Visit https://ebookluna.

com to download the full version and

explore more ebooks

Power Electronic System Design: Linking Differential

Equations, Linear Algebra, and Implicit Functions
1st edition - eBook PDF

_____ Click the link below to download _____

https://ebookluna.com/download/power-electronic-system-
design-linking-differential-equations-linear-algebra-
and-implicit-functions-ebook-pdf/

Explore and download more ebooks at ebookluna.com

Here are some recommended products that might interest you.
You can download now and explore!

Linear Algebra and Partial Differential Equations 1st

edition- eBook PDF

https://ebookluna.com/download/linear-algebra-and-partial-
differential-equations-ebook-pdf/

ebookluna.com

(eBook PDF) Differential Equations and Linear Algebra 3rd

Edition

https://ebookluna.com/product/ebook-pdf-differential-equations-and-
linear-algebra-3rd-edition/

ebookluna.com

(eBook PDF) Differential Equations and Linear Algebra 4th

Edition

https://ebookluna.com/product/ebook-pdf-differential-equations-and-
linear-algebra-4th-edition-2/

ebookluna.com

Differential Equations and Linear Algebra 4th Edition

(eBook PDF)

https://ebookluna.com/product/differential-equations-and-linear-
algebra-4th-edition-ebook-pdf/

ebookluna.com
(eBook PDF) Differential Equations and Linear Algebra 4th
Edition

https://ebookluna.com/product/ebook-pdf-differential-equations-and-
linear-algebra-4th-edition/

ebookluna.com

Differential Equations and Linear Algebra 4th Edition

(eBook PDF)

https://ebookluna.com/product/differential-equations-and-linear-
algebra-4th-edition-ebook-pdf-2/

ebookluna.com

(eBook PDF) Linear System Theory and Design 4th Edition

https://ebookluna.com/product/ebook-pdf-linear-system-theory-and-
design-4th-edition/

ebookluna.com

(eBook PDF) Power System Analysis and Design 6th Edition

https://ebookluna.com/product/ebook-pdf-power-system-analysis-and-
design-6th-edition/

ebookluna.com

Power System Analysis and Design 6th Edition J. Duncan

Glover - eBook PDF

https://ebookluna.com/download/power-system-analysis-and-design-ebook-
pdf/

ebookluna.com
POWER ELECTRONIC
SYSTEM DESIGN
Linking Differential
Equations, Linear Algebra,
and Implicit Functions

KENG C. WU
Switching Power, Inc. Ronkonkoma, NY, United States
POWER ELECTRONIC
SYSTEM DESIGN
Linking Differential
Equations, Linear Algebra,
and Implicit Functions
Elsevier
Radarweg 29, PO Box 211, 1000 AE Amsterdam, Netherlands
The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom 50
Hampshire Street, 5th Floor, Cambridge, MA 02139, United States

Copyright © 2021 Elsevier Inc. All rights reserved.

No part of this publication may be reproduced or transmitted in any form or by any means,
electronic or mechanical, including photocopying, recording, or any information storage
and retrieval system, without permission in writing from the publisher. Details on how to
seek permission, further information about the Publisher’s permissions policies and our
arrangements with organizations such as the Copyright Clearance Center and the
Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions.

This book and the individual contributions contained in it are protected under copyright
by the Publisher (other than as may be noted herein).

Notices
Knowledge and best practice in this field are constantly changing. As new research and
experience broaden our understanding, changes in research methods, professional practices,
or medical treatment may become necessary.

Practitioners and researchers must always rely on their own experience and knowledge in
evaluating and using any information, methods, compounds, or experiments described
herein. In using such information or methods they should be mindful of their own safety
and the safety of others, including parties for whom they have a professional responsibility.

To the fullest extent of the law, neither the Publisher nor the authors, contributors, or
editors, assume any liability for any injury and/or damage to persons or property as a matter
of products liability, negligence or otherwise, or from any use or operation of any methods,
products, instructions, or ideas contained in the material herein.

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library

Library of Congress Cataloging-in-Publication Data

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

ISBN: 978-0-323-88542-3

For Information on all Elsevier publications visit our website at

https://www.elsevier.com/books-and-journals

Publisher: Graham Nisbet

Acquisition Editor: Graham Nisbet
Editorial Project Manager: Grace Lander
Production Project Manager: Prasanna Kalyanaraman
Cover Designer: Brian Salisbury

Typeset by Aptara, New Delhi, India

Dedicated to
Grandson
Mitchell Patrick Wu Cayanni
Contents

About the Author xi

Preface xiii

1. Capacitor and inductor 1

1.1 Capacitor equation in differential form 1
1.2 Capacitor equation in integral form 2
1.3 Inductor equation in differential form 3
1.4 Inductor equation in integral form 4
1.5 Definition of inductance and Faraday’s law 4
1.6 Magnetic coupling and mutual inductance 5
1.7 Transformer equation 7
1.8 Nonideal capacitor, nonideal inductor, and equivalent circuit 10
1.9 Transformer equivalent circuits 11
1.10 Physical size of capacitor and inductor 13
1.11 Specifications for capacitor and inductor 15

2 First-order circuits 19
2.1 RC network with periodic drive source 19
2.2 Sawtooth (triangle ramp) generator 30
2.3 Full-wave rectifier with RC load 33
2.4 A brushless DC Motor with permanent magnets rotor 38
2.5 A BLDC motor speed detector 45
References 47

3 Current source 49
3.1 Semiconductor diode equation 49
3.2 Simple current source 50
3.3 Bob Widlar current source 54
3.4 Improved current source 58
3.5 Source impedance 60
3.6 555 timer 64
3.7 Precision current loop 70
3.8 Current-mode laser driver 74
3.9 LED array driver 76
3.10 JFET current source 77
3.11 MOSFET current source 78

vii
viii Contents

4 Second order 81
4.1 Form 81
4.2 Root 83
4.3 Time domain 85
4.4 Frequency domain 89
4.5 Parallel and serial resonance 92
4.6 Eigen value approach 103
4.7 RC filters and Sallen–Key filters 104
4.8 Power filters 111
4.9 Oscillator 113
4.10 Implicit function 120

5 Gain blocks 123

5.1 Class-A direct-coupled bipolar transistor amplifiers 123
5.2 Class-AB, B, C bipolar transistor amplifiers 129
5.3 Transformer-coupled transistor amplifiers 133
5.4 Class-D switch-mode power amplifiers 135
5.5 Pulse width modulator 139
5.6 Digital (clocked) window comparator 140
5.7 Linear operational amplifiers 142
5.8 Tuned amplifiers and implicit function 147
5.9 Composite nonlinear operational amplifiers 150
5.10 Unity-gain bandwidth of op-amp 153
5.11 Large signal gain of op-amp 156

6 Feedback approaches 167

6.1 Voltage feedback 167
6.2 Current feedback 170
6.3 PID feedback 175
6.4 State feedback 178
6.5 Feedback isolation 180

7 Control practices 189

7.1 Level control 189
7.2 Mode control 190
7.3 Zone control 192
7.4 Variable structures 193
7.5 Sensor 196
7.6 Open loop 198
 Contents ix

7.7 Close loop 200

7.8 Loop contention 203
7.9 Time control 204
7.10 Sequential time control 206

8 Linear regulator 213

8.1 Bipolar series voltage regulator 213
8.2 MOSFET series voltage regulator 223
8.3 Multiple implicit function approach 227
8.4 Design procedure for loop stability 228
8.5 Design procedure for error amplifiers 230
8.6 Current-mode laser driver design procedure 236
8.7 Shunt regulators 238

9 Switch-mode DC/DC converters 241

9.1 Power filter, inductor, and capacitor 243
9.2 Fundamental topologies 249
9.3 Operational dynamics of basic buck topology 254
9.4 Operational dynamics of basic boost topology 257
9.5 Operational dynamics of basic flyback converter 259
9.6 Cascaded converter—nonisolated 261
9.7 Isolated converter—forward converter 264
9.8 Isolated converter—half-bridge converter 269
9.9 Isolated converter—push–pull converter 272
9.10 Isolated converter—full-bridge converter 272
9.11 Isolated converter—quasi-resonant converter 273
9.12 Analog feedback 275
9.13 Close loop—analog 288
9.14 Close loop—digital 296

10 AC drives, rectification, and inductive loads 299

10.1 Reexamine RC-loaded rectifier 299
10.2 AC drive with unidirectional RL load 301
10.3 Half-wave AC drive with nonpulsating current feeding RL load 304
10.4 Full-wave AC drive with nonpulsating current feeding RL load 305
10.5 Phase-controlled AC drive with RL load 307
10.6 Phase-controlled AC drive with free-wheel diode and RL load 309
10.7 Phase-controlled full-wave AC drive with RL load 311
10.8 Three-phase circuits 313
x Contents

11 Rotation, three-phase synthesis, and space vector concepts 319

11.1 Magnetic field (flux) 319
11.2 Synthesis of three-phase sources and inverters 323
11.3 Vector concept 331

Appendix A Accelerated steady-state analysis for a parallel resonant

network fed by nonsinusoidal, half-wave rectified current 347

Appendix B Matrix exponential 349

Appendix C Example 4.7 MATLAB m-file 351

Appendix D Example 8.1 353

Appendix E A general mass-spring-dashpot second-order system;

first alternative 359

Appendix F A general mass-spring- dashpot second-order system;

second alternative 363

Appendix G A general mass-spring- dashpot second-order system;

third alternative 365

Appendix H Matrix exponential—Jordan form 367

Appendix I A step-by-step primer on digital power-supply design 369

Digital tides 369
Tumble to digital 369
Roadmap to digital 370
Navigate to digital filter 371
Work out a forward converter example 373
Implementation 377
Conclusion 380
References 381

Appendix J Motor winding driven by SCR phase-controlled

sine source 383
Index 385
About the Author

Keng C. Wu, a native of Chiayi( ), Dalin( ), Taiwan, received the B.S.

degree from Chiaotung University, Taiwan, in 1969 and the M.S. degree
from Northwestern University, Evanston, Illinois in 1973.
He was a lead member, technical staff, of Lockheed Martin, Moorestown,
NJ. He has published five books: Pulse Width Modulated DC-DC Converters
Chapman & Hall, January 1997; Transistor Circuits for Spacecraft Power System
Kluwer Academic Publishers, November 2002; Switch-mode Power Converters:
Design and Analysis Academic Press, Elsevier, November 2005; Power Recti-
fiers, Inverters, and Converter Lulu.com November 2008.; Power Converters with
Digital Filter Feedback Control, Elsevier, Academic Press, 2016. He holds a
dozen US patents, was awarded Author of the Year twice (2003 and 2006
at Lockheed Martin), and presented a 3-hour educational seminar at IEEE
APEC-2007 S17.

xi
 Visit https://ebookluna.com
now to explore a diverse
collection of ebooks available
in formats like PDF, EPUB, and
MOBI, compatible with all
devices. Don’t miss the chance
to enjoy exciting offers and
quickly download high-quality
materials in just a few simple
steps!
Preface

Years ago, Prof. Emeritus Chi-Tsong Chen, the author of Linear System
Theory and Design, a very successful textbook (Oxford University Press), met
the author at his Flushing, New York residence. In the meeting, and in the
preface of Signals and Systems – A Fresh Look his last publication (PDF form
free to all globally), Prof. Chen lamented that “Feedbacks from graduates
that what they learned in university is not used in industry prompted me to
ponder what to teach in signals and systems.”
Sadly,and based on long professional career serving RCA/GE/Lockheed
Martin space sector, the author can definitively confirm the fact Prof.
Chen was sad about. The less-than-desirable state had existed, and is still
present,in the form that many degree-holding engineers including electrical,
electronic, mechanical, and other specialties are falling short in applying
mathematical tools they were taught in college. Given electrical schematic
drawings, they were unable to formulate and express systems’ dynamics in
state variables and state transition using the first-order differential equations
and linear algebra technique. As a result, they were unable to boost their
productivity using software such as MATLAB.
This book intends to bridge the gap—what is taught in college and how
it is being applied in industry. In essence, this writing shall be considered
didactic.
It begins with Chapter one giving capacitors and inductors, two indis-
pensable energy storage components, an in-depth examination from the
view point of the first-order derivative, its corresponding integral form,
and its physical implications. Chapter two covers RC- and RL-type net-
works governed by a single differential equation. Key steps moving system
differential equations to Laplace transform in a frequency domain and to
a state-space transition form are introduced. Along the way, unconven-
tional approaches deriving Fourier series, explaining orthogonal property,
or treating boundary value problems are also explored. Chapter three covers
current sourcing circuits including current mirror, the workhorse of analog
integrated circuits, and precision current generator loops critical to instru-
mentation. Chapter four extends Chapter two to networks of second order
governed by two first-order differential equations. Procedures transforming
multiple differential equations to Laplace form, to state-transition form, and
to state-transition solution are shown. Chapter five examines circuit blocks

xiii
xiv Preface

and modules performing amplification, voltage-to-time window, duty cycle

modulation, etc. Chapter six covers feedback practices including voltage,
current, isolation, summative current, subtractive current, and state feedback.
Chapter seven discusses configurations of control loops including single loop,
multiple loop, open loop, closed loop, nested loop, loop contention, etc.
Chapter eight deals with linear regulators including series voltage regulator
and current shunt in parallel. Chapter nine explores switch-mode power
processing. Chapter ten presents complexities arising from inductive load fed
by rectified AC sources of single phase, multiple phases, and phase control.
Employing the concept of electromagnetic vectors in space, Chapter eleven
focuses on the formation of magnetic flux vector placed intentionally along
selected orientation, time-varying flux intensity, and rotational flux vector
that makes motor spin.
Considering the writer’s goal is to bridge materials taught in college and
applications of the material in actual industrial settings, the topics outlined
above and organized in that particular order are suitable for college seniors
and novice professionals in the industry. Following the material, and when
facing a real-world design schematic, readers will be able to (1) assign state
variables (circuit node voltage, inductor branch current), (2) write down
multiple differential equations, (3) place equation set in a state-transition
form, (4) select the approach one is more comfortable and confident, for the
time being, (5) obtain system response solutions corresponding to various
drives in different time frames, (6) stitch together a steady-state response
solution in closed-form analytical expressions.
Given time and practice, and when facing system order exceeding three,
most readers will quickly realize that state-transition equation and solution
invoking matrix operation delineated in linear algebra are more effective,
even elegant, in handling high-order systems.
This writer had definitely experienced that awareness, and expects all
reader to do the same.
As indicated in the subtitle of this writing, along the presentation,
mathematical notes are inserted where appropriateness is not violated. Quite
a few may be considered unconventional. This is done in the spirit of never
taking authority dogmatically—a true open mind respecting the unlimited
possibilities of viewing nature from multiple angles and a belief that what
 Preface xv

was said true in the past may not be true in the future when new discoveries
see the daylight.
On the backdrop of the above conviction, this author took additional
efforts to make this writing also available in Chinese language;thanks to pub-
lisher Elsevier for granting such translation right. Thanks are also extended
to Mr. , at ITRI (Industrial Technology Research Institute, Hsinchu Taiwan),
who had performed the translation, a very demanding task considering the
limitations of Chinese language in handling technical subjects.
With the advance of miniaturized electronic hardware and supercom-
puter equipped with mathematical co-processors, engineering design tasks
are now mostly carried out by the simulation and computation. The
implementation of both always requires design formulation in the form
of analytical expressions based on, in most cases, systems of differential
equations with coefficients depending on components/parts values.
In the course of almost four decades‘Ł‘™professional career in aerospace
industries, the author had definitely derived significant benefits from follow-
ing the path outlined above.
You, readers, can certainly do the same.
Keng C. Wu
Princeton, NJ.
Dec. 2020
CHAPTER 1

Capacitor and inductor

Two components, capacitors and inductors, play irreplaceable roles in elec-
trical power processing for their energy-storage properties.By presenting the
analytical equations governing both in either differential or integral forms,
this chapter illuminates the electromagnetic behaviors of those devices and
elucidates its physical significance when working with driving sources.

MATH. NOTE: In most calculus textbooks, derivatives and integrals are

introduced in the forms of f´(x) = dy/dx = df(x)/dx and ∫ydx = ∫f(x)dx,
given y = f(x) a two-dimensional plane curve and x is the independent vari-
able, with little physical meaning attached except the concept of “tangential
slope,” associated with the derivative, and “geometrical area,” associated with
the integral, employing the approach of limit. The independent variable x is
by no means restricted to signifying only space quantity.It certainly can stand
for time,and many other variables as well.The simple act of replacing dx with
dt, an infinitesimal time increment, introduces interesting, and important,
physical meaning to derivative f´(t) = dy/dt = df(t)/dt. As dt appears in the
numerator (inverse of time), derivative against time yields the dimension of
speed, velocity, and/or frequency; the temporal changes of a time-dependent
variable. ♣

1.1 Capacitor equation in differential form

Almost without exception, the action of capacitors is introduced in text-
books in a differential form; which links current through the device and
time rate of voltage change across it with a positive sign as shown.

MATH. NOTE: At a more fundamental level, the current is expressed as

the rate of charge carriers’ changes, i(t) = dQ(t)/dt, in which Q(t) = Cv(t) and
C, the capacitance and a constant within reason, is a function of geometry
and material property. ♣

i
dv(t ) +v
i(t ) C (1.1)
dt
Power electronic system design. Copyright © 2021 Elsevier Inc.
DOI: 10.1016/B978-0-32-388542-3.00004-2 All rights reserved. 1
2 Power electronic system design

v
i
t

Fig. 1.1 Terminal current and voltage of a capacitor.

What does this form tell us about a capacitive element?

1. When dv/dt = 0, that is when the device voltage reaches an extreme,
a maximum or a minimum, the corresponding device current crosses zero
value. Stated differently, the device’s time-domain current waveform makes
a zero-crossing at the time its corresponding voltage waveform peaks, or
bottoms out. In other words, and in a graphical form, terminal voltage and
through current for a capacitor must hold a relation as shown in Fig. 1.1.
2. As the voltage variable in Eq. (1.1) appears as a derivative, the current
variable does not change its value if Eq. (1.1) is rewritten as
d[VDC + v(t )]
i(t ) = C (1.2)
dt
In this form, one important property of capacitor stands out. That is, the
device sustains a DC (direct current) voltage, VDC , which however does not
contribute to its current. The significance of this attribute is that a capacitor
blocks DC current. Or, DC current does not flow through a capacitor. Only
AC (alternating) current does.
3. Capacitor allows the application of a DC voltage within limit; the
breakdown voltage.

1.2 Capacitor equation in integral form

Eq. (1.1) can of course be rewritten as
 
1 1 t
v(t ) = i(t )dt = V0 (t0 ) + i(τ )dτ (1.3)
C C t0
MATH. NOTE: This is actually a rewording of the previous MATH
NOTE, that is, charge is equal to the time integral of current. ♣
In contrast to the derivative form, the integral form (Eq. 1.3), in particular
the right-hand side, conveys an extremely important effect of the capacitive
state variable: voltage.
4. In a very straightforward manner, it declares the continuous nature of
capacitor voltage.
 Capacitor and inductor 3

t
v

Fig. 1.2 Terminal current and voltage of an inductor.

5. In addition, Eq. (1.3) implies that capacitor voltage is lagging its

current in time domain:voltage phase lags (The integral of a time-dependent
variable yields accumulation of its past acts.) Or, capacitor current is leading
its voltage across: current phase lead.
6. Furthermore, in a subtle way, Eq. (1.3) indicates that a capacitor stores
energy in stationary charge space: E-field.

1.3 Inductor equation in differential form

MATH. NOTE: At a more fundamental level, the inductive voltage will
be expressed as the rate of magnetic flux changes, v(t) = d(t)/dt, in which
(t) = Li(t) and L, the inductance and a constant within reason, is a function
of geometry and material property. This will be expounded in one of the
section to follow. ♣

By the same token, inductor and its electrical symbol was always intro-
duced by the following
i
di(t ) +v
v(t ) L (1.4)
dt

With reasoning similar to those given for capacitors, the following

electric behaviors are expected of an inductor.
1. When di/dt = 0, that is when the device current reaches an extreme,
a maximum or a minimum, the corresponding device voltage crosses zero
value. Stated differently, the device’s time-domain voltage waveform makes
a zero-crossing at the time its corresponding current waveform peaks, or
bottoms out. In other words, and in a graphical form, terminal voltage and
through current for an inductor must hold a relation as shown in Fig. 1.2.
2. As the current variable in Eq. (1.4) appears as a derivative, the voltage
variable does not change its value if Eq. (1.4) is rewritten as
d[IDC + i(t )]
v(t ) = L (1.5)
dt
4 Power electronic system design

In this form, one important property of inductor stands out. That is, the
device allows a DC current, IDC , which however does not contribute to its
voltage.
3. The allowed DC current is however constrained within a limit; the
magnetic core saturation and winding wire Ampere rating.

1.4 Inductor equation in integral form

Eq. (1.4) can also be rewritten as
 
1 1 t
i(t ) = v(t )dt = I0 (t0 ) + v(τ )dτ (1.6)
L L t0
Again in contrast to the derivative form, the integral form (Eq. 1.6), in
particular the right-hand side, conveys an extremely important effect of the
inductive state variable: current.
4. It states the continuous nature of inductive current.
5. In addition, (Eq. 1.6) implies that inductive current is lagging its voltage
in time domain: current phase lag. Or, inductive voltage is leading its through
current: voltage phase lead.
6. Furthermore, in a subtle way, Eq. (1.3) indicates that an inductor stores
energy in circulating charge: H-field.

1.5 Definition of inductance and Faraday’s law

Faraday’s law is also generally presented in a differential form and named
“Faraday’s Law of Induction”: An electromagnetic force, emf, is generated by
a time-varying magnetic flux, e = d/dt. This form may best be interpreted
as magnetic induction on a single turn (loop). For a multiple-loop device,
the case of a practical inductor with N turns, the total induced emf sums up
to v = Ne = d(N)/dt.
Placed in an integral form, Faraday’s law of induction with multiple
loops in series becomes λ = N = ∫vdt: Total flux linkage equals the time
integral of a driving (voltage) source. For a practical device consisting of
a magnetic core, which serves to guide and focus flux, and an N-loop
winding, which passes current, each loop experiences an identical flux, .
With known cross-sectional core area, Ac , supplied by core manufacturer,
the total flux linkage is expressed as λ = NAc B, in which B stands for flux
density. It is then understood in physics that, given a core material with
permeability μ = μo μr (μo vacuum permeability, μr relative permeability), flux
 Capacitor and inductor 5

density is proportional to the magnetic field intensity, H, generated by a

driving source in the form of current, I, while permeability serves as the
proportional constant. Therefore, the total flux linkage is further expressed
as λ = NAc μo μr H.
Here, another electromagnetic law kicks in. That is Ampere’s Law:
magneto-motive force Ni = ∫Hdlm , the contour integral of magnetic field
intensity along a closed path lm . The integral may be simplified to HLm for
a geometrically isotropic and materially homogeneous core.
This additional attribute allows us to step further and rewrite the total
flux linkage as λ = NAc μo μr Ni/Lm , or λ = (N2 μo μr Ac /Lm )i. Plugging into
Faraday’s law of induction, we have v = (N2 μo μr Ac /Lm )di/dt. Referring to
Eq. (1.4), inductance is therefore L = N2 μo μr Ac /Lm for an inductor with core
material property μo , μr and core geometry Ac ,Lm ; core area and magnetic
path length.
As given, parameters determining inductance value are grouped in two;
(1) inductance index equal to the product of material property and core area
to magnetic path length ratio, and (2) square of winding turn number. In
industrial sector, core inductance index is often assigned a symbol AL .

1.6 Magnetic coupling and mutual inductance

In previous section citing Faraday’s law of induction, “time-varying mag-
netic flux” is mentioned as the key to magnetic induction. However,
“variable magnetic flux” may be more fitting as there are basically three
mechanisms to create variable magnetic flux.
A flux source may be moved mechanically in space while the inductive
reception loop stays stationary. Or, the flux generator stays stationary while
the reception loop is moved mechanically. Either way, a relative spatial
motion exists between two entities and that is actually the bottom line
of magnetic induction. Hydropower, coal-fired power, and other turbine
generators all fall under this category. In conventional physics text, it is
named “generator action.”
But the third way involves no mechanical motion at all. Both the flux
source and the reception loop remain stationary in space while the flux
generation side is driven by a true time-varying electric source. By so doing,
a time-varying flux is also created. In conventional physics, it is named
“transformer action.”
Employing the “Transformer action,” multiple windings, each with
multiple turns, are built wrapping around a common ferrous core. One
6 Power electronic system design

i1 i2
v1 v2
N1 N2

Fig. 1.3 Ferrite cup-cored transformer and its electrical symbol.

B-H curve
0.4 0.4

0.2
B( H)
Tesla

0 0

−0.2

−0.4
−0.4
−400 −200 0 200 400
−400 H 400
Oersted

Fig. 1.4 A B-H curve example without hysteresis.

winding, the primary, serves as a flux source driven by an external voltage

while the rest, the secondary, act as output windings. Fig. 1.3 shows a ferrite
cup-cored transformer, a two-winding and a multi-winding transformer
symbols.
Based on the two-winding transformer, Ampere’s law establishes a mag-
netic field intensity (N1 i1 + N2 i2 )/Lm in the core. When only winding 1 is
driven while winding 2 remains open, i2 = 0, Eq. (1.7) holds

N1 i1 Aw1
λ1 = N1 1 = N1 Aw1 B = N1 Aw1 μoμr = N12 μoμr i1 = L1 i1
Lm Lm
(1.7)
where Aw1 stands for winding 1 cross-sectional area.
Under the same driving condition, Eq. (1.8), a cross-coupled flux linkage,
holds for winding 2, assuming Aw2 < Aw1 .

N1 i1 Aw2
λ2 = N2 2 = N2 Aw2 B = N2 Aw2 μoμr = N2 N1 μoμr i1
Lm Lm (1.8)
λ2 = M21 i1
 Capacitor and inductor 7

Next, we move driving source to winding 2 and open winding 1, i1 = 0.

Eq. (1.9), again a cross-coupled flux linkage, holds for winding 1. Readers
are cautioned to note the assumption that Aw2 < Aw1 .
N2 i2 Aw2
λ1 = N1 1 = N1 Aw2 B = N1 Aw2 μoμr = N1 N2 μoμr i2
Lm Lm (1.9)
λ1 = M12 i2
And, Eq. (1.10) holds for winding 2
N2 i2 Aw2
λ2 = N2 Aw2 B = N2 Aw2 μoμr = N22 μoμr i2 = L2 i2 (1.10)
Lm Lm
Clearly, a mutual inductance exists between the two windings. That is
Aw2
M = M12 = M21 = N1 N2 μoμr (1.11)
Lm
An interesting effect is also observed, that is the product of L1 and L2 .
Aw1 2 Aw2
L1 L2 = N12 μoμr N2 μoμr
Lm Lm
(1.12)
Aw1 Aw2 Aw1 2
= N1 N2 μoμr N1 N2 μoμr = M
Lm Lm Aw2
or,

Aw2 √
0<k= < 1, M = k L1 L2 (1.13)
Aw1
Here an imperfect attempt indicates the origin of coupling coefficient,
k. Imperfect because perfectionist theoretician may mount a challenge
against the use of winding area in place of core area. However, considering
the imprecise nature of magnetism, practitioners in the industrial sectors
are willing to accept Eq. (1.13) knowing there are other flux leakages
unaccounted for.

1.7 Transformer equation

In previous sections, the linkage, represented by symbol μo , μr , between flux
density and magnetic field intensity was freely invoked without qualification.
Unfortunately, nature always places limitation on everything. Permeability
of material is not exempted either. It is well understood, and documented,
that ferromagnetic materials exhibit highly nonlinear and hysteric properties
 Visit https://ebookluna.com
now to explore a diverse
collection of ebooks available
in formats like PDF, EPUB, and
MOBI, compatible with all
devices. Don’t miss the chance
to enjoy exciting offers and
quickly download high-quality
materials in just a few simple
steps!
8 Power electronic system design

when magnetized. The nonlinearity may be expressed in an analytic form

by borrowing the Fermi–Dirac (electron density) distribution, (Eq. 1.14), and
modifying it.
In solid-state physics, the Fermi–Dirac function is given as
1
p(E ) = E−E f
(1.14)
1+ e kT
−23
in which k = 1.380658 × 10 [Joule/degree], T temperature in degree
Kelvin, Ef Fermi energy.
With slight modifications based on Eq. (1.14), Eq. (1.15) describes the
nonlinear magnetic property for ferrous material.
⎛ ⎞
⎜ ⎟
⎜ 2 ⎟
B(H ) = Bsat ⎜ − 1⎟ (1.15)
⎝ −
H 
Bsat
⎠
2μoμr
1+e
in which Bsat is the saturation flux density, specific to a material, in Tesla
(MKS unit]; H the driving magnetic field intensity in Oersted (MKS unit].
For example, given μo = 4π × 10−7 [Henry/m, MKS uint], μr = 2500 [unit-
less], Bsat = 0.25 [Tesla], Eq. (1.15) yields Fig. 1.4 plot.
With further modifications, hysteretic behavior is also accounted for,
(Eq. 1.16), depending on either increasing or decreasing magnetic field.
⎛ ⎞
⎜ ⎟
⎜ 2 ⎟
B(H, Ho ) = Bsat ⎜ H −Ho 
− 1⎟ (1.16)
⎝ −
Bsat
⎠
2μoμr
1+e
In the case of increasing, the B-H curve traverses the right ascending leg,
while the left descending leg covers decreasing field (Fig. 1.5).
The main point of focus is that the magnetic flux density for any ferrous
material is limited to an upper bound, Bsat , beyond which the material can
no longer support additional flux change. Once a drive condition, that is,
the volt-second integral, exceeds the limit, the device/material permeability
drops to zero. As a result, the device incorporating the material no longer
offers inductance.Eq.(1.4) then states that,with near-zero inductance,a huge
di will also results. Such a surge current, limited only by winding resistance,
will lead to device destruction. And, this is exactly the reason the following,
flux density of transformer core, is discussed.
 Capacitor and inductor 9

B-H curve
0.4 0.4

0.2
B(H, 50)
Tesla 0
0
B(H, −50)

−0.2

−0.4
−0.4
−400 −200 0 200 400
−400 H 400
Oersted

Fig. 1.5 A B-H curve example with hysteresis.

MATH. NOTE: What Fig. 1.5 shows is a multivalue function. At a selected

H coordinate, there are two corresponding flux density values. At a selected
B coordinate, there are two corresponding field intensity values. ♣
Refer now to Fig. 1.3 transformer. Suppose an AC source v(t) = Va cos(ωt)
is driving the primary winding N1 . The integral form of Faraday’s law of
induction gives a total flux linkage
 √
Va 2 Vrms
λ = Va cos (ωt )dt = sin (ωt ) = sin (ωt ) (1.17)
ω ω
As discussed above and considering the allowable saturation flux density,
the flux linkage magnitude of Eq. (1.17) must satisfy Eq. (1.18)
√ √
2 Vrms 2 Vrms
|λ| = = ≤ N1 Ac Bsat (1.18)
ω 2π f
where f stands for a driving source frequency. In other words, the driving
source magnitude, in RMS, must be constrained to
2π
Vrms ≤ √ f N1 Ac Bsat ≈ 4.443 f N1 Ac Bsat (1.19)
2
Otherwise, the transformer core will be saturated if the drive exceeds the
limit.
In power processing, in addition to sinusoidal drive, symmetrical 50%
square wave drive, dotted trace in Fig. 1.6, is also a very popular source.
Readers are invited to prove that, in this case, the source magnitude is
10 Power electronic system design

Va
flux

−Va

Fig. 1.6 Square-wave drive and core flux function.

Va Va
0

−Va −Va

Fig. 1.7 Rectangular wave drives.

constrained by Eq. (1.20) and the core flux (per turn) traces the solid curve
given in Fig. 1.6.
Vrms ≤ 4 f N1 Ac Bsat (1.20)
At this point, a moment of caution is warranted. Both Eqs. (1.19) and
(1.20) invoke drive source in RMS quantity. This practice entails easily mis-
understood mistake. Specifically, Eq. (1.19) is applicable solely for sinusoidal
drive while Eq. (1.20) for symmetrical, 50%, square-wave drive. For drive
waves, such as those given in Fig. 1.7, conversion factors existing between
drive amplitude,Va ,and its equivalent RMS quantity will alter the numerical
factor, 4.443 or 4, that appears in Eqs. (1.19) and (1.20). Readers are again
invited to work out the flux profile for drive waves shown in Fig. 1.7.

1.8 Nonideal capacitor, nonideal inductor, and

equivalent circuit
So far, all discussions concerning capacitors and magnetic devices are carried
out in the context of idealistic, conceptual space on paper. To be of practical
use, both must be built and brought to face real world. Capacitors are made
of bilayer electrodes, which must extend externally via metal contacts, while
inductors of spiral metal wires. Therefore, serial resistances are inevitably
introduced. This is not all; even a short piece of wire has inductance, however
small. What else? Capacitance exists between adjacent wires. Surfaces of
device package also create sneaky, leaky path. Taking in all, a real capacitor
or inductor may be represented by a lumped model (Fig. 1.8).
As shown, a view begins to emerge that neither device would work
as purely as what it is theoretically. Depending on the frequency range it is
immersed in a capacitor/inductor actually acts like a complex RLC network
(Fig. 1.8). We will come back to this point later.
 Capacitor and inductor 11

(a) (b)
Fig. 1.8 Lumped model of real (a) capacitors and (b) inductors.

1.9 Transformer equivalent circuits

In the same spirit, two-wind transformer (Fig. 1.3) can be transformed into
equivalent circuits. However, before proceeding, the drawing will be slightly
modified (Fig. 1.9) to include effects of mutual inductance. Readers should
also refer to physics textbooks for the dot convention invoked in the figure.
For the time being and in the following, winding resistances are omitted
to simplify the derivation of equivalence.
One sees easily that N1 loop gives

di1 di2 di1 di1 di1 di2

v1 = L1 +M = L1 −M +M +M
dt dt dt dt dt dt
(1.21)
di1 d(i1 + i2 )
= (L1 − M ) +M
dt dt
And, N2 loop gives

di2 di1 di1 di2 di2 di2

v2 = L2 +M =M +M + L2 −M
dt dt dt dt dt dt
(1.22)
d(i1 + i2 ) di2
=M + (L2 − M )
dt dt
Both lead to the loop T-form equivalent circuit (Fig. 1.10).

i1 M i2
v1 v2
N1 N2

Fig. 1.9 Two-winding transformer with mutual inductance.

12 Power electronic system design

1:1
i1 L 1 - M L2 - M i2
v1 M v2
Ideal

Fig. 1.10 Loop T-form model for two-winding transformer.

1:1
i1
(L1L2 - M 2 )/ i2
v1 M
v2
(L1L2 - M 2 )/ (L1L2 - M 2 )/ Ideal
(L2-M ) (L1-M )

Fig. 1.11 Nodal π -form model for two-winding transformer.

Taking Laplace transform for the left most parts of Eqs. (1.21) and (1.22),
we reach a new equation set
v1 = sL1 I1 (s) + sMI2 (s) v2 = sMI1 (s) + sL2 I2 (s) (1.23)
Eq. (1.23) enables us to express I1 and I2 in terms of v1 and v2 .
v1 sM
v2 sL2 sL2 v1 − sMv2 (L2 − M )v1 + M (v1 − v2 )
I1 = = 2 = 
sL1 sM s L 1 L2 − M 2 s L 1 L2 − M 2
sM sL2 (1.24)
v1 (v1 − v2 )
=  +  
−M
s L1 L2M−M
2 2
s L1LL22−M

sL1 v1
sM v2 sL1 v2 − sMv1 (L1 − M )v2 + M (v2 − v1 )
I1 = = 2 = 
sL1 sM s L 1 L2 − M 2 s L 1 L2 − M 2
sM sL2 (1.25)
v2 (v2 − v1 )
=  +  
L1 L2 −M 2
s L1 L2M−M
2
s L1 −M

Eqs. (1.24) and (1.25) indicate a nodal π -form equivalent circuit

(Fig. 1.11).
Both the loop form and the nodal form transformer equivalent circuits
do not find them useful as it calls out all components and parameters.
 Capacitor and inductor 13

1:1
i1 (1-k 2)L 1 i2
v1 k 2L 1 v2
Ideal

Fig. 1.12 Hybrid τ -form model for two-winding transformer.

A keen observation changes the situation. By making L2 – M = 0 in

Fig. 1.10 and recognizing M = kæ(L1 L2 ), M = L2 becomes k2 L1 . As a result,
Fig. 1.10 becomes Fig. 1.12; the hybrid τ -form equivalent circuit.
This form has a loop and a node equation with two parameters, coupling
coefficient and primary inductance, involved.

1.10 Physical size of capacitor and inductor

Based on the fundamental principle dealing with electric charge storage
capacity of a pair of metallic plates or metalized films, the capacitance C of
such a mechanical construction is expressed as
A
C=ε (1.26)
d
in which ε represents the permittivity (dielectric constant) of separator
medium embedded between two plates, while A stands for plate area and d
plate separation, both spatial dimensions. Therefore, to a very good estimate,
the volumetric size for a capacitor will be at least equal to, or larger than,
dA; which is a simple product of two geometrical entities.
However, the process for sizing magnetic devices does not enjoy the
simplicity a capacitor is endowed with, as the construct of magnetic devices
always require some ferromagnetic filling a core shaped purposely to guide
magnetic flux such that the latter forms closed loops. In addition, space
must also be provided to accommodate multiple windings made of turns
of insulated metal wires. Fig. 1.13 shows one half of a ferrite core. The other
half is a mirror image.
As shown, the core has a center post giving a cross-sectional area Ac and
a winding window area Aw , within which coil(s) is(are) wound.
For a core wound with multiple windings, each winding in general
sustains different periodic voltage waveform and carries different currents,
also periodic. Eq. (1.18) can be generalized to specify how each winding will
be designed individually considering core flux limitation.

(Vrms ) j = K f f N j Ac Bmax (1.27)

14 Power electronic system design

Ac
Aw

Fig. 1.13 Half of a ferrite core; Ac = center post core area; Aw . (dotted line) = winding
window area, small filled circles = coil wires.

Here, Bmax is often chosen to be less than Bsat . Kf is a scaling factor relating
the RMS value of a periodic voltage and its time-domain magnitude.
Next, for each winding with peak current ij , a winding area (Aw )j is
conceptually assigned. Due to wire shape and unavoidable stacking in actual
build, only k(Aw )j is utilized. Given a desired inverse current density J [unit,
length2 /Amp],the winding peak current and its corresponding winding area
is associated by
k · (Aw ) j
ij = (1.28)
JN j
Eqs. (1.27) and (1.28) allow, for jth winding, its area product
(Vrms ) j JN j i j J (Vrms ) j i j
Ac (Aw ) j = · = (1.29)
K f f N j Bmax k kK f f Bmax
The total core cross-sectional area and winding area product covering all
winds is then
 J 
Ac A w = A c (Aw ) j = (Vrms ) j i j (1.30)
j
kK f f Bmax j

The summation on the right-hand side hints the total power handling
capacity of the device. The mathematical formulation may not be exactly
right, but it does give the flavor.
Therefore, referring to Fig. 1.13 and ignoring unit discrepancy, the
volumetric size of a magnetic device, in numerical term, may be considered
 Capacitor and inductor 15

almost twice of Eq. (1.30). Designers must check with core manufacturers
as to the accounting of the area2 number to avoid over, or under, count;
therefore over, or under, sizing a device.

1.11 Specifications for capacitor and inductor

In the following, only key parameters considered essential are tabulated
for both parts. Readers are to be reminded that a good understanding
of part specifications and ratings constitutes a major role in delivering a
good design, selecting reliable components, and making a high-quality final
product.

1.11.1 Capacitor

Parameter Digits, unit Example Comments

Value Numeral, pF, nF, or 100 pF, 100 μF
μF
Value, tolerance Numeral, % 5%, 10%
Value, thermal Numeral, %/°C, 0.02%/°C
ppm/CÚ
Voltage, breakdown Numeral, Volts 25 V, 100 V DCmax + AC peak
magnitude
Voltage, surge rate Numeral, 1 KV/ms Causes transient
(dv/dt) Volts/sec current and
localized
overheating
Current, steady Numeral, A or mA 0.5 A at 360 Hz Frequency
sinusoidal dependent
Current, single pulse Numeral, A μs; A 500 A 3 μs Causes localized
ns overheating
ESR (Equivalent Numeral, mOhm 7 mOhm Not a constant
Serial Resistance)
Reactance Numeral, Ohm = Xc (ω), A curve Frequency
1/(ωC) dependent
Dissipation factor Numeral, A curve ESR/Xc (ω),
dimensionless
Leakage current Numeral, pA, nA Insulation quality
Lead Polarity Not applicable Polarized or Electrolytic or Film
nonpolarized
Packaging Not applicable Radial or axial Mounting
Disc or tubular consideration
16 Power electronic system design

A c B–L m H curve
0.4

Weber(volt.second)=AcB
Ψ(H, 50∙500)

Ψ(H, −50∙500)

−0.4

−2×105 H 2×105
AmpereTurn=LmH

Fig. 1.14 A core specific Ac B–Lm H curve.

1.11.2 Inductor
The intricacy for specifying magnetic devices is, in this writer’s view, orders
of magnitude harder than that for capacitors. Here, we will begin with an
important reexamination of the B-H curves given in Fig. 1.4 and Fig. 1.5.
As far as this writer has been able to reach, the most existing literature
including journalistic articles and textbooks dealing with the subject cover
it solely in terms of B(flux density)-H(field strength) parameters. What was
not clearly mentioned in those presentations is the underlying significance
of presenting in B-H form. In a single statement, we proclaim that the
curve in B-H parameter form is “material specific.” It is characterizing a
specific ferrous material in terms of per unit volume. In other words, it is
independent of core geometry.
Therefore, those material property curves can be easily modified to be
“core specific,” Fig. 1.14, in which core geometry is now included; Ac core
cross-sectional area and Lm magnetic path length. The plot x-axis coordinate
is HLm (Ampere) while the y-axis is Ac B (flux, per turn, in Weber) (Readers
should ignore numerical figures in Fig. 1.14. It just shows that the inclusion
of core geometry will alter plot coordinate scales.)
By the same token, it can be further modified to be “device specific” or
“winding specific,” “terminal specific” in which the winding turn number,
N, is included. With that, the plot x-axis coordinate is Ni (Ampere turn)
while the y-axis is NAc B (total flux linkage in volt-second = ∫vdt).
In summary, three aspects are involved in specifying a magnetic device:
core material B-H, core geometry Ac B–HLm , and winding terminal NAc B-
Ni.
 Capacitor and inductor 17

1.11.3 Inductor core material

Parameter Digits, unit Example Comments
μi , zero-biased Numeral, 126, NIDC = 0 Referenced to μo
Permeability dimensionless = 4π × 10−7
μr , IDC -biased Numeral, 95, NIDC = 40, Referenced to μo
Permeability dimensionless Amp-Turn = 4π × 10−7
μo , free space 4π × 10−7 μ = μr • μo
Permeability

1.11.4 Inductor core Geometry

Parameter Digits, unit Example Comments
2
Ac , cross-sectional Numeral, SI (MKS) 0.67 cm
area meter2 or cgs cm2
lm , magnetic path Numeral, SI (MKS) 8.15 cm
length meter or cgs cm
Aw , winding area Numeral, SI (MKS) 2.13 cm2
meter2 or cgs cm2

1.11.5 Inductor winding

Parameter Digits, unit Example Comments
Fill factor 0< kf < 1, N/A 0.77, AWG #20 Wire gauge
dependent
lw , mean turn Numeral, m or 1.85 cm, Wire gauge
length cm AWG#24 dependent
N, turn number Numeral 13, AWG#22 Wire gauge
dependent
Many magnetic device manufacturers also often list the area product,
discussed in Section 1.10, to give designer a sense of how large the device
may be in handling a given power. This particular listing may be considered
to belong to the category of core geometry; but not exactly.
There are certainly other parameters, such as core losses unit volume
(a frequency-dependent parameter), thermal conductivity, etc. that are not
listed above; as this section presentation is not meant to be exhaustive.
Actually, this is the reason that efforts specifying magnetic devices properly
are very demanding.
 Visit https://ebookluna.com
now to explore a diverse
collection of ebooks available
in formats like PDF, EPUB, and
MOBI, compatible with all
devices. Don’t miss the chance
to enjoy exciting offers and
quickly download high-quality
materials in just a few simple
steps!
Other documents randomly have
different content
 Note 28, p. 70.
Lib. i. Od. 35.
Te semper anteit sæva Necessitas:
Clavos trabales et cuneos manu
Gestans ahenea; nec severus
Uncus abest liquidumque plumbum.

In this picture of Necessity drawn by Horace, perhaps the richest in

attributes of any to be found in the old poets, the nails, the clamps,
and the liquid lead, whether regarded as means of confinement or
implements of punishment, still belong to the class of poetical, rather
than allegorical, attributes. But, even so, they are too crowded; and
the passage is one of the least effective in Horace. Sanadon says:
“J’ose dire que ce tableau, pris dans le détail, serait plus beau sur la
toile que dans une ode héroïque. Je ne puis souffrir cet attirail
patibulaire de clous, de coins, de crocs, et de plomb fondu. J’ai cru en
devoir décharger la traduction, en substituant les idées générales aux
idées singulières. C’est dommage que le poëte ait eu besoin de ce
correctif.” Sanadon’s sentiment was fine and true, but he does not
give the right ground for it. The objection is not that these attributes
are the paraphernalia of the gallows, for he had but to interpret them
in their other sense to make them the firmest supports of
architecture. Their fault is in being addressed to the eye and not to
the ear. For all impressions meant for the eye, but presented to us
through the ear, are received with effort, and produce no great
degree of vividness. These lines of Horace remind me of a couple of
oversights on the part of Spence, which give us no very good idea of
the exactitude with which he has studied the passages he cites from
the old poets. He is speaking of the image under which the Romans
represented faith or honesty. (Dial. x.) “The Romans,” he says,
“called her ‘Fides;’ and, when they called her ‘Sola Fides,’ seem to
mean the same as we do by the words ‘downright honesty.’ She is
represented with an erect, open air, and with nothing but a thin robe
on, so fine that one might see through it. Horace therefore calls her
‘thin-dressed’ in one of his odes, and ‘transparent’ in another.” In
these few lines are not less than three gross errors. First, it is false
that “sola” was a distinct epithet applied to the goddess Fides. In the
two passages from Livy, which he adduces as proof (lib. i. sect. 21,
lib. ii. sect. 3), the word has only its usual signification,—the
exclusion of all else. In one place, indeed, the “soli” has been
questioned by the critics, who think it must have crept into the text
through an error in writing, occasioned by the word next to it, which
is “solenne.” In the other passage cited, the author is not speaking of
fidelity at all, but of innocence, Innocentia. Secondly, Horace, in one
of his odes (the thirty-fifth of the first book, mentioned above), is
said to have applied to Fides the epithet thin-dressed:
Te spes, et albo rara fides colit
Velata panno.

“Rarus,” it is true, can also mean thin; but here it means only rare,
seldom appearing, and is applied to Fidelity herself, not to her
clothing. Spence would have been right, had the poet said, “Fides
raro velata panno.” Thirdly, Horace is said to have elsewhere called
faith or honesty transparent, in the sense in which friends protest to
one another, “I wish you could read my heart.” This meaning is said
to be found in the line of the eighteenth ode of the First Book:
Arcanique Fides prodiga, pellucidior vitro.

How can a critic allow himself to be thus misled by a word? Is a faith,

“arcani prodiga,” lavish of secrets, faithfulness? is it not rather
faithlessness? And it is of faithlessness, in fact, that Horace says,
“She is transparent as glass, because she betrays to every eye the
secrets entrusted to her.”
 Note 29, p. 71.
Apollo delivers the washed and embalmed body of Sarpedon to
Death and Sleep, that they may bring him to his native country.
(Iliad, xvi. 681, 682.)
πέμπε δέ μιν πομποῖσιν ἅμα κραιπνοῖσι φέρεσθαι,
Ὕπνῳ καὶ Θανάτῳ διδυμάοσιν.

Caylus recommends this idea to the painter, but adds: “It is a pity
that Homer has given us no account of the attributes under which
Sleep was represented in his day. We recognize the god only by his
act, and we crown him with poppies. These ideas are modern. The
first is of service, but cannot be employed in the present case, where
even the flowers would be out of keeping in connection with the
figure of Death.” (Tableaux tirés de l’Iliade, de l’Odyssée d’Homère,
et de l’Enéide de Virgile, avec des observations générales sur le
costume, à Paris, 1757–58.) That is requiring of Homer
ornamentations of that petty kind most at variance with the nobility
of his style. The most ingenious attributes he could have bestowed on
Sleep would not have characterized him so perfectly, nor have
brought so vivid a picture of him before us, as the single touch which
makes him the twin brother of Death. Let the artist seek to express
this, and he may dispense with all attributes. The old artists did, in
fact, make Sleep and Death resemble each other, like twin-brothers.
On a chest of cedar, in the Temple of Juno at Elis, they both lay as
boys in the arms of Night. One was white, the other black; one slept,
the other only seemed to sleep; the feet of both were crossed. For so I
should prefer to translate the words of Pausanias (Eliac. cap. xviii. p.
422, edit. Kuhn), ἀμφοτέρους διεστραμμένους τοὺς πόδας, rather
than by “crooked feet,” as Gedoyn does, “les pieds contrefaits.” What
would be the meaning of crooked feet? To lie with crossed feet is
customary with sleepers. Sleep is thus represented by Maffei.
(Raccol. Pl. 151.) Modern artists have entirely abandoned this
resemblance between Sleep and Death, which we find among the
ancients, and always represent Death as a skeleton, or at best a
skeleton covered with skin. Caylus should have been careful to tell
the artists whether they had better follow the custom of the ancients
or the moderns in this respect. He seems to declare in favor of the
modern view, since he regards Death as a figure that would not
harmonize well with a flower-crowned companion. Has he further
considered how inappropriate this modern idea would be in a
Homeric picture? How could its loathsome character have failed to
shock him? I cannot bring myself to believe that the little metal
figure in the ducal gallery at Florence, representing a skeleton sitting
on the ground, with one arm on an urn of ashes (Spence’s Polymetis,
tab. xli.), is a veritable antique. It cannot possibly represent Death,
because the ancients represented him very differently. Even their
poets never thought of him under this repulsive shape.
 Note 30, p. 76.
Richardson cites this work as an illustration of the rule that the
attention of the spectator should be diverted by nothing, however
admirable, from the chief figure. “Protogenes,” he says, “had
introduced into his famous picture of Ialysus a partridge, painted
with so much skill that it seemed alive, and was admired by all
Greece. But, because it attracted all eyes to itself, to the detriment of
the whole piece, he effaced it.” (Traité de la Peinture, T. i. p. 46.)
Richardson is mistaken; this partridge was not in the Ialysus, but in
another picture of Protogenes called the Idle Satyr, or Satyr in
Repose, Σάτυρος ἀναπαυόμενος. I should hardly have mentioned
this error, which arose from a misunderstanding of a passage in
Pliny, had not the same mistake been made by Meursius. (Rhodi. lib.
i. cap. 14.) “In eadem tabula, scilicet in qua Ialysus, Satyrus erat,
quem dicebant Anapauomenon, tibeas tenens.”
Something of the same kind occurs in Winkelmann. (Von der
Nachahm. der Gr. W. in der Mal. und Bildh. p. 56.) Strabo is the only
authority for this partridge story, and he expressly discriminates
between the Ialysus and the Satyr leaning against a pillar on which
sat the partridge. (Lib. xiv.) Meursius, Richardson, and Winkelmann
misunderstood the passage in Pliny (lib. xxxv. sect. 36), from not
perceiving that he was speaking of two different pictures: the one
which saved the city, because Demetrius would not assault the place
where it stood; and another, which Protogenes painted during the
siege. The one was Ialysus, the other the Satyr.
 Note 31, p. 79.
This invisible battle of the gods has been imitated by Quintus
Calaber in his Twelfth Book, with the evident design of improving on
his model. The grammarian seems to have held it unbecoming in a
god to be thrown to the ground by a stone. He therefore makes the
gods hurl at one another huge masses of rock, torn up from Mount
Ida, which, however, are shattered against the limbs of the immortals
and fly like sand about them.
... οἱ δὲ κολώνας
χερσὶν ἀποῤῥήξαντες ἀπ’ οὔρεος Ἰδαίοιο
βάλλον ἐπ’ ἀλλήλους· αἳ δὲ ψαμάθοισιν ὁμοῖαι
ῥεῖα διεσκίδναντο θεῶν περὶ δ’ ἄσχετα γυῖα
ῥηγνύμενα διὰ τυτθά....

A conceit which destroys the effect by marring our idea of the size of
the gods, and throwing contempt on their weapons. If gods throw
stones at one another, the stones must be able to hurt them, or they
are like silly boys pelting each other with earth. So old Homer
remains still the wiser, and all the fault-finding of cold criticism, and
the attempts of men of inferior genius to vie with him, serve but to
set forth his wisdom in clearer light. I do not deny that Quintus’s
imitation has excellent and original points; but they are less in
harmony with the modest greatness of Homer than calculated to do
honor to the stormy fire of a more modern poet. That the cry of the
gods, which rang to the heights of heaven and the depths of hell,
should not be heard by mortals, seems to me a most expressive
touch. The cry was too mighty to be grasped by the imperfect organs
of human hearing.
 Note 32, p. 80.
No one who has read Homer once through, ever so hastily, will
differ from this statement as far as regards strength and speed; but
he will not perhaps at once recall examples where the poet attaches
superhuman size to his gods. I would therefore refer him, in addition
to the description of Mars just quoted, whose body covered seven
hides, to the helmet of Minerva, κυνέην ἑκατὸν πολίων πρυλέεσσ’
ἀραρυῖαν (Iliad, v. 744), under which could be concealed as many
warriors as a hundred cities could bring into the field; to the stride of
Neptune (Iliad, xiii. 20); and especially to the lines from the
description of the shield, where Mars and Minerva lead the troops of
the beleaguered city. (Iliad, xviii. 516–519.)
ἦρχε δ’ ἄρά σφιν Ἄρης καὶ Παλλὰς Ἀθήνη,
ἄμφω χρυσείω, χρύσεια δὲ εἵματα ἕσθην,
καλὼ καὶ μεγάλω σὺν τεύχεσιν, ὥς τε θεώ περ,
ἀμφὶς ἀριζήλω· λαοὶ δ’ ὑπ’ ὑπολίζονες ἦσαν.

... While the youths

Marched on, with Mars and Pallas at their head,
Both wrought in gold, with golden garments on,
Stately and large in form, and over all
Conspicuous in bright armor, as became
The gods; the rest were of an humbler size.—Bryant.

Judging from the explanations they feel called upon to give of the
great helmet of Minerva, Homer’s commentators, old as well as new,
seem not always sufficiently to have borne in mind this wonderful
size of the gods. (See the notes on the above-quoted passage in the
edition of Clarke and Ernesti.) But we lose much in majesty by
thinking of the Homeric deities as of ordinary size, as we are
accustomed to see them on canvas in the company of mortals.
Although painting is unable to represent these superhuman
dimensions, sculpture to a certain extent may, and I am convinced
that the old masters borrowed from Homer their conception of the
gods in general as well as the colossal size which they not
infrequently gave them. (Herodot. lib. ii. p. 130, edit. Wessel.)
Further remarks upon the use of the colossal, its excellent effect in
sculpture and its want of effect in painting, I reserve for another
place.
 Note 33, p. 82.
Homer, I acknowledge, sometimes veils his deities in a cloud, but
only when they are not to be seen by other deities. In the fourteenth
book of the Iliad, for instance, where Juno and Sleep, ἠέρα
ἐσσαμένω, betake themselves to Mount Ida, the crafty goddess’s
chief care was not to be discovered by Venus, whose girdle she had
borrowed under pretence of a very different journey. In the same
book the love-drunken Jupiter is obliged to surround himself and his
spouse with a golden cloud to overcome her chaste reluctance.
πῶς κ’ ἔοι, εἴ τις νῶϊ θεῶν αἰειγενετάων
εὕδοντ’ ἀθρήσειε....

She did not fear to be seen by men, but by the gods. And although
Homer makes Jupiter say a few lines further on,—
Ἥρη, μήτε θεῶν τόγε δείδιθι μήτε τιν’ ἀνδρῶν
ὄψεσθαι· τοῖόν τοι ἐγὼ νέφος ἀμφικαλύψω,
χρύσεον.

“Fear thou not that any god or man will look upon us,” that does not
prove that the cloud was needed to conceal them from the eyes of
mortals, but that in this cloud they would be as invisible to the gods
as they always were to men. So, when Minerva puts on the helmet of
Pluto (Iliad, v. 485), which has the same effect of concealment that a
cloud would have, it is not that she may be concealed from the
Trojans, who either see her not at all or under the form of Sthenelus,
but simply that she may not be recognized by Mars.
 Note 34, p. 87.
Tableaux tirés de l’Iliade, Avert. p. 5. “On est toujours convenu,
que plus un poëme fournissait d’images et d’actions, plus il avait de
supériorité en poésie. Cette réflexion m’avait conduit à penser que le
calcul des différens tableaux, qu’ offrent les poëmes, pouvait servir à
comparer le mérite respectif des poëmes et des poëtes. Le nombre et
le genre des tableaux que présentent ces grands ouvrages, auraient
été une espèce de pierre de touche, ou, plutôt, une balance certaine
du mérite de ces poëmes et du génie de leurs auteurs.”
 Note 35, p. 88.
What we call poetic pictures, the ancients, as we learn from
Longinus, called “phantasiæ;” and what we call illusion in such
pictures, they named “enargia.” It was therefore said by some one, as
Plutarch tells us (Erot. T. ii. edit. Henr. Steph. p. 1351), that poetic
“phantasiæ” were, on account of their “enargia,” waking dreams: Αἱ
ποιητικαὶ φαντασίαι διὰ τὴν ἐνάργειαν ἐγρηγορότων ἐνύπνια
εἰσίν. I could wish that our modern books upon poetry had used this
nomenclature, and avoided the word picture altogether. We should
thus have been spared a multitude of doubtful rules, whose chief
foundation is the coincidence of an arbitrary term. No one would
then have thought of confining poetic conceptions within the limits
of a material picture. But the moment these conceptions were called
a poetic picture, the foundation for the error was laid.
 Note 36, p. 89.
Iliad, iv. 105.
αὐτίκ’ ἐσύλα τόξον ἐΰξοον
καὶ τὸ μὲν εὖ κατέθηκε τανυσσάμενος, ποτὶ γαίῃ
ἀγκλίνας·...
αὐτὰρ ὁ σύλα πῶμα φαρέτρης, ἐκ δ’ ἕλετ’ ἰὸν
ἀβλῆτα πτερόεντα, μελαινέων ἕρμ’ ὀδυνάων·
αἶψα δ’ ἐπὶ νευρῇ κατεκόσμει πικρὸν ὀϊστὸν,
ἕλκε δ’ ὁμοῦ γλυφίδας τε λαβὼν καὶ νεῦρα βόεια·
νευρὴν μὲν μαζῷ πέλασεν, τόξον δὲ σίδηρον.
αὐτὰρ ἐπειδὴ κυκλοτερὲς μέγα τόξον ἔτεινεν,
λίγξε βιὸς, νευρὴ δὲ μέγ’ ἴαχεν ἆλτο δ’ ὀϊστὸς
ὀξυβελὴς, καθ’ ὅμιλον ἐπιπτέσθαι μενεαίνων.

To bend that bow the warrior lowered it

And pressed an end against the earth....
Then the Lycian drew aside
The cover from his quiver, taking out
A well-fledged arrow that had never flown,—
A cause of future sorrows. On the string
He laid that fatal arrow....
Grasping the bowstring and the arrow’s notch
He drew them back and forced the string to meet
His breast, the arrow-head to meet the bow,
Till the bow formed a circle. Then it twanged;
The cord gave out a shrilly sound; the shaft
Leaped forth in eager haste to reach the host.—Bryant.
 Note 37, p. 108.
Prologue to the Satires, 340.
That not in Fancy’s maze he wandered long,
But stooped to Truth and moralized his song.

Ibid. 148.
... Who could take offence
While pure description held the place of sense?

Warburton’s remark on this last line may have the force of an

explanation by the poet himself. “He uses pure equivocally, to signify
either chaste or empty; and has given in this line what he esteemed
the true character of descriptive poetry, as it is called,—a
composition, in his opinion, as absurd as a feast made up of sauces.
The use of a picturesque imagination is to brighten and adorn good
sense: so that to employ it only in description, is like children’s
delighting in a prism for the sake of its gaudy colors, which, when
frugally managed and artfully disposed, might be made to represent
and illustrate the noblest objects in nature.”
Both poet and commentator seem to have regarded the matter
rather from a moral than an artistic point of view. But so much the
better that this style of poetry seems equally worthless from
whichever point it be viewed.
 Note 38, p. 108.
Poétique Française, T. ii. p. 501. “J’écrivais ces réflexions avant
que les essais des Allemands dans ce genre (l’Eglogue) fussent
connus parmi nous. Ils ont exécuté ce que j’avais conçu; et s’ils
parviennent à donner plus au moral et moins au détail des peintures
physiques, ils excelleront dans ce genre, plus riche, plus vaste, plus
fécond, et infiniment plus naturel et plus moral que celui de la
galanterie champêtre.”
 Note 39, p. 115.
I see that Servius attempts to excuse Virgil on other grounds, for
the difference between the two shields has not escaped his notice.
“Sane interest inter hunc et Homeri clypeum; illic enim singula dum
fiunt narrantur; hic vero perfecto opere nascuntur; nam et hic arma
prius accipit Æneas, quam spectaret; ibi postquam omnia narrata
sunt, sic a Thetide deferuntur ad Achillem.” There is a marked
difference between this and the shield of Homer: for there events are
narrated one by one as they are done, here they are known by the
finished work; here the arms are received by Æneas before being
seen, there, after all has been told, they are carried by Thetis to
Achilles. (Ad. v. 625, lib. viii. Æneid.) Why? “For this reason,” says
Servius: “because, on the shield of Æneas, were represented not only
the few events referred to by the poet, but,—
... genus omne futuræ
Stirpis ab Ascanio, pugnataque in ordine bella,

“All the description of his future race from Ascanius, and the battles,
in the order in which they should occur.” It would have been
impossible for the poet, in the same short space of time occupied by
Vulcan in his work, to mention by name the long line of descendants,
and to tell of all their battles in the order of their occurrence. That
seems to be the meaning of Servius’s somewhat obscure words:
“Opportune ergo Virgilius, quia non videtur simul et narrationis
celeritas potuisse connecti, et opus tam velociter expedire, ut ad
verbum posset occurrere.” Since Virgil could bring forward but a
small part of “the unnarratable text of the shield,” and not even that
little while Vulcan was at work, he was obliged to reserve it till the
whole was finished. For Virgil’s sake, I hope that this argument of
Servius is baseless. My excuse is much more creditable to him. What
need was there of putting the whole of Roman history on a shield?
With few pictures Homer made his shield an epitome of all that was
happening in the world. It would almost seem that Virgil, despairing
of surpassing the Greek in the design and execution of his pictures,
was determined to exceed him at least in their number, and that
would have been the height of childishness.
 Note 40, p. 118.
“Scuto ejus, in quo Amazonum prœlium cælavit intumescente
ambitu parmæ; ejusdem concava parte deorum et gigantum,
dimicationem.”
“Her shield, on the convex side of which he sculptured a battle of
the Amazons, and on the concave side the contest of the gods and
giants.” (Plinius, lib. xxxvi. sect. 4.)
 Note 41, p. 122.
The first begins at line 483 and goes to line 489; the second
extends from 490 to 509; the third, from 510 to 540; the fourth, from
541 to 549; the fifth, from 550 to 560; the sixth, from 561 to 572; the
seventh, from 573 to 586; the eighth, from 587 to 589; the ninth,
from 590 to 605; and the tenth, from 606 to 608. The third picture
alone is not so introduced; but that it is one by itself is evident from
the words introducing the second,—ἐν δὲ δύω ποίησε πόλεις,—as
also from the nature of the subject.
 Note 42, p. 123.
Iliad, vol. v. obs. p. 61. In this passage Pope makes an entirely false
use of the expression “aerial perspective,” which, in fact, has nothing
to do with the diminishing of the size according to the increased
distance, but refers only to the change of color occasioned by the air
or other medium through which the object is seen. A man capable of
this blunder may justly be supposed ignorant of the whole subject.
 Note 43, p. 128.
Constantinus Manasses Compend. Chron. p. 20 (edit. Venet).
Madame Dacier was well pleased with this portrait of Manasses,
except for its tautology. “De Helenæ pulchritudine omnium optime
Constantinus Manasses; nisi in eo tautologiam reprehendas.” (Ad
Dictyn Cretensem, lib. i. cap. 3, p. 5.) She also quotes, according to
Mezeriac (Comment. sur les Epîtres d’Ovide, T. i. p. 361), the
descriptions given by Dares Phrygius, and Cedrenus, of the beauty of
Helen. In the first there is one trait which sounds rather strange.
Dares says that Helen had a mole between her eyebrows: “notam
inter duo supercilia habentem.” But that could not have been a
beauty. I wish the Frenchwoman had given her opinion. I, for my
part, regard the word “nota” as a corruption, and think that Dares
meant to speak of what the Greeks called μεσόφρυον, and the Latins,
“glabella.” He means to say that Helen’s eyebrows did not meet, but
that there was a little space between them. The taste of the ancients
was divided on this point. Some considered this space between the
eyebrows beauty, others not. (Junius de Pictura Vet. lib. iii. cap. 9, p.
245.) Anacreon took a middle course. The eyebrows of his beloved
maiden were neither perceptibly separated, nor were they fully
grown together: they tapered off delicately at a certain point. He says
to the artist who is to paint her (Od. 28):—
τὸ μεσόφρυον δὲ μή μοι
διάκοπτε, μήτε μίσγε,
ἐχέτω δ’ ὅπως ἐκείνη
τὸ λεληθότως σύνοφρυν
βλεφάρων ἴτυν κελαινήν.

This is Pauer’s reading, but the meaning is the same in other

versions, and has been rightly given by Henr. Stephano:—
Supercilii nigrantes
Discrimina nec arcus,
Confundito nec illos:
Sed junge sic ut anceps
Divortium relinquas,
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.

Let us accompany you on the journey of exploring knowledge and

personal growth!

ebookluna.com