Microstrip and Printed Antenna Design

Telecommunications 3rd Edition Randy

Bancroft
Visit to download the full and correct content document:
https://ebookmeta.com/product/microstrip-and-printed-antenna-design-telecommunic
ations-3rd-edition-randy-bancroft/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Microstrip Antenna Design for Wireless Applications 1st

Edition Praveen Kumar Malik

https://ebookmeta.com/product/microstrip-antenna-design-for-
wireless-applications-1st-edition-praveen-kumar-malik/

Microstrip Antenna Design for Wireless Applications 1st

Edition Praveen Kumar Malik Sanjeevikumar Padmanaban
Jens Bo Holm Nielsen

https://ebookmeta.com/product/microstrip-antenna-design-for-
wireless-applications-1st-edition-praveen-kumar-malik-
sanjeevikumar-padmanaban-jens-bo-holm-nielsen/

Fundamentals of Web Development 3rd Edition Randy

Connolly

https://ebookmeta.com/product/fundamentals-of-web-
development-3rd-edition-randy-connolly/

The Power of Comics and Graphic Novels Culture Form and

Context 3rd Edition Randy Duncan

https://ebookmeta.com/product/the-power-of-comics-and-graphic-
novels-culture-form-and-context-3rd-edition-randy-duncan/
Best Easy Day Hikes Orange County 3rd Edition Randy
Vogel Bruce Grubbs

https://ebookmeta.com/product/best-easy-day-hikes-orange-
county-3rd-edition-randy-vogel-bruce-grubbs/

The Art of Communication Improving Your Fundamental

Communication Skills 3rd Edition Randy Fujishin

https://ebookmeta.com/product/the-art-of-communication-improving-
your-fundamental-communication-skills-3rd-edition-randy-fujishin/

Kicad Like a Pro A Comprehensive Hands on Guide for

Learning the World s Favourite Open Source Printed
Circuit Board Design Tool 3rd Edition Dr Peter
Dalmaris
https://ebookmeta.com/product/kicad-like-a-pro-a-comprehensive-
hands-on-guide-for-learning-the-world-s-favourite-open-source-
printed-circuit-board-design-tool-3rd-edition-dr-peter-dalmaris/

Antenna Architectures for Future Wireless Devices Koul

https://ebookmeta.com/product/antenna-architectures-for-future-
wireless-devices-koul/

Metamaterials for Antenna Applications 1st Edition Amit

K. Singh

https://ebookmeta.com/product/metamaterials-for-antenna-
applications-1st-edition-amit-k-singh/
IET HEALTHCARE TECHNOLOGIES PBTE0830

Microstrip and Printed Antenna

Design
Microstrip and Printed Antenna
Design
Randy Bancroft

The Institution of Engineering and Technology

Published by The Institution of Engineering and Technology, London, United Kingdom

The Institution of Engineering and Technology is registered as a Charity in England & Wales
(no. 211014) and Scotland (no. SC038698).

© The Institution of Engineering and Technology 2019

First published 2004

Second edition 2009

Third edition 2019

This publication is copyright under the Berne Convention and the Universal Copyright
Convention. All rights reserved. Apart from any fair dealing for the purposes of research or
private study, or criticism or review, as permitted under the Copyright, Designs and Patents Act
1988, this publication may be reproduced, stored or transmitted, in any form or by any means,
only with the prior permission in writing of the publishers, or in the case of reprographic
reproduction in accordance with the terms of licences issued by the Copyright Licensing
Agency. Enquiries concerning reproduction outside those terms should be sent to the publisher
at the undermentioned address:

The Institution of Engineering and Technology Michael Faraday House Six Hills Way,
Stevenage Herts, SG1 2AY United Kingdom

www.theiet.org

While the authors and publisher believe that the information and guidance given in this work
are correct, all parties must rely upon their own skill and judgement when making use of them.
Neither the authors nor publisher assumes any liability to anyone for any loss or damage
caused by any error or omission in the work, whether such an error or omission is the result of
negligence or any other cause. Any and all such liability is disclaimed.

The moral rights of the authors to be identified as authors of this work have been asserted by
them in accordance with the Copyright, Designs and Patents Act 1988.

British Library Cataloguing in Publication Data

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

ISBN 978-1-78561-854-3 (hardback)

ISBN 978-1-78561-855-0 (PDF)

Typeset in India by MPS Limited

Printed in the UK by CPI Group (UK) Ltd, Croydon

Other volumes in this series:
Volume 9 Phase Noise in Signal Sources W.P. Robins
Volume 12 Spread Spectrum in Communications R. Skaug and J.F. Hjelmstad
Volume 13 Advanced Signal Processing D.J. Creasey (Editor)
Volume 19 Telecommunications Traffic, Tariffs and Costs R.E. Farr
Volume 20 An introduction to Satellite Communications D.I. Dalgleish
Volume 26 Common-channel Signalling R.J. Manterfield
Volume 28 Very Small Aperture Terminals (VSATs) J.L. Everett (Editor)
Volume 29 ATM: The broadband telecommunications solution L.G. Cuthbert
and J.C. Sapanel
Volume 31 Data Communications and Networks, 3rd Edition R.L. Brewster
(Editor)
Volume 32 Analogue Optical Fibre Communications B. Wilson, Z.
Ghassemlooy and I.Z. Darwazeh (Editors)
Volume 33 Modern Personal Radio Systems R.C.V. Macario (Editor)
Volume 34 Digital Broadcasting P. Dambacher
Volume 35 Principles of Performance Engineering for Telecommunication
and Information Systems M. Ghanbari, C.J. Hughes, M.C. Sinclair and J.P. Eade
Volume 36 Telecommunication Networks, 2nd Edition J.E. Flood (Editor)
Volume 37 Optical Communication Receiver Design S.B. Alexander
Volume 38 Satellite Communication Systems, 3rd Edition B.G. Evans (Editor)
Volume 40 Spread Spectrum in Mobile Communication O. Berg, T. Berg, J.F.
Hjelmstad, S. Haavik and R. Skaug
Volume 41 World Telecommunications Economics J.J. Wheatley
Volume 43 Telecommunications Signalling R.J. Manterfield
Volume 44 Digital Signal Filtering, Analysis and Restoration J. Jan
Volume 45 Radio Spectrum Management, 2nd Edition D.J. Withers
Volume 46 Intelligent Networks: Principles and applications J.R. Anderson
Volume 47 Local Access Network Technologies P. France
Volume 48 Telecommunications Quality of Service Management A.P. Oodan
(Editor)
Volume 49 Standard Codecs: Image compression to advanced video coding
M. Ghanbari
Volume 50 Telecommunications Regulation J. Buckley
Volume 51 Security for Mobility C. Mitchell (Editor)
Volume 52 Understanding Telecommunications Networks A. Valdar
Volume 53 Video Compression Systems: From first principles to
concatenated codecs A. Bock
Volume 54 Standard Codecs: Image compression to advanced video coding,
3rd Edition M. Ghanbari
Volume 59 Dynamic Ad Hoc Networks H. Rashvand and H. Chao (Editors)
Volume 60 Understanding Telecommunications Business A. Valdar and I.
Morfett
Volume 65 Advances in Body-Centric Wireless Communication:
Applications and state-of-the-art Q.H. Abbasi, M.U. Rehman, K. Qaraqe and A.
Alomainy (Editors)
Volume 67 Managing the Internet of Things: Architectures, theories and
applications J. Huang and K. Hua (Editors)
Volume 68 Advanced Relay Technologies in Next Generation Wireless
Communications I. Krikidis and G. Zheng
Volume 69 5G Wireless Technologies A. Alexiou (Editor)
Volume 70 Cloud and Fog Computing in 5G Mobile Networks E. Markakis,
G. Mastorakis, C.X. Mavromoustakis and E. Pallis (Editors)
Volume 71 Understanding Telecommunications Networks, 2nd Edition A.
Valdar
Volume 72 Introduction to Digital Wireless Communications Hong-Chuan
Yang
Volume 73 Network as a Service for Next Generation Internet Q. Duan and S.
Wang (Editors)
Volume 74 Access, Fronthaul and Backhaul Networks for 5G & Beyond M.A.
Imran, S.A.R. Zaidi and M.Z. Shakir (Editors)
Volume 76 Trusted Communications with Physical Layer Security for 5G
and Beyond T.Q. Duong, X. Zhou and H.V. Poor (Editors)
Volume 77 Network Design, Modelling and Performance Evaluation Q. Vien
Volume 78 Principles and Applications of Free Space Optical
Communications A.K. Majumdar, Z. Ghassemlooy, A.A.B. Raj (Editors)
Volume 79 Satellite Communications in the 5G Era S.K. Sharma, S.
Chatzinotas and D. Arapoglou
Volume 80 Transceiver and System Design for Digital Communications, 5th
Edition Scott R. Bullock
Volume 81 Applications of Machine Learning in Wireless Communications
R. He and Z. Ding (Editors)
Volume 84 Low Electromagnetic Emission Wireless Network Technologies:
5G and beyond M.A. Imran, F. Héliot and Y.A. Sambo (Editors)
Volume 86 Advances in Communications Satellite Systems Proceedings of the
36th International Communications Satellite Systems Conference (ICSSC-2018)
I. Otung, T. Butash and P. Garland (Editors)
Volume 905 ISDN Applications in Education and Training R. Mason and P.D.
Bacsich
Contents
Preface

1 Microstrip antennas
1.1 The origin of microstrip radiators

1.2 Microstrip antenna analysis methods

1.3 Microstrip antenna advantages and disadvantages

1.4 Microstrip antenna applications

References

2 Rectangular microstrip antennas

2.1 The transmission line model

2.2 The cavity model

2.2.1 The TM10 and TM01 modes

2.3 Radiation pattern of a linear rectangular patch

2.4 Quarter-wave microstrip antenna

2.5 Circular-polarized design

2.5.1 Single feed CP design
2.5.2 Dual-feed CP design
2.5.3 Quadrature (90°) hybrid
2.5.4 Impedance and axial ratio bandwidth

2.6 Efficiency

2.7 Design of microstrip antenna with dielectric cover

2.8 Design guidelines for rectangular microstrip antenna

2.9 Design guidelines for a circularly polarized microstrip antenna

2.10 Electromagnetically coupled rectangular microstrip antenna

2.11 Ultra-wide rectangular microstrip antennas

2.12 Rectangular microstrip antenna cross-polarization

2.12.1 MSA with shorted non-radiating edges

References

3 Circular microstrip antennas

3.1 Circular microstrip antenna properties

3.2 Directivity

3.3 Input resistance and impedance bandwidth

3.3.1 TM11, TM21, and TM02 impedance bandwidth

3.4 Gain, radiation pattern, and efficiency

3.4.1 TM11 efficiency
3.4.2 TM21 efficiency
3.4.3 TM02 efficiency

3.5 Circular microstrip antenna radiation modes

3.5.1 The TM11 bipolar mode
3.5.2 TM11 bipolar mode circular polarized design
3.5.3 The TM21 quadrapolar mode
3.5.4 The TM02 unipolar mode

3.6 Circular microstrip antenna cross-polarization

3.7 Annular microstrip antenna

3.8 Shorted annular microstrip antenna

References

4 Broadband microstrip antennas

4.1 Broadband microstrip antennas

4.2 Microstrip antenna broadbanding

4.2.1 Microstrip antenna matching with capacitive slot
 4.2.2 Microstrip antenna broadband matching with bandpass filter
4.2.3 Example microstrip antenna lumped-element broadband match
4.2.4 Lumped elements to T-line conversion
4.2.5 Real frequency technique broadband matching

4.3 Patch shape for optimized bandwidth

4.3.1 Patch shape bandwidth optimization using genetic algorithm

4.4 Broadband monopole pattern patch-ring

References

5 Dual-band microstrip antennas

5.1 Rectangular microstrip dual-band antenna

5.2 Multiple resonator dual-band antennas

5.2.1 Coupled microstrip dipoles
5.2.2 Stacked rectangular microstrip antennas

5.3 Dual-band microstrip antenna design using a diplexer

5.3.1 Example dual-band microstrip antenna using a diplexer

5.4 Multiband patch shaping using a genetic algorithm

References

6 Microstrip arrays
6.1 Planar array theory

6.2 Rectangular microstrip antenna array modeled with slots

6.3 Aperture excitation distribution

6.4 Microstrip array feeding methods

6.4.1 Corporate-fed microstrip array
6.4.2 Series-fed microstrip array
6.4.3 Series/parallel standing wave feed
6.4.4 Series/parallel matched tapped feed array
6.4.5 Feedline radiation and loss
6.4.6 Microstrip transmission line radiation

6.5 Mutual coupling

6.5.1 Mutual coupling between square MSAs
References

7 Printed antennas
7.1 Omnidirectional microstrip antenna
7.1.1 Low sidelobe omnidirectional MSA
7.1.2 Element shaping of OMA
7.1.3 Single-short omnidirectional microstrip antenna
7.1.4 Corporate-fed omnidirectional microstrip antenna

7.2 Tapered (Vivaldi) antenna

7.3 Microstrip-fed slot antenna

7.3.1 Slot antenna “fictitious resonance”

7.4 Stripline Series Slot Antenna

7.5 Inverted F antenna

7.6 Electrically small antennas

7.6.1 Electrically small antenna limitations
7.6.2 Meanderline antenna
7.6.3 Meanderline antenna radiation patterns
7.6.4 Half patch with reduced SC plane (PIFA)
7.6.5 Dual-band PIFA

7.7 Tapered balun printed dipole

7.8 Log-periodic balun dipole

7.9 Loop antenna and coupler for RFID

7.10 CPW flexible monopole

7.11 Characteristic mode antenna

References

8 Millimeter wave microstrip antennas

8.1 General millimeter wave design considerations

8.2 Corporate-fed patch arrays

8.2.1 28 GHz example
8.2.2 60 GHz example
References

Appendix A Microstrip antenna substrates

A.1 Microstrip antenna/transmission line substrates

A.2 Metal cladding

A.3 Dielectric materials

A.3.1 Plastics
A.3.2 Ceramics
A.3.3 Glass transition temperature (Tg)
A.3.4 Composite dielectric substrates
A.3.5 FR-4
A.3.6 Fiberglass
A.3.7 Dielectric foam

A.4 Radome materials

A.5 Water absorption

A.6 Dielectric films

A.7 Passive intermodulation

A.8 Solder mask and conformal coatings

References

Appendix B Numerical methods

B.1 Numerical integration

B.2 Evaluation of sums

B.3 Fixed-point iteration

B.4 Bisection algorithm

B.5 MSA Q-efficiency calculation

References

Appendix C Planar transmission lines

C.1 Microstrip transmission line design
C.2 Discontinuity compensation

C.3 Dielectric covered microstrip line

C.4 Twin strip transmission line

C.5 Parallel plate transmission line

References

Appendix D Antenna topics

D.1 Friis transmission formula

D.2 Wireless link range versus power input

D.3 Decibels
D.3.1 Historical origin of the decibel (dB)

D.4 Antenna gain and directivity

D.5 Attenuation and voltage standing wave ratio

D.5.1 Example 1
D.5.2 Example 2

D.6 Return Loss and Reflection Loss

D.7 Attenuators
D.7.1 Example 3

References

Appendix E Impedance matching techniques

E.1 The λ/8 transmission line transformer
E.1.1 Example: Combined λ/8 and λ/4 transformer matching
E.1.2 Dual λ/8 transmission line transformer

E.2 Q matching with λ/8 transmission line transformer

E.2.1 Example: Combined λ/8 transformer and Q matching

E.3 Single section series T-line impedance transformer

E.3.1 Example: Single section impedance match

E.4 Bramham–Regier two-section impedance transformer

E.4.1 Example: BR transmission line transformer design
E.5 Two-section Chebyshev impedance transformer
E.5.1 Example: λ/8 with two-section Chebyshev transformer

E.6 Bode-Fano limits/matching overview

References

Appendix F Baluns for printed antennas

F.1 Transmission line theory

F.2 L-C lattice balun

F.3 Coupled microstrip transmission line balun

F.4 Microstrip transmission line Marchand balun

F.5 Microstrip branchline (ladder) balun

References

Index
 Preface

The first and second edition of this book were written for designers of planar
microstrip antennas; the third edition expands on this. It includes details and
subjects that are useful to the physical realization of microstrip and printed
antennas that are generally not found in more academic works. An appendix on
balun design and analysis has been added, as a number of printed antennas require
a balanced feed and impedance matching. This book is written with commercial
applications in mind, but is also of use for aerospace and other fields. The designs
selected have been useful to the author over his career, along with ones that are
clearly of utility.
There are many very complex microstrip antenna designs in the literature, but
this volume emphasizes simplicity, and accessible mathematical designs, rather
than offering a geometry, and rationalizing the resonances and properties in an ad
hoc manner. This text includes the most recent work available from researchers in
the field of microstrip and printed antennas. This book is intended to be a
succinct, accessible, handbook, which provides useful, practical, simple, and
manufacturable antenna designs. It also includes references which allow the
reader to investigate more complex designs.
The third edition has a considerable number of additions to the material in
earlier editions, which I hope will make concepts presented clearer. The
efficiency analysis of rectangular and circular microstrip antennas is expanded
upon. Full-wave analysis software (HFSS) made it possible to separate out
radiative, conductor, and dielectric losses, and an extra loss that is thought to be
some manner of surface wave loss. The fields inside of copper conductors were
computed to produce as accurate modeling as possible. The impedance bandwidth
of quarter-wave microstrip antennas versus relative permittivity is more
thoroughly examined. Series/parallel feed structures in microstrip arrays have
been added to this edition. The Vivaldi antenna section has been rewritten,
demonstrating that common textbook design geometry limits the antenna's
intrinsic impedance bandwidth. A number of useful designs have been added, and
the appendices expanded.
This edition of the text uses updated nomenclature. First, the term dielectric
constant has been eliminated. The author has measured the relative permittivity of
numerous materials in transmission lines (waveguide and coax), cavities, and with
microstrip methods. Dielectric permittivity generally changes with frequency and
is not constant. The IEEE Standard Definitions of Terms for Radio Wave
Propagation (IEEE Std 211-1997) no longer accepts the term dielectric constant.*
It allows the terms permittivity, electric permittivity and dielectric permittivity. I
will also use relative permittivity for permittivity normalized to free space.
Another nomenclature change is increased, and more effective use of metric
quantities. Like the use of dielectric constant, centimeters are eschewed from this
text. The use of metric prefixes separated by 1000 (metric triads) is implemented.
The use of commas as triad separators has been eliminated, and thinspaces replace
them instead. In the case of a four digit number, either a space or no space is
implemented depending on context. If a table contains values up to four digits, the
space is generally suppressed. If the table has entries with five digits or more,
then usually the space is inserted into the four digit values. The prefix cluster
around unity: centi, deci, deca, and hecto are atavistic prefixes and should not be
used in a modern context.† It was realized that in metric building construction, the
use of millimeters as the everyday small metric unit essentially eliminated the use
of a decimal point on drawings, decreasing errors. The side of a house can be 23
000 mm and also immediately understood as 23 m. The use of centimeters is
forbidden in metric construction standards. It was noted by metrication expert Pat
Naughtin, that when comparing numbers in a table, choosing a metric prefix that
allows integer representation produces easy cognitive comparison of numbers in
tables. Microwave engineers have done this when they use 2450 MHz rather than
switching to 2.450 GHz. If a frequency sweep is from say 850 MHz to 2450
MHz, it makes no sense to introduce a cognitive discontinuity by switching to
GHz at 1000 MHz. There is no reason to immediately change metric prefixes at
1000. In everyday work, millimeters, grams, and milliliters allow one to use
integers to express length, mass, and volume. When possible, the thickness of
dielectric substrates will be given in micrometers ( m). This allows substrates to
be described with integers over the microwave and millimeter wave regions of the
electromagnetic spectrum.
The anachronistic term mil, is not a formally accepted unit of length, and will
not be used. The appropriate metric unit: micrometers or nanometers will be
employed. When an antenna design is articulated in millimeters, the dimensions
will all be in millimeters. The thickness of microwave substrates are best
designated in micrometers m as it produces integer values (similar to the pre-
metric integer values 10 mil, 20 mil, etc.). I have done this with comparison tables
in this book, but when substrate thicknesses are part of a design in millimeters,
they are presented as millimeters. For high-frequency work, micrometers can be
easier to work with than millimeters and should be used appropriately. The term
micron is a slang term for micrometer ( m) and should not be used in technical
papers. The term micron hides the dimensions under discussion with an argot.
Micron is sometimes thought to be an inch-based pre-metric quantity, introducing
a source of potential error.
When describing insertion loss, the use of the pre-metric dB/inch will not be
used. It is far preferable to use dB/100 mm. This produces larger numerical values
than dB/inch, which often have more utility. An insertion loss of 4 dB/100 mm is
effortlessly converted to 40 dB/1 000 mm or 40 dB/m. Converting 4 dB/in to
dB/foot, or dB/yard, requires more cognitive effort.
In the second edition of this book, in Appendix A, Table A.1, on p. 237, the
thickness of copper foil in terms of its weight per square foot is given. Typical
textbooks express 0.5 ounce copper as possessing a thickness of 0.0007 inches, or
0.01778 mm. Generally, the smallest unit of pre-metric measure in the US is the
inch,‡ which is then illegitimately assigned a metric prefix to produce a similar
number of decimal places, rather than thinking in, and using, the metric system. In
this edition, Table A.1 on page 291, is a metric-only table that relates grams per
square meter to foil thickness in micrometers. For instance 150 grams per square
meter is equivalent to a foil thickness of 16.87 m (0.000 664 inches). The
mixing of feet and inches used in the US designation is avoided, and the potential
confusion it creates. Manufacturers weigh the amount of copper foil they have
deposited, rather than measuring its thickness. The US designations are presented
in Table A.2.
When I was an undergraduate, I saw a directional coupler with a calibration
graph on it in KiloMegaCycles. We have made some progress, but there is much,
much room for metric improvement.
*Thiel,D.V., “Using the Right Terms,” IEEE Antennas and Propagation Magazine, October 2010, Vol.
52, No. 5, p. 192.
†Bancroft, R., The Dimensions of the Cosmos, Outskirts Press, 2016.
‡The smallest is the barleycorn which is 1/3 inch and used to define shoe sizes.
 Chapter 1
Microstrip antennas

1.1 The origin of microstrip radiators

The use of coaxial cable and parallel two wire (or “twin lead”) as a transmission
line can be traced to at least the 19th century. The realization of radio frequency
(RF) and microwave components using these transmission lines required
considerable mechanical effort in their construction. Paul Eiser envisioned using
copper foil to interconnect electronic components in the 1930s. The printed circuit
board (PCB) concept remained dormant until engineers required electronic
circuits which would withstand a shock of around 20 000 Gs in 1941. The
application of interest was the proximity fuse. The proximity fuse uses an antenna
at the top of a shell to detect its proximity with respect to a target. The internal
electronics of the proximity fuse utilized the first PCB [1].
The advent of commercially viable PCB techniques in the mid-twentieth
century led to the realization that printed circuit versions of RF transmission lines
could be developed. This would allow for a much simpler mass production of
microwave components. The printed circuit analog of a coaxial cable became
known as stripline. With a groundplane image providing a virtual second
conductor, the printed circuit analog of two wire (“parallel plate”) transmission
line became known as microstrip. For those not familiar with the details of this
transmission line, they can be found in Appendix C at the end of this book.
Microstrip geometries which radiate electromagnetic waves were originally
contemplated in the 1950s. The realization of radiators that are compatible with
microstrip transmission line is nearly contemporary with its introduction by Grieg
and Englemann in 1952 [2]. The earliest known realization of a microstrip-like
antenna, integrated with a microstrip transmission line was developed in 1953 by
Deshamps [3,4] (Figure 1.1). By 1955, Gutton and Baissinot patented a microstrip
antenna design [5].
Figure 1.1 Original conformal array designed by Deshamps [3] in 1953 fed with
a microstrip transmission line

Early microstrip lines and radiators were specialized devices developed in

laboratories. No commercially available PCBs with controlled dielectric
permittivity were developed during this period. The investigation of microstrip
resonators that were also efficient radiators languished. The theoretical basis of
microstrip transmission lines continued to be the object of academic inquiry [6].
Stripline received more interest as a planar transmission line at the time
because it supports a transverse electromagnetic wave and allowed for easier
analysis, design, and development of planar microwave structures. Stripline was
also seen as an adaptation of coaxial cable, and microstrip as an adaptation of two
wire transmission line. R.M. Barrett opined in 1955 that “The merits of these two
systems [stripline and microstrip] are essentially the merits of their respective
antecedents [coaxial cable and two wire] [7].” These viewpoints may have been
part of the reason microstrip did not achieve immediate popularity in the 1950s.
The development of microstrip transmission line analysis and design methods
continued in the mid to late 1960s by Wheeler [8] and Purcel et al. [9,10].
In 1969, Denlinger noted that rectangular and circular microstrip resonators
could efficiently radiate [11]. Previous researchers had realized that in some
cases, 50% of the power in a microstrip resonator would escape as radiation.
Denlinger described the radiation mechanism of a rectangular microstrip
resonator as arising from the discontinuities at each end of a truncated microstrip
transmission line. The two discontinuities, separated by a multiple of a guide half
wavelength, can be treated separately and then combined to describe the complete
radiator. It was noted that the percent of radiated power to the total input power
increased as the substrate thickness of the microstrip resonator increased.
These correct observations are discussed in greater detail in Chapter 2.
Denlinger's results only explored increasing the substrate thickness until
approximately 70% of the input power was radiated into space. Denlinger also
investigated radiation from a resonant circular microstrip disc. He observed that at
least 75% of the power was radiated by one circular resonator under study. In late
1969, Watkins described the fields and currents of the resonant modes of circular
microstrip structures [12].
The microstrip antenna concept finally began to receive closer examination in
the early 1970s when aerospace applications, such as spacecraft and missiles,
produced the impetus for researchers to investigate the utility of conformal
antenna designs. In 1972, Howell articulated the basic rectangular microstrip
radiator fed with microstrip transmission line at a radiating edge [13]. The
microstrip resonator with considerable radiation loss was now described as a
microstrip antenna. A number of antenna designers received the design with
considerable caution. It was difficult to believe that a resonator of this type could
radiate with >90% efficiency. The narrow bandwidth of microstrip antennas
seemed to severely limit the number of possible applications to which the antenna
could prove useful. By the late 1970s, many of these objections had not proven to
derail the use of microstrip antennas in numerous aerospace applications. By
1981, microstrip antennas had become so ubiquitous and studied they were the
subject of a special issue of the IEEE Transactions on Antennas and Propagation
[14].
Today a farrago of designs have been developed which can be bewildering for
designers who are new to the subject. This volume attempts to explain basic
concepts and present useful designs. It will also direct the reader who wishes to
research other microstrip antenna designs, which are not presented in this work, to
pertinent literature.
The geometry which defines a microstrip antenna is presented in Figure 1.2. A
conductive patch exists along the plane of the upper surface of a dielectric slab.
This conductor area, which forms the radiating element, is generally rectangular
or circular, but may be of any shape. The dielectric substrate has a groundplane
on its bottom surface.

Figure 1.2 Geometry of a microstrip antenna

1.2 Microstrip antenna analysis methods

It was known that the resonant length of a rectangular microstrip antenna is
approximately one-half wavelength, with the effective dielectric permittivity of
the substrate taken into account. Following the introduction of the microstrip
antenna, analysis methods were desired to determine the approximate resonant
resistance of a basic rectangular microstrip radiator. The earliest useful model
introduced to provide approximate values of resistance presented at the edge of a
microstrip antenna is known as the transmission line model, introduced by
Munson [15]. The transmission line model provides insight into the simplest
microstrip antenna design, but is not complete enough to be useful when more
than one resonant mode is present. In the late 1970s, Lo et al. developed a model
of the rectangular microstrip antenna as a lossy resonant cavity [16].
Microstrip antennas, despite their simple geometry, proved to be very
challenging to analyze using exact methods. In the 1980s, the method of moments
became the first numerical analysis method which was computationally efficient
enough so that contemporary computers could provide enough memory and CPU
speed to practically analyze microstrip antennas [17–20].
The improvement of computational power and memory size of personal
computers during the 1990s made numerical methods such as the finite difference
time domain (FDTD) method and finite element method (FEM), which require a
much more memory than moment method solutions, workable for everyday use
by designers. This volume will generally use Ansoft HFSS as a full wave analysis
method as well as FDTD [21,22].

1.3 Microstrip antenna advantages and disadvantages

The main advantages of microstrip antennas are
Low-cost fabrication;
Can easily conform to a curved surface of a vehicle or product;
Resistant to shock and vibration (most failures are at the feed probe solder
joint);
Many designs readily produce linear or circular polarization;
Considerable range of gain and pattern options (2.5–10.0 dBi);
Other microwave devices realizable in microstrip may be integrated with a
microstrip antenna with no extra fabrication steps (e.g. branchline hybrid
to produce circular polarization or corporate feed network for an array of
microstrip antennas);
Antenna thickness (profile) is small.

The main disadvantages of microstrip antennas are

Narrow bandwidth (5%–10% [2:1 voltage standing wave ratio] is typical
without special techniques);
Dielectric, conductor, and surface wave losses can be large for thin
patches, resulting in poor antenna efficiency;
Sensitivity to environmental factors such as temperature and humidity.
1.4 Microstrip antenna applications
A large number of commercial needs are met using microstrip and printed
antennas. These include the ubiquitous global positioning system (GPS), Zigbee,
Bluetooth, WiMax, Wi-Fi applications, 802.11a,b,g, and others. The most popular
microstrip antenna is certainly the rectangular patch (Chapter 2). GPS
applications, such as asset tracking of vehicles, and marine uses, have created a
large demand for antennas. The majority of these are rectangular patches, which
have been modified to produce right-hand circular polarization (RHCP), and
operate at 1.575 GHz. Numerous vendors offer patches designed using ceramics
with a high relative dielectric permittivity ( ) to reduce the
rectangular microstrip antenna to as small a footprint as possible for a given
application. Some manufactured patches are provided ready for circuit board
integration with low-noise amplifiers. Rectangular patch antennas are also used
for Bluetooth automotive applications (2.4 GHz) with RHCP.
In recent years, Satellite Digital Audio Radio Services have become a viable
alternative to AM and FM commercial broadcasts in automobiles. The system has
strict radiation pattern requirements which have been met with a combination of a
printed monopole and TM21 mode annular microstrip antenna, which has been
altered with notches to produce left-hand circular polarization at 2.338 GHz [23].
The annular microstrip antenna is addressed in Chapter 3.
Wireless local area networks (WLAN) provide short-range high-speed data
connections between mobile devices (such as a laptop computer) and wireless
access points. The range for wireless data links is typically around 30–90 m (100–
300 feet) indoors and 600 m (2000 feet) outdoors. Wireless data links use the
IEEE Standards 802.11a,b,g. The majority of WLANs use the unlicensed 2.4 GHz
band (802.11b and 802.11g). The 802.11a standard uses the 5 GHz unlicensed
frequency band. Multiband-printed antennas that are integrated into ceiling tiles
use a microstrip diplexer (Chapter 5) to combine the signal from Global System
for Mobile communication cell phones (860 MHz band), personal
communications services cell phones (1.92 GHz band), and 802.11a WLAN
service (2.4 GHz band) provided by two integrated microstrip dipoles [24].
WLAN systems sometimes require links between buildings that have wireless
access points. This is sometimes accomplished using microstrip-phased arrays at
5 GHz (Chapter 6).
In other applications, such as warehouse inventory control, a printed antenna
with an omnidirectional pattern is desired (Chapter 7). Omnidirectional microstrip
antennas are also of utility for many WiMax applications (2.3, 2.5, 3.5, and 5.8
GHz are some of the frequencies currently of interest for WiMax applications)
and for access points. Microstrip-fed printed slot antennas have proven useful to
provide vertical polarization and integrate well into laptop computers (Chapter 7)
for WLAN.
The advantages of using antennas in communication systems will continue to
generate new applications which require their use. Antennas of course have the
advantage of communication mobility without any required physical connection.
They are the device which has enabled all the “wireless” systems that have
become ubiquitous in our society. The use of transmission line, such as coaxial
cable or waveguide, may have an advantage in transmission loss for short lengths,
but as the distance increases the transmission loss between antennas becomes
lower than any transmission line and, in some applications, outperform cables for
shorter distances [25]. The amount of material cost for wired infrastructure also
encourages the use of antennas in many modern communication systems.

References
[1] H. Schlesinger, The Battery, Harper, 2010, pp. 228–229.
[2] Grieg, D.D., and Englemann, H.F., “Microstrip—A Transmission Technique
for the Kilomegacycle Range,” Proceedings of The IRE, 1952, Vol. 40, No.
10, pp. 1644–1650.
[3] Deschamps, G.A., “Microstrip Microwave Antennas,” The Third Symposium
on The USAF Antenna Research and Development Program, University of
Illinois, Monticello, Illinois, October 18–22, 1953.
[4] Bernhard, J.T., Mayes, P.E., Schaubert, D., and Mailoux, R.J., “A
Commemoration of Deschamps' and Sichak's `Microstrip Microwave
Antennas': 50 Years of Development, Divergence, and New Directions,”
Proceedings of the 2003 Antenna Applications Symposium, Monticello,
Illinois, September 2003, pp. 189–230.
[5] Gutton, H., and Baissinot, G., “Flat Aerial for Ultra High Frequencies,”
French Patent No. 703113, 1955.
[6] Wu, T.T., “Theory of the Microstrip,” Journal of Applied Physics, March
1957, Vol. 28, No. 3, pp. 299–302.
[7] Barrett, R.M., “Microwave Printed Circuits—A Historical Survey,” IEEE
Transactions on Microwave Theory and Techniques, Vol. 3, No. 2, March
1955, pp. 1–9.
[8] Wheeler, H.A., “Transmission Line Properties of Parallel Strips Separated by
a Dielectric Sheet,” IEEE Transactions on Microwave Theory of Techniques,
Vol. MTT-13, March 1965, pp. 172–185.
[9] Purcel, R.A., Massé, D.J., and Hartwig, C.P., “Losses in Microstrip,” IEEE
Transactions on Microwave Theory and Techniques, Vol. 16, No. 6, June
1968, pp. 342–350.
[10] Purcel, R.A., Massé, D.J., and Hartwig, C.P., Errata: “Losses in Microstrip,”
IEEE Transactions on Microwave Theory and Techniques, Vol. 16, No. 12,
December 1968, p. 1064.
[11] Denlinger, E.J., “Radiation from Microstrip Radiators,” IEEE Transactions
on Microwave Theory of Techniques, Vol. 17, April 1969, pp. 235–236.
[12] Watkins, J., “Circular Resonant Structures in Microstrip,” Electronics
Letters, October 1969, Vol. 5, No. 21, pp. 524–525.
[13] Howell, J.Q., “Microstrip Antennas,” IEEE International Symposium Digest
on Antennas and Propagation, Williamsburg, Virginia, December 11–14,
1972, pp. 177–180.
[14] IEEE Transactions on Antennas and Propagation, January 1981.
[15] Munson, R.E., “Conformal Microstrip Antennas and Microstrip Phased
 Arrays,” IEEE Transactions on Antennas and Propagation, January 1974,
Vol. 22, No. 1, pp. 235–236.
[16] Lo, Y.T., Solomon, D., and Richards, W.F., “Theory and Experiment on
Microstrip Antennas,” IEEE Transactions on Antennas and Propagations,
1979, Vol. AP-27, pp. 137–149.
[17] Hildebrand, L.T., and McNamara, D.A., “A Guide to Implementational
Aspects of the Spatial–Domain Integral Equation Analysis of Microstrip
Antennas,” Applied Computational Electromagnetics Journal, March 1995,
Vol. 10, No. 1, pp. 40–51, ISSN 1054–4887.
[18] Mosig, J.R., and Gardiol, F.E., “Analytical and Numerical Techniques in the
Green's Function Treatment of Microstrip Antennas and Scatterers,” IEE
Proceedings, March 1983, Vol. 130, Pt. H, No. 2, pp. 175–182.
[19] Mosig, J.R., and Gardiol, F.E., “General Integral Equation Formulation for
Microstrip Antennas and Scatterers,” IEE Proceedings, Vol. 132, Pt. H, No.
7, December 1985, pp. 424–432.
[20] Mosig, J.R., “Arbitrarily Shaped Microstrip Structures and Their Analysis
with a Mixed Potential Integral Equation,” IEEE Transactions on Microwave
Theory and Techniques, February 1988, Vol. 36, No. 2, pp. 314–323.
[21] Tavlov, A., and Hagness, S.C., Computational Electrodynamics: The Finite-
Difference Time-Domain Method, Second Edition, Artech House, 2000.
[22] Tavlov, A., Ed., Advances in Computational Electrodynamics: The Finite-
Difference Time-Domain Method, Artech House, 1998.
[23] Licul, S., Petros, A., and Zafar, I., “Reviewing SDARS Antenna
Requirements,” Microwaves & RF, September 2003.
[24] US Patent No. US 6,307,525 B1 2017-04-07.
[25] Milligan, T., Modern Antenna Design, McGraw Hill, 1985, pp. 8–9.
 Chapter 2
Rectangular microstrip antennas

2.1 The transmission line model

The rectangular patch antenna is very probably the most popular microstrip
antenna geometry implemented by designers. Figure 2.1 shows the geometry of
this antenna type. A rectangular metal patch of width and length
is separated by a dielectric material from a groundplane by a distance h. The two
ends of the antenna (located at 0 and b) can be viewed as radiating due to fringing
fields along each edge of the antenna width W (= a). The two radiating edges are
separated by a distance L (= b). The two edges along the sides of length L are
often referred to as non-radiating edges.

Figure 2.1 Rectangular microstrip patch geometry used to describe the

transmission line model. The patch antenna is fed along the centerline
 of the antenna’s dimension along the -axis (i.e. x = a/2). The feed
point is located at which is chosen to match the antenna with a
desired impedance. The radiation originates from the fringing electric
field at either end of the antenna. These edges are called radiating
edges, and the other two sides (parallel to the -axis) are non-
radiating edges

Numerous full wave analysis methods have been devised for the rectangular
microstrip antenna [1–4]. Often these advanced methods require a considerable
investment of time and effort to implement and are thus not convenient for
computer-aided design (CAD) implementation.
The two analysis methods for rectangular microstrip antennas which are most
popular for CAD implementation are the transmission line model and the cavity
model. In this section, we will address the least complex version of the
transmission line model. The popularity of the transmission line model may be
gauged by the number of extensions to this model which have been developed
[5–7].
The transmission line model provides a very lucid conceptual picture of the
simplest implementation of a rectangular microstrip antenna. In this model, the
rectangular microstrip antenna consists of a microstrip transmission line
terminated with a pair of loads at either end [8,9].
As presented in Figure 2.2(a), the resistive loads at each end of the
transmission line represent loss due to radiation. At resonance, the imaginary
components of the input impedance seen at the driving point cancel, and
therefore, the driving point impedance becomes exclusively real.
Figure 2.2 (a) The transmission line model of a rectangular microstrip antenna is
a transmission line with two terminating loads. A driving point is
chosen along the antenna length L which can be represented as a sum
of L1 and L2. The two transmission line sections each contribute to the
driving point impedance. The antenna is readily analyzed using a pair
of edge admittances separated by two sections of a transmission
line of characteristic admittance Y0. (b) The microstrip antenna may
be fed at one of its radiating edges using a transmission line. In this
 case, the transmission line model is augmented with a feed line of
characteristic admittance of length connected to a radiating
edge. The driving point admittance Y is then computed at the end of
this feed line

The driving point or feed point of an antenna is the location on an antenna

where a transmission line is attached to provide the antenna with a source of
microwave power. The impedance measured at the point where the antenna is
connected to the transmission line is called the driving point impedance or input
impedance. The driving point impedance at any point along the centerline of
a rectangular microstrip antenna can be computed using the transmission line
model. The transmission line model is most easily represented mathematically
using the transmission line equation written in terms of admittances, as presented
in the following equation:

where is the input admittance at the end of a transmission line of length L(=
b), which has a characteristic admittance of and a phase constant of
terminated with a complex load admittance . In other words, the microstrip
antenna is modeled as a microstrip transmission line of width W(= a), which
determines the characteristic admittance, and is of physical length L(= b) and
loaded at both ends by an edge admittance , which models the radiation loss.
This is shown in Figure 2.2(a).
Using (2.1), the driving point admittance at a driving point
located between the two radiating edges is expressed as

where is the complex admittance at each radiating edge, which consists of an

edge conductance and edge susceptance , as related in (2.3). The two loads
are separated by a microstrip transmission line of characteristic admittance :

Approximate values of and Be may be computed using 2.4 and 2.5 [10]:
 The effective relative dielectric permittivity ( ) is given as

The fringing field extension, normalized to the substrate thickness h, is

The value Δl is the line extension due to the electric field fringing at the edge
of the patch antenna. The physical size of a resonant microstrip patch antenna
would be were it not for the effect of fringing at the end of the rectangular
microstrip antenna.1 We may use equation (2.7) to correct for this effect and
compute the physical length of a rectangular microstrip antenna, which will
resonate at a desired design frequency .
The first feed method is called a coaxial probe feed,2 as shown in Figure
2.3(a). The outer shield of a coaxial transmission line is connected to the
groundplane of the microstrip antenna. Metal is removed from the groundplane
which is generally the same radius as the inside of the outer coaxial shield. The
coaxial center conductor then passes through the dielectric substrate of the patch
antenna and connects to the patch. Feeding the antenna in the center (i.e. at a/2)
suppresses the excitation of a mode along the width of the antenna. This feed
symmetry enforces the purest linear polarization along the length of the patch
which can be achieved with a single direct feed.
Figure 2.3 Common methods used to feed a rectangular microstrip antenna. (a)
Coaxial feed probe. (b) Microstrip transmission line feed along a
nonradiating edge. (c) Microstrip transmission feed along a radiating
edge. (d) Microstrip feed line into a cutout in a radiating edge which
is inset to a 50 Ω driving point

The second feed method, shown in Figure 2.3(b), drives the antenna with a
microstrip transmission line along a non-radiating edge. This feed method is
modeled in an identical manner to the coaxial probe feed when using the
transmission line model; in practice, it can often excite a mode along the width of
the patch when and cause the antenna to radiate with an elliptical
polarization. The advantage of this feed method is that it allows one to use a 50 Ω
microstrip transmission line connected directly to a 50 Ω driving point
impedance, which eliminates the need for impedance matching.
The third feed method, Figure 2.3(c), is to drive the antenna at one of its
radiating edges with a microstrip transmission line. This disturbs the field
distribution along one radiating edge which causes slight changes in the radiation
pattern. The impedance of a typical resonant rectangular ( ) microstrip
antenna at a radiating edge is around 200 Ω. This edge resistance at
resonance. In general, one must provide an impedance transformation to 50 Ω for
this feed method. This is often accomplished using a quarter-wave impedance
transformer between the radiating edge impedance and a 50 Ω microstrip feed
line. A quarter-wave transformer has a larger bandwidth than the antenna element
and therefore does not limit it. It is possible to widen a rectangular microstrip
antenna ( ) so the edge resistance at resonance is 50 Ω. In this special case,
no impedance transformer is required to feed the antenna with a 50 Ω microstrip
transmission line at a radiating edge.
A fourth feed method, illustrated in Figure 2.3(d), is to cut a narrow notch out
of a radiating edge far enough into the patch to locate a 50 Ω driving point
impedance. The removal of the notch perturbs the patch fields. A study by Basilio
et al. indicates that a probe-fed patch antenna has a driving point resistance which
follows an , a patch with an inset feed is measured to follow an
function, where [11]. More complex
relationships have also been developed [12]. One can increase the patch width,
which increases the edge conductance, until at resonance, and the edge impedance
is 50 Ω. The inset distance into the patch goes to zero which allows one to
directly feed a patch for this special case using a 50 Ω microstrip line at a
radiating edge. The patch width is large enough in this case to increase the
antenna gain considerably.
Equation (2.8) may be used to compute the resonant length (L) of a
rectangular microstrip antenna. The cutback value is given by (2.7) and the
effective relative permittivity, , is given by (2.6):

Equation (2.2) provides a predicted input impedance at the desired design

frequency . Numerical methods for obtaining the roots of an equation such as
the Bisection method (Appendix B) may be used with (2.2) to determine the value
of and L2 which correspond to a desired input resistance value. The initial
guesses are along b at ( ) and ( ). The
predicted position of a desired driving point impedance to feed the antenna is
generally close to measurement as long as the substrate height is not larger than
about 0.1 . A good rule of thumb for an initial guess to the location of a 50 Ω
feed point, when determining the position in an empirical manner, is 1/3 of the
distance from the center of the antenna to a radiating edge, inward from a
radiating edge.
Early investigation of the rectangular microstrip antenna, viewed as a linear
transmission line resonator, was undertaken by Derneryd [13]. The input
impedance characteristics of the transmission line model were altered by
Derneryd in a manner which allows for the influence of mutual conductance
between the radiating edges of the patch antenna. This model further allows for
the inclusion of higher order linear transmission line modes.
In 1968, an experimental method to investigate the electric fields surrounding
objects was developed, which used a liquid crystal sheet (LCS) backed with a
resistive thin film material [14,15]. Derneryd used a liquid crystal field detector of
this type to map the electric field of a narrow microstrip antenna. Derneryd’s
results are reproduced in Figure 2.4 along with thermal (electric field magnitude)
plots produced using the finite difference time domain (FDTD) method. The
FDTD patch analysis used a = 10.0 mm, b = 30.5 mm, ϵr = 2.55, , and
. The feedpoint location is 5.58 mm from the center of the patch antenna
along the centerline. The groundplane is 20 mm × 42 mm.
Figure 2.4 Electric field distribution surrounding a narrow patch antenna as
computed using FDTD and measured using a liquid crystal sheet: (a)
patch without fields, (b) 3.10 GHz, (c) 6.15 GHz, and (d) 9.15 GHz.
After Derneryd [13]

Figure 2.4(a) is the antenna without an electric field present. Figure 2.4(b),
just to the right of (a), is Derneryd’s element analyzed with a thermal LCS, which
shows the first (lowest order) mode of this antenna. The frequency for this first
mode is reported to be 3.10 GHz. A sinusoidal source at 3.10 GHz with FDTD
was used to model this antenna. The FDTD plot is of the total magnitude of the
electric field in the plane of the antenna. The FDTD simulation thermal plot is
very similar to the shape of the measured LCS thermal pattern. We see two
radiating edges at either end of the antenna in the lowest mode with two non-
radiating edges on the sides.
Figure 2.4(c) has Derneryd’s measured LCS results with the antenna driven at
6.15 GHz. The LCS visualization shows the next higher order mode one would
expect from transmission line theory. The electric field seen at either side of the
center of the patch antenna along the non-radiating edges still contribute little to
the antenna’s radiation. In the far field3, the radiation contribution from each side
of the non-radiating edges cancel. The FDTD thermal plot result in Figure 2.4(c)
is once again very similar in appearance to Derneryd’s LCS thermal measurement
at 6.15 GHz.
The next mode is reported by Derneryd to exist at 9.15 GHz. The measured
LCS result in Figure 2.4(d) and the theoretical FDTD thermal plot once again
have good correlation. As before, the radiation from the non-radiating edges will
cancel in the far field.
The LCS method of measuring the near fields of microstrip antennas is still
used, but other photographic and probe measurement methods have been
developed as an aid to the visualization of the fields around microstrip antennas
[16–19].

2.2 The cavity model

The transmission line model is conceptually simple, but has a number of
drawbacks. The transmission line model is often inaccurate when used to predict
the impedance bandwidth of a rectangular microstrip antenna for thin substrates.
The transmission line model also does not take into consideration the possible
excitation of modes which are not along the linear transmission line. The
transmission line model assumes that the currents flow in only one direction along
the transmission line. In reality, currents transverse to these assumed currents can
exist in a rectangular microstrip antenna. The development of the cavity model
addressed these difficulties.
The cavity model, originated in the late 1970s by Lo et al., views the
rectangular microstrip antenna as an electromagnetic cavity, with electric walls at
the groundplane and the patch, and magnetic walls at each edge [20,21]. The
fields under the patch are the superposition of the resonant modes of this two-
dimensional radiator.4 Equation (2.10) expresses the ( ) electric field under the
patch at a location (x, y) in terms of these modes. This model has undergone a
considerable number of refinements since its introduction [22,23].
The fields in the lossy cavity are assumed to be the same as those which will
exist in a thin cavity of this type. It is assumed that in this configuration where (
) only a vertical electric field ( ) will exist, which is assumed to be
constant along , and only horizontal magnetic field components ( and )
exist. The magnetic field is transverse to the -axis (Figure 2.5), and the modes
are described as TMmn modes (m and n are integers). The electric current on the
rectangular patch antenna is further assumed to equal zero normal to each edge.
Because the electric field is assumed to be constant along the direction, one
can multiply (2.10) by h to obtain the voltage difference from the patch to the
groundplane. The driving point current can be mathematically manipulated to
produce the ratio of voltage to current on the left side of (2.10). This creates an
expression which can be used to compute the driving point impedance (2.15) at an
arbitrary point as illustrated in Figure 2.5:
Figure 2.5 Rectangular microstrip patch geometry used for cavity model analysis

The cavity walls are slightly larger electrically than they are physically due to
the fringing field at the edges; therefore, we extend the patch boundary outward
and the new dimensions become and which are
used in the mode expansion. The effect of radiation and other losses is
represented by lumping them into an effective dielectric loss tangent as defined in
(2.19):

The driving point impedance at ( ) may be calculated using

where is the width of the feed probe.

The effective loss tangent for the cavity is computed from the total Q of the
cavity:

The total quality factor of the cavity consists of four components: , the
dielectric loss; , the conductor loss; , the radiation loss; and , the
surface wave loss:

where is the energy stored:

The power radiated into space is [24]:

where is the input (driving point) voltage.
The Q of the surface wave loss is related to the radiation quality factor
[25]:
 The cavity model is conceptually accessible and readily implemented, but its
accuracy is limited by assumptions and approximations, which are only valid for
electrically thin substrates. The self-inductance of a coaxial probe used to feed the
rectangular microstrip antenna is not included in this model. The cavity model is
generally accurate in its impedance prediction and is within 3% of measured
resonant frequency for a substrate thickness of or less. When it is thicker
than this, anomalous results may occur [26].

2.2.1 The TM10 and TM01 modes

When a rectangular microstrip antenna has its dimension a wider than dimension
b and is fed along the centerline of dimension b, only the TM10 mode is driven.
When it is fed along the centerline of dimension a, only the TM01 mode is driven.
When the geometric condition is met, the TM10 mode is the lowest
order mode and possesses the lowest resonant frequency of all the time harmonic
modes. The TM01 mode is the next highest order mode and has the next lowest
resonant frequency (Figure 2.6).

Figure 2.6 When (a > b), the TM10 mode is the lowest order mode (lowest
resonant frequency) for a rectangular microstrip antenna. The TM01
mode has the next highest resonant frequency

When the situation is reversed, TM01 becomes the mode with the
lowest resonant frequency and TM10 has the next lowest resonant frequency.
If , the two modes TM10 and TM01 maintain their orthogonal nature,
but have identical resonant frequencies.
The integer mode index m of TMmn is related to half-cycle variations of the
electric field under the rectangular patch along a. Mode index n is related to the
number of half-cycle electric field variations along b. In the case of the TM10
mode, the electric field is constant across any slice through b (i.e. the
direction), and a single half-cycle variation exists in any cut along a (i.e. the
direction). Figure 2.4 shows a narrow patch driven in the TM01, TM02, and TM03
mode according to cavity model convention.
One notes that the electric field is equal to zero at the center of a rectangular
patch for both the TM10 and TM01 modes. This allows a designer the option of
placing a shorting pin in the center of the rectangular patch without affecting the
generation of either of the two lowest order modes. This added shorting pin or via
forces the groundplane and rectangular patch to maintain an equivalent DC
electrostatic potential. In many cases, the build-up of static charge on a patch
antenna is undesirable from an electrostatic discharge (ESD) point of view, and a
via may be placed in the center of the rectangular patch to address the problem.
Figure 2.7(a) shows the general network model used to represent a rectangular
microstrip antenna. The TM00 mode is the static (DC) term of the series [27]. As
described previously, the TM10 and/or TM01 are two lowest order modes which
are generally driven in most applications. When this is the case, the other higher
order modes are below cut-off and manifest their presence as an infinite number
of small inductors which add in series with the driving point impedance. The
convergent sum of these inductances may be lumped into a single series inductor
which represents the contribution of the higher order modes to the driving point
impedance. As the substrate thickness h of a microstrip patch increases, the
contribution of the equivalent series inductance of the higher order modes to the
driving point impedance becomes larger and larger, which produces a larger and
larger mismatch, until the patch antenna can no longer be matched by simply
choosing an appropriate feedpoint location. The cavity model does not include the
small amount of intrinsic self-inductance introduced by a coaxial feed probe [26].
Increasing the thickness of the substrate also increases the impedance bandwidth
of the element. These two properties (impedance bandwidth and match) may need
to be traded off in a design.
Figure 2.7 Network models used to represent a rectangular microstrip antenna:
(a) general model and (b) narrow band model which is valid for the
TM10 mode

The cavity model is accurate enough to use for many engineering designs. Its
advantage is that it is expressed with closed-form equations, which allow efficient
computation and ease of implementation. Its disadvantage is its accuracy when
compared with more rigorous methods.
The cavity model equations presented previously were implemented for a
rectangular patch antenna with a = 34.29 mm and a resonant length of b = 30.658
mm (TM01). The feed point is 7.595 mm from the center of the patch,
and 7.734. The dielectric thickness is (0.120 inch) with and
(these values are in Table 2.1). The measured maximum return loss of a
patch fabricated using these dimensions is 30.99 dB at 2.442 GHz. The FDTD
method was also used to analyze this patch antenna. The impedance results for the
cavity model, FDTD, and measurement is presented in Figure 2.8. The cavity
model predicts a maximum return loss at 2.492 GHz, which is about a 2% error
versus measurement. FDTD predicts 2.434 GHz which is a 0.33% error. These
resonance values are presented in Table 2.2. The cavity model predicts a larger
bandwidth for the first resonance than is actually measured, and it is fairly good at
predicting the next higher resonance but then deviates significantly. The
groundplane size of the fabricated antenna, also used in the FDTD analysis, is
63.5 mm × 63.5 mm with the dielectric flush to each groundplane edge.

Table 2.1 A 2.45 GHz linear microstrip antenna (groundplane dimensions = 63.5
mm × 63.5 mm)

a b h ϵr tan δ
34.29 mm 30.658 mm 3.048 mm 3.38 0.0027 a/2 7.733 mm

Figure 2.8 Comparison of predicted negative return loss of a rectangular

microstrip patch (of parameters in Table 2.1) by the cavity model and
FDTD to measurement

Table 2.2 Microstrip antenna resonance values

Analysis method Resonance frequency

Cavity model 2.492 GHz
Measurement 2.442 GHz
FDTD 2.434 GHz

2.3 Radiation pattern of a linear rectangular patch

The transmission line model, combined with the measured and computed thermal
plots, suggests a model for the computation of radiation patterns of a rectangular
microstrip patch antenna in the TM01 mode. The fringing fields at the edge of a
microstrip antenna which radiate are centered about each edge of the antenna.
This implies that the radiation pattern would be comparable to a pair of radiating
slots centered about each radiating edge of the patch driven in phase. These slots
can be viewed as equivalent to slots in a groundplane with a uniform electric field
across them. This is illustrated in Figure 2.9(a). In Figure 2.9(b) is an FDTD
thermal plot of the magnitude of the electric field distribution of a microstrip
antenna cut through the – plane. We can see that the two radiating edges, and
the fields which radiate, form a semi-circle about each edge. The electric field
extends outward from each edge along the dielectric substrate about the same
amount as the dielectric thickness.

Figure 2.9 (a) Top view of a rectangular microstrip patch with a pair of
equivalent slots located at a distance a apart. The electric fields
across the slots radiate in phase. (b) Side view FDTD thermal plot of
the electric field for the patch analyzed in Figure 2.8 fed with a square
coaxial cable. This plot demonstrates that the radiating electric fields
 are approximately constant at each radiating edge of the patch and
extend for a distance which is nearly the thickness of the substrate.
(Note that the virtual short circuit at the center of the patch, under the
antenna, is clearly visible)

The radiating slots have a length b and are estimated to be of h (the substrate
thickness) across. The two slots form an array. When the dielectric substrate is air
, the resonant length a is nearly . When a pair of radiation sources
have this spacing in free space, the array produces a maximum directivity. As the
dielectric permittivity increases, the resonant length of the patch along a
decreases, which decreases the spacing between the radiating slots. The slots no
longer optimally add broadside to the rectangular microstrip antenna, which
decreases the directivity and hence increases the pattern beamwidth.
The electric field from a single slot with a voltage across the slot of is
given as [28]:

For two slots spaced at a distance a apart the E-plane radiation pattern is

The H-plane pattern is independent of the slot spacing a and is given by

The angle θ is measured from the -axis and is measured from the -axis.
The directivity of a microstrip antenna can be approximated by starting with
the directivity of a single slot [10]:
 In the case of a microstrip antenna with a pair of radiating slots, the directivity
is

where is the zeroth-order Bessel function with argument x.

where is the radiation resistance,

The integrals of (2.45) and (2.47) may be accurately evaluated numerically

with Gaussian quadrature (Appendix B). The directivity estimates and pattern
functions do not take groundplane effects into account and are often lower than
measured. These equations are very useful for estimating the directivity and
radiation pattern of a rectangular microstrip antenna. It is always best to use a
more powerful technique of analysis such as FDTD or finite element method
(FEM) to refine the pattern prediction of a given design.
Figure 2.10 shows measured E- and H-plane patterns of the 2.45 GHz
microstrip antenna of Table 2.1 plotted with results from the slot pair model, and
results using FDTD. The FDTD method results were obtained using a single
frequency square coaxial source and the patterns calculated using the surface
equivalence theorem [29,30]. One can see the measured and FDTD results are
very similar for the upper hemisphere in both E- and H-plane patterns. Equation
(2.46) was used to compute the directivity for the slot model. The E-plane slot
model pattern results are close for ±45° but begin to deviate at low angles. The H-
plane slot model is close up to about ±60°. The slot model does not take
groundplane affects into account, but is clearly very accurate considering the
simplicity of the model used.
Figure 2.10 Comparison of the measured and predicted radiation pattern of the
2.45 GHz linear microstrip antenna of Table 2.1 using FDTD
analysis, and the slot model for the TM01 mode

The important parameter which determines the directivity of a microstrip

antenna is the relative dielectric permittivity, , of the substrate. When the
substrate is air ( 1.0), the two antenna edges are approximately half of a free
space wavelength apart ( ). This spacing produces an array spacing for the
slot model which produces maximum directivity. It is possible to achieve a
directivity of almost 10 dB with an air loaded rectangular microstrip patch
antenna. As the dielectric permittivity of the substrate is increased, the slots
become closer to each other in terms of free space wavelengths, and no longer
array to produce as high a directivity as in the free space case. As the substrate
dielectric permittivity of a rectangular microstrip antenna increases, the directivity
of a patch antenna decreases. Table 2.3 presents a comparison of the directivity
predicted by the slot model and FDTD for a square microstrip antenna. For low
values of relative dielectric permittivity , the slot model is within
approximately 1 dB. For , the directivity of the slot model is still within
about 1.5 dB. The slot model can be useful for estimating directivity.

Table 2.3 Directivity (dB) of a square linear microstrip antenna versus (2.45
GHz, h = 3.048 mm, tan δ = 0.0005) [square groundplane dimensions
for FDTD = 63.5 mm × 63.5 mm (antenna centered)]

Slot directivity estimate vs. FDTD

ϵr Slot model FDTD
1.0 8.83 8.00
2.6 6.56 7.11
4.1 5.93 6.82
10.2 5.24 6.54
20.0 5.01 6.45

2.4 Quarter-wave microstrip antenna

Understanding the electric field distribution under a rectangular microstrip
antenna allows us to develop useful variations of the original /2 rectangular
microstrip antenna design. In the case where a microstrip antenna is fed to excite
the TM01 mode exclusively, a virtual short circuit plane exists in the center of the
antenna, parallel to the x-axis, centered between the two radiating edges. This
virtual shorting plane can be replaced with a physical metal shorting plane to
create a rectangular microstrip antenna which is half its original length
(approximately ), as illustrated in Figure 2.11. Only a single radiating edge
remains with this design, which reduces the radiation pattern directivity compared
with a half-wavelength patch. This rectangular microstrip antenna design is
Another random document with
no related content on Scribd:
however be of a different kind. The Muses, like other nymphs, were worshipped in
grottoes as guardian genii of fountains, and Pliny, Hist. Nat., XXXVI, 21, writes of
‘erosa saxa in aedificiis, quae musaea vocant, dependentia ad imaginem specus
arte reddendam,’ where the suggestion is of a rustic grotto like that in the Boboli
Gardens. Such grottoes, natural or artificial, might fittingly be decked with shells
and coloured stones and any bright inlay that offered itself. If incrustations of the
kind we call mosaic were actually met with in these haunts of the Muses, the work
might readily be called by a name suggestive of these same nymphs, and this might
be applied later on to tesselated work in general. There is however no proof, either
in Pliny or elsewhere, that what we call mosaic was actually so used, and it has
been questioned by more than one authority whether there is really any connection
between the word ‘mosaic,’ in its various forms, and the Muses. An oriental
derivation has even been suggested for the term.
Dr Albert Ilg, in an exceedingly learned paper on the subject in the Wiener
Quellenschriften, Neue Folge, V, 158 f., offered an entirely new explanation of the
word ‘mosaic,’ which he maintained had in its original sense nothing to do with
inlaid work at all, but rather with gilding. He connected it with a root ‘mus’ or
‘mos,’ with a sense of ‘beating’ or ‘grinding,’ and instanced the mediaeval Latin
term ‘mosnerium,’ which Ducange notices as equivalent to ‘molendinum,’ ‘mill.’
‘Musivum opus’ would refer on this view to the gilding process in which the gold is
ground to powder or beaten out; and Ilg affirmed ‘Musaicum im alten Sinne kann
nur eigentlich Vergoldung, nicht das moderne Mosaik, bezeichnen.’ If the word at
first meant ‘gilded work’ it would later on be extended to what we know as ‘mosaic,’
because of the use in mediaeval mosaics of the familiar gold background. The
argument of Dr Ilg is not convincing, and the question must be considered still
open. Theophilus, for example, Lib. II, c. 12, uses ‘musivum opus’ for inlaid work in
which there is no question of gold.
143. Possibly what we call ‘mother of pearl.’
144. See Note on ‘The Sassi, della Valle, and other Collections,’ etc., postea, p.
102 f. The mosaic here noticed is unfortunately lost. Lanciani, The Golden Days of
the Renaissance in Rome, 1906, p. 234, states that he has searched for it in vain.
145. See Note 5, ante, p. 27.
146. Mosaics made up of small cubes of coloured or gilded glass are abundant
in early Christian and Byzantine times, but were also used, though sparingly, by
the Romans from the time of Augustus downwards. See Pliny, Hist. Nat., XXXVI,
189, who fixes the time of their introduction.
147. Egg-shell mosaic. See Note, postea, p. 136.
148. See Chapters XV and XVI of the ‘Introduction’ to Painting. The pavement
of the cathedral of Siena exhibits a large collection of such mosaics in black and
white executed in different technical processes.
 149. See Note on ‘Ideal Architecture’ at the close of the ‘Introduction’ to
Architecture, postea, p. 138.
150. That is, about 4½ inches.
151. About 15½ inches.
152. See note on ‘The Nature of Sculpture,’ at the close of the ‘Introduction’ to
Sculpture, postea, p. 179.
153. ‘Working from manner.’ Vasari refers here to what artists call ‘treatment,’
which is a process of analysis and grouping, applied to appearances in nature
where the eye sees at first little more than a confused medley of similar forms that
are perhaps constantly changing. Under such an aspect the hair as well as the folds
of drapery on the human figure presented themselves to the early Greek sculptor,
and it was a long time before he learned to handle them aright. In the case of the
hair he had no help in previous work, for in Egyptian statues it is often covered, or
is replaced by a formal wig, and in Assyrian art the hair is very severely though
finely conventionalized. It was not until the age of Pheidias that the Greeks learned
how to suggest the soft and ample masses of the hair, and at the same time to
subdivide these into the distinct curls or tresses, each one ‘solid,’ as Vasari
requires, but individually rendered with the minuter markings which suggest the
structure and ‘feel’ of the material. The Italians started of course with this
treatment or ‘manner’ already an established tradition founded on antique
practice. In the mediaeval sculpture of the thirteenth and fourteenth centuries in
France and England the hair is often very artistically rendered.
154. This paragraph opens up a subject of much artistic interest, on which see
Note on ‘Sculpture Treated for Position,’ at the close of the ‘Introduction’ to
Sculpture, postea, p. 180 f.
155. For Vasari, a practical artist, to commit himself to the statement that
figures are made nine heads high, is somewhat extraordinary, for eight heads, the
proportion given by Vitruvius (III, 1) is the extreme limit for a normal adult, and
very few Greek statues, let alone living persons, have heads so small. The recently
discovered ‘Agias’ by Lysippus, at Delphi, is very nearly eight heads high. The
‘Doryphorus’ at Naples not much more than seven. The ‘Choisseul Gouffier Apollo’
about seven and a half, etc. Vasari seems to have derived his curious mode of
reckoning from Filarete, who in Book 1 of his Treatise on Architecture measures a
man as follows: Head = 1 head, neck = ½, breast = 1, body = 2, thighs = 2, legs = 2,
foot = ½, total nine heads. Alberti, Leonardo, Albrecht Dürer, and indeed almost
all the older writers on art, discourse on the proportions of the human figure.
156. See Note on ‘Waxen Effigies and Medallions,’ at the close of the
‘Introduction’ to Sculpture, postea, p. 188.
157. One objection to an armature of wood is that the material may swell with
the damp of the clay and cause fissures. Iron is objectionable because the rust
discolours the clay. Modern sculptors often use gas-piping in the skeletons of their
models, as this is flexible and will neither rust nor swell.
158. Baked flour used to be employed by plasterers to keep the plaster they
were modelling from setting too rapidly. See the Introduction by G. F. Robinson to
Millar’s Plastering Plain and Decorative, London, 1897. The former used rye
dough with good effect for the above purpose.
159. The tow or hay tied round the wood affords a good hold for the clay,
which is apt to slip on anything smooth.
160. This method of producing drapery is not very artistic.
161. See Note on ‘Proportionate Enlargement’ at close of the ‘Introduction’ to
Sculpture, postea, p. 190.
162. See Note on ‘The Use of Full-sized Models’ at the close of the
‘Introduction’ to Sculpture, postea, p. 192.
163. The carvers’ tools described by Vasari are the same that appear to have
been in use in ancient Greece (see the article by Professor E. Gardner already
referred to), that are figured in the Encyclopédie of the eighteenth century, and are
now in use. Fig. 2, E to J, ante, p. 48, shows a set of them actually employed in a
stone carver’s workshop at Settignano near Florence.
164. Actual polish of the surface of a marble figure is to be avoided, as the
reflections from it where it catches the light destroy the delicacy of the effect of
light and shade. Greek marbles were not polished, save in some cases where the
aim seems to have been to imitate the appearance of shining bronze, but the
Greeks finished their marbles more smoothly than the sculptors of to-day, most of
whom prefer a ‘sensitive’ surface on which the marks of the last delicate chiselling
can be discerned. Michelangelo’s Dead Christ in the ‘Pietà’ of St. Peter’s, his most
finished piece of marble work, may almost be said to show polish, and Renaissance
marbles generally are quite as smoothly finished as antiques. In the case of
coloured marbles, used for surface decoration in plain panels, polish is of course
necessary in order that the colour and veining may appear, but it does not follow
from this that a self-coloured marble, carved into the similitude of a face or figure,
should be polished.
165. English terminology for the different kinds of reliefs, and for sculpture
generally, is very deficient, and many Italian terms are employed. It may be noted
that Vasari’s ‘half relief’ (mezzo rilievo) is the highest kind he mentions, and would
correspond to what is called in English ‘high relief.’
166. See Note on ‘Italian and Greek Reliefs,’ at the close of the ‘Introduction’
to Sculpture, postea, p. 196.
167. Donatello’s flat, or ‘stiacciati’ reliefs are deservedly famous. The difficulty
here is to convey the impression of solid form of three dimensions with the
slightest possible actual salience. The treatment of the torso of the Christ in the
marble ‘Pietà’ of the Victoria and Albert Museum is a good example.
 168. The antique vessels of so-called ‘Arezzo’ ware are called Aretine vases.
Messer Giorgio was in duty bound to take some note of the ancient pottery of his
native city for it was from this that the Vasari derived their family name. According
to the family tree given in a note to the Life of an ancestor of the historian (Opere,
ed. Milanesi, II, 561), the family came from Cortona, and the first who settled in
Arezzo was the historian’s great-grandfather, one Lazzaro, an artist in ornamental
saddlery. He had a son, Giorgio, who practised the craft of the potter, and was
especially concerned with the old Roman Aretine vases the technique of which he
tried to reproduce. Hence he was called ‘Vasajo,’ ‘the vase maker,’ from which
came the family appellation Vasari.
This ancient Aretine ware ‘must be regarded as the Roman pottery par
excellence’ (Waters, History of Ancient Pottery, Lond., 1905, II, 480). It is
practically the same ware that is known by the popular but unscientific term
‘Samian,’ and consists in cups and bowls and dishes usually of a small size of a fine
red clay, ornamented with designs in low relief, produced by the aid of stamps or
moulds. It is these relief ornaments that Vasari had in his mind when he wrote the
words in the text. Arezzo is noticed by Pliny and other ancient writers as a great
centre for the fabrication of this sort of ware, and Vasari tells us how his
grandfather, Giorgio the ‘vasajo,’ discovered near the city some kilns of the ancient
potters and specimens of their work. Very good specimens of Aretine ware are to
be seen in the Museum at Arezzo, and the fabrique is represented in all important
collections of ancient pottery.
169. See Note on ‘The Processes of the Bronze Founder’ at the close of the
‘Introduction’ to Sculpture, postea, p. 199, which the reader who is unacquainted
with the subject, will find it useful to read forthwith. The best commentary on
Vasari’s and Cellini’s account of bronze casting is to be found in the French
Encyclopédie, where there is a description, with numerous illustrations, of the
casting in 1699 of Girardon’s great equestrian statue of Louis XIV, destined for the
Place Vendôme. It was claimed at the time to be the largest known single casting in
the world, and represents in their utmost elaboration the various processes
described by Vasari. Some of the illustrations are here reproduced, and will help to
render clearer the descriptions in the text.
170. Plate VII shows a section or two of a piece-mould round a portion of a
figure. It will be noticed that the pieces are so planned that they will all come away
easily from the model and not be held by any undercut projections. The small
pieces are then all enclosed in an outer shell divided into two halves, and called in
French ‘chape’ answering to the ‘cappa’ of Vasari’s text. Plate VIII, A, shows the
model of the Louis XIV statue as piece-moulded.
171. In the case of a heavy casting such an armature is necessary, and must be
carefully constructed to give support at all points. The armature within the core of
the horse of Louis XIV is shown in Plate VIII, D.
172. Vasari here describes a method of constructing the indispensable shell of
wax which is to be replaced by the bronze. The hollow piece-mould is lined section
by section with wax and a core is then formed to fill the rest of the interior and
touch the inner surface of the wax at every point. The plaster mould is then
removed and the wax linings of each of its sections are applied, each in its proper
place, to the core, and fixed thereon by skewers. There is then a complete figure in
wax, but, as this is made up of very many pieces, it has to be gone over carefully to
smooth over the joins and secure unity of surface. Cellini’s plan seems a better one.
He lines his hollow mould with a sort of paste or dough, and then fills up with the
core. The dough is then removed and wax is poured in in its place, thus forming a
continuous skin and securing a more perfect unity in the waxen shell.
173. On Plate VIII at B we see the core covered with the skin of wax and
carefully gone over and finished in every part. The system of pipes with which it is
covered are the ‘vents’ that Vasari notices in § 62, and also the channels through
which the melted wax is to escape and the molten bronze to enter, as noticed in §§
63, 64.
174. Vasari actually says that it must be put ‘al fuoco’ ‘to the fire,’ but it is clear
that he does not mean that heat is at once to be applied to it. If this were done the
wax would all be melted off the core too soon, before it was covered by the outer
skin. It is only when the wax has been securely enclosed between the core and the
outer skin that heat is needed to melt it away and leave its place free for the molten
metal.
175. Plate VIII, C, shows this outer armature, with the ends of the transverse
rods holding core and envelope together.
176. ‘Give passage to the metal.’ Their essential purpose is to allow for the
escape of air which would be dangerous if driven by the metal into a confined
space.
177. It should be understood that, in the process Vasari has in mind, the
melted metal is introduced at the bottom of the mould so as to rise in it and expel
before it the air. It is not poured in at the top. Hence the metal enters at the same
orifice at which the wax flows out.
178. Plate VIII, D, gives a section through the model in the casting-pit, when
all is ready for the actual operation of introducing the molten metal. The wax has
all been run out, and the outline of the figure and of the horse is marked by a
double line with a narrow space between. It is this space that will be filled by the
bronze which will be introduced through numerous channels so that it may be
distributed rapidly and evenly over the whole surface it is to cover. When in the pit
the mould is packed all round with broken bricks or similar material, so that ‘the
bronze may not strain it,’ nor cause it to shift.
179. The wax has already been carefully weighed, and in order to estimate how
much bronze will be required for the cast a rough calculation is made based on the
amount of wax.
 180. The subject of the composition of bronze and of other alloys of copper is a
complicated one, for the mixtures specified or established by analysis are very
varied. Normally speaking, bronze is a mixture of copper with about ten per cent.
of tin, brass of copper with twenty to forty per cent. of zinc. Vasari’s proportions
for bells and for cannon are pretty much what are given now. In the Manuel de
Fondeur (Manuels Roret) Paris, 1879, II, p. 94, eight to fifteen per cent. of tin are
prescribed for cannon, fifteen to thirty per cent. for bell metal, the greater
percentage of tin with the copper resulting in a less tough but harder and so
sharper sounding metal. It will be noted however that for statuary metal Vasari
specifies a mixture not of copper and tin but of copper and brass, that is, copper
and zinc. Brass is composed of, say, twenty-five per cent. of zinc and seventy-five
per cent. of copper, so that a mixture of two thirds, or sixty-six per cent., of copper
with one third, or thirty-three per cent., of brass would work out to about ten parts
of zinc to ninety of copper, and this agrees with classical proportions. The Greeks
used tin for their bronzes, but various mysterious ingredients were supposed to be
mingled in to produce special alloys. The Romans used zinc, or rather zinciferous
ores such as calamine, with or in place of tin, and this is the tradition that Vasari
follows.
A recent analysis of the composition of the bronze doors at Hildesheim, dating
from 1015 A.D., gives about seventy-six parts copper, ten lead, eight tin, four zinc;
and of the ‘Bernward’ pillar ascribed to about the same date, seventy copper,
twenty-three tin, and five lead. These differences may surprise us, but metal
casting in those days was a matter of rule of thumb, and we may recall Cellini’s
account of his cramming all his household vessels of pewter into the melting pot to
make the metal flow for casting his ‘Perseus.’
181. Vasari’s account of the making of dies for medals and of the process of
striking these is clear, and agrees with the more elaborate directions contained in
the seventh and following chapters of Cellini’s Trattato dell’ Oreficeria. Cellini
however, unlike Vasari, was a practical medallist, and he goes more into detail. The
process employed was not the direct cutting of the matrices or dies with chisels,
nor, as gems are engraved, by the use of the wheel and emery (or diamond)
powder, but the stamping into them of the design required by main force, by
means of specially shaped hard steel punches on which different parts of the design
had been worked in relief. The steel of the matrix or die had of course to be
previously softened in the fire, or these punches would have made no impression
on it. When finished it was again hardened by tempering. It may be noticed that
the dies from which Greek coins were struck were to all appearance engraved as
gems were engraved by the direct use of cutting tools or tools that, like the wheel,
wore away the material with the aid of sand or emery.
The two matrices, or dies, for the obverse and reverse of the medal, being now
prepared, the medal is not immediately struck. In the case of the Greek coin a
bean-shaped piece, or a disk, of plain metal, usually of silver, called a ‘blank’ or
‘flan,’ was placed between the two dies and pressed into their hollows by a blow or
blows of the hammer, so that all that was engraved on them in intaglio came out on
the silver in relief. Vasari’s process is more elaborate. A sort of trial medal is first
struck from the matrices in a soft material such as lead or wax, and this trial medal
is reproduced by the ordinary process of casting in the gold or silver or bronze
which is to be the material of the final medal. This cast medal has of course the
general form required, but it is not sharp nor has it a fine surface. It is therefore
placed between the matrices and forcibly compressed so as to acquire all the finish
of detail and texture desired.
182. Plaster, or stucco, is sometimes regarded as an inferior material only to
be used when nothing better can be obtained. It should not however be judged
from the achievements of the domestic plasterer of to-day, who has to trust
sometimes to the wall-paper to keep his stuff from crumbling away. Plaster as used
by the ancients, and through a good part of the mediaeval and Renaissance periods
up to the eighteenth century, is a fine material, susceptible of very varied and
effective artistic treatment. It was made by the Greeks of so exquisite a quality that
it was equivalent to an artificial marble. It could be polished, so Vitruvius tells us,
till it would reflect the beholder’s face as in a mirror, and he describes how the
Roman connoisseurs of his time would actually cut out plain panels of Greek
stucco from old walls and frame them into the plaster work of their own rooms,
just as if they were slabs of precious marble. (De Architectura, VII, iii, 10.) Vitruvius
prescribes no fewer than six successive coats of plaster for a wall, each laid on
before the last is dry, the last coat being of white lime and finely powdered marble.
By the Villa Farnesina at Rome some Roman, or more probably Greek, plaster
decoration was discovered a few years ago that surpassed any work of the kind
elsewhere known. We find there the moulded or stamped ornament Vasari
describes, as well as figure compositions modelled by hand, while the plain
surfaces are in themselves a delight to the artistic eye.
Among the best and best known stucco work, in figures and ornaments, of the
later Italian Renaissance, may be ranked that at Fontainebleau by Primaticcio and
other artists from the peninsula who were invited thither by François I, for the
decoration of the ‘Galerie François I’ and the ‘Escalier du Roi.’
183. The composition of these two mucilages is given by Theophilus, in the
Schedula, Book one, chapter 17, and also by Cennini, Trattato, chapters 110–112.
Soft cheese from cows’ milk must, according to the earlier recipe, be shredded
finely into hot water and braised in a mortar to a paste. It must then be immersed
in cold water till it hardens, and then rubbed till it is quite smooth on a board and
afterwards mixed with quick lime to the consistency of a stiff paste. Panels
cemented with this, says Theophilus, will be held so fast when they are dry that
neither moisture nor heat will bring them apart. Vasari does not seem to have such
faith in the mucilage, and prefers that made from boiling down shreds of
parchment and other skins. The twelfth century writer knows how to make this
also. See chapter eighteen of the first Book of the Schedula.
 184. Every museum contains examples of these delicate German carvings in
hard materials.
185. In a Note to the ‘Introduction’ to Architecture, ante, p. 128 f., an account
was given of some sculptures in travertine on the façade of the church of S. Luigi
dei Francesi at Rome by a ‘Maestro Gian’ who has been conjecturally identified as a
certain Jean Chavier or Chavenier of Rouen who worked at Rome in the first
quarter of the sixteenth century. Vasari in this place introduces an artist of the
name of ‘Maestro Janni francese,’ and the question at once arises whether he is the
same person as the ‘Maestro Gian’ of Rome.
The statue here described is to be seen in the church of the Annunziata at
Florence, but not where Vasari saw it. It has been placed for about the last half
century in the spacious round choir, where it occupies a niche in the wall of the
second chapel to the left as one faces the high altar. It has been painted white in
the hope that it may be mistaken for marble, and this characteristic performance
dates from about 1857. Certain fissures observable show however that it is of wood,
and one of the Frati remembers it when it was as Vasari saw it ‘nello stesso colore
del legname.’ The work is shown on Plate IX. We have been unable to discover
anything certain about the artist. The figure, which is in excellent preservation,
speaks for itself. The Saint has a tight fitting cap over his head and curling hair and
beard. His eyes are almost closed as he looks down with a somewhat affected air at
his wounded leg to which the finger of his right hand is pointing. The other hand
holds a staff, round which the drapery curls and over the top of which it is caught.
This drapery bears out Vasari’s description of it as ‘traforato’ ‘cut into.’ It is floridly
treated with the sharp angles common in the carving of the fifteenth and sixteenth
centuries in Germany, Flanders, and parts of France. M. Marcel Reymond, who has
kindly given his opinion on the photographs submitted to him, has written about it
as follows: ‘Le St. Roch, par la surcharge de vêtements, l’excès de reliefs, l’agitation
des draperies, se rattache à l’art français tel qu’il s’était constitué au xivme siècle, et
tel qu’il s’était continué jusqu’au xvime siècle, notamment dans le Bourgogne et la
Champagne.’ He does not consider the two ‘Maîtres Jean’ the same person. ‘Ce
sont sans doute deux artistes du xvime siècle, l’un travaillant la pierre, le travertin,
l’autre travaillant le bois. C’est leur aptitude à travailler ces deux matières, que les
artistes italiens travaillaient moins bien que les français qui a retenu l’attention de
Vasari sur eux et qui leur a fait attribuer une place si importante dans les préfaces
de Vasari.’ Our study of the originals at Rome and Florence has led us to the same
opinion. The S. Rocco is Gothic in feeling, the ‘Salamander’ and other pieces at
Rome are Renaissance. The Roman ‘Maestro Gian’ may be credited with an Italian
style, but Vasari does not show much critical acumen when he sees ‘la maniera
italiana’ in the S. Rocco of the Florentine Janni.
186. The first two sections, §§ 74, 75, of this chapter were added by Vasari in
the second edition. They contain his contribution to the philosophy of the graphic
art. It will be noted that his word ‘Disegno’ corresponds alike to our more general
word ‘design’ and the more special term ‘drawing.’
 187. This remark of Vasari is significant of the change in architectural practice
between the mediaeval and modern epochs. That the architect is a man that sits at
home and makes drawings, while practical craftsmen carry them out, is to us a
familiar idea, but the notion would greatly have astonished the builders of the
French Gothic cathedrals or the Florentines of the fourteenth century. In
mediaeval practice the architect was the master of the work, carrying the scheme of
the whole in his head, but busy all the time with the actual materials and tools, and
directing progress rather from the scaffolding than from the drawing office. On the
tombstone of the French architect of the thirteenth century, Hughes Libergier, at
Reims, he is shown with the mason’s square, rule, and compasses about him; while
in the relief that illustrates ‘Building’ on Giotto’s Campanile at Florence we see the
master mason directing the operations of the journeymen from a position on the
structure itself. In the present day there is a strong feeling in the profession that
this separation of architect and craftsman, which dates from the later Renaissance,
is a bad thing for art, and that the designer should be in more intimate touch with
the materials and processes of building.
188. It is characteristically Florentine to regard painting as essentially the
filling up of outlines, and to colour in staccato fashion with an assorted set of tints
arranged in gradation. To the eye of the born painter outlines do not exist and
nature is seen in tone and colour, while colours are like the tones of a violin infinite
in gradation, not distinct like the notes of a piano. With the exception of the
Venetians and some other North Italians such as Correggio and Lotto, the Italians
generally painted by filling outlines with local tints graded as light, middle, and
dark, and the Florentines were pre-eminent in the emphasis they laid on the well-
drawn outline as the foundation of the art. Since the seventeenth century the
general idea of what constitutes the art of painting has suffered a change and
Vasari’s account of Florentine practice, in which he was himself an expert, is all the
more interesting. Vasari’s point of view is that of the frescoist. In that process,
which, as we shall see, had to be carried out swiftly and directly so as to be finished
at one sitting, it was practically necessary to have the various tints in their
gradations mixed and ready to hand. The whole method and genius of oil painting,
as moderns understand it, is different, and its processes much more varied and
subtle.
189. The innumerable sketches and finished drawings that have come down to
us from the hands of Florentine artists testify to the importance given in the school
to preliminary studies for painting, and any collection will furnish examples of the
different methods of execution here described. Drawings by Venetian masters, who
felt in colour rather than in form, are not so numerous or so elaborate.
190. That is to say, by observation of aerial as well as linear perspective.
191. This practice is noticed in the case of more than one artist of whom Vasari
has written the biography. Tintoretto is one. See also postea, p. 216.
192. See the Note on ‘Fresco Painting’ at the close of the ‘Introduction’ to
Painting, postea, p. 287.
 193. Michelangelo’s greatest tour de force in foreshortening, much lauded by
Vasari in his Life of the master, is the figure of the prophet Jonah on the end wall
of the Sistine chapel. It is painted at the springing of the vault, on a surface that is
inclined sharply towards the spectator, but the figure is so drawn as to appear to be
leaning back in the opposite direction.
194. Correggio is responsible for many of the forced effects of drawing in the
decorative painting of vaults and ceilings in later times, but the Umbrian Melozzo
da Forlì in his painting of the Ascension of Christ, now destroyed save for the
fragments in the Quirinal and in the sacristy of St. Peter’s at Rome, may have the
doubtful honour of beginning the practice of foreshortening a whole composition,
so that the scene is painted as it would appear were we looking up at it from
underneath.
195. This truth, about the mutual influence of colours in juxtaposition, was
well put by Sir Charles Eastlake when he wrote, in his Materials for a History of
Oil Painting, ‘flesh is never more glowing than when opposed to blue, never more
pearly than when compared with red, never ruddier than in the neighbourhood of
green, never fairer than when contrasted with black, nor richer or deeper than
when opposed to white.’
196. Vitruvius describes the fresco process in his seventh Book. See Note on
‘Fresco Painting’ at the end of the ‘Introduction’ to Painting, postea, p. 287. This
chapter is one of the most interesting in the three ‘Introductions.’
197. Travertine, next to marble, makes when burnt the whitest lime (see § 30,
ante, p. 86). From this lime the fresco white, called bianco Sangiovanni, is made,
and Cennini gives the recipe for its preparation in his 58th chapter. The ordinary
lead white (biacca) cannot be used in fresco.
198. The word ‘tempera’ is used by Vasari and other writers as a noun
meaning (1) a substance mixed with another, as a medium with pigments (2) a
liquid in which hot steel is plunged to give it a particular molecular quality (ante, p.
30) (3) the quality thus given to the steel (ante, p. 32), while (4) it has come to
mean in modern times, as in the heading of this Note, a particular kind of painting.
It is really to be regarded as the imperative of the verb ‘temperare,’ which alike in
Latin and in Italian means ‘to divide or proportion duly,’ ‘to qualify by mixing,’ and
generally ‘to regulate’ or ‘to discipline.’ ‘Tempera’ thus means strictly ‘mix’ or
‘regulate.’ It is used in the latter sense in metallurgy, as the liquid which Vasari
calls (ante, p. 30) a ‘tempera’ (translated ‘tempering-bath’) regulates the amount of
hardness or elasticity required in the metal, and the quality the steel thus receives
is called (ante, p. 32) its ‘temper.’ In the case of painting the ‘tempera’ is the
binding material mixed with the pigment to secure its adhesion to the ground
when it is dry. The painting process is, in Italian, painting ‘a tempera’ ‘with a
mixture,’ and our expression ‘tempera painting’ is a loose one. For the form of the
word we may compare ‘recipe,’ also employed as a substantive but really an
imperative meaning ‘take.’
 Strictly speaking any medium mixed with pigments makes the process one ‘a
tempera.’ Many substances may be thus used, some soluble in water, as size, gum,
honey, and the like; others insoluble in water, such as drying oils, varnishes, resins,
etc., while the inside of an egg which is in great part oleaginous may have a place
between. It is not the usage however to apply the term ‘tempera’ to drying oils or
varnishes, and a distinction is always made between ‘tempera painting’ and ‘oil
painting.’ See Note on ‘Tempera Painting,’ postea, p. 291.
 199. This practice of covering wooden panels with linen and laying over this
the gesso painting ground was in use in ancient Egypt. In fact the methods
described by Cennini of preparing and grounding panels are almost exactly the
same as those used in ancient Egypt for painting wooden mummy-cases. Even the
practice, so much used in early Italian art, of modelling details and ornaments in
relief in gesso and gilding them, is common on the mummy-cases. On the subject
of gesso see Note 5 on p. 249.
200. Vasari’s expression ‘rosso dell’ uovo o tempera, la quale è questa’ calls
attention to the fact, to which his language generally bears testimony, that he
looked upon the yolk of egg medium as the tempera par excellence. When he uses
the term ‘tempera’ alone he has the egg medium in his mind, and the size medium
is something apart. See this chapter throughout.
201. Tempera painting has had a far longer history and more extensive use
than any other kind. The technique predominated for all kinds of painting among
the older Oriental peoples and in classical lands, and was in use both on walls and
on panels in Western Europe north of the Alps during the whole mediaeval period,
while south of the Alps and at Byzantium it was to a great extent superseded for
mural painting by fresco, but remained in fashion for panels till the end of the
fifteenth century. After the fifteenth century the oil medium, as Vasari remarks,
superseded it entirely for portable pictures, and partly for work on walls and
ceilings, but in our own time there has been a partial revival of the old technique.
See Note on ‘Tempera Painting,’ postea, p. 291.
The whole question of the different vehicles and methods used in painting at
various periods is a difficult and complicated one, and too often chemical analysis
fails to give satisfactory results owing to the small amount of material available for
experiment. Berger, in his Beiträge zur Entwicklungs-Geschichte der Maltechnik,
an unfinished work that has already run to a thousand pages, goes elaborately into
the subject, but has to admit that many points are still doubtful. It makes
comparatively little difference what particular medium is used in tempera
painting, but it is of great importance to decide whether a particular class of work
is in tempera or in fresco. In connection with this Berger has reopened the old
controversy as to the technique of Pompeian wall paintings, which have been
accepted as frescoes, on the authority of Otto Dönner, for a generation past. There
are difficulties about Pompeian work and it is well that the question has again been
raised, but Berger goes much too far when he attempts to deny to the ancients the
knowledge and use of the fresco process. The evidence on this point of Vitruvius is
quite decisive, as he, and Pliny after him, refer to the process of painting on wet
plaster in the most unmistakeable terms. See Note on ‘Fresco Painting, postea, p.
287.
202. This passage about the early painters of Flanders occurs just as it stands,
with some trifling verbal differences, in Vasari’s first edition of 1550. The best
commentary on it is, first, the account of the same artists in Guicciardini’s
Descrittione di Tutti i Paesi Bassi, first published at Antwerp in 1567, and next,
Vasari’s own notes on divers Flemish artists which he added at the end of the Lives
in the second edition of 1568 (Opere, ed. Milanesi, VII, 579 f.). He there made
certain additions and corrections from Guicciardini, the most noteworthy of which
is the mention of Hubert van Eyck, whom Vasari ignores in this passage of the
Introduction, but who is just referred to by Guicciardini at the end of his sentences
on the younger brother—‘A pari a pari di Giovanni andava Huberto suo fratello, il
quale viveva, e dipingeva continuamente sopra le medesime opere, insieme con
esso fratello.’ Vasari however in the notes of 1568 goes much farther than this, and,
though he does not call Hubert the elder brother, he seems to ascribe to him
personally the supposed ‘invention’—‘Huberto suo fratello, che nel 1510 (sic) mise
in luce l’ invenzione e modo di colorire a olio’ (Opere, l.c.). ‘John of Bruges’ is of
course Jan van Eyck. Vasari writes of him at the end of the Lives as ‘John Eyck of
Bruges.’ Vasari’s statement in this sentence is of great historical importance, for it
is the first affirmation of a definite ‘invention’ of oil painting, and the first
ascription of this invention to van Eyck. As van Eyck’s own epitaph makes no
mention of this, and as oil painting was practised long before his time, Vasari’s
statement has naturally been questioned, and on the subject the reader will find a
Note at the close of the ‘Introduction’ to Painting, postea, p. 294.
203. It was long supposed that this picture was the ‘Epiphany’ preserved
behind the High Altar of the Church of S. Barbara, Naples, but Crowe and
Cavalcaselle, History of Painting in North Italy, II, 103, pronounce this ‘a feeble
and injured picture of the eighteenth century.’
204. Frederick of Urbino (there were not two of the name as Vasari supposes)
seems to have had a bathroom decorated with secular compositions by the Flemish
master. Facio, whose tract De Viris Illustribus, written in the middle of the
fifteenth century, was printed at Florence in 1745, writes, p. 46, of ‘Joannes
Gallicus’ (who can be identified as Jan van Eyck) who had painted certain ‘picturae
nobiles’ then in the possession of Cardinal Octavianus, with ‘representations of fair
women only slightly veiled at the bath.’ Such pictures were considered suitable
decorations for bath chambers. There is a curious early example of mediaeval date
in the Schloss Runkelstein near Botzen in the Tyrol, in the form of wall paintings
round a bathroom on one side of which nude figures are seen preparing to enter
the water, while on two other walls spectators of both sexes are seen looking in
through an open arcade. The pictures here referred to by van Eyck are now lost,
but by a curious coincidence attention has just been directed to an existing copy of
one of them, of which Facio gives a special notice. The copy occurs in a painting by
Verhaecht of Antwerp, 1593–1637, that represents the picture gallery of an
Antwerp connoisseur at about the date 1615. There on the wall is seen hanging the
van Eyck, that corresponds closely to the full description given by Facio. The
painting by Verhaecht was shown at Burlington House in the Winter Exhibition,
1906–7, and in the ‘Toison d’Or’ Exhibition at Bruges in 1907. See also the
Burlington Magazine, February, 1907, p. 325. It may be added that the Cardinal
Octavianus mentioned above was a somewhat obscure prelate, who received the
purple from Gregory XII in 1408.
 205. The latest editors of Vasari (Opere, ed. Milanesi, I, 184) think this may be
a picture in the Museum at Naples, ascribed there to an apocryphal artist
‘Colantonio del Fiore.’ Von Wurzbach says it is by a Neapolitan painter influenced
by the Flemings.
206. Roger van der Weyden, more properly called, as by Guicciardini and by
Vasari in 1568, ‘Roger of Brussels.’ In 1449 he made a journey to Italy, and stayed
for a time at Ferrara, which under the rule of the art-loving Este was very
hospitable to foreign craftsmen. He was in Rome in 1450 and may have visited
Florence and other centres. His own style in works subsequent to this journey
shows little of Italian influence.
207. Hans Memling. ‘No Flemish painter of note,’ remark Crowe and
Cavalcaselle, Early Flemish Painters, p. 256, ‘produced pictures more attractive to
the Italians than Memling.’ The Portinari, for whom Memling worked, were
Florentine merchants who had a house at Bruges, the commercial connection of
which with Tuscany was very close. In his Notes on Flemish Painters at the end of
the Lives, Vasari says that the subject of ‘a small picture in the possession of the
Duke’ which is probably the one here mentioned, was ‘The Passion of Christ.’ If
this be the case, it cannot be the beautiful little Memling now in the Uffizi, No. 703,
for the subject of this is ‘The Virgin and Child.’ It might possibly however be the
panel of ‘The Seven Griefs,’ a Passion picture in the Museum at Turin. On the other
hand, Passavant thought the Turin panel was the ‘Careggi’ picture that Vasari goes
on to mention. See Note on p. 268 of Crowe and Cavalcaselle’s work.
208. The German editors of Vasari identified Lodovico da Luano with the
well-known painter Dierich Bouts of Louvain, but the name Ludovico (Chlodwig,
‘Warrior of Renown’) is not the same etymologically as Dierich (Theodoric, ‘Prince
of the People’). It is to be noted that in Guicciardini we find a mention of ‘Dirich da
Louano,’ who is undoubtedly Dierich Bouts (the surname is derived from St.
Rombout the patron of Haarlem, where the painter, who is also called ‘Dirick van
Haarlem’ [see below], was born) and also a mention of Vasari’s ‘Ludovico da
Luvano.’ A scrutiny however of the sentence in Guicciardini, where the last-
mentioned name occurs, shows that it is copied almost verbatim from our text of
Vasari. (Vasari [1550]:—‘Similmente Lodovico da Luano & Pietro Christa, &
maestro Martino, & ancora Giusto da Guanto, che fece la tavola della comunione
de’l Duca d’ Vrbino, & altre pitture; & Vgo d’ Anuersa, che fe la tauola di Sancta
Maria Nuoua di Fiorenza’; Guicciardini:—‘Seguirono a mano a mano Lodouico da
Louano, Pietro Crista, Martino d’ Holanda, & Giusto da Guanto, che fece quella
nobil’ pittura della comunione al Duca d’ Vrbino, & dietro a lui venne Vgo d’
Anuersa, che fece la bellissima tauola, che si vede a Firenze in santa Maria nuoua’).
Vasari is accordingly responsible for this ‘Ludovico da Luano,’ whose name is duly
chronicled in von Wurzbach’s ‘Niederländisches Künstler-Lexicon, Leipzig, 1906,
II, p. 69, on the authority of Guicciardini alone, and who is called in M. Ruelens’s
annotations to the French edition of Crowe and Cavalcaselle ‘Louys de Louvain
(peintre encore inconnu).’ Subsequently Guicciardini mentions also a ‘Dirich d’
Harlem,’ who can be none other than the same Dierick Bouts, and Vasari, as a
return favour, copies back all three Diericks into his Notes at the end of the edition
of 1568. The first ‘Ludovico’ may be merely due to a mistake in the text of Vasari
carelessly adopted by Guicciardini. Vasari’s copyist may have written ‘Ludovico’ in
place of the somewhat similar ‘Teodorico.’ There was however a certain Ludovicus
Dalmau or Dalman (D’Alamagna?), a Flemish painter who worked at Barcelona in
Spain about 1445 (von Wurzbach, sub voce) who may be meant, though there is no
indication of a connection between him and Louvain.
209. Pietro Crista is of course Petrus Christus or Christi of Bruges, an imitator,
though as Mr Weale has shown not an actual pupil, of the van Eycks. Von
Wurzbach says that Guicciardini was the first to mention his name, but Vasari in
1550 already knows him. As an explanation of the surname it has been suggested
that the artist’s father may have had a reputation as a painter or carver of Christ-
figures, so that Petrus would be called ‘son of the Christ-man.’
210. The name Martin belongs to painters of two generations in Ghent, and
von Wurzbach thinks it is the earlier of these, Jan Martins, apparently a scholar of
the van Eycks, who is referred to here, and called by Guicciardini (see above), and
by Vasari in 1568, ‘Martino d’ Holanda.’ There was a later and better known Martin
of Ghent called ‘Nabor Martin.’ The more famous ‘Martins,’ ‘of Heemskerk,’ and
‘Schongauer,’ when referred to by Vasari, have more distinct indications of their
identity. See, e.g., Opere, V, 396.
211. Justus of Ghent worked at Urbino, where he finished the altar piece
referred to by Vasari in 1474. The ‘other pictures’ may be a series of panels painted
for the library at Urbino, on which Crowe and Cavalcaselle have an interesting
paragraph, op. cit. p. 180.
212. Hugo of Antwerp is Hugo van der Goes, whose altar piece painted for S.
Maria Nuova at Florence has now been placed in the Uffizi.
213. Vasari’s stories about the connection with oil painting of Antonello da
Messina, Domenico Veneziano, and Andrea dal Castagno have of course been
subjected to a good deal of hostile criticism. Those about the two latter artists are
in the meantime relegated to the limbo of fable, but the case of Antonello da
Messina is somewhat different, and we are not dependent in his case on Vasari
alone. He certainly did not visit Flanders in the lifetime of Jan van Eyck, for this
artist died before Antonello was born, but von Wurzbach accepts as authentic a
visit on his part to Flanders between 1465 and 1475, and sees evidence of what he
learned there in his extant works (Niederländisches Künstler-Lexicon, sub voce,
‘Antonello’).
214. ‘Terre da campane,’ ‘bell earths.’ There seem to be two possible meanings
for the phrase. It may refer to the material used for the moulds in bell casting, or to
the clay from which are made the little terra-cotta bells by which children in Italy
set great store on the occasion of the mid-summer festival. This last is improbable.
Baldinucci, Vocabolario del Disegno, sub voce ‘Nero di Terra di Campana,’
says that this is a colour made out of a certain scale that forms on moulds for
casting bells or cannon, and that it is good with oil, but does not stand in fresco.
Lomazzo also mentions the pigment.
215. ‘L’abbozza’ evidently refers to the first or underpainting, not to the sketch
in chalk, for in the first edition the passage has some additional words which make
this clear. They run as follows: ‘desegnando quella: e così ne primi colori l’abozza,
il che alcuni chiamono imporre.’
216. With the above may be compared ch. 9 of Book VII of L. B. Alberti’s De Re
Aedificatoria.
217. The matter in our § 87 was added in the edition of 1568. Though Vasari
declared so unhesitatingly for fresco as the finest of all processes of painting, he
tells us that he used oil for a portion of his mural work in the Palazzo Vecchio at
Florence, when he prepared it for the residence of Duke Cosimo, and we shall
notice later his praise of tempera (postea, p. 291). Vasari describes how he painted
in oil on the walls of a refectory at Naples (Opere, VII, 674), and gives us an
interesting notice of his experiments in the technique about the year 1540 at the
monastery of the Camaldoli, near Arezzo, where he says ‘feci esperimento di unire
il colorito a olio con quello (fresco) e riuscimmi assai acconciamente’ (Opere, VII,
667). The technique required proper working out, for it was not a traditional one.
The most notable instance of its employment before the end of the fifteenth
century is in the case of the ‘Last Supper’ by Leonardo da Vinci at Milan. A
commission of experts has recently been examining the remains of this, the most
famous mural painting in the world, and has ascertained that the original process
employed by Leonardo was not pure oil painting but a mixed process in which oil
played only a part. The result at any rate, as all the world is aware, was the speedy
ruin of the work, which now only tells as a design, there being but little of its
creator’s actual handiwork now visible.
Some words of the Report are of sufficient interest to be quoted. ‘Pur troppo,
dunque, la stessa tecnica del maestro aveva in sè il germe della rovina, ben presto,
infatti, avvertita nelle sue opere murali. Spirito indagitore, innovatore, voglioso
sempre di “provare e riprovare” egli voile abbandonare i vecchi, sicuri e
sperimentati sistemi, per tentare l’ esito di sostanze oleose in miscela coi colori.
Perchè nemmeno può dirsi ch’ ei dipingesse, in questo caso, semplicemente, ad
olio come avrebbe fatto ogni altro mortale entrato nell’ errore di seguire quel
metodo anche pei muri. Egli tentò invece cosa affato nuova; poichè, se da un lato
appaiono tracce di parziali e circoscritte arricciature in uso pel fresco, dall’ altro, la
presenza delle sostanze oleose è accertata dalla mancanza di adhesione dei colori
con la superficie del muro e dalle speciali screpolature della crosta o pelle formata
dai colori stessi, non che dal modo con quale il dipinto si è andato e si va
lentamente disgregando e sfaldando.’ Bollettino d’ Arte del Ministero della
Pubblica Istruzione, Roma, 1907, I, p. 17.
Another famous instance of the use of oil paint in mural work about a
generation later is to be found in the Sala di Costantino in the Vatican, where
Raphael’s pupils have left two of the decorative figures by the side of the Popes
executed in that medium. One (Urbanity) is close to the door leading to the Chapel
of Nicholas V, the other is on the wall containing the battle, and is in better
preservation than the first which is covered with wrinkles. The oil paint gives a
certain depth and richness of effect, but there is the fatal disadvantage that the
painting does not look a part of the wall as is the case with work done in fresco.
The fresco is really executed in the material of the ground, whereas oils and
varnishes have nothing in common with lime and earths, and the connection of
structure and decoration is broken. One of the most successful pieces of work of
the kind is the painting of ‘Christ at the Pillar’ by Sebastian del Piombo in S. Pietro
in Montorio at Rome. The work, which is executed on a cylindrical surface, is
rather shiny, an appearance which in mural painting is to be avoided, and it has
darkened somewhat, though this defect is not very apparent and the experiment
has on the whole succeeded well. Vasari’s Life of Fra Sebastiano contains a good
deal of information about this particular technique, which was essayed in the later
age of Italian painting more often than is sometimes imagined. It needs hardly to
be said that this oil painting on the actual plaster of the wall is a different thing
from the modern process of painting on canvas in the studio and then cementing
the completed picture on to the wall. Mural painting on canvas was introduced by
the Venetians in the fifteenth century, for at Venice atmospheric conditions seem
to have been unfavourable to the preservation of frescoes, and the Venetians
preferred canvas to plaster for their work in oils. It would be interesting to know
whether the canvas was ever fixed in situ before the painter commenced
operations, as from the point of view of the preservation of decorative effect this
would be of importance. Vasari’s story about Tintoretto’s proceedings at the Scuola
di S. Rocco (Opere, VI, 594) is evidence that canvases were painted at home and
put up on walls or ceilings when finished. Of course if a wall be covered with
canvas before the painting begins the canvas is to all intents and purposes the wall
itself, grounded in a certain way.
218. The use of canvas for the purpose in view was, as Vasari mentions below,
very common at Venice, where as early as about 1476, if we believe Vasari (Opere,
III, 156), Gentile Bellini executed in this technique the large scenic pictures with
which he adorned the Hall of Grand Council in the Ducal Palace. Such a process
would come naturally enough to Italian painters as well as to the Flemings, for they
had been accustomed from time immemorial to paint for temporary purposes on
banners and draperies, after a fashion of which Mantegna’s decorative frieze on
fine canvas at Hampton Court is a classic example. Canvas had however been
actually used for pictures even in ancient Egypt. Not only was the practice of
stretching linen over wooden panels to receive the painting ground in use there in
the time of the New Empire, but some of the recently discovered mummy-case
portraits from Egypt, of the earliest Christian centuries, are actually on canvas.
There is an example in the National Gallery. At Rome painting on canvas is
mentioned by Pliny (Hist. Nat., XXXV, 51) and Boethius (de Arithmetica, Praef., I)
says that ‘picturae ... lintea operosis elaborata textrinis ... materiam praestant.’ The
Netherland painters of the fifteenth century nearly always painted on panel, but
canvas was sometimes used, as by Roger van der Weyden in his paintings for the
Town Hall at Brussels.
219. Vasari prescribes ‘due o tre macinate’ of white lead for mixture with the
flour and nut oil for the priming of canvas. A ‘macinata’ was the amount placed at
one time on the ‘macina’ or stone for grinding colours. Berger suggests ‘handfuls’
as a translation, but the amount would be small, as for careful grinding only one or
two lumps of the pigment would be dealt with at one time.
220. The Ducal Palace, that adjoins S. Marco, is probably the building in
Vasari’s mind. The Library of S. Marco, Sansovino’s masterpiece, might also be
meant, as this was called sometimes the Palace of S. Marco. We must remember
however that, as noticed before, ante, p. 56, this building, at the time of Vasari’s
visit to Venice, was still unfinished.
221. On panels and canvases as used at Venice Vasari has an interesting note
at the beginning of his Life of Jacopo Bellini (Opere, III, 152). This was a subject
that would at once appeal to his practical mind when he visited the city. He notices
incidentally that the usual woods for panels were ‘oppio’ acer campestris, maple;
or ‘gattice,’ the populus alba of Horace, but that the Venetians used only fir from
the Alps. (Cennini, c. 113, recommends poplar or lime or willow. Pliny, Hist. Nat.,
XVI, 187, speaks of larch and box, and Ilg says that northern painters generally used
oak.) The Venetian preference for canvas, Vasari says, was due to the facts that it
did not split nor harbour worms, was portable, and could be obtained of the size
desired; this last he notes too in our text. Berger (Beiträge, IV, 29), gives the
meaning of ‘Grossartigkeit’ to the word ‘grandezza’ used above by Vasari, but of
course it only means material size, not ‘grandeur’ in an aesthetic sense.
222. See ‘Introduction’ to Architecture, § 13, ante, p. 54. The stone is a species
of slate. Slate is suitable for painting on. See Church’s Chemistry of Paints and
Painting, 1890, p. 21.
223. Greek paintings on marble panels have come down to us from various
periods of ancient art. Some early Attic specimens on tombstones are in the
museums of Athens, and at Herculaneum there was found an interesting painting
on marble of a group of Greek heroines playing at knuckle bones. A much earlier
slab with a figure of a warrior is in the Acropolis Museum at Athens.
224. These chiaroscuri or monochromes are characteristic of the later
Renaissance. They may either be frankly decorative, and in this form obey the rules
of all other pictorial enrichment; or they may have an illusive intention, and be
designed to produce the appearance on a flat wall of architectural members or
sculptured or cast-bronze reliefs. In this case, when on monumental buildings and
permanent, they are insincere and opposed to sound decorative principles, though
on temporary structures they are quite in place. Vasari was a famous adept at the
construction and adornment of such fabrics, which were in great demand for the
numerous Florentine pageants and processions. See his letters, passim.
 225. There are examples of painted imitations of bronze in Michelangelo’s
frescoes on the vault of the Sistine. The medallions held by the pairs of decorative
figures of youths on the cornice are painted to represent reliefs in this metal.
Raphael’s Stanze and Loggie also furnish instances, and there are good examples
on the external façade of the Palazzo Ricci at Rome.
226. The clay or earth that Vasari speaks of forms the body of the ‘distemper’
or ‘gouache,’ as it would be called respectively in Britain and in France, and takes
the place of the ‘whitening’ used in modern times. Baldinucci in his Vocabolario
explains ‘Terra di cava o Terretta’ as ‘the earth (clay) with which vessels for the
table are made, that mixed with pounded charcoal is used by painters for
backgrounds and monochromes, and also for primings, and with a tempera of size
for the canvases with which are painted triumphal arches, perspectives, and the
like.’ It is of very fine and even texture, and Baldinucci says it was found near St.
Peter’s at Rome, and also in great quantity at Monte Spertoli, thirteen miles from
Florence.
227. This process of wetting the back of the canvas is to be noted. The chief
inconvenience of the kind of work here spoken of is that it dries very quickly, and
dries moreover very much lighter than when the work is wet. Hence it is an
advantage to keep the ground wet as long as possible till the tints are properly
fused, so that all may dry together. Wetting the back of the canvas secures this end.
The technique that Vasari is describing is the same as that of the modern theatrical
scene-painter, and would be called ‘distemper painting.’ The colours are mixed
with whitening, or finely-ground chalk, and tempered with size. The whitening
makes them opaque and gives them ‘body,’ but is also the cause of their drying
light. F. Lloyds, in his Practical Guide to Scene Painting and Painting in
Distemper, Lond. 1879, says (p. 42) ‘In the study of the art of distemper painting, a
source of considerable embarrassment to the inexperienced eye is that the colours
when wet present such a different appearance from what they do when dry.’
228. Does Vasari mean by ‘tempera’ yolk of egg? It has this sense with him
sometimes, as in the heading of chapter VI.
229. Cennini in his 67th chapter gives directions for preparing the mixed
colour he calls verdaccio. It was a compound of white, dark ochre, black and red.
230. The principle of sgraffito-work, that is the scratching through a thin
superimposed coat to bring to view an under layer of a different colour, seems to
have been established first in pottery making, and in this connection the Italians
called it ‘Sgraffiato.’ The adoption of the process for the decoration of surfaces of
plaster or cement was an innovation of the Renaissance, and Vasari appears to
have been the first writer who gives a recipe for it. According to his account in the
Lives, it was a friend of Morto da Feltro, the Florentine Andrea di Cosimo, who
first started the work, and Vasari describes the process he employed in phrases
that correspond with the wording of the present chapter (Opere, ed. Milanesi, V,
207). A modern expert describes the process as follows: ‘A wall is covered with a
layer of tinted plaster, and on this is superimposed a thin coating of white plaster.
The outer coat is scratched through, and the colour behind it is revealed. Then all
the white surface outside the design is cut away, and a cameo-like effect given to
the design. This is the art of Sgraffito as known to the Italian Renaissance’
(Transactions, R.I.B.A., 1889, p. 125). The process dropped out of use after a
while, but was revived in Germany in the middle of the nineteenth century, mainly
through the agency of the architect Gottfried Semper, the author of Der Stil. It is
sometimes used in our own country both on monumental and on domestic
buildings, and as it is simple and cheap and permanent it is well fitted for modern
use in our climate. The back of the Science School in Exhibition Road, S.
Kensington, was covered with sgraffiti by the pupils of the late F. W. Moody about
1872. They would be the better now for a cleansing with the modern steam-blast.
231. See the Notes on ‘Enriched Façades,’ and ‘Stucco “Grotesques,”’ at the
close of the ‘Introduction’ to Painting, postea, pp. 298, 299.
232. This passage presents some difficulty. It runs ‘Dunque, quelle che vanno
in campo bianco, non ci essendo il campo di stucco per non essere bianca la calce,
si dà per tutto sottilmente il campo di bianco.’ Vasari seems to have in his mind the
difference between ordinary plaster made, as he has just described, of ‘lime mixed
with sand in the ordinary fashion,’ which would not be white, and what he calls
‘stucco,’ by which term is probably meant the finer plaster made of white lime from
travertine and marble dust. Ordinary plaster has accordingly to be coated with
white before the work begins.
233. Examples of this whimsical style of decoration are abundant in the
Pompeian wall paintings, and the mind of Vitruvius was much exercised about
their frivolity and want of meaning (De Architectura, VII, v).
234. Vasari is not very clear in his account of these methods of work, but it is
enough to know that both by the ancients, and at the time of the Renaissance,
colour was used largely in connection with these reliefs, and the combination could
of course take several forms. In the loggia of the Villa Farnesina, where Raphael
worked with his assistants, there are painted panels in fresco framed in mouldings
of stucco, modelled plaster figures in white against a coloured ground, coloured
stuccoes against coloured fields, and tinted bands separating the framed plaster
medallions. The same kind of work is found in the Loggie of the Vatican, the Doria
Palace at Genoa, and other localities innumerable. Plate XII shows a characteristic
section of the decoration of the Vatican Loggie.
235. As in the work described at the close of ch. XII (the beginning of the
present section).
236. The word ‘bolus’ is derived from the Greek βῶλος, a lump or clod, and
means, according to Murray’s Dictionary, a pill, or a small rounded mass of any
substance, and also a kind of reddish clay or earth, used medically for its astringent
properties, that was brought from Armenia, and called by the pharmacologist ‘bole
armeniac.’ Its use in the arts is due to its unctuous character, which made gold
adhere to it. See below. In mediaeval illuminations a ‘bolus’ or small lump of a