reconnect voltage set point would be 24.25 V. If the voltage drop during discharge is 0.75 V
for a 48-V system, the LVD would be set to drop out at 41.25 and reconnect at 48.25 V.
5.4.5.3 Alarm and Control Set Points Alarm and control set points in power-
board controllers will depend on the specific application, but typical values are shown in
Tables 5.5(a), 5.5(b), and 5.5(c) for various battery technologies and configurations.
These settings should be used as a starting point and may need to be adjusted for specific
installations. For example, the battery manufacturer may recommend float and equalize
settings outside the ranges shown, and low- or high-voltage alarm settings may need to be
adjusted to compensate for specific equipment and battery settings. If a controller does
not have a separate battery-on discharge alarm, the low-voltage alarm may be used to in-
dicate battery discharge.
Fig. 5.24 Low-voltage disconnect voltage sensing location.

Table 5.5(a) Controller Settings (VRLA, No Temperature-Compensated Charging), at 25°Ca

Parameter 48-V System 24-V System
Midrange VRLA float voltage (2.25 V/cell) 54.00 V 27.00 V
Maximum VRLA float voltage (2.27 V/cell) 54.48 V 27.24 V
Minimum VRLA float voltage (2.23 V/cell) 53.5 V 26.75 V
High-voltage shutdown 56.0–56.5 V 28.0–28.25 V
High-voltage alarm 55.5 V 27.75 V
Low-voltage alarm 47.0 V 23.5 V
Equalize voltage (if used)—higher than float voltage by 0.7 V 0.35 V
Battery-on-discharge alarm—lower than float voltage by 2.0 V 1.0 V
Low-voltage disconnect See Section 5.4.5.2
a
Consult battery manufacturer before using minimum and maximum float voltages shown.
Table 5.5(b) Controller Settings (VRLA with Temperature-Compensated Charging) at 25°Ca

Midrange VRLA float voltage (2.25 V/cell) 54.00 V 27.00 V
Maximum VRLA float voltage (2.27 V/cell) 55.48 V 27.24 V
Minimum VRLA float voltage (2.23 V/cell) 52.5 V 26.75 V
High-voltage shutdown 56.0–56.5 V 28.0–28.25 V
Equalize voltage (if used)—higher than float voltage by 0.0–0.7 V 0.0–0.35 V
a
Table 5.5(c) Controller Settings (Lead–Calcium VLA) at 25°Ca

Midrange VLA float voltage (2.17 V/cell) 52.08 V 26.04 V
Maximum VLA float voltage (2.22 V/cell) 53.28 V 26.64 V
Minimum VLA float voltage (2.15 V/cell) 51.60 V 25.8 V
High-voltage shutdown 57.6 V 28.8 V
Equalize voltage (2.35 v/cell) 56.4 V 28.2 V
a
240 SYSTEM DESIGN
5.4.6 Distribution Options

The discharge bus normally has circuit breakers or fuses for primary distribution or at
least provisions for extending the bus to where the primary overcurrent devices are locat-
ed. All new powerboards should have spaces for at least six overcurrent devices for pri-
mary distribution:
앫 One pair (A and B) minimum to feed the secondary distribution system directly as-
sociated with a switching system, if applicable
앫 One pair (A and B) minimum to feed the secondary distribution system associated
with transmission and other systems not normally powered through the switching
system distribution
앫 One pair (A and B) minimum to feed future but as yet unidentified loads
This is an absolute minimum requirement. The spaces should be equipped as necessary

for the loads to be served, including initial and growth during the planning period. In most
cases, because of the difficulty in expanding in-service powerboards, ultimate load values
will be used. Figure 5.25 shows a lineup of primary distribution bays.
In powerboards up to around 200 A bus rating, one or more circuit breaker or fuse pan-
els are installed in the power distribution frame. These overcurrent devices may feed
loads directly or indirectly through secondary fuse panels at the top of equipment frames.
Where loads are fed directly from the powerboard, at least 20 overcurrent positions
should be provided in new installations (this is sufficient to feed 10 dual-bus loads or sec-
ondary distribution systems). Where only the minimum is initially provided, the power-
board should have provisions for expanding the number of circuit breaker or fuse panels.
These panels typically have a maximum current rating of 100 to 200 A per panel. It is de-
sirable to install empty panels so that only the circuit breakers need to be installed as
needed. Powerboards larger than approximately 1200 A should be equipped with at least
6 fuse or circuit breaker panels.
From a modern system design standpoint, it is best to abandon the historical practice of
using large overcurrent devices to feed concentrated or centralized loads. A tripped over-
current device will disable a large amount of equipment. If possible, the loads should be
Fig. 5.25 Primary distribution bays—rear view. Discharge busbar connections are at the top and
terminations for overcurrent devices are below. (Photo courtesy of Schultz Brothers Electric Compa-
ny.)
split into increments of no more than approximately 100 to 200 A each. If it is necessary
to feed larger centralized loads, such as the legacy end office circuit-switching systems
made by Lucent, Nortel, and Siemens, the primary distribution may have to include 200-,
400-, or 600-A circuit breakers or fuses to comply with vendor requirements. The largest
individual overcurrent device rating used in telecommunications usually is 600 A, al-
though larger sizes are sometimes used in battery circuits.
Molded case circuit breakers are the most common overcurrent device in modern dc
power systems. Circuit breaker current and voltage ratings are affected by altitude and
may require derating or correction (Table 5.6). Other interrupting devices (e.g., low-volt-
age disconnects and switches) may have similar derating or correction requirements.
5.4.7 Powerboard Checklist

The following checklist can be used as a basic specification. Suggested wording is pro-
vided in parentheses, and variables such as dimensions and cable lengths are shown in
brackets:
1. Application: (Specify system application such as “Attended telecommunica-

tions central office, Unattended telecommunications central
office, or Controlled Environment Vault.”)
2. Ambient temperature and (Specify expected ambient temperature range and altitude of
altitude: site.)
3. Nominal operating voltages: (Specify all voltages required, such as –48 V and +24 V.)
4. Main bus current rating: (Specify current rating and short-circuit current withstand rating
of main charge/discharge buses, such as “[xxx] A bus with
insulated return bar and [xxx] A short-circuit withstand
rating.”)
5. Frame configuration: (Specify equipment frame configuration, such as “Deadfront
powerboard in (1) or (2) [xxx] ft × 23 in. boxed frames,
equipped with insulating covers for all exposed busbars and
energized equipment on rear and sides, front and rear access
(prefer minimum depth with cable turning sections and cable
supports on both sides”.)
6. Controls and accessories (Specify required items such as low-voltage disconnect and
LVD configuration, voltage sensing configuration, such as
(a) selectable remote or local battery voltage sensing for
rectifiers with terminal block for remote sensing leads
(factory set to local sensing), (b) control and alarm panel
with serial and Ethernet/IP interface for monitoring and
control, (c) digital voltmeter/ammeter panel to measure total
Table 5.6 Molded Case Circuit Breaker Altitude Correction Factorsa

Altitude Rated Continuous Rated Voltage
(ft/m) Current (A) (V)
< 6,600/2,000 1.00 1.00
8,500/2,600 0.99 0.95
13,000/3,900 0.96 0.80
a
Use linear interpolation for intermediate altitudes. Source: From [16].
242 SYSTEM DESIGN
bus and individual distribution buses, (d) alarm system with

cables for external connections, [xxx] ft, (e) temperature-
compensated charging control, (f) factory recommended
voltage and alarm settings for [VLA or VRLA] batteries, (g)
battery monitoring and data logging, (h) craft interfaces, and
(i) data logging.)
7. Termination bars (Specify termination bar requirements, such as “Battery/
rectifier cable termination and return bars with lug landings to
accommodate factory rectifier terminations and at least (Qty)
[xxxx] kcmil copper conductors with 2-hole lugs on each bar for
field battery lead terminations. Include lug landings for at least
(Qty) [xxxx] kcmil dc grounding conductor with 2-hole lugs on
return busbar.” Adjust number of conductor terminations to
suit actual installation.)
8. Grounding provisions (Specify ground bar configuration, such as “Insulated ground
termination bar with space for [xxx] 2-hole lugs.)
9. Seismic Seismic zone or description
10. Primary distribution (Specify –48 V or +24 V or both circuit breaker or fuse
distribution panels, dual feed bus (A & B) with bus meter
shunts, trip handle guard, alarms, and return buses. Also
specify the required overcurrent protective device quantities
and ratings. Be sure to specify spaces for extra circuit
breakers or fuses if required for future growth.)
11. Secondary distribution (If required, this will include the same basic information as for
primary distribution.)
12. Temperature-compensated (If VRLA, specify details, such as “Provide (Qty) temperature
charging probes with [xxx] ft cable lengths, and provide ambient temper-
ature probe.”)
13. Rectifier system (Specify requirements such as “Modular rectifier system with N
+ 1 redundancy and [xxx] A working capacity, factory set to
float at [xx.xx] V. Input ac rating suitable for 208 or 240 V (or
480 V) service voltages. Rectifier alarms and automatic high-
voltage shutdown required.”) (See also Section 5.3.)
14. Converters (Specify DC–DC converters, if required. Include redundancy
requirements, quantity, bus configuration, rating, input and out-
put voltages, mounting arrangements, and distribution panels.)
15. Inverters (Specify inverters, if required. Include redundancy require-
ments, quantity, bus configuration, rating, input and output
voltages, mounting arrangements, and distribution panels.)
16. Spares (Specify fuse and other consumable spares, such as “For every
fuse rating equipped, provide minimum of [xxx] spare fuses;
provide factory recommended spares for all consumables (e.g.,
fan filters), if any.”)
17. Documentation (Specify job documents, such as “Two complete sets of docu-
mentation for installation, operation and maintenance of all
components provided under this specification, including frame
elevation drawings, termination bar configurations, hole pat-
terns, and wiring diagrams”)
5.6 VOLTAGE CONVERSION SYSTEMS DESIGN 243
5.5 SECONDARY DISTRIBUTION DESIGN
Secondary distribution systems are similar to primary, the main differences are their loca-
tion and overcurrent device ratings (secondary distribution systems usually have smaller
ratings). The specification for secondary distribution systems includes
앫 Overcurrent protection device type (fuse or circuit breaker, or combination)

앫 Bus configuration (single bus or dual bus)
앫 Bus rating
앫 Input wire termination requirements
앫 Output wire termination requirements
앫 Mounting arrangements
앫 Alarm requirements
앫 Filtering requirements
Typical mounting arrangements for small fuse and circuit breaker panels in equipment
frames are shown in Figure 5.26. A secondary distribution bay is shown in Figure 5.27.
5.6 VOLTAGE CONVERSION SYSTEMS DESIGN
5.6.1 The DC–DC Converters

Converter system design includes
앫 Capacity
Fig. 5.26 Fuse and circuit breaker panel mounting arrangements. Midmounted fuse or circuit
breaker panels can be dangerous unless the busbars are completely enclosed. Also, midmounted
circuit breaker panels must have handle guards to prevent accidental tripping.
244 SYSTEM DESIGN
Fig. 5.27 Secondary distribution bay. The return busbars are at the top and the individual distribu-
tion busbars for each circuit breaker panel are below (the panels shown are empty). (Photo courtesy
of Schultz Brothers Electric Company.)
앫 Redundancy requirements
앫 Bus configuration
앫 Location
앫 Input circuits
앫 Output circuits
Converter system capacity is determined using the same basic techniques as the primary
voltage system (Section 5.1). The various bus configurations are discussed in Chapter 3,
DC Power System Components. A typical converter system will use N + 1 converters
with their outputs connected to a common bus with dual-bus (A and B) distribution.
Generally, but not always, converter systems are conveniently located close to their
loads. The actual location will take into account the wire size and quantity required for the
input and output circuit wiring. When the converters are used to convert from a higher to
a lower voltage (e.g., from 48 to 24 V), the current on the inputs will be lower than the
outputs and smaller wiring generally is needed on the input than output. In this case, it
would be advantageous to locate the converters closer to the loads. For the opposite situa-
tion, converting from a lower to higher voltage, the converters could be located closer to
the source. The wiring costs for either situation will have labor and material cost compo-
nents and may influence the location.
Table 5.7 DC–DC Converter Input and Output Voltage Drops

Input/Output Input Voltage Output Voltage
Voltage (Vdc) Drop (V) Drop (V)
48/48 2.0 2.0
48/24 2.0 1.0
24/48 1.0 2.0
24/24 1.0 1.0
Input and output circuits are designed according to the methods described in Section
5.8.2. There is nothing unique about converter circuits except that converters generally
operate over a wider input voltage range than the load equipment they serve. Table 5.7
shows input and output voltage drops for common system voltages; the input voltage
drops may be adjusted for the actual input voltage ranges of the converters used in a par-
ticular application.
5.6.1.1 Economic Analysis A converter system should not be chosen without first
comparing its costs to the costs of a separate powerboard, rectifier system, and battery
plant. These costs include material and installation labor costs, operating and maintenance
costs, as well as building space costs. A battery system significantly adds to the mainte-
nance and disposal costs. A separate battery system almost always requires more space
than a converter system but almost always will have lower material and installation labor
costs when the load is more than 100 to 200 A. When comparing a converter system to a
conventional plant with a battery, it is necessary to consider that a converter system will
affect the primary voltage system’s battery reserve time and distribution (the converter
system is an added dc load). If the secondary voltage system has its own battery, there
will be no effect on the primary voltage system.
5.6.2 Inverter Systems

Inverter system design includes
앫 Capacity (VA)
앫 Configuration
앫 Redundancy requirements
앫 Input voltage and voltage range
앫 Output voltage and voltage range
앫 Output frequency and frequency range
앫 Input circuits
앫 Output circuits
Inverter input voltage and range must be compatible with the dc system voltage. Gen-
erally, inverters operate over a fairly wide range and operate at lower and higher voltages
than most network equipment. The output voltages of inverters used in the United States
correspond to the common ac utilization voltages described in Section 5.8.5. Small invert-
er systems up to around 1500 VA usually operate at 120 V, while larger inverter systems
operate at 120/240 V single-phase or 208Y/120 V three-phase. Inverter operating fre-
quency usually is 60 Hz ±3 Hz or ±2 Hz, although many modern types can provide a
246 SYSTEM DESIGN
much tighter tolerance (e.g., ±0.01%). Older inverters are acoustically quite noisy but
modern types are considerably quieter.
The dc input current should be obtained from manufacturers’ data sheets whenever
possible. Alternately, if the efficiency is known, the input current can be calculated from
(VA)(PF)
Idc = amperes (5.17)
VdcMin(Eff)
where Idc = dc input current (A)

VA = output volt–ampere rating (VA)
VdcMin = minimum dc input voltage (V)
Eff = efficiency at full load
PF = power factor
Generally, modern inverters operate at voltages as low as 40.0 to 42.0 Vdc input, but they
have relatively low efficiency (in the range of below 70% to above 90%). Inverter dc in-
put circuit design is the same as other dc load equipment circuits (Section 5.8.2).
The inverter operating configuration—active standby, passive standby, and continuous
operation—depends on operational requirements. Both passive and active standby have
been used in the past, but the most common configuration presently is continuous opera-
tion. With continuous operation, there is no connection to the commercial electrical ser-
vice, which reduces the likelihood of ac power disturbances damaging the inverter and as-
sociated loads and also simplifies the inverter system bonding and grounding.
The first step in determining the inverter system capacity is to estimate the total load.
Most inverter loads in telecommunications applications are continuous, although some
may include motor starting and intermittent loads (such as test equipment). All ac-pow-
ered equipment has a label indicating the load in ac amperes at a specified operating volt-
age or the load in volt–amperes or watts. For ac power supplies labeled in amperes, the
apparent power in volt–amperes can be calculated by multiplying the ac voltage by the ac
current. As discussed in Chapter 2, Electricity Review, the load in watts is related to the
load in volt–amperes by the load power factor.
Next, determine the load characteristics. The ac equipment connected to telecommuni-
cations inverter systems are broadly categorized as electronic loads. It may be possible
that lighting and motor loads are connected, but that is unlikely. Electronic loads consist
mostly of apparatus with switch-mode power supplies (SMPS), typically operational sup-
port systems such as servers, desktop computers (PCs), monitors, printers, and local area
network (LAN) equipment. Electronic, lighting, and motor loads have different initial
startup (surge) requirements. The startup currents for electronic and motor loads are con-
siderably higher than the corresponding running loads, and lighting startup loads depend
on the type of lighting technology (incandescent or gas discharge and the type of ballast
used with gas discharge lighting).
Some switch-mode power supplies, particularly older products, have a low input pow-
er factor due to relatively high peak current and distorted input current waveform (high
harmonics content). Many modern inverters, especially some high-frequency types, do
not run well when heavily loaded by low power factor equipment and have to be derated.
This is handled by using a 30 to 50% power factor when converting the load in watts to
load in volt–amperes. For example, if an older switch-mode power supply is rated 250 W
and is to be powered by a high-frequency inverter, the equivalent load rating is 250 W ÷
0.3 = 833 VA. Inverters with a true (or close to true) sine wave output waveform usually
have an output transformer that reduces the effect of low load power factor. The estimat-
ed equivalent power supply load for this example is 250 W ÷ 0.5 = 500 VA.
In practice, desktop computer and monitor power supplies seldom operate a full rated
load. For example, even though the power supply in a PC is rated 350 W, it may only
draw 150 W when running. The extra capacity usually is required on initial startup of the
PC or when the PC motherboard is completely filled with high-power interface cards.
Caution must be exercised when a high-frequency inverter is used to supply motor
loads, such as compressor motors in transmission line dehydrators or pumps. Some types
of inverters cannot deliver the required starting current. Low-frequency, transformer-iso-
lated inverters or inverters specially designed for motor loads are better for this application.
The required inverter capacity for motor starting can be found by multiplying the inverter
output voltage by the motor’s locked rotor (or surge) current, typically 6 to 10 times its full-
load running current. For example, if a motor has 2.0 A full-load running current at 120
Vac, the locked rotor current is approximately 2.0 A × 6 = 12 A and the required inverter
capacity is 12 A × 120 Vac = 1440 VA. If the load consists of more than one motor, they
should be prevented from starting at the same time when powered by the inverter system.
Example 5.13 Various ac equipment items presently are connected to 120-Vac power
strips in a small central office and are to be connected to an inverter system. The load in-
formation tabulated below is taken directly from the equipment nameplates. The total
load current measured with an ac clamp-on ammeter is 8.5 A. Determine the required in-
verter capacity.
Item Description
1 Remote access modem hub, 30-port—Nameplate 100 W at 100–240 Vac
2 Desktop server, email—Nameplate 230 W, 6.0 A
3 Desktop server, web—Nameplate 230 W, 6.0 A
4 Desktop server, router—Nameplate 230 W, 6.0 A
5 Color monitor—Nameplate 1.3–0.7 A at 100–125 Vac
6 Color monitor—Nameplate 1.5 A maximum at 100–240 Vac
7 Color monitor—Nameplate 1.5–0.6 A at 100–240 Vac
8 Dot matrix printer—Nameplate 0.8 A at 100–120 Vac
9 Dot matrix printer—Nameplate 2.0 A at 120 Vac
10 Modem—Nameplate 16 VA
11 Router—Nameplate 1.2–0.6 A at 100–240 Vac
12 10BaseT Ethernet hub, 8-port—Nameplate 10 W
13 Ethernet switch, 16-port—Nameplate 25 VA at 100–240 Vac
14 Data service unit (DSU), 56 kb/s—Nameplate 0.15 A at 115 Vac
15 Data service unit (DSU), 64 kb/s—Nameplate 0.22 A at 120 Vac
Solution The above load information is transferred to Table 5.8, where nameplate infor-
mation (if available) is shown unshaded and calculated or estimated information is shown
shaded. Assumptions are provided in notes below the table.
The total estimated load as indicated in the table is 1971 VA. The measured running
load is 8.5 A × 120 Vac = 1020 VA, or about one-half the estimate. This indicates that the
248
Table 5.8 Inverter Load Tabulation Form with Example Dataa
Nameplate Nameplate Starting Max. Max.
Load Current Voltage Nameplate Nameplate Surge Starting Running Running
Description Type (A) (Vac) (VA) (W) Factor (VA) (VA) (W)
Remote access modem hub, 30-port SMPS — 100–240 — 100 2 200 ~111 100
Desktop server, email SMPS 6.0 — — 230 2 460 ~256 ~230
Desktop server, web SMPS 6.0 — — 230 2 460 ~256 ~230
Desktop server, router SMPS 6.0 — — 230 2 460 ~256 ~230
Color monitor SMPS 1.3–0.7 120–240 — — 2 ~280 156 ~140
Color monitor SMPS 1.5 120–240 — — 2 ~320 180 ~160
Color monitor SMPS 1.5–0.6 120–240 — — 2 ~320 180 ~160
Dot matrix printer SMPS 0.8 100–120 — — 2 ~172 96 ~86
Dot matrix printer SMPS 2.0 120 — — 2 ~432 240 ~216
Modem SMPS — — 16 — 2 ~28 16 ~14
Router SMPS 1.2–0.6 100–240 — — 2 ~260 144 ~130
Ethernet hub, 8-port SMPS — — — 10 2 ~20 11 ~10
Ethernet switch, 16-port SMPS — 100–240 25 — 2 ~46 25 ~23
Data service unit, 56 kb/s SMPS 0.15 115 — — 2 ~32 18 ~16
Data service unit, 65 kb/s SMPS 0.22 120 — — 2 ~46 26 ~23
Total Estimated ~3,536 ~1,971 ~1,378
Actual Measured N/A 1,020 ~920
a
The “~” indicates an estimated or calculated value. 0.9 power factor is assumed for all power supplies; (therefore, Max. Running W = 0.9 × Max. Running VA).
5.7 OTHER DESIGN CONSIDERATIONS 249
equipment internal power supplies are somewhat oversized. The estimated starting VA,
which assumes all loads start at the same time (in this case, a likely scenario) is about 3.5
times the actual running VA, although experience shows this to be somewhat overstated.
In this application, an inverter rated around 2000 to 2500 VA would serve the existing
load equipment; however, the load is likely to increase over time. The actual inverter size
would include not only the existing loads but a prediction of inverter load growth. A mod-
ular system with three or four 1000-VA inverter modules would serve existing loads and
provide redundancy and some growth. If four modules are used initially, additional in-
verter modules will be required to meet N + 1 redundancy requirements when the actual
load approaches 3000 VA.
Inverter ac output circuit design is the same as other ac branch circuits (Section 5.8.5).
Since the inverter is a power source, the output circuit wiring must have overcurrent pro-
tection. Most load equipment operates at 120 Vac and is connected to an inverter system
through common 15- or 20-A duplex receptacles (e.g., NEMA17 5-15R or 5-20R, two-
pole, three-wire grounding, straight blade, 125-V receptacles). These circuits can be pro-
tected by 15- or 20-A circuit breakers, depending on the rating of the receptacle, and
wired with 12 AWG wire.
Inverters typically have an output circuit breaker matched to the inverter output VA
rating. Where this output circuit breaker is larger than 20 A or a modular systems is used
in which the output exceeds the rating of 15- or 20-A branch circuits, the system cannot
directly feed regular duplex receptacles. In this case, the output is connected to an ac cir-
cuit breaker panelboard that, in turn, feeds individual branch circuits (Fig. 5.28). Some in-
verter systems are available with an accessory ac distribution panel that is integrated with
an inverter maintenance bypass switch. This simplifies branch circuit installation by elim-
inating the need for a separate panel.
In some inverters, the output is current limited in such a way that there is not enough
fault current to trip protective devices in the event of a fault; others are specifically design
to coordinate with their output protection device. Generally, inverter distribution circuits
should use circuit breakers provided or recommended by the inverter manufacturer.
A maintenance bypass switch may be used to isolate the inverters and connect the
loads directly to the commercial ac power source during maintenance (Fig. 5.29). The by-
pass switch normally applies only to passive standby or active standby operation. When
an inverter system is set up for continuous operation, it is not connected to the commer-
cial ac source.
5.7 OTHER DESIGN CONSIDERATIONS
5.7.1 Powerline Filter Applications

Powerline filters are required wherever electromagnetic interference (EMI) is coupled
into dc power circuits by rectifier systems, load equipment, or other noise sources. The
primary coupling mechanisms are conduction, radiation, and induction. Noise from recti-
fiers and load equipment usually is by conduction, and the coupling can be into or out of
the equipment through the power circuit wiring. Radiation coupling usually is from wire-
less equipment such as radio transmitters or the oscillators and power supplies in load
17
National Electrical Manufacturers Association.
250 SYSTEM DESIGN
Inverter ac
Panelboard
or Circuit
Breaker
Distribution
Panel
Fig. 5.28 Inverter output distribution.
Fig. 5.29 Inverter maintenance bypass switch S1 consists of six contacts, S1-1 through S1-6. The
switch is shown in the maintenance position in which S1-2 is closed and all other contacts are open.
In this position, no dc is applied to the inverter modules and the inverter ac output and transfer
switch are isolated from all external connections. When the maintenance bypass switch is in the on-
line position, all contacts except S1-2 are closed, and the inverter system is in active standby opera-
tion. The transfer switch connects inverter ac output to the load through S1-3. If the inverter system
fails, the transfer switch connects the commercial ac input to the load through S1-1 and S1-3. When
the maintenance bypass switch is in the test position, all contacts except S1-3 are closed and dc is
applied to the inverter modules, but inverter ac output is isolated from the load. The test position is
used to verify inverter operation.
equipment. The radio transmitters do not have to be in the same room or building as the
power circuits. Induction coupling is from nearby disturbing circuits that run parallel to
the disturbed circuit.
Central offices in which the metallic components, such as equipment frames and
grounding electrode systems, are not well bonded will have more noise problems than
central offices in which great attention has been given to this aspect of equipment installa-
tion. Also, some equipment is more susceptible to noise than others and, in some cases,
the equipment manufacturer will recommend powerline filtering.
There are no industry guidelines that describe the conditions under which powerline
filters are installed, so their use and application becomes one of judgment and experi-
ence. All equipment installed in the vicinity of radio stations and known powerful noise
sources should have powerline filters. Filters should be installed as close as possible to
the equipment that requires filtering. An ideal location for the filter is in the fuse or cir-
cuit breaker panels typically located at the top of each equipment frame. Most equip-
ment frame fuse and circuit breaker panels can be purchased with optional powerline fil-
ters.
Equipment installed in the United States must meet the conducted and radiated limits
specified in 47CFR, part 15 of the FCC rules and regulations [17]. In some cases, NEBS
requirements may be more stringent. As a result of these requirements, equipment de-
ployed within the last 20 years is much less noisy (EMI) than older equipment.
5.7.2 Building Codes

Building codes have the effect of law when they have been adopted by governmental ju-
risdictions, such as states, counties, and municipalities. The codes relevant to dc power
system design are18
앫 NFPA 70 National Electrical Code (NEC)

앫 International Fire Code (IFC)
앫 International Building Code (IBC)
5.7.2.1 NFPA 70 National Electrical Code (NEC) The purpose of the NEC [18]
is to specify minimum safety requirements for electrical installation and not to serve as an
electrical design manual. It does not apply to central offices or to spaces used exclusively
for telecommunications unless company policy requires it to be used:
Article 90.2(b) Not Covered—”(4) Installations of communications equipment under the ex-
clusive control of communications utilities located outdoors or in building spaces used exclu-
sively for such installations.” [18, page 70-23]
Although the NEC does not apply as stated above, all dc power systems should be made
to comply with it wherever possible. Where the NEC does apply or is used, local county
or municipal modifications, if any, also apply. The NEC historically has been a difficult
document to use and interpret. An extensive rewrite in 2002 improved this situation
18
Not listed here is the National Electrical Safety Code (NESC), ANSI C2. The NESC has no specific require-
ments for telecommunications dc power systems, but it does have requirements for dc power systems used in
electrical power facilities.
252 SYSTEM DESIGN
slightly, but disputes over interpretations still arise and interpretations by the Authority
having jurisdiction (AHJ) will hold precedence.
5.7.2.2 International Fire Code (IFC, formerly Uniform Fire Code) Certain
sections of the IFC [19] apply to telecommunications dc power systems under limited cir-
cumstances. Specifically, Chapter 6, Section 608—Stationary Lead–Acid Battery Sys-
tems and Section 609—Valve-Regulated Lead–Acid (VRLA) Battery Systems address
battery systems with more than 50 gal (189 liters) of electrolyte capacity. The IFC has no
requirements for battery systems with less than 50 gal. Systems exceeding the 50-gal
threshold must meet ventilation (see Section 5.7.4), spill control and other requirements.
Spill control includes containing and neutralizing spilled electrolyte to a pH between 7.0
and 9.0 for 100% of capacity of the largest VLA cell or battery module and for 3% of the
capacity of the largest VRLA cell or monoblock cells. For example, if the largest VLA
cell or module holds 4 gal of electrolyte, there must be a method to control and neutralize
a spill of 4 gal. If the largest VRLA cell or block contains 4 gal, control and neutralization
must be available for 0.12 gal.
Although not a code (unless adopted by a government jurisdiction) and not directly re-
lated to the IFC, NFPA 76—Recommended Practice for the Fire Protection of Telecom-
munications Facilities [20] provides guidelines related to fire in telecommunications fa-
cilities. Chapter 6, Large Telecommunications Facilities applies to telecommunications
equipment space >2500 ft2 and includes fire detection, fire suppression, and limitation of
combustibles. It details the requirements for each major area:
앫 Telecommunications equipment spaces

앫 Cable entrance facilities
앫 Power areas
앫 Main distribution frames
앫 Standby engine areas
앫 Technical support areas
앫 Administrative areas
NFPA 76, Chapter 7, Small Telecommunications Facilities, applies to telecommunica-

tions equipment spaces between 500 and 2500 ft2 and includes fire detection, fire suppres-
sion, and limitation of combustibles but does not provide much detail.
NFPA 76 does not cover spaces <500 ft2 even though there are a large number of re-
mote enclosures (huts) and controlled environment vaults (CEVs) in the industry that fall
within this dimension. It is up to the designer to determine if fire detection and fire sup-
pression is to be provided. Small enclosures historically have not had fire problems and a
catastrophe in one does not normally affect a large number of telecommunications users.
5.7.2.3 International Building Code (IBC) The IBC [13] covers all aspects of
building construction and indirectly affects the installation of dc power systems, particu-
larly batteries because of their heavy floor loading.
5.7.3 Power Equipment Spaces

Battery systems normally are located in a room separate from network or electronic
equipment, but this is impractical in small installations such as CEVs and remote equip-
ment enclosures. When not located in a separate room, barriers or some type of mechani-
cal protection should be provided to prevent inadvertent contact and damage to live com-
ponents. In low voltage systems, these barriers typically are clear or opaque plastics such
as Plexiglas, Acrylite, Lucite, and Lexan (Fig. 5.30).
Batteries require adequate space for routine maintenance, removal, and replacement
and for air to freely circulate around them for cooling. Typical clearance requirements are
shown in Figure 5.31. Properly ventilated battery rooms are not considered hazardous
areas and special electrical wiring is not required.
The environmental requirements of dc power system equipment are similar to most
network equipment. Some of the parameters that should be addressed are
앫 Absolute high and low temperatures

앫 Temperature rate of change
앫 Humidity (noncondensing)
앫 Altitude (derating at higher altitudes)
Estimates of the heat dissipated by rectifiers, dc–dc converters, and system monitors and
controllers can be based on the equipment data sheets or based on estimates. Generally,
the rectifiers, inverters, and converters will be responsible for most heat gain. Tables 5.9
and 5.10 may be used as a starting point to estimate the heat gain from rectifiers and other
power system equipment.
5.7.4 Battery Room Ventilation

The telecommunications industry and other industries in the United States do not general-
ly agree on the amount of ventilation required in battery rooms, and there are considerable
variations in the ways ventilation calculations are made. This section will review the is-
sues and methods in present use.
Fig. 5.30 Insulating plastic covers on the back of a power bay. (Photo courtesy of Schultz Brothers
Electric Company.)
254 SYSTEM DESIGN
Fig. 5.31 Adequate space must be provided around a battery to allow free air movement and for
movement during seismic events. Space for removal or replacement should be at least 30 in. Clear-
ance from energized parts generally is 36 in.
Table 5.9 Approximate Heat Gain from Rectifiers (Based on 91.6% Efficiency)
48-V Rectifier 24-V Rectifier
Rating (A) Rating (A) Watts Btu/h
25 50 125 425
50 100 250 850
100 200 500 1,700
200 400 1,000 3,400
400 800 2,000 6,800
800 1,600 4,000 14,000
1,600 3,200 8,200 28,000
Table 5.10 Approximate Additional Heat Gain from Miscellaneous Power

Equipment (Not Including Rectifiers)
dc Power System dc Power System
Rating @ 48 V Rating @ 24 V Watts Btu/h
1,000 A 2,000 A 500 1,700
2,000 A 4,000 A 1,000 3,400
4,000 A 8,000 A 2,000 6,800
6,000 A 12,000 A 3,000 10,200
10,000 A 20,000 A 5,000 17,000
Under normal conditions, VLA cells release considerably more hydrogen gas than
VRLA, but a VRLA cell with a failed valve acts as a VLA cell and vents as much gas as a
VLA cell. Battery ventilation systems must be designed to account for abnormal condi-
tions.
Hydrogen is a very flammable gas. Fire or an explosion may occur in spaces where the
atmosphere is within the explosive limits of a flammable gas. Explosive limits (or flam-
mable limits) are expressed in percentage by volume and are defined as the minimum and
maximum concentrations of a flammable gas between which ignitions can occur. Concen-
trations below the lower explosive limit (LEL) are too lean to burn while those above the
upper explosive limit (UEL) are too rich. For hydrogen in air at ordinary temperature and
pressure, LEL = 4.0% and UEL = 74.2%. Therefore, for practical applications the ventila-
tion must keep the concentration well below the LEL of 4%. A safety factor in the range
of 2.0 to 5.0 is used, which limits the evolved hydrogen concentration to no more than 0.8
to 2% by volume. For example, if a room volume is 1000 ft3, the hydrogen volume would
be limited to 10 ft3 for 1% concentration.
Hydrogen gas evolution can be estimated from [21]
T
冢冣
QH = qH INCell 1 + liters/h
273
(5.18a)
where QH = hydrogen gas evolution rate (liters/h) Note: 1000 liters/h = 1 m3/h
qH = volume of hydrogen gas evolved per cell per ampere-hour (0.42 liters/Ah)
I = current through cells (A)
NCell = number of cells in battery system
T = cell temperature (°C)
When the cell temperature is 25°C (77°F), Eq. (5.18a) becomes
QH = 0.46 × I × NCell liters/h (5.18b)
which is equivalent to the rate of hydrogen evolution given in [22]. To convert to cubic
meters/hour, multiply liters/hour by 0.001. In more familiar units
QH = 0.00027 × I × NCell ft3/min (CFM) (5.18c)
which is equivalent to the rate of hydrogen evolution given in [23].

The foregoing equations for hydrogen evolution give a higher gas evolution rate than
is experienced in practice because only a portion of the current causes hydrogen gas to be
generated. However, it is a useful starting point for ventilation calculations.
When the cells are fully charged and the voltage across them is the normal float volt-
age, the current I is the expected float current. Float current varies with the type of bat-
tery, float voltage, age, and temperature. Some typical values are shown in Table 5.11.
Battery room ventilation systems are not designed for normal float conditions but for
abnormal conditions. There are at least four scenarios for abnormal operation of VLA and
VRLA batteries:
앫 One or more cells in a string shorts out, exposing the remaining cells to a higher
than normal float voltage and forcing them into an overcharge condition (barring
more specific requirements, assume 10% of the cells are shorted for ventilation de-
sign purposes).
256 SYSTEM DESIGN
Table 5.11 Some Typical Float Currents for Fully Charged Lead–Acid Batteriesa
Charge Current at 25°C (mA/100 Ah)
Charge Voltage
(V/cell) Pb–Sb Pb–Ca Pb–Sn
1.215 SG 1.215 SG 1.240 SG
2.15 15–60 — —
2.17 19–80 4 —
2.20 26–105 6 11
2.23 37–150 8 18
2.25 45–185 10 27
2.27 60–230 12 —
2.33 120–450 24 60
2.37 195–700 38 95
2.41 300–1100 58 —
a
Data from [24].
앫 Rectifier system output voltage is incorrectly set too high or drifts out of adjustment,
exposing all cells to a higher than normal float voltage and overcharge condition.
앫 A rectifier fails in a mode that causes its voltage to increase and the high-voltage
shutdown circuits fail to operate.
앫 A cell fails, is removed from the string, and a shorting bar is installed in its place to
keep the battery in service; unless the float voltage is decreased, this exposes the re-
maining cells to overcharge as in the first scenario above.
If any of these scenarios occur, the current through the cells is considerably higher than
the normal float current, and the hydrogen evolution also is considerably higher.
One estimate for abnormal operation that is commonly used is 0.01 A/Ah (1 A/100 Ah)
of cell capacity [that is, I = 0.01 × (AH8hour) A]. Using this value for float charge conditions
QH = 2.7 × 10–6 × (AH8hour)NCell CFM (float condition) (5.18d)
To convert from CFM to liters/hour, multiply by 1699.

The worst-case float current would be the maximum available rectifier system current;
however, the maximum current is not ordinarily used to calculate hydrogen evolution be-
cause it is highly improbable that a battery or cell failure will sink the entire current avail-
able from the rectifier system. Instead, it is common practice to use one-fourth of the max-
imum available rectifier current. With this adjustment, Eq. (5.18b) and Eq. (5.18c) become
IMax
冢冣
QH = 0.46 NCell liters/h
4
(5.18e)
IMax
冢冣
= 0.00027 NCell CFM
4
(5.18f)
where IMax is the maximum current available from the rectifier system when in overload
(amperes).
Now that a method has been described to determine the rate at which hydrogen gas is
evolved from a battery system, it is necessary to calculate the required ventilation rate. To
limit the gas concentration to a particular concentration, the ventilation rate is
QH
QVentRate = (5.19)
GLimit
where QVentRate = fan airflow rate (liters/h or CFM)

QH = hydrogen gassing rate (liters/h or CFM)
GLimit = hydrogen gas concentration limit by volume (0.01–0.02)
Example 5.14 Determine (1) the hydrogen evolved during normal (float) and abnormal
conditions and (2) the fan capacity to limit hydrogen concentration to 1% by volume dur-
ing abnormal operation for the following system. Provide the answers in liters/hour and
CFM for a cell temperature of 25°C (77°F):
Remote equipment enclosure: 8 ft (2.4 m) wide × 12 ft (3.7 m) long × 9 ft (2.7 m) high

48-V, 24-cell, 150-Ah battery
Rectifier capacity: 40 A (four 10-A rectifiers in N + 1 configuration, N = 3)
Equipment load 12 A
Solution The float current during normal operation is assumed to be 0.01 A/100 Ah ×
150 Ah = 0.015 A. Therefore,
QH = 0.46NCellI = 0.46 × 24 × 0.015 = 0.166 liters/h

= 0.00027 × 24 × 0.015 = 0.000097 CFM
During abnormal operation, the calculations give
IMax 40
冢冣冢冣
QH = 0.46 NCell = 0.46 24 = 110.4 liters/h
4 4
40
冢冣
= 0.00027 24 = 0.065 CFM
4
To maintain a concentration no greater than 1% by volume,
QH 110.4
QVentRate = = = 11,040 liters/h
GLimit 0.01
0.065
= = 6.5 CFM
0.01
As a point of reference, a typical residential bathroom fan is rated 50 to 100 CFM.

Hydrogen is very easily diluted by air, and in small installations (< 100 Ah) it is not al-
ways necessary to use fans if the space is equipped with inlet and outlet air vents. For nat-
ural ventilation, a common expression for the area of each vent (inlet and outlet) is
AVent 2800QH mm2 where QH is in m3/h

258 SYSTEM DESIGN
or
AVent 0.0178QH ft2 where QH is in ft3/min (CFM) (5.20)
Battery rooms and spaces should have inlet vent louvers in the lower half of doors
or walls and outlet louvers near the ceiling to outside air. The inlet vent area should be 1.5
times the outlet vent area to prevent fan starvation and to ensure free airflow (Fig. 5.32).
In larger installations, it is good practice to use at least one fan. If the battery room is
air conditioned, the exhaust air from the battery room should not be returned to the air dis-
tribution system. Battery room exhaust systems or vents always should be directed out-
doors. The exhaust requirements can be problematic in some climates, both hot and cold,
and in those cases only a portion of the air is exhausted, with make-up air provided for
that portion.
NFPA 111-2001, clause 5.3.2 [25] requires two air changes per hour in VLA battery
rooms. Although no guidance is given for VRLA battery rooms, if the NFPA requirements
are used, they should be applied to VRLA to be on the safe side. Note that these require-
ments do not take into account that a small battery may be installed in a big room or a big
battery in a small room. Two air changes in one hour means that the ventilation system must
move twice the room volume in one hour. To meet this requirement, the fan capacity is
QVentRate = 2(WRoomLRoomHRoom) m3/h (5.21a)
Fig. 5.32 Battery room ventilation layout.

where WRoom = room width (m)

LRoom = room length (m)
HRoom = room height (m)
To convert QFan in m3/h to ft3/min, multiply by 0.589, or use

2
QVentRate = —
60 (WRoomLRoomHRoom) CFM (5.21b)
where WRoom = room width (ft)

LRoom = room length (ft)
HRoom = room height (ft)
Example 5.15 Determine the fan capacity in liters/hour and CFM for the space in Exam-
ple 5.14 if the requirements of NFPA 111 are followed.
Solution The room volume is 24.5 m3 or 864 ft3. To achieve two air changes per hour,
the ventilation rate must be 49 m3/h or 1728 ft3/h, or in the specified units, 49,000 liters/h
or 28.8 CFM.
Other criteria have been used to calculate battery room ventilation rate. For example,
where the IFC applies and detailed calculations are not made to show otherwise, the re-
quired ventilation rate must be at least 1 ft3 per minute (CFM) per ft2 of floor space con-
tinuous, or as necessary to limit hydrogen to 1% by volume.
Example 5.16 Determine the fan capacity in CFM using IFC requirements for the space
in Example 5.14.
Solution The space area is 8 ft × 12 ft = 96 ft2 and the required ventilation rate is 1
CFM/ft2 × 96 ft2 = 96 CFM (163,105 liters/h).
A method that is frequently used outside the United States determines the amount of
ventilation required to maintain the LEL and then applies a safety factor to increase the
ventilation by a factor of 5, as in
QVentRate = DqHSNCellI liters/hour (5.22a)
where D = dilution factor for hydrogen in air at LEL = 100%/3.8% (Note: The LEL is
3.8% rather than 4.0% in this equation)
S = safety factor = 5
I = charge current = 0.02 A per ampere-hour of capacity
Substituting the above values gives
QVentRate = 26.3 × 0.42 × 5 × NCell × I = 55 × NCell × I liters/hour (5.22b)
Example 5.17 Determine the fan capacity in liters/hour and CFM for the space in Exam-
ple 5.14 using Eq. (5.22b).
Solution Using the criteria given, the charging current for a 150-Ah battery is 0.02 ×
150 = 3 A. Therefore,
260 SYSTEM DESIGN
QVentRate = 55 × 24 × 3 = 3960 liters/h
To convert to CFM, multiply by 0.0005886, giving 2.3 CFM.

Another industry practice is to use the most stringent of the following criteria:
1. One air change every 4 h

2. Two CFM per battery string
3. Twenty CFM per person when the space is occupied
4. Applicable codes
Example 5.18 Using the first three criteria above, determine the required ventilation rate
for the space in Example 5.14.
Solution
Criteria 1 One air change every 4 hours is equivalent to 0.25 air changes per hour.
Since the space volume is 24.5 m3 or 864 ft3, the ventilation rate is 6125 liters/h or
3.6 CFM.
Criteria 2 Since there is one battery string, the required rate is 3398 liters/h or 2
CFM.
Criteria 3 Assuming the space is occupied by one person, the required rate is 33,979
liters/h or 20 CFM.
Example 5.19 Compare the results from the previous five examples.
Solution
Example liters/h ft3/min Remarks

5.14 11,040 6.5 Safety factor = 4 and IMax/4
5.15 49,000 28.8 NFPA 111-2001, 2 air changes/h
5.16 163,105 96 IFC, 1 CFM/ft2 of floor space
5.17 3,960 2.3 Safety factor = 5 and I = 0.02 A/Ah
5.18.1 6,125 3.6 Industry practice, 0.25 air change/h
5.18.2 3,398 2 Industry practice, 2 CFM/string
5.18.3 33,979 20 Industry practice, space occupied
As the previous examples show, the required ventilation rate can vary by a factor of 50
depending on the method used to calculate it.
Ventilation design should include a design margin or redundancy, such as vent louvers
and fans sized twice as large as calculations indicate or using a minimum of two fans. It
may be desirable to have two fans, one that ventilates the room when unoccupied by
workers and another that can be manually turned on when workers are present. Also, in
cold or hot climates, it may be desirable to run the fan intermittently (say every other
hour) to conserve energy required for air conditioning. In this case, the fan should be
sized with the appropriate safety factor (2 to 5) such that the LEL cannot be reached dur-
ing the periods when the fan is off.
A hydrogen detector may be desirable in a battery installation, particularly in large
ones. The following alarm thresholds are commonly used:
앫 Minor alarm: 10% LEL

앫 Major alarm: 25% LEL
Hydrogen detectors require relatively frequent calibration, and some types are easily con-
taminated by ordinary materials used in battery rooms. Three manufacturers of hydrogen
gas detectors that are marketed to the telecommunications industry are Arrgh Manufactur-
ing Co., (www.arrgh.com), General Monitors (www.generalmonitors.com), and RKI In-
struments, Inc. (www.rkiinstruments.com).
5.7.5 Alternating Current Power

The ac distribution in dc power system spaces should include ac circuits for
앫 Rectifier system
앫 Inverter system (only if connected to the commercial ac system as in active or pas-
sive standby operation)
앫 Local lighting
앫 Convenience outlets
앫 Local heating, ventilating, and air conditioning equipment
앫 Other dedicated ac load equipment in the space
The ac distribution in telecommunications facilities (not just dc power system spaces) in-
cludes essential and protected circuits (see Chapter 1 for definitions) and the various ac
loads normally are assigned as shown in Table 5.12.
Table 5.12 Minimum Equipment Load List for ac Essential and Protected Buses
Component Essential Protected
Network elements (switches, transmission, alarm systems) ✓
Network element workstations ✓
Internal telephone systems supporting network elements ✓ ✓
Uninterruptible power systems (UPS) ✓
Waveguide pressurization and dehydrators ✓
Standby engine–generator set auxiliaries (louvers, fuel pumps, battery ✓
chargers, etc.)
Standby engine–generator set fuel transfer pumps ✓
Furnace pumps ✓
Air conditioners serving network equipment ✓
Water pumps for fire suppression ✓
Fire alarm reporting systems ✓
Building security systems (card readers, gate/door opening devices) ✓
Building automation controls ✓
Elevator (one per building bank, check local requirements) ✓
Equipment area lighting ✓
Exit and stairwell lighting ✓
Exterior security lighting ✓
Tower lighting ✓
262 SYSTEM DESIGN
REFERENCES (PART I)
1. ANSI/T1.315-2001, Voltage Levels for DC-Powered Equipment Used in the Telecommunica-

tions Environment, Alliance for Telecommunications Industry Solutions, 2001.
2. Rural Utilities Service Bulletin 1751E-302, Power Requirements for Digital Central Office
Equipment, Rural Utilities Service, Telecommunications Program, United States Department
of Agriculture, 1993.
3. GR-63-CORE, NEBS™ Requirements: Physical Protection, Telcordia, 2002.
4. ANSI/EIA-310-D-1992, Cabinets, Racks, Panels, and Associated Equipment, Electronic Indus-
tries Association, Washington, DC, 1992.
5. ANSI/T1.336-2003, Engineering Requirements for a Universal Telecom Framework, Alliance
for Telecommunications Industry Solutions, 2003.
6. E. L. Grant, et al., Principles of Engineering Economy, Wiley, New York, 1982.
7. T. J. Morgan, Telecommunications Economics, Technology Ltd., Stonehouse, Gloucester, 1976.
8. S. C. Littlechild, Elements of Telecommunications Economics, Peter Peregrinus Ltd., New
York, 1982.
9. Engineering Economy, A Manager’s Guide to Economic Decision Making, American Tele-
phone & Telegraph Co., McGraw-Hill, New York, 1977.
10. B. W. Jones, Inflation in Engineering Economics Analysis, Wiley, New York, 1982.
11. ANSI/T1.328-2001, Protection of Telecommunications Links from Physical Stress and Radia-
tion Effects and Associated Requirements for DC Power Systems (A Baseline Standard), Al-
liance for Telecommunications Industry Solutions, 2001.
12. ANSI/IEEE Std 493™-1997, IEEE Recommended Practice for the Design of Reliable Industri-
al and Commercial Power Systems (IEEE Gold Book), Institute of Electrical and Electronics
Engineers, 1997.
13. International Building Code®, International Code Council®, Falls Church, VA, 2003.
14. SEI/ASCE 7-02, Minimum Design Loads for Buildings and Other Structures, American Soci-
ety of Civil Engineers—Structural Engineering Institute, 2002.
15. ANSI/IEEE Std 946™-1992, IEEE Recommended Practice for the Design of DC Auxiliary
Power Systems for Generating Stations, Institute of Electrical and Electronics Engineers, 1992.
16. NEMA Standards Publication AB 3-2001, Molded Case Circuit Breakers and Their Applica-
tion, National Electrical Manufacturers Association, 2001.
17. Code of Federal Regulations (CFR) Title 47, Part 15, Federal Communications Commission,
2005.
18. NFPA-70-2005, National Electrical Code®, National Fire Protection Association, 2005.
19. International Fire Code®, International Code Council®, Falls Church, VA, 2003.
20. NFPA 76-2005, Recommended Practice for the Fire Protection of Telecommunications Facili-
ties, National Fire Protection Association, 2005.
21. ANSI/T1.330-2002, Valve-regulated Lead-acid Batteries Used in Telecommunications Envi-
ronment, Alliance for Telecommunications Industry Solutions, 2002.
22. ANSI/IEEE Std 484™-2002, IEEE Recommend Practice for Installation Design and Installa-
tion of Vented Lead-Acid Batteries for Stationary Applications, Institute of Electrical and Elec-
tronics Engineers, 2002.
23. ANSI/IEEE Std 1184™-1994, Batteries for Uninterruptible Power Systems, Institute of Elec-
trical and Electronics Engineers, 1994.
24. M. W. Migliaro, Stationary Battery Calculations Associated with Maintenance & Testing, 3rd
ed., ESA Consulting Engineers, Jupiter, FL, 2002.
25. NFPA 111-2001, Standard on Stored Electrical Energy Emergency and Standby Power Sys-
tems, National Fire Protection Association, 2001.
5.8 CIRCUIT DESIGN 263
PART II DIRECT CURRENT POWER SYSTEMS CIRCUIT DESIGN
This part describes guidelines for designing dc circuits and dc power system bonding and
grounding.
5.8 CIRCUIT DESIGN
Basic circuit design procedures include a few simple rules:
앫 Each equipment shelf, chassis, or assembly must be served by one or two dedicated
power circuits, each protected by its own overcurrent protection device.
앫 Circuit conductors must be large enough to safely carry the current.
앫 Circuit conductors must be large enough to keep the voltage drop below the desired
threshold.
5.8.1 General Considerations

5.8.1.1 Circuit Configurations All circuits must have an overcurrent device, and
dedicated circuits are used for all equipment [Fig. 5.33(a)]. Loads do not share circuit
overcurrent protection devices, and daisy chaining is not used [Figs. 5.33(b) and 5.33(c)].
5.8.1.2 Temperature Considerations Temperatures affect both voltage drop and

current rating calculations. For design purposes, the following temperatures commonly
are used:
Voltage drop: 30°C (86°F) or 75°C (167°F)

Current rating:
Cable bus: 75°C (167°F) or 90°C (194°F)
Rigid bus: 70°C (158°F)
The 30°C (86°F) design temperature for voltage drop calculations reflects common prac-
tice in telecommunications and represents the wire temperature when self-heating is
small. The higher values for current rating calculations also reflect common practice.
5.8.1.3 Conductors The basic properties of copper conductors are shown in Table
5.13. Both coarse-strand wire (ASTM B8-04, Class B, often called building wire) and
fine-strand wire (ASTM B172-01a, Class I, often referred to as flexible, welding19 or lo-
comotive cable) are used in telecommunication applications. Other types may be used if
they are properly rated and installed. Fine-strand wire usually is more expensive than
coarse-strand, but installation labor savings in sizes larger than 4/0 AWG usually cancel
the added material cost. Class I fine-strand wire also is used to allow movement during a
seismic event. For example, rigid busbars may be used as the main conductors with Class
19
Even though it has been used on some battery systems and even by battery manufacturers for interconnecting
cables, welding cable typically is not listed by independent electrical testing laboratories, and, when it is, its use
usually is limited to welding machines. If electrical testing laboratory approval is important in a particular appli-
cation or as a matter of company policy, welding cable would not be a good choice.
264 SYSTEM DESIGN
(a)
(b)
(c)
Fig. 5.33 (a) Dedicated circuits each with its own overcurrent device. (b) Circuit overcurrent protec-
tion device sharing and load paralleling not allowed. (c) Circuit daisy chaining not allowed.
I wire for the final connections at the battery and powerboard charge bus. Some installa-
tions use Class B wire in the main run and transition by splice to Class I wire before the
cable is terminated at the battery or rectifier frame.
The stranding of building wire is shown in Table 5.14 for conductors from 18 AWG to
2000 kcmil. The stranding of fine-strand wire used in telecommunications applications
for the range of 8 AWG to 2000 kcmil also is shown in Table 5.14. In telecommunica-
tions applications, fine-strand wire normally is used only in sizes larger than 4/0 AWG,
although some companies use it for all dc power wiring including wires as small as 8
AWG.
An important consideration in the selection of the wire insulation and jacket is the
combustion products when the wire burns. The materials should not support flame and
their smoke should have minimum toxicity and should not be corrosive. Low-smoke insu-
lations and jackets are preferred for central office installations.
A number of different insulation types have been used over the years. Type RHW,
RHW-2, and RHH insulated wires commonly are used in dc power systems as are THW
Table 5.13 Basic Properties of Stranded Copper Conductors (Building Wire)

Size Conductor
DC Resistance (/1000 ft)b
(AWG or Diameter Cross Section
kcmil) (in.) (Circular Mils)a 20°C 30°C 75°C 90°C
(1) (2) (3) (4) (5) (6) (7)
18 — 1,620 6.54 6.79702 7.95362 8.33915
16 — 2,580 4.10 4.26113 4.98622 5.22791
14 0.073 4,110 2.58 2.68139 3.13767 3.28976
12 0.092 6,530 1.63 1.69406 1.98232 2.07841
10 0.116 10,380 1.02 1.06009 1.24047 1.30060
8 0.146 16,510 0.64 0.66515 0.77834 0.81606
6 0.184 26,240 0.403 0.41884 0.49011 0.51387
4 0.232 41,740 0.253 0.26294 0.30769 0.32260
3 0.260 52,620 0.201 0.20890 0.24445 0.25630
2 0.292 66,360 0.159 0.16525 0.19337 0.20274
1 0.332 83,690 0.127 0.13199 0.15445 0.16194
1/0 0.372 105,600 0.100 0.10393 0.12162 0.12751
2/0 0.418 133,100 0.0795 0.08262 0.09668 0.10137
3/0 0.470 167,800 0.0630 0.06548 0.07662 0.08033
4/0 0.528 211,600 0.0500 0.05197 0.06081 0.06376
250 0.575 250 kcmil 0.0423 0.04396 0.05144 0.05394
350 0.681 350 kcmil 0.0302 0.03139 0.03673 0.03851
500 0.813 500 kcmil 0.0212 0.02203 0.02578 0.02703
750 0.998 750 kcmil 0.0141 0.01465 0.01715 0.01798
1,000 1.152 1,000 kcmil 0.0106 0.01102 0.01289 0.01352
1,250 1.289 1,250 kcmil 0.00847 0.00880 0.01030 0.01080
1,500 1.412 1,500 kcmil 0.00705 0.00733 0.00857 0.00899
2,000 1.632 2,000 kcmil 0.00529 0.00550 0.00643 0.00675
a
Equivalent cross section (circular mils) is the sum of the circular mil area of the individual conductor strands.
The overall wire diameter is slightly larger. The equivalent diameter (mils) is the square root of the equivalent
cross section.
b
Resistances at 20°C are from ASTM B-8 [27]; other resistances are adjusted to temperatures shown.
266 SYSTEM DESIGN
and THW-2. However, these types of wire insulations do not always have low smoke and
flame characteristics. Care must be exercised when wire is selected from manufacturers’
catalogs to ensure the correct insulation type (low smoke). Wires with thin insulations,
such as THHW, THWN, and THHN, have been used in some installations, but the associ-
ated slick nylon outer jacket makes the wire more difficult to handle and secure during
and after installation and their smoke products are undesirable.
For purposes of current rating calculations, the wire types are broadly classified as un-
jacketed and jacketed. The thicknesses of the insulations including jackets (if used) for
various wire sizes and insulation types are shown in Table 5.15. Wire current ratings are
shown in Table 5.16 for various installation conditions and physical layouts as follows:
Table 5.16(a) Single copper wire in free air

Table 5.16(b) Alternate current ratings for single copper wire in free air (based on
NEC)
Table 5-16(c) Two bundled copper wires
Table 5.16(d).1 Single unjacketed copper wires with no separation in one or more
layers, 75°C
Table 5.14 Copper Conductor Strand Properties and Weights

Coarse Strand (ASTM B8-04 Class B) Fine Strandb (ASTM B-172 Class I)
______________________________________ _______________________________________
AWG Strand Weight with AWG Strand Weight with
or Diameter Insulation or Diameter Insulation
kcmil Strands (mils) (lb/ft)a kcmil Strands (mils) (lb/ft)a
18 7 15 — — — — —
16 7 19 — — — — —
14 7 24 0.022 — — — —
12 7 30 0.031 — — — —
10 7 38 0.045 — — — —
8 7 49 0.074 8 41 20.1 0.082
6 7 61 0.108 6 63 20.1 0.117
4 7 77 0.161 4 105 20.1 0.175
3 7 87 0.198 3 133 20.1 —
2 7 97 0.244 2 161 20.1 0.253
1 19 66 0.319 1 210 20.1 —
1/0 19 74 0.392 1/0 266 20.1 0.423
2/0 19 84 0.484 2/0 342 20.1 0.538
3/0 19 94 0.599 3/0 418 20.1 —
4/0 19 106 0.743 4/0 532 20.1 0.790
250 37 82 0.889 250 637 — —
350 37 97 1.217 350 882 20.1 1.302
500 37 116 1.703 500 1,225 20.1 1.831
750 61 111 2.540 750 1,862 20.1 2.752
1,000 61 128 3.343 1,000 2,527 20.1 —
1,250 91 117 4.310 1,250 3,059 20.1 —
1,500 91 128 5.110 1,500 3,724 20.1 —
2,000 127 126 6.910 2,000 4,921 20.1 —
a
Weight varies slightly with the type of insulation.
b
Variations exist between different types of fine-strand wire; the table above is based on wire designed for use in
telecommunications central offices.
Table 5.15 Insulated Wire Propertiesa

Insulation Thickness (Mils)b
Conductor
AWG or Diameter RHH, RHW,
kcmil (in.) Hookup THW RHW-2 Jacketed
18 0.046 15–30 — — —
16 0.058 15–30 — — —
14 0.073 15–30 30 45 60
12 0.092 15–30 30 45 60
10 0.116 15–30 30 45 60
8 0.146 — 45 60 90
6 0.184 — 60 60 90
4 0.232 — 60 60 90
3 0.260 — 60 60 90
2 0.292 — 60 60 90
1 0.332 — 80 80 125
1/0 0.372 — 80 80 125
2/0 0.418 — 80 80 125
3/0 0.470 — 80 80 125
4/0 0.528 — 80 80 125
250 0.575 — 95 95 160
350 0.681 — 95 95 160
500 0.813 — 95 95 160
750 0.998 — 110 110 175
1,000 1.152 — 110 110 175
1,250 1.289 — 125 125 220
1,500 1.412 — 125 125 220
2,000 1.632 — 125 125 220
a
Hookup wire 300–600 V rating, all other wire 600 V rating.
b
Outside diameter of insulated wire = Conductor diameter + 2 × Insulation thickness.
Table 5.16(d).2 Single jacketed copper wires with no separation in one or more
layers, 75°C
Table 5.16(d).3 Single unjacketed copper wires with no separation in one or more
layers, 90°C
Table 5.16(d).4 Single jacketed copper wires with no separation in one or more
layers, 90°C
Table 5.16(e).1 Rounded bundle of unjacketed copper wires, 75° C
Table 5.16(e).2 Rounded bundle of jacketed copper wires, 75° C
Table 5.16(e).3 Rounded bundle of unjacketed copper wires, 90° C
Table 5.16(e).4 Rounded bundle of jacketed copper wires, 90° C
The descriptions of the installation conditions, physical layouts, and current rating cal-
culations are provided in Section 5.8.1.6.
Rigid busbars are used where it is necessary to carry large currents over relatively long
distances and typically are found in systems or circuits exceeding approximately 800 to
1000-A capacity. In most applications, rigid busbars are uninsulated (bare), although they
usually are taped or covered with an insulating cover where subject to accidental contact.
Rigid copper busbar physical and electrical properties including current ratings are
268 SYSTEM DESIGN
Table 5.16(a) Calculated Current Ratings in Free Air for Unjacketed and Jacketed Single Copper
Wires at 75°C and 90°C with 30°C Ambient Temperature
Insulation Unjacketed Unjacketed Insulation Jacketed Jacketed
Thickness DC Current DC Current Thickness DC Current DC Current
AWG or Unjacketed Rating Rating Jacketed rating rating
kcmil (in.) 75°C (A) 90°C (A) (in.) 75°C (A) 90°C (A)
(1) (2) (3) (4) (5) (6) (7)
14 0.045 34 39 0.060 34 40
12 0.045 44 51 0.060 45 52
10 0.045 59 68 0.060 60 69
8 0.060 79 92 0.090 81 93
6 0.060 105 122 0.090 107 124
4 0.060 141 164 0.090 143 165
3 0.060 163 190 0.090 165 191
2 0.060 190 220 0.090 191 221
1 0.080 221 256 0.125 222 257
1/0 0.080 257 299 0.125 258 298
2/0 0.080 299 347 0.125 299 346
3/0 0.080 348 404 0.125 347 402
4/0 0.080 406 471 0.125 404 467
250 0.095 453 525 0.160 448 518
350 0.095 566 657 0.160 558 646
500 0.095 718 834 0.160 706 816
750 0.110 943 1,094 0.175 923 1,068
1,000 0.110 1,146 1,330 0.175 1,120 1,295
1,250 0.125 1,330 1,541 0.220 1,284 1,482
1,500 0.125 1,509 1,749 0.220 1,454 1,679
2,000 0.125 1,842 2,135 0.220 1,771 2,045
shown in Table 5.17 (single busbar) and Table 5.18 (parallel busbars). Table 5.18 consists
of:
Table 5-18(a) ¼ in. Parallel Busbars

Table 5-18(b) ⅜ in. Parallel Busbars
Table 5-18(c) ½ in. Parallel Busbars
Table 5-18(d) ¾ in. Parallel Busbars
Table 5.19 provides alternate ratings for some busbar configurations and is based on
ANSI/T1.311 [28]. Refer to table notes for additional details on using these tables. Table
5.19 does not cover as many combinations, and the current ratings generally are higher
than Tables 5.17 and 5.18, particularly for parallel bars. However, Table 5.19 values have
been used successfully in the telecommunications industry for many years. Also, Table
5.19 provides current ratings for the situations where the busbar is mounted in such a way
that its ability to dissipate heat is diminished, such as when the bars are spaced less than
their thickness, run with the long edge horizontal, or run in the vertical direction. When
there is doubt or any discrepancies in the current ratings in this book, the more conserva-
tive (lower) current ratings should be used.
Busbars used in electrical power applications generally are referred to as copper No.
110 busbars, also called electrolytic tough pitch (ETP) or ETP-110 or C11000 (copper
Table 5.16(b) DC Current Ratings in Free Air Calculated from the NEC [29] for Copper Wire at
75°C and 90°C
AWG or Skin Effect/Proximity NEC Free Air dc NEC Free Air DC
kcmil Effect Factor, FSPa Current Rating 75°C (A)b Current Rating 90°C (A)b
(1) (2) (3) (4)
14 1.00 30 35
12 1.00 35 40
10 1.00 50 55
8 1.00 70 80
6 1.00 95 105
4 1.00 125 140
3 1.00 145 165
2 1.00 170 190
1 1.00 195 220
1/0 1.00 230 260
2/0 1.00 265 300
3/0 1.00 310 350
4/0 1.00 360 405
250 1.00 405 455
350 1.00 505 570
500 1.01 626 707
750 1.02 801 903
1,000 1.03 963 1,087
1,250 1.05 1,118 1,260
1,500 1.07 1,257 1,418
2,000 1.11 1,537 1,732
a
From [30].
b
The dc current ratings in columns (3) and (4) are ac ampacities from Table 310.16 of [29] multiplied by the
factors in column (2).
that has been refined electrolytically). The busbar is 99.90% pure copper by weight and
has 101% IACS (International Annealed Copper Standard) electrical conductivity and
10.371 -CM/ft resistivity at 20°C.
5.8.1.4 Parallel Conductors It is frequently necessary to use parallel conductors

to reduce voltage drop, increase current rating, or both (Fig. 5.34). Where wire conductors
are connected in parallel, their length must be the same to ensure current sharing. The
conductors must be the same size and type and same material, and they must be terminat-
ed using the same materials and methods. Where parallel conductors are to be used, they
must be used on both sides of the circuit (feed and return)—in other words, if the feed
uses parallel wires, the return also must use parallel wires.
The resistance of parallel conductors equals the resistance of one conductor divided by
the number of conductors in parallel:
RCond
RP = (5.23)
NCond
where RP = resistance of parallel conductors

RCond = resistance of one of the conductors
NCond = number of conductors connected in parallel
270 SYSTEM DESIGN
Table 5.16(c) Calculated Current Ratings for Two Bundled Unjacketed and Jacketed Copper
Wires at 75°C and 90°C
Insulation Unjacketed Unjacketed Insulation Jacketed Jacketed
Thickness DC Current DC Current Thickness DC Current DC Current
AWG or Unjacketed Rating Rating Jacketed Rating Rating
kcmil (in.) 75°C (A) 90°C (A) (in.) 75°C (A) 90°C (A)
(1) (2) (3) (4) (5) (6) (7)
14 0.045 29 34 0.060 30 35
12 0.045 38 45 0.060 39 46
10 0.045 51 59 0.060 52 61
8 0.060 69 80 0.090 71 83
6 0.060 92 107 0.090 94 110
4 0.060 123 143 0.090 126 146
3 0.060 142 165 0.090 145 168
2 0.060 165 192 0.090 168 195
1 0.080 193 225 0.125 197 228
1/0 0.080 225 262 0.125 228 264
2/0 0.080 261 303 0.125 264 306
3/0 0.080 304 353 0.125 306 355
4/0 0.080 353 411 0.125 355 412
250 0.095 395 459 0.160 397 460
350 0.095 494 574 0.160 494 572
500 0.095 625 726 0.160 622 721
750 0.110 820 953 0.175 814 943
1,000 0.110 995 1,156 0.175 985 1,141
1,250 0.125 1,156 1,343 0.220 1,135 1,314
1,500 0.125 1,310 1,521 0.220 1,284 1,486
2,000 0.125 1,595 1,852 0.220 1,560 1,805
The circular mil area of parallel conductors equals the circular mil area of one of the con-
ductors times the number of conductors in parallel:
CMP = CMCondNCond (5.24)
where CMP = circular mil area of the parallel conductors

CMCond = circular mil area of one of the conductors
NCond = number of conductors connected in parallel
It is not normal practice to use parallel conductors in sizes smaller than 1/0 AWG, but
there are many exceptions to this practice, particularly in smaller systems. However, there
is no reason to parallel conductors smaller than 2 AWG even in small systems. Parallel
conductors smaller than 1/0 AWG are not allowed in installations covered by the Nation-
al Electrical Code.20
5.8.1.5 Voltage Drop Calculations The dc voltage drop calculations are some-
what simpler than ac voltage drop calculations because with dc there is no skin effect or
reactance due to operating frequency. Two methods are used:
앫 The first method is based on a simple factor (voltage drop factor, sometimes called
20
2005 NEC, Article 310.4. [29].
Table 5.16(d).1 Current Ratings for Unjacketed Single (1/C) Copper Wires at 75°C with No Separation in One or More Layers—30°C Ambient Temperaturea
Conductor dc Resistance Free Air Current Rating Unjacketed 75°C (A)
AWG Cross Wire 75°Cc DC Current
Apparent (Calculated) Wire Mass Fill Depth
or Section Diameterb ( per Ratingd
kcmil (CM) (in.) 1000 ft) 75°C (A) 1 in. 1.5 in. 2 in. 2.5 in. 3 in. 3.5 in. 4 in.
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
14 4,110 0.163 3.13767 34 6.5 5.1 4.2 3.6 3.2 2.8 2.6
12 6,530 0.182 1.98232 44 9.1 7.1 5.9 5.1 4.4 4.0 3.6
10 10,380 0.206 1.24047 59 13 10 8.4 7.2 6.4 5.7 5.2
8 16,510 0.266 0.77834 79 21 17 14 12 10 9.3 8.4
6 26,240 0.304 0.49011 105 31 24 20 17 15 13 12
4 41,740 0.352 0.30769 141 45 35 29 25 22 20 18
3 52,620 0.380 0.24445 163 54 42 35 30 26 24 21
2 66,360 0.412 0.19337 190 66 51 43 37 32 29 26
1 83,690 0.492 0.15445 221 88 69 57 49 43 39 35
1/0 105,600 0.532 0.12162 257 108 84 69 60 52 47 43
2/0 133,100 0.578 0.09668 299 131 102 85 73 64 57 52
3/0 167,800 0.630 0.07662 348 161 125 104 89 78 70 63
4/0 211,600 0.688 0.06081 406 197 153 127 109 96 86 78
250 250,000 0.765 0.05144 453 238 185 153 132 116 104 94
350 350,000 0.871 0.03673 566 321 250 207 178 156 140 127
500 500,000 1.003 0.02578 718 442 343 284 244 215 192 174
750 750,000 1.218 0.01715 943 657 511 423 363 320 286 259
1,000 1,000,000 1.372 0.01289 1,146 854 664 550 472 415 372 337
1,250 1,250,000 1.539 0.01030 1,330 1,072 833 690 593 521 467 423
1,500 1,500,000 1.662 0.00857 1,509 1,269 986 817 702 617 552 501
2,000 2,000,000 1.882 0.00643 1,842 1,659 1,289 1,068 917 807 722 654
a
No diversity. Free air current rating is shown for reference.
b
From NFPA 70-2005 (NEC®) Table 8 and Table 310.13 [28, nec]
c
dc resistance from ASTM B8-04 [27, ASTM B8], adjusted to 75°C
d
From Table 5.16(a)
271
Table 5.16(d).2 Current Ratings for Jacketed Single Copper Wires (1/C) at 75°C with No Separation in One or More Layers—30°C Ambient Temperaturea
272
Conductor DC Resistance Free Air DC Current Rating Jacketed 75°C (A)
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
14 4,110 0.193 3.13767 34 7.7 6.0 5.0 4.3 3.7 3.4 3.0
12 6,530 0.212 1.98232 45 11 8.3 6.8 5.9 5.2 4.6 4.2
10 10,380 0.236 1.24047 60 15 12 9.6 8.3 7.3 6.5 5.9
8 16,510 0.326 0.77834 81 26 20 17 14 13 11 10
6 26,240 0.364 0.49011 107 37 29 24 20 18 16 15
4 41,740 0.412 0.30769 143 53 41 34 29 26 23 21
3 52,620 0.440 0.24445 165 63 49 41 35 31 27 25
2 66,360 0.472 0.19337 191 76 59 49 42 37 33 30
1 83,690 0.582 0.15445 222 105 81 67 58 51 46 41
1/0 105,600 0.622 0.12162 258 126 98 81 70 61 55 50
2/0 133,100 0.668 0.09668 299 152 118 98 84 74 66 60
3/0 167,800 0.720 0.07662 347 184 143 118 102 89 80 73
4/0 211,600 0.778 0.06081 404 223 173 144 123 108 97 88
250 250,000 0.895 0.05144 448 279 217 180 154 136 121 110
350 350,000 1.001 0.03673 558 369 287 238 204 180 161 146
500 500,000 1.133 0.02578 706 499 388 321 276 243 217 197
750 750,000 1.348 0.01715 923 728 565 468 402 354 317 287
1,000 1,000,000 1.502 0.01289 1,120 935 727 602 517 455 407 369
1,250 1,250,000 1.729 0.01030 1,284 1,204 936 775 666 586 524 475
1,500 1,500,000 1.852 0.00857 1,454 1,414 1,099 910 782 688 615 558
2,000 2,000,000 2.072 0.00643 1,771 1,826e 1,419 1,175 1,010 888 795 720
a
b
From NFPA 70-2005 (NEC) Table 8 and Table 310.13 [28].
c
dc resistance from ASTM B8-04 [27, ASTM B8], adjusted to 75°C.
d
From Table 5.16(a).
e
Use free air current rating.
Table 5.16(d).3 Current Ratings for Unjacketed Single Copper Wires (1/C) at 90°C with No Separation in One or More Layers—30°C Ambient Temperaturea
Conductor DC Resistance Free Air DC Current Rating Unjacketed 90°C (A)
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
14 4,110 0.163 3.28976 39 7.5 5.8 4.8 4.1 3.6 3.2 2.9
12 6,530 0.182 2.07841 51 11 8.2 6.7 5.8 5.1 4.5 4.1
10 10,380 0.206 1.30060 68 15 12 9.7 8.3 7.3 6.5 5.9
8 16,510 0.266 0.81606 92 25 19 16 14 12 11 9.6
6 26,240 0.304 0.51387 122 35 27 23 20 17 15 14
4 41,740 0.352 0.32260 164 52 40 33 28 25 22 20
3 52,620 0.380 0.25630 190 63 49 40 34 30 27 24
2 66,360 0.412 0.20274 220 77 59 49 42 37 33 30
1 83,690 0.492 0.16194 256 102 79 65 56 49 44 40
1/0 105,600 0.532 0.12751 299 125 96 80 68 60 54 49
2/0 133,100 0.578 0.10137 347 152 118 97 83 73 65 59
3/0 167,800 0.630 0.08033 404 186 144 119 102 90 80 72
4/0 211,600 0.688 0.06376 471 228 176 146 125 110 98 89
250 250,000 0.765 0.05394 525 276 213 176 151 133 119 107
350 350,000 0.871 0.03851 657 371 287 238 204 179 160 145
500 500,000 1.003 0.02703 834 510 395 326 280 246 220 199
750 750,000 1.218 0.01798 1,094 760 588 486 417 366 327 296
1,000 1,000,000 1.372 0.01352 1,330 987 764 631 541 475 425 385
1,250 1,250,000 1.539 0.01080 1,541 1,239 959 792 679 597 533 483
1,500 1,500,000 1.662 0.00899 1,749 1,466 1,135 938 804 706 631 571
2,000 2,000,000 1.882 0.00675 2,135 1,916 1,484 1,226 1,051 923 825 747
a
b
c
dc resistance from ASTM B8-04 [27], adjusted to 90°C.
d
From Table 5.16(a).
273
Table 5.16(d).4 Current Ratings for Jacketed Single Copper Wires (1/C) at 90°C with No Separation in One or More Layers—30°C Ambient Temperaturea
274
Conductor DC Resistance Free Air DC Current Rating Jacketed 90°C (A)
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
14 4,110 0.193 3.28976 40 8.9 6.9 5.7 4.9 4.3 3.8 3.5
12 6,530 0.212 2.07841 52 12 9.5 7.9 6.7 5.9 5.3 4.8
10 10,380 0.236 1.30060 69 17 13 11 9.5 8.3 7.5 6.7
8 16,510 0.326 0.81606 93 30 23 19 17 15 13 12
6 26,240 0.364 0.51387 124 42 33 27 23 20 18 17
4 41,740 0.412 0.32260 165 61 47 39 33 29 26 24
3 52,620 0.440 0.25630 191 73 56 47 40 35 31 28
2 66,360 0.472 0.20274 221 88 68 56 48 42 38 34
1 83,690 0.582 0.16194 257 121 94 77 66 58 52 47
1/0 105,600 0.622 0.12751 298 146 113 93 80 70 63 57
2/0 133,100 0.668 0.10137 346 175 136 112 96 85 76 68
3/0 167,800 0.720 0.08033 402 212 165 136 117 102 91 83
4/0 211,600 0.778 0.06376 467 258 200 165 141 124 111 100
250 250,000 0.895 0.05394 518 322 250 206 177 155 139 126
350 350,000 1.001 0.03851 646 427 330 273 234 206 184 166
500 500,000 1.133 0.02703 816 576 446 369 316 278 248 225
750 750,000 1.348 0.01798 1,068 841 651 538 461 405 362 328
1,000 1,000,000 1.502 0.01352 1,295 1,081 837 691 593 521 465 421
1,250 1,250,000 1.729 0.01080 1,482 1,392 1,078 890 763 670 599 542
1,500 1,500,000 1.852 0.00899 1,679 1,634 1,265 1,045 896 787 703 637
2,000 2,000,000 2.072 0.00675 2,045 2,110e 1,633 1,350 1,157 1,016 908 822
a
b
c
dc resistance from ASTM B8-04 [27], adjusted to 90°C.
d
From Table 5.16(a).
e
Use free air current rating.
Table 5.16(e).1 Calculated Current Ratings for Rounded Bundles of 14–4/0 AWG Unjacketed
Copper Wiresa
Insulation DC Current Rating Unjacketed 75°C (A)
Thickness
Number of Unjacketed Wires in Bundle
AWG Unjacketed
or kcmil (in.) 4 6 8 10 12 14 16
(1) (2) (3) (4) (5) (6) (7) (8) (9)
14 0.045 22 19 18 16 15 14 14
12 0.045 29 25 23 21 20 19 18
10 0.045 38 33 30 28 26 25 23
8 0.060 52 45 41 38 35 33 32
6 0.060 68 59 53 49 46 43 41
4 0.060 90 78 70 65 60 57 54
3 0.060 103 89 80 7 69 65 62
2 0.060 119 102 92 85 79 75 71
1 0.080 139 120 108 99 93 88 83
1/0 0.080 160 138 124 114 107 101 96
2/0 0.080 184 158 142 131 122 115 109
3/0 0.080 211 182 163 150 140 132 125
4/0 0.080 243 209 187 172 160 151 143
a
Maximum temperature 75°C with ambient temperature 30°C. Applies only to bundles in which all wires are the
same size.
Table 5.16(e).2 Calculated Current Ratings for Rounded Bundles of 14–4/0 AWG Jacketed
Copper Wiresa
Insulation DC Current Rating Jacketed 75°C (A)
Thickness
Number of Jacketed Wires in Bundle
AWG Jacketed
or kcmil (in.) 4 6 8 10 12 14 16
(1) (2) (3) (4) (5) (6) (7) (8) (9)
14 0.060 24 20 18 17 16 15 14
12 0.060 31 26 24 22 21 20 19
10 0.060 40 35 31 29 27 26 24
8 0.090 55 48 43 40 37 35 33
6 0.090 72 62 56 52 48 46 43
4 0.090 94 81 73 67 63 59 56
3 0.090 107 93 83 77 72 68 64
2 0.090 123 106 96 88 82 77 74
1 0.125 146 126 113 104 97 91 87
1/0 0.125 167 144 129 119 111 104 99
2/0 0.125 191 164 147 135 126 119 113
3/0 0.125 219 188 168 155 144 136 129
4/0 0.125 251 215 193 177 164 155 147
a
same size.
276 SYSTEM DESIGN
Table 5.16(e).3 Calculated Current Ratings for Rounded Bundles of 14–4/0 AWG Unjacketed
Copper Wiresa
Insulation DC Current Rating Unjacketed 90°C (A)
Thickness
Number of Unjacketed Wires in Bundle
AWG Unjacketed
or kcmil (in.) 4 6 8 10 12 14 16
(1) (2) (3) (4) (5) (6) (7) (8) (9)
14 0.045 26 23 20 19 18 17 16
12 0.045 34 29 27 24 23 22 21
10 0.045 44 38 35 32 30 28 27
8 0.060 60 52 47 43 41 38 37
6 0.060 79 68 62 57 53 50 48
4 0.060 104 90 81 75 70 66 63
3 0.060 119 103 93 85 80 75 72
2 0.060 137 118 106 98 91 86 82
1 0.080 161 139 125 114 107 101 96
1/0 0.080 185 159 143 131 123 116 110
2/0 0.080 212 183 164 150 140 132 125
3/0 0.080 244 209 188 172 161 151 143
4/0 0.080 280 240 215 197 184 173 164
a
same size.
Table 5.16(e).4 Calculated Current Ratings for Rounded Bundles of 14–4/0 AWG Jacketed
Copper Wiresa
Insulation DC Current Rating Jacketed 90°C (A)
Thickness
Number of Jacketed Wires in Bundle
AWG Jacketed
or kcmil (in.) 4 6 8 10 12 14 16
(1) (2) (3) (4) (5) (6) (7) (8) (9)
14 0.060 27 24 21 20 19 18 17
12 0.060 35 31 28 26 24 23 22
10 0.060 46 40 36 33 31 29 28
8 0.090 64 55 50 46 43 41 39
6 0.090 83 72 65 60 56 53 50
4 0.090 109 94 84 78 72 68 65
3 0.090 124 107 96 89 83 78 74
2 0.090 142 123 110 101 95 89 85
1 0.125 168 145 130 119 111 105 99
1/0 0.125 193 166 149 136 127 120 114
2/0 0.125 220 189 69 155 145 136 129
3/0 0.125 252 216 194 178 165 156 148
4/0 0.125 289 247 221 203 189 177 168
a
same size.
Table 5.17 Rigid Copper Bus Properties—1 Bara,b
DC Current
Dimensions DC Resistance Rating
_____________________________________ __________________ ________
Thickness Width Area Weight 30°C 70°C 70°C
(in.) (in.) (in.2) (kcmil) (lb/ft) (/ft) (/ft) (A)
1– 1
4
–
2 0.125 159 0.483 68.39 81.82 241
1– –3
4 4 0.188 239 0.726 45.52 54.47 322
1–
4 1 0.250 318 0.966 34.19 40.91 404
1–
4 11–2 0.375 477 1.450 22.76 27.23 568
1–
4 2 0.500 637 1.930 17.15 20.52 724
1–
4 21–2 0.625 796 2.410 13.72 16.41 875
1–
4 3 0.750 955 2.900 11.43 13.68 1,029
1–
4 31–2 0.875 1,114 3.380 9.77 11.69 1,206
1–
4 4 1.000 1,273 3.860 8.55 10.23 1,323
1–
4 5 1.250 1,592 4.830 6.84 8.18 1,616
1–
4 6 1.500 1,910 5.800 5.71 6.83 1,901
1–
4 8 2.000 2,546 7.730 4.27 5.11 2,495
1–
4 10 2.500 3,183 9.660 3.42 4.09 3,043
1–
4 12 3.000 3,820 11.600 2.85 3.41 3,605
3– –3
8 4 0.281 358 1.090 30.45 36.44 419
3–
8 1 0.375 477 1.450 22.76 27.23 518
3–
8 11–2 0.563 716 2.170 15.17 18.16 728
3–
8 2 0.750 955 2.900 11.43 13.68 915
3–
8 21–2 0.938 1,194 3.620 9.11 10.91 1,111
3–
8 3 1.125 1,432 4.350 7.64 9.14 1,287
3–
8 31–2 1.313 1,671 5.060 6.53 7.81 1,466
3–
8 4 1.500 1,910 5.800 5.71 6.83 1,643
3–
8 5 1.875 2,387 7.260 4.55 5.45 2,004
3–
8 6 2.250 2,865 8.690 3.80 4.55 2,367
3–
8 8 3.000 3,820 11.600 2.85 3.41 3,056
3–
8 10 3.750 4,775 14.500 2.28 2.72 3,759
3–
8 12 4.500 5,730 17.400 1.90 2.28 4,409
1–
2 1 0.500 637 1.930 17.15 20.52 632
1–
2 11–2 0.750 955 2.900 11.43 13.68 863
1–
2 2 1.000 1,273 3.860 8.55 10.23 1,058
1–
2 21–2 1.250 1,592 4.830 6.84 8.18 1,292
1–
2 3 1.500 1,910 5.800 5.71 6.83 1,534
1–
2 31–2 1.750 2,228 6.760 4.88 5.84 1,726
1–
2 4 2.000 2,546 7.730 4.27 5.11 1,908
1–
2 5 2.500 3,183 9.660 3.42 4.09 2,355
1–
2 6 3.000 3,820 11.600 2.85 3.41 2,799
1–
2 8 4.000 5,093 15.500 2.14 2.56 3,575
1–
2 10 5.000 6,366 19.300 1.71 2.05 4,365
1–
2 12 6.000 7,639 23.200 1.42 1.70 5,178
–3 4 3.000 3,820 11.600 2.85 3.41 2,443
4
–3 5 3.750 4,775 14.500 2.28 2.72 2,920
4
–3 6 4.500 5,730 17.400 1.90 2.28 3,452
4
–3 8 6.000 7,639 23.200 1.42 1.70 4,427
4
–3 10 7.500 9,549 29.000 1.14 1.37 5,428
4
–3 12 9.000 11,459 34.800 0.95 1.14 6,426
4
a
1 /ft = 1 × 10–6 /ft. For example, 68.39 /ft = 68.39 × 10–6 /ft.
b
Current ratings apply when the busbar is oriented with its long edge vertical, and bar spacing bar thickness,
and busbar is run in the horizontal plane. Current ratings based on 30°C rise above 40°C ambient temperature.
Source: Data from www.copper.org (Copper Development Association, Inc.).
277
278 SYSTEM DESIGN
Table 5.18(a) Rigid Copper Busbar Properties—1–4-in. Parallel Bars as Indicateda,b

DC Current
_____________________________________ __________________ ________
(in.) (in.) (in.2) (kcmil) (lb/ft) /ft /ft (A)
2 Parallel Bars
1– 1
4
–
2 0.250 318 0.966 34.195 40.910 434
1– –3
4 4 0.375 477 1.452 22.760 27.235 579
1–
4 1 0.500 637 1.932 17.095 20.455 727
1–
4 11–2 0.750 955 2.900 11.380 13.615 1,023
1–
4 2 1.000 1,273 3.860 8.575 10.260 1,303
1–
4 21–2 1.250 1,592 4.820 6.860 8.205 1,575
1–
4 3 1.500 1,910 5.800 5.715 6.840 1,852
1–
4 31–2 1.750 2,228 6.760 4.885 5.845 2,171
1–
4 4 2.000 2,546 7.720 4.275 5.115 2,381
1–
4 5 2.500 3,183 9.660 3.420 4.090 2,908
1–
4 6 3.000 3,820 11.600 2.855 3.415 3,422
1–
4 8 4.000 5,093 15.460 2.135 2.555 4,492
1–
4 10 5.000 6,366 19.320 1.710 2.045 5,477
1–
4 12 6.000 7,639 23.200 1.425 1.705 6,490
3 Parallel Bars
1– 1
4
–
2 0.375 477 1.449 22.797 27.273 603
1– –3
4 4 0.563 716 2.178 15.173 18.157 804
1–
4 1 0.750 955 2.898 11.397 13.637 1,010
1–
4 11–2 1.125 1,432 4.350 7.587 9.077 1,421
1–
4 2 1.500 1,910 5.790 5.717 6.840 1,810
1–
4 21–2 1.875 2,387 7.230 4.573 5.470 2,188
1–
4 3 2.250 2,865 8.700 3.810 4.560 2,572
1–
4 31–2 2.625 3,342 10.140 3.257 3.897 3,015
1–
4 4 3.000 3,820 11.580 2.850 3.410 3,307
1–
4 5 3.750 4,775 14.490 2.280 2.727 4,039
1–
4 6 4.500 5,730 17.400 1.903 2.277 4,752
1–
4 8 6.000 7,639 23.190 1.423 1.703 6,238
1–
4 10 7.500 9,549 28.980 1.140 1.363 7,607
1–
4 12 9.000 11,459 34.800 0.950 1.137 9,013
4 Parallel Bars
1– 1
4
–
2 0.500 637 1.932 17.098 20.455 772
1– –3
4 4 0.750 955 2.904 11.380 13.618 1,029
1–
4 1 1.000 1,273 3.864 8.548 10.228 1,293
1–
4 11–2 1.500 1,910 5.800 5.690 6.808 1,819
1–
4 2 2.000 2,546 7.720 4.288 5.130 2,317
1–
4 21–2 2.500 3,183 9.640 3.430 4.103 2,800
1–
4 3 3.000 3,820 11.600 2.858 3.420 3,292
1–
4 31–2 3.500 4,456 13.520 2.443 2.923 3,860
1–
4 4 4.000 5,093 15.440 2.138 2.558 4,233
1–
4 5 5.000 6,366 19.320 1.710 2.045 5,170
1–
4 6 6.000 7,639 23.200 1.428 1.708 6,083
1–
4 8 8.000 10,186 30.920 1.068 1.278 7,985
1–
4 10 10.000 12,732 38.640 0.855 1.023 9,737
1–
4 12 12.000 15,279 46.400 0.713 0.853 11,537
Table 5.18(a) Rigid Copper Busbar Properties—1–4-in. Parallel Bars as Indicateda,b (Continued)
DC Current
_____________________________________ __________________ ________
5 Parallel Bars
1– 1
4
–
2 0.625 796 2.415 13.678 16.364 941
1– –3
4 4 0.938 1,194 3.630 9.104 10.894 1,254
1–
4 1 1.250 1,592 4.830 6.838 8.182 1,576
1–
4 11–2 1.875 2,387 7.250 4.552 5.446 2,217
1–
4 2 2.500 3,183 9.650 3.430 4.104 2,824
1–
4 21–2 3.125 3,979 12.050 2.744 3.282 3,413
1–
4 3 3.750 4,775 14.500 2.286 2.736 4,012
1–
4 31–2 4.375 5,570 16.900 1.954 2.338 4,704
1–
4 4 5.000 6,366 19.300 1.710 2.046 5,159
1–
4 5 6.250 7,958 24.150 1.368 1.636 6,301
1–
4 6 7.500 9,549 29.000 1.142 1.366 7,414
1–
4 8 10.000 12,732 38.650 0.854 1.022 9,732
1–
4 10 12.500 15,915 48.300 0.684 0.818 11,867
1–
4 12 15.000 19,099 58.000 0.570 0.682 14,061
6 Parallel Bars
1– 1
4
–
2 0.750 955 2.898 11.398 13.637 1,061
1– –3
4 4 1.125 1,432 4.356 7.587 9.078 1,415
1–
4 1 1.500 1,910 5.796 5.698 6.818 1,778
1–
4 11–2 2.250 2,865 8.700 3.793 4.538 2,501
1–
4 2 3.000 3,820 11.580 2.858 3.420 3,186
1–
4 21–2 3.750 4,775 14.460 2.287 2.735 3,851
1–
4 3 4.500 5,730 17.400 1.905 2.280 4,527
1–
4 31–2 5.250 6,685 20.280 1.628 1.948 5,307
1–
4 4 6.000 7,639 23.160 1.425 1.705 5,821
1–
4 5 7.500 9,549 28.980 1.140 1.363 7,108
1–
4 6 9.000 11,459 34.800 0.952 1.138 8,364
1–
4 8 12.000 15,279 46.380 0.712 0.852 10,980
1–
4 10 15.000 19,099 57.960 0.570 0.682 13,388
1–
4 12 18.000 22,918 69.600 0.475 0.568 15,863
7 Parallel Bars
1– 1
4
–
2 0.875 1,114 3.381 9.770 11.689 1,206
1– –3
4 4 1.313 1,671 5.082 6.503 7.781 1,608
1–
4 1 1.750 2,228 6.762 4.884 5.844 2,020
1–
4 11–2 2.625 3,342 10.150 3.251 3.890 2,842
1–
4 2 3.500 4,456 13.510 2.450 2.931 3,620
1–
4 21–2 4.375 5,570 16.870 1.960 2.344 4,376
1–
4 3 5.250 6,685 20.300 1.633 1.954 5,144
1–
4 31–2 6.125 7,799 23.660 1.396 1.670 6,031
1–
4 4 7.000 8,913 27.020 1.221 1.461 6,614
1–
4 5 8.750 11,141 33.810 0.977 1.169 8,078
1–
4 6 10.500 13,369 40.600 0.816 0.976 9,505
1–
4 8 14.000 17,825 54.110 0.610 0.730 12,477
1–
4 10 17.500 22,282 67.620 0.489 0.584 15,214
1–
4 12 21.000 26,738 81.200 0.407 0.487 18,027
(continued)
280 SYSTEM DESIGN
DC Current
_____________________________________ __________________ ________
8 Parallel Bars
1– 1
4
–
2 1.000 1,273 3.864 8.549 10.228 1,327
1– –3
4 4 1.500 1,910 5.808 5.690 6.809 1,769
1–
4 1 2.000 2,546 7.728 4.274 5.114 2,222
1–
4 11–2 3.000 3,820 11.600 2.845 3.404 3,126
1–
4 2 4.000 5,093 15.440 2.144 2.565 3,982
1–
4 21–2 5.000 6,366 19.280 1.715 2.051 4,813
1–
4 3 6.000 7,639 23.200 1.429 1.710 5,659
1–
4 31–2 7.000 8,913 27.040 1.221 1.461 6,634
1–
4 4 8.000 10,186 30.880 1.069 1.279 7,276
1–
4 5 10.000 12,732 38.640 0.855 1.023 8,886
1–
4 6 12.000 15,279 46.400 0.714 0.854 10,455
1–
4 8 16.000 20,372 61.840 0.534 0.639 13,725
1–
4 10 20.000 25,465 77.280 0.428 0.511 16,735
1–
4 12 24.000 30,558 92.800 0.356 0.426 19,829
9 Parallel Bars
1– 1
4
–
2 1.125 1,432 4.347 7.599 9.091 1,447
1– –3
4 4 1.688 2,149 6.534 5.058 6.052 1,930
1–
1 4 2.250 2,865 8.694 3.799 4.546 2,424
1–
11–24 3.375 4,297 13.050 2.529 3.026 3,410
1–
2 4 4.500 5,730 17.370 1.906 2.280 4,344
1–
21–24 5.625 7,162 21.690 1.524 1.823 5,251
1–
3 4 6.750 8,594 26.100 1.270 1.520 6,173
1–
31–24 7.875 10,027 30.420 1.086 1.299 7,237
1–
4 4 9.000 11,459 34.740 0.950 1.137 7,937
1–
5 4 11.250 14,324 43.470 0.760 0.909 9,693
1–
6 4 13.500 17,189 52.200 0.634 0.759 11,406
1–
8 4 18.000 22,918 69.570 0.474 0.568 14,972
1–
10 4 22.500 28,648 86.940 0.380 0.454 18,256
1–
12 4 27.000 34,377 104.400 0.317 0.379 21,632
10 Parallel Bars
1– 1
4
–
2 1.250 1,592 4.830 6.839 8.182 1,568
1– –3
4 4 1.875 2,387 7.260 4.552 5.447 2,090
1–
4 1 2.500 3,183 9.660 3.419 4.091 2,626
1–
4 11–2 3.750 4,775 14.500 2.276 2.723 3,694
1–
4 2 5.000 6,366 19.300 1.715 2.052 4,706
1–
4 21–2 6.250 7,958 24.100 1.372 1.641 5,688
1–
4 3 7.500 9,549 29.000 1.143 1.368 6,687
1–
4 31–2 8.750 11,141 33.800 0.977 1.169 7,840
1–
4 4 10.000 12,732 38.600 0.855 1.023 8,599
1–
4 5 12.500 15,915 48.300 0.684 0.818 10,501
1–
4 6 15.000 19,099 58.000 0.571 0.683 12,356
1–
4 8 20.000 25,465 77.300 0.427 0.511 16,220
1–
4 10 25.000 31,831 96.600 0.342 0.409 19,778
1–
4 12 30.000 38,197 116.000 0.285 0.341 23,435
DC Current
_____________________________________ __________________ ________
11 Parallel Bars
1– 1
4
–
2 1.375 1,751 5.313 6.217 7.438 1,664
1– –3
4 4 2.063 2,626 7.986 4.138 4.952 2,219
1–
1 4 2.750 3,501 10.626 3.108 3.719 2,787
1–
11–24 4.125 5,252 15.950 2.069 2.475 3,922
1–
2 4 5.500 7,003 21.230 1.559 1.865 4,996
1–
21–24 6.875 8,754 26.510 1.247 1.492 6,038
1–
3 4 8.250 10,504 31.900 1.039 1.244 7,099
1–
31–24 9.625 12,255 37.180 0.888 1.063 8,322
1–
4 4 11.000 14,006 42.460 0.777 0.930 9,128
1–
5 4 13.750 17,507 53.130 0.622 0.744 11,147
1–
6 4 16.500 21,008 63.800 0.519 0.621 13,117
1–
8 4 22.000 28,011 85.030 0.388 0.465 17,218
1–
10 4 27.500 35,014 106.260 0.311 0.372 20,995
1–
12 4 33.000 42,017 127.600 0.259 0.310 24,877
12 Parallel Bars
1– 1
4
–
2 1.500 1,910 5.796 5.699 6.818 1,761
1– –3
4 4 2.250 2,865 8.712 3.793 4.539 2,348
1–
4 1 3.000 3,820 11.592 2.849 3.409 2,949
1–
4 11–2 4.500 5,730 17.400 1.897 2.269 4,149
1–
4 2 6.000 7,639 23.160 1.429 1.710 5,286
1–
4 21–2 7.500 9,549 28.920 1.143 1.368 6,388
1–
4 3 9.000 11,459 34.800 0.953 1.140 7,511
1–
4 31–2 10.500 13,369 40.560 0.814 0.974 8,805
1–
4 4 12.000 15,279 46.320 0.713 0.853 9,657
1–
4 5 15.000 19,099 57.960 0.570 0.682 11,794
1–
4 6 18.000 22,918 69.600 0.476 0.569 13,877
1–
4 8 24.000 30,558 92.760 0.356 0.426 18,216
1–
4 10 30.000 38,197 115.920 0.285 0.341 22,212
1–
4 12 36.000 45,837 139.200 0.238 0.284 26,319
a
b
Source: Data from www.copper.org (Copper Development Association, Inc.) with adjustments.
K factor) to arrive at the circular mil area required to keep voltage drop below some
specified threshold at a given temperature.
앫 The second method is based on a conventional application of Ohm’s law.
Both methods yield equivalent results, and both methods require knowledge of the con-
ductor length as determined from the conductor path distance. The path distance is the
one-way distance from termination to termination, including vertical and horizontal runs,
as shown in Figure 5.35. Since dc power conductors always are installed in a paired
282 SYSTEM DESIGN
Table 5.18(b) Rigid Copper Busbar Properties—38–-in. Parallel Bars as Indicateda,b

DC Current
_____________________________________ __________________ ________
2 Parallel Bars
3– –3
8 4 0.5625 716 2.18 15.225 18.220 754
3–
8 1 0.750 955 2.90 11.380 13.615 932
3–
8 11–2 1.125 1,432 4.34 7.585 9.080 1,310
3–
8 2 1.500 1,910 5.80 5.715 6.840 1,646
3–
8 21–2 1.875 2,387 7.24 4.555 5.455 2,000
3–
8 3 2.250 2,865 8.70 3.820 4.570 2,316
3–
8 31–2 2.625 3,342 10.12 3.265 3.905 2,640
3–
8 4 3.000 3,820 11.60 2.855 3.415 2,958
3–
8 5 3.750 4,775 14.52 2.275 2.725 3,608
3–
8 6 4.500 5,730 17.38 1.900 2.275 4,260
3–
88 6.000 7,639 23.20 1.425 1.705 5,501
3–
10
8 7.500 9,549 29.00 1.140 1.360 6,766
3–
128 9.000 11,459 34.80 0.950 1.140 7,936
3 Parallel Bars
3– –3
8 4 0.844 1,074 3.27 10.150 12.147 1,048
3–
8 1 1.125 1,432 4.35 7.587 9.077 1,294
3–
8 11–2 1.688 2,149 6.51 5.057 6.053 1,819
3–
8 2 2.250 2,865 8.70 3.810 4.560 2,286
3–
8 21–2 2.813 3,581 10.86 3.037 3.637 2,778
3–
8 3 3.375 4,297 13.05 2.547 3.047 3,217
3–
8 31–2 3.938 5,013 15.18 2.177 2.603 3,666
3–
8 4 4.500 5,730 17.40 1.903 2.277 4,108
3–
8 5 5.625 7,162 21.78 1.517 1.817 5,011
3–
8 6 6.750 8,594 26.07 1.267 1.517 5,916
3–
88 9.000 11,459 34.80 0.950 1.137 7,640
3–
10
8 11.250 14,324 43.50 0.760 0.907 9,398
3–
128 13.500 17,189 52.20 0.633 0.760 11,023
4 Parallel Bars
3– –3
8 4 1.125 1,432 4.36 7.613 9.110 1,341
3–
8 1 1.500 1,910 5.80 5.690 6.808 1,656
3–
8 11–2 2.250 2,865 8.68 3.793 4.540 2,328
3–
8 2 3.000 3,820 11.60 2.858 3.420 2,926
3–
8 21–2 3.750 4,775 14.48 2.278 2.728 3,556
3–
8 3 4.500 5,730 17.40 1.910 2.285 4,118
3–
8 31–2 5.250 6,685 20.24 1.633 1.953 4,693
3–
8 4 6.000 7,639 23.20 1.428 1.708 5,258
3–
8 5 7.500 9,549 29.04 1.138 1.363 6,414
3–
8 6 9.000 11,459 34.76 0.950 1.138 7,573
3–
88 12.000 15,279 46.40 0.713 0.853 9,780
3–
10
8 15.000 19,099 58.00 0.570 0.680 12,029
3–
128 18.000 22,918 69.60 0.475 0.570 14,109
5 Parallel Bars
3– –3
8 4 1.406 1,790 5.45 6.090 7.288 1,635
3–
8 1 1.875 2,387 7.25 4.552 5.446 2,019
3–
8 11–2 2.813 3,581 10.85 3.034 3.632 2,837
Table 5.18(b) Rigid Copper Busbar Properties—3–8-in. Parallel Bars as Indicateda,b (Continued)
DC Current
_____________________________________ __________________ ________
5 Parallel Bars
3–
8 2 3.750 4,775 14.50 2.286 2.736 3,567
3–
8 21–2 4.688 5,968 18.10 1.822 2.182 4,334
3–
8 3 5.625 7,162 21.75 1.528 1.828 5,019
3–
8 31–2 6.563 8,356 25.30 1.306 1.562 5,719
3–
8 4 7.500 9,549 29.00 1.142 1.366 6,408
3–
8 5 9.375 11,937 36.30 0.910 1.090 7,817
3–
8 6 11.250 14,324 43.45 0.760 0.910 9,230
3–
8 8 15.000 19,099 58.00 0.570 0.682 11,919
3–
8 10 18.750 23,873 72.50 0.456 0.544 14,661
3–
8 12 22.500 28,648 87.00 0.380 0.456 17,195
6 Parallel Bars
3– –3
8 4 1.688 2,149 6.54 5.075 6.073 1,844
3–
8 1 2.250 2,865 8.70 3.793 4.538 2,277
3–
8 11–2 3.375 4,297 13.02 2.528 3.027 3,201
3–
8 2 4.500 5,730 17.40 1.905 2.280 4,024
3–
8 21–2 5.625 7,162 21.72 1.518 1.818 4,889
3–
8 3 6.750 8,594 26.10 1.273 1.523 5,662
3–
8 31–2 7.875 10,027 30.36 1.088 1.302 6,452
3–
8 4 9.000 11,459 34.80 0.952 1.138 7,230
3–
8 5 11.250 14,324 43.56 0.758 0.908 8,819
3–
86 13.500 17,189 52.14 0.633 0.758 10,413
3–
8
8 18.000 22,918 69.60 0.475 0.568 13,447
3–
108 22.500 28,648 87.00 0.380 0.453 16,540
3–
12 8 27.000 34,377 104.40 0.317 0.380 19,400
7 Parallel Bars
3– –3
8 4 1.969 2,507 7.63 4.350 5.206 2,096
3–
8 1 2.625 3,342 10.15 3.251 3.890 2,588
3–
8 11–2 3.938 5,013 15.19 2.167 2.594 3,638
3–
8 2 5.250 6,685 20.30 1.633 1.954 4,573
3–
8 21–2 6.563 8,356 25.34 1.301 1.559 5,556
3–
8 3 7.875 10,027 30.45 1.091 1.306 6,434
3–
8 31–2 9.188 11,698 35.42 0.933 1.116 7,332
3–
8 4 10.500 13,369 40.60 0.816 0.976 8,216
3–
8 5 13.125 16,711 50.82 0.650 0.779 10,022
3–
86 15.750 20,054 60.83 0.543 0.650 11,833
3–
8
8 21.000 26,738 81.20 0.407 0.487 15,281
3–
108 26.250 33,423 101.50 0.326 0.389 18,796
3–
12 8 31.500 40,107 121.80 0.271 0.326 22,045
8 Parallel Bars
3– –3
8 4 2.250 2,865 8.72 3.806 4.555 2,305
3–
8 1 3.000 3,820 11.60 2.845 3.404 2,847
3–
8 11–2 4.500 5,730 17.36 1.896 2.270 4,001
3–
8 2 6.000 7,639 23.20 1.429 1.710 5,030
3–
8 21–2 7.500 9,549 28.96 1.139 1.364 6,112
3–
8 3 9.000 11,459 34.80 0.955 1.143 7,078
(continued)
284 SYSTEM DESIGN
DC Current
_____________________________________ __________________ ________
8 Parallel Bars
3–
8 31–2 10.500 13,369 40.48 0.816 0.976 8,066
3–
8 4 12.000 15,279 46.40 0.714 0.854 9,037
3–
8 5 15.000 19,099 58.08 0.569 0.681 11,024
3–
8 6 18.000 22,918 69.52 0.475 0.569 13,016
3–
8 8 24.000 30,558 92.80 0.356 0.426 16,809
3–
8 10 30.000 38,197 116.00 0.285 0.340 20,675
3–
8 12 36.000 45,837 139.20 0.238 0.285 24,250
9 Parallel Bars
3– –3
8 4 2.531 3,223 9.81 3.383 4.049 2,515
3–
8 1 3.375 4,297 13.05 2.529 3.026 3,106
3–
8 11–2 5.063 6,446 19.53 1.686 2.018 4,365
3–
8 2 6.750 8,594 26.10 1.270 1.520 5,487
3–
8 21–2 8.438 10,743 32.58 1.012 1.212 6,667
3–
8 3 10.125 12,892 39.15 0.849 1.016 7,721
3–
8 31–2 11.813 15,040 45.54 0.726 0.868 8,799
3–
84 13.500 17,189 52.20 0.634 0.759 9,859
3–
5
8 16.875 21,486 65.34 0.506 0.606 12,026
3–
68 20.250 25,783 78.21 0.422 0.506 14,199
3–
8 8 27.000 34,377 104.40 0.317 0.379 18,337
3–
10 8 33.750 42,972 130.50 0.253 0.302 22,555
3–
12 8 40.500 51,566 156.60 0.211 0.253 26,454
10 Parallel Bars
3– –3
8 4 2.813 3,581 10.90 3.045 3.644 2,724
3–
8 1 3.750 4,775 14.50 2.276 2.723 3,364
3–
8 11–2 5.625 7,162 21.70 1.517 1.816 4,729
3–
8 2 7.500 9,549 29.00 1.143 1.368 5,944
3–
8 21–2 9.375 11,937 36.20 0.911 1.091 7,223
3–
8 3 11.250 14,324 43.50 0.764 0.914 8,365
3–
8 31–2 13.125 16,711 50.60 0.653 0.781 9,532
3–
84 15.000 19,099 58.00 0.571 0.683 10,681
3–
5
8 18.750 23,873 72.60 0.455 0.545 13,029
3–
68 22.500 28,648 86.90 0.380 0.455 15,383
3–
8 8 30.000 38,197 116.00 0.285 0.341 19,865
3–
10 8 37.500 47,746 145.00 0.228 0.272 24,434
3–
12 8 45.000 57,296 174.00 0.190 0.228 28,659
11 Parallel Bars
3– –3
8 4 3.094 3,939 11.99 2.768 3.313 2,892
3–
8 1 4.125 5,252 15.95 2.069 2.475 3,571
3–
8 11–2 6.188 7,878 23.87 1.379 1.651 5,020
3–
8 2 8.250 10,504 31.90 1.039 1.244 6,310
3–
8 21–2 10.313 13,130 39.82 0.828 0.992 7,667
3–
8 3 12.375 15,756 47.85 0.695 0.831 8,879
3–
8 31–2 14.438 18,382 55.66 0.594 0.710 10,119
3–
8 4 16.500 21,008 63.80 0.519 0.621 11,338
3–
8 5 20.625 26,261 79.86 0.414 0.495 13,830
DC Current
_____________________________________ __________________ ________
11 Parallel Bars
3–
8 6 24.750 31,513 95.59 0.345 0.414 16,329
3–
8 8 33.000 42,017 127.60 0.259 0.310 21,087
3–
8 10 41.250 52,521 159.50 0.207 0.247 25,938
3–
8 12 49.500 63,025 191.40 0.173 0.207 30,423
12 Parallel Bars
3– –3
8 4 3.375 4,297 13.08 2.538 3.037 3,060
3–
8 1 4.500 5,730 17.40 1.897 2.269 3,778
3–
8 11–2 6.750 8,594 26.04 1.264 1.513 5,311
3–
8 2 9.000 11,459 34.80 0.953 1.140 6,676
3–
8 21–2 11.250 14,324 43.44 0.759 0.909 8,112
3–
8 3 13.500 17,189 52.20 0.637 0.762 9,394
3–
8 31–2 15.750 20,054 60.72 0.544 0.651 10,705
3–
8 4 18.000 22,918 69.60 0.476 0.569 11,995
3–
8 5 22.500 28,648 87.12 0.379 0.454 14,632
3–
8 6 27.000 34,377 104.28 0.317 0.379 17,276
3–
8 8 36.000 45,837 139.20 0.238 0.284 22,310
3–
8 10 45.000 57,296 174.00 0.190 0.227 27,442
3–
8 12 54.000 68,755 208.80 0.158 0.190 32,186
a
b
Table 5.18(c) Rigid Copper Busbar Properties—1–2-in. Parallel Bars as Indicateda,b

DC Current
_____________________________________ __________________ ________
2 Parallel Bars
1–
2 1 1.000 1,273 3.860 8.575 10.260 1,138
1–
2 11–2 1.500 1,910 5.800 5.715 6.840 1,553
1–
2 2 2.000 2,546 7.720 4.275 5.115 1,905
1–
2 21–2 2.500 3,183 9.660 3.420 4.090 2,326
1–
2 3 3.000 3,820 11.600 2.855 3.415 2,761
1–
2 31–2 3.500 4,456 13.520 2.440 2.920 3,107
1–
2 4 4.000 5,093 15.460 2.135 2.555 3,435
1–
2 5 5.000 6,366 19.320 1.710 2.045 4,239
1–
2 6 6.000 7,639 23.200 1.425 1.705 5,038
1–
2 8 8.000 10,186 31.000 1.070 1.280 6,435
1–
2 10 10.000 12,732 38.600 0.855 1.025 7,857
1–
2 12 12.000 15,279 46.400 0.710 0.850 9,321
(continued)
286 SYSTEM DESIGN
Table 5.18(c) Rigid Copper Busbar Properties—1–2-in. Parallel Bars as Indicateda,b (Continued)
DC Current
_____________________________________ __________________ ________
3 Parallel Bars
1–
2 1 1.500 1,910 5.790 5.717 6.840 1,581
1–
2 11–2 2.250 2,865 8.700 3.810 4.560 2,156
1–
2 2 3.000 3,820 11.580 2.850 3.410 2,646
1–
2 21–2 3.750 4,775 14.490 2.280 2.727 3,231
1–
2 3 4.500 5,730 17.400 1.903 2.277 3,834
1–
2 31–2 5.250 6,685 20.280 1.627 1.947 4,315
1–
2 4 6.000 7,639 23.190 1.423 1.703 4,771
1–
2 5 7.500 9,549 28.980 1.140 1.363 5,888
1–
2 6 9.000 11,459 34.800 0.950 1.137 6,997
1–
2 8 12.000 15,279 46.500 0.713 0.853 8,937
1–
2 10 15.000 19,099 57.900 0.570 0.683 10,912
1–
2 12 18.000 22,918 69.600 0.473 0.567 12,945
4 Parallel Bars
1–
2 1 2.000 2,546 7.720 4.288 5.130 2,023
1–
2 11–2 3.000 3,820 11.600 2.858 3.420 2,760
1–
2 2 4.000 5,093 15.440 2.138 2.558 3,387
1–
2 21–2 5.000 6,366 19.320 1.710 2.045 4,136
1–
2 3 6.000 7,639 23.200 1.428 1.708 4,908
1–
2 31–2 7.000 8,913 27.040 1.220 1.460 5,523
1–
2 4 8.000 10,186 30.920 1.068 1.278 6,106
1–
2 5 10.000 12,732 38.640 0.855 1.023 7,537
1–
2 6 12.000 15,279 46.400 0.713 0.853 8,956
1–
2 8 16.000 20,372 62.000 0.535 0.640 11,440
1–
2 10 20.000 25,465 77.200 0.428 0.513 13,967
1–
2 12 24.000 30,558 92.800 0.355 0.425 16,570
5 Parallel Bars
1–
2 1 2.500 3,183 9.650 3.430 4.104 2,466
1–
2 11–2 3.750 4,775 14.500 2.286 2.736 3,364
1–
2 2 5.000 6,366 19.300 1.710 2.046 4,127
1–
2 21–2 6.250 7,958 24.150 1.368 1.636 5,041
1–
2 3 7.500 9,549 29.000 1.142 1.366 5,981
1–
2 31–2 8.750 11,141 33.800 0.976 1.168 6,731
1–
2 4 10.000 12,732 38.650 0.854 1.022 7,442
1–
2 5 12.500 15,915 48.300 0.684 0.818 9,186
1–
2 6 15.000 19,099 58.000 0.570 0.682 10,916
1–
2 8 20.000 25,465 77.500 0.428 0.512 13,942
1–
2 10 25.000 31,831 96.500 0.342 0.410 17,023
1–
2 12 30.000 38,197 116.000 0.284 0.340 20,195
6 Parallel Bars
1–
2 1 3.000 3,820 11.580 2.858 3.420 2,782
1–
2 11–2 4.500 5,730 17.400 1.905 2.280 3,795
1–
2 2 6.000 7,639 23.160 1.425 1.705 4,657
1–
2 21–2 7.500 9,549 28.980 1.140 1.363 5,687
1–
2 3 9.000 11,459 34.800 0.952 1.138 6,748
1–
2 31–2 10.500 13,369 40.560 0.813 0.973 7,594
DC Current
_____________________________________ __________________ ________
6 Parallel Bars
1–
2 4 12.000 15,279 46.380 0.712 0.852 8,396
1–
2 5 15.000 19,099 57.960 0.570 0.682 10,363
1–
2 6 18.000 22,918 69.600 0.475 0.568 12,315
1–
2 8 24.000 30,558 93.000 0.357 0.427 15,730
1–
2 10 30.000 38,197 115.800 0.285 0.342 19,205
1–
2 12 36.000 45,837 139.200 0.237 0.283 22,784
7 Parallel Bars
1–
2 1 3.500 4,456 13.510 2.450 2.931 3,161
1–
2 11–2 5.250 6,685 20.300 1.633 1.954 4,313
1–
2 2 7.000 8,913 27.020 1.221 1.461 5,292
1–
2 21–2 8.750 11,141 33.810 0.977 1.169 6,462
1–
2 3 10.500 13,369 40.600 0.816 0.976 7,668
1–
2 31–2 12.250 15,597 47.320 0.697 0.834 8,630
1–
2 4 14.000 17,825 54.110 0.610 0.730 9,541
1–
2 5 17.500 22,282 67.620 0.489 0.584 11,776
1–
2 6 21.000 26,738 81.200 0.407 0.487 13,994
1–
2 8 28.000 35,651 108.500 0.306 0.366 17,875
1–
2 10 35.000 44,563 135.100 0.244 0.293 21,824
1–
2 12 42.000 53,476 162.400 0.203 0.243 25,891
8 Parallel Bars
1–
2 1 4.000 5,093 15.440 2.144 2.565 3,478
1–
2 11–2 6.000 7,639 23.200 1.429 1.710 4,744
1–
2 2 8.000 10,186 30.880 1.069 1.279 5,821
1–
2 21–2 10.000 12,732 38.640 0.855 1.023 7,108
1–
2 3 12.000 15,279 46.400 0.714 0.854 8,435
1–
2 31–2 14.000 17,825 54.080 0.610 0.730 9,493
1–
2 4 16.000 20,372 61.840 0.534 0.639 10,495
1–
2 5 20.000 25,465 77.280 0.428 0.511 12,954
1–
2 6 24.000 30,558 92.800 0.356 0.426 15,394
1–
2 8 32.000 40,744 124.000 0.268 0.320 19,662
1–
2 10 40.000 50,930 154.400 0.214 0.256 24,006
1–
2 12 48.000 61,115 185.600 0.178 0.213 28,480
9 Parallel Bars
1–
2 1 4.500 5,730 17.370 1.906 2.280 3,794
1–
2 11–2 6.750 8,594 26.100 1.270 1.520 5,175
1–
2 2 9.000 11,459 34.740 0.950 1.137 6,350
1–
2 21–2 11.250 14,324 43.470 0.760 0.909 7,755
1–
2 3 13.500 17,189 52.200 0.634 0.759 9,202
1–
2 31–2 15.750 20,054 60.840 0.542 0.649 10,356
1–
2 4 18.000 22,918 69.570 0.474 0.568 11,449
1–
2 5 22.500 28,648 86.940 0.380 0.454 14,132
1–
2 6 27.000 34,377 104.400 0.317 0.379 16,793
1–
2 8 36.000 45,837 139.500 0.238 0.284 21,449
1–
2 10 45.000 57,296 173.700 0.190 0.228 26,189
1–
2 12 54.000 68,755 208.800 0.158 0.189 31,069
(continued)
288 SYSTEM DESIGN
DC Current
_____________________________________ __________________ ________
10 Parallel Bars
1–
2 1 5.000 6,366 19.300 1.715 2.052 4,110
1–
2 11–2 7.500 9,549 29.000 1.143 1.368 5,607
1–
2 2 10.000 12,732 38.600 0.855 1.023 6,879
1–
2 21–2 12.500 15,915 48.300 0.684 0.818 8,401
1–
2 3 15.000 19,099 58.000 0.571 0.683 9,969
1–
2 31–2 17.500 22,282 67.600 0.488 0.584 11,219
1–
2 4 20.000 25,465 77.300 0.427 0.511 12,404
1–
2 5 25.000 31,831 96.600 0.342 0.409 15,309
1–
2 6 30.000 38,197 116.000 0.285 0.341 18,193
1–
2 8 40.000 50,930 155.000 0.214 0.256 23,237
1–
2 10 50.000 63,662 193.000 0.171 0.205 28,371
1–
2 12 60.000 76,394 232.000 0.142 0.170 33,658
11 Parallel Bars
1–
2 1 5.500 7,003 21.230 1.559 1.865 4,363
1–
2 11–2 8.250 10,504 31.900 1.039 1.244 5,952
1–
2 2 11.000 14,006 42.460 0.777 0.930 7,302
1–
2 21–2 13.750 17,507 53.130 0.622 0.744 8,918
1–
2 3 16.500 21,008 63.800 0.519 0.621 10,582
1–
2 31–2 19.250 24,510 74.360 0.444 0.531 11,909
1–
2 4 22.000 28,011 85.030 0.388 0.465 13,167
1–
2 5 27.500 35,014 106.260 0.311 0.372 16,251
1–
2 6 33.000 42,017 127.600 0.259 0.310 19,312
1–
2 8 44.000 56,023 170.500 0.195 0.233 24,667
1–
2 10 55.000 70,028 212.300 0.155 0.186 30,117
1–
2 12 66.000 84,034 255.200 0.129 0.155 35,729
12 Parallel Bars
1–
2 1 6.000 7,639 23.160 1.429 1.710 4,616
1–
2 11–2 9.000 11,459 34.800 0.953 1.140 6,297
1–
2 2 12.000 15,279 46.320 0.713 0.853 7,726
1–
2 21–2 15.000 19,099 57.960 0.570 0.682 9,435
1–
2 3 18.000 22,918 69.600 0.476 0.569 11,195
1–
2 31–2 21.000 26,738 81.120 0.407 0.487 12,600
1–
2 4 24.000 30,558 92.760 0.356 0.426 13,930
1–
2 5 30.000 38,197 115.920 0.285 0.341 17,193
1–
2 6 36.000 45,837 139.200 0.238 0.284 20,432
1–
2 8 48.000 61,115 186.000 0.178 0.213 26,097
1–
2 10 60.000 76,394 231.600 0.143 0.171 31,863
1–
2 12 72.000 91,673 278.400 0.118 0.142 37,800
a
b
Current ratings apply when the busbar is oriented with its long edge vertical, and bar spacing bar thick-
ness, and busbar is run in the horizontal plane. Current ratings based on 30°C rise above 40°C ambient tem-
perature.
Table 5.18(d) Rigid Copper Busbar Properties—3–4-in. Parallel Bars as Indicateda,b

DC Current
_____________________________________ __________________ ________
2 Parallel Bars
–3 4 6.00 7,639 23.2 1.425 1.705 4,397
4
–3 5 7.50 9,549 29.0 1.140 1.360 5,256
4
–3 6 9.00 11,459 34.8 0.950 1.140 6,214
4
–3 8 12.00 15,279 46.4 0.710 0.850 7,969
4
–3 10 15.00 19,099 58.0 0.570 0.685 9,770
4
–3 12 18.00 22,918 69.6 0.475 0.570 11,567
4
3 Parallel Bars
–3 4 9.00 11,459 34.8 0.950 1.137 6,107
4
–3 5 11.25 14,324 43.5 0.760 0.907 7,299
4
–3 6 13.50 17,189 52.2 0.633 0.760 8,630
4
–3 8 18.00 22,918 69.6 0.473 0.567 11,068
4
–3 10 22.50 28,648 87.0 0.380 0.457 13,569
4
–3 12 27.00 34,377 104.4 0.317 0.380 16,066
4
4 Parallel Bars
–3 4 12.00 15,279 46.4 0.713 0.853 7,817
4
–3 5 15.00 19,099 58.0 0.570 0.680 9,343
4
–3 6 18.00 22,918 69.6 0.475 0.570 11,047
4
–3 8 24.00 30,558 92.8 0.355 0.425 14,167
4
–3 10 30.00 38,197 116.0 0.285 0.343 17,368
4
–3 12 36.00 45,837 139.2 0.238 0.285 20,564
4
5 Parallel Bars
–3 4 15.00 19,099 58.0 0.570 0.682 9,527
4
–3 5 18.75 23,873 72.5 0.456 0.544 11,387
4
–3 6 22.5 28,648 87.0 0.380 0.456 13,463
4
–3 8 30.0 38,197 116.0 0.284 0.340 17,266
4
–3 10 37.5 47,746 145.0 0.228 0.274 21,168
4
–3 12 45.0 57,296 174.0 0.190 0.228 25,063
4
6 Parallel Bars
–3 4 18.0 22,918 69.6 0.475 0.568 10,749
4
–3 5 22.5 28,648 87.0 0.380 0.453 12,847
4
–3 6 27.0 34,377 104.4 0.317 0.380 15,189
4
–3 8 36.0 45,837 139.2 0.237 0.283 19,480
4
–3 10 45.0 57,296 174.0 0.190 0.228 23,881
4
–3 12 54.0 68,755 208.8 0.158 0.190 28,276
4
7 Parallel Bars
–3 4 21.0 26,738 81.2 0.407 0.487 12,214
4
–3 5 26.2 33,423 101.5 0.326 0.389 14,599
4
–3 6 31.5 40,107 121.8 0.271 0.326 17,260
4
–3 8 42.0 53,476 162.4 0.203 0.243 22,136
4
–3 10 52.5 66,845 203.0 0.163 0.196 27,138
4
–3 12 63.0 80,214 243.6 0.136 0.163 32,131
4
8 Parallel Bars
–3 4 24.0 30,558 92.8 0.356 0.426 13,436
4
–3 5 30.0 38,197 116.0 0.285 0.340 16,058
4
–3 6 36.0 45,837 139.2 0.238 0.285 18,986
4
(continued)
290 SYSTEM DESIGN
Table 5.18(d) Rigid Copper Busbar Properties—3–4-in. Parallel Bars as Indicateda,b (Continued)
DC Current
_____________________________________ __________________ ________
8 Parallel Bars
–3 8 48.0 61,115 185.6 0.178 0.213 24,350
4
–3 10 60.0 76,394 232.0 0.143 0.171 29,852
4
–3 12 72.0 91,673 278.4 0.119 0.143 35,345
4
9 Parallel Bars
–3 4 27.0 34,377 104.4 0.317 0.379 14,657
4
–3 5 33.7 42,972 130.5 0.253 0.302 17,518
4
–3 6 40.5 51,566 156.6 0.211 0.253 20,712
4
–3 8 54.0 68,755 208.8 0.158 0.189 26,563
4
–3 10 67.5 85,944 261.0 0.127 0.152 32,566
4
–3 12 81.0 103,132 313.2 0.106 0.127 38,558
4
10 Parallel Bars
–3 4 30.0 38,197 116.0 0.285 0.341 15,879
4
–3 5 37.5 47,746 145.0 0.228 0.272 18,978
4
–3 6 45.0 57,296 174.0 0.190 0.228 22,438
4
–3 8 60.0 76,394 232.0 0.142 0.170 28,777
4
–3 10 75.0 95,493 290.0 0.114 0.137 35,279
4
–3 12 90.0 114,592 348.0 0.095 0.114 41,771
4
11 Parallel Bars
–3 4 33.0 42,017 127.6 0.259 0.310 16,856
4
–3 5 41.2 52,521 159.5 0.207 0.247 20,146
4
–3 6 49.5 63,025 191.4 0.173 0.207 23,819
4
–3 8 66.0 84,034 255.2 0.129 0.155 30,548
4
–3 10 82.5 105,042 319.0 0.104 0.125 37,450
4
–3 12 99.0 126,051 382.8 0.086 0.104 44,341
4
12 Parallel Bars
–3 4 36.0 45,837 139.2 0.238 0.284 17,833
4
–3 5 45.0 57,296 174.0 0.190 0.227 21,314
4
–3 6 54.0 68,755 208.8 0.158 0.190 25,200
4
–3 8 72.0 91,673 278.4 0.118 0.142 32,318
4
–3 10 90.0 114,592 348.0 0.095 0.114 39,621
4
–3 12 108.0 137,510 417.6 0.079 0.095 46,912
4
a
b
Table 5.19 Alternate Values of DC Current Ratings for Rigid Copper Busbar—1–4-in. and 1–2-in.
Parallel Bars as Indicateda,b,c
DC Current Rating DC Current Rating
__________________ _________________
Dimensions 70°C 70°C Dimensions 70°C 70°C
Thickness Width Vertical Other Thickness Width Vertical Other
(in.) (in.) (A) (A) (in.) (in.) (A) (A)
1 Bar
1– 1
4 2 731 721 –
2 2 1,088 1,073
1– 1
4 3 1,040 1,019 –
2 3 1,525 1,494
1– 1
4 4 1,342 1,296 –
2 4 1,951 1,887
1– 1
4 6 1,931 1,820 –
2 6 2,783 2,623
1– 1
4 8 2,506 2,292 –
2 8 3,596 3,289
2 Parallel Bars
1– 1
4 2 1,301 1,259 –
2 2 1,961 1,902
1– 1
4 3 1,834 1,735 –
2 3 2,715 2,577
1– 1
4 4 2,350 2,163 –
2 4 3,445 3,182
1– 1
4 6 3,352 2,937 –
2 6 4,861 4,275
1– 1
4 8 4,325 3,583 –
2 8 6,236 5,189
3 Parallel Bars
1–
4 2 1,865 1,787 — — — —
1–
4 3 2,616 2,432 — — — —
1– 1
4 4 3,342 2,996 –
2 4 4,918 4,437
1– 1
4 6 4,745 3,992 –
2 6 6,902 5,848
1– 1
4 8 6,105 4,770 –
2 8 8,824 6,950
4 Parallel Bars
1–
4 2 2,426 2,313 — — — —
1–
4 3 3,394 3,123 — — — —
1– 1
4 4 4,328 3,819 –
2 4 6,384 5,673
1– 1
4 6 6,130 5,026 –
2 6 8,933 7,392
1– 1
4 8 7,872 5,916 –
2 8 11,395 8,659
5 Parallel Bars
1– 1
4 4 5,312 4,637 –
2 4 7,847 6,915
1– 1
4 6 7,512 6,048 –
2 6 10,960 8,921
1– 1
4 8 9,634 7,041 –
2 8 13,960 10,340
6 Parallel Bars
1– 1
4 4 6,295 5,452 –
2 4 9,309 8,148
1– 1
4 6 8,891 7,064 –
2 6 12,980 10,445
1– 1
4 8 11,395 8,154 –
2 8 16,520 12,005
7 Parallel Bars
1– 1
4 6 10,270 8,076 –
2 6 15,000 11,960
1– 1
4 8 13,150 9,259 –
2 8 19,080 13,660
8 Parallel Bars
1– 1
4 6 11,645 9,086 –
2 6 17,020 13,475
1– 1
4 8 14,905 10,360 –
2 8 21,635 15,310
9 Parallel Bars
1– 1
4 6 13,020 10,095 –
2 6 19,040 14,985
1– 1
4 8 16,660 11,455 –
2 8 24,190 16,955
10 Parallel Bars
1– 1
4 6 14,400 11,100 –
2 6 21,060 16,495
1– 1
4 8 18,415 12,545 –
2 8 26,745 18,600
(continued)
292 SYSTEM DESIGN
Table 5.19 Alternate Values of DC Current Ratings for Rigid Copper Busbar—1–4-in. and 1–2-in.
Parallel Bars as Indicateda,b,c (Continued)
DC Current Rating DC Current Rating
__________________ _________________
Dimensions 70°C 70°C Dimensions 70°C 70°C
Thickness Width Vertical Other Thickness Width Vertical Other
(in.) (in.) (A) (A) (in.) (in.) (A) (A)
11 Parallel Bars
1–
4 6 15,775 12,105 — — — —
1–
4 8 20,170 13,640 — — — —
12 Parallel Bars
1–
4 6 17,150 13,110 — — — —
1–
4 8 21,925 14,725 — — — —
a
Some current ratings in this table differ slightly from the current ratings given in previous tables. The differ-
ences are due to historic precedent.
b
Current ratings in the “Vertical” column apply when the busbars are oriented with their long edge vertical, and
bar spacing bar thickness, and busbar is run in the horizontal direction.
c
Current ratings in the “Other” column apply with the busbars are oriented with their long edge horizontal, or
when bar spacing < bar thickness, or when the bar is run in the vertical direction.
d
Current ratings based on 30°C rise above 40°C ambient temperature.
Source: Data from [28].
arrangement with a feed and return leads following the same path, the actual conductor
length for a given circuit, or loop length, is twice (2×) the path distance. Because of bus
mechanical arrangements, it is possible the feed and return conductors have slightly dif-
ferent lengths. In this case, the estimated path distance would be two times the average of
the feed and return distances.
As mentioned in Section 5.2.2.2, industry standards suggest no more than 2.0 V total
drop between the battery terminals and load equipment for 48-V systems and 1.0 V for
24-V systems. These voltage drops are not fixed throughout any given system (the volt-
age drop can be less or more in any particular circuit) but are target maximum values that
are used to rationalize circuit design. The target values will be considered specified maxi-
mum values for the purposes of circuit design in this book.
Since the important parameter is the total voltage drop from the battery terminals to the
load equipment terminals, the voltage drop in each segment (battery, primary distribution,
and secondary distribution circuits) can be designed for any value as long as the sum of
the voltage drops does not exceed the maximum.
The voltage drop between the battery terminals and the discharge bus normally is in
the range of 0.25 to 0.75 V for 48-V systems, with 0.5 V being a good starting value for
design purposes (one-half these values for 24-V systems). These values work particularly
well if the battery is near the powerboard discharge bus and the loads are farther away.
Where the situation is reversed, and the battery is far away and the load is near, the bat-
tery circuit usually is designed for a higher voltage drop to reduce conductor size. Since
only the total voltage drop is specified, the designer has flexibility in circuit segment de-
sign. Typical values for each circuit segment are shown in Table 5.20.
For the first voltage drop calculation method
FVD LLoad LLoop
CM = (5.25)
VDrop
Fig. 5.34 Parallel busbars. (Photo courtesy of Schultz Brothers Electric Company.)
Fig. 5.35 Cable path distance.
Table 5.20 Typical Circuit Segment Voltage Drops

Circuit Segment 48-V Systems (V) 24-V Systems (V)
(1) Battery 0.25–0.75 0.13–0.38
(2) Primary 0.5–1.5 0.25–0.75
(3) Secondary 0.5–1.5 0.25–0.75
(4) Maximum total (1) + (2) + (3) 2.0 1.0
(5) Rectifier (maximum) 1.0 0.5
294 SYSTEM DESIGN
where CM = conductor area (CM)

FVD = copper wire resistivity at desired design temperature in -CM/ft (11.1 at
30°C and 12.9 at 75°C)
ILoad = load current (A)
LLoop = conductor loop length, feed + return (ft)
VDrop = voltage drop (V)
For rectifier output circuits, the load current, ILoad, used in Eq. (5.25) is the maximum
possible output current as follows:
앫 Power rated rectifiers Output power divided by the lowest possible output voltage.
앫 Current rated rectifiers Maximum rated output current.
앫 Rectifiers with overload capability Rated output current plus overload current ca-
pability.
앫 Where a modular rectifier shelf or chassis is used, the circuit conductor current rat-
ing is based on the maximum rated capacity of the shelf.
Example 5.20 Calculate the cross-sectional area and conductor size so that the voltage
drop does not exceed 0.5 V on a 50-ft circuit loop (feed + return) when the load current is
55 A and the conductor temperature is 30°C.
Solution For the specified situation:
FVD = 11.1
I = 55 A
LLoop = 50 ft
VDrop = 0.5 V
The CM cross-sectional area is
11.1 × 55 A × 50 ft
CM = = 61,050 CM
0.5 V
Therefore, a single set of conductors or parallel set of conductors having a total cross sec-
tion 61,050 CM will meet the 0.5-V voltage drop requirement. The nearest conductor
size (exceeding the minimum CM requirement) is 2 AWG with a cross section of 66,360
CM (from Table 5.13). The actual voltage drop with this conductor can be determined by
solving Eq. (5.25) for VDrop:
FVD ILLoop
VDrop = (5.26)
CM
11.1 × 55 A × 50 ft
= = 0.46 V
66,360 CM
Example 5.21 Calculate the cross-sectional area and conductor size so that the voltage
drop does not exceed 0.5 V on a 50-ft circuit loop when the load current is 55 A and the
conductor temperature is 75°C.
FVD = 12.9
I = 55 A
LLoop = 50 ft
VDrop = 0.5 V
Solution
12.9 × 55 A × 50 ft
CM = = 70,950 CM
0.5 V
Referring to Table 5.13, it is seen that 1 AWG conductors with a cross-sectional area of
83,690 CM meet the requirements. The actual voltage drop is
12.9 × 55 A × 50 ft
VDrop = = 0.42
83,690 CM
Because copper conductors have higher resistance at higher temperatures, the voltage
drop at 75°C (167°F) is higher than at 30°C (86°F) and a larger wire is needed to keep it
at or below the specified value.
The second method mentioned above a variation of Ohm’s law, that is, the voltage
drop across a resistance is directly proportional to the resistance and the current through
the resistance, or
VDrop = IR (5.27)
where VDrop = voltage drop (V)

I = load current (A)
R = circuit resistance ()
If parallel conductors are used in a circuit, the resistance would be of the parallel combi-
nation. Since circuit resistance is proportional to conductor length, Eq. (5.27) can be
rewritten as
VDrop = IRUnit LLoop (5.28)
where VDrop = voltage drop (V)

I = load current (A)
RUnit = circuit unit resistance (/ft)
LLoop = circuit conductor loop length (ft)
Usually, the maximum voltage drop is known and it is necessary to calculate the maxi-
mum circuit resistance. In this case, the voltage drop equation is rearranged to solve for
RUnit as in
VDrop
RUnit = (5.29)
ILLoop
296 SYSTEM DESIGN
Once the unit resistance is known, it only is necessary to choose a conductor whose resis-
tance does not exceed the calculated value.
Example 5.22 Calculate the conductor size required to keep voltage drop below 0.5 V
on a 50-ft circuit loop (feed + return) when the load current is 55 A and the conductor
temperature is 30°C.
Solution
I = 55 A
LLoop = 50 ft
VDrop = 0.5 V
0.5 V
RUnit = = 0.000182 /ft
55 A × 50 ft
Entering Table 5.13 at a temperature of 30°C (86°F), it is seen that 2 AWG has a unit re-
sistance of 0.000166 /ft, which is less than the maximum calculated above. The actual
voltage drop is calculated from
VDrop = 55 A × 0.000166 /ft × 50 ft = 0.46 V
Note that this result agrees with the results in Example 5.21, which was calculated using
the voltage drop factor.
5.8.1.6 Current Rating in Cable Racks Current rating (current-carrying capacity

or ampacity) is the current that a wire may safely carry under the specified conditions
without overheating. Current rating varies with ambient temperature, allowable tempera-
ture rise of the wire, and installation conditions. Wire current ratings usually are specified
at 60°C (140°F), 75°C (167°F), or 90°C (194°F) with the latter two being the most com-
mon in telecommunications applications. In some cases, the temperature rating of the ter-
mination may determine the maximum temperature rating of the wire insulation. For ex-
ample, some circuit breakers and switch devices are rated for continuous load only if the
connecting wires have a 90°C rating. Similarly, a 90°C wire current rating cannot be used
on a 75°C termination—the wire must be used at its 75°C current rating.
Power wiring in modern telecommunications systems is continuously loaded, and the
continuous load current normally is increased by 25% (1.25 multiplier) to provide a de-
sign margin in current rating calculations. This design margin includes uncertainty in the
current-carrying capacity of the wire for different installation conditions. Therefore, the
required current ratings for the different circuits are
IAmpacity = 1.25ILoad distribution and load (branch) circuit21 (5.30a)
IAmpacity = 1.25IRectSystem battery circuit (5.30b)
IAmpacity = 1.25IRectChassis rectifier circuit (5.30c)
21
A branch circuit consists of circuit conductors between the final overcurrent protection device and the load
equipment.
where ICurrent rating = conductor current rating (A)

ILoad = load current (A)
IRectSystem = maximum current from rectifier system (A) [depending on system
design, this usually is the ultimate rectifier system current (IRSU)]
IRectChassis = maximum current from an individual rectifier chassis or rectifier
shelf assembly (A)
Example 5.23 A standalone rectifier is rated 50 A at 48 V and has 10% overload capa-
bility. Determine the output conductor current rating.
Solution The maximum available current from this rectifier is 50 × 1.1 = 55 A. There-
fore, the required current rating is
IAmpacity = 1.25IRectChassis = 1.25 × 55 A = 69 A
Example 5.24 A modular rectifier shelf holds four 3.0-kW rectifiers and the rectifiers
have no overload capability. The lowest operating voltage of the rectifiers is 42.0 V. De-
termine the output conductor current rating.
Solution The maximum available current from each rectifier is 3000 kW ÷ 42.0 V =
71.4 A, and the maximum available shelf current is 4 × 71.4 A = 286 A. Therefore, the re-
quired current rating is
IAmpacity = 1.25IRectShelf = 1.25 × 286 A = 358 A
The dc current ratings for different wire configurations and installation conditions (Fig.
5.36) are derived in the following subsections, including
앫 Wires in free air (separated from other wires by at least their diameter) in a single
layer
앫 Two bundled wires (two wires touching each other) in free air in a single layer
앫 Wires with no separation and laid flat in single layer, stacked in multiple layers, or
randomly laid
앫 Wires bundled in a rounded configuration in free air in a single layer
Many telecommunications power circuits use parallel wires to meet current rating and
voltage drop requirements. The current rating of parallel wires is the algebraic sum of in-
dividual wire current ratings, or
n
Ip = 冱 II np (5.31)
1
where Ip = parallel current rating (A)

II = current rating of individual wires for the specified physical configuration (A)
np = number of parallel wires
When determining the current rating of parallel wires, it is necessary to know the physical
configuration. For example, if the parallel wires are installed in free air, then the table for
298 SYSTEM DESIGN
Fig. 5.36 Common cable rack installation conditions. In single- or multiple-layer installations, the
wire layers do not always extend across the entire width of the cable rack. Wire sizes shown in each
configuration are common and are not meant to limit the sizes actually used in the field.
free air current rating would be used, and if the wires are bundled, the current rating tables
for bundled configurations would be used. Parallel wires always are the same size.
In all configurations discussed in this section, the current ratings are for wires on lad-
der-type cable racks commonly found in telecommunications facilities, and the installa-
tion conditions are such that the heat convected and radiated from the wires is freely dissi-
pated (i.e., there are no heat flow restrictions). This generally means there are no cable
racks or other obstructions above or below that can block the heat flow, and external heat
sources (such as lighting fixtures) are sufficiently separated from the cable racks that they
have no effect on wire temperatures. For purposes of consistent treatment, the current rat-
ings in this section are developed from basic heat flow and energy balance principles and
methods accepted in the electrical industries (U.S. and international).
Current ratings under specified conditions are derived by determining the heat generat-
ed in a wire from resistive losses. For a single wire in free air, the heat flows by conduc-
tion from the copper conductor through the insulation to its surface. The heat is then dissi-
pated through radiation and convection to the ambient air. The conductor in the wire is
operated at its maximum specified temperature, such as 75°C (167°F) or 90°C (194°F).
For purposes of calculation, the ambient is assumed to be 30°C (86°F), and factors are
provided to compensate for other ambient temperatures.
The insulation outside surface temperature will be somewhat lower than the copper
conductor temperature and somewhat higher than the ambient temperature. One of the
steps in calculating current rating is determining the temperature rise of the insulation sur-
face with respect to the ambient. Depending on the installation conditions, the tempera-
ture rise above ambient varies from around 35°C to 52°C for common power wires used
in telecommunications.
Where wires are close together, bundled or layered, each wire influences the tempera-
ture rise of the other wires. This complicates the calculations and it is necessary to make
simplifying assumptions. Specific simplifying assumptions are described in each subsec-
tion but, in all cases, it is assumed that:
앫 Wires are installed on horizontal cable racks.

앫 No heat is conducted from the wires to the cable rack.
앫 No air movement takes place around the cable racks except as caused by natural
convection (there is no forced air movement).
앫 Wires are Class B (concentric coarse strand) and Class I (bunch or rope fine strand)
constructions with conventional insulation or insulation and outer jacket (no shield-
ing or armoring).
앫 The thermal resistance of insulation and jackets is the same.
앫 The thermal resistance of the copper conductor is so small compared to other ther-
mal resistances (insulation, jackets, air) that it can be neglected.
앫 Where more than one wire is involved, all wires are assumed to be the same size
and to carry the same current.
앫 Wires are single conductors and are not cabled into multiple conductor cables under
a common sheath or common jacket.
In load equipment that uses an isolated return (see Section 5.9.2), the feed and return
conductors of a given circuit will carry the same current but in opposite directions. In old-
er systems with equipment that uses an integrated return, the return conductors may carry
less current than the feed conductors (the steel infrastructure and frame ground conduc-
tors carry some if not most of the return current). However, regardless of the type of re-
turn used, a conservative design will assume that currents are equal in the feed and return
conductors.
The ac ampacities used in the electrical industries are traceable to work published in
1957 by Neher and McGrath [31] in the American Institute of Electrical Engineers
(AIEE) Transactions.22 Both IEEE Standard 835, IEEE Standard Power Cable Ampacity
Tables [32], and NFPA 70, National Electrical Code (NEC) [29], ampacity tables refer to
[31]; however, the ampacity values in these documents are for ac at 60 Hz and, as such,
are not directly usable in dc circuits. The ac ampacities include skin effects and proximity
effects and losses due to alternating current that do not apply to direct currents.
Because there must be thermal equilibrium between the heat generated inside the wires
and the heat dissipated by the wires, the wire currents may be determined by solving the
22
The AIEE merged with the Institute of Radio Engineers (IRE) in 1963 to form the Institute of Electrical and
Electronics Engineers (IEEE).
300 SYSTEM DESIGN
heat transfer equations for a given maximum conductor temperature and ambient temper-
ature. The phrase wire mass is used to indicate stacked, layered, or bundled wires with no
intentional separation.
To allow easy cross reference to current engineering references on power cables
[33–36], the equations in this section are based on length dimensions in meters. Since al-
most all wire catalogs and data sheets from manufacturers in the United States still use di-
mensions in inches and feet, the necessary conversion factors are:
To convert length in meters to inches, multiply by 39.37 in./m

To convert length in inches to meters, multiply by 0.0254 m/in.
To convert length in meters to feet, multiply by 3.281 m/ft
To convert length in feet to meters, multiply by 0.3048 ft/m
Other common length units, such as millimeter and centimeter can be used if caution is
exercised in their application. Thermal coefficients, such as thermal resistivity and heat
transfer coefficients, must be in compatible units, and care must be taken with tempera-
tures. Where a temperature difference (such as a temperature rise) is used in a calculation,
the temperatures may be in Kelvin (K) or Celsius (°C) as long as both are the same units.
The subscript in the equations indicates where temperatures must be in Kelvin (e.g., TsK).
Note that a temperature in Celsius (°C) is converted to absolute temperature in Kelvin (K)
by adding 273.15 (i.e., K = °C + 273.15).
The heat generated per unit length in a single wire or group of wires is
Ww = nI w2 Rw (5.32)
where Ww = heat dissipated per unit length by the wire (W/m)

Iw = dc current in the wire (A)
Rw = dc resistance per unit length of the wire (/m)
n = number of wires in a bundle or layered configuration
= 1 for single wire
> 1 for bundled or layered configurations
To convert Rw from ohms/meter to ohms/feet, multiply by 0.3048 m/ft

To convert Rw from ohms/feet to ohms/meter, multiply by 3.281 ft/m
For thermal equilibrium, the heat generated by resistive losses in the wire must equal
the heat dissipated by convection and radiation, or
Ww = Wconv + Wrad (5.33)
where Wconv = heat dissipated by convection (W/m)

Wrad = heat dissipated by radiation (W/m)
The temperature rise from ambient to the surface of a hot wire or wire bundle is pro-
portional to the heat dissipated by convection and radiation. The concept is analogous to
voltage drop across a resistance due to current flow through the resistance. In the case of
heat flow, the temperature difference (or thermal potential difference) across a thermal re-
sistance is due to the heat flow through the resistance. The relationship for heat flow is
Tmax – Ta = (Wconv + Wrad)(TRTotal) (5.34)
where Tmax = maximum allowed temperature (K)

Ta = ambient temperature (K)
TRTotal = Total thermal resistance from the hottest location to ambient air (m-K/W)
The thermal resistance is a measure of how a material or medium resists heat flow.
Nonconducting materials in a wire, such as the insulation and jacket (if any) and the air
surrounding the wire, resist heat flow away from the conductor. The following subsec-
tions provide analysis of the thermal resistances for the various physical wire configura-
tions. The current ratings are then found by iterative calculations using a programmable
calculator or a personal computer running a math analysis program. Example code is pro-
vided in Appendix B.
Wires in Free Air Individual wires installed on ladder-type cable racks can be consid-
ered to be in free air if they are separated by a distance wire outside diameter. The heat
in the conductor due to resistive losses flows to the inside surface of the insulation,
through the insulation to the outside surface, and then to the ambient air. The thermal re-
sistance of the copper conductor is very small compared to the insulation. The tempera-
ture rise from the conductor to the ambient air is
Tmax – Ta = Ww(TRins + TRsa) (5.35)
where Tmax = maximum allowed temperature (K)

TRins = thermal resistance of the insulation (m-K/W)
TRsa = thermal resistance from the surface to ambient air (m-K/W)
The thermal resistance of the cylindrical insulation shell is given by [33]
ins 2tins
TRins = ln 1 +
2 冢
Dc 冣 (5.36)
where ins = thermal resistivity of insulation (m-K/W) (typically 3.5 to 5.0 m-K/W for
thermoset, thermoplastic, and rubberlike insulation materials)23
tins = insulation thickness (m)
Dc = conductor diameter (m)
The external thermal resistance of the air surrounding the insulation surface is given by [33]
1
TRsa = (5.37)
Dehcr(Ts – Ta)1/4
where De = external or outside diameter of the wire (m)

hcr = heat transfer coefficient (W/m2-K5/4)
23
Thermoset is a plastic material that undergoes a nonreversible chemical reaction when it is cured by heat (or
pressure) and cannot be melted by subsequent heating. A thermoplastic material hardens when cooled but will
soften when reheated.
302 SYSTEM DESIGN
Ts – Ta = temperature rise of wire surface above ambient (K)

Ts = temperature at conductor surface (K)
Ta = ambient air temperature (K)
The heat transfer coefficient, hcr, in Eq. (5.37) includes the effects of both convection and
radiation. For a wire in free air, the heat transfer coefficient may be found from [33]
Z
冢
hcr = g + E
(De) 冣 (5.38)
where Z = 0.21
g = 0.60
E = 3.94
The temperature rise of the surface with respect to the ambient air, (Ts – Ta), can be
calculated from the following iterative process (see also Appendix B):
1. Define the coefficient KA from
KA = Dehcr(TRins) (5.39)
2. Substitute KA from step 1 in
Tmax – Ta
(Ts – Ta)n+1 = (5.40)
1 + KA(Ts – Ta)1/4
n
3. Start the iteration by substituting a convenient value for (Ts – Ta)n and calculate a
value for (Ts – Ta)n+1
4. Use this as a new value for (Ts – Ta)n and recalculate step 2
5. Repeat step 4 until the absolute value of the difference between (Ts – Ta)n+1 and (Ts
– Ta)n 0.001.
Example 5.25 Find the surface temperature of a jacketed 4/0 AWG wire operated at a
maximum temperature of 90°C and an ambient temperature of 30°C. Assume the thermal
resistivity of the insulation is 5.0 m-K/W.
Solution The following dimensions apply to jacketed 4/0 AWG wire (from Table 5.15):
Conductor diameter, Dc = 0.528 in. = 0.528 in. × 0.0254 m/in. = 0.01341 m

Insulation thickness, t = is 0.125 in. = 0.125 in. × 0.0254 m/in. = 0.003175 m
Wire outside diameter, De = 0.778 in. = 0.778 in. × 0.0254 m/in. = 0.01976 m
The calculated heat transfer coefficient is
0.21
hcr = + 3.94 = 6.1517 W/m2-K5/4
(0.01976)0.6
The calculated value of TRins is

5.0 2 × 0.003175
冢
TRins = ln 1 + = 0.3085 m-K/W
2 0.01341 冣
The calculated value of KA in step 1 is
KA = × 0.01976 × 6.1517 × 0.3085 = 0.1178
Starting with an insulation temperature rise (Ts – Ta)n = 30 K, and substituting KA in step 2
gives
363.15 – 303.15
(Ts – Ta)n+1 = = 47.03 K
1 + 0.1178 × 301/4
Substituting this value for (Ts – Ta)n gives
363.15 – 303.15
(Ts – Ta)n+1 = = 45.85 K
1 + 0.1178 × 47.031/4
Continuing with the iteration until the difference between the results is 0.001 yields a
surface temperature rise of 45.92 K above ambient (since this is a temperature rise, it is
the same as a rise of 45.92°C). The results of this calculation indicate the temperature
profile; that is, the conductor temperature is 90°C, the outer insulation surface is (30 +
45.92 =) 75.92°C, and the ambient air temperature is 30°C.
Once the surface temperature is found using the above technique, the external thermal
resistance, TRsa, is found from Eq. (5.37), and the insulation thermal resistance, TRins, is
found from Eq. (5.36). The current rating is then calculated by combining Eq. (5.32) and
Eq. (5.35) and solving for the current, or
冪莦莦
Tmax – Ta
Iw = (5.41)
R (TR + TR )
w ins sa
Example 5.26 Determine the current rating of the jacketed 4/0 AWG wire in Example
5.25.
Solution The following information is known:
Tmax = 90°C (363.15 K)

Ta = 30°C (303.15 K)
TRins = 0.3085 m-K/W
The dc resistance from Table 5.13 for 4/0 AWG wire is 0.0000500 /ft at 20°C. This val-
ue must be converted to resistance in ohms/meter at 90°C. To convert from ohms/feet to
ohms/meter, multiply by 3.281, gving 0.0001641 /meter at 20°C. To find the resistance
at 90°C, use the conversion method described in Chapter 2, or
Rdc(90) = Rdc(20)[1 + 0.00393(90 – 20)] = 0.0001641[1 + 0.00393(90 – 20)]

= 0.0002093 /m at 90°C
304 SYSTEM DESIGN
The thermal resistance from the wire surface to the ambient air is
1
TRsa = = 1.0059 m-K/W
× 0.01976 × 6.1517(45.92)1/4
Substituting the known values in Eq. (5.39) gives
冪莦莦
363.15 – 303.15
Idc = = 467 A
0.0002093(0.3085 + 1.0059)
The foregoing procedures are used to develop the current ratings for single unjacketed
and jacketed wires in free air [Table 5.16(a)].
In some design problems, it is necessary to refer to the NEC for ampacity data. How-
ever, to be useful for dc, the ac ampacities in the NEC must be converted to dc current rat-
ings by multiplying factors that account for skin effect and proximity effect. The factors
are calculated from
冪莦
Rac
FSP = (5.42)
R dc
where FSP = skin effect/proximity effect multiplying factor

Rac = ac resistance per unit length (/m)
Rdc = dc resistance per unit length (/m)
Values for Rac and Rdc or their ratios may be found in [30] and [37]. The factors for ad-
justing ac to dc current ratings (FSP) and calculation results are shown in Table 5.16(b). It
should be noted that the dc current ratings in Table 5.16(b), as derived from NEC tables,
and the factors shown are lower than the calculated current ratings in Table 5.16(a). The
differences arise because the NEC ac ampacities were determined from [31] using differ-
ent assumptions than in the foregoing calculations and then adjusted for dc. Where it is
necessary to observe NEC requirements, the lower values should be used.
Two Bundled Wires Where two wires are bundled together (with lacing twine or cable
ties) with no separation between the two but separated from other wires by at least their di-
ameter, the heat dissipated by one wire is influenced by the other wire, and it is necessary
to reduce the current ratings from their free air values. The reduction is reflected in the heat
transfer coefficient of the bundled configuration. The combined heat transfer coefficient (in
W/m2-K5/4) includes the effects of convection and radiation and is of the familiar form
Z
hcr = g + E (5.43)
(De)
where [33] Z = 0.29
g = 0.50
E = 2.35
The procedures in the previous subsection for single wires in free air are used to develop the
current ratings for the two-wire bundle but with the new heat transfer coefficient [Table
5.16(c)]. These current ratings should be used only on two bundled wires if it can be en-
sured that other wires will not be bundled to them in future projects. If not, the current rat-
ings for rounded wire bundles or layered wires, as appropriate, should be used instead.
Wires with No Separation in One or More Layers Generally, when wires are operat-
ed at or less than 25% load factor (actual load current divided by rated free air current rat-
ing), resistive losses and self-heating are negligible and the wires can be installed in mul-
tiple layers with reduced or no spacing. At higher load factors, the self-heating is higher
and the current ratings must be reduced or the installation conditions must be changed to
compensate. The heat dissipation can be improved by increasing the wire spacing, in-
stalling fewer wires on the cable rack, derating their current rating, or a combination.
Where these changes are not practical, it is necessary to use current ratings based on a
wire mass as developed in this subsection.
The analytical methods for calculating cable tray (cable rack) ac current ratings were
proposed by Stolpe [38] in 1970. Stolpe’s method did not take into account load diversity
(i.e., variation in the loading of the wires of different circuits). Harshe and Black [39] pro-
posed an improved thermal model in 1994. These methods were incorporated in ampacity
tables in ANSI/ICEA P-54-440, Ampacity of Cables Installed in Cable Trays [40].24
Leake [41] proposed further modifications in 1996. The following discussion uses these
heat flow models.
The analysis of heat flow on a cable rack is simplified by assuming the heat flow is out
the top and bottom of the wire mass, and there is no heat flow out the sides of the wire
mass (i.e., the heat flow is one dimensional).
A convenient representation of the heat flow parameters is shown in Figure 5.37. As
current flows through a wire, it generates heat because of resistive losses. The heat flows
by conduction through the conductors and insulations to the surface of the overall wire
mass and then flows from the surface by convection and radiation to the surrounding air.
By limiting the wire currents to those determined by solving the heat transfer equations
for a given maximum wire mass temperature and a given ambient temperature, the wire
24
ICEA—Insulated Cable Engineers Association; this document also is known as NEMA (National Electrical
Manufacturers Association) WC 51.
Fig. 5.37 One-dimensional heat flow diagram. The illustration shows a cross section of the wire
mass; neat stacking of the wires is not a requirement for this analysis. Heat is assumed to flow to the
top and bottom surfaces only and not through the sides (a conservative assumption).
306 SYSTEM DESIGN
mass will reach thermal equilibrium and the hottest wire will not exceed its maximum
specified temperature. In principle, the wires can be any combination of sizes and can be
neatly stacked and laced or randomly laid in the cable rack.
The heat generated in any single wire is
Ww = I w2 Rw (5.44)
where Ww = heat dissipated per unit length by the wire (W/m)

Iw = current in the wire (A)
Rw = dc resistance per unit length of the wire (/m)
The total heat generated by all wires is

n
2
Wwm = I c1 2
Rc1 + I c2 2
Rc2 + · · · + I cnRcn = 冱 I n2 Rn (5.45)
1
where Wwm is the total heat per unit length and the subscripts on the right side indicate the
individual currents and resistances. This representation allows different or the same cur-
rents in the wires.
The total heat generated in the wire mass must equal the heat dissipated at its top and
bottom surfaces by convection and radiation, or
Wwm = Wconv + Wrad (5.46)
where Wconv = heat dissipated by convection (W/m)

Wrad = heat dissipated by radiation (W/m)
The wire mass is assumed to generate heat uniformly throughout its cross section. As-
suming no heat flows out the sides and equal amounts flow out the top and bottom of the
wire mass, the temperature at a point, x, between the top and bottom of the wire mass is
(Fig. 5.37)25
Wwmeff Wwmeff Wwmeff

T(x) = x2 + dx + Ts = (dx – x2) + Ts (5.47)
Awm × 2 Awm × 2 Awm × 2
where T(x) = temperature at distance x between the top and bottom of the wire mass (K)
Wwm = total heat per unit length generated in the wire mass (W/m)
Awm = cross-sectional area of the wire mass in which the heat is generated (m2)
eff = effective thermal resistivity of the wire mass including conductors, insula-
tions, and air in the interstices (m-K/W)
d = depth of wire mass (m)
Ts = surface temperature of cable mass (K)
The cross-sectional area of the wire mass is
Awm = wd (5.48)
25
For derivation, see, for example, [42, 43].
where Awm = total cross-sectional area (m2)

w = width of the wire mass (m)
d = apparent depth of the wire mass (m)
Substituting Eq. (5.48) in Eq. (5.47) and rearranging gives
Wwmeff
T(x) – Ts = (dx – x2) (5.49)
2wd
To find the location of maximum temperature in the cable mass, Eq. (5.49) is differentiat-
ed with respect to x and the result set equal to zero
T(x) Wwmeff
= (d – 2x) = 0 (5.50)
x 2wd
Solving for x gives x = d/2 for the location of maximum temperature, which is the center-
line of the cable mass, as would be expected.
The maximum temperature rise is found by substituting x = d/2 in Eq. (5.49), or
Wwm eff d2 d2 Wwm eff d2 Wwm effd

2冢
Tmax – Ts = – = =
2wd 4 2wd 冣 4 8w
(5.51)
The heat transferred by convection is
Wconv = hconv(Ts – Ta)As (5.52)
where Wconv = heat transferred per unit length by convection from the wire mass (W/m)
hconv = convection heat transfer coefficient from the overall wire mass to the sur-
rounding air (W/m2-K)
Ts = temperature of the cable mass surface (K)
As = total surface area per unit length of the top and bottom of the wire mass
surface (m2/m)
= 2 × width × length m2/m
In Eq. (5.52), the heat transfer coefficient only includes the effects of convection and has
different units than the heat transfer coefficient in the previous two subsections. The con-
vection heat transfer coefficient, hconv (in W/m2-K), may be calculated from [41]
hconv = 0.101(Ts – Ta)1/4 (5.53)
Alternately, a fixed value of 2.4 W/m2-K (0.223 W/ft2-K) may be used (as in ANSI/ICEA
P-54-440 [40]).
The heat transferred by radiation is
As (T 4sK – T 4aK)

Wrad = (5.54)
l
where Wrad = heat transferred per unit length by radiation from the cable mass (W/m)
308 SYSTEM DESIGN
= Stefan–Boltzmann constant
= 5.67 × 10–8 W/m2-K4 (5.27 × 10–9 W/ft2-K4)
= thermal emissivity coefficient of the cable mass surface
= 0.8 (based on field measurements)
TsK = absolute average cable mass surface temperature (K)
TaK = absolute ambient temperature (K)
The total heat transferred from the wire mass surface to the surrounding air is
Wsa = Wconv + Wrad (5.55)
and
Wsa = hconv(Ts – Ta) As + As (T 4sK – T 4aK) (5.56)
where Wsa is the total heat transferred per unit length from the wire mass surface to the
surrounding air (W/m).
For thermal equilibrium, the heat transferred from the surface to the surrounding air,
Wsa, must equal the heat generated in the wire mass, or
Wsa = Wwm (5.57)
The heat flow equations are solved by setting the ambient temperature and the maximum
temperature of the wire mass and then calculating the allowable heat generated per unit
length (W/m). The calculated heat generated then is divided by the cross-sectional area of
the wire mass to determine the allowable heat flux (heat per unit volume in W/m2-m). The
allowable heat flux applies to any wire or cross section in the wire mass. Since the cross-
sectional area of each wire size is known, the allowable heat generated per unit length by
each wire size can be found and the allowable current rating then can be calculated. Pro-
cedures are provided later in this subsection.
In some installations, the wires may be randomly laid in a cable rack rather than neatly
layered or stacked. For this situation, the concept of an apparent, or calculated, wire mass
depth is used. The apparent depth equals the sum of all the wire cross-sectional areas di-
vided by the width, or
n
Fpack
dapp =
w 冱1 niD i2 (5.58)
where dapp = apparent wire mass depth (m)

Fpack = wire packing factor
w = width of wire mass (m)
Di = outside diameter of each wire (m)
ni = number of wires of each diameter
The wire packing factor takes into account the interstices between the wires (Fpack = /4
= 0.786) or ignores them (Fpack = 1.0). With the latter value the wire cross sections
are taken as squares with side length equal to the wire diameter and is the value used
in ANSI/ICEA P-54-440 [40] to develop ac ampacities for wires and cables in cable
trays.
Where Fpack = 1.0, the apparent depth is
Fig. 5.38 Example where apparent cable rack fill depth (d) is < wire diameter (D).
n1D 21 + n2D 22 + · · · + nnDn2

dapp = (5.59)
w
Example 5.27 Determine the apparent depth of 10 unjacketed 750-kcmil wires in one
layer and separated by one diameter each in a 24-in. cable rack (Fig. 5.38).
Solution The usable cable rack width in this example is 23 in. Since the units are in
inches, the calculations will be made without converting to another unit system. The nom-
inal wire outside diameter is 1.22 in., and the apparent depth is
10(1.22)2
dapp = = 0.65 in.
23
In this example, the apparent cable rack fill depth is less than the diameter of the wires.
Example 5.28 Determine the apparent wire depth of thirty-six 750-kcmil wires in three
layers of 12 each with no spacing between wires on an 18-in. cable rack (Fig. 5.39).
Solution The actual width of the wire layers is 12 × 1.22 in. = 14.64 in. The apparent
depth is
36(1.22)2
dapp = = 3.66 in.
14.64
It should not be surprising that, in this example with neatly stacked layers, the apparent
depth is the same as the physical depth of the three layers (3 × 1.22 in. = 3.66 in.).
In terms of the apparent depth, the maximum temperature rise within the cable mass is
Fig. 5.39 Example of thirty-six 750-kcmil wires in three layers.

310 SYSTEM DESIGN
Wmaxeff dapp
Tmax – Ts = (5.60)
8w
where Wmax = maximum heat dissipated by cable mass per unit length (W/m)
Tmax = maximum temperature within the cable mass (K)
Ts = average temperature at the surface of the cable mass (K)
w = width of the cable mass (m), usually assumed to be cable tray width less
0.0254 m (1 in.)
eff = effective thermal resistivity of the cable mass (m-K/W)
dapp = apparent depth of the cable mass (m)
The maximum allowable heat flux in the wire mass is

Wmax
qw = (5.61)
Awm
where qw = heat flux (W/m2/m)

Wmax = maximum allowable heat generated per unit length by the wire mass (W/m)
Awm = cross-sectional area of wire mass in terms of apparent depth (m2)
The allowable heat generated in any wire is
Ww = qw Aw (5.62)
where Ww = allowable heat generated per unit length in an individual wire (W/m)
Aw = apparent cross-sectional area of wire (m2)
and
Aw = Dw2 (5.63)
where Dw it the outside diameter of insulated wire (in meters).

The current rating of a given wire size is found from
冪莦
Ww
Iw = (5.64)
Rw
where Iw = allowable current rating of an individual wire (A)

Ww = allowable heat generated per unit length in the wire (W/m)
Rw = dc resistance per unit length at maximum allowed wire temperature (/m)
The following is a set of iterative procedures that may be used to determine the current
rating of various wire sizes (see also Appendix B):
1. Determine the apparent wire mass depth from Eq. (5.59).

2. Set the required values for the ambient temperature (e.g., 30°C) and maximum
wire temperature (e.g., 75°C or 90°C).
3. Assume Wwm is initially zero.
4. Set a temporary variable, Wtemp, equal to Wwm.

5. Calculate the wire mass surface temperature Ts from Eq. (5.60). The first iteration
will yield a surface temperature equal to the maximum wire temperature used in
step 2.
6. Calculate the convection heat transfer coefficient hconv from Eq. (5.53).
7. Substitute the values in Eq. (5.56) and solve for the heat dissipated from the sur-
face Wsa.
8. Find the difference between Wtemp in step 4 and Wsa in step 7.
9. If the absolute value of the difference is > 0.001, increase Wwm by one-half the dif-
ference and repeat steps 4 through 9; if the difference is 0.001, the heat generat-
ed Wwm and surface temperature Ts are from steps 4 and 5, respectively.
10. Determine the heat flux, qw (in W/m2-m) from Eq. (5.61) by dividing Wwm by the
cross-sectional area of the wire mass from Eq. (5.48).
11. Determine the allowable heat generation, Ww, of a given wire size from Eq. (5.62)
by multiplying the heat flux, qw, from step 10 by the apparent cross-sectional area
of the wire Aw from Eq. (5.63).
12. Determine the allowable current for the given wire size from Eq. (5.64).
The foregoing techniques are used to calculate the current ratings of unjacketed and
jacketed 1/C wires in cable racks [Table 5.16(d)] using the following parameters:
앫 Cable rack width (w) = 24 in. (0.6096 m)

앫 Surface emissivity of cable mass () = 0.8
앫 Cable mass temperature (Tmax) = 75 and 90°C
앫 Ambient temperature (Ta) = 30°C
앫 Thermal resistivity of wire mass(eff)26 = 4.0 m-K/W (13.12 ft-K/W)27
앫 Apparent cable rack fill depth (dapp) = 0.0254 to 0.1016 m (1 to 4 in.)
Examination of the tables reveals that a few of the calculated cable rack current ratings
for large wires and shallow depths exceed the free air current ratings. To be conservative,
the table current ratings should be reduced to the free air current ratings to minimize the
possibility of hot spots in the cable layer.
The current ratings are based on the wire diameters shown in the tables. For wire of the
same gauge and construction but slightly different outside diameter, the current rating can
be determined from [40]
Dx
Ix = ITable (5.65)
DTable
where Ix = current rating of wire with different diameter (A)

ITable = current rating from Table 5.16(d) (A)
Dx = outer diameter of wire (same dimension units as DTable)
DTable = outer diameter of wire from table (same dimension units as Dx)
26
A constant overall thermal resistivity is used, which is consistent with [40]; however, in practice, the thermal
resistivity is slightly lower for large wires and slightly higher for small wires but the overall effect is small. For
a detailed discussion of thermal resistivity as applied to power cables, see [34, 36].
27
Value from [40].
312 SYSTEM DESIGN
Table 5.21 Allowable Heat Flux (watts/unit cross-sectional area/unit

length) for 30°C Ambient Temperature and 75°C and 90°C Maximum
Wire Temperaturesa
Allowable Heat Allowable Heat
Apparent Cable Flux, 75°C Flux, 90°C
Rack Fill Depth Maximum Maximum
(in.) (W/in.2/ft) (W/in.2/ft)
0.5 11.185 15.809
1 4.996 6.997
1.5 3.016 4.195
2 2.070 2.864
2.5 1.527 2.105
3 1.182 1.624
3.5 0.946 1.296
4 0.777 1.062
a
To convert from W/in.2/ft to W/m2/m, multiply by 1.966 × 10–4.
To convert from W/m2/m to W/in.2/ft, multiply by 5.085 × 103.
The calculated allowable heat fluxes for various apparent cable rack fill depths are giv-
en in Table 5.21. These values can be used to calculate the allowable current ratings for
wire diameters and resistances different from the current rating table values.
Example 5.29 Unjacketed 1/C 750-kcmil wires are installed in a 24-in. cable rack to an
apparent cable rack fill depth of 4 in. Assume all wires are equally loaded. Confirm the
current rating at 90°C given in Table 5.16(d).
Solution The following information is taken from Table 5.13 and Table 5.15 for 750-
kcmil wire and from Table 5.21 for 4-in. cable rack fill depth:
앫 dc resistance = 0.01798 /1000 ft

앫 Conductor diameter = 0.998 in.
앫 Insulation thickness = 0.110 in.
앫 Wire outside diameter = 0.998 + 2 × 0.110 = 1.22 in.
앫 Effective cross-sectional area (D2) = 1.488 in.2
앫 Allowable heat flux = 1.062 W/in.2/ft
The allowable heat generation is 1.488 in.2 × 1.062 W/in.2/ft = 1.580 W/ft; therefore,
冪莦冪莦
Ww 1.580
Iw = = = 296 A –3
R w0.01798 × 10
which agrees with the table value (except for small rounding error).
The foregoing current rating calculations are based on equal loading of all wires. If
some of the wires are lightly loaded, the heat generated by them will contribute less to
the overall temperature rise of the wire mass, and it is possible to operate the other
wires at a heavier loading. Such a condition can be described by a diversity factor,
which, for the purposes of this discussion, is the ratio of heat generated by the actual re-
Table 5.22 Current Rating Factors for Various Diversities and Cable Rack Fill Depths
Current Rating Multiplying Factor
Cable Rack Fill Depth
Diversity
Factor 1 in. 1.5 in. 2 in. 2.5 in. 3 in. 3.5 in. 4 in.
1.0 1.00 1.00 1.00 1.00 1.00 1.00 1.00
0.9 1.04 1.04 1.03 1.03 1.03 1.03 1.03
0.8 1.09 1.08 1.07 1.07 1.06 1.06 1.05
0.7 1.14 1.13 1.11 1.10 1.10 1.09 1.08
0.6 1.20 1.18 1.16 1.15 1.14 1.13 1.12
0.5 1.28 1.24 1.22 1.20 1.18 1.17 1.16
0.4 1.37 1.32 1.28 1.26 1.23 1.21 1.20
0.3 1.49 1.41 1.36 1.32 1.29 1.27 1.25
0.2 1.64 1.53 1.46 1.41 1.36 1.33 1.31
0.1 1.87 1.70 1.59 1.51 1.46 1.41 1.37
sistive losses in the wires to the heat generated when all wires are operated at their max-
imum ratings, or
Wres
FDiv = (5.66)
Wmax
where Wres = estimated heat dissipation based on actual resistive losses (W/ft)
Wmax = maximum allowable heat dissipation of the entire cable mass (W/ft)
Diversity factors and corresponding current rating multiplying factors are given in Table
5.22. Care should be taken when using low diversity factors in single-layer or shallow
arrangements to avoid overloading individual wires and causing hot spots.
The conductors in many telecommunications power circuits are oversized to meet volt-
age drop requirements and their actual current loading is a fraction of their current rating.
For example, say 4/0 AWG wires are used in a 200-A circuit (as determined by the over-
current protection) to limit voltage drop. The free air current rating of jacketed 4/0 AWG
wire at 75°C is 404 A (from Table 5.19), so the load factor is 200 A/404 A = 0.50, or 50%.
In many situations, the equipment manufacturer may require an overcurrent device that is
sized 150 to 200% of the continuous load to prevent false tripping. Ignoring voltage drop
for the moment, the wire used in such a circuit would be sized for the overcurrent protec-
tion and would have a load factor of 67 to 50%. The basic circuit design rules in this book
are based on a load factor no higher than 80%. Where the load factor of a circuit is 25% or
less, the heat generated by that circuit can be ignored in the current rating calculations.
Heavy loading and high load factors should be avoided in initial installations. Subse-
quent system expansion invariably will use existing cable racks and new wires probably
will be installed on top of or adjacent to existing wires. If this is done without considera-
tion of the original design, the current ratings of the existing lower layer possibly could be
decreased below the current they were originally designed to carry.
Example 5.30 Forty unjacketed 1/C 750-kcmil wires are uniformly stacked in an 18-in.
cable rack with 26 wires loaded to 215 A each and 14 wires loaded to 350 A each. Deter-
mine the current rating of the wires at 90°C.
314 SYSTEM DESIGN
Solution The unit area for unjacketed 1/C 750-kcmil wire is 1.488 in.2 and the total area
of all wires is 40 × 1.488 in.2 = 59.5 in.2. Assuming the full width of the cable rack (less 1
in.) is used, the apparent cable depth is 59.5 in.2 ÷ 17 in. = 3.5 in. The basic current rating
of 750-kcmil wire in a cable mass with calculated depth of 3.5 in. is taken from the near-
est depth value in Table 5.16(d).2 (3.5 in.), or 317 A. The maximum allowable heat flux
for 3.5-in. fill depth from Table 5.21 is 1.296 W/in.2/ft and the total allowable heat gener-
ation is Wwm = 1.296 W/in.2/ft × 59.5 in.2 = 77.1 W/ft. With the current loading given, the
actual heat generated is (0.01798 ÷ 1000 ft) × (350 A)2 = 2.203 W/ft for each heavily
loaded wire and (0.01798 ÷ 1000 ft) × (215 A)2 = 0.831 W/ft for each lightly loaded
wire. The total heat generated based on this current loading is (26 × 0.831 W/ft) + (14 ×
2.203 W/ft) = 52.4 W/ft. The diversity factor is (52.4 W/ft) ÷ (77.1 W/ft) = 0.7 (rounded).
Table 5.23 summarizes the calculations so far.
From Table 5.22, the current rating factor is 1.09 for 3.5-in. cable rack fill depth and
0.7 diversity factor. This factor is applied to the current rating from Table 5.16(d).2, or
1.09 × 317 A = 346 A. Therefore, with diversity taken into account, the current rating of
the heavily loaded wires is slightly less than their actual loading.
Wires in Rounded Bundles Where a group of wires is bundled together for manage-
ment purposes on cable racks, the heat is transferred in the same ways as described in the
previous subsection. For purposes of calculating the current ratings of wires in a rounded
bundle, all wires are assumed to be the same size and construction (either all jacketed or
all unjacketed) and to carry the same current. The bundled configuration includes air gaps
between the wires, or interstices. Although in practice the bundle may have an irregular
shape, it is assumed to be circular for the analysis that follows. All wire bundles in
telecommunications circuits contain an even number of wires (each circuit in the bundle
consists of paired feed and return wires). The wires are assumed to be laid straight with no
crossovers. The analytical development below generally follows [44].28
The maximum temperature rise from the bundle surface to the center, in terms of the
overall thermal resistance of the wire bundle, is
Tmax – Ts = Ww(TRb) (5.67)
where Tmax = maximum temperature at the wire bundle center (K)

Ts = temperature at the surface of the wire bundle (K)
Ww = heat generated per unit length in the wire bundle (W/m)
TRb = thermal resistance of wire bundle (m-K/W)
Assuming the heat is generated uniformly throughout the wire bundle, the radial tempera-
ture distribution of a cylindrical heat source is given by [42]
Ww eff r2
T(r) = 1 –
4 冢
r 2eb
+ Ts 冣 (5.68)
where eff = effective thermal resistivity of the wire bundle including the conductors, in-
sulations, and air gaps (m-K/W)
28
In [44], a wire located in the bundle center is assumed to be the hottest and the ampacity calculation includes
the effects of this conductor and its insulation. However, because bundles consisting of even numbers of wires
do not necessarily have a center conductor, this effect is ignored in this subsection.
Table 5.23 Example Loading with Three Layers of 750-kcmil Wire on 18-in. Cable Rack
Unit Heat Total Heat
Unit Area Total Area Load Current Generated Generated
Wire Size Quantity (in.2) (in.2) (A) (W/ft) (W/ft)
750 kcmil 26 1.488 38.7 215 0.831 21.6
750 kcmil 14 1.488 20.8 350 2.203 30.8
Total 40 59.5 52.4
52.4
Apparent depth (in.) = Total area ÷ 18 in. = 3.3 Diversity factor = = 0.7
77.1
= 4.0 m-K/W (13.12 ft-K/W)

reb = wire bundle outside radius (m)
= Deb/2 (m)
The maximum temperature occurs at the center where r = 0 [this may be confirmed by dif-
ferentiating Eq. (5.68) with respect to r, setting the result equal to zero and solving for r].
Substituting r = 0 gives the maximum temperature rise with respect to the bundle surface, or
eff
Tmax – Ts = Ww (5.69)
4
Therefore, the thermal resistance of the wire bundle is
eff
TRb = (5.70)
4
The total cross-sectional areas of wire insulations, copper conductors, and air gaps influ-
ence the effective thermal resistivity of the wire bundle. For calculation purposes, it is
considered to be the same as described in the previous subsection for wires with no sepa-
ration in one or more layers.
The temperature rise of the wire bundle surface over the ambient air is
Ts – Ta = Wb(TRsa) (5.71)
where Wb is the heat generated per unit length of wire bundle (W/m).
The thermal resistance of the air surrounding the wire bundle surface is similar to that
previously given for a cylindrical surface, or
1
TRsa = (5.72)
Debhcr(Ts – Ta)1/4
where TRsa = thermal resistance from the bundle surface to the surrounding air (m-K/W)
Deb = diameter of the wire bundle (m)
hcr = combined heat transfer coefficient including convection and radiation
(W/m2-K5/4)
Ts – Ta = temperature rise of wire surface above ambient (K)
Ts = temperature at conductor surface (K)
Ta = ambient air temperature (K)
316 SYSTEM DESIGN
The combined heat transfer coefficient for a circular bundle of wires in free air, hcr, in
W/m2-K5/4, is similar to the coefficient for a single wire, or
Z
hcr = g + E (5.73)
(Deb)
where Z = 0.21
g = 0.60
E = 3.94
Substituting Eq. (5.72) in Eq. (5.71) gives
1
Ts – Ta = Wb (5.74)
Deb hcr(Ts – Ta)1/4
Combining terms gives
Wb
(Ts – Ta)5/4 = (5.75)
Debhcr
and solving for the temperature difference gives
W b4/5 1
Ts – Ta = = Wb (5.76)
(Deb hcr)4/5 (Deb hcrW 1/4
b )
4/5
Therefore,
1
TRsa = (5.77)
(DebhcrW 1/4
b )
4/5
To calculate the wire bundle diameter, Deb, it is assumed that the bundle cross-sectional
area is the sum of the cross-sectional areas of the individual wires with a packing factor of
1.0 (to be consistent with [40]). Therefore,
D 2eb
Ab = = nDw2 (5.78)
4
where Ab = cross-sectional area of wire bundle (m2)

Dw = outside diameter of insulated wire (m)
n = number of wires in bundle
and
冪莦 D
n
Deb = 2 w (5.79)
The total temperature rise from the ambient air to the center of the wire bundle (assuming
the heat is generated uniformly throughout the bundle) is
Tmax – Ta = Wb(TRT) (5.80)
where
TRT = TRb + TRsa (5.81)
Substituting Eq. (5.70) and Eq. (5.77) in Eq. (5.81) and then substituting the result in Eq.
(5.80) gives
eff 1
冤
Tmax – Ta = Wb +
4 (DebhcrW 1/4
b )
4/5 冥 (5.82)
The maximum allowable heat flux in the wire bundle is
Wb
qw = (5.83)
Ab
where qw = heat flux per unit length (W/m2-m)

Wb = maximum allowable heat generated per unit length in the wire bundle (W/m)
Ab = cross-sectional area of wire bundle (m2)
Since the heat flux is the same in any cross section within the bundle, the allowable heat
generated per unit length in any individual wire is
Ww = qwAw (5.84)
where Ww = allowable heat generated per unit length in an individual wire (W/m)
Aw = apparent cross-sectional area of wire (m2)
and
Aw = Dw2 (5.85)
where Dw it the outside diameter of insulated wire (meters).

The current rating of a given wire size is found from
冪莦
Ww
Iw = (5.86)
Rw
where Iw = allowable current rating of an individual wire (A)

Ww = allowable heat generated per unit length in the wire (W/m)
Rw = dc resistance per unit length at maximum allowed wire temperature (/m)
The wire current rating in the bundled configuration can be calculated as follows (these
procedures are similar to those described in the previous subsection):
1. Determine the wire bundle equivalent diameter from Eq. (5.79).

2. Calculate the combined heat transfer coefficient hcr from Eq. (5.73).
3. Determine the required values for the ambient temperature (for example, 30°C)
and maximum wire temperature (e.g., 75 or 90°C).
4. Define a temporary variable, Wtemp, and set it to a small positive value (but not
zero).
318 SYSTEM DESIGN
5. Set Wb equal to Wtemp.

6. Using Wb from step 5, calculate a new value for Wtemp from
Tmax – Ta
Wtemp =
eff 1
+
4 (DebhcrWb1/4)4/5
7. Find the difference between Wtemp in step 6 and Wb in step 5.
8. If the absolute value of the difference is > 0.001, increase Wtemp by one-half the
difference and repeat steps 5 through 8; if the difference is 0.001, the heat gen-
erated, Wb (in W/m), is from step 5.
9. Determine the heat flux, qw (in W/m2/m), from Eq. (5.83) by dividing Wb by the
cross-sectional area of the wire bundle Ab.
10. Determine the allowable heat generated, Ww, in a given wire size from Eq. (5.84)
by multiplying the heat flux, qw, from step 9 by the apparent cross-sectional area
of the wire Aw from Eq. (5.85).
11. Determine the allowable current rating for the given wire size from Eq. (5.86).
Example 5.31 Find the current rating of eight bundled 2 AWG unjacketed wires if the
allowable insulation temperature is 90°C and the ambient temperature is 30°C.
Solution
Step 1 The outside diameter of unjacketed 2 AWG wire is 0.412 in. (see Table 5.15).
Therefore,
冢冪莦
8
冣
Deb = 2 0.412 = 1.31 in. = 0.03327 m
Step 2 The combined heat transfer coefficient is
Z 0.21
冤
hcr = g +E =
(Deb) 冥冢
0.033270.6 冣
+ 3.94 = 5.558 W/m2-K5/4
Step 3 The temperature variables are Tmax = 90°C (363.15 K) and Ta = 30°C (303.15 K).
Step 4 Initial Wtemp = 30 W/m.
Step 5 Wb = Wtemp = 30 W/m.
Step 6 New
Tmax – Ta 363.15 – 303.15

Wtemp = =
eff 1 4.0 1
+ +

4 (DebhcrW b )
1/4 4/5
4 ( × 0.03327 × 5.558 × 300.25)0.8
= 54.52366 W/m
Step 7 Wtemp – Wb = 54.52366 – 30 = 24.523666.

Step 8 Since 24.523666 > 0.001,
24.523666
Wtemp = + 54.523666 = 66.785499 W/m
2
At this point, steps 5 through 8 are repeated until the difference between Wb and Wtemp is
0.001, which occurs when Wb is close to 60.055.
Step 9 The area of the wire bundle and heat flux are
D2eb 0.033272
冢冣冢
Ab = = = 0.000869 m2
4 4 冣
Wb 60.055
qw = = = 69,079.9 W/m2-m
Ab 0.000869
Step 10 The wire cross-sectional area is
0.412 × 0.3048
冢冣 = 0.0001095 m
2
Aw = Dw2 = 2
12
The heat generated in the wire is
Ww = qwAw = 69,079 × 0.0001095 = 7.565 W/m
Step 11 The dc resistance of 2 AWG copper wire at 90°C (from Table 5.13) is 0.20274
/1000 ft, or 0.00066519 /m.
The allowable current rating is
冪莦冪莦
Ww 7.565
Iw = = = 106.6 A
R 0.00066519
w
The foregoing techniques are used to calculate the current ratings of unjacketed and
jacketed wire bundles in cable racks [Table 5.16(e)] using the following parameters:29
앫 Maximum wire temperature (Tmax) = 75 and 90°C (348.15 and 363.15 K)

앫 Ambient temperature (Ta) = 30°C (303.15 K)
앫 Thermal resistivity of wire bundle (eff) = 4.0 m-K/W (13.12 ft-K/W)
Ambient Temperature Correction To determine temperature correction factors when

the wire configurations are operated at ambient temperatures other than 30°C, it is neces-
sary to briefly analyze how the ambient temperature affects heat dissipation of a given
wire configuration. The basic relationship between temperature rise, heat generated, and
thermal resistance is
Tmax – Ta = Wconf (TRT) = I 2nR(TRT) (5.87)
29
The current rating found in Example 5.31 differs slightly (by 0.6 A) from the value given in Table 5.16(e).3
due to rounding errors in the manual solution shown.
320 SYSTEM DESIGN
where Tmax = maximum wire temperature (K)

Wconf = heat generated per unit length in the wire configuration (W/m)
TRT = total thermal resistance of wire configuration (m-K/W)
n = number of wires in the configuration
I = current carried by each wire (A)
R = dc resistance per unit length of wire (/m)
The ratio of current rating at one ambient temperature (T2) to the current rating at another
ambient temperature (T1) is
冪莦莦
IT (Tmax – T2)R1(TRT1)
2 = (5.88)
IT1 (T – T )R (TR )
max 1 2 T2
The resistance at one temperature is related to the resistance at another temperature by
R2 = R1[1 + 1(T2 – T1)] (5.89)
where 1 is the temperature coefficient of resistance for copper at temperature T1 (/°C).

By substitution
冪莦莦莦
IT (Tmax – T2)R1(TRT1)
2 = (5.90)
IT1 (T – T )R [1 + (T – T )](TR )
max 1 1 1 2 1 T2
Temperature rise has subsidiary effects on the thermal resistances and can be ignored in
this analysis. Therefore, assuming no change in thermal resistances over the temperatures
of interest
冪莦莦
IT Tmax – T2
2 = (5.91)
IT1 (Tmax – T1)[1 + 1(T2 – T1)]
Ambient temperature correction factors for the current ratings in Table 5.16(a) through
Table 5.16(e) are developed by substituting the following values in Eq. (5.91):
Tmax = 75 or 90°C
T1 = 30°C
T2 = new ambient temperature (°C)
1 = 0.00378 /K at 30°C
Table 5.24 shows the resulting correction factors.
Example 5.32 Four 500-kcmil unjacketed wires are installed in a single layer on a cable
rack and are separated from each other by a distance equal to their diameter. The wire in-
sulation and terminations are rated 90°C, and the ambient temperature is 40°C. Determine
the wire current rating.
Solution The wires are in free air. From Table 5.16(a) for 500 kcmil unjacketed wire,
Table 5.24 Ambient Temperature Correction Factorsa

Ambient Temperature (T2) 75°C Current Rating 90°C Current Rating
(1) (2) (3)
21–25°C (70–77°F) 1.06 1.05
26–30°C (78–86°F) 1.00 1.00
31–35°C (87–95°F) 0.93 0.95
36–40°C (96–104°F) 0.87 0.90
41–45°C (105–113°F) 0.79 0.84
46–50°C (114–122°F) 0.72 0.79
51–55°C (123–131°F) 0.64 0.73
56–60°C (132–140°F) 0.55 0.67
61–65°C (141–149°F) 0.44 0.61
66–70°C (150–158°F) 0.31 0.54
71–75°C (159–167°F) — 0.46
76–80°C (168–176°F) — 0.37
a
For ambient temperatures from 21–80°C (70–176°F), multiply the current ratings in
Table 5.16(a) through Table 5.16(e) by the factors given in the table.
the current rating at 30°C is 834 A. From Table 5.24 for 90°C wire temperature and 40°C
ambient temperature, the correction factor is 0.9. Therefore, the current rating of the 500-
kcmil wire in free air at 40°C ambient temperature is (834 A × 0.9 =) 751 A.
Parallel Wires The current rating of parallel wires is the algebraic sum of individual
wire current ratings, where the individual wire current ratings depend on the installation
conditions. For example, if the paralleled wires all are in free air, then the free air current
rating would apply; if the paralleled wires are in a rounded bundle with other wires, the
current rating for each wire in the bundle would apply. The current rating of parallel bus-
bars is treated differently and is discussed in the next subsection.
N
Ip = 冱 II = II NCond (5.92)
1
where Ip = parallel current rating (A)

II = current rating of individual wire (A)
NCond = number of parallel wires
Example 5.33 Each side of a circuit consists of four jacketed 3/0 AWG wires in parallel
(i.e., four parallel wires in the feed side and four parallel wires in the return side). These
wires are neatly stacked in a 12-in. cable rack (effective width 11 in.) with 12 other jack-
eted 3/0 AWG wires. Determine the current rating at 75°C of the individual wires and the
parallel combination.
Solution The diameter of jacketed 3/0 AWG wire is 0.470 + 2 × 0.080 = 0.630 in. There
are a total of 20 wires and the apparent depth of the wire mass is
nD2i 20 × 0.630
dapp = = = 0.722 in., rounded to 1 in.
w 11
322 SYSTEM DESIGN
From Table 5.16(d).1 for jacketed 3/0 AWG and 1-in. depth, the current rating of each
wire is 161 A. The total current rating of four wires in parallel is 4 × 161 A = 644 A.
Rigid Busbar Current Ratings Since most busbar installations use bare, uninsulated
busbar, they do not have the same thermal limitations as insulated wire. Nevertheless, the
temperature rise for busbar system design normally is limited to 30°C above an ambient
of 40°C (70°C busbar temperature). Also, the current rating of parallel busbars does not
follow a simple linear relationship; that is, the current rating of N parallel bars is not N
times the current rating of a single bar. For example, the multiplication factor for four
bars is not 4.0 but is 3.2.
The proximity and emissivity of the parallel busbars determines how much heat they
dissipate for a given temperature rise above ambient. Outer bars obstruct convected and
radiated heat from the inner bars. Also, the current density in the bars is distorted by the
magnetic field interactions due to their close proximity (proximity effect). Emissivity is
the ratio of the radiation emitted by a surface to the radiation emitted by a blackbody at
the same temperature. A busbar with a higher emissivity can radiate more heat and subse-
quently has a higher current rating. For example, for a 30°C rise, the current rating of a
1/4 in. × 4 in. copper bar with 0.15 emissivity is approximately 1100 A, but if the emis-
sivity is increased to 0.4, the current rating is increased to 1400 A. Surface coatings and
treatments may be applied to copper busbars to increase or decrease their emissivity. For
example, the emissivity may be increased by applying a flat or a matt black finish. Table
5.25 compares the emissivity for various busbar surface conditions, and Fig. 5.40 shows
its effects on the current rating of single and parallel busbars.
The busbar current ratings shown in Table 5.17 for single bars are based on ac ampac-
ity data from the Copper Development Association, Inc. (CDA) [45], adjusted to dc cur-
rent rating as follows:
Idc = Iac兹S
苶 (5.93)
where Idc = dc current rating (A)

Iac = ac ampacity (A), data from CDA, Inc., at www.copper.org
S = skin effect ratio at 60 Hz, data from CDA, Inc., at www.copper.org
The busbar current ratings in Table 5.18 apply to parallel bar combinations and are based
on the single bar current ratings from Table 5.17 multiplied by the factors from Table
5.26. The mechanical design of busbar arrangements must account for the changes in
length that occur as the bars expand and contract with temperature changes. If not com-
pensated for in the mechanical design, the expansion and contraction can lead to stress
and damage to the busbar supporting structure and joints and splices and even to the bus-
bars themselves. The change in length is
Table 5.25 Copper Busbar Surface Emissivity Comparison

Busbar Surface Condition Emissivity
Bright metal 0.1
Partially oxidized 0.3
Heavily oxidized 0.7
Dull nonmetallic paint 0.9
Fig. 5.40 Effects of emissivity on single and multiple parallel busbar current rating. (Data from
Storm Copper Components Company.)
LT = LT0T0(T – T0) (5.94)
where LT = length change

LT0 = length at base temperature T0
T0 = coefficient of linear expansion (approximately 17.3 × 10–6/°C or 9.61 ×
10–6/°F over the range of ordinary engineering work)
For example, a 100-ft section of copper busbar will expand in length by 0.052 ft, or 0.62
in., if the busbar temperature rises 30°C.
Table 5.26 Current Rating Multiplying Factors for Parallel

Copper Busbar Spaced Equal to Their Thickness
Number of Parallel Bars Multiplication Factor
2 1.8
3 2.5
4 3.2
5 3.9
6 4.4
7 5.0a
8 5.5
9 6.0a
10 6.5
11 6.9a
12 7.3a
a
Values are interpolated or extrapolated using second-order polynomial
trend line analysis.
Source: Basic data from [45].
324 SYSTEM DESIGN
To minimize stress on the busbar and its supports, expansion joints or flexible sections
should be used on straight runs greater than approximately 30 ft. Flexible sections also are
used where overhead horizontal busbar runs connect with vertical busbars to equipment
below. The joints may use laminated thin copper strips or leaves, braided copper conduc-
tors, or Z-bent sections (Fig. 5.41). The expansion joints and flexible sections must have
the same total current rating as the busbars.
5.8.1.7 Conductor Terminations and Taps For extraordinarily long runs be-
tween the batteries and the charge and discharge buses, wire sizes and quantities may be
larger than can be conveniently terminated at either end. In such cases, the larger wires
can be tapped with smaller pigtails to make the actual bus connections. Another method is
to terminate the larger wires on a nearby collector or termination busbar and then run
short, smaller wire to the final termination. The pigtails must have adequate current rating
and, if kept short, their voltage drop will have negligible contribution. For seismic instal-
lations, the collector bars provide a convenient point to transition from busbar or coarse-
strand (Class B) wire to flexible (Class I) wire for termination at the battery or bus to pre-
vent strain on the terminations during a seismic event.
5.8.1.8 Fault Currents The battery system contributes more current during a fault
than any other component. The only other source of significant fault current is the rectifi-
er system, but, depending on the location of the fault, rectifier currents generally are an
order of magnitude smaller than battery fault currents.
Figure 5.42 shows typical short-circuit current magnitudes available at the terminals of
telecommunications batteries. Variations can be expected for different brands and for
models within a brand. As a rule of thumb, a battery can supply a fault current in amperes
Fig. 5.41 Copper busbar expansion joints (edge-on views).

Fig. 5.42 Battery short-circuit current vs. battery capacity.
approximately equal to 10× the 1- or 2-min rating to 1.75 V/cell. Since such short time
ratings are not always available on data sheets for telecommunications batteries, another
useful rule of thumb is 7.5× the ampere-hour rating of the battery (at its 8-h rate). Battery
manufacturers may also provide the maximum short-circuit current for a particular battery
type.
Where a battery system consists of more than one string, each string contributes fault
current on an additive basis. For example, if three identical strings are connected in paral-
lel and each string can supply 2000-A fault current, the total fault current is 6000 A. The
equivalent resistance of batteries in parallel is calculated the same as conductor resis-
tances in parallel. For example, three identical strings, each with 0.06- equivalent resis-
tance, have a parallel resistance of 0.02 .
The highest fault current occurs at the cell or battery terminals. The resistance of the
fault reduces the fault current, and if the fault occurs some distance from the battery, the
circuit conductor resistances significantly reduce the available fault current at the fault lo-
cation. Also, intervening circuit components, such as overcurrent protective devices, in-
troduce additional resistance that will reduce fault current levels. Of these, the circuit con-
ductor resistances usually have the largest effect. For worst-case calculations, the fault
resistance is assumed to be 0 .
Circuit inductance and capacitance affect fault rise time but they do not affect steady-
state dc fault currents. If it is necessary to analyze fault rise time, circuit capacitance can
be ignored. The inductance of two parallel conductors is
d
冢冣
L = 10–7 1 + 4 ln l henries
r
(5.95)
326 SYSTEM DESIGN
where L = inductance per unit length (H)

d = distance between centers of the two conductors (m)
r = radii of the two conductors (assumed to be the same) (m)
l = conductor length (m)
The fault current as a function of time is
V
i(t) = I0 + (1 – e–Rt/L) A (5.96)
R
where I0 = circuit current at the instant of fault initiation (A)

V = source voltage (V)
R = circuit resistance ()
t = time (s)
L = circuit inductance (H)
The ratio L/R is the circuit time constant, denoted by the symbol . When = t (i.e., the ra-
tio L/R = t), the current will change by (1 – e–1 =) 63.2% from its steady-state value.
Example 5.34 Determine the resistance, inductance, and time constant for two bundled
parallel insulated 4/0 AWG conductors (nominal inner diameter 0.53 in. and outer diame-
ter 0.70 in.) that are shorted 20 m from their source (assume the source has infinite am-
perage capability). If the initial current is 100 A and voltage is 50 V, plot the fault current
versus time until the current reaches its steady-state value.
Solution From Table 5.13 for 4/0 AWG at 30°C, the unit resistance is R = 0.00005197
)/ft, and the total resistance is R = 0.00005197 × 20 × 2 × 3.281 = 6.82 × 10–3 . From
the information given
d = 0.7 in. (0.0178 m)

r = 0.53/2 in. = 0.265 in. (0.00673 m)
and
d 0.0178
冢 r 冣冢 0.00673 冣
L = 10–7 1 + 4 ln l = 10–7 1 + 4 ln 20 = 9.77 × 10–6 H
The time constant is
L 9.77 × 10–6
= = = 1.43 × 10–3 s = 1.43 ms
R 6.82 × 10–3
The fault current plot is shown in Fig. 5.43. The steady-state fault current in this example
is
V 50
ISS = = = 7331 A
R 6.82 × 10–3
Fig. 5.43 Plot of fault current vs. time for Example 5.34.
When t/ = 1, the fault current has increased from its initial value of 100 A to 63.2% of its
final steady-state value, or 4,633 A.
For steady-state fault analysis, only the circuit resistance is relevant. An equivalent cir-
cuit of a typical system is shown in Figure 5.44.
There are two basic types of faults—line-to-line and line-to-ground. A line-to-line
fault includes both the feed and return conductors in the path from the fault to the current
source (battery or rectifiers). A line-to-ground fault includes the ungrounded feed conduc-
tors from the fault to the current source, but only those return conductors from the current
source to the point where they are bonded to ground. The remainder of the return fault
path usually is not known in any detail, particularly in a common bonding network
(CBN).
The dc power system design and installation directly affects the fault risk and the fault
current magnitude. Figure 5.45 shows four fault scenarios for a typical dc power system.
Referring to Figure 5.45, any conductive object, such as a tool, may cause fault 1,
shorting the current-carrying negative and positive buses (line-to-line fault). For purposes
of analysis, the short circuit can occur anywhere along a parallel path consisting of the
bus conductors, across rectifier output terminals and battery terminals, and distribution
equipment and hardware.
Fault 2 is similar to fault 1 but the short occurs between the negative bus in a –48-V
system (or positive bus in a +24-V system) and the equipment and frame grounds (line-to-
ground fault). Such a fault has a higher probability of occurring than other scenarios be-
cause of the many opportunities for inadvertent grounding. This type of fault frequently is
328 SYSTEM DESIGN
Fig. 5.44 The dc power system equivalent circuit (positive grounded system).
seen during cable mining operations, where old cables are being removed and the removal
process damages the insulation of in-service cables. If an overcurrent protection device is
installed at the battery, it would clear faults 1 and 2.
Fault 3 occurs between the load side of the primary or secondary distribution protec-
tion device and the return bus (line–line). Fault 4 occurs between the load side of the pri-
mary or secondary distribution device and the equipment and frame grounds
(line–ground). Faults 3 and 4 differ from 1 and 2 in that overcurrent protection devices in
the primary or secondary distribution circuits (depending on the fault location) should
clear faults 3 and 4. The primary and secondary overcurrent devices are normally coordi-
Fig. 5.45 Typical fault scenarios.

nated so they clear a fault with the minimum disturbance to system loads (i.e., they dis-
connect the minimum amount of equipment necessary to remove the fault from the sys-
tem). The overcurrent devices and the distribution conductors feeding the loads are locat-
ed close to each other and there are many opportunities for inadvertent short circuits
during installation or removal.
Example 5.35 Calculate the fault current resulting from the fault shown in Figure 5.46.
Assume the available fault current at the terminals of a 200-Ah, 48-V (24-cell) battery is
2000 A, the circuit conductor resistance from the battery to the powerboard is 0.013
(equivalent to a total of 50 ft of 4 AWG conductor), the circuit conductor resistance from
the powerboard bus to a line-to-line fault is 0.0105 (equivalent to a total of 40 ft of 4
AWG wire), and the fault resistance is zero.
Solution By definition, a line-to-line fault is a short circuit between the feed and return
leads and ground circuit resistances are not involved. The equivalent resistance of the bat-
tery is 48 V/2000 A, or 0.024 . The total resistance to the fault is the battery equivalent
resistance (0.024 ) plus the battery feed and return circuit conductor resistance (0.013
) plus the distribution feed and return circuit conductor resistance (0.0105 ) plus the
fault resistance (0 ), or 0.0475 . The fault current will be 48 V/0.0475 = 1010 A, a
50% reduction from the current available at the battery terminals.
The foregoing example assumes a line-to-line fault. In this case, the circuit conductor
resistances include both the feed and return conductors between the battery and the fault.
If the fault occurs line-to-ground, the fault return path and its resistance are not well de-
fined. The return path includes the parallel resistance of all conductive components from
the fault to the point where the battery circuit is bonded to ground. Since the battery cir-
cuit normally is not bonded to ground at the battery itself, the battery circuit conductor re-
sistances (both feed and return) usually are included in the fault path. The fault path also
can include cable supporting and equipment frame infrastructure, conductive building
components, and equipment grounding conductors and will depend on the type of bond-
ing network involved [common bonding network (CBN) or isolated bonding network
Fig. 5.46 Example 5.28 line-to-line fault.

330 SYSTEM DESIGN
(IBN)].30 For worst-case calculations, the return path resistance is assumed to be 0 , and
that is the value to be used in calculations when more accurate information is not avail-
able.
Example 5.36 Calculate the fault current resulting from the fault shown in Figure 5.47.
Assume the same conditions as the previous example except that the fault is line-to-
ground.
Solution In this case, the circuit conductor resistance from the fault to the powerboard
bus is 0.0053 (equivalent to 20 ft of 4 AWG wire), the battery circuit resistance is 0.013
(as before), and the fault resistance is zero. The resistance of the grounded metallic
components, including all connections to the ground bus, is assumed to be zero (worst-
case). The equivalent resistance of the battery is 0.024 (as before), and the total resis-
tance to the fault in this case is 0.0423 . The ground fault current is 48 V/0.0423 =
1,136 A, a 12% increase compared to the example line-to-line fault.
5.8.1.9 Short-Circuit Heating of Conductors Calculations of the conductor

temperature rise during a short circuit normally assume that all the heat generated is ab-
sorbed by the conductor and none is lost by convection and radiation as would be for a
conductor under continuous load. The temperature rise depends on the specific heat of the
copper conductor material and its mass. Specific heat is the amount of heat energy per
unit mass required to raise the temperature by one degree Celsius. The specific heat of
copper increases as its temperature increases, and at normal ambient temperatures it is
about 385 J/kg-K (0.092 Btu/lb-°F) and at 300°C is about 410 J/kg-K (0.098 Btu/lb-°F).
The energy required to raise the temperature over small ranges is approximately
J = sm T joules
where s = specific heat of copper (385 J/kg-K at normal ambient temperatures)

m = mass (kg)
T = temperature increase (K)
Example 5.37 Determine the energy required to raise a 50-ft piece of 2 AWG copper
conductor by 40°C from a normal ambient temperature. If the temperature is raised this
amount over a 10-s period and none of the heat is radiated or convected away from the
conductor, determine the current in the conductor.
Solution The equivalent diameter of a 2 AWG conductor is 0.292 in., or 7.42 mm, and
the length is 50 ft, or 15,240 mm. The cross-sectional area is (d2/4) = 43.2 mm2, and the
volume is 43.2 mm2 × 15,240 mm = 658,428 mm3, or 0.000658 m3. The density of copper
is 8230 kg/m3, so the mass of the conductor in question is 0.000658 m3 × 8230 kg/m3, or
5.42 kg (about 12 lb).
Assuming that the specific heat is constant over the temperature range, the energy is
approximately 385 J/kg-K × 5.42 kg × 40°C = 83,450 J. Since 1 J = 1 Ws, the energy in
electrical units is 83,450 Ws, or 23.2 Wh. The power required to raise the temperature
30
See Chapter 1 for a brief discussion and [46, 47] for a detailed discussion of common and isolated bonding
networks.
Fig. 5.47 Example 5.29 line-to-ground fault.
40°C over a 10-s period is 83,450 Ws ÷ 10 s = 8345 W. The power is dissipated in the
conductor resistance and is given by P = I2R.
The resistance of 2 AWG at 30°C is 0.00016525 /ft and the total conductor resis-
tance is 50 ft × 0.00016525 /ft = 0.00826 . Assuming it is constant over the tempera-
ture range of interest (it actually increases by about 16%), the current is approximately
冪莦R = 冪莦
P 8345
I= = 1005 A
0.00826
The foregoing concepts of specific heat and the assumptions concerning heat radiation
and dissipation can be generalized. The time required for the temperature to rise during a
fault is approximately [45]
A 2
冢冣
t = 0.051 (兹1
I
苶苶+苶.0
0苶0苶7
苶6
苶苶T
苶 – 1) s (5.97)
where t = maximum short-circuit time (s)

A = conductor cross-section area (mm2) (to convert CM to mm2, multiply CM by
0.000507)
I = Current (kA = A/1000)
T = temperature increase (°C)
Example 5.38 Check the previous example using Eq. (5.97).
Solution The conductor cross-sectional area and current were found to be 43.2 mm2 and
1005 A, respectively. For a 40°C temperature rise
43.2
冢冣 (兹1苶苶+苶.0
2
t = 0.051 0苶0苶7
苶6
苶苶×苶0
4苶 – 1) = 13.4 s
1.005
The 30% difference indicates the approximate nature of this analysis.

332 SYSTEM DESIGN
If the allowable temperature rise is = 300°C, which is the maximum allowable for
copper busbar under fault conditions, then Eq. (5.97) can be simplified to
A
冢冣s
2
t = 41.4 × 10–3 (5.98)
I
The value of t obtained from the above equation always should be greater than the re-
quired short-circuit withstand time. The withstand time typically is less than 1 s for cir-
cuits with overcurrent protection but can be several minutes for large battery circuits
with no overcurrent device. Where A/I > 4, the rate of temperature rise is approximate-
ly [45]
I
冢冣 °C/s
2
= 5.2 × 10–3 (5.99)
A
Example 5.39 Determine rise time to 300°C in a 4/0 AWG copper conductor if the cur-
rent is 6000 A.
Solution The cross-sectional area of a 4/0 AWG conductor is 211,600 CM, or 107.3
mm2. The rise time to 300°C is approximately
107.3
冢冣 = 13.2 s
2
t = 4.14 × 10–2
6
5.8.1.10 Selective Coordination of Overcurrent Protective Devices Three

methods have been used to perform selective coordination analysis in power systems:
앫 Overlaying printed delay characteristic curves on a light table and hand plotting on
log–log paper
앫 Using computer software to display the curves and select overcurrent devices
앫 For fuses, using published selectivity ratios
No computer software is available for low-voltage fuses or circuit breakers used in

telecommunications applications, nor are published fuse selectivity ratios available. This
leaves hand plotting as the only alternative.
Selective coordination analysis requires published delay characteristic curves for dc.
The ac delay characteristic curves cannot be used for devices that are rated both ac and dc
without first verifying with the manufacturer. Also, delay characteristic curves normally
represent average conditions and do not show the minimum/maximum time values for
any given current amplitude. This means that, absent specific information from the device
manufacturer, a number of assumptions have to be made.
The following discussion covers fuse–fuse coordination, circuit breaker–circuit break-
er coordination, and fuse–circuit breaker coordination.
Fuse–Fuse Coordination For a given type of fuse (TPL, TPS, or TPN), selective coor-
dination generally can be achieved by using a 2 : 1 ratio in their current ratings. For ex-
ample, a 400-A TPL fuse generally will coordinate properly with a 200-A TPL fuse, and a
70-A TPS fuse generally will coordinate properly with a 35-A TPS fuse. The same can be
said about coordination between TPL and TPS fuse types (a 100-A TPL fuse will coordi-
nate with a 50-A TPS fuse; see Fig. 5.48).
Circuit Breaker–Circuit Breaker Coordination As mentioned in Chapter 3, DC Pow-

er System Components, circuit breakers with the delay 52 curve are most common in
telecommunications, and that curve will be used here. Because of the trip time variations
for any given current value, two circuit breakers with even a high ratio of current ratings
will not coordinate for all fault currents. For example, as seen in Figure 5.49, a 50-A cir-
cuit breaker does not coordinate for any fault currents with a 100-A circuit breaker be-
cause the delay curves overlap. The same 50-A circuit breaker will barely coordinate with
a 200-A circuit breaker at 250 A (the lower trip curve for the 200-A circuit breaker brush-
es against the upper trip curve for the 50-A circuit breaker at 250 A), but for currents
above 350 A, the curves for the two breakers overlap.
Fuse–Circuit Breaker Coordination Fuses do not coordinate with delay 52 circuit

breakers unless the current ratios are > 4, and even then the delay curves can overlap at
high current levels (Fig. 5.50).
Fig. 5.48 Fuse–fuse coordination (100-A TPL and 50-A TPS).

Fig. 5.49 Circuit breaker–circuit breaker coordination (200, 100, and 50 A delay 52).
Fig. 5.50 Fuse–circuit breaker coordination (200 A delay 52 circuit breaker, 100-A TPL fuse, and
50-A TPS fuse).
334
5.8.2 Circuit Design Procedures

The circuit design procedures described in this section take into account the foregoing
discussions on wire and busbar conductors. Each circuit is designed with its specific re-
quirements in mind. The basic steps are to independently determine:
앫 Conductor size needed to meet voltage drop requirements

앫 Conductor size needed to meet current rating requirements
The conductor size selected for that circuit is the larger of the two. Other circuit design
considerations include selecting wire or rigid busbar and determining conductor termina-
tion requirements and routing. Some additional considerations are required for battery cir-
cuits, and these are described in the next section.
Manufacturers usually recommend the overcurrent device rating for circuits feeding
their equipment. Where the manufacturer does not recommend a circuit rating but does
provide load amperage or wattage ratings, the overcurrent device is chosen to avoid false
tripping. Generally, the rating is 150 to 200% of the maximum load current. In some
cases, the manufacturer may specify that the overcurrent device not exceed a certain rat-
ing, thus providing a boundary of ratings that must be observed. The conductor current
rating must be at least as high as the overcurrent device rating.
The voltage drop used to size circuit conductors is the segment voltage drop. Each seg-
ment must be coordinated with the voltage drop in the other segments so that the end-end
circuit does not exceed maximum total voltage drop (Fig. 5.51).
5.8.2.1 Circuit Design Summary The following steps allow systematic circuit de-
sign:
Rectifier and Distribution Circuits (Table 5.27):

1. List the normal and peak load currents for each circuit in columns (2) and (3), re-
spectively. Normal currents are for reference and are at normal system operating
voltage. Peak load currents are used to size conductors and are at an average dis-
charge voltage somewhat less than normal operating voltage. See Section 5.1.4.2.
2. Indicate the design voltage drop for each circuit in column (4). The voltage drop in
any given circuit segment depends on how the total voltage drop from battery ter-
minals to load equipment terminals is allocated. See Section 5.8.1.5.
3. Calculate the minimum wire size that meets the peak current rating requirement for
each circuit and list in column (5). Wire current rating = 1.25 × column (3). See
Section 5.8.1.6.
4. Calculate the minimum wire size that meets the voltage drop requirement for each
circuit and list in column (6).
5. Select the larger of the two wire sizes for each circuit and indicate in column (7).
6. Indicate the circuit breaker or fuse rating for each circuit in column (8). Standard
fuse and circuit breaker sizes are tabulated in Tables 5.28 and 5.29. Check that the
fuse or circuit breaker panel will accept the size in question.
Battery Circuits (see also Section 5.8.3) (Table 5.30):

1. Indicate the maximum battery circuit current in column (2). The maximum battery
circuit current occurs during battery recharge after a discharge and equals the max-
336
Fig. 5.51 Segment voltage drops. The total voltage drop of all segments must not exceed the total
specified for the system operating voltage.
Table 5.27 Distribution Circuits
Fuse or
Minimum Wire Size
Normal Peak Design Circuit
Circuit Load Load Voltage Drop Voltage Selected Breaker
Number Current (A) Current (A) (V) Current Drop Wire Size Size (A)
(1) (2) (3) (4) (5) (6) (7) (8)
1
2
3
4
5
6
7
8
9
10
11
12
Totals
Table 5.28 Standard Alarm Indicating Fuse Ratings

Rating (A) Rating (A)
GMT Color Type 70 Color
18/100 Yellow 1/10 Gray & white
1/4 Violet 15/100 Red & white
3/8 White & gray 18/100 Yellow
1/2 Red 2/10 Black
65/100 Black 1/4 Violet or violet & white
3/4 Brown 1/2 Red
1 Gray 3/4 Brown
1– 1–
21 3 White 1 Pink
1– 1– 1
– 11
21 2 White & yellow White
–
2 3
2 Orange 2 Orange
3 Blue 3 Blue
1– 1– 1
– 31
23 2 White & blue Black & white
–
2 2
4 White & brown 5 Green & black
5 Green 6 Green & white
1– 1–
27 2 Black & white 8 Brown & white
10 Red & white 10 Violet & yellow
a
12a Yellow & green
a
15a Red & blue
4AB (125 V and 250 V)b TPL
1, 2, 3, 5, 8, 10, 15, 20, 25, 30 A 70, 100, 150, 200, 225, 250, 300, 400, 500, 600 A
TPN
1, 3, 5, 6, 10, 15, 20, 25, 30, 35, 40, 45, 50, 60, 70, 80, 90, 100, 110, 150, 175, 200, 225, 250, 300,
350, 400, 450, 500, 600 A
TPS
1, 2, 3, 5, 6, 10, 15, 20, 25, 30, 35, 40, 50, 60, 70 A
a
Do not use unless fuse holder is rated 15 A.
b
Use in existing fuse panels only; do not use 4AG glass cartridge fuses.
337
338 SYSTEM DESIGN
Table 5.29 dc Circuit Breaker Ratings (Typical for Telecommunications Applications)a

1 20 50 125 300 700
3 25 60 150 350 800
5 30 70 175 400 1000
7.5 35 80 200 450 1200
10 40 90 225 500 1600
15 45 100 250 600 2000
a
Not all sizes may be available from a given manufacturer, and sizes not listed may be available from some man-
ufacturers.
imum rectifier system current. The battery will draw the maximum rectifier current
until its voltage increases and its current acceptance decreases. Also, the low-volt-
age disconnect, if equipped, will engage, reconnecting the load, when the bus volt-
age increases to the LVD reconnect threshold. At this point, the rectifier system
current may again increase to its maximum value. In any case, the maximum recti-
fier system current determines the current rating of the battery circuit conductors.
2. Indicate the design voltage drop for the battery circuit in column (3). This is the
maximum allowable voltage drop in the battery circuit segment during discharge.
See Section 5.8.1.5.
3. Calculate the minimum wire size that meets the current rating requirement for each
battery string. Under maintenance conditions, any battery string may be subjected
to the maximum rectifier system current; therefore, no current sharing by the bat-
tery strings is assumed. Indicate this wire size in column (4). Wire current rating =
1.25 × column (2). See Section 5.8.1.6.
4. Calculate the minimum wire size that meets the voltage drop requirement for each
battery circuit and list in column (5).
5. Select the larger of the two wire sizes for each battery string and indicate in column
(6). If the conductor lengths are the same to all battery strings, all circuits would use
the same conductor size and have the same voltage drop. However, battery strings
may be different distances and different conductor sizes may be needed in practice.
6. Calculate the actual voltage drop in each battery string circuit based on the selected
conductor size in column (6). If the voltage drops do not match within 5%, increase
conductor sizes or adjust conductor lengths to match. Recheck voltage drops.
Table 5.30 Battery Circuits

Minimum Wire Size
Design Selected Battery Circuit
String Battery Voltage Drop Current Voltage Wire Connection Method
Number Current (A) (V) rating Drop Size and Rating (A)
(1) (2) (3) (4) (5) (6) (7)
1
2
3
4
5
7. Indicate the battery connection method in column (7) for reference (see Section
5.8.3.1).
5.8.3 Battery Circuits

In addition to the design steps just described for distribution circuits, battery circuits re-
quire several other considerations, which are described in the following sections.
5.8.3.1 Battery Connection Methods A number of methods are available for ter-
minating and connecting battery circuits. The method chosen for any particular applica-
tion depends on a number of factors. There is no single best method.
Current can flow in either direction in a battery circuit depending on whether the bat-
tery is charging or discharging. The bi-directional characteristic complicates the circuit
design in terms of overcurrent protection and service reliability. In general, there are two
competing objectives in battery circuit design:
앫 Minimize the possibility of human injury and building and equipment damage dur-
ing battery circuit electrical faults.
앫 Minimize battery service interruptions.
A battery circuit consists of a connection device at or close to the battery terminals,

circuit conductors, and another connection device at the powerboard or battery charge bus
(Fig. 5.52). The connection device can be a simple bolted direct connection (rigid busbar
with Z-bent section or fine-strand wire with terminal connector lugs) or can include a
switch or an overcurrent protection device. A switch or overcurrent protection device, if
used, can be connected to only the ungrounded conductors in the circuit, using a one-pole
device, or both the ungrounded and grounded conductors, using a two-pole device. Where
a one-pole device is used, the grounded return circuit conductors (positive polarity on
–48-Vdc systems and negative polarity on +24-Vdc systems) use a bolted direct connec-
tion at both ends.
Fig. 5.52 Battery circuit connections.

340 SYSTEM DESIGN
Table 5.31 shows various devices and whether they provide overload protection, short-
circuit protection, and disconnection and whether they are subject to false tripping and
have remote disconnection capability. Although shown in the table, fuses seldom are used
because of the need for isolation switches and the time involved in replacing them (com-
pared to resetting a circuit breaker). However, current-limiting fuses may be required as
part of a circuit breaker or switch assembly where high fault currents exceed the circuit
breaker interrupt rating or switch withstand rating.
Remotely controlled disconnects are used to trip a circuit breaker or molded case
switch from a remote location, typically at the entrance to the battery room or at the
powerboard. Local fire codes may require a remote disconnect or the local fire marshal
may request it be installed. Circuit breakers and switches are available that are specifical-
ly designed for battery circuit applications, and they can be ordered with other features
such as remote position indication and remote reset and for mounting on the battery rack
or wall.31
The battery circuit can carry current in two directions—from the battery during dis-
charge or from the powerboard charge bus during recharge (Fig. 5.53). It is difficult to de-
sign reliable bi-directional overcurrent protection on such a circuit, especially when the
currents in the two directions can be significantly asymmetric. It is for this reason that
overcurrent protection seldom is used on battery circuits unless electrical codes or other
engineering considerations require it. If used, the overcurrent protection device or devices
must be rated for operation with bi-directional currents.
Where multiple battery strings are used, each string should be equipped with a dedicat-
ed disconnect switch so an individual string may be isolated during maintenance. An in-
stallation with a single battery string may or may not have a disconnect switch, and the
disconnect may or may not have overcurrent protection capabilities.
Figure 5.54(a) shows two possible bus and overcurrent protection configurations with
one-pole overcurrent protection at the battery end of the circuit and disconnect switches at
the bus end. Figure 5.54(b) shows similar battery configurations with one-pole disconnect
switches at the battery end and direct connection at the bus end. Figure 5.54(c) shows
configurations with two-pole circuit breakers and two-pole disconnect switches, respec-
tively.
5.8.3.2 Battery Disconnect Switches and Circuit Breakers Battery discon-

nect switches generally are found in three forms—enclosed knife switch, also known as
a safety switch, fused contactor [Fig. 5.55(a)] and molded case switch [Fig. 5.55(b)].
The latter looks similar to a molded case circuit breaker, which also may be used as a
battery disconnect [Fig. 5.55(c)]. The knife switch and molded case switch, by defini-
tion, do not have a built-in overload trip mechanism. However, unlike the enclosed
knife switch, the molded case switch may be equipped with an instantaneous trip mech-
anism (and called an instantaneous trip circuit breaker). The instantaneous trip mecha-
nism monitors the load current and opens during a short circuit but not during an over-
load.32 The enclosed fuse contactor is a large relay that uses a fuse to limit fault
currents.
31
See, for example, Airpax Power Protection Products of Cambridge, MD, Electric Equipment & Engineering
Company of Denver, CO, and Cutler-Hammer of Moon Township, PA.
32
Note that a molded case thermal circuit breaker has an overload sensing mechanism that monitors the thermal
energy passing through it and opens when the rated current is exceeded for the period of time specified by its
time–current or delay characteristic curve. See Chapter 3, DC Power System Components.
Table 5.31 Battery Connection Devices

Short Potential False Remote
Device Overload Circuit Disconnect Tripping Disconnect
Direct connection No No No No No
Circuit breaker without shunt trip Yes Yes Yes Yes No
Circuit breaker with shunt trip Yes Yes Yes Yes Yes
Fuse without switch Yes Yes No Yes No
Fuse with switch Yes Yes Yes Yes No
Enclosed knife switch No No Yes No No
Molded case switch without No No Yes No No
instantaneous current trip
Molded case switch with No Yes Yes No No
instantaneous current trip
Molded case switch with shunt trip No Yes Yes Noa Yes
a
A molded case switch with shunt trip is subject to inadvertent or accidental tripping (caused by a failure in the
trip circuitry or human error).
A disconnect switch normally is located adjacent to or on the battery string. It must be

able to carry (withstand) the available fault current, but it is not required to interrupt the
fault current unless the switch could open during a fault. Three parameters are used to
specify a disconnect switch:
앫 Voltage rating
앫 Nominal operating current rating
앫 Withstand current rating.
Switches used in 24- and 48-Vdc telecommunications systems are rated 60-Vdc class or
Fig. 5.53 Battery circuit current directions.

(a)
(b)
Fig. 5.54 (a) One-pole overcurrent protection at battery. (Upper) One circuit breaker for each bat-
tery string; (lower) one circuit breaker for entire battery system (not recommended). (b) One-pole dis-
connect switches at battery. (Upper) One disconnect switch for each battery string; (lower) collector
bar with one disconnect for each battery string.
(c)
Fig. 5.54 (c) Two-pole circuit breakers and two-pole disconnect switches at battery. (Upper) Two-
pole circuit breaker for each battery string; (lower) two-pole disconnect switch for each battery
string.
higher. Operating current rating will depend on the required currents during battery dis-
charge and recharge. The withstand current rating will depend on the short-circuit cur-
rents available from the battery system. The withstand current is the current the device
will carry and remain operable without damage (however, a device subjected to its rated
withstand current probably will require replacement and, in fact, such replacement may
be recommended by the device manufacturer).
A circuit breaker is similar to a molded case switch, but the circuit breaker also in-
cludes an overload trip mechanism. A similar set of parameters is used to specify a circuit
breaker:
앫 Voltage rating
앫 Nominal operating current rating
앫 Interrupt rating
The voltage and operating current ratings are the same as described for the switch. The in-
terrupt rating is similar to the withstand current rating, but in this case it is the current the
device can interrupt during a fault (a circuit breaker must be able to safely interrupt a fault
in progress, whereas a switch is not expected to interrupt a fault in progress).
344 SYSTEM DESIGN
The interrupt rating is the highest current the device can interrupt without damaging
the device or its surrounding structure or injuring people. When the device tries to inter-
rupt a higher-than-rated current, the device may (1) successfully interrupt the current, (2)
weld closed, (3) open without extinguishing the arc, or (4) open but not be able to dissi-
pate the arc energy. The first is the goal of a successful design while the latter three are
unacceptable and are safety risks.
5.8.3.3 Battery Connection Device Selection and Sizing The selection of a

battery connection device should account for the following factors:
앫 Probability of a fault occurring

앫 Costs of the connection device
앫 Consequences of a fault
앫 System availability requirements
앫 Battery current under normal discharge conditions
앫 Rectifier charging current
앫 Fault current magnitude
앫 Coordination with downstream protection devices
The probability of a short or failure in the battery circuit is lower where circuit conductors
are not very long and are not exposed to physical damage (such as when they are enclosed
in conduit) and are insulated or covered at their terminations. The risk is higher on long
runs and where terminations and conductors are exposed.
An overcurrent protection device at the battery should be considered in the following
situations:
앫 Probability of a fault occurring is high.

앫 The costs or adverse operational and safety consequences resulting from an un-
cleared fault are high.
앫 The consequences of false tripping are not significantly adverse.
앫 NEC requirements apply to the installation.
앫 The facility is unattended in a remote location or may require a long response time.
An overcurrent protection device normally is not used in the following situations:
앫 Risk of a fault occurring is low, such as in most central office environments.

앫 The consequences of false tripping are significantly adverse.
앫 Circuit equipment and installation costs are an overriding consideration.
앫 Circuit conductors and terminations are protected from accidental short circuits un-
der all operating and maintenance conditions.
If the connection device consists of a switch, with or without overcurrent protection, it

should be capable of remote position indication through a contact closure. Some circuit
breaker types provide an indication only when tripped open due to an overload and not
when manually opened. While this feature eliminates alarms when the switch needs to be
manually opened, it does not guard against problems with the switch being opened and
then forgotten.
One of the basic requirements of electrical circuit design is that conductors be protect-
ed from overcurrent and that the overcurrent protection device be located at the current
source. In most central office applications, to which the National Electrical Code (NEC)
does not apply, the battery connection device does not include overcurrent protection.
Historically, there have been few catastrophic (although not zero) failures in properly de-
signed and maintained battery circuits in central office environments. However, battery
circuit overcurrent protection is required in applications covered by the NEC.
The 2005 edition of the NEC [29] Article 90.2(B) lists installations not covered by the
NEC Par. 90.2(B)(4): “Installations of communications equipment under the exclusive
control of communications utilities located outdoors or in building spaces used exclusive-
ly for such installations.” An example of such an installation is a network operator’s cen-
tral office. Battery installations in all other facilities are covered by article 480 of the 2005
edition of the NEC. Article 480.3 says “Wiring and equipment supplied from storage bat-
teries shall be subject to the requirements of this Code applying to wiring and equipment
operating at the same voltage, unless otherwise permitted by 480.4.” Article 480.4 applies
only to starting and ignition batteries for “prime movers” such as engines. Article 240.4
applies to protection of conductors and states that “Conductors . . . shall be protected
against overcurrent in accordance with their current ratings specified in 310.15, unless
otherwise permitted or required in 240.4(A) through (G).” The cited paragraphs, 240.4(A)
through 240.4(G), do not apply to telecommunications applications.
Protection of battery circuits is discussed in additional detail in [48]. Overcurrent pro-
tection devices in battery circuits are subject to false tripping and almost always reduce
overall system availability. A battery circuit overcurrent protective device that has tripped
open may go undetected until an ac service outage occurs that affects the rectifiers. There-
fore, if an overcurrent device is used in a battery circuit, it should have a trip indicator and
contacts that can be connected to an alarm indicating and reporting system.
The minimum current rating of the battery connection device and battery circuit con-
ductors is determined by the worst-case battery current. If the system is equipped with a
low-voltage disconnect, the worst-case battery charge current occurs when the battery is
fully discharged and ac power is restored after an outage but before the low-voltage dis-
connect reconnects the load. In this situation battery current acceptance is highest and rec-
tifiers provide the maximum possible current to the battery (the rectifiers operate in their
current-limiting mode because a discharged battery looks like a short circuit to the recti-
fiers). The total rectifier current will depend on the number and rating of the rectifiers.
When calculating charging currents, it is necessary to take into account the maximum
possible rectifier capacity including the installation of rectifiers to fill initially unused
slots in modular rectifier shelves and the addition of rectifier shelves as the load grows.
Note that the worst-case battery charge current always is higher than the worst-case dis-
charge current because the rectifiers must have extra capacity, as much as or higher than
50 to 100%, to simultaneously recharge the batteries and power the loads.
Where the system is not equipped with a low-voltage disconnect, the rectifier system
provides current to both the dead battery and load equipment. However, since most mod-
ern load equipment uses switch-mode power supplies, the equipment load will be small
during the initial charging stage (when bus voltage is below the threshold for the power
346 SYSTEM DESIGN
(a)
Fig. 5.55 Battery disconnect devices: (a) enclosed battery disconnect contactor mounted on chan-
nel strut on the side of a battery rack (upper); cabinet door open showing components (lower). (Pho-
tos courtesy of Argus Technologies.)
supplies), and the battery system will accept almost all rectifier current as described in the
previous paragraph. The load current will increase as equipment power supplies are acti-
vated. The initial conditions dictate that battery circuits carry the maximum available rec-
tifier system current.
All rectifiers used in telecommunication applications have integral output circuit
breakers, and the fault current from the rectifiers will be limited to the overload capability
of the rectifier output circuits or the circuit breaker rating, whichever is smallest. There-
(b)
(c)
Fig. 5.55 Battery disconnect devices: (b) Molded case switch. (Photo courtesy of Electric Equip-
ment & Engineering Company.) (c) Enclosed circuit breaker.
fore, an external circuit breaker is not normally used in the battery circuit at the discharge
bus.
The currents described above are significantly higher than normal battery circuit oper-
ating currents. During normal conditions the battery circuit carries only float current. The
float current depends on the battery technology and capacity rating but is in the few mil-
liamperes to few hundred milliamperes range per 100-Ah battery capacity. Float current
increases with battery age but never approaches charge or discharge current magnitudes.
If an overcurrent protection device is used in the battery circuit, the protection device
348 SYSTEM DESIGN
rating usually is at least 150 to 200% of the worst-case charge current (total rectifier cur-
rent capacity at current limit). Where overload protection is avoided because of the poten-
tial for false tripping, a short-circuit protection device (instantaneous trip circuit breaker)
may be desirable to protect against short-circuit currents.
Example 5.40 Determine the battery overcurrent device rating for the circuit in Figure
5.56 and the following conditions:
앫 Equipment load = 40 A (worst case)

앫 Modular rectifier system = (2) online 50-A rectifiers [total capacity of (4) 50-A rec-
tifiers]
앫 Battery capacity 200 Ah
앫 Battery reserve time = 4 h
Solution The 200-Ah battery used in this example can deliver 46 A at 77°F (25°C) for 4
h. From Figure 5.42 the short-circuit current available from a 200-Ah battery is approxi-
mately 1500 A. The configuration provides N + 1 rectifier redundancy such that if one
rectifier fails, the other can carry the load plus provide some charging current to the bat-
tery. Because the rectifier system is modular, it will be easy to add two more rectifiers for
a nominal output current capability of 200 A.
Because a battery overcurrent device is used in this example, it is oversized by 150%
to prevent false tripping. The maximum current in the battery circuit during recharge
will be 200 A (assuming the rectifiers are adjusted to limit output current to 50 A).
Therefore, the minimum current rating for the battery circuit protection device in this
example would be 300 A (1.5 × 200 A). In this case, the required rating is a standard
size, but if the calculated rating is not a standard size, the rating must be increased to
the next larger standard size. In this example, the device must be able to interrupt a fault
current of at least 1500 A, and the conductors between the battery and the charge bus
must be rated at least 300 A.
Good design practice (and NEC requirements, where applicable) requires that conduc-
tor current rating be increased to at least 125% of the continuous current and that temper-
ature and other correction factors be applied when determining conductor size. If an over-
sized protection device is used to prevent false tripping, the conductors must be oversized
to match the protection device rating.
Fig. 5.56 Circuit for Example 5.33.

All devices in the battery circuit must be able to safely carry fault current. They must
have the mechanical strength to withstand electromagnetic reaction forces during a fault
as well as have the electrical and temperature ratings to withstand the fault current
levels.
The choice between a one-pole and two-pole battery disconnection device is based on
the operational and maintenance requirements. If it is desirable to simply provide a means
to isolate the ungrounded circuit from the battery for cell replacement or capacity testing,
a one-pole device is sufficient. However, if it is desirable to completely isolate both bat-
tery terminals so that insulation resistance measurements may be made between the bat-
tery terminals and ground, a two-pole device is required.
5.8.3.4 Battery Overcurrent Protection Device Characteristics and Coordi-

nation Typical ratings for devices used in battery circuits are 200, 400, 600, 1200,
1600 and 2000 A (other values are available). If a battery circuit has overcurrent devices
at each end, it usually is impossible to predict which one will open when there is a fault in
the battery circuit (i.e., a fault somewhere between the protection device at the battery and
the protection device at the discharge bus). Similarly, with parallel battery strings and an
overcurrent protection device on each string, it is almost impossible to ensure coordinated
operation.
In addition to the coordination issue, an operational problem exists with overcurrent
protection (circuit breakers) on parallel battery strings. If one of the battery strings is
taken offline for maintenance and allowed to self-discharge or is discharged for any oth-
er reason, large currents will flow into the battery string when its circuit breaker is
closed. This may cause the circuit breaker to immediately trip or it may trip later from
the high charge currents. Such a problem can be eliminated if the offline battery string
is precharged to the same potential as the online strings prior to closing the circuit.
Chapter 6, System Installation and Maintenance, describes a simple precharge tech-
nique.
Example 5.41 A system consists of three battery strings. Either one main disconnect
switch or three switches (one for each battery string) are to be installed. Determine the
switch ratings for both scenarios for the circuit in Figure 5.57 and the following condi-
tions:
앫 Initial and ultimate equipment load = 170 A (worst case)

앫 Initial rectifier system = (1) rectifier shelf with (4) online 50-A rectifiers
앫 Ultimate rectifier system = (2) rectifier shelves with (8) online 50-A rectifiers
앫 Initial and ultimate battery = (3) strings at 680-Ah capacity each
Solution Single main disconnect switch—The dc power system is designed to meet ini-
tial and ultimate requirements. Two rectifier shelves are installed but only (4) rectifiers
are initially equipped. Three battery strings are initially installed to meet the reserve re-
quirement. Each 680-Ah battery string can deliver 85 A at 77°F (25°C) for 8 h. From Fig-
ure 5.42, the short-circuit current available from each battery string is approximately 5000
A. The maximum charging current available to the battery is initially 200 A (4 × 50 A)
but ultimately 400 A (8 × 50 A).
350 SYSTEM DESIGN
Fig. 5.57 Circuit for Example 5.41, showing one main battery disconnect switch (upper drawing) or
three individual switches (one for each string, lower drawing).
A suitable main battery circuit disconnect switch would have a minimum current rat-
ing of 400 A and a short-circuit current withstand rating greater than 15,000 A (three
times the fault current available from an individual battery string).
Individual Battery Disconnect Switches—If a battery disconnect device is to be used
on each of the three battery strings, each one would require the same minimum current
rating (400 A) as a single main device because of the possibility of two batteries being of-
fline (disconnected for maintenance). If the online battery became discharged, it would
initially sink the full rectifier system current. An individual disconnect device on each
battery string will have to carry the fault current available from the other battery strings or
just the string associated with it (depending on the location of the fault). All conductors
between each battery and the charge bus must be rated 400 A minimum.
5.8.3.5 Parallel Strings Conductors connecting different battery strings to the dis-
charge bus should be balanced by sizing the conductors for equal resistance (and therefore
equal voltage drop) between the battery terminals and the bus. If they are unbalanced, the
string with the least voltage drop to the bus will provide more than its share of current
during each discharge and thus will have a shorter life and shorter reserve time than the
other battery strings.
To minimize and balance the voltage drop to parallel battery strings, the largest practi-
cal conductor size should be used for the most distant battery string. The conductor sizes
for the strings nearer to the discharge bus are then selected so the voltage drop in each is
approximately equal to that of the most distant string.
앫 If the same size conductors are used to connect each parallel string, the conductors
should be the same length (within 5%).
앫 The resistances of the battery circuits should be the same (within 5%). Different
conductor sizes may be used to achieve this requirement (note: in the case of paral-
lel conductors on a given battery string, all parallel conductors must be the same
size).
Since conductors are available only in limited and discrete sizes, it normally is not possi-
ble to make the voltage drops exactly the same in all runs. Conductors also must meet cur-
rent rating requirements regardless of voltage drop requirements.
Rather than use individual runs from each battery to the charge/discharge bus, collec-
tor bars may be used near the battery strings [Figs. 5.58(a) and 5.58(b)]. With this
arrangement, it is easier to control the conductor length and resistance of the individual
(a)
(b)
Fig. 5.58 (a) Battery circuit collector bar schematic and (b) battery circuit collector bar application.
Note spill containment kit near each battery rack. (Photo courtesy of Schultz Brothers Electric Com-
pany.)
352 SYSTEM DESIGN
battery circuits. Generally, the drop circuits from the battery collector bars have the same
current rating as the main run to the powerboard charge bus; however, for the wiring from
the battery terminal to the battery collector busbar, it is common practice to use 4/0 AWG
Class I (fine-strand) power wire as a basic size on battery strings up to around 1800-Ah
and 350-kcmil Class I wire on larger battery strings. See Table 5.32.
5.8.4 Power Cable Supporting Structure

5.8.4.1 Structural Considerations The cable racks and their attachments to the
building must support the weight of the power wires. Cable racks look like steel ladders
and sometimes are called ladder racks (Fig. 5.59). They may be run horizontally or verti-
cally. The normal length of a cable rack section is 9 ft 8.5 in. Cable racks vary from 5 to
24 in. wide, although sizes outside this range will be encountered in the field or found in
manufacturers’ catalogs. Cable racks wider than 24 in. are equipped with additional rein-
forcement on the first and every other rung or cross member. Bar-type cable racks, with
cross members welded to the top of the stringers, generally are not used.
The stringers, or rectangular side rails, can be tubular (hollow) or solid steel. Solid
steel stringers are stronger and should be used in all applications except where cable loads
are small and site logistics (shipping and handling) require light-weight materials. For ex-
ample, a 12-in. section of tubular steel cable rack weighs 23 lb (10.4 kg) while the same
size in solid steel weighs 61 lb (27.7 kg). Cable racks generally are supported at 5-ft inter-
vals, although 4-ft intervals may be used where extra support is required, and 6-ft inter-
vals may be used where site physical and dimensional constraints require them. Two
stringer dimensions are available—–38 in. × 1.5 in. and –38 in. × 2 in.; the 2 in. is used in most
applications, especially where structural strength is of concern. Many small installations
use 1.5 in. The choice sometimes is dictated by logistics. Table 5.33 shows the weights
for various stringer configurations.
In small central offices, it is common to install only one cable rack tier that is shared by
both signal and power cables. In larger central offices, two or three tiers usually are in-
stalled, one or two for power wires and another for signal cables, although the latter some-
times may be used to carry secondary distribution wires for short distances (Fig. 5.60).
Cable rack structural design can be handled by calculations or by simply limiting the
wire weight or cable pileup height. Either way the goal is to limit the deflection (sag) in
the loaded cable rack while providing a factor of safety against overloading.
The cable rack stringers essentially are two uniformly loaded, rectangular beams sup-
ported at intervals (Fig. 5.61). A simple analysis assumes that the stringer length is equal
Table 5.32 Battery Circuit Wiring from the Battery Terminal to Battery
Collector Busbar—Common Practice
Battery Circuit Conductors
Battery String Capacity Quantity in Conductor Size
(Ah at 8 h Rate) Parallel (AWG or kcmil)
< 450 1 4/0
450–900 2 4/0
900–1,800 4 4/0
> 1,800 4 350
Fig. 5.59 Cable rack section.
Table 5.33 Cable Rack Weights (Painted)a

1.5-in. Tubular 1.5-in. Solid 2-in. Tubular 2-in. Solid
Stringer Stringer Stringer Stringer
Cable Rack Weight Weight Weight Weight
Width (in.) (lb) (lb) (lb) (lb)
5 16.0 39.1 16.5 51.4
6 17.0 40.0 17.5 52.5
9 20.0 42.0 20.5 58.0
10 21.0 42.5 21.5 59.0
11 22.0 44.0 22.5 60.0
12 22.5 45.0 23.0 61.0
15 23.0 45.8 23.5 62.5
18 25.0 48.1 25.5 64.0
20 27.0 55.0 27.5 68.0
21 27.5 56.0 28.0 70.0
24 29.0 58.0 29.5 74.0
25 30.0 63.0 30.5 75.0
27 32.0 65.0 32.5 78.0
30 35.0 68.0 35.5 80.0
36 40.0 70.0 40.5 92.0
a
To convert to kilograms, divide weight in pounds by 2.2.
Fig. 5.60 Two-tier cable racks.

354 SYSTEM DESIGN
Fig. 5.61 Free-body diagram of showing loads on one of two stringers.
to the support interval (e.g., 48, 60, or 72 in.) and that the cross members transmit the load
uniformly to the stringers. The deflection of a simple rectangular beam with a uniform
load is
5WL3
d = in. (5.100)
384EIM
where d = deflection (in.)

W = total uniform weight (lb)
E = modulus of elasticity (29 × 106 lb/in.2 for mild steel)33
IM = moment of inertial (in.4)
L = distance between supports (in.)
The moment of inertia, IM, for a solid rectangular shape with the long side vertical is34
bh3
IM = in.4 (5.101)
12
where b = width of the base of the stringer (typically 0.375 in.)

h = height of the stringer (typically 1.5 or 2 in.)
For a stringer that is 0.375 × 2 in., the moment of inertia is 0.25 in.4. There are two
stringers in a cable rack and both share a common centroidal axis. Therefore, the total
moment of inertia for the two stringers is twice that for a single stringer, or 0.5 in.4.
The above deflection equation can be used for narrow cable racks (12 in. or less)
where the load can be assumed to be the same on the two stringers. This is not necessarily
true for wider racks where the load could be asymmetric.
Generally, the maximum loaded deflection is limited to 1/360 of the span length be-
tween supports. Table 5.34 shows the deflections for the typical support intervals. It is
necessary to calculate the failure load (1) to make sure that the weights do not cause high-
er deflections and (2) to determine safe loads, which include a safety factor. The maxi-
mum allowable bending moment can be determined from
MMax = FY S (5.102)
33
The modulus of elasticity is a measure of a material’s stiffness.
34
The moment of inertia is a measure of a body’s resistance to angular acceleration.
Table 5.34 Cable Rack Deflections

Support Interval Deflection = 1/360 × Interval
48 in. 0.13 in.
60 in. 0.17 in.
72 in. 0.20 in.
where MMax = maximum moment (lb-in.)

FY = yield point (35 × 103 lb/in.2 for mild steel)35
S = section modulus (in.3)36
The maximum moment for a uniformly loaded beam is
WMaxL wMaxL2
MMax = = lb-in. (5.103)
8 8
where WMax = maximum total load (lb)

wMax = maximum uniform load (lb/in.)
The section modulus for a simple rectangular cross section is
bh2
S = in.3 (5.104)
6
For the 0.375-in. × 2-in. stringer, S = 0.25 in.3, and the maximum moment, MMax = 35 ×
103 lb/in.2 × 0.25 in.3 = 8750 lb-in. It is now a simple matter to determine the maximum
loading for each of the common support intervals, or spans, by rearranging Eq. (5.103), or
8
WMax = MMax (5.105)
L
For a 48-in. support interval (span), the calculations give a maximum total load to failure
on one stringer of 1458 lb. Since this load is uniformly distributed along the full span
length, the maximum uniform load to failure on one stringer of a 48-in. span is 365 lb/ft.
The cable rack consists of two stringers, so the maximum loads to failure of the cable rack
would be twice the individual stringer values.
The simplifying assumptions do not account for the effect of the cross members on the
stringer strength nor do they account for any twisting, different mounting configurations,
uncertainty in installation practices and materials, and future installations where cables
may be added without detailed analysis of their effect. The calculated maximum loads to
failure are never used in practice. To account for uncertainty and simplifications, the val-
ues are adjusted by a safety factor. With two stringers and a safety factor of 2.0, the max-
imum useful loads for the cable rack are the same as the maximum failure loads for a sin-
gle stringer without any safety factor. Table 5.35 summarizes the calculated maximum
35
The yield point is the stress at which material deformation first increases markedly without any increase in the
applied load.
36
The section modulus is a measure of the capacity of a section to resist any bending moment that may be applied.
356 SYSTEM DESIGN
Table 5.35 Maximum Loads for 3–8-in. 2-in. Solid Stringer Cable Rack Including
Safety Factor = 2.0a
Support Interval Maximum Uniform Load Maximum Total Load
(in.) (lb/ft) (lb)
48 365 1,458
60 233 1,167
72 162 972
a
Loads include cable rack weight.
loads including a safety factor of 2.0, and Figure 5.62 shows the deflection of narrow ca-
ble racks for uniform distributed loads. It is unlikely that in any practical installation the
cable rack loads would approach the maximums shown except where many layers of large
power cables are installed in large central offices (Fig. 5.63).
One method of determining pileup is based on the assumption that the power wire is
750 kcmil, which has an outside diameter (including insulation) of approximately 1.35 in.
(34.3 cm) and weighs approximately 2.7 lb/ft (4.0 kg/m) of length. Obviously, not all
power wire is 750 kcmil but it represents a worst-case condition. For design estimating
purposes the useful width on a cable rack is 1 in. less than its nominal width (0.5 in. on
each side is taken up by the mounting hardware). Therefore, a 12-in. cable rack would
have 11 in. of useful width. For this example, a single layer of eight 750-kcmil wires
could be placed on a 12-in. cable rack (100% fill). Table 5.36 shows the allowed pileup
according to common industry practice.
Fig. 5.62 Deflections and loads for 3–8 in. × 2 in. solid stringer cable rack assuming uniform loading.
Fig. 5.63 Power cable layering in a large central office. (Photo courtesy of Schultz Brothers Electric
Company.)
For 12-in. and wider cable racks, the allowed pileup is 7 in., or five layers of 750-kcmil
power wire. A 12-in. cable rack could hold a total of forty 750-kcmil power wires with a
weight of 40 × 2.7 lb/ft, or 108 lb/ft. This forms the basis for Table 5.37, which shows the
weights on fully loaded cable racks of various dimensions. It should be noted that the al-
lowed loads in this table are much less than the previous structural calculations indicate
are possible and therefore are more conservative.
Table 5.36 Maximum Power Cable Pileup, Common Practice

Pileup (in.)
Cable Rack 60-in. Support 72-in. Support
Width (in.) Interval Interval
5 5 5
12–20 7 6
Table 5.37 Power Cable Weight on Fully Loaded Cable Racks (Assuming 100% Fill)
Cable Rack Support Interval
Cable Rack 5 ft (60 in.) 6 ft (72 in.)
Width ______________________________ _____________________________
(in.) Uniform (lb/ft) Total (lb) Uniform (lb/ft) Total (lb)
12 108 540 86.4 518
15 135 675 108 648
20 189 945 151.2 907
358 SYSTEM DESIGN
In one-tier installations, power and signal cables share a common cable rack; therefore,
weight calculations must take into account the weight of both cable types. There are no
average weights for power wire but signal cables weigh about 0.8 lb/in.2/ft.
Cable rack supports, such as threaded rod and inserts, drop-in anchors, and framing
channel (e.g., Unistrut framing channel), must have adequate strength for the application.
For ordinary applications, detailed support analysis is normally not required. However,
where cable rack loading is higher than normal, Table 5.38 can be used to determine the
spacing of cable rack supports.
5.8.4.2 General Rules for Power Cable Rack Layout

1. Cable rack layouts should distribute the cable loads across the equipment area, min-
imize congestion, and include capacity for not only initial cables but a reasonable
estimate of future cables.
2. To minimize the possibility of power cable slippage on vertical cable racks, limit
continuous vertical lengths to three floors and provide a 20-ft (minimum) horizon-
tal section every third floor (e.g., 4th, 7th, 10th, . . .). This requires that cable holes
be offset by at least 20 ft (Fig. 5.64) or a horizontal loop be provided that is at least
20 ft long. Cable racks are considered continuous even if cable hole sheathing or
lining interrupts the cable rack through the hole.
3. Where power cables exit the cable rack infrastructure and drop to terminations,
they should not be unsupported for more than 36 in.
4. Cable racks should be placed at the front of equipment frames to accommodate rear
cabled equipment.
5. Cable racks should clear building columns, walls, and other surfaces by at least 5
in.
6. A vertical clearance of at least 4 in. should be provided between the ultimate cable
pileup and any obstruction for installation and removal access.
7. A minimum working space of 18 in. should be provided on at least one side of the
cable rack for installation and removal access.
Table 5.38 Loads for Cable Rack Supports

Support Type Safe Load (lb)
–5-in.—11 threaded insert 1,200
8
Ceiling insert (poured in place) 1,200
–5-in. drop-in anchor 480
8
–3-in. drop-in anchor 300
8
–3-in. lag screw in 2-in. wood 300
8
Embedded ceiling channel (at any single point) 2,000
Embedded ceiling channel (where two or more loads have 2 ft 0 in. spacing) 2,000 Total
Auxiliary framing channel (paired 2 in. × 9–16 in. × 3–16 in. steel)
Support span (interval) 2 ft 0 in. 2,000
Support span (interval) 2 ft 0 in.–3 ft 0 in. 1,500
Support span (interval) 3 ft 0 in.–5 ft 0 in. 1,000
Support span (interval) 5 ft 0 in.–7 ft 0 in. 700
Support span (interval) 7 ft 0 in.–8 ft 0 in. 500
Fig. 5.64 Cable hole offset every third floor.
8. Elevation changes should be gradual, typically 45°, to avoid sharp bends in power
cables (Fig. 5.65).
5.8.5 Alternating Current Circuits for Rectifier Inputs and Inverter Outputs
The ac distribution circuit design methods are beyond the scope of this book and can be
found elsewhere (e.g., see [49–51]); however, basic ac ampacity and voltage drop calcula-
tions are described in this section.
The ac input circuits to rectifier systems and output circuits from inverter systems are
designed the same as other ac branch circuits. The ac wiring always is contained in con-
duit or enclosed in wireway (Fig. 5.66) and never directly installed in cable racks (conduit
Fig. 5.65 Cable rack elevation change (side view).

360 SYSTEM DESIGN
Fig. 5.66 Rectifier ac input circuit wiring in conduit and wireway. (Photos Courtesy of Power-One.)
sometimes is installed in cable rack). Except where the ac loads are located a long dis-
tance from the source, voltage drop normally is not a design issue, and the circuit is de-
signed based on ampacity and protection requirements. If necessary, the voltage drop can
be calculated as described here (refer also to Chapter 2, Electricity Review).
The ac voltage drop depends on several more factors than dc, including circuit reac-
tance because of the operating frequency and type of conduit. Table 5.39 provides bal-
anced three-phase, line–line voltage drop per 10,000 A-ft at 60°C (140°F) conductor tem-
perature [the table is reasonably accurate to 75°C (167°F)] for various lagging load power
factors and for both copper and aluminum wire and magnetic (steel) and nonmagnetic
(aluminum or nonmetallic) conduits.
The values can be used directly for three-phase rectifiers. However, most modular
switch-mode rectifiers are single-phase devices that operate line–line at nominal voltages
of 208 or 240 Vac. Small rectifiers (5- to 10-A output) may operate line–neutral at 120
Vac. For single-phase line–line circuits, multiply the values in the table by 1.18, and for
line–neutral circuits, multiply by 0.577. The table shows only wire sizes up to 250 kcmil,
which covers most individual rectifier circuits. The voltage drops for larger wires can be
found in the data source reference.
To use the ac voltage drop table for rectifier ac input circuits:
1. Determine the full-load current and power factor of the rectifier from its data sheet
or as described below. For modern rectifiers, the power factor usually is around 0.9
at full load.
2. Determine the circuit distance in feet.
3. Determine the type of conduit in the circuit (magnetic or nonmagnetic).
4. Determine the wire size and material (copper or aluminum) to be used in the cir-
cuit; the wire size, at least initially, is based on the required ampacity (discussed be-
low).
5. Divide the product of the line current and cable route distance by 10,000 (e.g., if the
current is 50 A on a run of 100 ft, the product is 5000 A-ft, and the result of division
Table 5.39 Balanced Three-Phase, Line–Line Voltage Drop per 10,000 A-fta,b
Wire Size (AWG or kcmil)
Load PF
c c c c
Lagging 14 12 10 8 6 4 2 1 1/0 2/0 3/0 4/0 250
Copper Conductors in Magnetic Conduit
1.00 53 33 21 13 8.4 5.3 3.4 2.6 2.1 1.7 1.4 1.1 0.92
0.95 50 32 20 13 8.2 5.3 3.5 2.8 2.3 1.9 1.5 1.3 1.1
0.90 48 30 19 12 8.0 5.2 3.4 2.8 2.3 1.9 1.6 1.3 1.2
0.80 43 27 17 11 7.3 4.8 3.2 2.6 2.3 1.9 1.6 1.4 1.2
0.70 38 24 15 9.9 6.6 4.4 3.0 2.5 2.1 1.8 1.5 1.3 1.2
Copper Conductors in Nonmagnetic Conduit
1.00 53 33 21 13 8.4 5.3 3.3 2.6 2.1 1.6 1.3 1.0 0.88
0.95 50 32 20 13 8.2 5.3 3.4 2.7 2.2 1.8 1.5 1.1 1.0
0.90 48 30 19 12 7.9 5.1 3.3 2.7 2.2 1.8 1.5 1.1 1.1
0.80 43 27 17 11 7.2 4.7 3.1 2.5 2.1 1.7 1.4 1.1 1.1
0.70 38 24 15 9.7 6.4 4.3 2.8 2.4 2.0 1.6 1.4 1.1 1.1
Aluminum Conductors in Magnetic Conduit
1.00 52 33 21 13 8.4 5.2 4.2 3.3 2.6 2.1 1.7 1.4
0.95 50 32 20 13 8.2 5.3 4.2 3.4 2.7 2.3 1.8 1.6
0.90 48 30 19 12 7.9 5.1 4.1 3.4 2.7 2.3 1.9 1.6
0.80 43 27 17 11 7.3 4.7 3.9 3.2 2.6 2.2 1.8 1.6
0.70 37 24 15 10 6.5 4.3 3.6 2.9 2.4 2.1 1.7 1.6
Aluminum Conductors in Nonmagnetic Conduit
1.00 52 33 21 13 8.4 5.2 4.2 3.3 2.6 2.1 1.7 1.4
0.95 50 32 20 13 8.2 5.2 4.2 3.4 2.7 2.2 1.8 1.5
0.90 48 30 19 12 7.9 5.0 4.1 3.3 2.6 2.2 1.8 1.5
0.80 42 27 17 11 7.2 4.6 3.8 3.1 2.5 2.1 1.7 1.5
0.70 37 24 15 9.9 6.4 4.2 3.4 2.8 2.3 1.7 1.6 1.4
a
Data source [49].
b
For single-phase line–line circuits, multiply by 1.18 and for one- or three-phase line-neutral circuits, multiply
by 0.577.
c
Solid conductors; all others are stranded.
362 SYSTEM DESIGN
by 10,000 is 0.5 (10,000 A-ft). Note: The distance used here is not total conductor
length but route distance.
6. Enter the voltage drop table in the section appropriate for the conduit type, conduc-
tor size and type, and power factor. Read across to the intersection with the wire
size and note the voltage drop. This is the three-phase line–line voltage drop for
10,000 A-ft.
7. Multiply the voltage drop from step 6 by the actual ampere-feet divided by 10,000
from step 5. This is the actual line–line voltage drop on a balanced three-phase cir-
cuit.
8. If the circuit is single-phase line–line (most often the case except for small recti-
fiers), multiply the voltage drop found in step 7 by 1.18. If the circuit is line–neu-
tral, multiply the voltage drop found in step 7 by 0.577.
The best place to obtain rectifier input current values are from the manufacturer or data
sheet. The value used in all calculations (voltage drop and ampacity) should be worst
case, that is, at lowest input ac voltage, highest output dc voltage and highest dc output
current. The output current must include any overload capability of the rectifier (many
modern switch-mode rectifiers are rated in watts and the output current depends on the
output voltage; other switch-mode and some ferroresonant rectifiers limit at 100% of
nameplate rating, but many older rectifiers limit at 110%). The input current also can be
calculated from the rectifier efficiency characteristics (typically 80 to 90%) using the fol-
lowing:
VdcMaxIdcMax
IacMax = A (for single-phase rectifier) (5.106)
VacMinL-LEff(PF)
VdcMaxIdcMax
IacMax = A (for three-phase rectifier) (5.107)
苶
VacMinL-LEff(PF)兹3
where IacMax = maximum ac input current

VdcMax = maximum dc output voltage
IdcMax = maximum dc output current
VacMinL-L = minimum ac input voltage, line–line
Eff = rectifier efficiency at full load
PF = rectifier power factor
Example 5.42 Determine the input current for a rectifier that is connected line–line to
a 120/240 Vac, single-phase service. The rectifier output is rated 50 A at 56.0 Vdc but
has 10% overload capability. Rectifier efficiency is 0.9 and power factor is 0.9 at full
load.
Solution Because of its overload capability, the maximum rectifier output current will
be 55 A. Also, the minimum ac input voltage will be 216 Vac, assuming the worst-case
total voltage drop in the nominal service voltage is 10%, or 24 Vac. From Eq. (5.106), the
input current is
56.0 V × 55 A
IacMax = = 17.6 A
216 V × 0.9 × 0.9
As noted in the above example, the nominal service voltage is not the voltage used in
the calculations. The industry standard for electric power system voltage ratings is
C84.1—Electric Power Systems and Equipment—Voltage Ratings (60 Hz) [52], which
specifies three voltage ranges:
Range A—Service Voltage. Electric supply systems shall be so designed and operated
that most service voltages will be within the limits specified for Range A. The oc-
currence of service voltages outside of these limits should be infrequent.
Range A—Utilization Voltage. User systems shall be so designed and operated that
with service voltages within Range A limits, most utilization voltages will be with-
in the limits specified for this range. Utilization equipment shall be designed and
rated to give fully satisfactory performance throughout this range.
Range B—Service and Utilization Voltages. Range B includes voltages above and be-
low Range A limits that necessarily result from practical design and operating con-
ditions on supply and user systems, or both. Although such conditions are a part of
practical operations, they shall be limited in extent, frequency, and duration.
The voltage ranges A and B are illustrated in Figure 5.67. Voltages ranges at the top of
the figure are on a 120-V base and can be linearly scaled for other voltages (e.g., to con-
vert to a 240-V base, multiply all values by 2 and to 208 V, multiply by 1.732). These
ranges are at the bottom of the figure.
Limiting the voltage drop from the ac service equipment to the load equipment to no
more than 5% is achievable in most buildings without excessively large wire. This volt-
age drop includes the drop from the ac service equipment to any ac subpanels and from
the subpanels to the rectifiers. Use 3% maximum ac voltage drop for circuits from ac sub-
panels to the rectifiers unless a different value makes more sense. Table 5.40 shows max-
Fig. 5.67 Service and utilization voltage ranges. (Data source: ANSI C84.1 [52].)
364 SYSTEM DESIGN
Table 5.40 Line–Line Voltage Drops for Common Service Voltages

Service Voltage 3% L–L Voltage 5% L–L Voltage
(Vac) Drop (Vac) Drop (Vac)
120/240, one-phase 7.2 12.0
208Y/120, three-phase 6.2 10.4
240/120, three-phase 7.2 12.0
480Y/120, three-phase 14.4 24.0
imum target voltage drop values at 3 and 5% for the common service voltages. One ac cir-
cuit is required for each rectifier; more than one rectifier never shares a circuit.
Example 5.43 The rectifier from the previous example is located 100 ft (cable route dis-
tance) from its ac distribution panel (also called load center) and is to be connected sin-
gle-phase line–line. The circuit conductors are 10 AWG copper in PVC conduit. The ser-
vice voltage is 120/240 Vac, single phase.
Solution The rectifier input current from the previous example is 17.6 A. Multiplying
the current by the circuit distance and dividing by 10,000 gives 1760 A-ft/10,000, or
0.176. Assuming the rectifier power factor is 0.9, and since the conductors are in nonmag-
netic conduit, the voltage drop is 19 V for a 10 AWG circuit per 10,000 A-ft. Multiplying
19 V by 0.176 gives 3.3 V for the three-phase voltage drop. Since this is a single-phase
rectifier, the three-phase voltage drop is multiplied by 1.18, giving 4.0 V. Note that if the
actual service voltage is at the low end of range A service voltage (228 V from Fig. 5.67),
the actual utilization voltage would be 224 V, which is well within range A utilization
voltage (also from Fig. 5.67).
Example 5.44 An old 24-V, 400-A, three-phase rectifier is to be recycled from ware-
house storage to a site to augment an existing rectifier system because of unanticipated
load growth. The rectifier is designed to provide 10% overload indefinitely, has 0.8 effi-
ciency, and the power factor is 0.85 lagging at full load. The rectifier manual recommends
an ac input conductor size of 4 AWG copper. The circuit will be connected to a
208Y/120-V load center and will use steel conduit. The circuit route distance is 150 ft.
Determine the maximum ac input current and voltage drop and whether the circuit meets
range A utilization voltage range.
Solution For purposes of calculating the rectifier input current, the maximum rectifier
output voltage is assumed to be 28.0 Vdc, and the minimum input voltage is assumed to
be at the bottom of range A service voltage, or 198 Vac. The maximum output current can
be 440 Adc, including overload capability. Therefore, from Eq. (5.107), the input current
is
28.0 V × 440 A
IacMax = = 52.8 A
苶
198 V × 0.8 × 0.85兹3
Multiplying the current by the circuit distance and dividing by 10,000 gives 7920 A-
ft/10,000, or 0.792. From Table 5.39 and 4 AWG copper in magnetic conduit, the voltage
drop is 4.8 and 5.2 V per 10,000 A-ft for power factors of 0.8 and 0.9, respectively. Using
linear interpolation, the voltage drop at a power factor of 0.85 is half-way between, or 5.0
V per 10,000 A-ft. Multiplying 5.0 V by 0.792 gives 4.0 V for the three-phase voltage
drop. If the actual service voltage is 198 Vac, the actual utilization voltage would be 194
V, which is within range A utilization voltage shown in Fig. 5.67.
Rectifier loads are continuous so the ac input current must be multiplied by 1.25 to de-
termine the required conductor ampacity. Table 5.41 shows the allowable ampacities for
common building wire up to 250 kcmil.
Example 5.45 Confirm that the conductor size used in Example 5.44 has the required
ampacity for the application. The insulation type is THW, the ambient temperature is
30°C (86°F), and there are no more than three conductors in the raceway.
Solution The required ampacity is found by multiplying the maximum rectifier input
current by 1.25, or 52.8 A × 1.25 = 66 A. The allowable ampacity from Table 5.41 for 4
AWG copper under the stated conditions is 85 A, which exceeds the required ampacity.
Therefore, the 4 AWG is adequate in the application.
Table 5.41 Allowable ac Ampacities for 600-V Building Wire (only commonly used insulation types and
wire sizes to 250 kcmil are shown)a
Temperature Rating of Conductor
Copper Aluminum
_____________________________________ ______________________________________
60°C (140°F) 75°C (167°F) 90°C (194°F) 60°C (140°F) 75°C (167°F) 90°C (194°F)
_____________________________________ ______________________________________
RHH, THW-2, RHH, THW-2,
THHN, THHW, THHN, THHW,
THW-2, THW-2,
THWN-2, THWN-2,
Wire Size RHW, THHW, USE-2, XHH, RHW, THHW, USE-2, XHH,
AWG or THW, THWN, XHHW, THW, THWN, XHHW,
kcmil) TW, UF XHHW, USE XHHW-2 TW, UF XHHW, USE XHHW-2
14 20 20 25 — — —
12 25 25 30 20 20 25
10 30 35 40 25 30 35
8 40 50 55 30 40 45
6 55 65 75 40 50 60
4 70 85 95 55 65 75
3 85 100 110 65 75 85
2 95 115 130 75 90 100
1 110 130 150 85 100 115
1/0 125 150 170 100 120 135
2/0 145 175 195 115 135 150
3/0 165 200 225 130 155 175
4/0 195 230 225 130 155 175
250 215 255 290 170 205 230
a
Conditions: No more than three current-carrying conductors in raceway or cable and ambient temperature 30°C (86°F).
Refer to Article 310.15 in [29] for guidance when these conditions are not met.
Source: Data from NEC Table 310.16 [29].
366 SYSTEM DESIGN
For the situation where there are more than three current-carrying conductors in a race-
way or cable, the allowable ampacities from Table 5.41 are reduced by the factors shown
in Table 5.42.
Example 5.46 Determine whether the conductors in Example 5-45 have adequate am-
pacity if the conduit also contains six current-carrying conductors for two other three-
phase rectifiers.
Solution In this case, there are a total of nine current-carrying conductors, three from the
added rectifier and six from the two existing rectifiers. From Table 5.42, the allowable
conductor ampacity has to be adjusted by a factor of 0.7. For 4 AWG, the allowable am-
pacity under these conditions is 85 A × 0.7 = 59.5 A. Since the required ampacity is 66 A,
these conductors are too small. In this case, a 3 AWG or more likely a 2 AWG conductor
would be used (because 2 AWG is more readily available).
Circuit conductors must be protected at their source, and the overcurrent protection
must not exceed the allowable conductor ampacity. Circuit breakers (not fuses) are used
in modern ac distribution systems to serve branch circuit loads such as rectifiers. Com-
mon circuit breaker sizes are shown in Table 5.43 (this table shows ratings to 225 A; larg-
er standard ratings are available—see Article 240.6(A) of [29]). This table also shows the
maximum allowed circuit breaker size for smaller conductors (14, 12, and 10 AWG).
Example 5.47 Determine the circuit breaker rating for the rectifier input circuit of Ex-
ample 5.43. The insulation type is THW with a maximum conductor temperature rating of
75°C.
Solution The full-load current of the rectifier was found to be 17.6 A. The required cir-
cuit ampacity is 1.25 × 17.6 = 22 A. The conductor size was given as 10 AWG. From
Table 5.43 this conductor has a basic allowable ampacity of 35 A, which exceeds the re-
quired ampacity. Table 5.43 shows the highest circuit breaker rating for 10 AWG conduc-
tors is 30 A, which is suitable on this circuit. Since the single-phase rectifier is connected
line–line with two current-carrying conductors, a two-pole circuit breaker is required.
Example 5.48 Determine the circuit breaker size for the rectifier input circuit of Exam-
ple 5.45. The insulation type is THW with a temperature rating of 75°C.
Table 5.42 Adjustment Factors for More Than Three Current-

Carrying Conductors in a Raceway or Cable
Number of Current-Carrying
Conductors Adjustment Factor
4–6 0.80
7–9 0.70
10–20 0.50
21–30 0.45
31–40 0.40
41 0.35
Source: Data from NEC Table 310.15(B)(2)(a) [29].
5.9 DIRECT CURRENT POWER SYSTEM BONDING AND GROUNDING 367
Table 5.43 Standard Circuit Breaker Ratings up to 225 A for ac

Circuits
Circuit Breaker Size (A) Smallest Conductor Size (AWG)
15 14
20 12
25 —
30 10
35 —
40 —
45 —
50 —
60 —
70 —
100 —
110 —
125 —
150 —
175 —
200 —
225 —
Source: Data from Article 240.4(D) and 240.6(A) of [29].
Solution The required circuit ampacity was found to be 66 A. The nearest circuit break-
er rating that exceeds the required ampacity is 70 A. Since the 4 AWG conductors have
allowable ampacity of 85 A at 75°C, they are adequately protected by a 70-A circuit
breaker. Also, since there are three current-carrying conductors in the circuit, a three-pole
circuit breaker is required.
5.9 DIRECT CURRENT POWER SYSTEM BONDING AND GROUNDING
The design of an overall telecommunications grounding system is beyond the scope of

this book; however, this section will describe the bonding and grounding of the dc power
system.37
5.9.1 Building Principal Ground

The building principal ground (BPG) is the mechanical and electrical interface between
the interior and exterior grounding systems and is the ground reference point for the facil-
ity. Numerous names have been given to this functional element, such as main ground bar
(MGB), master ground bar, and CO ground bus.
The BPG bar (BPGB) is a plain copper or tin or silver-plated copper busbar and should
be at least 1–4 in. thick and 4 or 6 in. high. Depending on the purpose and number of wire
terminations, the bar can be between 6 and 48 in. long, although longer bars have been
used in practice. The BPG should be as close as possible to where the grounding electrode
37
Information on telecommunications bonding and grounding may be found in [46, 47, 53], and grounding in-
formation not specific to public telecommunications systems may be found in [54–56].
368 SYSTEM DESIGN
system conductor enters the building. The dc power system grounding connection may be
made directly to the BPG if it is on the same floor of the facility or through a floor ground
bar (FGB) if on a different floor.
The BPG is partitioned (conceptually, not physically) for connection purposes into
four segments called P-A-N-I (Fig. 5.68). This layout simplifies grounding audits and vis-
ual inspections and is thought to reduce interference between systems connected to it (but
this thinking has not been proven with rigorous analysis).
5.9.2 Direct Current Power Equipment Bonding and Grounding

In modern telecommunications equipment, the dc power return is isolated (insulated) from
the equipment frame ground. Such an arrangement is called an isolated return. The return
circuit conductors from the equipment are connected to an insulated return bus termination
bar in the powerboard. The return bus termination bar is bonded to the equipment grounding
conductor at one point only, typically the BPGB or a floor ground bar, using the dc ground-
ing (DCG) conductor. In older systems and some modern equipment designed for enterprise
applications, the dc power return is bonded internally to the equipment frame ground or it
may be bonded externally to ground at more than one point. The return conductors are con-
nected to the return bus termination bar in the powerboard and then to the equipment ground-
ing conductor as for the isolated return described above. This bonding arrangement is called
an integrated return. Figure 5.69 shows a simple schematic for the two arrangements.
All power equipment frames, cabinets, battery racks, cable racks, and other metallic
enclosures and supports in the dc power system area are bonded to the common bonding
network (CBN) or the isolated bonding network (IBN) depending on where they are lo-
cated. This bond preferably is called dc equipment grounding conductor (DCEG conduc-
Fig. 5.68 Building principal ground bar (BPGB) segmentation.

Fig. 5.69 Isolated (upper) and integrated (lower) return for load equipment chassis or frame
assembly.
tor) but also has been called frame ground conductor (FGC) and equipment grounding
conductor (EGC).
The powerboard return bus is bonded to the grounding system using a dc grounding
conductor (DCG conductor). Figure 5.70 shows the configuration for a typical central of-
fice in which the powerboard serves both IBN and CBN equipment or just CBN equip-
ment. The sizes of the DCG and DCEG conductors are discussed in Section 5.9.3.
A typical bonding configuration for a multifloor central office with both CBN and IBN
equipment is shown in Figure 5.71. In this case
앫 Only one single-point connection bar (SPCB) is used for all IBN equipment.
앫 The SPCB can be separate from the dc power system or a section of the insulated dc
power return bar.
앫 All IBN equipment should be within one floor of the dc power system.
앫 The dc power return bar is insulated from the framework and bonded to the SPCB
by the dc system grounding (DCG) conductor.
앫 The dc power system framework is bonded to the floor ground bar (FGB) of the
floor on which the dc power system is located.
370 SYSTEM DESIGN
dc Return Bar
Insulated from Equipment Frame
dc Ground (DCG) Conductor
dc Equipment
Ground (DCEG)
Conductor
dc Power System
serves CBN and IBN
Equipment
Fig. 5.70 Typical powerboard bonding. Note that the DCG conductor may be bonded to the power-
board frame (shown as alternate DCG) or directly to the BPGB. The DCEG is required in either case.
dc Power
Equipment
Fig. 5.71 Bonding a mixture of IBN and CBN equipment.

A typical configuration used in multifloor central offices with only IBN equipment is
shown in Figure 5.72. In this case
앫 There is only one single-point connection bar (SPCB) and it is located on floor N
for all equipment.
앫 All IBN equipment is located within one floor of the SPCB (A, B, C).
앫 Location a is preferred for dc power system, but b and c are acceptable.
by the DCG conductor.
앫 The dc power system framework is bonded to the SPCB if it is on the same floor as
the SPCB or to the FGB (or BPGB) if on a different floor.
The SPCB sometimes is called a ground window bar (GWB), although other names have
been used throughout the years. An example is shown in Figure 5.73. The SPCB forms a
window (SPC window, or SPCW), which conceptually is < 6 ft diameter through which
grounding conductors pass. In a mixed CBN and IBN environment, the SPCB and dc
power return bar can be the same bar. If so, the bond from the return bar to the FGB also
serves as the SPCW bond.
Fig. 5.72 Bonding of IBN equipment only.

372 SYSTEM DESIGN
dc Return Bar
dc Return
Connections
Fig. 5.73 Single-point connection bar (SPCB) can be part of the dc return bar.
Many CBN and IBN bonding arrangements have been used since 1980. The current
standard is given in [46]. Generally, if industry practices in effect at the time the equip-
ment was installed were followed, they are acceptable today (retrofit generally is not re-
quired). However, many systems may have been installed properly, but the bonding
arrangements were inadvertently corrupted by subsequent installations and related activi-
ty. Some acceptable alternatives based on past practices are [47]:
앫 The dc power system may be located several floors from the IBN equipment.
앫 Only one SPCB may be used for all IBN equipment (rather than separate SPCB for
each separate IBN installation).
앫 The SPC usually is located within one floor of the IBN equipment.
by the DCG conductor.
앫 The dc power system framework is bonded to the FGB of the floor on which the dc
power system is located.
앫 If CBN equipment is served by the same dc power system serving the IBN equip-
ment, the –48-V feeders to the CBN should be routed within 3 ft of the SPCB and
the CBN dc power return conductors connected to the SPCB.
Some other considerations are:
앫 There should be only one SPCW associated with each dc power system.
앫 If the dc power system is shared between IBN and CBN equipment, the preferred
location for the SPCW is near the power system or as part of the power system bat-
tery return bus.
앫 If the dc power system is dedicated to the IBN equipment, the preferred location for
the SPCW is the IBN equipment room.
앫 To overcome termination space limitations, single-point collector bars may be lo-
cated within the SPCW or < 3 ft away for connection to the SPCB.
앫 Older equipment frames not insulated from the floor and with multiple grounding
connections are considered a mesh bonding network (MBN), which is an extension
of a CBN. MBN dc power return leads usually are connected to the surrounding
CBN at a single point only, but they are permitted to be connected to system frame-
work at multiple points (return current will follow multiple paths).
5.9.3 Grounding Conductor Size

In isolated return systems, the DCG and DCEG conductors are based on industry require-
ments [47]. They are sized to safely carry the fault current and to provide a low-imped-
ance fault current path. There is a perception that larger conductors provide an added ben-
efit of better noise control, but there is no agreement on the conductor size for this
purpose. The DCG or DCEG conductor is the larger of
1. 6 AWG (copper).
2. Conductor size with resistance that achieves a voltage drop 90% of system nom-
inal voltage (48 or 24 V) during a fault, assuming the fault current is 10× the rat-
ing of the largest overcurrent device feeding the frame or frame assembly, or
0.9Vnom
Rmax (5.108)
10Iopd
where Rmax = maximum resistance of fault current path ()

Vnom = nominal source voltage (V)
Iopd = current rating of overcurrent protective device (A)
In most installations, the powerboard is the first point of distribution and is not be-
ing fed by an overcurrent device in the same manner as load equipment. In this
case, the current in the above calculation is based on the largest overcurrent device
in the powerboard.
3. Conductor size is based on the largest dc overcurrent device feeding the equipment
frame or frame assembly (Table 5.44).
Table 5.44 dc Equipment Grounding (DCEG) and dc Grounding

(DCG) Conductora
Overcurrent Device Rating (A) Conductor Size (Copper)
15 14 AWG
20 12 AWG
30 10 AWG
40 10 AWG
60 10 AWG
100 8 AWG
200 6 AWG
300 4 AWG
400 3 AWG
500 2 AWG
600 1 AWG
800 1/0 AWG
1,000 2/0 AWG
1,200 3/0 AWG
1,600 4/0 AWG
2,000 250 kcmil
a
This table is identical to Table 250.122 in the NEC [29].
374 SYSTEM DESIGN
Where the equipment frames use an integrated return, they are considered properly
bonded to ground if the return conductors and grounding conductors are common and
their total cross-sectional area is at least as large as the cross-sectional area of the conduc-
tors that feed the frames.
Example 5.49 A 48-V powerboard bus is rated 800 A and the largest circuit breaker size
is 225 A. The DCG conductor is 110 ft long. Determine the DCG conductor size.
Solution The conductor will be the larger of

First Test 6 AWG
0.9V 0.9 × 48 V
Second Test RMax = = 0.0192
10Iopd 10 × 225 A
Based on the calculated maximum conductor resistance, the unit resistance must be
0.0192 ÷ 110 ft = 0.000175 /ft. From Table 5.13 (at 30°C in column 4), a copper con-
ductor that meets the requirement is 2 AWG copper (unit resistance of 0.000165 /ft).
Third Test From Table 5.44, with an overcurrent device rating of 225 A, the DCG con-
ductor is 4 AWG.
The three tests above give 6 AWG, 2 AWG, and 4 AWG. The 2 AWG conductor is the
largest so it is the minimum conductor size for this particular application.
5.9.4 Inverter Bonding and Grounding

The inverter output grounding configuration depends on whether or not the inverter is
connected to the commercial ac service. If it is completely isolated from the commercial
ac service (continuous operation, See Chapter 3, DC Power System Components), the in-
verter is considered a separately derived system and the output neutral terminal is bonded
directly to the nearest frame ground bar or building structural member (Fig. 5.74). In a
separately derived system all circuit conductors on the ac side, including the grounded
(neutral) conductor, are independent of the building ac supply conductors. If the inverter
is connected to the commercial ac service (active standby operation or passive standby
dc Input
Fig. 5.74 Inverter bonding and grounding.

operation), the neutral terminal in the inverter is bonded to the grounded (neutral) conduc-
tor at the service or ac distribution panel and is not bonded directly to any local ground at
the inverter. The inverter frame always is bonded to the nearest frame ground bar. The
size of the bonding conductors is as discussed in Section 5.9.3.
If the inverter is located in an IBN, it normally is bonded and grounded according to
IBN equipment manufacturer requirements. In this case, the IBN inverter typically feeds
equipment in the same frame as the inverter. Where the NEC applies to nonexclusive use
environments, Articles 250.20 and 250.30 apply to bonding and grounding.
5.9.5 The DC–DC Converter System Bonding and Grounding

If the converter system consists of two or more converters connected in parallel, size the
converter system grounding conductor based on the largest overcurrent device in the sys-
tem. The output is bonded to the converter frame and to nearest frame ground bar or ex-
tension of grounding electrode system (Fig. 5.75). The size of the conductor is as dis-
cussed in Section 5.9.3.
5.9.6 Battery Rack Bonding and Grounding

Each battery rack or enclosure is bonded to the nearest floor ground bar or directly to the
BPGB with a 6 AWG conductor [28] (Fig. 5.76).
5.9.7 Cable Racking Bonding and Grounding

All cable racks are bonded and grounded to ensure a low-resistance path for ground fault
currents. A 6 AWG or larger strap is used to bond sections of cable rack where they are
spliced or joined (Fig. 5.77), and the cable rack is then bonded to the nearest floor ground
bar or the BPGB.
5.9.8 Equipment Frame Bonding and Grounding

It is common practice to install a small (typically 1 in. × 1–4 in., 318.3 kcmil) copper busbar
at the top of each equipment frame. The busbars are spliced to form a continuous busbar
dc Output to
dc Converter
Input Loads
Fig. 5.75 The dc–dc converter bonding and grounding.

376 SYSTEM DESIGN
Fig. 5.76 Battery rack bonding.
from one end of the lineup to the other. This busbar is bonded to the nearest floor ground
bar or BPGB and serves as a termination busbar for the DCEG conductor from each chas-
sis or shelf in the equipment frame. Any one of three methods may be used to bond the
frame grounding busbar to the grounding system (floor ground bar or BPGB):
Method 1 Run an individual conductor from the bar on each frame to the grounding
system [Fig. 5.78(a)]. The conductor should be sized according to the largest overcurrent
protective device feeding the equipment in the frame as described in Section 5.9.3. This
conductor is never smaller than 6 AWG. In many installations, a small fuse or circuit
breaker panel at the top of the frame is fed by a 30- to 80-A overcurrent device; if this is
the largest overcurrent device feeding equipment in the frame, then the DCEG conductor
size would be based on it. This method can be used with or without the small busbar at the
top of each frame.
Fig. 5.77 Cable rack bonding. Note the paint has been removed where the bonding strap connec-
tor lugs are fastened to the cable rack stringers.
(a)
(b)
Fig. 5.78 Equipment frame bonding and grounding, showing the elevation view of a lineup of five
equipment frame and the various bonding methods. (a) Equipment frame bonding using an individual
conductor, (b) equipment frame bonding using a bulk conductor. (continued)
Method 2 Run a large “bulk” conductor from the grounding system to the end of the
equipment frame lineup and bond to the equipment frame grounding busbar [Fig.
5.78(b)]. Although the minimum size of the bulk conductor is determined according to
Section 5.9.3, it is common practice to use a conductor with a cross-sectional area equal
to the sum of the cross-sectional areas of individual conductors that could be used in its
place (method 1). Generally, this DCEG conductor is limited to 2/0 or 4/0 AWG unless
the calculations in Section 5.9.3 require a larger conductor.
378 SYSTEM DESIGN
(c)
Fig. 5.78 Equipment frame bonding and grounding, showing the elevation view of a lineup of five
equipment frame and the various bonding methods. (c) Equipment frame bonding using a tapped ex-
press conductor.
Method 3 Run a large “express” conductor from the grounding system to the equip-
ment frame lineup and along the top of the lineup in the cable rack and then run a short
tap from the express conductor to each frame [Fig. 5.78(c)]. The express conductor is
sized as in method 2 above, and the tap conductors are sized as in method 1 above. This
method is often used when the lineup has eight equipment frames or less or, for longer
lineups, it may be used for each group of eight frames. Like method 1, this method can be
used with or without the small busbar at the top of each frame, but the busbar still is use-
ful for terminating small bonding leads from equipment shelves.
The selected method should take into account maintenance requirements as well as fu-
ture addition or removal of equipment and equipment frames. Method 3 has the most flex-
ibility in larger facilities.
REFERENCES (PART II)
26. ASTM B8-04, Standard Specification for Concentric-Lay Stranded Copper Conductors, Hard,
Medium-Hard, or Soft, ASTM International, 2004.
27. ASTM B172-01a, Standard Specification for Rope-Lay-Stranded Copper Conductors Having
Bunch-Stranded Members, for Electrical Conductors, ASTM International, 2001.
28. ANSI/T1.311-2002, DC Power Systems—Telecommunications Environment Protection,
ATIS, 2002.
29. NFPA-70-2005, National Electrical Code®, National Fire Protection Association, 2005.
30. H. Pender and W. Del Mar, Electrical Engineers’ Handbook, 4th ed., Wiley, New York, 1949.
REFERENCES (PART II) 379
31. J. H. Neher and M. H. McGrath, The Calculation of the Temperature Rise and Load Capability
of Cable Systems, Paper 57-660, American Institute of Electrical Engineers Transactions, Oc-
tober 1957.
32. ANSI/IEEE Std 835™-1994, IEEE Standard Power Cable Ampacity Tables, IEEE, New York,
1994.
33. IEC 60287, Electric Cables—Calculation of the Current Rating—Part 2-1: Thermal Resis-
tance—Calculation of Thermal Resistance, Ed. 1.1, International Electrotechnical Commission
(IEC), Nov. 2001.
34. G. J. Anders, Rating of Electric Power Cables, Ampacity Computations for Transmission, Dis-
tribution, and Industrial Applications, IEEE Press Power Engineering Series, New York, 1997.
35. G. J. Anders, Rating of Electric Power Cables in Unfavorable Thermal Environment, IEEE
Press/Wiley Interscience, Hoboken, NJ, 2005.
36. R. Bartnikas, Ed. and K. D. Srivastava, Co-Ed., Power and Communication Cables—Theory
and Applications, IEEE Press—Wiley Interscience, Hoboken, NJ, 2000.
37. D. Fink and J. Carroll, Standard Handbook for Electrical Engineers, 10th ed., McGraw-Hill,
New York, 1968.
38. J. Stolpe, Ampacities for Cables in Randomly Filled Cable Trays, IEEE Transactions on Power
Apparatus and Systems, Vol. PAS-90, 1970.
39. B. Harshe and W. Black, Ampacity of Cables in Single Open-Top Cable Trays, IEEE Transac-
tions on Power Delivery, Vol. 9, No. 4, Oct. 1994.
40. ANSI/ICEA P-54-440 (NEMA WC 51-2003), Ampacities of Cables installed in Cable Trays,
Insulated Cable Engineers Association, Jan. 2004.
41. H. Leake, Sizing of Cables in Randomly-Filled Trays with Consideration for Load Diversity,
IEEE Transactions on Power Delivery, Vol. 12, No. 1, Jan. 1997.
42. J. P. Holman, Heat Transfer, 9th ed., McGraw-Hill, New York, 2002.
43. J. H. Lienhard, A Heat Transfer Textbook, Phlogiston Press, Lexington, MA, 2005.
44. H. Brakelmann, P. Lauter, and G. Anders, Current Rating of Multicore Cables, Industry Appli-
cations Conference Record, Vol. 4, 39th IAS Annual Meeting, IEEE, 2004.
45. Publication No. 22, Copper for Busbars, Copper Development Association—Herts, UK, June
1996 (http://www.cda.org.uk/megab2/elecapps/pub22/index.htm).
46. ANSI/T1.313-2003, Electrical Protection for Telecommunications Central Offices and Similar
Type Facilities, ATIS, 2003.
47. ANSI/T1.333-2001, Bonding and Grounding of Telecommunications Equipment, ATIS, 2001.
48. ANSI/IEEE Std 1375™-1998, IEEE Guide for protection of Stationary Battery Systems, Insti-
tute of Electrical and Electronics Engineers, Piscataway, NJ, 1998.
49. ANSI/IEEE Std 241™-1990, IEEE Recommended Practice for Electric Power Systems in
Commercial Buildings (IEEE Gray Book), Institute of Electrical and Electronics Engineers,
Piscataway, NJ, 1990.
50. D. Beeman, ed., Industrial Power Systems Handbook, McGraw-Hill, New York, 1955.
51. J. Stallcup, Stallcup’s Electrical Design Book, Grayboy Publishing, North Richland Hills, TX,
1999.
52. ANSI C84.1-1995 (R2001), Electric Power Systems and Equipment—Voltage Ratings (60
Hz), National Electrical Manufacturers Association, 2001.
53. Bulletin 1751F-810, Electrical Protection of Digital & Lightwave Telecommunications Equip-
ment, Rural Utilities Service (RUS)—US Dept. of Agriculture, 1997.
54. ANSI/IEEE Std 142™-1991, IEEE Recommended Practice for Grounding of Industrial and
Commercial Power Systems (IEEE Green Book), Institute of Electrical and Electronics Engi-
neers, Piscataway, NJ, 1991.
380 SYSTEM DESIGN
55. MIL-HDBK-419, Grounding, Bonding, & Shielding for Electronic Equipments and Facilities,
Vols. 1 & 2, U.S. Dept. of Defense, Washington, DC, 1987.
56. J. P. Simmons, Soares Book on Grounding, International Association of Electrical Inspectors,
Richardson, TX, 1999.
CHAPTER 6
SYSTEM INSTALLATION
AND MAINTENANCE
PART I DIRECT CURRENT POWER SYSTEM INSTALLATION
This part covers telecommunications dc power system installation requirements that may
lead to field questions or field problems and does not cover every aspect of dc power sys-
tem installation. Except where industry standards are referenced, the discussions in this
part represent various industry practices and the author’s own experience and, as such,
are presented as guidelines. The methods can be changed to suit company practices or
specific installations, but the methods used in the field must always reflect good judgment
and common sense. Some companies may have practices that differ significantly from
those described here.
6.1 INSTALLATION SAFETY GUIDELINES
앫 Use factory insulated or taped tools when installing power equipment. Use a high-
quality electrical tape; do not use cheap electrical tape like that available at automo-
tive parts stores or department stores.
앫 Use only battery-powered or double-insulated power tools (double-insulated power
tools should be regularly tested).
앫 Remove all jewelry, watches, metal-rimmed eye glasses, and belt buckles when per-
forming any power system installation work.
앫 Wear safety goggles or a face shield and gloves when working with batteries.
앫 Wear safety glasses or face shields when working with hand tools.
앫 Do not use or place metal-framed ladders, metal desks, or metal chairs in the imme-
diate vicinity of energized power equipment.
앫 Use nonconductive rulers (e.g., carpenter’s folding wood ruler) in the vicinity of en-
ergized equipment; use a metal measuring tape only for floor measurements.
앫 Be alert and cautious when working in the vicinity of rotating equipment and tools
especially when wearing loose clothing, neck ties, and long hair.

382 SYSTEM INSTALLATION AND MAINTENANCE
앫 Protect working equipment from dust and debris with antistatic, fire-retardant tarps
when drilling and vacuuming in the vicinity of equipment. Use tarps that meet the
requirements of NFPA-701 [1] and are designed for telecommunications antistatic
applications
앫 Use fire-retardant pegboard to protect fan-cooled equipment; remove the pegboard
immediately after completion of the daily work.
앫 Place insulated blankets that meet the requirements of ASTM D1048-05 [2] over
energized busbars and components and when working near energized busbars, ter-
minations, and other power equipment.
앫 Remove drill shavings and other debris with a shop-type vacuum cleaner. The vacu-
um components that may contact electronic equipment should be made from insu-
lating rubber or plastic materials.
앫 Place fiberboard or PVC floor mats on the floor to prevent scuffing and scratching
power wire insulation during installation.
앫 Use electrostatic discharge (ESD) protective devices (e.g., grounded wrist straps)
when handling and storing circuit packs and other sensitive electronic equipment.
앫 Use a vacuum cleaner or a drill equipped with a vacuum attachment when drilling
floors, ceiling, and walls.
6.2 MATERIAL COMPATIBILITY
All metals used in the electrical system must be compatible to prevent galvanic corrosion
(Table 6.1).1 Generally, aluminum wiring is not used in telecommunications dc power
systems because telecommunications workers seldom are trained in its correct application
and installation. Occasionally, aluminum busbar will be found in existing installations.
Connections to them must use connector lugs designed for the purpose (e.g., connector
lugs suitable for both copper and aluminum conductors).
6.3 CIRCUIT WIRING
The dc power circuits always use paired conductor configurations. The following infor-
mation is required to properly install circuit wiring:
앫 Circuit route or path

앫 Wire length
앫 Conductor size and bend radius
앫 Terminal connector lug type, width, fastener hole diameter and hole spacing
앫 Compression (crimping) tool and die that matches the terminal connector lug
앫 Fastener torque values
앫 Securing methods
1
Galvanic corrosion (also called galvanic action) is caused by the electrical contact of dissimilar metals in the
presence of moisture (electrolyte) causing the metal higher in the galvanic series (the more anodic or least noble)
to corrode sacrificially.
6.3 CIRCUIT WIRING 383
Table 6.1 Material Compatibility

Compatible Bracket, Terminal,
Conductor Material and Hardware Material
Copper Copper
Bronze and brassa
Silicon bronze
Stainless steel
Steel
Aluminum Aluminum
Steel
Iron
a
Because brass hardware has comparatively low strength, it should be used
only where provided in factory assemblies.
Source: From [3].
6.3.1 Circuit Routing and Wire Length

To minimize the amount of material and installation effort, circuit wiring should follow
the most direct and shortest route; however, this may lead to excessive pileup and conges-
tion on some cable racks especially at cable rack junctions. Therefore, routes should be
planned to distribute the power wiring as evenly as possible on the cable racks even if it
means a longer run. If the route to be followed is longer than originally used in the circuit
design calculations, it will be necessary to recalculate the voltage drop and possibly in-
crease the wire size to keep the drop below the design value. The conductors of any given
circuit, including all paralleled feed and return conductors, must follow the same route.
Minimizing the length of low-voltage dc circuits is critical to maintaining voltage drop
within the design requirements, especially in 24-Vdc systems.
6.3.2 Conductor Size and Bend Radius

Conductor sizes are determined on the basis of current rating and voltage drop criteria as
described in Chapter 5, System Design. Where insulated wires are installed in cable racks
and formed around corners and cable rack exit points, bending them too sharply may
damage the conductors and insulation. The minimum bend radius for insulated wire is
shown in Table 6.2 for ASTM B8 Class B (concentric-lay coarse-strand) and ASTM
B172 Class I (bunch- or rope-lay fine-strand) wires.
Measurement procedure for cable bend radius after wire is placed on the cable rack:
앫 Place a framing square over the inside edge of the 90° bend (Fig. 6.1).
앫 Measure lengths A and B from the outside corner of the square to where the wire in-
tersects the straight edge; A and B should be approximately the same.
앫 A and B should be minimum bend radius for the wire size.
A properly bent wire will show no visible deformation of the insulation. Improperly
bent wire usually is obvious even if there is no deformation (Fig. 6.2). A conduit hickey
never should be used to bend insulated wire because it may damage the insulation.
Table 6.2 Training Bend Radius for Stranded Wirea

Wire Size Recommended Minimum Bend
(AWG or kcmil) Bend Radius (in.) Radius (in.)
(1) (2) (3)
14 11–2 1
–
8
12 11–2 1
–
2
1
10 2 –
2
8 21–2 1
6 2–34 1
4 3 1
2 3–34 11–2
1 4 11–2
1/0 41–2 11–2
2/0 51–4 31–2
3/0 51–2 31–2
4/0 51–2 31–2
250 61–4 5
300 6–34 5
350 71–2 5
400 8 5
500 9 5
600 9 7
750 10 7
1,000 11 10
a
Measured on the inside of the bend. Column (2) based on 8X wire outside diameter; Col-
umn (3) for Class B only [3]. This table applies to wires installed on cable racks and does
not apply to wires installed in conduit or pulled over sheaves or other curved surfaces
during installation, in which case larger bend radii may be required to limit sidewall pres-
sure and to prevent insulation damage.
Fig. 6.1 Measuring wire bend radius with a carpenter’s framing square. Place the square on the in-
side edge of the bend as shown and measure lengths A and B from the outside corner of the square
to where the cable intersects the straight edge.
6.4 TERMINAL CONNECTOR LUGS 385
Fig. 6.2 Unacceptable wire bend radius at a battery terminal. (Photo courtesy of M.W. Migliaro.)
6.3.3 Cable Rack Fill

Generally, when wire is operated at or less than 25% load factor (actual load current di-
vided by current rating), self-heating is negligible, and the wires can be installed with no
separation and in multiple layers. When wire is to be operated at its free-air ampacity,
only a single wire layer may be installed on a cable rack and the axial spacing between
wire centers must be greater than two times their outer diameter (spacing between outer
insulations equal to the wire outer diameter). For other installation conditions, such as re-
duced separation, wires touching, multiple layers, and bundled wires, the wires must be
operated at derated ampacity values. See Chapter 5, System Design.
6.4 TERMINAL CONNECTOR LUGS
Five or six parameters must be known to properly specify connector lugs (Fig. 6.3):
앫 Conductor and termination material (usually copper–copper)

앫 Conductor size and strand configuration
앫 Connector type (compression or mechanical)
앫 Lug width
앫 Hole diameter
앫 Hole spacing (two-hole lug only)
Generally, connector lugs should be listed by a nationally recognized testing laborato-

ry (NRTL)2 and have the proper dimensions. Both irreversible compression and mechani-
2
NRTL is an organization recognized by the U.S. Occupational Safety and Health Administration (OSHA) to
perform safety tests and list, label, or accept equipment or materials. Familiar examples of NRTLs are Under-
writers Laboratory (UL) and Intertek Testing Services NA, Inc. (ITSNA, formerly ETL Testing Laboratories,
Inc.). A complete listing of NRTLs may be found at http://www.osha.gov/dts/otpca/nrtl/index.html#nrtls.
Fig. 6.3 Two-hole and one-hole connector lug dimensions.
cal connector terminal lugs may be used, but irreversible compression are preferred. The
following sections describe both types in greater detail.
The conductor and connector lug materials in most installations will be copper–cop-
per, and standard connector lugs rated for these materials are readily available. The fas-
tener materials used with connector lugs and conductors also must be compatible as dis-
cussed in Section 6.2.
For the same conductor size (AWG or kcmil), fine-strand conductor diameter is larger
than coarse-strand and a larger connector lug may be required. For example, consider 4/0
AWG stranded conductors of different classes as specified in ASTM standards (Table
6.3).
The diameters of Class A, B, C, and D wires are very close and the same connector lug
may be used on any one of them. However, Class G and above are significantly larger,
and it is unlikely that a connector designed for, say, Class B (coarse) stranding will work
for Class I (fine) stranding. Generally, but not always, the connector lug for Class I wire is
two sizes larger than for the same AWG or kcmil Class B wire.
6.4.1 Compression-Type Connector Lugs

Compression-type connector lugs use either a circumferential crimp or indent crimp to
hold the conductor in the connector barrel. Generally, circumferential crimps are recom-
mended for low-voltage applications. Many tools used to install connector lugs are ratch-
et-types that, once the compression begins, do not release until the lug is properly
Table 6.3 Diameter of Different 4/0 AWG Stranded Conductor Classes

ASTM ASTM B-8 ASTM B-173 ASTM B-172
Classa A B C D G H I K M
No. Strands 7 19 37 61 133 259 532 2107 5320
Diameter (in.) 0.522 0.528 0.529 0.530 0.599 0.601 0.613 0.627 0.645
a
Fine-strand conductors, Class G through M, generally are not recommended for use as grounding and bonding
conductors because the strands do not adequately carry surge currents.
crimped. When a circumferential crimp is made, the tool die index is embossed on the
barrel to allow the connector and die combination to be easily verified. Crimped barrels
may be covered with heat-shrink tubing; however, clear tubing should be used so the
crimp may be examined later during routine site inspections.
The shape of the crimp varies with the connector manufacturer and may be square
(also called diamond or box), hexagonal, or other shape. It is the practice of some compa-
nies to use only hexagonal crimps. Although compression lugs cost more than mechanical
lugs because of the tooling, the bond between the connector and conductor is superior to
mechanical lugs. The compression provides a gas-tight molecular bond between the lug
and conductor with little chance for oxidation of the conductor material (Fig. 6.4).
For indent connectors, the tool compresses the connector barrel and wire strands by
deeply indenting the barrel on one side. Indent connectors can be used for any application
except small insulated ring- or fork-lug terminals and splices. Indent compression works
well for terminating fine-strand wire; however, the tool must be properly matched to the
connector lug.3
6.4.2 Mechanical-Type Connectors

Mechanical-type connector lugs use a wrench-tight bolt (or screw) and clamp to hold the
conductor in the connector barrel (Fig. 6.5). Mechanical connectors have a tendency to
loosen over time due to thermal expansion and contraction even if properly tightened dur-
ing installation. Generally, mechanical-type connectors should be avoided unless the ter-
mination is specifically designed for use only with a mechanical connector. In some
cases, mechanical connectors will be used when the correct compression connector can-
not be located or obtained in time for the job.
Set screw-type connectors (Fig. 6.6) should not be used with fine-strand wire because
the set screw will damage the individual strands and will not adequately hold the conduc-
tor. The only kind of mechanical connector that should be used with fine-strand wire is
one that completely encases all the strands (such as a V-shape clamp). The conductor size
should be toward the lower end of the connector’s range. For example, say the ranges of
two connectors are 300 to 500 kcmil and 500 to 800 kcmil. The recommended connector
for Class I 500-kcmil wire would be the larger 500- to 800-kcmil connector.
6.4.3 Connector Lug Installation Guidelines

앫 Use the proper connector lug for the conductor size being terminated as specified in
the manufacturer’s catalogs.
앫 Use irreversible compression-type connector lugs with fine-strand wire wherever
possible.
앫 Do not use abrasive paper or pads on plated lugs; unplated copper lugs may be
cleaned with abrasive paper or pads (see Section 6.6.11 for methods).
앫 Use flat washers under bolt (screw) heads and under all nuts unless specified other-
wise by the equipment manufacturer.
앫 Use two-hole terminal connector lugs for all connections unless the lug landing
does not accommodate a two-hole lug.
3
Indent compression (nest indentors) should not be used on aluminum wire.
Fig. 6.4 Compression-type connector lugs. (Top photo) The upper lug was crimped to a fine-strand
wire and then cut for examination. Note that the crimp was hexagonal and that the individual strands
are almost invisible. The lower lug shows embossing from the crimping operation. (Bottom photo)
The lug in the middle was subjected to tension during a pull test that exceeded its rated pull-out
force. Note that the middle lug has a standard barrel and the lower lug has a long barrel. The upper
lug is the same one shown in the top photo.
Fig. 6.5 Mechanical connector lugs. Four small clamp-type mechanical connector lugs are shown
with two short-barrel compression lugs. For size comparison, the two compression lugs are for 4
AWG and 3–8-in. stud (left) and 6 AWG and 1–4-in. stud (right).
Fig. 6.6 Set-screw mechanical connector lugs used with coarse-strand wire. Set-screw types dam-
age both fine and coarse strands and are not recommended in low-voltage applications. Other defi-
ciencies shown in this photo are the one-hole lug and no apparent oxidation inhibitor. (Photo cour-
tesy of M.W. Migliaro.)
앫 Where a one-hole lug must be used, install it with flat and lock washers under the
nut (with lock washer closest to the nut) or use a self-locking nut instead of a lock
washer. If attaching to a captive nut, use a flat washer under the bolt (screw) head.
앫 Choose a bolt or screw length such that when installed the threaded end does not
protrude beyond the nut more than one diameter.
앫 Use long-barrel lugs wherever possible (the extra conductor-barrel contact area is
important in low-voltage applications).
앫 Never attempt to crimp more than one stranded wire in a connector lug unless the
lug is specifically designed for more than one wire.
앫 Never attempt to crimp solid wire in a connector lug unless the lug is specifically
designed for solid wire.
앫 Do not remove strands from a larger wire to fit a smaller connector lug.
앫 Do not fold a smaller wire to fit a larger connector lug.
앫 Strip only enough insulation such that the gap between the wire insulation and the
1
16 in. (shiner length), and never allow the insulation
connector barrel end is less than ––
to extend into the connector barrel. Some lugs have a built-in stripping gauge.
앫 When stripping 8 AWG and larger wire, use a sharp electrician’s knife and a whit-
tling motion so that the strands are not damaged; alternately, use a properly adjust-
ed stripping tool specifically made for the conductor size and with cutting blades
adjusted properly for the insulation thickness.
앫 Always use a stripping tool on 10 AWG and smaller wire.
앫 Scuff untinned conductors with a wire brush (do not scuff or sand tinned conduc-
tors) and then coat the conductor with oxidation inhibitor before crimping (Table
6.4) (some connector lugs are factory filled with oxidation inhibitor in which case
the conductor does not need to be coated).
앫 Insert the conductor to the full depth of the connector lug barrel before compressing
(some connector lugs have an inspection hole).
앫 Do not crimp in the tab area.

앫 Larger lugs require more than one crimp (the number of crimps is marked on the
barrel by bands); do not overlap crimps.
앫 Apply the first crimp nearest the connector tab and crimp successively toward the
barrel end—do not skip. On splice connectors, apply the first crimp adjacent to the
center and crimp successively in both directions toward the barrel ends.
앫 Use a flashing cutter to remove the flashing left on the barrel where the die jaws
come together.
앫 Generally, place only one lug on a lug landing location; however, where necessary
to place more than one connector lug on a lug landing because of space constraints,
various stacking means are available (Fig. 6.7).
앫 To provide a neat appearance, do not cut wire to final length and do not install the
connector lugs until the wires have been formed and secured.
6.4.4 Insulated Terminal Connector Lugs for Small Wire

Color-coded insulated terminal lugs may be used only with the proper conductor sizes
(Table 6.5). These types of lugs are not designed for solid conductors (the barrels are soft
copper and cannot adequately grip a solid conductor). However, there may be situations
where it is necessary to use solid wire with ordinary soft copper insulated terminal lugs. If
so, follow these steps:
1. Pull the barrel insulator off the lug and discard.

2. Slip a piece of heat-shrink tubing over the wire (the tubing must be large enough to
slip over the barrel).
3. Strip and insert the conductor. The shiner length should be –– 1
32 in. and the conduc-
1 1
tor should protrude 32 to 16 in. beyond the barrel toward the ring or fork tab.
–– ––
4. Crimp with a tool made for uninsulated terminal lugs (either oval crimp or indent
crimp).
5. Solder the crimped conductor to the lug being sure to fill the barrel with solder.
6. Slip the heat-shrink tubing over the barrel and shrink with a heat gun.
Table 6.4 Oxidation Inhibitors

Brand Name Manufacturer Address Remarks
Penetrox-E Burndy www.fciconnect.com Cu–Cu, contains copper dust
Noalox Ideal www.idealindustries.com Cu–Cu, contains zinc dust
NO-OX ID Sanchem www.sanchem.com Soft, wax based with small amount
“A-Special” of solvent. This type of oxida-
tion inhibitor typically is used on
terminal pads or busbars (flat
mating surfaces) and is not used
to coat conductors before instal-
lation in a connector lug barrel
Cual-Gel Penn-Union www.penn-union.com Cu–Cu, contains copper dust
6.5 FASTENERS 391
Fig. 6.7 Stacking configurations for terminal connector lugs.
Table 6.5 Insulated Terminal Lugs

Terminal Lug Stranded Conductor
Insulator Color Size Range (AWG)
Red 22–18
Blue 16–14
Yellow 12–10
6.4.5 Insulated Terminal Lug Installation Guidelines

앫 Strip wire insulation such that the shiner length ––
1
32 in. unless the barrel insulator is
bulged to accommodate the insulation; in this case, strip the insulation so that it
may be inserted in the insulator but not in the lug barrel.
앫 Strip insulation and insert wire so that it protrudes between –– 1 1
32 and 16 in. beyond the
––
end of the barrel toward the ring or fork tab.

앫 Apply only one barrel crimp and one insulator crimp using a crimping tool that has
a full-cycle ratchet mechanism. Some tools simultaneously crimp both barrel and
insulator.
6.5 FASTENERS
The main application parameters for fasteners (bolts, screws, nuts, and washers) are
앫 Diameter
앫 Length
앫 Thread pitch
앫 Material strength
앫 Material compatibility
Bolts are headed, externally threaded fasteners that are meant to be tightened by turn-
ing a nut. Screws are headed, externally threaded fasteners that are meant to be tightened
into a preformed internal thread or by forming its own thread while being turned. The two
terms are often used interchangeably, but in common usage a screw generally is 1–4 in. or
less in diameter and has a slotted or cross-recessed (e.g., Phillips) head.
Nonmetric fastener dimensions (diameter and length) in the United States are in inches
and thread pitch is in inches/thread (e.g., a –38-16 designation indicates –38-in. diameter bolt
1
16 in./thread). The corresponding tightening torque values are speci-
with a thread pitch of ––
fied in inch-pounds (in.-lb) or foot-pounds (ft-lb.) Metric dimensions frequently are used
in new equipment and some equipment may have a mixture of metric and nonmetric.
Metric fastener size designations begin with M or MJ followed by nominal diameter and
thread pitch, both in millimeters and separated by “x” (e.g., M6x1 indicates 6 mm nomi-
nal diameter and 1 mm thread pitch). For metric fasteners the default thread pitch is
coarse (also called regular) and may be omitted from the designation (e.g., M6 indicates 6
mm nominal diameter and standard coarse thread pitch). Metric torque values usually are
specified in newton-meters (N-m).
Fasteners must be properly tightened. When fasteners are tightened by applying torque
to the head or nut, the thread helix converts the torque to a tensile force (tension or pre-
load) and the bolt stretches. The preload clamps the material between the head and nut
and holds the joint together. In electrical applications the parts in the joint may be subject-
ed to mechanical forces from vibration, shock (shipping, handling, installation, and seis-
mic events) and thermal cycling, and proper clamping is essential to keep them from loos-
ening and to ensure a low-resistance connection.
Torque values can be determined from the fastener’s characteristics and required
clamping force
Tq = KFCD in.-lb (6.1)
where Tq = torque (in.-lb)

K = coefficient of friction (0.2 for ordinary fasteners used in electrical work)
FC = clamping force (lb)
D = nominal diameter (in.)
The clamping force generally is 75% of the proof load force on the fastener, or
FC = 0.75SpAS (6.2)
where Sp = proof load (lb/in.2)—see below

AS = bolt or screw stress cross-sectional area (in.2)
The proof load is the tensile load a fastener must support without permanently deform-
ing. Proof loads generally are a high fraction of the minimum yield strength of the fasten-
er material. For example, the proof load of SAE Grade 2 fasteners, which are made from
low or medium carbon steel, is approximately 96% of the yield strength (Table 6.6).
6.5 FASTENERS 393
Table 6.6 SAE Grade 2 Mechanical Properties for 1–4–3–4 in. Diameter Fastenersa
Minimum Strengths (lb/in.2)
Proof Load Tensile Strength Yield Strength
55,000 74,000 57,000
a
Grade 2 bolts do not have any grade identification marks on the head.
Combining Eqs. (6.1) and (6.2) gives
T = K × 0.75Sp ASD in.-lb (6.3)
Torque values for nonmetric fasteners are given in Table 6.7 and for metric fasteners in
Table 6.8. These values may be used when more specific information is not available
from the fastener manufacturer; however, manufacturers always should be consulted
when in doubt.
The following conversion factors may be used to convert torque value units:
앫 To convert from inch-pound to newton-meter, multiply in.-lb by 0.11298

앫 To convert from newton-meter to inch-pound, multiply N-m by 8.8507
앫 To convert from newton-meter to foot-pound, multiply N-m by 0.7376
앫 To convert from inch-pound to foot-pound, multiply in.-lb by 0.08333
앫 To convert from foot-pound to inch-pound, multiply ft-lb by 12
Table 6.7 Suggested Torque Values for Nonmetric Fasteners, Applicable to Unlubricated,
Unplated Steel, Stainless Steel and Silicon-Bronze Fastenersa
Nominal Tensile Hex-Head Slotted-Head Hex-Head
Diameter Stress Area Cap Bolt Cap Screw Cap Bolt
(in.) (in.2) (in.-lb) (in.-lb) (in.-lb)
Fastener Size [7] [7] [4] Calculated Calculated
(1) (2) (3) (4) (5) (6)
6-32 UNC 0.1380 0.00909 — 10 —
8-32 UNC 0.1640 0.0140 — 19 —
10-24 UNC 0.1900 0.0175 — 27 —
12-24 UNC 0.2160 0.0242 — 43 —
1/4-20 UNC 0.2500 0.0318 80 66 66
5/16-18 UNC 0.3125 0.0524 180 — 135
3/8-16 UNC 0.3750 0.0775 240 — 240
7/16-14 UNC 0.4375 0.1063 — — 384
1/2-13 UNC 0.5000 0.1419 480 — 585
9/16-12 UNC 0.5625 0.182 — — 845
5/8-11 UNC 0.6250 0.226 660 — 1,165
3/4-10 UNC 0.7500 0.334 1,050 — 2,067
a
The torque values given in column (4) are cited in NEMA CC 1-2002 [4] and should be used as defaults for the
bolt sizes shown; columns (5) and (6) are calculated from Eq. (6.3). Differences between column (5) and (6) are
due to different proof load force assumption.
Table 6.8 Suggested Torque Values for Metric Fasteners Used on Electrical
Connections (Class 5—Equivalent to SAE Grade 2)
Tensile
Nominal Diameter Strength Area Torque
Cap Screw Size (mm) (mm2) (N-m)
(1) (2) (3) (4)
M2 2 2.1 0.2
M3 3 5.0 0.9
M4 4 8.8 2.0
M5 5 14.2 4.0
M6 6 20.1 6.9
M8 8 36.6 16.7
M10 10 58.0 33.0
M12 12 84.3 57.5
M16 16 157.0 142.9
M20 20 244.8 278.5
The accuracies of various tightening methods are shown in Table 6.9. Although a
torque wrench is not the most accurate method of tightening fasteners, it is adequate, in-
expensive, and simple to use in the field. A high-quality torque wrench (e.g., click-type,
micrometer-adjustable torque wrench shown in Fig. 6.8) should be used in all electrical
installation work; cheap automotive style (needle- or beam-type) torque wrenches should
not be used because they cannot be insulated and they are inaccurate. In some specific ap-
plications, for example, when tightening connections to large semiconductors, the turn-
of-nut method may be specified. In this case, the nut is finger tightened (or wrench tight-
ened to contact) and then a wrench is used to tighten a specified number of turns beyond.
6.5.1 Fastener Installation Guidelines

앫 Use oxidation inhibitor on the joints of all electrical connections.
앫 Use pan-head or hex-head cap screws.4
앫 Use SAE5 Grade 2 steel or equivalent for slotted machine and hex cap screws.
앫 Use fasteners and hardware made from silicon bronze, 316 stainless steel, or steel
that has a corrosion protective finish.
앫 Use lock washers, disk spring washers (Belleville washers), or self-locking nuts on
connections subject to vibration or thermal expansion and contraction.
앫 To minimize torque and the chance of stripping the threads on tapped copper busbars
< 1–4-in. thick, use only slotted No. 8-32, No. 10-32, 12-24, or 1–4-in. machine screws.
Where larger size screws or bolts are required, drill out the busbar and use nuts.
6.5.2 Belleville Washers

Belleville washers originally were developed for use with relatively soft aluminum busbar
to overcome connection problems due to creep of the busbar material (cold-flow under
4
Cap screw is a general term used to describe hex-head, slotted-head, square-head, and socket-head screws.
5
Society of Automotive Engineers, or SAE International (www.sae.org).
6.5 FASTENERS 395
Table 6.9 Accuracy of Bolt Tightening Methods [1]

Method Accuracy (%) Method Accuracy (%)
By feel ± 35 Computer-controlled wrench
Torque wrench ± 25 Below yield point ± 15
Turn-of-nut ± 15 Yield-point sensing ±8
Preload indicating washer ± 10 Bolt elongation ± 3–5
Strain gauge ±1 Ultrasonic sensing ±1
pressure). However, they also are used in many other electrical applications, including
copper–copper connections that carry high currents and are subject to thermal cycling
(expansion and contraction).
A Belleville washer also is known as a cupped spring washer, constant-force disk
spring washer, compression washer, and conical washer. It has a slight conical or cup
shape, which gives the washer a spring characteristic. When the fastener and washer com-
bination is tightened, the washer is compressed or flattened a certain amount and exerts an
opposing force on the threads. This force compensates for any creep or expansion and
contraction in the joint materials.
Three parameters define the characteristics of the disk spring washer—torque, diame-
ter, and finish. The torque of the washer must match the torque of the bolt. If the washer is
compressed too much (a few percent) by overtorque, the locking action and the washer
will be ruined. Undertorque does not compress the washer enough and the connection will
be loose. A disk spring washer generates a clamping force along its outside edge or cup
perimeter. If the washer diameter is too large and overhangs the connection, it will not
clamp properly. Large flat washers never should be used under the disk spring washer to
compensate for overhang since the flat washer will deform rather than transmit the force
to the connection. However, in copper busbar applications, a hardened thrust washer fre-
quently is installed underneath the disk spring washer to eliminate face damage and in-
dentation (Fig. 6.9). Disk spring washers used in battery rooms should not have an elec-
troplated finish because of the possibility of hydrogen embrittlement.
Fig. 6.8 Click-type, micrometer adjustable torque wrench. This wrench is about 12 in. long and has
a 3–8-in. drive. The torque is set by rotating the handle to the desired micrometer setting.
Fig. 6.9 Belleville washer applications. The “cup” of the cone points toward the flat surface and the
“peak” points toward the fastener head (upper two drawings) or nut (lower photograph).
Disk spring washers are installed with their cup pointed toward the contact surface.
Most disk spring washers have a shallow cup (approximately 10 to 20 thousandths of an
inch), which may be difficult to discern. In this case, the washer may be laid on a flat sur-
face and viewed from the side (Fig. 6.10).
A little practice is required to correctly tighten a fastener assembly with a disk spring
washer, and extra washers should be ordered for any job to allow for ruining a few. As the
washer is compressed, an abrupt change will be felt in the wrench when the washer flat-
tens. Overtightening (overflattening) usually ruins the washer. Disk spring washers nor-
mally are compressed no more than 75% of their total possible deflection. When more
than one set of fasteners is used in a joint or connection, each fastener should be partially
tightening in turn and in multiple stages. Some manufacturers recommend that the washer
be flattened and then backed off slightly.
6.6 POWER WIRING INSTALLATION 397
Fig. 6.10 Belleville washer orientation.
6.6 POWER WIRING INSTALLATION
6.6.1 Routing Guidelines

앫 Run unprotected battery circuit wires (i.e., circuits with no overcurrent protection)
on a separate cable rack from protected circuit wires.
앫 Run power system control cables (e.g., voltage sensing and rectifier control cables)
with physical separation from dc power circuit wires.
앫 Run all circuit wires (feed and return) as pairs, side-by-side on the cable rack (Fig.
6.11).
앫 Lay wire evenly in the cable rack and wiring troughs (Fig. 6.12).
앫 Group small wires in bundles approximately the same size as larger wires in the
same layer, but, if laced, do not exceed the maximum number of wires per stitch
(see Section 6.6.3).
앫 Avoid twisting the wires.
앫 Avoid crossovers wherever possible.
앫 Plan and lay wire runs so they can easily exit the cable rack where required.
앫 During installation, run multiple wires simultaneously wherever possible.
Fig. 6.11 Power wire polarity arrangements on cable rack. Polarity shown for parallel wires in one
circuit or individual circuits in one layer (upper and middle) and two layers (lower). For more than two
layers, extend the sequence shown in the lower drawing.
Fig. 6.12 Even forming and lacing of power wires. Note sloped “waterfall” cable rack arrangement
where an elevation change was needed. (Photo courtesy of Schultz Brothers Electrical Company.)
앫 When forming wire vertically down toward equipment shelves from the cable rack,
start forming at the top and bundle downward.
앫 Form wires in such a way that outside wires do not disappear into the bundle.
앫 Do not bundle power wires with signal cables.
앫 Where only one level of cable rack is available, and a larger number of signal cables
than power wires needs to be run, run the power wires on stand-off brackets mount-
ed on the cable rack stringers; in the reverse situation, run the signal cables on
stand-off brackets.
앫 Never route wires through or between cable rack cross members; always route
wires over the side of the cable rack stringers.
앫 Never route power wires through cable rack wall brackets.
앫 To avoid accidental shorts during installation, always tape the bare ends of cut pow-
er wires before running.
앫 Do not fill a cable hole more than 75% of its cross-sectional area, and do not place
cables closer than 3 in. to edges of cable holes (there must be adequate space for
fire-stopping materials).
앫 On vertical runs in equipment frames, route the dc load circuit wires closest to the
framework for lower mounted equipment shelves or chasses, progressively moving
the circuit wires outward for higher mounted shelves.
6.6.2 Guidelines for Securing and Physically Protecting Power Wiring

앫 Provide a neat appearance by banding the wire leaving cable racks and entering
equipment frames.
앫 Secure power wires on every cable rack cross member (every 9 in.) on vertical runs
(any run that is > 60° from horizontal is considered vertical) and on at least every
fourth cross member (every 36 in.) on horizontal runs; it is typical practice to secure
on every other cross member on horizontal runs.
앫 Secure wires where they leave the cable rack (Fig. 6.13).
앫 Support 1/0 AWG and smaller power wire no farther than 24 in. and 2/0 AWG and
larger wire no farther than 36 in. after leaving a cable rack and entering an equip-
ment frame. Secure the wires at the first support at the top of the equipment frame.
앫 Secure power wires within 36 in. of their termination point.
앫 Use lacing twine or cable ties (tie-wraps) to secure power wires; see respective
guidelines.
앫 Protect all wires with sheet fiber (vulcanized fiber sheet or “fish paper”) wherever
passing over cable rack stringers or coming into contact with sharp edges.
앫 Protect soft rubber and neoprene insulated power wires from compressive damage
and sharp edges with sheet fiber where it is secured (Fig. 6.14); wrap with two lay-
1 1
64-in. or one layer of 32-in. sheet fiber. Insulated wires with a “textile” (cotton
ers of –– ––
braid) jacket or a cross-linked polyethylene (XLPE) jacket normally do not require

sheet fiber protection where they are secured, but extra protection should be provid-
ed where they can rub metal edges.
앫 Install permanent labels on all power wires near their termination; use tie-wraps or
lacing twine to fasten the labels.
앫 Do not use a pileup larger than 7 in. on vertical runs.
앫 If a vertical run exceeds three floors, break it up with a 20 ft horizontal run (see
Chapter 5, System Design).
앫 To prevent sag at corners or junctions, install a 1–8-in. × 1-in. flat steel bar diagonally
across the corner and secure with lacing twine (Fig. 6.15).
Vulcanized fiber sheet, commonly called fish paper, may be obtained in sheets, rolls,
and hollow tubes and is specified in ASTM D710-97, Standard Specification for Vulcan-
Fig. 6.13 Securing power wires. (Photo courtesy of Schulz Brothers Electric Company.)
Fig. 6.14 Protecting insulation with vulcanized fiber card stock (fish paper). Note sheet fiber cards
wrapped around wires adjacent to bolt heads and held in place with cable ties.
ized Fibre Sheets, Rods and Tubes Used for Electrical Insulation.6 Fish paper originally
was made from old cotton rags that were treated with zinc chloride to dissolve the cellu-
lose. The paper obtained from this process does not have a definite fiber structure but is
hard and tough in both directions and does not delaminate. Although most commonly
used as electrical insulation (it has a dielectric strength of 200 to 400 V/mil), its purpose
here is to provide physical protection of the wire insulation wherever the insulation could
be abraded or deformed.
6.6.3 Lacing (Stitching or Sewing) Power Wires

Two lacing stitches are most commonly used: “Kansas City” and “Chicago” stitches.
Generally, the Kansas City stitch is used to secure power wires to cable rack cross mem-
bers, cable brackets, and supports, and the Chicago stitch is used to band wires where not
secured. Both use the same starting stitch method as shown in Figure 6.16(a). Subsequent
stitches differ. Refer to Figure 6.16(b) for the Kansas City stitch on the first layer and Fig-
ure 6.16(c) for the second and subsequent layers. Refer to Figure 6.16(d) for the Chicago
stitch and Figure 6.16(e) for ending either the Kansas City or Chicago stitches with a
square knot or Hawthorne knot. Finally, refer to Figure 6.16(f) for a method to splice dou-
ble twine.
Lacing twine generally is installed in multiples of two strands. The required number of
lacing twine strands and number of wires held by each stitch depend on the wire size and
whether the wires are run horizontally or vertically (Table 6.10).
6
The name fish paper is thought to come from its original use as a tabletop surface in the London fish markets.
Fig. 6.15 Plan view of wire support at turns and corners using a locally fabricated steel bar.
6.6.4 Using Cable Ties on Power Wiring

Cable ties (tie-wraps) are controversial because they are sometimes used improperly. If
applied and installed properly, they can be a time saver and will provide a trouble-free
and safe installation. The two main objections to their use with power wire is the possibil-
ity of insulation creep (cold-flow) under the tie and personnel injury and insulation dam-
age caused by the sharp edges left by improperly cut tails. Although both problems can be
mitigated by proper installation and application, some company practices do not allow ca-
ble ties except in limited circumstances.
Although cable ties are available in a number of different materials, the most common
in electrical installations are nylon and vinyl. Nylon is hygroscopic (absorbs moisture
from the air) whereas vinyl is nonhygroscopic and performs better in both low- and high-
humidity environments.
6.6.5 Cable Tie Application and Installation Guidelines

앫 Do not overtighten and do not tighten so much that the wire insulation is deformed.
앫 Use the cable tie manufacturer’s recommended tool for tensioning and cutting; gen-
erally, the tool is set to 18 lb tension and cuts off the tail flush with the tie head. If
the manufacturer’s tool is not available or cannot be used, ties may be tightened by
hand; however, do not overtighten the tie and always use a flush cutting tool and
check for a sharp edge at the “eye” of the tie. If a sharp edge exists and it cannot be
placed in a position that eliminates the possibility of personnel injury or other prob-
lem, quickly touch the sharp protrusion with the flat side of a hot soldering iron tip
(being careful to keep the hot soldering iron parts from touching anything else).
Note: Some fire detection systems may detect smoke from hot soldering irons.
앫 Use cable ties to band wires where they leave a cable rack and between L-brackets
mounted on cable rack stringers.
앫 Use cable ties to band fiber card stock to cable rack cross members and stringers,
but be sure to locate the cable tie head where it cannot contact the insulation on oth-
er wires.
(a)
(b)
Fig. 6.16 (a) Starting stitch for Kansas City and Chicago stitching methods. (b) Kansas City stitch-
ing method for securing wire to a cable rack. The method is illustrated on a bundle of four wires in
two layers, but it also can be used on individual wires or groups of two in one layer.
(c)
(d)
Fig. 6.16 (c) Kansas City stitching method for securing second and subsequent layers of wire to a
cable rack. (d) Chicago stitching method for banding wires. (continued)
403
(d)
Fig. 6.16 (d) Chicago stitching method for banding wires.

(d)
(e)
(f)
Fig. 6.16 (d) Chicago stitching method for banding wires. (e) Dead-end knot (ending stitch) for
Kansas City and Chicago stitching methods. (f) Splicing method for double twine strands.
405
Table 6.10 Securing Horizontal and Vertical Power Wire Runs—Typical Practicea
Secure at
Wire Size ... Number Wires Ultimate Maximum
(AWG or Cross of Twine per Number of Pileup
kcmil) Member Strands Stitch Layers (in.)
Horizontal Runs
14–8 Every 2 2 Any Any
6–1 Every 2 4 Any Any
1/0 Every 2 2 Any Any
2/0–4/0 Every other 2 2 Any Any
250–350 Every other 2 2 Any Any
400–750 Every other 4 2 Any Any
Vertical Runs
14–8 Every 2 Any Any 7
6–1 Every 2 1 Any 7
1/0 Every 4 1 3 7
2/0–4/0 Every 4 1 3 7
250–350 Every 4 1 >3 7
400–750 Every 4 1 Any 7
a
Some variations exist; for example, on horizontal runs, power wires may be secured on every other or every
third or every fourth cable rack cross member regardless of wire size. However, on vertical runs, most practices
require that wires be secured on every cross-member.
앫 Use cable ties to secure power wiring within a power equipment frame or other
equipment frame; apply in a crisscross arrangement for best results (Fig. 6.17).
Note: It is common practice to use twine rather than cable ties on the top bracket
where the wires first enter the top of the frame.
앫 Wrap all rubber and neoprene insulated power wire with protective sheet fiber cards
before securing the wire to a cable rack with a cable tie; secure the sheet fiber to the
wire with cable ties or twine.
6.6.6 Drop Wires from Equipment Frame Fuse Panels

Small hookup wires dropped from a fuse panel at the top of an equipment frame are easi-
er to handle and bundle when the wire pairs are twisted. The twisting also helps minimize
noise coupling from one power circuit to another.
The procedure shown in Figure 6.18 uses an aviation safety-wire twisting tool avail-
able from any aviation tool supply company.
6.6.7 Wire Taps

Taps generally are used to reduce the conductor size where the main conductor is too large
for a termination and in some equipment frame bonding and grounding arrangements (see
Chapter 5, System Design). Where it is necessary to tap a smaller power (nongrounding)
conductor to a larger main conductor, the smaller conductor must be protected by the over-
current protection device feeding the larger main conductor. For example, say a small sec-
ondary distribution fuse panel is fed by a 50-A circuit breaker in the primary distribution
system and the input terminations on the fuse panel accept conductors no larger than 6
Fig. 6.17 Cable tie installation on a support by crisscrossing on the front.
Fig. 6.18 Twisting small hookup wire (16, 14, 12, and 10 AWG) using an aviation safety-wire twist-
ing tool (Milbar brand shown). Secure the wire in a vise or clamp at the far end. Comb out any twists
and crossovers. Place the free end in the jaws of the tool, squeeze the handle, and set the lock. Pull
the twisting handle several times until the two wires have the desired twist. Do not overtwist such
that the insulation is deformed; generally, use 1-in. twist length for 16 AWG, 1.25-in. twist length for
14 AWG, 1.5-in. twist length for 12 AWG, and 2-in. twist length for 10 AWG hookup wire.
AWG. However, because of the circuit length and voltage drop requirements, a 2 AWG
conductor is required. In this case, short pieces, or pigtails, of 6 AWG (no more than 1 or 2
ft long) would be tapped to the main 2 AWG feed and return conductors at the fuse panel
end. The 6 AWG is adequately protected by the 50-A circuit breaker, and the pigtails are
short enough to have no appreciable effect on the circuit voltage drop.
6.6.8 Tap Installation Guidelines

앫 Where the tap wire is smaller than the main wire, be sure the smaller wire is pro-
tected by the overcurrent device.
앫 Only use irreversible compression-type H-taps; do no use mechanical-type H-taps
(for the same reason that mechanical-type connector lugs should not be used), and
do not use C-taps (because C-taps do not have a separate crimp station for the tap
conductor).
앫 Crimp H-taps with the tooling recommended by the connector manufacturer, typi-
cally a circular die (do not use a hex die).
앫 Use an insulating cover designed for the tap connector.
앫 Place taps between cable rack cross members to minimize cable rack pileup (Fig.
6.19).
6.6.9 Repairing Damaged Insulation

If wire insulation is damaged in any way (scratched, chipped, dented, or ripped), it must
be repaired. If the damage is not too severe, wrap with a minimum of two half-lapped lay-
ers of rubber tape followed by two half-lapped layers of electrical tape. Extend the rubber
and electrical tape a minimum of 2 in. past the damaged section. If the damage is severe,
cut out the damaged insulation and replace with a section from a loose piece of the same
size wire.7 Cover the replaced section with one layer of half-lapped electrical tape and
heat-shrink tubing or two layers of half-lapped electrical tape or two layers of heat-shrink
tubing. Do not use electrical tape sold at automotive parts stores or department stores un-
less it is a name brand.
6.6.10 Busbars
Busbars are supported by auxiliary framing using hardware designed for the purpose. If
copper busbars have to be bent on the jobsite, the bend radius must be at least equal to the
bar thickness (Table 6.11). For example, a 1–4-in. busbar must have at least 1–4-in. radius on
the inside of the bend. Bends that are too sharp will crack the metal on the outside of the
bend and crush it on the inside of the bend.
6.6.11 Busbar Installation Guidelines

앫 Splice busbars and install T-junctions and L-joints with clamps; drill busbars only
where terminal connector lugs are to be installed.
앫 Clean all bare contact surfaces with a fine-grain abrasive paper (aluminum oxide
sandpaper) or scouring pad just prior to assembly; sand in a circular motion across
7
This type of repair is not recommended for wires used with systems operating at voltages higher than 60 Vdc.
Fig. 6.19 (Left) Application of taps to reduce wire size. (Right) Cross-section views of typical H-tap
connectors.
the metal grain, clean off sandpaper grit and metal oxides and coat with oxidation
inhibitor compound before connecting or assembling; wipe off excess oxidation in-
hibitor after assembly. If the parts are not assembled within a few minutes, cover
with paper until assembled.
앫 Do not use abrasives on lead- or silver-plated or tinned contact surfaces but clean
with a mild cleaner and dry cloth and coat with oxidation inhibitor.
앫 Some busbars may be factory coated with a clear sealant, which makes the busbar ap-
pear clean and bright; this sealant must be removed from the contact area with sand
paper or scouring pad and the area coated with oxidation inhibitor before assembly.
앫 Tighten busbar clamps by drawing up all four corners until snug. Start at one corner
and draw the fastener a little tighter. Cross to the diagonal corner and tighten the
fastener a little tighter. Tighten the fastener in the adjacent corner and then
crossover and repeat. Continue tightening each fastener a small amount using the
same sequence until the required torque values are reached. Clamps are slightly
convex and will bend flat when properly tightened. Overtorquing will bow out the
clamp or bend the ears and cause a loss of contact surface area. Do not use clamps
that have a concave busbar contact surface or bent ears because they are defective.
Table 6.11 Minimum Bend Radius for Copper Busbar (Measured

on the Inside of the Bend)a
Busbar Thickness (in.) Minimum Bend Radius (in.)
1
– 1
–
4 4
–3 –3
8 8
1
– 1
–
2 2
–3 –3
4 4
a
Generally, the bending radius should be at least equal to the bar thick-
Fig. 6.20 Pal nut: Install with flat side toward nut.
앫 Install locknuts (self-locking nuts) or a combination of ordinary hex nuts and Pal
nuts (Fig. 6.20) on all busbar clamps. Torque the hex nuts before applying the Pal
nuts. Install the Pal nut with the smooth (flat) side in, open side out; run to the regu-
lar nut until snug and then one-quarter turn beyond. Generally, if a connection has a
Pal nut and needs to be taken apart, the old Pal nut should be replaced by a new one.
앫 Use zinc-plated (SAE J429 Grade 5) or cadmium-plated and chromate-treated hard-
ware or silicon-bronze (ASTM B-99) or copper finished hardware with American
Unified National Coarse (UNC) free-fit (Class 2 fit).
앫 Busbars 1–4 in. and thicker may be drilled and tapped.
앫 For untapped connections, oversize drilled holes in busbar by –– 1 1
–– 1
64 to 32 in. (16 in. also
––
is acceptable).
앫 Support busbars at 6-ft intervals or less; install supports as close as possible to right
angle turns and risers. In high-risk seismic areas, support busbars on both sides of a
mechanical splice.
앫 Over passageways, provide a minimum busbar height of 7 ft above floor level.
앫 Provide at least 3-in. clearance between busbars of different voltages and between
busbars and cable racks, auxiliary framing, threaded rods, equipment frames, and
other conductive apparatus. Where possible, increase the clearance to 12 in. If the
3-in. minimum clearance cannot be achieved, the clearance can be reduced to 1–2 in.
for 24- and 48-Vdc systems but only in low-risk seismic areas.
앫 Wrap busbars with sheet fiber card stock and tape where protection is required to
prevent contact with lighting fixture diffuser covers and frames (preferably, remove
or relocate the fixture).
앫 Where splicing, joining, or extending busbars, overlap the busbars by at least the
busbar width but no more than the busbar width plus 2 in.
앫 Install all busbars arranged in an L or T at the joint so that the adjoining bars are
even and flush wherever possible; if not possible, use a maximum extension of 1–2 in.
for the L-joint and 2 in. for the T-joint.
앫 Where exposed busbars are located outside the power system room or space, such
as above or below secondary power distribution frames, cable racks, and auxiliary
framing, protect with noncombustible covers or wrap with two half-lapped layers of
plastic insulating tape (e.g., Scotch 33 or 88 vinyl electrical tape).
앫 It is good practice to tape the grounded return busbars where they are close to ex-
posed live terminals, studs, and equipment. Similarly, it is good practice to tape live
busbars where they are close to grounded terminals, studs, and equipment.
앫 Install busbars with the long edge vertical wherever possible.
앫 Observe current ratings corresponding to the busbar orientation (vertical or horizon-
tal); see Chapter 5 (System Design).
6.7 POWER EQUIPMENT FRAMES 411
6.7 POWER EQUIPMENT FRAMES
Power equipment may be installed in any common frame type, including unequal flange
and channel-type frames, box-frames, and cabinets. Power equipment frames must be
able to support substantial loads (equipment weight and cable rack, overhead structure,
and seismic loads), which usually precludes the use of light-duty frames in power applica-
tions.
6.7.1 Frame Installation Guidelines

앫 Bolt adjacent frames together using hardware made for the frames. Where unequal
flange frames are used in the same lineup with channel-type frames, adapters must
be used to bolt the two different types together or the different frame types should
be separated by at least 2 in. and independently supported.
앫 Use a top angle on all channel-type and unequal flange equipment frames, and in-
stall a 1-in. rigid steel conduit underneath the top angle along the full length of the
lineup to stiffen the lineup (this is effective for stiffening a lineup consisting of dif-
ferent frame types such as channel and unequal flange).
앫 Support mechanically isolated frames and frames at each end of a lineup with two
top supports; other frames in the lineup may be supported by one top support
앫 Plumb frames to the tolerances shown in Table 6.12
앫 Anchor frames to the floor with a minimum of two fasteners, one in the front and
one diagonally across in the back; repeat the established pattern throughout the line-
up. Use four fasteners in seismic installations. The anchors must be suitable for the
seismic area of the installation
Table 6.12 Equipment Frame Lineup Tolerance

Maximum Deviation of Top
Frame Height from Vertical (in.)
4 ft 6 in. 1
––
16
1
4 ft 6 in.–7 ft 0 in. –
8
3
7 ft 0 in.–9 ft 0 in. ––
16
1
> 9 ft 0 in. –
4
Lineup length Maximum deviation from

straight horizontal line (in.)
1
Any ––
16
Maximum deviation between

adjacent frames (in.)
1
Any –
4
Maximum deviation from level

(in./ft of length)
1
Any (Note: no more than 1-in. ––
16
shim stack on any frame)
6.8 AUXILIARY FRAMING
Auxiliary framing consists of paired channels or bars (Fig. 6.21) installed as a structural
grid to provide mechanical support for equipment frames, cable racks, and other appara-
tus installed above equipment frames (Fig. 6.22). Two auxiliary framing levels may be re-
quired where the ceiling height is more than 5 ft above the equipment frames. The upper
level effectively lowers the ceiling height and uses a truss for bracing against lateral loads
and movement (Fig. 6.23). The upper level also may be used to support cable racks in a
multilevel cable racking scheme.
In some buildings, the ceiling may not be designed to provide resistance to lateral
loading during a seismic event. In this case, the auxiliary framing may be attached to
load-bearing walls and building columns, or a floor-mounted, cross-braced stanchion sys-
tem may be used.
6.8.1 Auxiliary Framing Installation Guidelines

앫 Use no more than 5-ft spacing between auxiliary framing levels. Run the low-level
framing in continuous lengths at right angles to the frames and equipment it sup-
ports.
앫 Support auxiliary framing with –58-in. diameter threaded rod attached to the ceiling or
higher level framing.
앫 For maximum rigidity, install the auxiliary framing in 20-ft stock lengths wherever
possible; however, use shorter lengths to stagger splices.
앫 Install all paired parallel bars and channels in equal lengths.
앫 Install auxiliary framing supports on 5-ft intervals where possible, but where neces-
sary to avoid obstructions extend to no more than 6-ft intervals.
앫 Use clip-type or through-bolt splices in low seismic areas with no more than one
splice between supports; use through-bolt splices in high seismic areas. Where
through-bolt splicing is used, drill or punch before it is installed. If after drilling or
punching, the holes do not line up with the auxiliary framing splice, enlarge or elon-
gate the holes, but use flat washers on the splice bolts. Do not exceed 1–4-in. gap be-
tween the ends of spliced auxiliary framing.
Fig. 6.21 Cross-section views: Parallel auxiliary channels and bars. The parallel components act as
beams to support cable racks, busbars, and other apparatus, which may be suspended above or be-
low. They also may support equipment frames from the top.
413
Fig. 6.22 Plan view: Basic layout for auxiliary framing on a 5-ft grid.
Fig. 6.23 Elevation view: Truss arrangement for multilevel auxiliary framing (the type shown is
called a Warren truss).
앫 Do not splice auxiliary framing past the last support (in other words, splices must be
supported on both sides).
앫 In low-risk seismic installations, provide at least 1–2-in. clearance between the ends of
the auxiliary framing and any building component, such as columns and walls; in
high-risk seismic installations provide at least 5-in. clearance.
앫 Install supports so that the horizontal distance between the last hanger or support
and the cable rack or other equipment load is no more than 36 in.
앫 Install auxiliary framing cross-bracing with 30° to 60° slope, preferably 45° (Table
6.13).
앫 Where the auxiliary framing is used to support only cable rack, install cross-bracing
as required to ensure rigidity.
앫 For personnel protection, install finishing caps and clips on all paired framing bars
and channels and on any exposed ends of framing installed or protruding below the
main auxiliary framing.
앫 To prevent bending when moving auxiliary bars and channels from one location to
another, carry them with the edge down and not flat.
앫 File rough or sharp edges at cuts and paint before installing.
앫 Measure the heights of the auxiliary framing from the floor to a common reference,
such as to the bottom of the paired channels or parallel bars.
앫 Cover threaded rods with vulcanized fiber tubes (ASTM D710-97) or PVC tubing
to prevent insulation damage during power wire installation. The tubes should be
long enough to protect the cabling to the ultimate pileup height.
앫 If the threaded rod end protrudes beyond the fasteners more than one diameter of
the rod, cut the rod end except where the protrusion is between the top nut and the
ceiling.
앫 Turn rods into ceiling inserts so that it contacts all the threads of the insert but do
6.9 CABLE RACKS 415
Table 6.13 Auxiliary Framing Cross-Bracing

Distance Between Auxiliary
Framing and Equipment Frames Cross-Brace Type
1 ft 6 in. Single rod
3 1
1 ft 6 in.–4 ft 10 in. 2 in. × 2 in. × ––
16 in. (or 4 in.)
–
thick angle brace
not turn so deep that it binds against the ceiling; install a nut and washer on the out-
side of the insert.
6.8.2 Additional Requirements for High-Risk Seismic Areas

앫 To prevent fasteners from sliding off, extend the end of the auxiliary framing at
least 3 in. beyond the last support point (hanger rod, brace, or cable rack attach-
ment) where possible; otherwise, install a –38-in. stop-bolt through the ends of the
auxiliary framing.
앫 Install stiffening clips on paired channels and parallel bars on no more than 24-in.
centers between auxiliary framing supports. Stiffening clips may be omitted at loca-
tions where an earthquake brace, cable rack, or other clipped fastening, 1–2 in. or larg-
er, is located.
앫 Extend conduits supported on auxiliary framing to wall-mounted switches, lights,
and outlets with flexible conduit or cable.
6.9 CABLE RACKS
Cable racks typically are installed in two levels, one for power and one for signal cables;
however, in small installations one level may be installed for both (Fig. 6.24). If two lev-
els are used, the upper layer should be separated by 12 to 18 in. from the ultimate pileup
on the lower level. Cable racks may be installed directly on the auxiliary framing, sup-
ported by threaded rods attached to the auxiliary framing, or suspended above it.
6.9.1 Cable Rack Installation Guidelines

앫 Install all cable rack level and aligned with existing cable rack or equipment.
앫 Install sufficient side bracing to ensure rigidity and prevent cable rack sway.
앫 Install J-bolts on both cable rack stringers at each end of the cable rack run and stag-
gered on alternate sides at each support in between; in high-risk seismic areas in-
stall J-bolts on both sides at each support.
앫 Where cable rack is spliced, keep the gap below –58–in.
앫 Except for elevation changes, do not install more than one cable rack splice be-
tween any pair of supports; do not splice beyond the last support.
앫 Support cable racks at 5-ft intervals wherever possible, but 6-ft intervals may be
used if necessary.
416
Fig. 6.24 Plan view: Basic layout for cable rack fastened to auxiliary framing.
6.10 SLEEVES, WALL OPENINGS, AND FLOOR OPENINGS 417
앫 Do not extend cable rack more than 36 in. beyond the last support.
앫 In low-risk seismic installations, provide at least 1–2-in. clearance between the cable
rack and any building component, such as columns and walls; in high-risk seismic
installations provide at least 5-in. clearance.
앫 Do not extend cable rack through a floor penetration (it cannot be adequately fire
stopped).
앫 Never support a cable rack by its cross members.
앫 Support vertical cable racks at the floor and install intermediate supports at no more
than 5-ft intervals.
앫 Place fastening clips or J-bolts on either the inside or outside of the cable rack
stringers.
앫 Place horizontal cable rack with the cross members at the top. A cable rack with
solid stringers, smaller than 25 in. in width, may be inverted to gain additional ca-
bling height where there are obstructions. Never invert cable rack that has rein-
forced cross members.
앫 Install unprotected battery circuit conductors on dedicated power cable racks wher-
ever possible.
앫 Do not use a cable rack larger than 20 in. on vertical runs.
앫 Provide at least 4-in. hand and arm clearance between the maximum pileup on a ca-
ble rack and any obstruction.
앫 Provide at least 18-in. working space for installation access on at least one side of
the cable rack.
앫 Install cable rack toward the front of equipment frames so that cables can be routed
over the side and down to the terminations at the rear of the frame.
6.10 SLEEVES, WALL OPENINGS, AND FLOOR OPENINGS
Temporarily close openings at the end of each workday or sooner if no further cabling
work is anticipated that same day:
Install the cover on one side (bottom for floor openings).

Fill the opening with fire-stopping bags to prevent any draft through the opening.
Install the other cover using a minimum of two screws.
Permanently close openings after all cables have been installed:
Cut the covers to accommodate the new cables; make straight cuts so the covers will
be close to the cables but not touch them.
Install one cover (bottom for floor openings).
Fill the opening with fire-stopping bags, overlapping each layer until the opening is full.
Install the other cover and seal between the cover and cables with sealing compound.
When a floor opening is uncovered to run cable, protect personnel by roping off the area
and placing signs similar to “Danger—Cable Opening.”
6.11 BASIC INSTALLATION SEQUENCE
The following sequence may not apply to all installations and should be altered to suit
field conditions or the particular type of equipment being installed.
1. Install ceiling-supported auxiliary framing.

2. Install powerboard equipment frames and remaining auxiliary framing.
3. Install all cable support systems, including cable rack and conduit systems.
4. If the charge and discharge busbars are separate assemblies, install them at the re-
quired locations.
5. Assemble battery rack (VLA) or place battery modules (steel encased VRLA).
The height of free-standing racks and the stacking height of battery modules may
be limited in seismic installations. Place cells using the procedures specified in
the battery manufacturer’s documentation. Install and torque cell interconnecting
straps on the cells. Do not connect the battery string to the charge and discharge
bus bars at this time.
6. Run all wiring between the charge bus and battery and between the discharge bus
and primary distribution. Do not connect the battery at this time.
7. Run the ac input power leads to each rectifier and run the dc output power leads
from each rectifier or rectifier shelf to the charge bus. Do not turn on ac power to
the rectifier system at this time.
8. Install external control and alarm cables. If remote battery voltage sensing is re-
quired, run the leads to the battery but do not connect them to the battery termi-
nals at this time; if the voltage sensing circuit is equipped with a fuse or circuit
breaker at the battery, all terminations may be made, but the fuse should be re-
moved or the circuit breaker turned off. If local battery voltage sensing is speci-
fied, connect the sense leads in the powerboard according to the manufacturer’s
instructions.
9. Run distribution leads from the power system overcurrent protection devices (cir-
cuit breakers or fuses) and return bus to the loads.
10. Run all remaining system interconnecting leads and alarms, but do not make the
final connections at the battery terminals at this time.
11. Perform a battery initial charge with a separate rectifier and temporary wiring.
12. Check out and set up power system controller according to the manufacturer’s in-
structions.
DANGER—The next step applies power to the battery system. Before con-
tacting any uninsulated conductor surfaces, always use a voltmeter to en-
sure that no voltage, or the expected voltage, is present.
13. Before connecting the battery, turn off or (if modular) unplug all rectifiers, open
all load distribution circuit breakers, and remove all load distribution fuses and
their associated alarm fuses.
14. Connect the circuit wiring between the battery terminals and the termination bus-
bars by first connecting to the termination busbars and then to the battery termi-
nals. To prevent sparking when the final connections are made to the battery, tem-
6.12 BATTERY SYSTEM INSTALLATION 419
porarily connect a regular household 60- to 100-W incandescent lightbulb as

shown in Figure 6.25.
15. Turn on the rectifiers, one at a time. Test each rectifier according to the procedures
given in the manufacturer’s rectifier manual.
16. Verify the operation of the fuse and circuit breaker alarms as follows: Short the
alarm leads on each circuit breaker and verify that the power alarm lamp on the
controller front panel or alarm monitor lights. To verify fuse alarm operation, in-
sert a blown fuse in each alarm fuse position. Verify that the appropriate fuse
alarm lamp lights.
17. Verify the operation of capacitor precharging circuits as follows: Place the circuit
breaker with a load precharge switch in the off position. Remove the load circuit
conductor. Connect a voltmeter between the circuit breaker output terminal and
return. Press the circuit breaker precharge switch and look for battery voltage
reading on the multimeter. Release the precharge switch and disconnect the multi-
meter. Reconnect the wiring previously removed.
18. If the power system is equipped with a low-voltage disconnect, test it according to
the manufacturer’s instructions.
WARNING—Follow the manufacturer’s power-up instructions before ap-

plying power to load equipment.
19. Connect the loads, one at a time, by turning on the load circuit breakers or insert-
ing the load (and associated alarm) fuses for each circuit.
6.12 BATTERY SYSTEM INSTALLATION
The applicable industry installation standards are [5] for VRLA and [6] for VLA battery
systems. Battery installations require consideration of structural and space requirements,
among other things. Adequate space must be provided around a battery for cooling air
Fig. 6.25 Using a 60- to 100-W lightbulb to raise battery voltage when connecting a battery to an
energized bus. The bulb acts as current-limiting resistor to provide a small charging current to the
battery string. The voltage difference between the power system bus and the battery will cause cur-
rent to flow through the bulb and light it dimly. When the bulb extinguishes completely, the battery
and bus voltages are the same and the connection may be made without a spark.
movement and to allow space to replace a cell or the entire battery. Structural require-
ments include the effects of the battery weight on the floor and during seismic events.
Another installation consideration is that VRLA batteries require less floor space than
VLA batteries because they are more compact and can be stacked (although stacking
height may be limited in seismic installations), but VRLA floor loading usually is higher.
Floor loading restrictions may limit the amount of space that can be saved in some build-
ings.
Vented lead–acid cells always are installed vertically (vent at the top) but many VRLA
cells may be installed either horizontally or vertically. Some VRLA battery types [e.g.,
AGM (absorbed glass mat) in steel modules] must be installed horizontally to ensure
proper wicking of the electrolyte in the cell separators. Manufacturer’s installation in-
structions should be checked to determine the proper orientation.
Good practice is to locate the battery rack or frame at least 6 in. from a wall but 12 in.
of clearance may be needed in seismic areas (Fig. 6.26). The cells themselves should have
at least 8 in. of clearance from a wall. At least one VRLA battery manufacturer does not
require clearance at the back of a battery, but an installation should never be made in this
manner without first discussing it with the manufacturer and also determining if it meets
seismic requirements.
In addition to spacing from a wall, adequate clearance must be provided for working
space in front of the battery and for removal (Fig. 6.27 shows an unacceptable installa-
tion). Many VRLA installations use a horizontal cell arrangement, and space must be pro-
vided to not only make electrical measurements and check fastener torque but also to pull
a cell or monobloc horizontally from a cell cage during replacement. The minimum clear-
ance typically is 30 in. in the direction of cell removal, but more may be required by a
particular battery type. Clearances between a battery rack and other equipment and build-
ing structure is shown in Table 6.14. Where the National Electrical Code applies, the
clearance requirements of Article 110 must be used (36-in. clear space from any ener-
gized component).
6.12.1 Cell Installation Guidelines

앫 Do not mix lead–calcium and lead–antimony VLA cells within a battery string or
within a battery system even in an emergency.
Fig. 6.26 Wall clearance requirements.

Fig. 6.27 Unacceptable battery installation in an access node equipment enclosure.
앫 Do not mix VRLA and VLA battery strings within a battery system even in an
emergency because of the different float voltages and charge and discharge charac-
teristics. Exception: To ensure battery service continuity when converting a facility
from lead–calcium VLA to VRLA, the voltage of the VLA may be raised to equal
the required VRLA float voltage, the VRLA string connected in parallel with the
VLA string, and then the VLA disconnected, all in one operation lasting a few min-
utes.
앫 Do not mix cells with different ampere-hour capacity within a string even in an
emergency.
Table 6.14 Minimum Clearances for Battery Racks

Minimum
Location Clearance (in.)
From battery rack to adjacent or parallel battery rack 36
From a double-row battery rack to a wall or permanent obstruction 36
From the back of a single-row battery rack to a parallel wall or permanent 8
obstruction
From a battery rack to equipment frames 36
From the end of a battery rack (6 ft or longer) to a wall or permanent obstruction 36
From the back of any cell to a wall or permanent obstruction 6
앫 Be sure flame arrestor vents are in place on VLA cells when making connections at
the battery. The flame arrestor vent funnel must extend below the surface of the
electrolyte.
앫 VLA cells exposed to prolonged agitation, such as during shipment, can build up a
dangerous concentration of hydrogen gas at the top of the cell in the space between
the electrolyte and cover. Use extreme caution and care when handling cells that
were just delivered. Take particular care to prevent static electricity from discharg-
ing near the filling funnels and vents even with the shipping plugs in place. The gas
concentration will dissipate after the cells have rested for 24-h. Take normal pre-
cautions after the 24 h period.
앫 Inspect newly delivered cells for signs of electrolyte leakage and any obvious ship-
ping damage before removing from the shipping containers. Look for dampness or
discoloration of the shipping containers and pallets (Fig. 6.28). If hoisting straps or
ropes have been exposed to electrolyte spillage, they could be unsafe to use. Be sure
to document any adverse findings.
앫 When handling cells where electrolyte spillage may occur, wear protective clothing:
앫 Chemical safety goggles
앫 Rubber gloves
앫 Coveralls
앫 Rubber apron
앫 Overshoes
앫 Remove the shipping cartons and closely examine the cells for physical damage to
the cell container (jar) and cover (top) and for missing electrolyte that has exposed
the plates (the latter applies to VLA only) (Fig. 6.29).
앫 Where more than one string has been delivered in the same shipment, the cells in a
given string will be factory matched by colored stickers (Fig. 6.30). There are four
Fig. 6.28 Cell shipping container with obvious discoloration from spilled electrolyte. (Photo cour-
tesy of M.W. Migliaro.)
Fig. 6.29 New cells still on the shipping pallet with one cell uncovered. (Photo courtesy of M.W.
Migliaro.)
basic colors and each represents a 10-mV difference in float voltages (it is neces-
sary to contact the manufacturer to determine the voltage corresponding to each
color). Generally, keep cells of a like color together and do not mix with cells with
an unlike color from another string. However, it normally is acceptable to use an-
other color cell in a group if it is within ±10 mV of the base voltage of the group.
For example, say the stickers are ordered white, yellow, green, and orange. It would
be acceptable to use a white or green cell with a yellow cell but not an orange cell
with a yellow cell. Similarly, a green cell should not be used with a white cell.
Fig. 6.30 Float voltage matching stickers are used when more than one battery string is in a ship-
ment. (Photo courtesy of M.W. Migliaro.)
앫 If the electrolyte has been spilled from a VLA cell and the plates exposed while
moving from the delivery truck into the battery room or storage location, the cell
can be saved if the exposure is less than 20 min and a small amount of electrolyte is
transferred from a number of other cells until the plates are covered (use the cells
with the highest electrolyte levels). If the exposure is more than 20 min, the cell
probably is permanently damaged.
앫 It is not unusual to have some electrolyte leakage from VLA cells during shipment
(from sloshing) that collects on the cover and sides of the cell. Neutralize the elec-
trolyte with a solution of 1-lb baking soda to 1-gal water. Never use solvents, min-
eral spirits, commercial detergents, ammonia, or other cleaning compounds or oils,
waxes, or polishes. Wet a cleaning cloth with the baking soda–water solution and
wring out sufficiently to prevent dripping. With the shipping caps in place, wipe the
container (jar) and cover taking care not to allow the solution into the cell (soda so-
lution can destroy the cell). If a part that is coated with oxidation inhibitor shows
signs of acid contamination, clean the parts thoroughly with a dry cloth and neutral-
ize with the soda solution. After neutralizing, rinse the cell and parts with clean wa-
ter. If a white residue appears afterwards, the cell should be rinsed again. Remove
the rinse water with a clean dry cloth. If oxidation inhibitor was removed from any
parts, recoat with a thin film.
앫 After all VLA cells are on the battery rack or VRLA modules are stacked, the cell
interconnecting straps can be installed.
앫 Clean all electrical contact areas prior to strap installation. It may be necessary to
remove oxidation inhibitor, clean with a cleaning pad, and recoat.
앫 Check all cell posts and terminals to be sure they have a thin coat of oxidation in-
hibitor (Fig. 6.31). As each interconnecting strap is installed, apply a thin coat on
that portion of the strap that will be in contact with the post. Also coat the contact
portion of the bolt head, stud, and nut before installing.
Fig. 6.31 Cell interconnecting straps.

REFERENCES (PART I) 425
앫 Almost all battery manufacturers use 316 stainless steel fasteners (bolts, nuts, wash-
ers) at cell post connections. Do not substitute.
앫 Stamped flat washers should be assembled with the smooth (“rounded) side toward
the cell post and the “sharp” side pointed away from the cell post.
앫 Do not tighten interconnecting strap fasteners beyond finger-tight until all straps are
in place. Once all straps are in place, tighten all fasteners to the manufacturer’s rec-
ommended initial torque value—do not overtorque or the cell post may be dam-
aged. At the time of final acceptance, tighten the fasteners to the manufacturer’s re-
torque values (retorque values may be the same as or different than initial values).
앫 Although different practices exist in the field, the following cell numbering scheme
may be used on 2-tier, 24-cell battery racks (the actual numbering scheme is not as
important as using the same scheme throughout the organization):
앫 Lower tier: Cells 1 to 12
앫 Upper tier: Cells 13 to 24
앫 Again, although different practices exist in the field, the lowest numbered cell is at
the positive end of the battery [6].
앫 In high-seismic areas, install factory-provided cell separators between cells.
앫 In low-seismic areas, install cell separators between cells if factory provided. If no
separators are provided, install cells so they do not touch each other or adjacent
framework and with a typical spacing of –38 to –58 in. Where two rows are installed on
one rack, use a spacing of –34 in. between rows.
앫 Never use foam packing material as cell separators.
6.12.2 Battery Circuit Wiring

Battery circuit wiring is sized based on current rating and voltage drop as discussed in
Chapter 5, System Design. In high-seismic risk areas, slack must be provided in all cable
runs from busbar to the battery terminals to allow 6 in. of movement. Refer to Part III,
Battery System Maintenance, for information and procedures for initial charging, testing,
and acceptance.
REFERENCES (PART I)
1. NFPA 701-2004, Standard Methods of Fire Tests for Flame Propagation of Textiles and Films,
National Fire Protection Association (NFPA), 2004 (www.nfpa.org).
2. ASTM D1048-05, Standard Specification for Rubber Insulating Blankets, ASTM, 2005
(www.astm.org).
3. ANSI/T1.311-1998, DC Power Systems—Telecommunications Environment Protection, ATIS,
1998.
4. NEMA CC 1-2002, Electrical Power Connection for Substations, National Electrical Manufac-
turers Association, 2002 (www.nema.org).
5. ANSI/IEEE Std 1187™-2002, IEEE Recommended Practice for Installation Design and Installa-
tion of Valve-Regulated Lead-Acid Storage Batteries for Stationary Applications, 2002.
6. ANSI/IEEE Std 484™-2002, IEEE Recommended Practice for Installation Design and Installa-
tion of Vented Lead-Acid Batteries for Stationary Applications, 2001.
7. E. Oberg, F. D. Jones, H. L. Horton, H. H. Ryffel et al., Machinery’s Handbook, 27th ed., Indus-
trial Press, New York, 2004 (www.industrialpress.com).
PART II SYSTEM MAINTENANCE
This part covers basic maintenance requirements for telecommunications dc power sys-
tems.
6.13 OVERALL MAINTENANCE REQUIREMENTS
Maintenance consists of inspection, overhaul, repair, adjustment, preservation, and parts

replacement. The main objective of scheduled maintenance is to reduce the number of
avoidable breakdowns and thus increase power system and network availability. The
costs of scheduled maintenance are high but not as high as unscheduled maintenance.
Whereas scheduled maintenance increases equipment reliability and system availability,
the lack of maintenance always decreases reliability and availability. The failure of dc
power systems in a telecommunications facility affects virtually every part of a network
operator’s operation at that facility. The cost of network outages caused by dc power sys-
tem failures usually amount to many times the actual cost of repair.
Maintenance can be categorized as proactive and reactive. Proactive maintenance con-
sists of routine and preventive maintenance. Routine maintenance helps to ensure contin-
ued operation of the dc power system, whereas preventive maintenance attempts to pre-
vent failures and is particularly important with battery systems. Reactive maintenance is
maintenance performed after a failure (also called breakdown maintenance) and consists
of the actions taken after a failure to restore equipment operation.
6.13.1 Maintenance Schedules

Maintenance schedules take into account that too frequent inspections are a waste of mon-
ey and that insufficient inspections allow components or subsystems to deteriorate and
jeopardize network operation. Deterioration of dc power system equipment is normal and
begins as soon as the equipment is installed. If this deterioration is not checked and cor-
rected, it can cause failures. Deterioration includes alterations that are made haphazardly
and without overall engineering coordination. The inspection frequency depends on the
equipment duty cycle and loading, predicted life and age, overload history, and other per-
tinent factors and is usually based on equipment manufacturers’ instructions and industry
standards.
The development of maintenance schedules involves establishing instructions, proce-
dures, and methods to ensure that the equipment and system components operate without
failure. The program should include procedures and instructions for thoroughly servicing
all equipment and components. These procedures should reference manufacturer’s oper-
ating and maintenance manuals to the extent possible. The maintenance program should
include shutdown procedures, safeguards, equipment interlocking and lock-out proce-
dures, the meaning of alarms and response to them, and methods of recording data and re-
porting unusual conditions to the proper authority.
6.13 OVERALL MAINTENANCE REQUIREMENTS 427
6.13.2 Maintenance Analysis

Equipment failures should be analyzed to determine the reasons for failure. The cause
may be obvious. Although reliability can be built into dc power equipment, it generally
requires continuous upkeep to retain it. Continuous upkeep is particularly important to
batteries. Unfortunately, the tendency to ignore regular maintenance generally prevails
over regularly scheduled maintenance because regular maintenance may be considered
unnecessary, too expensive, or of lower priority than other activities.
6.13.3 Tools and Test Equipment

앫 Multimeter, 0.02% dc accuracy, 1-mV dc resolution, 0.2% ac accuracy, 100-mV ac
resolution
앫 Clamp-on ammeter, 1 or 2% accuracy—ac/dc
앫 Noncontact thermometer
앫 Screwdriver assortment (Phillips No. 0, 1, 2, and 3 and slotted 1–8-in., 16 3
–– -in., 1–4-in., and
5
16-in. blade)
––
앫 Nut driver assortment (1–8 in. through 16

9
–– in.)
앫 Torque wrenches, adjustable click-type (40 to 200 in.-lb and 15 to 250 ft-lb)
앫 Socket set, nonmetric ( 16
3
–– in. through –34 in., drive size to match torque wrench)
앫 Socket set, metric (5 through 20 mm, drive to match torque wrench)
앫 Torque screwdriver, cam-over torque-limiting clutch-type (5 to 40 in.-lb)
6.13.4 Test Equipment and Tool Calibration

All test equipment, including torque wrenches, should be periodically calibrated (Table
6.15), and calibration records should be kept showing date and results of all calibrations
and tests.
6.13.5 Fuse Replacements

A common field problem occurs when a fuse is replaced. Through ignorance, inadequate
markings, or lack of spare fuses of the proper rating, technicians sometimes use a replace-
ment fuse rating that is too small or too large. Within certain maximum and minimum rat-
ings, fuse dimensions usually are the same; for example, it is possible, but never acceptable,
to substitute a 20-A fuse where a 5- or 10-A fuse should be used. Often when a correctly rat-
ed replacement is not available (or the correct fuse rating blows too often), a higher rating
is substituted although this never is acceptable practice. The only way to prevent these
Table 6.15 Test Equipment Calibration Frequency

Type of Test Equipment Calibration Interval Remarks
Field test sets and tools 6–12 months Depends on handling and use
Laboratory test sets 12 months
Leased specialty equipment 12 months
kinds of problems is to properly mark all fuse positions, equip the site with an adequate sup-
ply of spare fuses of the proper ratings, and to train operating personnel of the hazards in-
volved in using incorrect fuse sizes. The most important hazards are injury or fire.
Many older fuse panels with fuse holders rated < 30 A use 1–4-in. diameter × 11–4-in. long
AGC (glass) or ABC and MDL (nonglass, such as bakelite, fiber, ceramic) cartridge fus-
es. These fuses also are known as type 3AG and 3AB. Glass cartridge fuses should be
avoided because of the obvious cleanup difficulty if they are broken. Also, they may ex-
plode if accidentally used in an application that supplies higher fault currents than the
fuse is rated to handle.
9 1
Many Northern Telecom (Nortel Networks) fuse panels require –– 32-in. diameter × 1 4-in.
–
long type ABS cartridge fuses, but this is not marked on the panel (these fuses also are
known as type 4AG or 4AB). A common mistake is to install 1–4-in. diameter AGC or ABC
fuses in these panels. Since the fuse holder does not properly clamp the fuse, it will over-
heat and eventually blow the fuse for no apparent reason. In the process of replacing the
fuse many times, the fuse holder will be ruined by arcing. The problem is compounded by
the fact that ABS fuses are hard to obtain.
Type GMT alarm indicating fuses are available to up 15 A, but many GMT fuse panels
are rated 10 A maximum per fuse position. Therefore, craft personnel must be careful to
never install a fuse rated higher than 10 A in these fuse panels. Unfortunately, even
though some of these panels are marked with the maximum fuse rating, the label is hidden
on the side or on the back of the fuse panel where it is unreadable (a textbook example of
poor engineering and manufacturing design).
6.14 ANNUAL SYSTEM MAINTENANCE CHECK
앫 Review maintenance logs and alarm logs.

앫 Review overall system operation with site personnel, and record and report any rec-
ommended power system enhancements or equipment operation changes.
앫 Check each system for general operation; perform rectifier system maintenance as
described in the next section and battery system maintenance as described in Part
III of this chapter.
앫 Check the power equipment room environment for temperature, dust, and humidity
control.
앫 Check the operation of room vents and fans.
앫 Visually check all equipment, cable racks, and supporting structures for abuse,
smoke damage, and general mounting security (particularly important in seismical-
ly active areas).
앫 Open all doors, drawers, and covers and examine all power equipment frames for for-
eign objects, dead animals, and nests. Dust and vacuum the interiors of all dc power
equipment frames using only rubber or plastic vacuum attachments (never attempt to
disassemble and clean subassemblies while in service). Exterior panels can be
cleaned with a mild cleaning solution. Transparent panels should be removed and
cleaned with a mild soap and water solution or a cleaner specifically made for the
panel material. When finished, restore all doors and covers to their secured condition.
앫 Check and test the dc power system controller using its built-in diagnostics (if
equipped and only if the tests are not service affecting).
6.15 RECTIFIER SYSTEM MAINTENANCE 429
앫 Examine all power wiring for signs of overheating, discoloration, abrasions, burns,
and damaged insulation; pay particular attention to locations where the wiring is se-
cured to metallic components such as cable racks and cable brackets.
앫 Examine all visible components for signs of overheating, swelling, and leaking.
앫 Scan all power circuit connections and components using a noncontact thermome-
ter; if any abnormal temperatures are found, check for loose connections and re-
torque as necessary.
앫 Replace or clean air filters at regular intervals. Site conditions will determine how
often the filters should be serviced, but generally, they will need to be cleaned or re-
placed at least every 6 months in clean environments. If more frequent cleaning or
replacement is required, consider upgrading the facility’s dust control measures.
앫 Check system alarms for proper thresholds (overcurrent alarm, undercurrent alarm,
system minor alarm, system major alarm, system critical alarm, fuse and circuit
breaker alarms, high- and low-voltage alarms, low-voltage disconnect alarm).
앫 Simulate a system or component failure to check overall alarm system operation
(but avoid any simulation that could possibly affect service).
앫 Check all meters (voltmeters and ammeters) for proper calibration.
앫 Verify that the bonding connections between the dc power equipment frames and
racks and the floor ground bar are secure and in good condition (small installations
may have only a building principal ground bar).
앫 Verify that the bonding connection between the dc power system return bar and the
grounding system is secure and in good condition.
앫 Measure and record the dc load currents at the main shunt and all distribution
shunts; look for unexpected changes (either higher or lower).
앫 Check building and other metallic components for corrosion or discoloration; for
example, discolored door hinges and air conditioning vents indicate the presence of
corrosive chemicals or hydrocarbons (smoke). Hydrogen embrittlement can occur
on hardware that uses electroplated finishes, particularly in battery rooms.
앫 Resolve any previous outstanding problems.
6.15 RECTIFIER SYSTEM MAINTENANCE
Modern rectifier systems require little maintenance, but the rectifiers at most remote or
unattended sites require at least a periodic cleaning. Particular care must be exercised
when working on rectifiers because of the dangerous ac voltages on their inputs. Almost
all procedures described in this section, except references to ac voltages and currents, also
apply to dc–dc converter systems.
6.15.1 Annual Rectifier System Check

앫 Before performing any maintenance, verify proper operation of the rectifier system.
앫 If rectifiers are equipped with fans, check their operation. Note: Some fans operate
only when the internal rectifier or ambient temperature is above a preset threshold.
In this case, the rectifier or rectifier controller may be equipped with a fan test
switch to verify operation.
앫 Measure and record ac input voltage and current for each rectifier.
앫 Measure and record rectifier dc output voltage and current. The output voltage
should be set accurately for the battery technology used (see Part III, Battery Sys-
tem Maintenance).
앫 Check rectifier load sharing (see Section 6.15.2).
앫 Check rectifier temperature-compensated charging if equipped (see Section 6.15.3).
앫 Measure and record ac ripple voltage across the battery terminals with a true-rms ac
voltmeter (a true-rms meter accurately measures ac waveforms with harmonics).
앫 Measure and record ac ripple current on the battery circuit with a true-rms clamp-on
ammeter set to measure ac.
앫 Compare rectifier panel meter readings with a calibrated multimeter. Many modern
modular rectifiers are equipped with a simple low-resolution bar-graph display used
for both voltage and current, so its accuracy does not need to be measured.
앫 Examine all visible rectifier components for signs of overheating, swelling, and
leaking.
앫 To the extent possible, scan the internal components, circuitry, and all rectifier-re-
lated cabling and connections with a noncontact thermometer. Note: Active semi-
conductor components inside a rectifier may be designed to run hot and if the recti-
fier operation is otherwise normal, high measured temperatures of these devices
normally are no cause for alarm.
앫 Scan the rectifier ac load center (ac distribution panel) with a noncontact thermome-
ter and look for overheated circuit breakers. Overheated circuit breakers can be
caused by overload or by a poor electrical connection to the panelboard bus or by a
poor connection to the load circuit wiring.
앫 Clean the rectifiers with a vacuum. If the rectifiers are standalone types, turn off the
ac input at the circuit breaker panel before cleaning and verify no input voltage is
present; if the rectifiers are modular, remove them one at a time and clean. Do not
turn off or remove any rectifiers unless the remaining rectifiers are capable of carry-
ing the load.
6.15.2 Rectifier Load Sharing Tests

The rectifiers should share the load in proportion to their rating (within approximately
10%) according to
IL
IRNL = IRNC (6.4)
IRTC
where IRNL = load current on rectifier N (A)

IRNC = full load capacity of rectifier N (A)
IL = load current (A)
IRTC = total rectifier full-load capacity (A)
For example, if the total load on two 50-A rectifiers and one 100-A rectifier (total rectifi-
er full-load capacity of 200 A) is 100 A, the two small rectifiers each should carry 25 A
and the large rectifier should carry 50 A.
6.15 RECTIFIER SYSTEM MAINTENANCE 431
Modern modular rectifiers use a common controller for output voltage and load shar-
ing, and there are no separate load sharing adjustments. However, standalone rectifiers
generally are self-controlled and require independent adjustment. First, the load share
control lead output voltage must be adjusted using the procedures specified in the manu-
facturer’s instruction manual. Then, the float output voltage of each rectifier is adjusted so
that they properly share the load.
The following adjustment procedures are easy when there are three or less rectifiers
but can be used with a larger number. The best way to make these adjustments is to set the
rectifiers to the proper output voltage and then very slightly adjust each one until they all
share the load in proportion to their rating. If one rectifier is taking more than its share of
the load (“hogging”), its output voltage is slightly higher than the others. Very slightly re-
duce this rectifier’s output voltage to reduce its load. If one rectifier is not carrying any
load at all, its output voltage is slightly lower than the others. Very slightly increase its
output. Repeat this basic procedure for all rectifiers until they are proportionally sharing
the load. All output voltage changes should be made in very small increments. Once all
rectifiers are sharing the load, check the battery terminal voltage. If it is low, then start
with one rectifier and very slightly increase its output. If the voltage is high, then start
with one rectifier and very slightly decrease its output. Repeat for all rectifiers until the
proper voltage is measured at the battery terminals and all rectifiers are sharing the load.
It may be necessary to repeat this procedure several times.
The sharing accuracy generally is better when the rectifiers are operating above 50%
of their output rating (typically they will share within 10 to 20% of each other). The recti-
fiers may not share the load at all when operating at less than 10% of rating, and one or
more rectifiers may show a low-current alarm even though the rectifiers are fully opera-
tional. This situation is encountered at sites where the load is relatively small and the bat-
tery reserve time is relatively long. A large rectifier system capacity is required to
recharge the battery and, once it is recharged, the rectifier output current decreases
enough that the load sharing function is inaccurate.
6.15.3 Temperature-Compensated Charging Tests

If temperature-compensated charging is used, the temperature of the cell on which the
sensor is mounted needs to be determined to ensure the rectifiers are adjusted properly.
The sensor temperature should be checked with a noncontact thermometer such as an in-
frared (IR) thermometer or an insulated contact thermometer. The amount of compensa-
tion in the rectifier will determine the rectifier setting. For example, if the rectifier com-
pensates by reducing its output voltage by 3.0 mV/°C/cell and the cell temperature is
28°C, the compensation is (28° – 25°) × 3.0 mV/cell = 9 mV (0.009 V) per cell. If the bat-
tery consists of 24 cells and the desired voltage is 54.48 V at 25°C, the new voltage is
54.48 – (0.009 × 24) = 54.26 V at 28°C. Similar calculations can be made for tempera-
tures in degrees Fahrenheit. For example, if the rectifier compensation is 2 mV/°F/cell
and the temperature of a 12-cell battery is 80°F, the new voltage is (80° – 77°) × 2
mV/cell = 6 mV (0.006 V)/cell. If the desired voltage is 27.24 V at 77°F, the new voltage
is 27.24 – (0.006 × 12) = 27.17 V at 80°F.
Temperature-compensated charging may be tested by directing warm air from a hair
dryer or heat gun onto the battery temperature sensor while simultaneously measuring the
battery float voltage [for safety, do not use an all-metal heat gun, and hold the heat source
at least 12 in. (300 mm) from the sensor so it is not overheated]. As the temperature of the
sensor rises, the voltage should drop. There will be a time delay and the voltage drop may
be small, so the voltmeter will have to be watched closely. When the voltage drops, the
rectifier current also will drop until the battery discharges to the new voltage. When the
heat source is removed and the sensor cools down, the voltage should return to a value
normal for the cell temperature.
6.16 INVERTER SYSTEM MAINTENANCE
Particular care must be exercised when working on inverters because of the dangerous ac
output voltages.
6.16.1 Annual Inverter System Check

앫 Before making any measurements, verify proper operation of the inverter system.
앫 Measure and record dc input voltage and current for each inverter.
앫 Measure and record inverter ac output voltage, current, and frequency.
앫 Check inverter load sharing.
앫 Compare inverter panel meter readings with a calibrated multimeter.
앫 If inverters are equipped with fans, check their operation. Note: Some fans operate
only when the internal or ambient temperature is above a preset threshold. In this
case, the inverter or inverter controller may be equipped with a fan test switch to
verify operation.
앫 Examine all visible components for signs of overheating, swelling, and leaking.
앫 To the extent possible, scan the inverter internal components, subassemblies, and
circuitry with a noncontact thermometer; if any abnormal temperatures are found,
check for loose connections and retorque as necessary. Note: Active semiconductor
components inside an inverter may be designed to run hot and if the inverter opera-
tion is otherwise normal, high measured temperatures of these devices normally are
not cause for alarm.
앫 To the extent possible, scan all inverter power circuit connections with a noncontact
thermometer; if any abnormal temperatures are found, take the inverter offline (see
next section), check for loose connections and retorque as necessary.
앫 Scan the inverter ac output load centers or outlet panels with a noncontact ther-
mometer and look for overheated connections.
6.16.2 Offline Inverter Maintenance

앫 Do not perform offline maintenance unless the inverter loads can be powered from
an alternate source or else the loads themselves may be shut down.
앫 Transfer inverter loads to an alternate source and shut down the inverter system as
detailed in the manufacturer’s instruction manual.
앫 Verify that all input and output voltage sources are securely locked-out and isolated
for maintenance.
DANGER—Prior to examining or cleaning the inside of the inverters, use

an ac multimeter to check all areas of the inverter that may be energized
6.17. GENERAL REQUIREMENTS 433
and be cautious and alert at all times. Every circuit or connection is “hot”
(energized) unless proven otherwise. Always use the “one hand in the pock-
et” rule.8
앫 Verify that all electrical connections are properly tightened.

앫 Verify secure connections on all capacitors.
앫 Clean the inverters with a vacuum. If the inverters are standalone types, turn off the
dc input at the circuit breaker or fuse panel before cleaning and verify no input or
output voltage is present; if the inverters are modular, remove them one at a time
and clean.
앫 Clean the inside of inverters with a vacuum cleaner being careful of areas that may
have voltage present.
앫 If necessary for calibration or testing, restore only input power.
앫 If the inverter system is equipped with a maintenance bypass panel, check the oper-
ation of all switch positions (typically, offline, online, maintenance, and test).
앫 Once the inverter system is properly adjusted and tested, restore inverters to full op-
eration as detailed in the manufacturer’s instruction manual and verify operation.
PART III BATTERY SYSTEM MAINTENANCE
This part describes the requirements for operating and maintaining vented lead–acid
(VLA) and valve-regulated lead–acid (VRLA) batteries used in telecommunications ap-
plications. The first section discusses common requirements that apply to both battery
technologies followed by individual sections for VLA and VRLA. Excellent sources of
information on battery problems and maintenance are [1] and [2].
6.17. GENERAL REQUIREMENTS
Three general statements can be made about battery maintenance:
앫 There is nothing a user can do to prevent a battery from eventually wearing out, but
a good battery maintenance program will increase the odds of attaining design life.
앫 Network operators can pay now or pay later for battery maintenance, but paying
now increases reliability and paying later decreases reliability.
앫 Just about any battery maintenance statement, rule of thumb, or procedure has ex-
ceptions. Always read, understand, and follow manufacturer’s recommendations.
6.17.1 Test Equipment

Only a few test sets are required for basic battery operation and maintenance—a non-
metallic flashlight, voltmeter, and an accurate thermometer and, for VLA batteries, a hy-
drometer. However, additional types of test equipment, such as a clamp-on dc ammeter,
8
When working on electrical equipment of any type, keep one hand in a pocket to prevent a harmful electrical
current from flowing from one hand through the upper body and heart to the other hand.
microohmmeter, cell internal resistance, impedance or conductance tester, and insulation

resistance tester, will greatly aid battery maintenance and increase the chances the net-
work operator will attain battery system design goals.
Measuring and adjusting battery voltage with an accurate 4½ digit (or 6000 count) dig-
ital voltmeter or multimeter is critical to long battery life. A multimeter is preferred be-
cause it can be used for other electrical testing. Accuracy should be at least ± 0.1% with a
resolution of 0.1 mV, and the voltmeter should be calibrated at least annually.
A high-quality thermometer is necessary to measure cell and ambient temperature.
Traditional glass–mercury or glass–alcohol thermometers are available from battery man-
ufacturers. Thermocouple or thermistor temperature sensors with digital readouts are suit-
able if they can be safely immersed in battery acid and do not contaminate the battery.
Perhaps the most useful thermometer type is an accurate noncontact IR thermometer or
camera. It can be used to measure not only cell and ambient temperatures but also the
temperatures of electrical connections and wiring; however, it is important that these de-
vices be calibrated at intervals recommended by the manufacturer. Although bulb-type
hydrometers may be used, electronic hydrometers also measure electrolyte temperature
and automatically correct the specific gravity reading to the reference temperature.
Increasingly important for battery maintenance is a test set that measures cell internal
impedance, resistance, or conductance, which is a good indicator of cell condition. It is
particularly useful in VRLA installations because the internal components of VRLA cells
cannot be visually examined to determine their condition. Cell testers are relatively ex-
pensive but will help locate faulty cells and therefore increase the overall reliability of
both VLA and VRLA installations. Also, they usually have other test functions that allow
measurement of the cell interconnection resistances and other low-resistance connections.
Cell testers variously measure internal impedance, resistance, or conductance. Exten-
sive published test data indicate that these parameters yield equivalent results [3, 4]. It has
been shown that impedance, resistance, or conductance measurements do not directly in-
dicate cell capacity but changes in measured values can be used to predict cell failure.
6.17.2 Records
An important aspect of battery operation and maintenance is keeping proper and updated
records (Fig. 6.32). Form 1 (Vented Lead–Acid) and Form 2 (Valve-Regulated
Lead–Acid) at the end of this chapter may be used for regular battery maintenance; Form
3 may be used for cell and interconnection resistance records. These forms may be modi-
fied to take into account different battery post and terminal lug designs.
6.17.3 Cleanliness
Battery cleaning kits are available from most battery system manufacturers and other
manufacturers specializing in these types of products. The kits include tubs, containers,
and squirt bottles to hold the cleaning and neutralizing solution, bristle brushes and towels
as well as personal protective gear.
Keep battery connections clean, bright, and corrosion-free and lightly coated with a
corrosion inhibitor. Any corrosion that does form must be cleaned off terminals; other-
wise it will spread into areas between posts and connectors and will develop into a high-
resistance connection causing heat and wasted capacity. Any corrosion that has blue,
green, mustard yellow, or maroon color indicates that copper is involved in the corro-
Fig. 6.32 Battery system records. The informal nature of the records in this photo does not detract
from their usefulness.
sion process. The copper is coming from somewhere (such as post or connector) and
the source must be found and corrective action taken. Keep the battery and surrounding
parts clean, dry, and free of acid. Sulfuric acid electrolyte absorbs moisture, and spilled
electrolyte does not dry up or evaporate so it must be neutralized before it is cleaned
up.
Neutralize spilled electrolyte with a solution consisting of 1 lb (0.1 kg) of bicarbonate
of soda (baking soda) to 1 gal (1 liter) of clean water, then rinse with distilled water and
dry with a soft lint-free cloth or towel (be sure to dispose of properly). Do not allow the
soda–water solution to get into the cells.
6.17.4 Safety Practices and Tools

Battery system testing must be performed or supervised only by qualified persons. Even
relatively small telecommunications batteries store large amounts of chemical and electri-
cal energy, and failure by personnel to be alert and knowledgeable will lead to injury and
equipment damage. Although special tools are not usually required, they must be handled
and used carefully:
앫 Use insulated or taped tools when working on batteries and electrical connections
(Fig. 6.33). Use a high-quality electrical tape; do not use cheap electrical tape like
that available at automotive parts stores or department stores.
앫 Always use a good quality micrometer-style, click-type torque wrench to tighten
all electrical fasteners; do not use cheap automotive torque wrenches with wing-
style torque indicators. If the torque wrench is not insulated, be sure to tape it be-
fore use.
앫 The torque values in most battery installations are in inch-pound units; however,
some newer battery installations may use metric fasteners in which case the torque-
wrench should be calibrated in metric units (newton-meters or kilogram-meters).
앫 See also Section 6.1.
Fig. 6.33 Insulated tools. (Photo courtesy of M.W. Migliaro.)
When working on battery systems and handling electrolyte, wear protective equipment
at all times, including
앫 Acid-resistant gloves
앫 Acid-resistant aprons
앫 Goggles or face shields (be sure the face or eye shields do not have bare metallic
frames or rims that can cause a battery short if dropped)
Avoid splashing the electrolyte. Electrolyte can seriously injure eyes and skin and dam-
age clothing and equipment and must always be handled carefully. The eyes in particular
should be guarded with face shields or protective glasses. If acid is splashed into the eyes
or anywhere on the skin, flood with water for at least 15 min. For electrolyte contact with
the eyes get medical attention. Do not use bicarbonate of soda solution on the skin or eyes
because it may aggravate the burn.
Provide a combination eye-wash, face, and body spray unit within 25 ft of the battery
room or battery system. These units can be permanently mounted and connected to the fa-
cility’s potable water system or can be a portable pressurized unit. Keep the area in front
of the washing station clear.
Remove all metallic jewelry (large metal belt buckles, rings, watches, bracelets, neck-
laces) when working around batteries and electrical circuits. Remove metallic pens and
other implements from shirt pockets. Use a nonmetallic flashlight for battery inspection,
and follow the “one-hand-in-pocket” rule (see Section 6.16.2).
Never lay any conductive object on top of a battery. Avoid work methods that could
cause battery circuit interruption, short circuits, or arcing in the battery vicinity.
Post “No Smoking” and “No Open Flame” signs where they are clearly vis-
ible to anyone entering the battery room area. Keep doors leading out of
the battery room clear at all times.
Mount a 10-lb, Class C fire extinguisher just inside the battery room door.9 Carbon diox-
ide (CO2) extinguishers are not recommended because the cold temperature can cause
thermal shock and crack cell containers resulting in a spill hazard, adding to the original
fire problem and leaving an even larger mess to clean up.
Vented lead–acid batteries generate a highly explosive mixture of hydrogen and oxy-
gen when gassing, and sparks, open flame, or lighted cigarettes never should be permitted
near any kind of battery. Although VRLA batteries are sealed and use recombinant tech-
niques, they can generate and release explosive gasses during normal and overcharge con-
ditions. Hydrogen concentration in the battery room or area must be kept below 2% by
volume at all times.
Battery rooms must be ventilated regardless of the battery technology used. In small
enclosed battery installations, vents or louvers that provide natural air circulation through
the battery room or enclosure are sufficient. Larger installations require forced air ventila-
tion (fans) that runs continuously or periodically.
Periodically measure airflow and total combustible gas in the battery areas to ensure
there is adequate air movement for diffusing hydrogen gas.
Battery cells and modules are very heavy and should be handled with a lifting device
such as a small crane, cherry picker, or elevator lift. Certain types of cells require special
lifting devices (Fig. 6.34).
6.17.5 Acid Electrolyte Spill Management

Spill management applies to both VLA and VRLA batteries. The electrolyte in VRLA
batteries is immobilized, but it still can leak out through faulty post seals and container
cracks and pinholes. However, the magnitude of acid spills from VRLA batteries seldom
is of the same magnitude as spills from VLA batteries.
All battery installations should have a spill management kit for safely neutralizing,
cleaning up, and disposing of spilled electrolyte and should include personal protective
gear (Fig. 6.35). Battery system manufacturers and other manufacturers can provide com-
plete kits. A typical kit consists of:
앫 Personal protective gear

앫 Acid-resistant clothing
앫 Acid-resistant gloves
앫 Acid-resistant boots or socks
앫 Face shield or goggles
앫 Equipment and material to neutralize and clean up electrolyte
앫 Booms and pillows
앫 Disposal bags and ties
앫 Acid-absorbent towels and pads
앫 Salvage drum
The absorbent material, disposal bags, and drum must be large enough to hold the
amount of spilled liquid, which in the worst case will equal the electrolyte volume
9
Class C fires involve electrical equipment, and the agent used in Class C extinguishers is not electrically con-
ductive. A circle containing the letter “C” identifies Class C extinguishers.
Fig. 6.34 Lifting device for round cells.
(quarts, gallons, or liters) in each cell, as determined from the battery data sheets, times
the number of cells. Generally, one kit is provided per battery string.
6.17.6 Electrical Connections

All electrical connections must be properly torqued, including battery terminals, cell in-
terconnecting straps, and bus and terminal fasteners. Some VRLA modules are marked
with the correct torque values (Fig. 6.36). Particular care must be taken to prevent over-
tightening fasteners at battery posts as these are made from relatively soft material that
can be crushed and deformed (Fig. 6.37). If the fasteners loosen because of temperature
cycling or slow compression (cold flow) of the lead terminal posts, the interconnecting
strap resistance will increase over time. A high resistance is apparent on discharge be-
cause the straps will get very hot.10
Regular examination of the cell interconnecting straps and hardware is essential. If
corrosion is apparent (usually around the positive post), the connections need to be disas-
sembled and cleaned. The corrosion may be severe enough to crack the hardware, partic-
10
In the early 1980s at a site in Alaska, the author had checked the torques on all cell interconnect straps on a
new 24-cell battery string installed by others. About 6 months later, there was a commercial power outage and
the batteries started to discharge. When the site technician walked into the darkened battery room, he was sur-
prised to see the cell interconnect straps glowing red. The connections had loosened in 6 months and the high re-
sistance produced enough heat to glow.
Fig. 6.35 Spill management kits. (Upper photo ( Whitham D. Reeve; lower photo courtesy M.W.
Migliaro.)
Fig. 6.36 Terminal torque label on VRLA module.

Fig. 6.37 Battery post deformation caused by fastener overtorque. Note that the battery post is de-
formed in the shape of an anvil (and is called “anviling”). (Photo courtesy M.W. Migliaro.)
ularly the fastener nuts, or the cell cover (Fig. 6.38). A coating of antioxidant compound
should be applied to the posts and straps after they are cleaned. A stiff, nonmetallic scrub-
brush should be used to clean battery components (a metallic brush, steel wool, and
emery cloth should not be used as they may leave incompatible material that will corrode
even further).
6.17.7 Cell Ground Faults

Over the life of a VLA or VRLA battery, cells can develop cracks in the container or at
the seam of the cover and container, or the containers may have pinholes from the manu-
facturing process. Also, containers can be cracked during shipping, handling, and installa-
tion. Any electrolyte that leaks out through the openings can provide a conductive fault
path to ground via the battery frame or directly to the concrete floor for large, freestand-
ing cells (Fig. 6.39). The resulting fault may cause the cell containers to slowly burn at
first, but the smoldering fire may easily escalate into a major fire.
Fig. 6.38 Cracked cell cover caused by corrosion in the post seal. (Photo courtesy M.W. Migliaro.)
Fig. 6.39 Battery internal faults. In this illustration, a simple 4-cell battery system with a negative
ground experiences a ground fault in cell 3, possibly due to a cracked case or other problem that
causes electrolyte to leak out and provide a current path to ground.
Since one terminal of all telecommunications battery systems is connected to ground, a

cell ground fault cannot be detected unless each individual cell is connected to a monitor-
ing system. Cell ground fault currents normally are not large enough to trip battery circuit
overcurrent protection devices (if equipped). Battery circuit overcurrent devices only pro-
tect against faults external to the battery system, whereas cell ground faults are internal to
the battery system. Battery strings that are connected through a one-pole disconnect or
overcurrent device may be opened manually by a push-button operated shunt-trip mecha-
nism if a fault is discovered, but the one-pole device will do nothing to interrupt the
ground fault path. The only way to manually interrupt the ground fault is to have a two-
pole disconnect or molded case circuit breaker with a shunt trip mechanism in the battery
circuit that disconnects both battery terminals, including the grounded terminal (Fig.
6.40).
6.17.8 Measurements
6.17.8.1 Voltage Measurements Battery and cell voltage readings should be tak-
en on a regular basis. Rectifier output voltage also must be measured on a regular basis
and whenever changes have been made to the ac electrical service or building wiring. For
example, if the building electrical service voltage is changed or if the electric utility
changes taps on its service transformers or voltage regulators, rectifier voltage may need
adjustment. In the case of service voltage change, the input transformer taps in older rec-
tifiers may need to be changed. Most modern rectifiers operate over a wide input voltage
range and do not have transformer taps.
Fig. 6.40 Two-pole battery disconnect. Either a manual disconnect switch or molded case circuit
breaker with shunt trip mechanism can be used to disconnect both battery terminals during mainte-
nance or to isolate a battery with an internal ground fault.
When measuring cell voltages, the probes should be placed directly on the cell posts or
terminals and not on the cell interconnect straps. When measuring battery terminal volt-
age, the probes should be placed directly on the terminal plates or lugs.
6.17.8.2 Temperature Measurements All cells of a VLA battery should be at

approximately the same temperature—a 3°C (5°F) spread from the coldest to the
warmest is acceptable. Heat sources such as sunlight and portable heaters must be
blocked or moved so they do not raise the temperature of individual cells. If the tem-
perature spread of the cells exceeds 3°C (5°F) (e.g., if the upper row of a two-tier VLA
battery rack is warmer), the room ventilation or air circulation may be inadequate.
Never intentionally allow electrolyte temperature to exceed 38°C (100°F). Record the
room ambient temperature before measuring cell temperatures.
Battery installations with several tiers, such as many VRLA installations, may exhibit
a wider temperature spread from bottom to top. In this case, the warmest tier (usually the
highest) should never exceed 28°C (82°F).
If an accurate IR thermometer is available, measure the temperature of the battery con-
nections (terminals and cell-interconnecting straps) when load current is flowing such as
during a discharge test or when the battery discharges during a commercial ac power out-
age. The temperature of loose or improperly torqued connections will be higher than the
other connections.
6.17.8.3 Connection Resistance Measurements Cell interconnecting straps

are made of lead-coated copper. Connections include the strap as well as the contact to
the cell terminal post and typically are designed for 20- to 30-mV drop at the anticipat-
ed range of discharge currents. The connection resistance can be calculated from Ohm’s
law, or
VDrop
RConnection = ohms (6.5)
IDischarge
Example 6.1 The 3-h discharge current for a 620-Ah battery is 170 A. If the intercon-
necting straps are designed for 20-mV drop, determine their resistance.
Solution Using Ohms’s law
VDrop 0.020 V
RConnection = = = 118 × 10–6 = 118
IDischarge 170 A
Connection resistance measurements can be dangerous; people making the measurements

must be alert, methodical, and cautious. Connection resistances are very small, measured
in microohms. To obtain meaningful measurements it is important to use consistent and
proper methods. Measurements are made from post to post of connected cells and from
the post to the terminal lug on the first and last cells. The microohmmeter scale should be
set to the lowest resistance scale and the probes should be held perpendicular to the termi-
nal or post. There are many manufacturers of instruments designed to measure very small
resistances, and some are specifically designed for battery applications.
NEVER place the probes across a cell or cells (across positive and negative posts or
terminals of a given cell) with the meter set to measure resistance. This may damage the
meter and the cell and may cause injury.
Cell post and terminal configurations vary. Figure 6.41 shows some common configu-
rations and the recommended measurement points. Proper and improper methods for con-
nection resistance measurements are shown in Figures 6.42 and 6.43. Some VRLA cells
have a plastic cover over the cell post that makes measurement difficult if not impossible
(Fig. 6.44). In this case it is tempting to measure from bolt head to bolt head, but the mea-
surement so obtained is useless because the measurement includes the high-resistance
stainless steel bolt and is not a true cell-to-cell measurement.
Measurements should start at the battery terminal connected to cell No. 1, which in
many telecommunications installations is the cell at the positive end, and move to the op-
posite end of the battery. Ensure the battery is on float charge before beginning the mea-
surements and not on equalize or being discharged. Record all measurements on Form 3
at the end of this chapter.
1. Take the first measurement on cell No. 1 between the terminal lug and the first post
[Figs. 6.41(a) to 6.41(d)]. This measurement will be about one-half of the cell inter-
connecting strap readings in steps 2 and 3 below.
2. Take the second measurement between opposite polarity posts of cells No. 1 and
No. 2 (do not put the probes on the interconnecting straps). This measurement will
include the resistance of two bolted connections (one on each post) and the cell-in-
terconnecting strap or straps [Figs. 6.41(e) to 6.41(h)] and will be about twice the
value of the first and last measurements at the battery terminals (steps 1 and 4).
(a)
(b)
(c)
(d)
Fig. 6.41 Battery cell post and terminal configurations and associated resistance measurements.
(a) Post–flag terminal—Each post has an offset plate (flag) mounted on it or part of it for fastening in-
terconnection straps or terminal conductors. Measure resistance from terminal lug to terminal post.
(b) Single terminal—Terminal conductors are fastened directly to each battery terminal. Measure re-
sistance from terminal lug to terminal post. (c) Single post plate terminal—Each battery terminal has a
plate for fastening terminal conductors. Measure resistance from each terminal lug to the cell post.
(d) Cable–plate–post—Each battery terminal has two parallel posts with one large plate mounted on
them for fastening terminal conductors. Measure resistance from each terminal lug to plate and from
plate to each cell post.
(e)
(f)
(g)
(h)
Fig. 6.41 Battery cell post and terminal configurations and associated resistance measurements. (e)
Single post—One or two parallel interconnection straps (two shown) may be equipped depending on
cell capacity. Measure resistance from the positive terminal post A of one cell to the negative terminal
post B of the adjacent cell. (f) Parallel post—Each cell polarity has two posts internally connected in
parallel. Measure resistance from the positive terminal post A of one cell to the negative terminal post
C of the adjacent cell and from positive terminal post B to the negative terminal post D of the adjacent
cell. The two resistances should be approximately the same. (g) Double Post—Each cell polarity has
two closely spaced posts internally connected in parallel, and each post has a set of fasteners for the
interconnection straps. Measure resistance from the positive terminal post A of one cell to the nega-
tive terminal post C of the adjacent cell and from positive terminal post B to the negative terminal post
D of the adjacent cell. The two resistances should be approximately the same. (h) Triple Post—Each
cell polarity has three closely spaced posts internally connected in parallel, and each post has a set of
fasteners for the interconnection straps. Measure resistance from the positive terminal post A of one
cell to the negative terminal post D of the adjacent cell, from positive terminal post B to the negative
terminal post E of the adjacent cell, and from positive terminal post C to the negative terminal post F
of the adjacent cell. The three resistances should be approximately the same.
(a)
(b)
Fig. 6.42 Connection resistance measurement methods. (a) Proper method—Probes should be
placed on the cell post and not on the interconnection strap. See also Figure 6.43. (b) Improper
method—Connecting the meter across the cell may result in cell or meter damage and injury.
Fig. 6.43 Correct test probe connection at cell post. Note that the probe is touching the lead metal
on the cell post. The second probe is visible behind the labeled test probe. (Photo courtesy of M.W.
Migliaro.)
3. Make subsequent measurements between positive and negative posts of adjacent

cells in a manner similar to step 2, moving from cell to cell.
4. Take the last measurement between the last post on the last cell and the battery ter-
minal lug fastened to it.
5. Analyze the measurements as follows: The positive and negative battery terminal
resistances should be no higher than approximately one-half of the cell interconnec-
tion strap resistances. The higher of the two terminal measurements should be no
Fig. 6.44 Plastic cover on VRLA cells. When the cell interconnect strap is in place, the lead metal
on the cell post is completely covered and not accessible for resistance measurements. A measure-
ment from bolt head to bolt head, if attempted, is useless. (Photo courtesy of M.W. Migliaro.)
more than 20% higher than the lower measurement. The resistance of any intercon-
nection strap should not be 20% higher than the lowest resistance, as in
RHigh 1.2RLow (6.6)
6. If a high resistance is found, measure from each post to its cell-interconnecting

strap to determine which of the two connections is bad. Mark all high-resistance
connections for repair (step 7).
7. Retorque problem connections to the manufacturer’s specifications. If the resis-
tance still does not meet the requirements of step 5, disassemble, clean, and apply
corrosion inhibitor compound and reassemble. Retest the repaired connection and
record the resistance in the “as-left” column on Form 3. Disassembly of connec-
tions on an active battery string must be done with caution and planning so as to not
jeopardize network equipment operation.
The resistances of the interconnection straps and battery terminals should be on the order
of 100 or less; actual values will depend on the amount of contact area and size of the
inter-connection straps (smaller battery installations will have higher resistances than
larger installations). Measurements taken upon completion of the original installation will
be used as a baseline for comparison to later measurements. See Section 6.18.6 (VLA)
and 6.19.5 (VRLA).
6.17.8.4 Cell Internal Impedance Measurements Internal impedance, resis-

tance, or conductance can indicate cell internal condition and applies to both VLA and
VRLA batteries. Measuring the internal impedance of a VRLA cell allows the user to mon-
itor two important failure modes, which are grid corrosion and dry out. Battery manufactur-
er’s literature should list the normal expected values. If a manufacturer is reluctant to pro-
vide this information or says it is not available, batteries should not be purchased from them.
Internal impedance, resistance, or conductance measurements should not be made until
after the cell interconnection resistances are checked (and repaired if necessary). Depend-
ing on how the measurements are made, a high connection resistance may falsely indicate
a high internal cell resistance.
There is some disagreement in the industry as to exact values and tolerances of cell in-
ternal measurements, but Table 6.16 may be used to establish a reference program for
tracking trends. A cell reference value (CRV) is determined and then used as a benchmark
for future measurements. It is important that written records be kept starting on the instal-
lation date and periodically thereafter. Initial values are as important as trends throughout
the battery system life.
Although it is possible to measure cell internal impedance while the battery is being dis-
charged (by dividing the cell voltage by the discharge current), this type of test may not pro-
vide repeatable results. Commercial test sets are available to measure internal impedance,
resistance, or conductance and provide repeatable results during normal float operation.
6.17.9 Monthly, Quarterly, and Yearly Inspections

Inspections are required to determine the general condition of safety equipment and are in
addition to specific inspections and measurements of power system components (batteries
and rectifiers) listed in respective sections that follow.
6.18 VENTED LEAD–ACID BATTERIES 449
Table 6.16 Cell Internal Impedance, Resistance, or Conductance Measurements

Description Conductance Resistance or Impedance
Cell reference value (CRV) Average of 10 highest (24-cell Average of 10 lowest (24-cell
battery) or 5 highest (12-cell battery) or 5 lowest (12-cell
battery) cells battery) cells
For a battery string that is new Measurements of all cells Measurements of all cells
or only a couple years old should fall within ±5% of should fall within ±5% of
the CRV the CRV
For an older, fully functional Measurements of all cells Measurements of all cells
battery string should fall within ±10% of should fall within ±10% of
the CRV the CRV
Battery string is in good All cells are above 80% of All cells are below 125% of
condition if the CRV the CRV
Additional testing may be The cell is 60–80% of the The cell is 125–167% of the
required on a suspect cell if CRVa CRVb
Cell is unserviceable and The cell is 60% of the CRV The cell is 167% of the CRV
requires replacement if
a
This range is for guidance. It is possible that a cell measuring < 70% of the average conductance has < 80% rat-
ed capacity and should be replaced.
b
This range is for guidance. It is possible that a cell measuring > 130% of the average resistance or impedance
has < 80% rated capacity and should be replaced.
앫 Check for availability and condition of all safety equipment, including gloves,
aprons, and face shields.
앫 Check for a full gallon (4 liters) of labeled acid-neutralizing solution.
앫 Check operation of the eyewash station or portable eyewash equipment.
앫 Check operation and cleanliness of the body wash station.
앫 Check that a Class C fire extinguisher is available and that it has been inspected and
tested according to schedule.
앫 Check that insulated tools and utensils are available.
앫 Check the hydrometer for cleanliness and cracked rubber parts (VLA installations
only).
6.18 VENTED LEAD–ACID BATTERIES
6.18.1 General
This section outlines the requirements for operating and maintaining vented lead–acid
(VLA) batteries. The battery manufacturer and industry standards are the best places to
get detailed operation and maintenance information; however, this section provides suffi-
ciently detailed information for most VLA batteries.11
6.18.2 Specific Gravity Measurements

6.18.2.1 Instruments Three types of hydrometers are available for measuring the
electrolyte specific gravity in VLA cells—float type, syringe type, and electronic display
11
See also IEEE Std 450 [5].
(Fig. 6.45). The float type can be installed in the sampling port in one corner of the cell
cover. The syringe type has a rubble nozzle, weighted float, glass tube, and a rubber bulb
at the top. Both hydrometers have a weighted float with divisions marked off on the upper
portion. The float sinks into the electrolyte far enough to displace a volume of electrolyte
equal in weight to the weight of the float itself. The higher the specific gravity, the heav-
ier the electrolyte and the less the float sinks. The marks on the float are graduated in in-
crements of 1 point, or 0.001, over a specific gravity range of 1.100 to 1.300.
At sites with a large number of cells, it is desirable to have two high-accuracy hydrom-
eters and to frequently check them against each other; however, most sites have only one
hydrometer. Bulb-type hydrometers should be replaced every 2 or 3 years.
Specific gravity in a cell can vary with depth and is known as stratification. Stratifica-
tion occurs when the sulfuric acid that is formed by high initial charging currents settles
to the bottom because it is heavier than water. Gassing upon full charge tends to mix the
electrolyte and reduce stratification.
The electrolyte level in a cell is lowered by evaporation and water electrolysis. The
sulfuric acid in the electrolyte does not evaporate, so the specific gravity increases as the
level decreases. For small VLA cells, a small difference in level can make a several-point
difference in specific gravity. To avoid dry-out, the electrolyte level should never be al-
lowed to decrease below the low-level mark on the cell. Electrolyte should not be added
to a cell unless directed to by the manufacturer.
The loss of water from the electrolyte due to evaporation and charging affects the spe-
cific gravity. In a fully charged cell, if the electrolyte level is 1–4 in. (6 mm) below the high-
level mark, the specific gravity will be approximately 6 points (0.006) higher than with
the electrolyte at the high mark (the lower electrolyte level has a higher concentration of
acid because it is less diluted with water). Therefore, when measuring specific gravity, the
electrolyte level with respect to the high mark should be noted so that the specific gravity
can be properly evaluated. Battery manufacturers can provide a more accurate correction
for each battery type.
The electrolyte level is affected by the charging rate and amount of gassing. Gassing
electrolyte has a slightly higher volume and will show a higher level. If the electrolyte
Fig. 6.45 Electronic display specific gravity measurement. (Photo courtesy of M.W. Migliaro.)
level is at the high-level mark during float, it will rise above that mark when the battery is
being equalized. Such a rise is not a problem unless the level is so high that the electrolyte
overflows.
When measuring VLA electrolyte specific gravity, the electrolyte level should be at
the high-level mark on the cell. If water needs to be added, specific gravity measurements
should not be made for at least 72 h to allow the water to mix with the electrolyte (some
cell types require longer mixing times). Tap water should not be used to adjust electrolyte
level as it contains impurities that could damage the plates.
Specific gravity measurements require very little practice to perform correctly. Cells
used in telecommunications applications usually are marked with the rated specific gravi-
ty and temperature. The main problem will be accidentally spilling or dripping electrolyte
from the hydrometer nozzle. Any spilled electrolyte must be neutralized and cleaned up.
The specific gravity changes 1 point inversely with each 1.67°C (3°F) change in tem-
perature. The reference temperature for specific gravity measurements is 25°C (77°F), so
all measurements must be corrected to this temperature. Every VLA battery room should
be equipped with a thermometer for measuring room ambient temperature and another for
measuring cell electrolyte temperature. Two types of cell thermometers are available for
measuring electrolyte temperature—floating and nonfloating. The floating type can be in-
stalled in the sampling port in one corner of the cell cover. The nonfloating type is in-
stalled in the cell vent. The temperature of a cell rises very little during discharge but ris-
es rapidly during recharge.
The battery room ambient temperature ideally is 25°C (77°F) but a range of 16 to 27°C
(60 to 80°F) is acceptable. When the ambient temperature is consistently above 32°C
(90°F), due to inadequate air conditioning and it is not possible to install better air condi-
tioning equipment, the battery manufacturer may recommend the specific gravity be re-
duced.
To minimize measurement errors, use a long nozzle syringe and take samples about
one-third down from the top of the plates if possible. Some cells have an electrolyte with-
drawal tube for this purpose. Bubbles in the hydrometer electrolyte cause errors, and read-
ings should not be taken sooner than 15 min after gassing has stopped.
6.18.2.2 Specific Gravity Measurement Procedures

1. Clean the hydrometer glass barrel and float with soap and water if necessary. Rinse
thoroughly with clean distilled or demineralized water and dry so the hydrometer
components do not contaminate the electrolyte. Because of possible contamination,
never use a hydrometer on lead–acid cells that has been used on nickel–cadmium
(NiCd) cells.
2. Hold the hydrometer vertically and squeeze the rubber bulb before immersing it in
the electrolyte. Insert the nozzle into the electrolyte withdrawal tube or opening and
release the bulb. Draw enough electrolyte into the hydrometer barrel so the float
freely floats without touching the sides or top of the syringe. To avoid spilling the
electrolyte, do not remove the nozzle from the electrolyte withdrawal tube while
reading it.
3. Read the specific gravity on the hydrometer scale at the flat surface of the elec-
trolyte (Fig. 6.46).
4. Slowly squeeze the rubber bulb to release the electrolyte back into the cell. Be sure
to completely empty the barrel; otherwise, the electrolyte from one cell may affect
the measurement of the electrolyte from another cell.
Fig. 6.46 Hydrometer: Read the specific gravity at the flat surface of the electrolyte, not on the
raised edges at the float or inside of tube.
5. If the hydrometer is a different temperature than the cell electrolyte, repeat steps 2
to 4 several times until the float reaches the same temperature as the electrolyte.
6. Correct the specific gravity measurements to 25°C (77°F) before recording (Table
6.17). When temperature readings are in Celsius, add 0.6 points (0.0006) to the
reading for every 1°C the pilot cell temperature is above 25°C, and subtract 0.6
points from the reading for every 1°C the pilot cell temperature is below 25°C.
When temperature readings are in Fahrenheit, add one point (0.001) to the reading
for every 3°F the pilot cell temperature is above 77°F, and subtract one point from
the reading for every 3°F below 77°F. If an electronic hydrometer is used, the in-
strument may automatically correct the specific gravity measurement for tempera-
ture, but this must be confirmed before recording the measurement.
Table 6.17 Specific Gravity Temperature Correction Factors

Electrolyte Electrolyte
Temperature Correction Temperature Correction
32.2°C (90°F) +0.004 23.3°C (74°F) –0.001
31.7°C (89°F) +0.004 22.8°C (73°F) –0.001
31.1°C (88°F) +0.003 22.2°C (72°F) –0.001
30.6°C (87°F) +0.003 21.7°C (71°F) –0.002
30.0°C (86°F) +0.003 21.1°C (70°F) –0.002
29.4°C (85°F) +0.002 20.6°C (69°F) –0.002
28.9°C (84°F) +0.002 20.0°C (68°F) –0.003
28.3°C (83°F) +0.002 19.4°C (67°F) –0.003
27.8°C (82°F) +0.001 18.9°C (66°F) –0.003
27.2°C (81°F) +0.001 18.3°C (65°F) –0.004
26.7°C (80°F) +0.001 17.8°C (64°F) –0.004
26.1°C (79°F) 0.000 17.2°C (63°F) –0.004
25.6°C (78°F) 0.000 16.7°C (62°F) –0.005
25.0°C (77°F) 0.000 16.1°C (61°F) –0.005
24.4°C (76°F) 0.000 15.6°C (60°F) –0.005
23.9°C (75°F) 0.000 23.3°C (74°F) –0.006
7. Record all readings (Form 1 at the end of this chapter) and keep them for the life of
the battery. The spread in specific gravity across all cells in a battery string (the dif-
ference between the highest and the lowest) normally should not exceed 0.020 (the
tolerance on nominal specific gravity is ±0.010) with the electrolyte level at the
high mark.
8. Rinse the hydrometer with clean distilled or demineralized water and return it to its
holder.
Example 6.2 Determine the temperature-corrected specific gravity if the measured spe-
cific gravity is 1.207 at 21°C (70°F).
Solution From Table 6.17, the correction factor is –0.002 so the corrected specific grav-
ity is 1.207 – 0.002 = 1.205.
Example 6.3 Determine the temperature-corrected specific gravity if the measured spe-
cific gravity is 1.212 at 31°C (87°F).
Solution From Table 6.17, the correction factor is +0.003 so the corrected specific grav-
ity is 1.212 + 0.003 = 1.215.
6.18.3 Periodic Measurements and Inspections

Some measurements, such as cell temperature, are based on a pilot cell. The pilot cell
serves as a proxy or representative for the other cells in the battery and its use saves time
measuring battery systems with large number of cells; each battery string will have its
own pilot cell (Fig. 6.47). The pilot cell is chosen from one of the cells with the lowest
specific gravity and lowest voltage, a new pilot cell is chosen at least once a year.
6.18.3.1 Initial Installation Initial installation measurements can be made coinci-

Fig. 6.47 Battery string pilot cell. Note the engraved plastic tag slipped over the vent in the middle
of the picture. When a new pilot cell is chosen, the tag is moved to the new cell. (Photo courtesy of
M.W. Migliaro.)
dentally with the acceptance tests (Section 6.18.6) and should be performed no sooner
than one week after the battery has had its initial freshening charge. A complete set of
measurements should be taken to establish a baseline for a new installation including:
앫 Ambient temperature.
앫 Internal impedance, resistance, or conductance of each cell. Note: Some users con-
sider this an optional test unless the cells have factory-made connections that are in-
accessible for maintenance, such as a three-cell monobloc with internal lead-burned
connections between the cells.
앫 Specific gravity and temperature of each cell.
앫 Voltage of each cell and voltage across the battery terminals.
앫 Cell interconnecting strap resistances and battery terminal resistances.
앫 Battery float current.
6.18.3.2 Monthly Measurements

앫 Check the ambient temperature and ensure ventilation devices (fans and vents) are
operable.
앫 Measure and record pilot cell specific gravity, voltage, and temperature.
앫 Examine electrolyte level in all cells and add water if necessary.
앫 Measure and record float voltage at battery terminals.
앫 Examine the overall battery, battery rack, and battery room or area for appearance
and cleanliness. Remove combustible materials such as empty boxes from the room.
앫 Measure and record rectifier output voltage and current.
앫 Examine cell containers and covers for cracks, distortion, and electrolyte leaks.
앫 Examine cell and battery terminals and connections for corrosion.
앫 Examine racks and fasteners for corrosion and overall structural integrity (shake
test).
앫 Measure and record float charging current.

앫 Check battery monitoring system operation, if equipped.
6.18.3.3 Quarterly Measurements

앫 All monthly work items.
앫 Measure and record the specific gravity and temperature of 10% of the total number
of cells; rotate these cells so that readings are taken on different cells each quarter.
앫 Measure and record voltage of each cell.
6.18.3.4 Yearly Measurements

앫 All quarterly work items.
앫 Measure and record the voltage, specific gravity, and temperature of every cell.
앫 If equipped with clear containers, examine every cell for internal condition, plate
and separator deterioration and distortion, excessive sediment at the bottom of cells,
proper plate color (positive plates normally are a deep chocolate brown to black
color and negative plates are a medium gray color), post distortion or lifting, and
any other visual attributes considered out of the ordinary. Note: Some plate growth
is normal over time but it normally should not exceed 5%.
앫 Examine cell interconnections and terminal connections to ensure they are clean
and are coated with a thin layer of corrosion inhibiter.
앫 Verify that all cells are properly numbered (in telecommunications applications, the
cell closest to the positive end of the battery normally is No. 1).
앫 Verify that all cells are equipped with flame arrester vents and plastic dust caps. Ex-
amine vents and clean or replace if clogged (see Section 6.18.7).
앫 For seismic installations, examine the battery to ensure a spacer is present between
each cell and that the spacer is in good condition and not broken, distorted, or
cracked.
앫 Measure and record cell interconnection and terminal resistances and compare val-
ues to those obtained on initial installation. If any connection resistance has in-
creased more than 20%, disassemble, clean, apply corrosion inhibiter, reassemble,
retorque, and remeasure.
앫 Measure and record the internal impedance, resistance, or conductance of each cell
and compare to initial and previous measurements (some users consider this optional).
Table 6.18 Typical Float and Equalize Voltages at 25°C (77°F)—Vented Lead–Acid Batteries
Cell Voltage 24-V, 12-Cell 48-V, 24-Cell
Type (V/cell) System (V) System (V)
Float
Equalize
6.18.4 Battery Charging and Measurement Procedures

6.18.4.1 Initial Freshening Charge New batteries and batteries that have been
stored for 3 months or more should be given an initial freshening charge. Unless the man-
ufacturer specifies a different voltage, normal equalizing voltage should be used (Table
6.18). If the battery is connected to a power system, care must be taken to ensure the
charge voltage used does not exceed the maximum rated voltage of the load equipment.
The initial charge voltage should be applied until each cell gasses freely and equally and
the specific gravity and cell voltage stops rising. Gassing can be determined easily by ex-
amining cells that have transparent or clear translucent containers but is not possible with
opaque cell containers (such as hard rubber). In the latter case, specific gravity, cell volt-
age, or charge current measurements are used to determine the end of charge condition.
Just before the end of the initial charge, record the cell voltages and then change to the
normal float voltage. Allow the cells to stabilize for 72 h, measure the cell specific gravi-
ties, correct them to 25°C (77°F), and record the corrected value.
6.18.4.2 Float Charge Batteries are continuously float charged at a constant volt-
age during normal operation. Use Table 6.18 to determine the appropriate float voltage
unless the manufacturer specifies a different value. Measure the voltage across the battery
terminals with an accurate digital voltmeter or multimeter and compare the reading with
the rectifier and powerboard bus voltmeters. If necessary, adjust the rectifier output volt-
age to provide the required float voltage at the battery terminals, and adjust the rectifier
and powerboard voltmeters (if adjustable) to agree with the digital voltmeter. Do not rely
on the rectifier voltmeter to set the battery voltage because the voltmeters on rectifiers
usually do not have the required accuracy and resolution; the voltage at the battery termi-
nals is the important parameter and not the voltage at the rectifier.
6.18.4.3 Equalizing Charge Equalizing charges should not be performed on a rou-

tine basis; however, if one of the conditions below is met, apply an equalizing charge at
the recommended voltage (Table 6.18 unless recommended otherwise by the manufactur-
er). Failure to equalize cells when needed can lead to problems such as overdischarge and
internal physical damage or cell polarity reversal when the battery is discharged. Ensure
that the electrolyte level in all cells is near the high-level mark before beginning the
equalizing charge. Lead–calcium batteries normally do not require periodic equalizing if
floated between 2.20 and 2.25 V per cell.
Conditions for equalizing charge:
앫 Following a discharge that reduces the battery capacity by 10% or more.

앫 If the temperature-corrected specific gravity of any cell is more than 10 points
(0.010) below the nominal full charge value when the battery is on float.
앫 If the voltage of any cell is more than 0.05 V (lead–calcium) or 0.03 V (lead–anti-
mony) below the average cell voltage when the battery is on float. The average cell
voltage is found by dividing the float voltage measured at the battery terminals by
the number of cells in the battery string.
앫 If the electrolyte level in any cell falls to or below the minimum fill line, distilled or
demineralized water must be added to restore the level to the maximum fill line, and
an equalizing (mixing) charge can be performed to restore specific gravity. A mixing
charge is important where low temperatures may cause the water in the cells to freeze.
Fig. 6.48 Typical water consumption at 25°C (77°F). To convert from pints per cell per month to
liters per cell per month, multiply by 0.473.
앫 If a cell is undercharging, it will not consume water at the same rate as normal cells.
In this case, an equalize charge should restore it to normal. Figure 6.48 shows typi-
cal water consumption.
Terminate the equalize charge when all conditions below are met.
앫 Every cell gasses freely and equally.

앫 The specific gravity of all low cells has stopped rising as determined by two consec-
utive specific gravity measurements over the last one-eighth of the charging period.
앫 The voltage difference between the highest and lowest cells is no greater than
recorded for the initial charge.
The equalizing charge may be started and stopped by an automatic equalize timer if the
powerboard is so equipped. Use the battery manufacturer’s recommended charge time for
the cell type or, if that information is not available, use Table 6.19. Take care to ensure
the equalizing voltage does not exceed the maximum equipment voltage.
When an equalizing charge is stopped and float voltage is reapplied to a battery, heavy
gassing will end shortly thereafter. About 15 or 20 min afterwards measure and record the
specific gravity of every cell. If two cells with the lowest specific gravity (checked over
the last one-eighth of the charging period) have not stopped rising, continue the equaliz-
ing charge.
Table 6.19 Typical Equalize Charge Duration for Various Charge Voltages
Cell Voltage 24-V, 12-Cell 48-V, 24-Cell Time
Type (V/cell) System (V) System (V) (h)a
VLA (lead–antimony)—1.215 SG 2.24 26.88 53.76 80
2.27 27.24 54.48 60
2.30 27.60 55.20 48
2.33 27.96 55.92 36
2.36 28.32 56.64 30
2.39 28.68 57.36 24
VLA (lead–calcium)—1.215 SG 2.24 26.88 53.76 222
2.27 27.24 54.48 166
2.30 27.60 55.20 105
2.33 27.96 55.92 74
2.36 28.32 56.64 50
2.39 28.68 57.36 34
a
If the battery temperature is in the range 4–16°C (40–60°F), the duration should be doubled and if below 4°C
(40°F), the duration should be quadrupled.
If some cells are worn out or otherwise faulty, an equalize charge may not bring them
to the rated specific gravity, and it may be incorrectly assumed that additional electrolyte
is needed to raise the specific gravity in that cell. However, battery manufacturers gener-
ally do not recommend that electrolyte be added to in-service cells. If in doubt about the
cell’s condition, contact the manufacturer for guidance.
6.18.4.4 Check Charge A check charge may be used at unattended sites to deter-
mine if a full equalizing charge is necessary. With a check charge, the battery is manually
placed on equalize charge. After 15 or 20 min, the battery voltage should be stable at
which point the individual cell voltages are measured. If the highest and lowest cell volt-
ages (while on equalize) differ by no more than 0.04 V, the battery does not require equal-
izing. The battery should be equalized if the voltage differences exceed 0.04 V. Measure-
ments should be made with a digital voltmeter accurate to at least 0.01 V. Alternately, the
charge current can be measured at 2-h intervals to see if it has stabilized.
The check charge is effective in determining the equality of cell charge. However, the
test should not be substituted for specific gravity or float current measurements that deter-
mine the state of cell charge. All cells must be within 3°C (5°F) and the ambient tempera-
ture must be above 13°C (55°F).
6.18.5 Troubleshooting Battery Problems

6.18.5.1 General This section describes VLA troubleshooting principles and how to
determine if the care is adequate and correct. Correct charging is critical for long battery
life and reliable service, and the most important part of battery care is a proper charging
program. Accurate records of physical condition and electrical measurements are required
to determine if the charging program is correct.
The following discussion describes cell conditions with both proper and improper op-
eration and maintenance. Manufacturers should be consulted to confirm diagnoses before
action is taken to replace a battery or cell. Table 6.20 summarizes various problems that
Table 6.20 Battery Problems and Possible Causes

Problem Possible Cause
Gassing
앫 Lack of gassing while on charge 앫 Internal short between plates causing the cell
to discharge internally as fast as it is being
charged
Specific Gravity or Voltage
앫 Lower specific gravity or voltage than 앫 Excessive internal losses resulting from
other cells consistent undercharging
앫 Loss of capacity over time is shown by a
gradual decrease in specific gravity of the cells
Color
앫 Patches of white lead-sulfate either on the 앫 Cells standing idle or undercharging for
positive or negative plates extended periods
앫 Antimony deposits (dark-slate patches on 앫Charging at too high a rate or an aged cell
the negative plates, usually near the terminal) nearing the end of its service life
앫 White top layer of sediment 앫 Undercharging
앫 Lumpy brown sediment 앫 Overcharging, cycling, or high temperature
앫 All white sediment with no visible layers 앫 Overcharging after prolonged low float
voltage
앫 Large flaking on the interplate collector bar 앫 Battery on float charge for extended periods
at insufficient float voltage and without
equalizing charge
Plate Problems
앫 Cracks on the edges of the positive plate grids 앫 Cell at end of life
앫 Light-colored sulfating spots on edges of plates 앫 Cell at end of life
below the cracks mentioned above
앫 Excessive sediment in the bottom of the case 앫 Cell at end of life
앫 “Mossing” on the tops of negative plates 앫 Cell at end of life
앫 Buckling of positive plates 앫 Excessive sulfating caused by undercharging
or excessive temperature
Water Problems
앫 Excessive water consumption 앫 Excessive charging rates, high operating
temperatures, or a leaking cell
앫 Unusually low water consumption 앫 Insufficient charging
Capacity Problems
앫 Failure to supply rated ampere-hours 앫 Discharged condition, excessive sulfating, or
loss of active material from positive plates
앫 Cells may be worn out or active material
may be gone from positive plates
may be encountered in the field. Additional detailed descriptions are provided in the fol-
lowing sections.
6.18.5.2 Surface Charge Phenomenon When a battery has been on float charge
for a long time and then is discharged under load, the voltage initially drops rapidly (so-
called coupe de fouet) because plugged pores on the surface of the plates partially block
ion transfer. The voltage may drop below the low-voltage alarm threshold and in severe
cases may drop below the low-voltage disconnect (LVD) setting, if equipped. If the LVD
does not trip, the battery voltage usually will increase to above the low-voltage alarm
threshold. The battery will then operate normally until its capacity is exhausted.
If the battery is exercised (partially discharged) on a routine basis, the voltage dip can
be reduced or eliminated. Turning off the rectifiers for at least 15 min and allowing the
battery to discharge into the load exercises the battery. The first few times this procedure
is performed, disable the LVD to prevent an inadvertent trip and be sure to reenable the
LVD after the test. The first time the battery is exercised, the procedure should be per-
formed several times in succession until the voltage stays above the alarm setting. Always
allow the battery to fully recharge (the charging current reduces to the normal float value)
before turning off the rectifiers again for the next cycle. A battery should not be exercised
too often because even 15-min discharge will affect its life.
Each battery has its own characteristics, and the exercise frequency should be adjusted
so the voltage drop does not cause the low-voltage alarm. Start at a monthly cycle and ex-
periment with increasing the time between exercises. The proper time between exercises
exists when the voltage drop is just above the alarm threshold. Some alarm systems may
be equipped with a time delay and may be set to ignore a low-voltage alarm for a preset
time period giving the battery time to recover. The time delay will have to be determined
experimentally for each installation.
6.18.5.3 Initial Troubleshooting If any cells seem to be in trouble, give the whole
battery an equalizing charge, and then measure the specific gravity of all cells. If all cells
gas evenly and the specific gravity of every cell is normal, all the battery needed was the
equalizing charge. Otherwise, record all low specific gravities and give an extra thorough
equalizing charge.
Measure and compare the temperature of all cells. Excessively sulfated cells can run
hot enough to cause damage if not corrected. Investigate for impurities and measure for
internal short circuits any cells that still do not gas with the extra equalize charging.
6.18.5.4 Cell Replacement If a faulty cell is to be replaced, replace it with one

in good condition and of the same make, type, rating, and approximate age (the latter is
difficult with an older battery because spare cells of the same age seldom, if ever, are
available). Avoid installing a new cell in series with older cells unless no other alterna-
tive is available. A 12- or 24-cell battery may be operated temporarily without one cell
by removing the cell and installing temporary jumpers between the adjacent cells. The
float and equalizing voltages must be reduced to that required by an 11- or 23-cell bat-
tery. If the installation uses parallel battery strings, operating with one cell removed
from only one string cannot be done; instead, a cell must be removed from all battery
strings.
The capacity of the battery with one cell removed will be reduced due to the lower dis-
charge voltage range. For example, say load equipment stops operating when the battery
terminal voltage reaches 42.0 V (the voltage at the load equipment will be somewhat low-
er due to voltage drop from the battery to the load). For a 24-cell battery, this corresponds
to a final cell voltage of 1.75 V/cell, or the voltage at which a telecommunications battery
is considered fully discharged. If one cell is temporarily removed, the 23-cell battery will
be fully discharged when its voltage reaches 40.25 V (23 cells × 1.75 V/cell). However,
the equipment will have stopped operating at 42.0 V (1.83 V/cell) so the 23-cell battery
never fully discharges to 1.75 V/cell. In this case, increasing the end-of-discharge (final)
cell voltage from 1.75 to 1.83 V/cell decreases the overall capacity by approximately 10
to 15% (this reduction corresponds to the recommended design margin used to size the
battery; see Chapter 5, System Design).
6.18.5.5 Charging It is important to know when a cell is fully charged. A cell is ful-
ly charged when, during equalize, the cell is gassing, specific gravity has stopped rising,
and specific gravity, corrected for temperature, remains constant for two successive read-
ings. Hydrometer readings must be corrected for any changes in cell temperature that
have occurred between readings. The two successive readings should be taken during the
last one-eighth of the charging period (usually 1, 2, or 3 h apart).
Proper Charging If cells are undercharged, service will be poor and battery life short.
If overcharged, service with will be good initially but battery life short. Proper charging
means slight overcharging just enough to cause the least possible sedimentation and a
minimum of heavy gassing. This condition requires very little makeup water in lead–calci-
um VLA cells. No perceptible sedimentation or plate buckling occurs if the charging rate
is such that cells are not allowed to gas vigorously. Sedimentation starts with gassing and
is proportional to the total amount of gas liberated.
Appearance of Normal Cells The edges of normal positive plates will not show any
sulfation, cracks, or plate growth. The edges of negative plates should be uniformly gray.
When a normal cell with correct float charging is examined with a flashlight, there should
be no sparkling from lead sulfate crystals.
No visible change occurs when the cell is discharged a normal amount (i.e., not
overdischarged). If the charging program is correct, sediment accumulates very slowly. It
should never be white or lumpy as shown in Figure 6.49. The charging program may pro-
duce a very small amount of fine, dark-brown sediment.
Chemical Changes A fully charged cell has brown lead dioxide on the positive plates
and gray sponge lead on the negative plates. On discharge, electric current converts active
Fig. 6.49 Sediment at the bottom of the cell due to improper charging. (Photo courtesy of M.W.
Migliaro.)
materials in the positive and negative plates to normal lead sulfate. This process leaves
the electrolyte weak at the end of the discharge. Lead sulfate is white in color but cannot
be seen on plates unless the cell is overdischarged, which produces oversulfating. This
condition first makes the plate color lighter and finally mottled white in patches or white
all over.
Charging the cell reverses this process, converting lead sulfate in the plates to lead
dioxide and sponge lead and producing sulfuric acid, which restores the specific gravity
to normal. As the charge nears completion, only a little lead sulfate remains to be convert-
ed to lead. The charging current begins to separate water into oxygen and hydrogen (elec-
trolysis), which bubbles to the top of the electrolyte and forms a mixture of very explosive
gases.
A detailed description of the chemical reactions is given in Chapter 4 (Telecommuni-
cations Batteries).
Effect of Impurities on Float Voltage Compared to lead–antimony, lead–calcium

cells have the advantage of low internal losses, which remain constant for the life of the
cell. Fully charged lead–antimony cells discharge internally by an action between the ac-
tive material and the grid. Impurities may hasten this action and may result in visible or
invisible changes on the plates, depending on the types of impurities present.
No metal objects should be put into the cell electrolyte at any time, except a cadmium or
other compatible reference electrode. Impurities may prevent a proper float voltage from
maintaining float charge and may prevent an equalizing charge from equalizing the cell.
High-Rate Overcharging After a battery is fully charged, continuation of charging

current at a high rate damages the positive plates. Violent gassing forms bubbles in the in-
terior of the active material. The resulting pressure forces gasses through the porous ac-
tive material breaking off particles. These particles rise with the bubbles and may result in
a muddy red or brown electrolyte color (although it is rare to see this). Some of this fine
sediment settles on and, if allowed to accumulate, short circuits the negative plates. The
sediment is converted to gray sponge lead and results in a growth of mosslike sediment on
the top edges of the negative plates. The “mossing” indicates that a high-rate overcharg-
ing previously occurred and can cause the cell to overheat during charging.
Low-Rate Overcharging At lower overcharge rates, bubbling is reduced and sediment

falls to the bottom of the cell. Overcharging at a very slow rate disturbs electrolyte so lit-
tle that fine brown sediment falls in a vertical line, forming tiny ridges on top of the sedi-
ment. Ridged sediment is a good indication that the recent overcharging was not at high
rates. Overcharging should be kept at a minimum, and ridges should be small.
Undercharging If the battery is undercharged, unconverted sulfate remains on the

plates too long and hardens. The longer the plates stay in the undercharged condition, the
harder the sulfate becomes and the more difficult it is to reconvert. When new, the sulfate
is easily converted back to soft active materials by a normal charge, but a long overcharge
is required to remove it after becoming hard. Sulfate accumulates unnoticed, a little on
each charge, if charging is not enough to eliminate all sulfate. This residue buildup con-
tinues until a substantial portion of the cell capacity is lost. The remedy is to slightly in-
crease float voltage to give a slight overcharge. This procedure must be put into practice
while the battery is new and followed for the life of the battery. Prolonged undercharging
also leads to large flaking on the internal interplate connection strap.
The pressure of expanding active material due to sulfate buildup can break plate sepa-
rators and cause shorts inside the cell. If charged at too low a rate, the hardened sulfate is
thrown out of plates and settles in white ridges on the cell bottom. At higher rates, the
gassing distributes the sediment evenly without ridges. An oversulfated cell has high in-
ternal resistance and requires extra voltage across the cell, which also causes it to develop
higher temperatures on charge. Buckled or cracked plates cannot be repaired by sulfate
removal, but the cell may be used as long as it has satisfactory capacity.
6.18.5.6 Overdischarge The plates suffer greatly when overdischarged. The cell
voltage should not be allowed to drop below 1.75 V. Specific gravity should not be al-
lowed to decrease below the limit given by the manufacturer, which is different for vari-
ous types and sizes of cells. Normal discharge converts active materials to normal lead
sulfate, which requires only slightly more space than active materials. Overdischarge
forms more lead sulfate in the pores of the active material than they are able to hold. This
process may expand and bend or buckle the plates or crack the grids. In some instances,
sufficient pressure is created to crack or puncture separators. Overdischarge may lead to
cell polarity reversal, which permanently damages the cell.
6.18.5.7 Sediment The history of each cell is shown by the sedimentation because
successive layers are laid down in colored strata at the bottom of the cell. These layers can
be seen edgewise against the inside of the cell container. Fine, dark brown or black layers
indicate periods of excessive charging (current too high or charge too long). Lumpy gray
layers indicate times the battery was overdischarged. A layer of white sulfate from subse-
quent charges generally covers these layers. A considerable amount of sediment and sliv-
ers will be found initially in some batteries. This condition is a normal result of the form-
ing process. Some additional sediment and slivers will be dislodged in shipment and will
accumulate at the bottom of the case of these batteries during the first few equalizing
charges. With this exception, a perfectly charged battery should have little fine brown or
gray-white sediment and no lumps. If some experimenting is done with the charging pro-
gram, slight undercharging may result in a white sulfate layer. This layer indicates that
the float voltage should be slightly increased.
6.18.5.8 Battery Watering

Water Requirements As cells are charged, a small quantity of water in the electrolyte
is broken down into hydrogen and oxygen by the charging current. The gases are dissipat-
ed through the vents. As this process takes place, the electrolyte level gradually falls until
water must be added. Commercially available “demineralized” water is equal to commer-
cial or locally distilled water and may be substituted for it. Do not add water above the
maximum level mark. Specific gravity readings will be inaccurate until the water has had
time to mix with the rest of the electrolyte, which can take weeks in some cells. General-
ly, at least 72 h should elapse from the time water is added to a cell to the time specific
gravity measurements are made on that cell.
Always use glass, plastic, or rubber containers to store battery water; never store it in
any kind of metallic container. Keep records of water consumption and compare the long-
term average with Figure 6.48.
Water Replacement Rate for Lead–Antimony Cells Lead–antimony cells begin

their lives with low water consumption, but water consumption increases as much as five
times toward the end of their lives. Very little water evaporates from capped cells; loss is
caused by gassing and is proportional to the amount of charge the battery receives. Heavy
gassing requires frequent water additions. When equalizing a battery, the water should be
added just before or at the beginning so that gassing will ensure thorough mixing before
specific gravity readings are taken. Proper charging minimizes excess gassing and the
need for adding distilled water.
Water Replacement Rate for Lead–Calcium Cells Water additions two or three
times a year should be sufficient if the lead–calcium cells are properly charged. Frequent
water additions to lead–calcium cells indicate improper charging. The electrolyte in all
cells should be maintained within 1–4 in. (6 mm) below the high-level mark. Because of
greater purity of their components, lead–calcium cells require only about one-tenth the
water needed by equivalent size lead–antimony cells. This low requirement remains con-
stant during the entire battery life.
6.18.5.9 Specific Gravity Adjustments Specific gravity should not be adjusted

until it is definitely established to be wrong and, even then, only upon recommendation
by the battery manufacturer. Before adjusting for low specific gravity, make sure that it
cannot be raised by equalize charging. Continue the charge until the specific gravity
shows no rise and then charge for 3 more hours. Do not adjust specific gravity on a cell
that does not gas on charge—the cell probably has failed.
To increase specific gravity, remove some electrolyte from the cell and replace it with
pure, 1.300 specific gravity sulfuric acid (30% concentrated acid and 70% water by vol-
ume, or 39% concentrated acid and 61% water by weight). Recharge until all cells gas for
an hour. Repeat the procedure if the gravity is still not normal. To lower the gravity, re-
move some of the electrolyte and replace it with distilled or demineralized water. Specific
gravity measurements should be made after waiting at least 72 h for the acid or water to
mix in the cell.
6.18.5.10 Internal Short Circuits A short circuit through a separator may be

caused by:
앫 Sulfated plates—Insufficient charging causes material in the plates to become most-

ly lead sulfate. The lead sulfate expands and, if the grid does not crack to relieve the
strain, the plate will distort or buckle. The buckling is most pronounced, and shorts
are most likely to occur, at the four corners of the positive plates.
앫 Impurities in the electrolyte caused by using contaminated water or dirty utensils.
앫 Plate contact—Excessive overcharging causes the grid to be partially converted to
lead dioxide, which reduces its mechanical strength and allows positive and nega-
tive plate contact.
A short in a cell can be detected by falling specific gravity and falling cell voltage over
time. In some cases, a gray discoloration occurs at the point of the short. If a short is long-
standing, disintegration of the positive plate will occur at the point of contact with the
negative plate because of the conversion of positive plate material to negative.
6.18.5.11 Sulfation
Normal and Oversulfation During cell discharge, “normal” sulfate is formed, which is
required to produce current. If recharging is not performed in a reasonable time, the sul-
fate fills the pores of the plates and makes the active material dense and hard. This condi-
tion is referred to as “oversulfated.”
Normal lead sulfate formed on discharge is in a form that a charge will easily recon-
vert. When a battery is oversulfated, plates are less porous than normal and absorb a
charge with difficulty. With this condition, an ordinary charge will not reconvert the en-
tire sulfate and specific gravity remains below normal. The active material of oversulfated
negative plates is light in color and either hard and dense or granular and gritty and easily
disintegrated. Oversulfated negative plates require a prolonged charge to restore. The
plates in an individual cell may become oversulfated by external grounding, by an inter-
nal short, or by exposure to air because electrolyte level was neglected and water was not
added. Prolonged low float charging may also cause oversulfation.
Treating Oversulfation A battery or cell that is oversulfated should be fully equalize

charged in the regular way until specific gravity stops rising. Then, one of the weakest
cells should be discharged through a load bank at the normal 8-h discharge rate to a final
voltage of 1.75 V. The battery is not oversulfated if the representative cell gives normal
capacity—that is, about 100% rated capacity for a fairly new battery or down to 80% for a
battery nearing the end of its expected life.
If the required capacity is not obtained, possible oversulfation should be treated as fol-
lows:
1. In cases where one or more individual cells have become oversulfated and the rest
of the battery is in good condition, these cells should be treated separately after re-
moving them from the circuit.
2. Recharge the removed cells at half the 8-h discharge rate. Record specific gravity
readings and temperature at regular intervals (3 to 5 h) during the charge to deter-
mine if the specific gravity has peaked. Maintain constant electrolyte level by
adding water after each reading. Do not add water before taking readings.
3. Continue the charge and record the specific gravity readings until no further rise oc-
curs in any cell for 10 h. If the temperature reaches 38°C (100°F), reduce the cur-
rent or temporarily interrupt the charge so as not to exceed this temperature. When
the specific gravity has peaked, terminate the charge and record the specific gravity
reading of each cell.
4. Replace the cells if they again fail the capacity check.
6.18.6 Vented Lead–Acid Battery Testing

The following information does not replace specific instructions by the battery manufac-
turer. Follow the manufacturer’s information in case of a conflict between these guide-
lines and the manufacturer’s information.
6.18.6.1 Acceptance Testing Acceptance tests should be performed at least one

week after the battery has been operating at the proper float voltage or given an initial
charge according to the manufacturer’s instructions. Factory tests do not replace on-site
acceptance tests because there may be shipping and handling damage to the battery prior
to installation. The acceptance test should be at least a 3-h and preferably an 8-h discharge
test and should discharge at least 90% of rated capacity.
Maintain accurate records of the acceptance test, including make, model, and serial
number of all equipment used and the test results. These initial records will be used as a
baseline for later comparisons.
1. Conduct the test only after all terminal and cell interconnection resistances have
been measured and any discrepancies corrected.
2. Install an accurate ammeter, voltmeter, and thermometer, and use an accurate stop-
watch to record elapsed time.
3. Record the following parameters just prior to the test.
앫 Specific gravity and voltage of each cell
앫 Temperature of the electrolyte of 10% or more of the cells to establish an aver-
age temperature (an IR thermometer may be used)
앫 Battery terminal voltage
4. Disconnect the battery from the rest of the dc power system and connect it to a load
bank (Fig. 6.50) so that constant current can be maintained equal to the 3- or 8-h
rating of the battery at the initial battery temperature.
5. Using the average temperature from step 3, calculate the end-of-discharge voltage.
An end-of-discharge voltage of 1.75 V/cell applies to an average cell temperature
of 25°C (77°F). If the average cell temperature is not 25°C (77°F), determine the
new end-of-discharge voltage as follows: For each 10°C above 25°C, add 0.006 V
to 1.75 V, and for each 10°C below 25°C, subtract 0.006 V from 1.75 V. Similarly,
for each 10°F above 77°F, add 0.01 V to 1.75 V, and for each 10°F below 77°F,
subtract 0.01 V from 1.75 V.
6. Measure and record individual cell voltages and the battery terminal voltage. The
measurements should be taken after applying the load at the beginning of the test.
Repeat the measurements at one-half hour intervals until the last hour of the test.
Fig. 6.50 Load bank for capacity testing a battery system. (Photo courtesy of M.W. Migliaro.)
During the last hour take measurements at shorter intervals such as 10 min. Mea-
sure individual cell voltage between respective terminals of like polarity (positive
to positive) so the voltage drop of the cell interconnection straps are included in the
measurement.
7. If an individual cell is approaching reversal of its polarity (voltage drops to 1.0
V/cell or less), terminate the test and consult the manufacturer.
8. Maintain the discharge rate until the battery terminal voltage decreases to 1.75 V/cell
times the number of cells (21.00 V for a 12-cell and 42.00 V for a 24-cell battery).
9. Calculate the capacity from
Actual Time
Capacity25°C(%) = × 100 (6.7)
Rated Time × KT
where KT is the temperature correction factor (Table 6.21 or 6.22).
6.18.6.2 Capacity Testing Capacity tests are very similar to acceptance tests de-
scribed above. Frequent VLA capacity tests are not recommended (each discharge/charge
cycle wears out the battery). However, a capacity test on initial installation will establish
a baseline for a new battery, and capacity tests at 5-year intervals will indicate if the bat-
tery is aging normally. Additional capacity tests may be needed to determine if a battery
is nearing the end of its useful life or for warranty claim purposes. Absent specific proce-
dures from the manufacturer, use the following test procedures. Perform yearly tests when
the measured capacity drops below 90% of rated capacity, and replace the battery when it
has reached 80% of rated capacity. Although capacity tests are very similar to acceptance
tests previously described, capacity tests are separately described in this section because
they could involve network equipment that is in operation and extra precautions are nec-
essary to ensure that equipment operation is not jeopardized.
1. Equalize the battery and then return it to float service for between 3 and 30 days.
2. Take precautions to ensure that a battery failure will not jeopardize other equip-
ment. Disconnect the battery to isolate it from the rest of the dc power system.
4. Check all battery connections visually and with a microohmmeter to ensure con-
nections are clean and low resistance. An IR thermometer may be used to check the
connections just after the discharge test has begun. The temperature will be higher
on poor connections. If poor connections are found, stop the test and repair them
before continuing.
5. Record the following parameters just prior to the test:
앫 Specific gravity and voltage of each cell
앫 Temperature of the electrolyte of 10% or more of the cells to establish an aver-
age temperature (an IR thermometer may be used)
6. Use the average temperature measured in step 5 to calculate the end-of-discharge
voltage. An end-of-discharge voltage of 1.75 V/cell applies to an average cell tem-
perature of 25°C (77°F). If the average cell temperature is not 25°C (77°F), deter-
mine the new end-of-discharge voltage as follows: For each 10°C above 25°C, add
0.006 V to 1.75 V, and for each 10°C below 25°C, subtract 0.006 V from 1.75 V.
Similarly, for each 10°F above 77°F, add 0.01 V to 1.75 V, and for each 10°F be-
low 77°F, subtract 0.01 V from 1.75 V.
7. Connect the battery to a load bank. Discharge the battery through the load bank for
3 h at the rated 3-h discharge current (or for 8 h at the rated 8-h discharge current)
until it reaches the end-of-discharge voltage determined in step 6. Record the volt-
age of each cell every half-hour during discharge. Watch closely during the last
hour to determine the exact time when the end-of-discharge voltage is reached. If
the end-of-discharge voltage is reached before 3 (or 8) h has elapsed, the test should
be stopped and the capacity should be computed. If the measured capacity is be-
tween 80 and 90% of rated capacity, the battery should be tested again in one year.
The entire battery should be replaced as soon as possible after capacity drops below
80%.
Actual Time
Capacity25°C(%) = × 100 (6.8)
Rated Time × KT
where KT is the temperature correction factor (Table 6.21 or 6.22).
As with acceptance tests, maintain accurate records of all capacity tests, including make,
model, and serial number of all test equipment used and the test results. These records
should be compared to the baseline records and previous tests to determine the battery
performance trend.
6.18.7 Flame Arrester Vents

Article 480.9 of the National Electrical Code [6] requires each vented cell to be equipped
with a flame arrester designed to prevent destruction of the cell by ignition of gases inside
the cell by an external spark or flame. The diffuser material in flame arresters can become
Table 6.21 Temperature Correction Factors for Battery Capacity Tests (Celsius)a
Initial Temperature Initial Temperature Initial Temperature
Temperature Correction Temperature Correction Temperature Correction
(°C) Factor KT (°C) Factor KT (°C) Factor KT
5 0.684 22 0.966 30 1.045
10 0.790 23 0.977 31 1.054
15 0.873 24 0.986 32 1.063
16 0.888 25 1.000 33 1.072
17 0.902 26 1.006 34 1.081
18 0.916 27 1.015 35 1.090
19 0.929 28 1.025 40 1.134
20 0.942 29 1.036 45 1.177
21 0.954
a
These factors apply to batteries with 1.215 specific gravity (consult the manufacturer for batteries with differ-
ent specific gravities).
Source: IEEE Std 450-2002 [5].
6.19. VALVE-REGULATED LEAD–ACID BATTERIES 469
Table 6.22 Temperature Correction Factor for Battery Capacity Tests (Fahrenheit)a
Initial Temperature Initial Temperature Initial Temperature
Temperature Correction Temperature Correction Temperature Correction
(°F) Factor KT (°F) Factor KT (°F) Factor KT
40 0.670 73 0.975 85 1.040
45 0.735 74 0.980 86 1.045
50 0.790 75 0.985 87 1.050
55 0.840 76 0.990 88 1.055
60 0.882 77 1.000 89 1.060
65 0.920 78 1.002 90 1.065
66 0.927 79 1.007 95 1.090
67 0.935 80 1.011 100 1.112
68 0.942 81 1.017 105 1.140
69 0.948 82 1.023 110 1.162
70 0.955 83 1.030 115 1.187
71 0.960 84 1.035 120 1.210
72 0.970
a
These factors apply to batteries with 1.215 specific gravity (consult the manufacturer for batteries with differ-
ent specific gravities).
Source: IEEE Std 450-2002 [5].
partially clogged from electrolyte spray if cells are overfilled with water or have been ex-
cessively overcharged.
Flame arresters should have dust caps and be in good condition. Examine them annual-
ly and replace those having clogged pores or clean them as follows:
1. Immerse the flame arrester vent several times in a plastic bucket full of fresh clean
water.
2. Remove the water from the vent after each immersion by vigorous shaking or using
compressed air.
3. Refill the bucket with clean water for every 15 flame arresters that are cleaned.
4. Do not use any cleaning or neutralizing agents in the water because dry residue
may clog the pores of the diffuser materials.
5. Thoroughly dry the flame arresters so that the plain water used to clean them can-
not drip into the cells and contaminate them.
6.19. VALVE-REGULATED LEAD–ACID BATTERIES
6.19.1 General
This section outlines the requirements for operating and maintaining valve-regulated
lead–acid (VRLA) batteries. As with VLA, the battery manufacturer and industry stan-
dards are the best place to get detailed operation and maintenance information; however,
this section provides sufficiently detailed information for most VRLA batteries.12
12
See also IEEE Std 1188 [7].
6.19.2 Basic VRLA Characteristics

Valve-regulated lead–acid (VRLA) batteries are built in multicell assemblies (modules)
or blocs (monoblocs). Because the cases are usually made of opaque plastic material, vis-
ual inspection of the internal components is impossible. The hydrogen and oxygen nor-
mally are not vented but are recombined, and the cells are sealed, making specific gravity
readings and water addition impossible. These cells are typically lead–calcium pasted-
plate-type cells with the electrolyte immobilized in gel form (gel-cells) or in fiberglass
mats (absorbed glass matt, AGM, cells).
These cells are not flooded and do not dissipate heat as effectively as VLA cells. This
characteristic can lead to thermal runaway if ambient and cell temperatures are not care-
fully controlled. Numerous cases have occurred in which a battery has burst into flame or
cracked and spewed electrolyte throughout the room. Maintaining the cells as close as
possible to 25°C (77°F) is imperative. Ambient temperature should be maintained so as to
not exceed 22°C (72°F).
Air circulation must be sufficient to eliminate cell temperature differences. Do not al-
low sunlight or other heat sources to raise the temperature of individual cells. The maxi-
mum cell temperature spread (warmest to coldest cell) should not exceed 3°C (5°F), and
the hottest cell should not be more than 3°C (5°F) above ambient. Lower temperatures re-
duce capacity, and higher temperatures reduce service life and increase the chances of
thermal runaway. As a rule of thumb, battery capacity will be decreased by about 10% for
every 10°C (18°F) below 25°C (77°F), and battery service life will be decreased by about
50% for every 8°C (15°F) a battery is continuously operated above 25°C (77°F).
6.19.3 Battery Charging and Measurement Procedures
6.19.3.1 Initial Freshening Charge

Valve-regulated lead–acid batteries and modules are typically shipped fully charged and
do not require an initial charge unless recommended by the manufacturer. If VLRA bat-
teries are stored for more than a couple months before installation, they should be given a
“freshening” charge immediately after the installation is completed.
Unless the manufacturer specifies a different voltage, normal equalizing voltage should
be used (Table 6.23). If the battery is connected to a power system, care must be taken to
ensure the voltage used does not exceed the maximum rated voltage of the load equipment.
The initial charge voltage should be applied until the cell voltages stop rising. Just
Table 6.23 Typical Float and Equalize Voltages at 25°C (77°F)—Valve Regulated Lead–Acid
Batteries
Cell Voltage 12-Cell System 24-Cell System
Type (V/cell) (V) (V)
Float
VRLA—1.300 SG 2.25–2.27 27.00–27.24 54.00–54.48
a
Equalize
VRLA—1.300 SG 2.30–2.35 27.60–28.20 55.20–56.40
a
VRLA batteries should not be equalized or operated at elevated voltages unless recommended by the manufac-
turer.
before the end of the initial charge, record the cell voltages and then change to the nor-
mal float voltage. After allowing the cells to stabilize for 72 h, measure the cell volt-
ages.
6.19.3.2 Float Charge Valve-regulated lead–acid cells are typically floated at the
voltages shown in Table 6.23; some variations exist, as indicated. Correct battery float
voltage is critical for VRLA cells. To reduce the chance of thermal runaway, the float
voltage must be temperature compensated such that the higher the temperature the lower
the float voltage.
When VRLA cells are operated on float at normal full charge, no net chemical reaction
occurs and almost all the overcharge energy results in heat generation. If the environment
is such that the heat produced can be dissipated, no thermal runaway problems occur. If
the heat generation rate exceeds the dissipation rate, the battery temperature rises, charge
acceptance increases, and more current is required to maintain the float voltage. The high-
er current results in more heat generation, which raises the battery temperature further,
and the cycle continues until the battery destroys itself. Elevated ambient temperature
above 22°C (72°F) or cell or rectifier malfunction will aggravate this condition.
As cells approach full charge, charging current decreases. The battery is fully charged
when the charging current has not changed more than 10% for more than 3 h. If the charg-
ing voltage has been set higher than the required float voltage to reduce the charging time,
then the charging voltage must be reduced to normal float value after the charging current
has stabilized. Never exceed the manufacturer’s recommended maximum charging volt-
age.
6.19.3.3 Equalizing Charge Equalizing charges are not normally performed on

VRLA cells because the cells are sufficiently charged at the normal float voltage; howev-
er, an equalizing charge may be necessary if a faster than normal recharge time is needed
after a discharge. Manufacturers’ data sheets normally provide a recommended voltage
range, and the float voltage normally is set to a midrange value. If an equalizing voltage is
used, it is set to the high end of the recommended range. The risk of VRLA thermal run-
away is higher during equalize charging.
6.19.4 Periodic Measurements and Inspections

Regular visual examinations of VRLA cells are very important to check for bloated or
cracked cases (Fig. 6.51), leaking seals (Fig. 6.52), and warped cell interconnection
straps.
Where temperature-compensated charging is used, the battery voltage will vary with
the temperature, and measurements should be corrected to the baseline temperature of
25°C (77°F). Temperature-compensated charging adjusts rectifier voltage by 2.5 to 4.5
mV/°C (0.7 to 2.5 mV/°F). For example, if the battery temperature (or the cell used to
sense battery temperature) is 31°C (88°F) and the rectifiers are set to 4.5 mV/°C (2.5
mV/°F) compensation, the voltage of a 24-cell battery should read 27 mV lower than at
25°C (77°F).
The VRLA battery float current is an important indicator of impending cell failure.
The float current in normal cells does not increase appreciably throughout the battery’s
life. Any significant increase indicates one or more cells that are in the failure mode.
Fig. 6.51 Consequences of excessive plate growth in VRLA cells. (Photos courtesy of M.W.
Migliaro.)
Fig. 6.52 Post seal leak on cell 21 of a VRLA battery. (Photo courtesy of M.W. Migliaro.)
6.19.4.1 Initial Installation Measurements Initial installation measurements can

be made coincidentally with the acceptance tests (Section 6.19.5) and should be per-
formed no sooner than one week after the battery has had its initial freshening charge. A
complete set of measurements should be taken to establish a baseline for a new installa-
tion including
앫 Internal impedance, resistance or conductance of each cell or monobloc
앫 Voltage of each cell or monobloc
앫 Battery voltage measured at the battery terminals
앫 Ambient temperature and temperature of each cell or monobloc
앫 Cell or monobloc interconnecting strap resistances and terminal resistances
앫 Battery float current
6.19.4.2 Monthly Measurements

앫 Examine each cell or module for general appearance and cleanliness, cracked cov-
ers and containers, electrolyte leakage, and excessive bloating (some types have a
“normal bloat range”).
앫 Examine the overall battery, battery rack, and battery room or area for appearance
and cleanliness. Remove combustible materials such as empty boxes from the room.
앫 Measure and record pilot cell voltage and temperature.
앫 Measure and record float voltage at battery terminals.
앫 Measure and record float current.
앫 Measure and record rectifier output voltage and current.
앫 Examine cell or module and battery terminals and connections for corrosion.
앫 Examine racks and fasteners for corrosion and overall structural integrity (shake
test).
앫 Check the ambient temperature and ensure ventilation devices (fans and vents) are
operable.
앫 Check battery monitoring system operation, if equipped.
6.19.4.3 Quarterly Measurements

앫 All monthly work items.
or monobloc and compare the results with the initial records. Changes in the inter-
nal resistance of 20% or greater should be considered significant. Contact the bat-
tery manufacturer and follow its recommendations.
앫 Measure and record the temperature of the negative terminal of each cell or module.
앫 If the battery has a discharge rate of 1 h or less, measure and record the cell inter-
connecting strap resistances on a minimum of 10% of all connections. If an upward
trend is noted from the initial readings, measure all connection resistances and take
corrective action. Test different connections each quarter.
6.19.4.4 Semiyearly Measurements In addition to the quarterly work items,

measure and record the voltage of each cell or monobloc and compare with previous mea-
surements for trending purposes. Measurements should be from cell post to cell post.
6.19.4.5 Yearly Measurements

앫 All quarterly and semiyearly work items.
앫 Measure and record the interconnecting strap resistances of each cell or monobloc
and of the battery terminals.
앫 Measure and record the ac ripple current and voltage at the battery terminals.
앫 Test the temperature-compensated charging system.
or monobloc and compare to initial and previous measurements for trends.
앫 Measure and record the ground fault resistance of each cell or monobloc.
6.19.5 Value-Regulated Lead–Acid Battery Testing

6.19.5.1 Acceptance Testing Acceptance tests for VRLA batteries are very similar
to VLA batteries. Acceptance tests should be performed at least one week after the battery
has been operating at the proper float voltage or given an initial charge according to the
manufacturer’s instructions. Factory tests do not replace on-site acceptance tests because
there may be shipping and handling damage to the battery prior to installation. The accep-
tance test should be at least a 3-h and preferably an 8-h discharge test and should discharge
at least 90% of rated capacity. The operating temperature of the battery will affect the bat-
tery capacity; manufacturers’ data sheets must be consulted for correction factors.
Maintain accurate records of the acceptance test, including make, model and serial
number of all equipment used and the test results. These initial records will be used as a
baseline for later comparisons.
1. Conduct the test only after all terminal and cell interconnection resistances have
been measured and any discrepancies corrected.
앫 Temperature of 10% or more of the cells to establish an average temperature (an
IR thermometer may be used)
앫 Individual cell or monobloc voltages
4. Disconnect the battery from the rest of the dc power system and connect it to a load
bank so that constant current can be maintained equal to the 3- or 8-h rating of the
battery at the initial battery temperature.
5. Using the average temperature from step 3, calculate the end-of-discharge voltage.
An end-of-discharge voltage of 1.75 V/cell applies to an average cell temperature
of 25°C (77°F). If the average cell temperature is not 25°C (77°F), determine the
new end-of-discharge voltage as follows: For each 10°C above 25°C, add 0.006 V
to 1.75 V, and for each 10°C below 25°C, subtract 0.006 V from 1.75 V. Similarly,
for each 10°F above 77°F, add 0.01 V to 1.75 V, and for each 10°F below 77°F,
subtract 0.01 V from 1.75 V.
6. Measure and record individual cell voltages and the battery terminal voltage. The
measurements should be taken after applying the load at the beginning of the test.
Repeat the measurements at one-half hour intervals until the last hour of the test.
During the last hour take measurements at shorter intervals such as 10 min. Mea-
sure individual cell voltage between respective terminals of like polarity (positive
to positive) so the voltage drop of the cell interconnection straps are included in the
measurement.
7. If an individual cell is approaching reversal of its polarity (voltage drops to 1.0
V/cell or less) or the voltage of a monobloc (6- or 12-V module) is more than 2 V
lower than the others, terminate the test and consult the manufacturer.
8. Maintain the discharge rate until the battery terminal voltage decreases to 1.75
V/cell times the number of cells (21.00 V for a 12-cell and 42.00 V for a 24-cell
battery).
Actual Time
Capacity25°C(%) = × 100 (6.9)
Rated Time × KT
where KT is the temperature correction factor (obtain from battery manufacturer).
6.19.5.2 Capacity Tests As with VLA, frequent VRLA capacity tests are not rec-
ommended. However, tests at 1- or 2-year intervals will indicate if the battery is aging
normally. Additional capacity tests may be needed to determine if a battery is nearing the
end of its useful life or for warranty claim purposes. Absent specific procedures from the
manufacturer, use the following test procedures. If the test shows the measured capacity
has dropped below 90% rated capacity, then tests should be made every 6 months to 1
year. When the measured capacity has reached 80% of rated capacity, it should be re-
placed. Although capacity tests are very similar to acceptance tests previously described,
capacity tests are separately described in this section because they could involve network
equipment that is in operation, and extra precautions are necessary to ensure that equip-
ment operation is not jeopardized.
1. Charge the battery at the high end of its recommended float voltage range and then
return it to float service for between 3 and 30 days.
2. Take precautions to ensure that a battery failure will not jeopardize other equip-
ment. Disconnect the battery to isolate it from the rest of the dc power system.
4. Check all battery connections visually and with a microohmmeter to ensure con-
nections are clean and low resistance. An IR thermometer may be used to check the
connections just after the discharge test has begun. The temperature will be higher
on poor connections. If poor connections are found, stop the test and repair them
before continuing.
앫 Temperature of 10% or more of the cells to establish an average temperature (an
IR thermometer may be used)
앫 Individual cell or monobloc voltages
6. Use the average temperature measured in step 5 to calculate the end-of-discharge

voltage. An end-of-discharge voltage of 1.75 V/cell applies to an average cell tem-
perature of 25°C (77°F). If the average cell temperature is not 25°C (77°F), deter-
mine the new end-of-discharge voltage as follows: For each 10°C above 25°C, add
0.006 V to 1.75 V, and for each 10°C below 25°C, subtract 0.006 V from 1.75 V.
Similarly, for each 10°F above 77°F, add 0.01 V to 1.75 V, and for each 10°F be-
low 77°F, subtract 0.01 V from 1.75 V.
7. Connect the battery to a load bank. Discharge the battery through the load bank for 3
h at the rated 3-h discharge current (or for 8 h at the rated 8-h discharge current) un-
til it reaches the end-of-discharge voltage determined in step 6. Record the voltage of
each cell or monobloc every half-hour during discharge. Watch closely during the
last hour to determine the exact time when the end-of-discharge voltage is reached.
If the end-of-discharge voltage is reached before 3 (or 8) h has elapsed, the test
should be stopped and the capacity should be computed. If the measured capacity is
between 80 and 90% of rated capacity, the battery should be tested again in one year.
The entire battery should be replaced as soon as possible after capacity drops below
80%.
Actual Time
Capacity25°C(%) = × 100 (6.10)
Rated Time × KT
where KT is the temperature correction factor (obtain from battery manufacturer).
As with acceptance tests, maintain accurate records of all capacity tests, including make,
model, and serial number of all test equipment used and the test results. These records
should be compared to the baseline records and previous tests to determine the battery
performance trend.
6.20 BATTERY FAILURE CASE STUDIES
6.20.1 Cell Explosion in UPS Battery

In Figure 6.53 corrosion in the post connection of cell 38 eventually caused the post to
open at the connection between cell 38 and 39. Cell 38 exploded, breaking the partition
wall to cell 39, which then exploded as well. This corrosion could not be visually detected
because the connection that corroded was made by the manufacturer in a “well” in the jar
cover and was encased in epoxy.
6.20.2 Lightning Damage to 12-Cell Wireless Site Battery

Figure 6.54 shows lightning damage to a 12-cell wireless site battery in the southern
United States. The lightning currents found their way to the battery causing high over-
pressures presumably from rapid heating of the internal materials. The vent caps were
blown out of seven cells and there was obvious damage to eight cells, exposing the
plates. Interestingly, the site was not contaminated with electrolyte and there was no
other damage.
6.20 BATTERY FAILURE CASE STUDIES 477
Fig. 6.53 Cell explosion in UPS battery. (Photo and description courtesy of M.W. Migliaro.)
6.20.3 Corroded Interconnecting Strap

The corrosion shown in Figure 6.55 occurred adjacent to the cell post and was noticed by
a maintenance engineer who said the post had a large amount of “blue-green stuff” on it.
6.20.4 Corrosion Around Terminal Post

The corrosion around the terminal post shown in Figure 6.56 has forced the post seal cover
off a cell in a poorly maintained battery. This particular problem was pointed out to mainte-
nance personnel, but when the site was visited 18 months later, there had been no changes.
Fig. 6.54 Lightning damage to 12-cell wireless site battery. (Photo courtesy of Gary Trent.)
Fig. 6.55 Corroded interconnecting strap. (Photo and description courtesy of M.W. Migliaro.)
Fig. 6.56 Corrosion around the terminal post. (Photo and description courtesy of M.W. Migliaro.)
Fig. 6.57 Excessive oxidation inhibitor compound. (Photo courtesy of M.W. Migliaro.)
478
6.20 BATTERY FAILURE CASE STUDIES 479
6.20.5 Excessive Oxidation Inhibitor Compound

Thick applications of the compound as shown in Figure 6.57 not only are difficult to clean
but also hide potential problems.
Fig. 6.58 Post melting. (Photo and description courtesy of M.W. Migliaro.)
Fig. 6.59 Plate bowing in cell in 19-year-old battery. (Photo and description courtesy of M.W.
Migliaro.)
6.20.6 Post Melting

The battery shown in Figure 6.58 was a new telecommunications battery that was being
acceptance tested, but the user failed to check the intercell connection resistance
and torque before the test. Some connections were loose and a number of cell posts
melted.
6.20.7 Plate Bowing in Cell in 19-Year-Old Battery

The associated plate growth shown in Figure 6.59 caused two jars to crack and spill elec-
trolyte.
Fig. 6.60 Plate sulfation in shorted cell. (Photo and description courtesy of M.W. Migliaro.)
Fig. 6.61 VRLA Thermal runaway. (Photo courtesy of M.W. Migliaro.)

REFERENCES (PART III) 481
6.20.8 Plate Sulfation in Shorted Cell

The cell in Figure 6.60 was found to have a crack in the separator between two plates.
6.20.9 VRLA Thermal Runaway

Figure 6.61 shows a VRLS battery with thermal runaway, cause unknown.
REFERENCES (PART III)
1. M. Migliaro, Stationary Batteries: A Pictorial Reference for Maintenance and Testing, The Bat-
tery Connector, Inc., 2001 (marco@esaconsulting.com).
2. M. Migiliaro, M., Stationary Battery Calculations Associated with Maintenance & Testing, 3rd
ed., 1995.
3. Intelec® 2001, 23rd International Telecommunications Energy Conference, Oct. 14–18, 2001,
Edinburgh International Conference Centre, Edinburgh, UK (IEEE catalog number 01CH37336).
4. Intelec® 2002, 24th Annual Telecommunications Energy Conference, Sept. 29–Oct. 3, 2002,
Montreal, Quebec, Canada (IEEE catalog number 02CH37374).
5. ANSI/IEEE 450™-2002, IEEE Recommended Practice for Maintenance, Testing, and Replace-
ment of Vented Lead-Acid Batteries for Stationary Applications, Institute of Electrical and Elec-
tronics Engineers, Piscataway, NJ, 2002.
6. NFPA 70-2005, National Electrical Code®, National Fire Protection Association, 2005.
7. ANSI/IEEE 1188™-1996, IEEE Recommended Practice for Maintenance, Testing, and Re-
placement of Valve Regulated Lead-Acid (VRLA) Batteries for Stationary Applications, Institute
of Electrical and Electronics Engineers, Piscataway, NJ, 1996.
VLA BATTERY MAINTENANCE REPORT—FORM 1
Period From: ______________________________________ To: ___________________________________________________________

Location: _________________________________________ Date Installed: __________________________________________________
String No.: ___________________ No. Cells: ____________ Specific Gravity: _________________________________________________
Battery Manufacturer: ________________________________________ Model: ________________________________________________
Type: ____________________________________________ Amp-Hour Capacity: _____________________________________________
Notes:
1. Rotate pilot cell yearly Form © Whitham D. Reeve
2. All readings taken with battery stabilized on float Electronic version of this form available
3. Do not add water prior to taking readings from wreeve@ieee.org
Monthly Readings Quarterly Readings

Charging Pilot Cell No. Date:
Electrolyte Ambient
Date V I Level V Temp. SG Temp Cell Voltage Specific Gravity
1
2
3
4
5
Corrective Action 6
Water Addition to: (circle cell numbers at right) 7
Date: Qty Added (total): 8
Reason for equalize charge: 9
Date: Voltage: Duration: 10
13
14
Monthly Visual Examination (indicate Y/N) 15
Positive plates dark brown/free from sulfate crystals ________________ 16
Positive plates/internal hardware free from flaking ________________ 17
Terminal post seals okay ________________
18
Flame arrester vents and dust caps okay ________________
Connections free of corrosion ________________ 19
Cell containers and covers in good condition ________________ 20
Rack in good condition/free from corrosion ________________ 21
Ventilating fans and vents working ________________ 22
Insulated tools available/in good conditions ________________
23
Hydrometer and thermometer in good condition ________________
Goggles, face shield, gloves, apron available ________________ 24
Minimum 1-gallon labeled soda solution ________________
Eye wash, body spray clean and working ________________
Class C fire extinguisher available and inspected ________________
Circle one in each column

Sediment Negative Plates Grids
Temperatures of 10% of Cells
Fine Scarce Brown Dark Mossy tops Distorted
No. No. No.
Coarse Medium Gray Medium Clean Cracked
No. No. No.
Lumps Excessive White White Spots Good
List instruments by type and serial number used to make measurements: _____________________________________________________
________________________________________________________________________________________________________________
Comments (if additional space is required, write on back): __________________________________________________________________
________________________________________________________________________________________________________________
________________________________________________________________________________________________________________
________________________________________________________________________________________________________________
________________________________________________________________________________________________________________
Name: ______________________________________________ Signed: _____________________________________________________
482
VRLA BATTERY MAINTENANCE REPORT—FORM 2
Period From: ______________________________________ To: ___________________________________________________________

Notes:
1. Rotate pilot cell yearly.
2. All readings taken with battery stabilized on float. Form © Whitham D. Reeve
3. Individual cell voltages should vary no more than 0.03 V from average cell voltage. Electronic version of this form available
4. Cell temperatures should be no greater than 5°F above ambient temperature. from wreeve@ieee.org
Monthly Readings Quarterly Readings

Charging Pilot Cell No. Date:
Ambient
Date V I V Temp. Temp Cell Voltage Temperature Notes
1
2
3
4
5
Corrective Action 6
Water Addition to: (circle cell numbers at right) 7
Date: Qty Added (total): 8
Reason for equalize charge: 9

13
14
Monthly Visual Examination (indicate Y/N) 15
Cells not bloated ________________ 16
Terminal post seals okay ________________ 17
Connections free of corrosion ________________
18
Cell containers and covers in good condition ________________
Rack in good condition/free from corrosion ________________ 19
Ventilating fans and vents working ________________ 20
Insulated tools available/in good conditions ________________ 21
Hydrometer and thermometer in good condition ________________ 22
Goggles, face shield, gloves, apron available ________________
23
Minimum 1-gallon labeled soda solution ________________
Eye wash, body spray clean and working ________________ 24
Class C fire extinguisher available and inspected ________________
Annual Examination (indicate Y/N)

Cell inter-connection strap torque okay ________________
Temperatures of 10% of Cells

No. No. No.
No. No. No.
________________________________________________________________________________________________________________
________________________________________________________________________________________________________________
________________________________________________________________________________________________________________
________________________________________________________________________________________________________________
________________________________________________________________________________________________________________
Name: ______________________________________________ Signed: _____________________________________________________
483
CELL IMPEDANCE & INTER-CONNECTION RESISTANCE REPORT—FORM 3
Measurement Date: _________________________________

Period From: ______________________________________ To: ___________________________________________________________
Notes:
1. See back of this form for instructions and post orientation and numbering. Form © Whitham D. Reeve
2. Use additional sheets for double- and triple-post cells. Electronic version of this form available
from wreeve@ieee.org
Parallel No. 1 Parallel No. 2 Parallel No. 1 Parallel No. 2 Internal

As-Found As-Found As-Left As-Left Impedance
Meter probes between (micro-ohms) (micro-ohms) (micro-ohms) (micro-ohms) (ohms)
Lug–Cell No. 1 Post Cell 1
Post/Cells 1–2 Cell 2
Cell No. 24 Post—Lug
Averages
________________________________________________________________________________________________________________
________________________________________________________________________________________________________________
________________________________________________________________________________________________________________
________________________________________________________________________________________________________________
________________________________________________________________________________________________________________
Name: ______________________________________________ Signed: _____________________________________________________
CELL IMPEDANCE & INTER-CONNECTION RESISTANCE REPORT—FORM 3 485
Start at the battery terminal connected to cell no. 1, which normally is the cell closest to ground potential, and move from there to
the opposite end of the battery. Ensure the battery is on float charge before beginning the measurements and not on equalize or be-
ing discharged. Record all measurements on this form.
1. Take the first measurement on cell No. 1 between the terminal lug and the first post. This measurement will be the resistance between
the cell post and battery terminal lug and will be about one-half of the cell interconnecting strap readings in 2 and 3 below.
2. Take the second measurements between opposite polarity posts (not interconnecting straps) of cells No. 1 and No. 2. The subse-
quent measurements will be between positive and negative posts of adjacent cells to include the cell interconnecting straps and hard-
ware connections. These measurements will include the resistance of two bolted connections (one on each post) and the cell-inter-
connecting strap and will be about twice the value of the first and last measurements at the battery terminals. If the cells have two
parallel posts for each polarity, measure both and record on the form as “Parallel No. 1” and “Parallel No. 2”.
3. Measure the remaining interconnections as in step 2, moving from cell to cell.
4. Take the last measurement between the last post on the last cell and the battery terminal lug fastened to it.
5. If a comparatively high resistance is found, measure from each post to its cell-interconnecting strap to determine which of the two con-
nections is bad. Mark this and all high-resistance connections for later repair.
6. Configure the test set to measure cell impedance. Measure cell No. 1 between its two posts.
7. Measure the remaining cells as in step 6, moving from cell to cell.
APPENDIX A
COMMENTS ON HISTORIC dc POWER

SYSTEM CONFIGURATIONS
A.1 INTRODUCTION
The battery and rectifier configuration described in this book at one time was called a
“straight cell system” as compared to two other configurations called a “counter cell sys-
tem” (also called “counter emf cell system” or “cemf cell system”) and an “end-cell sys-
tem” (also called “emergency cell system”).1 These two configurations were used with
electromechanical (step-by-step, or SXS) switching systems. They were replaced when
the SXS switching systems were replaced first by stored program-controlled (SPC) and
then digital switching systems in the public network in the early 1970s and early 1980s,
respectively.
The SXS switching systems operated at a nominal –48 Vdc and required relatively
tight voltage regulation (~10%) as compared to the later technologies, which were de-
signed to work over a somewhat wider voltage range (~20%). Although the tighter regu-
lation could be provided by using an extra-large battery and only partially discharging it,
in most cases either cemf cells or end cells were used to improve load voltage regulation
under all conditions of battery recharge and discharge. Thus, at the expense of higher
complexity, the two configurations helped to mitigate two limitations of the straight cell
system:
앫 Relatively poor voltage regulation during battery charging, equalizing, and dis-
charging
앫 Not fully discharging the battery capacity due to relatively high load minimum volt-
age limits
A.2 COUNTER emf CELL SYSTEM
The cemf cells originally were an alkaline battery cell with nickel electrodes that devel-
oped a relatively constant voltage drop (1.5 to 2 V). The cemf cells were later made from
1
Counter electromotive force (cemf) or counter voltage.

488 COMMENTS ON HISTORIC dc POWER SYSTEM CONFIGURATIONS
selenium diodes and later still from regular silicon diodes in series and parallel arrange-
ments to achieve the desired voltage drop and current capacity.
The cemf cells were placed in series with the load under normal float conditions, thus
“countering” the rectifier voltage, and shorted during battery discharge (Fig. A.1). When
the battery was recharging and the bus voltage increasing, the cemf cells were automati-
cally switched in again. In many systems, cemf cells were not inserted in the load circuit
to reduce the voltage except when the battery was charged at the equalizing voltage of
56.4 V (2.35 V/cell). The cemf cells also were used in individual load circuits (rather than
the main bus) to control the voltage to only certain loads.
A.3 END-CELL SYSTEM
End cells were ordinary lead–acid cells automatically switched in series with the load to
add voltage under discharge conditions (Fig. A.2). The end cells usually were switched in
steps to avoid exceeding the allowable load upper voltage limit. Thus, a 24-cell battery
was increased to 25 cells and then 26 cells when on discharge. A quick calculation shows
that a 26-cell battery has an end-of-discharge (final) voltage of 45.5 V (1.75 V/cell). This
allowed the SXS equipment to operate while simultaneously deriving maximum discharg-
ing capacity from the battery.
A.4 SUMMARY
Neither the cemf nor end-cell configurations are used in modern dc power systems be-
cause:
앫 Both systems required relaying and interlocking schemes that added complexity
and were subject to failure.
Fig. A.1 The cemf cell system schematic. During normal float operation, the cemf cells are in series
with the load circuit and the system output voltage at the primary distribution system is the rectifier
float voltage less the voltage drop across the cemf cells (~1.5 V each). When the battery is dis-
charged and its voltage decreases, the cemf cells are shorted by the contactors and the full battery
voltage is impressed on the load circuits. A stepped arrangement can be used (shown) to more
closely control the voltage delivered to the loads. When the battery is recharged, the cemf cells are
automatically switched back into the circuit by opening the contactors.
A.4 SUMMARY 489
Fig. A.2 End cell system schematic. During normal float operation, the end cells are switched out of
the load circuit by open contactors 1 and 2, and the primary distribution system is connected to the
main battery through contactor 0. When the battery discharges and its voltage decreases, the end
cells are automatically switched in by contactors 1 and 2, thus increasing the load voltage by about 2
V with each step. When the main battery is recharged, the end cells are switched out and indepen-
dently recharged by end-cell rectifiers. Contactors 0, 1, and 2 are interlocked to prevent shorting the
end cells.
앫 Both systems used break-before-make relay contactors or switches that either mo-
mentarily interrupted the dc current or quickly changed the load voltage, which
caused transients and noise from the momentary arcing. While acceptable in SXS
switching systems, these characteristics were unacceptable in SPC and digital
switching systems.
앫 The relatively tight voltage control required by SXS switching systems was not re-
quired by SPC and digital switching systems.
The SPC and digital switching systems were deployed in the network in total replace-
ment and cap-and-grow schemes. In total replacement schemes, the dc power system used
with the SXS system was completely replaced with a new straight cell system.2 In cap-
and-grow schemes the old SXS system was retained in service while the access lines were
migrated to the new SPC system over a period of years, thus shrinking the SXS and grow-
ing the SPC. Meanwhile, the old dc power system (either cemf or end cell) was main-
tained along side a new straight cell system for the SPC equipment. Eventually, the SXS
and its power system were decommissioned.
2
Actually, many SPC switching systems used two battery plants, one for the switching matrix and “talk battery”
at 48 Vdc and another at 24 Vdc for the computer and common control equipment. Both were straight cell sys-
tems. Some SPC systems used dc–dc converters to derive the 24 Vdc from the main 48-Vdc plant.
APPENDIX B
MATHCAD WORKSHEET EXAMPLES
The following worksheet examples show a few of the many ways to use commercial soft-
ware, such as Mathcad by Mathsoft Engineering & Education, Inc. (www.mathsoft.com),
to perform iterative ampacity calculations. Although the basic calculation procedures are
similar for all physical wire configurations, different procedural techniques are used in
the examples to show the flexibility available.
These examples are not intended to show the best or most efficient programming
method; they are intended to allow the reader to duplicate the calculations and to calculate
the ampacities for wire sizes not covered.
B.1 SINGLE WIRE IN FREE AIR
Example shows 4/0 AWG jacketed 1/C with 90°C maximum wire temperature and 30°C
ambient temperature.
Tmax := 90 Ta := 30 Dc := 0.528 tins := 0.125 De := Dc + 2 tins
冤冢冣冥
.21
DeltaT := Tmax – Ta ins := 5 hcr := + 3.94
25.4 .6
De ·
1000
tins
冢
ins · ln 1 + 2 ·
Trins :=
Dc 冣
2·
25.4
Ka := · De · · hcr · Trins err:= 0.00001 Rw := 0.06376
1000

492 MATCHED WORKSHEET EXAMPLES
冤
冤
Ts(err) := x씯2 1
Diffr 씯 .001 Trsa := 25.4
while |Diffr| > err · De · · hcr · (Ts(.0001)).25
1000
DTs 씯 x4
y 씯 DTs.25
冤冥
DeltaT DeltaT
冤冥
.25 5
x 씯 I :=
(1 + Ka · y) Rw
· (TRins + TRsa)
1000 · 0.3048
Diffr 씯 x – y
y4 I = 467
B.2 TWO BUNDLED WIRES
Tmax := 75 Ta := 30 Dc := 0.528 tins := 0.125 De := Dc + 2 tins
冤冢冣冥
.29
DeltaT := Tmax – Ta ins := 5 hcr := + 2.35
25.4 .5
De ·
1000
tins
冢
ins · ln 1 + 2 ·
Trins :=
Dc 冣
2·
25.4
Ka := · De · · hcr · Trins err:= 0.0001 Rw := 0.06081
1000
冤
冤
Ts(err) := x씯2
1
Diffr 씯 .001 Trsa :=
25.4
while |Diffr| > err · De · · hcr · (Ts(0.0001)).25
DTs 씯 x4 1000
y 씯 DTs.25
冤冥
DeltaT DeltaT
冤冥
.25
x 씯 I := 5
(1 + Ka · y) Rw
· 0.3048 · (TRins + TRsa)
1000
Diffr 씯 x – y
y4 I = 355
B.4 WIRES IN ROUNDED BUNDLE 493
B.3 WIRES WITH NO SEPARATION IN ONE OR MORE LAYERS
ambient temperature. w
Tmax := 90 Ta := 30 d := .5, 1 . . 7 w := 24 As := 2 · Fdiv := 1 Dw := 0.73
12
0.06376
:= 0.8 eff := 13.13 := 5.3 · 10–9 err := .00001 Rw := Aw := Dw2
1000
Wwm(d) := Wwm 씯 0
Diffr 씯 .0001
while |Diffr| > err
Wwm 씯 Wwm + Diffr · .0
Wtemp 씯 Wwm · Fdiv
(Wwm · eff · d)
Ts 씯 Tmas – 冤 (8 · w) 冥
hconv 씯 .101 · (ts – Ta).25
Diffr 씯 [hconv · As · (Ts – ta)] + [ · As · · [(Ts + 273.15)4
– (Ta + 273.15)4]] – Wtemp
Wwm
Wwm(d) 0 0
qw(d) :=
d·w 0 15.809 0 363
Ww(d) := qw(d) · Aw 1 6.997 1 242
2 4.195 2 187
冪莦
Ww(d)
I(d) := 3 2.864 3 155
Rw
4 2.105 4 133
5 1.624 5 116
qw(d) = 6 1.296 I(d) = 6 104
7 1.062 7 94
8 0.888 8 86
9 0.755 9 79
10 0.650 10 74
11 0.566 11 69
12 0.498 12 65
13 0.442 13 61
494 MATCHED WORKSHEET EXAMPLES
B.4 WIRES IN ROUNDED BUNDLE
0.3048
Tmax := 90 Ta := 30 Dc := 0.528 tins := 0.125 Dw := (Dc + 2 · tins) ·
12
n := 2, 4 . . 16 Dw := 0.73
DT := Tmax – Ta ins := 16.4 · 0.3048 eff := 13.12 · 0.3048 Ab(n) := n · Dw2
Aw := Dw2
0.06376
Rw := err := .0001
1000 · .3048
冢冣
冪莦 · Dw
n 58.400
Wb(n) := Deb 씯 2 · 69.645
76.700
.21
hcr 씯 + 3.94 Wb(n) =
81.867
Deb0.6 85.944
Wtemp 씯 0.001 89.308
Diffr 씯 .001 92.167
while |Diffr| > err 94.651
Wstemp 씯 Wtemp + Diffr - .5
Wb 씯 Wtemp
冤冤冥
DT
Wb(n)
Wtemp 씯 eff 1 _______ qb(n) :=
+ _______________
4 – ( · Deb · hcr · Wb.25)4/5 冥 Ab(n)
Diffr 씯 Wb – Wtemp
冢冣
Wb 7.478 × 104
4.459 × 104
冢冣
374 3.274 × 104
Ww(n) := qb(n) · Aw 289 2.621 × 104
247 qb(n) =
2.201 × 104
冪莦
Ww(n) 221 1.906 × 104
I(n) := I(n) =
203
Rw 1.686 × 104
189 1.515 × 104
177
168
INDEX
ac power system, 8, 9 capacity, 23, 210

prime, 8 capacity calculations, 217
standby, 8 cell configurations, 152
Access network, 3 charge acceptance, 179
Alarm circuits, 118 cycle, 160
Alternating current, 66 design margin (uncertainty factor), 217
applications, 66 discharge factor, 212, 215
Alternating current circuits, 72 electrochemistry, 156
with capacitance, 74 end-of-life factor (aging factor), 216
with inductance, 73 for long-term loading, 15
other forms, 76 for short-term loading, 15
with resistance only, 72 heat damage, 181
with resistance, inductance, and capacitance, lead–acid, 14
74 modular stacking, 180
wiring, 81 number of strings, 218
Alternating current circuits for rectifier inputs parallel connection, 153
and inverter outputs, 359 rack structures, 220
Alternating current power, 261 reserve time, 211
Ampacity, 20 Series connection, 153
ATIS Network, Interface, Power and telecommunications, 151
Protection Committee, 26 temperature factor, 216
Automatic equalize controller, 121 temperature performance, 174
Auxiliary framing, 412 temperature probes, 181
installation guidelines, 412 thermal runaway, 179
requirements for high-risk seismic areas, 415 types, 151
valve-regulated lead–acid (VRLA), 14
Basic requirements for telecommunications vented, 162
power systems, 1 vented lead–acid (VLA), 14
Batteries, 5, 14 VLA cells, 154
active components, 152 voltage drop, 213

496 INDEX
Battery charging and measurement procedures, International Fire Code (IFC, formerly
456 Uniform Fire Code), 252
Battery circuit wiring, 425 NFPA 70 National Electrical Code (NEC),
check charge, 458 251
equalizing charge, 456 Bus arrangements, 8
float charge, 456 Busbar orientation, 45
Initial Freshening Charge, 456 Busbars, 111, 408
Battery equalize control, 120 charge, 111
Battery failure case studies, 476 copper, 112
Battery life, 176 discharge, 111
end-of-life definition, 176 installation guidelines, 408
life-limiting and failure mechanisms, 177 sizes, 114
Battery racks, 222
Battery room ventilation, 253 Cable and rigid bus supporting structure, 10
Battery system installation, 419 Cable racks, 415
Battery system maintenance, 433 installation guidelines, 415
acid electrolyte spill management, 437 Capacitor fuse, 91
cell ground faults, 440 Capacitor precharge circuit, 122, 123
cell internal impedance measurements, capacity calculations, 217
448 “Carrier hotel,” 3
cleanliness, 434 Cell Impedance and Inter-Connection
connection resistance measurements, 442 Resistance Report, 484
electrical connections, 438 Cell impedance and conductance, 183
general requirements, 433 Cell installation guidelines, 420
monthly, quarterly, and yearly Inspections, Cells, 155
448 active components, 155
records, 434 conductance, 183
safety practices and tools, 435 containers, 161
temperature measurements, 442 equivalent circuit, 183
test equipment, 433 gassing, 187
voltage measurements, 441 Impedance, 183
Battery systems, 7 ventilation, 187
checklist, 222 Central office, 3
choice of technology, 208 Charge acceptance, 179
design, 207 Charge bus, 7
Battery voltage and discharge curves, 165 Charging, 461
cell voltages, 165 appearance of normal cells, 461
discharge, 167 chemical, 461
discharge factor, 172 effect of impurities on float voltage, 462
float and equalize charge voltages, 166 high-rate overcharging, 462
overdischarge, 174 low-rate overcharging, 462
self-discharge, 172 undercharging, 462
specific gravity, 165 Circuit breaker alarm circuits, 121
Battery watering, 463 Circuit breaker interrupt ratings, 125
water replacement rate for lead–antimony Circuit breakers, 127, 130
cells, 463 direct circuit resistance, 135
water replacement rate for lead–calcium magnetic, 133
cells, 464 short-circuit capability, 134
water requirements, 463 thermal, 130
Belleville washers, 394 thermal-magnetic, 131
Building codes, 251 time delay characteristics, 134
International Building Code (IBC), 252 transient (false or nuisance) tripping, 134
INDEX 497
trip and switching configurations, 135 equipment frame bonding and grounding,
Circuit redundancy, 11 375
Circular mil, 40 grounding conductor Size, 373
Codes, 24, 29 inverter bonding and grounding, 374
International Building Code (IBC), 29 Direct current power system elements, 5
International Fire Code (IFC), 29 Direct current power system installation, 381
National Electrical Code (NEC), 29 basic installation sequence, 418
National Electrical Safety Code (NESC), 29 cable rack fill, 385
U.S. Dept. of Labor, Occupational Safety & circuit routing and wire length, 383
Health Administration (OSHA), 29 circuit wiring, 382
Collocation, 3, 207 conductor size and bend radius, 383
Conductors, 20, 33 material compatibility, 382
rigid busbar, 39 safety guidelines, 381
wire, 39 Direct current power system loads, 17
Controlled ferroresonant transformer rectifier, Direct current power systems circuit design,
95 263
Converter system capacity, 244 ambient temperature correction factors,
Counter emf cell system, 487 321
Current density, 52 battery circuits, 339
Current flow, 31, 32 battery connection device selection and
Current shunt, 112, 114 sizing, 344
battery connection methods, 339
dc power system components, 85 battery disconnect switches and circuit
dc power system design considerations, 189 breakers, 340
basic design steps, 190 battery overcurrent protection device
collocation requirements, 193 characteristics and coordination, 349
current drain studies, 192 circuit breaker–circuit breaker coordination,
floor space requirements and availability, 333
191 circuit configurations, 263
maintenance requirements, 193 circuit design procedures, 335
new technologies deployment, 193 conductor terminations and taps, 324
planning interval, 190 conductors, 263
system operating voltages and ranges, 193 current rating in cable racks, 296
wire center forecasts, 192 fault currents, 324
DC–DC converters, 135, 243 fuse–circuit breaker coordination, 333
basic circuit components, 141 general considerations, 263
common-bus configuration, 143 parallel conductors, 269
conceptual schematic, 140 parallel strings, 350
economic analysis, 245 power cable supporting structure, 352
split-bus configuration, 143 rigid busbar current ratings, 322
Design loads, 18 selective coordination of overcurrent
Design margin (uncertainty factor), 217 protective devices, 332
Detailed requirements and considerations, 2 short-circuit heating of conductors, 330
Direct current power system bonding and temperature considerations, 263
grounding, 367 voltage drop calculations, 270
battery rack bonding and grounding, 375 wires with no separation in one or more
building principal ground, 367 layers, 305
cable racking bonding and grounding, 375 Direct current power systems component
DC–DC converter system bonding and design, 189
grounding, 375 Direct current powerline filters, 89
direct current power equipment bonding and Discharge bus, 7
grounding, 368 Discharge factor (DF), 172
498 INDEX
Distribution system, 7 redundant, 146

primary, 8
secondary, 8 Kirchhoff’s laws, 55
current law, 55
Electrical Generating Systems Association voltage law, 55
(EGSA), 27
Electrical power, 49 Lighting, 10
Electricity, 31 Load characteristics, 194, 196, 197
Electromotive force, 31 estimating equipment loads, 195
End-cell system, 488 Load conditions, 206
End-of-life factor, 216 Load equipment, 10
Energy storage, 14 Low-voltage disconnects, 116
Equipment frame, 10
Estimating equipment loads, 195 Mathcad worksheet examples, 491
European Telecommunications Standards single wire in free air, 491
Institute (ETSI), 30 two bundled wires, 492
wires in rounded bundle, 494
Fasteners, 391 wires with no separation in one or more
Belleville washers, 394 layers, 493
disk spring washers, 396 Monitoring and control systems, 110, 117
installation cuidelines, 394 Multiplate cell, 155
Fine-strand wire, 42
Flame arrester vents, 468 National Fire Protection Association (NFPA),
Floor openings, 417 29
Floor structural components, 10 Noise and transients in dc power systems, 86
Fuse alarm circuits, 120 Noise characteristics, 86
Fuses, 127 Nominal system and operating voltages, 85
cartridge, 129
configurations, 128 Ohm’s law, 44
direct current resistance, 130 Operating voltages, 193
operating parameters, 128 Overcurrent protection, 22, 122
short-circuit capability, 129 Overdischarge, 174, 463
telecommunications, 130 Overloads, 124
time delay characteristics, 129
types, 128 Phase, 69
POTS (plain old telephone service), 3
Grounding system, 10 Power, 71
in alternating current circuits, 71
Historic dc power system configurations, 487 Power cable supporting structure, 352
HVAC, 10 general rules for power cable rack layout,
358
Institute of Electrical and Electronics structural considerations, 352
Engineers (IEEE), 26 Power equipment frames, 411
Internal short circuits, 464 installation guidelines, 411
International Telecommunications Union, 29 Power equipment spaces, 252
Inverter system maintenance, 432 Power sources and loads, 12
annual check, 432 alternating current, 14
offline, 432 essential, 14
Inverter systems, 245 nonessential, 14
Inverters, 144 prime, 12
circuit description, 147 protected, 14
configurations, 145 standby, 13
INDEX 499
Power systems applications in the Rectifiers, 5, 22, 91

telecommunications network, 3 control functions, 108
Power wiring installation, 397 controlled ferroresonant transformer, 92
busbars, 408 current limitation, 102
cable tie application and installation efficiency, 103
guidelines, 401 features, 99
drop wires from equipment frame fuse high-voltage shutdown, 106
panels, 406 internal protection, 108
guidelines for securing and physically load sharing, 99
protecting power wiring, 398 operation mode, 99
lacing (stitching or sewing) power wires, parallel operation, 99
400 power factor correction, 105
repairing damaged insulation, 408 remote control, 99
routing guidelines, 397 silicon-controlled rectifier, 92
tap installation guidelines, 408 switch-mode rectifier, 92
using cable ties on power wiring, 401 technology, 92, 98
wire taps, 406 temperature-compensated charging, 100
Powerboards, 10, 109 visual indicators and controls, 106
bus ratings, 109 voltage regulation, 103
busbar mounting arrangements, 109 voltage sensing, 99
Powerboard design, 233 Reference plane, 32
alarm and control set points, 238 Relative conductance, 35
bus type and amperage rating, 233 Resistance, 33
checklist, 241 in parallel, 37
conductor terminations, 235 in series, 36
control options, 237 units, 35
current shunts, 237 Ripple voltages, 87
distribution options, 240 Rural Utilities Service (RUS), 27
equalizer control, 237
low-voltage disconnects, 238 Secondary distribution design, 243
voltage rating, 233 Secondary distribution systems, 122
Powerline filters, 89 Sediment, 463
applications, 249 Series–parallel circuit, 48
dc, 89 Short circuits, 126
Practices, 24, 28 Short-circuit currents, 184
Primary distribution system, 122 Sine wave voltage and frequency, 67
Single-phase voltages, 77
Rectifier system design, 223 Sleeves, 417
ac input circuit requirements, 227 Specific gravity, 165
battery voltage sensing, 227 adjustments, 464
capacity, 225 measurement procedures, 451
checklist, 232 Standards, practices, and codes, 24
dc output circuit requirements, 232 ATIS Network, Interface, Power and
output configurations, 224 Protection Committee, 26
rectifier quantity, 225 Electrical Generating Systems Association,
technology, 223 27
Rectifier system maintenance, 429 Institute of Electrical and Electronics
annual check, 429 Engineers (IEEE), 26
load sharing Tests, 430 Rural Utilities Service, 27
temperature-compensated charging tests, Telcordia, 27
431 Telecommunications Industry Association,
Rectifier systems, 91 27
500 INDEX
Sulfation, 464 periodic measurements and inspections, 471

normal, 464 quarterly measurements, 473
oversulfation, 464 semiyearly measurements, 473
treating, 465 yearly measurements, 474
Switch-mode rectifier, 97, 98 Valve-regulated lead–acid battery (VRLA)
System maintenance, 381, 426 cells, 164
annual check, 428 advantages, 164
fuse replacements, 427 comparison to VLA cells, 164
maintenance analysis, 427 disadvantages, 165
maintenance schedules, 426 Vented lead–acid batteries (VLAs), 449
overall maintenance requirements, 426 maintenance report, 482
test equipment and tool calibration, 427 monthly measurements, 454
tools and test equipment, 427 periodic measurements and inspections, 453
quarterly measurements, 455
Telcordia, 27 specific gravity measurements, 449
Telecommunications Industry Association yearly measurements, 455
(TIA), 27 Vented lead–acid (VLA) battery cells, 154
Telecommunications terminal, 3 advantages, 164
Temperature coefficient of resistance, 35 comparison to VLRA cells, 164
Temperature factor, 216 disadvantages, 164
Terminal connector lugs, 385 Vented lead–acid battery (VLA) testing, 465
compression type, 386 acceptance testing, 465
installation guidelines, 387 capacity testing, 467
insulated terminal lug installation guidelines, Ventilation, 253
391 VLA, see Vented lead–acid batteries
lugs for small wire, 390 Voice-over-asynchronous transfer mode
mechanical type, 387 (VoATM), 3
Thermal runaway, 179 Voice-over-Internet protocol (VoIP), 3
Three-phase voltages, 77 Voice-over-packet, 3
Transients, 88 Voltage and current in-phase, 71
Triac-controlled ferroresonant transformer Voltage and current out-of-phase by angle ,
rectifier, 96 71
Troubleshooting battery problems, 458 Voltage conversion, 135
cell replacement, 460 Voltage conversion systems design, 243
charging, 461 Voltage converter, 8
surface charge phenomenon, 459 Voltage drop, 19
Voltage magnitudes, 16
Valve-regulated lead–acid batteries (VRLAs), alternating current, 16
469 direct current, 17
acceptance testing, 474 Voltage regulation, 103
basic characteristics, 470 Voltage source, 32
battery charging and measurement alternating current, 32
procedures, 470 batteries, 32
capacity tests, 475 rectifiers, 33
equalizing charge, 471 voltage conversion devices, 33
float charge, 471 VRLA, see Valve-regulated lead–acid batteries
initial freshening charge, 470
maintenance report, 483 Wall openings, 417
monthly measurements, 473

DC Power-System-Design-for-Telecommunications PDF

Uploaded by

Copyright:

Available Formats

DC Power-System-Design-for-Telecommunications PDF

Uploaded by

Document Information

Original Title

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

DC Power-System-Design-for-Telecommunications PDF

Uploaded by

Copyright:

Available Formats

DC POWER SYSTEM

IEEE Press Telecommunications Handbook Series

A JOHN WILEY & SONS, INC., PUBLICATION

IEEE Press Editorial Board

J. B. Anderson R. J. Herrick F. M. B. Pereira

Kenneth Moore, Director of IEEE Book and Information Services (BIS)

IEEE Press Telecommunications Handbook Series

A JOHN WILEY & SONS, INC., PUBLICATION

Published by John Wiley & Sons, Inc. Published simultaneously in Canada.

Library of Congress Cataloging-in-Publication Data is available.

Printed in the United States of America.

3 dc Power System Components 85

4 Telecommunications Batteries 151

5 System Design 189

6 System Installation and Maintenance 381

6.7 Power Equipment Frames 411

Appendix A Comments on Historic dc Power System Configurations 487

Appendix B Mathcad Worksheet Examples 491

1.1 BASIC REQUIREMENTS FOR TELECOMMUNICATIONS

1.2 APPLICATIONS REVIEW

Telecommunications dc power systems and modern network equipment in the United

Table 1.1 Basic Requirements for Telecommunications Power Systems

dc Power System Design for Telecommunications. By Whitham D. Reeve 1

Table 1.2 Detailed Requirements and Considerations

Fig. 1.1 Power systems applications in the telecommunications network.

1.3 DIRECT CURRENT POWER SYSTEM ELEMENTS

A telecommunications dc power system consists of seven basic elements, shown schemat-

앫 Rectifier (charger) system

a. Power the loads when commercial ac power is available.

Fig. 1.3 Simplified schematic of a telecommunications dc power system.

Fig. 1.4 Basic elements of a telecommunications dc power system.

Fig. 1.6 Typical charge and discharge bus arrangements.

Fig. 1.8 Block diagram of a hybrid power supply system.

Fig. 1.10 Powerboard frame assembly.

A common method of improving the reliability and performance of dc power systems

1.4 POWER SOURCES AND LOADS

1.4.1 Prime Power Sources

앫 Electrical equipment failures (e.g., damage to service transformers, distribution

1.4.2 Standby Power Sources

Table 1.3 Prime Movers and Other Energy Sources

1.4.3 Alternating Current Power System Loads

1.4.4 Energy Storage

Batteries used in telecommunications are classified as secondary (rechargeable) and

앫 Batteries for short-term loading (< 1 h)

1.4.5 Voltage Magnitudes

1.4.6 Direct Current Power System Loads

Facility ac Power System

Fig. 1.14 Decentralized power for network equipment shelves.

1.5 GENERAL DESIGN CONSIDERATIONS

A telecommunications dc power system is required to provide the specified voltage range

1.5.1 Design Loads

1.5.2 Voltage Drop

Fig. 1.15 dc power system circuits.

Fig. 1.16 Cascaded secondary distribution circuits.

1.5.4 Overcurrent Protection

Fig. 1.18 Overcurrent protection.

Fig. 1.19 Primary distribution (left) and secondary distribution (right).

1.5.6 Battery Capacity

1.6 STANDARDS, PRACTICES, AND CODES

RT = R0[1 + 0(TT – T0)] (2.2)

RT = R0[1 + (TT – T0)] = 0.00252[1 + 0.00393(75 – 20)]