Powering Laser

Diode Systems

Gregoriy A. Trestman
SPIE Terms of Use: This SPIE eBook is DRM-free for your convenience. You may install this eBook on any
device you own, but not post it publicly or transmit it to others. SPIE eBooks are for personal use only.
For details, see the SPIE Terms of Use. To order a print version, visit SPIE.
Tutorial Texts Series
. Powering Laser Diode Systems, Grigoriy A. Trestman, Vol. TT112
. Optics Using MATLAB®, Scott W. Teare, Vol. TT111
. Plasmonic Optics: Theory and Applications, Yongqian Li, Vol. TT110
. Design and Fabrication of Diffractive Optical Elements with MATLAB®, A. Vijayakumar and Shanti
Bhattacharya, Vol. TT109
. Energy Harvesting for Low-Power Autonomous Devices and Systems, Jahangir Rastegar and Harbans S.
Dhadwal, Vol. TT108
. Practical Electronics for Optical Design and Engineering, Scott W. Teare, Vol. TT107
. Engineered Materials and Metamaterials: Design and Fabrication, Richard A. Dudley and Michael A.
Fiddy, Vol. TT106
. Design Technology Co-optimization in the Era of Sub-resolution IC Scaling, Lars W. Liebmann, Kaushik
Vaidyanathan, and Lawrence Pileggi, Vol. TT104
. Special Functions for Optical Science and Engineering, Vasudevan Lakshminarayanan and L. Srinivasa
Varadharajan, Vol. TT103
. Discrimination of Subsurface Unexploded Ordnance, Kevin A. O’Neill, Vol. TT102
. Introduction to Metrology Applications in IC Manufacturing, Bo Su, Eric Solecky, and Alok Vaid, Vol.
TT101
. Introduction to Liquid Crystals for Optical Design and Engineering, Sergio Restaino and Scott Teare, Vol.
TT100
. Design and Implementation of Autostereoscopic Displays, Byoungho Lee, Soon-gi Park, Keehoon Hong,
and Jisoo Hong, Vol. TT99
. Ocean Sensing and Monitoring: Optics and Other Methods, Weilin Hou, Vol. TT98
. Digital Converters for Image Sensors, Kenton T. Veeder, Vol. TT97
. Laser Beam Quality Metrics, T. Sean Ross, Vol. TT96
. Military Displays: Technology and Applications, Daniel D. Desjardins, Vol. TT95
. Interferometry for Precision Measurement, Peter Langenbeck, Vol. TT94
. Aberration Theory Made Simple, Second Edition, Virendra N. Mahajan, Vol. TT93
. Modeling the Imaging Chain of Digital Cameras, Robert D. Fiete, Vol. TT92
. Bioluminescence and Fluorescence for In Vivo Imaging, Lubov Brovko, Vol. TT91
. Polarization of Light with Applications in Optical Fibers, Arun Kumar and Ajoy Ghatak, Vol. TT90
. Digital Fourier Optics: A MATLAB Tutorial, David G. Voeltz, Vol. TT89
. Optical Design of Microscopes, George Seward, Vol. TT88
. Analysis and Evaluation of Sampled Imaging Systems, Richard H. Vollmerhausen, Donald A. Reago, and
Ronald Driggers, Vol. TT87
. Nanotechnology: A Crash Course, Raúl J. Martin-Palma and Akhlesh Lakhtakia, Vol. TT86
. Direct Detection LADAR Systems, Richard Richmond and Stephen Cain, Vol. TT85
. Optical Design: Applying the Fundamentals, Max J. Riedl, Vol. TT84
. Infrared Optics and Zoom Lenses, Second Edition, Allen Mann, Vol. TT83
. Optical Engineering Fundamentals, Second Edition, Bruce H. Walker, Vol. TT82
. Fundamentals of Polarimetric Remote Sensing, John Schott, Vol. TT81
. The Design of Plastic Optical Systems, Michael P. Schaub, Vol. TT80
. Fundamentals of Photonics, Chandra Roychoudhuri, Vol. TT79
. Radiation Thermometry: Fundamentals and Applications in the Petrochemical Industry, Peter Saunders,
Vol. TT78
. Matrix Methods for Optical Layout, Gerhard Kloos, Vol. TT77
. Fundamentals of Infrared Detector Materials, Michael A. Kinch, Vol. TT76
. Practical Applications of Infrared Thermal Sensing and Imaging Equipment, Third Edition, Herbert
Kaplan, Vol. TT75
. Bioluminescence for Food and Environmental Microbiological Safety, Lubov Brovko, Vol. TT74
. Introduction to Image Stabilization, Scott W. Teare and Sergio R. Restaino, Vol. TT73
. Logic-based Nonlinear Image Processing, Stephen Marshall, Vol. TT72
. The Physics and Engineering of Solid State Lasers, Yehoshua Kalisky, Vol. TT71

(For a complete list of Tutorial Texts, see http://spie.org/publications/books/tutorial-texts.)

Library of Congress Cataloging-in-Publication Data

Names: Trestman, Grigoriy A., author.

Title: Powering laser diode systems / Grigoriy A. Trestman.
Description: Bellingham, Washington : SPIE Press, [2017] | Series: Tutorial texts in
optical engineering ; volume TT 112 | Includes bibliographical references and index.
Identifiers: LCCN 2017015231 (print) | LCCN 2017017794 (ebook) | ISBN
9781510608467 (pdf) | ISBN 9781510608474 (epub) | ISBN 9781510608481 (mobi) |
ISBN 9781510608450 | ISBN 9781510608450q (softcover) | ISBN 1510608451q
(softcover)
Subjects: LCSH: Semiconductor lasers. | Diodes, Semiconductor.
Classification: LCC TA1700 (ebook) | LCC TA1700.T73 2017 (print) |
DDC 621.36/61–dc23
LC record available at https://lccn.loc.gov/2017015231

Published by
SPIE
P.O. Box 10
Bellingham, Washington 98227-0010 USA
Phone: +1 360.676.3290
Fax: +1 360.647.1445
Email: books@spie.org
Web: http://spie.org

Copyright © 2017 Society of Photo-Optical Instrumentation Engineers (SPIE)

All rights reserved. No part of this publication may be reproduced or distributed in

any form or by any means without written permission of the publisher.

The content of this book reflects the work and thought of the author. Every effort has
been made to publish reliable and accurate information herein, but the publisher is not
responsible for the validity of the information or for any outcomes resulting from
reliance thereon.

Printed in the United States of America.

First Printing.
For updates to this book, visit http://spie.org and type “TT112” in the search field.
 To the memory of my father:
engineer and scientist Abram Trestman
 Introduction to the Series
Since its inception in 1989, the Tutorial Texts (TT) series has grown to cover
many diverse fields of science and engineering. The initial idea for the series
was to make material presented in SPIE short courses available to those who
could not attend and to provide a reference text for those who could. Thus,
many of the texts in this series are generated by augmenting course notes with
descriptive text that further illuminates the subject. In this way, the TT
becomes an excellent stand-alone reference that finds a much wider audience
than only short course attendees.
Tutorial Texts have grown in popularity and in the scope of material
covered since 1989. They no longer necessarily stem from short courses;
rather, they are often generated independently by experts in the field. They are
popular because they provide a ready reference to those wishing to learn
about emerging technologies or the latest information within their field. The
topics within the series have grown from the initial areas of geometrical optics,
optical detectors, and image processing to include the emerging fields of
nanotechnology, biomedical optics, fiber optics, and laser technologies.
Authors contributing to the TT series are instructed to provide introductory
material so that those new to the field may use the book as a starting point to
get a basic grasp of the material. It is hoped that some readers may develop
sufficient interest to take a short course by the author or pursue further
research in more advanced books to delve deeper into the subject.
The books in this series are distinguished from other technical
monographs and textbooks in the way in which the material is presented.
In keeping with the tutorial nature of the series, there is an emphasis on the
use of graphical and illustrative material to better elucidate basic and
advanced concepts. There is also heavy use of tabular reference data and
numerous examples to further explain the concepts presented. The publishing
time for the books is kept to a minimum so that the books will be as timely
and up-to-date as possible. Furthermore, these introductory books are
competitively priced compared to more traditional books on the same subject.
When a proposal for a text is received, each proposal is evaluated to
determine the relevance of the proposed topic. This initial reviewing process
has been very helpful to authors in identifying, early in the writing process, the
need for additional material or other changes in approach that would serve to
strengthen the text. Once a manuscript is completed, it is peer reviewed to
ensure that chapters communicate accurately the essential ingredients of the
science and technologies under discussion.
It is my goal to maintain the style and quality of books in the series and to
further expand the topic areas to include new emerging fields as they become
of interest to our reading audience.
James A. Harrington
Rutgers University
vii
Contents
Preface xiii
Acknowledgments xvii

1 Introduction 1
1.1 Photonics Revolution and Power Electronics 1
1.2 Audience 2
1.3 Equation Derivation, Numerical Calculations, and Units 2
1.4 Schematic Plotting and Circuit Computer Simulation 4
1.5 Power-Supply Development 4
2 Power-Electronics Design Rules to Save Time and Avoid Frustration 7
2.1 Start with a Clearly Written Technical Specification 7
2.2 Attentively Study the Characteristics of the Load 8
2.3 Choose the PS Topology Carefully 9
2.4 Choose a Topology that Efficiently Uses the
Parasitic Elements of the Circuit Components 9
2.5 Use MathCad for Circuit Calculation and Mathematical Modeling 10
2.6 Perform a Computer Simulation of the Power-Electronics Circuits 10
2.7 Verify the Correctness of the Component Kit
and the Soldering Quality 11
2.8 Test Offline Power Sources Carefully 11
2.9 Offline Devices Not Connected to the AC Mains Do Not Work 12
2.10 Skillful Measurements Obtain the Required Results 12
2.11 LD-Driver Design Priorities 12
3 Similarities and Differences between LDs and LEDs 15

4 Laser Diodes: Electrical Loads and Driving Requirements 17

4.1 Electrical Load Types 17
4.2 Characteristics of Laser Diodes 18
4.3 Aging and Temperature Effects on LDs 19
4.4 Ideal and Real Sources of Electrical Energy 20
4.5 Requirements for LD Powering 21

ix
x Contents

5 Primary and Secondary Sources of Electrical Energy 23

5.1 Primary Sources of Electrical Energy 24
5.1.1 Electrical batteries 24
5.1.2 Electrical-energy generation and distribution 24
5.2 Secondary Sources of Electrical Energy 25
5.2.1 Active secondary converters 25
5.3 Passive Voltage-to-Current Converters 29
5.4 Passive Linear DC VtoI Converter with a Ballasting Resistor 30
5.5 Passive Sinusoidal AC VtoI Converter with a Ballasting Inductor 31
5.6 Passive Sinusoidal AC VtoI Converter with a Ballasting Capacitor 34
5.7 Capacitors and Inductors in Laser Systems 35
6 High-Frequency Switch-Mode Passive DC-to-DC Converters
in LD Systems 37
6.1 Passive Converter with a Ballasting Inductor and Capacitive Output 37
6.2 Converter Circuit Description and Calculation 38
6.3 Conduction Losses and Optimization in VtoV Converters 44
6.4 Switching Losses in Converters with Pulse-Width Modulation 45
6.5 Resonant and Quasi-resonant Converters 48
6.6 Reducing Switching Losses in HF VtoV Converters 48
6.7 Utilizing Converter Parasitic Elements 51
6.8 Transformer in a HF Converter 51
7 High-Frequency Switch-Mode LD Driver Topology 57
7.1 Calculating the Filter Inductor 59
8 Powering Medium-Power LD Systems 61
8.1 Powering a LDD with a Battery 61
8.2 Powering a LDD with a One-Phase AC Line 61
8.3 One-Phase Compound Converter as a LDD 70
9 Powering High-Power LD Systems 71
9.1 Three-Phase AC Line 71
9.2 LD Bars and Stacks 72
9.3 Powering Multiple LD Strings from a Single Inverter 76
9.4 Powering RGB Lasers in Movie Projection Systems 76
9.5 Programmable LED Color-Lighting Systems 76
9.6 Powerful Modular DC-Voltage Source 79
10 LD Drivers Based on Other Topologies 81
10.1 Buck Converter as a LDD 81
10.2 Active Linear DC VtoI Converter 84
11 Passive Switch-Mode Voltage-to-Power Converters in Laser Systems 87
11.1 Passive VtoP Converter as a LDD 87
11.2 Passive VtoP Converter as a Capacitor Charger 91
Contents xi

12 Powering Other Components of LD Systems 95

12.1 Thermoelectric Cooler 95
12.2 High-Voltage Power Supply with a Pockels Cell 96
Appendix: LTspice Circuit-Simulation Examples 101
A.1 VtoV Converter Circuit 101
A.2 VtoI Converter Circuit and Load-Curve Plotting 101
A.3 VtoP Converter Circuit 105
A.4 VtoP Converter as a Capacitor Charger Circuit 105
References 109

Index 111
Preface
In February 2015, SPIE Press offered Dr. Ilya Bystryak and me the
opportunity to write a book based on our professional development course
“Powering and Integration of Laser Diode Systems,” presented by both of
us at the Photonics West conference.1 The course material was the result
of our lengthy collaboration, initially in the Soviet Union, where we worked
for the same research laboratory for an extended period of time, and later in
the U.S.
Professional collaboration and the sharing of ideas are important factors
of our successful careers in the field of power electronics. Such communication
multiplies the engineering effectiveness of each other. The best-known
examples of similar fruitful cooperation are the joint work of Steve Jobs
and Steve Wozniak, Larry Page and Sergey Brin, and Bill Gates and Paul
Allen. Constant, critical face-to-face or telephonic discussions around ongoing
projects and emerging problems enables the fast and efficient resolution of
these challenges. This form of cooperation is rarely achieved by project formal
discussions at scheduled company meetings, design reviews, etc. Therefore,
I recommend to my young colleagues to find a partner and a formidable
opponent to cooperate with and to maintain this kind of professional
relationship throughout their career.
Because the book is based on an instructional course, I decided to present
it to a certain extent as a textbook; however, because strict technicalities are
often not very useful and boring to read, the material is not written in the
traditional textbook style. This Tutorial Text discusses the competent design
and skilled use of laser diode drivers (LDDs) and power supplies (PSs) for the
electrical components of laser diode systems. It is intended to help power-
electronic design engineers during the initial design stages: the choice of the
best PS topology, the calculation of parameters and components of the PS
circuit, and the computer simulation of the circuit. Readers who use laser
diode systems for research, production, and other purposes will also benefit.
The book will help readers avoid errors when creating laser systems from
ready-made blocks, as well as understand the nature of the “mystical failures”
of laser diodes (and possibly prevent them).

xiii
xiv Preface

The following questions guided the compilation of this book:

• What is the basis of choosing one or another technical solution?
• Which converter topologies are better suited than others for powering
certain LD system parts?
• Why should soft-switching topologies be used when developing switch-
mode PSs?
• Why must one contend with the parasitic elements of the circuit com-
ponents in some topologies, whereas other topologies can use parasitic
elements beneficially?
• How does the choice of topology influence LD and LDD reliability?
• How can the right topology protect against overload and avoid the
premature death of the LD (which usually costs more than the LDD)?
• Which topologies generate less electromagnetic interference (EMI) noise
and therefore require fewer measures of protection against penetration
EMI in the AC mains and the environment?
I tried not to overload the book with complex mathematical formulas.
In my experience, most converter calculations are straightforward, and
well-known mathematical formulations of the electrical circuit laws are
sufficient. Excessive math does not promote but rather hinders understanding
the essence of things.
The electrical characteristics of laser diodes and the specifics of the sources
for their powering are discussed herein. It has been shown that laser diodes
require electrical sources with the characteristic of either a constant current or
a constant power source. Because the primary sources of electrical energy are
voltage sources, the methods of converting voltage sources into a constant
current or a constant power source are the main subject of the book.
Detailed attention is paid to the passive methods of conversion (Chapter 5).
Passive conversion of the voltage to the current has several advantages when
designing LDDs:
• Because there is no need for a feedback loop in passive converters, their
circuits are less complicated than those of active converters.
• An inherent current limit prevents an overcurrent that could destroy the
laser diode.
• It is simple to create powerful sources with the direct parallel connection
of conversion modules.
• Multiple laser diode strings can be driven by a single converter.
• Pulse modulation of the laser diode current in the pulsed laser system is
simple.
Modern power supplies are based on switching converters. The main
problems that afflict these converters are switching losses and EMI noise
generation, which can be eliminated or reduced by soft-switching. Soft-switching
Preface xv

is the hallmark of professional PS design, so Chapter 6 addresses the design of

quasi-resonant circuits and the calculation of lossless snubbers, which allows one
to organize soft-switching.
Another problem facing the power-electronics designer involves the
parasitic elements of the converter circuit. Section 6.7 demonstrates the proper
topology that derives a benefit from parasitic elements instead of fighting
them.
The initial stages of PS design are reflected in the book via examples.
MathCad is used for calculations, and LTspice is used for electrical schematic
plots and the computer simulations. In some complex cases, the purely
mathematical calculation of circuits is impossible, so a combination of
calculation and simulation is necessary.
Readers should understand the basic concepts and definitions used in
power electronics. Sufficient context is provided to explain the topics at hand,
but other publications should be consulted for an overview of laser-diode
power design. Furthermore, some common design issues are not covered here,
e.g., the design of synchronous rectifiers in PSs with a low voltage output,
EMI filters, gate drivers for MOSFETs, and how to protect PSs from
instabilities in the AC line. A list of references is provided for further reading
regarding these matters.
Grigoriy A. Trestman
Reseda, California
April 2017
Acknowledgments
Two very important people helped me write this book. One of them is
my engineering colleague and long-time friend, Dr. Ilya Bystryak. Our
collaborative work in the past and our endless discussions and knowledge
exchanges provided the necessary climate for solidifying the ideas described in
this book. Another is my son, Arkady Trestman, who helped immensely with
editing the content of the book, reviewing the style, and making sure the book
is applicable to a broad audience of power-electronics and laser engineers.
I also want to thank my wife, Lidiya, for supporting and encouraging me
throughout the writing process.

xvii
Chapter 1
Introduction

1.1 Photonics Revolution and Power Electronics

At the beginning of the new millennium, the invention of high-intensity and
blue light-emitting diodes (LEDs) started a revolution in lighting technology.
Their unique properties are leading to the replacement of all other light
sources in general illumination devices by LEDs. The characteristics of the
new LEDs are a high efficiency of electrical energy conversion into light, high
reliability, long life, compact size, low cost, color variety, environmental
friendliness, ability to be manufactured by very-well-developed semiconductor
technologies, etc.
A similar revolution is occurring for laser engineering and technology.
Diode lasers are not only replacing other laser types in well-known
technologies but also creating new applications. The rapid progress in fiber
optics further accelerated the advancement of industrial laser technologies.
Optical fibers helped to increase the power delivered by a laser beam for up to
several tens of kilowatts, making lighting technologies preferable for industrial
applications.
Laser diode (LD) radiation characteristics depend primarily on the electric
power provided to the LD by an electrical power source, i.e., a laser diode
driver (LDD). As mentioned in the Preface, LEDs and LDs require power
sources with the characteristic of a constant current source and, in some cases,
constant power source. There has not been enough attention and coverage
given to current and power sources in the contemporary technical literature.
One of the goals of this book is to address that deficiency.
Whereas the maximum power of LED fixtures is limited to several
hundred watts, laser-system power had reached several tens of kilowatts. Due
to this difference, the conversion topologies and constructions of LDDs
significantly differ from LED drivers. LDDs can be created using a variety of
electrical conversion topologies. However, the reliability, efficiency, size,
LDD cost, and ultimately the success or failure of the LDD project all depend
primarily on the correct choice of the conversion topology. This book presents
the effective and reliable conversion topologies for LDDs.

1
2 Chapter 1

Mid- and high-power LDs are very expensive devices. The conversion
topologies offered in the book have an inherent current limit. This feature
provides the LD protection from any AC line instabilities or LDD
malfunctioning. Unlike LEDs, LDs are widely used in the pulsed mode and
thus require pulsed LDDs. Another goal of this book is to familiarize readers
with the effective pulsed LDDs.
The book also covers some other important aspects of power-electronics
design, such as soft switching in high-frequency (HF) switched-mode
converters, advantageous usage of parasitic elements of electrical circuit
components, paralleling conversion modules to increase power in industrial
LDDs, etc.

1.2 Audience
This book focuses on the powering of LD systems. It was written for two
categories of professionals: (a) power-electronics design engineers involved in
the development of LDDs and power supplies (PSs) for electrical components
of laser systems, and (b) professionals who build and/or operate laser systems.
Companies or organizations that build or use LD systems have two
alternatives when it comes to system creation. The first is to develop specific
LDDs and PSs. This case requires a significant investment of finances and
time, and it requires power-electronics-engineering resources experienced in
the development of current and power sources (which is not always possible).
The second alternative is to assemble the system from commercially available
components. In many cases, companies prefer the latter.
LDDs and PSs for LD subsystems are very important parts of LD systems
that directly affect the system’s general reliability and the expected lifetime of
the most expensive part of the system, i.e., the LD assembly. System
developers and users should understand how to choose the appropriate power
sources, which often involves compiling a list of questions for vendors and
knowing the limitations of the standard PSs on the market. These topics are
also discussed in the book.
The issue specific to LDs is their dangerous sensitivity to the slightest
current overloads in comparison to the other laser types. There is a joke
among laser specialists that “all LD system users can be divided into two
groups: those who already burned down an expensive LD and those who will
eventually.” I hope that this book will help readers avoid belonging to either
of these groups.

1.3 Equation Derivation, Numerical Calculations, and Units

The engineering math software MathCad 152 was used to derive the equations
that describe electrical circuit operation and to numerically calculate these
Introduction 3

circuit examples. These formulas and examples can be copied from the book
and pasted into MathCad software for use with variations for calculations of
your circuits. Most graphics were also plotted using MathCad.
To interpret the formulas in the book correctly, it is necessary to mention
several MathCad specifics:
• To define a variable definition, use the symbol :¼. A variable can be set
either numerically or symbolically. For example,

x :¼ 2; y :¼ x2  1:

• To evaluate an expression with previously defined variables, the standard

evaluation symbol ¼ is used. Following the previous expressions,

y ¼ 3:

• To block the evaluation of an expression that has some undefined

variables but is required to derive the following equations, the “Disable
Equation” operator is used. The small black box next to the equation
indicates that evaluation is turned off:

dI load V
:¼ dc ▪.
T3  T2 Lf

Numerical calculation of circuit parameters helps clarify circuit

behavior and limitations. Variables are defined in the Initial Data section
of the MathCad program, and the evaluation results are in the Calculation
section. There are numerical examples for most of the circuits discussed in
this book.
All of the examples of numerical calculation of circuits use units defined
by the International Unit System (SI):
• Current in amperes (A),
• Voltage in volts (V),
• Power in watts (W),
• Energy in joules (J),
• Charge in coulombs (C),
• Resistance in ohms (ohm),
• Capacitance in farads (F),
• Inductance in henrys (H),
• Frequency in hertz (Hz),
• Time in seconds (s), and
• Temperature in degrees Celsius (°C).
4 Chapter 1

1.4 Schematic Plotting and Circuit Computer Simulation

Schematic plotting and circuit computer simulation were performed with the
simulation software LTspice IV by Linear Technology Inc.3 For clarity,
excessive details in schematics, such as the parameters of the component and
those of the voltage and current sources, as well as the simulation directives,
are omitted. These parameters should be added to make circuits workable in
LTspice IV (see the Appendix). If you are interested in obtaining the files with
calculations and simulations of circuits highlighted in this book, you can order
them directly from the author.

1.5 Power-Supply Development

There is an absurd saying, “The easiest things are curing people and governing
the state—everyone knows how to do it.” Similarly, some electronics
engineers and managers think that PS design is a trivial task. In the author’s
professional experience, people who are not specialists in power electronics,
such as digital-electronic engineers, have sometimes been assigned to develop
PSs, even for mass production, and the results have always been negative.
To a non-power-electronics engineer, it may seem that PS design and
prototyping may commence after referring to books and online sources about
power-conversion technology and finding some seemingly suitable schematics.
In addition to the aid of a “developer,” there is also modern computer
simulation software. A working schema can be obtained in a short amount of
time without a detailed understanding of circuit operation and mathematical
calculations of circuit components. After assembling the prototype and
varying the circuit component values, it may be possible to produce a device
that fulfills a specified task.
This device might be suited for laboratory experimentation, but it would be
hardly possible to manufacture it on the industrial scale due to the
inconsistencies of this “development” for the requirements of the PS
specification and certification compliance by such agencies as the FCC, UL,
CE, VDE, etc. It will also be difficult to achieve a highly reliable performance.
A developed PS often cannot be placed in a given enclosure because its
efficiency is too low and the temperature of the components is too high. Large
heatsinks are needed to reduce the temperature to an acceptable level. The
selected topology may be very “noisy,” and the “designer” cannot identify the
sources of noise and how to reduce them.
It should be clear that without preliminary consideration of the thermal
and EMI problems, the design should not be attempted. A designer should
consider a different topology to combat these problems. What if the PS spec
requires a manufacturing warranty for two or even five years? The designer
must now factor in PS long-term reliability. Which components could fulfill
the warranty period? Again, it is necessary to dig into the literature and look
Introduction 5

for answers. The list of emerging issues and challenges is very long, including
ensuring PS safety, protecting against power surges in the AC mains,
preventing load failure when the PS malfunctions, etc.
It is possible to consider all of these issues, to find ways of solving them,
and learn how to design reliable PSs. However, digital-design engineers would
need to abandon their specialty for several years and completely immerse
themselves in power-conversion technology, namely the design of PSs.
Qualified experts in digital design who are not passionate about power
electronics would do better to focus on their field. (The exception is a PS with
microprocessor control. All a digital designer must do is develop the
microprocessor control and nothing more.)
Readers who are adamant about becoming a power-electronics engineer
will benefit from the useful rules and tips contained in this book. However,
this one text is not enough to become a professional power-electronics
designer. There are many other good books devoted to the subject matter,
which can be found online.4 Beginners should start with those light on
mathematics; the calculation of converters is rather simple and is based on
well-known mathematical formulations of electric-circuits laws.
Chapter 2
Power-Electronics Design
Rules to Save Time and Avoid
Frustration
So you have decided to specialize in power electronics (or you are working in
the field already), and you are tasked with developing a power source. The
following rules and recommendations will help you avoid unnecessary stress,
frustration, and undesirable consequences.

2.1 Start with a Clearly Written Technical Specification

It is a good practice to have clear specifications before starting a design.
Carefully study the specification of the PS and understand the customer’s or
management’s requirements. For example, a manager may casually mention
that “something” needs to be developed. In most cases, when this “something”
is created according to the manager’s verbal input, either the result does not
match the manager’s original vision, the instructions were not complete, or the
critical parameters of the intended device were omitted. The design must be
restarted, but the culprit is usually not the manager. Written specifications
prevent this scenario from happening. If something is not articulated or
understood, try to find clarification and correct the spec before starting
development.
A designer joke states, “If, in the process of developing a device, it is
difficult to implement some of the requirements of the spec, try to change the
spec.” It would seem like an absurd recommendation, but every joke has some
truth. Sometimes a customer makes excessive demands that are unnecessary
for a particular application and does not understand how they complicate a
particular PS design or increase the dimensions and raise the cost. For
example, the requirement of an unnecessary, very small output-voltage ripple
results in a large and expensive output filter, which precludes the PS from
fitting inside the predetermined housing or leads to cost overrun.

7
8 Chapter 2

This recommendation is specific for custom designs and not very

applicable to PSs produced by various manufacturers and intended to be
installed as a subcomponent in OEM equipment, where the PS electrical
characteristics and dimensions are strictly limited.

2.2 Attentively Study the Characteristics of the Load

Once you have a clear and concise specification that covers all points, it is
time to start developing the PS, which converts the electric energy of the
battery or AC mains into energy with the characteristics required by the load.
Therefore, a clear understanding of the electrical characteristics of the load is
important, especially the current–voltage characteristic (IVC), representing
the load voltage dependence on the load current. It is also important to know
the following load characteristics:
• maximum allowable load voltage and current,
• load dynamic requirements (the speed of the PS reaction on the load or
AC-line parameter changes),
• the effect of changes in temperature, vibration, pressure, humidity, and
other external factors on the load characteristics, and
• the load characteristic changes due to load aging, thermal cycling, etc.
The following is an example of how load aging can affect PS reliability.
I was designing a linear fluorescent-lamp, instant-start-type electronic ballast
with a five-year warranty period. The electronic ballast is a current source that
keeps constant the alternating current in the lamp despite the voltage drop on
the lamp electrodes. In a new 4-ft lamp at a 180-mA current, a nominal
voltage drop on the lamp’s electrodes is 140 V. In this case, the lamp power is
25.2 W. With a 90% offline efficiency, the AC-line power consumption is
28 W. At this point, it seems like the development of a 30-W ballast would be
sufficient.
The issue is that the electron emission capability of fluorescent lamp
electrodes gradually decreases during the lamp’s operating period, which
increases the lamp voltage drop. By the time the lamp reaches its end of life,
the voltage drop is two times the nominal voltage, whereas the ballast keeps
the same lamp current. The ballast power consumption increases two times,
and it will fail if it was developed to survive 30 W. The problem is made more
difficult because an instant-start ballast ages the lamp, mostly due to the
number of starts and not the actual operating hours. No one knows when
aging starts to affect the lamp voltage. To overcome this issue, a 60-W ballast
was developed.
A laser diode (LD) is not a simple load, either. LDs are complex devices
that are very sensitive to electric power provided by LDDs and can be easily
damaged by the current surges, electrostatic discharges, etc. Some LD
Power-Electronics Design Rules to Save Time and Avoid Frustration 9

characteristics are discussed later in this book. Due to the large variety of LD
assemblies on the market, power-electronics designers should carefully study the
specifics of each LD assembly and develop a corresponding LD driver; this
method ensures that a LDD would provide not only reliable LD operation but
also prevent failure of the driver itself during the warranty period.

2.3 Choose the PS Topology Carefully

The next step involves choosing the optimal electrical topology of the
converter for a particular application. This is a critical step because the
success or failure of the project ultimately depends on it. The same conversion
task could be solved by using several different topologies; the differences
between them lie in the converter performance characteristics.
Most of the modern conversion topologies are high-frequency (HF)
switch-mode convention topologies. Compared to linear-mode conversion
topologies (used during the early power-electronics years), they have much
higher efficiency, and PSs based on them are much smaller and lighter. The
fundamental issues of switch-mode topologies are losses in power-switching
devices (transistors and diodes) and electromagnetic noise generated by fast
switching actions. To significantly reduce these harmful effects, use topologies
with natural soft-switching or generate soft-switching with lossless snubbers.
Lossless snubber schemes can increase the overall converter efficiency,
i.e., energy accumulated in the snubber is transmitted to the load. This book
examines various methods of obtaining soft-switching.

2.4 Choose a Topology that Efficiently Uses the Parasitic

Elements of the Circuit Components
Electrical components used in converter circuits are not ideal. For example,
• The capacitor has a parasitic inductance and resistance of electrodes.
• The inductor has parasitic winding capacitance and winding resistance.
• The transformer also has parasitic primary and secondary winding
capacitances and resistances. There are parasitic capacitances between
windings. Another parasitic parameter of a transformer is a leakage
inductance, which is caused by a non-ideally coupled magnetic field
between transformer windings.
• The power-switching device, a metal–oxide field effect transistor (MOS-
FET), has junction parasitic capacitances and a parasitic body diode.
Some topologies make you contend with these parasitic elements, whereas
others allow you to utilize them beneficially. In converters, parasitic elements
function as real components. In some cases, the use of parasitic elements
allows you to reduce the number of installed parts (discussed in Chapter 10).
10 Chapter 2

2.5 Use MathCad for Circuit Calculation and Mathematical

Modeling
After the converter topology is selected, the scheme of a power converter is
defined. There are a number of ways that engineers approach the next step.
Some start generating computer simulations or even PS prototypes right
away. They try to obtain the output characteristics of the simulated circuit
close to the desired result by varying the parameters of the circuit components.
Others start by calculating the schematic parameters and the components,
followed by building a computer simulation of the circuit and then
prototyping. This book recommends the latter approach.
To calculate the circuit, the extremely convenient and easy-to-learn
MathCad software2 is recommended for several reasons. First, circuit
calculations in MathCad can be regarded as a mathematical model of the
PS. By varying the input parameters of the PS within a given spec in the
software (minimum and maximum voltage, current, conversion frequency,
etc.), you can immediately see how the changes affect PS characteristics. You
can also vary the values of the critical circuit components within the selected
tolerances and see the effect. Based on the calculated data, you can make the
necessary adjustments. Second, after one project is proved by the simulation
and testing of the experimental device, the MathCad calculations can be
reused in similar projects by tweaking the input parameters. In this case, the
entire MathCad program development could be avoided.
The advantage that MathCad has over other methods of power-
electronic-scheme calculations, such as Microsoft Excel spreadsheets, is its
visibility. The calculation is as familiar as a regular form of the interlaced
equations of mathematically described laws of the electrical circuit. For
greater clarity, a comment can be placed on each line of calculations to serve
as a reminder of a particular formula’s purpose. It is also useful to include in
the MathCad program a general description of the scheme, comments about
the original data, and the comments about the obtained results.
Because the goal of circuitry calculations is the computation of component
values at the beginning of the process, you should create a schematic of the
circuit that names the input and output parameters and the reference
designators of the components. MathCad is convenient for plotting graphs.
One graph can display curves calculated by the formulas, as well as the
experimental data, to help you analyze the reasons for deviations. This book
presents examples of MathCad calculations based on these recommendations.

2.6 Perform a Computer Simulation of the Power-Electronics

Circuits
The next design stage creates a computer simulation of the circuit. Design
engineers have different attitudes toward computer simulation—some reject
simulation, and some overuse it. More than 15 years ago, simulation software
Power-Electronics Design Rules to Save Time and Avoid Frustration 11

lacked accuracy, and the experimental results were at odds with the simulated
results. A simulation run required additional circuit elements that were not
required in the real circuit, e.g., small resistors. Personal computers were slow,
and simulations required significant amounts of time to run. These issues were
irritating enough that many engineers refused to use simulation because it was
much faster to assemble and test the prototype circuits.
Most of these drawbacks no longer apply, and simulation has become a
powerful tool for the preliminary study of circuits. However, other hazards
remain. Some engineers omit calculation steps and attempt to solve design
problems by varying the component parameters in the simulation program.
This approach is rarely successful.
The next question asks to what extent a PS must be simulated. PSs are
devices of varying complexity. Some are simple devices powered by batteries
and have one converter with (or without) a simple feedback circuit. Others are
complex devices powered by AC mains and have several converters and
complex analog or digital feedback circuits. It may be unrealistic to simulate
the complete scheme of a complex PS.
A positive feature of simulation is its ability to quickly verify the quality of
the scheme calculation. If there was a significant difference between the
calculation and the results of the simulation, it is necessary to stop and try to
find the sources of these discrepancies.
LTspiceIV software by Linear Technology Inc. is one of the most convenient,
fast working, and accurate tools to simulate power converter circuits. This
software is free to download on the company’s website.3 It is very easy to master
and only requires a few hours to get comfortable with simulating circuits in it.
Also, Linear provides examples of circuit simulations for many of the chips they
manufacture. The case studies of Linear chips simulations are also beneficial.
(The author has no personal or financial ties to Linear Technology Inc.)

2.7 Verify the Correctness of the Component Kit and the

Soldering Quality
When starting work on the first PS prototype, there is no confidence in the
correctness of the developed scheme. Negligence during the preparation of
the component’s kit and soldering connections can make the launch of the
prototype time-consuming. If the prototype burns out, you would have to
start over. Furthermore, it is difficult to identify “cold” soldering because it
looks normal. It is thus essential to eliminate this negligence during the
prototyping process.

2.8 Test Offline Power Sources Carefully

A high voltage in an offline PS (up to several hundred volts) can deliver a very
painful electric shock or even lead to death. Turning off the PS does not
guarantee the absence of high voltage in some of its components even though
12 Chapter 2

the AC line is cut off. High-quality capacitors can hold a charge for a long
time. An engineer might forget to disconnect the device from the AC line
before making changes to the scheme, which might be under the main voltage
(the presence of which is not evident).
It is advised to keep a voltmeter constantly connected to the rectified
mains voltage (bus voltage) to protect the laboratory staff and yourself. Check
the voltmeter to ensure that there is no voltage on the bus before attempting
anything with the circuit. A better option would shunt the input of the
voltmeter with a push-button contact connected in series with a several-
thousand-ohms two-watt resistor. Press the contact to discharge the
electrolytic capacitors before altering the circuit.

2.9 Offline Devices Not Connected to the AC Mains Do Not Work

This rule might sound humorous, but it has a certain meaning. Remember this
rule when testing, especially during troubleshooting. Someone may start
looking for a fault in a circuit, forgetting to check whether the PS is plugged
into the AC line, the fuse is blown, or the circuit breaker is tripped.

2.10 Skillful Measurements Obtain the Required Results

Another rule that might sound like a joke, this one has a twofold meaning.
Unethical scientists claim this rule when it is necessary to adjust the
experimental results to fit the theory. However, setting aside the irony, there
are times when it is not easy to use measurement equipment and interpret the
data. Extra efforts should be made to master the measuring device.
As an example, the correct usage of a frequency response analyzer, e.g., an
AP300 manufactured by Ridley Engineering Inc., becomes critical when the
PS feedback circuit is difficult to stabilize. The analyzer allows one to measure
the gain and the phase shift versus the frequency in the feedback loop.
However, sometimes the initial attempt to obtain the obvious results does not
succeed; it takes time to study the analyzer manual and talk to the vendor’s
application engineer to determine the correct method of measuring and
interpreting the results.

2.11 LD-Driver Design Priorities

The final general topic—design priorities of LDDs—can be arranged in the
following order:
• Compliance
The LDD should comply with the client’s spec because otherwise the
customer would not be able to use it. Along with the electrical-power
requirements, the spec should include the requirements for the EMI level
Power-Electronics Design Rules to Save Time and Avoid Frustration 13

(varies by country) and the safety features so that the appropriate

compliance agencies will allow the LDD to be sold and used.
• Reliability
Reliability has two aspects. First, it is the absence of LDD failures
during the warranty period. A failure can have costly consequences: in
industrial equipment, it can halt the production cycle or damage the
product. The second aspect involves the higher cost of mid- and high-
power LD assemblies compared to the LDDs that power them. As
mentioned earlier, LDs are very sensitive to the slightest overloads,
which differentiates the design of LDDs from power sources designed to
supply other types of electrical devices. LDDs must prevent LD damage
due to any instabilities in the AC mains (surges, bursts, and sags) or due
to LDD malfunctions. One of the main goals of this book is to highlight
the design patterns for such LDDs.
• Convenience for production
Manufacturability is one of the most essential conditions for mass-
produced LDDs. If the underlying technology used to manufacture the
LDD is complex, a decline in quality and subsequent LDD failure could
result, i.e., it affects the reliability.
• Cost
Cost appears at the end of the list, but for devices produced by
competing companies, it remains an important factor. A consumer is
usually motivated to acquire a similar product at a lower price.
Therefore, mass-production companies perform several cycles of cost
reduction after completing the initial design phase. It is a normal
practice, but the product reliability should be verified during each
iteration of cost reduction.
Chapter 3
Similarities and Differences
between LDs and LEDs
Although they have similarities, light-emitting diodes (LEDs) and LDs have
significant differences in their construction and application. LED systems are
used for surface illumination, as a backlight in LCD TVs, and as indicators.
The main purpose of LEDs in general lighting is the creation of a certain level
of light and color on an illuminated surface. The power of conventional LED
fixtures ranges from a few watts to several hundred watts. A LED system is
usually simple and comprises only a single LED (or one or more LED strings)
and a LED driver enclosed in the lighting fixture. Due to their relatively small
power, LED systems usually use heatsinks to maintain an acceptable
operating temperature, and they do not require forced cooling. Most LEDs
have a single p-n junction. A white LED consists of a semiconductor structure
that generates blue light and a phosphor coating that converts the blue light
into white light. Practical Lighting Design with LEDs, by R. Lenk and
C. Lenk,5 is an excellent source of information about modern LED
characteristics and applications, as well as LED driver-circuit design.
The semiconductor structure and the mechanical construction of LDs are
much more complex than LEDs. LDs are usually built on semiconductor
heterostructures. They have mirrors on semiconductor surfaces (facets) to
create the laser optical resonator.
The application area and power range of LD systems are much broader
than LED systems. They are used for material processing, communication,
information technologies, transportation, defense, medicine, and biotech. LDs
are intended to provide the required power and stability of a laser beam. The
power of LD systems ranges from a few watts to several tens of kilowatts.
Mid- and high-power LDs require forced air or water for cooling, as well as
other accessories, to form a laser system.
The laser beam characteristics and LD system reliability largely depend
on the electric power provided to the LD by an electric power source, i.e., a
LDD. Due to some similarities between LDs and LEDs, their powering

15
16 Chapter 3

methods have some resemblance; however, in general, they are significantly

different.
LDs and LEDs are semiconductor diodes and thus have the electrical
characteristics of the diodes. They cannot be powered from a constant voltage
source and require power supplies with the characteristics of a constant
current source. In some cases, the preferred electrical supply type is a constant
power source. However, on this point, the similarity of the drivers ends.
One of the main factors that determines the variance in the design
approach of LD and LED drivers is the difference in their failure mechanism.
In general, LED degradation and failure happen due to overheating of
the semiconductor structure or, in the case of white LEDs, overheating of
the phosphor coating. The LD semiconductor structure also suffers from
overheating. However, LD degradation usually starts with damage to the
facet mirror when the beam intensity, proportional to the LD current, exceeds
the facet destruction level. Exceeding the LD destructive threshold leads to
instant and terminal LD failure. A minor overcurrent produces local facet
defects that affect the LD characteristics, e.g., beam color shift, lasing
threshold increase, reduction of efficacy, etc. Another factor that determines
the alternative approach to LDD design is the higher power level of LD
systems in comparison to LED systems. Electrical conversion topologies that
are acceptable for low-power LED drivers are not suitable for LDDs.
When considering the LDD design, it is necessary to emphasize the
following circumstances. LDDs can be created using a variety of conversion
topologies. However, their reliability, efficiency, size, and cost depend
primarily on the choice of the conversion topology. Chapters 7–9 offer the
most efficient and reliable conversion topologies for constant-current,
constant-power, and pulsed-type LDDs.
Medium- and high-power LD assembly is a part of the laser system, and
the system components require electrical power to operate. Power supplies for
LD system components are described in Chapter 12. The variety of LD
systems is quite large, and it is impossible to give individual consideration for
each case in one book; thus, the focus here is on the most-effective solutions
for powering LDs and LD system components. With some variations, they
could apply to most cases.
Chapter 4
Laser Diodes: Electrical Loads
and Driving Requirements

4.1 Electrical Load Types

This chapter addresses the particular problems of powering laser diodes. This
task involves two questions:
1. What are the characteristics of a LD as an electrical device?
2. What should be the characteristics of a LDD as an electrical source
for LD power?
A LD is a specific electrical load and requires a power source with certain
characteristics. To better understand how to power a LD, consider the three
types of electrical loads:
• Resistive load. The ideal load of this type follows Ohm’s law: the load
voltage drop is directly proportional to the current flowing through the
load. The current–voltage characteristic (IVC) of the ideal resistive load
is represented by a straight line that originates at the beginning of the
I–V coordinates [Fig. 4.1(a)]. An example of this load is the filament of
an incandescent lamp. The IVC of the real resistive load differs from the
straight line because the load resistance changes during operation due to
heating. However, for a PS designer this fact can be neglected, and a
load with a voltage drop that is in some degree proportional to the load
current can be referred to as a resistive load.
• Constant voltage load. The ideal load belonging to this class maintains a
constant voltage when the current increases [Fig. 4.1(b)]. Again, the
voltage across a real load can have a small positive or negative change
with a current increase. Gas-discharge lamps, LEDs, and LDs fall into
this category, as well as energy-storing capacitors and electrical
batteries. Because devices with a constant voltage are the most
important objects of this book, their characteristics are considered in
detail later in this chapter.

17
18 Chapter 4

Figure 4.1 IVCs of ideal loads: (a) resistive load, (b) constant voltage load, and
(c) constant current load.

• Constant current load. The ideal load belonging to this class maintains a
constant current when the load voltage increases [Fig. 4.1(c)]. To this
load type belong energy-storage inductors and some electrical motors.
The characteristics of real loads are close to ideal with some degree of
approximation. They change with the operational temperature. Other factors,
such as humidity, electrical stress, and aging, also have an effect.

4.2 Characteristics of Laser Diodes

In LED and LD systems, diodes radiate light with the passage of a forward
current. This process is described in technical literature very well, and thus it
will not be covered here. Only the LD characteristics that are applicable to
LDD design are addressed.
The most-important electrical characteristic of a LD is the IVC; a typical
IVC is depicted in Fig. 4.2. The specifics of a LD IVC are as follows:

Figure 4.2 Typical IVC of a LD.

Laser Diodes: Electrical Loads and Driving Requirements 19

• In the region of working currents, the IVC has a small positive

differential impedance (dV/dI)—with a large increase in the diode
forward current, the diode forward voltage remains fairly constant. A
LD has the electrical characteristic of a constant voltage load.
• To produce light of different colors, the LDs must be made from
different semiconductor materials. The shorter the wavelength of the
generated light is, the larger the forward voltage drop on the LD with
the same value as the forward current. The forward voltage drop of LDs
emitting light in the near-IR is 1–2 V, and 3–4 V in the near-UV. The
forward voltage drop for other color LDs is within these boundaries.
• A LD does not generate a laser beam below the lasing-threshold-current
level. Exceeding the destructive-current level, even for a short amount of
time, can lead to LD damage or catastrophic failure. The working currents
of a LD are between the lasing-threshold-current and destructive-current
levels.
• Unlike silicon diodes, a LD cannot handle a large reversed voltage. The
LD reversed-distractive voltage is several volts. A LDD must protect the
LD from both forward damaging overcurrent and reversed destructive
overvoltage.
• The LD forward voltage drop has a negative dependence on the LD
temperature in the range of several mV/°C. The intensity of the laser
beam, to a certain extent, is proportional to the operational current.
• To produce significant power in industrial or military lasers, LD
assemblies called diode bars and stacks are used. Several laser diodes are
built in a bar as an integrated circuit and connected electrically in
parallel. For further power increases, several bars are connected
electrically in series to form a stack assembly. The voltage on the stack
assembly can reach several tens of volts. Chapter 9 discusses effective
methods of powering bars and stacks.

4.3 Aging and Temperature Effects on LDs

LDs degrade with operating time: defects in semiconductors appear, facet
mirrors oxidize, and so on. These processes are hastened by driving-current
surges, surges in the laser-beam intensity, excessive operational temperature,
thermal cycling, moisture, and other factors. However, the ability to keep LD
operational parameters under the values recommended by vendor specifica-
tions will prolong a LD’s life up to several tens of thousands of hours. Because
the LD lifetime depends on the operating temperature, the LDD efficiency is
essential when a LDD is mechanically integrated with a LD.
An elevated temperature affects not only the LD lifetime but also the laser
operational characteristics, i.e., laser lighting efficiency (efficacy), generated-
light frequency, and lasing threshold. As the LD temperature increases, the
20 Chapter 4

efficacy rapidly drops, and the lasing threshold and light frequency shift.
Effective cooling is necessary to keep the LD parameters stable. Although a
passive heatsink usually provides adequate cooling for low-power LDs, mid-
and high-power LDs require active cooling that uses thermoelectric (TE)
coolers or forced air or water.

4.4 Ideal and Real Sources of Electrical Energy

With the characteristics of LDs defined, the next step clarifies which electrical
sources are suitable to power LDs. Because power sources supply a load with
energy, they have negative differential impedance. The most important
characteristic for a power-electronics engineer to consider is the PS output
IVC, also called the PS load characteristic. It reflects the PS output
parameters’ behavior with the load variation.
Three ideal types of electrical sources are
• Constant voltage source (CVS). The ideal PS with a CVS output
characteristic keeps the output voltage constant despite variations in the
load and other factors (temperature, aging, etc.). A CVS has a negative
differential output impedance dV/dI, which approaches zero (–0).
• Constant current source (CCS). The ideal CCS keeps the output current
constant despite variations in the load and other factors. It has a
negative differential output impedance dV/dI, which approaches infinity
(–`).
• Constant power source (CPS). The ideal CPS keeps the output power
constant despite variations in the load and other factors. The product of
the output current to voltage is constant (Pout ¼ Vout  Iout ¼ constant).
On the IVC graphs (Fig. 4.3) the ideal CVS is presented as parallel to the
abscissa axis straight line [Fig. 5.1(a)], the ideal CCS is the straight line
parallel to the ordinate axis [Fig. 4.3(b)], and the ideal CPS is a hyperbolic
curve [Fig. 4.3(c)].

Figure 4.3 IVCs of ideal electrical sources: (a) CVS, (b) CCS, and (c) CPS.
Laser Diodes: Electrical Loads and Driving Requirements 21

The electrical load characteristics of real electrical sources are close to

ideal in some degree of approximation. First, real sources are characterized by
internal impedance. For instance, the load-absent voltage on the real CVS
electrodes is characterized by the open-circuit voltage (Voc). Because of the
internal impedance, the voltage on the electrodes of a real CVS drops with an
increase in the load current and reaches zero at the short output. The current
at this point is called the short-circuit current (Isc).

4.5 Requirements for LD Powering

As mentioned earlier, a LD IVC has a small positive impedance (constant
voltage load) in the region of the working currents. Figure 4.4 illustrates why
a CV-type load cannot be powered by a CVS: temperature increases have an
adverse effect on the LD IVC. Suppose a LD is powered by an ideal CVS with
output voltage Vc [Fig. 4.4(a)]. Powering the LD causes the semiconductor die
temperature to increase. Laser diodes have a negative dependence on the
forward voltage drop of the LD temperature in the range of 2–4 mV/°C. This
dependence will lead to a big current increase (dI) in the LD. The current
increase can produce further temperature increases and cause thermal
runaway and LD failure. The other problem of powering a constant voltage
load with a CVS: even a small ripple or drift (dV) of the CVS output voltage
[Fig. 4.4(b)] will lead to unacceptably large pulsations or drift of the load
current (dI).
When powering a LD from a CCS, the temperature increase leads to a
decrease in the voltage drop (dV) on the LD [Fig. 4.4(c)]. Neither an increase
in LD power dissipation nor thermal runaway occur. Because the LD voltage
ripple caused by the current ripple of CCS is minuscule, its effect on the laser
optical power is negligible. If the LD is powered by a CPS, the electrical
power in the LD remains constant at any moment of time.

Figure 4.4 Powering a LD with (a)–(b) a CVS and (c) a CCS. (a) The LD temperature
increase during operation produces a large current increase (dI) in the LD. (b) Even a small
voltage ripple (dV) in the CVS leads to a significant current ripple (dI) in the LD. (c) The
LD temperature increase does not produce a large voltage decrease dV in the LD powered
by a CCS.
22 Chapter 4

In conclusion, a LDD must have the output characteristic of either a CCS

or CPS but not a CVS. The general rules are thus:
• A resistive load can be powered by a CVS, CCS, or CPS.
• A constant voltage load can be powered by either a CCS or CPS.
• A constant current load can be powered by either a CVS or CPS.
As the earlier examples illustrate, plotting a LDD IVC and a LD IVC on the
same graph provides a good visual representation of their collaborative work.
This book uses such graphical representations because they are visually
superior to spreadsheet data representations.
Chapter 5
Primary and Secondary
Sources of Electrical Energy
As mentioned earlier, a LD is an electrical device that needs a source of
electrical energy with the output characteristic of a CCS or CPS for power.
This chapter examines the types of available sources.
There are a variety of devices designed to produce and supply electricity to
consumers. They transform carriers of nonelectric energy, e.g., from oil, gas,
chemical reactions, water, wind, sunlight, etc., into electrical energy (the
technical literature refers to them as the “primary sources of electrical
energy”). This transformed energy is either direct current (DC) or alternating
current (AC).
Most applications of this energy use two types of primary sources: electric
batteries (used in mobile objects such as vehicles and flashlights) and AC
generators with an associated AC grid (in which the energy from many
primary sources is combined to power stationary equipment). There are also
mobile generators, which generate electricity with parameters that match an
AC grid.
The primary sources of electrical energy can only directly power
limited types of electrical devices, such as the incandescent bulbs in
flashlights powered by batteries. Some AC motors and incandescent lamps
are powered directly from an AC line. Most electrical devices (electrical
loads) require either AC sources with a different AC voltage and/or
frequency, or DC sources with an output characteristic of a CVS, CCS, or
CPS.
This arrangement involves an intermediate device connected between
the primary electric source and the load that converts electrical energy
from the primary source into electrical energy that has the characteristics
needed by the load. These devices are called “secondary sources of
electrical energy.” In the American technical literature, they are usually
called power supplies (PSs). In relation to powering LDs, secondary
sources are identified as LDDs.

23
24 Chapter 5

An understanding of the characteristics of electrical sources and electrical

loads, as well as the interaction between them, is necessary for competent
LDD design. One of the primary goals of this book is to present the design
principles and characteristics of the secondary sources of electrical energy
devoted to powering LDs. The following section considers the characteristics
of primary sources.

5.1 Primary Sources of Electrical Energy

5.1.1 Electrical batteries
Batteries convert the energy of a chemical reaction into electrical energy
and provide a load with direct current. This conversion process is reversible,
e.g., rechargeable batteries. A battery consists of a single cell or multiple
cells connected together to store certain energy and provide a necessary
voltage and current.
Batteries are characterized by the open-circuit voltage—the voltage on
the battery electrodes in the absence of a load. When the load is connected
and current starts to flow, part of the voltage drops due to the internal
resistance, and the voltage on the battery electrodes is reduced below the
open-circuit value (it also gradually decreases during operation in the
absence of charging). Due to internal resistance during the charging
process, the voltage on the battery electrodes is higher than the open-circuit
value. During operation, which includes charging and discharging cycles,
the battery voltage can change up to two or more times. Nonetheless, a
battery can be considered as a constant DC voltage source. The design of a
LDD with such significant input voltage variation is not an easy task;
switch-mode converters are usually used to stabilize the voltage at the
LDD input.

5.1.2 Electrical-energy generation and distribution

Power stations generate electrical energy by converting nonelectrical
energy into electrical energy, and regulate and keep constant the sinusoidal
AC voltage and frequency (120 Vrms, 60 Hz in the U.S.). Energy from
multiple stations is combined and distributed to customers through an
AC-power distribution grid. However, the voltage of the AC grid is not
precisely constant. In the U.S., the standard voltage can vary
10% of the
nominal 120 V.
Furthermore, devices connected to the AC grid are occasionally affected
by voltage surges above or sags below these standard limits. These surges
occur due to lightning striking power lines, the starting or stopping of
powerful electrical equipment located nearby, etc. Regardless, an AC grid can
be considered as a constant AC voltage source.
Primary and Secondary Sources of Electrical Energy 25

5.2 Secondary Sources of Electrical Energy

5.2.1 Active secondary converters
Secondary sources of electrical energy can be divided into two groups. The
first group includes controlled PSs, also called active PSs. Their output
parameters are regulated and remain stable in the range determined by the PS
specifications. An active PS contains a feedback loop that guarantees the
stability of the controlled output parameter (voltage/current/power) despite
changes in the input voltage, load parameters, temperature, and other factors.
Most of the PSs used to power electronic equipment belong to this group. The
second group comprises uncontrolled PSs, also called passive PSs. The output
parameters of a passive PS (voltage, current) change with the input voltage,
load parameters, temperature, and other factors. However, in some
applications these changes are not critical, and passive PSs could be used.
An example of such a PS is a magnetic ballast (an inductor with several
lamp-starting components) to directly power fluorescent lamps from an AC
line. The possibility of using this simple PS is based on human vision:
the human eye is not very sensitive to brightness changes. Variations in the
fluorescent lamp brightness due to a change in the AC line voltage are in the
acceptable range.
Any real PS has internal impedance formed by the circuit elements (the
resistance and inductance of conductors, and the parasitic capacitances
between wires and components). A voltage drop on these elements during
operation reduces the PS output voltage. In DC sources, only resistances
affect the output voltage. To summarize the effect of many internal
resistances, a PS is characterized by the equivalent internal series resistance
Rint [Fig. 5.1(a)]. It acts as a real resistor Rint in series with the PS output.
Because the Rint voltage on the connected load connected to the CVS linearly
decreases with the decreasing load resistance Rload, the load current increases
and reaches the short-circuit value Isc at Rload ¼ 0.

Figure 5.1 (a) Block diagram and (b) IVC of a real passive (unregulated) DC-voltage
source.
26 Chapter 5

An output IVC or load curve of this voltage source is illustrated as a

straight line that originates at the open-circuit voltage Voc point on the y axis
and ends at the short-circuit current point Isc on the x axis [Fig. 5.1(b)]. As
mentioned earlier, a feedback loop in active PSs is used to regulate the
controlled parameter in order to eliminate the dependence of the output
parameter on parasitic elements.
Figure 5.2(a) shows an active DC voltage source powered by an
unregulated constant DC voltage source CVS with internal resistance Rint
and open-circuit voltage Voc. The output voltage senses with a voltage
resistive divider R1, R2. The voltage sense signal is compared with the
precision signal from the voltage reference source in the comparator (high-
gain operational amplifier) and the differential signal driving the active
regulating element that adjusts the load current to maintain a constant output
voltage Vc.
Due to the high gain, a feedback loop can keep the constant output
voltage Vc close to an ideal constant voltage given the load current changes.
However, when the output current increases, it can reach the maximum
current regulation level Ireg max [Fig. 5.3(b)]. At this point, an active CVS
becomes uncontrolled, and the output voltage gradually drops following the

Figure 5.2 (a) Block diagram and (b) output IVC of a real active CVS.
Primary and Secondary Sources of Electrical Energy 27

Figure 5.3 (a) Block diagram and (b) IVC of a real active CCS.

load curve of an unregulated CVS until at 0-V resistance it reaches 0 V and the
short-circuit current Isc.
An active CCS [Fig. 5.3(a)] is created by adding an active regulating
circuit to an unregulated CVS. An active regulating circuit includes a current
sense resistor Rs that converts the load current into a voltage form of the
current sense signal. The current sense signal is compared with the precision
signal from the voltage reference source with the differential signal-driving
active regulating element, which adjusts the load voltage to keep the output
current Ic constant. This active CCS can keep the constant output current Ic
close to ideal with the load resistance increase. It can reach the maximum
regulation voltage level Vreg max. At this point, the CCS becomes
unregulated, the output current starts to decrease, and the output voltage
still increases following the load curve of the CVS and reaches the open-circuit
voltage Voc at infinite load resistance.
An active CPS comprises an active regulating circuit and an unregulated
CVS, as depicted in Fig. 5.4(a). It includes a voltage divider R1, R2 to get
the voltage sense signal and a current sense resistor Rs to get the current
28 Chapter 5

Figure 5.4 (a) Block diagram and (b) IVC of a real active CPS.

sense signal. These signals are multiplied by a multiplier. The product of the
current-to-voltage signals is compared with the precision voltage reference
signal in a comparator, which drives the active regulating element and keeps
the output power constant.
An active CPS reaches the regulation limits with a load resistance decrease
as well as with a load resistance increase [Fig. 5.4(b)]. With the load resistance
decreases, it goes out of power control at the Ireg max, and the output voltage
follows the load curve of the CVS with a further resistance decrease, up to Isc.
When the load resistance increases and the output voltage reaches Vreg max,
the CPS goes out of power regulation, the output voltage increases more
slowly, and follows the load curve of the CVS with further resistance
increases, up to Voc.
Summarizing the explanations in this section,
• Depending on the conversion topology the load curve of unregulated
CVS could have a different shape.
• Different types of active secondary sources (PSs) can be made from
unregulated primary CVSs.
Primary and Secondary Sources of Electrical Energy 29

Figure 5.5 Load curve of an unregulated CVS and the regulation area. Depending on the
conversion topology, the load curve could have a different shape.

• Within the regulation area, an active control takes power over the
conversion topology specifics, and the PS behaves as a voltage, current,
or power source in the presence of the corresponding feedback.
• On the IVC graph, the PS regulation area is confined between the
voltage and current axis and the load curve of uncontrolled CVS
(shaded area in Fig. 5.5).

5.3 Passive Voltage-to-Current Converters

Along with active voltage-to-current (VtoI) conversion methods, there is a
way to create a VtoI converter with an output that is approximately close to
the characteristic of a CCS without active feedback. Passive converters are
very important: PSs created based on passive VtoI converters provide many
benefits for powering LDs and LD subsystems.
A voltage source can be converted to a current source by limiting the CVS
output current with the aid of an additional passive element connected
between the CVS and load (Fig. 5.6). The current-limiting element is also
called a ballasting element in the technical literature; this term will be used
throughout the book. Note that for a stable CVS and stable ballasting
elements, the output current will also be stable.
When converting a DC CVS into a DC CCS (DC VtoI converter), only
the resistor can serve as a ballasting element. In this case, a significant part of

Figure 5.6 Block diagram of a passive VtoI converter.

30 Chapter 5

the power dissipates in the ballasting resistor. A resistor can also be used as
ballast when converting an AC CVS into an AC CCS (AC VtoI converter),
but in order to minimize power losses it is better to avoid resistive elements
and use instead reactive ballasting elements, e.g., inductors or capacitors.
Sinusoidal AC CVSs can use either an inductor or capacitor as a ballasting
element. However, as will be explained later in this chapter, a square-wave
AC CVS can only use an inductor.

5.4 Passive Linear DC VtoI Converter with a Ballasting Resistor

To make a DC VtoI converter, the ballasting resistor should be added in series
with the CVS output. Figure 5.7(a) depicts a DC VtoI converter with an
external resistor Rext. Because the circuit does not need active control, it is
categorized as a passive VtoI converter.
The dependence of the load voltage Vload on the load current Iload (the
load curve) for a passive VtoI converter with a resistive ballasting resistor is
defined by

V load ðI load Þ :¼ V oc  I load ðRint þ Rext Þ▪ . (5.1)

In VtoI converters, usually Rext ≫ Rint, and the equation for a load curve
neglecting Rint is

V load ðI load Þ :¼ V oc  I load ðRext Þ▪ . (5.2)

The load curve of a passive VtoI converter is depicted in Fig. 5.7(b). It
illustrates how the external resistor Rext affects the PS load curve. The closer
the CCS output characteristic is to the ideal CCS characteristic (the output
impedance of ideal CCS ¼ –`), the higher the voltage Voc of the CVS and the
greater the resistor Rext value; see Fig. 5.8.
Assume that a LED must be powered at point C (IVC of LED represented
by the curve A–B), i.e., by the current ILED. In this case, the LED voltage will
be VLED. The higher the open-circuit voltage (Voc2 . Voc1) is, the larger the
resistor (Rext2 . Rext1) must be to maintain the same ILED. However, the

Figure 5.7 (a) Schematic and (b) load curves of a passive DC VtoI converter.
Primary and Secondary Sources of Electrical Energy 31

Figure 5.8 IVCs of DC VtoI converters with different open-circuit voltages and ballasting
resistors.

closer the IVC of the VtoI converter is to the IVC of the ideal CCS, the larger
the voltage drop Voc2 – VLED and the power loss PRext in the ballasting resistor
Rext:
Prext :¼ ðV oc  V LED ÞI LED ▪ . (5.3)

In order to calculate the current-limiting resistor Rext, the Voc of the CVS, the
load voltage VLED, and the load current ILED must be known. Based on
Eq. (5.2), the value of the ballasting resistor is

Rext :¼ ðV oc  V LED Þ∕I LED ▪ . (5.4)

The power loss in the ballasting resistor of a DC VtoI converter often

significantly exceeds the power delivered to the load. Therefore, resistors
are used as ballasting components only in very light loads, such as
indicator LEDs, neon lamp indicators, and He-Ne lasers. A resistor is also
used as a ballasting element in circuits with Zener diodes to obtain
reference voltages.

5.5 Passive Sinusoidal AC VtoI Converter with a Ballasting

Inductor
The inductor Lb added in series with the output of AC CVS acts as a
current-limiting element in the AC VtoI converter (Fig. 5.9). To calculate
the IVC characteristic of this type of converter, suppose there is a
sinusoidal AC CVS with frequency F and open-circuit root-mean-square
(RMS) voltage Voc.
32 Chapter 5

Figure 5.9 A passive AC VtoI converter with an inductor Lb as a ballasting element and a
resistive load Rload.

The voltage Voc of the CVS drops on the circuit impedance XZ:

V oc :¼ I load X Z ▪ , (5.5)

where the circuit impedance XZ is the square root of the sum of the squares of
the inductor impedance XLb and the load impedance Rload:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
X Z :¼ X 2Lb þ R2load ▪ , (5.6)

where the inductor impedance at the CVS AC frequency F is

X Lb :¼ 2pF Lb ▪ , (5.7)

and Lb is the inductance of inductor.

For variable Rload, equations for Iload(Rload) and Vload(Rload) can be
acquired, and a graph for the IVC of a VtoI converter can be drawn (Fig. 5.10):

V oc
I load ðRload Þ :¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ▪, (5.8)
X 2Lb þ R2load

V oc
V load ðRload Þ :¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Rload ▪: (5.9)
X 2Lb þ R2load

The normalized graph is built on the assumption that the AC voltage

Voc ¼ 100 Vrms, the CVS frequency F ¼ 100 Hz, and the inductance of
inductor Lb ¼ 1 henry.
The following conclusions can be drawn based on Fig. 5.10:
• An inductor limits the short-circuit current of the converter (in this
example, at 4 A).
Primary and Secondary Sources of Electrical Energy 33

Figure 5.10 Load curve of a sinusoidal AC VtoI converter with an inductive ballasting
element.

• Up to 20% of the Voc, the IVC is close to the characteristic of the ideal
current source. In this area, the CVS with a ballasting inductor acts as a
passive CCS.
• In applications where the proximity of the IVC to the ideal current
source characteristic is not critical, this circuit can be considered as a
VtoI converter up to 50% of the open-circuit voltage.
• From 0–50% of the load current, the circuit acts as a nonideal voltage
source (the load voltage drops from 100% to 88%).
• The area from 88% and 50% of the load voltage is a transition between a
CVS and CCS.
Disregarding any losses in the wires and core, an inductor can be
considered as a lossless ballasting element.
The formula for a ballasting inductor Lb can be derived by combining
Eqs. (5.6) and (5.7):
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi V 2oc V 2load
X 2Z  R2load I 2load
Lb :¼ ▪ :¼ ▪: (5.10)
2pF 2pF

In the past, current-limiting inductors are widely used in magnetic

ballasts for fluorescent lamps. Inductors had a large number of wire turns
wound around the iron core. Losses in the wire and the core reached
15–20% of the input power. An illustration of a passive AC VtoI converter
with a ballasting inductor and a fluorescent lamp appears in Fig. 5.11.
34 Chapter 5

Figure 5.11 Performance of an AC VtoI converter with a ballasting inductor and a

fluorescent-lamp load.

5.6 Passive Sinusoidal AC VtoI Converter with a Ballasting

Capacitor
Sinusoidal AC voltage sources can use a capacitor as a ballasting element in a
passive VtoI converter. The circuit schematic is depicted in Fig. 5.12. To
calculate the ballasting capacitor value, consider an AC CVS with an open-
circuit RMS voltage Voc and frequency F. The following expressions describe
the ballasting circuit:
V oc :¼ I load X Z ▪ , (5.11)

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
X Z :¼ X 2Cb þ R2load ▪, (5.12)

Figure 5.12 An AC VtoI converter with a capacitor Cb as a passive ballasting element.

Primary and Secondary Sources of Electrical Energy 35

Figure 5.13 Block diagram of an instant-start electronic ballast powering several

fluorescent lamps FL1 to FLN with a HF AC CVS through ballasting capacitors Cb1 to CbN.

X Cb :¼ 1∕ð2pF C b Þ▪ , (5.13)

where XCb is the impedance of capacitor Cb at frequency F.

A combination of Eqs. (5.11) to (5.13) produce the following expression
for the value of the ballasting capacitor Cb:
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
V 2oc  V 2load
C b :¼ 1∕2pF ▪: (5.14)
I 2load

The use of capacitors for ballasting is popular in instant-start electronic

ballasts (Fig. 5.13), where the rectified AC line voltage is first converted
into a stable (600 Vrms), high-frequency (42 kHz) sinusoidal voltage. This
alternative voltage is converted into individual alternative currents for each
lamp by ballasting capacitors Cb1 to CbN, which have a high impedance at
42 kHz.

5.7 Capacitors and Inductors in Laser Systems

In laser systems, capacitors and inductors are used for the following purposes:
• In AC VtoI converters, inductors (and rarely capacitors) are used as
ballasting elements to convert the AC voltage to the alternative current.
• In HF switch-mode DC-voltage-to-DC-voltage (VtoV) converters, an
output capacitor is used to reduce the load voltage ripple.
• In HF switch-mode VtoI converters, an output inductor is used to
reduce the load current ripple.
Capacitors (and rarely inductors) are used as energy-storing elements in
pulsed systems. These elements periodically accumulate an electrical energy in
the form of a capacitor charge (or inductor current) during the charging cycle
of pulsed system operation and afterward release this energy into the load
during discharge.
36 Chapter 5

Because the capacitor voltage cannot be changed instantly, the capacitor

can be considered as a constant voltage load during the charging cycle. This
fact is why a capacitor charger should be CCS or CPS. During the discharge
cycle, a capacitor functions as a voltage source powering a LD.
The LD is a constant voltage load, and thus it cannot be powered directly
from the capacitor because the current in the LD is not limited and can
destroy the LD. The application of pulses from the storage capacitor to the
LD requires the current limiters to be either a passive type (e.g., resistors,
inductors), a pulse-forming network (combination of inductors and capaci-
tors), or an active current-limiting type, e.g., a transistor with current
feedback.
The current in the inductor cannot be changed instantly, so it can be
considered as a constant current load while charging and as a constant current
source while discharging. Due to these features, an inductive-energy-storage
element can be connected to the LD directly during discharge; a CVS or CPS
is needed to charge such an element.
Chapter 6
High-Frequency Switch-Mode
Passive DC-to-DC Converters
in LD Systems

6.1 Passive Converter with a Ballasting Inductor

and Capacitive Output
This chapter addresses the most important type of DC VtoV converter topology
that can power LD subsystems. It converts a CVS with one level of voltage into
a CVS with another voltage level using an intermediate DC-to-AC voltage
converter (called an inverter in the power-electronics literature). Specifically, the
topology is a HF switch-mode quasi-resonant converter with a transformer that
matches the output and input voltage levels, a ballasting inductor in the
inverter, and a capacitive output to reduce the output voltage ripple. It
possesses many remarkable features, including the following:
• The higher the conversion frequency is, the smaller the size of the
HF converter. Modern inverters use a HF square-wave AC voltage that
is generated by chopping the DC voltage using modern power-switching
devices, such MOSFETs or insulated-gate bipolar transistors (IGBTs).
• Because this topology is a quasi-resonant converter, it allows pulse
width modulation (PWM) at a constant conversion frequency for output
voltage regulation. It also permits soft-switching in a wide regulation
range (defined later in this chapter).
• Operation at a constant frequency simplifies the construction of an
electromagnetic interference (EMI) filter.
• A ballasting inductor Lb and a resonant capacitor Cr form a resonant
tank (Fig. 9.1), which provides low-loss switching transitions.
• Due to the transformer, it is easy to match the load voltage with the
converter input voltage.
• In offline PSs, the transformer galvanic isolation of the load from the
AC line is an important safety feature.

37
38 Chapter 6

• The parasitic elements of a circuit—such as a transformer leakage

inductance, body diode, and drain-source capacitance of MOSFETs—
are effectively used for conversion purposes.
The use of an inductor instead of a capacitor at the converter output
makes this topology one of the best for the mid- and high-power VtoI
converters that can be used as LDDs. This application is discussed in detail in
Chapter 7. In VtoI converters, this topology has all of the aforementioned
benefits plus the following:
• The IVC of a passive VtoI converter in the voltage range below 50%
of the Voc is close to the CCS characteristic. In some LD-power
applications, a LDD based on this topology can be used without current
feedback.
• Due to the inherent current limit, the topology provides reliable LD
protection from current overloads.
• Also due to its inherent current limit, the topology allows the direct
parallel connection of conversion modules for the power increase in
industrial LDDs. No active current-sharing circuitry is required.
• The topology allows one to drive several LD strings from a single
converter.
The addition of a fast switching device, such as a MOSFET in parallel to a
LD, allows for direct PW modulation of the LD current in pulsed LD systems.

6.2 Converter Circuit Description and Calculation

This section examines the operation of a converter with capacitive output and
circuit calculations. Figure 6.1 depicts a schematic of the converter. The
converter topology is a half-bridge (HB) inverter with DC voltage split
capacitors C1, C2; power switches Q1, Q2 with anti-parallel diodes D1, D2;
transformer T; ballasting inductor Lb on the transformer’s primary side;
resonant capacitor Cr; bridge rectifiers D3–D6; and filter capacitor Cf.

Figure 6.1 Schematic of a HF switch-mode DC VtoV converter.

High-Frequency Switch-Mode Passive DC-to-DC Converters in LD Systems 39

The circuit operates similarly to a standard HB inverter. A PWM

controller through MOSFET gate drivers alternatively turns on and off the
MOSFET Q1 and Q2. It creates a HF square-wave voltage in node A that,
through the ballasting inductor Lb, applies to the primary winding W1 of the
transformer T. The voltage from the transformer’s secondary winding W2 is
rectified by the diode bridge D3–D6 and is delivered to the load. The filter
capacitor Cf reduces the output voltage ripple. The resonant capacitor Cr and
snubber capacitor Csn are used to organize soft-switching in the inverter.
The details of the circuit operation and waveforms are discussed during
derivation of the formulas that define the circuit operation. A circuit calculation
includes the following:
• values of the main circuit components (ballasting inductor Lb, resonant
capacitor Cr, filter capacitor Cf, and transformer T),
• peak and RMS values of the currents and voltages in the circuit,
• voltage ripple at converter output, and
• converter load characteristic (IVC).
To simplify the analysis and calculation, use the following assumptions:
• All of the elements of the converter are ideal. The voltage drops on
power switches Q1, Q2 and the bridge rectifier diodes (D3–D6) during
current conduction equal 0.
• Transformer T produces the ratio NTr ¼ W2/W1 equals 1, where W1 and
W2 are the number of turns in the primary and secondary windings.
• The CVS voltage Vdc and the output converter output voltage Vout
remain constant during the conversion cycle. Capacitors C1 and C2 are
so large that their voltages stay constant during the conversion cycle.
• The power switches Q1, Q2 turn on and off opposite each other with a
duty ratio of 50%.
The derived formulas and circuit calculation are performed with MathCad
software. For a better understanding of the circuit operation, the following
example uses normalized input and output data (see Table 6.1) in SI units.

Table 6.1 Initial data for a circuit calculation. The value of the output voltage
peak-to-peak ripple equals 1% of the nominal output voltage Vnom (dV :¼ 1).
CVS voltage Vdc :¼ 200
Acting HB voltage (Vin ¼ 0.5Vdc) Vin :¼ 100
Switching frequency Fsw :¼ 100 · 103
Switching period Tsw :¼ 1/Fsw ¼ 1  10–5
Nominal output power Pnom :¼ 100
Nominal output voltage Vnom :¼ 100
Nominal output current Inom :¼ Pnom/Vnom ¼ 1
Efficiency h :¼ 1
Turns ratio of transformer NTr :¼ W2/W1▪; NTr :¼ Vnom/Vin ¼ 1
40 Chapter 6

The first step calculates Isc, the value of the output current at the short-
circuit condition (Rload ¼ 0). An accurate calculation requires knowledge of the
converter IVC (load curve), but the derivation of an analytical expression for the
IVC in this case is impossible. The IVC for our numerical example (Fig. 6.2) was
obtained using computer simulation in LTspiceIV and verified experimentally.
Based on the load curve graph, the short-circuit current equals 1.135 of Inom:

I sc :¼ 1.135I nom ¼ 1.135: (6.1)

The current in the inductor Lb has a triangular shape (Fig. 6.3), and the
peak current is

Figure 6.2 Load curve of a DC VtoV converter: Vdc ¼ 200 V, Vin ¼ 100 V, Fsw ¼ 100 Hz,
Lb ¼ 110 uH, Cr ¼ 8 nF, Cf ¼ 10 uF, Csn ¼ 0.6 nF, and W2/W1 ¼ 1.

Figure 6.3 Waveforms of one conversion cycle at the short-circuit output (Rload ¼ 0):
(a) voltage in the node A and (b) current in the inductor Lb.
High-Frequency Switch-Mode Passive DC-to-DC Converters in LD Systems 41

I PKcs :¼ 2I sc ¼ 2.27: (6.2)

Figure 6.3 depicts the waveforms of the voltage in the node A (Fig. 6.3b)
and the current in the inductor Lb for one conversion cycle at the short-circuit
output condition.
Because the voltage Vin is applied to the inductor Lb during a quarter of
the switching period (T 0–T1), the current in Ls linearly rises and reaches the
peak value
I PKcs :¼ V in ðT sw ∕4Lb Þ▪ : (6.3)
From Eq. (6.3), the inductor Lb value is
Lb :¼ ðV in T sw Þ∕4I PKcs ¼ 1.101  10–4 : (6.4)
The resonant frequency Fr of the resonant tank formed by the inductor
Lb and capacitor Cr equals 1.75 of Fsw (the rationale for this value is explained
in Section 6.6):

F r :¼ 1.75F sw ¼ 1.75  105 : (6.5)

The period of the resonant tank is

T r :¼ 1∕F r ¼ 5.714  10–6 : (6.6)

The equation for the resonant frequency of the tank is
pffiffiffiffiffiffiffiffiffiffiffi
F r :¼ 1∕ð2p Lb C r Þ▪ : (6.7)
Based on Eq. (9.7), the resonant capacitor value is

C r :¼ 1∕ð4p2 Lb F 2r Þ ¼ 7.51  10–9 : (6.8)

In order to design the transformer T and inductor Lb, it is necessary to
know the maximum RMS current values in their windings. These currents are
maximum at the short-circuit condition IRMCsc. The ratio of the transformer
windings NTr ¼ 1, and these currents have the same values in the inductor and
in the primary and secondary windings of the transformer:
I PKsc
I RMSsc :¼ p ffiffiffi ¼ 1.1311: (6.9)
3
Now we analyze converter operation at the nominal output. Figure 6.4
depicts significant waveforms that help explain the converter operation.
Suppose at the moment T0 the switch Q2 is on and capacitor Cr is charged
to –Vo. The current in the inductor Lb is zero, and diodes D3 and D6 have
stopped conducting. In a quarter of the resonant tank (Lb, Cr) period (T0–T1),
the resonant capacitor Cr recharges from –Vo to Vo. Because the bridge diodes
42 Chapter 6

Figure 6.4 Waveforms of one conversion cycle at nominal output: (a) voltage at the node
A, (b) voltage across the resonant capacitor Cr, (c) current in the inductor Lb at Vo ¼ Vin,
(d) current in the inductor Lb at Vo . Vin, (e) current in the inductor Lb at Vo , Vin, and
(f) output voltage Vo with ripple voltage dV.

D3–D6 are not conducting during this time, the resonant tank is not loaded,
and the resonant oscillation happens. At T1, the current in the inductor
Lb reaches the maximum of IPKnom, which equals 2 Vin over the square root of
the characteristic impedance of the resonant tank:

2V in
I PKnom :¼ qffiffiffiffi ¼ 1.652: (6.10)
Lb
Cr

At moment T1, the diodes D4, D5 start to conduct, and they clamp the Cr
voltage to Vo. In the absence of a clamp, the Cr voltage can resonate up to 3Vo
(the dotted line in Fig. 6.4). During the voltage clamping time, the current in the
inductor Lb is flat until the moment T2, when the switch Q2 turns off. The current
that accumulated in the inductor Lb recharges the parasitic MOSFET drain-
source capacitors and snubber capacitor Csn (see Fig. 6.1). The inductor
Lb current then switches to D1 (moment T2) and flows into capacitors C1 and Co
(which are connected in series for this current). Because capacitors are large and
High-Frequency Switch-Mode Passive DC-to-DC Converters in LD Systems 43

maintain a constant voltage Vo ¼ Vin (the small ripple voltage on Cf is neglected),

the current in Lb linearly drops to 0 (moment T3). After that, the described
processes repeat for the next half of the switching cycle.
The next step calculates the value of the filter capacitor Cf, which provides
the required ripple reduction of the output voltage. During T0–T1, the resonant
capacitor Cr recharges, and the power from the CVS is not delivered to the
load. The filter capacitor Cf discharges to the load, and the voltage on it drops.
The capacitor Cf discharge time equals a quarter of the resonant tank
period:

dT :¼ T r ∕4 ¼ 1.429  10–6 : (6.11)

The formula for the peak-to-peak ripple voltage calculation is

dV :¼ ðI nom dTÞ∕C f ▪ : (6.12)

Based on Eq. (6.12), the capacitor Cf value is

C f :¼ ðI nom dTÞ∕ðdV Þ ¼ 1.429  10–6 : (6.13)

The described circuit is very convenient for creating step-up or step-down

voltage converters. The output voltage can be stabilized by using voltage
feedback and the PWM of the power switches.
A DC VtoV step-down converter can be used to create the isolated low-
voltage bus (24 V/48 V) from the rectified AC line voltage in cost-effective
offline LD systems.1 The bus voltage intended to power several system devices
(main controller, temperature controllers, TE cooler, optical positioning
modules, shutters, etc.) uses the point-of-load method. This approach can
avoid using front-end circuits for each subsystem device. The DC VtoV step-
up topology is also effective when creating a high-voltage PS to power a
Pockels cell (Section 12.2).
Now let us see if this topology can be used in a LDD. At first glance, the
converter characteristic in the voltage range that is below 50% of the open-
circuit voltage (Fig. 6.2) seems to be close to the CCS and thus is suitable for
powering a LD. However, the output filter capacitor caused a problem that
makes this topology unacceptable for a LDD.
An explanation of this situation involves the static and dynamic output
impedances of the converter. The static impedance (the output impedance of
the converter for slow load changes) is highly negative in a voltage range
below 50% of the open-circuit voltage. It may seem that the converter could
be used as a passive constant-current source, but because there is a capacitor
at the converter output, the output dynamic impedance (the output impedance
of the converter for the fast load changes) is barely negative. In other words,
the converter for fast load changes acts as a constant voltage source and is not
suitable for powering LDs.
44 Chapter 6

For example, if a single diode shorted in the LD string powered by this

converter, then the output capacitor will cause a large current spike in the LD
string and damage all of the other diodes. Another shortcoming of this circuit
is the inability to direct PWM of the LD current by connecting the power
switch in parallel to the LD string and the large output capacitor. The
discharge of the capacitor will damage the switch.
Consider an alternative method of LD string current PWM that starts and
stops the converter operation. In this case, fast PWM transitions are not
possible because it takes time to establish stable converter oscillation at
converter start, charging a large output capacitor and reaching a steady
current in the LD string. Discharging the output capacitor on the LD string
after the converter has stopped can also damage string diodes.
One more alternative for PWM in a converter with an output capacitor
involves connecting a power switch in series with the LD string. In this case,
when the switch disconnects the LD string, due to static impedance, the
converter acts as a current source and charges the output capacitor to a voltage
higher than the LD string nominal voltage. Upon subsequent switch activation,
a high current pulse from the output capacitor will hit the LD string and
damage it. Because laboratory PSs (which can work in static constant current
mode) contain large output capacitors, these flaws are inherent. Laser engineers
should be very careful when using laboratory PSs to power LDs.
The solution to these problems requires eliminating the output capacitor
and using an inductor connected in series with the converter output (see
Chapter 7). Both the static and the dynamic output impedances of such a
converter are highly negative. Therefore, it can be used as a LDD without
limitations. Another benefit of this topology is the possibility for direct PWM
of the LD current in pulsed systems by placing a power switch in parallel with
the LD string. However, before considering this type of LDD, the following
section discusses the losses in switch-mode converters and how to use the
parasitic elements of the converter circuit.

6.3 Conduction Losses and Optimization in VtoV Converters

As mentioned, modern LDDs are based on the HF switch-mode conversion
topologies, which can dramatically increase converter efficiency and reduce size
compared to linear mode converters. To a certain extent, the volume of a
converter is inversely proportional to the conversion frequency. However,
nothing is gained for free, and a designer must deal with the negative
consequences of HF switch-mode conversion, e.g., power losses in the converter
and EMI noise generation due to the fast transitions of currents and voltages.
There are two types of losses in HF converters: conduction losses and
switching losses. Both types affect converter parameters such as efficiency,
operating frequency, component temperature, and reliability. Every component
High-Frequency Switch-Mode Passive DC-to-DC Converters in LD Systems 45

with a conductive current has a resistance. Wires that connect components

and wires used for windings in magnetics also have a resistance. Power-switching
devices, e.g., MOSFETs, are characterized by so-called “drain-source on-
resistance.” When the current flows through these wires and devices, some of the
power dissipates and heats them. Conducting losses in switching PSs are pro-
portional to the square of the RMS currents flowing through wires and power
switches.
Consider the conditions in which the RMS current in the inverter in Fig. 6.1
is minimal. As is evident from the current waveforms [Figs. 6.4(c)–(e)], the
smallest RMS current (and thus the lowest conduction losses in the Lb and the
windings of transformer T) for a given power has a flat top [Fig. 6.4(c)].
When designing a VtoV converter with a nominal output voltage Vo that
is different than the input voltage Vin, the calculation of the transformer
should be performed so that the input voltage Vin after being transformed by
the secondary transformer T equals Vo. Given this condition, there are
minimal possible RMS currents, as well as conduction losses in the power
switches Q1, Q2 and bridge diodes D3–D6. Note that the voltage drop across
the rectifier diodes D3–D6 are neglected, but the final design should account
for it, especially for converters with a low output voltage.
The next step when reducing conduction losses is the choice of optimal
wires and semiconductor devices. The wires used in magnetics have standard
calculation methods that produce good results. There are also well-known
recommendations for choosing wire thicknesses and PCB traces in the HF
part of the conversion circuit. In the case of MOSFETs, optimization means
finding a compromise between the drain-source on-resistance of the MOSFET
and its cost. Optimization of the output bridge rectifier is related to converters
with a small output voltage. Optimization methods include the following,
which are covered in detail in the technical literature:
• Splitting the secondary winding of a transformer into two windings with
a center tap, and using two rectifying diodes instead of four.
• Using Shottky diodes, which have a smaller forward-voltage drop than
standard diodes.
• Using MOSFETs as rectifiers (synchronous rectification).

6.4 Switching Losses in Converters with Pulse-Width

Modulation
Switching losses have negative consequences in switch-mode converters:
• They create switching-power losses that, in addition to conduction losses,
heat the wires and parts of the converter. The heating effect of switching
losses limits the maximum operational frequency of the converter. The
need to reduce the operational frequency also necessitates the use of larger
46 Chapter 6

circuit-reactive components (inductors, transformers, and capacitors),

which leads to a larger overall converter size.
• They produce EMI noise, which requires filters to keep the penetration
of conductive EMI noise in the AC line and the ambient radiated EMI
noise in the ambient environment under the limits specified by the U.S.
Federal Communication Commission.
Switching losses greatly depend on the converter topology. They can be
eliminated by choosing the corresponding topology or minimized with
additional circuitry (snubbers). The resulting devices are called soft-switching
converters. Devices in which the switching losses are not eliminated or
minimized are called hard-switching converters. The circuit diagrams in
Fig. 6.5 and the waveforms in Fig. 6.6 illustrate switching losses in the power-
switching MOSFET Q1 and how snubbers can reduce them.
During one switching period, the transistor Q1 switches two times: once
when Q1 turns on (the current in the device Q1 increases, and the voltage drop
on the device decreases) and again when it turns off (the current in the device
Q1 decreases, and the voltage drop on device increases). Before Q1 turns on, the
drain-source voltage Vds equals Vdc, and the drain-source current Ids equals
zero. When Q1 turns on, the voltage Vds starts to drop, and the current Ids starts
to rise. The product of the voltage Vds to the current Ids is instantaneous power
in Q1.

Figure 6.5 Circuit diagrams for illustration switching losses in the power-switching device
MOSFET Q1: (a) without snubber components; (b) with current snubber components Lsn, D1,
R1; and (c) with voltage snubber components Csn, D2, R2.
High-Frequency Switch-Mode Passive DC-to-DC Converters in LD Systems 47

Figure 6.6 Voltage and current waveforms that illustrate hard- and soft-switching in a
MOSFET Q1 (the actual waveforms could look different).

In Fig. 6.6, the areas under the crossing trajectories of the voltage and
current (dark shaded areas) reflect energy loss in one switching cycle of Q1. The
fastest areas are turn-on and turn-off transitions, and the smaller ones are the
switching losses. The switching times of the MOSFET drain-source current
Ids and voltage Vds depend on the power of the MOSFET gate drive signal.
However, after reaching a certain level, the improvement of the gate drive
signal cannot hasten the MOSFET transitions any further. The transitions are
limited by the HF characteristics of the MOSFET: the standard switching times
are in tens of nanoseconds.
As mentioned earlier, another problem associated with fast voltage and
current transitions is the subsequent EMI generation. The faster the voltage
transition is, the broader the common-mode noise spectrum. Faster drain-
source current transitions mean a broader differential mode noise spectrum.
Broader noise spectrums are harder to suppress.
An important part of the switch-mode PS design process is the reduction
of the switching losses and the EMI generation. It is possible to reduce
switching losses and generate EMI simultaneously by slowing the speed of the
voltage or current transitions.
In order to slow the speed of the current rise in the switch Q1 at turn-on, it
is necessary to place an inductor Lsn in series [Fig. 6.5(b)]. As can be seen from
Fig. 6.6, Lsn slows the speed of the current rise and reduces turn-on losses
(lightly shaded area) in the switch.
Likewise, to reduce turn-off losses, a capacitor Csn (with additional diode
D2) is placed in parallel with Q1 [Fig. 6.5(c)]. The charging of the Csn slows the
speed of the Q1 Vds rise and reduces Q1 turn-off losses (lightly shaded area in
Fig. 6.6). The parasitic drain-source capacitance Cds of a MOSFET acts the
same way. Cds is sometimes enough for an acceptable level of Q1 turn-off losses.
The inductor Lsn with additional components D1 and R1 in Fig. 6.5(b), which is
used to remove energy from Lsn and prepare it for the next switching cycle, form
the so-called current snubber (also called a turn-on snubber). Similarly, the
capacitor Csn with additional components D2 and R2 in Fig. 6.5(c), which is used to
remove energy from Csn, form the voltage snubber (also called a turn-off snubber).
48 Chapter 6

These types of snubbers carry over switching losses from the semiconduc-
tor switching devices to the resistors, but losses still dissipate in the converter
circuit. The snubbers that move this energy to the load or return to the CVS are
called non-dissipative or lossless snubbers. Non-dissipative snubbers help to
increase the efficiency of the switching converters. Sections 6.6 and 8.2 show
how to design and calculate nondissipative snubbers in each particular case.
The slower the transition speed of the current or voltage is (which is
provided by the corresponding snubbers), the smaller the switching losses
and the EMI generation. However, if too much time is left for transitions, it
will limit the maximum frequency of the converter. The compromise is a
transition time ~5 times larger than the MOSFET switching time, indicated by
manufacturer specifications.

6.5 Resonant and Quasi-resonant Converters

It is possible to design switch-mode converters in which, due to the electrical
resonance, the current or the voltage (or both) at the power switch reaches
zero before the next switching cycle. Note that corresponding switching losses
in these circuits are absent. These converter types are called zero-current
switching (ZCS) or zero-voltage switching (ZVS) converters. Converters with
both ZCS and ZVS are called resonant converters. At first glance, they look
attractive; however, resonant converters have shortcomings that make them
unappealing in many applications. One of their primary issues is an inability
to work at a constant frequency with PWM control. At PWM, the converter
loses either ZVS or ZCS capability, and attempts to control the frequency will
complicate the EMI filter. Further problems of resonant converters are
discussed in the literature.7,8
Other converter types that work at a constant frequency and use PWM
control but have ZVS or ZCS only during part of the switching period (either
at turn-on or turn-off) are called quasi-resonant converters. Snubbers can
reduce the total switching losses and EMI to acceptable levels even in
converters working at very high frequencies. These quasi-resonant converters
have many benefits and are used widely in power electronics.

6.6 Reducing Switching Losses in HF VtoV Converters

The converter depicted in Fig. 6.1 (Fig. 6.8 includes circuit parasitic elements)
is a quasi-resonant version. One of the main benefits of this converter is how
easy it is to organize soft-switching in Q1 and Q2 during PWM operation by
adding only one snubber capacitor Csn in parallel with one of the power
switches, either Q1 or Q2.
Figure 6.7 shows the voltage and current waveforms in the power
MOSFET Q1. Assume that the converter switching frequency is 100 kHz
High-Frequency Switch-Mode Passive DC-to-DC Converters in LD Systems 49

Figure 6.7 Waveforms illustrating soft-switching in a HF VtoV converter.

(with a switching period of 10 ms); the MOSFET turn-off fall time, i.e., the
current in the MOSFET drops to 0, is 20 ns. In the initial time T0, the gate
drive signal of the switch Q1 is high, and the switch is turned on. The current
from the capacitor C1 flows through Q1, Lb, and the primary winding W1 of
transformer T. The Q1 drain-source voltage (voltage across the snubber
capacitor Csn) is 0. At the moment T1, the Q1 gate-drive voltage goes to zero,
and the switch Q1 starts turning off. A 20-ns current in Q1 reduces to 0
(the moment T3).
What happens to the Q1 drain-source voltage while it turns off? As soon as
the gate-drive voltage falls to 0, Vds starts to increase and, in the absence of
a snubber capacitor Csn, reaches the Vdc level at the moment T2. Rise time
T1–T2 is defined by the value of the current in the inductor Lb and the sum of
Q1 and Q2 drain-source parasitic capacitances CdsQ1 and CdsQ2 (see Fig. 6.8)
because at turn-off they act in parallel (CdsQ1 is charging, and CdsQ2 is
discharging). If these capacitances are small, the Q1 drain voltage quickly rises
while Q1 current is still high. This situation produces high switching losses in Q1.
In this case, the addition of a snubber capacitor Csn reduces the voltage rise
and Q1 turn-off losses.
When choosing a snubber capacitor, follow a practical rule: to reduce
turn-off losses to an acceptable amount, the total capacitance (CdsQ1, CdsQ2,
and Csn) is such that the growth time of the MOSFET drain voltage is five
times greater than the spec value of the MOSFET current fall time. For
instance, if the spec value is 20 ns, then the growth time should be ~100 ns
(T1–T4 time).
50 Chapter 6

Figure 6.8 Schematic of a quasi-resonant VtoV converter with parasitic circuit elements:
leakage inductance Ls of transformer, secondary interwinding capacitance Csec, and body
diodes D1, D2 and drain-source capacitances CdsQ1, CdsQ2 of MOSFETs Q1 and Q2.

As soon as the voltage across capacitor Csn reaches Vdc [the moment T2 in
Fig. 6.4(c)], the body diode D2 of the MOSFET Q2 turns on, and the current
of the inductor Lb starts to flow through it into the capacitor C2, linearly
dropping to zero [moment of time T3 in Fig. 6.4(c)]. During any moment
when the diode D2 conducts current [period T2–T3 in Fig. 6.4(c)], the switch
Q2 can be turned on without losses (ZVS turn-on) because the voltage value
on it is the voltage of the directly biased Q2 body diode.
This process is shown in Fig. 6.7 for the switch Q1. At T5, the Q2 gate
drive signal drops to 0. The current of Lb recharges the capacitors CdsQ1,
CdsQ2, and Csn. The Q1 drain-source voltage linearly reduces to 0. At the
moment T6, Lb current switches to Q1 body diode. At the moment T7, the Q1
gate-drive signal turns on Q1. Because the current flows through Q1 body
diode at this moment, the Q1 turns on without losses (ZVS turn-on). After
that, the described operations are repeated for the next switching cycle.
In order to avoid cross-conduction in the power switches Q1 and Q2
(which could damage them), the following switch turn-on can be delayed by
the time required to recharge the snubber capacitor. The adjustable delay is a
built-in function in modern PWM controllers.
Consider which factors limit the conversion frequency. The first limiting
factor is the switching speed of the power circuit elements (MOSFETs and
output rectifier diodes). The second factor is the switching losses, and the third
factor is the parasitic parameters of the circuit components.
The last question to resolve is the relation of the converter switching
frequency Fsw and the resonant frequency Fr of the resonant tank, comprising
Lb and Cr. While the capacitor Cr [To–T1 in Fig. 6.4(c)] recharges, all rectifier
bridge diodes D3–D6 are off, and the current from the converter input does not
reach the output. Therefore, the closer the switching frequency Fsw is to the
resonant frequency Fr, the greater the RMS currents in the conductors and
the lower the converter efficiency. On the other hand, the large decrease in the
High-Frequency Switch-Mode Passive DC-to-DC Converters in LD Systems 51

conversion frequency leads to an increase in the dimensions of the converter

components and the converter size and weight. The optimal converter
operation achieved at the Fr/Fsw ratio is close to 1.75.

6.7 Utilizing Converter Parasitic Elements

As mentioned in Section 2.4, electrical components in converter circuits is not
ideal. Some topologies force power-electronics designers to contend with
problems caused by parasitic elements, whereas other topologies can derive a
benefit from these elements.
The parasitic elements of a VtoV converter can be utilized as shown in
the circuit in Fig. 6.8. The transformer leakage inductance Ls is summarized
by the ballasting inductor Lb. In some cases, it is possible to construct a
transformer with such leakage inductance that there is no need for a separate
ballasting inductor. Also, the energy stored in Ls can be used to recharge the
parasitic drain-source capacitors of the MOSFET switches and help reach
soft-switching. The construction of transformers with excessive leakage
inductance is described in this section.
The same applies to the parasitic interwinding secondary winding
capacitance Csec of a high-voltage (HV) transformer. Due to the high number
of secondary turns, this capacitance could be large. Its value is summarized by
the resonant capacitor Cr, and if it is sufficient, there is no need to install a
separate capacitor. If it is too high, it must be taken into account during HV
converter design. The capacitance of the primary winding is usually negligibly
small.
When discussing topology, other parasitic elements of circuits play an
active role and can be used to reduce the number of components. For
instance, MOSFET parasitic diodes (body diodes) D1, D2 are also useful. As
described earlier, they act as commutation diodes. Because this current for a
period T2–T3 [Fig. 6.4(c)] decreases linearly and slowly to zero, there is no
diode reverse-recovery problem, and MOSFET body diodes (which are not
very fast in conventional MOSFETs) work quite well in converters with a
switching frequency up to a couple hundred kHz.
Also, as mentioned, the MOSFET drain-source parasitic capacitance acts
as a capacitive snubber and reduces the drain-voltage growth rate when the
MOSFET is turning off, thereby reducing turn-off losses. In some cases, this
capacitance is enough for soft-switching action, and no snubber capacitor is
required.

6.8 Transformer in a HF Converter

The transformer is one of the main components in power converters. Unlike
other circuit elements that have a variety of standard parts, in most cases the
power transformer is specific for a particular converter and should be
52 Chapter 6

designed by a power-electronics engineer. Many literature sources devoted to

transformer design are available, so this book will focus only on the
transformer features that are relevant to the converters described herein.
A transformer serves to reconcile the levels of the load voltage with the
converter input voltage. It also provides an isolation barrier between the load
and AC line (see Fig. 6.9). As mentioned earlier, a transformer is not an ideal
component. Variations from the ideal transformer are characterized by the
parasitic elements. The higher the conversion frequency is, the larger the effect
of these parasitic elements on the converter operation.
The electrical model of a real transformer with parasitic elements is
depicted in Fig. 6.9. The main transformer parameters (including parasitic
elements) are
• core material, size, and configuration;
• number of primary winding turns W1 and secondary winding turns W2;
• parasitic capacitance Cw1 of the primary winding and the parasitic
winding capacitance Cw2 of the secondary winding;
• magnetizing inductance Lu of the primary winding; and
• leakage inductance Ls between the primary and secondary windings
(shown on the primary side of the transformer).
These inductances are used in different converters for various purposes.
For instance, in a flyback converter with an isolated output, the magnetizing
inductance of the primary winding serves as a storage element during the
energy-storage stage of the conversion cycle. This energy transfers to the load
through the secondary winding during the energy-releasing phase of the
conversion cycle. A magnetic core with a corresponding air gap length and the
required number of primary winding turns produces the required magnetizing
inductance of the primary winding. In the forward, half-bridge, and full-
bridge converters, energy storage usually is not required, and no air gap is
used in the transformer.
Leakage inductance (LI) represents the imperfect magnetic coupling of the
secondary winding with the primary. Depending on the converter topology, LI
can have a positive or adverse effect on converter operation. For instance, in a
flyback converter, LI plays a negative role. During the energy-accumulation

Figure 6.9 Electrical model of a real transformer.

High-Frequency Switch-Mode Passive DC-to-DC Converters in LD Systems 53

stage of the conversion cycle, the energy accumulates not only in the
magnetizing inductance of the primary winding but also in the LI. This energy
is not transferred into the load; in order to prepare the converter for the next
conversion cycle, it must be either dissipated or returned to the primary source.
Both cases lead to energy losses and a reduction in the converter efficiency.
Flyback converter designers try to minimize LI. Flyback transformer
construction has two shortcomings: the difficulty in providing isolation
between the windings, and the large between-windings stray capacitance,
which provides the noise pass from the secondary to the primary side of the
converter. To reduce this capacitance, a Faraday shield (a single-turn, earth-
grounded piece of foil) is placed between the primary and secondary windings.
The necessity for a gap in the transformer core, improved isolation, and
Faraday shield complicates the transformer design.
The LI is useful in the HB VtoV converter presented in this section. It
depends on the configuration of the coupling windings and their respective
position. The LI between two windings can vary in the desirable range by
changing the number and distribution of winding turns and mutual
orientation. The LI does not depend on of the magnetic core parameters,
and it has good repetition from sample to sample. It makes the operation of
the power supply (utilizing the transformers with the LI) predictable and
usable in mass production.
Accurate calculation of LI is usually not possible because are too many
factors to consider. Much faster and efficient is the method of building a
transformer sample and measuring the LI. However, the calculation formulas
for the LI of several transformer configurations were developed; the most
significant ones for this book’s topic are discussed in this section.
The minimal LI is achieved in the transformer where the primary and
secondary are wound with bifilar wires and evenly distributed on the toroid
core. Formulas for toroidal transformer LI calculations can be found in
Ref. 9. The LI calculation methods for transformers made on an EE core and
pot core come from the Transformer and Inductor Design Handbook.10
Figure 6.10 shows the conventional transformer winding configuration where
the secondary winding is placed over the primary with tape insulation between
them. The formula of this type of LI transformer is
 
4pðMLTÞN p b1 þ b2
Ls :¼ cþ · 109▪ , (6.14)
a 3
where a is the winding length, b1 is the winding build of the primary, b2 is the
winding build of the secondary, c is the insulation thickness, MLT is the mean
length of the winding turns, and Np is the number of primary turns. The
dimensions are in centimeters, and inductance is in henrys.
Figure 6.11 shows the pot core sectionalized transformer where secondary
and primary windings are wound side by side. Compared to the conventional
54 Chapter 6

Figure 6.10 Conventional winding configuration of a transformer on an EE core.

Figure 6.11 Pot core transformer with a sectional winding configuration.

configuration, this transformer has excessive LI. Because the bobbin has an
insulation barrier, no tape is necessary between the secondary and primary
windings. The LI formula for this type is

4pðMLTÞN p  a þ a2 
Ls :¼ cþ 1 · 109▪ , (6.15)
b 3
where a1 is the primary winding length, a2 is the secondary winding length, b is
the winding build, c is the insulation barrier thickness, MLT is the mean
length of the winding turns, and Np is the number of primary turns. The
dimensions are in centimeters, and inductance is in henrys.
The maximum LI can be reached in a UU-core transformer with the
primary and secondary windings placed on the opposite legs of the core
(Fig. 6.12). To get the required LI, the turns of the primary winding can be
distributed: part of the primary is on one leg of the core, and the other part of
the other leg is under the secondary winding.
High-Frequency Switch-Mode Passive DC-to-DC Converters in LD Systems 55

Figure 6.12 Configuration of a HV transformer with air insulation built on a UU-core.

An exact LI calculation formula for the UU-core transformer is not

available, but it is simple to make a sample and measure the LI. Due to their
exceptional isolation properties, transformers built on U-cores are preferred in
HV power supplies.
Chapter 7
High-Frequency Switch-Mode
LD Driver Topology
The problems involved when using a converter with a capacitor at the output as
a LDD (Fig. 6.1), and the solution to these challenges, i.e., using an inductor in
series with the converter output, were discussed in Chapter 6. The topology with
the inductor at the output makes it one of the best options for a LD driver.
Figure 7.1 depicts a schematic of the LDD. The output filter inductor Lf filters
the rectified current ripple and provides a high-dynamic-output impedance,
whereas the ballasting inductor Lb provides a high-static-output impedance.
In the normalized example, a LD string has the nominal load voltage
Vnom ¼ 100 V and a nominal load current Inom ¼ 1 A. Eliminating the output
capacitor and adding the inductor to the output changes the load curve of the
converter. With the output inductor Lf ¼ 10 mH and the same remaining
circuit parameters as in the Section 6.2 numerical example (Vdc ¼ 200 V,
Vin ¼ 100 V, Fsw ¼ 100 kHz, Lb ¼ 110 uH, Cr ¼ 8 nF, and W2/W1 ¼ 1), the
output IVC of the converter is depicted in Fig. 7.2.
In order to power a LD string with the nominal load voltage Vnom ¼ 100 V
and nominal load current Inom ¼ 1 A, a converter with such an IVC is not
suitable because for this string it is an IVC area with a low negative
impedance. A transformer with a turn ratio W2/W1 ¼ 2 is necessary to have
a LDD with an IVC at Vnom ¼ 100 V in the high-negative-impedance area.
The capacitance on the secondary side of the transformer is reflected at
the primary side as a square of the secondary-to-primary turn ratio. To keep
the same resonant frequency of the Lb–Cr resonant circuit, the Cr should be
reduced four times.
The new circuit parameters are Vdc ¼ 200 V, Vin ¼ 100 V, Fsw ¼ 100 kHz,
Lb ¼ 100 uH, Cr ¼ 2 nF, Lf ¼ 10 mH, and W2/W1 ¼ 2. The output IVC of this
circuit is depicted in Fig. 7.3. This graph also shows a hypothetical IVC of
the LD string with a nominal current of 1 A and a nominal voltage of 100 V.
Also shown are the hypothetical LD-lasing-threshold current and the
hypothetical destructive current level of the LD string.

57
58 Chapter 7

Figure 7.1 Schematic of a LD driver with an inductive output and direct LD current PW
modulator.

Figure 7.2 IVC of converter with inductive output, where Vdc ¼ 200 V, Vin ¼ 100 V, Fsw ¼
100 kHz, Lb ¼ 110 uH, Cr ¼ 8 nF, Lf ¼ 10 mH, and W2/W1 ¼ 1.

Figure 7.3 illustrates how a LDD with such characteristics provides

reliable LD overcurrent protection due to the inherent current limit, even if
the distractive current is only 20% larger than nominal. In the area of 20–50%
of the open circuit voltage, the IVC is pretty close to the characteristic of the
ideal current source. In the case of powering this LDD from the stabilized
CVS, many practical cases do not require feedback to stabilize the current.
If necessary, the feedback loop with PW modulation can be applied to the
converter. The additional circuitry comprises the current sensing resistor R,
High-Frequency Switch-Mode LD Driver Topology 59

Figure 7.3 The static LDD load curve (Vdc ¼ 200 V, Vin ¼ 100 V, Fsw ¼ 100 kHz, Lb ¼ 100 uH,
Lf ¼ 10 mH, W2/W1 ¼ 2, Cr ¼ 2 nF) and the hypothetical LD string IVC. The graph also shows
the hypothetical LD-lasing-threshold current and the hypothetical destructive current level of the
LD string.

reference signal generator, and PWM control, as shown in Fig. 7.1. The
dependence of the load current on the duty ratio of the PWM signal is linear.
A switch Q3 connected in parallel with the LD string and driven by the
PWM signal generator is used for the direct PWM of the LD string current in
a broad range of frequencies and duty ratios. By using the MOSFET as a Q3
switch, very fast on and off transitions (in the range of tens of nanoseconds)
can be achieved. It is critical to install the MOSFET switch in the closest
proximity to the LD string when switching large currents. Very fast switching
of large currents causes significant voltage spikes even on the small
inductances of short wires. These spikes are dangerous for LDs. The parasitic
body diode D3 of the MOSFET Q3 shunts the LD string in the reversed
polarity and protects LDs from reverse voltage spikes.

7.1 Calculating the Filter Inductor

The following equations provide the initial data used to calculate the filter
inductor Lf:
• Inverter switching frequency: Fsw :¼ 100  103
• Inverter switching period: Tsw :¼ 1/Fsw ¼ 1  10–5
• Load voltage: Vload :¼ 100
• Load current: Iload :¼ 1
60 Chapter 7

Figure 7.4 Waveforms of the pulsing voltage Vrec at the output of rectifier bridge D4–D7
(node A) and the ripple of the load current ILoad.

The peak-to-peak load ripple current in Lf equals 10% of the nominal

current:
dLf pp :¼ 0.1 · I load :
A circuit simulation was performed in LTspiceIV to calculate the Lf ripple
current. The waveform of the pulsing voltage Vrec at the output of rectifier
bridge D4–D6 (node A in Fig. 7.1) and the waveform of the load ripple current
Iload are depicted in Fig. 7.4. As seen from the waveforms, when the rectified
voltage Vrec exceeds the load voltage Vnom, the corresponding diodes of
the bridge rectifier start conducting the current, increasing in Lf. When
Vrec , Vnom, the diodes do not conduct, and the current in Lf decreases.
The time Tdec, when Vrec , Vload, is approximately a quarter of the
switching period Tsw:
T dec :¼ T sw ∕4 ¼ 2.5  10−6 :
During this time, the Lf discharges on the load, and the peak-to-peak
current in Lf drops as described by the following equation:
dI pp :¼ ðV load ∕Lf ÞT dec ▪ :
Based on this equation, the required Lf inductance value is

Lf :¼ ðV load ∕dI pp ÞT dec ¼ 2.5  103 :

Chapter 8
Powering Medium-Power
LD Systems
8.1 Powering a LDD with a Battery
As mentioned previously, the proximity of the load curve (Fig. 7.3) of a
passive VtoI converter (Fig. 7.1) in the voltage range below 50% of the open-
circuit voltage is sufficiently close to the CCS; it can be suitable for some
LD-powering applications. The primary condition of a stable current output
of a LDD without current feedback is the stability of the LDD input voltage.
If the LDD is powered by a battery (the voltage of which varies during
operation more than two times), then a conventional switch-mode boost-type
converter is required to provide a stable voltage Vdc to the LDD input
(Fig. 8.1). A Vdc at least 10% higher than the maximum battery voltage Vbat is
sufficient for the boost converter operation. Methods of achieving soft-
switching in a boost-type converter are described later in this chapter.

8.2 Powering a LDD with a One-Phase AC Line

When powering a LDD from the AC line, the rectification of the line voltage
with a diode bridge rectifier is the first step of converting the AC line voltage to
DC voltage (Fig. 8.2). In the absence of a capacitor Cf, the output voltage of
a one-phase rectifier periodically drops to zero [dotted line in Fig. 8.3(a)]
and cannot be used for the rest of the conversion. A large electrolytic capacitor
Cf can be used at the rectifier output to support the rectified voltage level [solid
line in Fig. 8.3(a)] when the input voltage drops.
However, rectification with a large capacitor creates an issue where the
current is consumed from the AC line only when the level of rectified voltage
exceeds the voltage across the capacitor Cf. The AC line has the characteristic
of the voltage source, and the capacitor is a constant voltage load. In this
combination, the amplitude of the current pulses is limited only by the small
impedance of the AC line and the input EMI filter; therefore, these pulses
have a large amplitude. The bigger the capacitor Cf (for a 120-Hz ripple

61
62 Chapter 8

Figure 8.1 Boost converter for powering a passive LDD from a battery.

Figure 8.2 Front end of a passive LDD powered by a one-phase AC line.

Figure 8.3 Waveforms of the rectified voltage and current of a one-phase rectifier: (a) the
voltage in the absence (dotted line) and in the presence (solid line) of an electrolytic
capacitor; and (b) the AC line current in the absence (bold line) and in the presence (thin
line) of the PFC circuit.
Powering Medium-Power LD Systems 63

reduction, a bigger Cf is better) is, the shorter the duration of the current
pulses [bold line in Fig. 8.3(b)]. The effect of these current pulses on the AC
line is negative: they distort the AC line voltage sinusoid (harmonic
distortion), generate significant differential mode noise, and lead to a low
power factor (PF).11
These adverse effects can be eliminated by a power factor correction
(PFC) circuit. Figure 8.4 depicts a block diagram of a LDD powered by a
one-phase AC line with a PFC circuit. The PFC circuit is usually an active
boost, switch-mode-type converter. The PFC circuit forces the AC line current
[thin line in Fig. 8.3(b)] to coincide in phase and shape with the AC line
voltage. It also produces very low harmonic distortion and a PF close to
100%. For the AC line, a LDD with a front-end PFC circuit looks like a
resistive load. However, the role of the PFC circuit is not limited just by these
functions. Additional advantages include the following:
• A PFC circuit produces a stable constant DC voltage that can be used at
the input of a passive LDD.
• Designing a PFC with an output voltage higher than the global highest
AC line peak voltage makes it possible to create a LDD with a
universal input, i.e., such a LDD could be plugged into any AC line
worldwide.
• A high-power laser system with a high PF saves significant money
because the customer pays the electrical company for the apparent
power. The apparent power of low PFs is much greater than the power
consumed by the load.
Because a PFC brings so many benefits for offline LDDs, there is no
reason to avoid it. PFC controllers for one-phase PSs are produced by many
companies. Their operational principles and circuit calculation are very well

Figure 8.4 PFC circuit as the front end of an offline LDD.

64 Chapter 8

described in application notes and the technical literature.11 Therefore, this

book focuses only on the items necessary for LDD designs and on switching-
loss reductions in the PFC circuit. The design of three-phase PFC circuits is
much more complicated and outside the scope of this book.
As mentioned, a boost converter is a common type of PFC converter with
an output voltage Vout
10% higher than the maximum peak of the rectified
line voltage (Fig. 8.4). When the power switch Q1 is on, the energy from the
AC line accumulates in the inductor L1. When Q1 turns off, this energy is
transferred to the load through the diode D1. A filter capacitor Cf reduces the
voltage ripple.
The PFC controller receives the signals proportional to the rectified line
voltage, the input current, and the output voltage. The PFC controller
generates the PWM signal, which controls the switch Q1 and forces the input
current to follow the input voltage in phase and shape. The PFC controller
maintains the output voltage constant simultaneously.
A boost converter can operate in two modes: continuous conduction
mode (CCM), where the current in L1 and D1 does not drop to zero before
Q1 turns on [Fig. 8.5(a)], and discontinuous conduction mode (DCM), where
the current in L1 and D1 drops to zero before Q1 turns on [Fig. 8.5(b)]. Each
mode has advantages and disadvantages. Because the amplitude of the current
swing is small in CCM, for the same power consumption the RMS current in
the circuit and associated conduction losses are lower compared to DCM,
where the current swing is 100%. Furthermore, the generation of differential
mode noise is much smaller in CCM than in DCM.
The main problems of a CCM converter involve the losses when Q1 turns
on: losses in Q1 (it turns on at the full drain voltage) and the reverse recovery
losses in the diode D1. In DCM, reverse recovery losses do not occur because
the D1 current reduces to zero before Q1 turns on. These losses and the
circuitry for their reduction are discussed later in this section.

Figure 8.5 Current in the inductor L1 during one half of the AC line period: (a) in a CCM
PFC converter and (b) in a DCM PFC converter.
Powering Medium-Power LD Systems 65

The circuitry of a DCM PFC is less complicated than the CCM version,
but conduction losses and differential-mode noise generation are larger.
An engineering compromise uses a DCM PFC circuit in PSs with a power less
than 300 W and a CCM PFC in PSs with higher power.
Consider the switching losses in the boost converter operating in CCM,
starting with the diode reverse recovery losses. In CCM, the energy stored in
L1 is not completely released into the load, and the current in L1 and diode D1
continues to flow before Q1 turns on. At this moment, the Q1 drain voltage is
Vout (minus the voltage drop on the forward-biased diode D1). When Q1 turns
on, the anode D1 voltage drops to zero while the cathode stays at the Vout
potential. Because the diode D1 stores the charge in the p-n junction, the
reverse polarity on its electrodes creates a reverse current that flows from the
cathode to the anode until the charge is removed completely (Fig. 8.6).
The faster the speed of the forward current drop is, the larger the peak of
the reverse current Irrpeak and the corresponding reverse recovery energy
losses. If the speed is limited only by the inductance of wires between Q1drain
and D1anode, then the reverse recovery losses are high. An additional inductor
Lsn can be used to reduce these losses (Fig. 8.7). As seen from the pulse shape
of the diode D1 current, this action leads to the reduction of the speed of
forward current drop. When the forward current reaches zero, the only small
charge left resides in the D1 p-n junction. This action of the Lsn dramatically
reduces the reverse recovery losses. The Lsn, with the components D2 and R2
that help discharge the current in the inductor and prepare it for the next
switching cycle, form the current snubber.
It is possible to use a silicon carbide (SiC) diode instead of a silicon
diode. It is a Shottky-type diode that does not have the reverse recovery
problem, and a current snubber does not seem necessary. However, the SiC
diode has a junction capacitance that is significantly larger than a silicon

Figure 8.6 Reverse recovery losses in diode D1 without and with current snubber.
66 Chapter 8

Figure 8.7 CCM PFC circuit with dissipative current and voltage snubbers.

diode that causes a problem similar to the reverse recovery issue. The
current snubber not only reduces the D1 reverse recovery losses but also
helps to reduce the turn-on losses in Q1 and the differential-mode noise
associated with the fast switching of the L1 current into Q1 (refer back to
Fig. 6.6).
Now consider what happens at Q1 turn-off. At the turn-off moment,
the transistor Q1 conducts the full L1 current. The speed of rising of the Q1
drain voltage is limited only by the Q1 drain-source parasitic capacitance Cds.
It leads to high turn-off losses and the generation of big common mode noise.
To reduce these losses and noise used the voltage snubber consisting of the
snubber capacitor Csn, diode D3, and resistor R3. The capacitor Csn charges
through diode D3 in parallel with Cds. The Csn capacitance reduces the speed
of the Q1 drain voltage rise and thereby decreases the turn-off losses in Q1 and
the common mode noise. During the next Q1 turn-on cycle, Csn discharges to
the resistor R3 and prepares Csn for the next turn-off cycle.
The current and voltage snubbers, described earlier, transfer the energy of
the switching losses from the semiconductor devices Q1 and D1 to the resistors
R2 and R3. However, this energy still dissipates in the PFC converter. To
increase the efficiency of the PFC, this energy should be utilized. Lossless
snubbers are described next.
The following equations calculate the power losses during turn-on and
turn-off and the component values of the current and voltage snubbers.
During one cycle of the AC line, the voltage and the pulse width in the
MOSFET Q1 change in a wide range. Thus, the accurate calculation of losses
in dissipative snubbers in a CCM PFC circuit is a difficult task. The
calculation is restricted to the estimate; the peak values of voltages and
currents in the snubbers were obtained by using LTspice simulation (Fig. 8.8).
Powering Medium-Power LD Systems 67

Figure 8.8 Waveforms in the circuit in Fig. 8.7 (from top to bottom): current in D1, Q1 drain-
source current, and Q1 drain-source voltage (V[n008]).

Initial data
AC line RMS voltage: Vline :¼ 120
PFC converter output voltage: Vout :¼ 400
PFC converter output voltage: Iout :¼ 1
PFC converter output power: Pout :¼ Vout · Iout ¼ 400
Converter switching frequency: Fsw :¼ 100  103
Switching period: Tsw :¼ 1/Fsw ¼ 1  10–5
Diode D1 reverse recovery time: Trr :¼ 50  10–9
MOSFET drain-source parasitic capacitance: Cds :¼ 0.1  10–9
PFC inductor: L1 :¼ 500  10–6
From circuit simulation waveforms (Fig. 8.8),
Peak current in D1 at Q1 turn-on: ID1pk :¼ 5
Peak current in Lsn at Q1 turn-on: ILsnpkon :¼ 6.3
Q1 drain-source peak voltage at turn-on: VQ1dspkon :¼ 400
Q1 drain-source peak current at turn-off: IQ1dspkoff :¼ 6.3
Q1 drain-source peak voltage at turn-off: VQ1dspkoff :¼ 600

Current snubber calculation

The values of the current in the snubber inductor Lsn and the value of the
discharge resistor R2 are obtained. When switch Q1 is off, the full output current
flows through L1 and D1. When switch Q1 turns on, the diode D1 conducts
current in the reverse direction, and the output voltage is applied to L2.
68 Chapter 8

Let the time of current reduction in the snubber inductor L2 equal 3Trr:

dT red :¼ 3T rr ¼ 1.5  107 ,

and the discharge time of inductor L2 on the resistor R2 equals 0.1 of the PFC
switching period:

T L2dch :¼ 0.1T sw ¼ 1  106 :

The speed of the current reduction in the snubber inductor L2 is

dI red ∕dT red :¼ V Q1dspkon ∕Lsn ▪ :

Based on these equations,

Lsn :¼ V Q1dspkon dT red ∕I D1pk ¼ 1.2  105 :

The discharge time L2 on resistor R3 (current reduction in 2.78 times) is
T Lsn2dch :¼ Lsn ∕R2 ▪ :
From this equation,
R2 :¼ Lsn ∕T Lsndch ¼ 12:
The power dissipated in resistor R2 is

Lsn ðI Lsnpkon Þ2
PR2 :¼ F sw ¼ 23.814:
2

Voltage snubber calculation

Suppose that the MOSFET Q1 turn-off time TFEToff :¼ 25  10–9. A snubber
capacitor C3 is used, which slows the Q1 drain voltage rise to 3TFEToff:

dT Q1of f :¼ 3T FEToff ¼ 7.5  108 :

When Q1 is on, a full output current Iout flows through it. When Q1 turns
off, this current charges the MOSFET drain-source capacitance Cds and
snubber capacitor Csn. It reduces Q1 dV/dT per the following formula:

dV Q1of f ∕dTQ1of f :¼ I Q1dspkof f ∕ðCsn þ Cds Þ▪ :

Because the voltage on the MOSFET drain rises from zero to VQ1dspkoff,
dV Q1of f :¼ V Q1dspkof f :
Powering Medium-Power LD Systems 69

Based on these equations, the required Csn value is

I Q1dspkof f · dT Q1of f
C sn :¼  C ds ¼ 6.875  1010 :
V Q1dspkof f

When Q1 turns on, C1 discharges on resistor R3. The discharging time

(voltage reduction in 2.72 times) is

T Csndch :¼ C sn R3 ▪ :

Let the Csn discharge time equal 0.1 of Tsw:

T Csndch :¼ 0.1T sw :

Based on these equations,

R3 :¼ T Csndch ∕C sn ¼ 1.455  103 :

The power dissipating in resistor R3 is

C sn ðV Q1dspkof f Þ2
PR3 :¼ F sw ¼ 12.375:
2

The total power dissipating in the resistors of the current and voltage
snubbers is

Ptot :¼ PR2 þ PR3 ¼ 36.189:

The power loss due to switching in percentage of the converter output

power is

Ploss% :¼ ðPtot ∕Pout Þ · 100 ¼ 9.047:

Dissipative snubbers help to reduce the noise and losses in the switching
semiconductors by transferring losses in resistors. Lossless snubbers are used
to increase the efficiency of a PFC and other nonresonant PWM converters.
Figure 8.9 depicts a schematic of a CCM PFC with lossless current and
voltage snubbers. The addition of one capacitor C4 (for proper circuit
operation, the C4 value should be much larger than Csn) and one diode D4
helps to forward the energy accumulated in the capacitor Csn and in the
inductor Lsn to the output capacitor C2 and further to the load of the PFC.
References 12 and 13 are recommended for more details regarding lossless
snubber operation.
70 Chapter 8

Figure 8.9 CCM PFC circuit with a lossless current and voltage snubber.

Figure 8.10 Schematic of a compound converter: PFC and HB inverter combination.

8.3 One-Phase Compound Converter as a LDD

The combination of a PFC converter, which has a stabilized output voltage of
450 Vdc, with a HB inverter is a way to build a LDD with a universal AC
input (Fig. 8.10), i.e., one that can plug into any one-phase AC line in the
world. This LDD can be considered as a compound converter.8 The PFC and
inverter in this converter share common components. In this case, the
common components are two electrolytic capacitors C1 and C2 connected in
series; they serve as PFC output capacitors and voltage-splitting capacitors in
the HB inverter.
The HB inverter equalizes the voltage across electrolytic capacitors,
and there is no need to use balancing resistors in parallel with capacitors.
Also, up to 100 kHz of the inverter switching frequency, electrolytic
capacitors do not need film or ceramic capacitors in parallel for HF bypass.
The stabilized output voltage of the PFC provides an opportunity to create a
passive LDD.
Chapter 9
Powering High-Power
LD Systems

9.1 Three-Phase AC Line

A three-phase AC line is used to power multi-kilowatt laser systems. The
block diagram of the three-phase LDD depicted in Fig. 9.1. It consists of
an EMI filter, a three-phase bridge rectifier, a full-bridge (FB) inverter with a
FB diode rectifier, an output inductor, and a control circuit.
The essential features of this circuit are as follows:
• Due to the overlapping of the phase voltages, the rectified voltage
Vdc of the three-phase bridge rectifier has a 360-Hz ripple frequency
(in the U.S.). The RMS value of the ripple voltage dV is
5% of the
average rectified voltage Vave (Fig. 9.2). Such a small voltage ripple
excludes the necessity of a large electrolytic capacitor on the bridge
rectifier output. Only a film or ceramic capacitor Cf is required to
provide a bypass for the HF current of the inverter. Usually, the
electrolytic capacitor is the main part that affects the PS reliability,
and its elimination prolongs the LDD lifetime.
• In the absence of an electrolytic capacitor, the PF . 94%, and the total
harmonic distortion (THD) , 32%; in some cases, it is not necessary to
have a PFC circuit.
• In the absence of the PFC circuit, the bus voltage varies with the
changes of the AC main voltage (10%). For the LD current,
stabilization can be used with current feedback and PW modulation
of the inverter.
• A DC-blocking capacitor Cbl is needed to avoid saturating transformer
T1 in the FB inverter.
• When using the formulas in Section 6.2 to calculate the parameters of a
FB inverter, note that the acting voltage Vin equals the full rectified
voltage Vdc.

71
72 Chapter 9

Figure 9.1 Three-phase LDD block diagram.

Figure 9.2 Ripple voltage at the output of three-phase rectifier.

9.2 LD Bars and Stacks

In some industrial lasers, the beam power reaches several tens of kilowatts.
Because the beam power of a laser diode emitter does not exceed several
watts, in order to obtain such significant power, the beams of many laser diodes
are combined into a single beam with the help of optical devices (lenses and
mirrors) or optical fibers. Industrial systems use different electrical combina-
tions of multiple LDs, which require correspondingly powerful LDDs.
One method of producing high light power involves a diode bar—a
semiconductor chip that contains a one-dimensional array of laser diode
emitters and combines their light with the help of optical fibers. Because the
diodes are formed in the same semiconductor structure, they have identical
IVCs. Due to their proximity to each other, they also have the same
temperature during operation. These circumstances provide a unique
opportunity to connect the diode emitters in the bar in parallel and power
them without using additional components to share the currents.
The number of diodes in the bar does not usually exceed 50 units, and the
output power does not exceed 200 W. With a normal ratio of light power to
electrical power (efficacy) of 50%, this bar requires 400 W of electrical power.
Most industrial laser diodes emit light in the near-IR region (780–980 nm) and
Powering High-Power LD Systems 73

have a voltage drop of 1.8–2.0 V. For further power increases, the diode
bars are combined in stacks, i.e., the bars are electrically connected in series.
If 15 bars are connected in series, the total electrical power of the LDD will
be 6000 W.
There are two ways to build such a high-power driver for a LD stack. The
first way uses a single VtoI converter with powerful semiconductor devices, a
large transformer, inductors, and capacitors. The disadvantages of this
approach are as follows:
• Whenever new power requirements arise, a new LDD must be developed.
• High-power, slow semiconductor devices limit the maximum conversion
frequency, which leads to bad dynamic characteristics in the LDD.
• Due to the low conversion frequency, a large output filter is required.
• Moreover, one of the main drawbacks of a single converter is reliability.
A converter failure completely stops LDD system operation.
The only advantage of this approach is its minimal number of parts.
The second way to build a powerful LDD connects many relatively-
low-power, passive VtoI converter modules in parallel. Due to their current
intrinsic limitations, these converters can be connected in parallel directly. The
advantages of this approach are thus:
• Using pre-developed modules significantly decreases the design time of
the new, more-or-less powerful LDD.
• Redundant modules and replaced failed modules guarantee reliable
LDD operation.
• A very high effective-conversion frequency can be achieved by operating
the modules with a time shift; thus, a modular LDD has a very fast
dynamic response.
• A high effective-conversion frequency significantly reduces the require-
ments of the output-current filter.
• Inexpensive semiconductor devices and standard low-cost magnetic
cores can be used to lower the cost of a modular system compared to a
single converter LDD.
Figure 9.3 depicts a block diagram of a modular LDD. A modular LDD
is designed in such a way that N modules can handle the full power, and Nþ1
modules can support uninterruptable LDD operation if one module fails. The
inputs of all of the modules are connected to a DC voltage source Vdc
(rectified AC line) through individual fuses F1 – FNþ1, one of which
disconnects a failed module from the DC source. This topology allows for
“hot” replacement of failed modules.
Because each module has an intrinsic, passive current limit, the outputs of
all of the modules are connected directly in parallel. The current of a LD stack
can reach several hundred amps. A current transducer should be used with the
74 Chapter 9

Figure 9.3 Block diagram of a modular LDD containing Nþ1 passive VtoI converters
powering a single LD stack.

Hall effect when sensing such a large current. It converts the LD stack current
into an isolated voltage signal. In the comparator, this signal is compared
to the reference voltage. The signal from the comparator reaches the con-
troller, which contains the converter (voltage to PWM signal), the generator
of time-shifted signals (Fig. 9.4), and the signal distributor. These signals,
sent through FET drivers, control the MOSFETs of the corresponding
modules.

Figure 9.4 Time-shifted FET driver signals of parallel-connected passive VtoI converters.
Powering High-Power LD Systems 75

As with a single converter, the output current can be directly PW

modulated by the power switch Q5 connected in parallel with the LD stack. In
this case, two problems arise:
• The fast switching (in several tens of nanoseconds) of high currents
(several tens or even hundreds of amperes) leads to high voltage spikes
even on small inductances of wires connected between the LD stack and
power switch Q5. These voltage spikes can damage either the LD stack
or power switch. To prevent this issue, the power switch must be placed
as close as possible to the LD stack.
• The electrical efficiency of the LDD can suffer if the switch Q5 works at
a low-duty-ratio PWM; the switch usually stays on and dissipates high
power, which leads low overall efficiency of the LDD.
For instance, if the switching current is 100 A (Isw :¼ 100), then the on-
resistance of the MOSFET is 50 milliohms (Ron :¼ 0.05). The 10% duty ratio
(90% of the time current flows in the MOSFET, 10% in the LD stack) is
DC :¼ 0.1. The power dissipating in the MOSFET is almost half a kilowatt:

Psw :¼ I 2sw · Ron ð1  0.1Þ ¼ 450:

One solution to this problem forms the power switch with the parallel
connection of several MOSFETs. This arrangement reduces the on-resistance
of Q5 to an acceptable level. Another solution is illustrated in Fig. 9.5. At the
initial moment T0, all modules are disabled, and the gate drive signal of switch

Figure 9.5 Operational logic of the circuit shown in Fig. 9.2 to obtain high efficiency with a
low-duty-ratio pulsed mode.
76 Chapter 9

Q5 is low. At moment T1, the gate drive signal turns on the switch Q5.
However, because the modules are disabled, current does not flow in Q5. At
moment T2, the converter modules are enabled, and current starts to flow in
Q5. At moment T3, Q5 turns off, and the current switches to the LD stack.
Current flows in the LD stack up at the moment T4 when Q5 turns on again.
At moment T5, a signal disables the converter modules. After that, at moment
T6, the switch Q5 turns off and stays off until the start of the next cycle at
moment T7. Because power switch Q5 conducts current only during short
times (T2–T3 and T4–T5), the system has high efficiency in low-duty-ratio
operation mode. Due to the very fast MOSFET Q5 switching capability, the
circuit provides fast switching of the current in the LD stack.

9.3 Powering Multiple LD Strings from a Single Inverter

Laser companies use different approaches to create high-power industrial
lasers. While some companies use direct laser-beam generation in a single
stack, others use a large number of LD strings for the optical pumping of
optical fiber amplifiers or solid state amplifiers, and then combine the light of
the fibers in a single beam.
The electrical schematic in Fig. 9.6 shows the LDD structure for powering
multiple LD strings. It also has a modular structure; each module contains a
FB inverter powering the M number of LD strings. Ballasting inductors
Lb1–LbM limit the currents in corresponding strings. A programmable
controller (through PWM signal generators) regulates the currents in the
groups of strings powered by the corresponding inverter.

9.4 Powering RGB Lasers in Movie Projection Systems

In a laser movie-projection system, the red, green, and blue lasers generate
beams of corresponding colors. The screen image is formed by an image
processor with the help of the corresponding optics.14 The LDD maintains
constant currents in the LD strings. In this application, the topology that
allows one to power several strings from a single inverter is very effective.
Figure 9.7 illustrates a schematic of an inverter that powers RGB strings
in a laser projection system. The HB inverter is formed by capacitors C1, C2
and MOSFETS Q1, Q2, which work with a 50% duty ratio. Different-color
LDs require different nominal currents for operation. These currents are set
by the values of the corresponding ballasting inductors Lb1, Lb2, Lb3.

9.5 Programmable LED Color-Lighting Systems

This section diverges from LD powering to address another application of this
conversion topology: programmable color LED illumination. The schematic
depicted in Fig. 9.7 can be used very effectively for this purpose by adding
Powering High-Power LD Systems 77

Figure 9.6 Block diagram of a modular LDD for powering multiple LD strings.

power switches Q3–Q5 in parallel to each LED string (Fig. 9.8). These switches
provide LED light dimming by direct PWM of the LED.
The PWM dimming of LED light has advantages over analog dimming:
it does not create a color shift and has higher efficiency. At a PWM
frequency above several hundred hertz, light flicker is not visible. Due to the
very fast switching of MOSFETs, the duty ratio of the PWM can vary
practically from 0–100%, i.e., the LED light varies from full brightness to
complete darkness due to dimming. The light in each channel is controlled
by a programmable controller that provides dynamic color change. In both
cases, the galvanic isolation of the light-emitting strings from the AC line
through transformers is an important safety feature that also simplifies
control.
78 Chapter 9

Figure 9.7 Powering RGB strings in a laser projection system.

Figure 9.8 Programmable color LED lighting source.

Powering High-Power LD Systems 79

Figure 9.9 Block diagram of a powerful, modular DC CVS based on Nþ1 passive
converters with intrinsic current limiting.

9.6 Powerful Modular DC-Voltage Source

A powerful modular DC voltage source (Fig. 9.9) can be constructed similarly
to a powerful VtoI modular converter (Fig. 9.3). In this source, outputs of
Nþ1 time-shifted passive inverters with an intrinsic current limit are
connected to the single output-filter capacitor Cf. Precise voltage control is
provided by a feedback circuit, which uses the voltage signal from a resistive
voltage divider (R1, R2). As with a VtoI converter, active current sharing
between modules is not required.
Chapter 10
LD Drivers Based on Other
Topologies

10.1 Buck Converter as a LDD

At first glance, the buck converter (Fig. 10.1) looks like a very attractive
topology for creating a LDD. Compared to other topologies, it contains a
minimal number of parts: just one power switch Q1, one filter inductor Lf, and
one freewheeling diode D1. It can work in a continuous conduction mode with
PWM and regulate the average current within a wide range. By using a
corresponding inductor Lf, the current ripple in the LD could be minuscule.
Moreover, the converter has a high dynamic output impedance due to this
inductor.
However, the analysis shows that basic schematic of buck converter has
many shortcomings; in order to meet the LDD requirements, it needs
additional circuitry, as shown in Figs. 10.2 and 10.3. The following list
describes these deficiencies and the means to overcome them:
• The main drawback of a buck converter is the lack of intrinsic current
limiting because the passive buck converter is a voltage source with low
static output impedance. To provide the characteristic of the current
source is needed a current feedback loop. The feedback loop is shown in
Fig. 10.2 and contains a current sensing resistor Rs, reference voltage
source, and PWM controller with a comparator. A buck converter does
not have isolated output and might not be acceptable because of safety
reasons.
• A buck converter does not have an isolated output and might not be
acceptable for safety reasons.
• To prevent a catastrophic LD failure due to overcurrent, the LDD
should contain an instant-acting crowbar circuit, e.g., the MOSFET Q2.
The crowbar puts a short circuit on the LD string when the current in it
reaches the maximum safe level. A current signal for the crowbar circuit
could be sensed from the same Rs. After that, a fast-acting fuse F1 or

81
82 Chapter 10

Figure 10.1 Basic schematic of a buck converter used as a LDD.

Figure 10.2 Schematic of a buck converter with the additional circuitry required for LD
driving.

Figure 10.3 Buck converter with lossless voltage and current snubbers.

circuit breaker, which disconnects the buck converter from the CVS, is
necessary in order to prevent failures of the MOSFETs Q1 and Q2.
• When the CVS voltage Vdc is four or more times higher than the
nominal voltage of the LD string Vst, a buck converter becomes
inefficient because of the large pulsed currents in the circuit, which
create significant conduction losses and EMI. When powering a LDD
from the AC line, the rectified voltage Vdc is usually much higher than
Vst and better suited to the topology of the converter with a matching
transformer.
LD Drivers Based on Other Topologies 83

• Because a buck converter works in continuous conduction mode,

MOSFET Q1 and diode D1 experience switching losses and also reverse
recovery losses in diode D1. Similar to the boost converter, these
switching losses can be minimized by using lossless snubbers. The basic
schematic13 of a buck converter with lossless current and voltage
snubbers is depicted in Fig. 10.3. The operation of snubbers is not
covered in depth here because similar analysis has been performed for
lossless snubbers in a boost-type PFC circuit. It is enough to note that
snubbers further complicate the buck converter circuitry.
• Connecting converter modules in parallel to get a large current in
industrial applications can be difficult because they do not have an
intrinsic current limit. This issue is usually resolved by using active
current sharing in the master–slave configuration of converters.
A buck converter used for LD driving involves primarily a calculation of
the inductor Lf value. The initial data for this calculation are the value of the
Lf current and level of current ripple, the converter input and output voltages,
the converter switching frequency, and the duty ratio. However, some
engineers use buck converters for LD driving, in which case the converter
design (Fig. 10.1) starts by calculating the inductor Lf value to obtain the
specified level of LD string current ripple (Fig. 10.4).

Initial data
DC supply voltage: Vdc :¼ 100
Lf current ILf :¼ 1
Maximum Lf peak-to-peak ripple
current is 5% of the Lf current: dI :¼ 0.05
LD string voltage: Vst :¼ 50
Converter switching frequency: Fsw :¼ 100  103
Converter switching period: Tsw :¼ 1/Fsw ¼ 1  105
To simplify calculations, this example uses PWM with a 50% duty ratio
(DC :¼ 0.5).

Calculation
When FET Q1 turns on, the voltage applied to Lf is

V Lf :¼ V dc –V st ¼ 50:

The duration of the voltage applied to Lf is

dT Lf :¼ T sw · DC ¼ 510–6 :
84 Chapter 10

Figure 10.4 Significant buck converter waveforms: (a) voltage in the node A (Fig. 10.1)
and (b) Lf current.

The peak value of the ripple current in Lf is

dI :¼ ðV Lf ∕Lf ÞdT Lf ▪ :
From this equation, the inductor Lf value is

Lf :¼ ðV Lf ∕dI ÞdT Lf ¼ 510–3 :

10.2 Active Linear DC VtoI Converter

In VtoI converters, the power transistor can work not only in switch-mode
operation (switching between two states, either off or on) but also linearly.
Linear-mode operation of a VtoI converter requires current feedback; thus,
the circuit is classified as an active converter. Linear converters are considered
much less efficient than switch-mode converters. However, it is possible to
create a highly efficient circuit using a linear converter by adding the active
switch-mode VtoV converter at the front of the linear converter. The
combination of a switch-mode converter and a linear converter represents the
compound converter.
Figure 10.5 illustrates a block diagram of a high-efficiency compound
converter. The converter circuit consists of two parts. The first part is a
standard linear current sink, based on a MOSFET transistor Q1 with a
feedback loop, which comprises a current-sensing resistor Rs, reference
voltage source Vref1, and comparator U1. When switch Sw1 is in the top
position, the signal from resistor Rs is compared in the comparator U1 with
the signal Vdc1 from the reference voltage source, and Q1 maintains a constant
current in the LD string.
The second part of the circuit consists of a switch-mode VtoV converter
with a voltage feedback loop, including voltage divider R2, R3, voltage
reference source Vref2, and comparator U2. The goal of a VtoV converter is to
keep the minimal level of stable voltage Vst required to operate the linear
LD Drivers Based on Other Topologies 85

Figure 10.5 High-efficiency compound VtoI converter consisting of a switch-mode VtoV

converter and a linear current sink.

Figure 10.6 A switch-mode VtoV converter and linear current sink operating as a LDD for
powering a LD string. Vnom and Inom are the nominal voltage and current of the LD string,
respectively.

current sink in VtoI mode. The total circuit efficiency could be relatively high.
Figure 10.6 depicts the I-V characteristics of a switch-mode VtoV converter
and a linear current sink acting as a LDD for powering a LD string.
The described compound converter can also be used for LD pulsed
operation. In the pulsed mode, the PWM signal acts as a reference voltage
source (switch Sw1 in the bottom position). The peak voltage of the PWM
signal determines the level of current in the LD string during the Q1 on-time.
When the PWM signal drops to zero, Q1 disconnects the LD string from the
voltage source, and the current in the LD string stops.
The main shortcoming of this compound converter is the lack of an
intrinsic current limit. If the current feedback malfunctions due to noise in the
circuit, the current of the current sink is not restricted, and it can lead to a
catastrophic failure of the LD string.
Chapter 11
Passive Switch-Mode
Voltage-to-Power Converters
in Laser Systems
11.1 Passive VtoP Converter as a LDD
Some applications require a stabilized laser beam power. For this purpose,
many low-power LDs have a photodiode attached to the rear facet. Light
penetrates the rear facet mirror and is converted by the photodiode into a
signal proportional to the laser beam power. This signal is used to stabilize the
laser beam power, but in some LD systems it can be preferable to stabilize the
electrical power of the LD assembly.
Another application where a CPS is beneficial involves charging energy-
storing capacitors in pulsed laser systems. The advantage of a CPS-type
capacitor charger compared to a CCS type is that it provides a constant load
on the primary source of electrical energy during the charging cycle. Unlike
the CPS, a CCS-type capacitor charger increases the power consumption from
the primary source with an increased storage capacitor voltage.
The electrical power in a CPS can be stabilized by an active feedback loop
using the product of the LD current signal and the voltage signal according to
the block diagram depicted in Fig. 5.4(a). However, there is a passive means
(without feedback) of power stabilization. The working principal of this
converter is based on charging a capacitor with a constant frequency up to a
certain voltage and the subsequent discharge into the load.
If a voltage-to-power (VtoP) converter with a constant switching
frequency Fsw is powered by a CVS with output voltage Vdc, and the storage
capacitor value is C, the energy stored in the capacitor is

C · V 2dc
E c :¼ ▪,
2

87
88 Chapter 11

and the power delivered into the load is

C · V 2dc
Pload :¼ F sw ▪:
2

Figure 11.1 illustrates a schematic of a passive VtoP converter. The

configuration is similar to a HB inverter. Capacitors C1, C2 are used to
accumulate and release energy into the load. To fix the voltage, and thereby
the energy, in capacitors, they are shunted by diodes D1 and D2. The
recharging process of capacitors C1, C2 is completed before the next
MOSFETs switch. Due to the clamping action of the diodes D1, D2, the
voltage on the capacitors is limited by the voltage Vdc, providing stable energy
storage in the capacitors.
The filter inductor Lf should be sufficient to provide a low current ripple
to the load. The small inductor Ls provides soft-switching in MOSFETs
Q1 and Q2. The current accumulated in Ls recharges the parasitic MOSFET
drain-source capacitors CQ1, CQ2. Therefore, before the corresponding
FET turns on, the voltage on it drops to zero (ZVS turn-on). The leakage
inductance of transformer T1 can act as Ls. Energy stored in Ls is
not delivered to the load. However, it is usually too small and can be
neglected.
As before, the formulas are derived and circuit calculations are made
using a numerical example. To simplify the math, we will ignore voltage drop
on diodes and MOSFETs. Assume that a pulse generator works at a constant
frequency Fsw with a duty ratio slightly less than 50% to avoid cross-
conduction in the MOSFETs.

Figure 11.1 Schematic of a constant-power LDD based on a passive VtoP converter.

Passive Switch-Mode Voltage-to-Power Converters in Laser Systems 89

Initial data
CVS voltage: Vdc :¼ 200
Converter switching frequency: Fsw :¼ 100  103
Converter switching period: Tsw :¼ 1/Fsw ¼ 1  10–5
Nominal LD string (load) voltage: Vload :¼ 50
Nominal LD string current: Iload :¼ Pload / Vload▪
The peak-to-peak load current ripple (the same ripple level in the filter
inductor Lf) is 5% of the nominal load current:

dI load :¼ 0.05I load ▪ :

The energy-dosing capacitors C1 and C2 are equal:

C 1 :¼ 10  10–9 ; C 2 :¼ 10  10–9 :

Transformer T1 becomes a ratio

W 2 ∕W 1 :¼ 1▪ :

Calculation
The waveforms of voltage and current in Fig. 11.2 illustrate converter operation.
At the first moment T0, the voltage on the gate of the MOSFET Q1 is high; Q1 is
on, and the voltage of the node A is zero. Before the moment T1, the MOSFET
Q1 turns off; and at the moment T1, the MOSFET Q2 turns on. Capacitor
C2 discharges to zero, and the voltage on the capacitor C1 reaches Vdc.

Figure 11.2 Voltage and current waveforms in a VtoP converter.

90 Chapter 11

During period T1–T2, the capacitors C1, C2 are recharging, and the energy
is delivered through the transformer T1, bridge rectifier D3–D6, and filter
inductor Lf to the load (LD string). This energy is

ðC 1 þ C 2 ÞV 2dc
E :¼ ¼ 4  104 : (11.1)
2
Due to the large Lf inductance and relatively small capacitance of C1 and
C2, the recharge current is constant, and during period T1–T2 the voltage rise
in the node A is linear.
In the moment T2, the voltage at the node A reaches Vdc and stays at this
level until the moment T3. During period T2–T3, energy is not delivered to the
load from the CVS, and the current in the load supported by the energy
accumulates in Lf and decreases.
The recharging process of the capacitors C1, C2 repeats during period
T3–T4 when Q2 turns off and Q1 turns on. Because capacitors recharge two
times during one converter switching cycle, the power delivered to the load is

Pload :¼ 2EF sw ¼ 80: (11.2)

The nominal current of LD string is

I load :¼ Pload ∕V load ¼ 1.6:

The peak-to-peak current ripple in the load is

dI load :¼ 0.05I load ¼ 0.08:

The last item to calculate is the value of the filter inductor Lf. During time
T3–T2, Lf discharges on the LD string. Because the LD string is a constant
voltage load, the current drops linearly with speed:

dI load V
:¼ dc ▪, (11.3)
T3  T2 Lf

where

T 3 –T 2 :¼ ð1∕4ÞT sw ▪ : (11.4)

Based on Eqs. (11.3) and (11.4),

T sw · V dc
Lf :¼ ¼ 6.25  103 : (11.5)
4dI load
Passive Switch-Mode Voltage-to-Power Converters in Laser Systems 91

Equations (11.1) and (11.2) show that there are three ways to regulate the
converter output power Pload:
• Change the CVS voltage Vdc. In this case, Pload is proportional to the
square of Vdc.
• Change the switching frequency of the converter Fsw. In this case, Pload
is linearly proportional to Fsw.
• Use the corresponding capacitors C1, C2 for each project.
The dependence of the output power on the load voltage, presented in
Fig. 11.3, deserves further attention. The graphic data was obtained using
LTspice circuit simulation software.
The voltage drops in the MOSFETs, diodes D1, D2, and rectifier bridge
diodes D3–D6 were neglected while calculating the circuit using MathCad, so
the value of the power in the LTspice simulation is 3 W lower than calculated
(80 W). The graph shows that the power is fairly stable in the range of load
voltages from 40–88%. A passive VtoP converter will provide constant power
to LD strings with voltages in this range. Like their VtoI counterparts, passive
VtoP converters can be directly connected in parallel to increase the output
power.

11.2 Passive VtoP Converter as a Capacitor Charger

Returning to the VtoP graph in Fig. 11.3, as soon as Vload reaches 88% of the
nominal load voltage, the power delivered to the load starts to drop; at 98% of
the nominal load voltage, the converter switches to the constant voltage mode.
This characteristic allows for the creation of a very simple and efficient
capacitor charger that does not require active control circuitry (Fig. 11.4).

Figure 11.3 Dependence of the VtoP converter output power on the load voltage.
92 Chapter 11

Figure 11.4 Schematic of a capacitor charger based on a passive VtoP converter.

Figure 11.5 Output power and load voltage for one charging cycle in a VtoP capacitor
charger. The graph is based on a LTspice simulation of the circuit in Fig. 11.4 with the
following parameters: Vdc ¼ 200 V, C1 ¼ C2 ¼ 10 nF, Fsw ¼ 100 kHz, Cst ¼ 100 uF, W2/W1 ¼ 1,
and Lf ¼ 1 mH.

If the converter is developed with an output voltage equal to the nominal

voltage of the capacitor, upon reaching this voltage it switches into the CV
mode and maintains this voltage on the capacitor’s electrodes (Fig. 11.5). An
alternative to the passive VtoP capacitor charger is a passive VtoI charger,
shown in Fig. 11.6.
Figure 11.7 displays graphs of the output power and load voltage of a
VtoI capacitor charger with components that provide the same charging time
(7 ms) as a VtoP charger. Compared to the 100-W peak power of a VtoP
charger, the peak power of a VtoI charger reaches 150 W. Consequently, a
VtoI charger should be developed for a peak power 1.5 times larger than a
VtoP charger.
Reference 15 recommends the passive VtoP converter topology as a very
efficient electrical battery charger. Even at a deep discharge, the battery
Passive Switch-Mode Voltage-to-Power Converters in Laser Systems 93

Figure 11.6 Schematic of a capacitor charger based on a passive VtoI converter.

Figure 11.7 Output power and load voltage for one charging cycle in a passive VtoI
capacitor charger. The graph is based on a LTspice simulation of the circuit in Fig. 11.6 with
the following parameters: Vdc ¼ 200 V, C1 ¼ C2 ¼ 10 uF, Lb ¼ 155 uH, Cr ¼ 8 nF, Fsw ¼
100 kHz, Cst ¼ 100 uF, W2/W1 ¼ 0.85, and Lf ¼ 10 mH.

voltage does not drop below 30% of the nominal open circuit voltage, so the
next charging cycle starts in a constant power mode and stays in this mode
almost to the end of the charging cycle. This feature provides a constant
power load on the AC mains. As with a capacitor, upon reaching the nominal
battery voltage, the charging process stops, and the VtoP circuit maintains the
nominal voltage on the battery electrodes without additional control circuitry.
Poon, Pong, and Tse15 mentions that this circuit has “the inherent resistive
input characteristic that can be exploited for power factor correction[. . . ]when
the input is connected to AC mains.”
Chapter 12
Powering Other Components
of LD Systems
12.1 Thermoelectric Cooler
LDs need cooling to extend their life and maintain stable laser-beam char-
acteristics during operation. For low-power LDs, passive heatsinks are usually
sufficient. Medium- and high-power LD systems require forced cooling.
High-power lasers use a circulating liquid for cooling; however, liquid cooling
systems are beyond the scope of this book.
For mid-power LDs, thermoelectric coolers (TECs) offer a simple and
reliable solution. They are a kind of semiconductor diode that operates based
on the Peltier effect: the direct current flows through a p-n junction, which
forces one side of the junction to cool down and the other side heat up. The
cold side is used to cool designated objects, while the heat from other side is
removed with the help of heatsink.
The advantages of TE coolers in LD systems over other coolers include
• the absence of moving parts and associated mechanical vibration,
• precise temperature control (within fractions of a degree) by regulating
the TEC current,
• a wide variety of sizes and shapes,
• very long lifetime, and
• no liquid media or chlorofluorocarbons.
Diodes in TECs are mechanically placed in parallel between two ceramic
plates to form cold and hot plates. Electrically, the diodes are connected in
series, forming a diode string. Like a LD string, the TEC needs to be powered
by a PS with the characteristic of the current source, which is why the same
type of VtoI converter (Fig. 7.1) for powering TECs can be used. Temperature
control can be achieved by either an average current regulation (in this case, a
PWM of the inverter’s MOSFETs Q1 and Q2) or direct TEC current PWM by
using a switch Q3 parallel to the diode string. Temperature sensing should
involve a corresponding thermal sensor.

95
96 Chapter 12

12.2 High-Voltage Power Supply with a Pockels Cell

Another device, broadly used in laser systems, is the Pockels cell with a
corresponding HV driver. It is used for pulse modulation of a laser beam to
obtain a very short single pulse or very short repetitive light pulses. The
Pockels cell rotates the polarization of the laser beam when a high voltage is
applied to its electrodes. Such rotation allows the laser beam to pass through
the Pockels cell. The cell driver consists of the HV CVS and HV switches
(MOSFETs Q1, Q2). Figure 12.1 depicts a schematic of the differential-type
driver for a Pockels cell, and its operation logic is described in Ref. 16.
This section focuses on the HV CVS topology, designed to power a
Pockels cell. The VtoV converter topology, described in Chapter 6, is a good
fit for creating the HV PS. A numerical example of HV circuit calculations is
also provided.
In order to obtain DC voltages above 3 kV, modern switch-mode
converters use step-up transformers and voltage multipliers. Figure 12.2
shows a block diagram of this type of converter. It includes some previously
discussed parts: a HB inverter with a ballasting inductor Lb, transformer
T1, and a resonant capacitor Cr. It also has a half-wave voltage multiplier
circuit, which was developed by Cockroft and Walton in 1932. Figure 12.2
depicts a three-stage multiplier. The voltage multiplier is formed by the
cascade of AC capacitors Cac1–Cac3 and the cascade of DC capacitors
Cdc1–Cdc3, connected by the rectifier diodes D1.1–D3.2. Theoretically, the
number of stages is not limited.
Because the secondary transformer winding acts as a pulsed voltage, the
Cac capacitors recharge during each switching cycle, and the diodes transfer
the charges from the Cac to the Cdc capacitors. Each multiplier stage doubles
the peak voltage of the secondary winding.

Figure 12.1 HV Pockels cell driver.

Powering Other Components of LD Systems 97

Figure 12.2 Block diagram of a HV VtoV converter.

D. Major developed the most accurate formulas for calculating the main
HV multiplier parameters, i.e., the regulation and the ripple of the output
voltage.17 The formula for regulation expresses the dependence of the
multiplier output voltage Vout on the multiplier parameters:
 3 
I out 2n n2 n
V out :¼ 2 ⋅ n ⋅ V secpk  þ þ ▪, (12.1)
F sw C ac 3 2 3

where Vsecpk is the secondary peak voltage, Iout is the voltage multiplier output
current, Fsw is the converter switching frequency, Cac is the capacitance of AC
capacitors (all AC capacitors are equal), and n is the number of stages in the
voltage multiplier.
The first part on the right side of the equation is a multiplier output
voltage at no load condition (Iout ¼ 0). The second part is the regulation of the
multiplier output voltage (output voltage reduction due to load current Iout).
The regulation voltage increases linearly with the load current increase.
A second formula expresses the dependence of output peak-to-peak
voltage ripple Voutrip from multiplier parameters. The output voltage ripple is
also proportional to the output current:
 
I out nþ1
V outrip :¼ ·n ▪, (12.2)
F sw C dc 2

where Cdc is the capacitance of the DC capacitors (all DC capacitors are

equal).
98 Chapter 12

To simplify the design procedure, it is better to start with the values of the
DC capacitors Cdc equaling the AC capacitors Cac. If the output voltage
ripple exceeds the PS specification requirements, there are two alternatives to
reduce the ripple: either use larger DC capacitors or add a RC or LC filter.
Figure 12.2 shows a schematic with an output LC filter.
The specifics and practical limits of HV design are as follows:
• A HV between components and wires with different voltage potentials
causes a breakdown and a corona discharge, which destroy the HV PS.
• Adequate spacing is necessary between components and wires. The
spacing in the air (clearance distance) and along the surface (creepage
distance) should not exceed 3 kV/cm. This arrangement avoids the need
for an encapsulant or oil for isolation.
• Because the secondary winding of the HV transformer contains a vast
number of turns, its parasitic capacitance is significant. It acts in parallel
with the resonant capacitor Cr and can limit the maximum switching
frequency of the converter.
• The ability to use a transformer with a large LI simplifies the insulation
between the primary and secondary windings because they can be
placed on the opposite legs of a UU-core (Fig. 6.12).
• A high voltage across the secondary winding requires a safe creepage
distance between the first and the last layers. To ensure a safe distance
without an encapsulant, a universal-type winding is used for the
secondary. The rule states that the voltage across the secondary should
not exceed a 3-kV peak.
The following information can be used to calculate a HV converter for
powering a Pockels cell. Common HV PSs for Pockels cells have a voltage up
to 5 kV and a power up to 100 W, which will be applied to this example.

Initial data
Nominal output power: Pnom :¼ 100
Nominal output voltage: Vnom :¼ 5000
Nominal output current: Inom :¼ Pnom/Vnom ¼ 0.02
Switching frequency: Fsw :¼ 100  103
The output voltage regulation is 10% ofVnom: Vreg :¼ 0.1Vnom ¼ 500.
The output peak-to-peak voltage ripple is 1%
of Vnom: Vrip :¼ 0.01Vnom ¼ 50.

Calculation
Because the maximum secondary winding voltage should not exceed a 3000-V
peak, a one-stage voltage multiplier n ¼ 1 should be used to obtain 5000 V of
output voltage. The formula for the multiplier regulation voltage17 is
Powering Other Components of LD Systems 99

 3 
I nom 2n n2 n
V reg :¼ þ þ ▪:
F sw · C ac 3 2 3

From this formula, the value of the AC capacitors is

 3 
I nom 2n n2 n
C ac :¼ þ þ ¼ 6  1010 ▪:
F sw · V reg 3 2 3

To get a 5000-V nominal voltage for the multiplier output at the nominal
load, the peak voltage of the transformer secondary winding should be at least

V nom þ V reg
V sec :¼ ¼ 2750:
2n

To create a margin, let Vsecpk :¼ 3000. The calculation of the peak-to-peak

ripple voltage at the multiplier output Vripcalc starts with the value of the DC
capacitors equal to the value of the AC capacitors:

C dc :¼ C ac ¼ 6  10–10 ,

 
I nom nþ1
V ripcalc :¼ ·n ¼ 333:
F sw · C dc 2

Because the calculated voltage ripple Vripcalc exceeds the spec ripple Vrip
by more than seven times, use DC capacitors that are ten times larger
(Cdc10 ¼ 10Cdc) and calculate ripple again:

Figure 12.3 IVC of voltage multiplier.

100 Chapter 12

 
I nom nþ1
V ripcalc :¼ ·n ¼ 33:
F sw · 10C dc 2
This voltage ripple meets the specified requirement.
Now we plot the load curve (IVC) of the multiplier using the MathCad
graph plotting tool. In the range of output current from 0–0.2 A, the output
current is stepped-up with the 0.01-A interval
I out :¼ 0,0.02,: : : ,0:2,
and according to Eq. (12.1),
 3 
I out 2n n2 n
V out ðI out Þ :¼ 2n · V secpk  þ þ : (12.3)
F sw · C ac 3 2 3
Figure 12.3 depicts an IVC graph. As seen from the chart, by using a
voltage feedback loop with PWM of the inverter it is possible to obtain a
stable output voltage of 5000 V with an output current up to 40 mA. At the
nominal 20 mA current, the stable output voltage Vnom has a sufficient margin
to stay in regulation.
Appendix
LTspice Circuit-Simulation
Examples

A.1 VtoV Converter Circuit

Figure A.1 illustrates a VtoV converter circuit simulation, based on Fig. 6.1,
with the following conditions: Nominal load R4 ¼ 100 ohms, Vnom ¼ 100 V,
V1 ¼ 200 V, Vin ¼ 100 V, Fsw ¼ 100 Hz, Tsw ¼ 10 us, Lb ¼ 110 uH, Cr ¼ 8 nF,
Cf ¼ 10 uF, Csn ¼ 0.6 nF, and the number of turns ratio L3/L2 ¼ 1. The transient
analysis ranges from 0–5 ms.

A.2 VtoI Converter Circuit and Load-Curve Plotting

Figure A.2 illustrates a VtoI converter circuit simulation, based on Fig. 12.1,
with the following conditions: LD string load equivalent of a 100-V Zener
1N5378B, Vdc ¼ 200 V, Vin ¼ 100 V, Fsw ¼ 100 kHz, Tsw ¼ 10 us, Lb ¼ 100 uH,
Cr ¼ 2 nF, Lf ¼ 10 mH, Csn ¼ 0.6 nF, and L3/L2 ¼ 2 (inductance L3/inductance
L2 ratio ¼ 40 mH/10 mH ¼ 4). The transient analysis ranges from 0–5 ms.
To plot the load curve of a VtoI converter instead of a Zener D1, use
the resistor Rload. Vary the Rload values from 0.1 ohms to 10000 ohms
according to Table A.1. For each Rload value, run the simulation and record
the measurements of the load current Iload and the load voltage Vload in
Table A.2. In Mathcad, create vectors Iload and Vload by copying and pasting
their values from Fig. A.2 into the corresponding placeholders. The Mathcad
graph tool can create an xy plot (Fig. A.3) by placing Iload in the x-axis
placeholder and Vload in the y-axis placeholder.

101
102 Appendix

Figure A.1 VtoV converter circuit simulation. Waveforms from top to bottom: MOSFET Q1
and Q2 gate drive signals, Q1 drain voltage, Q1 drain current, current in inductor Lb, and
output voltage ripple. The waveforms are zoomed from 4.978 ms to 5.000 ms.
LTspice Circuit-Simulation Examples 103

Figure A.2 VtoV converter circuit simulation. Waveforms from top to bottom: Q1 drain
voltage, Q1 drain current, current in inductor Lb, and D1 current ripple. The waveforms are
zoomed from 4.978 ms to 5 ms.
104 Appendix

Table A.1 Values for circuit simulation.

Rload Iload Vload
0.1 1.150 0.115
1 1.100 1.100
5 1.060 5.300
10 1.040 10.400
15 1.020 15.300
20 1.010 20.200
25 1.000 25.000
50 0.993 49.650
75 0.990 74.250
100 0.984 98.400
150 0.941 141.150
200 0.857 171.400
250 0.756 189.000
500 0.425 212.500
1000 0.218 218.000
5000 0.044 220.000
10000 0.023 230.000

Table A.2 Vector values from Table A.1 for Mathcad.

LTspice Circuit-Simulation Examples 105

Figure A.3 Mathcad xy plot.

A.3 VtoP Converter Circuit

Figure A.4 illustrates a VtoP converter circuit simulation, based on Fig. 11.1,
with the following conditions: LD string load equivalent is a 51-V Zener
1N5369B,Vdc ¼ 200 V, Fsw ¼ 100 kHz, Tsw ¼ 10 us, transformer T leakage
inductance Ls ¼ 10 uH, Lf ¼ 10 mH, and L3/L2 ¼ 1. The transient analysis
ranges from 0–5 ms.

A.4 VtoP Converter as a Capacitor Charger Circuit

Figure A.5 illustrates a simulation of a VtoP converter serving as a capacitor
charger circuit, based on Fig. 11.4, with the following conditions: storage
capacitor Cst ¼ 100 uF,Vdc ¼ 200 V, Fsw ¼ 100 kHz, Tsw ¼ 10 us, transformer
T leakage inductance Ls ¼ 10 uH, Lf ¼ 1 mH, and L3/L2 ¼ 1. The transient
analysis ranges from 0–5 ms.
106 Appendix

Figure A.4 VtoP converter circuit simulation. Waveforms from top to bottom: Q1 drain
voltage, node n009 voltage, and current in inductor Ls. The waveforms are zoomed from
4.978 ms to 5 ms.
LTspice Circuit-Simulation Examples 107

Figure A.5 Simulation of a VtoP converter serving as a capacitor charger circuit.

Waveforms for the Cst charging cycle: Cst voltage V[n003] (thin curve), and power in Cst:
V[n003]I(Cst). Waveforms are in the range of 0–10 ms.
References
1. I. M. Bystryak and G. A. Trestman, Powering and Integration of Laser
Diode Systems, professional development course, Photonics West,
San Francisco, CA (2015).
2. http://www.ptc.com/engineering-math-software/mathcad/free-trial
3. http://www.linear.com
4. http://www.smpstech.com/books/best-power-supply-books.html
5. R. Lenk and C. Lenk, Practical Lighting Design with LEDs, John Wiley &
Sons, Hoboken, NJ (2011).
6. D. Hodgson and B. Olsen, Protecting Your Laser Diode, application
note #3, ILX Lightwave Corp., Bozeman, MT (2003).
7. A. I. Pressman, Switching Power Supply Design, 2nd ed., McGraw-Hill,
New York (1998).
8. R. Lenk. Practical Design of Power Supplies, IEEE Press, John Wiley &
Sons, Hoboken, NJ (1998, 2005).
9. http://d4magnetics.com/page5/page5.html
10. W. T. McLyman, Transformer and Inductor Design Handbook, 3rd ed.,
Marcel Dekker, Inc., New York (2004).
11. PFC handbook. ON semiconductor (2014).
12. K. M. Smith and K. M. Smedley, “Lossless, Passive Soft-switching Methods
for Inverters and Amplifiers,” IEEE PESC Proceedings, 1–9 (1997).
13. X. Wu, X. Jin, L. Huang, and G. Feng, “A lossless snubber for DC/DC
converters and its application in PFC,” Third International Power
Electronics and Motion Control Conf. Proc. 3, 1144–1149 (2000).
14. https://www.christiedigital.com/en-us/display-technology/laser-projection/
rgb-laser-projection
15. N. K. Poon, B. M. H. Pong, and C. K. Tse, “A Constant-Power Battery
Charger with Inherent Soft-switching and Power Factor Correction,”
IEEE Trans. Power Electronics 18(6), 1262–1269 (2003).
16. http://www.fustpulse.com/pdf/5046_manual.pdf
17. D. Major, “Behavior of E.H.H rectifier units under steady state conditions
and under transient conditions,” AEI research report, TP/R5625.

109
Index
A E
alternating current (AC), 23 electrical characteristics, 8
electromagnetic interference (EMI)
B filter, 37
ballasting capacitor, 35 electromagnetic interference (EMI)
ballasting inductor, 33 noise, 46
battery, 24 elevated temperature, 19
buck converter, 81, 82
bus voltage, 12, 43 F
Faraday shield, 53
C feedback loop, 25, 26, 100
circuit calculation, 10, 39 filter inductor, 59
compliance, 12 flyback converter, 52, 53
computer simulation, 10
conduction losses, 45 G
constant current load, 18 generator of time-shifted signals, 74
constant current source (CCS), 20
constant power source (CPS), 20 H
constant voltage load, 17 half-bridge (HB) inverter, 38
constant voltage source (CVS), 20
continuous conduction mode I
(CCM), 64 inverter, 70, 71
cross-conduction, 50
current ripple, 21 L
current–voltage characteristic laser diode (LD), 8
(IVC), 8 stack assembly, 19
string, 44, 57
D laser diode driver (LDD), 1
differential impedance, 19 design priorities, 12, 13
diode bar, 19, 72 modular, 73
direct current (DC), 23 leakage inductance (LI), 52
discontinuous conduction mode light-emitting diode (LED), 1, 15
(DCM), 64 LTspice, 4, 101

111
112 Index

M S
magnetic ballast, 25 Shottky-type diode, 65
MathCad, 2, 10 silicon carbide (SiC) diode, 65
modular DC voltage source, 79 snubber, 47, 48, 51, 66, 67
capacitor, 49
O dissipative, 69
open-circuit voltage, 24 lossless, 9, 69
specifications, 7
P switch-mode converter, 96
p-n junction, 65, 95 switching losses, 45–47, 65
parasitic elements, 51, 52
Peltier effect, 95 T
Pockels cell, 96, 98 thermoelectric cooler (TEC), 95
power factor correction (PFC) topology, 9, 51
circuit, 63–66 switch-mode convention, 9
power supply (PS), 23 total harmonic distortion
prototype, 11 (THD), 71
pulse width modulation
(PWM), 37 V
voltage-to-current (VtoI)
Q conversion, 29, 30, 34
quasi-resonant converter, 37 voltage-to-power (VtoP)
converter, 87
R voltage-to-voltage (VtoV)
reliability, 13 converter, 35
resistance, 9
resistive load, 17 Z
resonant tank, 41 zero-current switching (ZCS), 48
reverse recovery losses, 64, 65 zero-voltage switching (ZVS), 48
 Grigoriy Trestman has over 40 years of experience in
developing switch-mode power supplies for gas discharge
lasers, electronic ballasts for fluorescent lamps, drivers for
laser diodes, and LEDs. He holds a Masters in Physics and a
Ph.D. in Photochemistry.
Trestman worked as a Senior Research Scientist in the
Department of Research and Development of Tajik State
University (USSR) for 20 years. After immigrating to the
United States, he worked as a Power-Electronics Design Engineer for several
U.S. companies (Glassman High Voltage Inc., Converter Power Inc., Philips
Color Kinetics Inc.) for another 20 years. For the past 13 years, he worked as
a Staff Engineer for Osram Sylvania Inc. He is currently a power-electronics
consultant.
Trestman is a co-author and instructor of several professional development
courses: “Powering and Integration of Laser Diode Systems;” “Electronic
Ballast Design;” “Electrical Ballasts for AC, DC, and Pulsed Loads;” and
“Powering Lasers and Laser Systems.” He holds eight U.S. patents and
14 patents in the former U.S.S.R. for various aspects of power conversion,
electrical supply of CW, and pulse devices.