TRANSFORMER

AND INDUCTOR DESIGN

HANDBOOK
Third Edition, Revised and Expanded

COLONEL WM. T. MCLYMAN

Kg Magnetics, Inc.
Idyllwild, California, U.S.A.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Although great care has been taken to provide accurate and current information, neither the author(s) nor the publisher, nor anyone else
associated with this publication, shall be liable for any loss, damage, or liability directly or indirectly caused or alleged to be caused by
this book. The material contained herein is not intended to provide specific advice or recommendations for any specific situation.

Trademark notice: Product or corporate names may be trademarks or registered trademarks and are used only for identification and
explanation without intent to infringe.

Library of Congress Cataloging-in-Publication Data

A catalog record for this book is available from the Library of Congress.

ISBN: 0-8247-5393-3

This book is printed on acid-free paper.

Headquarters
Marcel Dekker, Inc.
270 Madison Avenue, New York, NY 10016, U.S.A.
tel: 212-696-9000; fax: 212-685-4540

Distribution and Customer Service

Marcel Dekker, Inc.
Cimarron Road, Monticello, New York 12701, U.S.A.
tel: 800-228-1160; fax: 845-796-1772

Eastern Hemisphere Distribution

Marcel Dekker AG
Hutgasse 4, Postfach 812, CH-4001 Basel, Switzerland
tel: 41-61-260-6300; fax: 41-61-260-6333

World Wide Web

http://www.dekker.com

The publisher offers discounts on this book when ordered in bulk quantities. For more information, write to Special Sales/Professional
Marketing at the headquarters address above.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

Neither this book nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical, including
photocopying, microfilming, and recording, or by any information storage and retrieval system, without permission in writing from the
publisher.

Current printing (last digit):

10 9 8 7 6 5 4 3 2 1

PRINTED IN THE UNITED STATES OF AMERICA

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 ELECTRICAL AND COMPUTER ENGINEERING
A Series of Reference Books and Textbooks

FOUNDING EDITOR

Martin O. Thurston
Department of Electrical Engineering
The Ohio State University
Columbus, Ohio

1. Rational Fault Analysis, edited by Richard Saeks and S. R. Liberty

2. Nonparametric Methods in Communications, edited by P. Papantoni-Kazakos
and Dimitri Kazakos
3. Interactive Pattern Recognition, Yi-tzuu Chien
4. Solid-State Electronics, Lawrence E. Murr
5. Electronic, Magnetic, and Thermal Properties of Solid Materials, Klaus Schroder
6. Magnetic-Bubble Memory Technology, Hsu Chang
7. Transformer and Inductor Design Handbook, Colonel Wm. T. McLyman
8. Electromagnetics: Classical and Modern Theory and Applications, Samuel See//
and Alexander D, Poularikas
9. One-Dimensional Digital Signal Processing, Chi-Tsong Chen
10. Interconnected Dynamical Systems, Raymond A. DeCar/o and Richard Saeks
11. Modern Digital Control Systems, Raymond G. Jacquot
12. Hybrid Circuit Design and Manufacture, Roydn D. Jones
13. Magnetic Core Selection for Transformers and Inductors: A User's Guide to
Practice and Specification, Colonel Wm. T. McLyman
14. Static and Rotating Electromagnetic Devices, Richard H. Engelmann
15. Energy-Efficient Electric Motors: Selection and Application, John C. Andreas
16. Electromagnetic Compossibility, Heinz M. Schlicke
17. Electronics: Models, Analysis, and Systems, James G. Gottling
18. Digital Filter Design Handbook, Fred J. Taylor
19. Multivariable Control: An Introduction, P. K. Sinha
20. Flexible Circuits: Design and Applications, Steve Gurley, with contributions by
Carl A. Edstrom, Jr., Ray D. Greenway, and William P. Kelly
21. Circuit Interruption: Theory and Techniques, Thomas E. Browne, Jr.
22. Switch Mode Power Conversion: Basic Theory and Design, K. Kit Sum
23. Pattern Recognition: Applications to Large Data-Set Problems, Sing-Tze Bow
24. Custom-Specific Integrated Circuits: Design and Fabrication, Stanley L. Hurst
25. Digital Circuits: Logic and Design, Ronald C. Emery
26. Large-Scale Control Systems: Theories and Techniques, Magdi S. Mahmoud,
Mohamed F. Hassan, and Mohamed G. Darwish
27. Microprocessor Software Project Management, Eli T. Fathi and Cedric V. W.
Armstrong (Sponsored by Ontario Centre for Microelectronics)
28. Low Frequency Electromagnetic Design, Michael P. Perry
29. Multidimensional Systems: Techniques and Applications, edited by Spyros G.
Tzafestas
30. AC Motors for High-Performance Applications: Analysis and Control, Sakae
Yamamura

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 31. Ceramic Motors for Electronics: Processing, Properties, and Applications,
edited by Relva C. Buchanan
32. Microcomputer Bus Structures and Bus Interface Design, Arthur L. Dexter
33. End User's Guide to Innovative Flexible Circuit Packaging, Jay J. Miniet
34. Reliability Engineering for Electronic Design, Norman B. Fuqua
35. Design Fundamentals for Low-Voltage Distribution and Control, Frank W.
Kussy and Jack L. Warren
36. Encapsulation of Electronic Devices and Components, Edward R. Salmon
37. Protective Relaying: Principles and Applications, J. Lewis Blackburn
38. Testing Active and Passive Electronic Components, Richard F. Powell
39. Adaptive Control Systems: Techniques and Applications, V. V. Chalam
40. Computer-Aided Analysis of Power Electronic Systems, Venkatachari
Rajagopalan
41. Integrated Circuit Quality and Reliability, Eugene R. Hnatek
42. Systolic Signal Processing Systems, edited by Earl E. Swartzlander, Jr.
43. Adaptive Digital Filters and Signal Analysis, Maurice G. Bel/anger
44. Electronic Ceramics: Properties, Configuration, and Applications, edited by
Lionel M. Levinson
45. Computer Systems Engineering Management, Robert S. Alford
46. Systems Modeling and Computer Simulation, edited by Nairn A. Kheir
47. Rigid-Flex Printed Wiring Design for Production Readiness, Walter S. Rigling
48. Analog Methods for Computer-Aided Circuit Analysis and Diagnosis, edited
by Takao Ozawa
49. Transformer and Inductor Design Handbook: Second Edition, Revised and
Expanded, Colonel Wm. T. McLyman
50. Power System Grounding and Transients: An Introduction, A. P. Sakis
Meliopoulos
51. Signal Processing Handbook, edited by C. H. Chen
52. Electronic Product Design for Automated Manufacturing, H. Richard Stillwell
53. Dynamic Models and Discrete Event Simulation, William Delaney and Erminia
Vaccari
54. FET Technology and Application: An Introduction, Edwin S. Oxner
55. Digital Speech Processing, Synthesis, and Recognition, SadaokiFurui
56. VLSI RISC Architecture and Organization, Stephen B. Furber
57. Surface Mount and Related Technologies, Gerald Ginsberg
58. Uninterruptible Power Supplies: Power Conditioners for Critical Equipment,
David C. Griffith
59. Polyphase Induction Motors: Analysis, Design, and Application, Paul L.
Cochran
60. Battery Technology Handbook, edited by H. A. Kiehne
61. Network Modeling, Simulation, and Analysis, edited by Ricardo F. Garzia and
Mario R. Garzia
62. Linear Circuits, Systems, and Signal Processing: Advanced Theory and
Applications, edited by Nobuo Nagai
63. High-Voltage Engineering: Theory and Practice, edited by M. Khalifa
64. Large-Scale Systems Control and Decision Making, edited by Hiroyuki Tamura
and Tsuneo Yoshikawa
65. Industrial Power Distribution and Illuminating Systems, Kao Chen
66. Distributed Computer Control for Industrial Automation, Dobrivoje Popovic
and Vijay P. Bhatkar
67. Computer-Aided Analysis of Active Circuits, Adrian loinovici
68. Designing with Analog Switches, Steve Moore

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 69. Contamination Effects on Electronic Products, CarlJ. Tautscher
70. Computer-Operated Systems Control, Magdi S. Mahmoud
71. Integrated Microwave Circuits, edited by Yoshihiro Konishi
72. Ceramic Materials for Electronics: Processing, Properties, and Applications,
Second Edition, Revised and Expanded, edited by Relva C. Buchanan
73. Electromagnetic Compatibility: Principles and Applications, David A. Weston
74. Intelligent Robotic Systems, edited by Spyros G. Tzafestas
75. Switching Phenomena in High-Voltage Circuit Breakers, edited by Kunio
Nakanishi
76. Advances in Speech Signal Processing, edited by Sadaoki Furui and M.
Mohan Sondhi
77. Pattern Recognition and Image Preprocessing, Sing-Tze Bow
78. Energy-Efficient Electric Motors: Selection and Application, Second Edition,
John C. Andreas
79. Stochastic Large-Scale Engineering Systems, edited by Spyros G. Tzafestas
and Keigo Watanabe
80. Two-Dimensional Digital Filters, Wu-Sheng Lu and Andreas Antoniou
81. Computer-Aided Analysis and Design of Switch-Mode Power Supplies, Yim-
Shu Lee
82. Placement and Routing of Electronic Modules, edited by Michael Pecht
83. Applied Control: Current Trends and Modern Methodologies, edited by Spyros
G. Tzafestas
84. Algorithms for Computer-Aided Design of Multivariable Control Systems,
Stanoje Bingulac and Hugh F. VanLandingham
85. Symmetrical Components for Power Systems Engineering, J. Lewis Blackburn
86. Advanced Digital Signal Processing: Theory and Applications, Glenn Zelniker
and Fred J. Taylor
87. Neural Networks and Simulation Methods, Jian-Kang Wu
88. Power Distribution Engineering: Fundamentals and Applications, James J.
Burke
89. Modern Digital Control Systems: Second Edition, Raymond G. Jacquot
90. Adaptive IIR Filtering in Signal Processing and Control, Phillip A. Regalia
91. Integrated Circuit Quality and Reliability: Second Edition, Revised and
Expanded, Eugene R. Hnatek
92. Handbook of Electric Motors, edited by Richard H. Engelmann and William H.
Middendorf
93. Power-Switching Converters, Simon S. Ang
94. Systems Modeling and Computer Simulation: Second Edition, Nairn A. Kheir
95. EMI Filter Design, Richard Lee Ozenbaugh
96. Power Hybrid Circuit Design and Manufacture, Haim Taraseiskey
97. Robust Control System Design: Advanced State Space Techniques, Chia-Chi
Tsui
98. Spatial Electric Load Forecasting, H. Lee Willis
99. Permanent Magnet Motor Technology: Design and Applications, Jacek F. Gieras
and Mitchell Wing
100. High Voltage Circuit Breakers: Design and Applications, Ruben D. Garzon
101. Integrating Electrical Heating Elements in Appliance Design, Thor Hegbom
102. Magnetic Core Selection for Transformers and Inductors: A User' s Guide to
Practice and Specification, Second Edition, Colonel Wm. T. McLyman
103. Statistical Methods in Control and Signal Processing, edited by Tohru Kata-
yama and Sueo Sugimoto
104. Radio Receiver Design, Robert C. Dixon
105. Electrical Contacts: Principles and Applications, edited by Paul G. Slade

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 106. Handbook of Electrical Engineering Calculations, edited by Arun G. Phadke
107. Reliability Control for Electronic Systems, Donald J. LaCombe
108. Embedded Systems Design with 8051 Microcontrollers: Hardware and Soft-
ware, Zdravko Karakehayov, Knud Smed Christensen, and Ole Winther
109. Pilot Protective Relaying, edited by Walter A. Elmore
110. High-Voltage Engineering: Theory and Practice, Second Edition, Revised and
Expanded, Mazen Abdel-Salam, Hussein An/'s, Ahdab EI-Morshedy, and
Roshdy Radwan
111. EMI Filter Design: Second Edition, Revised and Expanded, Richard Lee
Ozenbaugh
112. Electromagnetic Compatibility: Principles and Applications, Second Edition,
Revised and Expanded, David Weston
113. Permanent Magnet Motor Technology: Design and Applications, Second Edi-
tion, Revised and Expanded, Jacek F. Gieras and Mitchell Wing
114. High Voltage Circuit Breakers: Design and Applications, Second Edition,
Revised and Expanded, Ruben D. Garzon
115. High Reliability Magnetic Devices: Design and Fabrication, Colonel Wm. T.
McLyman
116. Practical Reliability of Electronic Equipment and Products, Eugene R, Hnatek
117. Electromagnetic Modeling by Finite Element Methods, Joao Pedro A. Bastos
and Nelson Sadowski
118. Battery Technology Handbook: Second Edition, edited by H. A. Kiehne
119. Power Converter Circuits, William Shepherd and Li Zhang
120. Handbook of Electric Motors: Second Edition, Revised and Expanded, edited
by Hamid A. Toliyat and Gerald B. Kliman
121. Transformer and Inductor Design Handbook: Third Edition, Revised and
Expanded, Colonel Wm. T. McLyman

Additional Volumes in Preparation

Energy-Efficient Electric Motors: Third Edition, Revised and Expanded, AH

Emadi

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 To My Wife, Bonnie

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Foreword

Colonel McLyman is a well-known author, lecturer, and magnetic circuit designer. His previous books on
transformer and inductor design, magnetic core characteristics, and design methods for converter circuits
have been widely used by magnetics circuit designers.

In this book, Colonel McLyman has combined and updated the information found in his previous books.
He has also added several new subjects such as rotary transformer design, planar transformer design, and
planar construction. The author covers magnetic design theory with all of the relevant formulas along with
complete information on magnetic materials and core characteristics. In addition, he provides real-world,
step-by-step design examples.

This book is a must for engineers working in magnetic design. Whether you are working on high "rel"
state-of-the-art design or high-volume or low-cost production, this book is essential. Thanks, Colonel, for a
well-done, useful book.

Robert G. Noah
Application Engineering Manager (Retired)
Magnetics, Division of Spang and Company
Pittsburgh, Pennsylvania, U.S.A.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Preface

I have had many requests to update Transformer and Inductor Design Handbook, because of the way
power electronics has changed over the past few years. This new edition includes 21 chapters, with new
topics such as: The forward converter, flyback converter, quiet converter, rotary transformers, and planar
transformers, with even more design examples than the previous edition.

This book offers a practical approach, with design examples for design engineers and system engineers in
the electronics and aerospace industries. Transformers are found in virtually all electronic circuits. This
book can easily be used to design lightweight, high-frequency aerospace transformers or low-frequency
commercial transformers. It is, therefore, a design manual.

The conversion process in power electronics requires the use of transformers, components that frequently
are the heaviest and bulkiest item in the conversion circuit. Transformer components also have a significant
effect on the overall performance and efficiency of the system. Accordingly, the design of such
transformers has an important influence on overall system weight, power conversion efficiency, and cost.
Because of the interdependence and interaction of these parameters, judicious trade-offs are necessary to
achieve design optimization.

Manufacturers have, for years, assigned numeric codes to their cores to indicate their power-handling
ability. This method assigns to each core a number called the area product, Ap, that is the product of its
window area, Wa, and core cross-section area, Ac. These numbers are used by core suppliers to summarize
dimensional and electrical properties in their catalogs. The product of the window area, Wa, and the core
area, Ac, gives the area Product, Ap, a dimension to the fourth power. I have developed a new equation for
the power-handling ability of the core, the core geometry, Kg. Kg has a dimension to the fifth power. This
new equation provides engineers with faster and tighter control of their design. It is a relatively new
concept, and magnetic core manufacturers are now beginning to include it in their catalogs.

Because of their significance, the area product, Ap, and the core geometry, Kg, are treated extensively in this
handbook. A great deal of other information is also presented for the convenience of the designer. Much of
the material is in tabular form to assist the designer in making the trade-offs best suited for a particular
application in a minimum amount of time.

Designers have used various approaches in arriving at suitable transformer and inductor designs. For
example, in many cases a rule of thumb used for dealing with current density is that a good working level is
1000 circular mils per ampere. This is satisfactory in many instances; however, the wire size used to meet
this requirement may produce a heavier and bulkier inductor than desired or required. The information

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 presented here will make it possible to avoid the use of this and other rules of thumb, and to develop a more
economical and better design. While other books are available on electronic transformers, none of them
seems to have been written with the user's viewpoint in mind. The material in this book is organized so that
the student engineer or technician—starting at the beginning of the book and continuing through the end—
will gain a comprehensive knowledge of the state of the art in transformer and inductor design.

No responsibility is assumed by the author or the publisher for any infringement of patent or other rights of
third parties that may result from the use of circuits, systems, or processes described or referred to in this
handbook.

Acknowledgments

In gathering the material for this book, I have been fortunate in having the assistance and cooperation of
several companies and many colleagues. I wish to express my gratitude to all of them. The list is too long
to mention them all. However, there are some individuals and companies whose contributions have been
especially significant. Colleagues who have retired from Magnetics include Robert Noah and Harry
Savisky, who helped so greatly with the editing of the final draft. Other contributions were made by my
colleagues at Magnetics, Lowell Bosley and his staff for sending up-to-date catalogs and sample cores. I
would like to thank colleagues at Micrometals Corp., Jim Cox and Dale Nicol, and George Orenchak of
TSC International. I would like to give special thanks to Richard (Oz) Ozenbaugh of Linear Magnetics
Corp. for his assistance in the detailed derivations of many of the equations and his efforts in checking the
design examples. I also give special thanks to Steve Freeman of Rodon Products, Inc., for building and
testing the magnetics components used in the design examples.

I am also grateful to: Dr. Vatche Vorperian of Jet Propulsion Laboratory (JPL) for his help in generating
and clarifying equations for the Quiet Converter; Jerry Fridenberg of Fridenberg Research, Inc., for
modeling circuits on his SPICE program; Dr. Gene Wester of JPL for his input; and Kit Sum for his
assistance in the energy-storage equations. I also thank the late Robert Yahiro for his help and
encouragement over the years.

Colonel Wm. T. McLyman

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 About the Author

Colonel Wm. T. McLyman recently retired as a Senior Member of the Avionics Equipment Section of the
Jet Propulsion Laboratory (JPL) affiliated with the California Institute of Technology in Pasadena,
California. He has 47 years of experience in the field of Magnetics, and holds 14 United States Patents on
magnetics-related concepts. Through his 30 years at JPL, he has written over 70 JPL Technical
Memorandums, New Technology Reports, and Tech-Briefs on the subject of magnetics and circuit designs
for power conversion. He has worked on projects for NASA including the Pathfinder Mission to Mars,
Cassini, Galileo, Magellan, Viking, Voyager, MVM, Hubble Space Telescope, and many others.

He has been on the lecture circuit for over 20 years speaking in the United States, Canada, Mexico, and
Europe on the design and fabrication of magnetic components. He is known as a recognized authority in
magnetic design. He is currently the President of his own company, Kg Magnetics, Inc., which specializes
in power magnetics design.

He recently completed a book entitled, High Reliability Magnetic Devices: Design and Fabrication
(Marcel Dekker, Inc.). He also markets, through Kg Magnetics, Inc., a magnetics design and analysis
software computer program called "Titan" for transformers and inductors (see Figure 1). This program
operates on Windows 95, 98, 2000, and NT.

iKG Maanehcs Main Menu " '-' ' '

Cadit — - • Analysis ;- - Spec III-
, AC Inductor Transformer Mag Amp. |; Coupled Ind )

OC Inductor DC Inductor Mag. Current i. Sinewave Inverter

: Transformer ACInductoi : Short CraA j' Current Transformer ,

•jr — --Spec! • . - —Spec II ;

; Continuous Discontinuous ! ; Magnetic Espwtions
j Forward Cow,
Boost Boost
j Three Ph8«« Resistor Trta
', Buck Buck
, Mag Amp. ii Corn Mode Ind
BucfoBoostlso. Buck/Boost Im
:
: OC Ind Resistance i Single Layer tnd
Inv. Buck/Boost Inv. Buck/Boost
Low Current Ind
.• i. _ .- , ,
^p^olaKy i v
I AfrCore i Rotary Transformer j •! •i • i
Core Conversiai 1 :
>

w. I

Figure 1. Computer Design Program Main Menu.

Colonel Wm. T. McLyman, (President)

Kg Magnetics, Inc.
Idyllwild, California 92549, U.S.A.
www.kgmagnetics.com; colonel@kgmagnetics.com

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Contents
Foreword Robert G. Noah
Preface
About the Author
Symbols
Chapter 1
Fundamentals of Magnetics
Chapter 2
Magnetic Materials and Their Characteristics
Chapter 3
Magnetic Cores
Chapter 4
Window Utilization, Magnet Wire, and Insulation
Chapter 5
Transformer Design Trade-Offs
Chapter 6
Transformer-Inductor Efficiency, Regulation, and Temperature Rise
Chapter 7
Power Transformer Design
Chapter 8
DC Inductor Design Using Gapped Cores
Chapter 9
DC Inductor Design Using Powder Cores
Chapter 10
AC Inductor Design
Chapter 11
Constant Voltage Transformer (CVT)
Chapter 12
Three-Phase Transformer Design
Chapter 13
Flyback Converter, Transformer Design
Chapter 14
Forward Converter, Transformer Design, and Output Inductor Design
Chapter 15
Input Filter Design

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Chapter 16
Current Transformer Design
Chapter 17
Winding Capacitance and Leakage Inductance
Chapter 18
Quiet Converter Design
Chapter 19
Rotary Transformer Design
Chapter 20
Planar Transformers
Chapter 21
Derivations for the Design Equations

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Symbols

a regulation, %
Ac effective cross section of the core, cm'2
Ap area product, cm4
At surface area of the transformer, cm'2
Aw wire area, cm2
AW(B) bare wire area, cm2
Aw(i) insulated wire area, cm
Awp primary wire area, cm2
Aws secondary wire area, cm
A-T amp turn
AWG American Wire Gage
B flux, tesla
Bac alternating current flux density, tesla
AB change in flux, tesla
Bdc direct current flux density, tesla
Bm flux density, tesla
Bmax maximum flux density, tesla
B0 operating peak flux density, tesla
Bpi( peak flux density, tesla
Br residual flux density, tesla
Bs saturation flux density, tesla
C capacitance
Cn new capacitance
Cp lumped capacitance
CM circular mils
DAWG wire diameter, cm
D(min) minimum duty ratio
D(max) maximum duty ratio
Dx dwell time duty ratio
E voltage
ELine line to line voltage
Ephase line to neutral voltage
Energy energy, watt-second

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 ESR equivalent series resistance
t| efficiency
f frequency, Hz
F fringing flux factor
Fm magneto-motive force, mmf
F.L. full load
G winding length, cm
y density, in grams-per-cm2
e skin depth, cm
H magnetizing force, oersteds
Hc magnetizing force required to return flux to zero, oersteds
AH delta magnetizing force, oersteds
H0 operating peak magnetizing force
Hs magnetizing force at saturation, oersteds
I current, amps
Ic charge current, amps
AI delta current, amps
Idc dc current, amps
Ijn input current, amps
ILine input line current, amps
Iphase input phase current, amps
Im magnetizing current, amps
I0 load current, amps
lo(max) maximum load current, amps
lo(min) minimum load current, amps
IP primary current, amps
Is secondary current, amps
Is(Fhase) secondary phase current, amps
Is(Line) secondary line current, amps
J current density, amps per cm2
KC copper loss constant
K<. quasi-voltage waveform factor
Ke electrical coefficient

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Kf waveform coefficient
Kg core geometry coefficient, cm
KJ constant related to current density
Ks constant related to surface area
Ku window utilization factor
Kup primary window utilization factor
Kus secondary window utilization factor
Kvoi constant related to volume
Kw constant related to weight
L inductance, henry
Lc open circuit inductance, henry
Lp primary inductance, henry
1 is a linear dimension
L(crt) critical inductance
X density, grams per cm3
lg gap, cm
lm magnetic path length, cm
lt total path length, cm
mks meters-kilogram-seconds
MLT mean length turn, cm
mmf magnetomotive force, Fm
MPL magnetic path length, cm
mW/g milliwatts-per-gram
|a permeability
Ui initial permeability
UA incremental permeability
um core material permeability
u0 permeability of air
ur relative permeability
ue effective permeability
n turns ratio
N turns
N.L. no load

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 NL inductor turns
Nn new turns
Np primary turns
Ns secondary turns
P watts
Pcu copper loss, watts
Pfe core loss, watts
Pg gap loss, watts
(|> magnetic flux
Pin input power, watts
PL inductor copper loss, watts
P0 output power, watts
Pp primary copper loss, warts
Ps secondary copper loss, watts
PX total loss (core and copper), watts
P, total apparent power, watts
PVA primary volt-amps
R resistance, ohms
Rac ac resistance, ohms
RCU copper resistance, ohms
Rdc dc resistance, ohms
Re equivalent core loss (shunt) resistance, ohms
Rg reluctance of the gap
Rm reluctance
Rmt total reluctance
RO load resistance, ohms
RO(R) reflected load resistance, ohms
Rm(equiv) reflected load resistance, ohms
Rp primary resistance, ohms
RR ac/dc resistance ratio
Rs secondary resistance, ohms
R, total resistance, ohms
p resistivity, ohm-cm

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 51 conductor area/wire area
52 wound area/usable window
83 usable window area/window area
S4 usable window area/usable window area + insulation area
Snp number of primary strands
Sns number of secondary strands
SVA secondary volt-amps
T total period, seconds
t0ff off time, seconds
ton on time, seconds
tw dwell time, seconds
Tr temperature rise, °C
U multiplication factor
VA volt-amps
Vac applied voltage, volts
Vc control voltage, volts
VC(pk) peak voltage, volts
Vd diode voltage drop, volts
V;n input voltage, volts
Vin(max) maximum input voltage, volts
Vjn(,nin) minimum input voltage, volts
Vn new voltage, volts
V0 output voltage, volts
Vp primary voltage, volts
Vp(rms) primary rms voltage, volts
secondary line to line voltage, volts
secondary line to neutral voltage, volts
Vr(pk) peak ripple voltage
Vs secondary voltage, volts
AVCc capacitor voltage, volts
AVcR capacitor ESR voltage, volts
AVp delta primary voltage, volts
AVS delta secondary voltage, volts

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 W watts
W/kg watts-per-kilogram
Wa window area, cm2
Wap primary window area, cm2
Was secondary window area, cm2
Wa(efi) effective window area, cm2
w-s watt-seconds
Wt weight, grams
Wtcu copper weight, grams
Wtfe iron weight, grams
XL inductive reactance, ohms

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Chapter 1

Fundamentals of Magnetics

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Introduction

Considerable difficulty is encountered in mastering the field of magnetics because of the use of so many
different systems of units - the centimeter-gram-second (cgs) system, the meter-kilogram-second (mks)
system, and the mixed English units system. Magnetics can be treated in a simple way by using the cgs
system. There always seems to be one exception to every rule and that is permeability.

Magnetic Properties in Free Space

A long wire with a dc current, I, flowing through it, produces a circulatory magnetizing force, H, and a
magnetic field, B, around the conductor, as shown in Figure 1-1, where the relationship is:

B = fi0H, [gauss]
H = ^—, [oersteds] H

1
Bm=—T, [gauss]
cm

Figure 1-1. A Magnetic Field Generated by a Current Carrying Conductor.

The direction of the line of flux around a straight conductor may be determined by using the "right hand
rule" as follows: When the conductor is grasped with the right hand, so that the thumb points in the
direction of the current flow, the fingers point in the direction of the magnetic lines of force. This is based
on so-called conventional current flow, not the electron flow.

When a current is passed through the wire in one direction, as shown in Figure l-2(a), the needle in the
compass will point in one direction. When the current in the wire is reversed, as in Figure l-2(b), the
needle will also reverse direction. This shows that the magnetic field has polarity and that, when the
current I, is reversed, the magnetizing force, H, will follow the current reversals.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Compass

(b)
Figure 1-2. The Compass Illustrates How the Magnetic Field Changes Polarity.

Intensifying the Magnetic Field

When a current passes through a wire, a magnetic field is set up around the wire. If the conductors, as
shown in Figure 1-3, carrying current in the same direction are separated by a relatively large distance, the
magnetic fields generated will not influence each other. If the same two conductors are placed close to each
other, as shown in Figure 1-4, the magnetic fields add, and the field intensity doubles.

B2
r [energy density] [1-1]

If the wire is wound on a dowel, its magnetic field is greatly intensified. The coil, in fact, exhibits a
magnetic field exactly like that of a bar magnet, as shown in Figure 1-5. Like the bar magnet, the coil has a
north pole and a neutral center region. Moreover, the polarity can be reversed by reversing the current, I,
through the coil. Again, this demonstrates the dependence of the magnetic field on the current direction.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Magnetic Field

Current Carrying Conductor

Large distance between conductors.
Figure 1-3. Magnetic Fields Produced Around Spaced Conductors.

Magnetic Field

Current Carrying Conductors

Conductors are in Close Proximity

Figure 1-4. Magnetic Fields Produced Around Adjacent Conductors.

The magnetic circuit is the space in which the flux travels around the coil. The magnitude of the flux is
determined by the product of the current, I, and the number of turns, N, in the coil. The force, NI, required
to create the flux is magnetomotive force (mmf). The relationship between flux density, B, and magnetizing
force, H, for an air-core coil is shown in Figure 1-6. The ratio of B to H is called the permeability, \i, and
for this air-core coil the ratio is unity in the cgs system, where it is expressed in units of gauss per oersteds,
(gauss/oersteds).

^:=l „ [1-2]
If the battery, in Figure 1-5, were replaced with an ac source, as shown in Figure 1-7, the relationship
between B and H would have the characteristics shown in Figure 1-8. The linearity of the relationship
between B and H represents the main advantage of air-core coils. Since the relationship is linear, increasing
H increases B, and therefore the flux in the coil, and, in this way, very large fields can be produced with
large currents. There is obviously a practical limit to this, which depends on the maximum allowable
current in the conductor and the resulting rise.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Magnetic Field \ / /
\
an
1r V V V —^ N
/
/, i VT7

tI
4

I = Direct Current Magnetic Path

Figure 1-5. Air-Core Coil with dc excitation

B, (tesla)

I I I I I I I I I I I H, ( oersteds)

Figure 1-6. Relationship Between B and H with dc Excitation.

'(\ i i\^' i ) \/
\\ \
\
1
v
» . \ -•! 000000
I.-.- —IT'- II - ' ' ' J' 11
*^ f /

Magnetic Field '/ ,

uuuuu
\\
i'' r
l\ \ \ V
\ —
I = Alternating Current Magnetic Path

Figure 1-7. Air-Core Coil Driven from an ac Source.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Fields of the order of 0.1 tesla are feasible for a 40° C temperature rise above room ambient temperature.
With super cooled coils, fields of 10 tesla have been obtained.

1—|—I 1 H, ( oersteds)

Figure 1-8. Relationship Between B and H with ac Excitation.

Simple Transformer
A transformer in its simplest form is shown in Figure 1-9. This transformer has two air coils that share a
common flux. The flux diverges from the ends of the primary coil in all directions. It is not concentrated or
confined. The primary is connected to the source and carries the current that establishes a magnetic field.
The other coil is open-circuited. Notice that the flux lines are not common to both coils. The difference
between the two is the leakage flux; that is, leakage flux is the portion of the flux that does not link both
coils.

Leakage Flux Mutual Flux

^ ^, ^ <^
i 1 . f Primary \ , Secondary »
\\ \ v _ n n n n 0nnn /y
~\^^-
_f^\\\
FiplH // ^" <^J
1 l /I (j \j u E^\J
\\ u u u
\\ ^\
—J) y
—*' /
I = Alternating Current
/- —\
f] \l (.) c^)

Figure 1-9. The Simplest Type of Transformer.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Magnetic Core
Most materials are poor conductors of magnetic flux; they have low permeability. A vacuum has a
permeability of 1.0, and nonmagnetic materials, such as air, paper, and copper have permeabilities of the
same order. There are a few materials, such as iron, nickel, cobalt, and their alloys that have high
permeability, sometimes ranging into the hundreds of thousands. To achieve an improvement over the air-
coil, as shown in Figure 1-10, a magnetic core can be introduced, as shown in Figure 1-11. In addition to its
high permeability, the advantages of the magnetic core over the air-core are that the magnetic path length
(MPL) is well-defined, and the flux is essentially confined to the core, except in the immediate vicinity of
the winding. There is a limit as to how much magnetic flux can be generated in a magnetic material before
the magnetic core goes into saturation, and the coil reverts back to an air-core, as shown in Figure 1-12.

\ <|> \ <+~~ Magnetic Flux \ I/

\ o

J I I I I H, (oersteds)

Figure 1-10. Air-Core Coil Emitting Magnetic Flux when Excited.

Magnetic Flux is Contained within Core Q

Magnetic Core H, (oersteds)

Figure 1-11. Introduction of a Magnetic Core.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Magnetic Flux
^ \ ^v \\ I [ / /

-^- >

\ Top View
B, (tesla) „ H

Magnetic Core

H, (oersteds)
Figure 1-12. Excited Magnetic Core Driven into Saturation.

Fundamental Characteristics of a Magnetic Core

The effect of exciting a completely demagnetized, ferromagnetic material, with an external magnetizing
force, H, and increasing it slowly, from zero, is shown in Figure 1-13, where the resulting flux density is
plotted as a function of the magnetizing force, H. Note that, at first, the flux density increases very slowly
up to point A, then, increases very rapidly up to point B, and then, almost stops increasing. Point B is
called the knee of the curve. At point C, the magnetic core material has saturated. From this point on, the
slope of the curve is:

= 1, [gauss/oersteds] [1-3]
H

The coil is now behaving as if it had an air-core. When the magnetic core is in hard saturation, the coil has
the same permeability as air, or unity. Following the magnetization curve in Figure 1-14, Figures 1-15
through Figures 1-16 show how the flux in the core is generated from the inside of the core to the outside
until the core saturates.

B, tesla
C

8
Q

i i i i i i i i H
Magnetizing Force
Figure 1-13. Typical Magnetization Curve.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 1= 0
Magnetic Core

Mean Magnetic Path Length

Winding

Figure 1-14. Magnetic Core with Zero Excitation.

Magnetic Core

Magnetic Field
at Low Excitation

o
Winding

Figure 1-15. Magnetic Core with Low Excitation.

Magnetic Core

Magnetic Field
at High Excitation

O
Winding

Figure 1-16. Magnetic Core with High Excitation.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Hysteresis Loop (B-H Loop)

An engineer can take a good look at the hysteresis loop and get a first order evaluation of the magnetic
material. When the magnetic material is taken through a complete cycle of magnetization and
demagnetization, the results are as shown in Figure 1-17. It starts with a neutral magnetic material,
traversing the B-H loop at the origin X. As H is increased, the flux density B increases along the dashed
line to the saturation point, Bs. When H is now decreased and B is plotted, B-H loop transverses a path to
Br, where H is zero and the core is still magnetized. The flux at this point is called remanent flux, and has a
flux density, Br.

The magnetizing force, H, is now reversed in polarity to give a negative value. The magnetizing force
required to reduce the flux Br to zero is called the coercive force, Hc. When the core is forced into
saturation, the retentivity, Brs, is the remaining flux after saturation, and coercivity, Hcs, is the magnetizing
force required to reset to zero. Along the initial magnetization curve at point X, the dashed line, in Figure
1-17, B increases from the origin nonlinearly with H, until the material saturates. In practice, the
magnetization of a core in an excited transformer never follows this curve, because the core is never in the
totally demagnetized state, when the magnetizing force is first applied.

The hysteresis loop represents energy lost in the core. The best way to display the hysteresis loop is to use
a dc current, because the intensity of the magnetizing force must be so slowly changed that no eddy currents
are generated in the material. Only under this condition is the area inside the closed B-H loop indicative of
the hysteresis. The enclosed area is a measure of energy lost in the core material during that cycle. In ac
applications, this process is repeated continuously and the total hysteresis loss is dependent upon the
frequency.

(tesla)

*- H
(oersteds)

Figure 1-17. Typical Hysteresis Loop.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Permeability
In magnetics, permeability is the ability of a material to conduct flux. The magnitude of the permeability at
a given induction is the measure of the ease with which a core material can be magnetized to that induction.
It is defined as the ratio of the flux density, B, to the magnetizing force, H. Manufacturers specify
permeability in units of gauss per oersteds.

Permeability = — ,
D
I [1-4]
H _ oersteds]

The absolute permeability, u0 in cgs units is unity 1 (gauss per oersteds) in a vacuum.

gauss |~ tesla
cgs: fja=l, (10*)
[oersteds |_ oersteds
[1-5]
s
mks: /J0 = 0.4^(10~
v ), —
' [ meter J

When B is plotted against H, as in Figure 1-18, the resulting curve is called the magnetization curve. These
curves are idealized. The magnetic material is totally demagnetized and is then subjected to gradually
increasing magnetizing force, while the flux density is plotted. The slope of this curve, at any given point
gives the permeability at that point. Permeability can be plotted against a typical B-H curve, as shown in
Figure 1-19. Permeability is not constant; therefore, its value can be stated only at a given value of B or H.

There are many different kinds of permeability, and each is designated by a different subscript on the
symbol u.

(i0 Absolute permeability, defined as the permeability in a vacuum.

Hi Initial permeability is the slope of the initial magnetization curve at the origin. It is measured
at very small induction, as shown in Figure 1-20.
UA Incremental permeability is the slope of the magnetization curve for finite values of peak-to-
peak flux density with superimposed dc magnetization as shown in Figure 1-21.
|ie Effective permeability. If a magnetic circuit is not homogeneous (i.e., contains an air gap), the
effective permeability is the permeability of hypothetical homogeneous (ungapped) structure
of the same shape, dimensions, and reluctance that would give the inductance equivalent to the
gapped structure.

Uj Relative permeability is the permeability of a material relative to that of free space.

un Normal permeability is the ratio of B/H at any point of the curve as shown in Figure 1-22.
umax Maximum permeability is the slope of a straight line drawn from the origin tangent to the
curve at its knee as shown in Figure 1-23.
UP Pulse permeability is the ratio of peak B to peak H for unipolar excitation.
um Material permeability is the slope of the magnetization curve measure at less than 50 gauss as
shown in Figure 1-24.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Permeability 1-13

B, tesla B, tesla

OJ 0)
Q Q
x x
J3
E E

I I I I I I I I H H
Magnetizing Force Magnetizing Force

Figure 1-18. Magnetizing Curve.

B, tesla

Permeability

0>
Q B-H Curve
g
ID
PH

1—1 H
Magnetizing Force
Figure 1-19. Variation of Permeability (j. along the Magnetizing Curve.

B, tesla

<u
Q

i i i i I i I I I
Magnetizing Force

Figure 1-20. Initial Permeability.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 H (oersteds)

Figure 1-21. Incremental Permeability.

B, tesla

a
u
Q
X!

'\ I I I I I I I I
Magnetizing Force

Figure 1-22. Normal Permeability.

B, tesla

Q
x
E

I I I I I I I I I
H
Magnetizing Force

Figure 1-23. Maximum Permeability.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 B, tesla

Q
s
x

'\ i i i i i i I i H
Magnetizing Force

Figure 1-24. Material Permeability.

Magnetomotive Force (mmf) and Magnetizing Force (H)

There are two force functions commonly encountered in magnetics: magnetomotive force, mmf, and
magnetizing force, H. Magnetomotive force should not be confused with magnetizing force; the two are
related as cause and effect. Magnetomotive force is given by the equation:

mmf = 0.4;rM, [gilberts] [1-6]

Where, N is the number of turns and I is the current in amperes. Whereas mmf is the force, H is a force
field, or force per unit length:

mmf [gilberts
H= • = oersteds [1-7]
MPL cm
Substituting,

H= , [oersteds]
L J [1-8]
MPL
Where, MPL = magnetic path length in cm.

If the flux is divided by the core area, Ac, we get flux density, B, in lines per unit area:

B = -£-, [gauss] [1-9]

The flux density, B, in a magnetic medium, due to the existence of a magnetizing force H, depends on the
permeability of the medium and the intensity of the magnetic field:

/jH, [gauss] [1-10]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The peak, magnetizing current, Inl, for a wound core can be calculated from the following equation:

H(MPL)
I = —^-'-, [amps] [1-11]

Where H0 is the field intensity at the peak operating point. To determine the magnetizing force, H0, use the
manufacturer's core loss curves at the appropriate frequency and operating flux density, B0, as shown in
Figure 1-25.

B (tesla)

DC
B
5,000 Hertz
10,000 Hertz

-H f-f- H (oersteds)

-B,

Figure 1-25. Typical B-H Loops Operating at Various Frequencies.

Reluctance

The flux produced in a given material by magnetomotive force (mmf) depends on the material's resistance
to flux, which is called reluctance, Rm. The reluctance of a core depends on the composition of the material
and its physical dimension and is similar in concept to electrical resistance. The relationship between mmf,
flux, and magnetic reluctance is analogous to the relationship between emf, current, and resistance, as
shown in Figure 1-26.

emf (£) = IR = Current x Resistance

[1-12]
mmf (fm) = ^Rm = Flux x Reluctance

A poor conductor of flux has a high magnetic resistance, Rm. The greater the reluctance, the higher the
magnetomotive force that is required to obtain a given magnetic field.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Current, I
Flux,

Magnetomotive Electromotive
Force, (mmf) Force, emf

Magnetic Core
Reluctance, Rm Resistance, R

Figure 1-26. Comparing Magnetic Reluctance and Electrical Resistance.

The electrical resistance of a conductor is related to its length 1, cross-sectional area Aw, and specific
resistance p, which is the resistance per unit length. To find the resistance of a copper wire of any size or
length, we merely multiply the resistivity by the length, and divide by the cross-sectional area:

R = —, [ohms] [1-13]

In the case of magnetics, 1/ia. is analogous to p and is called reluctivity. The reluctance Rm of a magnetic
circuit is given by:

4-=-^- t1-14!

Where MPL, is the magnetic path length, cm.

Ac is the cross-section of the core, cm .
ur is the permeability of the magnetic material.
ti0 is the permeability of air.

A typical magnetic core is shown in Figure 1-27 illustrating the magnetic path length MPL and the cross-
sectional area, Ac, of a C core.
Magnetic Core

Magnetic Path Length, (MPL)

Iron Cross-section, Ac

Figure 1-27. Magnetic Core Showing the Magnetic Path Length (MPL) and Iron Cross-section Ac.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Air Gap

A high permeability material is one that has a low reluctance for a given magnetic path length (MPL) and
iron cross-section, Ac. If an air gap is included in a magnetic circuit as shown in Figure 1-28, which is
otherwise composed of low reluctivity material like iron, almost all of the reluctance in the circuit will be at
the gap, because the reluctivity of air is much greater than that of a magnetic material. For all practical
purposes, controlling the size of the air gap controls the reluctance.

Magnetic Core

Magnetic Path Length, (MPL)

-*— Gap, L

Iron Cross-section, Ac

Figure 1-28. A Typical Magnetic Core with an Air Gap.

An example can best show this procedure. The total reluctance of the core is the sum of the iron reluctance
and the air gap reluctance, in the same way that two series resistors are added in an electrical circuit. The
equation for calculating the air gap reluctance, Rg, is basically the same as the equation for calculating the
reluctance of the magnetic material, Rm. The difference is that the permeability of air is 1 and the gap
length, lg, is used in place of the magnetic path length (MPL). The equation is as follows:

But, since uc = 1, the equation simplifies to:

[1-16]

Where:
lg is the gap length, cm.
Ac is the cross-section of the core, cm2.
u0 is the permeability of air.

The total reluctance, Rmt, for the core shown in Figure 1-28 is therefore:

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 mi m g

MPL [1-17]

Where ur is the relative permeability, which is used exclusively with magnetic materials.

_ n _ B gauss
[1-18]
pa //„//' L oersteds

The magnetic material permeability, unl, is given by:

Hm=Hrlio [1-19]

The reluctance of the gap is higher than that of the iron even when the gap is small. The reason is because
the magnetic material has a relatively high permeability, as shown in Table 1-1. So the total reluctance of
the circuit depends more on the gap than on the iron.

Table 1-1. Material Permeability

Material Permeability, \\,m

Material Name Permeability
Iron Alloys 0.8K to 25K
Ferrites 0.8K to 20K
Amorphous 0.8K to 80K

After the total reluctance, Rt, has been calculated, the effective permeability, u,e, can be calculated.

[1-20]
/, = / g + M P L
Where 1, is the total path length and u.e is the effective permeability.

[1-21]

Simplifying yields:

r^o t'o r^r

Then:

l g
_...". I
MPL_

[1-23]
/ S +MPL
He =
lg_ MPL
Ho HoHr

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 If 18 « MPL, multiply both sides of the equation by (u r u 0 MPL)/ ( u r u 0 MPL).

FL1-241
i^j

MPL
The classic equation is:

[1-25]

Introducing an air gap, lg, to the core cannot correct for the dc flux, but can sustain the dc flux. As the gap
is increased, so is the reluctance. For a given magnetomotive force, the flux density is controlled by the
gap.

Controlling the dc Flux with an Air Gap

There are two similar equations used to calculate the dc flux. The first equation is used with powder cores.
Powder cores are manufactured from very fine particles of magnetic materials. This powder is coated with
an inert insulation to minimize eddy currents losses and to introduce a distributed air gap into the core
structure.

The second equation is used, when the design calls for a gap to be placed in series with the magnetic path
length (MPL), such as a ferrite cut core, a C core, or butt stacked laminations.

[gauss]i
r [1 27]
-
'
Substitute (MPLum) /(MPLuJ for 1:

// //
[1-28]
1+ w * L +U
"" MPL MPLn '"'"MPL

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Then, simplify:

[1-29]
MPL
+ /„

MPL
[gauss] [1-30]
MPL MPL /
+/
<",„

Then, simplify:

[1-31]

Types of Air Gaps

Basically, there are two types of gaps used in the design of magnetic components: bulk and distributed.
Bulk gaps are maintained with materials, such as paper, Mylar, or even glass. The gapping materials are
designed to be inserted in series with the magnetic path to increase the reluctance, R, as shown in Figure 1-
29.

Magnetic Core

Magnetic Path Length, (MPL)

Gapping materials, such as:

paper, mylar, and glass.

Iron Cross-section, Ar

Figure 1-29. Placement of the Gapping Materials.

Placement of the gapping material is critical in keeping the core structurally balanced. If the gap is not
proportioned in each leg, then the core will become unbalanced and create even more than the required gap.
There are designs where it is important to place the gap in an area to minimize the noise that is caused by
the fringing flux at the gap. The gap placement for different core configurations is shown in Figure 1-30.
The standard gap placement is shown in Figure 1-30A, C, and D. The EE or EC cores shown in Figure 1-
3OB, are best-suited, when the gap has to be isolated within the magnetic assembly to minimize fringing
flux noise. When the gap is used as shown in Figure 1-30A, C, and D, then, only half the thickness of the
calculated gap dimension is used in each leg of the core.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Gap is across entire El surface. Gap is in the center leg.
Flux,
I i I
Gap, !„ = 1/2
o

Gap, !„ = 1
E-E & EC Type Cores
B

Flux, O

Gap, 1_ = 1/2

Toroidal Core C Core

C D

Figure 1-30. Gap Placement using Different Core Configurations.

Fringing Flux

Introduction

Fringing flux has been around since time began for the power conversion engineer. Designing power
conversion magnetics that produce a minimum of fringing flux has always been a problem. Engineers have
learned to design around fringing flux, and minimize its effects. It seems that when engineers do have a
problem, it is usually at the time when the design is finished and ready to go. It is then that the engineer will
observe something that was not recognized before. This happens during the final test when the unit
becomes unstable, the inductor current is nonlinear, or the engineer just located a hot spot during testing.
Fringing flux can cause a multitude of problems. Fringing flux can reduce the overall efficiency of the
converter, by generating eddy currents that cause localized heating in the windings and/or the brackets.
When designing inductors, fringing flux must to be taken into consideration. If the fringing flux is not
handled correctly, there will be premature core saturation. More and more magnetic components are now
designed to operate in the sub-megahertz region. High frequency has really brought out the fringing flux
and its parasitic eddy currents. Operating at high frequency has made the engineer very much aware of
what fringing flux can do to hamper a design.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Material Permeability,
The B-H loops that are normally seen in the manufacturers' catalogs are usually taken from a toroidal
sample of the magnetic material. The toroidal core, without a gap, is the ideal shape to view the B-H loop
of a given material. The material permeability, um, will be seen at its highest in the toroidal shape, as shown
in Figure 1-31.

B (tesla)

Normal B-H Loop

Sheared B-H Loop

Figure 1-31. The Shearing of an Idealized B-H Loop Due to an Air Gap.

A small amount of air gap, less than 25 microns, has a powerful effect by shearing over the B-H loop. This
shearing over of the B-H loop reduces the permeability. High permeability ferrites that are cut, like E cores,
have only about 80 percent of the permeability, than that of a toroid of the same material. This is because
of the induced gap, even though the mating surfaces are highly polished. In general, magnetic materials
with high-permeability, are sensitive to temperature, pressure, exciting voltage, and frequency. The
inductance change is directly proportional to the permeability change. This change in inductance will have
an effect on the exciting current. It is very easy to see, that inductors that are designed into an LC, tuned
circuit, must have a stable permeability, ue.

2
Ac A/u (l
L= [henrys] [1-32]
MPL

Air Gaps

Air gaps are introduced into magnetic cores for a variety of reasons. In a transformer design a small air
gap, lg, inserted into the magnetic path, will lower and stabilize the effective permeability, ue.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 [1-33]

'(MPL )

This will result in a tighter control of the permeability change with temperature, and exciting voltage.
Inductor designs will normally require a large air gap, lg, to handle the dc flux.

0.47r/V/ rfr (l(r 4 )

I _ _ ^ [cm] [1-34]

Whenever an air gap is inserted into the magnetic path, as shown in Figure 1-32, there is an induced,
fringing flux at the gap.

Core
I V ! t !t
T i*!Ti

Core

Minimum Gap Small Gap Large Gap

Figure 1-32. Fringing Flux at the Gap.

The fringing flux effect is a function of gap dimension, the shape of the pole faces, and the shape, size, and
location of the winding. Its net effect is to shorten the air gap. Fringing flux decreases the total reluctance
of the magnetic path and, therefore, increases the inductance by a factor, F, to a value greater than the one
calculated.

Fringing Flux, F

Fringing flux is completely around the gap and re-enters the core in a direction of high loss, as shown in
Figure 1-33. Accurate prediction of gap loss, Pg, created by fringing flux is very difficult to calculate.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Fringing flux

Figure 1-33. Fringing Flux, with High Loss Eddy Currents.

This area around the gap is very sensitive to metal objects, such as clamps, brackets and banding materials.
The sensitivity is dependent on the intensity of the magnetomotive force, gap dimensions and the operating
frequency. If a metal bracket or banding material is used to secure the core, and it passes over the gap, two
things can happen: (1) If the material ferromagnetic is placed over the gap, or is in close proximity so it
conducts the magnetic field, this is called "shorting the gap." Shorting the gap is the same as reducing the
gap dimension, thereby producing a higher inductance, than designed, and could drive the core into
saturation. (2) If the material is metallic, (such as copper, or phosphor bronze), but not ferromagnetic, it
will not short the gap or change the inductance. In both cases, if the fringing flux is strong enough, it will
induce eddy currents that will cause localized heating. This is the same principle used in induction heating.

Gapped, dc Inductor Design

The fringing flux factor, F, has an impact on the basic inductor design equations. When the engineer starts
a design, he or she must determine the maximum values for Bdc and for Bac, which will not produce
magnetic saturation. The magnetic material that has been selected will dictate the saturation flux density.
The basic equation for maximum flux density is:

, [tesla] [1-35]
MPL

The inductance of an iron-core inductor, carrying dc and having an air gap, may be expressed as:

[henrys] [1-36]
MPL

The inductance is dependent on the effective length of the magnetic path, which is the sum of the air gap
length, lg, and the ratio of the core magnetic path length to the material permeability, (MPL/um). The final
determination of the air gap size requires consideration of the fringing flux effect which is a function of the

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 gap dimension, the shape of the pole faces, and the shape, size, and location of the winding. The winding
length, or the G dimension of the core, has a big influence on the fringing flux. See, Figure 1-34 and
Equation 1-37.

/ \ «*-
~\ 1
\ F
1 J>

E
1 '
1
1
T a
V J

* r * D
G D

Figure 1-34. Dimensional, Call Out for C and E Cores.

The fringing flux decreases the total reluctance of the magnetic path length and, therefore, increases the
inductance by a factor of F to a value greater than that calculated. The fringing flux factor is:

[1.37]

After the inductance has been calculated using Equation 1-36, the fringing flux factor has to be incorporated
into Equation 1-36. Equation 1-36 can now be rewritten to include the fringing flux factor, as shown:

L =F , [henrys] [1-38]

The fringing flux factor, F, can now be included into Equation 1-35. This will check for premature, core
saturation.

, [tesla] [1-39]
MPL

Now that the fringing flux factor, F, is known and inserted into Equation 1-38. Equation 1-38 can be
rewritten to solve for the required turns so that premature core saturation will not happen.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 [turns] [1-40]
[o.47rA c F(lCr 8 )'

Fringing Flux and Coil Proximity

As the air gap increases, the fringing flux will increase. Fringing flux will fringe out away from the gap by
the distance of the gap. If a coil was wound tightly around the core and encompasses the gap, the flux
generated around the magnet wire will force the fringing flux back into the core. The end result will not
produce any fringing flux at all, as shown in Figure 1-35. As the coil distance moves away from the core,
the fringing flux will increase until the coil distance from the core is equal to the gap dimension.

Fringing Flux No Fringing Flux

Bobbin/Coil Form Magnetic Wire

Figure 1-35. Comparing a Tightly- Wound Coil, and a Coil Wound on a Coil Form.

Fringing Flux, Crowding

Flux will always take the path of highest permeability. This can best be seen in transformers with interleave
laminations. The flux will traverse along the lamination until it meets its mating, I or E. At this point, the
flux will jump to the adjacent lamination and bypass the mating point, as shown in Figure 1-36.
Laminations E and I

Flux Crowding

Minute Air Gap —&- ~=. j | ^3 I f rr

Flux

Interleave 1 x 1

Figure 1-36. Flux Crowding in Adjacent Laminations

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 This phenomena can best be seen by observing the exciting current at low, medium and high flux levels, as
shown in Figure 1-37. At low levels of excitation, the exciting current is almost square, due to the flux
taking the high permeability path, by jumping to the adjacent lamination, as shown in Figure 1-36. As the
excitation is increased, the adjoining lamination will start to saturate, and the exciting current will increase
and become nonlinear. When the adjacent lamination approaches saturation, the permeability drops. It is
then that the flux will go in a straight line and cross the minute air gap, as shown in Figure 1-36.

I, Excitation

Low Flux Medium Flux High Flux

Figure 1-37. Exciting Current, at Different Levels of Flux Density, B.

Fringing Flux and Powder Cores

Designing high frequency converters, using low permeability powder cores, will usually require very few
turns. Low perm power cores (less than 60), exhibit fringing flux. Powder cores with a distributed gap will
have fringing flux that shorts the gap and gives the impression of a core with a higher permeability.
Because of the fringing flux and a few turns, it is very important to wind uniformly and in a consistent
manner. This winding is done to control the fringing flux and get inductance repeatability from one core to
another, as shown in Figures 1-38 and 1-39.

Evenly Spaced Winding Random Wound

Fringing Flux

Winding Powder Cores Winding

Figure 1-38. Comparing Toroidal, Winding Methods.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Winding Evenly Wound Winding Randomly Wound
i Y'"'V 'V 'V 'Y"%Y'"''i /' r V'"'Y " Y""V'"'\

Bobbin Fringing Flux

'-....•''....•'"^L A...." '' *.. A A.

Powder EE Cores
Figure 1-39. Comparing EE Cores, Winding Methods.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Chapter 2

Magnetic Materials and Their Characteristics

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Introduction

The magnetic material is the paramount player in the design of magnetic components. The magnetics
design engineer has three standard words when making the normal design trade-off study: cost, size, and
performance. He will be happy to stuff any two into the bag. The magnetics engineer is now designing
magnetic components that operate from below the audio range to the megahertz range. He is normally
asked to design for maximum performance, with the minimum of his parasitic friends' capacitance and
leakage inductance. Today, the magnetic materials the engineer has to work with are silicon steel, nickel
iron (permalloy), cobalt iron (permendur), amorphous metallic alloys, and ferrites. These also have spin-off
material variants, such as moly-permalloy powder, sendust powder, and iron powder cores. From this group
of magnetic materials, the engineer will make trade-offs with the magnetic properties for his design. These
properties are: saturation Bs, permeability u, resistivity p (core loss), remanence Br, and coercivity Hc.

Saturation

A typical hysteresis loop of a soft magnetic material is shown in Figure 2-1. When a high magnetizing
force is encountered, a point is reached where further increase in, H, does not cause useful increase in, B.
This point is known as the saturation point of that material. The saturation flux density, Bs, and the required
magnetizing force, Hs, to saturate the core are shown with dashed lines.

Figure 2-1. Typical B-H or Hysteresis Loop of a Soft Magnetic Material.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Remanence Flux, Br, and Coercivity Hc
In Figure 2-1 the hysteresis loop clearly shows the remanence flux density, Br. The remanence flux is the
polarized flux remaining in the core after the excitation has been removed. The magnetizing force, -Hc, is
called coercivity. It is the amount of magnetizing force required to bring the remanence flux density back to
zero.

Permeability, fi

The permeability of a magnetic material is a measure of the ease in magnetizing the material. Permeability,
u, is the ratio of the flux density, B, to the magnetizing force, H.

= — , [permeability] [2-1]
H

The relationship between B and H is not linear, as shown in the hysteresis loop in Figure 2-1. Then, it is
evident that the ratio, B/H, (permeability), also varies. The variation of permeability with flux density, B, is
shown in Figure 2-2. Also, it shows the flux density at which the permeability is at a maximum.

u = Permeability

Magnetizing Force

Figure 2-2. Variation in Permeability n with B and H.

Hysteresis Loss, Resistivity, p, (core loss)

The enclosed area within the hysteresis, shown in Figure 2-1, is a measure of the energy lost in the core
material during that cycle. This loss is made up in two components: (1) the hysteresis loss and (2) eddy
current loss. The hysteresis loss is the energy loss when the magnetic material is going through a cycling
state. The eddy current loss is caused when the lines of flux pass through the core, inducing electrical
currents in it. These currents are called eddy currents and produce heat in the core. If the electrical
resistance of the core is high, the current will be low; therefore, a feature of low-loss material is high
electrical resistance. In the norm, when designing magnetic components, the core loss is a major design

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 factor. Core loss can be controlled by selecting the right material and thickness. Selecting the correct
material, and operating within its limits, will prevent overheating that could result in damage to the wire
insulation and/or the potting compound.

Introduction to Silicon Steel

Silicon steel was one of the first alloys to be used in transformers and inductors. It has been greatly
improved over the years and is probably, pound for pound, the most, widely used magnetic material. One
of the drawbacks in using steel in the early years was, as the material became older, the losses would
increase. With the addition of silicon to the steel, the advantages were twofold: it increased the electrical
resistivity, therefore reducing the eddy current losses, and it also improved the material's stability with age.

Silicon steel offers high saturation flux density, a relatively good permeability at high flux density, and a
moderate loss at audio frequency. One of the important improvements made to the silicon steel was in the
process called cold-rolled, grain-oriented, AISI type M6. This M6 grain-oriented steel has exceptionally
low losses and high permeability. It is used in applications requiring high performance and the losses will
be at a minimum.

Introduction to Thin Tape Nickel Alloys

High permeability metal alloys are based primarily on the nickel-iron system. Although Hopkinson
investigated nickel-iron alloys as early as 1889, it was not until the studies by Elmen, starting in about 1913,
on properties in weak magnetic fields and effects of heat-treatments, that the importance of the Ni-Fe alloys
was realized. Elmen called his Ni-Fe alloys, "Permalloys," and his first patent was filed in 1916. His
preferred composition was the 78Ni-Fe alloy. Shortly after Elmen, Yensen started an independent
investigation that resulted in the 50Ni-50Fe alloy, "Hipernik," which has lower permeability and resistivity
but higher saturation than the 78-Permalloy, (1.5 tesla compared to 0.75 tesla), making it more useful in
power equipment.

Improvements in the Ni-Fe alloys were achieved by high temperature anneals in hydrogen atmosphere, as
first reported by Yensen. The next improvement was done by using grain-oriented material and annealing
it, in a magnetic field, which was also in a hydrogen atmosphere. This work was done by Kelsall and
Bozorth. Using these two methods, a new material, called Supermalloy, was achieved. It has a higher
permeability, a lower coercive force, and about the same flux density as 78-Permalloy. Perhaps the most
important of these factors is the magnetic anneal, which, not only increases permeability, but also provides a
"square" magnetization curve, important in high frequency power conversion equipment.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 In order to obtain high resistance, and therefore lower core losses for high frequency applications, two
approaches have been followed: (1) modification of the shape of metallic alloys and (2) development of
magnetic oxides. The result was the development of thin tapes and powdered alloys in the 1920's, and thin
films in the 1950's. The development of thin film has been spurred by the requirements of aerospace, power
conversion electronics from the mid 1960's to the present.

The Ni-Fe alloys are available in thicknesses of 2 mil, 1 mil, 0.5 mil, 0.25 and 0.125 mil. The material
comes with a round or square B-H loop. This gives the engineer a wide range of sizes and configurations
from which to select for a design. The iron alloy properties for some of the most popular materials are
shown in Table 2-1. Also, given in Table 2-1, is the Figure number for the B-H loop of each of the
magnetic materials.

Table 2-1 Magnetic Properties for Selected Iron Alloys Materials.

Iron Alloy Material Properties

Material Initial Flux Density Curie dc, Coercive Density Weight Typical
Name Composition Permeability Tesla Temp. Force, He grams/cm Factor B-H Loop

Hi Bs °C Oersteds 8 X Figures
Silicon 3% Si 1.5K. 1.5-1.8 750 0.4-0.6 7.63 1.000 (2-3)
97% Fe

Supermendur* 49% Co 0.8 K 1.9-2.2 940 0.15-0.35 8.15 1.068 (2-4)

49% Fe
2%V
Orthonol 50% Ni 2K 1.42-1.58 500 0.1-0.2 8.24 1.080 (2-5)
50% Fe

Permalloy 79% Ni 12K-100K 0.66-0.82 460 0.02-0.04 8.73 1.144 (2-6)

17%Fe
4% Mo
Supermalloy 78% Ni 10K-50K 0.65-0.82 460 0.003-0.008 8.76 1.148 (2-7)
17%Fe
5% Mo
* Field Anneal.
x Silicon has unity weight factor.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 B, t ssla
1.6-

1.2-
Magnesil
DC Hysteresis Loop
0.8- (
0.4-
- 0.4 0.8 1.2 1
i
1.6 1.2 0.8 .4 - H, oersteds
- 0.4

1 " 0.8

^/ : ~ 1.2

1.6

Figure 2-3. Silicon B-H Loop: 97% Fe 3% Si.

1.6
Supermendur 1.2
DC Hysteresis Loop
0.8

0.4
0.4 0.8 1.2 1.6

1.6 1.2 0.8 0.4 H, oersteds

0.4

0.8

1.2

1.6

Figure 2-4. Supermendur B-H Loop: 49% Fe 49% Co 2% V.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Orthonol
DC Hysteresis Loop 0.8

0.4
0.2 0.4 0,6 , 0.*
0.8 0.6 0.4 0.:
H, oersteds

Figure 2-5. Orthonol B-H loop: 50% Fe 50% Ni.

B, tesla
0.8-r

Square Permalloy 80 j - -
DC Hysteresis Loop Of 4 - -

10.04 0.08 0.12 0.16

Figure 2-6. Square Permalloy 80 B-H loop: 79% Ni 17% Fe 4% Mo.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 B, tesla
0.8 -r

Supermalloy
DC Hysteresis Loop

0.04 0.08 0.12 0.16

Figure 2-7. Supermalloy B-H Loop: 78% Ni 17% Fe 5% Mo.

Introduction to Metallic Glass

The first synthesis of a metallic glass drawing wide attention among material scientists, occurred in 1960.
Klement, Willens and Duwez reported that a liquid, AuSi alloy, when rapidly quenched to liquid nitrogen
temperature, would form an amorphous solid. It was twelve years later that Chen and Polk produced
ferrous-based metallic glasses in useful shapes with significant ductility. Metallic glasses have since
survived the transition from laboratory curiosities to useful products, and currently are the focus of intensive
technological and fundamental studies.

Metallic glasses are generally produced, by liquid quenching, in which a molten metal alloy is rapidly
cooled, at rates on the order of 105 degrees/sec., through the temperature at which crystallization normally
occurs. The basic difference between crystalline (standard magnetic material) and glassy metals is in their
atomic structures. Crystalline metals are composed of regular, three-dimensional arrays of atoms which
exhibit long-range order. Metallic glasses do not have long-range structural order. Despite their structural
differences, crystalline and glassy metals of the same compositions exhibit nearly the same densities.

The electrical resistivities of metallic glasses are much larger, (up to three times higher), than those of
crystalline metals of similar compositions. The magnitude of the electrical resistivities and their
temperature coefficients in the glassy and liquid states are almost identical.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Metallic glasses are quite soft magnetically. The term, "soft," refers to a large response of the
magnetization to a small-applied field. A large magnetic response is desirable in such applications as
transformers and inductors. The obvious advantages of these new materials are in high frequency
applications with their high induction, high permeability and low core loss.

There are four amorphous materials that have been used in high frequency applications: 2605SC, 2714A,
2714AF and Vitroperm 500F. Material 2605SC offers a unique combination of high resistivity, high
saturation induction, and low core loss, making it suitable for designing high frequency dc inductors.
Material 2714A is a cobalt material that offers a unique combination of high resistivity, high squareness
ratio Br/Bs, and very low core loss, making it suitable for designing high frequency aerospace transformers
and mag-amps. The Vitroperm 500F is an iron based material with a saturation of 1.2 tesla and is well-
suited for high frequency transformers and gapped inductors. The high frequency core loss for the
nanocrystal 500F is lower than some ferrite, even operating at a high flux density. The amorphous
properties for some of the most popular materials are shown in Table 2-2. Also, given in Table 2-2, is the
Figure number for the B-H loop of each of the magnetic materials.

Table 2-2. Magnetic Properties for Selected Amorphous Materials.

Amorphous Material Properties

Material Major Initial Flux Density Curie dc, Coercive Density Weight Typical
Name Composition Permeability Tesla Temperature Force, He grams/cm" Factor B-H Loop

Mi Bs °C Oersteds 5 X Figures
2605 SC 81%Fe 1.5K 1.5-1.6 370 0.4-0.6 7.32 0.957 (2-8)
13.5%B
3.5% Si
2714A 66% Co 0.8K 0.5-0.65 205 0.15-0.35 7.59 0.995 (2-9)
15% Si
4% Fe
2714AF 66% Co 2K 0.5-0.65 205 0.1-0.2 7.59 0.995 (2-10)
15% Si
4% Fe
Nanocrystal 73.5% Fe 30K-80K 1.0-1.2 460 0.02-0.04 7.73 1.013 (2-11)
Vitroperm 500F* l%Cu
15.5% Si
* Vitroperm is the trademark of Vacuumschmelze.
x Silicon has a unity weight factor. See Table 2-1.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 B, Tesla
1.6

Metglas Type 2605SC 1.2

DC Hysteresis Loop
0.8

0.4
H, oersted
H—h H h- H 1 1 1 1 1—
0.6 0.4 0.2 0.2 0.4 0.6
O 4 H, oersted

0.8

1.2

-L 1.6

Figure 2-8. Amorphous 2605SC B-H Loop: 81% Fe 13.5% B 3.5% Si.

B, Tesla
0.6
0.5 t . f
Metglas Type 271 4 A 0.4
DC Hysteresis Loop
0.3
0.2 -
0.1
i i 0.01 0.03 0.05
i i u ... . 1 |1 1 1
1 |1 1t

0.05 0.03 0.01 . 0 1 H, oersted

- 0.2
- 0.3
- 0.4
J ) 0.5
- 0.6

Figure 2-9. Amorphous 2714A B-H Loop: 66% Co 15% Si 4% Fe.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Metglas Type 2714AF
DC Hysteresis Loop

Figure 2-10. Amorphous 2714AF B-H Loop: 66% Co 15% Si 4% Fe.

Vitroperm SOOF
10 Hz

Figure 2-11. Vitroperm SOOF B-H loop: 73.5% Fe 15.5% Si 1% Cu.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Introduction to Soft Ferrites

In the early days of electrical industry, the need for the indispensable magnetic material was served by iron
and its magnetic alloys. However, with the advent of higher frequencies, the standard techniques of
reducing eddy current losses, (using laminations or iron powder cores), was no longer efficient or cost
effective.

This realization stimulated a renewed interest in "magnetic insulators," as first reported by S. Hilpert in
Germany, in 1909. It was readily understood that, if the high electrical resistivity of oxides could be
combined with desired magnetic characteristics, a magnetic material that was particularly well-suited for
high frequency operation would result.

Research to develop such a material was being performed by scientists in various laboratories all over the
world, such as V. Kato, T. Takei, and N. Kawai in the 1930's in Japan, and by J. Snoek of the Philips'
Research Laboratories in the period 1935-1945 in The Netherlands. By 1945, Snoek had laid down the
basic fundamentals of the physics and technology of practical ferrite materials. In 1948, the Neel Theory of
ferromagnetism provided the theoretical understanding of this type of magnetic material.

Ferrites are ceramic, homogeneous materials composed of oxides; iron oxide is their main constituent. Soft
ferrites can be divided into two major categories; manganese-zinc and nickel-zinc. In each of these
categories, changing the chemical composition, or manufacturing technology, can manufacture many
different Mn-Zn and Ni-Zn material grades. The two families of Mn-Zn and Ni-Zn ferrite materials
complement each other, and allow the use of soft ferrites from audio frequencies to several hundred
megahertz. Manufacturers do not like to handle manganese-zinc in the same area, or building with nickel-
zinc, because one contaminates the other, which leads to poor performance yields. The basic difference
between Manganese-Zinc and Nickel-Zinc is shown in Table 2-3. The biggest difference is Manganese-
Zinc has a higher permeability and Nickel-Zinc has a higher resistivity. Shown in Table 2-4 are some of the
most popular ferrite materials. Also, given in Table 2-4, is the Figure number for the B-H loop of each of
the materials.

Table 2-3. Comparing Manganese-Zinc and Nickel-Zinc Basic Properties.

Basic Ferrite Material Properties

Initial Flux Density Curie dc, Coercive
Materials Permeability p
D Temperature, Force, Hc Resistivity
max
Hi Tesla °C Oersteds Q-cm
Manganese Zinc 750-15 K 0.3-0.5 100-300 0.04-0.25 10-100
Nickel Zinc 15-1500 0.3-0.5 150-450 0.3-0.5 106

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Manganese-Zinc Ferrites
This type of soft ferrite is the most common, and is used in many more applications than the nickel-zinc
ferrites. Within the Mn-Zn category, a large variety of materials are possible. Manganese-zinc ferrites are
primarily used at frequencies less than 2 MHz.

Nickel-Zinc Ferrites
This class of soft ferrite is characterized by its high material resistivity, several orders of magnitude higher
than Mn-Zn ferrites. Because of its high resistivity, Ni-Zn ferrite is the material of choice for operating
from 1-2 MHz to several hundred megahertz.

The material permeability, um, has little influence on the effective permeability, ue, when the gap dimension
is relatively large, as shown in Table 2-5.

Table 2-4. Magnetic Properties for Selected Ferrite Materials.

Ferrites Material Properties

*Magnetics Initial Flux Density Residual Flux Curie dc, Coercive Density Typical
Material Permeability Tesla Tesla Temperature Force, He grams/cm B-H Loop

Name m B s @15Oe Br °C Oersteds 8 Figures

K 1500 0.48T 0.08T >230 0.2 4.7 (2-12)

R 2300 0.50T 0.1 2T >230 0.18 4.8 (2-13)

P 2500 0.50T 0.1 2T >230 0.18 4.8 (2-13)

F 5000 0.49T 0.1 OT >250 0.2 4.8 (2-14)

W 10,000 0.43T 0.07T >125 0.15 4.8 (2-15)

H 15,000 0.43T 0.07T >125 0.15 4.8 (2-15)

*Magnetics, a Division of Spang & Company

Table 2-5. Permeability, and its Effect on Gapped Inductors.

Comparing Material Permeabilities

*Material Um Gap, inch Gap, cm **MPL, cm "e
K 1500 0.04 0.101 10.4 96
R 2300 0.04 0.101 10.4 98
P 2500 0.04 0.101 10.4 99
F 3000 0.04 0.101 10.4 100
*The materials are from Magnetics, a Division of Spang and Company
**Core , ETD44

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Bm, Tesla

K Material

0.5 l.O 1.5 2.0 2.5

25 °C Bm = 0460T ig IS oersted
100 °C Bm = 0.350T @ 15 oersted

Figure 2-12. Ferrite B-H loop, K Material at 25 and 100 ° C.

P & R Material

25 °C Bm - 0 500T @ 15
0.2 oersted
100 °C Bm = 0.375T @ 15 oersted

Figure 2-13. Ferrite B-H loop, P & R Material at 25 and 100 ° C.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Bm, Tesla
0.5 T

F Material

0.5 1.0 1.5 2.0 2.5

—i 1 \ 1 1
H, oersted

25 °C Bm = 0.490T @ 15 oersted
100 "C Bm = 0.340T @ 15 oersted

Figure 2-14. Ferrite B-H loop, F Material at 25 and 100 ° C.

Bm, Tesla

W & H Material

1.0 0.8 0.6 0.4 0.2 0.2 0.4 0.6 0.8 1.0

25 °C Bm = 0.430T @ 15 oersted
10
0.2 ° °C Bm = 0.220T @ 15 oersted

0.3

0.4

Figure 2-15. Ferrite B-H loop, W & H Material at 25 and 100 ° C.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Ferrite Cross Reference

The cross-reference, Table 2-6, has been put together using some of the leading ferrite manufacturers. The
ferrite materials have been organized and referenced under Magnetics materials. This is because Magnetics
has one of the broadest lines of standard ferrite materials.

Table 2-6. Ferrite Materials, Manufacturers' Cross-Reference.

Ferrite Material Cross Reference

Permeability 1500 2300 2500 3000 5000 10000 15000
Application Power Power Power Power Filter Filter Filter
Manufacturer's Material Designation
Magnetics 1 K 2 R 3 P F J W H
Ferroxcube 3F35 3F3 3C94 3C81 3E27 3E5 3E7
Permeability, U; 1400 2000 2300 2700 6000 10000 15000
Fair-Rite 78 75 76
Permeability, Uj 2300 5000 10000
Siemens N49 N87 N67 T41 T35 T38 T46
Permeability, u^ 1300 2200 2100 3000 6000 10000 15000
TDK Corp. PC50 PC40 PC44 H5A HP5 H5C2 H5C3
Permeability, u( 1400 2300 2400 3300 5000 10000 15000
MMG F44 F5 F5C F-10 F-39
Permeability, u< 1900 2000 3000 6000 10000
Ceramic Mag MN67 MN80 MN80 MN8CX MN60 MC25 MC15K
Permeability, Uj 1000 2000 2000 3000 6000 10000 15000
Tokin HBM B25 B3100 H5000 HI 2000
Permeability, u,
Ferrite Int. TSF-5099 TSF-7099 TSF-7070 TSF-8040 TSF-5000 TSF-010K
Permeability, u; 2000 2000 2200 3100 5000 10000
1 . High Frequency power material 250 kHz & up.
2. Lowest loss at 80°-100°C, 25 kHz to 250 kHz.
3. Lowest loss at 60°C-80°C.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Introduction to Molypermalloy Powder Cores

The nickel-iron (Ni-Fe) high permeability magnetic alloys (permalloy) were discovered in 1923 and 1927.
Permalloy alloys were successfully used in powder cores, greatly contributing to the carrier wave
communications of the time.

In the early 1940's, a new material, trademarked Molybdenum Permalloy Powder (MPP), was developed
into cores by the Bell Telephone Laboratory and the Western Electric Company. This new material was
developed for loading coils, filtering coils, and transformers at audio and carrier frequencies in the
telephone facility. The use of such cores has been extended to many industrial and military circuits. The
stability of permeability and core losses with time, temperature, and flux level, are particularly important to
engineers designing tuned circuits and timing circuits. This new material has given reliable and superior
performance over all past powder core materials.

Molybdenum permalloy powder, [2 Molybdenum (Mo)-82 Nickel (Ni)-16 Iron (Fe)], is made by grinding
hot-rolled and embrittled cast ingots; then, the alloy is insulated and screened to a fineness of 120 mesh for
use in audio frequency applications, and 400 mesh for use at high frequencies.

In the power conversion field, the MPP core has made its greatest impact in switching power supplies. The
use of MPP cores and power MOSFET transistors has permitted increased frequency, resulting in greater
compactness and weight reduction in computer systems. The power supply is the heart of the system.
When the power supply is designed correctly, using a moderate temperature rise, the system will last until it
becomes obsolete. In these power systems there are switching inductors, smoothing choke coils, common
mode filters, input filters, output filters, power transformers, current transformers and pulse transformers.
They cannot all be optimally designed, using MPP cores. But, in some cases, MPP cores are the only ones
that will perform in the available space with the proper temperature rise.

Introduction to Iron Powder Cores

The development of compressed iron powder cores as a magnetic material for inductance coils, stemmed
from efforts of Bell Telephone Laboratory engineers to find a substitute for fine iron-wire cores. The use of
iron powder cores was suggested by Heaviside, in 1887, and again, by Dolezalek in 1900.

The first iron powder cores of commercially valuable properties were described by Buckner Speed, in U.S.
Patent No. 1274952, issued in 1918. Buckner Speed and G.W. Elman published a paper in the A.I.E.E.
Transactions, "Magnetic Properties of Compressed Powdered Iron," in 1921. This paper describes a

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 magnetic material, which is well-suited to the construction of cores in small inductance coils and
transformers, such as those used in a telephone system. These iron powder cores were made from 80 Mesh
Electrolytic Iron Powder. The material was annealed, then, insulated by oxidizing the surface of the
individual particles. In this way, a very thin and tough insulation of grains of iron was obtained; this did not
break down when the cores were compressed. A shellac solution was applied to the insulated powder as a
further insulator and binder. This was how toroidal iron powder cores were manufactured by Western
Electric Company until about 1929. Today's iron powder cores are manufactured in much the same way,
using highly pure iron powder and a more exotic insulator and binder. The prepared powder is compressed
under extremely high pressures to produce a solid-looking core. This process creates a magnetic structure
with a distributed air-gap. The inherent high saturation flux density of iron, combined with the distributed
air-gap, produces a core material with initial permeability of less than 100, and with high-energy storage
capabilities.

The dc current does not generate core loss, but an ac or ripple current does generate core loss. Iron powder
material has higher core loss than some other, more expensive, core materials. Most dc-biased inductors
have a relatively small percentage of ripple current and, thus, core loss will be minimal. However, core loss
will sometimes become a limiting factor in applications with a relatively high percentage of ripple current at
very high frequency. Iron powder is not recommended for inductors with discontinuous current or
transformers with large ac flux swings.

Low cost, iron powder cores are typically used in today's, low and high frequency power switching
conversion applications, for differential-mode, input and output power inductors. Because iron powder
cores have such low permeability, a relatively large number of turns are required for the proper inductance,
thus keeping the ac flux at a minimum. The penalty for using iron powder cores is usually found in the size
and efficiency of the magnetic component.

There are four standard powder materials available for power magnetic devices: Mob/permalloy (MPP)
Powder Cores with a family of curves, as shown in Figure 2-20; High flux (HF) Powder Cores with a family
of curves, as shown in Figure 2-21; Sendust Powder Cores, *(Kool Mu), with a family of curves, as shown
in Figure 2-22; and Iron Powder Cores, with a family of curves, as shown in Figure 2-23. The powder cores
come in a variety of permeabilities. This gives the engineer a wide range in which to optimize the design.
The powder core properties for the most popular materials are shown in Table 2-7. Also, given in Table 2-
7, is the Figure number for the B-H loop of each of the powder core materials. In Table 2-8 is a listing of
the most popular permeabilities for each of the powder core materials.

*Trademark of Magnetics Division, Spang and Company.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table 2-7. Powder Core Material Properties.

Powder Core Material Properties

Material Initial Flux Density Curie dc, Coercive Density Typical
Name Composition Permeability Tesla Temperature Force, He grams/cm B-H Loop
Hi Bs °C Oersteds 6 Figures
MPP 80% Ni 14-550 0.7 450 0.3 8.5 (2-16)
20% Fe
High Flux 50% Ni 14- 160 1.5 360 1 8 (2-17)
50% Fe
Sendust 85% Fe 26-125 1 740 0.5 6.15 (2-18) .
(Kool Mu) 9% Si
6% Al
Iron Powder 100%Fe 4.0- 100 0.5- 1.4 770 5.0-9.0 3.3-7.2 (2-19)

Table 2-8. Standard Powder Core Permeabilities.

Standard Powder Core Permeabilities

Powder Material MPP High Flux Sendust Iron Powder
(Kool Mu)
Initial Permeability, u,j
10 X
14 X X
26 X X X
35 X
55 X
60 X X X X
75 X X
90 X
100 X
125 X X X
147 X X
160 X X
173 X
200 X
300 X
550 X

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 B, Tesla
1.0 T
Molypermalloy MPP
125 Perm

50 100 150 200 250

250 200 150 100 50

Figure 2-16. Molypermalloy Powder Core, 125 Perm.

High Flux HF
125 Perm

50 100 150 200 250

250 200 150 100 50

Figure 2-17. High Flux Powder Core, 125 Perm.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Kool Mu
125 Perm

50 100 150 200 250

250 200 150 100 50

Figure 2-18. Sendust (Kool MH) Powder Core, 125 Perm.

Tesla

Iron Powder-52
75 Perm

50 100 150 200 250

i 1 1
250 200 150 100 50

Figure 2-19. Iron Powder (-52) Core, 75 Perm.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 i i i i i i
MPP Powder Cores

DC Magnetizing Force (Oersteds)

i i i i i i i i I I i i M
100 1000

Figure 2-20. Permeability Versus dc Bias for Molypermalloy Powder Cores.

100 I I I I I I
High Flux Powder Cores

DC Magnetizing Force (Oersteds)

I I I i i i i i i i i i
1000

Figure 2-21. Permeability Versus dc Bias for High Flux Powder Cores.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 100 i i i i r \ \T i l T l I l l r
Sendust Powder Cores (Kool Mp,)

80

a
1 60

40
a

&

20

DC Magnetizing Force (Oersteds)

I I I I I I I l l I I I I I I I
1.0 10 100 1000

Figure 2-22. Permeability Versus dc Bias for Sendust Powder Cores.

100 I I I I ! I I
Iron Powder Cores

80

J2
03
01
8 60
o
OH

40
B

20

DC Magnetizing Force (Oersteds)

J I l l l I l l l l l l I I J I
1.0 10 100 1000

Figure 2-23. Permeability Versus dc Bias for Iron Powder Cores.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Core Loss
The designer of power magnetic components, such as transformer and inductors, requires specific
knowledge about the electrical and magnetic properties of the magnetic materials used in these components.
There are two magnetic properties that are of interest, the dc and the ac. The dc B-H hysteresis loop is a
very useful guide for comparing the different types of magnetic materials. It is the ac magnetic properties
that are of interest to the design engineer. One of the most important ac properties is the core loss. The ac
core loss is a function of the magnetic material, magnetic material thickness, magnetic flux density Bac,
frequency f, and operating temperature. Thus, the choice of the magnetic material is based upon achieving
the best characteristic using the standard trade-off such as cost, size, and performance.

All manufacturers do not use the same units when describing their core loss. The user should be aware of
the different core loss units when comparing different magnetic materials. A typical core loss graph is
shown in Figure 2-24. The vertical scale is core loss, and the horizontal scale is flux density. The core loss
data is plotted at different frequencies, as shown in Figure 2-24.

100

Frequency #1
Frequency #2
Frequency #3
10
Frequency #4

0>
k.
o
U

1.0

0.1
0.01 0.1 1.0
Flux Density

Figure 2-24. Typical Graph for Plotting Core Loss at Different Frequencies.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Vertical Scale
Here is a list of core loss units used by manufacturers:
1. watts per pound
2. watts per kilogram
3. milliwatts per gram
4. milliwatts per cubic centimeter (cmj)

Horizontal Scale
Here is a list of flux density units used by manufacturers:
1. gauss
2. kilogauss
3. tesla
4. millitesla
The data can be plotted or presented in either hertz or kilohertz.

Core Loss Equations

Manufacturers are now presenting the core loss in the form of an equation such as:

watts / kilogram =kf(m)B(n) [2-2]

Here, again, the units will change from one manufacturer to another. In the following tables the
manufacturers core loss data has been organized with the same units for all core losses. The data was
modified to put the data in metric units, gauss to tesla and, watts per pound to watts per kilogram. The
coefficients for Magnetics Inc. molypermalloy powder cores, (MPP), using Equation [2-2] are shown in
Table 2-9. The coefficients for Magnetics Inc. High Flux powder cores, (HF), using Equation [2-2] are
shown in Table 2-10. The coefficients for Magnetics Inc. Sendust powder cores, (Kool-Mu), using
Equation [2-2] are shown in Table 2-11. The coefficients for iron alloy materials using Equation [2-2] are
shown in Table 2-12.

Table 2-9. Core Loss Coefficients for MPP Powder Cores.

Core Loss Equation Factors

Magnetics MPP Powder Cores
Material Permeability Coefficient Coefficient Coefficient
V k (m) (n)
MPP 14 0.005980 1.320 2.210
MPP 26 0.001190 1.410 2.180
MPP 60 0.000788 1.410 2.240
MPP 125 0.001780 1.400 2.310
MPP 147-160-173 0.000489 1.500 2.250
MPP 200-300 0.000250 1.640 2.270
MPP 550 0.001320 1.590 2.360

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table 2-10. Core Loss Coefficients for High Flux Powder Cores.

Core Loss Equation Factors

Magnetics HF Powder Cores
Material Permeability Coefficient Coefficient Coefficient
U k (m) (n)
High Flux 14 4.8667(1 0"7) 1.26 2.52
High Flux 26 3.0702(10"7) 1.25 2.55
7
High Flux 60 2.0304(1 0' ) 1.23 2.56
7
High Flux 125 1.1627(10' ) 1.32 2.59
7
High Flux 147 2.3209(10" ) 1.41 2.56
High Flux 160 2.3209(10'7) 1.41 2.56

Table 2-11. Core Loss Coefficients for Sendust Powder Cores.

Core Loss Equation Factors

Magnetics Kool-Mii Powder Cores
Material Permeability Coefficient Coefficient Coefficient
U k (m) (n)
Sendust 26 0.000693
Sendust 60 0.000634
Sendust 75 0.000620 1.460 2.000
Sendust 90 0.000614
Sendust 125 0.000596

Table 2-12. Core Loss Coefficients for Iron Alloy Cores.

Core Loss Equation Factors

Iron Alloy
Material Thickness Frequency Coefficient Coefficient Coefficient
mil's Range k (m) (n)
50/50 Ni-Fe 1.00 0.0028100 1.210 1.380
2.00 0.0005590 1.410 1.270
4.00 0.0006180 1.480 1.440
Supermendur 2.00 0.0236000 1.050 1.300
4.00 400 Hz 0.0056400 1.270 1.360
Permalloy 80 1.00 0.0000774 1.500 1.800
2.00 0.0001650 1.410 1.770
4.00 0.0002410 1.540 1.990
Supermalloy 1.00 0.0002460 1.350 1.910
2.00 0.0001790 1.480 2.150
4.00 0.0000936 1.660 2.060
Silicon 1.00 0.0593000 0.993 1.740
2.00 0.0059700 1.260 1.730
4.00 0.0035700 1.320 1.710
12.00 0.0014900 1.550 1.870
14.00 0.0005570 1.680 1.860
24 M27 non-or 50-60 H/ 0.0386000 1.000 2.092

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The coefficients for amorphous materials, using Equation [2-2] are shown in Table 2-13. The coefficients
for Magnetics ferrite materials, using Equation [2-2] are shown in Table 2-14. The coefficients for
Micrometals iron powder materials, using Equation [2-3] are shown in Table 2-15.

2
watts/kilogram = k Bl [2.3]

Table 2-13. Core Loss Coefficients for Amorphous Materials.

Core Loss Equation Factors

Amorphous
Material Thickness Coefficient Coefficient Coefficient
mils k (m) (n)
6
2605SC 0.80 8.79(10" ) 1.730 2.230
2714A 0.80 10.1(10-6) 1.550 1.670
Vitroperm 500 0.80 0.864(10-6) 1.834 2.112

Table 2-14. Core Loss Coefficients for Magnetics Ferrites Materials.

Core Loss Equation Factors

Magnetic's Ferrite Core Materials
Material Frequency Coefficient Coefficient Coefficient
Range k (m) (n)
4
K f< 500kHz 2.524(10" ) 1.60 3.15
K 500kHz < = f < 1.0 MHz 8.147(10'8) 2.19 3.10
19
K f=>1.0MHz 1.465(10" ) 4.13 2.98
4
R f < 100kHz 5.597(10' ) 1.43 2.85
5
R 100kHz<=f< 500kHz 4.316(10" ) 1.64 2.68
6
R f=> 500kHz 1.678(10" ) 1.84 2.28
P 3
f < 100kHz 1.983(10' ) 1.36 2.86
5
P 100kHz<=f< 500kHz 4.855(10' ) 1.63 2.62
15
P f=> 500kHz 2.068(10' ) 3.47 2.54
2
F f =< 10kHz 7.698(10" ) 1.06 2.85
5
F 10kHz<f<100kHz 4.724(1 0" ) 1.72 2.66
5
F 100kHz<=f< 500kHz 5.983(10' ) 1.66 2.68
6
F f=> 500kHz 1.173(10" ) 1.88 2.29
J f=< 20kHz 1.091(10"3) 1.39 2.50
f> 20kHz 8
J 1.658(10' ) 2.42 2.50
3
W f=< 20kHz 4.194(10" ) 1.26 2.60
w f> 20kHz 8
3.638(10" ) 2.32 2.62
4
H f=< 20kHz 1.698(10" ) 1.50 2.25
H 5
f> 20kHz 5.3720(10" ) 1.62 2.15

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table 2-15. Core Loss Coefficients for Iron Powder Cores.

Core Loss Equation Factors

Micrometals Iron Powder Cores
Material Permeability Coefficient Coefficient Coefficient Coefficient
u (a) (b) (c) (d)
Mix-08 35 0.01235 0.8202 1.4694 3.85(10"7)
Mix- 18 55 0.00528 0.7079 1.4921 4.70(1 0"7)
Mix-26 75 0.00700 0.4858 3.3408 2.71(10"6)
Mix-52 75 0.00700 0.4858 3.6925 9.86(10'7)

Selection of Magnetic Materials

Transformers used in static inverters, converters, and transformer-rectifier, (T-R), supplies, intended for
aerospace and electronics industry power applications, are usually square loop tape, toroidal design. The
design of reliable, efficient, and lightweight devices for this use has been seriously hampered by the lack of
engineering data, describing the behavior of both the commonly used and more exotic core materials, with
higher-frequency square wave excitation.

A program was carried out at the Jet Propulsion Laboratory, JPL, to develop these data from measurements
of the dynamic B-H loop characteristics of the tape core materials presently available from various industry
sources. Cores were produced in both toroidal and C forms, and were tested in both ungapped (uncut) and
gapped (cut) configuiations. This section describes the results of that investigation.

Typical Operation

A transformer used for inverters, converters, and transformer-rectifier suppliers operates from a power bus,
which could be dc or ac. In some power applications, a commonly used circuit is a driven transistor switch
arrangement, such as that shown in Figure 2.25.

Tl

2.4kHz
Square-wave

Figure 2-25. Typical Driven Power MOSFET Inverter.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 One important consideration affecting the design of suitable transformers is that care must be taken to
ensure that operation involves balance drive to the transformer primary. In the absence balance drive, a net
dc current will flow in the transformer primary, which causes the core to saturate easily during alternate
half-cycles. A saturated core cannot support the applied voltage, limited mainly by its on resistance. The
resulting high current, in conjunction with the transformer leakage inductance, results in a high-voltage
spike during the switching sequence, which could be destructive to the power MOSFET. To provide
balanced drive, it is necessary to exactly match the MOSFETs for RDS(on)- But, this is not always
sufficiently effective. Also, exact matching of the MOSFETs is a major problem in a practical sense.

Material Characteristics
Many available core materials approximate the ideal square loop characteristic, illustrated by the B-H
curve, as shown in Figure 2-26. Representative, dc B-H loops for commonly available core materials, are
shown in Figure 2-27. Other characteristics are tabulated in Table 2-16.

Many articles have been written about inverter and converter transformer design. Usually, authors'
recommendations represent a compromise among material characteristics, such as those tabulated in Table
2-16, and displayed in Figure 2-27. These data are typical of commercially available core materials that are
suitable for the particular application.

B, (tesla)

+Saturation

-H H

-Saturation

-B, (tesla)

Figure 2-26. Ideal Square B-H Loop.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 B
2.2
Supermendur

Sq. Permalloy

Supermalloy

H
0.4 0.3 0.2 0.1 0 0. 0.2 0.3 0.4
H, amp-turn/cm

Figure 2-27. Typical dc B-H Loops of Magnetic Materials.

As can be seen, the material that provides the highest flux density (supermendur) would result in the
smallest component size, and tins would influence the choice if size were the most important consideration.
The ferrite materml, (see the ferrite curve in Figure 2-27), has the lowest flux density. This results in the
largest transformer. Magnetic materials selected for transformers or inductors cannot be chosen by flux
alone. There are other parameters, such as frequency and core configuration that must be taken into
consideration.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table 2-16. Magnetic Core Material Characteristics.

Magnetic Core Material Characteristics

Material Initial Flux Density Curie dc, Coercive Density
Name Composition Permeability Tesla Temperature Force, He grams/cm

Hi Bs °C Oersteds 5
Magnesil 3% Si 1.5 K 1.5-1.8 750 0.4-0.6 7.3
97% Fe

Supermendur* 49% Co 0.8 K 1.9-2.2 940 0.15-0.35 8.15

49% Fe
2%V
Orthonol 50% Ni 2K 1.42-1.58 500 0.1-0.2 8.24
50% Fe

Sq. Permalloy 79% Ni 12K-100K 0.66-0.82 460 0.02-0.04 8.73

17%Fe
4% Mo
Supermalloy 78% Ni 10K-50K 0.65-0.82 460 0.003-0.008 8.76
17%Fe
5% Mo
Amorphous 81%Fe 3K 1.5-1.6 370 0.03-0.08 7.32
2605-SC 13.5%B
3.5% Fe
Amorphous 66% Co 20K 0.5-0.58 205 0.008-0.02 7.59
2714A 15% Si
4% Fe
Ferrite MnZn 0.75-15K 0.3-0.5 100-300 0.04-0.25 4.8

* Field Anneal.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Usually, inverter transformer design is aimed at the smallest size, the highest efficiency and adequate
performance under the widest range of environmental conditions. Unfortunately, the core material that can
produce the smallest size has the lowest efficiency, and the highest efficiency materials result in the largest
size. Thus, the transformer designer must make tradeoffs between allowable transformer size and the
minimum efficiency that can be tolerated. Then, the choice of core material will be based upon achieving
the best characteristic on the most critical or important design parameter, and upon acceptable compromises
on the other parameters.

After analysis of a number of designs, most engineers choose size rather than efficiency as the most
important criterion, and select an intermediate loss factor on core material for their transformers.
Consequently, as the frequency is increased, ferrites have become the most popular material.

Magnetic Material Saturation Defined

To standardize the definition of saturation, several unique points on the B-H loop are defined, as shown in
Figure 2-28.
B

:: Bi

H
0 1A
H n i•

Figure 2-28. Defining the B-H Loop.

The straight line through (H0, 0) and (Hs, Bs) may be written as:

The line through (0, B2) and (Hs, Bs) has essentially zero slope and may be written as:

[2-5]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Equations [2-1] and [2-2] together define saturation conditions as follows:

Solving Equation [2-3] for Hs yields:

- [2-7]

Where by definition:

By definition, saturation occurs when the peak exciting current (B) is twice the average exciting current (A)
as shown in Figure 2-29. Analytically, this means that:

[2-9]

c
Ji *
y J
Magnetizing Q y— 1
i
\
Current

^—N
By definition, saturation occurs when B = 2A

Figure 2-29. Defining the Excitation Current.

Solving Equation [2-1] for HI, we obtain:

A [2.10]

To obtain the pre-saturation dc margin (AH), Equation [2-4] is subtracted from Equation [2-6]:

AH=Hs-Hl = [2-11]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The actual unbalanced dc current must be limited to:

7, fc < [amperes] [2-12]

Where, N is the number of turns

MPL is the mean magnetic path length.

Combining Equations [2-7] and [2-8] gives:

, [amperes] [2-13]

As mentioned earlier, in an effort to prevent core saturation, the drive to the switching power MosFet must
be symmetrical and the power MosFet on resistance Ros(on) must be matched. The effect of core saturation,
using an uncut or ungapped core, is shown in Figure 2-30, which illustrates the effect on the B-H loop
transversed with a dc bias. Figure 2-31 shows typical B-H loops of 50-50 nickel-iron material excited from
an ac source, with progressively reduced excitation; the vertical scale is 0.4 T/cm. It can be noted that the
minor loop remains at one extreme position within the B-H major loop after reduction of excitation. The
unfortunate effect of this random minor loop positioning is that, when conduction begins again in the
transformer winding after shutdown, the flux swing could begin from the extreme ends rather than from the
normal zero axis. The effect of this is to drive the core into saturation, with the production of spikes that
can destroy transistors.

B, (tesla)

Figure 2-30. B-H Loop with dc Bias.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Figure 2-31. Typical Square Loop Material with dc Bias.

Test Conditions

The test fixture, schematically illustrated in Figure 2-32, was built to affect comparison of dynamic B-H
loop characteristics of various core materials. Cores were fabricated from various core materials in the
basic core configuration, designated No. 52029 for toroidal cores, manufactured by Magnetics Inc. The
materials used were those most likely to be of interest to designers of inverter or converter transformers.
Test conditions are listed in Table 2-17.

Tl
SI

-O Gnd
-O Vertical
Current Probe ,~ Horizontal

Figure 2-32. B-H Loop with dc Bias.

Winding Data were derived from the following:

V 104)
N= ^ '—, [turns] [2-14]
4.0BafAc'

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table 2-17. Materials and Test Conditions.

Test Conditions
Bs Turns Frequency MPL
Core Number* Trade Name Tesla N kHz cm
52029-2A Orthonol 1.45 54 2.4 9.47
52029-2D Sq. Permalloy 0.75 54 2.4 9.47
52029-2F Supermalloy 0.75 54 2.4 9.47
52029-2H 48 Alloy 1.15 54 2.4 9.47
52029-2K Magnesil 1.6 54 2.4 9.47
*Magnetics toroidal cores.

The test transformer, represented in Figure 2-32, consists of 54-turn primary and secondary windings, with
square wave excitation on the primary. Normally, switch SI is open. With switch SI closed, the secondary
current is rectified by the diode to produce a dc bias in the secondary winding.

Cores were fabricated from each of the materials by winding a ribbon of the same thickness on a mandrel of
a given diameter. Ribbon termination was affected by welding in the conventional manner. The cores were
vacuum impregnated, baked, and finished as usual.

Figures 2-33 - 2-36 show the dynamic B-H loops obtained for various core materials. In each of these
Figures, switch SI was in the open position, so there was no dc bias applied to the core and windings.

B, (tesla)

- Saturation

-H H

-Saturation

-B, (tesla)

Figure 2-33. Magnesil (K) B-H Loop, B = 0.5 T/cm, H = 100 ma/cm.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 B, (tesla)

Saturation

-Saturation

-B, (tesla)

Figure 2-34. Orthonol (2A) B-H Loop, B = 0.5 T/cm, H = 50 ma/cm.

B, (tesla)

• Saturation

-Saturation

-B, (tesla)

Figure 2-35. Square Permalloy (2D) B-H Loop, B = 0.2 T/cm, H = 20 ma/cm.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 B, (tesla)

Saturation

-Saturation

-B, (tesla)

Figure 2-36. Supermalloy (2F) B-H Loop, B = 0.2 I/cm, H = 10 ma/cm.

Figures 2-37 to 2-40 show the dynamic B-H loop patterns obtained for various core materials when the test
conditions are included in a sequence, in which SI was in open condition (A), then in closed condition (B),
and then, opened again, in condition (C). It is apparent from this data that, with a small amount of dc bias,
the minor dynamic B-H loop can traverse the major B-H loop from saturation to saturation. Note that after
the dc bias has been removed, the minor B-H loops remained shifted to one side or the other. Because of
the ac coupling of the integrator to the oscilloscope, the photographs in these figures do not present a
complete picture of what really happens during the flux swing.

B, (tesla)

-B, (tesla)

Figure 2-37. Magnesil (2K) B-H Loop, B = 0.3 T/cm, H = 200 ma/cm.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 B, (tesla)

B c
-H -H+ H

-B, (tesla)

Figure 2-38. Orthonol (2A) B-H Loop, B = 0.2 T/cm, H = 100 ma/cm.

B, (tesla)

-B, (tesla)

Figure 2-39. Sq. Permalloy (2D) B-H Loop, B = 0.1 T/cm, H = 20 ma/cm.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 B, (tesla)

-B, (tesla)

Figure 2-40. Supermalloy (2F) B-H Loop, B = 0.1 T/cm, H = 10 ma/cm.

Magnetic Material Saturation Theory

The domain theory of the nature of magnetism is based on the assumption that all magnetic materials consist
of individual molecular magnets. These minute magnets are capable of movement within the material.
When a magnetic material is in its unmagnetized state, the individual magnetic particles are arranged at
random, and effectively neutralize each other. An example of this is shown in Figure 2-41, where the tiny
magnetic particles are arranged in a disorganized manner. (The north poles are represented by the darkened
ends of the magnetic particles.) When a material is magnetized, the individual particles are aligned or
oriented in a definite direction, as shown in Figure 2-42.

Domains

Toroidal Core

Figure 2-41. Magnetic Domains, Arranged in a Random Manner.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Toroidal Core

Domains

Figure 2-42. Magnetic Domains, Aligned in a Definite Direction.

The degree of magnetization of a material depends on the degree of alignment of the particles. The external
magnetizing force can continue to affect the material up to the point of saturation, the point at which
essentially all of the domains are lined up in the same direction.

In a typical toroidal core, the effective air gap is less than 10~6 cm. Such a gap is negligible in comparison
to the ratio of mean length to permeability. If the toroid was subjected to a strong magnetic field (enough to
saturate), essentially all of the domains would line up in the same direction. If suddenly the field were
removed at Bm, the domains would remain lined up, and be magnetized along that axis. The amount of flux
density that remains is called the residual flux, Br. The result of this effect was shown earlier in Figures 2-
37 through 2-40.

Air Gap Effect

An air gap introduced into the core has a powerful demagnetizing effect, resulting in a "shearing over" of
the hysteresis loop, and a considerable decrease in permeability of high-permeability materials. Direct
current excitation follows the same pattern. However, the core bias is considerably less affected than the
magnetization characteristics by the introduction of a small air gap. The magnitude of the air gap effect
also depends on the length of the mean magnetic path and on the characteristics of the uncut core. For the
same air gap, the decrease in permeability will be less with a greater magnetic flux path, but more
pronounced in a high-permeability core with a low coercive force

Effect of Gapping
Figure 2-43 shows a comparison of a typical toroidal core B-H loop, without and with a gap. The gap
increases the effective length of the magnetic path. When voltage E is impressed across primary winding,
Np, of a transformer, the resulting current, Im, will be small because of the highly inductive circuit, as shown
in Figure 2-44. For a particular core size, maximum inductance occurs when the air gap is minimum.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 B, (teslas)
Bmax

{H-HhH-

(a)
Without Gap

-Bmax

-B, (teslas) -B, (teslas)

Figure 2-43. Comparing Magnetic Materials with and Without a Gap.

Tl
SI
R2
2.4kHz
Square-wave 54T 54T
CR1

Figure 2-44. Implementing dc Unbalance.

When SI is closed, an unbalanced dc current flows in the secondary, Ns turns, and the core is subjected to a
dc magnetizing force, resulting in a flux density that may be expressed as:

**=- MPL
, [tesla] [2-15]

In converter and inverter design, this dc flux is augmented by the ac flux swing, which is:

B = , [tesla] [2-16]
KffAcN

If the sum of Bdc and Bac shifts operations above the maximum operating flux density of the core material,
the incremental permeability, (|iac), is reduced. This condition lowers the impedance and increases the flow

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 of magnetizing current, Im. This condition can be remedied by introducing into the core assembly an air gap
which causes a decrease in dc magnetization in the core. However, the size of the air gap that can be
incorporated has a practical limitation. Since the air gap lowers impedance, it results in increased
magnetizing current Im, which is inductive. The resultant voltage spikes produced by such currents apply a
high stress to the switching transistors and may cause failure. This stress can be minimized by tight control
of lapping and etching of the gap to keep the gap to a minimum.

From Figure 2-43, it can be seen that the B-H curves depict maximum flux density, Bm, and residual flux,
Br, for ungapped and gapped cores, and that the useful flux swing is designated, AB, which is the difference
between, Bm and Br. It will be noted, in Figure 2-43(a), that Br approaches Bnl, but, in Figure 2-43(b), there
is a much greater, AB, between them. In either case, when excitation voltage is removed at the peak of the
excursion of the B-H loop, flux falls to the Br point. It is apparent that introducing an air gap reduces Br to
a lower level, and increases the useful flux density. Thus, insertion of an air gap in the core eliminates, or
markedly reduces, the voltage spikes produced by the leakage inductance, due to the transformer saturation.

Two types of core configurations were investigated in the ungapped and gapped states. Figure 2-45 shows
the type of toroidal core that was cut, and Figure 2-46 shows the type of C core that was cut. Toroidal cores
are virtually gapless, when conventionally fabricated. To increase the gap, the cores were physically cut in
half, and the cut edges were lapped, acid etched to remove cut debris, and banded to form the cores. A
minimum air gap, on the order of less than 25 urn, was established.

Gap Placement

Figure 2-45. Typical cut Toroidal Core.

As will be noted from Figures 2-47 through 2-50, which show the B-H loops of the uncut and cut cores, the
results obtained indicated that the effect of gapping was the same for both the C cores and the toroidal cores
subjected to testing. It will be noted, however, that gapping of the toroidal cores produced a lowered
squareness characteristic for the B-H loop, as shown in Table 2-18. This data was obtained from Figures 2-
47 through 2-50. AH values extracted from the same figures, as shown in Figure 2-51, are tabulated in
Table 2-19.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Gap Placement

Figure 2-46. Typical Middle Cut C Core.

B, (tesla) B, (tesla)
\ i
— Uncut

-H H -H

B = 0.5 T/cm, H = 100 ma/cm B = 0.5 T/cm, H = 500 ma/cm

Figure 2-47. Magnesil (K) B-H Loop, Uncut and Cut with Minimum Gap.

B, (tesla) B, (tesla)

H -H

B = 0.5 T/cm, H = 50 ma/cm B = 0.5 T/cm, H = 100 ma/cm

Figure 2-48. Orthonal (A) B-H Loop, Uncut and Cut with Minimum Gap.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 B, (tesla) B, (tesla)

H -H

B = 0.2 T/cm, H = 20 ma/cm B = 0.2 T/cm, H = 100 ma/cm

Figure 2-49. Square Permalloy (D) B-H Loop, Uncut and Cut with Minimum Gap.

B, (tesla) B, (tesla)

H -H

B = 0.2 T/cm, H = 10 ma/cm B = 0.2 T/cm, H = 50 ma/cm

Figure 2-50. Supermalloy (F) B-H Loop, Uncut and Cut with Minimum Gap.

Table 2-18.

ComparingI Br/Bm on Uncut and Cut Cores.

Bs Turns Uncut Cut
Core Number* Trade Name Tesla N Br/Bm Br/Bm
52029-2A Orthonol 1.45 54 0.96 0.62
52029-2D Sq. Permalloy 0.75 54 0.86 0.21
52029-2F Superpermalloy 0.75 54 0.81 0.24
52029-2K Magnesil 1.60 54 0.93 0.22
*Magnetics toroidal cores.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 B, (teslas) B, (teslas)

(a)
Without Gap

-H I 1111 111111

AH,op

-B, (teslas) -B, (teslas)

Figure 2-51. Defining AHop and AH.

Table 2-19. Comparing AH and AHop on Uncut and Cut Cores.

Comparing AH and AHop on Uncut and Cut Cores.

Amp-turns/cm
Material Bm Bac Bdc Uncut Cut
*Trade Name Tesla Tesla Tesla AHop AH AHop AH
Orthonol 1.44 1.15 0.288 0.0125 0 0.895 0.178
Sq. Permalloy 0.73 0.58 0.146 0.0100 0.005 0.983 0.178
Supermalloy 0.63 0.58 0.136 0.0175 0.005 0.491 0.224
Magnesil 1.54 1.23 0.310 0.0750 0.025 7.150 1.780
*Magnetics Cores.

A direct comparison of cut and uncut cores was made electrically by means of two different test circuits.
The magnetic material used in this branch of the test was Orthonol. The frequency was 2.4 kHz, and the
flux density was 0.6 T. The first test circuit, shown in Figure 2-52, was a driven inverter operating into a
30-W load, with the power MOSFETs, operating into and out of saturation. Drive was applied

continuously. SI controls the supply voltage to Ql and Q2.

With switch SI closed, transistor Ql was turned on and allowed to saturate. This applied voltage, E-VDs(on),
across the transformer winding. Switch SI was then opened. Then, the flux in transformer, T2, dropped to

the residual flux density, Br. Switch SI was closed again. This was done several times in succession to
catch the flux in an additive direction. Figures 2-53 and 2-54 show the inrush current measured at the
center tap of T2.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Tl Rl T2

2.4kHz
Square-wave SOW

Current Probe

Figure 2-52. Inverter Inrush Current Test Fixture.

Figure 2-53. Typical Inrush Current of a Uncut Core in a Driven Inverter.

rot

Figure 2-54. Resulting Inrush Current using a Cut Core.

It will be noted, in Figure 2-53, that the uncut core saturated, and the inrush current was limited only by
circuit resistance and power, MOSFETs Ros(on)- Figure 2-54 shows that saturation did not occur in the case
of the cut core. Thus, the high inrush current and transistor stress were virtually eliminated.

The second test circuit arrangement is shown in Figure 2-55. The purpose of this test was to excite a
transformer and measure the inrush current, using a current probe. A square wave power oscillator was
used to excite transformer, T2. Switch, SI, was opened and closed several times to catch the flux in an
additive direction. Figures 2-56 and 2-57 show inrush current for an uncut and cut core, respectively.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 2.4kHz
Square-wave

Current Probe
Figure 2-55. Transformer Rectifier Inrush Current Measurement.

J3 ^ftp^ W

Figure 2-56. Inrush Current of a Transformer using a Uncut Core.

Figure 2-57. Inrush Current of a Transformer using a Cut Core.

A small amount of air gap, less than 25 ^m, has a powerful effect on the demagnetizing force, but little
effect on the core loss. This small air gap decreases the residual magnetism by "shearing over" the
hysteresis loop, which eliminates the problem of the core tending to remain saturated.

A typical example of the merits of the cut core occurred in the checkout of a Mariner spacecraft. During the
checkout of a prototype science package, a large (8 A, 200 s) turn-on transient was observed. The normal
running current was 0.06 A, fused with a parallel-redundant 1/8-A fuse, as required by the Mariner Mars
design philosophy. With the 8-A inrush current, the 1/8-A fuses were easily blown. This did not happen,
on every turn-on, but only when the core would "latch up" in the wrong direction for turn-on. Upon
inspection, the transformer turned out to be a 50-50 nickel-iron toroid. The design was changed from a
toroidal core to a cut core with a 25 um, air gap. The new design was completely successful in eliminating
the 8-A turn-on transient.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Composite Core Configuration

A composite core configuration has been developed for transformers that combine the protective feature of
a gapped core with the much lower magnetizing current requirement of an uncut core. The uncut core
functions, under normal operating conditions, and the cut core takes over during abnormal conditions to
prevent high switching transients aii'd their potentially destructive effect on the transistors.

This configuration is a composite of cut and uncut cores assembled together concentrically, with the uncut
core nested within the cut core. The uncut core has high permeability, and thus requires a very small
magnetizing current. On the other hand, the cut core has a low permeability and thus requires a much
higher magnetization current. The uncut core is designed to operate at a flux density that is sufficient for
normal operation of the converter. The uncut core may saturate under the abnormal conditions previously
described. The cut core then takes over and supports the applied voltage so that excessive current does not
flow. In a sense, it acts like a ballast resistor in some circuits to limit current flow to a safe level.

Figures 2-58 and 2-59 show the magnetization curves for an uncut core and a composite core of the same
material at the same flux density. The much lower, Br characteristic of the composite compared to the uncut
core is readily apparent.

The desired features of the composite core can be obtained more economically by using different materials
for the cut and uncut portions of the core. It was found that when the design required high nickel (4/79), the
cut portion could be low nickel, (50/50), and because low nickel has twice as high a flux density as high
nickel, the core was made of 66% high nickel, and 33% low nickel.

B, (tesla)

-H H

-B, (tesla)
Figure 2-58. Uncut Core Excited at 0.2 T/cm.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 B, (tesla)

4-

-H ++++ H

:/

-B, (tesla)
Figure 2-59. Both Cut and Uncut Cores Excited at 0.2 T/cm.
Figure 2-60 shows cut and uncut cores that have been impregnated to bond the ribbon layers together. The
uncut core was first trimmed to fit within the inner diameter of the cut core by peeling off a wrap or two of
the ribbon steel. The two cores are assembled into a composite core (Figure 2-61, on the right).

Square Permalloy

Orthonol 50/50
Figure 2-60. Composite Cores Ready for final Assembly.

Orthonol 50/50
Square Permalloy 80

OD ID

Figure 2-61. Composite Cores Assembled in Final Form.

To ensure uniform characteristics for gapped cores, a gap dimension of 50 urn is recommended, because
variations produced by thermal cycling will not affect this gap greatly. In the composite core, the gap is
obtained by inserting a sheet of paper Mylar or Kapton film material between the core ends during banding.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The same protective feature can be accomplished in transformers with laminated cores. When laminations
are stacked by interleaving them one-by-one, the result will be a minimum air gap, as shown in Figure 2-62
by the squareness of the B-H loop. Shearing over of the B-H loop, or decreasing the residual flux, as shown
in Figure 2-63, is accomplished by butt joining half the laminations in the core-cross section, which
introduces a small, additional air gap.

B, (tesla)

-H H

-B, (tesla)
Figure 2-62. B-H Loop with Laminations Stacked 1x1 Interleaved.

B, (tesla)

-H H

-B, (tesla)

Figure 2-63. B-H Loop with Laminations Stack Half 1x1 and Half Butt Stack.

Table 2-20 is a compiling of composite cores manufactured by Magnetics Inc., alongside their standard
dimensional equivalent cores. Also, included in Table 2-20, is the cores' area product, Ap, and the core
geometry Kg, which is discussed in Chapter 7.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table 2-20. Composite Core Listing Along with the Area Product and Core Geometry.

Magnetics Inc. Composite Cores

Composite Standard Core A
P Kg
Number Number (cm4) (cm5)
01605-2D 52000 0.0728 0.00105
01754-2D 52002 0.1440 0.00171
01755-2D 52076 0.2850 0.00661
01609-D2 52061 0.3890 0.00744

01756-2D 52106 0.4390 0.00948

01606-2D 52094 0.6030 0.02210
01761-2D 52318 0.7790 0.02600
01757-2D 52029 1.0900 0.02560

01760-2D 52188 1.1520 0.05120

02153-2D 52181 1.2200 0.04070
01758-2D 52032 1.4550 0.04310
01607-2D 52026 2.1800 0.08740

01966-2D 52030 2.3370 0.06350

01759-2D 52038 2.9100 0.14000
01608-2D 52035 4.6760 0.20600
01623-2D 52425 5.2550 0.26200
01624-2D 52169 7.1300 0.41800
Ac = 66% Square Permalloy 4 / 79.
Ac = 33% Orthonol 50 / 50.
lg = 2 mil Kapton.

Summary

Low-loss tape-wound toroidal core materials, that have a very square hysteresis characteristic, (B-H loop),
have been used extensively in the design of spacecraft transformers. Due to the squareness of the B-H
loops of these materials, transformers designed with them tend to saturate quite easily. As a result, large
voltage and current spikes, which cause undue stress on the electronic circuitry, can occur. Saturation
occurs when there is any unbalance in the ac drive to the transformer, or when any dc excitation exists.
Also, due to the square characteristic, a high residual flux state, (high Br), may remain when excitation is
removed. Reapplication of excitation in the same direction may cause deep saturation, and an extremely
large current spike, limited only by source impedance and transformer winding resistance, can result. This
can produce catastrophic failure.

With the introduction of a small, (less than 25 um), air gap into the core, the problems described above can
be avoided while retaining the low-loss properties of the materials. The air gap has the effect of "shearing
over" the B-H loop of the material so that the residual flux state is low and the margin between operating,

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 flux density, and saturation, flux density is high. The air gap thus has a powerful demagnetizing effect upon
the square loop materials. Properly designed transformers, using cut toroid or C core square-loop materials,
will not saturate upon turn-on, and can tolerate a certain amount of unbalanced drive or dc excitation.

It must be emphasized, however, that because of the nature of the material and the small size of the gap,
extreme care and control must be taken in performing the gapping operation. Otherwise, the desired
shearing effect will not be achieved, and the low-loss properties will be lost. The cores must be very
carefully cut, lapped, and etched to provide smooth, residue-free surfaces. Reassembly must be performed
with equal care.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Chapter 3

Magnetic Cores

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table of Contents

1. Introduction
2. Core Type and Shell Type Construction
3. Types of Core Materials
4. Eddy Currents and Insulation
5. Laminations
6. Annealing and Stress-Relief
7. Stacking Laminations and Polarity
8. Flux Crowding
9. Exciting Current
10. Tape Wound C, EE, and Toroidal Cores
11. Tape Toroidal Cores
12. Toroidal, Powder Core
13. Stacking Factors
14. Design and Dimensional Data for El Laminations
15. Design and Dimensional Data for UI Laminations
16. Design and Dimensional Data for LL Laminations
17. Design and Dimensional Data for DU Laminations
18. Design and Dimensional Data for Three Phase Laminations
19. Design and Dimensional Data for Tape Wound C Cores
20. Dimensional Outline for Tape Wound EE Cores
21. Design and Dimensional Data for Tape Wound Toroidal Cores
22. Design and Dimensional Data for EE and El Ferrite Cores
23. Design and Dimensional Data for EE and El Planar, Ferrite Cores
24. Design and Dimensional Data for EC, Ferrite Cores
25. Design and Dimensional Data for ETD, Ferrite Cores
26. Design and Dimensional Data for ETD/(low profile), Ferrite Cores
27. Design and Dimensional Data for ER, Ferrite Cores
28. Design and Dimensional Data for EFD, Ferrite Cores
29. Design and Dimensional Data for EPC, Ferrite Cores
30. Design and Dimensional Data for PC, Ferrite Cores
31. Design and Dimensional Data for EP, Ferrite Cores
32. Design and Dimensional Data for PQ, Ferrite Cores
33. Design and Dimensional Data for PQ/(low profile), Ferrite Cores
34. Design and Dimensional Data for RM, Ferrite Cores

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 35. Design and Dimensional Data for RM/(Iow profile), Ferrite Cores
36. Design and Dimensional Data for DS, Ferrite Cores
37. Design and Dimensional Data for UUR, Ferrite Cores
38. Design and Dimensional Data for UUS, Ferrite Cores
39. Design and Dimensional Data for Toroidal, Ferrite Cores
40. Design and Dimensional Data for Toroidal, MPP Powder Cores
41. Design and Dimensional Data for Toroidal, Iron Powder Cores
42. Design and Dimensional Data for Toroidal, Sendust Powder Cores
43. Design and Dimensional Data for Toroidal, High Flux Powder Cores
44. Design and Dimensional Data for EE, Iron Powder Cores
45. Design and Dimensional Data for EE, Sendust Powder Cores
46. References

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Introduction

The key ingredient in a magnetic device is the magnetic field (flux) created when current is passed through
a coiled wire. The ability to control (channel, predict, conduct), the magnetic field (flux) is critical to
controlling the operation of the magnetic device.

The ability of a material to conduct magnetic flux is defined as permeability. A vacuum is defined as
having a permeability of 1.0 and the permeability of all other materials is measured against this baseline.
Most materials such as air, paper, and wood are poor conductors of magnetic flux, in that they have low
permeability. If wire is wound on a dowel, it exhibits a magnetic field exactly, as shown in Figure 3-1.
There are a few materials, such as iron, nickel, cobalt, and their alloys that have high permeabilities,
sometimes ranging into the hundreds of thousands. These materials and their alloys are used as the base
materials for all core materials.

Magnetic Flux, O

Current, I

Coil

Dowel

Figure 3-1. Air Core with an Intensified Magnetic Field.

The main purpose of the core is to contain the magnetic flux and create a well-defined, predictable path for
the flux. This flux path, and the mean distance covered by the flux within the magnetic material, is defined
as the Magnetic Path Length (MPL) (see Figure 3-2). The Magnetic Path Length and permeability are vital
keys in predicting the operation characteristic of a magnetic device. Selection of a core material and
geometry are usually based on a compromise between conflicting requirements, such as size, weight,
temperature rise, flux density, core loss, and operating frequency.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 agnetic Path Length

Flux, <t

Magnetic Core

Figure 3-2. Magnetic Core Confines the Magnetic Field.

Core Type and Shell Type Construction

There are two types of construction for magnetic cores, core type and shell type. The shell type
construction is shown in Figure 3-3, and the core type construction is shown in Figure 3-4. In the shell type,
shown in Figure 3-3, the core surrounds the coil. Here the magnetic fields are around the outside of the
coil. The advantage of this configuration is that it requires only one coil. In the core type of construction,
shown in Figure 3-4, the coils are outside of the core. A good example of this is a toroid, where the coil is
wound on the outside of a core.

\
Flux, E-I Core

Coil

Figure 3-3. Shell Type Construction: the Core Surrounds the Coil.

Flux, <X> CCore

Coils

Figure 3-4. Core Type Construction the Coil Surrounds the Core.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Types of Core Materials

Magnetic cores are made of three basic materials. The first is bulk metal, the second is powdered materials,
and the third is ferrite material.

The bulk metals are processed from the furnace into ingots. Then, the material is put into a process of hot
and cold rolling. The rolling process produces a sheet of material with a thickness ranging from 0.004 to
0.031 inches that can be punched into laminations. It can be further rolled to thicknesses ranging from
0.002 to 0.000125 inches, then slit and wound into tape cores, such as C cores, E cores and toroids.

The powder cores, such as powder molypermalloy and powdered iron materials, are die-pressed into
toroids, EE cores and slugs. Powder core processing starts at the ingot, then goes through various steps of
grinding until the powder is the right consistency for the required performance. Normally, powder cores are
not machined after processing.

Ferrites are ceramic materials of iron oxide, alloyed with oxides or carbonate of manganese, zinc, nickel,
magnesium, or cobalt. Alloys are selected and mixed, based on the required permeability of the core.
Then, these mixtures are molded into the desired shape with pressure of approximately 150-200 tons per
square inch and fired at temperatures above 2000 degrees F. After the parts are made, they are usually
rumbled to remove burrs and sharp edges, which are characteristic of this process. Ferrites can be
machined to almost any shape to meet the engineer's needs.

Eddy Currents and Insulation

Transformers, operating at moderate frequency, require the reduction of eddy current losses in the magnetic
material. To reduce the eddy current losses to a reasonable value requires electrical steel to have adequate
resistivity. Also, it needs to be rolled to a specific thickness, and it needs effective electrical insulation or
coating of the magnetic material.

If an alternating voltage is applied to the primary winding, as shown in Figure 3-5, it will induce an
alternating flux in the core. The alternating flux will, in turn, induce a voltage on the secondary winding.
This alternating flux also induces a small alternating voltage in the core material. These voltages produce
currents called eddy currents, which are proportional to the voltage. The magnitude of these eddy currents
is also limited by the resistivity of the material. The alternating flux is proportional to the applied voltage.
Doubling the applied voltage will double the eddy currents. This will raise the core loss by a factor of four.
Eddy currents not only flow in the lamination itself, but could flow within the core as a unit, if the
lamination is not properly stamped, and if the lamination is not adequately insulated, as shown in Figure 3-
6.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Magnetic Core
Applied Voltage

Secondary Voltage

Figure 3-5. Applied Alternating Voltage Induces an Alternating Flux.

There are two eddy currents, as shown in Figure 3-6, ia and ib. The intralaminar eddy current, ia, is
governed by flux, per lamination and resistance of the lamination. It is, therefore, dependent on lamination
width, thickness, and volume resistivity.
Insulation, (Coating)

Eddy Current, ia

Eddy Current, z

Figure 3-6. Insulation is Required Between Laminations to Reduce Eddy Currents.

The interlaminar eddy current, ib, is governed by total flux and resistance of the core stack. It is primarily
dependent upon stack width and height, the number of laminations, and the surface insulation resistance,
per lamination.

The magnetic materials used for tape cores and laminations are coated with an insulating material. The
insulating coating is applied to reduce eddy currents. The American Iron and Steel Institute (AISI) has set
up insulation standards for transformer steels used in different applications. High permeability, nickel-iron
cores are very strain sensitive. Manufacturers of these cores normally have their own proprietary,
insulating material.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Laminations
Laminations are available in scores of different shapes and sizes. The punch press technology for
fabricating laminations has been well-developed. Most lamination sizes have been around forever. The
most commonly used laminations are the El, EE, FF, UI, LL, and the DU, as shown in Figure 3-7. The
laminations differ from each other by the location of the cut in the magnetic path length. This cut
introduces an air gap, which results in the loss of permeability. To minimize the resulting air gap, the
laminations are generally stacked in such a way the air gaps in each layer are staggered.

El, Laminations EE, Laminations FF, Laminations

UI, Laminations LL, Laminations DU, Laminations

Figure 3-7. Commonly Used, Lamination Shapes.

There are bobbins and brackets for almost all standard stacking dimensions. Most of the El lamination is
the scrapless. The name, scrapless, is derived from shapes that are punched with minimum waste, as shown
in Figure 3-8.
i A I

El, Laminations E, Laminations

Laminations

Figure 3-8. Typical, Scrapless El Lamination.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Annealing and Stress-Relief
One of the most important parameters in transformer steels is permeability. Any stress or strain of the
magnetic materials will have an impact on the permeability. The resulting stress could cause higher
magnetizing current, or a lower inductance. When the transformer is being assembled (in the stacking
process), and a lamination is bent (does not return to its original shape), that lamination has been stressed
and should be replaced.

Some of the important magnetic properties are lost due to stress and strain after stamping, shearing and
slitting. These properties that have been lost or seriously reduced can be restored to the magnetic materials
by annealing. Basically, stress relief is accomplished by heating (annealing) the magnetic material to
prescribed temperature, (depending on the material), followed by cooling to room temperature. The entire
annealing process is a delicate operation. The annealing must be done under controlled conditions of time,
temperature and the ambient atmosphere that will avoid, even minute, adverse changes in the chemistry of
the steel.

Stacking Laminations and Polarity

The edges of the magnetic material that have been stamped, sheared, or slit, will have a burr, as shown in
Figure 3-9. The quality of the equipment will keep the burr to a minimum. This burr now gives the
lamination a polarity. When a transformer is being stacked, the lamination build is normally sized by
dimensions, or it just fills the bobbin.
Lamination
Worn Die

Expanded View

Bun-
=t
Figure 3-9. Expanded View, Showing Lamination Burr.

If the laminations are stacked correctly, all of the burred ends will be aligned. If the laminations are
stacked randomly, such as the burr ends facing each other, then, the stacking factor would be affected. The
stacking factor has a direct impact on the cross-section of the core. The end result would be less iron. This
could lead to premature saturation, as an increase in the magnetizing current, or a loss of inductance.

There are several methods used in stacking transformer laminations. The most common technique used in
stacking laminations is the alternate method. The alternate method is where one set of laminations, such as
an E and an I, are assembled. Then, the laminations are reversed, as shown in Figure 3-10. This technique,
used in stacking, provides the lowest air gap and the highest permeability. Another method for stacking

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 laminations is to interleave two-by-two, also shown in Figure 3-10. The second method of stacking would
be in groups of two or more. This is done to cut assembly time. The loss in performance in stacking, other
than one by one, is the increase in magnetizing current and a loss of permeability.

Laminations E and I

Interleave 1 x 1 Interleave 2 x 2

Figure 3-10. Methods for Stacking Laminations.

Flux Crowding
When laminations are stacked, as shown in Figure 3-11, there is flux crowding. This flux crowding is
caused by the difference in spacing between the E, I, and the adjacent lamination. The adjacent lamination
has a minimum air gap, which translates into a higher permeability.

Laminations E and I

Flux Crowdingg , i

\ , -| 1
\
> i i\i
Minute Air Gap —^~ —' %!/
:= ===

-
jli izi |1{ ^ 53S

-\
i
I
i i i
= =
Flux = =

II
"
^ \ r L
—
J

Figure 3-11. Flux Crowding, when Laminations are Interleaved.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Exciting Current

The flux will skirt the low permeability air gap and migrate into the adjacent lamination, causing flux
crowding in that lamination. Eventually, this crowding will cause saturation in that portion of the
lamination, and the excitation current will rise. After that portion of the lamination has saturated, the flux
will migrate back to the lower permeability segment of the lamination from where it left. This effect can be
easily viewed by observing the B-H loops at low and high flux densities, and comparing them with a
toroidal core of the same material, with a minimum air gap, as shown in Figure 3-12. The B-H loop, along
with the magnetizing current, Im, of a toroidal core, is shown in Figure 3-12A. The toroidal core, with its
inherit minimum air gap, will have almost a square of current. Using the same material in lamination form
will exhibit a B-H loop, and a magnetizing current, Im, similar to Figure 3-12B operating at low flux
densities. Increasing the excitation will cause premature saturation of the lamination, as seen by the non-
linear, exciting current, as shown in Figure 3-12C

AB
H

Figure 3-12. Comparing the Exciting Currents and Three B-H Loops.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Most finished transformers or inductors will have some sort of bracket, such as an L bracket, end bells, a
channel bracket or maybe a bolt through the mounting holes to the chassis. When transformers are being
assembled, there is a certain amount of attention that has to be used to get proper performance. The
insulation material used to coat the lamination is normally very durable, but it can be scratched off and
degrade the performance. When brackets are used in the transformer assembly, as shown in Figure 3-13,
care must be taken on how the bolts and brackets are put together. The transformer assembly bolts, shown
in Figure 3-13, should be the recommended size for the mounting hole and use all of the required hardware.
This hardware should include the correct bolt size and length, and correct surface washer, lock washer and
nut. Also, included in this hardware, should be fiber shoulder washers and proper sleeving to cover the bolt
threads. If insulating hardware is not used, there is a good chance of a partial, shorted turn. The continuity
for this partial turn can be created through the bolts and bracket, or the bolts, bracket, and the chassis. This
partial shorted turn will downgrade the performance of the transformer.

Sleeving Laminations
Shoulder Washer

Bolt

Air Gap Material

Fringing Flux

Mounting Bracket Mounting Bracket

Butt Stack

Figure 3-13. Lamination Mounting Hardware.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Tape Wound C, EE, and Toroidal Cores
Tape wound cores are constructed by winding around a mandrel, a magnetic material in the form of a
preslit tape, as shown in Figure 3-14. This tape material comes in all of the iron alloys, plus the amorphous
materials. The tape thickness varies from 0.0005 inch (0.0127 mm) to 0.012 inch (0.305 mm). The
advantage of this type of construction is that the flux is parallel with the direction of rolling of the magnetic
material. This provides the maximum utilization of flux with the minimum of magnetizing force. There
are two disadvantages in this type of construction. When the core is cut in half, as shown in Figure 3-15,
the mating surface has to be ground, lapped, and then, acid-etched. This is done to provide a smooth
mating surface with the minimum of air gap and the maximum of permeability. The other disadvantage is
when the cores are reassembled, the method used is normally done with a band and buckle, and this
procedure requires a little skill to provide the right alignment and correct tension, as shown in Figure 3-16.
The C cores are impregnated for strength, prior to being cut. The cut C core can be used in many
configurations in the design of a magnetic component, as shown in Figure 3-17. The EE cores are
constructed in the same way as C cores, but they have an additional overwind, as shown in Figure 3-18.
The assembled three-phase transformer is shown in Figure 3-19.

Magnetic Material (Tape)

Magnetic Material (Tape)

C Core Construction Toroidal Core

Mandrel Construction "

Mandrel

Figure 3-14. Tape Cores Being Wound on a Mandrel.

Cut C Core
Mating Surface

Figure 3-15. Two Halves of a Cut C Core.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Cut C Core
Banding Material

Buckle

Figure 3-16. Banding the Cut C Core.

Single Core, Single Coil Single Core, Dual Coils Dual Cores, Single Coil
Core

Coil
c
Coil Coil
core y

Coil
<-ore

—'—
ji

Figure 3-17. Three Different C Core Configurations.

Overwind 7^ ^ \ Band -
^/JEV —T —-\
C Core
X

-•
"\ s~ s Buckle
"\ ( r 1I
Coil Cnil Coil
^

/V / Core —*( ( ) I
V.
V

J V ^ J

Figure 3-18. Three-Phase, Cut EE Core. Figure 3-19. Typical, Assembled EE Cut Core.

Tape Toroidal Cores

Tape toroidal cores are constructed in the same way as tape C cores, by winding the magnetic material
around a mandrel, in the form of a preslit tape. This tape material comes in all of the iron alloys, plus the
amorphous materials. The tape thickness varies from 0.000125 inch (0.00318 mm) to 0.012 inch (0.305
mm). The tape toroid is normally offered in two configurations, cased and encapsulated, as shown in
Figure 3-20. The cased toroid offers superior electrical properties and stress protection against winding.
The encapsulated cores are used when not all of the fine magnetic properties are important to the design,
such as in power transformers.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Enclosure
Cased Toroid Caseless Toroid

Figure 3-20. Outline of a Cased and a Caseless Toroidal Core.

Toroidal, Powder Core

Powder cores, as shown in Figure 3-21, are very unique. They give the engineer another tool to speed the
initial design. Powder cores have a built-in air gap. They come in a variety of materials and are very stable
with time and temperature. The cores are manufactured with good engineering aids. Manufacturers
provide catalogs for their cores, listing not only the size, but also permeability and Millihenrys per 1000
turns. The data is presented to the engineer in such a way that it takes the minimum amount of time to have
a design that will function.

OD

Figure 3-21. Outline of a Powder Toroidal Core.

Stacking Factors
The standard stacking factors for tape cores, wound cut cotes and laminations are shown in Table 3-1.

Table 3-1. Standard Stacking Factors.

Thickness Laminations
Tape Cores Wound Cut Cores (S.F.)2
mils Butt Stack Interleave 1x1
0.125 0.250 0.062
0.250 0.375 0.141
0.500 0.500 0.250
1.000 0.750 0.830 0.562
2.000 0.850 0.890 0.722
4.000 0.900 0.900 0.900 0.800 0.810
6.000 0.900 0.900 0.850 0.810
12.000 0.940 0.950 0.884
14.000 0.940 0.950 0.950 0.900 0.902
18.000 0.950 0.900 0.810
25.000 0.950 0.920 0.846

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for El Laminations
Laminations are still one of the most widely-used cores in power conversion. The dimensional outline for
El laminations and an assembled transformer is shown in Figure 3-22. Dimensional data for El laminations
is given in Table 3-2; design data is given in Table 3-3.
i . G i I D

Coil

E and I, Laminations Channel Bracket Assembly

Figure 3-22. El Lamination Outline.

Table 3-2. Dimensional Data for El Laminations.

El, Laminations, (Tempel) 14 mil

Part D E F G Part D E F G
No. cm cm cm cm No. cm cm cm cm
EI-375 0.953 0.953 0.794 1.905 EM 12 2.857 2.857 1.429 4.286
EI-021 1.270 1.270 0.794 2.064 El- 125 3.175 3.175 1.588 4.763
EI-625 1.588 1.588 0.794 2.381 EI-138 3.493 3.493 1.746 5.239
EI-750 1.905 1.905 0.953 2.857 El- 150 3.810 3.810 1.905 5.715
EI-875 2.223 2.223 1.111 3.333 El- 175 4.445 4.445 2.223 6.668
El- 100 2.540 2.540 1.270 3.810 EI-225 5.715 5.715 2.858 8.573

Table 3-3. Design Data for 14 mil El Laminations.

El, Laminations, (Tempel) 14 mil

Part wtcu wtfe MLT MPL wa Ac wa AP Kg A,
No. grams grams cm cm Ac cm cm2 cm4 cm5 cm
EI-375 36.1 47.2 6.7 7.3 1.754 0.862 1.512 1.303 0.067 46.2
EI-021 47.6 94.3 8.2 8.3 1.075 1.523 1.638 2.510 0.188 62.1
EI-625 63.5 170.0 9.5 9.5 0.418 2.394 1.890 4.525 0.459 83.2
EI-750 108.8 296.0 11.2 11.4 0.790 3.448 2.723 9.384 1.153 120.0
EI-875 171.0 457.0 13.0 13.3 0.789 4.693 3.705 17.384 2.513 163.0
El- 100 254.0 676.0 14.8 15.2 0.790 6.129 4.839 29.656 4.927 212.9
El-112 360.0 976.0 16.5 17.2 0.789 7.757 6.124 47.504 8.920 269.4
El- 125 492.0 1343.0 18.3 19.1 0.789 9.577 7.560 72.404 15.162 333.0
EI-138 653.0 1786.0 20.1 21.0 0.789 11.588 9.148 106.006 24.492 403.0
EI-150 853.0 2334.0 22.0 22.9 0.789 13.790 10.887 150.136 37.579 479.0
El- 175 1348.0 3711.0 25.6 26.7 0.789 18.770 14.818 278.145 81.656 652.0
EI-225 2844.0 7976.0 32.7 34.3 0.789 31.028 24.496 760.064 288.936 1078.0

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for UI Laminations
The dimensional outline for UI laminations and an assembled transformer is shown in Figure 3-23.
Dimensional data for UI laminations is given in Table 3-4; design data is given in Table 3-5.

o o i
e
1, H
1

W,
G Coil#l Coil#2

1
1
o o 1 H
e e
^ ^

F F F D Side View End View

UI, Laminations UI Transformer Assembled

Figure 3-23. UI Lamination Outline.

Table 3-4. Dimensional Data for UI Laminations.

UI, Standard Laminations 14 mil

Part D E F G H Part D E F G H
No. cm cm cm cm cm No. cm cm cm cm cm
50UI 1.270 1.270 1.270 3.810 1.270 125UI 3.175 3.175 3.175 9.525 3.175
60UI 1.429 1.429 2.223 5.398 1.429 150UI 3.810 3.810 3.810 11.430 3.810
75UI 1.905 1.905 1.905 5.715 1.905 180UI 4.572 4.572 4.572 11.430 4.572
100UI 2.540 2.540 2.540 7.620 2.540 240UI 6.096 6.096 6.096 15.240 6.096

Table 3-5. Design Data for 14 mil UI Laminations.

UI, Standard Laminations 14 mil

Part wtcu Wtfe MLT MPL wa Ac wa AP KK A,
4
No. grams grams cm cm Ac cm2 cm2 cm cm5 cm2
50UI 132 173 7.68 15.24 3.159 1.532 4.839 7.414 0.592 110
60UI 418 300 9.81 18.10 6.187 1.939 11.996 23.263 1.839 209
75U1 434 585 11.22 22.86 3.157 3.448 10.887 37.534 4.614 247
100UI 1016 1384 14.76 30.48 3.158 6.129 19.355 118.626 19.709 439
125UI 1967 2725 18.29 38.10 3.158 9.577 30.242 289.614 60.647 685
150UI 3413 4702 22.04 45.72 3.158 13.790 43.548 600.544 150.318 987
180UI 4884 7491 26.28 50.29 2.632 19.858 52.258 1037.740 313.636 1296
240UI 11487 17692 34.77 67.06 2.632 35.303 92.903 3279.770 1331.997 2304

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for LL Laminations
The dimensional outline for LL laminations and an assembled transformer is shown in Figure 3-24.
Dimensional data for LL laminations is given in Table 3-6; design data is given in Table 3-7.

o O i
i H
t

Coil#l Coil#2
\
i
o O 1 H
-* *-• -^ *- -^ fc-

E F E D
LL, Laminations LL Transformer Assembly
Figure 3-24. LL Lamination Outline.

Table 3-6. Dimensional Data for 14 mil LL Laminations.

LL, Standard Laminations 14 mil

Part D E F G H Part D E F G H
No. cm cm cm cm cm Number cm cm cm cm cm
141L 0.635 0.635 1.270 2.858 0.635 104L 1.270 1.270 1.984 5.555 1.270
108L 1.031 1.031 0.874 3.334 1.111 105L 1.270 1.270 1.905 6.826 1.270
250L 1.031 1.031 0.874 5.239 1.111 102L 1.429 1.429 1.588 5.398 1.429
101L 1.111 1.111 1.588 2.858 1.111 106L 1. 429 1.429 2.223 5.398 1.429
7L 1.270 1.270 1.270 3.810 1.270 107L 1.588 1.588 2.064 6.350 1.588
4L 1.270 1.270 1.905 3.810 1.270

Table 3-7. Design Data for 14 mil LL Laminations.

LL, Standard Laminations 14 mil

Part wtcu wtft MLT MPL wa Ac wa AP Kg A,
2 2 4
No. grams grams cm cm Ac cm cm cm cm5 cm2
141L 63.8 31.3 4.9 10.8 9.473 0.383 3.629 1.390 0.043 55.2
108L 61.2 97.9 5.9 12.7 2.884 1.010 2.913 2.943 0.201 70.3
250L 96.1 127.1 5.9 16.5 4.532 1.010 4.577 4.624 0.316 92.0
101L 118.5 115.9 7.3 13.3 3.867 1.173 4.536 5.322 0.340 97.3
7L 132.2 173.9 7.7 15.2 3.159 1.532 4.839 7.414 0.592 109.7
4L 224.0 185.2 8.7 16.5 4.737 1.532 7.258 11.121 0.785 141.9
104L 344.9 228.0 8.8 20.2 7.193 1.532 11.020 16.885 1.176 180.2
105L 401.3 256.5 8.7 22.5 8.488 1.532 13.004 19.925 1.407 199.4
102L 268.6 284.1 8.8 19.7 4.419 1.939 8.569 16.617 1.462 167.6
106L 418.6 302.1 9.8 21.0 6.187 1.939 11.996 23.263 1.839 208.8
107L 475.2 409.5 10.2 23.2 5.474 2.394 13.105 31.375 2.946 235.8

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for DU Laminations

The dimensional outline for DU laminations and an assembled transformer is shown in Figure 3-25.
Dimensional data for DU laminations is given in Table 3-8; design data is given in Table 3-9.

ti o 0 O o
H
ii H = 2E
t I
^t
^a
G
\r
A
tk
H r /
' O
^ : | =

+*
E F E «—
•4

DU, Laminations DU Transformer Assembly

Figure 3-25. DU Lamination Outline.

Table 3-8. Dimensional Data for 14 mil DU Laminations.

DU, Standard Laminations 14 mil

Part D E F G H Part D E F G H
No. cm cm cm cm cm No. cm cm cm cm cm
DU-63 0.159 0.159 0.318 0.794 0.318 DU-39 0.953 0.953 0.953 2.858 1.905
DU-124 0.318 0.318 0.476 1.191 0.635 DU-37 0.953 0.953 1.905 3.810 1.905
DU-18 0.476 0.476 0.635 1.588 0.953 DU-50 1.270 1.270 2.540 5.080 2.540
DU-26 0.635 0.635 0.635 1.905 1.270 DU-75 1.905 1.905 3.810 7.620 3.810
DU-25 0.635 0.635 0.953 2.064 1.270 DU-1 125 2.858 2.858 5.715 11.430 5.715
DU-1 0.635 0.635 0.953 3.810 1.270 DU-125 3.175 3.175 5.080 10.160 6.350

Table 3-9. Design Data for 14 mil DU Laminations.

DU, Standard Laminations 14 mil

Part wtcu wtre MLT MPL wa Ac wa AP KB A,
No. grams grams cm cm Ac cm2 cm2 cm4 cm3 cm2
DU-63 1.4 0.6 1.5 3.2 10.500 0.024 0.252 0.006 0.00003 4.2
DU-124 4.9 4.3 2.4 5.2 5.906 0.096 0.567 0.054 0.0009 11.8
DU-18 11.9 13.5 3.3 7.3 4.688 0.215 1.008 0.217 0.0057 23.4
DU-26 17.0 28.9 3.9 8.9 3.159 0.383 1.210 0.463 0.0180 33.9
DU-25 31.1 30.4 4.4 9.9 5.133 0.383 1.966 0.753 0.0260 44.3
DU-1 57.3 42.4 4.4 13.3 9.634 0.383 3.630 1.390 0.0479 60.9
DU-39 55.3 104.5 5.7 13.3 3.158 0.862 2.722 2.346 0.1416 76.2
DU-37 186.0 124.5 7.2 17.2 8.420 0.862 7.258 6.256 0.2992 134.3
DU-50 443.9 287.8 9.7 22.8 8.422 1.532 12.903 19.771 1.2524 238.0
DU-75 1467.0 985.2 14.2 34.3 8.420 3.448 29.032 100.091 9.7136 537.1
DU-1 125 4880.0 3246.0 21.0 51.4 8.421 7.757 65.322 506.709 74.8302 1208.0
DU-125 3906.0 3966.0 21.3 41.4 5.389 9.577 51.610 494.275 88.9599 1147.0

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for Three Phase Laminations

The dimensional outline for 3Phase El laminations and an assembled transformer is shown in Figure 3-26.
Dimensional data for 3Phase El laminations is given in Table 3-10; design data is given in Table 3-11.

t e e e
ET o o
G
F1
i
- ,. D

o o Coil #1 Coil #2 Coil #3

Wi

o o e e e
3Phase Laminations 3Phase Transformer Assembly

Figure 3-26. El Three Phase Laminations Outline.

Table 3-10. Dimensional Data for 14 mil El Three Phase Laminations.

3Phase, Standard Laminations, Thomas & Skinner 14 mil

Part D E F G Part D E F G
No. cm cm cm cm No. cm cm cm cm
0.250EI 0.635 0.635 0.871 2.858 l.OOOEI 2.540 2.540 3.810 7.620
0.375EI 0.953 0.953 1.270 3.175 1.200EI 3.048 3.048 3.048 7.620
0.500EI 1.270 1.270 1.588 3.493 1.500EI 3.810 3.810 3.810 9.525
0.562EI 1.427 1.427 1.588 5.398 1.800EI 4.572 4.572 4.572 11.430
0.625EI 1.588 1.588 1.984 5.634 2.400EI 6.096 6.096 6.096 15.240
0.875EI 2.223 2.223 2.779 6.111 3.600EI 9.144 9.144 9.144 22.860

Table 3-11. Design Data for 14 mil El Three Phase Laminations.

3Phase, Standard Laminations, Thomas & Skinner 14 mil

Part wtcu wtfe MLT wa Ac wa AP KR At
No. grams grams cm 2AC cm2 cm2 cm4 cm5 cm
0.250EI 57 54 4.3 3.251 0.383 2.49 1.43 0.051 53
0.375EI 134 154 6.2 2.339 0.862 4.03 5.21 0.289 102
0.500EI 242 324 8.2 1.810 1.532 5.54 12.74 0.955 159
0.562EI 403 421 8.8 2.213 1.936 8.57 24.88 2.187 207
0.625EI 600 706 10.1 2.334 2.394 11.18 40.13 3.816 275
0.875EI 1255 1743 13.9 1.809 4.693 16.98 119.53 16.187 487
1 .OOOEI 2594 2751 16.7 2.368 6.129 29.03 266.91 39.067 730
1.200EI 2178 3546 17.6 1.316 8.826 23.23 307.48 61.727 725
1.500E1 4266 6957 22.0 1.316 13.790 36.29 750.68 187.898 1132
1 .800EI 7326 12017 26.3 1.316 19.858 52.26 1556.61 470.453 1630
2.400EI 17230 28634 34.8 1.316 35.303 92.90 4919.66 1997.995 2899
3.600EI 58144 96805 52.2 1.316 79.432 209.03 24905.75 15174.600 6522

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for Tape Wound C Cores
The dimensional outline for C cores is shown in Figure 3-27. Dimensional data for C cores is given in
Table 3-12; design data is given in Table 3-13.

Figure 3-27. Tape C Core Dimensional Outline.

Table 3-12. Dimensional Data for Tape C Cores.

C Cores, Magnetic Metals, 2 mil

Part D E F G Part D E F G
No. cm cm cm cm No. cm cm cm cm
ML-002 0.635 0.476 0.635 1.588 ML-014 1.270 1.270 1.270 3.969
ML-004 0.635 0.635 0.635 2.223 ML-016 1.905 1.270 1.270 3.969
ML-006 1.270 0.635 0.635 2.223 ML-018 1.270 1.111 1.588 3.969
ML-008 0.953 0.953 0.953 3.016 ML-020 2.540 1.588 1.588 3.969
ML-010 1.588 0.953 0.953 3.016 ML-022 2.540 1.588 1.588 4.921
ML-012 1.270 1.111 1.270 2.858 ML-024 2.450 1.588 1.905 5.874

Table 3-13. Design Data for Tape C Cores.

C Cores, Magnetic Metals, 2 mil

Part wtcu
vv wtfe MLT MPL wa Ac wa AP K8 A,
No. grams grams cm cm Ac cm2 cm2 cm4 cm5 cm2
ML-002 13.0 13.0 3.6 6.4 3.747 0.269 1.008 0.271 0.0080 21.0
ML-004 19.8 22.6 3.9 8.3 3.933 0.359 1.412 0.507 0.0184 29.8
ML-006 27.2 45.2 5.4 8.3 1.967 0.718 1.412 1.013 0.0537 37.5
ML-008 58.4 72.5 5.7 11.8 3.556 0.808 2.874 2.323 0.1314 63.6
ML-010 73.5 120.8 7.2 11.8 2.134 1.347 2.874 3.871 0.2902 74.7
ML-012 95.1 121.7 7.4 12.7 2.891 1.256 3.630 4.558 0.3109 87.1
ML-014 137.7 170.4 7.7 15.6 3.513 1.435 5.041 7.236 0.5408 112.1
ML-016 160.5 255.6 9.0 15.6 2.341 2.153 5.041 10.854 1 .0443 126.8
ML-018 176.2 149.1 7.9 15.6 5.019 1.256 6.303 7.915 0.5056 118.9
ML-020 254.5 478.4 11.4 17.5 1.756 3.590 6.303 22.626 2.8607 182.0
ML-022 315.6 530.5 11.4 19.4 2.177 3.590 7.815 28.053 3.5469 202.0
ML-024 471.7 600.1 11.9 21.9 3.117 3.590 11.190 40.170 4.8656 244.8

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Dimensional Outline for Tape Wound EE Cores
The dimensional outline for EE cores is shown in Figure 3-28. Dimensional data for EE cores is given in
Table 3-14; design data is given in Table 3-15.

1
Ac

Figure 3-28. Tape EE Core Dimensional Outline.

Table 3-14. Dimensional Data for Tape EE Cores.

3Phase E Cores, National-Arnold Magnetics, 14 mil

Part D E F G Part D E F G
No. cm cm cm cm No. cm cm cm cm
CTA-25 1.905 1.905 1.905 2.858 CTA-12 3.810 2.540 2.381 6.350
CTA-22 3.175 1.429 1.905 5.239 CTA-20 5.715 2.540 2.540 6.350
CTA-17 3.175 1.746 1.905 6.350 CTA-03 4.445 2.540 3.493 9.843
CTA-14 3.175 2.223 2.381 4.763 CTA-15 5.080 3.493 2.540 7.620

Table 3-15. Design Data for Tape EE Cores.

3Phase E Cores, National-Arnold Magnetics, 14 mil

Part wtcu Wtfe MLT wa AC wa AP KR A,
No. grams grams cm 2AC cm cm cm4 cm5 cm2
CTA-25 326 686 11.2 0.789 3.448 5.44 28.16 3.461 261
CTA-22 682 1073 12.8 1.158 4.310 9.98 64.53 8.686 324
CTA-17 867 1422 13.4 1.148 5.266 12.10 95.56 14.977 400
CTA-14 916 1803 15.1 0.846 6.705 11.34 1 14.06 20.203 468
CTA-12 1391 2899 17.3 0.822 9.194 15.12 208.50 44..43S 613
CTA-20 1834 4420 21.3 0.585 13.790 16.13 333.64 86.347 737
CTA-03 3717 4597 20.3 1.602 10.730 34.38 553.15 117.079 993
CTA-15 2266 6544 22.0 0.574 16.860 19.35 489.40 150.340 956

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for Tape Wound Toroidal Cores
The dimensional outline for tape wound Toroidal cores is shown in Figure 3-29. Dimensional data for
cased tape wound Toroidal cores is given in Table 3-16; design data is given in Table 3-17.

Enclosure
Wa

OD

Cased Toroid Caseless Toroid

Figure 3-29. Tape Toroidal Core Dimensional Outline.

Table 3-16. Dimensional Data for Tape Toroidal Cores.

Toroidal Tape Cores, Mag netics 2 mil Iron Alloy (cased and coated)
Part OD ID HT Part OD ID HT Part OD ID HT
No. cm cm cm No. cm cm cm No. cm cm cm
52402 1.346 0.724 0.610 52057 2.134 1.359 0.610 52061 2.781 1.664 0.927
52107 1.651 1.041 0.610 52000 2.134 1.041 0.610 52004 3.429 2.286 0.927
52153 1.499 0.724 0.610 52155 1.659 0.884 0.927 52076 2.794 1.334 0.762
52056 1.816 1.041 0.610 52176 2.134 1.041 0.927 52007 2.794 1.334 0.927

Table 3-17. Design Data for Tape Toroidal Cores.

Toroidal Tape Cores, Magnetics 2 mil Iron Alloy (cased)

Part wtcu wtfc MLT MPL wa AC wa AP Kg A,
No. grams grams cm cm Ac cm cm2 cm4 cm5 cm2
52402 2.84 0.50 2.16 3.25 18.727 0.022 0.412 0.00906 0.0000388 9.80
52107 6.76 0.70 2.30 4.24 38.682 0.022 0.851 0.01872 0.0000717 15.50
52153 3.20 1.10 2.20 3.49 9.581 0.043 0.412 0.01770 0.0001400 11.20
52056 7.40 1.50 2.40 4.49 19.791 0.043 0.851 0.03660 0.0002592 16.80
52057 13.80 1.80 2.70 5.48 33.744 0.043 1.451 0.06237 0.0003998 23.70
52000 8.10 3.30 2.70 4.99 9.895 0.086 0.851 0.07320 0.0009384 20.60
52155 6.10 2.60 2.80 3.99 7.140 0.086 0.614 0.05278 0.0006461 16.00
52176 9.70 6.50 3.20 4.99 4.977 0.171 0.851 0.14554 0.0031203 23.30
52061 28.70 9.10 3.70 6.98 12.719 0.171 2.175 0.37187 0.0068597 40.30
52004 61.70 11.70 4.20 8.97 24.000 0.171 4.104 0.70184 0.0113585 62.20
52076 17.20 9.50 3.50 6.48 7.244 0.193 1.398 0.26975 0.0060284 34.60
52007 18.50 12.70 3.70 6.48 5.440 0.257 1.398 0.35920 0.0099305 36.40

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for EE Ferrite Cores

The dimensional outline for EE ferrite cores is shown in Figure 3-30. Dimensional data for EE ferrite cores
is given in Table 3-18; design data is given in Table 3-19.

A i L
c

G
1 \
ii
\
E B A \
\r
>\ \
\
\ \r
Vra
•^ ^
^
c fe-
D
EE Ferrite Cores Perspective View

Figure 3-30. Dimension Outline for EE Ferrite Cores.

Table 3-18. Dimensional Data for EE Ferrite Cores.

EE, Ferrite Cores (Magnetics)

Part A B C D E G Part A B C D E G
No. cm cm cm cm cm cm No. cm cm cm cm cm cm
EE-187 1.930 1.392 1.620 0.478 0.478 1.108 EE-21 4.087 2.832 3.300 1.252 1.252 2.080
EE-2425 2.515 1.880 1.906 0.653 0.610 1.250 EE-625 4.712 3.162 3.940 1.567 1.567 2.420
EE-375 3.454 2.527 2.820 0.935 0.932 1.930 EE-75 5.657 3.810 4.720 1.880 1.880 2.900

Table 3-19. Design Data for EE Ferrite Cores.

EE, Ferrite Cores (Magnetics)

Part Wvv
tcu w
tfe MLT MPL a Ac waw AP
Kg A, *AL
4
No. grams grams cm cm Ac cm2 cm cm4 cm5 cm2 mh/lK
EE-187 6.8 4.4 3.8 4.01 2.219 0.228 0.506 0.116 0.0028 14.4 500
EE-2425 13.9 9.5 4.9 4.85 2.068 0.384 0.794 0.305 0.0095 23.5 767
EE-375 36.4 33.0 6.6 6.94 1.875 0.821 1.539 1.264 0.0624 45.3 1167
EE-21 47.3 57.0 8.1 7.75 1.103 1.490 1.643 2.448 0.1802 60.9 1967
EE-625 64.4 103.0 9.4 8.90 0.808 2.390 1.930 4.616 0.4700 81.8 2767
EE-75 111.1 179.0 11.2 10.70 0.826 3.390 2.799 9.487 1.1527 118.0 3467
*This AL value has been normalized for a permeability of IK. For a close approximation of AL for other
values of permeability, multiply this AL value by the new permeability in kilo-perm. If the new
permeability is 2500, then use 2.5.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for EE and El Planar, Ferrite Cores
The dimensional outline for EE and El planar ferrite cores is shown in Figure 3-31. Dimensional data for
EE and El planar ferrite cores is given in Table 3-20; design data is given in Table 3-21.

Matting Set E o r l
A
c

B A *

•« >-
C
EE or El Planar Ferrite Cores Perspective View

Figure 3-31. Dimension Outline for EE, El Planar Ferrite Cores.

Table 3-20. Dimensional Data for EE, El Planar Ferrite Cores.

EE&EI/LP, Ferrite Cores (Magnetics)

Part A B C D E G Part A B C D E G
No. cm cm cm cm cm cm No. cm cm cm cm cm cm
EI-41805 1.800 1.370 0.638 1.000 0.398 0.188 EI-43208 3.175 2.490 0.953 2.032 0.635 0.305
EE-41805 1.800 1.370 0.796 1.000 0.398 0.376 EE-43208 3.175 2.490 1.270 2.032 0.635 0.610
EI-42216 2.160 1.610 0.867 1.590 0.508 0.297 EI-44310 4.318 3.440 1.395 2.790 0.813 0.533
EE-42216 2.160 1.610 1.144 1.590 0.508 0.610 EE-44310 4.318 3.440 1.906 2.790 0.813 1.066

Table 3-21. Design Data for EE, El Planar Ferrite Cores.

EE&EI/LP, Ferrite Cores (Mag netics)

Part W
vv
tcu wtfe MLT MPL wa Ac wa AP KB A, *AL
4
No. grams grams cm cm Ac cm2 cm cm cm5 cm mh/lK
EI-41805 1.5 4.1 4.7 2.03 0.2269 0.401 0.091 0.0366 0.00124 10.4 1737
EE-41805 3.1 4.9 4.7 2.42 0.4564 0.401 0.183 0.0715 0.00248 11.6 1460
EI-42216 3.8 10.4 6.5 2.58 0.2035 0.806 0.164 0.1319 0.00651 17.8 2592
EE-42216 7.8 13.0 6.5 3.21 0.4169 0.806 0.336 0.2709 0.01337 20.5 2083
EI-43208 8.9 22.0 8.9 3.54 0.224 1.290 0.289 0.3649 0.02126 33.4 3438
EE-43208 17.8 26.0 8.9 4.17 0.4388 1.290 0.566 0.7299 0.04253 37.9 2915
EI-44310 29.7 58.0 11.9 5.06 0.3084 2.270 0.700 1.5892 0.12085 65.4 4267
EE-44310 59.4 70.8 11.9 6.15 0.6167 2.270 1.400 3.1784 0.24170 75.3 3483
This AL value has been normalized for a permeability of 1 K. For a close approximation of AL for other values of
permeability, multiply this AL value by the new permeability in kilo-perm. If the new permeability is 2500, then
use 2.5.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for EC, Ferrite Cores

The dimensional outline for EC ferrite cores is shown in Figure 3-32. Dimensional data for EC ferrite
cores is given in Table 3-22; design data is given in Table 3-23.

u—
\
G
\

B A

X
W
vv

D
EC Ferrite Core Perspective View

Figure 3-32. Dimension Outline for EC Ferrite Cores.

Table 3-22. Dimensional Data for EC Ferrite Cores.

EC, Ferrite Cores (Magnetics)

Part A B C D E G
No. cm cm cm cm cm cm
EC-35 3.450 2.270 3.460 0.950 0.950 2.380
EC-41 4.060 2.705 3.901 1.161 1.161 2.697
EC-52 5.220 3.302 4.841 1.340 1.340 3.099
EC-70 7.000 4.450 6.900 1.638 1.638 4.465

Table 3-23. Design Data for EC Ferrite Cores.

EC, Ferrite Cores (Magnetics)

Part wtcu wtfe MLT MPL wa AC wa Ap Kg A, *AL
2
No. grams grams cm cm Ac cm cm2 4
cm cm5 cm2 rnh/lK
EC-35 35.1 36.0 6.3 7.59 2.213 0.710 1.571 1.115 0.050 50.2 1000
EC-41 55.4 52.0 7.5 8.76 1.964 1.060 2.082 2.207 0.125 67.6 1233
EC-52 97.8 111.0 9.0 10.30 2.156 1.410 3.040 4.287 0.267 106.5 1680
EC-70 256.7 253.0 11.7 14.10 2.927 2.110 6.177 13.034 0.941 201.7 1920
*This AL value has been normalized for a permeability of 1 K. For a close approximation of AL for other
values of permeability, multiply this AL value by the new permeability in kilo-perm. If the new permeability
is 2500, then use 2.5.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for ETD, Ferrite Cores
The dimensional outline for ETD ferrite cores is shown in Figure 3-33. Dimensional data for ETD ferrite
cores is given in Table 3-24; design data is given in Table 3-25.

G \

B A

4 *
C D
ETD Ferrite Core Perspective View

Figure 3-33. Dimension Outline for ETD Ferrite Cores.

Table 3-24. Dimensional Data for ETD Ferrite Cores.

ETD, Ferrite Cores (Ferroxcube)

Part A B C D E G Part A B C D E G
No. cm cm cm cm cm cm No. cm cm cm cm cm cm
ETD-29 3.060 2.270 3.160 0.980 0.980 2.200 ETD-49 4.980 3.610 4.940 1.670 1.670 3.540
ETD-34 3.500 2.560 3.460 1.110 1.110 2.360 ETD-54 5,450 4.120 5.520 1.890 1.890 4.040
ETD-39 4.000 2.930 3.960 1.280 1.280 2.840 ETD-59 5.980 4.470 6.200 2.165 2.165 4.500
ETD-44 4.500 3.250 4.460 1.520 1.520 3.220

Table 3-25. Design Data for ETD Ferrite Cores.

ETD, Ferrite Cores (Ferroxcube)

Part wtcu wtfe MLT MPL wa Ac wa AP KR At *AL
No. grams grams cm cm Ac cm2 cm2 cm4 cm5 cm mh/lK
ETD-29 32.1 28.0 6.4 7.20 1.865 0.761 1.419 1.0800 0.0517 42.5 1000
ETD-34 43.4 40.0 7.1 7.87 1.757 0.974 1.711 1.6665 0.0911 53.4 1182
ETD-39 69.3 60.0 8.3 9.22 1.871 1.252 2.343 2.9330 0.1766 69.9 1318
ETD-44 93.2 94.0 9.4 10.30 1.599 1.742 2.785 4.8520 0.3595 87.9 1682
ETD-49 126.2 124.0 10.3 11.40 1.627 2.110 3.434 7.2453 0.5917 107.9 1909
ETD-54 186.9 180.0 11.7 12.70 1.609 2.800 4.505 12.6129 1.2104 133.7 2273
ETD-59 237.7 260.0 12.9 13.90 1.410 3.677 5.186 19.0698 2.1271 163.1 2727
*This AL value has been normalized for a permeability of IK. For a close approximation of AL for other values of
permeability, multiply this AL value by the new permeability in kilo-perm. If the new permeability is 2500, then use
2.5.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for ETD/(low profile), Ferrite Cores

The dimensional outline for ETD/lp low profile ferrite cores is shown in Figure 3-34. Dimensional data for
ETD/lp low profile ferrite cores is given in Table 3-26; design data is given in Table 3-27.

•w
vv
a ~
c D
ETD-lp Ferrite Core Perspective View

Figure 3-34. Dimension Outline for ETD/lp Ferrite Cores.

Table 3-26. Dimensional Data for ETD/lp Ferrite Cores.

ETD/lp, Ferrite Cores (TSC Ferrite International)

Part A B C D E G
No. cm cm cm cm cm cm
ETD34(lp) 3.421 2.631 1.804 1.080 1.080 0.762
ETD39(lp) 3.909 3.010 1.798 1.250 1.250 0.762
ETD44(lp) 4.399 3.330 1.920 1.481 1.481 0.762
ETD49(lp) 4.869 3.701 2.082 1.631 1.631 0.762

Table 3-27. Design Data for ETD/lp Ferrite Cores.

ETD/lp, Ferrite Cores (TSC Ferrite International)

Part W
vv
tcu W tfc MLT MPL wa Ac wa Ap Kg A, *AL
2 2 4
No. grams grams cm cm Ac cm cm cm cm5 cm 2
mh/lK
ETD34(lp) 15.1 32.7 7.2 4.65 0.609 0.970 0.591 0.5732 0.0310 33.1 2382
ETD39(lp) 20.0 46.3 8.4 5.03 0.559 1.200 0.671 0.8047 0.0461 39.6 2838
ETD44(lp) 24.6 72.1 9.5 5.40 0.420 1.730 0.727 1.2583 0.0914 48.4 3659
ETD49(lp) 29.1 95.0 10.4 5.85 0.374 2.110 0.789 1.6641 0.1353 58.2 4120
*This AL value has been normalized for a permeability of 1 K. For a close approximation of AL for other values
of permeability, multiply this AL value by the new permeability in kilo-perm. If the new permeability is 2500,
then use 2.5.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for ER, Ferrite Cores
Surface Mount Device, SMD
The dimensional outline for ER ferrite cores is shown in Figure 3-35. Dimensional data for ER ferrite
cores is given in Table 3-28; design data is given in Table 3-29.

c ER Ferrite Core
D
Perspective View

Figure 3-35. Dimension Outline for ER Ferrite Cores.

Table 3-28. Dimensional Data for ER Ferrite Cores.

ER, Ferrite Cores (Ferroxcube)

Part A B C D E G Part A B C D E G
No. cm cm cm cm cm cm No. cm cm cm cm cm cm
ER9.5 0.950 0.750 0.490 0.500 0.350 0.320 ER42 4.200 3.005 4.480 1.560 1.550 3.090
ER11 1.100 0.870 0.490 0.600 0.425 0.300 ER48 4.800 3.800 4.220 2.100 1.800 2.940
ER35 3.500 2.615 4.140 1.140 1.130 2.950 ER54 5.350 4.065 3.660 1.795 1.790 2.220

Table 3-29. Design Data for ER Ferrite Cores.

ER, Ferrite Cores (Ferroxcube)

Part wtcu Wtfe MLT MPL wa Ac wa AP KB A, *AL
No. grams grams cm cm Ac cm2 cm2 cm4 cm5 cm mh/lK
ER9.5 0.6 0.7 2.700 1.42 0.842 0.076 0.0640 0.00486 0.000054 3.0 435
ER11 0.7 1.0 3.200 1.47 0.650 0.103 0.0670 0.00688 0.000090 3.7 609
ER35 56.7 46.0 7.300 9.08 2.190 1.000 2.1900 2.19037 0.120340 62.4 1217
ER42 72.9 96.0 9.100 9.88 1.189 1.890 2.2480 4.24867 0.352444 81.0 2000
ER48 120.7 128.0 11.500 10.00 1.185 2.480 2.9400 7.29120 0.626245 100.1 2478
ER54 101.9 122.0 1 1 .400 9.18 1.052 2.400 2.5250 6.06060 0.512544 96.2 2652
*This AL value has been normalized for a permeability of 1 K. For a close approximation of AL for other values
of permeability, multiply this AL value by the new permeability in kilo-perm. If the new permeability is 2500,
then use 2.5.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for EFD, Ferrite Cores
Surface Mount Device, SMD
The EFD cores, (Economic Flat Design), offer a significant advance in power transformer circuit
miniaturization. The dimensional outline for EFD ferrite cores is shown in Figure 3-36. Dimensional data
for EFD ferrite cores is given in Table 3-30; design data is given in Table 3-31.

EFD Ferrite Core Perspective View

Figure 3-36. Dimension Outline for EFD Ferrite Cores.

Table 3-30. Dimensional Data for EFD Ferrite Cores.

EFD, Ferrite Cores (Ferroxcube)

Part A B C D E G H
No. cm cm cm cm cm cm cm
EFD- 10 1.050 0.765 0.270 1.040 0.455 0.750 0.145
EFD- 15 1.500 1.100 0.465 1.500 0.530 1.100 0.240
EFD-20 2.000 1.540 0.665 2.000 0.890 1.540 0.360
EFD-25 2.500 1.870 0.910 2.500 1.140 1.860 0.520
EFD-30 3.000 2.240 0.910 3.000 1.460 2.240 0.490

Table 3-31. Design Data for EFD Ferrite Cores.

EFD, Ferrite Cores (Ferroxcube)

Part wtcu wtfe MLT MPL wa Ac wa AP K
s At *AL
2 2 4 5 2
No. grams grams cm cm Ac cm cm cm cm cm mh/lK
EFD- 10 0.8 0.90 1.8 2.37 1.611 0.072 0.116 0.00837 0.00013 3.3 254
EFD- 15 3.0 2.80 2.7 3.40 2.093 0.150 0.314 0.04703 0.00105 7.3 413
EFD-20 6.8 7.00 3.8 4.70 1.616 0.310 0.501 0.15516 0.00506 13.3 565
EFD-25 11.5 16.00 4.8 5.70 1.171 0.580 0.679 0.39376 0.01911 21.6 957
EFD-30 17.0 24.00 5.5 6.80 1.267 0.690 0.874 0.60278 0.03047 28.9 913
*This AL value has been normalized for a permeability of IK. For a close approximation of AL for other
values of permeability, multiply this AL value by the new permeability in kilo-perm. If the new permeability
is 2500, then use 2.5.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for EPC, Ferrite Cores
Surface Mount Device, SMD
The dimensional outline for EPC ferrite cores is shown in Figure 3-37. Dimensional data for EPC ferrite
cores is given in Table 3-32; design data is given in Table 3-33.

EPC Ferrite Core Perspective View

Figure 3-37. Dimension Outline for EPC Ferrite Cores.

Table 3-32. Dimensional Data for EPC Ferrite Cores.

EPC, Ferrite Cores (TDK)

Part A B C D E G H
No. cm cm cm cm cm cm cm
EPC- 10 1.020 0.760 0.340 0.810 0.500 0.530 0.190
EPC- 13 1.325 1.050 0.460 1.320 0.560 0.900 0.205
EPC- 17 1.760 1.430 0.600 1.710 0.770 1.210 0.280
EPC- 19 1.910 1.580 0.600 1.950 0.850 1.450 0.250
EPC-25 2.510 2.040 0.800 2.500 1.150 1.800 0.400
EPC-27 2.710 2.160 0.800 3.200 1.300 2.400 0.400
EPC-30 3.010 2.360 0.800 3.500 1.500 2.600 0.400

Table 3-33. Design Data for EPC Ferrite Cores.

EPC, Ferrite Cores (TDK)

Part W
vv
tcu wtfe MLT MPL wa AC wa AP Kg A, *AL
4
No. grams grams cm cm AC cm cm cm cm cm" mh/lK
EPC- 10 0.5 1.1 1.9 1.78 0.735 0.094 0.069 0.00647 0.000128 2.9 416
EPC- 13 2.0 2.1 2.5 3.06 1.768 0.125 0.221 0.02756 0.000549 5.9 363
EPC-17 4.9 4.5 3.4 4.02 1.750 0.228 0.399 0.09104 0.002428 10.2 479
EPC- 19 6.9 5.3 3.7 4.61 2.330 0.227 0.529 0.12014 0.002981 12.1 392
EPC-25 14.8 13.0 5.0 5.92 1.804 0.464 0.837 0.38837 0.014532 20.6 650
EPC-27 18.8 18.0 5.1 7.31 1.890 0.546 1.032 0.56347 0.024036 26.8 642
EPC-30 21.9 23.0 5.5 8.16 1.833 0.610 1.118 0.68198 0.030145 31.5 654
This AL value has been normalized for a permeability of IK. For a close approximation of AL for other values of
permeability, multiply this AL value by the new permeability in kilo-perm. If the new permeability is 2500, then use
2.5.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for PC, Ferrite Cores

The dimensional outline for PC ferrite pot cores is shown in Figure 3-38. Dimensional data for PC ferrite
pot cores is given in Table 3-34; design data is given in Table 3-35.

Ferrite Pot Core Perspective View

Figure 3-38. Dimension Outline for PC Ferrite Cores.

Table 3-34. Dimensional Data for PC Ferrite Cores.

PC, Ferrite Cores (Magnetics)

Part A B C E G Part A B C E G
No. cm cm cm cm cm Number cm cm cm cm cm
PC-40905 0.914 0.749 0.526 0.388 0.361 PC-42616 2.550 2.121 1.610 1.148 1.102
PC-41408 1.400 1.160 0.848 0.599 0.559 PC-43019 3.000 2.500 1.880 1.350 1.300
PC-41811 1.800 1.498 1.067 0.759 0.720 PC-43622 3.560 2.990 2.200 1.610 1.460
PC-42213 2.160 1.790 1.340 0.940 0.920 PC-44229 4.240 3.560 2.960 1.770 2.040

Table 3-35. Design Data for PC Ferrite Cores.

PC, Ferrite Cores (Magnetics)

Part wtcu Wtf, MLT MPL wa Ac wa AP K
A, *AL
cm*5
4
No. grams grams cm cm AC cm2 cm2 cm cm2 mh/lK
PC-40905 0.5 1.0 1.9 1.25 0.650 0.100 0.065 0.00652 0.000134 2.8 455
PC-41408 1.6 3.2 2.9 1.97 0.631 0.249 0.157 0.03904 0.001331 6.8 933
PC-41811 3.5 7.3 3.7 2.59 0.620 0.429 0.266 0.11413 0.005287 11.1 1333
PC-42213 6.2 13.0 4.4 3.12 0.612 0.639 0.391 0.24985 0.014360 16.4 1633
PC-42616 10.1 20.0 5.3 3.76 0.576 0.931 0.536 0.49913 0.035114 23.1 2116
PC-43019 16.7 34.0 6.3 4.50 0.550 1.360 0.748 0.97175 0.080408 31.9 2700
PC-43622 26.7 57.0 7.5 5.29 0.499 2.020 1.007 2.03495 0.220347 44.5 3400
PC-44229 55.9 104.0 8.6 6.85 0.686 2.660 1.826 4.85663 0.600289 67.7 4000
"This AL value has been normalized for a permeability of IK. For a close approximation of AL for other values of
permeability, multiply this AL value by the new permeability in kilo-perm. If the new permeability is 2500, then use
2.5.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for EP, Ferrite Cores
The EP ferrite cores are typically used in transformer applications. The shape of the assembly is almost
cubical, allowing high package densities on the PCB. The dimensional outline for EP ferrite cores is shown
in Figure 3-39. Dimensional data for EP ferrite cores is given in Table 3-36; design data is given in Table
3-37.

A, D

EP Ferrite Core Perspective View

Figure 3-39. Dimension Outline for EP Ferrite Cores.

Table 3-36. Dimensional Data for EP Ferrite Cores.

EP, Ferrite Cores (Magnetics)

Part A B C D E G
No. cm cm cm cm cm cm
EP-07 0.920 0.720 0.635 0.740 0.340 0.500
EP-10 1.150 0.920 0.760 1.030 0.345 0.720
EP-13 1.250 0.972 0.880 1.290 0.452 0.899
EP-17 1.798 1.160 1.100 1.680 0.584 1.118
EP-20 2.400 1.610 1.495 2.139 0.899 1.397

Table 3-37. Design Data for EP Ferrite Cores.

EP, Ferrite Cores (Magnetics)

Part wtcu wtfe MLT MPL wa Ac wa AP KB A, *AL
No. grams grams cm cm AC cm2 cm2 cm4 cm5 cm2 mh/lK
EP-07 1.4 1.4 1.8 1.57 0.922 0.103 0.095 0.00979 0.00022 3.5 413
EP-10 1.6 2.8 2.1 1.92 1.832 0.113 0.207 0.02339 0.00049 5.7 400
EP-13 2.0 5.1 2.4 2.42 1.200 0.195 0.234 0.04558 0.00148 7.7 667
EP-17 11.6 11.6 2.9 2.85 0.950 0.339 0.322 0.10915 0.00510 13.7 1033
EP-20 7.4 27.6 4.2 3.98 0.637 0.780 0.497 0.38737 0.02892 23.8 1667
*This AL value has been normalized for a permeability of 1 K. For a close approximation of AL for other
values of permeability, multiply this AL value by the new permeability in kilo-perm. If the new permeability is
2500, then use 2.5.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for PQ, Ferrite Cores

The PQ ferrite cores, (Power Quality), feature round center legs with rather small cross-sections. The
dimensional outline for PQ ferrite cores is shown in Figure 3-40. Dimensional data for PQ ferrite cores is
given in Table 3-38; design data is given in Table 3-39.

s -i

^
\ *-- i L

ii

E B A

1r

r
PQ Ferrite Core Perspective View

Figure 3-40. Dimension Outline for PQ Ferrite Cores.

Table 3-38. Dimensional Data for PQ Ferrite Cores.

PQ, Ferrite Cores (TDK)

Part A B C D E G Part A B C D E G
No. cm cm cm cm cm cm No. cm cm cm cm cm cm
PQ20/16 2.050 1.800 1.620 1.400 0.880 .030 PQ32/30 3.200 2.750 3.035 2.200 1.345 2.130
PQ20/20 2.050 1.800 2.020 1.400 0.880 .430 PQ35/35 3.510 3.200 3.475 2.600 1.435 2.500
PQ26/20 2.650 2.250 2.015 1.900 1.200 .150 PQ40/40 4.050 3.700 3.975 2.800 1.490 2.950
PQ26/25 2.650 2.250 2.475 1.900 1.200 .610 PQ50/50 5.000 4.400 4.995 3.200 2.000 3.610
PQ32/20 3.200 2.750 2.055 2.200 1.345 .150

Table 3-39. Design Data for PQ Ferrite Cores.

PQ, Ferrite Cores (TDK)

Part W
vv
tcu Wtfe MLT MPL wa Ac wa Ap KE A, *AL
4
No. grams grams cm cm Ac cm cm2 cm cm5 cm2 mh/lK
PQ20/16 7.4 13.0 4.4 3.74 0.765 0.620 0.474 0.294 0.0167 16.9 1617
PQ20/20 10.4 15.0 4.4 4.54 1.061 0.620 0.658 0.408 0.0227 19.7 1313
PQ26/20 31.0 31.0 5.6 4.63 0.508 1.190 0.604 0.718 0.0613 28.4 2571
PQ26/25 17.0 36.0 5.7 5.55 0.716 1.180 0.845 0.997 0.0832 32.6 2187
PQ32/20 18.9 42.0 6.6 5.55 0.475 1.700 0.808 1.373 0.1417 36.3 3046
PQ32/30 35.5 55.0 6.7 7.46 0.929 1.610 1.496 2.409 0.2326 46.9 2142
PQ35/35 59.0 73.0 7.5 8.79 1.126 1.960 2.206 4.324 0.4510 60.7 2025
PQ40/40 97.2 95.0 8.4 10.20 1.622 2.010 3.260 6.552 0.6280 77.1 1792
PQ50/50 158.5 195.0 10.3 11.30 1.321 3.280 4.332 14.209 1.8120 113.9 2800
*This AL value has been normalized for a permeability of IK. For a close approximation of AL for other values of
permeability, multiply this AL value by the new permeability in kilo-perm. If the new permeability is 2500, then
use 2.5.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for PQ/(low profile), Ferrite Cores

The PQ/lp cores are a cut down version of the standard PQ cores. The PQ/lp cores have a substantially
reduced total height. The dimensional outline for PQ ferrite cores is shown in Figure 3-41. Dimensional
data for PQ ferrite cores is given in Table 3-40; design data is given in Table 3-41.

\s i I
1 t
-* -

B A

1r

D
^
i
PQ Ferrite Core, low profile Perspective View

Figure 3-41. Dimension Outline for PQ/lp Ferrite Cores.

Table 3-40. Dimensional Data for PQ/lp Ferrite Cores.

PQ/1 >, Ferrite Cores (Ferrite International)

Part A B C D E G
No. cm cm cm cm cm cm
PQ20-14-141p 2.125 1.801 1.352 1.400 0.884 0.762
PQ26-16-141p 2.724 2.250 1.630 1.900 1.199 0.762
PQ32-17-221p 3.302 2.751 1.670 2.200 1.348 0.762
PQ35-17-261p 3.612 3.200 1.738 2.601 1.435 0.762
pQ40-18-281p 4.148 3.701 1.784 2.799 1.491 0.762

Table 3-41. Design Data for PQ/lp Ferrite Cores.

PQ/lp, Ferrite Cores (TSC Ferrite International)

Part wttu wtfe MLT MPL wa Ac wa AP KS A, *AL
2
No. grams grams cm cm
AC cm cm2 cm
4
cm5 cm2 mh/lK
PQ20-14-141p 5.4 12.5 4.4 3.2
0.563 0.620 0.349 0.217 0.0123 15.4 1948
PQ26-16-191p 7.9 28.0 5.6 3.9
0.336 1.190 0.400 0.477 0.0407 25.4 3170
PQ32-17-221p 12.5 39.4 6.6 4.8
0.315 1.700 0.535 0.909 0.0937 32.9 3659
PQ35-17-261p 17.8 44.9 7.4 5.3
0.343 1.960 0.672 1.318 0.1389 40.4 3893
PQ40-18-281p 24.9 63.5 8.3 5.8
0.419 2.010 0.842 1.692 0.1637 48.0 3850
This AL value has been normalized for a permeability of IK. For a close approximation of AL for other values of
permeability, multiply this AL value by the new permeability in kilo-perm. If the new permeability is 2500, then use
2.5.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for RM, Ferrite Cores

The RM cores, (Rectangular Modular), were developed for high Printed Circuit Board, (PCB), packing
densities. The dimensional outline for RM ferrite cores is shown in Figure 3-42. Dimensional data for RM
ferrite cores is given in Table 3-42; design data is given in Table 3-43.

RM Ferrite Core Perspective View

Figure 3-42. Dimension Outline for RM Ferrite Cores.

Table 3-42. Dimensional Data for RM Ferrite Cores.

RM, Ferrite Cores (TDK)

Part A B C E G Part A B C E G
No. cm cm cm cm cm No. cm cm cm cm cm
RM-4 0.963 0.815 1.04 0.38 0.72 RM-10 2.415 2.165 1.86 1.07 1.27
RM-5 1.205 1.04 1.04 0.48 0.65 RM-12 2.925 2.55 2.35 1.26 1.71
RM-6 1.44 1.265 1.24 0.63 0.82 RM-14 3.42 2.95 2.88 1.47 2.11
RM-8 1.935 1.73 1.64 0.84 1.1

Table 3-43. Design Data for RM Ferrite Cores.

RM, Ferrite Cores (TDK)

Part wtcu
vv wtfe MLT MPL wa Ac wa AP Ks A, *AL
No. grams grams cm cm Ac cm cm cm4 cm5 cm2 mh/lK
RM-4 1.1 1.7 2.0 2.27 1.121 0.140 0.157 0.0219 0.0006 489 5.9
RM-5 1.6 3.0 2.5 2.24 0.768 0.237 0.182 0.0431 0.0016 869 7.9
RM-6 2.9 5.5 3.1 2.86 0.710 0.366 0.260 0.0953 0.0044 11.3 1130
RM-8 7.3 13.0 4.2 3.80 0.766 0.640 0.490 0.3133 0.0191 20.2 1233
RM-10 13.2 23.0 5.3 4.40 0.709 0.980 0.695 0.6814 0.0502 29.6 1833
RM-12 24.4 42.0 6.2 5.69 0.788 1.400 1.103 1.5440 0.1389 44.6 2434
RM-14 39.9 70.0 7.2 6.90 0.830 1.880 1.561 2.7790 0.2755 62.8 2869
*This AL value has been normalized for a permeability of IK. For a close approximation of AL for other
values of permeability, multiply this AL value by the new permeability in kilo-perm. If the new permeability
is 2500, then use 2.5.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for RM/(low profile), Ferrite Cores
Surface Mount Device, SMD
The RM/lp ferrite cores are a cut down version of the standard RM cores. The dimensional outline for
RM/lp ferrite cores is shown in Figure 3-43. Dimensional data for RM/lp ferrite cores is given in Table 3-
44; design data is given in Table 3-45.

RM Ferrite Core, low profile Perspective View

Figure 3-43. Dimension Outline for RM/lp Ferrite Cores.

Table 3-44. Dimensional Data for RM/lp Ferrite Cores.

RM/lp, Ferrite Cores (Ferroxcube)

Part A B C E G Part A B C E G
No. cm cm cm cm cm No. cm cm cm cm cm
RM4/ILP 0.980 0.795 0.780 0.390 0.430 RM8/ILP 1.970 1.700 1.160 0.855 0.590
RM5/ILP 1.230 1.020 0.780 0.490 0.360 RM10/ILP 2.470 2.120 1.300 1.090 0.670
RM6S/LP 1.470 1.240 0.900 0.640 0.450 RM12/ILP 2.980 2.500 1.680 1.280 0.900
RM7/1LP 1.720 1.475 0.980 0.725 0.470 RM14/ILP 3.470 2.900 2.050 1.500 1.110

Table 3-45. Design Data for RM/lp Ferrite Cores.

RM/lp, Ferrite Cores (Ferroxcube)

Part W
vv
tcu Wtfe MLT MPL wa Ac wa AP KK A, AL
No. grams grams cm cm Ac cm2 cm2 cm4 cm5 cm2 mh/lK
RM4/ILP 0.6 1.5 2.0 1.73 0.770 0.113 0.087 0.00984 0.00022 5.0 609
RM5/ILP 0.9 2.2 2.5 1.75 0.525 0.181 0.095 0.01727 0.00049 6.9 1022
RM6S/LP 1.5 4.2 3.1 2.18 0.433 0.312 0.135 0.04212 0.00169 9.6 1380
RM7/ILP 2.3 6.0 3.6 2.35 0.444 0.396 0.176 0.06979 0.00306 12.7 1587
RM8/ILP 3.7 10.0 4.2 2.87 0.449 0.554 0.249 0.13810 0.00733 16.9 1783
RM10/ILP 6.4 17.0 5.2 3.39 0.426 0.809 0.345 0.27915 0.01736 25.0 2435
RM12/ILP 11.9 34.0 6.1 4.20 0.439 1.250 0.549 0.68625 0.05627 37.8 3087
RM14/ILP 19.5 55.0 7.1 5.09 0.463 1.680 0.777 1.30536 0.12404 52.5 3652
*This AL value has been normalized for a permeability of 1 K. For a close approximation of AL for other values of
permeability, multiply this AL value by the new permeability in kilo-perm. If the new permeability is 2500, then
use 2.5.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for DS, Ferrite Cores

The DS ferrite cores are similar to standard Pot Cores. These cores have a large opening to bring out many
strands of wire, which is convenient for high power and multiple outputs. The dimensional outline for DS
ferrite cores is shown in Figure 3-44. Dimensional data for DS ferrite cores is given in Table 3-46; design
data is given in Table 3-47.

DS Ferrite Core Perspective View

Figure 3-44. Dimension Outline for DS Ferrite Cores.

Table 3-46. Dimensional Data for DS Ferrite Cores.

DS, Ferrite Cores (Magnetics)

Part A B C D E G
No. cm cm cm cm cm cm
DS-42311 2.286 1.793 1.108 1.540 0.990 0.726
DS-42318 2.286 1.793 1.800 1.540 0.990 1.386
DS-42616 2.550 2.121 1.610 1.709 1.148 1.102
DS-43019 3.000 2.500 1.880 1.709 1.351 1.300
DS-43622 3.561 2.985 2.170 2.385 1.610 1.458
DS-44229 4.240 3.561 2.960 2.840 1.770 2.042

Table 3-47. Design Data for DS Ferrite Cores.

DS, Ferrite Cores (Magnetics)

Part W
vv
tcu wtfe MLT MPL wa Ac wa AP KB A, *AL
No. grams grams cm cm Ac cm cm2 cm
4
cm
5
cm
2
mh/lK
DS-42311 4.7 10.0 4.5 2.68 0.770 0.378 0.291 0.110 0.00368 16.2 1487
DS-42318 9.1 13.0 4.6 3.99 1.366 0.407 0.556 0.227 0.00800 21.1 1267
DS-42616 10.1 15.0 5.3 3.89 0.855 0.627 0.536 0.336 0.01593 23.1 1667
DS-43019 16.7 22.0 6.3 4.62 0.778 0.960 0.747 0.717 0.04380 31.9 1933
DS-43622 26.6 37.0 7.5 5.28 0.802 1.250 1.002 1.253 0.08404 44.2 2333
DS-44229 56.0 78.0 8.6 7.17 1.028 1.780 1.829 3.255 0.26917 67.7 2800
*This AL value has been normalized for a permeability of 1 K. For a close approximation of AL for other values of
permeability, multiply this AL value by the new permeability in kilo-perm. If the new permeability is 2500, then
use 2.5.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for UUR, Ferrite Cores

The UUR ferrite cores feature round legs with rather small cross sections. The round legs allow easy
winding with either wire or foil. U cores are used for power, pulse and high-voltage transformers. The
dimensional outline for UUR ferrite cores is shown in Figure 3-45. Dimensional data for UUR ferrite cores
is given in Table 3-48; design data is given in Table 3-49.

c wa D

/
/
j
r /\
1 X
F
A 1
G AC
r

m
1

\ J i.

UUR Ferrite Cores Perspective View

Figure 3-45. Dimension Outline for UUR Ferrite Cores.

Table 3-48. Dimensional Data for UUR Ferrite Cores.

UUR, Ferrite Cores (Magnetics)

Part A C D F G
No. cm cm cm cm cm
UUR-44121 4.196 4.120 1.170 1.910 2.180
UUR-44119 4.196 4.180 1.170 1.910 2.680
UUR-44125 4.196 5.080 1.170 1.910 3.140
UUR-44130 4.196 6.100 1.170 1.910 4.160

Table 3-49. Design Data for UUR Ferrite Cores.

UUR, Ferrite Cores (Magnetics)

Part wtcu Wtfe MLT MPL wa Ac wa AP Kg A, *AL
2 4 5
No. grams grams cm cm Ac cm cm cm cm cm mh/lK
UUR-44121 119.0 55.0 8.0 11.3 4.215 0.988 4.164 4.114 0.202 98.5 616
UUR-44119 146.2 54.0 8.0 12.1 5.619 0.911 5.119 4.663 0.211 102.9 710
UUR-44125 171.3 64.0 8.0 13.3 6.070 0.988 5.997 5.925 0.291 116.1 702
UUR-44130 227.0 75.0 8.0 15.3 8.043 0.988 7.946 7.850 0.386 134.9 610
*This AL value has been normalized for a permeability of 1 K. For a close approximation of AL for other
values of permeability, multiply this AL value by the new permeability in kilo-perm. If the new permeability is
2500, then use 2.5.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for UUS, Ferrite Cores

The UUS ferrite cores feature square or rectangular legs. U cores are used for power, pulse and high-
voltage transformers. The dimensional outline for UUS ferrite cores is shown in Figure 3-46. Dimensional
data for UUS ferrite cores is given in Table 3-50; design data is given in Table 3-51.

J L

1i
F
A
W.
ir (•J

1r

UUS Ferrite Core Perspective View

Figure 3-46. Dimension Outline for UUS Ferrite Cores.

Table 3-50. Dimensional Data for UUS Ferrite Cores.

UUS, Ferrite Cores (Ferroxcube)

Part A C D F G
No. cm cm cm cm cm
U10-08-03 1.000 1.640 0.290 0.435 1.000
U20- 16-07 2.080 3.120 0.750 0.640 1.660
U25-20-13 2.480 3.920 1.270 0.840 2.280
U30-25-16 3.130 5.060 1.600 1.050 2.980
U67-27-14 6.730 5.400 1.430 3.880 2.540
U93-76-16 9.300 15.200 1.600 3.620 9.600

Table 3-51. Design Data for UUS Ferrite Cores.

UUS, Ferrite Cores (Ferroxcube)

Part W
vv
tcu Wtfe MLT MPL wa Ac wa Ap K
E A, *AL
No. grams grams cm cm Ac cm 2
cm 2
cm 4
cm5 cm2 mh/lK
Ul 0-08-03 3.5 1.8 2.2 3.8 5.370 0.081 0.435 0.0352 0.000510 8.1 213
U20-16-07 16.4 19.0 4.4 6.8 1.896 0.560 1.062 0.5949 0.030661 29.5 826
U25-20-13 41.6 47.0 6.1 8.8 1.841 1.040 1.915 1.9920 0.135669 51.1 1261
U30-25-16 83.9 86.0 7.5 11.1 1.943 1.610 3.129 5.0380 0.430427 82.5 1609
U67-27-14 435.0 170.0 12.4 17.3 4.831 2.040 9.855 20.1050 1.321661 240.2 1652
U93-76-16 1875.2 800.0 15.2 35.4 7.757 4.480 34.752 155.6890 18.386023 605.3 1478
*This AL value has been normalized for a permeability of 1 K. For a close approximation of AL for other values of
permeability, multiply this AL value by the new permeability in kilo-perm. If the new permeability is 2500, then use
2.5.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for Toroidal, Ferrite Cores
The toroidal ferrite core has the best possible shape from the magnetic point of view. The magnetic flux
path is completely enclosed within the magnetic structure. The toroidal structure fully exploits the
capabilities of a ferrite material. The dimensional outline for toroidal ferrite cores is shown in Figure 3-47.
Dimensional data for toroidal ferrite cores is given in Table 3-52; design data is given in Table 3-53.

Wa —*" HT. -*—

,^fi, A

\O.D
i k

\
I.D.
r
©
ir
Toroidal Ferrite Core Perspective View

Figure 3-47. Dimension Outline for Toroidal Ferrite Cores.

Table 3-52. Dimensional Data for Toroidal Ferrite Cores.

Toroidal, Ferrite Z Coated Cores (Magnetics)

Part OD ID HT Part OD ID HT
No. cm cm cm No. cm cm cm
TC-40907 1.016 0.495 0.768 TC-42206 2.286 1.295 0.691
TC-41005 1.016 0.411 0.529 TC-42908 2.990 1.811 0.806
TC-41206 1.334 0.452 0.691 TC-43806 3.925 1.790 0.691
TC-41306 1.334 0.729 0.691 TC-43610 3.689 2.212 1.065
TC-41605 1.664 0.812 0.521 TC-43813 3.925 1.790 1.334
TC-42106 2.134 1.193 0.691 TC-48613 8.738 5.389 1.334

Table 3-53. Design Data for Toroidal Ferrite Cores.

Toroidal, Ferrite Cores (Magnetics)

Part w
vv
tcu Wtfe MLT MPL wa Ac Wa AP Ks A, *AL
No. grams grams cm cm Ac cm2 cm2 cm4 cm5 cm2 mh/lK
TC-41005 0.8 1.2 1.7 2.07 1.243 0.107 0.133 0.014196 0.000366 5.3 657
TC-40907 1.4 1.6 2.0 2.27 1.422 0.135 0.192 0.025980 0.000687 6.6 752
TC-41206 1.2 3.3 2.2 2.46 0.724 0.221 0.160 0.035462 0.001443 8.6 1130
TC-41306 3.2 2.4 2.2 3.12 2.856 0.146 0.417 0.060939 0.001638 10.2 591
TC-41605 4.0 2.8 2.2 3.68 3.386 0.153 0.518 0.079231 0.002240 12.8 548
TC-42106 11.2 5.4 2.8 5.00 4.840 0.231 1.118 0.258216 0.008482 22.7 600
TC-42206 13.7 6.4 2.9 5.42 5.268 0.250 1.317 0.329283 0.011221 25.8 600
TC-42908 33.7 12.9 3.7 7.32 7.196 0.358 2.576 0.922167 0.035869 44.6 630
TC-43806 38.0 29.4 4.2 8.97 4.006 0.628 2.516 1.580357 0.093505 61.2 878
TC-43610 63.6 26.4 4.7 8.30 6.742 0.570 3.843 2.190456 0.107283 68.5 883
TC-43813 47.2 51.7 5.3 8.30 2.188 1.150 2.516 2.893966 0.252394 71.0 1665
TC-48613 740.1 203.0 9.1 21.50 12.197 1.870 22.809 42.652794 3.496437 348.0 1091
This AL value has been normalized for a permeability of IK. For a close approximation of AL for other values of
permeability, multiply this AL value by the new permeability in kilo-perm. If the new permeability is 2500, then use
2.5.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for Toroidal, MPP Powder Cores
The dimensional outline for MPP powder cores is shown in Figure 3-48. Dimensional data for MPP
powder cores is given in Table 3-54; design data is given in Table 3-55. For more information, see Chapter
2.

HT.-
t
OD

Figure 3-48. Dimension Outline for Toroidal MPP Powder Cores.

Table 3-54. Dimensional Data for Toroidal MPP Powder Cores.

MPP Powder Cores, Magnetics 60 mu (coated)

Part OD ID HT Part OD ID HT Part OD ID HT
No. cm cm cm No. cm cm cm No. cm cm cm
55021 0.699 0.229 0.343 55381 1.803 0.902 0.711 55076 3.670 2.150 1.135
55281 1.029 0.427 0.381 55848 2.110 1.207 0.711 55083 4.080 2.330 1.537
55291 1.029 0.427 0.460 55059 2.360 1.334 0.838 55439 4.760 2.330 1.892
55041 1.080 0.457 0.460 55351 2.430 1.377 0.965 55090 4.760 2.790 1.613
55131 1.181 0.584 0.460 55894 2.770 1.410 1.194 55716 5.170 3.090 1.435
55051 1.346 0.699 0.551 55071 3.380 1.930 1.143 55110 5.800 3.470 1.486
55121 1.740 0.953 0.711 55586 3.520 2.260 0.978

Table 3-55. Design Data for Toroidal MPP Powder Cores.

MPP Powder Cores, Magnetics 60 mu (coated)
Part wtcu wtfe MLT MPL wa Ac wa AP Kg A, AL
2
No. grams grams cm cm Ac cm cm cm4 cm5 cm mh/lK
55021 0.10 0.553 1.10 1.36 0.723 0.047 0.034 0.001610 0.000027 2.30 24
55281 0.70 1.307 1.40 2.18 1.729 0.075 0.130 0.009757 0.000204 4.80 25
55291 0.70 1.645 1.60 2.18 1.376 0.095 0.130 0.012359 0.000301 5.10 32
55041 0.90 1.795 1.60 2.38 1.500 0.100 0.150 0.014998 0.000375 5.60 32
55131 1.50 1.993 1.70 2.69 2.759 0.091 0.250 0.022735 0.000492 6.90 26
55051 2.50 2.886 2.00 3.12 3.175 0.114 0.362 0.041279 0.000961 9.30 27
55121 6.10 6.373 2.50 4.11 3.563 0.192 0.684 0.131267 0.003985 16.00 35
55381 5.60 7.670 2.60 4.14 2.634 0.232 0.611 0.141747 0.005099 16.30 43
55848 11.10 8.836 2.80 5.09 4.898 0.226 1.107 0.250092 0.008001 22.70 32
55059 15.20 14.993 3.20 5.67 4.097 0.331 1.356 0.448857 0.018406 28.60 43
55351 17.90 18.706 3.50 5.88 3.727 0.388 1.446 0.561153 0.024969 31.40 51
55894 22.30 33.652 4.10 6.35 2.320 0.654 1.517 0.992423 0.062916 39.80 75
55071 46.20 44.086 4.50 8.15 4.263 0.672 2.865 1.925420 0.114179 58.30 61
55586 61.40 32.806 4.40 8.95 8.681 0.454 3.941 1.789128 0.074166 64.40 38
55076 60.20 48.692 4.80 8.98 5.255 0.678 3.563 2.415897 0.137877 68.00 56
55083 85.30 86.198 5.70 9.84 3.910 1.072 4.191 4.492709 0.336608 87.50 81
55439 101.90 170.140 6.80 10.74 2.106 1.990 4.191 8.340010 0.971244 112.60 135
55090 136.90 122.576 6.40 11.63 4.497 1.340 6.026 8.075211 0.677485 117.20 86
55716 169.30 132.540 6.40 12.73 5.917 1.251 7.402 9.260268 0.720435 133.10 73
55110 233.30 164.500 7.00 14.300 6.474 1.444 9.348 13.498792 1.111049 164.70 75

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for Toroidal, Iron Powder Cores

The dimensional outline for Iron powder cores is shown in Figure 3-49. Dimensional data for Iron powder
cores is given in Table 3-56; design data is given in Table 3-57. For more information, see Chapter 2.

Figure 3-49. Dimension Outline for Toroidal Iron Powder Cores.

Table 3-56. Dimensional Data for Toroidal Iron Powder Cores.

Iron Powder Cores, Micrometals 75 mu (coated)

Part OD ID HT Part OD ID HT Part OD ID HT
No. cm cm cm No. cm cm cm No. cm cm cm
T20-26 0.508 0.224 0.178 T50-26 1.270 0.770 0.483 T130-26 3.300 1.980 1.110
T25-26 0.648 0.305 0.244 T60-26 1.520 0.853 0.594 T 132-26 3.300 1.780 1.110
T26-26 0.673 0.267 0.483 T68-26 1.750 0.940 0.483 T131-26 3.300 1.630 1.110
T30-26 0.780 0.384 0.325 T80-26 2.020 1.260 0.635 T141-26 3.590 2.240 1.050
T37-26 0.953 0.521 0.325 T94-26 2.390 1.420 0.792 T 150-26 3.840 2.150 1.110
T38-26 0.953 0.445 0.483 T90-26 2.290 1.400 0.953 T175-26 4.450 2.720 1.650
T44-26 1.120 0.582 0.404 T 106-26 2.690 1.450 1.110

Table 3-57. Design Data for Toroidal Iron Powder Cores.

Iron Powder Cores, Micrometals 75 mu (coated)
Part wtcu wtre MLT MPL wa AC wa AP Kg A, AL
No. grams grams cm cm AC cm2 cm2 cm4 cm5 cm2 mh/lK
T20-26 0.10 0.19 0.70 1.15 1.713 0.023 0.039 0.000900 0.000010 1.2 18.5
T25-26 0.24 0.39 0.90 1.50 1.973 0.037 0.073 0.002700 0.000038 2.0 24.5
T26-26 0.26 0.93 1.30 1.47 0.644 0.090 0.058 0.005030 0.000130 2.6 57
T30-26 0.47 0.77 1.14 1.84 1.933 0.060 0.116 0.006940 0.000140 3.1 33.5
T37-26 0.97 1.04 1.28 2.31 3.328 0.064 0.213 0.013630 0.000270 4.5 28.5
T38-26 0.85 1.74 1.50 2.18 1.360 0.114 0.155 0.017700 0.000520 4.8 49
T44-26 1.46 1.86 1.50 2.68 2.687 0.099 0.266 0.026320 0.000670 6.2 37
T50-26 2.96 2.50 1.80 3.19 4.071 0.112 0.456 0.052120 0.001300 8.8 33
T60-26 4.40 4.89 2.20 3.74 3.053 0.187 0.571 0.106800 0.003680 12.2 50
T68-26 5.36 5.30 2.17 4.23 3.877 0.179 0.694 0.124150 0.004090 14.4 43.5
T80-26 11.66 8.31 2.63 5.14 5.394 0.231 1.246 0.287880 0.010100 21.4 46
T94-26 17.44 15.13 3.10 5.97 4.373 0.362 1.583 0.573000 0.026770 29.6 60
T90-26 18.37 15.98 3.40 5.78 3.894 0.395 1.538 0.607740 0.029600 29.4 70
T 106-26 23.05 29.94 3.93 6.49 2.504 0.659 1.650 1.087660 0.072990 38.0 93
T 130-26 48.33 40.46 4.40 8.28 4.408 0.698 3.077 2.148800 0.135810 56.9 81
T 132-26 39.05 44.85 4.40 7.96 3.089 0.805 2.487 2.002190 0.145990 53.9 103
T131-26 32.75 47.83 4.40 7.72 2.357 0.885 2.086 1.845820 0.147960 51.7 116
T141-26 62.70 45.70 4.60 9.14 5.743 0.674 3.871 2.608887 0.154516 66.6 75
T 150-26 62.55 58.24 4.85 9.38 4.091 0.887 3.629 3.218620 0.235550 71.6 96
T 175-26 128.04 105.05 6.20 11.20 4.334 1.340 5.808 7.782300 0.672790 107.4 105

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for Toroidal, Sendust Powder Cores
The dimensional outline for Sendust powder cores is shown in Figure 3-50. Dimensional data for Sendust
powder cores is given in Table 3-58; design data is given in Table 3-59. For more information, see Chapter
2.

HT.- ^^

OD

I
Figure 3-50. Dimension Outline for Toroidal Sendust Powder Cores.

Table 3-58. Dimensional Data for Toroidal Sendust Powder Cores.

Sendust Powder Cores, Magnetics 60 mu (coated)

Part OD ID HT Part OD ID HT Part OD ID HT
No. cm cm cm No. cm cm cm No. cm cm cm
77021 0.699 0.229 0.343 77381 1.803 0.902 0.711 77076 3.670 2.150 1.135
77281 1.029 0.427 0.381 77848 2.110 1.207 0.711 77083 4.080 2.330 1.537
77291 1.029 0.427 0.460 77059 2.360 1.334 0.838 77439 4.760 2.330 1.892
77041 1.080 0.457 0.460 77351 2.430 1.377 0.965 77090 4.760 2.790 1.613
77131 1.181 0.584 0.460 77894 2.770 1.410 1.194 77716 5.170 3.090 1.435
77051 1.346 0.699 0.551 77071 3.380 1.930 1.143 77110 5.800 3.470 1.486
77121 1.740 0.953 0.711 77586 3.520 2.260 0.978

Table 3-59. Design Data for Toroidal Sendust Powder Cores.

Sendust Powder Cores, Magnetics 60 mu (coated)

Part wtcu wtfe MLT MPL wa Ac wa AP KR A, AL
No. grams grams cm cm Ac cm2 cm2 cm4 cm5 cm2 mh/lK
77021 0.10 0.448 1.10 1.36 0.723 0.047 0.034 0.001610 0.000027 2.30 24
77281 0.70 1.148 1.40 2.18 1.729 0.075 0.130 0.009757 0.000204 4.80 25
77291 0.70 1.442 1.60 2.18 1.376 0.095 0.130 0.012359 0.000301 5.10 32
77041 0.90 1.666 1.60 2.38 1.500 0.100 0.150 0.014998 0.000375 5.60 32
77131 1.50 1.706 1.70 2.69 2.759 0.091 0.250 0.022735 0.000492 6.90 26
77051 2.50 2.490 2.00 3.12 3.175 0.114 0.362 0.041279 0.000961 9.30 27
77121 6.10 5.524 2.50 4.11 3.563 0.192 0.684 0.131267 0.003985 16.00 35
77381 5.60 6.723 2.60 4.14 2.634 0.232 0.611 0.141747 0.005099 16.30 43
77848 11.10 8.052 2.80 5.09 4.898 0.226 1.107 0.250092 0.008001 22.70 32
77059 15.20 13.137 3.20 5.67 4.097 0.331 1.356 0.448857 0.018406 28.60 43
77351 17.90 15.970 3.50 5.88 3.727 0.388 1.446 0.561153 0.024969 31.40 51
77894 22.30 29.070 4.10 6.35 2.320 0.654 1.517 0.992423 0.062916 39.80 75
77071 46.20 38.338 4.50 8.15 4.263 0.672 2.865 1.925420 0.114179 58.30 61
77586 61.40 28.443 4.40 8.95 8.681 0.454 3.941 1.789128 0.074166 64.40 38
77076 60.20 42.619 4.80 8.98 5.255 0.678 3.563 2.415897 0.137877 68.00 56
77083 85.30 73.839 5.70 9.84 3.910 1.072 4.191 4.492709 0.336608 87.50 81
77439 101.90 149.608 6.80 10.74 2.106 1.990 4.191 8.340010 0.971244 112.60 135
77090 136.90 109.089 6.40 11.63 4.497 1.340 6.026 8.075211 0.677485 117.20 86
77716 169.30 111.477 6.40 12.73 5.917 1.251 7.402 9.260268 0.720435 133.10 73
77110 233.30 144.544 7.00 14.300 6.474 1.444 9.348 13.498792 1.111049 164.70 75

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for Toroidal, High Flux Powder Cores
The dimensional outline for High Flux powder cores is shown in Figure 3-51. Dimensional data for High
Flux powder cores is given in Table 3-60; design data is given in Table 3-61. For more information, see
Chapter 2.

HT.—I
t
OD

I
Figure 3-51. Dimension Outline for Toroidal High Flux Powder Cores.

Table 3-60. Dimensional Data for Toroidal High Flux Powder Cores.
High Flux Powder Cores, Magnetics 60 mu (coated)
Part OD ID HT Part OD ID HT Part OD ID HT
No. cm cm cm No. cm cm cm No. cm cm cm
58021 0.699 0.229 0.343 58381 1.803 0.902 0.711 58076 3.670 2.150 1.135
58281 1.029 0.427 0.381 58848 2.110 1.207 0.711 58083 4.080 2.330 1.537
58291 1.029 0.427 0.460 58059 2.360 1.334 0.838 58439 4.760 2.330 1.892
58041 1.080 0.457 0.460 58351 2.430 1.377 0.965 58090 4.760 2.790 1.613
58131 1.181 0.584 0.460 58894 2.770 1.410 1.194 58716 5.170 3.090 1.435
58051 1.346 0.699 0.551 58071 3.380 1.930 1.143 58110 5.800 3.470 1.486
58121 1.740 0.953 0.711 58586 3.520 2.260 0.978

Table 3-61. Design Data for Toroidal High Flux Powder Cores.

High Flux Powder Cores, Magnetics 60 mu (coated)

Part wtcu Wtfe MLT MPL wa Ac wa AP KB A, AL
No. grams grams cm cm Ac cm cm2 cm4 cm cm mh/lK
58021 0.10 0.504 1.10 1.36 0.723 0.047 0.034 0.001610 0.000027 2.30 24
58281 0.70 1.222 1.40 2.18 1.729 0.075 0.130 0.009757 0.000204 4.80 25
58291 0.70 1.598 1.60 2.18 1.376 0.095 0.130 0.012359 0.000301 5.10 32
58041 0.90 1.692 1.60 2.38 1.500 0.100 0.150 0.014998 0.000375 5.60 32
58131 1.50 1.880 rr.70 2.69 2.759 0.091 0.250 0.022735 0.000492 6.90 26
58051 2.50 2.726 2.00 3.12 3.175 0.114 0.362 0.041279 0.000961 9.30 27
58121 6.10 6.016 2.50 4.11 3.563 0.192 0.684 0.131267 0.003985 16.00 35
58381 5.60 7.238 2.60 4.14 2.634 0.232 0.611 0.141747 0.005099 ^ 16.30 43
58848 11.10 8.366 2.80 5.09 4.898 0.226 1.107 0.250092 0.008001 22.70 32
58059 15.20 14.100 3.20 5.67 4.097 0.331 1.356 0.448857 0.018406 28.60 43
58351 17.90 17.672 3.50 5.88 3.727 0.388 1.446 0.561153 0.024969 31.40 51
58894 22.30 31.772 4.10 6.35 2.320 0.654 1.517 0.992423 0.062916 39.80 75
58071 46.20 41.548 4.50 8.15 4.263 0.672 2.865 1.925420 0.114179 58.30 61
58586 61.40 30.926 4.40 8.95 8.681 0.454 3.941 1.789128 0.074166 64.40 38
58076 60.20 45.966 4.80 8.98 5.255 0.678 3.563 2.415897 0.137877 68.00 56
58083 85.30 81.310 5.70 9.84 3.910 1.072 4.191 4.492709 0.336608 87.50 81
58439 101.90 160.740 6.80 10.74 2.106 1.990 4.191 8.340010 0.971244 112.60 135
58090 136.90 115.620 6.40 11.63 4.497 1.340 6.026 8.075211 0.677485 117.20 86
58716 169.30 125.020 6.40 12.73 5.917 1.251 7.402 9.260268 0.720435 133.10 73
58110 233.30 155.100 7.00 14.300 6.474 1.444 9.348 13.498792 1.111049 164.70 75

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for EE, Iron Powder Cores

The dimensional outline for EE iron powder cores is shown in Figure 3-52. Dimensional data for EE iron
powder cores is given in Table 3-62; design data is given in Table 3-63. For more information, see Chapter
2.

L
AC '
ii \
G
\
t 1

E B A \
\ F

> \ F

^ \ '

\V a _^ ^
^
c ^
D
EE Iron Powder Cores Perspective View

Figure 3-52. Dimension Outline for EE Iron Powder Cores.

Table 3-62. Dimensional Data for EE Iron Powder Cores.

EE, Iron Powder Cores (Micrometals) 75 mu Mix-26

Part A B C D E G Part A B C D E G
No. cm cm cm cm cm cm No. cm cm cm cm cm cm
DIN- 16-5 1.640 1.130 1.630 0.462 0.462 1.200 DIN-42-15 4.280 3.070 4.220 1.500 1.200 3.070
EI-187 1.910 1.430 1.610 0.475 0.475 1.160 DIN-42-20 4.280 3.070 4.220 2.000 1.200 3.070
EE-24-25 2.540 1.910 1.910 0.635 0.635 1.270 EI-625 4.740 3.180 3.940 1.570 1.570 2.420
EI-375 3.490 2.540 2.910 0.953 0.953 1.960 DIN-55-21 5.610 3.860 5.540 2.080 1.730 3.830
EI-21 4.130 2.860 3.410 1.270 1.270 2.140 EI-75 5.690 3.810 4.760 1.890 1.890 2.900

Table 3-63. Design Data for EE Iron Powder Cores.

EE, Iron Powder Cores (Micrometals) 75 mu Mix-26

Part wtcu Wtfe MLT MPL wa Ac wa AP Ks A, AL
2
No. grams grams cm cm AC cm cm2 cm4 cm5 cm2 mh/lK
DIN- 16-5 4.7 5.3 3.3 3.98 1.790 0.224 0.401 0.090 0.00243 11.5 58
EI-187 7.5 5.5 3.8 4.10 2.451 0.226 0.554 0.125 0.00297 14.4 64
El -24-25 14.3 12.2 5.0 5.10 2.010 0.403 0.810 0.326 0.01062 23.5 92
EI-375 37.1 40.1 6.7 7.40 1.714 0.907 1.555 1.411 0.07624 46.8 134
EI-21 50.2 80.8 8.2 8.40 1.071 1.610 1.725 2.777 0.21852 63.3 210
DIN-42-15 91.3 112.4 8.9 10.40 1.586 1.810 2.870 5.196 0.42050 84.4 195
DIN-42-20 101.5 149.6 9.9 10.40 1.191 2.410 2.870 6.918 0.67054 92.9 232
EI-625 65.2 141.1 9.4 9.5 0.785 2.480 1.948 4.831 0.50894 82.4 265
DIN-55-21 167.9 283.7 11.6 13.2 1.133 3.600 4.079 14.684 1.82699 141.3 275
EI-75 110.7 245.8 11.2 11.5 0.778 3.580 2.784 9.9667 1.27615 119.3 325

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design and Dimensional Data for EE, Sendust Powder Cores

The dimensional outline for EE Sendust cores is shown in Figure 3-53. Dimensional data for EE Sendust
powder cores is given in Table 3-64; design data is given in Table 3-65. For more information, see Chapter
2.

A
c '
i
i \
G
ii
\
E B A \
\r

>, \ \r
\
X 1

y, a •^ ^-
^
c te-
D
EE Sendust Powder Cores Perspective View

Figure 3-53. Dimension Outline for EE Sendust Powder Cores.

Table 3-64. Dimensional Data for EE Sendust Powder Cores.

EE, Sendust Powder Cores (Magnetics) 60 mu

Part A B C D E G Part A B C D E G
No. cm cm cm cm cm cm No. cm cm cm cm cm cm
EI-187 1.910 1.430 1.610 0.475 0.475 1.160 DIN-42-15 4.280 3.070 4.220 1.500 1.200 3.070
EE-24-25 2.540 1.910 1.910 0.635 0.635 1.270 DIN-42-20 4.280 3.070 4.220 2.000 1.200 3.070
EI-375 3.490 2.540 2.910 0.953 0.953 1.960 DIN-55-21 5.610 3.860 5.540 2.080 1.730 3.830
EI-21 4.130 2.860 3.410 1.270 1.270 2.140

Table 3-65. Design Data for EE Sendust Powder Cores.

EE, Sendust Powder Cores (Magnetics) 60 mu

Part wtcu Wtfe MLT MPL wa Ac wa AP KK A, AL
No. grams grams cm cm Ac cm cm2 cm4 cm5 cm mh/lK
EI-187 7.5 6.4 3.8 4.01 2.451 0.226 0.554 0.125 0.00297 14.4 48
El -24-25 14.3 13.1 5.0 4.85 2.010 0.403 0.810 0.326 0.01062 23.5 70
EI-375 37.1 40.8 6.7 6.94 1.714 0.907 1.555 1.411 0.07624 46.8 102
EI-21 50.2 82.6 8.2 7.75 1.071 1.610 1.725 2.777 0.21852 63.3 163
DIN-42-15 91.3 126.0 8.9 9.84 1.586 1.810 2.870 5.196 0.42050 84.4 150
DIN-42-20 101.5 163.0 9.9 9.84 1.191 2.410 2.870 6.918 0.67054 92.9 194
DIN-55-21 167.9 302.0 11.6 12.3 1.133 3.600 4.079 14.684 1.82699 141.3 219

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.
*********************************************
uioo ' pduisi • AVAV A\
0301-0^909 II '
**********************************************
***********************************************
UIO 3 ' J 3Ul3mi ' A\ A\ A\.
108Z6 VO
***********************************************
66£0-e8009 II 'i
66£ X°H 'O'd
P3UOIJEUJ3JUI
***********************************************
/.0836 VO 'UII
***********************************************
uioo'SnBds.(U)soii3iigBra ----- t
'£££0-969-2117-1
Vd 'i
ZZtll
pUB 30IJJO 3UIOJJ
 Chapter 4

Window Utilization, Magnet Wire,

and
Insulation

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table of Contents
1. Window Utilization Factor, Ku
2. Si, Wire Insulation
3. S2, Fill Factor
4. S3, Effective Window
5. S4, Insulation Factor
6. Summary
7. Window Utilization Factor, Ku for Bobbin Ferrites
8. Circular mil and Square mil
9. Magnet Wire
10. Magnet Wire, Film Insulation
11. Wire Table
12. Solderable Insulation
13. Bondable Magnet Wire
14. Base Film Insulation
15. Bonding Methods
16. Miniature Square Magnet Wire
17. Multistrand Wire and Skin Effect
18. Reduce Skin Effect in Transformers
19. Calculating Skin Effect in Inductors
20. Multistrand Litz Wire
21. Proximity Effect
22. Proximity Effect in Transformers
23. Multiple Layer High Frequency Transformer and High Loss
24. Proximity Effect Using Dowell Curves
25. Specialty Wire
26. Triple Insulated Wire
27. Triple Insulated Litz
28. Polyfilar Magnetic Wire
29. Standard Foils
30. The Use of Foils
31. Calculating, MLT
32. Calculating, MLT (toroid)
33. Copper Resistance
34. Copper Weight
35. Electrical Insulating Materials
36. References

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Window Utilization Factor, Ku

The window utilization factor is the amount of copper that appears in the window area of the transformer or
inductor. The window utilization factor is influenced by five main factors:

1. Wire insulation, S i.
2. Wire lay fill factor, layer or random wound, S2.
3. Effective window area (or when using a toroid, the clearance hole for passage of the
shuttle), S3.
4. Insulation required for multiplayer windings, or between windings, S4.
5. Workmanship, (quality).

These factors, multiplied together, will give a normalized window utilization of Ku = 0.4, as shown in
Figure 4-1.

Core Window Area

Area Taken By:

Bobbin
Tube
Margin Copper Area
Wrapper Insulation
Layer Insulation
Magnet Wire Insulation
Fill Factor

Figure 4-1. Window Area Occupied by Copper.

The window utilization factor, Ku, of the available core window is that space occupied by the winding
(copper), and is calculated from areas Si, S2, S3, and S4:

Ku=StS2S3S4 [4-1]
Where:
51 = conductor area/wire area
52 = wound area/usable window area
53 = usable window area/window area
54 = usable window area/usable window area + insulation

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 In which:
Conductor area, Aw(B) = copper area.
Wire area, Aw = copper area + insulation area.
Wound area = number of turns x wire area of one turn.
Usable window area = available window area - residual area, that results from the particular winding
technique used.
Window area = available window area.
Insulation area = area used for winding insulation.

Si, Wire Insulation

In the design of high-current or low-current transformers, the ratio of the conductor area to the total wire
area can vary from 0.941 to 0.673, depending on the wire size. In Figure 4-2, the thickness of the insulation
has been exaggerated to show how the insulation impacts the overall area of the wire.

It can be seen, in Figure 4-2, that, by using multi-strands of fine wire to reduce the skin effect, it will have a
significant impact on the window utilization factor, Ku. Si is not only dependent upon wire size, but it is
also dependent upon insulation coating. Table 4-1 shows the ratio of bare magnet wire to the magnet wire
with insulation for single, heavy, triple, and quad insulation. When designing low-current transformers, it is
advisable to re-evaluate Si because of the increased amount of insulating material.

[4-2]

Insulation

0.00787 cm 0.259 cm

AWG #40 AWG #10

Copper Area

Figure 4-2. Comparing Insulation with Different Wire Gauges.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table 4-1

Magnetic Wire Data (Nominal)

Size Bare Ratio Ratio Ratio Ratio
AWG Area (cm ) Bare/Single Bare/Heavy Bare/Triple Bare/Quad
10 0.1019 0.961 0.930 0.910 0.880
15 0.0571 0.939 0.899 0.867 0.826
20 0.0320 0.917 0.855 0.812 0.756
25 0.0179 0.878 0.793 0.733 0.662
30 0.0100 0.842 0.743 0.661 0.574
35 0.0056 0.815 0.698 0.588 0.502
40 0.0031 0.784 0.665 0.544 0.474

S2, Fill Factor

S2 is the fill factor, or the wire lay, for the usable window area. When winding a large number of turns
tightly on a smooth surface, the winding length exceeds the calculated value from the wire diameter by 10 to
15%, depending on the wire gauge. See Figure 4-3. The wire lay is subjected to wire tension, and wire
quality, such as continuous wire diameter and the winding technique depending on the skill of the operator.
The wire lay factor relationship for various wire sizes for layer wound coils is shown in Table 4-2, and for
random wound coils in Table 4-3. The Tables list the outside diameter for heavy film magnetic wire, 10 —
44 AWG.

Table 4-2

Wire Lay Factor For Layer Wound Coils

Insulated Wire Insulated Wire Wire Lay
AWG OD (inch) OD (cm) Factor
10 to 25 0.1051-0.0199 0.2670 - 0.0505 0.90
26 to 30 0.0178-0.0116 0.0452 - 0.0294 0.89
31 to 35 0.0105-0.0067 0.0267 - 0.0170 0.88
36 to 38 0.0060 - 0.0049 0.0152-0.0124 0.87
39 to 40 0.0043 - 0.0038 0.0109-0.0096 0.86
41 to 44 0.0034 - 0.0025 0.00863 - 0.00635 0.85
Heavy film magnetic wire.

Table 4-3

Wire Lay Factor For Random Wound Coils

Insulated Wire Insulated Wire Wire Lay
AWG OD (inch) OD (cm) Factor
10 to 22 0.1051-0.0276 0.267-0.0701 0.90
23 to 39 0.0623 - 0.0109 0.0249 - 0.0043 0.85
40 to 44 0.0038 - 0.0025 0.0096 - 0.00635 0.75
Heavy film magnet wire.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Calculated turns
Actual turns

QQQQQQQQQQQQQOQQ
Winding Length

Figure 4-3. Capable Turns per Unit Length.

There are two ideal winding arrangements shown in Figure 4-4 and Figure 4-5. The square winding is
shown in Figure 4-4 and the hexagonal winding is shown in Figure 4-5. The simplest form of winding is
done by a coil being wound, turn-by-turn and layer-upon-layer, as shown in Figure 4-4. The square winding
pattern has a theoretical fill factor of 0.785.

Wire Area = 0.785

Winding Build

Figure 4-4. Theoretically, the Square Winding Pattern Fill Factor 0.785.

A seemingly, better fill factor can be achieved by using the hexagonal winding in Figure 4-5, compared to
the square winding in Figure 4-4. In this type of winding, the individual wires do not lie exactly above each
other, as in the square winding pattern. Instead, the wires lie in the grooves of the lower layer, as shown in
Figure 4-5. This style of winding produces the tightest possible packing of the wire. The hexagonal style of
winding will yield a theoretical fill factor of 0.907.

The fill factor, using the square winding pattern of 0.785, would be nearly impossible to achieve by hand
winding without some layer insulation. Any layer insulation will reduce the fill factor even further. The fill
factor, using the hexagonal winding pattern of 0.907, is just as hard to get. Hand winding, using the
hexagonal technique, will result in the following: The first layer goes down with almost complete order. In
the second layer, some disordering has occurred. With the third and fourth layer, disordering really sets in
and the winding goes completely awry. This type of winding performs well with a small number of turns,
but, with a large number of turns, it becomes randomly wound.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 0.866(0)

Winding Build
99999999999JUO
99
JJUULXAJ-OJJQQT

Winding Length

Figure 4-5. Theoretically, the Hexagonal Winding Pattern Fill Factor 0.907.

The ideal winding on a rectangular bobbin is shown in Figure 4-6. Then, when winding rectangular bobbins
or tubes, the actual winding height in the region covered by the core will be greater than the calculated
winding height or build due to the bowing of the windings. See Figure 4-7. The amount of bowing depends
on the proportions of the winding and the height of the winding. Usually, the available winding build
should be reduced by 15 to 20%, or 0.85x the winding build. When winding on a round bobbin or tube, this
bowing effect is negligible.

The conclusion is, in comparing the square winding pattern used in the layer wound coil with its insulation
with the hexagonal winding pattern and its awry winding pattern, both seem to have a fill factor of about
0.61. But there is always the hundred to one exception, such as, when a design happens to have the right
bobbin, the right number of turns, and the right wire size. This normally only happens when the design is
not critical.

Winding Build /
*
^*
Rectangular Core

Figure 4-6. Ideal Winding on a Rectangular Bobbin.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 To minimize this bowing effect and to insure a minimum build for either random or layer winding, the
round bobbin, shown in Figure 4-8, will provide the most compact design. It can be seen, in Figure 4-8,
that the round bobbin provides a uniform tension, all 360 degrees around the bobbin, for both layer and
random windings. The other benefit, in using a round bobbin, is the reduction and minimizing of the
leakage inductance caused from the bowing.

Bowing Rectangular Core

Figure 4-7. Bowing in Transformer Windings.

Winding Build

Round Core

Figure 4-8. A Round Bobbin Insures Minimum Bowing.

83, Effective Window

The effective window, S3, defines how much of the available window space may actually be used for the
winding. The winding area available to the designer depends on the bobbin or tube configuration.
Designing a layer winding that uses a tube will require a margin, as shown in Figure 4-9. The margin
dimensions will vary with wire size. See Table 4-4. It can be seen, in Figure 4-9 and Table 4-4, how the
margin reduces the effective window area. When transformers are constructed, using the layer winding
technique, there is an industry standard for layer insulation thickness. This thickness is based on the
diameter of the wire, as shown in Table 4-5.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Tube Layer Insulation Wrapper .

Winding Build
OOaOOOOQOOOOQ
-Winding Length -
Margin

Figure 4-9. Transformer Windings with Margins.

Table 4-4
Winding Margins Versus AWG
Margin
AWG cm inch
10-15 0.635 0.25
16-18 0.475 0.187
19-21 0.396 0.156
22-31 0.318 0.125
32-37 0.236 0.093
38-up 0.157 0.062

Table 4-5
Layer Insulation Thickness
Insulation Thickness
AWG cm inch
10-16 0.02540 0.01000
17- 19 0.01780 0.00700
20-21 0.01270 0.00500
22-23 0.00760 0.00300
24-27 0.00510 0.00200
28-33 0.00381 0.00150
34-41 0.00254 0.00100
42-46 0.00127 0.00050

A single bobbin design, as shown in Figure 4-10, offers an effective area, Wa, between 0.835 to 0.929 for
laminations, and 0.55 to 0.75 for ferrites; a two bobbin configuration, as shown in Figure 4-11, offers an
effective area, Wa, between 0.687 to 0.873 for the tape C cores.

The toroid is a little different. The term, S3, defines how much of the available window space can actually
be used for the winding. In order to wind the toroidal core, there has to be room to allow free passage of
the shuttle. If half of the inside diameter is set aside for the shuttle, then, there will be 75% of the window

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 area, (Wa), left for the design which is a good value for the effective window area factor, S3 = 0.75, as
shown in Figure 4-12. The toroid would fall into all of the above categories.

Bobbin Channel Bracket

Coil

Lamination Mounting Foot

Figure 4-10. Transformer Construction with Single Bobbin.

Bobbin

Coil #1 Coil #2

Tape C Core
Mounting Bracket

Figure 4-11. Transformer Construction with Dual Bobbins.

0.5 ID Clearance For Shuttle

Effective Window Area
Core
Effective Window area Wa(eff) = (0.75)(Tt)(ID)2/4

Figure 4-12. Effective Winding Area of a Toroidal Core.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 84, Insulation Factor

The insulation factor, S4, defines how much of the usable window space is actually being used for
insulation. If the transformer has multiple secondaries with significant amounts of insulation, S4 should be
reduced by 5 to 10% for each additional secondary winding, partly because of the added space occupied by
insulation and partly because of the poorer space factor.

The insulation factor, S4, is not taken into account in Figure 4-12. The insulation factor, S4, is to be 1.0.
The window utilization factor, Ku, is highly influenced by insulation factor, S4, because of the rapid buildup
of insulation in the toroid, as shown in Figure 4-13.

In Figure 4-13, it can be seen that the insulation buildup is greater on the inside, than on the outside. For
example, in Figure 4-13, if 1.27 cm (1/2") wide tape was used with an overlap of 0.32 cm (1/8") on the
outside diameter, the overlap thickness would be four times the thickness of the tape. It should be noted
that the amount of overlap depends greatly on the size of the toroid and the required tape. In the design of
toroidal components, and using the 0.5 ID remaining for passage of the shuttle, there is normally enough
room for the wrapper.

Dacron Wrapper

4
Wound Ht.

Figure 4-13. Wrapped Toroid.

Summary
The author hopes to have cleared up some of the mystery of how the window utilization factor, Ku, was
derived and that the magic of 0.4 is now clear. All the different parts making up window utilization, have
been explained, hopefully eliminating confusion and simplifying the complexity of the subject.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 As stated at the beginning of this chapter, a good approximation for the window utilization factor is Ku =
0.4.

8, = conductor area/wire area = 0.855, #20 AWG

82 = wound area/usable window area = 0.61
53 = usable window area/window area = 0.75
54 = usable window area/usable window area + insulation = 1
Ku =8,5,8,5,
1 2 3 4

Ku = (0.855)(0.61)(0.75)(1.0) = 0.391 * 0.4

Being a very conservative number, it can be used in most designs. It is an important factor in all designs of
magnetic components.

Window Utilization Factor, Ku for Bobbin Ferrites

In high frequency power electronics, the majority of the designs will use some kind of bobbin ferrite. The
main reasons for using ferrites is its high frequency performance and cost. The window utilization factor,
Ku, for bobbin ferrites is not as high as it is for iron alloy materials, such as laminations and C cores.
Design engineers, who have been using bobbin ferrite materials, know the drawback in the window
utilization factor, Ku. Once this problem is understood, then, the problem should go away.

Ferrite materials are fired in kilns like ceramic pottery. There is a certain amount of shrinkage after firing,
and the amount varies from one manufacturer's process to another. The amount of shrinkage could vary as
much as 15 to 30%, as shown in Figure 4-14. The ferrite manufacturers try to keep a tight control on the
amount of shrinkage, because these cores must meet a dimensional tolerance after firing. Even though the
shrinkage is under tight control, the tolerances on the end product are much larger than the iron alloy,
stamped laminations. The end result is the bobbin has to slip on and meet all of the minimum and
maximum dimensional tolerances.

A
c
B
\
\
B

Wa

D C D
EE Ferrite Cores Before Firing EE Ferrite Cores After Firing

Figure 4-14. Ferrite EE Core Before and After Firing.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 This dimensional tolerance has an impact on the winding area of the bobbin, clearly shown in Table 4-6.
This smaller winding area reduces the power handling capability of the core. Operating at high frequency
will also reduce the power handling capability of the core because of the skin effect. The skin effect
requires the use of multistrands of fine wire in place of a large single strand. The selection of the correct
wire size to minimize the skin effect at given frequency is shown in Equations [4-5] through [4-9]. Also
shown is an example of the largest wire size that should be used when operating at 100kHz. Reevaluate the,
Ku, Equation [4-3] so that it can operate at 100kHz, using a #26 wire, and using a cut ferrite core.

51 = conductor area/wire area = 0.79, #26 AWG

52 = wound area/usable window area = 0.61
53 = usable window area/window area = 0.6
54 = usable window area/usable window area + insulation = 1

i ~ 1 2 3 4
[4-4]
Ku =(0.79)(0.61)(0.6)(1.0) = 0.289

Table 4-6. Effective Window Area.

Effective Window Area

Core Window Bobbin Ratio
2 2
cm cm B/W
RM-6 0.260 0.150 0.577
RM-8 0.456 0.303 0.664
RM-12 1.103 0.730 0.662
PQ-20/16 0.474 0.256 0.540
PQ-26/25 0.845 0.502 0.594
PQ-35/35 2.206 1.590 0.721
EFD-10 0.116 0.042 0.362
EFD-15 0.314 0.148 0.471
EFD-25 0.679 0.402 0.592
EC-35 1.571 0.971 0.618
EC-41 2.082 1.375 0.660
EC-70 6.177 4.650 0.753
Laminations
El- 187 0.529 0.368 0.696
EI-375 1.512 1.170 0.774
EI-21 1.638 1.240 0.757

Circular mil and Square mil

There are engineers that use circular mils (CM)/amp or square mils/amp. This is the reciprocal current
density. The norm is to use amps/cm2, which is a true current density. There have been some requests to
define circular mils and square mils. First, let's define a mil, which is .001 inch. Figure 4-15 shows the area
of a square mil, and the area of a circular mil.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 One Square mil = 0.001

0.001

One Circular mil =

Figure 4-15. Comparing Circular-Mils and Square-Mils.

To convert Square mils to Circular mils , multiply by 1.2732.

To convert Circular mils to Square mils , multiply by 0.7854.
To convert Circular mils to Square centimeters , multiply by 5.066x10~6-
To convert Square mils to Square centimeters , multiply by 6.45x10"°
Note:
Designers have for many years used this rule of thumb:

500CM/Amp « 400Amps/cm2
lOOOCM/Amp ~ 200Amps/cm2

Magnet Wire

Standard magnet wire is available in three different materials, shown in Table 4-7. The most common is
copper, but aluminum and silver are available. Aluminum magnet wire is one-third the weight of copper for
the same size conductor and one-half the weight for the same conductivity. Aluminum magnet wire is a
little more difficult to terminate, but it can be done. Silver magnet wire has the highest conductivity, easy to
solder to, and weighs 20% more than copper.

Table 4-7

Magnet Wire Material Properties

Density Resistivity Weight Resistance Temperature
3
Material Symbol grams/cm (j,Q/cm Factor Factor Coefficient
Copper Cu 8.89 1.72 1 1 0.00393
Silver Ag 10.49 1.59 1.18 0.95 0.00380
Aluminum Al 2.703 2.83 0.3 1.64 0.00410

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Magnet Wire, Film Insulation

It is the design engineer's responsibility to ensure that the selected magnet wire used in the design is
compatible with the environmental and design specification. The environmental specification will set the
ambient temperature. The maximum operating temperature of the magnet wire is obtained by summing the
maximum ambient temperature, plus the temperature rise of the magnetic component. After the maximum
temperature has been obtained, see Table 4-8 for the Temperature Class. The magnet wire insulation guide
listing in Table 4-7, is only a partial list fromNEMA, Standard MW 1000.

The maximum operating temperature is the "Achilles Heel" to the magnet wire. Standard magnet wire is
rated by temperature. The range is from 105°C to 220°C, as shown in Table 4-8. The insulation film of the
magnet wire is on the surface of the copper wire. This insulation film is the most vulnerable to thermal
overloads, so the selection of the insulation film is very critical for long life. When magnet wire is
subjected to thermal overloads, or a high ambient temperature above its rated temperature, the life of the
magnet wire is greatly reduced, as shown in Figures 4-16 and 4-17. The engineer must be very careful of
hot spots so as not to degrade the service life of the magnetic component.

Table 4-8

Magnet Wire Insulation Guide

NEMA
Temperature Insulation Dielectric Standard
Class Type Constant MW 1000
105°C Polyurethane* 6.20 MW-2-C
105°C Formvar 3.71 MW-15-C
130°C Polyurethane -Nylon* 6.20 MW-28-C
155°C Polyurethane- 1 55 6.20 MW-79-C
180°C Polyester Solderable* 3.95 MW-77-C
200°C Polyester-amid-imide 4.55 MW-35-C
220°C Polyimide (ML) 3.90 MW-16-C
*Solderable insulations.

Wire Table
Table 4-9 is the Wire Table for AWG, 10 to 44, heavy film wire. The bare wire areas are given in cm2 in
column 2, and the circular mils are given in column 3 for each wire size. The equivalent resistance in
micro-ohms per centimeter (|j,Q/cm or 10"6 Q/cm) is given in column 4 for each wire size. Columns 5
through 13 relate to heavy, insulated film coating. The weight of the magnet wire is found in column 13, in
grams, per centimeter.

Table 4-10 provides the maximum outside diameter for magnet wire with single, heavy, triple, and quad
film insulation. The dimensional data is in centimeters and inches, for AWG 10 through 44.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table 4-9

Wire Table
Resistance Heavy Synthetics
AWG Bare Area (.ifi/cm Area Diameter Turns-Per Turns-Pcr Weight
cm (10~3) cir-mil
2
20°C ctn (10"3) cir-mil
2
cm Inch cm Inch cm2 Inch2 gm/cm
I 2 3 4 5 6 7 8 9 10 11 12 13
10 52.6100 10384.00 32.7 55.9000 11046.00 0.2670 0.105 3.9 10 11 69 0.46800
11 41.6800 8226.00 41.4 44.5000 8798.00 0.2380 0.094 4.4 11 13 90 0.37500
12 33.0800 6529.00 52.1 35.6400 7022.00 0.2130 0.084 4.9 12 17 108 0.29770
13 26.2600 5184.00 65.6 28.3600 5610.00 0.1900 0.075 5.5 13 21 136 0.23670
14 20.8200 4109.00 82.8 22.9500 4556.00 0.1710 0.068 6.0 15 26 169 0.18790
15 16.5100 3260.00 104.3 18.3700 3624.00 0.1530 0.060 6.8 17 33 211 0.14920
16 13.0700 2581.00 131.8 14.7300 2905.00 0.1370 0.054 7.3 19 41 263 0. 1 1 840
17 10.3900 2052.00 165.8 11.6800 2323.00 0.1220 0.048 8.2 21 51 331 0.09430
18 8.2280 1624.00 209.5 9.3260 1857.00 0.1090 0.043 9.1 23 64 415 0.07474
19 6.5310 1289.00 263.9 7.5390 1490.00 0.0980 0.039 10.2 26 80 515 0.05940
20 5.1880 1024.00 332.3 6.0650 1197.00 0.0879 0.035 11.4 29 99 638 0.04726
21 4.1160 812.30 418.9 4.8370 954.80 0.0785 0.031 12.8 32 124 800 0.03757
22 3.2430 640.10 531.4 3.8570 761.70 0.0701 0.028 14.3 36 156 1003 0.02965
23 2.5880 510.80 666.0 3.1350 620.00 0.0632 0.025 15.8 40 191 1234 0.02372
24 2.0470 404.00 842.1 2.5140 497.30 0.0566 0.022 17.6 45 239 1539 0.01884
25 1.6230 320.40 1062.0 2.0020 396.00 0.0505 0.020 19.8 50 300 1933 0.01498
26 1.2800 252.80 1345.0 1.6030 316.80 0.0452 0.018 22.1 56 374 2414 0.01185
27 1.0210 201.60 1687.0 1.3130 259.20 0.0409 0.016 24.4 62 457 2947 0.00945
28 0.8046 158.80 2142.0 1.0515 207.30 0.0366 0.014 27.3 69 571 3680 0.00747
29 0.6470 127.70 2664.0 0.8548 169.00 0.0330 0.013 30.3 77 702 4527 0.00602
30 0.5067 100.00 3402.0 0.6785 134.50 0.0294 0.012 33.9 86 884 5703 0.00472
31 0.4013 79.21 4294.0 0.5596 110.20 0.0267 0.011 37.5 95 1072 6914 0.00372
32 0.3242 64.00 5315.0 0.4559 90.25 0.0241 0.010 41.5 105 1316 8488 0.00305
33 0.2554 50.41 6748.0 0.3662 72.25 0.0216 0.009 46.3 118 1638 10565 0.00241
34 0.2011 39.69 8572.0 0.2863 56.25 0.0191 0.008 52.5 133 2095 13512 0.00189
35 0.1589 31.36 10849.0 0.2268 44.89 0.0170 0.007 58.8 149 2645 17060 0.00150
36 0.1266 25.00 13608.0 0.1813 36.00 0.0152 0.006 62.5 167 3309 21343 0.00119
37 0.1026 20.25 16801.0 0.1538 30.25 0.0140 0.006 71.6 182 3901 25161 0.00098
38 0.0811 16.00 21266.0 0.1207 24.01 0.0124 0.005 80.4 204 4971 32062 0.00077
39 0.0621 12.25 27775.0 0.0932 18.49 0.0109 0.004 91.6 233 6437 41518 0.00059
40 0.0487 9.61 35400.0 0.0723 14.44 0.0096 0.004 103.6 263 8298 53522 0.00046
41 0.0397 7.84 43405.0 0.0584 11.56 0.0086 0.003 115.7 294 10273 66260 0.00038
42 0.0317 6.25 54429.0 0.0456 9.00 0.0076 0.003 131.2 333 13163 84901 0.00030
43 0.0245 4.84 70308.0 0.0368 7.29 0.0069 0.003 145.8 370 16291 105076 0.00023
44 0.0202 4.00 85072.0 0.0316 6.25 0.0064 0.003 157.4 400 18957 122272 0.00020

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table 4-10

Dimensional Data for Film Insulated Magnetic Wire

Wire Maximum Diameter
Size Single-Insulation Heavy-Insulation Triple-Insulation Quad-Insulation
AWG Inches Centimeters Inches Centimeters Inches Centimeters Inches Centimeters
10 0.1054 0.2677 0.1071 0.2720 0.1084 0.2753 0.1106 0.2809
11 0.9410 2.3901 0.0957 0.2431 0.0969 0.2461 0.0991 0.2517
12 0.0840 0.2134 0.0855 0.2172 0.0867 0.2202 0.0888 0.2256
13 0.0750 0.1905 0.0765 0.1943 0.0776 0.1971 0.0796 0.2022
14 0.0670 0.1702 0.0684 0.1737 0.0695 0.1765 0.0715 0.1816
15 0.0599 0.1521 0.0613 0.1557 0.0624 0.1585 0.0644 0.1636
16 0.0534 0.1356 0.0548 0.1392 0.0558 0.1417 0.0577 0.1466
17 0.0478 0.1214 0.0492 0.1250 0.0502 0.1275 0.0520 0.1321
18 0.0426 0.1082 0.0440 0.1118 0.0450 0.1143 0.0468 0.1189
19 0.0382 0.0970 0.0395 0.1003 0.0404 0.1026 0.0422 0.1072
20 0.0341 0.0866 0.0353 0.0897 0.0362 0.0919 0.0379 0.0963
21 0.0306 0.0777 0.0317 0.0805 0.0326 0.0828 0.0342 0.0869
22 0.0273 0.0693 0.0284 0.0721 0.0292 0.0742 0.0308 0.0782
23 0.0244 0.0620 0.0255 0.0648 0.0263 0.0668 0.0279 0.0709
24 0.0218 0.0554 0.0229 0.0582 0.0237 0.0602 0.2520 0.6401
25 0.0195 0.0495 0.0206 0.0523 0.0214 0.0544 0.0228 0.0579
26 0.0174 0.0442 0.0185 0.0470 0.0192 0.0488 0.0206 0.0523
27 0.0156 0.0396 0.0165 0.0419 0.0172 0.0437 0.0185 0.0470
28 0.0139 0.0353 0.0148 0.0376 0.0155 0.0394 0.0166 0.0422
29 0.0126 0.0320 0.0134 0.0340 0.0141 0.0358 0.0152 0.0386
30 0.0112 0.0284 0.0120 0.0305 0.0127 0.0323 0.0137 0.0348
31 0.0100 0.0254 0.0108 0.0274 0.0115 0.0292 0.0124 0.0315
32 0.0091 0.0231 0.0098 0.0249 0.0105 0.0267 0.0113 0.0287
33 0.0081 0.0206 0.0088 0.0224 0.0095 0.0241 0.0102 0.0259
34 0.0072 0.0183 0.0078 0.0198 0.0084 0.0213 0.0091 0.0231
35 0.0064 0.0163 0.0070 0.0178 0.0076 0.0193 0.0082 0.0208
36 0.0058 0.0147 0.0063 0.0160 0.0069 0.0175 0.0074 0.0188
37 0.0052 0.0132 0.0057 0.0145 0.0062 0.0157 0.0067 0.0170
38 0.0047 0.0119 0.0051 0.0130 0.0056 0.0142 0.0060 0.0152
39 0.0041 0.0104 0.0045 0.0114 0.0050 0.0127 0.0053 0.0135
40 0.0037 0.0094 0.0040 0.0102 0.0044 0.0112 0.0047 0.0119
41 0.0033 0.0084 0.0036 0.0091 0.0040 0.0102 0.0043 0.0109
42 0.0030 0.0076 0.0032 0.0081 0.0037 0.0094 0.0038 0.0097
43 0.0026 0.0066 0.0029 0.0074 0.0033 0.0084 0.0035 0.0089
44 0.0024 0.0061 0.0027 0.0069 0.0030 0.0076 0.0032 0.0081

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 20,000

10,000
" Formvar
I 105°C Insulation
MW15-C
Bl

I
w 1,000

100
100 200 300
Film Insulation Temperature, °C

Figure 4-16. Thermal Endurance, for 105°C Formvar Insulation.

20,000

10,000

o
Polyimide (ML)
if 220°C Insulation
a2 MW16-C

1,000

100
100 200 300
Film Insulation Temperature, °C

Figure 4-17. Thermal Endurance for 220°C Polyimide Insulation (ML).

Solderable Insulation
Solderable insulation is a special film insulation that is used on magnet wire in low cost, high volume
applications. The magnet wire, with this solderable insulation, is wrapped around the terminal or pin, as
shown in Figure 4-18. Then the terminal can be dip-soldered at the prescribed temperature without prior
stripping. The ambient temperature range for this type of film insulation is 105°C to 180°C.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 There are drawbacks in using some of the solderable insulation magnet wire. Prior to using, check your
application with the wire manufacturer. Some solderable film insulation is not recommended where severe
overloads may occur. Some solderable film insulations are susceptible to softening, due to prolonged
exposure to strong solvents, such as alcohol, acetone, and methylethylketone.

Terminal
Dipped Solder Connection
Strain Relief
Solderable Insulation

Figure 4-18. Solderable Insulation on a Dip Solder Terminal.

Bondable Magnet Wire

Bondable, magnet wires are a film-coated, copper or aluminum, with an additional coating of a
thermoplastic adhesive. See Figure 4-19. They are used in applications where it is desirable to have the
bonding agent, such as a solvent, which will hold the coil form until it is oven-baked. Most adhesive
coatings can be softened with solvents or heat. If a coil is wound with an irregular shape, held in a form,
and then raised to the appropriate temperature, the coil will retain its shape. Bondable magnet wires have
applications such as armatures, field coils, and self-supporting coils.

Bondable Thermoplastic Adhesive

Film Insulation
Copper Wire

Figure 4-19. Typical Cross-Section of a Bondable Magnet Wire.

Base Film Insulation

All conventional film insulations may be adhesive-coated to achieve a bondable wire. However, care
should be taken in selecting wires, which are insulated with high temperature films, since the adhesive
coating may not withstand the equally high temperatures. See Table 4-11. The temperatures, in Table 4-11,
are for reference only. It is wise to always check with the manufacturer for the latest in materials and
application notes. The addition of the adhesive coating over the film insulation will result in an increase in
the finished diameter, by the same magnitude, as if going from a single to a heavy insulation.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table 4-11

Bondable Overcoats
Heat Solvents
Operating Activation Activating
Type Temperature Temperature Agents
Polyvinyl Butryal 105°C 120°-140°C Alcohol
Epoxy 130°C 130°-150°C Methylethylketone
Acetone
Polyester 130°C 130°- 150°C Methylethylketone
Nylon 155°C 180°-220°C None

Bonding Methods

Heat Bonding may be accomplished by the use of a temperature-controlled oven. Small components can
use a controlled hot air blower to bond the wires. In either case, caution should be used when handling the
coil while it is still hot, since deformation can take place.

Resistance Bonding is a method where a current is passed through the winding to achieve the desired
bonding temperature. This method generates a very even, heat distribution resulting in a good bonding
throughout the winding. Many coils can be resistance-bonded at the same time. The current required for
one coil, will be the same current required when many are connected in series. Just solder the coils in series
then adjust the applied voltage until the same current is reached.

Solvent Bonding is a method where the solvent activates the bonding material. This can be done, by
passing the wire through a solvent-saturated felt pad or a light spray application. There are many activating
solvents that can be used: denatured ethyl alcohol, isopropyl alcohol, methylethylketone and acetone. The
solvents should always be checked on with the manufacturer for the latest in materials and application
notes.

Miniature Square Magnet Wire

When product miniaturization calls for more copper in a given area, MWS Microsquare film, insulated
magnet wire, allows the design of compact coils to deliver more power in less space. See Table 4-12.
Microsquare magnet wire is available in both copper and aluminum. It is also available in a range of
solderable and high temperature film insulation. A cross-section of a number 26, heavy build, microsquare
magnet wire is shown in Figure 4-20.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 0.0445 cm Film Insulation
Copper Wire

0.0445 cm

Figure 4-20. Cross-Section of a #26, Heavy, Microsquare Magnet Wire.

Table 4-12

Micro-Square Magnetic Wire (Nominal Dimension)

Wire Bare Bare Wire Wire Copper Aluminum Single Heavy
Size Width Width Area Area Resistance Resistance Width Width
AWG cm Inch cm" sq-mils Q/cm Q/cm cm cm
15 0.1450 0.0571 0.019614 3041 0.0000879 0.000144 0.1483 0.1514
16 0.1290 0.0508 0.015228 2361 0.0001132 0.000186 0.1323 0.1354
17 0.1151 0.0453 0.011816 1832 0.0001459 0.000239 0.1184 0.1212
18 0.1024 0.0403 0.009675 1500 0.0001782 0.000293 0.1054 0.1080
19 0.0912 0.0359 0.007514 1165 0.0002294 0.000377 0.0940 0.0968
20 0.0813 0.0320 0.006153 954 0.0002802 0.000460 0.0841 0.0866
21 0.0724 0.0285 0.004786 742 0.0003602 0.000591 0.0749 0.0772
22 0.0643 0.0253 0.003935 610 0.0004382 0.000719 0.0668 0.0688
23 0.0574 0.0226 0.003096 480 0.0005568 0.000914 0.0599 0.0620
24 0.0511 0.0201 0.002412 374 0.0007147 0.001173 0.0536 0.0556
25 0.0455 0.0179 0.002038 316 0.0008458 0.001388 0.0480 0.0498
26 0.0404 0.0159 0.001496 232 0.0011521 0.001891 0.0427 0.0445
27 0.0361 0.0142 0.001271 197 0.0013568 0.002227 0.0389 0.0409
28 0.0320 0.0126 0.001006 156 0.0017134 0.002813 0.0348 0.0366
29 0.0287 0.0113 0.000787 122 0.0021909 0.003596 0.0312 0.0330
30 0.0254 0.0100 0.000587 91 0.0029372 0.004822 0.0277 0.0295

Multistrand Wire and Skin Effect

Electronic equipment now operate at higher frequencies, and the predicted efficiency is altered, since the
current carried by a conductor is distributed uniformly across the conductor, cross-section only, with direct
current, and at low frequencies. The flux generated by the magnet wire is shown in Figure 4-21. There is a
concentration of current near the wire surface at higher frequencies, which is termed the skin effect. This is
the result of magnetic flux lines that generate eddy currents in the magnet wire, as shown in Figure 4-22.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 External Flux, <j>
Internal Flux, (j)
Magnet Wire
Current, I

Figure 4-21. Flux Distribution in a Magnet Wire.

Eddy currents setup by

the internal flux, (|>, field.
Magnet wire cross-section

Note: The main current shown in the center

is being cancelled by the eddy currents. This
forces the current to the surface, which causes
surface crowding of the magnet wire.
Main current direction

Figure 4-22. Eddy Currents Generated in a Magnet Wire.

Reduce Skin Effect in Transformers

Skin effect accounts for the fact that the ratio of effective alternating current resistance to direct current is
greater than unity. The magnitude of this effect, at high frequency on conductivity, magnetic permeability,
and inductance, is sufficient to require further evaluation of conductor size, during design. The skin depth
is defined as the distance below the surface, where the current density has fallen to l/e or 37 percent of its
value at the surface.

6.62 [cml
e, is the skin depth [4-5]
/, isfrequencyin hertz
K, is equal to 1 for copper

When selecting the wire for high frequency, select a wire where the relationship between the ac resistance
and the dc resistance is 1.

R
R=1T
K
= 1 t4-6]
dc

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Using this approach, select the largest wire, operating at 100 kHz.

6 =

/100,000
= 0.0209, [cm]

The wire diameter is:

DAWG=2(£], [cm]
0^=2(0.0209), [cm] [4-8]
= 0-0418, [cm]

The bare wire area AW(B) is:

4B)

(3.14)(0.0418) ,
•i £- '-, [cm2] [4-9]
A
H{B) = 0-00137, [cm2]

The wire size closest to this area of 0.00137 is AWG #26 with 0.00128 cm2. (See Table 4-9).

Calculating Skin Effect in Inductors

Inductors have skin effect problems just like transformers. The skin effect depends on the amount of ac
current AI in the inductor. The high frequency inductor current has two components: the dc current, I<jc and
the ac current, AI. The dc current travels in the center of the conductor, and the ac travels on the surface of
the conductor, as shown in Figure 4-23.

Wire Diameter

I(jc Current Density AI Current Density

Skin Depth

Figure 4-23. DC Inductor High Frequency Current Distribution.

The following procedure is used to calculate the high frequency current density, AI, while using Figure 4-23
as a reference.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The skin depth equation is:

' [cm] [4-10]

I v/

Calculate the diameter of the copper conductor:

, [cm] [4-11]

Subtract two times the skin depth, e from the diameter, DAWG.

Dn=DAWC-2s, [cm] [4-12]

Calculate the new wire area, An.

An=*^-> [cm2] [4-13]

The high frequency wire area, AW(A1) is the difference between the wire area, Aw(B) and the new area, An.

-V)=-4 w ( f l )-4, [cm2] [4-14]

The ac current, AI in an inductor is a triangular waveform. The AImls current is:

-, [amps] [4-15]
»3

Calculate the current density for the delta rms current, AI^.

= rms_ ^ [amps.per.cm2 ]

The delta rms current, AI^s current density, J should be:

A/rms current density < Idc current density

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 A graph of skin depth, as a function of frequency, is shown in Figure 4-24. The relationship of skin depth
to AWG radius is shown in Figure 4-25, where Rac/Rdc= 1 is plotted on a graph of AWG versus frequency.

1.0 :

0.1
CJ

cx
•s
0.01

0.001
IK 10K 100K IMeg
Frequency, Hz

Figure 4-24. Skin Depth Versus Frequency.

50

40
Skin depth is more than the wire radius.

bJO
I
60
30
a
20
J
10
Skin depth is less than the wire radius.

IK 10K 100K IMeg

Frequency, Ffz

Figure 4-25. AWG Versus Frequency at Which Skin Depth Equals the Radius.

To illustrate how the AWG ac/dc resistance ratio changes with frequency, see Table 4-13. In Table 4-13, it
can be seen that when a converter operates at 100 kHz, the largest wire that should be used is a number 26,
with an ac/dc resistance ratio of 1.001.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table 4-13

AWG ac/dc Resistance Ratio at Common Converter Frequencies

25kHz 50kHz 100 kHz 200kHz
D(AWG) £ Rac £ Rac £ Rac £ Rac
AWG cm cm Rdc cm Rdc cm Rdc cm Rdc
12 0.20309 0.041868 1.527 0.029606 2.007 0.020934 2.704 0.014802 3.699
14 0.16132 0.041868 1.300 0.029606 1.668 0.020934 2.214 0.014802 2.999
16 0.12814 0.041868 1.136 0.029606 1.407 0.020934 1.829 0.014802 2.447
18 0.10178 0.041868 1.032 0.029606 1.211 0.020934 1.530 0.014802 2.011
20 0.08085 0.041868 1.001 0.029606 1.077 0.020934 1.303 0.014802 1.672
22 0.06422 0.041868 1.000 0.029606 1.006 0.020934 1.137 0.014802 1.410
24 0.05101 0.041868 1.000 0.029606 1.000 0.020934 1.033 0.014802 1.214
26 0.04052 0.041868 1.000 0.029606 1.000 0.020934 1.001 0.014802 1.078
28 0.03219 0.041868 1.000 0.029606 1.000 0.020934 1.000 0.014802 1.006
30 0.02557 0.041868 1.000 0.029606 1.000 0.020934 1.000 0.014802 1.000
AWG Copper, skin depth is at 20°C.

Multistrand Litz Wire

The term litz wire is extracted from the German word, meaning woven wire. Litz wire is generally defined,
as a wire constructed of individually, film insulated wires, braided together in a uniform pattern of twists
and length of lay. This multistrand configuration minimizes the power losses, otherwise encountered, in a
solid conductor, due to the skin effect. The minimum and maximum number of strand for standard litz wire
is shown in Table 4-14. Magnet wire suppliers will supply larger, twisted magnet wire on request.

Table 4-14

Standard Litz Wire

Minimum Approximate Maximum Approximate
AWG Strands AWG Strands AWG
30 3 25 20 17.0
32 3 27 20 19.0
34 3 29 20 21.0
36 3 31 60 18.5
38 3 33 60 20.5
40 3 35 175 18.0
41 3 36 175 18.5
42 3 37 175 19.5
43 3 38 175 21.0
44 3 39 175 21.5
45 3 40 175 22.5
46 3 41 175 23.5
47 3 42 175 25.0
48 3 43 175 25.5

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Proximity Effect
The operating frequency for power supplies is now in the range of 50 to 500 kHz. With it came along some
new tasks for the engineer to address skin effect and proximity effect. They are quite similar in that they
both generate eddy currents in the magnet wire. The eddy currents produced by these effects have the same
solution, keeping the ratio of the ac resistance, Rac, to the dc resistance, R<jc down:

RR=^- [4-17]
R
dc

The information provided here on proximity effect is taken from the five references provided at the end of
this Chapter. The references are excellent, providing an in-depth analysis of the losses due to proximity
effect, which is beyond the intent of this effort.

Proximity effect is caused by eddy currents induced in a wire due to the alternating magnetic field of other
conductors in the vicinity. The flux generated by the magnet wire is shown in Figure 4-26. The eddy
currents cause a distortion of the current density. This distortion is the result of magnetic flux lines that
generate eddy currents in the magnet wire, therefore enhancing the main current, I, on one side and
subtracting from the main current on the other, as shown in Figure 4-27. A magnet wire with its distorted
current density is shown in Figure 4-28.

External Flux, (j>

Internal Flux, <|>

Main Current, I
Magnet Wire

Figure 4-26. Flux Distribution in a Magnet Wire.

Eddy currents setup by

the internal flux, <j), field.
Magnet wire cross-section

Note: The main current shown in the center

is being distorted by the eddy currents. The
eddy current enhances on one side, and causes
crowding on the other.
Main current direction

Figure 4-27. Eddy Currents Generated in a Magnet Wire.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Surface Crowding
Magnet Wire

Figure 4-28. Magnet Wire, Showing Distorted Surface Crowding.

Proximity Effect in Transformers

Proximity effect has a minimum of impact on a transformer with a single layer secondary, as shown in
Figure 4-29 along with its low frequency magneto-motive force (mmf) diagram. Keeping the proximity
effect to a minimum requires the transformer to be designed with a minimum of layers. The selection of a
core with a long narrow window will produce a design with a minimum of layers, in the same way as
picking a core for a minimum of leakage inductance.

Spatial Illustration
0, mmf

Secondary-layer 1
Primary-layer 2
QQQQ
Primary-layer 1
Insulation

Core Center leg

Figure 4-29. Simple Transformer Showing mmf.

Multiple Layer High Frequency Transformers and High Loss

The proximity effect is outlined for a transformer having a secondary with three layers, evenly spaced, as
shown in Figure 4-30. A schematic diagram version of the transformer is shown in Figure 4-31, showing
the different magneto-motive force (mmf = Fm) potentials. It is assumed that the high frequency penetration
depth is 25%. The transformer has a 24 turn primary and a 24 turn secondary at 1 ampere. The
transformer, A-T or magneto-motive force, (mmf) or Fm, is equal to 24.
Insulation Spatial Illustration
0, mmf

Secondary layer 3
Secondary layer 2
Secondary layer 1
Primary-layer 3
Primary-layer 2
Primary-layer 1

Bobbin Center leg Magnet wire

Figure 4-30. A Simple Transformer, Showing the Magneto-Motive Force, mmf.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Fm = N I, [magneto-motive force, mmfj
[4-18]
cgs, Fm = 0.8,/V/, [magneto-motive force, mmfj

The schematic diagram as shown in Figure 4-31, is used as a guide to show how the proximity effect
impacts the layer wound transformers. The load current, I0, equals 1 amp, and the secondary will have three
identical layers, with each layer having eight turns. Due to the skin effect or penetration depth, each wire
uses only 25% of the available area. Therefore, the current will be crowded into 25% of the available
copper wire.

To the right of S3, the rrnnf is 0. At the left of S3, Fm = 8 A-T.

1. The magnet field, fa set up by the load current, I0 of 1 amp in layer S3 will generate a current, llg
in the winding layer, S2. It is in the opposite direction to the normal current flow and cancels the
load current, I0. The magneto-motive force, Fm, will generate 16 A-T or Ic = 2 amps to preserve
the original load current, I0, of 1 amp.
2. The magnet field, <t>2 set up by the load current, I0, plus the difference between Ic and Ig in S2 will
generate, 2Ig, in the winding layer, SI. This is in the opposite direction to the normal current flow
that cancels the load current, I0, out. The magneto-motive force, Fm will generate, 24 A-T or Ic = 3
amps, to preserve the original load current, I0, equals 1 amp.

Core Window

Primary

1st Layer 2nd Layer 3rd Layer

>- Load Current, Io = 1 amp

> Current, !„ induced by mmf in the adjacent winding due to the load current.
o
^ Current, Ic induced by mmf to counter the mmf current induced by the adjacent winding.

Figure 4-31. Transformer Schematic Diagram Showing mmf.

If the current in each layer is just the 1 amp, and limited in penetration, due to skin effect of only 25% of the
conductor's thickness, the ac to dc resistance ratio, RR, would be 4:1. The surface currents successive
layers become much larger, as discussed above. The winding currents are tabulated in Table 4-15. The
summation of the currents is given in Table 4-15. The current, Ig, is the adjacent winding induced current.
The current, Ic, is the counter current induced by the magneto-motive force, mmf.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table 4-15

Secondary Current Levels

Winding Io Ic lo + Ic lo + Ic I* Total Wire Current
amps 2
amps amps amps amps amps
S3 1 0 1 1 0 (I0 + I C ) 2 = 1

S2 1 1 2 4 1 (I0 + I c ) 2 +(I g ) 2 =5
SI 1 2 3 9 2 (I0 + I c ) 2 +(I g ) 2 =13

It can be seen, from the data in Table 4-15 that transformers with multiple layers operating at high
frequency could be a real problem with proximity effect. The eddy current losses caused by the proximity
effect go up exponentially as the number of layers. The selection of a core with a long winding length to a
winding height ratio, will reduce the number of layers to a minimum, as shown in Figure 4-32.

EE Cores. Bobbin Secondary

In
][ ^^^gj^jjjjjj;
[j[fv££ffiv<v> •»•»•> Bobbin

EE Cores
Primary -

Special Long Tongue EE Core. Standard EE Core Configuration

Figure 4-32. Comparing the Standard and the Special Long Tongue EE Cores.

Proximity Effect Using Dowell Curves

Dowell curves on proximity effect are shown in Figure 4-33. The vertical scale is, RR, the ratio of Rac to
RdC. The horizontal scale, K, is the ratio of the effective conductor height, or layer thickness, to the
penetration depth, e. On the right side of the curve it is labeled Number of Layers. These are segmented
layers. Segmented layers are when the secondary is interleaved with the primary, then, each separation is a
segment. The equation for K is:

K=
[4-19]
h = OM6DAWC

Where:

1p —. [4-20]
1

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The variables in Equation 4-20 are described in Figure 4-34. It can be seen that if the number of turns, N,
times the wire diameter, DAWG are equal to the winding length, lw, then, Equation 4-21 is simplified to:

[4-21]
h = 0.866Z),

100

10

Figure 4-33. Ratio of ac/dc Resistance Due to Proximity Effect.

TV = turns/layers

Magnet Wire COOOOOOO

t
Figure 4-34. Winding Layer Parameters.

Using the Dowell curves as shown in Figure 4-33. Compare the loss ratio between the transformer in
configuration A with two layers and transformer B that has the secondary interleaved with the primary, as
shown in Figure 4-35. With a skin effect penetration depth of 25%, it will yield a, K, factor of 4. Both
transformers, A and B, have the same A-T, but since the windings on transformer B are interleaved, it has
only half the low frequency magneto-motive force (mmf).

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 There is a vertical dotted line shown in Figure 4-33, where K = 4. Follow the dotted line up to where it
intersects 1 layer, then read the vertical column on the left, RR = 4. Now follow the dotted line up to where
it intersects 2 layer, then read the vertical column on the left, RR = 13. Transformer B with its interleaved
windings has a lower ac to dc resistance ratio, RR, by a factor 3.25.

0, mmf 0, mmf 0, mmf

Core

A B

Figure 4-35. Transformers Wound with Different Primary and Secondary Configurations.

The proximity effect, with its exponentially losses tends to be the dominant conductor loss mechanism in
high frequency magnetic components, particularly when windings are multi-layered.

Specialty Wire

There are a lot of new ideas out in the wire industry, if only the engineer had the time to evaluate these new
concepts to build confidence and apply them.

Triple Insulated Wire

Transformers designed to meet the IEC/VDE safety specification requirements for creepage and clearance
must adhere to one of the following specifications:

1. VDE0805 2. IEC950 3. EN60950 4. UL1950-3e 5. CSA 950-95

The engineer must be aware that one specification does not encompass all applications. For example, the
IEC has specifications for office machines, data-processing equipment, electromedical equipment,
appliances, and others.

Originally these IEC specifications were developed around linear 50 and 60 Hz transformers, and were not,
always, conducive to optimal designs for high frequency, such as switching power transformers. The
complexity of a standard, high frequency, switching type transformer, designed to the IEC/VDE safety
specification, is shown in Figure 4-36. In any switching transformer, coupling has the highest priority
because of the leakage flux.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Wrapper Insulation
Winding Area Primary
Bobbin Flange 3 Layers Insulation Minimum
Secondary Winding Area Positive Tape Barrier

Figure 4-36. Bobbin Cross-Section Design to Meet IEC/VDE Specifications.

The triple, insulated wire was developed to meet the above specification and eliminate the need for three
layers of insulating tape between primary and secondary. Also, the triple, insulated wire eliminates the need
for the creepage margin, and now, the whole bobbin can be used for winding. This wire can also be used as
hook-up wire, from the primary or secondary, to the circuits, without the use of sleeving or tubing.

The construction of the triple, insulated wire is shown in Figure 4-37. The temperature range for this type
of wire is from 105°C to 180°C. The dimensions for triple, insulated wire are shown in Table 4-16, using a
0.002 inch coat per layer. Other thicknesses are available. The manufacturer, Rubadue Wire Company, is
listed in the Reference section on page 4-41.

^\ ^ Copper conductor
1st insulation layer
2nd insulation layer
3rd insulation layer

Figure 4-37. Triple, Insulated Wire Construction.

Triple Insulated Litz

High frequency litz wire, shown in Figure 4-38, is also available as a triple insulated wire from
manufacturers. The insulation layers' thickness for litz wire comes in 0.002 and 0.003 inches.

1st insulation layer

Copper conductor
2nd insulation layer
Film Insulation
3rd insulation layer

Figure 4-38. Triple, Insulated Litz Wire.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table 4-16

Triple Insulated Wire (.002) Material

Bare Wire With Insulation
Area Diameter Diameter Resistance Diameter Diameter
AWG cm2(10"3) inch mm uQ/cm inch mm
16 13.0700 0.0508 1.2903 132 0.0628 1.5951
18 8.2280 0.0403 1.0236 166 0.0523 1.3284
19 6.5310 0.0359 0.9119 264 0.0479 1.2167
20 5.1880 0.0320 0.8128 332 0.0440 1.1176
21 4.1160 0.0285 0.7239 419 0.0405 1.0287
22 3.2430 0.0253 0.6426 531 0.0373 0.9474
23 2.5880 0.0226 0.5740 666 0.0346 0.8788
24 2.0470 0.0201 0.5105 842 0.0321 0.8153
25 1.6230 0.0179 0.4547 1062 0.0299 0.7595
26 1.2800 0.0159 0.4039 1345 0.0279 0.7087
27 1.0210 0.0142 0.3607 1687 0.0262 0.6655
28 0.8046 0.0126 0.3200 2142 0.0246 0.6248
29 0.6470 0.0113 0.2870 2664 0.0233 0.5918
30 0.5067 0.0100 0.2540 3402 0.0220 0.5588
32 0.3242 0.0080 0.2032 5315 0.0200 0.5080
34 0.2011 0.0063 0.1600 8572 0.0183 0.4648
36 0.1266 0.0050 0.1270 13608 0.0170 0.4318
38 0.0811 0.0040 0.1016 21266 0.0160 0.4064

Polyfilar Magnetic Wire

Poly or multiple strands of magnet wire, bonded together, can be used in many high frequency transformer
and inductor applications. Round polyfilar magnet wire is shown in Figure 4-39, and square polyfilar is
shown in Figure 4-40. Both can be used in place of foil in some applications. Polyfilar magnet wire can be
used as a foil type winding, such as a low voltage, high current, or even a Faraday shield. The polyfilar,
magnet wire strip width can be easily increased or decreased by adding or removing wires to provide the
proper strip width to fit a bobbin. It is relatively easy to wind. Polyfilar wire has complete insulation, and it
does not have the sharp edge problem that could cut insulation in the way foil does. It is not recommended
to wind a transformer with polyfilar magnet wire in order to have an exact center tap, unless it is just a few
turns, because of the penalty in capacitance. If the use of polyfilar is necessary, then use a magnet wire with
a film insulation that has a low dielectric constant. See Table 4-8.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Bondable Thermal Adhesive
Copper conductor
Film Insulation

Figure 4-39. Polyfilar, Strip-Bonded, Round Magnet Wire.

Bondable Thermal Adhesive

Copper conductor
Film Insulation

Figure 40. Polyfilar, Strip-Bonded, Square Magnet Wire.

Standard Foils

The biggest advantage for using foil over magnet wire is the fill factor. The design of a high current, high
frequency, dc to dc converter is common place. The main reason for going to high frequency is the
reduction in size. The power transformer is the largest component in the design. When designing high
frequency transformers, the design equations relate to a very small transformer. When operating
transformers at high frequencies, the skin effect becomes more and more dominate, and requires the use of
smaller wire. If larger wire is required, because of the required current density, then, more parallel strands
of wire will have to be used (litz wire). The use of small wire has a large effect on the fill factor.

When using foil, the gain in the fill factor is the biggest improvement over litz. To make a comparison, a
litz design is shown in Figure 4-41, and a foil design is shown in Figure 4-42. In the litz design, there is a
percentage of the winding area which cannot be used for the conductors. This lost area is made up of voids,
space between the wires, and the insulation film on the wire. The foil wound coil, shown in Figure 4-42,
can be designed to make optimum use of the available winding area. Each turn of the foil can extend,
within limits, edge-to-edge of the bobbin or tube. The insulation required between layers is at a minimum,
as long as the foil has been rolled to remove the sharp burr as shown in Figure 4-46.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Winding Build

Winding Length

Figure 4-41. Layer Winding, Using Litz Magnet Wire.

Winding Build

I Winding Length

Figure 4-42. Layer Winding, Using Foil with Insulation.

The Use of Foils

Designing transformers and inductors, with foil, is a very laborious task, especially if the engineer only does
it now and then. A monumental job, in itself, is finding out where to get the materials. Foil has its
advantages, mainly, in high current, high frequency, and a high density environment.

The window utilization factor, Ku, can be greater than 0.6, under the right conditions, without a lot of
force. The standard foil materials used, by transformer engineers, are copper and aluminum. The engineer
has a good selection of standard thicknesses as shown:

1.0 mil, 1.4 mil, 2.0 mil, 5.0 mil, and 10 mil

The engineer will find other thicknesses available, but standard thicknesses should be considered first. Be
careful of using a nonstandard thickness. What you might be using could be from an overrun, and could
create problems for you. Foil comes in standard widths, in inches, as shown:

0.25, 0.375, 0.50, 0.625, 0.75, 1.0, 1.25, 1.50, 2.00, 2.50, 3.00, 4.00 (inches)

Standard widths are the widths that are most readily available. There are also different styles of pre-fab
foils, as shown in Figures 4-43, 4-44, and 4-45.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Cuffed Conductor Backed Conductor

Figure 4-43. Pre-fab Foils.

Backed Multiple Conductor Sandwiched Conductor

Figure 4-44. Pre-fab Foils.

Jacketed Conductor Jacketed Multiple Conductor

Figure 4-45. Pre-fab Foils.

Although special slitting is done all the time, there is normally a minimum buy. When slitting is done,
special care must be attended to, with the sharp edges, as shown in Figure 4-46. The cut edge should be
rolled after slitting it, at least two times, to remove the sharp burrs that could cut through the insulation.
Therefore it is wise not to use insulation between layers of less than 1 mil.

Sharp edge caused by slitting.

Figure 4-46. Foil with Sharp Edge Burrs after Slitting.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 When winding transformers or inductors with foil, special care must be taken with lead finishing. One of
the biggest problems about using foil is solder wicking. This wicking will puncture the insulation, resulting
in a shorted turn. The normal insulation used for foil is very thin. Winding with foil, the coil is still
subjected to bowing, only more so, as shown in Figure 4-7.

Foil used for winding transformers and inductors should be dead soft. There is another shortcoming about
using foil, and that is, the inherit capacitance build-up, as shown in Figure 4-47.

Wrapper
Bobbin

C Core

rnmnnn
Layer Capacitance

Figure 4-47. Foil Capacitance Equation.

The capacitance build-up is expressed:

[4-22]

K = Dielectric Constant
MLT = Mean Length Turn
N = Number of Turns
G = Foil Width, cm
d = Layer Insulation Thickness, cm

The dielectric constant K for different materials can be found in Table 4-17.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table 4-17

Dielectric Constants
Material K
Kapton 3.2-3.5
Mylar 3-3.5
Kraft Paper 1.5-3.0
Fish Paper 1.5-3.0
Nomex 1.6-2.9

Calculating, MLT
The Mean Length Turn, (MLT), is required to calculate the winding resistance and weight for any given
winding. The winding dimensions, relating to the Mean Length Turn, (MLT) for a tube or bobbin coil, are
shown in Figure 4-48.

F = Winding tube thickness

(MLT)!, first winding

(MLT), single winding
(MLT>2, second winding

4, single winding
MLT, = 2(D + 2F} + 2(E + 2F} + 7iB, first winding [4-23]
MLT2 =2(D + 2F) + 2(E + 2F) + n:(2B+C), second winding

Figure 4-48. Dimensions, Relating to the Winding Mean Length Turn, (MLT).

Calculating, MLT (toroid)

It is very difficult to calculate the Mean Length Turn (MLT) for a toroidal core that would satisfy all
conditions. There are just too many ways to wind a toroid. If the toroid were designed to be wound by
machine, then that would require a special clearance for a wire shuttle. If the toroid were designed to be

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 hand-wound, the wound inside diameter would be different. The fabrication of a toroidal design is
weighted heavily on the skill of the winder. A good approximation for a toroidal core, Mean Length Turn
(MLT), is shown in Figure 4-49.

Mean Length Turn (MLT)

Wound Toroid
Toroidal Core

Toroidal Core OD

MLT = 0.8(<9Z) + 2 ( H t ) ] , approximation [4-24]

Figure 4-49. Toroidal Mean Length Turn (MLT), is an Approximation.

Copper Resistance

The dc resistance calculation of a winding requires knowing the total length, 1, of the conductor, the cross-
sectional area, Aw, of the conductor, and the resistivity, p, of the conductor material. The value for the
resistivity, p, in uQ per cm for three different conductor materials can be found in Table 4-7.

Copper Weight

The weight calculation of a winding requires knowing the total length, 1, of the conductor, the cross-
sectional area, Aw, of the conductor, and the density, A,, of the conductor material. The value for the
density, X, in grams per cm' for three different conductor materials, can be found in Table 4-7.

,, [grams] [4-21]

Electrical Insulating Materials

The reliability and life span of a magnetic component depends on the stress level put upon the insulating
materials. If the design or workmanship is not incorporated, then, insulation will not help you.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 References

1. P.L. Dowell, "Effects of Eddy Currents in Transformer Windings," Proceedings IEE (UK), Vol. 113,
No.8, August 1966, pp 1387-1394.

2. B. Carsten, "High Frequency Conductor Losses in Switch Mode Magnetics," High Frequency Power
Converter Conference, 1986, pp 155-176.

3. L. Dixon, Eddy Current Losses in Transformer Windings and Circuit Wiring, Unitrode Corp. Power
Supply Seminar Handbook, Unitrode Corp., Watertown MA, 1988.

4. E. Snelling, Soft Ferrites, pp 341-348, Iliffe, London, 1969.

5. A.I. Pressman, Switching Power Supply Design, pp 298-317, McGraw-Hill, Inc., New York 1991.

6. E.C. Snelling, Soft Ferrites. CRC Press. Iliffe Books Ltd., 42 Russell Square, London, W.C.I, 1969.

7. Werner Osterland, "The Influence of Wire Characteristics on the Winding Factor and Winding Method,"
WIRE, Coburg, Germany. Issue 97, October 1968.

8. H.A. George, "Orthocyclic Winding of Magnet Wire Without Interleaving Materials,"

Insulation/Circuits, August 1976.

9. MWS Wire Industries, "Wire Catalog," Revised June, 1992, 31200 Cedar Valley Drive, Westlake
Village, CA 91362.

10. Alpha-Core Inc. (Special Foils), 915 Pembroke Street, Bridgeport, CT 06608 Phone: (203) 335 6805.

11. Industrial Dielectrics West, Inc., (Special Foils), 455 East 9th Street, San Bernardino, CA 92410 Phone:
(909)3814734.

12. Rubadue Wire Company, Inc., (Triple Insulated Wire), 5150 E. LaPalma Avenue, Suite 108, Anaheim
Hills, CA 92807 Phone: (714) 693 5512, Email: www.rubaduewire.com.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Chapter 5

Transformer Design Trade-Offs

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table of Contents

1. Introduction 5-3
2. The Design Problem Generally 5-3
3. Power Handling Ability 5-4
4. Relationship, Ap, to Transformer Power Handling Capability 5-4
5. Relationship, Kg, to Transformer Regulation and Power Handling Capability 5-5
6. Transformer Area Product, Ap 5-6
7. Transformer Volume and the Area Product, Ap 5-6
8. Transformer Weight and the Area Product, Ap 5-9
9. Transformer Surface Area and the Area Product, Ap 5-11
10. Transformer Current Density, J, and the Area Product, Ap 5-14
11. Transformer Core Geometry, Kg, and the Area Product, Ap 5-17
12. Weight Versus Transformer Regulation 5-20
13. References 5-20

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Introduction
The conversion process in power electronics requires the use of transformer components that are frequently
the heaviest and bulkiest item in the conversion circuit. They also have a significant effect upon the overall
performance and efficiency of the system. Accordingly, the design of such transformers has an important
influence on the overall system weight, power conversion efficiency, and cost. Because of the
interdependence and interaction of these parameters, judicious trade-offs are necessary to achieve design
optimization.

The Design Problem Generally

The designer is faced with a set of constraints that must be observed in the design on any transformer. One
of these constraints is the output power, P0 (operating voltage multiplied by maximum current demand) in
that the secondary winding must be capable of delivering to the load within specified regulation limits.
Another constraint relates to minimum efficiency of operation, which is dependent upon the maximum
power loss that can be allowed in the transformer. Still another constraint defines the maximum
permissible temperature rise for the transformer when it is used in a specified temperature environment.

One of the basic steps in transformer design is the selection of proper core material. Magnetic materials
used to design low and high frequency transformers are shown in Table 5-1. Each one of these materials
has its own optimum point in the cost, size, frequency and efficiency spectrum. The designer should be
aware of the cost difference between silicon-iron, nickel-iron, amorphous and ferrite materials. Other
constraints relate to volume occupied by the transformer and, particularly in aerospace applications, weight
minimization is an important goal. Finally, cost effectiveness is always an important consideration.

Depending upon the application, some of these constraints will dominate. Parameters affecting others may
then be traded off, as necessary, to achieve the most desirable design. It is not possible to optimize all
parameters in a single design because of their interaction and interdependence. For example, if volume and
weight are of great significance, reductions in both can often be affected, by operating the transformer at a
higher frequency, but with the penalty being in efficiency. When, the frequency cannot be increased,
reduction in weight and volume may still be possible by selecting a more efficient core material, but with
the penalty of increased cost. Thus, judicious trade-offs must be affected to achieve the design goals.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Transformer designers have used various approaches in arriving at suitable designs. For example, in many
cases, a rule of thumb is used for dealing with current density. Typically, an assumption is made that a
good working level is 200 amps-per-cm (1000 circular mils-per-ampere). This rule of thumb will work in
many instances, but the wire size needed to meet this requirement may produce a heavier and bulkier
transformer than desired or required. The information presented in this Chapter makes it possible to avoid
the assumption use of this and other rules of thumb, and to develop a more economical design with great
accuracy.

Table 5-1. Magnetic Materials and Their Characteristics

Magnetic Core Material Characteristics

Material Initial Flux Density Curie dc, Coercive Operating
Name Permeability Tesla Temperature Force, He Frequency
Hi Bs °C Oersteds f
Iron Alloys
Magnesil 1.5 K 1.5-1.8 750 0.4-0.6 <2kHz
Supermendur* 0.8 K 1.9-2.2 940 0.15-0.35 <lkHz
Orthonol 2K 1.42-1.58 500 0.1-0.2 <2kHz
Sq.Permalloy 12K-100K 0.66-0.82 460 0.02-0.04 < 25kHz
Supermalloy 10K-50K 0.65-0.82 460 0.003-0.008 < 25kHz
Amorphous
2605-SC 3K 1.5-1.6 370 0.03-0.08 < 250kHz
2714A 20K 0.5-0.58 >200 0.008-0.02 < 250kHz
Vitro perm 500 30K 1.0-1.2 >200 <0.05 < 250kHz
Ferrite
MnZn 0.75-15K 0.3-0.5 100-300 0.04-0.25 <2MHz
NiZn 15-1500 0.3-0.5 150-450 0.3-0.5 < 100MHz
* Field Anneal.

Power Handling Ability

For years, manufacturers have assigned numeric codes to their cores to indicate their power-handling
ability. This method assigns to each core a number called the area product, Ap. That is the product of the
window area, Wa, and the core cross-section, Ac. The core suppliers use these numbers to summarize
dimensional and electrical properties in their catalogs. They are available for laminations, C cores, fertile
cores, powder cores, and toroidal tape wound cores.

Relationship, Ap, to Transformer Power Handling Capability

Transformers
According to the newly developed approach, the power handling capability of a core is related to its area
product, Ap, by an equation, which may be stated as:

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 A= —, [cm4]
' KfKuBmJf [5_l}

Where:

Kf = waveform coefficient
4.0 square wave
4.44 sine wave

From the above, it can be seen that factors such as flux density, frequency of operation, and window
utilization factor Ku, define the maximum space which may be occupied by the copper in the window.

Relationship, Kg, to Transformer Regulation and Power Handling Capability

Although most transformers are designed for a given temperature rise, they can also be designed for a given
regulation. The regulation and power-handling ability of a core is related to two constants:

«=—^—, [%]
2K K
s * [5-2]

a = Regulation (%) [5-3]

The constant, Kg, (See Chapter 7) is determined by the core geometry, which may be related by the
following equations:

g t
* MLT [5-4]

The constant, Ke, is determined by the magnetic and electric operating conditions, which may be related by
the following equation:

Where:
Kf = waveform coefficient
4.0 square wave
4.44 sine wave

From the above, it can be seen that factors such as flux density, frequency of operation, and waveform
coefficient, have an influence on the transformer size. Because of their significance, the area product, Ap,
and the core geometry, Kg, are treated extensively in this handbook. A great deal of other information is
also presented for the convenience of the designer. Much of the information is in tabular form to assist
designers in making the trade-offs best suited for the particular application, in a minimum amount of time.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Transformer Area Product, Ap
The author has developed additional relationships between, Ap, numbers and current density, J, for given
regulation and temperature rise. The area product, Ap, is a length dimension to the fourth power, ( I4 ), as
shown in Figure 5-1.

Wa=FG, [cm2]
Ae=DE, [cm2] [5-6]
4
Ap=WaAc, [cm ]

A
W,,

•*-F-H E -* *• D

Figure 5-1. C Core Outline Showing the Window Area, Wa and Iron Area, Ac.

It should be noted. The constants for tape-wound cores, such as: KVO|, Kw, Ks, Kj and Kp will have a
tendency to jump around and not be consistent. This inconsistency has to do with the core being in a
housing, without true proportions.

Transformer Volume and the Area Product, Ap

The volume of a transformer can be related to the area product, Ap of a transformer, treating the volume, as
shown in Figures 5-2 to 5-4, as a solid quantity without any subtraction for the core window. The
relationship is derived according to the following reasoning: Volume varies in accordance with the cube of
any linear dimension, (1), whereas area product, Ap, varies as the fourth power:

Volume = Kf, [cm3] [5-7]

Ap=K2l\ [cm4] [5-8]

I4=^r [5-9]

\(0.25)

7 = 1^-1 [5-10]

[5-11]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Volume = /d-f t5-12]
K-,

K,
.(0.75)
[5-13]

The volume-area product, Ap, relationship is therefore:

Volume = KvolA(°7i\ [cm3] [5-14]

in which, Kvol, is a constant related to core configuration whose values are given in Table 5-2. These
values were obtained by averaging the values from the data taken from Tables 3-1 through Tables 3-64 in
Chapter3.

Table 5-2. Volume-Area Product Relationship.

Volume -Area Product Relationship

Core Type KVoi
Pot Core 14.5
Powder Core 13.1
Laminations 19.7
CCore 17.9
Single-coil C Core 25.6
Tape-wound Core 25.0

Height
Volume

Figure 5-2. Toroidal Transformer Outline, Showing the Volume.

Volume

Figure 5-3. El Core Transformer Outline, Showing the Volume.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Volume

Figure 5-4. C Core Transformer Outline, Showing the Volume.

The relationship between volume and area product, Ap, for various core types is graphed in Figures 5-5
through 5-7. The data for these Figures has been taken from Tables in Chapter 3.

1000

luOQ
CJ /
/
of
/
o /
> 10
x
1.0
^x Laminations

0. 01 0.1 1.0 10 , 100 100(

Area Product, (Ap, cnv)

Figure 5-5. Volume Versus Area Product, Ap for El Laminations.

1000

luoo
o /
/ :

a
"o
X /
> 10
x
1.0
_x
0. 01 0.1 1.0 10 „
cc Dres

100 100(
Area Product, (A_, cm4)

Figure 5-6. Volume Versus Area Product, Ap for C Cores.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 1000

100

1 10
Toroidal Cores

1.0
0.01 0.1 1.0 10 100 1000
Area Product, (Ap, cm4)

Figure 5-7. Volume Versus Area Product, Ap, for Toroidal MPP Cores.

Transformer Weight and the Area Product, Ap

The total weight of a transformer can also be related to the area product, Ap, of a transformer. The
relationship is derived according to the following reasoning: weight, W,, varies, in accordance with the
cube of any linear dimension 1, whereas area product, Ap, varies, as the fourth power:

W,=K,f, [grains] [5-15]

Ap=K2l\ [cm4] [5-16]

/4=^ [5-17]

s(0.25)
. I JI I

'=!TH [5-18]
A,

075

-^ I [5-20]

(0.75)
[5-21]
K

The weight-area product, Ap, relationship is therefore:

[5-22]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 in which, K w , is a constant related to core configuration, whose values are given in Table 5-3, These values
were obtained by averaging the values from the data taken from Tables 3-1 through Tables 3-64 in Chapter
3.
Table 5-3. Weight-Area Product Relationship.

Weight- Area Product Relationship

Core Type Kw
Pot Core 48.0
Powder Core 58.8
Laminations 68.2
CCore 66.6
Single-coil C Core ^ 76.6
Tape- wound Core 82.3

The relationship between weight and area product, Ap, for various core types is graphed in Figures 5-8
through 5-10. The data for Figures 5-8 through 5-10 has been taken from Tables in Chapter 3.

10,000

11000
2bfl

• ao
100

Laminations

10
0.01 0.1 1.0 10 100 1000
Area Product, (Ap, cm4)
Figure 5-8. Total Weight Versus Area Product, Ap, for El Laminations.

10,000

5 1000
x /•_
2 /
//
/
M
'| 100

C C cDres

10
0. 01 0.1 1.0 10 , 100 100(
Area Product, (Ap, cm4)
Figure 5-9. Total Weight Versus Area Product, Ap, for C Cores.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 lOOOF

g 100
2
60

GO

I 10
Toroidal Cores

1.0
0.01 0.1 1.0 10 100 1000
Area Product, (Ap, cnr)

Figure 5-10. Total Weight Versus Area Product, Ap, for Toroidal MPP Cores.

Transformer Surface Area and the Area Product, Ap

The surface area of a transformer can be related to the area product, Ap, of a transformer, treating the
surface area, as shown in Figure 5-11 through 5-13. The relationship is derived in accordance with the
following reasoning: the surface area varies with the square of any linear dimension ( 1 ) , whereas the area
product, Ap, varies as the fourth power.

A,=KJ\ [cm2] [5-23]

A =K2l\ [cm4] [5-24]

A",

(0.25)

/ = |-M [5-26]
A,

[5-27]

,=K<\-£-\ [5-28]
A2

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The relationship between surface area, A, and area product, Ap can be expressed as:

A,=K,A(*S] [5-30]

in which, Ks, is a constant related to core configuration, whose values are given in Table 5-4. These values
were obtained by averaging the values from the data taken from Tables 3-1 through Tables 3-64 in Chapter
3.
Table 5-4. Surface Area-Area Product Relationship.

Surface Area-Area Product Relationship

Core Type Ks
Pot Core 33.8
Powder Core 32.5
Laminations 41.3
CCore 39.2
Single-coil C Core 44.5
Tape-wound Core 50.9
The surface area for toroidal type transformers is calculated, as shown below.

Top and Bottom Surface = 2 — '— , [cm2 ]

I 4 J
Periphery Surface = (x(OD))(Height), [cm2] [5-31]

A, = —± '— + (x(OD)( Height), [cm2 ]

Surface Area Height

Figure 5-11. Toroidal Transformer Outline Showing the Surface Area.

The surface areas for C cores, Laminations and similar configurations are calculated as shown below.
There is a small amount of area that is deducted because the sides and the ends are not a complete square.

End = (Height) (Length), [cm2]

Top= (Length)(Width), [cm2]
[5-32]
Side = (Height) (Width), [cm2 ]
Surface Area = 2(End)+ 2(Top)+ 2(Side), [cm2]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Height, cm

Length, cm Width, cm

Figure 5-12. C Core Transformer Outline, Showing the Surface Area.

Height, cm

Length, cm
Width, cm

Figure 5-13. Typical EE or El Transformer Outline, Showing the Surface Area.

The relationship between surface area and area product, Ap, for various core types is graphed in Figures 5-
14 through 5-16. The data for these Figures has been taken from Tables in Chapter 3.

1000

1000
Area Product, (Ap, cnv)

Figure 5-14. Surface Area, A,, Versus Area Product, Ap, for El Laminations.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 1000
^s
rvP

^
J,^
10
J, °
^
0)
o ^^^
<g
W
3
10
Jl^
-x^
^

^ C Cores :

1.0
0.01 0.1 1.0 10 100 1000
Area Product, (Ap, cm4)

Figure 5-15. Surface Area, At, Versus Area Product, Ap, for C Cores.

1000

100
rf

10
&o
Toroidal Cores

1.0
0.001 0.01 0.1 1.0 10 100
Area Product, (A_, cm4)

Figure 5-16. Surface Area, At, Versus Area Product, Ap, for Toroidal MPP Cores.

Transformer Current Density, J, and the Area Product, Ap

The current density, J, of a transformer can be related to the area product, Ap, of a transformer for a given
temperature rise. The relationship can be derived as follows:

A,=K,A(™\ [cm2] [5-33]

Pcu = I2R, [watts] [5-34]

I = AWJ, [amps] [5-35]

Therefore,
[5-36]
And since,
[5.37]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 We have:

[5-38]

p
cu = A»j2 (MLT)yVp [5-39]
Since MLT has a dimension of length,

MLT = K5A(°-"} [5-40]

[5-41]

[5-42]

2
P [5-43]
Let:
[5-44]

Then assuming the core loss is the same as the copper loss for optimized transformer operation: (See
Chapter 6),

Pcu=K7Af75]J2=Pfe [5-45]

P^P^+Pfi [5-46]

A7 = ^8 [5-47]

- t5-48]
To simplify, let:

9 = l [5.49]
9
K.
Then,
= K9J2A(°25) [5-50]

Then, letting:

[5-52]

We have:
[5-53]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The relationship between current density, J, and area product, Ap, can, therefore, be expressed as:

[5-54]

The constant, Kj, is related to the core configuration, whose values are given in Table 5-5. These values
have been derived by averaging the values from the data taken from Tables 3-1 through Tables 3-64 in
Chapter3.

Table 5-5. Constant, Kj, for Temperature Increase of 25°C and 50°C.

Temperature Constant, Kj
Core Type Kj (A25°) Kj (A50°)
Pot Core 433 632
Powder Core 403 590
Laminations 366 534
CCore 322 468
Single-coil C Core 395 569
Tape-wound Core 250 365

The relationship between current density, J, and area product, Ap, for temperature increases of 25°C and
50°C is graphed in Figures 5-17 through 5-19 from data calculated of Tables 3-1 through 3-64 in Chapter 3.

1000
800
600
I
400
£>
GO
a
<u P0 = Output Power
2 200
Pcu,Copper Loss = Pfe, Iron Loss
I Laminations
u
100
0.01 0.1 1.0 10 100 1000
Area Product, A» (crrr)

Figure 5-17. Current Density, J, Versus Area Product, Ap, for El Laminations.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 1000
., 800

P0 = Output Power
Pcu ,Copper Loss = Pfe, Iron

100
o.oi o.i i.o 10 100 1000
Area Product, A« (cm*)

Figure 5-18. Current Density, J, Versus Area Product, Ap, for C Cores.

P.u .Copper Loss » P,, Iron Loss

0.01 0.1 1.0 10 100 1000

Area Product, An (cm-')

Figure 5-19. Current Density, J, Versus Area Product, Ap> for MPP Cores.

Transformer Core Geometry, Kg, and the Area Product, Ap

The core geometry, Kg, of a transformer can be related to the area product, Ap. The relationship is
according to the following: the core geometry, Kg, varies in accordance with the fifth power of any linear
dimension, (1), whereas area product, Ap, varies as the fourth power.

w
K = AKu ^Cm5-| |-5_55-j
g
MLT

[5-56]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 A =K2l4 [5-57]

From Equation 5-56,

[5-58]

Then,

=l H [5-59]

Substituting Equation 5-59 into Equation 5-57,

(0.8)

A=K
P ~ 2
[5-60]

Let:

[5-61]
A

Then,

[5-62]

The constant, Kp, is related to the core configuration, whose values are given in Table 5-6. These values
have been derived by averaging the values from the data taken from Tables 3-1 through Tables 3-64 in
Chapter3.

Table 5-6. Configuration Constant, Kp, for Area Product, Ap, and Core geometry, Kg.

Constant, Kp
Core Type KP
Pot Core 8.9
Powder Core 11.8
Laminations 8.3
CCore 12.5
Tape-wound Core 14.0

The relationship between area product, Ap, and core geometry, Kg, is graphed in Figures 5-20 through 5-22,
from the data taken from Tables 3-1 through Tables 3-64 in Chapter 3.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 1000

100

o
-§o
10

Laminations

1.0
0.01 0.1 10 10 100 1000
Core Geometry, (Kg cm3)

Figure 5-20. Area Product, Ap, Versus Core Geometry, Kg, for El Laminations.

100

10
ea Produc

C Cores

0.1
0.001 0.01 0.1 1.0 10 100
Core Geometry, (K™ air)

Figure 5-21. Area Product, Ap, Versus Core Geometry, Kg, for C Cores.

10

g
o
cu 1.0

rs
o

ca
o.i
*-4
< Toroidal Cores

o.oi
0.0001 0.001 0.01 0.1 1.0 10
Core Geometry, (Kg cnr>)

Figure 5-22. Area Product, Ap, Versus Core Geometry, Kg, for MPP Powder Cores.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Weight Versus Transformer Regulation

There are many design tasks where the transformer weight is very important in meeting the design
specification. The engineer will raise the operating frequency in order to reduce the size and weight. The
magnetic materials will be reviewed for performance at the operating frequency and at minimum and
maximum temperatures. When the idealized magnetic material has been found and the weight of the
transformer is still too high, then the only solution is to change the regulation. The regulation of a
transformer versus the weight is shown in Figure 5-23. There are times when the engineer would like to
know what the weight impact would be, if the regulation were to be increased or decreased.

10
Graph is normalized at 1 %

£top1.0
'53

0.1
0.1 1.0 10
Regulation, a %

Figure 5-23. Weight Versus Regulation.

References

1. C. McLyman, Transformer Design Tradeoffs, Technical Memorandum 33-767 Rev. 1, Jet Propulsion
Laboratory, Pasadena, CA.
2. W. J. Muldoon, High Frequency Transformer Optimization, HAC Trade Study Report 2228/1130,
May, 1970
3. R. G. Klimo, A. B. Larson, and J. E. Murray, Optimization Study of High Power Static Inverters and
Converters, Quarterly report No. 2 NASA-CR-54021, April 20, 1964, Contract NAS 3-2785.

20
Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.
 4. F. F. Judd and D. R. Kessler, Design Optimization of Power Transformers, Bell Laboratories,
Whippany, New Jersey IEEE Applied Magnetics Workshop, June 5-6, 1975

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Chapter 6

Transformer-Inductor Efficiency, Regulation,

and Temperature Rise

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table of Contents

1. Introduction
2. Transformer Efficiency
3. Maximum Efficiency
4. Transformer Dissipation, by Radiation and Convection
5. Temperature Rise Versus Surface Area, At, Dissipation
6. Surface Area, At, Required for Heat Dissipation
7. Required Surface Area, At
8. Regulation as a Function of Efficiency
9. References

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Introduction
Transformer efficiency, regulation, and temperature rise are all interrelated. Not all of the input power to
the transformer is delivered to the load. The difference between input power and output power is converted
into heat. This power loss can be broken down into two components: core loss, Pfe, and copper loss, Pcu.
The core loss is a fixed loss, and the copper loss is a variable loss that is related to the current demand of
the load. The copper loss increases by the square of the current and also is termed a quadratic loss.
Maximum efficiency is achieved when the fixed loss is equal to the quadratic loss at rated load.
Transformer regulation, a, is the copper loss, Pcu, divided by the output power, P0.

a =^(100), [%] [6-1]

*0

Transformer Efficiency

The efficiency of a transformer is a good way to measure the effectiveness of the design. Efficiency is
defined as the ratio of the output power, P0, to the input power, P;n. The difference between, P0, and, P^, is
due to losses. The total power loss, Pj;, in the transformer is determined by the fixed losses in the core and
the quadratic losses in the windings or copper. Thus,

Pz=Pfo+P,u, [watts] [6-2]

Where, Pfe is the core loss, and Pcu is the copper loss.

Maximum Efficiency
Maximum efficiency is achieved when the fixed loss is made equal to the quadratic loss, as shown by
Equation 6-12. A graph of transformer loss versus output load current is shown in Figure 6-1.

The copper loss increases as the square of the output power, P0, multiplied by a constant, K<.:

Pa,=KcP? [6-3]
Which may be rewritten as:
2
0 [6-4]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Since:
P^Po+Pz [6-5]

The efficiency can be expressed as:

?7=- [6-6]

Then, substituting Equation 6-4 into 6-6 gives:

/>„
[6-7]
P0 + Pfe + KP^ Pfe +P0+ KP*

And, differentiating with respect to P0:

dr) Pfe+i
'-J- [6-8]
dP

Then, to solve for the maximum, equate Equation 6-8 to 0.

-P0 (1 + 2KP0 ) + (Pfe +P0+ KPl ) = 0 [6-10]

-P0 -2KP* +Pfe +P0 +KP* = 0 [6-1 1]

Therefore,
Pfo=KP?=Pa, [6-12]

12
Maximum Efficiency Occurs When, Pcu = Pfe
W 10
PH

Copper (l2R) Losses -^..-•-

4 Fixed Losses
I

\ I
20 40 60 80 100 120
Output Load Current, %

Figure 6-1. Transformer Losses Versus Output Load Current.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Transformer Dissipation, by Radiation and Convection
Temperature rise in a transformer winding cannot be predicted with complete precision, despite the fact
that many techniques are described in the literature for its calculation. One reasonable accurate method for
open core and winding construction is based upon the assumption that core and winding losses may be
lumped together as:

And the assumption is made that thermal energy is dissipated uniformly throughout the surface area of the
core and winding assembly.

Transfer of heat by thermal radiation occurs when a body is raised to a temperature above its surroundings
and emits radiant energy in the form of waves. In accordance with Stefan-Boltzmann Law, (Ref. 1) this
transfer of heat may be expressed as:

Wr=KrE(T}-T?} [6-14]
Where:
Wr , is watts per square centimeter of surface
Kr = 5.70(lO~n}w/(cm2/K4)
e, is the emissivity factor
T2 , is the hot body temperature, K (kelvin)
7J, is the ambient or surrounding temperature, AT (kelvin)

Transfer of heat by convection occurs when a body is hotter than the surrounding medium, which is usually
air. The layer of air in contact with the hot body that is heated by conduction expands and rises, taking the
absorbed heat with it. The next layer, being colder, replaces the risen layer and, in turn, on being heated,
also rises. This transfer continues as long as the air, or other medium surrounding the body, is at a lower
temperature. The transfer of heat by convection is stated mathematically as:

Wc=KcF6(n]4P [6-15]

Where:
Wc , is the watts loss per square centimeter
Kc =2.17(l<r 4 )
F, is the air friction factor (unity for a vertical surface)
9, is the temperature rise, °C
P, is the relative barometic pressure (unity at sea level)
77, is the exponential value, which ranges from 1.0 to 1.25,
depending on the shape and position of the surface being cooled
The total heat dissipated from a plane vertical surface is expressed by Equations 6-13 and 6-15:

)
VP [6-16]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Temperature Rise Versus Surface Area, At, Dissipation

The temperature rise that can be expected for various levels of power loss is shown in the monograph of
Figure 6-2. It is based on Equation 6-16, relying on data obtained from Blume (1938) (Ref. 1) for heat
transfer affected by the combination of 55% radiation and 45% convection, from a surface having an
emissivity of 0.95, in an ambient temperature of 25°C, at sea level. Power loss (heat dissipation) is
expressed in watts per square centimeter of the total surface area. Heat dissipation, by convection from the
upper side of a horizontal flat surface, is on the order of 15-20% more than from vertical surface. Heat
dissipation, from the underside of a horizontal flat surface, depends upon area and conductivity.

1.0

CJ
0.1
a
a

0.01
Q
Emissivity 0.95
45% Convection
55% Radiation

0.001
10°C 100°C
AT = Temperature Rise, Degrees °C

Figure 6-2. Temperature Rise Versus Surface Dissipation.

(Adapted from L. F. Blume, Transformers Engineering, Wiley, New York, 1938, Figure 7.)

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Surface Area, At, Required for Heat Dissipation

The effective surface area, At, required to dissipate heat, (expressed as watts dissipated per unit area), is:

4 = — , [cm2] [6-17]

In which \y is the power density or the average power dissipated per unit area from the surface of the
transformer and, PE) is the total power lost or dissipated.

The surface area, A,, of a transformer can be related to the area product, Ap, of a transformer. The straight-
line logarithmic relationship, shown in Figure 6-3, has been plotted from the data in Chapter 3. The
derivation for the surface area, A,, and the area product, Ap, is in Chapter 5.

1000

100
of
<D

10
00

Laminations

1.0
0.01 0.1 1.0 10 100 1000
Area Product, (Ap, cnv)

Figure 6-3. Surface Area, A, Versus Area Product, Ap.

From this surface area, At, the following relationship evolves:

= , [cm2] [6-18]

And from Figure 6-3:

i// = 0.03, [watts-per-cm2 at 25°C]

i// = 0.07, [watts-per-cm2 at 50°C]

The temperature rise, Tr, equation in °C is:

[6-19]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Required Surface Area, At

There are two common allowable temperature rises for transformers above the ambient temperature. These
temperatures are shown in Figure 6-4. The surface area, A,, required for a 25°C and 50°C rise above the
ambient temperature for the total watts, dissipated. The presented data is used as a basis for determining
the needed transformer surface area, A,, in cm2.

1000

25°C Rise
6u
50°C Rise

100

10
0.1 1.0 10 100
Total Watts Loss,

Figure 6-4. Surface Area, A, Versus total Watts Loss for Temperature increases of 25°C and 50°C.

If the transformer is said to be homogeneous, and the thermal energy is dissipated uniformly throughout the
surface area of the core and winding assembly, then Figure 6-5 will give a good approximation for the
required time constant for a transformer to reach 63% of the final temperature. The temperature rise of a
typical transformer is shown in Figure 6-6.

10
i
o
ffi

I
W

o i.o
H
"e3

0.1
10 100 IK 10K 100K
Transformer Total Weight - Grams

Figure 6-5. Time Required to Reach 63% of Final Temperature.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Actual Temperature Rise in Transformer
Exponential Temperature Rise
9 = Temperature above ambient at any instant t
GQ = Initial temperature above ambient

1 2
Time in multiples of thermal time constant.

Figure 6-6. Transformer Temperature Rise Time.

Regulation as a Function of Efficiency

The minimum size of a transformer is usually determined either by a temperature rise limit, or by allowable
voltage regulation, assuming that size and weight are to be minimized. Figure 6-7 shows a circuit diagram
of a transformer with one secondary. Note that a = regulation (%).

Primary Secondary
n = NJNL = 1

Figure 6-7. Transformer Circuit Diagram.

The assumption is that distributed capacitance in the secondary can be neglected because the frequency and
secondary voltage are not excessively high. Also, the winding geometry is designed to limit the leakage
inductance to a level, low enough, to be neglected, under most operating conditions. The transformer
window allocation is shown in Figure 6-8.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 W
— = Primary = Secondary [6-20]

Figure 6-8. Transformer Window Allocation.

Transformer voltage regulation can now be expressed as:

..
L J
A V /'
[6-21]

in which, V0(N.L.), is the no load voltage and, VO(F.L.), is the full load voltage. For the sake of simplicity,

assume the transformer, in Figure 6-5, is an isolation transformer, with a 1:1 turns ratio, and the core
impedance, Re, is infinite.

If the transformer has a 1:1 turns ratio, and the core impedance is infinite, then:

/;„=/„, [amps]

R=RS, [ohms] [6-22]

With equal window areas allocated for the primary and secondary windings, and using the same current
density, J:

AV=I,,R=*V,=I0R,, [volts]
[6-23]
Then Regulation is:

a = ( 1 0 0 )+ (100), [%]
[6-24]

Multiply the equation by currents, I:

a =^*L(100) + ^^(100), L [%]

V.I,. V } V.I. V ' J
[6-25]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Primary copper loss is:

^ = A K , 4 , [watts] [6 _ 26]

Secondary copper loss is:

P,=W,I0, [watts] [6.27]

Total copper loss is:

PCU=PP+PS, [watts] [6 _ 2g]

Then, the regulation equation can be rewritten to:

a = ^-(100), [%]
^° [6-29]

References

1. Blume, L.F., Transformer Engineering, John Wiley & Sons Inc. New York, N.Y. 1938. Pages 272-
282.
2. Terman, F.E., Radio Engineers Handbook, McGraw-Hill Book Co., Inc., New York, N.Y., 1943.
Pages 28-37.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Chapter 7

Power Transformer Design

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table of Contents

1. Introduction
2. The Design Problem Generally
3. Power-Handling Ability
4. Output Power, P0, Versus Apparent Power, Pt, Capability
5. Transformers with Multiple Outputs
6. Regulation
7. Relationship, Kg, to Power Transformer Regulation Capability
8. Relationship, Ap, to Transformer Power Handling Capability
9. Different Cores Same Area Product
10. 250 Watt Isolation Transformer Design, Using the Core Geometry, Kg, Approach
11. 38 Watt 100kHz Transformer Design, Using the Core Geometry, Kg, Approach

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Introduction
The conversion process in power electronics requires the use of transformers and components that are
frequently the heaviest and bulkiest item in the conversion circuit. They also have a significant effect upon
the overall performance and efficiency of the system. Accordingly, the design of such transformers has an
important influence on the overall system weight, power conversion efficiency and cost. Because of the
interdependence and interaction of parameters, judicious tradeoffs are necessary to achieve design
optimization.

The Design Problem Generally

The designer is faced with a set of constraints that must be observed in the design on any transformer. One
of these constraints is the output power, P0, (operating voltage multiplied by maximum current demand).
The secondary winding must be capable of delivering to the load within specified regulation limits.
Another constraint relates to the minimum efficiency of operation, which is dependent upon the maximum
power loss that can be allowed in the transformer. Still another defines the maximum permissible
temperature rise for the transformer when it is used in a specified temperature environment.

One of the basic steps in transformer design is the selection of proper core material. Magnetic materials
used to design low and high frequency transformers are shown in Table 7-1. Each one of these materials
has its own optimum point in the cost, size, frequency and efficiency spectrum. The designer should be
aware of the cost difference between silicon-iron, nickel-iron, amorphous and ferrite materials. Other
constraints relate to the volume occupied by the transformer and, particularly in aerospace applications, the
weight, since weight minimization is an important goal in today's electronics. Finally, cost effectiveness is
always an important consideration.

Depending upon the application, certain ones of these constraints will dominate. Parameters affecting
others may then be traded off as necessary to achieve the most desirable design. It is not possible to
optimize all parameters in a single design because of their interaction and interdependence. For example, if
volume and weight are of great significance, reductions in both can often be affected, by operating the
transformer at a higher frequency, but, at a penalty in efficiency. When the frequency cannot be increased,
reduction in weight and volume may still be possible by selecting a more efficient core material, but, at the
penalty of increased cost. Thus, judicious trade-offs must be affected to achieve the design goals.

Transformer designers have used various approaches in arriving at suitable designs. For example, in many
cases, a rule of thumb is used for dealing with current density. Typically, an assumption is made that a
good working level is 200 amps-per-cm2 (1000 circular mils-per-ampere). This will work in many

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table 7-1 Magnetic Materials

Magnetic Material Properties

Material Trade Initial Flux Density Typical
Name Name Permeability Tesla Operating
Composition Hi Bs Frequency
Silicon 3-97 SiFe 1500 1.5-1.8 50-2k
Orthonol 50-50 NiFe 2000 1.42-1.58 50-2k
Permalloy 80-20 NiFe 25000 0.66-0.82 lk-25k
Amorphous 2605SC 1500 1.5-1.6 250k
Amorphous 2714A 20,000 0.5-6.5 250k
Amorphous Nanocrystalline 30,000 1.0-1.2 250k
Ferrite MnZn 0.75-15k 0.3-0.5 10k-2M
Ferrite NiZn 0.20-1. 5k 0.3-0.4 0.2M-100M

instances, but the wire size needed to meet this requirement may produce a heavier and bulkier transformer
than desired or required. The information presented in this volume makes it possible to avoid the use of
this assumption and other rules of thumb, and to develop a more economical design with great accuracy.

Power-Handling Ability

For years manufacturers have assigned numeric codes to their cores; these codes represent the power-
handling ability. This method assigns to each core a number that is the product of its window area, Wa, and
core cross-section area, Ac, and is called the area product, Ap.

These numbers are used by core suppliers to summarize dimensional and electrical properties in their
catalogs. They are available for laminations, C-cores, pot cores, powder cores, ferrite toroids, and toroidal
tape-wound cores.

The regulation and power-handling ability of a core is related to the core geometry, Kg. Every core has its
own inherent, Kg. The core geometry is relatively new, and magnetic core manufacturers do not list this
coefficient.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Because of their significance, the area product, Ap, and core geometry, Kg, are treated extensively in this
book. A great deal of other information is also presented for the convenience of the designer. Much of the
material is in tabular form to assist the designer in making trade-offs, best-suited for his particular
application in a minimum amount of time.

These relationships can now be used as new tools to simplify and standardize the process of transformer
design. They make it possible to design transformers of lighter weight and smaller volume, or to optimize
efficiency, without going through a cut-and-try, design procedure. While developed especially for
aerospace applications, the information has wider utility, and can be used for the design of non-aerospace,
as well.

Output Power, P0, Versus Apparent Power, Pt, Capability

Output power, P0, is of the greatest interest to the user. To the transformer designer, the apparent power, Pt,
which is associated with the geometry of the transformer, is of greater importance. Assume, for the sake of
simplicity, that the core of an isolation transformer has only two windings in the window area, a primary
and a secondary. Also, assume that the window area, Wa, is divided up in proportion to the power-handling
capability of the windings, using equal current density. The primary winding handles, P^, and the
secondary handles, P0, to the load. Since the power transformer has to be designed to accommodate the
primary, P^, and, P0, then,

By definition:

Pr=P.n+Po, [watts]
P
m =~> [watts]
n [7-1]
The primary turns can be expressed using Faraday's Law:

The winding area of a transformer is fully utilized when:

By definition the wire area is:

, / r ,.
[7-4]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Rearranging the equation shows:

[7-5]

Now, substitute in Faraday's Equation:

" " AcBaJK\J

Rearranging shows:

-, [cm4]
*«/•'*/*„
B JK K
acf f U [7_ 7 ]

The output power, P0, is:

P,=V,f,, [watts] [7.g]

The input power, Pj,,, is:

P"• = Ff / P', [watts]

L -, m
J r[7-9]

Then:

^ = ^ + P o , [watts] [7-10]

Substitute in, Pt:

P(l0 4 )
WaAc = ^ '—, [cm4] [7-11]

By definition, Ap, equals:

Ar=WaAc, [cm4] [7-12]

Then:

*,= /',,/„ > t cm4 ] t7-133

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The designer must be concerned with the apparent power, Pt, and power handling capability of the
transformer core and windings. P, may vary by a factor, ranging from 2 to 2.828 times the input power, Pjn,
depending upon the type of circuit in which the transformer is used. If the current in the rectifier
transformer becomes interrupted, its effective RMS value changes. Thus, transformer size is not only
determined by the load demand, but also, by application, because of the different copper losses incurred,
due to the current waveform.

For example, for a load of one watt, compare the power handling capabilities required for each winding,
(neglecting transformer and diode losses, so that Pjn = P0) for the full-wave bridge circuit of Figure 7-1, the
full-wave center-tapped secondary circuit of Figure 7-2, and the push-pull, center-tapped full-wave circuit
in Figure 7-3, where all the windings have the same number of turns, (N).

Figure 7-1. Full-Wave Bridge Secondary.

The total apparent power, Pt, for the circuit shown in Figure 7-1 is 2 watts.
This is shown in the following equation:

P,=Pin+Po, [watts] [7.14]

P,=2Pin, [watts] [7-15]

CR1
n

'2 ^ I0

' J .
CR2 1
Figure 7-2. Full-Wave, Center-Tapped Secondary.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 o—*

Figure 7-3. Push-Pull Primary, Full-Wave, Center-Tapped Secondary.

The total power, Pt, for the circuit shown in Figure 7-2, increased 20.7%, due to the distorted wave form of
the interrupted current flowing in the secondary winding. This is shown in the following equation:

P,=Pin+P0j2, [watts] [?_

16]

P, =Pin(\ +72), [watts]

[7-17]

The total power, Pt, for the circuit is shown in Figure 7-3, which is typical of a dc to dc converter. It
increases to 2.828 times, Pin, because of the interrupted current flowing in both the primary and secondary
windings.

P,=Pinj2+P042, [watts] [7-18]

P,=2Pinj2, [watts] [7_

19]

Transformers with Multiple Outputs

This example shows how the apparent power, Pt, changes with a multiple output transformers.

Output Circuit
5 V @ 10A center-tapped Vj = diode drop = 1 V
15 V@ 1A full-wave bridge V^ = diode drop = 2 V

Efficiency = 0.95

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The output power seen by the transformer in Figure 7-4 is:

P 01 =(5 + 1)(10), [watts]

P0} = 60, [watts]

And:

Po2=(r02+r<)(la,)> [watts]
P 0 2 =(15 + 2)(1.0), [watts]
P02=17, [watts] [7.21-j

PI
CR1
-*- Io

t |
CR2 V
y o ^
Pi
1

o $ Rr

Figure 7-4. Multiple Output Converter.

Because of the different winding configurations, the apparent power, Pt, the transformer outputs will have
to be summed to reflect this. When a winding has a center-tap and produces a discontinuous current, then,
the power in that winding, be it primary or secondary, has to be multiplied by the factor, U. The factor, U,
corrects for the rms current in that winding. If the winding has a center-tap, then the factor, U, is equal to
1.41. If not, the factor, U, is equal to 1.

For an example, summing up the output power of a multiple output transformer, would be:

[7-22]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Then:

P 2 =60(1.41) + 17(1), [watts]

/ > = 101.6, [watts] [1-2?,]

After the secondary has been totaled, then the primary power can be calculated.

P +P
p ^-£l

(60)
P, = / , L[wattsJ
(0.95)
Pm=8l, [watts] [7-24]

Then, the apparent power, Pt, equals:

P,=Pin(u) + P,, [watts]

/?=(81)(1.41) + (101.6), [watts]
/ > = 215.8, [watts] [?.25]

Regulation
The minimum size of a transformer is usually determined either by a temperature rise limit, or by allowable
voltage regulation, assuming that size and weight are to be minimized. Figure 7-5 shows a circuit diagram
of a transformer with one secondary.
Note that a = regulation (%).

Primary Secondary
n = NJ1SL
s P
=1

Figure 7-5. Transformer Circuit Diagram.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The assumption is that distributed capacitance in the secondary can be neglected because the frequency and
secondary voltage are not excessively high. Also, the winding geometry is designed to limit the leakage
inductance to a level, low enough, to be neglected under most operating conditions.

Transformer voltage regulation can now be expressed as:

V (N.L.)-K(F.L.), ,
a = - ^- \-V-
I/ frj T \ V
-(100),
/'
t%] L J
V FL
° ( ' -> [7-26]

in which, V0(N.L.), is the no load voltage and, V0(F.L.), is the full load voltage. For the sake of simplicity,
assume the transformer in Figure 7-5, is an isolation transformer, with a 1:1 turns ratio, and the core
impedance, Re, is infinite.

If the transformer has a 1 : 1 turns ratio, and the core impedance is infinite, then:

7
/n=/0> [amps]

With equal window areas allocated for the primary and secondary windings, and using the same current
density, J:

= / „ * „ = A F, =/„/?„ [volts] [7 _ 2g]

Regulation is then:

AK \V
a=—^(100) + —^(100), [%]
s
p [7-29]

Multiply the equation by currents, I:

'p'b ','0 [7_30]

Primary copper loss is:

Secondary copper loss is:

^ = A K s / o , [watts] [7.32]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Total copper loss is:

P C U = P P + P S , [watts] [?_33]

Then, the regulation equation can be rewritten to:

a =^-(100), [%]
P
° [7-34]

Regulation can be expressed as the power lost in the copper. A transformer, with an output power of 100
watts and a regulation of 2%, will have a 2 watt loss in the copper:

Pc cu=^-, [watts]
" 100 [7-35]

(100)(2)L
Pru=- '-^- , [watts]
100 [7-36]

p
cu =2' twatts] [7-37]

Relationship, Kg, to Power Transformer Regulation Capability

Although most transformers are designed for a given temperature rise, they can also be designed for a given
regulation. The regulation and power-handling ability of a core is related to two constants:

P.
a =•
2K K
* * [7-38]

a = Regulation (%) [7-39]

The constant, Kg, is determined by the core geometry, which may be related by the following equations:

WAlK,
K = " c » , [cm5]
MLT [7-40]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The constant, Ke, is determined by the magnetic and electric operating conditions, which may be related by
the following equation:

Where:
Kf = waveform coefficient
4.0 square wave
4.44 sine wave

From the above, it can be seen that factors such as flux density, frequency of operation, and the waveform
coefficient have an influence on the transformer size.

Relationship, Ap, to Transformer Power Handling Capability

Transformers
According to the newly developed approach, the power handling capability of a core is related to its area
product, Ap, by an equation which may be stated as:

4
P(l0
l
)
' ' , [cm4]

Where:

Kj~ = waveform coefficient

4.0 square wave
4.44 sine wave

From the above, it can be seen that factors such as flux density, frequency of operation, and the window
utilization factor, Ku, define the maximum space which may be occupied by the copper in the window.

Different Cores Same Area Product

The area product, Ap, of a core is the product of the available window area, Wa, of the core in square
centimeters, (cm2), multiplied by the effective, cross-sectional area, Ac, in square centimeters, (cm2), which
may be stated as:

Ar=WaAc, [cm4] [y _ 43]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Figures 7-6 through Figure 7-9 show, in outline form, three transformer core types that are typical of those
shown in the catalogs of suppliers.

A
c

Figure 7-6. Dimensional Outline of a C Core.

.
Wa i
i
.«—
E
i

G D

Figure 7-7. Dimensional Outline of a El Lamination.

OD

Figure 7-8. Dimensional Outline of a Toroidal Core.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 A w
c a
\
\ j L
s
JL

B A

ip
\

c
PQ Ferrite Core

Figure 7-9. Dimensional Outline of a PQ Ferrite Core.

250 Watt Isolation Transformer Design, Using the Core Geometry, Kg, Approach

The following information is the Design specification for a 250 watt isolation transformer, operating at 47
Hz, using the, Kg, core geometry approach. For a typical design example, assume with the following
specification:

1. Input voltage, Vin = 115 volts

2. Output voltage, V0 = 115 volts
3. Output current, I0 = 2.17 amps
4. Output power, P0 = 250 watts
5. Frequency, f = 47Hz
6. Efficiency, TI = 95%
7. Regulation, a = 5%
8. Operating flux density, Bac = 1.6 tesla
9. Core Material = Silicon M6X
10. Window utilization, Ku = 0.4
11. Temperature rise goal, Tr = 30°C

Step No. 1 Calculate the transformer apparent power, P(

— + 1 I, [watts]

P=250\ [watts]
{0.95
P,=513, [watts]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 2 Calculate the electrical conditions, Ke

Kf = 4.44, [sine wave]

Ke = 0.145(4.44)2 (47 )2 (1.6)2 (lO 4 )
* =1.62

Step No. 3 Calculate the core geometry, Kg.

[cm ]
*•-•£? '
^=31.7, [cm5]

Step No. 4 Select a lamination from Chapter 3, comparable in core geometry, Kg.

Lamination number El-150

Manufacturer Thomas and Skinner
Magnetic path length, MPL 22.9 cm
Core weight, Wtfe 2.334 kilograms
Copper weight, Wtcu 853 grams
Mean length turn, MLT 22 cm
Iron area, Ac 13.8 cm2
Window area, Wa 10.89 cm2
Area product, Ap 150 cm4
Core geometry, Kg 37.6 cm5
Surface area, A, 479 cm2

Step No. 5 Calculate the number of primary turns, Np using Faraday's Law.

N =—^ —, [turns]
' KfBaJAc

Np
(4.44)(1.6)(47)(13.8)'
Nf = 250, [turns]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 6 Calculate the current density, J.

J =
^(io 4 )—, [amps/cm' ]
KfKuBaJAp
513(l0 4 )
J = ~, r; rr^—rr—^ 7, [amps/cm 1
(4.44)(0.4)(1.6)(47)(150)
J = 256, [amps/cm2]

Step No. 7 Calculate the input current, !„,.

7 , , = - , [amps]
V
in>l
25
° r i
[amps]
'
4=2.28, [amps]

Step No. 8 Calculate the primary bare wire area, Awp(B).

_4 ,
J

= (2.28) ^
»/>(*) ~ 256 ' Lcm
4^=0.0089, [cm2]

Step No. 9 Select the wire from the Wire Table, in Chapter 4.

4^=0.00822, [cm2]
4^=0.00933, [cm2]

= 209, [micro-ohm/cm]
cm

Step No. 10 Calculate the primary resistance, Rp.

I, [ohms]
cm,
tfp=(22)(250)(209)(lO-6), [ohms]
Rp =1.15, [ohms]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 1 1 Calculate the primary copper loss, Pp.

Pp =(2.28) 2 (1.15), [watts]

P ;) =5.98, [watts]

Step No. 12 Calculate the secondary turns, Ns.

. - [turns]
vin ioo '
(250)(115)( 5 ^
Ns=\ . A . ' 1 + - , [turns]
(115) t IOOJ
Ns = 262.5 use 263, [turns]

Step No. 13 Calculate the secondary bare wire area, Aws(B).

J'
.(2.17)
256
AwsB
s(B} = 0.00804, [cm2]

Step No. 14 Select the wire from the Wire Table, in Chapter 4.

A
«P(B) = 0.00822, [cm2]
4^=0.00933, [cm2]

= 209, [micro-ohm/cm]
cm

Step No. 15 Calculate the secondary winding resistance, Rs.

6
Rs =MLT(^)| — |(lO" ), [ohms]
^ cm J v '
/? s =(22)(263)(209)(lO- 6 ), [ohms]
/?,=!. 21, [ohms]

Step No. 16 Calculate the secondary copper loss, Ps.

P,=llR,, [watts]
P S =(2.17) 2 (1.21), [watts]
^=5.70, [watts]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 17 Calculate the total primary and secondary copper loss, Pcu.

pcu = Pt> + ps, [watts]

Pcu=5.98 + 5.1, [watts]
/>,„=! 1.68, [watts]

Step No. 18 Calculate the transformer regulation, a.

a =^(100), [%]
o

(11.68)
<*=-, r
(250)
a = 4.67, [%]

Step No. 19 Calculate the watts per kilogram, W/K. Use the equation for this material in Chapter 2.

1.86
WIK = 0.000557 (f) (Bac)
1.51 /. ,\1 86
WIK = 0.000557 (47)' 5 ' (1.6)
= 0.860

Step No. 20 Calculate the core loss, Pfe.

Pfe=(W/K)(Wtfe)(lO-}), [watts]
P /e =(0.860)(2.33), [watts]
Pfe = 2.00, [watts]

Step No. 21 Calculate the total loss, Ps.

PZ =(11.68) + (2.00), [watts]

Pz= 13.68, [watts]

Step No. 22 Calculate the watts per unit area, y.

p
y = —, [watts/cm2]
4
w = ^ ' / , [watts/cm2 ]
(479)
^ = 0.0286, [watts/cm2]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 23 Calculate the temperature rise, Tr.

7>450(0.0286) (0826) ,
Tr = 23.9, [°C]

Step No. 24 Calculate the total window utilization, Ku.

K =-
A,,,

(263)(0.00822)
K = ^-- ---'- = 0. 199
(10.89)
N

-
=

(250)(0.00822)
K
un=~-7^-x- = ^0.189
"" (10.89)
Ku =(0.189) + (0.199)
K,, = 0.388

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 38 Watt 100kHz Transformer Design, Using the Core Geometry, Kg, Approach

The following information is the design specification for a 38 watt push-pull transformer, operating at
100kHz, using the Kg core geometry approach. For a typical design example, assume a push-pull, full-
wave, center-tapped circuit, as shown in Figure 7-4, with the following specification:

1. Input voltage, V(min) = 24 volts

2. Output voltage #1, V(oi, = 5.0 volts
3. Output current #1,1(0|) = 4.0 amps
4. Output voltage #2, V(o2) = 12.0 volts
5. Output current #2,1(o2) = 1.0 amps
6. Frequency, f = 100kHz
7. Efficiency, r| = 98%
8. Regulation, ot = 0.5%
9. Diode voltage drop, Vd = 1.0 volt
10. Operating flux density, Bac = 0.05 tesla
11. Core Material = ferrite
12. Window utilization, Ku = 0.4
13. Temperature rise goal, Tr = 30°C
14. Notes:

Using a center-tapped winding, U = 1.41

Using a single winding, U = 1.0

At this point, select a wire so that the relationship between the ac resistance and the dc resistance is 1:

The skin depth, E, in centimeters, is:

6.62
TT

6.62
E= i =, [cm ]
y 100, ooo
s = 0. 0209 , [cm]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Then, the wire diameter, DA\VG> is:

0^=2(0.0209), [cm]
DAlva= 0.0418, [cm]

Then, the bare wire area, Aw, is:

(3.1416)(0.0418)2
Aw=±-^- , '[cm2]

4 =0.00137, [cm2]

From the Wire Table 4-9 in Chapter 4, number 27 has a bare wire area of 0.001021 centimeters. This will
be the minimum wire size used in this design. If the design requires more wire area to meet the
specification, then the design will use a multifilar of #26. Listed Below are #27 and #28, just in case #26
requires too much rounding off.

Wire AWG Bare Area Area Ins. Bare/Ins. ufi/cm

#26 0.001280 0.001603 0.798 1345
#27 0.001021 0.001313 0.778 1687
#28 0.0008046 0.0010515 0.765 2142

Step No. 1 Calculate the transformer output power, P

P0 = Pol+Po2, [watts ]

P,i=I0^oi + Vd} [watts]

/ > „ , = 4(5+1) [watts ]

Pol =24, [watts ]

Po2 =I02^o2+Vd}, [watts ]

P02 = 1(12 + 2) [watts ]

Po2=14, [watts]

P 0 =(24 + 14) [watts ]

P0 =38, [watts ]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 2 Calculate the total secondary apparent power, Pts.

P,s=PKoi+P,so2, [watts]

^,=24(1.41), [watts]
/L, =33.8, [watts]
P«o*=P,*(U), [watts]
^ 0 2=14(1), [watts]
^2 =14, [watts]
^=(33.8 + 14), [watts]
pa = 47.8, [watts]

Step No. 3 Calculate the total apparent power, Pt.

Ptp = PinPa, [watts ]

Pt =P,P+Pts, [watts ]

.8, [watts ]

Pt =102.5, [watts ]

Step No. 4 Calculate the electrical conditions, Ke

K , = 4. 0, [square wave ]

Ke =0.1 45 (4.0)? (100000 )?(0.05) 2 (lO" 4 )

Ke =5800

Step No. 5 Calculate the core geometry, Kg.

K g =——, [cm 5 ]
2Kea

(102.5)
g L
2(5800)0.5'

K =0.0177, [cm 5 ]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 When operating at high frequencies, the engineer has to review the window utilization factor, K u , in
Chapter 4. When using a small bobbin ferrites, use the ratio of the bobbin winding area to the core window
area is only about 0.6. Operating at 100kHz and having to use a #26 wire, because of the skin effect, the
ratio of the bare copper area to the total area is 0.78. Therefore, the overall window utilization, Ku, is
reduced. To return the design back to the norm, the core geometry, Kg, is to be multiplied by 1.35, and
then, the current density, J, is calculated, using a window utilization factor of 0.29.

Kg = 0.0177 (1.35) [cm 5 ]

K g = 0.0239, [cm5]

Step No. 6 Select a PQ core from Chapter 3, comparable in core geometry Kg.

Core number PQ-2020

Manufacturer TDK
Magnetic material PC44
Magnetic path length , MPL 4.5 cm
Window height, G 1.43 cm
Core weight, W(fe 15 grams
Copper weight, W(cu 10.4 grams
Mean length turn, MLT 4.4 cm
Iron area, AC 0.62 cm2
Window area, Wa 0.658cm2
Area product, A 0.408 cm4
Core geometry, K 0.0227 cm5
Surface area, A( 19.7 cm2

Millihenrys per 1000 turns, AL 3020

Step No. 7 Calculate the number of primary turns, Np, using Faraday's Law.

Kf acJ
K BB AfAc '

Np = -
(4.0 XO. 05 )(100000 X°-62 )'

N = 19, [turns ]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 8 Calculate the current density, J, using a window utilization, Ku = 0.29.

J [amps / c m ]
= KfKuif BacJ
n f A
A
p
'

102.5(l0 4 )
(4. 0X0.29X0. 05X100000 )(0.408)'

J =433, [amps / c m 2 ]

Step No. 9 Calculate the input current, !;„.

38
<-« (24X0.98)'

/,.„ = 1.61, [amps ]

Step No. 10 Calculate the primary bare wire area, Awp(B).

(1.61X0.707)

'-, [cm 2 ]

Step No. 11 Calculate the required number of primary strands, Snp.

/I f D\
C _ r V^)
np
~ #26
0.00263
p
" ~ 0.00128

Step No. 12 Calculate the primary new ^0 per centimeter.

//Q/cm
(new)//L2 / cm =

1345
(new) wQ / cm =
2
(new)//Q / cm = 673

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 13 Calculate the primary resistance, Rp.

6
(icr
v
) [ohms]
'

Rp =(4. 4X19X673 ^lO'6) [ohms]

Rp= 0.0563, [ohms]

Step No. 14 Calculate the primary copper loss, Pp.

Pp =I2pRp, [watts ]

Pp =(1.61^(0.0563) [watts ]

Pp =0.146, [watts ]

Step No. 15 Calculate the secondary turns, N sl .

sl 1
v.m 'V • " ',
100 [turns ]

V s] = V0+ Vd, [volts]

Vsl=5+l, [volts]
Vsl=6, [volts]

""»'
Nsl =4.77 use 5, [turns ]

Step No. 16 Calculate the secondary turns, Ns2.

N „¥,

Vs2=12+2, [volts]

V s2 = 14, [volts ]

0.5 r
1+ [turns]
Tbo-
N J2 =11.1 use 1 1 , [turns ]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 17 Calculate the secondary bare wire area, A wsl .

A of max
, r
[cm2 ->]

(4XQ. 707) 2

/) v r a l =0.00653 , [cm 2 ]

Step No. 18 Calculate the required number of secondary strands, Sns|.

0.00653
Jl
" ~ 0.00128
•5,,, = 5 . 1 use 5

Step No. 19 Calculate the secondary, S[ new uQ per centimeter.

ns 1

1345
(new )/jQ.I cm =

(new ),uQ / cm = 269

Step No. 20 Calculate the secondary S] resistance, R s) .

6
), [ohms]

Rsl =(4.4)(5)(269)(l0^ 6 ), [ohms]

Rs} = 0.0059, [ohms]

Step No. 21 Calculate the secondary copper loss, Psl.

P,\=ti\R,i, [watts]

psl = (4. 0)2 (0.0059) [watts ]

ps , = 0. 0944 , [watts ]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 22 Calculate the secondary bare wire area, Aws2.

AWS2=-^, [cm 2 ]

, [cm 2 ]

Aws2= 0.00231, [cm 2 ]

Step No. 23 Calculate the required number of secondary strands, SnS2.

"s2 #26
0.00231
0.00128
5M2=1.8use 2

Step No. 24 Calculate the secondary, S2 new ufi per centimeter.

, , „ , uQ. I cm
(new )//Q / cm = —-
ns

1345
(new )/uQ. I cm = —-

(new )fJ.Q. I cm = 673

Step No. 25 Calculate the secondary, S2 resistance, Rs2.

RS2 = MLT (N, ,) (10 -6 ) [ohms ]

V cm J x '

RS2 = (4.4X1 0(673 )(lO-6) [ohms]

/?,2 = 0. 0326 , [ohms ]

Step No. 26 Calculate the secondary, S2 copper loss, Ps2.

p
S2 = fs2^s2, [watts ]

P J2 =(1.0)? (0.0326), [watts]

Ps2 =0.0326, [watts ]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 27 Calculate the total secondary copper loss, Ps.

Ps = Ps] +Ps2, [watts ]

Ps =0.0944 +0.0326, [watts ]

Ps =0.127, [watts ]

Step No. 28 Calculate the total primary and secondary copper loss, Pcu.

Pcu =Pp+Ps, [watts ]

Pcu =0.146 +0.127, [watts]

Pcu =0.273, [watts ]

Step No. 29 Calculate the transformer regulation, ex.

a =-£HL(IOO) [%]
"o

[%]
a =0.718, [%]

Step No. 30 Calculate the milliwatts per gram, mW/g. Use the equation for this material in Chapter 2.

mW / g = '

mW I g = 0.000318 (lOOOOO )'' 51 (0.05 f47

mW / g = 3.01

Step No. 31 Calculate the core loss, Pfe.

Pje =(mW I g^Wtfe )(l(r3) [watts ]

P / e =(3.0lXl5)(lO- 3 ) [watts ]

Pfe =0.045, [watts ]

Step No. 32 Calculate the total loss, Pz.

Pi. = PCU + Pfe , [wattS ]

P£ = (0.273 )+ (0.045 ) [watts ]

Pz =0.318, [watts ]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 33 Calculate the watts per unit area, \\i.

p
if/=—— , [watts / c m 2 ]

(0.318 n
Y , , [watts
L /cm 1
(19.7)

y = 0.0161, [watts / c m 2 ]

Step No. 34 Calculate the temperature rise, Tr.

7>450(</) (o - 826) , [°C]

r r =450(0.016l) ( ° 8 2 6 ) , [°C]

TV = 14.9, [°C]

Step No. 35 Calculate the total window utilization, Ku.

(10)(5)(0.00128)
T = v A / v---i = 0.0973
"' (0.658)
(11)(2)(0.00128)
al = ^ V —v-- = 0.0428
(0.658)
_

(38X2X000,28)
* (0.658)
A:U =(0.148) + (0.0973 + 0.0428)
AT.. = 0.288

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Chapter 8

DC Inductor Design Using Gapped Cores

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table of Contents

1. Introduction
2. Critical Inductance for Sine Wave Rectification
3. Critical Inductance for Buck Type Converters
4. Core Materials, Used in PWM Converters
5. Fundamental Considerations
6. Fringing Flux
7. Inductors
8. Relationship of, Ap, to Inductor's Energy-Handling Capability
9. Relationship of, Kg, to Inductor's Energy-Handling Capability
10. Gapped Inductor Design Example Using the Core Geometry, Kg, Approach
11. Gapped Inductor Design Example Using the Area Product, Ap, Approach

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Introduction
Designers have used various approaches in arriving at suitable inductor designs. For example, in many
cases, a rule of thumb used for dealing with current density is that a good working level is 200 amps-per-
cnr (1000 Cir-Mils-per-amp). This rule is satisfactory in many instances; however, the wire size used to
meet this requirement may produce a heavier and bulkier inductor than desired or required. The
information presented herein will make it possible to avoid the use of this and other rules of thumb and to
develop an economical and a better design.

Critical Inductance for Sine Wave Rectification

The LC filter is the basic method of reducing ripple levels. The two basic rectifier circuits are the full-
wave center-tap as shown in Figure 8-1 and the full-wave bridge, as shown in Figure 8-2. To achieve
normal inductor operation, it is necessary that there be a continuous flow of current through the input
inductor, LI.

Full Wave Center Tap

Figure 8-1. Full-Wave Center Tap with an LC filter.

Tl

Full Wave Bridge

Figure 8-2. Full-Wave Bridge with an LC filter.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The value for minimum inductance called critical inductance, L(crt) is:

= [henrys]
'
Where:
co = IK f
f = line frequency

The higher the load resistance, R,,, (i.e., the lower the dc load current), the more difficult it is to maintain a
continuous flow of current. The filter inductor operates in the following manner: When R<, approaches
infinity, under an unloaded condition, (no bleeder resistor), I0 = 0, the filter capacitor will charge to Vl pk ,
the peak voltage. Therefore, the output voltage will be equal to the peak value of the input voltage, as
shown in Figure 8-3.

>
iT L/crt)= Critical Inductance

o
V

-I
O

Load Current, IQ

Figure 8-3. Critical Inductance Point.

The ripple reduction from a single stage LC filter can be calculated, using Equation 8-2 and Figure 8-4.

V
r
r(pk)
-V
r , [volts-peak] [8-2]

CR1 'o ^/Vr(pk)

Figure 8-4. LC Filter Ripple Reduction.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Critical Inductance for Buck Type Converters
The buck type converter schematic is shown in Figure 8-5, and the buck type dc-to-dc converter is shown
in Figure 8.6. The buck regulator filter circuit shown in Figure 8-5 has three current probes. These current
probes monitor the three basic currents in a switch mode, buck output filter. Current probe A monitors the
power MOSFET, Ql, switching current. Current probe B monitors the commutating current through CR1.
Current probe C monitors the current through the output inductor, LI.

The typical filter waveforms of the buck converter are shown in Figure 8-7. The waveforms are shown
with the converter operating at a 0.5 duty ratio. The applied voltage, VI to the filter, is shown in Figure 8-
7A. The power MOSFET, Ql, current is shown in Figure 8-7B. The commutating current flowing through
CR1 is shown in Figure 8-7C. The commutating current is the result of Ql being turned off, and the field
in LI collapsing, producing the commutating current. The current flowing through LI is shown in Figure
8-7D. The current flowing through LI is the sum of the currents in Figure 8-7B and 8-7C.

VI LI
• \ i iirrenr Krone i
w w
T *
Ql 1-11
\J f

Cl
o CR1
C2 +
v
in ,- ->\ 2L f 0

O .1 *• ^/

Current Probe B

Figure 8-5. Buck Regulator Converter.

The critical inductance current is shown in Figure 8-8, 8-B and is realized in Equation 8-3. The critical
inductance current is when the ratio of the delta current to the output load current is equal to 2 = AI /10. If
the output load current is allowed to go beyond this point, the current will become discontinuous, as shown
in Figure 8-8, 8-D. The applied voltage, VI, will have ringing at the level of the output voltage, as shown
in Figure 8-8, 8-C. When the current in the output inductor becomes discontinuous, as shown in Figure 8-
8, 8-D, the response time for a step load becomes very poor.

When designing multiple output converters similar to Figure 8-6, the slaved outputs should never have the
current in the inductor go discontinuous or to zero. If the current goes to zero, a slaved output voltage will
rise to the value of VI. If the current is allowed to go to zero, then, there is no potential difference between
the input and output voltage of the filter. Then the output voltage will rise to equal the peak input voltage.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 VT(\-
, [henrys] [8-3]
21o(min)

V
[8-4]
V
(>> *(>

Figure 8-6. Push-Pull Buck Type Converter.

7-A

Current Probe C

Figure 8-7. Typical Buck Converter Waveforms, Operating at a 0.5 Duty Ratio.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Vj L T

Current Probe C
V 1 ton toff

I
r, !
r 1
V0

8-D 1 i
::
AI ^ —^-«=~" ^ ^^^^^ -- o
l
t
t Current Probe C

Figure 8-8. Buck Converter, Output Filter Inductor Goes from Critical to Discontinuous Operation.

Core Materials, Used in PWM Converters

Designers have routinely tended to specify Molypermalloy powder materials for filter inductors used in
high-frequency, power converters and pulse-width-modulators (PWM) switched regulators, because of the
availability of manufacturers' literature containing tables, graphs, and examples that simplify the design
task. Use of these cores may result in an inductor design not optimized for size and weight. For example,
as shown in Figure 8-9, Molypermalloy powder cores, operating with a dc bias of 0.3T, have only about
80% of the original inductance, with very rapid falloff at higher flux densities. When size is of greatest
concern then, magnetic materials with high flux saturation, Bs, would be first choice. Materials, such as
silicon or some amorphous materials, have approximately four times the useful flux density compared to
Molypermalloy powder cores. Iron alloys retain 90% of their original inductance at greater than 1.2T. Iron
alloys, when designed correctly and used in the right application, will perform well at frequencies up to
100kHz. When operating above 100kHz, then the only material is ferrite. Ferrite materials have a negative
temperature coefficient regarding flux density. The operating temperature and temperature rise should be
used to calculate the maximum flux density.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 To get optimum performance, together with size, the engineer must evaluate the materials for both, Bs, and
Bac. See Table 8-1. The operating dc flux has only to do with I2R losses, (copper). The ac flux, Bac, has to
do with core loss. This loss depends directly on the material. There are many factors that impact a design:
cost, size, temperature rise and material availability.

There are significant advantages to be gained by the use of iron alloys and ferrites in the design of power
inductors, despite certain disadvantages, such as the need for banding and gapping materials, banding tools,
mounting brackets, and winding mandrels.

Iron alloys and ferrites provide greater flexibility in the design of high frequency power inductors, because
the air gap can be adjusted to any desired length, and because the relative permeability is high, even at
high, dc flux density.

100

£ 80

70

60

0.001 0.01 0.1

Polarized Flux Density, tesla

Figure 8-9. Inductance Versus dc Bias.

Table 8-1. Magnetic Material Properties

Magnetic Material Properties

Material Initial Flux Density Curie Density
Name Composition Permeability Tesla Temp. grams/cm3

l^i Bs °C 5
Silicon 3-97 SiFe 1500 1.5-1.8 750 7.63
Orthonol 50-50 NiFe 2000 1.42-1.58 500 8.24
Permalloy 80-20 NiFe 25000 0.66-0.82 460 8.73
Amorphous 81-3.5 FeSi 1500 1.5-1.6 370 7.32
Amorphous 66-4 CoFe 800 0.57 250 7.59
Amorphous(u) 73-15FeSi 30000 1.0-1.2 460 7.73
Ferrite MnZn 2500 0.5 >230 4.8

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Fundamental Considerations

The design of a linear reactor depends upon four related factors:

1. Desired inductance, L
2. Direct current, Idc
3. Alternating current, AI
4. Power loss and temperature, Tr

With these requirements established, the designer must determine the maximum values for, Bdc, and, Bac,
that will not produce magnetic saturation. The designer must make trade-offs that will yield the highest
inductance for a given volume. It should be remembered the peak operating flux, Bpk, depends upon, Bdc +
Bac, in the manner in Figure 8-10.

=Bdc+-, [tesla] [8-5]

[teslal
',+

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 B (tesla)
All

H (oersteds)

Figure 8-10. Inductor Flux Density Versus 1^ + AI Current.

The inductance of an iron-core inductor carrying direct current and having an air gap may be expressed as:

[henrys] [8-9]
MPL

This equation shows that inductance is dependent on the effective length of the magnetic path, which is the
sum of the air gap length, lg, and the ratio of the core mean length to the material permeability, MPL/u m .
When the core air gap, lg, is large compared to the ratio, MPL/um, because of material permeability, um,
variations in um do not substantially affect the total effective magnetic path length or the inductance. Then
the inductance Equation [8-9] reduces to:

[henrys] [8-10]

Final determination of the air gap size requires consideration of the effect of fringing flux, which is a
function of gap dimension, the shape of the pole faces, and the shape, size, and location of the winding. Its
net effect is to shorten the air gap. Because of the fringing flux it is wise to lower the initial operating flux
density, 10 to 20%.

Fringing Flux

Fringing flux decreases the total reluctance of the magnetic path and therefore, increases the inductance by
a factor, F, to a value greater than that calculated from Equation 8-10. Fringing flux is a larger percentage
of the total for the larger gaps.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The fringing factor is:

2G
1
V
rn
J
[8-11]

Where G is the winding length, defined in Chapter 3. This equation is valid for laminations, C cores and
cut ferrites. Equation [8-11] is plotted in Figure 8-11.

3 -
X

' M

;? 2 -

Figure 8-11. Increase of Inductance with Fringing Flux at the Gap.

As the air gap increases, the flux across the gap fringes more and more. Some of the fringing flux strikes
the core, perpendicular to the strip or tape, and sets up eddy currents, which cause additional losses in the
core. If the gap dimension gets too large, the fringing flux will strike the copper winding and produce eddy
currents, generating heat, just like an induction heater. The fringing flux will jump the gap and produce
eddy currents, in both the core and winding, as shown in Figure 8-12.

The inductance, L computed in Equation [8-10], does not include the effect of the fringing flux. The value
of inductance, L' corrected for fringing flux is:

-, [henrys] [8-12]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Eddy currents

Core

Winding
©©©©©
©0©©©
Fringing Flux U/2

Magnetic Path

Figure 8-12. Fringing Flux Around the Gap of an Inductor.

The effect permeability may be calculated from the following equation:

[8-13]

Where, um, is the material permeability.

Inductors
Inductors that carry direct current are used frequently in a wide variety of ground, air, and space
applications. Selection of the best magnetic core for an inductor frequently involves a trial-and-error type
of calculation.

The author has developed a simplified method of designing optimum, dc carrying inductors with gapped
cores. This method allows the engineer to select the proper core that will provide correct copper loss, and
make allowances for fringing flux, without relying on trial-and-error and the use of the cumbersome
Hanna's curves.

Rather than discuss the various methods used by transformer designers, the author believes it is more useful
to consider typical design problems, and to work out solutions using the approach based upon newly
formulated relationships. Two gapped core designs will be compared. To compare their merits, the first
design example will use the core geometry, Kg, and the second design will use the area product, Ap.

Inductors, designed in this handbook, are banded together with phosphor bronze banding material, or held
together with aluminum brackets. The use of steel banding material; or brackets that bridge the gap are not
recommended, because the use of steel across the gap is called shorting the gap. When the gap is shorted,
the inductance will increase from the calculated value.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Relationship of, Ap, to Inductor's Energy-Handling Capability
The energy-handling capability of a core is related to its area product, Ap, by the equation:

2(Energy)(l0 4 )
A =— '-, [cm4] [8-14]
BmJKu

Where: Energy is in watt-seconds.

Bm is the flux density, tesla.
J is the current density, amps-per-cm2.
Ku is the window utilization factor. (See Chapter 4)

From the above, it can be seen that factors such as flux density, Bm, window utilization factor, Ku, (which
defines the maximum space that may be used by the copper in the window), and the current density, J,
which controls the copper loss, all impact the area product, Ap. The energy-handling capability of a core is
derived from:

LI2
Energy = , [watt-seconds] [8-15]

Relationship of, Kg, to Inductor's Energy-Handling Capability

Inductors, like transformers, are designed for a given temperature rise. They can also be designed for a
given regulation. The regulation and energy handling ability of a core is related to two constants:

a=± ^-, [%] [8-16]

Where, a, is the regulation, %:

The constant, Kg, is determined by the core geometry:

W A1 K
K s = " c ", [cm5] [8-17]
MLT

The constant, Ke, is determined by the magnetic and electrical operating conditions:

Ke=Q.145P0B2pk(lQ-4} [8-18]

The peak operating flux density, Bpj(, is:

Bpt=Btlc+^, [tesla] [8-19]

From the above, it can be seen that the flux density, Bpk, is the predominant factor governing size.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The output power, P0, is defined in Figure 8-13.

[8-20]

ijn I/QJ\ ^ v1 ^^^^ 1(02) ^1

j 111 , <j -<

Vjn i
CR1 y
(».)
Cl
/-- ->
!> a^
CR2
v
(02)
C2 +
(^ ^ 0

i i
1^1 n

Figure 8-13. Defining the Inductor Output Power.

Gapped Inductor Design Example Using the Core Geometry, Kg, Approach

Step No. 1 Design a linear dc inductor with the following specifications.

1. Inductance, L = 0.0025 henrys

2. dc current, I0 = 1.5 amps
3. ac current, AI = 0.2 amps
4. Output power, P0 = 100 watts
5. Regulation, a = 1.0 %
6. Ripple Frequency = 200 kHz
7. Operating flux density, Bm = 0.22 tesla
8. Core Material = ferrite
9. Window utilization, Ku = 0.4
10. Temperature rise goal, Tr = 25°C

Step No. 2 Calculate the peak current, Ipk.

-y, [amps]

2 > [amPs]
/
M = 1 - 6 > [amps]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 3 Calculate the energy-handling capability.

Ll\
Energy = —— , [watt-seconds]

(0.0025)(1.6)2
Energy = --——— , [watt-seconds]

Energy = 0.0032, [watt-seconds]

Step No. 4 Calculate the electrical conditions coefficient, Ke.

Ke = 0.145(lOO)(0.22) 2 (lO~ 4 )
K = 0.0000702

Step No. 5 Calculate the core geometry coefficient, Kg.

K,a
(0.0032)2 5
s
~(0.0000702)(1.0)'
A T g = 0.146, [cm5]

Step No. 6 Select an ETD ferrite core from Chapter 3. The data listed is the closest core to the calculated
core geometry, Kg.

1. Core Number = ETD-39

2. Magnetic Path Length, MPL = 9.22 cm
3. Core Weight, Wtfi = 60 grams
4. Mean Length Turn, MLT = 8.3 cm
5. Iron Area, Ac = 1.252 cm2
6. Window Area, Wa = 2.34 cm"
7. Area Product, Ap = 2.93 cm4
8. Core Geometry, Kg = 0.177 cm5
9. Surface Area, A, = 69.9 cm2
10. Material, P = 2500^1
11. Millihenrys-per-lk, AL = 3295 mh
12. Winding Length, G = 2.84 cm

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 7 Calculate the current density, J, using the area product equation, Ap.

J =
-„ , „- > [amps-per-cm ]
B
mApKu

2(0.0032)(l0 4 )
7 = -,-TT-r,—r > [amps-per-cm2 ]
(0.22)(2.93)(0.4)
J = 248, [amps-per-cm2 ]

Step No. 8 Calculate the rms current, Irms.

^ s = V ( l - 5 ) 2 + ( 0 . 2 ) 2 , [amps]
/„„,=!.51, [amps]

Step No. 9 Calculate the required bare wire area, Aw(B).

^ _ 'rim fcni 2 1
J

> [cm
(248)
AW(B} = 0.00609, [cm2]

Step No. 10 Select a wire from the Wire Table in Chapter 4. If the area is not within 10%, take the next
smallest size. Also, record the micro-ohms per centimeter.

AWG = #19
Bare, AW(B) = 0.00653, [cm2 ]
Insulated, Aw = 0.00754, [cm2]

—— = 264, [micro-ohm/cm]

Step No. 11 Calculate the effective window area, Wa(eff,. Using the window area found in Step 6. A typical
value for, S3, is 0.75, as shown in Chapter 4.

W.W=WaS>, [cm2]

«W)=( 2 - 34 )(°- 75 )>

^,=1.76, [cm2]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 12 Calculate the number turns possible, N, using the insulated wire area, Aw, found in Step 10. A
typical value for, 82, is 0.6, as shown in Chapter 4.

w,(eff}lf,s,2
N= " , [turns]
Aw
(1.76)(0.60)
±-.—^-- [turns]
(0.00754)
TV = 140, [turns]

Step No. 13 Calculate the required gap, lg.

0.4;r^ 2 4(lO- 8 )
/.= -Z -- — , [cm]

(1.26)(14Q)2(1.25)(1Q-8) (9.22}
I g = -;-7—---- - . [cm]
(0.0025) {2500 )
/ „ = < ) . 120, [cm]

Step No. 14 Calculate the equivalent gap in mils.

mils = cm(393.7)
mils = (0.120)(393.7)
mils = 47. 2 use 50

Step No. 15 Calculate the fringing flux factor, F.

VI .25 ^ 0.120 J
F=1.41

Step No. 16 Calculate the new number of turns, Nn, by inserting the fringing flux, ¥.

/g L
Nn = — -, [turns]
" ^ 0.4^4 F(lO- 8 )

AT, =116, [turns]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 17 Calculate the winding resistance, RL. Use the MLT from Step 6 and the micro-ohm per
centimeter from Step 10.

6
) l ( r , [ohms]

RL =(8.3)(ll6)(264)(l(T 6 ), [ohms]
RL = 0.254, [ohms]

Step No. 1 8 Calculate the copper loss, Pcu.

P f u =(l.5l) 2 (0.254), [watts]

Pca = 0.579, [watts]

Step No. 19 Calculate the regulation, a.

a = ^-(100), [%]
o

(0.579)
a=~, r
(100)
a = 0.579, [%

Step No. 20 Calculate the ac flux density, Bac.

[tesla]
(0.120)
v ;
'
1, 2500
Bac= 0.0167, [tesla]

Step No. 21 Calculate the watts per kilogram for ferrite, P, material in Chapter 2. Watts per kilogram can
be written in milliwatts per gram.

mW/g = (0.00004855) (200000)° 63) (0.0167 ) (262)

mW/g = 0.468

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 22 Calculate the core loss, Pfe.

Pfe = (mW/g)(Wtfe )(l<r 3 ), [watts]

P /e =(0.468)(60)(l(r 3 ), [watts]
/> /e =0.028l, [watts]

Step No. 23 Calculate the total loss, copper plus iron, Ps.

Pj. =(0.0281) + (0.579), [watts]

PL = 0.607, [watts]

Step No. 24 Calculate the watt density, vy. The surface area, A,, can be found in Step 6.

p
i// = —, [watts/cm2 ]
4
(0.607)
y/ = -.-^ , [watts/cm2 ]
(69.9)
^ = 0.00868, [watts/cm2]

Step No. 25 Calculate the temperature rise, Tr.

Tr =450(0.00868) ( ° 826) ,
Tr=8.92, [°C]

Step No. 26 Calculate the peak flux density, Bpi<.

p
/ q ^-, \ ' L j
(0.127)+-^=-
v ;
(2500)
Bit= 0.252, [tesla]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Note:
The big advantage in using the core geometry design procedure is that the wire current density is
calculated. When using the area product design procedure, the current density is an estimate, at best. In
this next design the same current density will be used as in the core geometry design.

Gapped Inductor Design Example Using the Area Product, Ap, Approach

Step No. 1 Design a linear dc inductor with the following specifications:

1. Inductance, L = 0.0025 henrys

2. dc current, I0 = 1.5 amps
3. ac current, AI = 0.2 amps
4. Output power, P0 = 100 watts
5. Current Density, J = 250 amps-per-cm2
6. Ripple Frequency = 200 kHz
7. Operating flux density, Bm = 0.22 tesla
8. Core Material = ferrite
9. Window utilization, Ku = 0.4
10. Temperature rise goal, Tr = 25°C

Step No. 2 Calculate the peak current, Ip^.

I I
Pk= o+ — , [amPs]

7 M =(1.5) + ^, [amps]
7^=1.6, [amps]

Step No. 3 Calculate the energy-handling capability.

LI\
Energy = ——, [watt-seconds]

(0.0025)(1.6)2
Energy = - ———, [watt-seconds]

Energy = 0.0032, [watt-seconds]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 4 Calculate the area product, Ap.

^2( E ne,gy)(loQ ^

2(0.0032)(l0 4 )
^ = (0.22)(248)(0.4)' ^ ]
4, =2.93, [cm4]

Step No. 5 Select an ETD ferrite core from Chapter 3. The data listed is the closest core to the calculated
area product, Ap.

1. Core Number = ETD-39

2. Magnetic Path Length, MPL = 9.22 cm
3. Core Weight, Wtft = 60 grams
4. Mean Length Turn, MLT = 8.3 cm
5. Iron Area, Ac , = 1.252 cm2
6. Window Area, Wa = 2.34 cm2
7. Area Product, Ap = 2.93 cm4
8. Core Geometry, Kg = 0.177 cm5
9. Surface Area, A, = 69.9 cm2
10. Material, P = 2500^
11. Millihenrys-per-lk, AL = 3295 mh
12. Winding Length, G = 2.84 cm

Step No. 6 Calculate the rms current, Irms.

/ ™ = V O - 5 ) 2 + ( ° - 2 ) 2 > [amps]
/
™ s = 1 - 5 1 ' [amps]

Step No. 7 Calculate the required bare wire area, AW(B).

=
™W(S) ~j~>

, [cm2]

, (fl) = 0.00609, [cm2]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 8 Select a wire from the Wire Table in Chapter 4. If the area is not within 10%, take the next
smallest size. Also, record micro-ohms per centimeter.

AWG = #19
Bare, Aw(tt} = 0.00653, [cm2 ]
Insulated, Aw = 0.00754, [cm2 ]

= 264, [micro-ohm/cm]
cm

Step No. 9 Calculate the effective window area, Wa(eff). Use the window area found in Step 6. A typical
value for, S3, is 0.75, as shown in Chapter 4.

Wa(eff)=WaS,, [cm2]
«V/)=( 2 - 34 )(°-75), [cm2]
WaW=\.16, [cm2]

Step No. 10 Calculate the number turns possible, N, using the insulated wire area, Aw found in Step 8. A
typical value for, 82, is 0.6, as shown in Chapter 4.

N = "(eff} 2
, [rums]
Aw
(1.76)(0.60)
ZA 115
vV = V-
(0.00754)r-> [to ]
N = 140, [turns]

Step No. 1 1 Calculate the required gap, lg.

lg= -Z -- —J, , [cm]

^

/ (1.26)(140)2(L25)(10-8) / 9.22 >|

g
(0.0025) '
/ g =0.120, [cm]

Step No. 12 Calculate the equivalent gap in mils.

mils = cm(393.7)
mils = (0.120)(393.7)
mils = 47.2 use 50

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 13 Calculate the fringing flux factor, F.

(0.120) l (2(2.84}
V
- '
11.25 ( 0.120

Step No. 14 Calculate the new number of turns, Nn, by inserting the fringing flux, F.

/ L
Nn = *—. r, [turns]
\ 0.4*4 ^K 8 )
I (0.120)(0.0025)

AT, =116, [turns]

Step No. 15 Calculate the winding resistance, RL. Use the MLT, from Step 5, and the micro-ohm per
centimeter, from Step 10.

), [ohms]

RL = (8.3)(ll6)(264)(lO~ 6 ), [ohms]
RL = 0.254, [ohms]

Step No. 16 Calculate the copper loss, Pcu.

Pa,=lL*L, [watts]
PC1I =(1.51) 2 (0.254), [watts]
Pm = 0.579, [watts]

Step No. 17 Calculate the regulation, a.

a =^(100), [%]
o

(0.579)
a = ^(100)^
a =0.579, [%]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 18 Calculate the ac flux density, Bac.

fteslal
p. )
•
(0.120)+
V -—
' 1,2500
Bac= 0.0167, [tesla]

Step No. 19 Calculate the watts per kilogram for ferrite, P, material in Chapter 2. Watts per kilogram can
be written in milliwatts per gram.

mW/g = (0.00004855)(200000)(1 63) (0.0167 )(

mW/g = 0.468

Step No. 20 Calculate the core loss, Pfe.

l O - 3 , [watts]
pfe = (0.468)(60)(lO~ 3 ), [watts]
Pfe= 0.0281, [watts]

Step No. 21 Calculate the total loss copper plus iron, PE.

P^Pfe+Pcv, [watts]
Pz =(0.0281) + (0.579), [watts]
PE = 0.607, [watts]

Step No. 22 Calculate the watt density, v|/. The surface area, A,, can be found in Step 5.

p
<//=—, [watts/cm2]
4
(0.607)
w = —, r-, [watts/cm" ]
(69.9)
V/ = 0.00868, [watts/cm2]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 23 Calculate the temperature rise, Tr.

Tr =450(0.00868) ( ° 826) ,
Tr = 8.92, [°C]

Step No. 24 Calculate the peak flux density, Bpk.

[tesla]
'
(0.127)+
V -—
' (2500)
Bpk = 0.252, [tesla]

Step No. 25 Calculate the effective permeability, |Lie. Knowing the effective permeability, the ETD-39
ferrite core can be ordered with a built in gap.

(2500)

9.22
jue = 74.5 use 75

Step No. 26 Calculate the window utilization, Ku.

K
=

_(116)(0.00653)
K =
" (234)
A:,, = 0.324

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Chapter 9

DC Inductor Design Using Powder Cores

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table of Contents

1. Introduction
2. Molybdenum Permalloy Powder Cores (MPP)
3. High Flux Powder Cores (HF)
4. Sendust Powder Cores (Magnetics Kool MH)
5. Iron Powder Cores
6. Inductors
7. Relationship of, Ap, to Inductor's Energy-Handling Capability
8. Relationship of, Kg, to Inductor's Energy-Handling Capability
9. Fundamental Considerations
10. Toroidal Powder Core Design Using the Core Geometry, Kg, Approach
11. Toroidal Powder Core Inductor Design, Using the Area Product, Ap, Approach

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Introduction
Powder cores are manufactured from very fine particles of magnetic materials. The powder is coated with
an inert insulation to minimize eddy current losses and to introduce a distributed air gap into the core
structure. The insulated powder is then compacted into toroidal and EE cores. The magnetic flux in a
toroidal powder core can be contained inside the core more readily than in a lamination or C core, as the
winding covers the core along the entire magnetic path length. The design of an inductor also frequently
involves consideration of the effect of its magnetic field on devices near where it is placed. This is
especially true in the design of high-current inductors for converters and switching regulators used in
spacecraft.

Toroidal powder cores are widely used in high-reliability military and space applications because of their
good stability over wide temperature ranges, and their ability to withstand high levels of shock, vibration,
and nuclear radiation without degradation. Other applications for these cores are:

1. Stable, high-Q filters operating in the frequency range of 1kHz to 1 MHz.

2. Loading coils used to cancel out the distributed capacitance in telephone cables.
3. Pulse transformers.
4. Differential mode EMI noise filters.
5. Flyback transformers.
6. Energy storage, or output inductors, in circuits with large amounts of dc current flowing.

Molybdenum Permalloy Powder Cores (MPP)

Molybdenum Permalloy Powder Cores (MPP) are manufactured from very fine particles of an 81 % nickel,
17% iron, and a 2% molybdenum alloy. The insulated powder is then compacted into EE and toroidal
cores. The toroidal cores range in size from 0.1 inch (0.254 cm) to 5 inches (12.7 cm) in the outside
diameter. MPP cores are available in permeabilities ranging from 14 up to 550. See Table 9-1

High Flux Powder Cores (HF)

High Flux Powder Cores (HF) are manufactured from very fine particles of a 50% nickel, and 50% iron.
The insulated powder is then compacted into EE and toroidal cores. The toroidal cores range in size from
0.25 inch (0.635 cm) to 3 inches (7.62 cm) in the outside diameter. HF cores are available in permeabilities
ranging from 14 up to 160. See Table 9-1.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Sendust Powder Cores (Magnetics Kool

Sendust powder cores are manufactured from very fine particles of an 85% iron, 9% silicon, and 6%
aluminum. The insulated powder is then compacted into EE and toroidal cores. The toroidal cores range in
size from 0.14 inch (0.35 cm) to 3 inches (7.62 cm) in the outside diameter. Sendust cores are available in
permeabilities ranging from 26 up to 125. See Table 9-1

Iron Powder Cores

The low cost iron powder cores are typically used in today's low and high frequency power switching
conversion applications for differential-mode, input and output, power inductors. The distributed air gap
characteristic of iron powder produces a core with permeability ranging from 10 to 100. This feature, in
conjunction with the inherent high saturation point of iron, makes it very difficult to saturate. While iron
powder cores may be limited in their use because of low permeability or rather high core loss at high
frequency, they have become a very popular choice in either EE or toroidal as a core material for high-
volume commercial applications. They are popular due to their low cost compared with other core
materials. The toroidal cores range in size from 0.3 inch (0.76 cm) to 6.5 inches (16.5 cm) in the outside
diameter. See Table 9-1

Table 9-1. Standard Powder Core Permeability

Standard Powder Core Permeabilities

Powder Material MPP High Flux Sendust Iron Powder
(Kool Mu)
Initial Permeability, Uj
10 X
14 X X
26 X X X
35 X
55 X
60 X X X X
75 X X
90 X
100 X
125 X X X
147 X X
160 X X
173 X
200 X
300 X
550 X

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Inductors
Inductors that carry direct current are used frequently in a wide variety of ground, air, and space
applications. Selection of the best magnetic core for an inductor frequently involves a trial-and-error type
of calculation.

The design of an inductor also frequently involves consideration of the effect of its magnetic field on other
devices in the immediate vicinity. This is especially true in the design of high-current inductors for
converters and switching regulators used in spacecraft, which may also employ sensitive magnetic field
detectors. For this type of design problem, frequently it is imperative that a toroidal core be used. The
magnetic flux in a powder core can be contained inside the core more readily than in a lamination or C
core, as the winding covers the core along the entire magnetic path length. The author has developed a
simplified method of designing optimum dc carrying inductors with powder cores. This method allows the
correct core permeability to be determined without relying on the trial and error method.

Relationship of, Ap, to Inductor's Energy-Handling Capability

The energy-handling capability of a core is related to its area product, Ap, by the equation:

2(Energy)(l0 4 )
Ap =— n — r v- '-, L[cm4J] [9-1]
L J
B
mJKu

Where: Energy is in watt-seconds.

Bm is the flux density, tesla.
J is the current density, amps-per-cm2.
Ku is the window utilization factor. (See Chapter 4)

From the above factors, such as flux density, Bm, window utilization factor, Ku, (which defines the
maximum space that may be used by the copper in the window), and the current density, J, which controls
the copper loss can be seen. The energy-handling capability of a core is derived from:

LJ2
Energy = , [watt-seconds] [9-2]

Relationship of, Kg, to Inductor's Energy-Handling Capability

Inductors, like transformers, are designed for a given temperature rise. They can also be designed for a
given regulation. The regulation and energy handling ability of a core is related to two constants:

g= ( nergy)
L J L J
If V
K
SKe

Where a is the regulation, %.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The constant, Kg, is determined by the core geometry:

W A2 K
K g = " c " , [cm5] [9-4]
MLT

The constant, Ke, is determined by the magnetic and electrical operating conditions:

[9-5]

The output power, P0, is defined in Figure 9-1.

P ~- Po(L2) ~-VI
r
(02)''(02)

LI L2
Ql
+O i
1
111
""I
Cl
CR1 y CR2 C2 +
V. n 7 f^ -> 2 ^ V
(02) f~ Vr

1 \
\ I \ i

Figure 9-1. Defining the Inductor Output Power.

The operating flux density, Bm, is:

, [tesla] [9-7]

From the above, it can be seen that the flux density, Bm, is the predominant factor in governing size.

Fundamental Considerations

The design of a linear reactor depends upon four related factors:

1. Desired inductance, L.
2. Direct current, Idc.
3. Alternating current, AI.
4. Power loss and temperature, Tr.

With these requirements established, the designer must determine the maximum values for, Bjc, and Bac,
that will not produce magnetic saturation and must make trade-offs that will yield the highest inductance
for a given volume. The core permeability chosen dictates the maximum dc flux density that can be
tolerated for a given design.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 If an inductance is to be constant with the increasing direct current, there must be a negligible drop in
inductance over the operating current range. The maximum H (magnetizing force) then is an indication of a
core's capability, as shown in Figure 9-2.

100 i i i i i r
Powder Cores

80
Low Permeability

I 60 High Permeability
cE
*2
I 40

20

i i i i i d.
1.0 10 100 1000
DC Magnetizing Force, (oersteds)

Figure 9-2. Typical Permeability Versus dc Bias Curves for Powder Cores.

Most manufacturers give the dc magnetizing force, H in oersteds:

OAxNI
MPL

Some engineers prefer amp-turns:

= 0.8//(MPL), [amp-turns] [9-9]

Inductance decreases with increasing flux density, B, and magnetizing force, H, for various materials of
different values of permeability. The selection of the correct permeability for a given design is made using
Equation [9- 10].

B (MPL)fl
A

It should be remembered the maximum flux, Bm, depends upon, Bdc + Bac, in the manner in Figure 9-3.

-, [tesla] [9-11]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 0.4;r7V/ ffc /,(lO- 4 )
, [tesla] [9-12]
MPL

B,,,= , [tesla] [9-13]

MPL

[tesla]
'

H (oersteds)

Figure 9-3. Inductor Flux Density Versus Idc + AI Current.

The flux density for the initial design for Molypermalloy powder cores should be limited to 0.3T maximum
for, Bdc + Bac, as shown in Figure 9-4.

100

90

5 80

70

60

0.001 0.01 0.1 1.0

Polarized Flux Density, tesla

Figure 9-4. Inductance Versus dc Bias.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Toroidal Powder Core Design Using the Core Geometry, Kg, Approach
This design procedure will work with all powder cores.

Step No. 1 Design a linear dc inductor with the following specifications.

1. Inductance, L ........................................................................................................... = 0.0025 henrys

2. dc current, I0 .............................. . ............................................................................. = 1.5 amps
3. ac current, AI ........................................................................................................... = 0.2 amps
4. Output power, P0 ..................................................................................................... = 100 watts
5. Regulation, a ........................................................................................................... = 1.0 %
6. Ripple Frequency .................................................................................................... = 20 kHz
7. Operating flux density, Bm ...................................................................................... = 0.3 tesla
8. Core Material .......................................................................................................... = MPP
9. Window utilization, Ku ............................................................................................ = 0.4
10. Temperature rise goal, Tr ....................................................................................... = 25°C

Step No. 2 Calculate the peak current, lp^.

, [amps]

, [amps]

Ipk =1.6, [amps]

Step No. 3 Calculate the energy-handling capability.

LI*
Energy = —— , [watt-seconds]

(0.0025)(1.6)2
Energy = --——— , [watt-seconds]

Energy = 0.0032, [watt-seconds]

Step No. 4 Calculate the electrical conditions coefficient, IQ.

^ e =0.145(lOO)(0.3) 2 (lO" 4 )
K =0.0001305

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 5 Calculate the core geometry coefficient, Kg.

(Energy) 5
g=
Kea '
(0.0032)2 5
?
• ~(0.0001305)(1.0)'
Kg= 0.0785, [cm5]

Step No. 6 Select a MPP powder core from Chapter 3. The data listed is the closest core to the calculated
core geometry, Kg.

1. Core Number = 55586

2. Magnetic Path Length, MPL = 8.95 cm
3. Core Weight, Wtfe = 34.9 grams
4. Mean Length Turn, MLT = 4.40 cm
5. Iron Area, Ac = 0.454 cm2
6. Window Area, Wa = 3.94 cm
7. Area Product, Ap = 1.79 cm4
8. Core Geometry, Kg = 0.0742 cm5
9. Surface Area, A, = 64.4 cm2
10. Permeability, n = 60
11. Millihenrys-per-lk, AL = 38 mh

Step No. 7 Calculate the current density, J, using the area product Equation, Ap.

J = ---- , [amps-per-cm 1
BmApKu
2(0.0032)(l0 4 )
J =-,—r;-r,—r> [amps-per-cm 1
(0.3)(1.79)(0.4)
J = 298, [amps-per-cm2 ]

Step No. 8 Calculate the rms current, \rnK.

/ M O = > /0-5) 2 +(0.2) 2 , [amps]

/„„,=!. 51, [amps]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 9 Calculate the required bare wire area, Aw(B).

A —
/ rms rpTYl
,\
W ( R\ — J I Ull I

(L51)
A rL C m2 iJ
^'(298)'
4^=0.00507, [cm2]

Step No. 10 Select a wire from the Wire Table in Chapter 4. If the area is not within 10%, take the next
smallest size. Also, record micro-ohms per centimeter.

AWG = #20
Bare, AW(B} =0.00519, [cm2]
Insulated, Aw = 0.00606, [cm2 ]
j
—- = 332, [micro-ohm/cm]
cm )

Step No. 11 Calculate the effective window area, Wa(eff). Use the window area found in Step 6. A typical
value for, 83, is 0.75, as shown in Chapter 4.

, [cm2]

Step No. 12 Calculate the number turns possible for, N. Use the insulated wire area, Avv, found in Step 10.
A typical value for, 83, is 0.6, as shown in Chapter 4.

W S
N = —"<e-y) , [turns]
Aw
(2.96)(0.60)
N = ^—&—-i, [turns]
(0.00606)
N = 293, [turns]

Step No. 13 Calculate the required core permeability, u.

= 5 m (MPL)(l0 4 )
i A

(0.30)(8.95)(l04)
'" A ~(l.26)(3.94)(298)(0.4)
^=45.4

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Note:
The permeability of 45.4 is close enough to use a 60u core. Also note that there are other permeabilities
available, See Table 9-1. Because of size, Chapter 3 has listed only 60u Tables for MPP, High Flux,
Sendust and a 75 |i Table for Iron powder. For cores with other than 60u, use the manufacturer's catalog.

Step No. 14 Calculate the number of turns, NL, required.

^ =1000 — — , [turns]
(1000)

f 2 5}
AT,1 = 1 0 0 0 , —
3 8 , [turns]
VI /
NL = 256, [turns]

Step No. 15 Calculate the winding resistance, RL. Use the MLT from Step 6 and the micro-ohm per
centimeter from Step 10.

RL =(MLT)(7V i )|^|(lO- 6 ), [ohms]

RL =(4.4)(256)(332)(lO" 6 ), [ohms]
RL = 0.374, [ohms]

Step No. 16 Calculate the copper loss, Pcu.

^ = C ^ , [watts]
P CU =(1.51) 2 (0.374), [watts]
Pcu = 0.853, [watts]

Step No. 17 Calculate the regulation, a.

a =^(100), [%]

a = V
(100)
a =0.853, [%]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 18 Calculate the ac flux density, Bac.

Bac = -- - - , [tesla]
MPL
(l.25)(256)f°(60)(lO-)
*=

B flc =0.0215, [tesla]

Step No. 19 Calculate the watts per kilogram for the appropriate MPP powder core material in Chapter 2.
Watts per kilogram can be written in milliwatts per gram.

mW/g = kf(m}B?£
mW/g = (0.0055 1) (20000)° U) (0.021 5) <2 ' 2)
mW/g = 0.313

Step No. 20 Calculate the core loss, Pfe.

P / e =(mW/ g )(w^)(lO- 3 ), [watts]

/> /c =(0.313)(34.9)(lO' 3 ), [watts]
Pfe= 0.011, [watts]

Step No. 21 Calculate the total loss copper plus iron, Pv.

/> =(0.011) + (0.853), [watts]

PZ = 0.864, [watts]

Step No. 22 Calculate the watt density, vj/. The surface area, At can be found in Step 6.

p
y/ = —, [watts/cm2 ]
4
(°-864r-,
w = -. ) r[watts/cm
, 2,]
(64.4)
i// = 0.0134, [watts/cm2]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 23 Calculate the temperature rise, Tr.

7>450(0.0134)(°826),
r r =12.8, [°C]

Step No. 24 Calculate the dc magnetizing force, H.

OAxN, I k
H = -L-^- , [oersteds]
MPL
(1.26)(256)(1.6)
H = ^-^-r^—'- , [oersteds]
(8.95)
H = 57.7, [oersteds]

Step No. 25 Calculate the window utilization, Ku.

„
A =

((256)(0.00519))
KU=
(3.94)
Ku = 0.337

Note:
The big advantage in using the core geometry design procedure is that the current density is calculated.
Using the area product design procedure, the current density is an estimate at best. In this next design the
same current density will be used as in core geometry.

Toroidal Powder Core Inductor Design, Using the Area Product, Ap, Approach

Step No. 1 Design a linear dc inductor with the following specifications.

1. Inductance, L = 0.0025 henrys

2. dc current, I0 = 1.5 amps
3. ac current, AI = 0.2 amps
4. Output power, P0 = 100 watts
5. Current Density, J = 300 amps-per-cm2
6. Ripple Frequency = 20 kHz
7. Operating flux density, Bm = 0.3 tesla
8. Core Material = MPP
9. Window utilization, Ku = 0.4
10. Temperature rise goal, Tr = 25°C

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 2 Calculate the peak current, lp^.

I 1
pk= o+ — > [amPs]

/ p t =(l.5) + ^, [amps]

/ M =1.6, [amps]

Step No. 3 Calculate the energy-handling capability.

Energy = —— , [watt-seconds]

(0.0025)(1.6)2
Energy = --——— , [watt-seconds]

Energy = 0.0032, [watt-seconds]

Step No. 4 Calculate the area product, Ap.

2(Energy)(lQ 4 )
A = ----, [cm]

= 2(0.0032)(104)
L m J
" (0.3)(300)(0.4)' °
4
4, =1.78, [cm ]

Step No. 5 Select a MPP powder core from Chapter 3. The data listed is the closest core to the calculated
core geometry, Kg.

1. Core Number .................................................................................................................. = 55586

2. Magnetic Path Length, MPL ......................................................................................... = 8.95 cm
3. Core Weight, W,fe ........................................................................................................... = 34.9 grams
4. Mean Length Turn, MLT ................................................................................................ = 4.40 cm
5. Iron Area, Ac .................................................................................................................. = 0.454 cm2
6. Window Area, Wa .......................................................................................................... = 3.94 cm2
7. Area Product, Ap ............................................................................................................ = 1.79 cm4
8. Core Geometry, Kg ......................................................................................................... = 0.0742 cm5
9. Surface Area, A, ............................................................................................................. = 64.4 cm2
10. Permeability, u ............................................................................................................. = 60
11. Millihenrys-per-lk, A L ............................................................................................... = 38 mh

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 6 Calculate the rms current, Irms.

^=V/o+A/2' [amPS]

/,ms=1.51, [amps]

Step No. 7 Calculate the required bare wire area, AW(B).

<Vo= [cm2]

W(B) ^y I J

AW(B)= 0.00507, [cm2]

Step No. 8 Select a wire from the Wire Table in Chapter 4. If the area is not within 10%, take the next
smallest size. Also, record the micro-ohms per centimeter.

AWG = #20
Bare, AW(B)= 0.00519, [cm2]
Insulated, 4,, =0.00606, [cm2 ]

—— = 332, [micro-ohm/cm]

Step No. 9 Calculate the effective window area, Wa(eff). Use the window area found in Step 5. A typical
value for, 83 is 0.75 as shown in Chapter 4.

W,~=WS,,
a e
\ JJ} a J
[cm2]
»W,=(3.94)(0.75), [cm2]
0^=2.96, [cm2]

Step No. 10 Calculate the number turns possible, N. Use the insulated wire area, Aw found in Step 8. A
typical value for, S2, is 0.6, as shown in Chapter 4.

W..^S,
-, [turns]
Aw
(2.96)(0.60)
N=±--^—-+, [turns]
(0.00606)
N = 293, [turns]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 11 Calculate the required core permeability, u.

5 m (MPL)(l0 4 )

(0.30)(8.95)(l0 4 )
(1.26)(3.94)(298)(0.4)

Note:
The permeability of 45.4 is close enough to use a 60u core. Also note there are other permeabilities
available, See Table 9-1. Because of size, Chapter 3 has listed only 60^ tables for MPP, High Flux,
Sendust and 75u, table for Iron powder. For cores with other than 60u, use the manufacturer's catalog.

Step No. 12 Calculate the number of turns, NL, required.

JVL =10001—^, [turns]

^ -^(1000)

1(25}
^=1000. — , [turns]

NL = 256, [turns]

Step No. 13 Calculate the winding resistance, RL. Use the MLT from Step 6 and the micro-ohm per
centimeter from Step 10.

RL =(MLT)(jV t )j^j(lO~ 6 ), [ohms]

RL =(4.4)(256)(332)(lO~ 6 ), [ohms]
RL = 0.374, [ohms]

Step No. 14 Calculate the copper loss, Pc

Pc^'L^L, [watts]
2
^=(1.51) (0.374), [watts]
Pcu = 0.853, [watts]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 15 Calculate the ac flux density, Bac.

Bac= -± - , [tesla]
MPL
(L25)(256)fM](60)(lO-)
Bac= -r-^T- '
(8.95)
Bac= 0.0215, [tesla]

Step No. 16 Calculate the watts per kilogram for the appropriate MPP powder core material in Chapter 2.
Watts per kilogram can be written in milliwatts per gram.

mW/g = (0.00551)(20000)°23) (0.0215)(1'2)

mW/g = 0.313

Step No. 17 Calculate the core loss, Pfe.

^ /e )(lO" 3 ), [watts]
/} e =(0.313)(34.9)(lO- 3 ), [watts]
Pfe= 0.011, [watts]

Step No. 18 Calculate the total copper loss plus iron, Py.

PI. = Pfe + Pcu > [watts]

/> =(0.011) + (0.853), [watts]
Pz = 0.864, [watts]

Step No. 19 Calculate the watt density, v|/. The surface area, At can be found in Step 5.

p
y/ = —, [watts/cm2]
4
(0-864) ^
(64.4) '
^ = 0.0134, [watts/cm2]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 20 Calculate the temperature rise, Tr.

Tr = 450(0.0134)(°826) , [°
?;= 12.8, [°C]

Step No. 21 Calculate the dc magnetizing force, H.

0.47T NLI,
H = -—^ , [oersteds]
MPL
(1.26)(256)(1.6)
H = ±-^-^—'- , [oersteds]
(8.95)
H = 57.7, [oersteds]

Step No. 22 Calculate the window utilization, K u .

N
'^
4

J(256)(0.00519))
(3.94)
K.. = 0.337

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Chapter 10

AC Inductor Design

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table of Contents

1. Introduction
2. Requirements
3. Relationship of, Ap, to the Inductor Volt-Amp Capability
4. Relationship of, Kg, to the Inductor Volt-Amp Capability
5. Fundamental Considerations
6. Fringing Flux
7. AC Inductor Design Example
8. Reference

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Introduction

The design of an ac inductor is quite similar to that of a transformer. If there is no dc flux in the core, the
design calculations are straightforward. The apparent power, Pt, of an inductor is the VA of the inductor;
that is, the product of the excitation voltage and the current through the inductor.

P,=yA, [watts] [10-1]

Requirements
The design of the ac inductor requires the calculation of the volt-amp (VA) capability. In some
applications the inductance is specified, and in others, the current is specified. If the inductance is
specified, then, the current has to be calculated. If the current is specified, then the inductance has to be
calculated. A series, ac inductor, LI, being used in a Ferroresonant Voltage Stabilizer is shown in Figure
10-1.

LI

Cl
V i n = 105-130V V0 =120V

Figure 10-1. Series ac Inductor, LI, as used in a Ferroresonant Voltage Stabilizer.

Relationship of, Ap, to the Inductor Volt-Amp Capability

The volt-amp capability of a core is related to its area product, Ap, by the equation that may be stated as
follows:

-, [cm4] [10-2]

Where:
Kf = wave form factor.
Ku = window utilization factor.
Bttc = operating flux density, T, tesla
/ = operating frequency, Hz
J = current density, amps/cm2

From the above, it can be seen that factors such as flux density, Bac, the window utilization factor, Ku,
(which defines the maximum space occupied by the copper in the window), and the current density, J, all
have an influence on the inductor area product, Ap.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Relationship, Kg, to the Inductor Volt-Amp Capability

Although most inductors are designed for a given temperature rise, they can also be designed for a given
regulation. The regulation and volt-amp ability of a core is related to two constants:

VA
a =
^* A ' [10-3]

a = Regulation (%) [10-4]

The constant, Kg, is determined by the core geometry, which may be related by the following equations:

= , [on']
MLT [10-5]

The constant, Ke, is determined by the magnetic and electric operating conditions, which may be related by
the following equation:

Where:
K-f = waveform coefficient
4.0 square wave
4.44 sine wave

From the above, it can be seen that factors such as flux density, frequency of operation, and the waveform
coefficient have an influence on the transformer size.

Fundamental Considerations

The design of a linear ac inductor depends upon five related factors:

1. Desired inductance
2. Applied voltage, (across inductor)
3. Frequency
4. Operating Flux density
5. Temperature Rise

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 With these requirements established, the designer must determine the maximum values for, Bac, which will
not produce magnetic saturation, and make trade-offs that will yield the highest inductance for a given
volume. The core material selected determines the maximum flux density that can be tolerated for a given
design. Magnetic materials and their operating flux levels are given in Chapter 2.

The ac inductor like a transformer, must support the applied voltage, Vac. The number of turns is calculated
from Faraday's Law, which states:
10
P
' , [turns] [10-7]
K B
f ,,cfAc

The inductance of an iron-core inductor, with an air gap, may be expressed as:

L= - s ^ , thenrys] [10-8]
MPL

Inductance is seen to be inversely dependent on the effective Magnetic Path Length, MPL, which is the
sum of the air gap length, lg, and the ratio of the Magnetic Path Length, MPL, to material permeability, um.

When the core air gap, lg, is larger compared to the ratio, MPL/iim, because of the high material
permeability, um, variations in, |tim, do not substantially affect the total effective Magnetic Path Length,
MPL, or the inductance, L. The inductance equation then reduces to:

0.4?rN 2 A (lO~ 8 )
L= -,, [henrys] [10-9]

Rearranging the equation to solve for the gap:

i(io- 8 )
, [cm] [10-10]

Fringing Flux

Final determination of the air gap requires consideration of the effect of fringing flux, which is a function
of gap dimension, the shape of the pole faces, and the shape, si/e, and location of the winding, as shown in
Figure 10-2 and Figure 10-3. Its net effect is to make the effective air gap less than its physical dimension.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 El Core Configuration

J
lj^>
_
T!Xr>n "/
X 3R p
s^
/ §
JK
H^
XB w
*^ w
BB
w ^s3H
i<
w

Fringing ¥\ux
s/
/ RHb XK X
w
Winding
^i H X XW
X X
Gap KX X

3\X8 3
K M
i^/
^

Figure 10-2. Fringing Flux Location on an El Core Configuration.

EE or Pot Core Configuration

Fringing Flux Fringing Flux

Winding

Gap Bobbin

Figure 10-3. Fringing Flux Location on an EE or Pot Core Configuration.

Fringing flux decreases the total reluctance of the magnetic path, and therefore increases the inductance by
a factor, F, to a value greater than that calculated from Equation [10-9]. Fringing flux is a larger percentage
of the total for larger gaps. The fringing flux factor is:

\2G
[10-11]
\T.
Where G is a dimension, (winding length), defined in Chapter 3. Equation [10-11] is valid for cut C cores,
laminations and cut ferrite cores.

The inductance, L, computed in Equation [10-9] does not include the effect of fringing flux. The value of
inductance, L', in Equation [10-12] does correct for fringing flux:

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 c
L' = — —± '-, [henrys] [10-12]

Now that the fringing flux, F, has been calculated, it is necessary to recalculate the number of turns using
the fringing flux, Factor F.

[10-13]

After the new turns, N(new), have been calculated, then, use Equation [10-13] with the new turns, N(new), and
solve for Bac. This check will provide the operating flux density, in order to calculate the core loss, Pfc, and
will also provide a check on core saturation margin.

^(io 4 )
*»= I, ' , [tesla] [10-14]
K N
f (^)fAc

The losses in an ac inductor are made up of three components:

1. Copper loss, Pcu
2. Iron loss, Pf e
3. Gap loss, Pg

The copper loss, Pcu, is I2R and is straightforward, if the skin effect is minimal. The iron loss, Pfe, is
calculated from core manufacturers' data. Gap loss, Pg, is independent of core material strip thickness and
permeability. Maximum efficiency is reached in an inductor, as in a transformer, when the copper loss, Pcu,
and the iron loss, Pfe, are equal, but only when the gap of the core is zero. The gap loss does not occur in
the air gap, itself, but is caused by magnetic flux, fringing around the gap, and reentering the core in a
direction of high loss. As the air gap increases, the flux across the gap fringes more and more, and some of
the fringing flux strikes the core, perpendicular to the laminations, and sets up eddy currents which cause
additional losses called gap loss, Pg. Also distribution of the fringing flux is affected by other aspects of the
core geometry, the proximity of the coils turns to the core, and whether there are turns on both legs. (See
Table 10-1). Accurate prediction of the gap loss depends on the amount of fringing flux. (See the
Reference at the end of this Chapter)

Pg=K,ElJB2ac, [watt] [10-15]

Where, E (as defined in Chapter 3) is the strip or tongue width, in cm.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table 10-1. Gap Loss Coefficient

Configuration Kj
Two-coil C core 0.0388
Single-coil C core 0.0775
Lamination 0.1550

When designing inductors where there is a choice of cores, always pick the core with the smallest ratio:

W,
= [smallest ratio] [10-16]

Comparing two cores with identical area products, Ap, for the same design specification, the core with a
minimum of window area will generate a minimum of fringing flux. If there is a design change and it
requires the use of the next larger core, it would be far more beneficial to double up on the core being used,
than to pick a larger core, as shown in Figure 10-4.

Next Larger Core

Doubling the Original Core

Original Core

Figure 10-4. Comparing Core Configurations.

For example, if the next larger core was selected, normally all of the core proportions will increase. This
means, the window, Wa, and the iron cross-section, Ac, would have both increased. (A larger core should
not be used as the fringing flux would also increase.) If you want to keep the fringing flux to a minimum,
then double up on the original core. Therefore the iron area, Ac, would double, but the window area, Wa,
will remain the same. This will reduce the, Wa/Ac, ratio, as shown in Equation 10-16. With an increase in
iron cross-section, Ac, the turns would have to decrease for the same window area, Wa. With a decrease in
turns, the gap would also decrease, resulting in less fringing flux.

When designing a transformer, the engineer will push the flux density as far he can without saturating the
core. That cannot be done with an ac inductor because you must leave a margin for the fringing flux factor.
One of the biggest problems in designing ac inductors is keeping the gap to a minimum. This problem
becomes acute when designing high frequency ac inductors. The problem in designing high frequency
inductors is the required turns to support the applied voltage, then gapping to provide the proper
inductance. This problem is minimized when using powder cores if the right permeability can be found.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 AC Inductor Design Example

Step No. 1 Design a linear ac inductor with the following specifications.

1. Applied voltage, VL = 120 volts

2. Line current, IL = 1.0 amps
3. Line frequency = 60 hertz.
4. Current density, J = 300 amp/cm2
5. Efficiency goal, r)(100) = 90%
6. Magnetic material = Silicon
7. Magnetic material permeability, um = 1500
8. Flux density, Bac = 1.4 tesla
9. Window utilization, Ku = 0.4
10. Waveform factor, Kf = 4.44
11. Temperature rise goal, Tr = 50°C

Step No. 2 Calculate the apparent power, Pt or VA of the inductor, L.

VA = V L I L , [watts]
K/4=(120)(1.0), [watts]
VA=UO, [watts]

Step No. 3 Calculate the area product, Ap.

A= ^—, [cm4]
" KfKJBacJ

A (120)(104) 4
p
(4.44)(0.4)(60)(1.4)(300)'
^=26.8, [cm4]

Step No. 4 Select an El lamination from Chapter 3. The closest lamination to the calculated area product,
Ap, is the El-100.

1. Core Number = EI-100

2. Magnetic Path Length, MPL = 15.2 cm
3. Core Weight, Wtfe = 676 grams
4. Mean Length Turn, MLT = 14.8 cm
5. Iron Area, Ac = 6.13 cm2
6. Window Area, Wa = 4.84 cm2
7. Area Product, Ap = 29.7 cm4

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 8. Core Geometry, Kg = 4.93 cm5
9. Surface Area, A, = 213 cm2
10. Winding length, G = 3.81 cm
11. Lamination tongue, E = 2.54 cm

Step No. 5 Calculate the number of inductor turns, NL.

K L (10 4 )
NL = ^ '—, [turns]
KfBmfAe'

N Q2Q)(io 4 )
L
(4.44)(1.4)(60)(6.13)'
NL = 525, [turns]

Step No. 6 Calculate the inductive reactance, XL.

XL=-, [ohms]

(120)
[ohms]
** (T5)'
^=120, [ohms]

Step No. 7 Calculate the required inductance, L.

X
L=—^-, [henrys]
271 f

L = —-——-, [henrys]
2(3.14)(60)
1 = 0.318, [henrys]

Step No. 8 Calculate the required gap, Lg.

(1.26)(525)2(6.13)(10-^ n 5 . 2 ^
~ \-\ > tcm]
0.318 I U500J
v
/ = 0.0568, [cm] or / = 22.4, [mils]: This would be in 10 mils each leg.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 9 Calculate the fringing flux, F.

/6.13 0.0568

Step No. 10 Using the fringing flux, recalculate the series inductor turns, N L(nevv)

=459
#i(«»-> >

Step No. 1 1 Using the new turns, recalculate the flux density, Bac.

BK= - --— , [tesla]

K N
f L(^)4f

S
(12Q)(1Q 4 )
~
(4.44)(459)(6.13)(60)'
=1.6, [tesla]

Step No. 12 Calculate the inductor bare wire area, Aw L(B)-

=~J>

_ (1.0)
~

4^=0.00333, [cm2]

Step No. 13 Select a wire from the Wire Table in Chapter 4.

4^=0.00324, [cm2]

=531, [micro-ohm/cm]
cm )

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 14 Calculate the inductor winding resistance, RL. Use the MLT from the core data found in Step 4
and the micro-ohm per centimeter found in Step 13.

RL=(MLi:)(Ns)(^](lO'6}, [ohms]
\ cm J
i^m J * '

RL =(l4.8)(459)(53l)(l(T 6 ), [ohms]
RL=3.6l, [ohms]

Step No. 15 Calculate the inductor winding copper loss, PL.

^ = ( O X [watts]
PL =(1.0) 2 (3.61), [watts]
/>=3.61, [watts]

Step No. 16 Calculate the watts-per-kilograms, W/K, for the appropriate core material. See Chapter 2.

Wl K = 0.000557/0 68)fis(186), [watts-per-kilogram]

68)
WIK = 0.000557(60)° (l .6)° 86) , [watts-per-kilogram]
WIK = 1.30, [watts-per-kilogram]

Step No. 17 Calculate the core loss in watts, P fe-

Pfe=(WIK}Wtfe, [watts]
/> /e =(l.30)(0.676), [watts]
Pfe = 0.878, [watts]

Step No. 18 Calculate the gap loss, Pg.

P^K.ElJBl, [watts]
Pg =(0.155)(2.54)(0.0568)(60)(1.6)2, [watts]
Pg = 3.43, [watts]

Step No. 19 Calculate the total inductor losses, P2.

F s =(3.6l) + (0.878) + (3.43), [watts]

Pz = 7.92, [watts]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 20 Calculate the inductor surface area watt density, \\i.

p
\l/ = —- , [watts-per-cm2 ]
4
(7.92) ,
y/ = --^ , [watts-per-cm ]

iff = 0.0372, [watts-per-cm2 ]

Step No. 2 1 Calculate the temperature rise, Tr.

Tr = 450(0.0372)(°826) , [°
Tr = 29.7, [°C]

Step No. 22 Calculate the window utilization, Ku.

N,
A,. =

_ ((459)(0.00324))
K =
" (JM)
K,. = 0.307

Reference
1. Ruben, L., and Stephens, D. Gap Loss in Current-Limiting Transformer. Electromechanical Design,
April 1973, pp. 24-126.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Chapter 11

Constant Voltage Transformer (CVT)

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table of Contents

1. Introduction
2. Constant-Voltage Transformer, Regulating Characteristics
3. Electrical Parameters of a CVT Line Regulator
4. Constant-Voltage Transformer, Design Equations
5. Constant-Voltage Transformer, Design Example
6. Series AC Inductor, Design Example
7. References

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Introduction
The constant-voltage transformer (CVT) has a wide application, particularly where reliability and inherent
regulating ability against line voltage changes are of prime importance. The output of a constant-voltage
transformer is essentially a square wave, which is desirable for rectifier output applications while, also
having good circuit characteristics. The main disadvantage to a constant-voltage transformer is efficiency
and regulation for frequency and load. The equations presented here for designing a constant-voltage
transformers at line frequency have been used at 400 Hz on aircraft, and as high as 20 kHz.

Constant-Voltage Transformer, Regulating Characteristics

The basic two-component (CVT) Ferroresonant regulator is shown in Figure 11-1. The inductor, LI, is a
linear inductor and is in series with Cl across the input line. The voltage across capacitor, Cl, would be
considerably greater than the line voltage, because of the resonant condition between LI and Cl.

LI

Cl
p
t
O
105- 130V ^-

ri V0 = 120V

Figure 11-1. Two Component Ferroresonant Voltage Stabilizer.

The voltage, Vp, can be limited to a predetermined amplitude by using a self-saturating transformer, Tl,
which has high impedance, until a certain level of flux density is reached. At that flux density, the
transformer saturates and becomes a low-impedance path, which prevents further voltage buildup across
the capacitor. This limiting action produces a voltage waveform that has a fairly flat top characteristic as
shown in Figure 11-2 on each half-cycle.

Figure 11-2. Primary Voltage Waveform of a Constant Voltage Transformer.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Electrical Parameters of a CVT Line Regulator
When the constant voltage transformer is operating as a line regulator, the output voltage will vary as a
function of the input voltage, as shown in Figure 11-3. The magnetic material used to design transformer,
Tl, has an impact on line regulation. Transformers designed with a square B-H loop will result in better
line regulation. If the output of the line regulator is subjected to a load power factor (lagging) with less
than unity, the output will change, as shown in Figure 11-4.

120

~o
> 118

O 117

i i i i i i i i i i i i i i
95 100 105 110 115 120 125 330
Input Voltage, Vjn

Figure 11-3. Output Voltage Variation, as a Function of Input Voltage.

50% Load

100% Load

1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3

Load Power Factor, (lagging)

Figure 11-4. Output Voltage Variation, as a Function of Load Power Factor.

If the constant voltage transformer is subjected to a line voltage frequency change the output voltage will
vary, as shown in Figure 11-5. The regulation of a constant-voltage transformer can be designed to be
better than a few percent. Capability for handling a short circuit is an inherent feature of a constant-voltage
transformer. The short-circuit current is limited and set by the series inductance, L. The regulation
characteristics at various lines and loads are shown in Figure 11-6. It should be noted that a dead short,
corresponding to zero output voltage, does not greatly increase the load current; whereas for most
transformers, this dead short would be destructive.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 5? 8.0
& 6.0
0 4.0
3 2.0
1 0
s 2-°
| 4.0
If 6.0
5 8.0 i i i i i i i i i i i i i
56 58 60 62 64 66
Line Frequency, Hz

Figure 11-5. Output Voltage Variation, as a Function of Line Frequency Change.

160
140

u 100
M 80
o 95V Input 115V 130V Input
> 60
I 40
O 20
i i i i i i
60 80 100 120 140 160 180
% Rated Load

Figure 11-6. Output Voltage Variation, as a Function of Output Voltage vs. Load.

Constant-Voltage Transformer, Design Equations

Proper operation and power capacity of a constant-voltage transformer (CVT) depends on components, LI
and Cl, as shown in Figure 11-7. Experience has shown that the, LC, relationship is:

LCa)2 =1.5 [11-1]

LI
CV1

Cl
V
m V,,

Figure 11-7. Basic, Constant Voltage Transformer Circuit.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The inductance can be expressed as:

[henrys] [11-2]
Ico
The capacitance can be expressed as:

S-*
[farads] [11-3]

Referring to Figure 11-7, assume there is a sinusoidal input voltage, an ideal input inductor, LI, and a
series capacitor, Cl. All voltage and currents are rms values. V;,, is the voltage value just before the circuit
starts to regulate at full load; RO(R), is the reflected resistance back to the primary, including efficiency; r| is
the efficiency, and, P0, is the output power.

/>„= — , [watts] [11-4]

[ohms] [11-5]

It is common practice for the output to be isolated from the input and to connect Cl to a step-up winding on
the constant-voltage transformer (CVT). In order to use smaller capacitor values, a step-up winding must
be added, as shown in Figure 11-8. The penalty for using a smaller capacitor requires the use of a steo-jp
winding. This step-up winding increases the VA or size of the transformer. This can be seen in Equation
11-6. The energy in a capacitor is:

Energy =
cv2 , [watt-seconds]
[11-6]
, [farads]

Figure 11-8. CVT, with a Capacitor Step-up Winding.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The secondary current, Is, can be expressed as:

/ , = - - , [amps] [11-7]

With the step-up winding, the primary current, Ip, is related to the secondary current by the following
Equation [Ref. 3]:

, [amps] [11-8]
V (l-3)
c j

The current, Ic, through the capacitor, is increased by, Kc, because of the effective higher frequency. Due to
the quasi-voltage waveform as shown in Figure 1 1-2, the equivalent ac impedance of the resonant capacitor
is reduced to some value lower than its normal sine wave value. This is due to an increase in odd
harmonics.

Ic=KcVccoC, [amps] [11-9]

Where, Kc, can vary from 1.0 to 1.5.

Empirically, it has been shown that for good performance the primary operating voltage should be:

, [volts] [11-10]

When the resonating capacitor is connected across a step-up winding, as is Figure 11-8, both the value of
the capacitor and the volume can be reduced. Cn, is the new capacitance value, and, Vn, is the new voltage
across the capacitor.

C V2 - C V2
^n% '-(I-2) 0-2)

The apparent power, Pt, is the sum of each winding, VA:

The line voltage regulation of a constant- voltage transformer is:

AK p =4.44Afl s 4/W p (l0 4 ), [volts] [11-12]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The output voltage regulation of a constant-voltage transformer, for a change in line voltage, is a function
of the squareness of the B-H loop, as shown in Figure 11-9. The saturation flux density, Bs, is dependent
on the annealing process of the magnetic material. It seems that each manufacturer has his own annealing
process, which has an impact on the saturation flux density, Bs.

B, (teslas)

Bs, High Line

B,,, Low Line

-H I-H-H

Figure 11-9. B-H Loop of a CVT at High and Low Line.

Constant- Voltage Transformer, Design Example

Design a constant-voltage line regulator (CVT) with the following specifications.

1. Input voltage ..... ....... ....................................................................................................... = 105-129 volts

2. Line frequency ............................................................................................................... = 60 hertz.
3. Output voltage, Vs .......................................................................................................... = 120 volts
4. Output VA ....................................................................................................................... = 250 watts.
5. Transformer current density, J ........................................................................................ = 300 amps/cm2
6. Capacitor voltage, Vc ...................................................................................................... = 440 volts
7. Capacitor coefficient, Kc ................................................................................................. = 1.5
8. Efficiency goal, t|(100) .................................................................................................. = 85%
9. Magnetic Material .......................................................................................................... = Silicon
10. Saturating flux density, Bs ............................................................................................ = 1.95 tesla
11. Window utilization, Ku ................................................................................................. = 0.4
12. Temperature rise goal, Tr ............................................................................................... = 50°C

Step No. 1 Calculate the primary voltage, Vp.

Vp =(105)(0.95), [volts]
Vf = 99.75, [volts]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 2 Calculate the reflected resistance, RO(R) back to the primary, including efficiency r|.

PO
\2 ,
(99.75) (0.85)
Ro(R)
, ., = ^-'—^ -'-L , [ohms]
J

/? o(R) =33.8, [ohms]

Step No. 3 Calculate the required capacitance, Cl.

C = --- , [farads]

C = -—- ---, [farads]

0.33(377)(33.8)
C = 238(lO~ 6 ), [farads]

Step No. 4 Calculate the new capacitance value, using the higher voltage, Vc.

(238(lO~ 6 ))(99.75) 2
°"3) (440)2
C (1 _ 3) =12.3(lO- 6 ), [farads]

A standard motor run capacitor is a 12.5 (.if/440v.

Step No. 5 Calculate the capacitor current, Ic.

Ic=\.5VccoC, [amps]
Ie =1.5(440)(377)(l2.5(lO- 6 )), [amps]
/ c =3.11, [amps]

Step No. 6 Calculate the secondary current, Is.

!,=%-, [amps]

25
/r s = —° , r[amps]i

Is = 2.08, [amps]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 8 Calculate the apparent power Pt.

P, = VA(l_2) + vA{2_3) + yA^5), [watts]

VA(l_2) = Vf Ip = (99.75)(4.35) = 434, [watts]
VA
(^) =(VC-VP}IC =(340)(3.11) = 1057, [watts]
VA(^i} = V S I S = (120)(2.08) = 250, [watts]
pt =(434) + (l057) + (250), [watts]
7^=1741, [watts]

Step No. 9 Calculate the area product, Ap.

A = - -'—, [cm4!
" KfKJBsJ
(1742)(10 4 )
A =
" (4.44)(0.4)(60)(1.95)(300)'
Ap=279, [cm4]

Step No. 10 Select an El lamination from Chapter Three. The closest lamination to the calculated area
product, Ap, is the El-175.

1. Core Number = El-175

2. Magnetic Path Length, MPL = 26.7 cm
3. Core Weight, Wtfe = 3.71 kilograms
4. Mean Length Turn, MLT = 25.6 cm
5. Iron Area, Ac = 18.8 cm2
6. Window Area, Wa = 14.8 cm2
7. Area Product, Ap = 278 cm4
8. Core Geometry, Kg = 81.7 cm5
9. Surface Area, A, = 652 cm2

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 1 1 Calculate the number of primary turns, Np.

FflO4)
= [tums]
^ '
(99.75)(l0 4 )
=
(4.44)(1.95)(60)(18.8)'
Np = 102, [turns]

Step No. 12 Calculate the primary bare wire area, Awp(B).

<„(«)= y> [cm2]

(4.35)
A Hp(B)
,,.=1 + [cm2]
(300)

Step No. 13 Select a wire from the Wire Table in Chapter 4.

4,^=0.0131, [cm 2 ]

—— =132, [micro-ohm/cm]
cm )

Step No. 14 Calculate the primary resistance, Rp. Use the MLT from the core data and the micro-ohm per
centimeter found in Step 13.

,) ^ (10-), [ohms]

#p =(25.6)(l02)(l32)(lO- 6 ), [ohms]
Rp = 0.345, [ohms]

Step No. 15 Calculate the primary copper loss, Pp.

P p =(4.35) 2 (0.345), [watts]

Pp = 6.53, [watts]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 16 Calculate the required turns for the step-up capacitor winding, Nc.

-V }
' "', [turns]
p
(!02)(440-99.75)
N=±-^-'-, [turns]
99.75
Nc = 348, [turns]

Step No. 17 Calculate the capacitor step-up winding bare wire area, AWC(B).

J
(3.11)

Am(B)= 0.0104, [cm2]

Step No. 18 Select a wire from the Wire Table in Chapter 4.

Au(B] =0.0104, [cm2]

\
=• 166, [micro-ohm/cm]
cm

Step No. 19 Calculate the capacitor winding resistance, R<.. Use the MLT from the core data and the micro-
ohm per centimeter found in Step 18.

4=(MLT)(;V c )|j£J(lO- 6 ), [ohms]

Rc =(25.6)(348)(l66)(lO~ 6 ), [ohms]
RC=IAS, [ohms]

Step No. 20 Calculate the capacitor step-up winding copper loss, Pc.

Pc=(lc}2Rc, [watts]
2
P C =(3.11) (1.48), [watts]
Pc = 14.3, [watts]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 21 Calculate the turns for the secondary, Ns.

NV
N,=-^-, [turns]
p
(102)(120)
TV = ^ '-+ '-, [turns]
99.75
Ns = 123, [turns]

Step No. 22 Calculate the secondary bare wire area, Aws(B).

_ (2.08)
ws(B}
~ (300)'
Am(g) = 0.00693, [cm2]

Step No. 23 Select a wire from the Wire Table in Chapter 4.

A JVC = #19
4^=0.00653, [cm 2 ]

=264, [micro-ohm/cm]
cm )

Step No. 24 Calculate the secondary winding resistance, Rs. Use the MLT from the core data and the
micro-ohm per centimeter found in Step 23.

ls — l<r , 7 [ohms]
\ cm j v
«s =(25.6)(l23)(264)(lO" 6 ), [ohms]
^=0.831, [ohms]

Step No. 25 Calculate the secondary winding copper loss, Ps.

P,=(1,)2R» [watts]
2
Ps =(2.08) (0.831), [watts]
Ps =3.59, [watts]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 26 Calculate the total copper loss, Pcu.

PCU=PP+P,+PO [watts]
^=(6.53) + (3.59) + (l4.3), [watts]
Pcu = 24.4, [watts]

Step No. 27 Calculate the watts-per-kilogram, W/K, for the appropriate core material. See Chapter 2.

WIK= 0.000557/1 '6s)Bf 86)

, [watts-per-kilogram]
WIK = 0.000557(60)° "8) (l.95)° 86) , [watts-per-kilogram]
W/K =1.87, [watts-per-kilogram]

Step No. 28 Calculate the core loss in watts, Pfe.

Pfe=(WIK)Wtft, [watts]
>
/ /c =(l.87)(3.7l), [watts]
Pfc = 6.94, [watts]

Step No. 29 Calculate the total losses, Pj>

Pz =(24.4) + (6.94), [watts]

^=31.34, [watts]

Step No. 30 Calculate the transformer surface watt density, \\i.

p
i// = —, [watts-per-cm2 ]
4
(31.34) ,
¥= t^\ > [watts-per-cm2]
(652)
\l/ = 0.048 1, [watts-per-cm2 ]

Step No. 3 1 Calculate the temperature rise, Tr.

Tr =450(0.0481) ( ° 826> ,
Tr = 36.7, [°C]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 32 Calculate the transformer efficiency, r\.

_^
(250 + 31.3)V '
77 = 88.9, [%]

Step No. 33 Calculate the window utilization, Ku.

+ ((348)(0.0104)) + ((123)(0.00653))
"" - (I£?j
K.. = 0.394

Series AC Inductor, Design Example

(Also See Chapter 9)

Step No. 33 Design a series, linear ac inductor with the following specifications:

1. Applied voltage = 129 volts

2. Line frequency = 60 hertz
3. Current density, J = 300 amp/cm2
4. Efficiency goal, r](100) = 85%
5. Magnetic material = Silicon
6. Magnetic material permeability, unl = 1500
7. Flux density, Bac = 1.4 tesla
8. Window utilization, Ku = 0.4
9. Waveform factor, Kf = 4.44
10. Temperature rise goal, Tr = 50°C

Step No. 34 Calculate the required series inductance, LI. See Figure 11-8.

11 = -^, [henrys]
2co

LI = \\ , [henrys]
2(377)
II = 0.0448, [henrys]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 35 Calculate the inductor reactance, XL.

XL = 2nfL\, [ohms]
XL = (6.28)(60)(0.0448) , [ohms]
^=16.9, [ohms]

Step No. 36 Calculate the short-circuit current, IL.

(129)
4L = 7-(,
(16.9)
IL = 7.63, [amps]

Step No. 37 Calculate the apparent power, P, or VA, of the input series inductor, LI. Use the high line
voltage of 129 volts and the normal running current, Ip, from Step 7.

K4=(129)(4.35), [watts]
VA = 561, [watts]

Step No. 38 Calculate the area product, Ap.

A=
" l/~
IS , A.
^f '—, [cm4]
/Y / DD JI

(561)(104)
J ;— : '_ CHI I
" (4.44)(0.4)(60)(1.4)(300)'
4, =125, [cm4]

Step No. 39 Select an El lamination from Chapter Three with the closest calculated area product, Ap.

1. Core Number = EI-138

2. Magnetic Path Length, MPL = 21 cm
3. Core Weight, Wtfe = 1.79 kilograms
4. Mean Length Turn, MLT = 20.1 cm
5. Iron Area, Ac = 11.6 cm2
6. Window Area, Wa = 9.15 cm"
7. Area Product, Ap = 106 cm4
8. Core Geometry, Kg = 24.5 cm5
9. Surface Area, A, = 403 cm2
10. Winding length, G = 5.24 cm
11. Lamination tongue, E = 3.49 cm

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 40 Calculate the number of inductor turns, NL.

,
* KfBaJAc'
(I29)(l0 4 )
^ = (4.44)(1.4)(60)(11.6)'
Wj, = 298, [turns]

Step No. 41 Calculate the required gap, Lg.

l
s - , [cm]

, 21 A
g , [cm]
I 0.0448 I U500;
/s = 0.276, [cm] or /g = 0.109, [mils]: This would be SOmils in each leg.

Step No. 42 Calculate the fringing flux, F.

v/11.6 0.276
F = 1.29

Step No. 43 Using the fringing flux, recalculate the series inductor turns, N L(new) .

i rz
:?—TT> [tun18]

^ i(1KW) =256, [rums]

Step No. 44 Using the new turns, recalculate the flux density, Bac.

(I29)(l0 4 )
flcc , [tesla]
" (4.44)(256)(11.6)(60)
Soc=1.63, [tesla]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 45 Calculate the inductor bare wire area, AwL(B).

»LW J
, [cm2]

_ (4.35)
«"-W ~ (300) '

AwL(B}= 0.0145, [cm2]

Step No. 46 Select a wire from the Wire Table in Chapter 4.

4^=0.01307, [cm2]
\
= 132, [micro-ohm/cm]
cm

Step No. 47 Calculate the inductor winding resistance, RL. Use the MLT from the core data and the micro-
ohm per centimeter found in Step 38.

RL = (MLT)(AO| —](lO-*), [ohms!

^ cm )
RL = (20.l)(256)(l32)(lO6 ), [ohms]
RL = 0.679, [ohms]

Step No. 48 Calculate the inductor winding copper loss, PL-

2
PL =(IL} RL, [watts]
2
PL =(4.35) (0.679), [watts]
PL = 12.8, [watts]

Step No. 49 Calculate the watts-per-kilogram, W/K, for the appropriate core material. See Chapter 2.

W/K = 0.000557/ 168) flf 86) , [watts-per-kilogram]

WIK = 0.000557(60)°68) (l.63)°'86), [watts-per-kilogram]
WIK = 1.34, [watts-per-kilogram]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 50 Calculate the core loss in watts, Pfe.

Pfe=(WIK}Wtfe, [watts]
P /C =(1.34)(1.79), [watts]
Pfi=2A, [watts]

Step No. 5 1 Calculate the gap loss, Pg.

Pg=K.,ElJBlc, [watts]
Pg =(0.155)(3.49)(0.276)(60)(1.63)2, [watts]
Pg = 23.8, [watts]

Step No. 52 Calculate the total losses, P£.

^=(12. 8) + (2.4) + (23. 8), [watts]

Pz = 39, [watts]

Step No. 53 Calculate the inductor surface area watt density, \\i.

p
= — , [watts-per-cm2 J
A
<
(39)
:, [watts-per-cm]

\l/ = 0.0968, [watts-per-cm2]

Step No. 54 Calculate the temperature rise, Tr.

7;=450(0.0968) ( ° 826) ,
r r = 65, [°C]
Step No. 55 Calculate the window utilization, Ku.

A.,, =

((2S6)(0.0131))
(9.15)
K,. = 0.367

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 References
1. Ruben, L., and Stephens, D. Gap Loss in Current-Limiting Transformer. Electromechanical Design,
April 1973, pp. 24-126.
2. H. P. Hart and R. J. Kakalec, "The Derivation and Application of Design Equations for Ferroresonant
Voltage Regulators and Regulated Rectifiers," IEEE Trans. Magnetics, vol. Mag-7, No.l, March 1971,
pp 205-211.
3. I. B. Friedman, "The Analysis and Design of Constant Voltage Regulators," IRE Trans. Component
Parts, vol. CP-3, March 1956, pp.11-14.
4. S. Lendena, "Design of a Magnetic Voltage Stabilizer." Electronics Technology, May 1961, pp. 154-
155.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Chapter 12

Three-Phase Transformer Design

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table of Contents

1. Introduction
2. Primary Circuit
3. Comparing Transformer, Physical Size
4. Phase Current, Line Current, and Voltage in a Delta System
5. Phase Voltage, Line Voltage, and Current in a Wye System
6. Comparing Multiphase and Single-Phase Power
7. Multiphase Rectifier Circuits
8. Area Product, Ap, and Core Geometry, Kg, for Three-Phase Transformers
9. Output Power Versus Apparent, Pt, Capability
10. Relationship, Kg, to Power Transformer Regulation Capability
11. Relationship, Ap, to Transformer Power Handling Capability
12. Three-Phase Transformer, Design Example

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Introduction
Three-phase power is used almost exclusively for generation, transmission, and distribution, as well as for
all industrial uses. It is also used on aircraft, both commercial and military. It has many advantages over
single-phase power. The transformer can be made smaller and lighter for the same power handling
capability, because the copper and iron are used more effectively. In circuitry, for conversion from ac to
dc, the output contains a much lower ripple amplitude, and a higher frequency component, which is 3 times
and 6 times the line frequency, and which, in turn, requires less filtering.

Primary Circuit
The two most commonly used primary circuits for three-phase transformers are the Star, or Y connection,
as shown in Figure 12-1, and the other being known as the Delta (A) connection, as shown in Figure 12-2.
The design requirement for each particular job dictates which method of connection will be used.

Line, 3 O

Figure 12-1. Three-Phase Transformer, Connected in Star.

Line, 1 O

Line, 2 O

Line, 3 O

Figure 12-2. Three-Phase Transformer, Connected in Delta.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Comparing Transformer, Physical Size
The schematic diagram in Figure 12-3, shows the connection of three single-phase transformers: (a)
Operating from a three-phase power source and a single three-phase transformer; and (b) Operating from a
three-phase power source connected in a delta-delta configuration. The single three-phase transformer, T4,
would be lighter and smaller than a bank of three single-phase transformers of the same total rating. Since
the windings of the three-phase transformer are placed on a common magnetic core, rather than on three
independent cores, the consolidation results in an appreciable saving in the copper, the core, and the
insulating materials.

Three-Phase, 3 Wire
Three-Phase, 3 Wire
Delta Input
Delta Output

Three-Phase, 3 Wire
Three-Phase, 3 Wire
Delta Output
Delta Input

(b)

Figure 12-3. Comparing Three Single-Phase Transformers Connected in Three-Phase Delta.

A cutaway view of a single-phase transformer, showing the window area and iron area of two types of core
configuration, is shown in Figure 12-4 and Figure 12-5. The El lamination, shown in Figure 12-4, is
known as a shell type, because it looks like the core surrounds the coil. The C core, shown in Figure 12-5,
is known as a core type, because it looks like the coil surrounds the core.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Winding
/"

Side View Top View

\
Core

Figure 12-4. Illustrating a Shell Type Transformer.

Winding

Side View Top View

\ Core

Figure 12-5. Illustrating a Core Type Transformer.

Cutaway views of a three-phase transformer are shown in Figure 12-6. These cross-sectional views show
the window and iron areas. The three-legged core is designed to take advantage of the fact that, with
balanced voltages impressed, the flux in each phase leg, adds up to zero. Therefore, no return leg is needed
under normal conditions. When the transformer is subjected to unbalanced loads, or unbalanced line
voltages, it may be best to use three single-phase transformers, because of the high-circulating currents.

Core
Side View

Figure 12-6. Cutaway View of a Three-Phase Transformer.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Phase Current, Line Current, and Voltage in a Delta System
In a Three-Phase Delta Circuit, such as the one shown in Figure 12-7, the line voltage and line current are
commonly called phase voltage and phase current. The line voltage, E(Line), will be the same as the actual
winding voltage of the transformer. However, the line current, I(Line). is equal to the phase current, I(phase),
times the square root of 3.

(Phase)

E
(Line, 1)

Figure 12-7. Voltage and Current Relationship of a Three-Phase Delta Circuit.

Phase Voltage, Line Voltage, and Current in a Wye System

The relationship between the Line voltage, and Line current, and the winding, or Phase voltage and Phase
current, in a Three-Phase Wye Circuit can be seen in Figure 12-8. In a Wye System, the voltage between
any two wires in the line will always be the square root of three times the phase voltage, E(PhaS!,,, between
the neutral, and any one of the lines.

""(phiise) -, [volts] [12-2]

E
(Line, 1)
O (Line)

Neutral
E E
(Line, 3) • (Line, 2) (Phase)

Figure 12-8. Voltage and Current Relationship for a Three-Phase Wye Circuit.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Comparing Multiphase and Single-Phase Power

Three-phase power distribution has a significant advantage over the single-phase. Most high power
equipment and industrial complexes will use three-phase power. One of the biggest advantages in using
three-phase power distribution has to do with smaller magnetic components handling the same power as
single-phase. This can be seen in aircraft, and shipboard equipment, as well as fixed ground installations.
One of the basic reasons for selecting three-phase is the transformer size. Another reason is, if dc is a
requirement, the capacitor and inductor filtering components are both smaller. The odd shape of a three-
phase transformer could be troublesome, as well as keeping balanced loads to minimize circulating
currents.

The single-phase, full wave bridge circuit is shown in Figure 12-9. The ripple voltage frequency is always
twice the line frequency. Only 50% of the total current flows through each rectifier. The three-phase,
Delta full wave bridge circuit is shown in Figure 12-10. The ripple voltage frequency is always 6 times the
line frequency. Only 33% of the total current flows through each rectifier. Looking at the ripple in Figure
12-10, it is obvious the LC components will be smaller.

O
Tl
V

V
dc

Figure 12-9. Single-Phase Full Wave Bridge.

Figure 12-10. Three-Phase, "Delta" Full Wave Bridge.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Multiphase Rectifier Circuits
Table 12-1 lists voltage and current ratios for the circuits, shown in Figures 12-11 through Figures 12-14
for inductive output filters. These ratios apply for sinusoidal ac input voltages. Values shown do not take
into consideration voltage drops which occur in the power transformer or rectifier diodes.

Table 12-1. Three-Phase Voltage and Current Ratios for Rectifier Circuits.

Three Phase Rectifier Circuit Data

Delta-Delta Full Wave, Figure 12-11
Item Factor
Primary VA 1.050 X dc watts output
Secondary VI leg 0.740 X average dc output voltage
Secondary I/leg 0.471 X average dc output current
Secondary VA 1.050 X dc watts output
Ripple Voltage % 4.200
Ripple Frequency 6f
Delta- Wye Full Wave, Figure 12-12
Item Factor
Primary VA 1.050 X dc watts output
Secondary Line to Line 0.740 X average dc output voltage
Secondary VI leg 0.428 to Neutral X average dc output voltage
Secondary I/leg 0.817 X average dc output current
Secondary VA 1.050 X dc watts output
Ripple Voltage % 4.200
Ripple Frequency 6f
Delta- Wye Half Wave, Figure 12-13
Item Factor
Primary VA 1.210 X dc watts output
Secondary Line to Line 0.740 X average dc output voltage
Secondary VI leg 0.855 to Neutral X average dc output voltage
Secondary I/leg 0.577 X average dc output current
Secondary VA 1.480 X dc watts output
Ripple Voltage % 18.000
Ripple Frequency 3f
Delta- Wye 6 Phase Half Wave, Figure 12-14
Item Factor
Primary VA 1.280 X dc watts output
Secondary Line to Line 1.480 X average dc output voltage
Secondary VI leg 0.740 to Neutral X average dc output voltage
Secondary I/leg 0.408 X average dc output current
Secondary VA 1.810 X dc watts output
Ripple Voltage % 4.200
Ripple Frequency 6f
Root mean square values to the average dc.
Sine-wave , infinite inductance, no transformer or rectifier losses.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 o

V,p(LL)

Figure 12-11. Three-Phase, "Delta-Delta" Full Wave Bridge.

Tl

XL)

Figure 12-12. Three-Phase, "Delta-Wye" Full Wave Circuit.

Tl

V,s(LL)

Figure 12-13. Three-Phase, "Delta-Wye" Half Wave Bridge.

Tl

Figure 12-14. Three-Phase, "Delta-Wye" Six Phase Star.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Area Product, Ap, and Core Geometry, Kg, for Three-Phase Transformers

The area product, Ap, of a three-phase core is defined differently than that for a single-phase core. The
window area, Wa, and iron area, Ac, for a single-phase transformer is shown in Figure 12-4 and 12-5. The
window area, Wa, and iron area, Ac, for a three-phase transformer is shown in Figure 12-6. The area
product, Ap, of a core is the product of the available window area, Wa, of the core in square centimeters
(cm2) multiplied by the effective, cross-section area, Ac, in square centimeters (cm2), which may be stated
as:

Single-phase: Ap = WaAc, [cm4] [12-3]

This is alright for a single-phase transformer. For three-phase transformers, because there are basically two
windows area, Wa, and three iron areas Ac, the window utilization is different, and the area product, Ap,
changes to:

fW }
Three-phase: Ap = 3 -^ Ac , [cm4 ] [12-4]

This reduces to:

Ap=l.5(WaAc), [cm 4 ] [12-5]

It is basically the same thing for the core geometry, Kg, for a single-phase transformer and the core
geometry, Kg, for a three-phase transformer. The core geometry, Kg, for a single-phase transformer is:

W A2 K
I "
MLT
|. [cm5] [12-6]

In the three-phase transformer, core geometry, Kg, is:

Three-phase : ^ = 3 f - [cm5] [12.7]

This reduces to:

*,=1.S
K
" c " ,
TL rr< '
[cm5
*-
MLT
jfT

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Output Power Versus Apparent Power, Pt, Capability

The apparent power, Pt, is described in detail in Chapter 7. The apparent power, Pt, of a transformer is the
combined power of the primary and secondary windings which handle, Pjn and P0, to the load, respectively.
Since the power transformer has to be designed to accommodate the primary, Pjn, and the secondary, P0:

P,=Pin+P0, [watts] [12-9]

Pin=^-, [watts] [12-10]

Substituting:

P,=— + P0, [watts] [12-11]

P = P - + 1 , [watts] [12-12]

The designer must be concerned with the apparent power handling capability, Pt, of the transformer core
and winding. The apparent power, Pt, varies with the type of circuit in which the transformer is used. If the
current in the rectifier is interrupted, its effective rms value changes. Transformer size is thus determined,
not only by the load demand, but also by current wave shape. An example of the primary and secondary,
VA, will be done to compare the power-handling capability required by each three-phase rectifier circuit in
Table 12-1 and Figures 12-1 1 through 12-14. This comparison will negate transformer and diode losses so
that P;,, = P0 (T) = 1) for all three-phase rectifier circuits.

1 . Delta-Delta, Full Wave, Figure 12-11.

.05 [watts] [12-13]

/ > = / > „ (2.1), [watts]

2. Delta- Wye, Full Wave, Figure 12-12.

=p^L +S [watts]

.05, [watts] [12-14]

P,=P0(2.i), [watts]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 3 . Delta-Wye, Half Wave, Figure 12-13.

= p ^ . +S [watts]

Pt=p — + 1.48 , [watts] [12-15]

Pt=P0(2.69), [watts]

4. Delta- Wye, 6 Phase Half Wave, Figure 12-14.

= p L +S [watts]

P=P\^ + 1.%1\, [watts] [12-16]

V 1 )
/?=P 0 (3.09), [watts]

Relationship, Kg, to Power Transformer Regulation Capability

Although most transformers are designed for a given temperature rise, they can also be designed for a given
regulation. The regulation and power-handling ability of a core are related to two constants:

a=—^—, [%1
ZA
*^' [12-17]

a = Regulation (%) [12-18]

The constant, Kg, is determined by the core geometry, which may be related by the following equations:

K=l.5\1 "-"'"- ! = -£-, [cm5]

MLT J 2Kea [12-19]

The constant, Ke, is determined by the magnetic and electric operating conditions, which may be related by
the following equation:

From the above, it can be seen that factors such as flux density, frequency of operation, and waveform
coefficient, have an influence on the transformer size.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Relationship, Ap, to Transformer Power Handling Capability

According to the newly developed approach, the power handling capability of a core is related to its area
product, Ap, by an equation which may be stated as:

Pl04 4
Ap= - , [cm
, [cm]
[\2-2\\

Ap=l.5(WaAc) =
^(1()4)
^ '—, [cm4]
KfKuBmJf [12_22]

Where:
Kf = waveform coefficient
Kf =4.0 square wave
Kf = 4.44 sine wave

From the above, it can be seen that factors such as flux density, frequency of operation, and window
utilization factor, Ku, define the maximum space, which may be occupied by the copper in the window.

Three-Phase Transformer, Design Example

The following information is the Design specification for a three-phase, isolation transformer, using the,
Kg, core geometry approach.

Design specification:

1. Input voltage, Vto 208 V, 3 Wire

2. Output voltage, V0 28 V
3. Output Current, I0 10 amps
4. Output Circuit Full Bridge
5. Input / Output Delta / Delta
6. Frequency, Three Phase, f 60 hertz
7. Efficiency, r|(100) 95 %
8. Regulation, a 5%
9. Flux Density, Bac 1.4 tesla
10. Magnetic Material Silicon M6X
11. Window Utilization Ku = (Kup + Kus) 0.4
12. Diode Drop, Vd 1.0 volt

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 1 Calculate the apparent power, Pt.

Pt=P0\— + l.05\, [warts]

[watts]

P = 3 0 0 | — + 1.051 [watts]
U-95 )
Pt = 647, [watts]

Step No. 2 Calculate the electrical conditions, Ke.

Ke = 2.86/ 2 5 2 (lO~ 4 )
^ e =2.86(60) 2 (l.4) 2 (lO- 4 )
Ke = 2.02

Step No. 3 Calculate the core geometry, Kg.

= -L—, [cm5]
2Kea
647

K g = 3 2 , [cm5]

Step No. 4 This data is taken from Chapter 3. The section is on, El, Three-Phase Laminations.

Core number 100EI-3P

Iron weight, Wtfe 2.751 kilograms
Mean length turn, MLT 16.7 cm
Iron area, Ac 6.129 cm"
Window area, Wa 29.0 cm
Area product, Ap 267cm4
Core geometry, Kg 39 cm
Surface area, A, 730 cm2

Step No. 5 Calculate the number of primary turns, Np, using Faraday's Law.

208(l0 4 )
f
4.44(1.4)(6.129)(60)'
/ p =910, [turns]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 6 Calculate the primary line current, I^UM)-

p
I p(Lme)
a »= -.„ ° ,' [amps]
L f J
J
>(£«!«)'/

300

[amps]
'
Step No. 7 Calculate the primary phase current, Ip(phase)-

0.506

Step No. 8 Calculate the primary bare wire area, AWp(B). The window area available for the primary is,
Wa / 4. The primary window utilization, Kup = 0.2.

4 910
I ( ) /
= 0.00159, [cm2]

Step No. 9 The selection of the wire would be from the Wire Table in Chapter 4.

AWG#25
Aw^= 0.001623, [cm2]
A
w(ins) = 0.002002, [cm2]

^ = 1062
cm

Step No. 10 Calculate the primary winding resistance. Use the MLT, from Step 4, and the micro-ohm,
per centimeter, found in Step 9.

RP=MLT(NP) ^ (ur6), [ohms]

v cm J
Rp =(!6.7)(910)(l062)(lO" 6 ), [ohms]
Rp=\6.\, [ohms]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 11 Calculate the total primary copper loss, Pp.

p> [watts]

Pp =3(0.292)2(16.1), [watts]
^=4.12, [watts]

Step No. 12 Calculate the secondary turns, Ns.

, [turns]
s
vp i o o ,'
Vs = (0.740)(F0 + 2Vd } = (0.740)(28 + 2) = 22.2
(910)(22.2)f 5 "|
Ns = V , A x ' 1 + — . [turns]
(208) i, IOOJ
Ns=\02, [turns]

Step No. 13 Calculate the secondary bare wire area, AWS(B).

.((0.2X29.0))
4»(B)-[ 4(102) J'

Step No. 14 The selection of the wire will be from the Wire Table in Chapter 4.

AWG#\6
4,^=0.01307, [cm2]
><»</«) =0.01473, [cm2]

^ = 132
cm

Step No. 15 Calculate the secondary winding resistance, Rs. Use the MLT, from Step 4, and the micro-
ohm per centimeter, found in Step 14.

/?, = MLT(A^) —- (lO~ 6 ), [ohms]

#S=(16.7)(102)(132)(10-6), [ohms]
= 0.225, [ohms]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 16 Calculate the secondary line current, IS(ime).

',(««) =(°- 471 K>

/,(/»») =(0.471)(10), [amps]

Step No. 17 Calculate the secondary phase current, Is(phase)-

s(line)
> LamPsJ

/ _4.71
i p rise j ,^ '
/
,(pha,e)=2.72,

Step No. 18 Calculate the total secondary copper loss, Ps.

J, [watts]

Ps =3(2.72) 2 (0.225), [watts]

Ps = 4.99, [watts]

Step No. 19 Calculate the transformer regulation, a.

a =^2- (100), [%]

'o

Pcu=Pp+Ps, [watts]
/>„, =4.12+4.99, [watts]
^=9.11, [watts]

a = 3.03, [%]

Step No. 20 Calculate the watts per kilogram.

Watts/kilogram = Kf(m]B%}
Watts/kilogram = 0.000557(60)° 68) (l.40)° 86)
Watts/kilogram = 1 .01

Step No. 21 Calculate the core loss, Pfe. Core weight, Wtfe, is found in Step 4.

pfe = Watts/Kilogram ( Wtfe ) , [watts]

Pfe= 1.01 (2.751), [watts]
Pfe = 2.78, [watts]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 22 Summarize the total transformer losses, P2.

Pz=Pp+Ps+Pfe, [watts]
Pz =4.12 + 4.99 + 2.78, [watts]
P z =11.89, [watts]

Step No. 23 Calculate the transformer efficiency, r|.

I = (100) [%]
? 77F
r L 0
'
+1<

-300- , Q
300 + 11.89V ;
ri = 96.2, [%]

Step No. 24 Calculate the watts per unit area, \\i. The surface area, At, is found in Step 4.

p
y/ = — , [watts per cm2 ]
4
1L89
i// = (-^- )
r^ , r [watts per cm
2,
(730)
(// = 0.0163, [watts per cm2 ]

Step No. 25 Calculate the temperature rise, Tr. The watts per unit area y is found in Step 24.

7;=450(0.0163)° 826 , [°C]

T; =15, [°C]

Step No. 26 Calculate the total window utilization, Ku. The window area is found in Step 4.

wtt wa
4(910)(0.001623) 4(l02)(0.01307)
A".. — -- 1 --
29 29
Ku =(0.204) + (0.184)
K,, = 0.388

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Chapter 13

Flyback Converter, Transformer Design

The author would like to thank Dr. V. Vorperian, Senior Engineer, Power and Sensor Electronics Group,
Jet Propulsion Laboratory (JPL), Richard Ozenbaugh of Linear Magnetics and Kit Sum, Senior
Consultant, for their help with the Flyback design equations.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table of Contents

1. Introduction
2. Energy Transfer
3. Discontinuous Current Mode
4. Continuous Current Mode
5. Continuous and Discontinuous Boundary
6. The Buck Converter
7. Discontinuous Current Buck Converter Design Equations
8. Continuous Current Buck Converter Design Equations
9. The Boost Converter
10. Discontinuous Current Boost Converter Design Equations
11. Continuous Current Boost Converter Design Equations
12. The Inverting Buck-Boost Converter
13. Discontinuous Current Inverting, Buck-Boost Design Equations
14. Continuous Current Inverting, Buck-Boost Design Equations
15. The Isolated Buck-Boost Converter
16. Discontinuous Current Isolated, Buck-Boost Design Equations
17. Continuous Current Isolated, Buck-Boost Design Equations
18. Design Example, Buck-Boost Isolated Converter Discontinuous Current
19. Design Example, Boost Converter, Discontinuous Current
20. Designing Boost Inductors for Power Factor Correction (PFC)
21. Standard Boost Flyback Converter
22. Boost PFC Converter
23. Design Example, (PFC) Boost Converter, Continuous Current
24. References

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Introduction
The principle behind Flyback converters is based on the storage of energy in the inductor during the
charging, or the "on period," ton, and the discharge of the energy to the load during the "off period," toff.
There are four basic types that are the most common, energy storage, inductor type converter circuits.

1. Step down, or buck converter.

2. Step up, or boost converter.
3. Inverting, buck-boost converter.
4. Isolated, buck-boost converter.

Energy Transfer

Two distinct modes of operation are possible for the Flyback switching converters, shown in Figure 13-1:
Discontinuous Mode All energy stored in the inductor is transferred to an output capacitor and load circuit
before another charging period occurs. This topology results in a smaller inductor size, but puts a larger
stress on the capacitor and switching device.
Continuous Mode Energy stored in the inductor is not completely transferred to the output capacitor and
load circuit before another charging period occurs.

The total period is:

T = - [13-1]

Discontinuous Current Continuous Current

\A
IL
rl
l
l w 'on toff
on
T T

Figure 13-1. Comparing Discontinuous and Continuous Current Waveforms.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Discontinuous Current Mode

In the Discontinuous Mode, a smaller inductance is required, but the penalty results in higher peak currents
in the switching transistor. As a consequence, the winding losses are increased because of the higher rms
values, due to the higher peak currents. This also results, in a higher ripple current and ripple voltage in the
input and output capacitor, and gives added stress to the switching transistor. The advantage of this circuit,
other than having a smaller inductor, is that when the switching device is turned on, the initial current is
zero. This means the output diode has completely recovered, and the switching device does not
momentarily turn on into a short. This diode recovery reduces the EMI radiation. The discontinuous mode
converter does not exhibit the right half plane zero. Without the right half plane zero, the loop is easy to
stabilize.

Continuous Current Mode

In the Continuous Mode, a larger inductor is required; this results in a lower peak current at the end of the
cycle than in a discontinuous system of equivalent output power. The Continuous Mode demands a high
current flowing through the switch during turn-on, and can lead to high switch dissipation. The continuous
mode converter does exhibit the right half-plane zero. With the right half-plane zero, the loop becomes
very difficult to stabilize for a wide range of input voltage. The relationship between the B-H loops for
continuous and discontinuous operation is shown in Figure 13-2.

B (tesla) B (tesla)
All

H (oersteds) H (oersteds)

Bac = AB/2

-- A B

Figure 13-2. Continuous (A) and Discontinuous (B), B-H Loops, Showing AB and AI.

Continuous and Discontinuous Boundary

When the load current increases, the control circuit causes the transistor to increase the "on time," ton. The
peak current in the inductor will increase, resulting in a steady reduction in the dwell time, tw. When the
load current increases to a critical level, tw becomes zero, and the discontinuous boundary is reached. If the

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 load current is further increased, the inductor current will no longer discharge to zero on every cycle, and
continuous current operation results.
The Buck Converter
The Buck Converter is shown in Figure 13-3. The output voltage of this converter is always less than the
input voltage. In the buck circuit, the transistor switch, Ql, is placed in series with the dc input voltage.
The transistor, Ql, interrupts the dc input voltage, providing a variable-width pulse, (duty ratio), to a simple
averaging, LC, filter. When the transistor switch, Ql, is closed, the dc input voltage is applied across the
output filter inductor, LI, and the current flows through the inductor to the load. When the switch is open,
the energy, stored in the field of the inductor, LI, maintains the current through the load. The
discontinuous voltage and current waveforms are shown in Figure 13-4, and the continuous waveforms in
Figure 13-5.

1 Cl J
Ql-ULL LI

C2
,__,
t

^ CR1 V
V-
v
m >• 7" S o

ICK,
n , n

Figure 13-3. Schematic of a Buck Switching Converter.

Discontinuous Current Buck Converter Design Equations

V rR i Discontinuous Current

- J-v,

Figure 13-4. Discontinuous, Current Buck Converter Waveforms.

Inductance, L:

•" o(max)

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Maximum duty ratio:

D,(max) Y [B-3]
yl in(min) J

Maximum on time:

Maximum off time:

[13-5]

The inductor peak current, I(pk):

[13-6]

Continuous Current Buck Converter Design Equations

VC,R1, Continuous Current

Figure 13-5. Continuous Current Buck Converter Waveforms.

Inductance, L:

L=V0T(\-Dmm)
[henrys] [13-7]
21.o(min)

Maximum duty ratio:

V
(max) [13-8]
yl in(mm) J

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Minimum duty ratio:

[13-9]

Maximum on time:

[13-10]

Maximum off time:

The inductor delta current, AI:

^"WU.K'-'W [1W2]

The inductor peak current,

A/
2

The Boost Converter

The Boost Converter is shown in Figure 13-6. The output voltage of this converter is always greater than
the input voltage. The boost converter stores energy in the inductor, LI, and then, delivers the stored
energy along with the energy from the dc source to the load. When the transistor switch, Ql, is closed,
current flows through inductor, LI, and the transistor switch, Ql, charging inductor, LI, but does not
deliver any current to the load. When the switch is open, the voltage across the load equals the dc input
voltage plus the energy stored in inductor, LI. The energy is stored in, LI, then discharges, delivering
current to the load. The discontinuous voltage and current waveforms are shown in Figure 13-7, and the
continuous waveforms in Figure 13-8.

w n 'i U -t-

Cl
LI
CR1 t
4 » C2
",„
J Ql ^ P- V
»

n i
1
n

Figure 13-6. Schematic of a Boost Switching Converter.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Discontinuous Current Boost Converter Design Equations

v Discontinuous Current
Vm

-t
-*--(Vvm)
Ipk
I0
t
l
l w
on
T

Figure 13-7. Discontinuous, Current Boost Converter Waveforms.

Inductance, L:

Anax -, [henrys] [13-14]

21o(max)

Maximum duty ratio:

y -V +V

Minimum duty ratio:

v -v.

Maximum on time:

[13-17]

Maximum off time:

[13-18]

The inductor peak current, I(pk):

IP (max)
7
(/>*) ~" ° [13-19]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Continuous Current Boost Converter Design Equations

V Continuous Current

-(V0-vm)

Figure 13-8. Continuous, Current Boost Converter Waveforms.

Inductance, L:

.
L= , [henrys] [13-20]

Maximum duty ratio:

A(max) [13-21]

Minimum duty ratio:

A_,_>=1- [13-22]

Maximum on time:

[13-23]

Maximum off time:

[13-24]

The inductor delta current, AI:

[13_25]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The inductor peak current, I(pk):

[13-26]
!-£>,max)

The Inverting Buck-Boost Converter

The inverting buck-boost converter is shown in Figure 13-9. It is a variation of the boost circuit. The
inverting converter delivers only the energy stored by the inductor, LI, to the load. The output voltage of
the inverting converter can be greater, or less than, the input voltage. When the transistor switch, Ql, is
closed, the inductor is storing energy, but no current is delivered to the load because diode, CRl, is back-
biased. When the transistor switch, Ql, is open, the blocking diode is forward-biased and the energy stored
in inductor, LI, is transferred to the load. The discontinuous voltage and current waveforms are shown in
Figure 13-10 and the continuous waveforms in Figure 13-11.

J O-
i Qi-LLL
V
"
• 1 | CRl
Cl 1 H
?= s C LI v. -s ^
vin
L C2

\
o
1 +

Figure 13-9. Schematic of an Inverting Buck-Boost Switching Converter.

Discontinuous Current Inverting, Buck-Boost Design Equations

V Discontinuous Current

Figure 13-10. Discontinuous, Current Inverting, Buck-Boost Converter Waveforms.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Inductance, L:

[B-27]
21o(max)

Maximum duty ratio:

^4 [13-28]
vo +v,+v.,
a .
/fl(min)

Minimum duty ratio:

D . =
vo + V,it + V.zw(max)
, ,

Maximum on time:

[13-30]

Maximum off time:

[13-31]

The inductor peak current, I^:

\ (max) in (min) '/J

Continuous Current Inverting, Buck-Boost Design Equations

V Continuous Current

Figure 13-11. Continuous, Current Inverting, Buck-Boost Converter Waveforms.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Inductance, L:

2^o(mm)

Maximum duty ratio:

V,
™x v , __ - [13-34]

Minimum duty ratio:

V
min / " \ [13-35]
K.+l^™*))

Maximum on time:

C(max) = TDm^ [13-36]

Maximum off time:

The inductor delta current, AI:

17T.,_,...,,A.
[13-38]
L

The inductor peak current, 1^:

V)

The Isolated Buck-Boost Converter

The Isolated Buck-Boost Converter is shown in Figure 13-12. This converter can provide line isolation,
and also has the capability of multiple outputs, which require only a diode and a capacitor; the filter
inductor is built-in. The isolated buck-boost converter is quite popular in low power applications because
of simplicity and low cost. This converter does not lend itself to the VDE specification because of the
required voltage insulation between primary and secondary. Care must be taken because this leakage
inductance could generate high voltage spikes on the primary. The discontinuous voltage and current
waveforms are shown in Figure 13-13, and the continuous waveforms in Figure 13-14.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 1 1| n w
C CR1 t
Cl £NS
C2 +
vm
(_j

7= s
v,

1
i O

Figure 13-12. Schematic of an Isolated, Buck-Boost Switching Converter.

Discontinuous Current Isolated, Buck-Boost Design Equations

Discontinuous Current

1 V-
v
in + V
v
<
- Vm
!
P
vc
/
o+vd

'im^s'

!
- s(pk)

-t w
on toff

Figure 13-13. Discontinuous Current, Isolated, Buck-Boost Converter Waveforms.

Primary inductance, Lp(maX):

(R , ,\T(D
\ m(eqmv.) J \ m
-, [henrys] [13-40]

Maximum on time:

-x)

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Maximum off time:

[13-42]

Total output power:

Maximum input power:

[13-44]

Equivalent input resistance:

(V. V
I iVi(min) I
R~in(equv.} [13-45]

The primary peak current, Ip(pk):

2P.f T
/..,..-.= "" [13-46]
ofl(max) /rt(min)

Continuous Current Isolated, Buck-Boost Design Equations

Qi Continuous Current
_ V-
v
m + vV
— — V in

rrn: !p(pk)

s
-^~ !s(pk)

•off
T

Figure 13-14. Continuous Current, Buck-Boost Converter Waveforms.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Inductance, L:

, D, . , } T
L=^-~—' ' "—, [henrys] [13-47]
m(min)

Minimum duty ratio:

in(m-dx)

Maximum on time:

C(max) = ^Amx [13-49]

Maximum off time:

taff(^}=T(\-Dmm} [13-50]

Minimum output power:

Minimum input power:

P
°(mm)

The inductor delta current, AI:

A/ = ^ '"(m'n) ( ' m X > ' [13-53]

LJ

The inductor peak current, \^:

[13-54]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design Example, Buck-Boost Isolated Converter Discontinuous Current

1 Input voltage nominal, V^ ............................................................................ = 28 volts

2. Input voltage minimum, Vin(min) .................................................................... = 24 volts
3. Input voltage maximum, Vjn(max) ................................................................... = 32 volts
4. Output voltage, V0| ....................................................................................... = 5 volts
5. Output current, I01 ......................................................................................... = 2 amps
6. Output voltage, Vo2 ........................................................................................ = 12 volts
7. Output current, Io2 ........................................................................................ = 0.5 amps
8. *Window utilization, Ku ............................................................................... = 0.29
9. Frequency, f ................................................................................................. = 100 kHz
10. Converter efficiency, r| ................................................................................. = 90%
11. Maximum duty ratio, D(,mx) ........................................................................... = 0.5
12. Dwell time duty ratio, D(w) ........................................................................... =0.1
13. Regulation, a ................................................................................................ = 1.0%
14. Operating flux density, Bm ............................................................................ = 0.25 tesla
15. Diode voltage, Vd .......................................................................................... = 1.0 volt
*When operating at high frequencies, the engineer has to review the window utilization factor, Ku. When
using a small bobbin ferrite, the ratio of the bobbin winding area to the core window area is only about 0.6.
Operating at 100kHz and having to use a #26 wire, because of the skin effect, the ratio of the bare copper
area is 0.78. Therefore, the overall window utilization, Ku, is reduced. The core geometries, Kg, in Chapter
3 have been calculated with a window utilization, Ku, of 0.4. To return the design back to the norm, the
core geometry, Kg is to be multiplied by 1.35, and then, the current density, J, is calculated, using a window
utilization factor of 0.29. See Chapter 4.

Skin Effect
The skin effect on an inductor is the same as a transformer. In the normal dc inductor, the ac current (ac
flux), is much lower, and does not require the use of the same, maximum wire size. This is not the case in
the discontinuous, current type, flyback converter, where all of the flux is ac and without dc. In the
discontinuous, flyback design, the skin effect has to be treated just like a high frequency transformer.

There are times when the larger wire is just too difficult to wind. Large wire is not only hard to handle, but
it does not give the proper lay. It is easier to wind with bi-filar or quad-filar wire, with the equivalent
cross-section.

Select a wire so that the relationship between the ac resistance and the dc resistance is 1 :

[13-55]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The skin depth in centimeters is:
6.62
, [cm]
V7
I00,000
s = 0.0209, [cm]

Then, the wire diameter is:

Wire Diameter = 2 ( £ ), [cm]

Wire Diameter = 2(0.0209), [cm]
Wire Diameter = 0.0418, [cm]
Then, the bare wire area Aw is:

A,,, = - -, [cm-]

Aw= 0.00137, [cm 2 ]

From the Wire Table in Chapter 4, Number 26 has a bare wire area of 0.00128 centimeters. This will be
the minimum wire size used in this design. If the design requires more wire area to meet the specification,
then, the design will use a multifilar of #26. Listed Below are #27 and #28, just in case #26 requires too
much rounding off.

Wire AWG Bare Area Area Ins. Bare/Ins. (iQ/cm

#26 0.001280 0.001603 0.798 1345
#27 0.001021 0.001313 0.778 1687
#28 0.0008046 0.0010515 0.765 2142

Step No. 1 Calculate the total period, T.

T =— [seconds]

1
T = [seconds!
100000
T = 10 [//sec]

Step No. 2 Calculate the maximum transistor on time, ton.

ton = 7Dmax [//sec]

6
/ On =(l0xl0- )(0.5) [//sec]
ton=5.0 [//sec]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 3 Calculate the secondary load power, P0].

P0i=Ioi(roi+Yd) [watts]
P o l =(2)(5 + l) [watts]
P0i=l2 [warts]

Step No. 4 Calculate the secondary load power, Po2.

P 02 =(0.5)(12 + 1) [watts]
Po2 = 6.5 [watts]

Step No. 5 Calculate the total secondary load power, P0(maX).

= P
^(max) o\ + Po2 [watts]

P0( max) =12 + 6.5 [watts]

P, > = 1 8 . 5 [watts]

Step No. 6 Calculate the maximum input current, Iu,(max).

o(max)
[amps]

18.5

/;,,(max) = 0.856

Step No. 7 Calculate the primary peak current, I p(p k).

Ip(pk) = -^ [amPs PeakJ

^ /n(min)'on(max)

2(l8.5)(lOxlO" 6 )
laP(pk}
iok\= ; r [amps peak]
(0.9)(24)(5xlO- 6 )
; peak]

Step No. 8 Calculate the primary rms current, Ip(pk)_

l^r [amps]

[amps]
• ^3(10)
1.40 [amps]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 9 Calculate the maximum input power, Pjn(max)-

18.5 r
^(™x)=^ [Watfe]

^( m ax)=20.6 [watts]

Step No. 10 Calculate the equivalent input resistance, Rm(equiv)-

/n(min) j

m(max)

Step No. 1 1 Calculate the required primary inductance, L.

[henry]

(28)(lOxlO" 6 )(0.5) 2
L = ^ ^-2 y [henry]

L = 35 [uh]

Step No. 12 Calculate the energy-handling capability in watt-seconds, w-s.

[w.s]

(35xl0^ 6 )(3.43) 2
Energy = --- [w-s]

Energy = 0.000206 [w-s]

Step No. 13 Calculate the electrical conditions, Ke.

Ke =(0.145)(l8.5)(0.25) 2 xlO~
K =0.0000168

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 14 Calculate the core geometry, Ke. See the design specification, window utilization factor, Ku.
o

(Energy)2
[cm5]
Kea
(0.000206)2

Kg = 0.00253 [cm5]
A: g = 0.00253 (1.35), [cm5]
Kg = 0.00342, [cm5]

Step No. 15 Select, from Chapter 3, an EFD core comparable in core geometry, Kg.

Core number .................................................................................................... EFD-20

Manufacturer ................................................................................................... Philips
Material .......................................................................................................... 3C85
Magnetic path length, MPL ............................................................................. = 4.7 cm
Core weight, W tfe ........................................................................................... = 7.0 grams
Copper weight, Wtcu .................. .................................................................... = 6.8 grams
Mean length turn, MLT ................................................................................... = 3.80 cm
Iron area, Ac .................................................................................................... = 0.31 cm3
Window Area, Wa...........................................................................................=0.501 cm2
Area Product, Ap ............................................................................................. = 0.155 cm4
Core geometry, Kg .......................................................................................... = 0.00506 cm5
Surface area, At ............................................................................................... = 13.3 cm
Core Permeability ........................................................................................... = 2500
Winding Length, G ........................................................................................ = 1.54 cm

Step No. 16 Calculate the current density, J, using a window utilization, Ku = 0.29.

2 (Energy) (lO 4 )
J=
- - '

2(0.000206)(l04)
J = ~, ^ r^ -T , [amps/cm2 ]
(0.25)(0.155)(0.29)
J = 367, [amps/cm2]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 17 Calculate the primary wire area, Apw(B).

* arms

Step No. 18 Calculate the required number of primary strands, Snp.

"" #26 (bare area)

s = (0.00381)
np
(0.00128)
Snp = 2.97 use 3

Step No. 19 Calculate the number of primary turns, Np. Half of the available window is primary, Wap 12.
Using the number of strands, Snp, and the area for #26.

> [turns]
3(#26(BareArea))
Ar (0.29)(0.25) r
NDp = v f A -^, [turns]
3(0.00128)
Np =18.9 use 19, [turns]

Step No. 20 Calculate the required gap, lg.

Ccm]
r n

_
* (0.000035)
lg = 0.0384, [cm]

Step No. 21 Calculate the equivalent gap in mils.

mils = cm(393.7)
mils = (0.0384)(393.7)
mils = 15

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 22 Calculate the fringing flux factor, F.

/._1| (0-0384) (2(1.54)^1

V(J3T I, 0.0384 J
F = 1.30

Step No. 23 Calculate the new number of turns, Nnp, by inserting the fringing flux, F.

/ L
N = -*—.--, [turns]
""

I (0.0384)(0.000035)
A' = M-^- ?—^, [turns]
"" -8'
Nnp =16, [turns]

Step No. 24 Calculate the peak flux density, Bp^.

. [tes,a]

(0.0384)+
v ; ~^-
1, 2500
Bpk= 0.223, [tesla]

Step No. 25 Calculate the primary, the new fo.Q/cm.

S
np
1345
(new)//Q / cm = -

= 448

Step No. 26 Calculate the primary winding resistance, Rp.

6
Rp=MLT(N
v np)\^]xW- [ohms]
\cm )
Rp =(3.8)(l6)(448)xlO~ 6 [ohms]
Rp = 0.0272 [ohms]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 27 Calculate the primary copper loss, Pp.

Pp=l2pRp [watts]
^=(1.4) 2 (.0272) [watts]
Pp = 0.0533 [watts]

Step No. 28 Calculate the secondary turns, N sl .

— [turns]

16(V
•5i ' ','"' '" " -/ [turns]
(24)(0.5)
<fsi =3.2 use 3 [turns]

Step No. 29 Calculate the secondary peak current, I s i (p k).

2(2.0)
[amps]

Step No. 30 Calculate the secondary rms current, Is(rmS).

(1-0.5-0.1)

Step No. 31 Calculate the secondary wire area, A swl(B) .

3 65
'

= 0-00995 [cm2]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 32 Calculate the required number of secondary strands, S ns i.

wire ^
(0.00995)
(0.00128)
S M ,=7.8 use

Step No. 33 Calculate the, Si secondary,

1345
cm = -
8
COT = 168

Step No. 34 Calculate the winding resistance, R si .

[ohms]
cm
~6
/?sl = 3.8(3)(l68)xlO~ 6 [ohms]
Rs] =0.00192 [ohms]

Step No. 35 Calculate the secondary copper loss, P s l .

Ps] = f ^ R s l [watts]
Psl =(3.65) 2 (.00192) [watts]
Psi = 0.0256 [watts]

Step No. 36 Calculate the secondary turns, Ns2.

d
Ns2 = ——— r-^ ~ [turns]

16(l2 + l)(l-0.5-0.l)
N,7 = — . . .— [turns]
(24)(0.5)
Ns2 = 6.9 use 7 [turns]

Step No. 37 Calculate the secondary peak current, IS2(pk)-

2L
[amps]

2(0.5)

(1-0.5-0.1)

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 38 Calculate the secondary rms current, Is2(rms).

[amps]

3
/
S 2(™ i ) =0.913 [amps]

Step No. 39 Calculate the secondary wire area, Asw2(B).

_ *s2(rms) 2

_ 0.913 2
5w2( B) -j f-i-i L J
' 367
ASW2(B)= 0.00249 [cm2]

Step No. 40 Calculate the required number of secondary strands, Sns2.

wireA
__ (0.00249)
s2
" "(0.00128)
Sns2 = 1.95 use 2

Step No. 41 Calculate the, S2 secondary, uQ/cm.

(S2 )//Q / cm =

1345

Step No. 42 Calculate the winding resistance, Rs2.

Rs2 =MLT(Ns2]\— jxlO' 6 [ohms]

\cm)
Rs2 =3.8(7)(672)xlO~ 6 [ohms]
«J2 = 0.0179 [ohms]

Step No. 43 Calculate the secondary copper loss, Ps2.

p
S2 = ^2Rs2 [watts]
Ps2 =(0.913) 2 (.0179) [watts]
Ps2 =0.0149 [watts]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 44 Calculate the window utilization, Ku.

[turns] = (NpSnp) [primary]

[turns] =(16) (3) =48 [primary]
[turns] =(N^Snsl) [secondary]
[turns] = (3)(8)=24 [secondary]
[turns] =(Ns2Sns2) [secondary]
[turns] =(7)(2)=14 [secondary]
Nt = 86 turns #26

Wa (0.501)
K = 0.220

Step No. 45 Calculate the total copper loss, Pcu.

PC,, = PP + Psi+?s2 [watts]

Pcu =(0.0533) + (0.0256) + (0.0149) [watts]
Pcu = 0.0938 [watts]

Step No. 46 Calculate the regulation, a, for this design.

a =-5^x1 00 [%

(0.0938)
a=^. -- -^
(18.5)
=

Step No. 47 Calculate the ac flux density, Bac.

(0.0384)+
v ;

Bac=0.m, [tesla]

Step No. 48 Calculate the watts per kilogram, WK.

ff^ = 4.855(lO- 5 )(/) ( '- 63) (fi flf ) (2 - 62) [watts/kilogram]

WK = 4.855(lO"5 )(100000)(1 63) (0.1 1 1)(2'62) [watts/kilogram]

WK = 21.6 [watts/kilogram] or [milliwatts/gram]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 49 Calculate the core loss, Pfe.

pfe=\ I\milliwatts
1
gram
P /e =(21.6)(7)xlO- 3 [watts]
Pfe= 0.151 [watts]

Step No. 50 Calculate the total loss, core Pfe and copper Pcu, in watts,

Pcu [watts]

P2 =(0.151) + (0.0938) [watts]

Ps = 0.245 [watts]

Step No. 5 1 Calculate the watt density,

p
\l/ = — [watts/cm2 ]
4
0.245 r , ,n
i// = - [watts/cm"
13.3
i// = 0.01 84 [watts/cm2]

Step No. 52 Calculate the temperature rise, Tr, in, °C.

Tr= 450(0.01 84)(a826)

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Design Example, Boost Converter, Discontinuous Current

1 Input voltage nominal, Vin = 28 volts

2. Input voltage minimum, Vju^n) = 26 volts
3. Input voltage maximum, Vin(max) = 32 volts
4. Output voltage, V0, = 50 volts
5. Output current, I01 = 1 amps
6. *Window utilization, Ku = 0.29
7. Frequency, f = 100 kHz
8. Converter efficiency, r| = 92 %
9. Dwell time duty ratio, D (w) = 0.1
10. Regulation, a = 1.0%
11. Operating flux density, Bm = 0.25 tesla
12. Diode voltage, Vd = 1.0 volts

*When operating at high frequencies, the engineer has to review the window utilization factor, Ku. When
using a small bobbin ferrite, the ratio of the bobbin winding area to the core window area is only about 0.6.
Operating at 100kHz and having to use a #26 wire, because of the skin effect, the ratio of the bare copper
area is 0.78. Therefore, the overall window utilization, Ku, is reduced. The core geometries, KK, in Chapter
3 have been calculated with a window utilization, Ku, of 0.4. To return the design back to the norm, the
core geometry, Kg is to be multiplied by 1.35, and then, the current density, J, is calculated, using a window
utilization factor of 0.29. See Chapter 4.

Skin Effect
The skin effect on an inductor is the same as a transformer. In the normal dc inductor, the ac current, (ac
flux), is much lower and does not require the use of the same maximum wire size. This is not the case in
the discontinuous, current type, flyback converter, where all of the flux is ac and without dc. In the
discontinuous, flyback design, the skin effect has to be treated just like a high frequency transformer.

There are times when the larger wire is just too difficult to wind. Large wire is not only hard to handle, but
it does not give the proper lay. It is easier to wind with bi-filar or quad-filar, wire with the equivalent
cross-section.

At this point, select a wire so that the relationship between the ac resistance and the dc resistance is 1:

Rn, .

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The skin depth in centimeters is:

6 62
' , r[CHI]
i

/100,000
s = 0.0209, [cm]

Then, the wire diameter is:

Wire Diameter = 2(s), [cm]

Wire Diameter = 2(0.0209), [cm]
Wire Diameter = 0.0418, [cm]

Then, the bare wire area Aw is:

-, [cm2]

(3.1416X0.0418)

Aw =0.00137, [cm2]

From the Wire Table in Chapter 4, Number 26 has a bare wire area of 0.001028 centimeters. This will be
the minimum wire size used in this design. If the design requires more wire area to meet the specification,
then, the design will use a multifilar of #26. Listed Below are #27 and #28, just in case #26 requires too
much rounding off.

Wire AWG Bare Area Area Ins. Bare/Ins. ufi/cm

#26 0.001280 0.001603 0.798 1345
#27 0.001021 0.001313 0.778 1687
#28 0.0008046 0.0010515 0.765 2142

Step No. 1 Calculate the total period, T.

T = —, [seconds]

1
T =- , [seconds]
100,000
= 10, [//sec]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 2 Calculate the maximum output power, P0.

P 0 =(50 + 1.0)(1.0), [watts]

P0=5l, [watts]

Step No. 3 Calculate the maximum input current, Ijn(max).

(51)

/
/ n (max)= 2 - 1 3 '

Step No. 4 Calculate the maximum duty ratio, D(max).

(V -V. : . > + V,
„

Step No. 5 Calculate the minimum duty ratio, D(mill).

( v - V.
D
^ -(\-
-

'(50)-
1}
(
£>(mjn) = 0.342

Step No. 6 Calculate the required inductance, L.

1.0)(lo(l(r 6 )Vo.45)(l-0.45--0.l) 2
Lmax = —— , [henrys]

, = 23.2 use 23, [uh]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 7 Calculate the peak current, l^. In a discontinuous current boost the peak current is, I^y = AI.

T
IPo(max) r -i
= [ a m p s ]
* ~ —
2(51)
(pk)
, [amps]
(0.92)((50)(0.342))
Vr6.48,
Step No. 8 Calculate the rms current, I(nns).

-, [amps]

Step No. 9 Calculate the total energy-handling capability in watt-seconds, w-s.

LI
lk
Energy = ——, [w-s]

(23.rlO~ 6 )(6.48) 2
Energy = , [w-s]

Energy = 0.000483, [w-s]

Step No. 10 Calculate the electrical conditions, Ke.

Ke = 0.145(5l)(0.25) 2 (lO~ 4 )
Ka = 0.0000462

Step No. 11 Calculate the core geometry, Kg.

.
(0.000483)2 5
s
~ (0.0000462)(1.0)'
^=0.00505, [cm5]
Kg =0.00505(1.35), [cm5]
AT. = 0.00682, [cm5]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 12 From Chapter 3, select a core that is comparable in core geometry, Kg.

Core number RM-6

Manufacturer TDK
Magnetic path length, MPL = 2.86 cm
Core weight, Wtfe = 5.5 grams
Copper weight, W,cu = 2.9 grams
Mean length turn, MLT =3.1 cm
Iron area, Ac =0.366 cm
Window Area, Wa = 0.260 cm2
Area Product, Ap = 0.0953 cm4
Core geometry, Kg = 0.0044 cm5
Surface area, At = 11.3 cm
Permeability, |am = 2500
Winding Length, G = 0.82

Step No. 13 Calculate the current density, J, using a window utilization, Ku = 0.29.

2(Energy)(l04}
J = , [amps/cm ]

2(0.000483)(l0 4 )
J = ~, rr TT——?, [amps/cm ]
(0.25)(0.0953)(0.29)
J = 1398, [amps/cm2]

Step No. 14 Calculate the wire area, A W(B) .

~ [cm']
2.51 2
|cm 1
HW 1398
4,B)= 0.00179 [cm2]

Step No. 15 Calculate the required number of strands, Sn.

S. =
#26 (bare area)

_ v(0.00180)/
" (0.00128)
» =1.41 use 2

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 16 Calculate the number of turns, N, using the number of strands, Sn, and the area for #26.

K W
N = —*-*-, [turns]
5 n #26
(0.29)(0.26)
N = ^ . A-J-, [turns]
2(0.00128)
yV = 29.5 use 30, [turns]

Step No. 17 Calculate the required gap, lg.

[cm]
(1.26)(3Q) 2 (0.366)(10 8 ) M. .
/8 = -;-—---- , f[cm]
(0.000023) V2500J
^=0.179, [cm]

Step No. 19 Calculate the equivalent gap in mils.

mils = cm(393.7)
mils = (0.179)(393.7)
mils = 70

Step No. 20 Calculate the fringing flux factor, F.

V0.366 0.179
F=1.66

Step No. 21 Calculate the new number of turns, Nn, by inserting the fringing flux, F.

—i*
,V - 0.4*
(0.179)(O.QQ0023)
"" ^j(l.26)(0.366)(l.66)(lO~ 8 ) '
TV =23, [turns]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 22 Calculate the peak flux density, Bpk.

QAxN
[teshl
•
(1.26)(23)(1.66)(6.48)(10-4)
Bnt = -7-^--, [tesla]
V(0.179)
' ^2500j
Bpk=0.m, [tesla]

Step No. 23 Calculate the new, uQ/cm.

/.£l/cm
(new)/A2 / cm =

1345
(new)//Q / cm =

(new)//Q / cm = 673

Step No. 24 Calculate the primary winding resistance, R.

xlO~ 6 [ohms]
cm j
# = (3.l)(23)(673)xlO" 6 [ohms]
R = 0.0480 [ohms]

Step No. 25 Calculate the copper loss, P cu .

Pcu =(2.51) 2 (.0480) [watts]

Pcu = 0.302 [watts]

Step No. 26 Calculate the regulation, a, for this design.

(0.302)
a=
(50)
a = 0.604, [%]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 27 Calculate the ac flux density in tesla, Bac.

[tesla]
8
Am

(l.26)(23)(l.66)(3.24)(lO- 4 )
5 = [tesla]
- - ^T^Y- '
2 6
v( O . l T ^ f ^ )
' UsooJ
flac = 0.0869, [tesla]

Step No. 28 Calculate the watts per kilogram, WK.

WK = 4.855(lO~ 5 )(/) (1 - 63) (5 fl( ,) (2 - 62) , [watts/kilogram]

WK = 4.855(lO~ 5 )(100000/L63) (0.0869/2 62) , [watts/kilogram]

W.K = 11.39, [watts/kilogram] or [milliwatts/gram]

Step No. 29 Calculate the core loss, Pfe .

( milliwatts | _^
Pfe=\ - K ^ I O 3 , [Watts]
V gram )
Pfe =(ll.39)(5.5)xl0^ 3 , [watts]
Pfe = 0.0626. [watts]

Step No. 30 Calculate the total loss, Pv, core, Pfe, and copper, Pcu.

Pz =(0.0626) + (0.302), [watts]

P± = 0.365, [watts]

Step No. 31 Calculate the watt density, \\i.

p
ij/ = —, [watts/cm2 ]
A
t
0.365 r , 2n
w= , I watts/cm ]
11.3
y/ = 0.0323, [watts/cm2]

Step No. 32 Calculate the temperature rise, Tr in, °C.

r r =450(0.0323) (0 ' 826) ,

Tr = 26.4, [°C]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Designing Boost Inductors for
Power Factor Correction (PFC)

Historically, the standard power supplies designed for electronic equipment have had a notoriously poor
power factor in the area of (0.5-0.6), and a correspondingly, high, harmonic current content. This design
approach utilizes a simple rectifier capacitor input filter that results in large current pulses drawn from the
line, that cause distorting of the line voltage and create large amounts of EMI and noise.

The regulating bodies, IEC in Europe and IEEE in the United States, have been working to develop a
standard for limiting harmonic current, in off-line equipment. The German standardization bodies have
established IEC 1000-2, and it is generally accepted as the standard for limiting harmonic currents in off-
line equipment.

Many new electronic products are required to have a near unity power factor and a distortion free, current
input waveform. The conventional ac-dc converters usually employ a full wave, rectifier-bridge, with a
simple filter to draw power from the ac line. The typical, rectifier capacitor, input bridge filter and
associated waveforms, as shown in Figure 13-15, are no longer good enough.

in I
CR1 CR2 Rectified Line Voltage

O
n

V rx
f^R T >
/
CR4 +
f- ~>
Cl
i
v
v
o
CR2 and CR3 Current

CR4 and CR1 Current

\
Line Current
O

Figure 13-15. Typical, Capacitor Input Bridge Rectifier Filter.

The line current waveform for equipment that utilizes off-line rectifier capacitor input filter, is shown in
Figure 13-15. The line current is supplied in narrow pulses. Consequently, the power factor is poor (0.5 -
0.6), due to a high harmonic distortion of the current waveform. The power supply can be designed with a
power factor approaching unity, by the addition of an input inductor, as shown in Figure 13-16. The
reasons why the input inductors are not designed into power supplies is very simple: cost, weight and bulk.
The inductance equation for, LI, is shown below.

Ll=- -, [henrys] [13-56]

o(min)

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Rectified Line Voltage

CR2 and CR3 Current

CR4 and CR1 Current

Line Current

Figure 13-16. A Typical, Inductor Input Bridge Rectifier Filter.

Standard Boost Flyback Converter

The standard dc-to-dc boost flyback converter is shown in Figure 13-6, along with the voltage and current
waveforms, shown in Figure 13-7 and 13-8. The boost converter has become the choice of many engineers
as the power stage in the active power factor corrector design. The basic circuit can be operated in either
the continuous or discontinuous mode.

Boost PFC Converter

The boost power factor correction converter is shown in Figure 13-17. The boost converter is the most
popular of the power factor pre-regulators. The boost converter can operate in two modes, continuous and
discontinuous. The current through the inductor, LI, is shown in Figure 13-18, for both continuous and
discontinuous operation. After examining the schematic, the advantages and disadvantages of the boost
converter can readily be seen. The disadvantage is the high output voltage to the load circuit and current
limit cannot be implemented. The advantage is that the circuit requires a minimum of parts and the gate
drive to, Ql, is referenced to ground.

Downstream
dc/dc
Converter
O

Figure 13-17. Boost PFC Converter.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Ipk

t o
Continuous Discontinuous

Figure 13-18. Current Through Inductor LI.

Design Example, (PFC) Boost Converter, Continuous Current

The following pages describe a step-by-step procedure for designing a continuous current boost inductor
for a Power Factor Correction (PFC) converter, as shown in Figure 13-17, with the following
specifications:

1. Output power, P0 = 250 watts

2. Input voltage range, Vin = 90 - 270 volts
3. Line frequency, f ( i ine) - 47 - 65 Hz
4. Output voIiage v V0 = 400 volts
5. Switching frequency, f = 100kHz
6. Inductor ripple current, AI = 20% of l^
1. Magnetic core = ETD
8. Magnetic material =R
9. Converter efficiency, r) = 95%
10. Inductor regulation, a = 1%
11. *Window utilization, Ku = 0.29
12. Operating Flux, Bm = 0.25 tesla

*When operating at high frequencies, the engineer has to review the window utilization factor, Ku. When
using a small bobbin ferrite, the ratio of the bobbin winding area to the core window area is only about 0.6.
Operating at 100kHz and having to use a #26 wire, because of the skin effect, the ratio of the bare copper
area is 0.78. Therefore, the overall window utilization, Ku, is reduced. The core geometries, Kg, in Chapter
3 have been calculated with a window utilization, Ku, of 0.4. To return the design back to the norm, the
core geometry, Kg, is to be multiplied by 1.35, and then, the current density, J, is calculated, using a
window utilization factor of 0.29. See Chapter 4.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Skin Effect
The skin effect on an inductor is the same as a transformer. In the normal dc inductor, the ac current (ac
flux), is much lower, and does not require the use of the same, maximum wire size. This is not the case in
the discontinuous, current type, flyback converter, where all of the flux is ac and no dc. In the
discontinuous, flyback design, the skin effect has to be treated just like a high frequency transformer.

There are times when the larger wire is just too difficult to wind. Large wire is not only hard to handle, but
it does not give the proper lay. It is easier to wind with bi-filar or quad-filar wire, with the equivalent
cross-section.

Select a wire so that the relationship between the ac resistance and the dc resistance is 1 :

The skin depth in centimeters is:

6.62
s= ,— , [cmj
v/
ioo,ooo
£ = 00209, [cm]

Then, the wire diameter is:

Wire Diameter = 2 (.?), [cm]

Wire Diameter = 2 (0.0209), [cm]
Wire Diameter = 0.041 8, [cm]
Then, the bare wire area, Aw, is:

(3.1416)(0.0418)2 ,
4,=---4-L, [cm2]

4, =0.00137, [cm2]

From the Wire Table in Chapter 4, Number 26 has a bare wire area of 0.00128 centimeters. This will be
the minimum wire size used in this design. If the design requires more wire area to meet the specification,
then, the design will use a multifilar of #26.

Wire AWG Bare Area Area Ins. Bare/Ins.

#26 0.001280 0.001603 0.798 1345

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 1 Calculate the input power, P^.

Pin=^, [watts]
i]
P = .250 L[watts]J
'" 0.95
Pla = 263, [watts]

Step No. 2 Calculate the peak input current, lp^.

^ = ^ > [amps]
in(min)

(263)(1.41)
1*=-
pt -9Q---.

/M=4.12, [amps]

Step No. 3 Calculate the input ripple current, AI.

A/ = 0.2/ M , [amps]
A/ = 0.2(4.12), [amps]
A/ = 0.824, [amps]

Step No. 4 Calculate the maximum duty ratio, D (max) .

(max)

400 -(90V2))

400
= 0-683

Step No. 5 Calculate the required boost inductance, L.

1
L-- --- , y[henrys]
L J
A//
(126.9)(0.683)
- --> [henrys]
(0.824)(100000)
1 = 0.00105, [henrys]

Step No. 6 Calculate the Energy required, Eng.

LI
Eng = —— , [watt-seconds]
2
(0.00105)(4.12)2
, [watt-seconds]

Eng = 0.00891, [watt-seconds]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 7 Calculate the electrical coefficient, Ke.

K e =0.145(250)(0.25) 2 (l(r 4 )
K = 0.000227

Step No. 8 Calculate the core geometry coefficient, Kg.

(0.00891)2
g
~(0.000227)(1)'
£ ? =0.35, [cm5]
Kg = 0.35(1.35), [cm5] Corrected
Kg=OA7, [cm5]

Step No. 9 From Chapter 3, select an ETD ferrite core, comparable in core geometry, Kg.

Core number = ETD-44

Manufacturer = Ferroxcube
Magnetic path length, MPL = 10.3 cm
Core weight, W t f e = 93.2 grams
Copper weight, Wtcu = 94 grams
Mean length turn, MLT = 9.4 cm
Iron area, Ac = 1.74 cm2
Window Area, Wa =2.79 cm"
Area Product, Ap = 4.85 cm
Core geometry, K B = 0.360 cm5
Surface area, A, = 87.9 cm2
Permeability, um = 2000
Millihenrys per 1000 turns, AL = 3365
Winding Length, G = 3.22

Step No. 10 Calculate the current density, J.

2(Eng)(l0 4 )
J = , [amps/cm2 ]

2(0.0089l)(l0 4 )
J=. ., £ '-, [amps/cm2]
(0.25)(4.85)(0.29)
J = 507, [amps/cm2 ]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 11 Calculate the rms current, Irms

/,.,„= ^, [amps]

lrms= [amps]
~JT'
Jrm=2.9l, [amps]

Step No. 12 Calculate the required bare wire area, AW(B).

2 91
AA »<«) ' , r[cm21]
,„> = 507

/4w(s) = 0.00574, [cm2]

Step No. 13 Calculate the required number of strands, Sn.

A {B ,
\o = ' Lr cm 2Jn
H
\.}

#26(bare area)
(0.00574)
S =± '-. [cm2]
(0.00128)
S n = 4 . 4 8 u s e 5 , [cm 2 ]

Step No. 14 Calculate the required number of turns, N, using the number of strands, Sn, and the area for #26.

W K
N= " " , [turns]
5,, #26
(2.79)(0.29)
N=±—-^ 2 [turns]
5(0.00128)
N = 126, [turns]

Step No. 15 Calculate the required gap, lg.

[cm]

'(1.257)(126)2(1.74)(10-8)
/ = , [cm]
0.00105
V
/ =0.331, [cm]

Change the gap to mils: 0.331 x 393.7 = 130 mils center or 65 mils per each outer leg.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 16 Calculate the fringing flux factor, F.

0.331V ( 6.44}
In
1.32 ) (033l)
F = 1.74

Step No. 17 Calculate the new turns using the fringing flux.

LL
N= -*—.-
-8 -, [turns]

(0.331)(0.00105)
N= -i-^--/--r , [turns]
V(l.257)(l.74)(l.74)(lO-")
N = 96, [turns]

Step No. 18 Calculate the peak flux, Bpk.

^
B. =F\ , [tesla]
' y
1f (1.257)(96)(4.12)(10 ')
, [tesla]
0.331
Bpt= 0.261, [tesla]

Step No. 19 Calculate the new, ud/cm.

/(Q/cm
(new)//i2 / cm =

(new)//fi / cm = —'•—

(new),uQ / cm =

Step No. 20 Calculate the winding resistance, R.

(lO~ 6 ), [ohms]
v
mj '
^ = (9.4)(96)(269)(l0^ 6 ), [ohms]
R = 0.243, [ohms]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 2 1 Calculate the winding copper loss, Pcu.

Pcu=(2.9l)2 (0.243), [watts]

Pcu = 2.06, [watts]

Step No. 22 Calculate the regulation, a.

a =^-100, [%]

(2.06)
a=\-flOO, [%]
(250)
a =0.824, [%]

Step No. 23 Calculate the ac flux density, Bac.

(I.257)(96)(0.412)(l0 4 )
B =- ^—^ ^ '-, [tesla]
0.331
fi,,r=0.0150, [tesla]

Step No. 24 Calculate the watts per kilogram, W/K, using R material in Chapter 2.

WIK = 4.316(lO' 5 )(/)' W (Bac)26S, [watts per kilogram]

W / K = 4.316(lO~ 5 )(lOOOOO)' M (0.0150) 268 , [watts per kilogram]
WIK = 0.0885, [watts per kilogram]

Step No. 25 Calculate the core loss, Pfe.

Pfe=Wtfe(\^}(WIK], [watts]
3
Pfe =(93.2)(lO~ )(0.0885), [watts]
Pfe = 0.0082, [watts]

Step No. 26 Calculate the total loss core loss, Pfe and copper loss, Pcu.

P = Pcu+Pfe, [watts]
)
/ = (2.03) + (0.0082), [watts]
P = 2.04, [watts]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 27 Calculate the watt density, vj/.

P , [watts per cm ]
\l/ = —
4
2.04 r 2n
w = - , [watts per cm I
87.9
if/ = 0.023, [watts per cm2 ]

Step No. 28 Calculate the temperature rise, Tr.

Tr =
Tr= 450(0.023)°
r r = 19.9, [°C]

Step No. 29 Calculate the window utilization, Ku.

(95)(5)(0.00128)
K =
" (2.79)
A:,, =0.218

References

1. Unitode Application Note U-132, Power Factor Correction Using TheUC3852 Controller on-time zero
current Switching Technique.

2. Unitode Application Note U-134, UC3854 Controlled Power Factor Correction Circuit Design

3. AlliedSignal Application Guide: Power Factor Correction Inductor Design for Switch Mode Power
Supplies using Powerlite C Cores.

4. PCIM August 1990, Active Power Factor Correction Using a Flyback Topology, James LoCascio and
Mehmet Nalbant/ Micor Linear Corporation.

5. Silicon General Application SG3561A Power Factor Controller.

6. SGS Thomson Application Note AN628/0593 Designing a High Power Factor Pre-regulator with the
L4981 Continuous Current.

7. IEEE, A Comparison Between Hysteretic and Fixed Frequency Boost Converter Used for Power
Factor Correction, James J. Spanger Motorola and Anup K. Behera Illinois Institute Technology.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Chapter 14

Forward Converter, Transformer Design,

and Output Inductor Design

The author would like to thank the late Dr. J. K. Watson, Professor of Electrical Engineering at the
University of Florida for his help with the Forward Converter design equations.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table of Contents

1. Introduction
2. Circuit Operation
3. Comparing the Dynamic B-H Loops
4. Forward Converter Waveforms
5. Transformer Design Using the Core Geometry, Kg, Approach
6. Forward Converter Output Inductor Design
7. Output Inductor Design Using the Core Geometry, Kg, Approach

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Introduction

When speaking of a forward converter, the circuit that comes to mind is the single-ended, forward
converter circuit, as shown in Figure 14-1. This single-ended, forward converter was developed about
1974 and has become one of the most popular and widely-used topology for powers under 200 W. The
single-ended, forward converter gets its name from a family of converters. A description of a forward
converter is that when current is flowing in the primary, there is current flowing in the secondary, and in
the load. The push-pull converter, full-bridge converter, and half-bridge converter are all, basically,
forward converters. The voltage stress on the single-forward converter is the same as it is on the push-pull
converter, 2Vjn. This circuit's main advantage, that is so appealing to engineers is its simplicity and parts'
count.

LI
CRl
^ r^v-v-^ — i0 +
j n
2 |— Is t
C2
4 "j CR3 +
•
Cl N2
V
c
Id
V
in 7*
^ "LJ3 1 t
1
t
JE L_ - n

-o -\

Figure 14-1. Schematic of a Single-Ended Forward Converter.

Circuit Operation

The basic circuit operation of this single-ended, forward converter is as follows: When the drive is applied
to, Ql, the secondary current, IS; will flow through, CR2, and, LI, and into the load. This process is due to
transformer action, (Tl). At the same time, the magnetizing current begins to build up in the transformer
primary. When the base drive to, Ql, is removed, then, Ql, turns off the magnetizing current that has built
up in the primary. The magnetizing current continues to flow through the demagnetizing winding, Nmag
and CRl. The demagnetizing winding, Nmag, has the same number of turns as the primary winding. So,
when the magnetizing field collapses, when Ql is turned off, diode CRl is clamped to the same voltage as
the applied voltage during the ton time. This means the transistor on time, ton, divided by the total time, T,
must not exceed 0.5 or 50%. Otherwise, the forward volt-seconds will exceed the reset volt-second

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 capability and the transformer will saturate. To ensure smooth transfer of the magnetizing current, the
primary and demagnetizing winding must be tightly coupled (bifilar). In a push-pull converter, the reset of
the core occurs naturally on each alternate half cycle.

Comparing the Dynamic B-H Loops

One of the main reasons for engineers to use the single-ended, forward converter circuit is the problem they
have with the push-pull converter core-saturating. The core saturation can be due to an imbalance of the
primary or secondary. The dynamic BH loops for the single-ended, forward converter and the push-pull
converter are shown in Figure 14-2.

B B
+R
D
sat, +B
sat
/f /f
1: If
I
i
AB
1' t AB
\ TT 1 \ \ \ TT
1 11 1

" Bac= AB/2 Bac= AB/2

i i

/ /
/
-B,,t _/ -Bsat _l/_

Single-Ended, Forward Converter Push-Pull Converter

Figure 14-2. The dynamic BH loop comparison.

The average input current for the single-ended, forward converter is about the same as the push-pull
converter, but the peak current is always greater than twice the average current. Operating the single-
ended, forward converter at low input voltages, the high peak currents could be a component problem. The
input filter and output filter for the single-ended, forward converter are always larger than the push-pull
converter, because it is operating at the fundamental frequency.

The waveforms shown in Figure 14-3 are typical waveforms of the single-ended forward converter. The
collector current, Ic, is shown in Figure (14-3-A), and the magnetizing current, Im, is shown in Figure (14-3-
B). The inductor, LI, current, IL, made up from the rectifier, CR2, and the commutating rectifier, CR3, are
shown in Figure (14-3-C). The collector voltage, Vc, is shown in Figure (14-3-D).

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Forward Converter Waveforms

Im
-CR1
(14-3-B)
0

(14-3-C) CR3
0

2V in
(14-3-D)
0

Figure 14-3. Typical Single-Ended Forward, Converter Waveforms.

Another version of the classic, forward converter is the double-ended, forward converter, shown in Figure
14-4. The double-ended, forward converter Las two transistors rather than one, compared to the single-
ended, forward converter, shown in Figure 14-1. The double-ended forward converter is more complicated
than the single-ended forward converter because one of the transistors is on the high side of the input
voltage, but it has some significant advantages. The series switching transistors are subjected to only the
input voltage, (Vin), rather than twice the input voltage, (2Vin). It also removes the need for a
demagnetizing winding. The demagnetizing current now flows through the primary, through, CR1, and
CR2, and back to the source, as shown in Figure 14-5. This demagnetizing path also provides a path for the
energy stored in the leakage inductance. The resulting spiking voltage, caused from the leakage
inductance, is now clamped to the input voltage, plus the two diode drops (V^ + 2Vd).

LI
CR3
CR1
Fl

C2
CR4 +
L H "^ V
v
N2
id o

t
1 n
J

Figure 14-4. Schematic of a Double-Ended Forward Converter.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 !Q1,Q2; Q1.Q2
j on! off

(14-5-A)
0
CR1, CR2
(14-5-B)
0
IL
(14-5-C) CR3 CR4
0

(14-5-D) V.in
0

Figure 14-5. Typical Double-Ended Forward, Converter Waveforms.

Note:
The design equations are from communication and work with the late Dr. J. K. Watson, Professor of
Electrical Engineering at the University of Florida.

Electrical coefficient is:

=0.145/ 2 A5 2 (l(T 4 ) [14-1]

Core geometry is:

[cm5]
'
Current density is:

Primary current is:

/„ =- ,, [amps] [14-4]
vin(mm)-\lID(max)

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Transformer Design Using the Core Geometry, Kg, Approach

The following information is the Design specification for a 30 watts, single-ended transformer, operating at
100kHz, using the, Kg, core geometry approach. For a typical design example, assume a single-ended
converter circuit, as shown in Figure 14-1, with the following specification:

1. Input voltage, V(m;n) = 22 volts

2. Input voltage, V(nom) = 28 volts
3. Input voltage, V(max) = 35 volts
4. Output voltage, V0 = 5.0 volts
5. Output current, I0 = 5.0 amps
6. Frequency, f = 100kHz
7. Efficiency, r) = 98%
8. Regulation, a = 0.5%
9. Diode voltage drop, Vd = 1.0 volt
10. Operating flux density, AB, (Bac = AB/2) = 0.1 tesla
11. Core Material = ferrite
12. Window utilization, Ku = 0.3
13. Temperature rise goal, Tr = 30°C
14. Maximum duty ratio, Dmax = 0.5
15. Notes:
Demag turns ratio, Nmag/Np =1
Demag power, Pmag =(0.1)P 0

Select a wire so that the relationship between the ac resistance and the dc resistance is 1:

/?„

The skin depth in centimeters is:

6.62
£
=-JT> [cml

6.62
e= / =•, [cm J
VIOO,QOO
£ = 0.0209, [cm]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Then, the wire diameter is:

Wire Diameter =2(e), [cm]

Wire Diameter =2(0.0209) [cm]
Wire Diameter = 0.0418 , [cm ]

Then, the bare wire area, Aw, is:

Aw =——, [cm 2 ]

A (3.1416)(Q.0418) Z ,
Av = - ^ ~, [cm ]

Aw =0.00137, [cm 2 ]

From the Wire Table, in Chapter 4, Number 26 has a bare wire area of 0.001280 centimeters. This will be
the minimum wire size used in this design. If the design requires more wire area to meet the specification,
then, the design will use a multifilar of #26. Listed Below are #27 and #28, just in case #26 requires too
much rounding off.
Wire AWG Bare Area Area Ins. Bare/Ins. uO/cm
#26 0.001280 0.001603 0.798 1345
#27 0.001021 0.001313 0.778 1687
#28 0.0008046 0.0010515 0.765 2142

Step No. 1 Calculate the transformer output power, P0.

P0 = Io(V0+Vd} [watts]
P0=5(5+l), [watts ]
P0 =30, [watts ]

Step No. 2 Calculate the input power, P;,,.

) [watts]
n
[watts]

/>,.„ =33.67, [watts ]

Step No. 3 Calculate the electrical coefficient, Ke.

Ke =0.145/ 2 AS 2 (lO- 4 )

Ke =0.145(100,000 ) 2 (0.l) 2 (lO" 4 )

Kf = 1450

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 4 Calculate the core geometry, Kg.

K = P<" °mm , [cm 5 ]

ocKe
(33.67X0.5) 5n
= [Cm ]
*' (0.5X1450)'
Kg = 0.0232, [cm 5 ]

When operating at high frequencies, the engineer has to review the window utilization factor, Ku. When
using small bobbin ferrites, the ratio of the bobbin winding area to the core window area is only about 0.6.
Operating at 100kHz and having to use a #26 wire, because of the skin effect, the ratio of the bare copper
area to the total area is 0.78. Therefore, the overall window utilization, Ku, is reduced. To return the design
back to the norm, the core geometry, Kg, is to be multiplied by 1.35, and then, the current density, J, is
calculated, using a window utilization factor of 0.29.

Kg = 0.0232 (1.35) [cm 5 ]

Kg = 0.0313, [cm 5 ]

Step No. 5 Select a EPC core from Chapter 3, comparable in core geometry Kg.

Core number EPC-30

Manufacturer TDK
Magnetic material PC44
Magnetic path length , MPL 8.2 cm
Window height, G 2.6 cm
Core weight, W(ft 23 grams
Copper weight, W(cu 22 grams
Mean length turn, MLT 5.5 cm
Iron area, AC 0.61 cm2
Window area, Wa 1.118 cm2
Area product, A 0.682 cm4
Core geometry, K 0.0301 cm5
Surface area, At 31.5 cm2

Millihenrys per 1000 turns, AL 1570

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 6 Calculate the number of primary turns, Np.

TV = --—-—---^ -c, [turns I

fAcAB
(22)(Q.5)(lQ4)
[tums]
(100, ooo)(o 61X0.1)''
N p =18.0, [turns ]

Step No. 7 Calculate the current density, J, using a window utilization, Ku = 0.29.

[amps / Cm2
fA<ABWaKtt
2(33.67XO-707)(l0 4 )
(100,000X0. 61X0.1X1. H8X0.29)'
=241, [amps / c m 2 ]

Step No. 8 Calculate the primary rms current, Ip.

in (min )W (max

(33.67)
" (22X0.707)'
/ = 2.16, [amps ]

Step No. 9 Calculate the primary bare wire area, Awp(B).

2
A wp(B)
,„, =-^,
j L[cm J]

_2.16 2
~~ 241 ' L°m J
= 0.00896, [cm 2 ]

Step No. 10 Calculate the required number of primary strands, NSP.

#26
0.00896
0.00128

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 1 1 Calculate the primary new ufi per centimeter.

cm
(new )//Q / cm = ——-
p
1345
(new )//Q / cm = —-—

(new)//Q/cm = 192

Step No. 12 Calculate the primary resistance, Rp.

6
RP=MLT:(N
v P
) —](io~
v
) [ohms]
Acmy '
Rp= (5.5X18X192^10^) [ohms]
Rp= 0.0190, [ohms]

Step No. 13 Calculate the primary copper loss, Pp.

p =!2pRp, [watts ]

Pp =(2.16) 2 (0.019) [watts ]

Pp = 0.0886, [watts ]

Step No. 14 Calculate the secondary turns, Ns.

ss
(0.5X22)1 100 '
[tums]
L J

Ns =9.87 use 10, [turns ]

Step No. 15 Calculate the secondary rms current, Is.

Is= [amps]
~L4~i'
/ =3.55, [amps ]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 16 Calculate the secondary bare wire area, Aws(B).

A
™(B) = ~^> tcm2]
AvS(*) = 0.0147, [cm 2 ]

Step No. 17 Calculate the required number of secondary strands, NSS.

NS =
#26
.
5
0.00128
NSS =11.48 use 11

Step No. 18 Calculate the secondary, new ufi per centimeter.

, , ,. . Liiil cm
(new)//Q/cm =-
AS,
1345
(new )fjQ I cm =

(new )fj£l / cm = 122

Step No. 19 Calculate the secondary winding resistance, Rs.

Rs = M L T ( A O l O -6 [ohms ]
Vcm/v s
^ = (5.5)(loXl22)(lO~ 6 ) [ohms]
^,=0.00671, [ohms]

Step No. 20 Calculate the secondary copper loss, Ps.

Ps =/, 2 /? s , [watts ]
Ps = (3.55 /(0. 00671 ) [watts ]
Ps = 0. 0846 , [watts ]

Step No. 2 1 Calculate the total primary and secondary copper loss, Pcu.

PCU =PP+PS, [watts ]

Pcu =0.0886 +0.0846, [watts]
Pcu =0.173, [watts ]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 22 Calculate the transformer regulation, a.

a =--(100) [%]
"o

(0.173"
(100)

a = 0.576, [%]

Step No. 23 Calculate the demag winding inductance,

[mh]

Ldemag =(l570)(l8) ! (lO- 6 ) [mh]

Ldemag =0.509, [mh]

Step No. 24 Calculate the time of, At. See Figure 14-6.

^(demag)

Figure 14-6. Magnetizing Current Waveform.

A =71Z)
' (max)' [seconds ]

T =—, [seconds ]

T = -, [seconds ]
100,000
lo(lO- 6 ) [seconds ]

= ^10(lO~ 6 jy0.5) [seconds]

[seconds ]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 25 Calculate the demag, winding delta current, Aldemag-

V • A?
-£-*- , [amps]
demag

- • (509
ag =0.217, [amps ]

Step No. 26 Calculate the demag, winding rms current, Idemag- This is the rms equation for a saw tooth
current.

, . . max r n
Idemag = A/J , [amps ]

Idemag =(0.217X0.408), [amps]

Idemag =0.089, [amps]

Step No. 27 Calculate the required demag, wire area, Aw(demag).

i demag r 2n
A , d <= , [cm ]

0.089 r ,
A ,, , = , [cm ]
»(*'»«*) 241
Aw(ilemag) = 0.000369, » #31 use a #26

Step No. 28 Calculate the window utilization, Ku.

' ' **...! Dv«-i/:i

N = (NpNSp ) + (NsNSs) + (NdemagNSdemag

N = 254
_ (254)(0.00l28)
1.H8
K,. =0.291

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 29 Calculate the milliwatts per gram, mW/g.

mW/g = 0.000318 (/)'5'(S,,J2747

mW/g = 0.000318 (1OOOOO) 15 '(0.05) 2747
mW/g = 3.01

Step No. 30 Calculate the core loss, Pfe.

Pfe =(mW I g^Vtfe Y!° 3 ) [watts ]

P /e =(3.0lX23)(lO- 3 ) [watts]
Pfe =0.069, [watts ]

Step No. 31 Calculate the total loss,

PI. = pcu + P/e, [watts ]

Pz = (0.173 )+ (0.069 ) [watts
Pz =0.242, [watts ]

Step No. 32 Calculate the watts per unit area, \\s.

p
y/=—-, [watts / c m 2 ]
A,
(0.242) ,
V= \3l 5y [watts / c m ' ]

(^ = 0.0077, [watts / c m 2 ]

Step No. 33 Calculate the temperature rise, Tr.

7;=450(i//) (o - 826) , [°C]

r r = 450 (0.0077 )(0'826), [°C]

7 r = 8 . 0 8 , [°C]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Forward Converter Output Inductor Design

Part 2 is designing the output inductor, LI, as shown in Figure 14-7. The output filter inductor for switch-
mode power supplies, (SMPS), probably has been designed more times than any other single component.
Presented here is a straight-forward approach for selecting the core and the proper wire size to meet the
specification.

The losses in the magnetic material will increase significantly when the converter is operating at a higher
frequency. However, the core loss in the output inductor of a switching regulator is much lower compared
to the core loss in the main converter transformer. The core loss in the output inductor is caused by the
change in current or AI, which induces a change in flux, as shown in Figure 14-7.

B (tesla)
>• AI

H (oersteds)

Figure 14-7. Typical output inductor BH Loop.

The single-ended, forward converter schematic is shown in Figure 14-8. This topology is appealing to
engineers for it simplicity and parts' count. The output filter circuit, shown in Figure 14-8, has three current
probes. These current probes monitor the three basic currents in a switch mode, converter output filter.
Current probe A monitors the transformer's secondary current. Current probe B monitors the commutating
current through CR3. Current probe C monitors the current through the output inductor, LI.

The typical secondary and filter waveforms of the forward converter are shown in Figure 14-8. The
waveforms are shown with the converter operating at a 0.5 duty ratio. The applied voltage, V1; to the filter,
is shown in Figure (14-9-A). The transformer's secondary current is shown in Figure (14-9-B). The
commutating current flowing through, CR3, is shown in Figure (14-9-C). The commutating current is the
result of Ql being turned off, and the field in, LI, collapsing, producing the commutating current. The
current flowing through LI is shown in Figure (14-9-D). The current flowing through LI is the sum of the
currents in Figure (14-9-B) and (14-9-C).

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 VI LI Current Probe C
Current Probe A
U PI TT* — \j
CR1
CR2
C2
CR3 +
2c s- ^
N2 v0
1 —c >

R Current Probe B

Figure 14-8. Typical single-ended, forward converter.

V
(14-9- A)

(14-9-B)

(14-9-C)

(14-9-D)

Current Probe C

Figure 14-9. Typical forward converter waveforms, operating at a 0.5 duty ratio.

The critical inductance current is shown in Figure (14-10-B). The critical inductance current is when the
ratio of the delta current to the output load current is equal to 2 = AI / I0. If the output load current is
allowed to go beyond this point, the current will become discontinuous, as shown in Figure (14-10-D). The
applied voltage, V,, will have ringing at the level of the output voltage, as shown in Figure (14-10-C).
When the current in the output inductor becomes discontinuous, as shown in Figure (14-10-D), the
response time for a step load becomes very poor.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 When designing multiple output converters, the slaved outputs should never have the current in the
inductor go discontinuous or to zero. If the current goes to zero, a slaved output voltage will rise to the
value of V|. If the current is allowed to go to zero, then, there is not any potential difference between the
input and output voltage of the filter. Then the output voltage will rise to equal the peak input voltage.

14-10-A

14-10-B

14-10-C

14-10-D

Current Probe C

Figure 14-10. Forward converter; output filter inductor goes from critical to discontinuous operation.

Output Inductor Design Using the Core Geometry, Kg, Approach

The following information is the design specification for a forward converter, 30 watt output filter design,
operating at 100kHz, using the Kg core geometry approach. For a typical design example, assume an
output filter circuit, as shown in Figure 14-1, with the following specifications:

1. Frequency, f = 100kHz
2. Output voltage, V0 = 5 volts
=
3. Output current, I0(max) 5.0 amps
=
4. Output current, I0(max) •••• 0-5 amps
5. Delta current, AI = 1.0 amps
6. Input voltage, V l(max) = 19 volts
7. Input voltage, V1(min) = 12 volts
8. Regulation, ot = 1.0%

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 9. Output power (V0 + Vd) (I0(max)), P0 = 30 watts
10. Operating flux density, Bpk = 0.3 tesla
11. Window utilization, Ku = 0.4
12. Diode voltage drop, Vd = 1.0 volt

This design procedure will work equally well with all of the various powder cores. Care must be taken
regarding maximum flux density with different materials and core loss.

The skin effect on an inductor is the same as a transformer. The main difference is that the ac flux is much
lower and does not require the use of the same maximum wire size. The ac flux is caused by the delta
current, AI, and is normally only a fraction of the dc flux. In this design the ac current and the dc current
will be treated the same.

At this point, select a wire so that the relationship between the ac resistance and the dc resistance is 1:

=1

The skin depth, z, in centimeters, is:

6.62
77
6.62
£ =' , [cm]
'100,000

s = 0.0209, [cm]

Then, the wire diameter, Dw, is:

Dw=2(s], [cm]
Dv= 2 (0.0209), [cm]
Av=0.0418, [cm]

Then, the bare wire area, Aw, is:

A,f =0.00137, [cm2]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 From the Wire Table in Chapter 4, Number 27 has a bare wire area of 0.001021 centimeters. This will be
the minimum wire size used in this design. If the design requires more wire area to meet the specification,
then the design will use a multifilar of #26. Listed Below are #27 and #28, just, in case #26 requires too
much rounding off.

Wire AWG Bare Area Area Ins. Bare/Ins. ufi/cm

#26 0.001280 0.001603 0.798 1345
#27 0.001021 0.001313 0.778 1687
#28 0.0008046 0.0010515 0.765 2142

Step No. 1 Calculate the total period, T.

r = — , [seconds]

T = - , [seconds]
100,000
7 = 10, [usec]

Step No. 2 Calculate the minimum duty ratio> Dmin.

D =—
mi n ^

£ mi ,,= 0.263

Step No. 3 Calculate the required inductance, L.

A/
6
(lOxlO" )(5.0 + 1.0)(l-0.263)
L= -_ - , [henrys]

1=44.2, [//h]

Step No. 4 Calculate the peak current, lp^.

, [amps]

Ipk = 5.5, [amps]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 5 Calculate the energy-handling capability in, watt-seconds.

LI*
Energy = ——, [watt-seconds]

(44.2x1 (T6)(5.5)2
Energy = - , [watt-seconds]

Energy = 0.000668, [watt-seconds]

Step No. 6 Calculate the electrical conditions, Ke.

Ke = OA45P0B

AT e =(0.145X30X0.3)?xlO~ 4

Ke =0.0000392

Step No. 7 Calculate the core geometry, Kg.

= (Energy J
S ws- L J
* K,a

(0.000668 )P
\V.VV\JVVO J 5
g ~ ir\ f\r\f\f\ic\^ LCm J
(0.0000392 \/l
)(1.n\
0)

Kg =0.01138 [cm 5 ]

Step No. 8 Select, from Chapter 4, a MPP powder core, comparable in core geometry, Kg.

Core number MP-55059-A2

Manufacturer Magnetics
Magnetic path length, MPL = 5.7 cm
Core weight, Wtfe = 16.0 grams
Copper weight, Wtcu = 15.2 grams
Mean length turn, MLT = 3.2 cm
Iron area, Ac = 0.331 cm2
Window Area, Wa = 1.356 cm2
Area Product, Ap = 0.449 cm4
Core geometry, Kg = 0.0184 cm5
Surface area, A, = 28.6 cm2
Permeability, p, = 60
Millihenrys per 1000 turns, AL = 43

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 9 Calculate the number of turns, N.

\L ,
N = 1000 U^i, [turns]
\\\I^(1000)
In 0442
N = 1000. , [turns]
V 43
N = 32, [turns]

Step No. 10 Calculate the rms current, lms.

™ = > / ( 5 . 0 ) 2 + ( l . O ) 2 , [amps]

Step No. 1 1 Calculate the current density, J, using a window utilization, Ku = 0.4.

J = - , [ amps-per-cm2 ]

, [amps-per-cm ]
(1.36)(0.4)
J = 300, [amps-per-cm2]

Step No. 12 Calculate the required permeability, Au.

A^=-^ ^ '-, [perm]

(0.3)(5.7)(l0 4 )
A
(1.26)(1.36)(300)(0.4)'
A// = 83.1, use 60 [perm]

Step No. 13 Calculate the peak flux density, Bp^.

^K)
" (MPL)
(1.26)(32)(5.5)(6Q)(10-4)
, [tesla]
(5.7)
Bpk= 0.233, [tesla]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 14 Calculate the required bare wire area, AW(B).

A =Z^ [cm 2]
J

5.1 r „
L
»w 300'
4,^=0.017, [cm2]

Step No. 15 Calculate the required number of strands, Sn.

5 = ' , [strands]
" #26
S = —'• , [strands]
0.00128
5, =13, [strands]

Step No. 16 Calculate the new, ufi, per centimeter.

/ x f£l/cm
(newj//Q/cm = -
n
1345
I(newJ/^Q/cm
N o / =

Step No. 17 Calculate the winding resistance, R.

6
R = (MLT:)N\ — |(iO"
v ;
), [ohms]
Umj
/? = (3.2)(32)(l03)(lO' 6 ), [ohms]
fl = 0.0105, [ohms]

Step No. 18 Calculate the winding copper loss, Pcu.

^ = C « , [watts]
pcu =(5.1) 2 (0.0105), [watts]
Pcu = 0.273, [watts]

Step No. 1 9 Calculate the magnetizing force in oersteds, H.

OAxNI,
H = -p— , [oersteds]
MPL
(1.26)(32)(5.5)
V-A _ A—1_ [oersteds]
5.7
H = 38.9, [oersteds]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 20 Calculate the ac flux density in tesla, Bac.

Bac = - - - , [tesla]
MPL
(1.26)(32)(0.5)(60)(10- 4 )
A
' A A '

Bac= 0.0212, [tesla]

Step No. 2 1 Calculate the regulation, a, for this design.

a =^-(100), [%]
*o

(0.273)

a = 0.91, [%]

Step No. 22 Calculate the watts per kilogram, WK, using MPP 60 perm powder cores coefficients, shown
in Chapter 2.

WK = 0.55l(lO~ 2 )/ (l -23)B% I2) , [watts-per-kilogram]

WK = 0.55l(lO~ 2 )(lOOOOO) ( ' 23) (0.0212)(2''2) , [watts-per-kilogram]

WK = 2.203, [watts-per-kilogram]

Step No. 23 Calculate the core loss, Pfe.

| milliwatts ]„, /, . , \
P = -v W ; 1 0-3 , [watts]
^ gram )
Pfe= (2.203) (16)(10-3), [watts]
Pfe = 0.0352, [watts]

Step No. 24 Calculate the total loss, Ps, core, Pfe, and copper, Pcu, in watts.

Pz=Pfe+Pcu> [watts]
Pj. =(0.0352) + (0.273), [watts]
Ps = 0.308, [watts]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 25 Calculate the watt density, v|/.

p
\y = — , [watts-per-cm2 ]
4
(0.308) 2n
ur = —,--r- , [watts-per-cm 1
(28.6)
iff = 0.0108, [watts-per-cm2 ]

Step No. 26 Calculate the temperature rise, in °C.

Tr = 450(0.0108)(°826) ,
Tr = 10.7, [°C]

Step No. 27 Calculate the window utilization, Ku.

J32)(13)(0.00128)
(1.356)
K,. = 0.393

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Chapter 15

Input Filter Design

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table of Contents

1. Introduction
2. Capacitor
3. Inductor
4. Oscillation
5. Applying Power
6. Resonant Charge
7. Input Filter Inductor Design Procedure
8. Input Filter Design Specification
9. References

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Introduction
Today, almost all modern equipment uses some sort of power conditioning. There are a lot of different
circuit topologies used. When you get to the bottom line, all power conditioning requires some kind of an
input filter. The input LC filter has become very critical in its design and must be designed not only for
EMI, but also for system stability, and for the amount of ac ripple current drawn from the source.

The input voltage supplied to the equipment is also supplied to other users. For this reason, there is a
specification requirement regarding the amount of ripple current seen at the source, as shown in Figure 15-
1. Ripple currents generated by the user induce a ripple voltage, Vz, across the source impedance. This
ripple voltage could impede the performance of other equipment connected to the same bus.

Other Users LI
Mn
[ —*" o^o
^source
Switching
v
^source ~~^ ln Cl
^~ ->, Converter

Figure 15-1. Simple, LC, Input Filter.

Capacitor
Switching regulators have required the engineer to put a significantly more analytical effort into the design
of the input filter. The current pulse, induced by the switching regulator, has had the most impact on the
input capacitor. These current pulses required the use of high quality capacitors with low ESR. The
waveforms, induced by the switching regulator, are shown in Figure 15-2. In the input inductor, LI, peak-
peak ripple current is, IL- In the capacitor, Cl, peak-peak, ripple current is, Ic. In the capacitor, Cl, peak-
peak, ripple voltage is, AVC. The equivalent circuit for the capacitor is shown in Figure 15-3. The voltage,
AVC, developed across the capacitor, is the sum of two components, the equivalent series resistance, (ESR),
and the reactance of the capacitor.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The voltage developed across the equivalent series resistance, (ESR), is:

F M = / C ( E S R ) , [volts] [15-1]

The voltage developed across the capacitance is:

[15-2]
•A1) )

The sum of the two voltages, AVCR and AV<x, is:

A Vc = A VCR + A Vcc, [volts] [15-3]

Ic 0

IL o _J.__X-^-

Figure 15-2. Typical Voltage, Current Waveforms.

AVf
Input from Inductor L -O Input switching converter

ESR AVCR

Cl +

Figure 15-3. Capacitor, Individual Ripple, Components.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Inductor

The input filter inductor is basically a straight-forward design. There are four parameters required to
achieve a good design: (1) required inductance, (2) dc current, (3) dc resistance, and (4) temperature rise.
The requirement for the input inductor is to provide a low ac ripple current to the source. The low ac ripple
current in the inductor produces an ac flux at a magnitude of about 0.025 tesla. This resulting low ac flux
will keep the core loss to a minimum. The input inductor losses will normally be 80 to 90% copper. A high
flux magnetic material is ideally suited in this application. Operating with a high dc flux and a low ac flux,
silicon, with its high flux density of 1.6 teslas, will produce the smallest size, as shown in Table 15-1.

Table 15-1. Most Commonly Used Input Filter Material.

Magnetic Material Properties

Material Operating Flux, Permeability
B, tesla "i
Silicon 1.5-1.8 1.5K
Permalloy Powder 0.3 14-550
Iron Power 1.2-1.4 35-90
Ferrite 0.3 1K-15K

Oscillation
The input filter can affect the stability of the associated switching converter. The stability problem results
from an interaction between the output impedance of the input filter and the input impedance of the
switching converter. Oscillation occurs when the combined positive resistance of the LC filter, and power
source exceed the negative dynamic resistance of the regulator's dc input. To prevent oscillation, the
capacitor's ESR, and the inductor's resistance must provide sufficient damping. Oscillation will not occur
when:

[15-4]

Where r| is the switching converter efficiency, Vjn(max) is the input voltage; P0 is the output power in watts, L
is the input inductor in henrys; where, C, is the filter capacitor in farads, RL is inductor series resistance in
ohms; Rs is the source resistance in ohms, and Rj (ESR), is the equivalent series resistance in ohms. If
additional damping is required, it can be done, by increasing the Rj (ESR), and/or RL. See Figure 15-4. The
series resistance, Rj, lowers the Q of the filter and kills the potential Oscillation.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Applying Power
The inrush current has always been a problem with this simple LC input filter. When a step input is
applied, such as a relay or switch SI as shown in Figure 15-5, there is always a high inrush current.

^source

-"source

Figure 15-4. Input Filter, with Additional Damping.

source

source

Figure 15-5. Input Filter Inrush Current Measurement.

When SI is closed, the full input voltage, Vin, is applied directly across the input inductor, LI, because Cl
is discharged. The applied input voltage, V^, (volt-seconds), to the input inductor, LI, and the dc current,
(amp-turns), flowing through it is enough to saturate the core. The inductor, LI, is normally designed, using
the upper limits of the flux density for minimum size. There are two types of core configurations commonly
used for input inductor design: powder cores and gapped cores. Some engineers prefer to design around
powder cores because they are simple and less hassle, while others design using gapped cores. It is strictly
a game of trade-offs. Tests were performed using three different core materials: (1) powder core, (2) ferrite
core, and (3) iron alloy. All three materials were designed to have the same inductance and the same dc
resistance. The three-inductor designs were tested to compare the inrush current under the same conditions.
The inrush current, AI, for all three materials is shown in Figure 15-6, using the test circuit, shown in Figure
15-5.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 As, shown in Figure 15-6, the inrush current for all three test inductors has about the same general shape
and amplitude. The changes in permeability, with dc bias, for both gapped and powder cores are shown in
Figure 15-7. Gapped cores have a definitely sharper knee while the powder cores roll off more gradually.
The advantage in using a gapped core over the powder core is the ability to use the full flux capacity of the
core up to the knee, before the permeability starts to droop.

Ferrite &
Iron Alloy
Powder Core

Vertical 2 amps/Div., Horizontal 500usec/Div.

Figure 15-6. Typical, Inrush Current for a Simple Input.

I
Gapped Ferrites and Iron Alloys
Exhibit this Sharp Knee.

10 100 1000
DC Magnetizing Force (Oersteds)

Figure 15-7. Comparing Gapped and Powder Cores, Permeability Change with DC Bias.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Resonant Charge
Most all types of electronic equipment are energized by either a switch or relay. This type of turn-on goes
for spacecraft, aircraft, computer, medical equipment, and automobiles. There are some power sources that
require some type of current limiting that does not follow the general rule. If the input voltage is applied via
a switch or relay to an input filter, as shown in Figure 15-8, a resonant charge condition will develop with
LI and Cl. The resulting resonant charge with LI and Cl could put a potential on Cl that could be as
much as twice the applied input voltage, as shown in Figure 15-9. The voltage rating of Cl must be high
enough to sustain this peak voltage without damage. The oscillating voltage is applied to the switching
converter.

A simple way to dampen this oscillation is to place one or two diodes across the input choke, as shown in
Figure 15-10. The reason for two diodes is the ripple voltage, Vc, might be greater than the threshold
voltage of the diode. As the voltage across Cl rises above the input voltage, Vin, due to the oscillation
diodes, CR1 and CR2 will become forward-biased, clamping the voltage across Cl to two diode drops
above the input voltage, V^, as shown in Figure 15-11.

LI
m
250
Cl
Vin = 28V V, 40 ohm
40 uf 20W

-O-

Figure 15-8. Typical, Simple LC, Input Filter.

60v--

50v--

40v--

1ms 2ms 3ms 4ms 5ms 6ms

Figure 15-9. Resonating Voltage, across Capacitor, Cl.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 LI

-O-
250 uh
Cl

V in = 28V vr 40z
40 20W

V
-O

Figure 15-10. Input Inductor with Clamp Diodes.

60v--

50v --

40v --

30v

20vH-

lOv -•

1ms 2ms 3ms 4ms 5ms 6ms

Figure 15-11. DC Voltage Across Cl, with the Clamp Diodes.

Input Filter Inductor Design Procedure

The input filter inductor, LI, for this design is shown in Figure 15-12.

nn •»- J^^X,
J- ' LI
source ^ Vz M i W
d
Switching
\r. CR1 CR2 +
ln /-•• -> rSUCK
E^source ~ Regulator
n

1 1
Figure 15-12. Input Filter Circuit.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The ac voltages and currents associated with the input capacitor, Cl, are shown in Figure 15-13.

! V ! loff !
1i
0 A C = Capacitor Current
AI
\r

DT D'T

AVC = AV C R ,AV C R »AV C C

Figure 15-13. Input Capacitor Voltage and Current Ripple.

The ac voltages and currents impressed on the input capacitor, Cl, are defined in Figure 15-14.

Input from Inductor, L1Q- -O Input to the Buck Converter

AIC

1 4
ESR AV,CR

Cl +
AVcc

Figure 15-14. Defining the Input Capacitor Voltage and Current Ripple.

The components of the inductor current, due to, AVCR and AVCC, are:

AVCR = Peak to Peak component due to capacitor, ESR.

A Vcc = Peak to Peak component due to capacitor.
AILR = Component of the inductor ripple current developed by &VCR.
AILC = Component of the inductor ripple current developed by A Vcc.

T), [amps]
[15-5]
A/,, = u l f i , [amps]
2L A4.

It will be considered that, AILR, dominates because of the capacitor, ESR, so:

L
\(DD'T), [amps] [15-6]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Input Filter Design Specification

1. Peak-Peak ripple voltage, AVcr = 0.5 volts

2. Peak-Peak ripple current to the source, AIL = 0.010 amps
3. Period, T = 10 usec
4. *Converter on-time duty cycle, D = ton/T = 0.5
5. *Converter off-time duty cycle, D"= t0g/T = 0.5
6. Regulation, a = 0.5%
7. Output power drawn from the filter network, P0 = 50 watts
8. Maximum current to the load, AIC = 4 amps
9. Average input current, 1^ = Iav = AIC D = 2 amps
10. The ripple frequency, f = 100kHz
11. The core RM ferrite, gapped, Bmax = 0.25 tesla
* The worse case time domain is where D and D" = 0.5

Step No. 1 Calculate the required inductance, L.

\V
L==-^-(DD'T), [henrys]

1 = 0.000125, [henrys]

Step No. 2 Calculate the energy-handling capability.

L(I )2
Energy = —-—— , [watt-second]

125(lO~ 6 )(2.0) 2
Energy = ----- , [watt-second]

Energy = 0.000250, [watt-second]

Step No. 3 Calculate the electrical coefficient, Ke.

A" e =0.145(50)(0.25) 2 (lO' 4 )

K. = 0.0000453

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 4 Calculate the core geometry, Kg.

(0.00025)2
is ________v_ /_ Fern 1
* ~(0.0000453)(0.5)'
Kg= 0.00275, [cm5]

Step No. 5 Select the comparable core geometry from the RM ferrite cores.

1. Core part number ...................................................................................................... = RM-6

2. Core geometry, Kg .................................................................................................... = 0.0044 cm5
3. Core cross-section, Ac ............................................................................................... = 0.366 cm"
4. Window area, Wa ...................................................................................................... = 0.260 cm2
5. Area product, Ap ....................................................................................................... = 0.0953 cm4
6. Mean length turn, MLT ............................................................................................ = 3.1 cm
7. Magnetic path length, MPL ...................................................................................... = 2.86 cm
8. Core weight, Wtfe ...................................................................................................... = 5.5 grams
9. Surface area, A, ......................................................................................................... = 11.3 cm"
10. Winding Length, G ................................................................................................... = 0.82 cm
11. Permeability, um ....................................................................................................... = 2500

Step No. 6 Calculate the current density, J, using the area product equation, Ap.

2(Energy)(l0 4 )
J = --- , [amps-per-cnr 1
BmApKu
2(0.00025)(l0 4 )
J = --—-—-- , [amps-per-cm ]
(0.25)(0.0953)(0.4)
J = 525, [amps-per-cm2 ]

Step No. 7 Calculate the required bare wire area, Aw(B).

(525)' —
= 0.00381, [cm2]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 8 Select a wire from the Wire Table in Chapter 4. If the area is not within 10%, take the next
smallest size. Also record micro-ohms per centimeter.

AWG = #21
Bare, AW(B) = 0.0041 1, [cm2]
Insulated, Aw = 0.00484, [cm2 ]

—— =419, [micro-ohm/cm]

Step No. 9 Calculate the effective window area, Wa(eff). Use the window area found in Step 5. A typical
value for, S3 is 0.75, as shown in Chapter 4.

Wa(eff)= 0.195, [cm2]

Step No. 10 Calculate the number turns possible, N. Use the insulated wire area, Aw, found in Step 8. A
typical value for, S2, is 0.6, as shown in Chapter 4.

W ^
N = ^^-, [rums]
AW
(0.195)(0.60)
N = ^--A , ; , [turns]
(0.00484)
N = 24, [rums]

Step No. 1 1 Calculate the required gap, lg.

, [cm]
*• i A. ;
(l.26)(24) 2 (0.366)(lO~ 8 )
, [cm]
*~ (0.000125)
lg = 0.0201, [cm]

Step No. 12 Calculate the equivalent gap in mils.

mils = cm(393.7)
mils = (0.0197)(393.7)
mils = 7.91 use 8

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 13 Calculate the fringing flux factor, F.

V0366 ^ 0.0201 J
F = 1.146

Step No. 14 Calculate the new number of turns, Nn, by inserting the fringing flux, F.

/ L
N = -*—.- -, [turns]
" \ 0.4* F l O - 8

(0.0201)(0.000125)
N = -i-^- j1 —r , [turns]
" y(l.26)(0.366)(l.!46)(lO- 8 )
Nn = 22, [turns]

Step No. 15 Calculate the winding resistance, RL. Use the, MLT, from Step 5 and the micro-ohm per
centimeter, from Step 8.

I, [ohms]

RL =(3.l)(22)(419)(lO" 6 ), [ohms]
RL = 0.0286, [ohms]

Step No. 16 Calculate the copper loss, Pcu.

pm =(2.0) 2 (0.0286), [watts]

P C1 ,=0.114, [watts]

Step No. 17 Calculate the regulation, a.

a =^-(100), [%]
0

(0.114)
a=
(io)
a = 0.228, [%

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 18 Calculate the ac flux density, Bac.

c 1 , [tesla]

" 4T)
(l.26)(22)(l.l4)f°fV 10 -)
Bac= ^-4 . [tesla]
(0.0197)+
V -~
' {25QOJ
Bac = 0.000758, [tesla]

Step No. 19 Calculate the watts-per-kilogram, for ferrite, P, material in Chapter 2. Watts per kilogram can
be written in milliwatts-per-gram.

mW/g = (0.00198)(100000)° 36) (0.000758)(286)

mW/g = 0.0000149

Step No. 20 Calculate the core loss, Pfe.

/%=(mW/g)(^)(lO- 3 ), [watts]
Pfe =(0.0000149)(5.5)(lO" 3 ), [watts]
Pfe = 0.082 (lO" 6 ), [watts]

Step No. 21 Calculate the total loss copper plus iron, Pj>

/>, = Pfe + Pcu, [watts]

PL =(0.000) + (0.114), [watts]
4=0.114, [watts]

Step No. 22 Calculate the watt density,y. The surface area, A,, can be found in Step 5.

^
iff = —, [watts/cm2 ]
4
(0.114)
w = —, -^, [watts/cm ]
(11.3)
y = 0.010, [watts/cm2]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step No. 23 Calculate the temperature rise, Tr.

Tr =450(0.010) ( ° S26) ,
?;=1 0.0, [°C]

Step No. 24 Calculate the peak flux density, Bpk.

(I.26)(22)(l.l4)(2.005)(l0~ 4 )

(0.0197)-
2500
Bpk = 0.304, [tesla]

Step No. 25 Calculate the window utilization, Ku.

K=-

= (0.00411)(22)
(0.260)
A:,, = 0.348

References

1. T. K. Phelps and W. S. Tate, "Optimizing Passive Input Filter Design," (no source).

2. David Silber, "Simplifying the Switching Regulator Input Filter," Solid-State Power Conversion,
May/June 1975.

3. Dan Sheehan, "Designing a Regulator's LC Input Filter: 'Ripple' Method Prevents Oscillation Woes,'
Electronic Design 16, August 2, 1979.

Note:
I would like to thank Jerry Fridenberg, for modeling the circuits in Figure 15-8 and 15-10, on his SPICE
program. The modeling results are shown in Figures 15-9 and 15-11.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Chapter 16

Current Transformer Design

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table of Contents

1. Introduction
2. Analysis of the Input Current Component
3. Unique to a Current Transformer
4. Current Transformer Circuit Applications
5. Current Transformer Design Example
6. Design Performance

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Introduction

Current transformers are used to measure, or monitor, the current in the lead of an ac power circuit. They
are very useful in high-power circuits, where the current is large, i.e., higher than the ratings of so-called
self-contained current meters. Other applications relate to overcurrent and undercurrent relaying for power
circuit protection, such as, in the power lead of an inverter or converter. Multiturn secondaries then
provide a reduced current for detecting overcurrent, undercurrent, peak current, and average current, as
shown in Figure 16-1.

Load

Figure 16-1. Simple, Secondary AC Current Monitor.

In current transformer designs, the core characteristics must be carefully selected because excitation
current, Ira, essentially subtracts from the metered current and effects the true ratio and phase angle of the
output current.

The simplified equivalent circuit of a current transformer, as shown in Figure 16-2, represents the important
elements of a current transformer, where the ratio of primary to secondary turns is:

N
«=—K [turnsratio] [16-1]
N
P

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Figure 16-2. Simplified, Equivalent Circuit for a Current Transformer.

Analysis of the Input Current Component

A better understanding of the current transformer behavior may be achieved by considering the applied
input current to the primary winding, in terms of various components. Only the ampere-turn component,
linNp, drives the magnetic flux around the core. The ampere -turn, ImNp, provides the core loss. The
secondary ampere -turns, ISNP, balance the remainder of the primary ampere -turns.

The exciting current, Im, in Figure 16-2, determines the maximum accuracy that can be achieved with a
current transformer. Exciting current, Im, may be defined as the portion of the primary current that satisfies
the hysteresis and eddy current losses of the core. If the values of Lc and Re- in Figure 16-2, are too low
because the permeability of the core material is low and the core loss is high, only a part of the current,
(Ip/n), will flow in the output load resistor, RO. The relationship of the exciting current, Im, to the load
current, I0, is shown in Figure 1 6-3 .

The exciting current is equal to:

QAxN
, [amps] [IM]

where H is the magnetizing force and material dependent, and MPL is the Magnetic Path Length.

lin to- lr» to-

O- VNAA-
RS

Figure 16-3. Input Current-Output Current Relationship.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The input current is comprised of two components: exciting current, Im, and the load current, I0.

Then:

I2m=lrn-I20, [amps] [16-4]

Im=Iin , [amps] [16-5]

The above equation has shown graphically in Figure 16-4, that the higher the exciting current, Im or core
loss, the larger the error. The magnetizing impedance, Rg, determines accuracy, because it shunts part of
the input current, !;„, away from the primary and thus, produces an error, as shown in Figure 16-4. Core
material with the lowest value of H achieves the highest accuracy.

Figure 16-4. Input Current I in Phase Relationship Diagram.

Unique to a Current Transformer

The current transformer function is different than that of a voltage transformer. A current transformer
operates with a set primary current and will try to output a constant current to the load, independent of the
load. The current transformer will operate into either a short circuit or a resistive load until the voltage
induced is enough to saturate the core or cause voltage breakdown. For this reason a current transformer
should never operate into an open circuit, as a voltage transformer should never operate into a short circuit.
The primary current of a current transformer is not dependent of the secondary load current. The current is
really injected into the primary by an external load current, Ijn. If the load current, I0, on the current
transformer is removed from the secondary winding, while the external load current, Ijn, is still applied, the
flux in the core will rise to a high level, because there is not an opposing current in the secondary winding
to prevent this. A very high voltage will appear across the secondary. A current transformer, like any other
transformer, must satisfy the amp-turn equation:

I. N,

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The secondary load, RO, secondary winding resistance, Rs, and secondary load current, I0, determine the
induced voltage of the current transformer.

VS=I0(RS+R0], [volts] [16-7]

If the secondary is designed for dc, then the diode drop must be taken into account.

Vs=I0(Rs+R0) + Vd, [volts] [16-8]

Simple form:
Vs=V0 + V d , [volts] [16-9]

The current ratio will set the turns ratio. The secondary, RO load will determine the secondary voltage, Vs.
The engineer would use Equation 16-10, to select the required core cross-section, Ac. It is now up to the
engineer to pick a core material that would provide the highest permeability at the operating flux density,
Bac.

4
I!n(R, -
+ R __ 2
KfBacfNs

The design requirements would dictate choosing a core material and operating flux density, Bac, that would
result in values of, Lc and Re, as shown in Figure 16-2, values which would be large enough to reduce the
current flowing in these elements to satisfy the ratio and phase requirements.

The inductance is calculated from the equation:

L = -—- -L '-, y[henrys]

J L [16-11]
J
MPL

is the equivalent core loss, (shunt), resistance. The current is, in phase, with the voltage.

R e = , [ohms] [16-12]

Where:
Hf-D RS+R0 [16-13]
n
And:

RS + R0 [16-14]
n

Then:
Ip=nls [16-15]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Or:
IpNp=IsNs [16-16]

Except for relatively low-accuracy industrial types, current transformers are wound on toroidal cores,
which virtually eliminate errors due to leakage inductance. Some errors may be compensated for by
adjusting the number of secondary turns.

Current Transformer Circuit Applications

Typical current transformer applications are shown in Figures 16-5 through 16-8.

Figure 16-5. Current Transformer, T2, used to Monitor, Ql and Q2, Drain Current.

Tl -O

I)

v
-O

Figure 16-6. Current Transformer, Tl, used to Monitor Line Current.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 RIO +5VRef
Current
Sense

Figure 16-7. Current Transformer, Tl, is used as a Level Detector.

Figure 16-8. Current Transformer, Tl, is used for regenerative drive.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Current Transformer Design Example

The following information is the design specification for a current transformer, as shown in Figure 16-9.

CR1-4

Figure 16-9. Current Monitoring Transformer with dc Output.

1. Primary 1 turn
2. Input current, I;n 0 - 5 amps
3. Output voltage, V0 0 - 5 volts
4. Output load resistance, R, 500 ohms
5. Operating frequency, f (square wave) 2500 hertz
6. Operating flux density, Bac 0.2 tesla
7. Core loss less than 3% (error)
8. Diode drop, Vd 1 volt
9. Magnetic material Supermalloy 2 mil
10. Waveform factor, Kf 4.0

Step 1. Calculate the secondary current, Is.

/,=—, [amps]

5
/Ts = '° , r[amps]n
500
^=0.01, [amps]

Step 2. Calculate the secondary turns, Ns.

/ /v
, [turns]

[turns]
(0.01)
Ns = 500, [turns]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step 3. Calculate the secondary voltage, Vs.

Vs=V0+2Vd, [volts]
Vs =5.0 + 2(1.0), [volts]
Vs = 7.0, [volts]

Step 4. Calculate the required core iron cross-section, Ac, using Faraday's Equation.

, [cm2]
(Kf}BacfNs

(7.0)(l0 4 )
A = _ ^_ A_ I_ r 2
c L
(4.0)(0.2)(2500)(500)'
Ac= 0.070, [cm2]

Step 5. Select a 2mil tape, toroidal core from Chapter 3 with an iron cross-section, Ac, closest to the value
calculated.

Core number ................................................................................................... 52000

Manufacturer .................................................................................................. Magnetics
Magnetic material ......................................................................................... 2mil Silcon
Magnetic path length, MPL ............................................................................. 4.99 cm
Core weight, Wtfc ........................................................................................... 3.3 grams
Copper weight, W(cu ....................................................................................... 8.1 grams
Mean length turn, MLT .................................................................................. 2.7 cm
Iron area, AC ................................................................................................... 0.086 cm2
Window area, Wa ........................................................................................... 0.851 cm"
Area product, A ............................................................................................ 0.0732 cm4
Core geometry, K ........................................................................................... 0.000938 cm5
Surface area, A( .............................................................................................. 20.6 cm2

Step 6. Calculate the effective window area, W a(eff) . A typical value for, S3, is 0.75, as shown in Chapter 4.

Wa(<ff)=WttS,, [cm2]

Wa(eff]= 0.638, [ c m2 ]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step 7. Calculate the secondary window area, Wa(Sec).

2
0.638
[
"*•"*•> 2
^fl(scc) =0.319, [cm2]

Step 8. Calculate the wire area, A w , with insulation , using a fill factor, S2 of 0.6.

(0319X^6)
(500)
/4W = 0.000383, [cm2]

Step 9. Select a wire area, Aw, with insulation from Wire Table in Chapter 4 for an equivalent AWG wire
size. The rule is that when the calculated wire area does not fall within 10% of those listed, in Wire Table,
then, the next smaller size should be selected.

AWG No. 33
4, = 0.0003662, [cm2]

Step 10. Calculate the secondary winding resistance, Rs using the Wire Table in Chapter 4, for uQ/cm; and
Step 5 for the MLT.

Rs = MLT(7V,.)[ — ](lO~ 6 ), [ohms]

Rs =(2.7)(500)(6748)(lO~ 6 ), [ohms]
Rs=9.ll, [ohms]

Step 11. Calculate the secondary output power, P0.

P0=I,(V0+2Vd), [watts]
P 0 =(0.01)(5.0 + 2(1.0)), [watts]
P0 = 0.070, [watts]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step 12. Calculate the acceptable core loss, Pfe.

„ ( core loss '

Pfe= 0.0021, [watts]

Step 13. Calculate the effective core weight, Wtfe(eff). Select the core weight correction factor, Kw, in
Chapter 2, for Supermalloy.

»W)= (3-3)0.148), [grams]

Wtfe(eff)=3.19, [grams]

Step 14. Calculate the allowable core loss, Pfe, in milliwatts per gram, mW/g.

/> , ,\
mW/g = —^10 J, [milliwatts per gram]

mW/g = ±— -^ (103), [milliwatts per gram]

mW/g = 0.554, [milliwatts per gram]

Step 15. Calculate the new flux density using the new core iron, cross-section, Ac.

(7.0)(l0 4 )
ac
"(4.0)(0.086)(2500)(500)'
Bac= 0.162, [tesla]

Step 16. Calculate the core loss, Pfe, in milliwatts per gram, mW/g.

mW/g = 0.000179(/)° 48) (Bac )(2 '5) , [milliwatts per gram]

mW/g = 0.000179(2500)° 48) (0.162)'2 '5) , [milliwatts per gram]

mW/g = 0.382, [milliwatts per gram]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step 17. Calculate the core loss, Pfe, in watts.

= — . V -5 J

P /e =3.79(0.382)(lO- 3 ), [watts]
P/e = 0.00145, [watts]

Step 18. Calculate the induced core error in, %.

Core loss induced error = -^-(100), [%]

o

Core loss induced error = °'00145 (100), [%]

Core loss induced error = 2.07, [%]

Design Performance
A current transformer was built and the data recorded in Table 16-1. It was plotted in Figure 16-10, with
an error of 3.4 %. The secondary winding resistance was 6.5 ohms.

Table 16-1
Current Transformer Electrical Data
lin
amps volts amps volts amps volts
0.250 0.227 1.441 1.377 3.625 3.488
0.500 0.480 2.010 1.929 3.942 3.791
0.746 0.722 2.400 2.310 4.500 4.339
1.008 0.978 2.693 2.593 5.014 4.831
1.262 1.219 3.312 3.181 5.806 5.606

5.0

4.0

~- 3.0

I
U 2.0

i.o
I I
1.0 2.0 3.0 4.0 5.0
Output Voltage, VQ (volts)

Figure 16-10. Current Transformer, Input Current versus Output Voltage.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Chapter 17

Winding Capacitance and Leakage Inductance

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table of Contents

1. Introduction
2. Parasitic Effects
3. Leakage Flux
4. Minimizing Leakage Inductance
5. Winding Capacitance
6. Winding Capacitance Turn-to-Turn
7. Winding Capacitance Layer-to-Layer
8. Capacitance Winding-to-Winding
9. Stray Capacitance
10. References

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Introduction
Operation of transformers at high frequencies presents unique design problems due to the increased
importance of core loss, leakage inductance, and winding capacitance. The design of high frequency power
converters is far less stringent than designing high frequency, wide-band audio transformers. Operating at
a single frequency requires fewer turns, and consequently, there is less leakage inductance and less
capacitance with which to deal. The equivalent circuit for a two-winding transformer is shown in Figure
17-1.
Lr
R

Figure 17-1. Equivalent Transformer Circuit.

High frequency designs require considerably more care in specifying the winding specification. This is
because physical orientation and spacing of the windings determine leakage inductance and winding
capacitance. Leakage inductance and capacitance are actually distributed throughout the winding in the
transformer. However, for simplicity, they are shown as lumped constants, in Figure 17-1. The leakage
inductance is represented by, Lp for the primary and, Ls for the secondary. The equivalent lumped
capacitance is represented by, Cp and Cs for the primary and secondary windings. The dc winding
resistance is, Rp, and Rs is for the equivalent resistance for the primary and secondary windings. Cw is the
equivalent lumped, winding-to-winding capacitance. R,, is the equivalent core-loss shunt resistance.

Parasitic Effects
The effects of leakage inductance on switching power supplies' circuits are shown in Figure 17-2. The
voltage spikes, shown in Figure 17-2, are caused by the stored energy in the leakage flux and will increase
with load. These spikes will always appear on the leading edge of the voltage switching waveform.

L,'(Leak
Energy = - [watt-seconds] [17-1]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Transformers designed for switching applications are normally designed to have minimum leakage
inductance, in order to minimize the voltage spikes, as shown in Figure 17-2. Also, leakage inductance can
be observed by the leading edge slope of the trapezoidal current waveform.

'on
Actual

Figure 17-2. Switching Transistor Voltage and Current Waveforms.

Transformers designed for power conversion are normally being driven with a square wave, characterized
by fast rise and fall times. This fast transition will generate high current spikes in the primary winding, due
to the parasitic capacitance in the transformer. These current spikes, shown in Figure 17-3, are caused by
the capacitance in the transformer; they will always appear on the lead edge of the current waveform and
always with the same amplitude, regardless of the load. This parasitic capacitance will be charged and
discharged every half cycle. Transformer leakage inductance and capacitance have an inverse relationship:
if you decrease the leakage inductance, you will increase the capacitance; if you decrease the capacitance,
you increase the leakage inductance. These are trade-offs that the power conversion engineer must make to
design the best transformer for the application.

T
P Light Load Heavy Load

Current Spike »•

^
2 3 4
ton ton
^
ton
•4 »- ton
-4 »•
^
Ideal Actual Ideal Actual

Figure 17-3. Transformer Capacitance Induced Current spike.

Leakage Flux

Leakage inductance is actually distributed throughout the windings of a transformer because of the flux set-
up by the primary winding, which does not link the secondary, thus giving rise to leakage inductance in
each winding without contributing to the mutual flux, as shown in Figure 17-4.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 However, for simplicity, leakage inductance is shown as a lumped constant in Figure 17-1, where the
leakage inductance is represented by Lp.

/
c=:^^^^^^=
^^^^
Leakage Flux — —*"Ls
-^ •s* -i
-r-^s r=- Primary
•^ c E5 ~
—
3
c
^ > Secondary
C >
EE Core ^ |=ii=^^^l
^

Figure 17-4. Leakage Flux.

In the layer-wound coil, a substantial reduction in leakage inductance, Lp and Ls, is obtained by
interweaving the primary and secondary windings. The standard transformer, with a single primary and
secondary winding, is shown in Figure 17-5, along with its leakage inductance, Equation [17-2]. Taking
the same transformer and splitting the secondary on either side of the primary will reduce the leakage
inductance, as shown in Figure 17-6, along with its leakage inductance, Equation [17-3]. The leakage
inductance can be reduced even more, by interleaving the primary and secondary, as shown in Figure 17-7,
along with its leakage inductance, Equation [17-4]. Transformers can also be constructed using the side-
by-side, sectionalized bobbin as shown in Figure 17-8, along with its leakage inductance, Equation [17-5].
The modified three section, side-by-side bobbin is shown in Figure 17-9, along with its leakage inductance
Equation [ 17-6].

Leakage flux
EE Core *- \
^ \ ^
££ t
Winding Area

Core Mating Surface— *- ->-*-

1
I
a = Winding length, cm
a b = Winding build, cm
Mutual Flux »-
Window — —<*- ^••••••••J
I1 \ '
c = Insulation thickness, cm

..., \
\v
H Secondary
Primary ~~*" J\ Insulation
-».

Figure 17-5. Conventional Transformer Configuration.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 EE Core »• ,- Primary
^"-
H~ i

|11
\Vinding Area |\
s
\
\ a = Winding length, cm
s,
Core Mating Surface —*- s a b = Winding build, cm
c = Insulation thickness, cm

W////M/A
Mutual Flux *-
Window — —\^*- 1EB^^^ \ _ Insulation
V \

b!-,
c
b2
c
•4 -b3 ^ N /2 Secondary each
s

Figure 17-6. Conventional, Transformer Configuration with Simple Interleave.

2
n(MLT\N
{ X
Lp= > " , [henrys] [17-3]

EE Core *- Np/2 Primary each

^

II
i

f^
a = Winding length, cm
Core Mating Surface —*-
\ a b = Winding build, cm
c = Insulation thickness, cm
Mutual Flux »•
\ Insulation
1
\

Window — —n*. l_^___£

\• V

Ib b b b \
Ns/2 Secondary each
c cc

Figure 17-7. Sectionalized, Transformer Configuration Primary and Secondary Interleave.

EE Core —*• ' Secondary

Bobbin ? \
/
y i
Winding Area • a
\r ± a = Winding length, cm
Core Mating Surface
Mutual Flux
Window
»•
-**•
s
< ... V..
I
1
f
=*=

^—
\> N
JT-
a
c

^ Insulation
x
Primary
b - Winding build, cm
c = Insulation thickness, cm

-H b
Figure 17-8. Pot Core, Sectionalized Transformer Configuration.

2
4x(MLT)N
V
^ " (l(T 9 ), [henrys] [17-5]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 EE Core Secondary
Bobbin
Winding Area 0
a = Winding length, cm
Core Mating Surface a b = Winding build, cm
c = Insulation thickness, cm
Mutual Flux
Window Insulation
Np/2 Primary

Figure 17-9. Modified, Pot Core Sectionalized, Transformer Configuration.

7r(MLT}Nl(
V Za\ «A
Lp= ' ' Zc + — v(KT 9 ), [henrys] [17-6]
b I 3 j

Minimizing Leakage Inductance

Magnetic core geometry has a big influence on leakage inductance. To minimize leakage inductance, the
primary winding should be wound on a long bobbin, or tube, with the secondary wound as close as
possible, using a minimum of insulation. Magnetic cores can have identical rating, but one core will
provide a lower leakage inductance than the other. A simple comparison would be two cores with the same
window area, but one core has twice the winding length. Only half the winding build is shown in Figure
17-10.

EE Cores. Bobbin Secondary

\ / Bobbin
EE Cores
Primary /

Special Long Tongue EE Core Standard EE Core Configuration

Figure 17-10. Comparing a Standard EE Core and a Special Long Tongue Core.

If layers must be used, the only way to reduce the leakage inductance is to divide the primary winding into
sections, and then to sandwich the secondary winding between them, as shown in Figure 17-7. This can
pose a real problem when designing around the European VDE specification, because of the required
creepage distance and the minimum insulation requirements between the primary and secondary.
Minimizing the leakage inductance on a push-pull converter design could be a big problem. A special
consideration is required symmetry in both the leakage inductance and dc resistance; this is in order to get a
balanced winding for the primary switching circuit to function properly.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The best way to minimize the leakage inductance, and to have a balanced dc resistance in a push-pull or
center-tapped winding, is to wind bifilar. Bifilar windings will drastically reduce leakage inductance. This
condition also exists on the secondary, when the secondary is a full-wave, center-tapped circuit. A bifilar
winding is a pair of insulated wires, wound simultaneously and contiguously, (i.e., close enough to touch
each other); Warning: do not use bifilar wire or the capacitance will go out of sight. Each wire constitutes a
winding; their proximity reduces leakage inductance by several orders of magnitude, more than ordinary
interleaving. This arrangement can be applied to the primary, to the secondary, or, it can be applied to the
primary and secondary together. This arrangement will provide the minimum leakage inductance.

Winding Capacitance

Operating at high frequency presents unique problems in the design of transformers to minimize the effect
of winding capacitance. Transformer winding capacitance is detrimental in three ways: (1) winding
capacitance can drive the transformer into premature resonance; (2) winding capacitance can produce large
primary current spikes when operating from a square wave source, (3) winding capacitance can produce
electrostatic coupling to other circuits.

When a transformer is operating, different voltage gradients arise almost everywhere. These voltage
gradients are caused by a large variety of capacitance throughout the transformer, due to the turns and how
they are placed throughout the transformer. When designing high frequency converters, there are several
factors that have a control over the turns: (1) the operating flux density or core loss; (2) the operating
voltage levels in the primary and secondary; (3) the primary inductance.

Keeping turns to a minimum will keep the capacitance to a minimum. This capacitance can be separated
into four categories: (1) capacitance between turns; (2) capacitance between layers; (3) capacitance
between windings; and (4) stray capacitance. The net effect of the capacitance is normally seen by the
lumped capacitance, Cp, on the primary, as shown in Figure 17-1. The lumped capacitance is very difficult
to calculate by itself. It is much easier to measure the primary inductance and the resonant frequency of the
transformer or inductor, as shown in Figure 17-11. Then, calculate the capacitance using Equation [17-7].
The test circuit, in Figure 17-11 functions as follows: The input voltage, VI, is held constant while
monitoring the voltage, V2, sweep through the frequency with the power oscillator. When the voltage, V2,
rises to a peak, and starts to decay at this peak voltage, the transformer or inductor is in resonance. At this
point the phase angle is also 0 degrees at resonance when looking at both the curves of VI and V2.

Ca = —~ = — — — , [farads] [17-7]
' '

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Power Audio Oscillator Dual Channel Oscilloscope

Vert. #1, VI
Vert. #2, V2

UUT

Vert #2 Input Voltage

Figure 17-11. Circuit for Measuring either a Transformer or Inductor Self Resonates.

For transformers designed to operate with a square wave, such as dc-to-dc converter, leakage inductance,
Lp, and the lumped capacitance, Cp, should be kept to a minimum. This is because they cause overshoot
and oscillate, or ring, as shown in Figure 17-12. The overshoot oscillation, seen in Figure 17-12A, has a
resonant frequency, f, that is controlled by, Lp and Cp. This resonant frequency could change and change
drastically after potting, depending on the material and its dielectric constant, as shown Figure 17-12B.

(17-12-A) (17-12-B)

Figure 17-12. Primary Voltage with Leading Edge Ringing.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Winding Capacitance Turn-to-Turn

The turn-to-turn capacitance, Cb, shown in Figure 17-13, should not be a problem if you are operating at
high frequency, low voltage power converters, due to the low number of turns. If the turn-to-turn
capacitance is important, then change the magnet wire insulation to one with a lower dielectric constant.
See Chapter 4.

<w> ' o
C
t

Primary
Cp
f
C
s~ -~^

t
f^ ~> N p
s s
cs
Secondary
J^
s~ ->
r» , r>

^
Figure 17-13. Capacitance Turn-to-Turn.

Winding Capacitance Layer-to-Layer

The capacitance between layers on the primary or secondary is the best contributor to the overall, lumped
capacitance, Cp. There are three ways to minimize the layer capacitance: (1) Divide the primary and
secondary windings into sections, and then sandwich the other winding between them, as shown in Figure
17-7; (2) The foldback winding technique, shown in Figure 17-14, is preferred to the normal U type
winding, even though it takes an extra step before starting the next layer. The foldback winding technique
will also reduce the voltage gradient between the end of the windings; (3) Increasing the amount of
insulation between windings will decrease the amount of capacitance. But remember, this will increase the
leakage inductance. If the capacitance is reduced, then the leakage inductance will go up. There is one
exception to this rule, and that is, if the windings are sandwiched or interleaved, it will reduce the winding
capacitance, but, it will increase the winding-to-winding capacitance.

Foldback Winding U Type Winding

Figure 17-14. Comparing the Foldback to the U Type Winding.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Transformers and inductors wound on toroidal cores can have capacitance problems, just as much if care is
not taken in the design at the beginning. It is difficult to control the winding capacitance on a toroidal core
because of its odd configuration, but there are ways to control the windings and capacitance. The use of
tape barriers to mark a zone for windings, as shown in Figure 17-15, offers a good way to control this
capacitance.
Tape Barrier

Winding area

- Toroidal Core

Figure 17-15. Tape Barrier for Winding Toroidal Core.

Another way to help reduce the capacitance effect on toroids is to use the progressive winding technique.
The progressive winding technique example is shown in Figure 17-16 and 17-17: Wind 5 turns forward and
wind 4 turns back, then wind 10 turns forward and keep repeating this procedure until the winding is
complete.

Continuous winding Back wind

Core

Figure 17-16. Progress Winding Top View.

Expanded View
Back Wind

Figure 17-17. Progress Winding Side View.

Capacitance Winding-to-winding

Balanced windings are very important in keeping down noise and common mode signals that could lead to
in-circuit noise and instability problems later on. The capacitance, from winding-to-winding, shown in
Figure 17-18, can be reduced, by increasing the amount of insulation between windings. This will decrease
the amount of capacitance, but again, this will increase the leakage inductance. The capacitance effect

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 between windings can be reduced, without increasing the leakage inductance noticeably. This can be done,
by adding a Faraday Shield or screen, as shown in Figure 17-19, between primary and secondary windings.

A Faraday Shield is an electrostatic shield, usually made of copper foil. The Faraday Shield is normally
added along with the insulation between primary and secondary. In some designs, the Faraday Shield can
consist of three independent insulated shields or just one. It all depends on the required noise rejection.

Primary Secondary

Figure 17-18. Capacitance, Cw, Winding-to-Winding.

Primary Secondary

Shield #1 Shield #2

Figure 17-19. Transformer with a Primary and Secondary Shield.

Stray Capacitance
Stray capacitance is very important to minimize because it too, can generate asymmetry currents and could
lead to high common mode noise. Stray capacitance is similar to winding-to-winding capacitance except
that the capacitance is between the winding next to the core, Cc, and the outer winding next to the
surrounding circuitry, Cs, as shown in Figure 17-20. Stray capacitance can be minimized by using a
balanced winding, or using a copper shield over the entire winding. A means for measuring leakage current
is shown in Figure 17-21. The winding-to-winding capacitance can be calculated, using Equations [17-8]
and [17-9].

C = -, [farads] [17-9]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Core -Surrounding Circuitry

Electrostatic

Figure 17-20. Transformer Winding with Stray Capacitance.

Power Audio Oscillator Digital Voltmeter

Frequency

1111IIIIIIII11111111
Voltage

Figure 17-21. Test Circuit for Measuring Primary and Secondary, ac Leakage Current.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 References
1. Grossner, N., "Transformer for Electronic Circuits." McGraw-Hill, New York, 1967.
2. Landee, R., Davis, D., and Albecht, A., "Electronic Designer's Handbook," McGraw-Hill, New York,
1957, p. 17-12.
3. Lee, R., "Electronic Transformer and Circuits," 2nd ed., John Wiley & Sons, New York, 1958.
"Reference Data for Radio Engineers," 4th ed., International Telephone and Telegraph Co., New York.
4. Richardson, I., The Technique of Transformer Design, Electro-Technology, January 1961, pp. 58-67.
Flanagan, W., "Handbook of Transformer Application." McGraw-Hill, New York, 1986.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Chapter 18

Quiet Converter Design

The author would like to thank Dr. V. Vorperian, Senior Engineer, Power and Sensor Electronics Group,
Jet Propulsion Laboratory (JPL), for his help with the Quiet Converter design equations.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table of Contents

1. Introduction
2. The Voltage-fed Converter
3. Regulating and Filtering
4. The Current-fed Converter
5. The Quiet Converter
6. Regulating and Filtering
7. Quiet Converter Waveforms
8. Technology on the Move
9. Window Utilization Factor, Ku
10. Temperature Stability
11. Calculating the Apparent Power, P,
12. Quiet Converter Design Equations
13. Transformer Design, Using the Core Geometry, Kg, Approach
14. Design Review
15. References

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Introduction
A few designers have known about the Resonant Converter described here for many years. This type of
Resonant Converter has been built mainly in the range of 200 watts to 2 kilowatts, and has been used as a
static inverter. However, it has remained relatively obscure in the general literature. The Quiet Converter
was developed at Jet Propulsion Laboratory (JPL), Division 38, to power very sensitive instruments. The
Quiet Converter produces a sinusoidal voltage across a parallel resonant tank. The dc output voltage is
obtained after rectification and filtering of the sinusoidal secondary voltage. The regulation is achieved by
controlling the duty-cycle of the switching transistors. A comparison of the standard type of PWM control
with the Quiet Converter and its amplitude modulation (AM), is shown in Figure 18-1. The inherent low
noise from this converter is how the nickname, Quiet Converter, came about. The low noise can easily be
reduced, even further, by the addition of a Faraday Shield and common-mode inductors. Programs at Jet
Propulsion Laboratory (JPL) that have successfully used the low noise environment of the Quiet Converter
are, WF/PC-II, Articulated Fold, Mirror Actuators, (Hubbell Space Telescope), MISR (Earth Orbiting
System), Raman, and Mars 05 ONC, CCD Camera.

Square wave with PWM control Sine wave with amplitude control

Figure 18-1. Comparing PWM and Amplitude Control.

The Voltage-fed Converter

The voltage-fed converter circuit is the most widely-used, converter topology. In a voltage-fed converter,
the powder source, Vjn is connected directly to the transformer through a transistor, Ql, as shown in Figure
18-2. When the transistor, Ql, is switched on, the full source voltage is applied to the transformer, Tl,
primary, (1-2). The transistor saturation will be ignored. Conversely, when Q2 is switched on, the full
source voltage is applied to the other half of the transformer, Tl, primary, (2-3).

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 LI

Figure 18-2. Typical, Voltage-fed Power Converter.

In Figure 18-2, the switching drive circuit alternately saturates and cuts off the semiconductors' switches,
Ql and Q2, causing an alternating voltage to be generated across the primary winding of transformer, Tl,
and then delivered to the secondary to be rectified and filtered before going to the load. The primary
source voltage, Vin is directly impressed onto the primary of the transformer, Tl, and therefore, the voltage
across the transformer, Tl, is always a square wave.

Regulating and Filtering

The most effective method of regulation for a voltage-fed converter is pulse width modulation (PWM). A
constant output voltage can be obtained for a changing input voltage, by reducing the on time, Ton of Q1
and Q2, as shown in Figure 18-3. The pulse width voltage is applied to the output filter, L1C2, averaging
circuit to provide the proper output voltage, V0.

A 0
Low Line

Vin

B 0 t

High Line

Figure 18-3. Primary Voltage of a PWM Controlled Converter.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The Current-fed Converter
The main difference between a voltage-fed converter and a current-fed converter is the series inductor, LI,
shown in Figure 18-3. The inductor, LI, is commonly called a feed-choke or series inductor. It has an
inductance large enough in value to maintain a continuous current through the circuit under all conditions
of line and load.

CR1 _^ T
M
PI °~
4 \
5 C2
+
-^
S^
6 v0

CR2
n
\
Figure 18-4. Typical Current-fed Power Converter Circuit.

The Quiet Converter

Simple additions to the circuit, in Figure 18-4, change the performance dramatically, and it becomes a
whole new converter. The new converter is shown in Figure 18-5. The changes are: 1. The transformer,
Tl, core material has been changed to molypermalloy powder core, (MPP). The reason for using a powder
core is because it has a built-in gap required for the tank circuit and these cores are available with
temperature stabilized permeability. The use of a gap ferrite would perform just as well, but the design
must be stable over temperature. 2. A commutating winding has been added to the series inductor, LI.
3. A capacitor, C3, was added for the required parallel tuned tank. The tuning capacitor, C3, should be of
high quality with a low ESR and stable. The capacitors that were used in the flight power supplies, were
plastic film, type CRH, to MIL-C-83421.

1.L1 2
L2
CR2
>j +
4
5 C2
V
S v
6 o

kj
PI
CR3 1
n-

Figure 18-5. Current-fed Parallel Resonant Converter.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 With properly designed components, the output voltage of transformer, Tl, will always be a sine wave.
The sine wave is accomplished by using a tuned parallel resonant tank circuit, (T1C3), to the natural
running frequency of the converter. The series inductor, LI, isolates the input dc source from the sine wave
voltage across the primary of the transformer, Tl.

Regulating and Filtering

The current-fed resonant converter, shown in Figure 18-5, requires a minimum of dead time, (dwell), for
the circuit to function properly. The series inductor, LI, when connected, as shown in Figure 18-4, requires
continuous conduction of both Ql and Q2, along with a small amount of overlap. In this way, there would
always be continuous current flowing in LI. If there is any disruption of current in the series inductor, LI,
no matter how small, it would destroy the switching transistors, Ql and or Q2.

In order to incorporate pulse width modulation (PWM), or a drive circuit that has inherent dead time that
neither transistor is conducting, there must be a means to commutate the current in the series inductor, LI.
Adding a winding to the series inductor, LI, is a simple way to commutate the current. When the current
flowing in winding, (1-2), is interrupted, the current will now be commutated to the added winding, (3-4).
This is done when connected with proper phasing, through a diode CR1, then, back to the dc source to
complete the path, as shown in Figure 18-5. Now, when either transistors, Ql or Q2, are interrupted, the
added winding of the series inductor, LI, commutates the current back into the dc source, thus preventing
the destruction of the switching transistors Ql and Q2.

Quiet Converter Waveforms

The current-fed, sine wave converter waveforms will be referenced from Figure 18-6. In Figure 18-7
through Figure 18-15, refer back to (A)-(J) points in Figure 18-6. The waveforms presented here are
copies drawn from an actual photo taken with an oscilloscope camera.

1.L1 2

-O

Figure 18-6. Quiet Converter Schematic with Reference Points.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 \
Dr ain t 3 Ground Volt,ige
iX" "*" X" "N

Af NN A/

(n, f
Horizontal = 5|isec/cm
^
Figure 18-7. Drain to Ground, Voltage Waveform of Ql and Q2.

The drain voltage waveform of Ql is shown in Figure 18-7. Waveform is taken between points A and G.
The converter is properly tuned to the natural frequency.

Drain to Ground Voltage

Horizontal -- 5usec/cm

Figure 18-8. Drain to Ground, Voltage Waveform of Ql and Q2.

The drain voltage waveform of Ql is shown in Figure 18-8, with minimum dead time. Waveform is taken
between points A and G. The converter is properly tuned to the natural frequency.

* 1
eadiiig Power Factc

/ K N

A/
Sfx
hi\l / NIN

Ho rizorital ='- 5|is ec/crn

Figure 18-9. Drain to Ground, Voltage Waveform of Ql and Q2.

The drain voltage waveform of Ql is shown in Figure 18-9. Waveform is taken between points A and G.
The converter is improperly tuned to the natural frequency. The resonant tank capacitor is too small in
value.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 I i
T a gging Power ] 7acto

<v N \[\
A A
'V N
\hi
I
H Drizontal == 5^ sec/c m

Figure 18-10. Drain to Ground, Voltage Waveform of Ql and Q2.

The drain voltage waveform of, Ql, is shown in Figure 18-10. Waveform is taken between points A and G.
The converter is improperly tuned to the natural frequency. The resonant tank capacitor is too large in
value.
\ ±
— Voltaj*e Across Primary or Tank Circuit —

^
N /
v~^ \
\\
/

/
/. A' "\ \ : /'
"^
i

V
^ ^
H<>iizo ital ~-- JSn
M 56C/C1
\\

Figure 18-11. Voltage Waveform Across Transformer Primary.

The primary voltage waveform is shown in Figure 18-11, across transformer, Tl. Waveform is taken
between points A and B. The converter is properly tuned to the natural frequency.

±
Voltage Across Secondary

\f ^^ \ / ^
\
L An \ \
y ii
.
/'"
^; J \
Hr)rizoiital == 5[is ec/cr
\[

Figure 18-12. Voltage Waveform Across Transformer Secondary.

The secondary voltage waveform of transformer, Tl, is shown in Figure 18-12. Waveform is taken
between points D and E. The converter is properly tuned to the natural frequency.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Jt
Rec tifie dOu tpuf v^olta ge ,

\/ x*"x

\,f 1 /
, V,

VVs
V .V
Horizontal = 5usec/cm
^
I.
Figure 18-13. Secondary Rectified Voltage Waveform at CR2 and CR3.

The secondary, rectified voltage waveform, at the cathodes of, CR2 and CR3, is shown in Figure 18-13.
Waveform is taken at point H. The converter is properly tuned to the natural frequency.

4- 1
Secondary Current

' ~ ™ ' —- '-- ™ -~.

Horizontal == 5usec/cm

Figure 18-14. Secondary Current Waveform.

The secondary current waveform is shown in Figure 18-14. The current waveform is taken at point F.

j 1 i
Current Through the Commutating Diode

n r n n
Horizontal = 5usec/cm

Figure 18-15. Current Through the Commutating Diode, CR1.

Commutating diode current waveform is shown in Figure 18-15. The current is through the series inductor
LI winding (3-4). Waveform is taken at point C. The converter is properly tuned to the natural frequency.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Technology on the Move

As technology moves ahead, instruments become more sophisticated, smaller in size, and require less
power. Less power normally relates to lower current. Lower current requires smaller wire to carry the
current. There is a practical point where the wire size can no longer be reduced, even though the current is
very small. Reliability is affected when the wire size becomes very small. It becomes a handling and
termination problem. If a larger wire size can be tolerated, and it does not impact the size a great deal,
then, the larger wire should be used. The smallest wire size that seems to be tolerable, depending on the
application, ranges from #35 to #39 AWG and this would be from a specialty house.

Window Utilization Factor, Ku

When designing a transformer or inductor, the window utilization factor, Ku, is the amount of copper that
appears in the window area. See Chapter 4. The window utilization factor, Ku, is influenced by five main
factors:

1. Wire insulation, S,.

2. Wire lay fill factor, S2.
3. Effective window area, S3.
4. Winding insulation, S4.
5. Workmanship.

These factors multiplied together will give a normalized window utilization factor of Ku = 0.4.

(5 3 )(S 4 ) = 0.4 [18-1]

The design of the current-fed sine wave converter is much more detailed and complex, compared to the
simple voltage-fed, square wave converter. The sole reason to use the Quiet Converter is because of its
inherent low noise, (EMI). The noise of the Quiet Converter can be, reduced even further by adding a
primary and a secondary Faraday Shield. When a Faraday Shield is added between the primary and
secondary, the transformer must be designed to accommodate the shield. Transformer size is, mainly
determined by the loads. The window utilization, Ku has to be adjusted during the design to accommodate
the Faraday Shield. When the core size is selected for the transformer, it will be a little larger core, do to
the added space required by the Faraday Shield.

After the preliminary design, the engineer will select the proper core size for the power transformer. The
core geometry, Kg will select the molypermalloy powder core size. After the molypermalloy powder core
size has been selected, the engineer will now select a core with a permeability best-suited for the
application. The molypermalloy powder cores come with a range of permeability from 14 to 550, all with
the same core geometry, Kg.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Temperature Stability

For the Quiet Converter to function properly over a wide temperature range, the components must be stable
over that temperature range. The components that control the oscillator frequency must be stable. The LC
tank circuit must be stable and not drift with temperature. Molypermalloy powder cores are offered with
stabilized permeability, with code letters M, W, and D from Magnetics, Inc. The W material temperature
stability is shown in Figure 18-16.

1)
00
0.2
•u
0. V/ - Material Stabilization +/- 0.25%
0.2
04
-60 -40 -20 0 20 40 60 80 100 120
Temperature, Degrees C

Figure 18-16. Typical, Stabilized Molypermalloy Material.

Calculating the Apparent Power, Pt

The apparent power, Pt is the power associated with the geometry of the transformer. The designer must
be able to make allowances for the rms power in each winding. The primary winding handles, P^, the
secondaries handle, P0, to the load. Since the power transformer has to be designed to accommodate the
primary power, Pin and the secondary, P0, then by definition,

P,=P»+PL> [watts]

Pin=, [watts] [18-2]

*7

P,=~ + Pi, [watts]

ri
77 = efficiency

The designer must be concerned with the apparent power-handling capability, P, of the transformer core
and winding. The apparent power, P, may vary by a factor ranging from 2 to 2.828 times the input power,
Pj n , depending upon the type of circuit in which the transformer is used. If the current in the transformer
becomes interrupted, such as a center-tapped secondary or push-pull primary, its effective rms value
changes. Transformer size is thus determined not only by the load demand, but also by application,
because of the different copper losses incurred owing to current waveforms.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Because of the different winding configurations, the apparent power, Pt, of the transformer will have to be
summed to reflect these differences. When the winding has a center tap and produces a discontinuous
current, then the power in that winding, whether it is primary or secondary, has to be multiplied by the
factor, U, to correct for the rms current in that winding. If the winding has a center tap, then, U = 1.41; if
not, then, U = 1. Summing the output power of a multiple-output transformer would be:

[18-3]

Quiet Converter Design Equations

The transformer secondary voltage, Vs, is:

V0 = Output voltage
Vd = Diode Drop

The maximum secondary true power, PS(max), is:

[watts]
V«))>
The minimum secondary true power, PS(min> is:

The secondary apparent power, Psa, is:

U = 1 .4 1 , center tapped winding
U = 1.0, single winding

If, there is more than one output, then, sum the total secondary maximum apparent load power, Psal.

^1=^01+^02 +........., [watts] [18-8]

If, there is more than one output, then, sum the total secondary maximum load power, Pot(max)-

^(max) = ^0, (m ax) + ^02 (max) + .......... . [Wa«s] [18-9]

If, there is more than one output, then, sum the total secondary minimum load power, P0t(min>

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The maximum reflected secondary load resistance, R<max), is:
R<max) = Resistance Value
r) = Efficiency

R,(max) -, [ohms] [18-11]

The required series inductor inductance, LI, is:

f = fundamental frequency

> [henrys]

The total period, T, is:

= —, [seconds] [18-13]

The maximum transistor on time, ton(max), is:

Transistor drive circuits such as a pulse width modulator (PWM), will have a minimum of dead time,
Dead time or dwell is shown in Figure 18-17.

Qi
•t
;
Dead Time or Dwell

Q2 l
on

Figure 18-17. Transistor Drive Waveforms, Showing Dead Time or Dwell.

The conversion ratio, Ka, is:

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The peak voltage, Vc(pk), on the resonant capacitor, C3, as shown in Figure 18-5, is:
Kb = 2, center tapped winding.
Kb = 1, single winding.

V = U\ " '" b
', [volts] [18-16]
2

The primary rms voltage, Vp(mis), is:

Kb = 2, center tapped winding.
Kb = 1, single winding.

[18-17]

The primary maximum reflected secondary current, Ips, is:

pon max)

The secondary reflected loads to the primary, RSR, is:

Kb = 2, center tapped winding.
Kb = I, single winding.

_ ii
Vp(rrns)
,, \ b, J
p(rrns) -, noir\i
r l
RSR= -!—!- , [ohms] [18-19]
sp

Note: The capacitance reactance affects the total percentage of harmonic distortion when:

coRSRC = \, « [12%], a>RSRC = 2, «

As a general rule:
[18-20]

The resonant capacitance, Cx, is:

QT, is a variable that provides the engineer a little latitude with the capacitance value. (1 < Qj < 3)

[18-21]

The reactance, Xcx, of capacitor, Cx, is:

Use a standard capacitor.

[ohms] [18 22]

' "

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The capacitor rms current, I CX (rms)> is;

. [amps)

The total primary current, 1^,™), is:

The primary tank inductance, Lx, is:

The total transformer apparent power, Pt, is:

P, =(Primary VA) + (Secondary VA) + (Capacitor VA), [watts]

p.=\

The core geometry, Kg, is:

Kf is the waveform factor = 4.44
Bac is the operating flux density and its value is an engineering judgment based on the frequency and core
material.

, [cm5] [18-27]
0.000029

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Transformer Design, Using the Core Geometry, Kg, Approach
The following information is the Design specification for a 2.2 watt push-pull transformer, operating at
32kHz, using the Kg, core geometry approach. For a typical design example, assume a push-pull, full wave
bridge circuit, with the following specification:

1. Input voltage, V(rnin) ................................................................................................. = 22 volts

2. Output voltage #1, VS(M ........................................................................................... = 5.0 volts
3. Output current #1, I s0 i( m ax) ......................................................................................= 0-2 amps
4. Output current #1, I s oi(min) ....................................................................................... = 0.1 amps
5. Output voltage #2, Vs02 ............................................................................................ = 12.0 volts
6. Output current #2, Iso2(max) ...................................................................................... = 0.1 amps
7. Output current #2, Iso2(min) ....................................................................................... = 0.05 amps
8. Frequency, f ............................................................................................................. = 32 kHz
9. Switching dead time, td ............................................................................................ = 0.625usec
10. Efficiency, r) .......................................................................................................... = 95%
11. Regulation, a ......................................................................................................... = 1.0 %
12. Diode voltage drop, Vd .......................................................................................... = 0.5 volt
13. Operating flux density, Bac .................................................................................... = 0.05 tesla
14. Core Material ........................................................................................................ = MPP
15. Window utilization, Ku .................................................... ...................................... = 0.4
16. Temperature rise goal, Tr ....................................................................................... = 15°C
17. Waveform coefficient, Kf ...................................................................................... = 4.44
18. Notes:

Using a center tapped winding, U = 1 .41

Using a single winding, U = 1.0

Step 1 . Calculate the total secondary voltage, Vs for each output.

), [volts]
K J 0 1 =(5.0) + (1.0) = 6.0, [volts]
K S 0 2 =(12) + (1.0) = 13.0, [volts]

Step 2. Calculate the maximum secondary true power, P S(max)-

= 6.0(0.2) = 1.2, [watts]

= 13.0(0.1) = 1.3, [watts]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step 3. Calculate the minimum secondary true power, Ps(inin).

^o, ( m ,n>= 6-0(0.1) = 0.6, [watts]

^02( m i n ) = 13-0(0.05) = 0.65, [watts]

Step 4. Calculate the secondary apparent power, Psa.

P M01 =6.0(0.2)(1.0) = 1.2, [watts]

Psa02 = 13.0(0.1)(1.0) = 1.3, [watts]

Step 5. Calculate the secondary total maximum apparent load power, Psay.

P s f l 2 =(l.2) + (l.3), [watts]

P sflS =2.5, [watts]

Step 6. Calculate the secondary total maximum load power, Pot(max)-

=
^rf(max) ^oOl(max) + Po02(mm)' [watts]

^ ( ™ x ) = ( l - 2 ) + (l-3), [watts]
fymax) = 2-5, [watts]

Step 7. Calculate the secondary total minimum load power, Pot(min)-

^(n™)=l-25, [watts]

Step 8. Calculate the secondary maximum reflected load resistance, R(max).

R(max> = Resistance Value
r) = Efficiency

of (min)

(22) 2 (0.95)
125 '

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step 9. Calculate the inductance of the series inductor, LI.

L1= V^
3co [henrys]
Ll
= i / w vT > [henrys]
3(2(3.14)(32000))
Ll = 0.000610, [henrys]

Step 10. Calculate the total period, T.

T = —, [seconds]

T= , [seconds]
32,000
T = 31.25, [micro-seconds]

Step 11. Calculate the maximum transistor on time, Ton(max). Dead time is shown in Figure 18-17.

5
> [usec]

^ (max) =15,

Step 12. Calculate the conversion ratio, Ka.

K=-
TSinl
T )

K (4(15)~(32.25))
' (32.25)54^^
v
'
v 32.25
OT Tt

AT,, =0.866

Step 13. Calculate the peak voltage, Vc(pk), on the resonant capacitor, C3, as shown in Figure 18-5.
Kb = 2, center tapped winding.
Kb = 1, single winding.

n(K
l
V, Kh\
Vc(pk)= °2'" b), [volts]

= (3.1415)(0.866)(22)(2)

^)= 59.85, [volts]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step 14. Calculate the primary rms voltage, Vp(mlS).
Kb = 2, center tapped winding.
Kb = 1, single winding.

^(-)=—JT^-' tvolts]
0.707(59.85)
Vp(rms)= \ -, [volts]
yp(rms)=2l.2, [volts]

Step 15. Calculate the primary maximum reflected secondary current, Ip

p
, o/(max) r -.
/„ = „ , [amps]

/ „ = 0.124, [amps]

Step 16. Calculate the secondary reflected loads to the primary, RSR.
Kb = 2, center tapped winding.
Kb = 1, single winding.

P
*«= 7' V ' , [ohms]
sp

(0.866)(21.2)(2)z2
/?„=- — ^- -, [ohms]
0.124
^ = 592, [ohms]

Note: The capacitance reactance effects the total percentage of harmonic distortion when:

coRSRC = l, ~ [12%], coRSRC = 2, «

As a general rule:
C = ^ ,, [farads]

Step 17. Calculate the resonant capacitance, Cx.

2
C = -;-- , [farads]
"
x
~ (6.28)(32000)(592)'
Cx =1.68(lO' 8 ), [farads]
Cx = 0.0168 use a 0.015, [micro-farads]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step 18. Calculate the reactance, Xcx, of capacitor, Cx. Use a standard capacitor. Let, Cx, equal 0.015(if.

, [ohms]

a -~i- ; -^ »
(6.28)(32,000)(0.015(lO-6))
Xcx = 332, [ohms]

Step 19. Calculate the capacitor current, I CX (nns)-

_(0.707)(59.85)
« ( n » ) - ^ >

/ cx(rms)
, , =0.127,' L[amps]
F J

Step 20. Calculate the total primary current, Itp(m,S).

Step 2 1 . Calculate primary tank inductance, Lx.

;—
2
'
, [henrys]

L = ---—.-r, [henrys]
(6.28) (32000) (0.015 (lO-
^=0.00165, [henrys]

Step 22. Calculate the total transformer apparent power, Pt.

Pt = (Primary VA) + (Secondary VA) + (Capacitor VA), [watts]

(Pl( ,(U\\ : ,

P, = -^^ + (P^) + KbV Ici , [watts]

P,=ll.6, [watts]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step 23.Calculate the core geometry, Kg. Bac is the operating flux density and its value is an engineering
judgment based on the frequency and core material.

, [cm5]
2
0.000029 (Kf)

[0.000029(4.44) (32,000) (0.05) (l)J

^=0.00793, [cm5]

Design Review
Conversion factor, Ka 0.866
Tank Capacitance, Cx 0.015 uf
Tank Capacitance Peak Voltage, Vcx 59.85 volts
Tank Capacitance rms Current, ICX(rms) 0.127 amps
Primary Inductance, Lx 0.00165 henrys
Series Inductor, LI 0.000610 henrys
Primary Reflected Current, Ips(mis) 0.124 amps
Primary rms Voltage, Vp(m)s) 21.2 volts
Primary Total rms Current, Itp(rms) 0.177 amps
Secondary Total Load Power, P0t(max) 2.5 watts
Transformer Total Apparent Power, Pt 11.6 watts
Transformer Core Geometry, Kg 0.0107 cm2

Step 24. From Chapter 3, select a MPP powder core, comparable in core geometry, Kg.

Core part number 55848-W4

Manufacturer Magnetics
Magnetic path length, MPL = 5.09 cm
Core weight, Wtfe = 9.4 gm
Copper weight, Wtcu =11.1 gm
Mean length turn, MLT = 2.8 cm
Iron area, Ac = 0.226 cm2
Window area, Wa = 1.11 cm2
Area product, Ap = 0.250 cm4
Core geometry, Kg = 0.008 cm5
Surface area, At = 22.7 cm2
Permeability, ja = 60
Millihenrys per 1000 turns, AL = 32

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step 25. Calculate the total number of primary turns, Ntp.

/, =1000 \-^L L[turns]

J
IP \\ , 1
^(1000)

V 32
|f, B*.]
N, = 226, round-down [turns]
Np =113, [each side center tap]

Step 26. Calculate the operating flux density, Bac.

. , [tesla]
K,NpfAc '
(21.2)(10 4 )
"c " (4.44)(1 13)(32000)(0.226) '
Bac = 0.0587, [tesla]

Step 27. Calculate the watts per kilogram, WK, using the MPP 60 perm loss equation in Chapter 2.

WK = 0.788(lO'3 )(/)° 41) (Bac ) (224) , [watts/kilogram]

WK = 0.788(lO"3)(32000)° 41) (0.0587)(224) , [watts/kilogram]

WK = 3.09, [watts/kilogram] or 3.09, [milliwatts/gram]

Step 28. Calculate the core loss, Pfc.

f milliwatts | TT , /. „ 1, \
Pfe = W,fe 10- , [watts]
^ grams )
P /e =(3.09)(9.4)(lO' 3 ), [watts]
Pfe = 0.0290, [watts]

Step 29. Calculate the volts per turn, K N/

N/V-

N
KNIV=—?-, [turns/volt]

' [turns/volt]
"" (21.2)'
KN/V = 5.33, [turns/volt]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step 30. Calculate the secondary number of turns Ns. a is regulation in percent. See Chapter 6.

ioo
,o, = KNIV VM K = (5.33)(6.0)(1.01) = 32, [turns]
.0)(l.Ol) = 70, [turns]

Step 31. Calculate the current density, J, using a window utilization, Ku = 0.4.

J = -'-- , [amps per cm2 ]

ApBmfKfKu'
(11.6)10"
J = --—- ,, -—-——- , [amps per cm" ]
(0.25)(0.0587)(32000)(4.44)(0.4)
7 = 139, [amps per cm 2 ]

Step 32. Calculate the secondary required wire area, Aws.

A oi
01 = y = 139 = 1.44*10-3, [cm2]
M

3
. - , [cm2]
139

Step 33. Then select the wire from the Wire Table, in Chapter 4. Record uQ/cm.

Awsm = 1.44xlO"3, use #26 = 1.28*1 (T3, [cm2]

#26,^ = 1345
cm
A,o2 = 0.7 1 9*1 (T3, use #29 = 0.647*10"3, [cm2]

#29,^ =
cm

Step 34. Calculate the primary required wire area, A wp.

. tp(rms) i
Aup = —-j—- ,, [cm
L 1J

A
(0.177)
u ,=~--.
m [cm2]
" 139
Awr= \.27x\Q-\ [cm2]

Step 35. Then select the wire from the Wire Table, in Chapter 4. Record u,fi/cm.

Awp = 1.27*1 (T3, use #26 = 1.28*1 0~3, [cm2]

#26,^ = 1345
cm

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step 36. Calculate the total secondary window utilization, Kuts.

(N0.A0,) (32)(0.00128)
A: s01 = v "' "m> = ±—^ 1 = 0.0369
" Wa 1.11

„ s _ V( N 002
, A . ,O, Jo)2 7 _ v(70)(0.000647)
A i = 0.0408
"- wa 1.11
Klltt=KusOI+Klls02= 0.0777

Step 37. Calculate the primary window utilization, Kup.

(226)(0.00128)

^=0.261

Step 38. Calculate the total window utilization, Ku.

Ku=Kllp+Kuts
Ku =0.261 + 0.0777
K,, = 0.339

Step 39. Calculate the primary, winding resistance, Rp.

x lO-6, [ohms]
cm
Rr =2.80(113)(1345)(10- 6 ), [ohms]
Rp = 0.426, [ohms]

Step 40. Calculate the primary, copper loss, Pp.

p> [watts]

Pp =(0.177) 2 (0.426), [watts]

^=0.0133, [watts]

Step 4 1 . Calculate the secondary, winding resistance, Rs.

| | ( l Ov - 6 ) , [ohms]
Vcmj '
RsOI =2.80(32)(l345)(lO" 6 ) = 0.121, [ohms]
Ri02 =2.80(70)(2664)(lO- 6 ) = 0.186, [ohms]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step 42. Calculate the secondary, copper loss, Ps.

Psm = (0.2)2 (0.121) = 0.00484, [watts]

Psm = (o.l) 2 (0.186) = 0.00186, [watts]

Step 43. Calculate the total secondary, copper loss, Pts.

P,s=Psm+P^ [watts]
Pts =0.00484 + 0.00186, [watts]
Pa = 0.0067, [watts]

Step 44. Calculate the total loss, core and copper, Pz.

Pz=P,,+Pa+Pfi, [watts]
Pz =(0.0133) + (0.0067) + (0.0290), [watts]
Pz = 0.049, [watts]

Step 45. Calculate the watts per unit area, \\i.

p
- r[watts per cm2-,]
yr =—=-,
4
(0.049)
w = —.-~- , [watts per cm" ]
(22.7)
y = 0.002 1 6, [watts per cm2 ]

Step 46. Calculate the temperature rise, Tr.

Tr =450(0.00216) ( ° 826) ,
Tr= 2. 83, [°C]

Step 47. Calculate the, Qt, of the tank.

Q,=
Q, = (6.28)(32000)(o.015(lO- 6 ))(592)
e,=i.79
For more information see Equation [18-20].

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 References

1. V. Vorperian, and C. McLyman, "Analysis of a PWM-Resonant DC-to-DC Converter." IEEE

transaction.
2. S. Lendena, "Current-Fed Inverter." 20th Annual Proceedings Power Sources Conference, May 24
1966.
3. S. Lendena, "Single Phase Inverter for a Three Phase Power Generation and Distribution System."
Electro-Optical-System, Contract #954272, from Jet Propulsion Laboratory, January 1976.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Chapter 19

Rotary Transformer Design

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table of Contents

1. Introduction
2. Basic Rotary Transformer
3. Square Wave Technology
4. Rotary Transformer Leakage Inductance
5. Current-fed Sine Wave Converter Approach
6. Rotary Transformer Design Constraints
7. References

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Introduction
There are many requirements to transfer signals and power across rotary interfaces. Most things that use
slip rings or brushes can be replaced with a rotary transformer. Science instruments, antennas and solar
arrays are elements needing rotary power transfer for certain spacecraft (S/C) configurations, such as a
spin, stabilized (S/C). Delivery of signals and power has mainly been done by slip rings. There are
problems in using slip rings for long life and high reliability: contact wear, noise, and contamination.
Contact wear will lead to a conductive path to ground. This conductive path will generate noise and upset
the original designed common-mode noise rejection. A simple slip ring assembly and a rotary transformer
are shown in Figure 19-1. High data rates and poor slip ring life forced the Galileo (S/C) to replace the
signal interface with rotary transformers. The use of a rotary transformer to transfer power on the Galileo
(S/C) was contemplated, but it was thought the impact on the (S/C) delivery was too great. The rotary
transformers on the Galileo (S/C) lasted the life of the spacecraft, from 1989 to 2003 without a glitch.

Slip Rings Rotary Transformer

Conductive /TVTVTVTA Power Out Stator

Rings Rotor
Power Out Through-Bore

Rotor

Brushes
Power In
Through-Bore

Power In
Figure 19-1. Comparing a Slip Ring Assembly and a Rotary Transformer.

Existing approaches to rotary power transfer use square wave converter technology. However, there are
problems caused by the inherent gap in a rotary transformer, coupled with the fast rate of change in the
square wave voltage. Undue stress is placed on the power electronics and the interface becomes a source of
Electromagnetic Interference (EMI) that impacts the overall system's operating integrity.

Basic Rotary Transformer

The rotary transformer is essentially the same as a conventional transformer, except that the geometry is
arranged so that the primary and secondary can be rotated, with respect to each other with negligible
changes in the electrical characteristics. The most common of the rotary transformers are the axial rotary

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 transformer, shown in Figure 19-2, and the flat plane, (pot core type), rotary transformer, shown in Figure
19-3. The power transfer is accomplished, electro-magnetically, across an air gap. There are no wearing
contacts, noise, or contamination problems due to lubrication or wear debris.

Primary Winding Secondary Winding

Stator
A

Power In \ Outer Bearing

, Flux Path Rotary
Platform

Figure 19-2. Pictorial of an Axial, type Rotary Transformer.

Primary Winding Gap

Power In
Flux Path

Bearing Rotary
Platform

Stator Power Out

Secondary Winding

Figure 19-3. Pictorial of a Flat Plane, type Rotary Transformer.

Square Wave Technology

The ideal converter transformer would have a typical square B-H loop, as shown in Figure 19-4. A
converter transformer is normally designed to have a minimum of leakage inductance. The voltage spikes
that are normally seen on the primary of a square wave converter transformer are caused by the leakage
inductance. To design a converter transformer to have a minimum of leakage inductance, the primary and
secondary must have a minimum of distance between them. Minimizing the leakage inductance will

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 reduce the need for power-wasting, snubber circuits. Although there are rotary power transformers
designed with the use of square wave converter technology, they are not, without problems.

B (tesla)

H (oersted)

Figure 19-4. Typical, Transformer BH Loop.

There are two basic problems not found in the normal transformer: (1) the inherent gap in a rotary
transformer is one problem, and (2) the required spacing between primary and secondary that leads to large
leakage inductance is the other. These problems, along with a square wave drive, are what leads to a high
loss, snubber circuit, and beccme a source of Electromagnetic Interference (EMI) that impacts the
adjoining systems operating integrity. The rotary transformer, because of its inherent gap, has a B-H loop
similar to an inductor, as shown in Figure 19-5. Basically, the transformer transforms power, and the
inductor stores energy in the gap. The rotary transformer does not have any of the traits of an ideal
transformer. It is, more accurately, a trans-inductor having a gap and a secondary, spaced away from the
primary.

B (tesla)

H (oersted)
Stored Energy

Figure 19-5. Typical, Rotary Transformer BH Loop.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Rotary Transformer Leakage Inductance

The rotary transformer has an inherent gap and spacing of the primary and secondary. The gap and spacing
in the rotary transformer result in a low primary magnetizing inductance. This low primary inductance
leads to a high magnetizing current. The leakage inductance, Lp, can be calculated for both axial and flat
plane using Equation 19-1. The axial rotary transformer winding dimensions are shown in Figure 19-6.
The flat plane rotary transformer winding dimensions are shown in Figure 19-7.

47i(MLT)N;
c +- (l(T 9 ), [henrys] [19-1]

„a, bl
^tntnr ^>
-1+- 1
55^^^^
Ss^^^s^
* .
g^^^^ I
T? ntnr ^

]
Thr ou gh -B ore
^b 2

a = Winding length, cm
_ b = Winding build, cm
c = Space between windings, cm

Figure 19-6. Axial Rotary Transformer, Showing Winding Dimensions.

Primary
Secondary

Through-Bore

a = Winding length, cm
b = Winding build, cm
Stator c = Space between windings, cm

Air Gap

Figure 19-7. Flat Plane Rotary Transformer, Showing Winding Dimensions.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Current-fed Sine Wave Converter Approach

The current-fed, sine wave converter topology is a good candidate to power the rotary transformer. The
design would be a current-fed, push-pull, tuned tank converter requiring a gapped transformer. A
comparison between a standard, square wave converter, shown in Figure 19-8, and a current-fed, sine wave
converter, is shown in Figure 19-9. Using the rotary transformer in this topology, the energy that is stored
in the rotary gap that causes so much trouble in the standard square wave driving a rotary transformer, is
recovered and is used in the tank circuit. There would not be any need of power-wasting snubbers using
the rotary transformer approach. See Chapter 18.

LI
CR1
n
4
C2
_CR3 ± •s
2 k 1
6

Snubber Circuit
CR2
_
I
f\

Figure 19-8. Typical, Voltage-fed, Square wave Converter Circuit with Snubbers.

LI 2
1 Tuned Tank Circuit ^
-L,

C2
CR4
v

CR3
-O.

Figure 19-9. Typical, Current- fed, Resonant Converter Circuit.

The current-fed sine wave converter requires a resonant, LC, tank circuit to operate properly. The primary
of the rotary transformer would be the ideal inductor, because of the inherent gap of the rotary transformer.
There are several advantages to incorporating the resonant tank circuit into the rotary transformer. First, it
minimizes the number of components in the power stage. Secondly, the output of the inverter is a natural
sine wave, as shown in Figure 19-10, and usually requires no additional filtering. Thirdly, energy stored in
the gap of the transformer is released when either power switch is turned off. This energy is commutated in
the resonant tank circuit. This provides the capability for direct exchange of power between the tank circuit

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 and the load. There is not a noticeable drive torque in a rotary transformer. The tuning or tank capacitor
must be of high quality, stable, and with low ESR.

Voltage Across Secondary

Horizontal = 1 0 usec/cm

Figure 19-10. Current-fed Converter, Secondary Sine Wave Secondary Voltage.

Rotary Transformer Design Constraints

The rotary transformer requirements pose some unusual design constraints compared to the usual
transformer design. The first is the relatively large gap in the magnetic circuit. This gap size depends on
the eccentric dimension and the tolerance of the rotating shaft. The gap results in a low primary
magnetizing inductance. Secondly, the large space separating primary and secondary windings results in
an unusually high primary-to-secondary leakage inductance. Thirdly, the large through-bore requirement
results in an inefficient utilization of the core material and copper, due to the fixed mean-length turn. This
large diameter results in requiring more copper area for the same regulation. Finally, the core has to be
more robust than the normal transformer because of the structural requirement. See Figure 19-11.

Winding Separation Air Gap

Stator -
Primary
Secondary Primary Secondary
Rotor Stator Rotor

Through-Bore
Through-Bore

Air Gap
t
Axial Flat Face-to-Face

Figure 19-11. Geometries of the Basic Type Rotary Transformers.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Rotary transformer dimensions are usually governed by the mechanical interface, in particular the relatively
large gap and the large through-bore, resulting in a long Mean Length Turn (MLT). The rotary transformer
is not an ideal magnetic assembly. A toroidal core is an ideal magnetic assembly. Manufacturers use test
data, taken from toroidal cores, to present magnetic material characteristics. The magnetic flux in a
toroidal core travels through a constant core cross-section, Ac, throughout the whole Magnetic Path Length,
MPL, as shown in Figure 19-12, and provides ideal magnetic characteristics. It can be seen that the core
cross-section throughout the rotary transformers, shown in Figure 19-13 and Figure 19-14, does not provide
constant flux density or an ideal magnetic assembly. The rotary transformers for the Galileo spacecraft
were about 10 cm in diameter, and manufactured by CMI (Ref 4.)

Magnetic Path Length, MPL

Iron Area, Ac

Perspective View
Figure 19-12. Typical Perspective View of a Toroidal Core.

A View
Magnetic Path Length Stator Rotor

A View Primary II Secondary Perspective View

Figure 19-13. Open View of a Flat Plane, Type Rotary Transformers.

Magnetic Path Length A View —, Stator

Primary Secondary
A View I Perspective View

Figure 19-14. Open View of an Axial Type Rotary Transformers.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 References

1. E. Landsman, "Rotary Transformer Design." Massachusetts Institute Technology, PCSC-70 Record,

pp. 139-152
2. L. Brown, "Rotary Transformer Utilization in a Spin Stabilized Spacecraft Power System." General
Electric, pp 373-376.
3. S. Marx, "A Kilowatt Rotary Power Transformer." Philco-Ford Corp., IEEE Transactions on
Aerospace and Electronic Systems Vol. AES-7, No. 6 November 1971.
4. Ceramic Magnetics, Inc. 16 Law Drive Fairfield, NJ 07006. Tel. (973) 227-4222.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Chapter 20

Planar Transformers

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table of Contents

1. Introduction
2. Planar Transformer Basic Construction
3. Planar Integrated PC Board Magnetics
4. Core Geometries
5. Planar Transformer and Inductor Design Equations
6. Window Utilization, Ku
7. Current Density, J
8. Printed Circuit Windings
9. Calculating the Mean Length Turn, MLT
10. Winding Resistance and Dissipation
11. PC Winding Capacitance
12. Planar Inductor Design
13. Winding Termination
14. PC Board Base Materials
15. Core Mounting and Assembly
16. References

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Introduction
The planar transformer, or inductor, is a low profile device that covers a large area, whereas, the
conventional transformer would be more cubical in volume. Planar Magnetics is the new "buzz" word in
the field of power magnetics. It took a few engineers with the foresight to come up with a way to increase
the power density, while at the same time increasing the overall performance, and also, making it cost
effective. One of the first papers published on planar magnetics was by Alex Estrov, back in 1986. After
reviewing this paper, you really get a feeling of what he accomplished. A whole new learning curve can be
seen on low profile ferrite cores and printed circuit boards if one is going to do any planar transformer
designs. It is an all-new technology for the transformer engineer. The two basic items that made this
technology feasible were the power, MOSFETs that increased the switching frequency and enabled the
designer to reduce the turns, and the ferrite core, which can be molded and machined into almost any shape.
After this paper was written the interest in planar magnetics seems to increase each year.

Planar Transformer Basic Construction

Here, shown in Figure 20-1 through Figure 20-4 are four views of a typical EE core, planar construction
method. The assembled planar transformers have very unique characteristics in their finished construction.
In the assembled planar transformer, every primary turn is at a precise location, governed by the PC board.
The primary is always the same distance from the secondary. This provides a tight control over the
primary to secondary leakage inductance. Using the same insulating material will always provide the same
capacitance between primary and secondary; in this way, all parasitics will be the same from unit to unit.
With this type of planar construction, the engineer will have a tight control over leakage inductance, the
resonant frequency, and the common-mode rejection. A tight control is necessary on all materials used.

Alignment Pins or Terminals Double E Ferrite Core Insulation

^

*— PC Boards
Secondary
Primary
Side View

Figure 20-1. Side View of a Typical EE Planar Transformer.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Double E Ferrite Core Insulation Alignment Pins or Terminals

• PC Boards
Secondary — Primary

End View

Figure 20-2. End View of a Typical EE Planar Transformer.

Double E Ferrite Core ^ Alignment Pins or Terminals

Q\ y
w
Insulation
>n — >~ r
-4— - PC Boards

Top View Primary

Secondary -
^-4 k. /'

6/ o

Figure 20-3. Top View of a Typical EE Planar Transformer.

Planar EE Cores

Insulation

Figure 20-4. A Perspective View of a Typical EE Planar Transformer.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Planar Integrated PC Board Magnetics
Planar transformers and inductors are now being integrated right on the main PC board. Design engineers
are pushing the operating frequency higher and higher to where it is commonplace to operate at frequency
range between 250-500kHz. As the frequency increases the power supplies are getting smaller and smaller.
To reduce the size of the power supply even further engineers are going to planar magnetics that are
integrated into the main PC board. An exploded view to show the multi-layers PC board of a planar
transformer that has been integrated into the main PC board is shown in Figure 20-5. The final assembly of
the same planar transformer is shown in Figure 20-6.

1/2 Planar EE Magnetic Core

Multi-layer PCB

1/2 Primary Layer 1

1/2 Secondary

1/2 Secondary

1/2 Primary

Inerconnection

1/2 Planar EE Magnetic Core

Figure 20-5. A Planar Transformer Integrated into the Main PC Board.

Multi-layer PCB

Figure 20-6. PC Board Planar Transformer in Final Assembly.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Core Geometries
The EE and El are not the only planar geometries now available. There are a few firms in the ferrite
industry that offer low profile versions of their standard cores, giving the engineer a few more choices in
his design. There are EE and El cores available from Magnetics Inc. as shown in Figure 20-7; there are ER
cores available from Ferroxcube, as shown in Figure 20-8; there are ETD-lp cores available from Ferrite
International, as shown in Figure 20-9; there are PQ-lp cores available from Ferrite International, as shown
in Figure 20-10; and there are RM-lp cores available from Ferroxcube, as shown in Figure 20-11. There
are several advantages, with cores with a round center post, such as PQ-lp, RM-lp, ETD-lp and ER. A
round center post results in a more efficient use of copper and a more efficient use of board space. There is
a company, Ceramic Magnetics, Inc. (CMI), that can modify any of these cores to your specification or
machine a special core for your application. The IEC has a new standard 62313 for planar cores that
supercedes standard 61860.

Matting Set E or I
EE or El Planar Ferrite Cores Perspective View

Figure 20-7. Magnetic Inc. EE and El Low Profile Planar Cores.

D
ER Ferrite Core Perspective View

Figure 20-8. Ferroxcube ER Low Profile Planar Cores.

(5 t
1 I

E B
\f
V 1

Fs
C D
ETD-lp Ferrite Core Perspective View

Figure 20-9. Ferrite International ETD Low Profile Planar Cores.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 B

D C
PQ Ferrite Core, low profile Perspective View

Figure 20-10. Ferrite International PQ Low Profile Planar Cores.

W
a ^
i k -*.-

E B G —i*~ -^
r

1r
;,-
-— ^-~
RM-lp Ferrite Core Perspective View

Figure 20-11. Ferroxcube RM Low Profile Planar Ceres.

Planar Transformer and Inductor Design Equations

The same design equations are used, as well as the criteria used to select the proper core, to design a planar
transformer as a conventional transformer. Faraday's Law is still used to calculate the required turns:

N= , [turns] [20-1]

The core power handling equation, Ap:

A = , [cm4] [20-2]
KfKJAcBacJ
The gapped inductor equation, L:

QAnN2A C
L= , . , [henrys] [20-3]
U—
\ p.)
The core energy handling equation, Ap:

2(Energy)
P
K..BJ

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Window Utilization, Ku
The window utilization factor in the conventional transformer is about 0.40. This means that 40% of the
window is filled with copper, the other 60% of the area is devoted to the bobbin or tube, to the insulation
both layer and wire, and to the winding technique. The window utilization is explained, in detail, in
Chapter 4. Designing a planar transformer and using the PC winding technique, reduces the window
utilization factor even further. The window utilization, Ku, comparison of the two different winding
techniques is shown in Figure 20-12.

Standard Transformer Typical Planar Transformer

Copper Area

Window Area, Wa

- Ku = 0.188

Figure 20-12. Comparing the Window Utilization of a Standard Transformer and a Planar Transformer.

A PC board window utilization, K u , calculation example will be as fellows:

The windings will be placed on a double-sided 2oz PC board 10 mils thick, giving a total thickness of 15.4
mils (0.0391 cm). The Mylar insulation material is between the PC boards, and between the PC boards,
and the core will add another 4 mils (0.0102 cm) to the thickness. This will give 19.4 mils (0.0493 cm) per
layer. There will be a 20 mil space (margin) between the edge of the board and the copper clad. The
copper width will be the window width of 0.551cm minus 2x the margin of 0.102. This will give a total
copper width of 0.449. The window utilization, Ku, will be summed in Table 20-1, using Figure 20-13 as a
guide.

Wa = 0.164cm 2
Core Center Leg Mylar Insulation Sheet = 0.0102 cm

Double Sided PC Board = 0.0391 cm

0.051 cm
Margin
0.051 cm
Margin -

0.551 cm
EI-42216, Window

Figure 20-13. Window Utilization of a Typical El Planar Transformer.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table 20-1

EI-422 1 6 Window Utilization

Window Height, cm 0.2970
Window Width, cm 0.5510
Window Area, cm 0.1640
PC Board Thickness with Copper, cm 0.0391
Sheet Insulator, cm 0.0102
Total Insul. 5+1 Layers Thick, cm 0.0612
Total Thickness 5 Layers, cm 0.2570
Copper Thickness 5 Layers, cm 0.0686
Copper Width, cm 0.4494
Total Copper Area, cm 0.0308
Window Utilization, Ku 0.1878

Current Density, J

One of the unknown factors in designing planar transformers is the current density, J. The current density
controls the copper loss (regulation) and the inherit temperature rise caused by the copper loss. The
temperature rise is normally controlled by the surface dissipation of the transformer. The size of a
transformer goes up by the cubic law, and the surface area goes up by the square law. Large transformers,
such as 60 Hz, are designed with a low current density, while 400 Hz are designed with higher current
density for the same temperature rise. There used to be an old rule of thumb, for a large transformer, you
use 1000 circular mils per amp, and for a small transformer, you use 500 circular mils per amp:

500CM/Amp ~ 400Amps/cm2, [400 Hertz Aircraft]

2
1000CM/Amp « 200Amps/cm , [60 Hertz]

Planar transformer designers handle the current density in a different way. When designing planar
transformer PC windings, designers use the same technology used by the printed, circuit board designers,
and that is the current rating for a given voltage drop and temperature rise. It is another way of saying the
same thing. The printed circuit boards are covered with a copper clad. The thickness of this copper is
called out in ounces, such as loz, 2oz, and 3oz. The weight in ounces comes from an area of one square
foot of material. So loz of copper clad would be 1 square foot, and have a thickness of 0.00135 inch; 2oz
would be 0.0027 inch; and 3oz would be 0.00405 inch. Tables have been made to show the current
capacity for a constant temperature rise with different line width. The design data for 1 oz copper is shown
in Table 20-2. The 2oz copper is shown in Table 20-3, and 3oz copper is shown in Table 20-4. Planar
transformer engineers are using the industrial guidelines for their selection of copper trace thickness and
line width, based on temperature rise. The first effort for a planar transformer, PC winding should be
around:

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 lOOCM/Amp « 2000Amps/cm2, [500 kHertz Planar Transformers]

If the current density is based on Table 20-1, with a line width of 0.06 inches, then use:

35CM/Amp « 5700Amps/cm2, [500 kHertz Planar Transformers]

Table 20-2. Design Data for 0.00135 Inch Thick Copper Clad.

*Printed Circuit Trace Data for loz Copper (Based on 10 Inches Long)
Line Line Resistance Copper Weight loz Temp. °C Increase above Amb. Vs.
Width Width micro-ohm Thickness 0.00 135 Current in Amperes
2
Inches mm per-mm cm **AWG 5° 20° 40°
0.0200 0.51 989.7 0.000174 35 1.00 3.00 4.00
0.0400 1.02 494.9 0.000348 32 2.25 5.00 6.50
0.0600 1.52 329.9 0.000523 30 3.00 6.50 8.00
0.0800 2.03 247.4 0.000697 29 4.00 7.00 9.50
0.1000 2.54 197.9 0.000871 28 4.50 8.00 11.00
0.1200 3.05 165.0 0.001045 27 5.25 9.25 12.00
0.1400 3.56 141.4 0.001219 26 6.00 10.00 13.00
0.1600 4.06 123.7 0.001394 26 6.50 11.00 14.25
0.1800 4.57 110.0 0.001568 25 7.00 11.75 15.00
0.2000 5.08 99.0 0.001742 25 7.25 12.50 16.60
*Data From: Handbook of Electronic Packaging.
**This is a close approximation to an equivalent AWG wire size.

Table 20-3. Design Data for 0.0027 Inch Thick Copper Clad.

*Printed Circuit Trace Data for 2oz Copper (Based on 10 Inches Long)
Line Line Resistance Copper Weight 2oz Temp. °C Increase above Amb. Vs.
Width Width micro-ohm Thickness 0.0027 Current in Amperes
2
Inches mm per-mm cm **AWG 5° 20° 40°
0.0200 0.51 494.9 0.000348 32 2.00 4.00 6.25
0.0400 1.02 247.4 0.000697 29 3.25 7.00 9.00
0.0600 1.52 165.0 0.001045 27 4.25 9.00 11.25
0.0800 2.03 123.7 0.001394 26 5.00 10.25 13.25
0.1000 2.54 99.0 0.001742 25 5.25 11.00 15.25
0.1200 3.05 82.5 0.002090 24 5.75 12.25 17.00
0.1400 3.56 70.7 0.002439 23 6.25 13.25 18.50
0.1600 4.06 61.9 0.002787 23 6.50 14.25 20.50
0.1800 4.57 55.0 0.003135 22 7.00 15.25 22.00
0.2000 5.08 49.5 0.003484 22 7.25 16.25 24.00
*Data From: Handbook of Electronic Packaging.
**This is a close approximation to an equivalent AWG wire size.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table 20-4. Design Data for 0.00405 Inch Thick Copper Clad.

*Printed Circuit Trace Data for 3oz Copper (Based on 10 Inches Long)
Line Line Resistance Copper Weight 3oz Temp. °C Increase above Amb. Vs.
Width Width micro-ohm Thickness 0.00405 Current in Amperes
2
Inches mm per-mm cm **AWG 5° 20° 40°
0.0200 0.51 329.9 0.000523 30 2.50 6.00 7.00
0.0400 1.02 165.0 0.001045 27 4.00 8.75 11.00
0.0600 1.52 110.0 0.001568 25 4.75 10.25 13.50
0.0800 2.03 82.5 0.002090 24 5.50 12.00 15.75
0.1000 2.54 66.0 0.002613 23 6.00 13.25 17.50
0.1200 3.05 55.0 0.003135 22 6.75 15.00 19.50
0.1400 3.56 47.1 0.003658 22 7.00 16.00 21.25
0.1600 4.06 41.2 0.004181 21 7.25 17.00 23.00
0.1800 4.57 36.7 0.004703 20 7.75 18.25 25.00
0.2000 5.08 33.0 0.005226 20 8.00 19.75 27.00
*Data From: Handbook of Electronic Packaging.
**This is a close approximation to an equivalent AWG wire size.

Printed Circuit Windings

There will be a few paths of mystery along the way when engineers first get started in the design of a planar
transformer. Therefore, it would be much easier to start on a simple design and use magnet wire, then
convert that into a truly all planar approach, using a PC winding board design. In this way the engineer
will slide up the learning curve slowly. There are several benefits to a printed circuit winding. Once the
printed winding board is finished and the layout is fixed, the winding will not vary and all of the parasitics,
including the leakage inductance, will be frozen. This is not necessarily true in conventional transformers.
There are two basic core configurations available to the engineer for planar design. The first configuration
is the EE or El with the rectangular center post. A typical high current and low current winding PC board
for E cores is shown in Figure 20-14.

J
^^^^^"^^^^^^^^^^^~' - — Inner Connecting Eyelets

- High Current Trace

PC Board

Low Current Trace

Figure 20-14. Typical Planar E Core Winding PC Board.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The second configuration is shown in Figure 20-15. These are four cores with round center legs. Winding
PC boards with round center legs are used on PQ-lp, RM-lp, ETD-lp and ER cores. There is an advantage
to cores with round center legs. Cores with round center leg will produce a round ID, OD resulting in a
more efficient use of copper.

Inner Connecting Eyelets

High Current Trace

— Low Current Trace

PC Board

Figure 20-15. Typical Circular Winding PC Board for Cores with Round Center Leg.

Calculating the Mean Length Turn, MLT

The Mean Length Turn (MLT), is required to calculate the dc winding resistance. With the winding
resistance known, the winding voltage drop can be calculated at rated load. The winding dimensions,
relating to the Mean Length Turn (MLT) for a rectangular winding, is shown in Figure 20-16, along with
the MLT equation, and a circular winding is shown in Figure 20-17, along with the MLT equation.

;•*-
C

, [mm] [20-5]

Figure 20-16. Dimensions, Relating to a Rectangular Winding, Mean Length Turn (MLT).

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 MLT = , [mm] [20-6]

Figure 20-17. Dimensions, Relating to a Circular Winding, Mean Length Turn (MLT).

Winding Resistance and Dissipation

The winding dc resistance and voltage drop will be calculated as follows:

Calculate the Mean Length Turn (MLT) using the winding board configuration and Equation in Figure 20-
17. Use the printed winding data in Table 20-5.

Table 20-5. PC Board Winding Data

PC Winding Data
Item Units
PC Board Turns Each Side 4
Winding Trace Thickness 0.0027 inches
Winding Trace Width 2.54 mm
Trace Resistance 99 HQ/mm
Winding Board, OD 31.5 mm
Winding Board, ID 14.65 mm
Winding Current, I 3 amps
PC Board Thickness 0.5 mm
PC Board Dielectric Constant, K 4.7

Step 1. Calculate the Mean Length Turn, MLT:

n(OD-
MLT = , [mm]

3.14(31.5 + 14.65)
MLT = -i-'-, [mm]

MLT = 72.5, [mm]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Step 2. Calculate the winding resistance, R:
6
R = MLT(N)\^-}(\0-
v }, [ohms]
'
/? = (72.5)(8)(99.0)(lO- 6 ), [ohms]
R = 0.057, [ohms]

Step 3. Calculate the winding voltage drop, Vw:

yw=IR, [volts]
F w =(3.0)(0.057), [volts]
F H ,= 0.171, [volts]

Step 4. Calculate the winding dissipation, Pw:

Pn=I2R, [watts]
2
PR =(3) (0.057), [watts]
Pw=0.5\3, [watts]

PC Winding Capacitance

The PC winding board traces will have capacitance, to the other side of the board as shown in Figure 20-18.
This capacitance could be to another winding, or a Faraday shield to ground.

PC Board
Upper Trace

Board Thickness
Board Capacitance

C
Under Side Trace

Figure 20-18. PC Board Trace Capacitance.

The formula for calculating the winding trace capacitance, to either another winding trace or ground plane,
is given in Equation 20-7.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 C
P=-
Where:
Cp = capacitance, [pf] [20-7]
K = dielectric constant
A = area of the trace, [mm2 ]
d= thickness of the PC board, [mm]

A typical square wave power converter, operating at 250kHz, will have extremely fast rise and fall times in
the order of 0.05 micro-seconds. This fast excursion will generate a fairly high current pulse depending on
the capacitance and source impedance.

The calculation of the winding capacitance is as follows:

Use the PC board winding data in Table 20-5, the outline drawing in Figure 20-19, and Equation 20-7:

- d = board thickness, (mm)

K = material dielectric constant

A = winding trace area = (trace width, mm)(MLT, mm)(N),

Figure 20-19. PC Board Winding Capacitance.

Step 1 . Calculate the winding trace area, A.

A = ( trace width, mm) ( MLT, mm) ( turns, N ) , [mm2 ]
y4=(2.54)(72.5)(8), [mm2]
A = 1473, [mm2]

Step 2. Calculate the winding capacitance, Cp.

0.0085(4.7)(1473)
- - [pf]
'
C =118, [pf]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Planar Inductor Design

Planar inductors are designed the same way as the conventional inductors. See Chapter 8. Planar inductors
use the same planar cores and PC winding board techniques as the transformers. The main difference is the
inductor will have a gap to prevent the dc current from prematurely saturating the core. It is normal to
operate planar magnetics at a little higher temperature than conventional designs. It is important to check
the maximum operating flux level at maximum operating temperature.

Fringing flux can be severe in any gapped ferrite inductor, but, even more so, on planar construction,
because of the printed winding board, as shown in Figure 20-20. When the flux intersects the copper
winding, eddy currents are generated, which produces hot spots and reduces the overall efficiency. The use
of a PC winding board, (flat traces), can give the eddy currents an added degree of freedom. The resulting
loss could be a disaster.

Double E Ferrite Core v

Fringing Flux
PC Boards Insulation

Figure 20-20. Fringing Flux Cutting Across PC Winding Boards.

Winding Termination

Making connections from a planar transformer to the outside world could be very clumsy, if not enough
thought is put in for termination. It has to be remembered that this is a high frequency transformer, and
skin effect, (ac resistance), has to be addressed. Because of the skin effect it is important the external leads
of the planar transformer must be keep as short as possible. Terminations are very important for currents of
one amp and above. A poor connection will only get worse. It is recommended to use plated-through holes
and eyelets, where possible, but cost will control that. If the transformer has many interconnections, or
only a few, there must be provisions made for those connections. When the PC winding boards are
stacked, and because of the high density, all connections and interconnections have to be done with
extended area pads, as shown in Figure 20-21. The PC winding boards require good artwork registry to
make sure the interconnections can be made between boards. Interconnections are usually done, by passing
a bus wire through a hole, and at the same time making the connection on the other board. If the solder
terminations are to be made on the board, then it is important to leave as much room as possible especially
if the connection is to be made with copper foil, as shown in Figure 20-22. When the PC windings have to
be paralleled, because of the increased current, the interconnecting jumpers will also have to be increased.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 PC Winding A PC Winding B

Figure 20-21. PC Winding Boards Showing Butterfly Pads.

Double E Ferrite Core

Bus Wire Interconnects

Plated Through Holes

PC Boards Insulation

Circuitry
Foil Lead
Solder Pad

PC Boards

Strain Relief Circuitry

Solder Pad
Exit Lead
PC Boards

Figure 20-22. PC Winding Boards, Showing Interconnections and Exit Leads.

PC Board Base Materials

PC Board materials are available in various grades, as defined by the National Electrical Manufacturers
Association (NEMA). The important properties for PC Board materials are tabulated in Table 20-6. It is
very important to choose the correct PC board material for your application. Planar transformers are
normally stressed to the last watt for a given temperature rise. This could give rise to hot spots at winding
terminations and cause PC Board discoloration. Due to their inherit design Planar transformers will have a
wide temperature delta, At. It would be wise to stay away from paper/phenolic materials and materials that
absorb moisture.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table 20-6. Properties of Typical Printer Circuit Board Materials

Properties of Typical Printed Circuit Board Materials

NEMA Grade
FR-1 FR-2 FR-3 FR-4 FR-5 G10 Gil
Material/Comments
Paper Paper Paper Glass/Cloth Glass/Cloth Glass/Cloth Glass/Cloth
Phenolic Phenolic Epoxy Epoxy Epoxy Epoxy Epoxy
Mechanical Strength good good good excellent excellent excellent excellent
Moisture Resistant poor good good excellent excellent excellent excellent
Insulation fair good good excellent excellent excellent excellent
Arc Resistance poor poor fair good good good good
Tool Abrasion good good good poor poor poor poor
Max. Cont. Temp. °C 105 105 105 130 170 130 170
Dielectric Constant, K 4.2 4.2 4.4 4.7 4.3 4.6 4.5

Core Mounting and Assembly

Core assembly and mounting should be strong and stable with temperature. One of the most viable
methods for securing core halves together is epoxy adhesive. There is one epoxy adhesive that has been
around a long time and that's 3M EC-2216A/B. This bonding technique is shown in Figure 20-23 and it
seems to work quite well. When the core halves are properly bonded with epoxy adhesive, there will be
little or no effect on the electrical performance. This means the epoxy adhesive added little or no gap to the
mating surface. Large temperature excursions are normal in planar magnetics. Care should be taken into
account for the coefficients of thermal expansion between the core and mounting surfaces. It has to be
remembered ferrite is a ceramic and is very brittle. Planar cores have a low silhouette with thin sections
that cannot absorb as much strain as other geometries. After the planar transformer has been assembled,
there should be a small amount of play in the PC winding assembly to guarantee there will be a minimum
of stress over temperature.

Epoxy Adhesive for

Securing Core Halves Planar EE Cores

Small Filet Bead for

Core Assembly Mounting

Figure 20-23. Epoxy Adhesive for Securing Transformer Assembly.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 References

1. Designing with Planar Ferrite Cores, Technical Bulletin FC-S8, Magnetics, Division of Spang and
Company 2001.
2. Brown, E., "Planar Magnetics Simplifies Switchmode Power Supply Design and Production," PCIM,
June 1992, pp. 46-52.
3. Van der Linde, Boon, and Klassens, "Design of High-Frequency Planar Power Transformer in
Multilayer Technology," IEEE Transaction on Industrial Electronics, Vol. 38, No. 2, April 1991, pp. 135-
141.
4. Bloom, E., "Planar Power Magnetics: New Low Profile Approaches for Low-Cost Magnetics Design,"
Magnetic Business & Technology, June 2002, pp. 26,27.
5. Charles A. Harper, Handbook of Electronic Packaging, McGraw-Hill Book Company, pp. 1-51-1-53.
6. Reference Data for Radio Engineers, Fourth Edition, International Telephone and Telegraph Corp.
March 1957, pp. 107-111.
7. PC Boards, Casco Circuits, Inc., 10039 D Canoga Ave., Chatsworth, CA 91311. Tel. (818) 882-0972.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Chapter 21

Derivations for the Design Equations

The author would like to thank Richard Ozenbaugh of Linear Magnetics for his help with the derivations.

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Table of Contents

1. Output Power, P0, Versus Apparent Power, Pt, Capability

2. Transformer Derivation for the Core Geometry, Kg
3. Transformer Derivation for the Area Product, Ap
4. Inductor Derivation for the Core Geometry, Kg
5. Inductor Derivation for the Area Product, Ap
6. Transformer Regulation

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Output Power, P0, Versus Apparent Power, Pt, Capability

Introduction

Output power, P0, is of the greatest interest to the user. To the transformer designer, the apparent power, Pt,
which is associated with the geometry of the transformer, is of greater importance. Assume, for the sake of
simplicity, that the core of an isolation transformer has only two windings in the window area, a primary
and a secondary. Also, assume that the window area, Wa, is divided up in proportion to the power-handling
capability of the windings, using equal current density. The primary winding handles, Pin, and the
secondary handles, P0, to the load. Since the power transformer has to be designed to accommodate the
primary, Pj,,, and, P0, then,

By definition:

Pt=P.n+Po, [watts]
P
m =—> [watts]

The primary turns can be expressed using Faraday's Law:

[tums]
[21-A2]

The winding area of a transformer is fully utilized when:

By definition the wire area is:

w=—, [cm2]
" J [21-A4]

Rearranging the equation shows:

J/ I Mv J / ,
' ' [21-A5]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Now, substitute in Faraday's Equation:

K.W = —-

Rearranging shows:

_ [cn<]
[21-A7]
The output power, P0, is:

P0=y,I,, [watts] [21.Ag]

The input power, Pin, is:

Pin=Vplp, [watts] [21_A9]

Then:

P,=P>n+Pa, [watts] [21-A10]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Transformer Derivation for the Core Geometry, Kg

Introduction
Although most transformers are designed for a given temperature rise, they can also be designed for a given
regulation. The regulation and power-handling ability of a core are related to two constants, Kg and Ke by
the equation:
P,=2KgKea, [watts] [21-Bl]

Where:
a = Regulation, [%]
The constant, Kg, is a function of the core geometry:
Kg=f(Ac,Wa,MLT) [21-B2]

The constant, Ke, is a function of the magnetic and electrical operating conditions:
Ke=g(f,Bm) [21-B3]
The derivation of the specific functions for, Kg and Ke, is as follows: First, assume there is a two-winding
transformer with equal primary and secondary regulation, as schematically shown in Figure 21-Bl. The
primary winding has a resistance of, Rp, ohms, and the secondary winding has a resistance of, Rs ohms:

Primary Secondary
n = Ns/Np = 1

Figure 21-Bl. Isolation Transformer.

AF \v
—i [21-B4]

The assumption, for simplicity, is that R^ is infinity (no core loss).

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 And:
/„ =/„ [21-B5]
Then:
[21-B6]

/ R
[21-B7]
p

Multiply the numerator and denominator by Vp:

a = 2 0 0 - [21-B8]

R VA
a = 200-^ [21-B9]
p
From the resistivity formula, it is easily shown that:

(MLT)yvP2
V
Rp = „, ' P [21-B10]

Where:
p=1.724(10" 6 )ohmcm
Kp is the window utilization factor (primary)
Ks is the window utilization factor (secondary)

Ku
1

Faraday's Law expressed in metric units is:

Vp=KffNpAcBm(\V*) [21-B12]
Where:
Kf = 4.0 for a square wave.
Kf = 4.44 for a sine wave.

Substituting Equation 21-BIO and 21-B12, for Rp and Vp, in Equation [21-B13]:

Ep2
VA= —a [21-B13]
200 Rp

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The primary VA is:

Simplify:
2 2
K]fA
f
B
c m
WK (l<r 10 )
VA= , "/ '-a [21-B15]
2(MLT)/7

Inserting 1.724(10'6) for p:

0.29K2f2A2B2WaK p (l(T
4
'
a [21-B16]
MLT

Let primary electrical equal:

is _ A O O r ^ 2 f2n2(tr\-4\ r^i r> 17T

^e = U . Z y A , / D I 1 U I L/l-r51/J

Let the primary core geometry equal:

W Ac2 K
K=
g " ", [cm5] [21-B18]
MLT

The total transformer window utilization factor is:

Kp+Ks=Ku
KU [21-B19]
p 2

When this value for Kp is put into Equation [21-B16], then:

VA = KeKga [21-B20]

Where:

Ke=0.l45K2f2B2m(lO-4) [21-B21]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The above VA is the primary power, and the window utilization factor, Ku, includes both the primary and
secondary coils.
W A2 K
K g = " c " , [cm5] [21-B22]
MLT

Regulation of a transformer is related to the copper loss, as shown in Equation [21-B23]:

a =^-(100), [%] [21-B23]

o

The total VA of the transformer is primary plus secondary:

Primary, VA = KeKga
plus [21-B24]
Secondary, VA = KeKga

The apparent power, Pt, then is:

P. = (Primary) KKn + (Secondary) Ke K, a

\ » * \ JJ [21-B25]
P=2KKa

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Transformer Derivation for the Area Product, Ap

Introduction
The relationship between the power-handling capability of a transformer and the area product, Ap can be
derived as follows.

Faraday's Law expressed in metric units is:

Where:
Kf = 4.0 for a square wave.
Kf = 4.44 for a sine wave.

The winding area of a transformer is fully utilized when:

By definition the wire area is:

Aw=—, [cm2]
J [21-C3]

Rearranging the equation shows:

[21-C4]

Now, substitute in Faraday's Equation:

I
KW = —
AcBaJKf\J

Rearranging shows:

[cm<]
[21-C6]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The output power, P0, is:

P0=V,I,, [watts] [21.c?]

The input power, P;n, is:

Pm=VpIp, [watts] [21 _ C8]

Then:

P,=Pin+P0, [watts] [21-C9]

Therefore:

WaAc = -i-'— , [cm4] [21-C10]

' BacfJKfKu

By definition:

Then:

A = ----— , [cm4] [21-C12]

' BacfJKfKu

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Inductor Derivation for the Core Geometry, Kg

Introduction
Inductors, like transformers, are designed for a given temperature rise. They can also be designed for a
given regulation. The regulation and energy-handling ability of a core are related to two constants, K g and
Ke, by the equation:

(Energy) 2 = KgKea, [21-D1]

Where:
a = Regulation, [%]
The constant, Kg, is a function of the core geometry:
Kx=f(Ac,Wa,MLl) [21-D2]
The constant, K,,, is a function of the magnetic and electrical operating conditions:
K.=s(P.,Bm) [21-D3]
The derivation of the specific functions for, Kg and Ke, is as follows: First, assume a dc inductor could be
an input or output as schematically shown in Figure 21-D1. The inductor resistance is RL-

% P^O
V1
pr>o 1+• ldc
r
1
111 i

C1
CR1 + CR2 C2 + <
vn 2^ ^ -> 2
k -> V <
^

1 \
- o>— r
Figure 21-D1. Typical Buck Type Switching Converter.

The output power is:

P.=I*Vo, [watts] [21-D4]

a = , [%] [21-D5]

The inductance equation is:

0.4^ N2 Ac (lO" 8 )
L= ^ '-, [henrys] [21-D6]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 The inductor flux density is:

QAxNI, (l(T 4 )
Bdc= -^ '-, [tesla] [21-D7]

Combine Equations [21-D6] and [21-D7]:

— = ^ '- [21-D8]

Solve for N:

, [turns] [21-D9]

From the resistivity formula, it is easily shown that:

_(MLT)N2p
RL = P
, [ohms] [21-D10]
WaKu
Where:
p=1.724(10- 6 )ohmcm

Combining Equations [21-D5] and [21-D10]:

KO /\
» WK*-"}
0 « j

Take Equation [21-D9] and square it:

f // V
—*M (lO 8 ) [21-D12]
^,fc A )

Combine Equations [21-D11] and [21-D12]:

« = l :/..,,. V||^r| (10'") [21-D13]

Combine and simplify:

/ i i*t-r T•^l
I (10'°) [21-D14]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Multiply the equation by Idc / Iac and combine:

a= -P (lO'°) [21-D15]

The energy equation is:

Energy = ——, [watt-second]

Q^
2Energy = Ll]c

Combine and simplify:

f(2Energy) 2 Vp(MLT) ..
a = ±- !- t-±- lO'° [21-D17]

The resistivity is:

p = 1.724 (10'6) [ohm cm] [21 -D18]

Combine the resistivity:

Solving for energy:

(Energy)2 =0.l45P0B2ac a [21-D20]

The core geometry equals:

W Ac 2K
K=
g " " , [cm5] [21-D21]
MLT

The electrical conditions:

[21-D22]

The regulation and energy-handling ability is:

(Energy)2 = KgKea [21-D23]

The copper loss is:

a =^-(100), [%] [21-D24]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Inductor Derivation for the Area Product, Ap

Introduction
The energy-handling capability of an inductor can be determined by the area product, Ap. The area
product, Ap, relationship is obtained by the following: (Note that symbols marked with a prime (such as
H"), are mks (meter-kilogram-second) units.)

dt dt

Combine and simplify:

dl

Flux density is:

<t> = BmAc [21-E3]

[21-E4]

'.+ —
I Am

[21-E5]
MPU

, [21-E6]
df

Combine Equations [21-E2] and [21-E6]:

" ^ ^ [21-E7]
01
'>!-
The energy equation is:

Energy = , [watt-seconds] [21-E8]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Combine Equations [21-E7] and [21-E8]:

LI2 'A'I2
Energy = [21-E9]
2 lf fMPL'

If Bm is specified:

/= M'-fT
v V A'm /y [21-E10]

Combine Equations [21-E7] and [21-E10]:

Energy = - [21-E11]

Combine and simplify:

Energy = - [21-E12]

The winding area of a inductor is fully utilized when:

[21-E13]

By definition the wire area is:

J [21-E14]

Combining Equations [21 -El 3] and [21 -El 4]:

KW=N-,
[21-E15]

Solving for I:

[21.E16]
N

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Rearrange Equation [21-El6]:

If W I .. p^gj^
g
\ Mn ) Bm

Now, substitute in Energy Equation [21-El 1]:

P
Energy = ^ '-— [21-El 8]
If,

Rearrange Equation [21-El8]:

Energy = | "" | [2i.E19]

r^o ) \ m

Combine and simplify:

Energy = I B"K'W'J' < I [21-E20]

Now, multiply mks units to return cgs.

J'=j(l04)
MPL =MPL(lO" 2 )

We can substitute into the energy equation to obtain:

Energy =B'»K"W"JA< (w4} [21-E21]

Solve for the area product:

2(Energy) [21-E22]
— -, [cm 1
BJK

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Transformer Regulation

The minimum size of a transformer is usually determined either by a temperature rise limit, or by allowable
voltage regulation, assuming that size and weight are to be minimized. Figure 21-F1 shows a circuit
diagram of a transformer with one secondary.
Note that a = regulation (%).

Primary Secondary
n = N./N
» pn = 1

Figure 21-F1. Transformer Circuit Diagram.

The assumption is that distributed capacitance in the secondary can be neglected because the frequency and
secondary voltage are not excessively high. Also, the winding geometry is designed to limit the leakage
inductance to a level low enough to be neglected under most operating conditions. The transformer
window allocation is shown in Figure 21-F2.

W
—- = Primary = Secondary [21-F1]

Primary

<y. Secondary

Figure 21-F2. Transformer Window Allocation.

Transformer voltage regulation can now be expressed as:

K,(N.L.)-K(F.L.) . .
= ^A-1—,\- -
(WO], [%]
L J
V / '

[21-F2]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.

 In which, VO(N.L.) is the no load voltage, and VO(F.L.) is the full load voltage. For the sake of simplicity,

assume the transformer in Figure 21-F1 is an isolation transformer, with a 1:1 turns ratio, and the core
impedance, R,,, is infinite.

If the transformer has a 1:1 turns ratio and the core impedance is infinite, then:

^ , = ^ > [°hms] [21-F3]

With equal window areas allocated for the primary and secondary windings, and using the same current
density, J:

A F, =/„*,= A r, =/,*,, [volts] [21_p4]

Regulation is then:

AV \V
a=—^(100) + —^(100), [%]
> ^ [21-F5]

Multiply the equation by currents, I:

a= VI
- ( 1 0 0 ) , [%]
/ ' L J
' ° [21-F6]

Primary copper loss is:

Pp=Wplm, [watts] [21_F?]

Secondary copper loss is:

P,=W,I,, [watt«] [21-F8]

Total copper loss is:

PCU=PP+PS, [watts] [21 _ p9]

Then, the regulation equation can be rewritten to:

a = 2 L ( l O O ) , [%] [21-F10]

Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved.