Charge-based CMOS Digital RF

Transmitters 1st Edition Pedro Emiliano

Paro Filho
Visit to download the full and correct content document:
https://textbookfull.com/product/charge-based-cmos-digital-rf-transmitters-1st-edition-
pedro-emiliano-paro-filho/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

ElasticSearch Cookbook, Second Edition Alberto Paro

https://textbookfull.com/product/elasticsearch-cookbook-second-
edition-alberto-paro/

Professional Node js Building Javascript Based Scalable

Software 1st Edition Teixeira Pedro

https://textbookfull.com/product/professional-node-js-building-
javascript-based-scalable-software-1st-edition-teixeira-pedro/

Data Science and Digital Business Fausto Pedro García

Márquez

https://textbookfull.com/product/data-science-and-digital-
business-fausto-pedro-garcia-marquez/

Endocrinology of Aging Emiliano Corpas

https://textbookfull.com/product/endocrinology-of-aging-emiliano-
corpas/
Management and Technological Challenges in the Digital
Age 1st Edition Pedro Novo Melo

https://textbookfull.com/product/management-and-technological-
challenges-in-the-digital-age-1st-edition-pedro-novo-melo/

Radiation Hardened CMOS Integrated Circuits for Time

Based Signal Processing Jeffrey Prinzie

https://textbookfull.com/product/radiation-hardened-cmos-
integrated-circuits-for-time-based-signal-processing-jeffrey-
prinzie/

Chipless RFID Based on RF Encoding Particle.

Realization, Coding and Reading System 1st Edition
Arnaud Vena

https://textbookfull.com/product/chipless-rfid-based-on-rf-
encoding-particle-realization-coding-and-reading-system-1st-
edition-arnaud-vena/

CMOS-Compatible Key Engineering Devices for High-Speed

Silicon-Based Optical Interconnections Jing Wang

https://textbookfull.com/product/cmos-compatible-key-engineering-
devices-for-high-speed-silicon-based-optical-interconnections-
jing-wang/

Low Power CMOS Digital Pixel Imagers for High Speed

Uncooled PbSe IR Applications 1st Edition Josep Maria
Margarit (Auth.)

https://textbookfull.com/product/low-power-cmos-digital-pixel-
imagers-for-high-speed-uncooled-pbse-ir-applications-1st-edition-
josep-maria-margarit-auth/
ACSP · Analog Circuits And Signal Processing

Pedro Emiliano Paro Filho

Jan Craninckx
Piet Wambacq
Mark Ingels

Charge-
based CMOS
Digital RF
Transmitters
Analog Circuits and Signal Processing

Series Editors
Mohammed Ismail, Dublin, USA
Mohamad Sawan, Montreal, Canada

More information about this series at http://www.springer.com/series/7381

Pedro Emiliano Paro Filho • Jan Craninckx
Piet Wambacq • Mark Ingels

Charge-based CMOS Digital

RF Transmitters

123
Pedro Emiliano Paro Filho Jan Craninckx
SSET department SSET department
IMEC SSET department IMEC SSET department
Leuven, Belgium Leuven, Belgium

Piet Wambacq Mark Ingels

SSET department SSET department
IMEC SSET department IMEC SSET department
Leuven, Belgium Leuven, Belgium

ISSN 1872-082X ISSN 2197-1854 (electronic)

Analog Circuits and Signal Processing
ISBN 978-3-319-45786-4 ISBN 978-3-319-45787-1 (eBook)
DOI 10.1007/978-3-319-45787-1

Library of Congress Control Number: 2016951714

© Springer International Publishing Switzerland 2017

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology
now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or
the editors give a warranty, express or implied, with respect to the material contained herein or for any
errors or omissions that may have been made.

Printed on acid-free paper

This Springer imprint is published by Springer Nature

The registered company is Springer International Publishing AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
If Education alone cannot transform society,
nor without it society changes.

Paulo Freire
To my family
the leaning,
the standing, and
the crawling ones.
Preface

The point where wireless communication and ubiquitous connectivity became an

essential part of our lives is already past. Generation after generation communi-
cation speed is being taken to unprecedented levels, requiring both state-of-the-art
hardware and software to handle a huge volume of data, delivered to an increasing
number of users in an overcrowded spectrum. To our delight, the challenges are
always plentiful.
With respect to the radio front-end, providing an extremely low noise emission
with an improved signal integrity is a key requirement to support high-order
modulation schemes (e.g., 64 QAM) in situations where anyone’s transmitter can be
interfering with a neighbor user or its own receiver in frequency-division full duplex
mode. Increasing power and/or area consumption is not an option in this case. On
the contrary, for an improved user experience the battery should last longer, and the
price per component should always go down, so that more and more features can
be added to a mobile/handheld device. Thus, making a better performing CMOS
radio front-end that consumes even less power and area is a hot research topic these
days, especially regarding the transmitter and PA designs, considered by many the
“battery killers” on most mobile devices.
A quick analysis of literature shows that with regard to CMOS transmitter
implementations, the state of the art is clearly divided into analog- and digital-
intensive architectures. In terms of out-of-band noise, analog-intensive architectures
are undoubtedly the best performing implementations. However, their improved
noise performance is typically achieved through extensive low-pass filtering along
the entire signal path, which has a significant impact on area consumption.
Digital-intensive implementations, on the other hand, are by far the most portable,
area-efficient, and scaling-friendly ones. However, the lack of filtering (for both
noise and aliases) makes it very challenging to meet the stringent out-of-band noise
requirements in SAW-less operation.
In this book, a novel digital-intensive transmitter architecture that can relax this
trade-off is described. Through the combination of charge-domain operation with
incremental signaling, this architecture gives the best of both worlds, providing the

ix
x Preface

reduced area and high portability of digital-intensive architectures with an improved

out-of-band noise performance given by intrinsic noise filtering capabilities.
Two implementations of the incremental charge-based TX are demonstrated, dif-
fering on how the charge-based DAC (QDAC) is implemented and the RF load being
driven: In the first realization, the RF load is the input capacitance of a PPA stage,
and the QDAC is implemented with a controllable capacitance that is alternately
pre-charged and connected to a charge reservoir. In the second implementation,
the ability of delivering more power using the charge-based architecture has been
investigated with a direct-launch architecture, where the 50  load representing the
PA input is directly driven with charge. The QDAC is implemented with a 12-bit
conductance array, which proves to be the most area-efficient implementation in
this case.
Prototyped using a 28 nm 0.9 V CMOS technology, both charge-based TX
realizations provide remarkable results in terms of noise performance, thanks
to their intrinsic noise filtering capability, improved sampling alias attenuation,
and reduced quantization noise. With an out-of-band noise spectral density of
159 dBc/Hz and a core area of 0.22 mm2 , the second implementation achieves—
to the author’s knowledge—the best out-of-band noise performance versus area
consumption when compared to other similar works. ACLR and EVM performance
are also among the best. As a result, this work paves the way for compact CMOS
SAW-less transmitter implementations enabling advanced wireless communication
systems, including 3G, 4G, and beyond.

Leuven, Belgium Pedro Emiliano Paro Filho

Jan Craninckx
Piet Wambacq
Mark Ingels
Acknowledgments

At the age of 16, I found myself with a book in hands, which tells the story of a
seagull named Jonathan Livingston Seagull (“Fernão Capelo Gaivota”). Jonathan
Seagull had a great drive to learn and improve his flying skills. His passion took
him to new places and ever higher altitudes, even setting him apart from his loved
ones eventually. Flying was an obsession for him, only shadowed by his willingness
to share what he has learned along this long—and many times lonesome—journey
toward greater understanding and self-awareness.
Four years ago, another seagull started a journey that in many ways resembles
the story of Jonathan Seagull, reason why this book came to my mind 16 years
later. These were years of hard work and great dedication to learning, teaching, and
creating. There were also various high-speed dives that turned out as great crashes
against the water, leaving the seagull adrift many times. But more than anything,
the lessons learned—eventually leading to few successful flights over beautiful
landscapes—and the people met along the way made every piece of this journey
worthwhile. For that, I’d like to thank:
Jan, Piet, and Mark, whose guidance, trust, and support not only made this
journey possible but also served me as inspiration, teaching with lifelong dedication
the values of serious and honest research.
All my colleagues and friends from IMEC, with whom I had the chance to share
fruitful moments over coffee and beer, and the brazilian community, which was
a great source of support during the last 4 years with great barbecues, feijoadas,
anniversaries, and, unfortunately, farewell parties.
Along the way, there were also people that somehow made all the difference,
and for whom I keep a special place in my heart: Adrian, a living proof that
companionship and intelligence cannot be measured in kilograms; Karlinha, Cadu,
and Guto, three great-hearted people always ready to offer a friendly shoulder; and
Oscar, Ricardão, André, Marcão, Cássio, Gigi, João, and Mihnea, great friends
whose discussions were always a good way to wash away microelectronics from
an overloaded mind.
Thank you all for the great time.

xi
xii Acknowledgments

Also, none of this would have been possible if it were not for the support and
affection received from abroad. Daily phone calls and countless countdowns to our
next encounter were the only way to keep us together at distance but never apart.
And we made it! Te amo linda.
Finally, my greatest gratitude to my beloved family. My dear father and mother,
to whom I dedicate this work. My siblings Paula, Roberta, and Renato, my brother-
in-law Rodrigo, and my two “not so little anymore” princesses Ana Luiza and
Laurinha.
You are the sunshine of my life.

How much more there is now to living! Instead of our drab slogging forth and back
to the fishing boats, there’s a reason to life! We can lift ourselves out of ignorance,
we can find ourselves as creatures of excellence and intelligence and skill. We can
be free!
We can learn to fly!
(Jonathan Livingston Seagull, by Richard Bach)
Contents

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 The Fear of Disconnection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Advanced Wireless Communication Systems . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Flexible Multi-Standard Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3.1 TX Frontend Key Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.4 High Performance TX Architectures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.5 Quadrature Direct-Conversion Transmitters . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2 Incremental-Charge-Based Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.2 Incremental-Charge-Based Operation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.2.1 Noise and Alias Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.3 Charge-Based Transmitter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.3.1 Power Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3 Capacitive Charge-Based Transmitter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.2 Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.2.1 Operating Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.2.2 CQDAC Operation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.2.3 Noise and Alias Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.2.4 Harmonic Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.3 Circuit Realization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
3.3.1 CQDAC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
3.3.2 Mixer and PPA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
3.3.3 LO Generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
3.3.4 Top-Level Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

xiii
xiv Contents

3.4 Measurement Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

3.4.1 Measurement Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
3.4.2 CQDAC Measurement Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
3.4.3 CQDAC TX Measurement Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
3.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
4 Resistive Charge-Based Transmitter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
4.2 Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
4.2.1 Operating Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
4.2.2 RQDAC Operation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
4.2.3 Noise and Alias Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
4.2.4 Harmonic Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
4.3 Circuit Realization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
4.3.1 RQDAC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
4.3.2 Mixer Design. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
4.3.3 LO Generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
4.3.4 Top-Level Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
4.4 Measurement Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
4.4.1 Measurement Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
4.4.2 RQDAC Measurement Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
4.4.3 RQDAC TX Measurement Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
4.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
5.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143

Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
List of Figures

Fig. 1.1 Example LTE signal (left) showing spectral mask,

adjacent channel and RX band (FDD), typically
separated from the carrier (fc) by tens of megahertz
(OFFSET). Captured modulated signal (right) with a
EVM of 1.6 % . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Fig. 1.2 Probability Density for QPSK and 64QAM
modulations (left). The increased probability at small
amplitudes also translates into large PAPRs, as shown
in Table 1.2 (right) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Fig. 1.3 Out-of-band noise calculations. Example extracted
from [Oka11]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Fig. 1.4 Typical analog-intensive direct conversion quadrature TX . . . . . . . . . 9
Fig. 1.5 Simplified polar architecture. The Cartesian to polar
conversion is typically implemented using a coordinate
rotation for digital computer (CORDIC) algorithm [Meh09]. . . . . . . 10
Fig. 1.6 Simplified Outphasing block diagram. Both V1 and
V2 are constant-envelope signals that can be amplified
using very non-linear PAs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Fig. 1.7 Spectrum content at the different stages of conventional
analog-intensive I/Q transmitters. The reconstruction
filter significantly attenuates both noise and spurs
at the RX-band . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Fig. 1.8 Simplified direct-conversion digital-intensive I/Q
transmitter (left). Current-based DACs combined with
the LO signal provide direct digital-to-RF conversion
(right) as proposed in [Elo07] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

xv
xvi List of Figures

Fig. 1.9 The direct digital-to-RF conversion hinders noise

and spurs filtering, making it challenging for
digital-intensive architectures to meet the stringent
SAW-less operation requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Fig. 2.1 Watermill mechanical system. The amount of power
transferred to rotating wheel should follow a wanted
input command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Fig. 2.2 Example control mechanism. A voltage-regulated
water pump is used to control the amount of water
pushed through the rotating wheel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Fig. 2.3 Alternative control mechanism. Instead of using a
water pump, the water flow is controlled by changing
the water level in a large reservoir . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Fig. 2.4 Top-level description of the reservoir-based operation.
Following the input command, the control system
determines how much water should be added or
subtracted from the reservoir . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Fig. 2.5 Example structure used to control the reservoir level.
According to required WTOTAL , the corresponding
number of buckets (#BUCKETS ) are pre-filled or
pre-emptied before being connected to the reservoir,
therefore increasing or decreasing the total amount stored . . . . . . . . . 21
Fig. 2.6 Minimum achievable water level in(de)crease. In this
alternative implementation, the water level resolution
(and hence quantization noise) can be improved by
either increasing the reservoir size, or reducing the
minimum bucket size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
Fig. 2.7 Simplified representation of a typical system where a
load Z has to be driven following a wanted (digital)
input command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Fig. 2.8 In the incremental charge-based architecture the output
voltage is changed by adding and subtracting charge
from the output load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Fig. 2.9 Charge-based operation. A charge calculation block
defines how much charge should be moved to/from the
output load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Fig. 2.10 Using zero-order hold the reconstruction spectrum is
shaped by a sinc.x/ function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
Fig. 2.11 sinc.x/2 alias attenuation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Fig. 2.12 L-fold linear interpolation for different values of L . . . . . . . . . . . . . . . . . 27
Fig. 2.13 Output spectrum of a wideband output signal sampled
at 80 MS/s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
Fig. 2.14 Noise and Alias filtering achieved with a second-order
reconstruction RC filter with a cutoff frequency of 15 MHz. . . . . . . . 29
List of Figures xvii

Fig. 2.15 The voltage drop across the QDAC terminals in

combination with the output current can be translated
into an equivalent DAC conductance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Fig. 2.16 Intrinsic noise filtering mechanism. The signal
dependence of the DAC GEQU makes sure that both
slow (1) and fast (2) transitions of the desired signal
are properly constructed at the output, facing no
attenuation. Uncorrelated noise, on the other hand, is
filtered by an equivalent cutoff frequency given by the
average conductance (3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Fig. 2.17 Single-order intrinsic noise filtering, showing a
0.5 Vpkpk 10 MHz single-tone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Fig. 2.18 Noise cutoff frequency versus signal output frequency . . . . . . . . . . . . . 33
Fig. 2.19 Noise cutoff frequency versus signal amplitude. . . . . . . . . . . . . . . . . . . . . 33
Fig. 2.20 Error accumulation can quickly compromise the
charge-based operation if subsequent blocks are not
re-adjusted to account for a wrong packet of charge . . . . . . . . . . . . . . . . 35
Fig. 2.21 “Self-healing” mechanism. The increased voltage
difference implied by a reduced charge “block”
increases the amount of charge transferred during the
following steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Fig. 2.22 Thanks to the “self-healing” mechanism, the
charge-based operation is completely resilient even to
large errors in the charge accumulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Fig. 2.23 Even when large errors are implied, the accumulated
error is quickly dissipated . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Fig. 2.24 Output Spectrum for several degrees of mismatch
between CLOAD and its assumed value . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Fig. 2.25 Harmonic distortion versus CLOAD mismatch. . . . . . . . . . . . . . . . . . . . . . . . 38
Fig. 2.26 Harmonic distortion versus QDAC gain mismatch . . . . . . . . . . . . . . . . . 39
Fig. 2.27 Simplified charge-based TX block diagram . . . . . . . . . . . . . . . . . . . . . . . . . 40
Fig. 2.28 RF spectrum of a charge-based TX driving a 200 fF
capacitance at 2 GHz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
Fig. 2.29 Simplified schematic where the RF load is translated to
the baseband side . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
Fig. 2.30 Noise cutoff frequency versus RF load resistance. . . . . . . . . . . . . . . . . . . 42
Fig. 2.31 Simplified schematic depicting QDAC Class-B operation. . . . . . . . . . 43
Fig. 2.32 Efficiency versus Backoff considering CBB only . . . . . . . . . . . . . . . . . . . . 43
Fig. 2.33 Example charge-based transmitter considered for
Efficiency calculations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
Fig. 2.34 QDAC TX Efficiency versus Backoff considering a
50  load and various CBB values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
xviii List of Figures

Fig. 2.35 Efficiency and noise cutoff frequency versus CBB at

2 dB backoff . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
Fig. 3.1 Architecture overview of the direct-conversion IQ
charge-based transmitter. Four CQDACs provide
each one of the differential I/Q components, driving
the pre-power amplifier (PPA) through a passive
switch-based mixer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
Fig. 3.2 CQDAC block, which delivers discrete packets
of charge at LO speed by controlling a variable
capacitance CDAC Œk, and to what voltage it is
pre-charged before connecting to CBB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
Fig. 3.3 Operating principle of the charge-based transmitter
architecture. Based on the digital I/Q input signal, the
amount of charge that should be transferred to both
baseband and RF nodes are calculated. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
Fig. 3.4 The charge transfer between the supply and the RF
node is divided into three non-overlapping phases,
namely precharge, share and LO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
Fig. 3.5 Detailed incremental charge-based operation. Since
the amount of charge taken by the RF node is also
accounted, the baseband voltage reaches the expected
final value VBB Œk at the end of the sampling period . . . . . . . . . . . . . . . . 52
Fig. 3.6 Conventional first-order RC filter. The fixed resistance
(RFIXED ) yields an also fixed cutoff frequency
[Eq. (3.9)], typically placed above the maximum
baseband frequency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
Fig. 3.7 CQDAC equivalent RC filter. The signal-dependent
resistance (RVARIABLE ) creates a time-varying
single-order low-pass filter whose bandwidth is
automatically adjusted according to the input signal. . . . . . . . . . . . . . . . 55
Fig. 3.8 PNOISE simulation of a 400 mVpp (13:97 dBV),
8 MHz single-tone sampled at 128 MS/s. Uncorrelated
noise is filtered by an equivalent cutoff frequency 2.8
lower than fBB , without attenuating the baseband signal . . . . . . . . . . . . 56
Fig. 3.9 Noise cutoff frequency versus baseband amplitude
for a 10 MHz single-tone sampled at 128 MS/s, and a
baseband capacitance of 50 pF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
Fig. 3.10 Noise cutoff frequency versus fBB for a 400 mVpp
single-tone sampled at 128 MS/s, and a baseband
capacitance of 50 pF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
Fig. 3.11 Baseband output spectrum of the charge-based CQDAC
versus a typical DAC implementation (CBB D 50 pF,
ABB D 400 mVpp). The quantization noise filtering
capability is clearly noted . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
List of Figures xix

Fig. 3.12 Quantization noise floor versus CUNIT for a fixed CBB
of 50 pF. As expected, the noise floor power density is
reduced by 12 dB when the unit capacitance is divided
by four . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
Fig. 3.13 CQDAC equivalent number of bits versus CUNIT , for
different baseband capacitances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
Fig. 3.14 Required number of DAC elements for various
baseband maximum amplitude and frequency (CBB D 50 pF) . . . . . 60
Fig. 3.15 Detailed timing diagram showing the “quasi” linear
interpolation between samples achieved through the
incremental charge of CBB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
Fig. 3.16 Alias relative power versus baseband sampling
frequency (fS ). For sampling frequencies lower than
fLO =2 the alias attenuation improves significantly . . . . . . . . . . . . . . . . . . 61
Fig. 3.17 Charge transfer and resulting voltage excursion using
two different switch implementations. The voltage
error introduced by the poor voltage settling is
accumulated at CBB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
Fig. 3.18 Baseband output spectrum using two different
settling conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
Fig. 3.19 Harmonic Distortion versus charge path time constant
as a fraction of the switch ON-period. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
Fig. 3.20 Top-level schematic of the CQDAC implementation . . . . . . . . . . . . . . . 65
Fig. 3.21 CQDAC equivalent number of bits for various
combinations of CUNIT and CBB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
Fig. 3.22 Different possible combinations of maximum swing
and frequency attained with 1024 DAC unit capacitances . . . . . . . . . . 67
Fig. 3.23 CQDAC unit cell schematic, showing the control logic
used to synchronize the different charge phases . . . . . . . . . . . . . . . . . . . . 67
Fig. 3.24 Equivalent RC time constants involved in each one of
the charge convey steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
Fig. 3.25 Error voltage mechanism induced by charge injection. . . . . . . . . . . . . . 69
Fig. 3.26 Charge injected by a 400 nm/50 nm transistor as a
function of its terminal’s voltage. The combination
of complementary NMOS and PMOS switches
reduces the amount of charge injected, as well as
creates a zero-crossing point where charge-injection is
completely cancelled . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
Fig. 3.27 Final schematic of the implemented unit cell showing
device sizes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
Fig. 3.28 CQDAC floorplanning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
Fig. 3.29 Bit-line in detail . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
Fig. 3.30 CQDAC TX block diagram showing the voltage
sampling passive mixer and PPA used in this implementation . . . . . 75
Fig. 3.31 PPA schematic depicting the external bias tee and 50  load . . . . . . 76
xx List of Figures

Fig. 3.32 Simulated PPA gain and efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

Fig. 3.33 PPA output noise spectral density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Fig. 3.34 Voltage sampling passive mixer operation. Each IQ
component is sampled for 1/4 of the LO period. In the
following analysis the mixer switch has a finite ON
resistance represented by RMIX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
Fig. 3.35 Equivalent RC time constants involved in each one of
the charge convey steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
Fig. 3.36 Switch resistance voltage dependence for example
NMOS, PMOS and NPMOS implementations . . . . . . . . . . . . . . . . . . . . . . 79
Fig. 3.37 Example spectrum showing the impact of the switch
resistance voltage dependence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
Fig. 3.38 Mixer switch schematic with device sizes in detail. The
LO signal is decoupled and biased with VBIASC and
VBIAS in order to increase the switch overdrive . . . . . . . . . . . . . . . . . . . . 80
Fig. 3.39 LO generation scheme. Quadrature LO signals with
adjustable duty-cycle are made from logic combination
between LO signals at fLO and 2  fLO frequencies. . . . . . . . . . . . . . . . . . . 82
Fig. 3.40 Top-level block diagram of the CQDAC TX prototype . . . . . . . . . . . . . 83
Fig. 3.41 Unity-gain voltage amplifier schematic. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
Fig. 3.42 Chip micrograph . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
Fig. 3.43 Overview of the measurement setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
Fig. 3.44 Dynamic measurement scheme used to determine
CUNIT and CBB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
Fig. 3.45 Measured charge injection error versus input code . . . . . . . . . . . . . . . . . 88
Fig. 3.46 Measured baseband spectrum for a 1 MHz single-tone,
with the mixer disabled . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
Fig. 3.47 CQDAC second and third-order harmonic distortion
for different sampling frequencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
Fig. 3.48 Measured RF spectrum of a 5 MHz baseband signal
sampled at 128 MS/s and transmitted at 1.024 GHz. . . . . . . . . . . . . . . . . 90
Fig. 3.49 TX spurious emission for different sampling frequencies . . . . . . . . . . 91
Fig. 3.50 Measured alias attenuation at multiple baseband
frequencies, from 1 to 20 MHz for a 128 MS/s baseband
sampling rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
Fig. 3.51 Measured output noise for a 5 MHz, 1 dBm output
power (7 dB backoff) single-tone transmitted at
1.024 GHz (baseband harmonics were removed for
clarity). Both LO and quantization noise are included . . . . . . . . . . . . . . 92
Fig. 3.52 Measured ACLR1/2 performance for 20 MHz BW . . . . . . . . . . . . . . . . . 93
Fig. 4.1 Architecture overview of the Cartesian resistive
charge-based DAC transmitter. Four QDACs provide
each one of the differential I/Q components, however
the RF load is now directly driven by the baseband
circuitry through passive switch-based mixer, without a PPA . . . . . . 98
List of Figures xxi

Fig. 4.2 Maximum charge capacity of both capacitive and

resistive QDACs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
Fig. 4.3 Resistive charge-based transmitter operating principle . . . . . . . . . . . . . 100
Fig. 4.4 Simplified schematic showing the parasitic output
capacitance CPAR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
Fig. 4.5 Simplified diagram of the RQDAC implementation . . . . . . . . . . . . . . . . 102
Fig. 4.6 RQDAC TX timing diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
Fig. 4.7 Simplified schematic including the equivalent RF
impedance seen from the baseband node . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
Fig. 4.8 PNOISE simulation of a 700 mVpp, 8 MHz single-tone
sampled at 500 MS/s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
Fig. 4.9 Noise cutoff frequency versus baseband amplitude
for a 8 MHz single-tone sampled at 500 MS/s, and a
baseband capacitance of 150 pF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
Fig. 4.10 Noise cutoff frequency versus baseband frequency for
a 500 mVpp single-tone sampled at 500 MS/s, and a
baseband capacitance of 150 pF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
Fig. 4.11 Noise cutoff frequency versus baseband capacitance
for a 500 mVpp, 8 MHz single-tone sampled at 500 MS/s . . . . . . . . . . 107
Fig. 4.12 RQDAC equivalent number of bits versus unit
resistance (RUNIT ) at multiple sampling frequencies, for
a baseband capacitance of 100 pF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
Fig. 4.13 RQDAC equivalent number of bits versus baseband
capacitance (CBB ) at multiple sampling frequencies, for
a unit resistance of 25 k . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
Fig. 4.14 Example baseband output spectrum showing the
RQDAC transmitter quantization noise filtering . . . . . . . . . . . . . . . . . . . . 109
Fig. 4.15 Required number of DAC elements versus signal
amplitude for two baseband maximum frequencies. . . . . . . . . . . . . . . . . 110
Fig. 4.16 Alias attenuation for different baseband sampling frequencies. . . . . 111
Fig. 4.17 Simplified schematic used to study the impact of the
unit cell switch on the transmitter harmonic performance . . . . . . . . . . 112
Fig. 4.18 Example baseband output spectrum for two
implementations with different switch resistances . . . . . . . . . . . . . . . . . . 113
Fig. 4.19 Harmonic distortion versus RQDAC switch resistance . . . . . . . . . . . . . 113
Fig. 4.20 Example baseband output spectrum for two
implementations with different ROFF =RON ratios . . . . . . . . . . . . . . . . . . . 114
Fig. 4.21 Harmonic distortion versus ROFF =RON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
Fig. 4.22 RF output spectrum for two example implementations
with different switch resistances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
Fig. 4.23 Harmonic distortion versus mixer switch resistance . . . . . . . . . . . . . . . . 116
Fig. 4.24 RQDAC architecture. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
Fig. 4.25 Equivalent noise cutoff frequency as a function of the
baseband capacitance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
xxii List of Figures

Fig. 4.26 Simulated noise spectral density versus the RUNIT , and
the required number of DAC elements for a 20 MHz
bandwidth with a peak output power of 7 dBm . . . . . . . . . . . . . . . . . . . . . 118
Fig. 4.27 RQDAC unit cell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
Fig. 4.28 Non-linear distortion mechanism due to OFF cells
parasitic loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
Fig. 4.29 RQDAC series/parallel bit construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
Fig. 4.30 Example spectrum showing the impact of reducing the
code-dependent parasitic loading at node P through
series/parallel resistor combination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
Fig. 4.31 Harmonic distortion of proposed RQDAC construction
for various baseband frequencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
Fig. 4.32 RQDAC floorplanning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
Fig. 4.33 Bit-line in detail . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
Fig. 4.34 Required switch resistance for a given mixer dynamic
performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
Fig. 4.35 Mixer switch schematic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
Fig. 4.36 Simulated harmonic performance for given mixer design . . . . . . . . . . 125
Fig. 4.37 Proposed LO generation block . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
Fig. 4.38 LO Phase Noise performance at minimum and
maximum duty cycle configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
Fig. 4.39 Top-level block diagram of the RQDAC TX prototype . . . . . . . . . . . . . 127
Fig. 4.40 Chip micrograph . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
Fig. 4.41 RQDAC TX measurement setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
Fig. 4.42 Measured RQDAC baseband output spectrum with
mixer OFF. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
Fig. 4.43 Measured RQDAC harmonic performance with mixer OFF . . . . . . . 131
Fig. 4.44 Example RF output spectrum for a 10 MHz single-tone
modulated at 900 MHz, with 0 dBm output power . . . . . . . . . . . . . . . . . . 131
Fig. 4.45 RF spurious emission versus baseband frequency at 0 dBm . . . . . . . . 132
Fig. 4.46 Mixer switch resistance voltage dependence used to
pre-distort the baseband voltage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
Fig. 4.47 RF spectrum before and after baseband pre-distortion. . . . . . . . . . . . . . 133
Fig. 4.48 CIM3 and CIM5 before and after pre-distortion. As
noted, the improvements are more pronounced at
2.4 GHz since calibration was performed at this LO frequency . . . . 134
Fig. 4.49 Single-tone spurious emission for various backoff
values and sampling frequencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
Fig. 4.50 Measured sampling aliases at both 250 and 500 MS/s . . . . . . . . . . . . . . 136
Fig. 4.51 Measured ACLR1/2 performance for 20 MHz BW at 2.4 GHz . . . . 136
Fig. 4.52 Measured QPSK, 16QAM and 64QAM constellation plots . . . . . . . . 137
Fig. 4.53 Measured out-of-band noise at maximum output power
for 900 MHz modulated carrier (10 MHz single-tone
sampled at 500 MS/s) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
List of Figures xxiii

Fig. 4.54 Measured out-of-band noise at maximum output power

for 2.4 GHz modulated carrier (10 MHz single-tone
sampled at 500 MS/s) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
Fig. 4.55 Measured out-of-band noise at different backoff values . . . . . . . . . . . . 139
List of Tables

Table 1.1 Modern wireless communication systems. . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Table 1.2 PAPR versus wireless standard . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Table 3.1 Performance summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Table 3.2 Comparison table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
Table 4.1 ACLR1/2 performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
Table 4.2 EVM performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
Table 4.3 Performance summary (900 MHz). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
Table 4.4 Performance summary (2.4 GHz) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
Table 4.5 Comparison table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141

xxv
List of Abbreviations

ACLR Adjacent-Channel Leakage Ratio

ACPR Adjacent-Channel Power Ratio
AMPS Advanced Mobile Phone Service
AM Amplitude Modulation
BER Bit-Error Rate
BW Bandwidth
CA Carrier Aggregation
CIM3 Third-order Counter Intermodulation
CIM5 Fifth-order Counter Intermodulation
CMOS Complementary Metal Oxide Semiconductor
CORDIC Coordinate Rotation for Digital Computer
CQDAC Capacitive Charge-based Digital-to-Analog Converter
CS Common Source
DAC Digital-to-Analog Converter
DNL Differential Nonlinearity
DPD Digital Pre-Distortion
EDGE Enhanced Data-rates for GSM Evolution
ET Envelope Tracking
EVM Error-Vector Magnitude
FDD Frequency-Division Duplexing
FDMA Frequency-Division Multiple Access
FM Frequency Modulation
GBW Gain-Bandwidth Product
GMSK Gaussian Minimum Shift Keying
GPRS General Packet Radio Service
GPS Global Positioning System
GSM Global System for Mobile communication
HD3 Third-order Harmonic Distortion
HD5 Fifth-order Harmonic Distortion
HD Harmonic Distortion
HSDPA High-Speed Downlink Packet Access

xxvii
xxviii List of Abbreviations

INL Integral Nonlinearity

LO Local Oscillator
LSB Least-Significant Bit
LTE Long Term Evolution
MIMO Multiple Input Multiple Output
MOM Metal-Oxide-Metal
MOS Metal Oxide Semiconductor
NFC Near Field Communication
NMOS N-type Metal Oxide Semiconductor
NMT Nordic Mobile Telephone
NOC Network-On-Chip
NTT Nippon Telegraph and Telephone
OFDM Orthogonal Frequency-Division Multiplexing
OOB Out-Of-Band (noise)
OQPSK Offset Quadrature Phase-Shift Keying
OSR Oversampling Ratio
OTA Operational Transconductance Amplifier
PAPR Peak-to-Average Power Ratio
PA Power Amplifier
PCB Printed Circuit Board
PMOS P-type Metal Oxide Semiconductor
PM Phase Modulation
PNOISE Periodic Steady-State Noise
PPA Pre-Power Amplifier
QAM Quadrature Amplitude Modulation
QDAC Charge-based Digital-to-Analog Converter
RAM Random Access Memory
RFDAC Typical denomination for direct Digital-to-RF Converter
RF Radio Frequency
RMS Root Mean Square
RQDAC Resistive Charge-based Digital-to-Analog Converter
SAW Surface Acoustic Wave
SC-R Switched-Capacitor Resistance
SDR Software-Defined Radio
SNR Signal-to-Noise Ratio
SPI Serial Peripheral Interface
SoC System-on-Chip
TDMA Time Division Multiple Access
TX Transmitter
UMTS Universal Mobile Telecommunications System
USB Universal Serial Bus
WCDMA Wideband Code-Division Multiple Access
WLAN Wireless Local-Area Network
ZOH Zero-Order Hold
Z Electrical Impedance
Another random document with
no related content on Scribd:
organs, as the lung, the testicle, the liver, the spleen, etc. The
dependence of miliary tuberculosis of the pia upon previously-
existing caseous or other inflammatory deposits in some part of the
body is acknowledged by most modern pathologists. Seitz3 states
that out of 130 cases, with autopsies, of adults, upon which his work
is based, such deposits were found in 93.5 per cent. General
constitutional weakness, either congenital or resulting from grave
disease or from overwork, from insufficient or unwholesome food,
and from bad hygienic surroundings, also favors the deposit of
tubercle in the meninges. Sometimes two or more predisposing
causes exist at once. Thus, a child born of tuberculous parents may
be fed with artificial diet instead of being nursed, or may live in a
house whose sanitary condition is bad. Hence the disease is
common among the poor, although by no means rare in the higher
classes of society. In some cases it is difficult or impossible to assign
any predisposing cause. A single child out of a numerous family may
be stricken with the disease, while the rest of the children, as well as
the parents and other ascendants, are healthy. For instance, while
writing this article I had under observation a little boy six years old
whose parents are living and healthy, with no pulmonary disease in
the family of either. The only other child, an older brother, is healthy.
While apparently in perfect health the child was attacked with
tubercular meningitis, and died in seventeen days with all the
characteristic symptoms of the disease. At the autopsy there was
found much injection of the cerebral pia everywhere, a large effusion
of lymph at the base of the brain and extending down the medulla,
abundance of miliary tubercles in the pia and accompanying the
vessels in the lateral regions of the hemispheres, lateral ventricles
distended with nearly clear fluid, ependyma smooth, choroid
plexuses covered with granulations, convolutions of brain much
flattened. Careful investigation, however, will usually enable us to
detect some lurking primary cause, either in the family predisposition
or in the history of the patient himself.
3 Die Meningitis Tuberculosa der Erwachsenen, von Dr. Johannes Seitz, Berlin, 1874,
p. 317.
Season appears to have but little influence on the production of the
disease. The largest number of cases is observed during winter and
spring, owing doubtless to the influence of the temperature and
weather, and to the exclusion from fresh air, in favoring the
development of tubercle and the scrofulous diathesis. Males, both
children and adults, are somewhat more frequently attacked than
females.

In regard to the exciting causes it may be said that where a

disposition to the deposit of tubercle exists, anything which tends to
lower the vitality of the individual is likely to hasten the event. In
infants with hereditary tendency to tubercle, an improper diet is
especially liable to develop meningeal tubercle. In older children,
besides unwholesome or insufficient food and unfavorable hygienic
surroundings, the acute diseases common to that period of life, such
as the eruptive fevers, intestinal disorders, whooping cough, etc.,
often act as immediate causes. Sometimes the development of the
disease may be traced to over-stimulation of the nervous system by
excessive study, often aided by imperfect ventilation or overheating
of the school-room. Caries of the temporal bone from disease of the
middle ear may act as an immediate cause of tubercular meningitis,
although simple meningitis is of course a more frequent result of that
condition. The disease has been known to follow injuries of the head
from blows or falls. In a larger number of cases the exciting cause is
not discoverable, especially when the meningeal affection is simply
an extension of the disease from some other part of the body, as the
lungs, the bronchial or mesenteric glands, etc. This is often the case,
both in adults and in children, when tubercular meningitis
complicates pulmonary consumption.

SYMPTOMS.—The disease is most frequently observed in children

between the ages of two and seven years. It is much less common in
adults, who are generally attacked between the ages of twenty and
thirty years. In the majority of cases the invasion of the malady is
preceded by a prodromic stage, usually occupying from a few days
to several weeks, though sometimes extending over a considerably
longer period. This stage probably represents the process of deposit
of miliary tubercles in the pia mater before their presence has given
rise to much structural change in the tissue. The characteristic
symptoms of the prodromic stage consist chiefly in an alteration of
the character and disposition of the patient, varying in extent in
different cases. In general, it may be said that he becomes sad,
taciturn, apathetic, irritable, indisposed to play, often sitting apart
from his companions, gazing in a strange way into vacancy. There is
diminution or loss of appetite and some emaciation. He is restless at
night, is disturbed by nightmare, or grinds his teeth. The digestion is
deranged. Usually there is constipation, but occasionally diarrhœa,
or these conditions may alternate with each other. Squinting and
twitching of the facial muscles are sometimes noticed. Headache
may occur early in this stage, but it is usually observed later, and it
then forms a prominent symptom. Vomiting is also frequent, usually
not preceded by nausea, sometimes provoked by sudden
movement, as in sitting up in bed, and is apt to occur when the
stomach contains little or no food. These symptoms vary much in
degree, and they are often so slight that they pass unnoticed by the
parents or friends. Occasionally the patient, if a child, will manifest a
strange perversity or an unusual disobedience, for which he is
perhaps punished under the belief that his misconduct is intentional.
In older children and in adults delirium, especially at night,
sometimes followed by delusions which may be more or less
permanent, is frequent at this stage. The above symptoms often
remit from time to time, and during the interval the patient may seem
to have recovered his health. The prodromic symptoms are rarely
altogether wanting in children, although they may have escaped
notice from lack of opportunity of observation on the part of the
physician. On the other hand, as Steffen4 justly observes, the most
characteristic symptoms may be present, and lead even an
experienced observer to a confident diagnosis of tubercular
meningitis during the early stage of a case of typhoid fever or of
cerebral congestion without tuberculosis.
4 “Meningitis Tuberculosa,” by A. Steffen, in Gerhardt's Handb. der Kinderkrankheiten,
5 B., 1ste Abth., 2te Hälfte, p. 465.
For convenience of description it is customary to divide the disease
proper, after the prodromal period, into three stages—viz. of
irritation, compression, and collapse. In some cases it is not difficult
to observe these divisions, but it must be borne in mind that in others
the symptoms do not follow any regular sequence, so that no
division is possible. In infants profound slumber may be the only
morbid manifestation throughout the entire disease. Steffen records
such a case, and I have seen two similar ones.

First Stage: The interval between the prodromic period and the first
stage is usually so gradual that no distinction between the two can
be detected. In other cases the disease is ushered in suddenly by
some striking symptom, such as an attack of general convulsions,
with dilated pupils and loss of consciousness. This is not often
repeated, though partial twitchings of the limbs or of the muscles of
the face may follow at intervals. In young children a comatose
condition, with unequal pupils, is apt to take the place of these
symptoms. The principal phenomena of the first stage are headache,
sensitiveness to light and sound, vomiting, and fever. The latter
varies much in intensity from time to time, but is not usually high, the
temperature seldom rising above 103° F., and usually, but not
always, higher at night than in the morning; but there is no
characteristic curve. The pulse varies in rate, but is usually slow and
irregular or intermittent. The respiration is irregular, with frequent
sighing. The tongue is dryish and covered with a thin white coat. The
bowels are costive. Delirium is frequent at night, and the sleep is
disturbed, the patient tossing about and muttering or crying out. The
eyes are half open during sleep. These symptoms become more
marked from day to day. The pain in the head is more frequent and
severe; the patient presses the hands to the forehead or rests the
head against some support if sitting up. During sleep he occasionally
utters a loud, sharp cry, without waking. There is increasing apathy,
and some intolerance of light, shown by an inclination to turn toward
the wall of the room or to lie with the face buried in the pillow. The
appetite is lost, the constipation becomes more obstinate, the
slowness and irregularity of the pulse persist. With the rapid
emaciation the belly sinks in, so that the spinal column can be easily
felt. Soon the child falls into a state of almost continual somnolence,
from which, however, he can be awakened in full consciousness,
and will answer correctly, generally relapsing again immediately into
slumber. His restlessness diminishes or ceases altogether, and he
lies continuously on the back with the head boring into the pillow. He
becomes more passive under the physician's examination, in strong
contrast to his previous irritability. At the end of a week or more from
the beginning of this stage symptoms of irritation of some of the
cerebral nerves begin to show themselves, in consequence of
pressure from the increasing exudation at the base of the brain and
into the ventricles. Strabismus (usually convergent), twitching of the
facial muscles and grimaces, grinding of the teeth, or chewing
movements of the mouth are noticed. The somnolence deepens into
sopor, from which it becomes more and more difficult to arouse the
patient, who gradually becomes completely insensible.

Notwithstanding the alarming and often hopeless condition which

this assemblage of symptoms indicates, intervals of temporary
amendment not unfrequently take place. The child may awake from
his lethargy, recognize those about him, converse rationally, take his
food with relish, and exhibit such symptoms of general improvement
that the parents and friends are led to indulge in fallacious hopes,
and sometimes the physician himself ventures to doubt the accuracy
of his diagnosis. Such hopes are of short duration; the unfavorable
symptoms always return after a brief interval. The duration of the first
stage may be reckoned at about one week.

Second Stage: This period is not separated from the preceding one
by any distinct change in symptoms. The patient lies in a state of
complete insensibility, from which he can no longer be aroused by
any appeal. The face is pale or of an earthen tint, the eyes are half
closed. If the anterior fontanel be still open, the integument covering
it is distended by the pressure beneath. Often one knee is flexed, the
opposite leg extended; one hand applied to the genitals, the other to
the head. Sometimes one leg or arm is alternately flexed and
extended. The head is apt to be retracted and bores into the pillow.
The pupils are dilated, though often unequal and insensible to light:
the sclerotica are injected; a gummy exudation from the Meibomian
glands forms on the edges of the lids. The patient sighs deeply from
time to time, and occasionally utters a loud, piercing cry. Paralysis,
and sometimes rigidity of one or more of the extremities, are often
observed, and occasionally there is an attack of general convulsions.
The pulse continues to be slow and irregular, the emaciation
progresses rapidly, and the abdomen is deeply excavated. The
discharges from the bladder and rectum are involuntary. The
average duration of the second stage is one week.

Third Stage: No special symptoms mark the passage of the second

stage into the third, which is characterized by coma, with complete
resolution of the limbs. The constipation frequently gives place to
moderate diarrhœa. The distended fontanel subsides, and often
sinks below the margin of the cranial bones. A striking feature of this
stage is a great increase in the rate of the pulse, the heart being
released from the inhibitory influence of the par vagum in
consequence of the complete paralysis of the latter from pressure.
The pulse varies in rapidity from 120 to 160 or more in the minute.
For the same reason the respiration also increases in frequency,
though not to the same degree. The eyelids are widely open; the
pupils are dilated and generally motionless, even when exposed to a
bright light. The eyes are rolled upward, so that only the lower half of
the iris is visible; the sclerotica is injected from exposure to the air
and dust. Convulsions may occur from time to time. Death
terminates the painful scene, usually in from twenty-four to forty-
eight hours, but sometimes the child lives on for days, unconscious,
of course, of suffering, though the afflicted parents and friends can
with difficulty be brought to believe it.

Certain points in the symptomatology of tubercular meningitis

demand especial consideration.

I have already observed that the division of the disease into definite
stages is purely arbitrary, and is employed here merely for
convenience of description; in fact, few cases pursue the typical
course. A period of active symptoms and another of depression can
often be observed, but these frequently alternate. Stupor and
paralysis may characterize the early stage, and symptoms of
irritation, with restlessness, screaming, and convulsions,
predominate toward the end. Certain characteristic symptoms may
be wholly or in part wanting, such as vomiting, constipation, or
stupor.

The temperature shows no changes which are characteristic of the

disease. Throughout its whole course it varies from time to time,
without uniformity, except that it usually rises somewhat toward
night. It seldom exceeds 102° or 103° F., unless shortly before
death, when it may rise to 104° F., or even higher, and may continue
to rise for a short time after death.

During the premonitory stage the pulse offers no unusual

characteristics. Its frequency is often increased, as is usual in any
indisposition during the period of childhood, but it preserves its
regularity. Toward the close of this period, and especially during the
first stage of the disease proper, a remarkable change takes place. It
becomes slow and irregular, the rate often diminishing below that in
health. The irregularity varies in character; sometimes the pulse
intermits, either at regular or irregular intervals. An inequality in the
strength of different pulsations is also observed. These peculiarities
of the circulation are due to the irritation of the medulla and the roots
of the par vagum, by which the inhibitory function of that nerve upon
the action of the heart is augmented. During the last period, on the
other hand, the increasing pressure on the vagus paralyzes its
function, and the heart, freed from its control, takes on an increased
action, the pulse rising to 120 beats, and often many more, in the
minute. Robert Whytt, in his interesting memoir,5 dates the beginning
of the second stage from the time that the pulse, being quick but
regular, becomes slow and irregular; the change again to the normal
frequency, or beyond it, marking the commencement of the third
stage.
5 “An Account of the Symptoms in the Dropsy of the Ventricles of the Brain,” in the
Works of Robert Whytt, M.D., published by his son, Edinb., 1768, p. 729.
In the early stage the respiration presents nothing abnormal, but
when the pulse becomes slow and irregular the breathing is similarly
affected. Sighing is very common in the prodromal period and first
stage. Toward the end of the second stage the increasing paralysis
of the respiratory centre gives rise to the phenomena known as the
Cheyne-Stokes respiration, consisting of a succession of respiratory
acts diminishing in force until there is a complete suspension of the
breathing, lasting from a quarter to three-quarters of a minute, when
the series begins again with a full inspiration. In general, the
variations in the rate of the respiration follow those of the pulse,
though the correspondence is not always exact.

In the early stage of the disease the pupils are usually contracted
and unequal. They are sluggish, but still respond to the stimulus of
light. At a later period they become gradually dilated, and react even
more slowly to light or not at all, the two eyes often differing in this
respect. Ophthalmoscopic examination frequently shows the
appearance of choked disc and commencing neuro-retinitis. In rare
cases tubercles are seen scattered over the fundus of the eye. They
are about the size of a small pin's head, of a yellowish color, and of
sharply-defined contour. Neuro-retinitis and choked disc are not, of
course, pathognomonic of tubercular meningitis, and choroidal
tubercles are so rarely seen as to be of little avail in diagnosis. In
fact, they are less frequent in this disease than in general
tuberculosis without meningitis. In twenty-six cases of tubercular
meningitis examined by Garlick at the London Hospital for Sick
Children they were found only once.6 The effect upon the conjunctiva
of the unclosed lids has been already described.
6 W. R. Gowers, M.D., Manual and Atlas of Medical Ophthalmoscopy, Philada., 1882,
p. 148. See, also, Seitz, op. cit., p. 347; Steffen, op. cit., pp. 452 and 472; and
“Tubercle of the Choroid,” Med. Times and Gazette, Oct. 21, 1882, p. 498.

The tongue is somewhat coated soon after the beginning of the

disease, and the breath is offensive. The appetite is lost, and there is
decided emaciation in many cases during the prodromic period. The
thirst is usually moderate. Vomiting is one of the most constant
symptoms during the first period, and its occurrence on an empty
stomach is characteristic of tubercular meningitis. It is not usually
preceded by nausea, and often takes place without effort, by mere
regurgitation, the rejected fluid consisting chiefly of bile mixed with
mucus. Although constipation is the most common condition in the
early stage, and is often rebellious to treatment, yet in some cases
diarrhœa is observed, which may mislead the physician in respect to
the diagnosis. From the beginning of the second stage, and
sometimes earlier, the discharges from the bowels and the bladder
are involuntary.

DURATION.—The duration of tubercular meningitis, apart from the

prodromic period, which often can hardly be determined, averages
from two weeks to two weeks and a half. In exceptional cases death
may take place in a few days or a week, and occasionally a patient
may linger for several weeks,7 the difference being apparently due to
the rapidity of the tubercular deposit and of the resulting
inflammation and exudation. The patient usually takes to his bed at
the beginning of the first stage, but he may be up during a part of the
day until the beginning of the second. In rare instances the child will
be about, and even out of doors, until a few days before death.
7 Such a case is reported by Michael Collins in the London Lancet, March 8, 1884.

PATHOLOGICAL ANATOMY.—The essential lesion of tubercular

meningitis consists in a deposit of miliary tubercles in the pia mater
of the brain, giving rise to inflammation of that membrane and
exudation of serum and pus. In the early stage both surfaces of the
pia are reddened and more or less thickened, and present an
opaline appearance, while between them—that is, in the meshes of
the pia—we find a colorless and transparent fluid which is effused in
greater or smaller amount, resembling jelly when viewed through the
arachnoid. These conditions are sometimes observable on the
convexity of the hemispheres, but are much more abundant on the
lateral surfaces, and especially at the base. More distinct evidence of
inflammation is shown by the presence of a yellowish or greenish-
yellow creamy deposit on the surface of the pia, consisting chiefly of
pus, which is also much more abundant at the base than elsewhere,
especially about the optic commissure, infundibulum, pons Varolii,
and the anterior surface of the medulla. The cranial nerves may be
deeply imbedded in the deposit, which often extends into the fissure
of Sylvius, gluing together the adjacent surfaces of the lobes, and
accompanies the vessels, forming narrow streaks along the sides of
the brain up to the convexity.

The miliary tubercles or granulations consist of semi-transparent

bodies, grayish or whitish in color, varying in size from that of the
head of the smallest pin, indeed almost invisible to the naked eye, to
that of a millet-seed (whence their name). Larger masses are
frequently seen, formed by the aggregation of smaller granulations.
The tubercles are usually found on the inner surface of the pia,
always in the immediate neighborhood of the blood-vessels, which
they accompany in their ramifications, and are also scattered, in
greater or less numbers, throughout the purulent exudation from the
surface of the pia. They are most abundant at the base of the brain,
ascending the sides along the course of the vessels. Sometimes,
though rarely, they are more abundant on the convexity. The total
number varies; it is usually very large, but sometimes only a limited
number exists, even in well-marked cases, and along with intense
inflammation of the pia. The granulations are found in different
degrees of development—sometimes all of them similar in color,
size, and consistency, at others in various stages of fatty
degeneration. The distribution may be symmetrical in the two
hemispheres or irregular. Under the microscope (after suitable
preparation of the part) the bacillus tuberculosus in considerable
numbers may be found in the pia, in places adjacent to the
arterioles.8
8 See a case reported by Y. Dawson in the London Lancet, April 12, 1884, in which
tubercles were visible only by the microscope with numerous bacilli.

The ventricles of the brain are usually distended with a clear or

opalescent, rarely bloody, fluid, the amount of which generally
corresponds to the intensity and extent of the meningeal
inflammation, although sometimes it is not above the normal
quantity. The two lateral ventricles are affected in an equal degree;
the third and fourth ventricles are more rarely implicated. According
to Huguenin,9 it is doubtful whether acute inflammation of the
ependyma takes place in tubercular meningitis. Steffen also10 says
that the ependyma is not inflamed, and that it is not the seat of the
deposit of tubercles. This latter statement is denied by other
authorities, and Huguenin is inclined to believe that they may exist in
that membrane. In the following case, under my care, abundant
granulations were found on the surface of the ependyma:
9 G. Huguenin, op. cit., p. 499.

10 Op. cit., p. 449.

Olaf M—— (male), æt. 8 years, born in Denmark, entered

Massachusetts General Hospital Sept. 13, 1881. Maternal
grandmother died of consumption; paternal grandfather lived to the
age of ninety-five years. One brother had some disease of hip.
Patient was the child of poor parents and lived in an unhealthy
suburb of Boston. During the two preceding winters he had a bad
cough. He was apparently well till four weeks before his entrance,
when he complained of bellyache, and became listless, but he was
out of doors ten days before he came to the hospital. It was noticed
that he was sensitive to sound. No vomiting, no diarrhœa, no
epistaxis, no cry; some cough. He had been somnolent, and was
observed to swing his arm over his head while asleep. June 14,
when first seen by me, he was lying on his back, unconscious, eyes
half closed, pupils dilated, jaw firmly closed, much emaciated, belly
retracted, left leg occasionally flexed and extended. No priapism.
The optic discs were reddened. June 15, there is some intelligence,
he answers questions; keeps one hand on the genitals. June 16,
pupils contracted, does not swallow. June 18, left eye divergent,
conjunctiva injected, whole surface livid, cries out occasionally. Died
at midnight.

FIG. 30.
Autopsy.—General lividity of surface, much emaciation. Much fine
arborescent injection on outer surface of dura mater. Numerous
Pacchionian bodies. Yellow matter beneath arachnoid along course
of vessels on each side of anterior lobes. Abundant fine granulations
along course of vessels on each anterior lobe, on upper margins of
median fissure, along fissure of Sylvius, and on choroid plexuses.
Very little lymph at base of brain. Six or eight ounces of serum from
lateral ventricles, and abundant fine transparent granules over
ependyma of both. Numerous opaque granulations in pia mater of
medulla oblongata. Surface of right pleura universally adherent.
Mucous membrane of bronchia much injected; a considerable
amount of pus flowed from each primary bronchus. No tubercles in
lungs nor in peritoneum. No ulcerations in intestines. No other
lesions.

The choroid plexuses are generally involved in the inflammatory

process, and are sometimes covered with yellow purulent
exudations. As in the above case, large numbers of tubercles may
be found in them, notwithstanding the opinion of Huguenin that their
number is always small.

The substance of the brain in the vicinity of the tubercular deposit is

generally found in a more or less œdematous condition, owing to the
obstruction of the circulation resulting from compression of the
vessels by the tubercles and effused lymph. Softening, sometimes
even to diffluence, not unfrequently occurs in the neighborhood of
the deposit, probably from ischæmia (necrobiosis). If there be any
considerable amount of exudation in the ventricles, the convolutions
are flattened by compression against the cranial bones.

The above-described lesions are not confined to the brain, but may
extend to the cerebellum, the pons, the medulla, and the spinal cord.
If examinations of the latter were more frequent in autopsies of this
disease, we should doubtless find, as has been done in some
instances, that the membranes often show the characteristic
alterations of tubercular meningitis, and even the presence of
granulations in the cord itself. The lesions may extend throughout
the cord, and are especially noticed in the dorsal region and in the
vicinity of the cauda equina. Their presence explains some of the
symptoms evidently due to spinal origin, such as retraction of the
head with rigidity of the neck and of the trunk, contractions of the
limbs, tetanic spasms, priapism, paralysis of the bladder and rectum,
etc., which are common in simple spinal meningitis.
The deposit of miliary tubercles in the pia mater, with little or no
accompanying meningitis, is met with in rare instances. The
tubercles are few in number, but vary in dimensions, being
sometimes united together in masses of considerable size, which are
frequently encysted. Beyond thickening and opacity of the
membrane, their presence seems to excite but little inflammatory
reaction, but they are generally accompanied by ventricular effusion
which by its pressure gives rise to characteristic symptoms.

The principal lesions found in other organs of the body consist of

tubercle in various stages of development, caseous matter, diseases
of the bones, etc. Miliary granulations are chiefly seen in the lungs,
peritoneum, intestinal mucous membrane, pleura, spleen, liver, and
kidneys. The bronchial and mesenteric glands often contain caseous
masses, some of which are broken down and suppurating. The
testicles sometimes present the same appearances. In adults, the
most frequent lesion which is found external to the brain is
pulmonary tuberculosis in a more or less advanced stage. Tubercles
are also sometimes present in the eye. Angel Money11 states that out
of 44 examinations made at the Hospital for Sick Children, London,
the meninges were the seat of gray granulations in 42. The choroid
(one or both) showed tubercles 14 times (right 3, left 5, both 6), and
11 times there were undoubted evidences of optic neuritis. Twice the
choroid was affected with tubercle when the meninges were free; in
one of these instances there was a mass of crude tubercle in the
cerebellum; in the other, although there were tubercles in the belly
and chest, there were none in the head. So that 12 times in 42 cases
of tubercles in the meninges there were tubercles in the choroid—i.e.
about 31 per cent.
11 “On the Frequent Association of Choroidal and Meningeal Tubercle,” Lancet, Nov.
10, 1883.

DIAGNOSIS.—In many cases tubercular meningitis offers but little

difficulty in the diagnosis. Although the symptoms, taken singly, are
not pathognomonic, yet their combination and succession, together
with their relation to the age, previous health, and antecedents of the
patient, are usually sufficient to lead us to a correct opinion. The
prodromic period of altered disposition (irritability of temper or
apathetic indifference), headache, constipation, vomiting, and
emaciation, followed by irregularity and slowness of the pulse,
sighing respiration, sluggishness and irregularity of the pupils; the
progress from somnolence to unconsciousness and coma; the
sudden lamentable cry; the convulsions and paralysis; the return of
rapid pulse and respiration in the last stage,—are characteristic of no
other disease. Our chief embarrassment arises during the insidious
approach of the malady, before its distinctive features are visible or
when some important symptom is absent. Its real nature is then apt
to be overlooked, and, in fact, in some cases it is impossible to
decide whether the symptoms are indicative of commencing cerebral
disease, or, on the other hand, are owing to typhoid fever, to a
simple gastro-intestinal irritation from error in diet, to worms in the
alimentary canal, to overwork in school, or to some other cause.
Under these circumstances the physician should decline giving a
positive opinion until more definite signs make their appearance. It
must be remembered that very important symptoms may be absent
in cases which are otherwise well marked. In all doubtful cases the
family history should, if possible, be obtained, especially whether
one or both parents or other near relatives have been consumptive
or have shown symptoms of scrofula or tuberculosis in any form, and
whether the patient himself has signs of pulmonary tuberculosis, of
enlarged or suppurating glands, or obstinate skin eruptions. The
presence or history of those conditions would add greatly to the
probability of tubercular meningitis.

The diseases for which tubercular meningitis is most liable to be

mistaken are acute simple meningitis, typhoid fever, acute gastro-
intestinal affections, eclampsia of infants and children, worms in the
intestines or stomach, the hydrencephaloid disease of Marshall Hall,
and cerebro-spinal meningitis.

Acute meningitis is distinguished from the tubercular disease by its

sudden invasion without prodromatous stage, by the acuteness and
intensity of the symptoms, the severity of the headache, the activity
of the delirium, the greater elevation of the temperature, and by its
brief duration, which rarely exceeds one week. In those exceptional
cases of tubercular meningitis in which the prodromal period is
absent or not observed and the course is unusually rapid, it would be
perhaps impossible to distinguish between the two diseases. A
family history of tubercle, or the discovery of the granulations in the
choroid by ophthalmoscopic examination, might save us from error
under such circumstances. The great rarity of idiopathic simple
meningitis should be remembered. Meningitis from disease of the
ear sometimes resembles the tubercular affection, but the history of
the attack, usually beginning with local pain and otorrhœa, will in
most cases prevent any confusion between the two forms of
disease.

The early period of typhoid often bears considerable resemblance to

that of tubercular meningitis. Headache, languor, restlessness, and
mild delirium are common to both. Typhoid can be distinguished by
the coated tongue, the diarrhœa, the enlargement of the spleen, the
tympanites, abdominal tenderness and gurgling, the eruption, and,
above all, by the characteristic temperature-curve, which, if
accurately observed, is conclusive. The course of typhoid fever is
comparatively uniform, while that of tubercular meningitis is often
extremely irregular. It should not be forgotten that the two diseases
may coexist.

The presence of worms in the alimentary canal may cause

symptoms somewhat like those of tubercular meningitis, and the
symptoms of the latter disease are occasionally erroneously
attributed to those parasites. The administration of an anthelmintic,
which should never be omitted in doubtful cases, will clear up all
uncertainty.

Cerebro-spinal meningitis is usually an epidemic, and therefore not

likely to be confounded with the tubercular disease. In sporadic
cases it can be recognized by its sudden onset and acute character,
by the eruption, and by the prominence of the spinal symptoms.
The so-called hydrencephaloid disease of Marshall Hall is a
condition of exhaustion and marasmus belonging to infancy, caused
by insufficient or unsuitable nourishment, by diarrhœa, and by the
injudicious depletive treatment so much in vogue in former times,
when the affection was much more common than at present. Some
of its symptoms, such as sighing respiration, stupor, pallor, and
dilated pupils, bear a certain resemblance to those of tubercular
meningitis, though it would be more easily confounded with chronic
hydrocephalus. The absence of constipation, headache,
convulsions, and vomiting, and the favorable results of suitable
nourishment and stimulants, serve to distinguish it from cerebral
disease.

Eclampsia, or sudden convulsion, is common in infants and young

children, and, since the occurrence of a fit may be the first or the
most striking symptom in tubercular meningitis, it is important to
ascertain its origin. In the majority of cases convulsions in children
arise from some peripheral irritation, such as difficult dentition,
worms in the alimentary canal, constipation, fright, etc., acting
through the reflex function of the spinal cord, which is unusually
sensitive in the early period of life. The absence of previous
symptoms, and the discovery of the source of the irritation, with the
favorable effect of its removal by appropriate treatment, will in most
cases suffice to eliminate structural disease of the brain. In others
we must withhold a positive opinion for a reasonable time in order to
ascertain whether more definite symptoms follow. Convulsions also
occasionally form the initial symptom of the eruptive fevers,
especially scarlatina. Here the absence of prodromal symptoms, and
the speedy appearance of those belonging to the exanthematous
affection, will remove all sources of doubt. Convulsions, with or
without coma, occurring in the early stage of acute renal
inflammations, may simulate the symptoms of tubercular meningitis.
An examination of the urine will show the true nature of the disease.

In addition to the above diseases there are some cerebral affections

of uncertain pathology which resemble tubercular meningitis, but
which are not generally fatal. As Gee justly remarks,12 “Every
practitioner from time to time will come across an acute febrile
disease accompanied by symptoms which seem to point
unmistakably to some affection of the brain, there being every
reason to exclude the notion of suppressed exanthemata or
analogous disorders. After one or several weeks of coma, delirium,
severe headache, or whatever may have been the prominent
symptom, the patient recovers, and we are left quite unable to say
what has been the matter with him. To go more into detail, I could not
do otherwise than narrate a series of cases which would differ from
each other in most important points, and have nothing in common
excepting pyrexia and brain symptoms. There is, generally,
something wanting which makes us suspect that we have not to do
with tubercular meningitis. Brain fever is as good a name as any
whereby to designate these different anomalies; cerebral congestion,
which is more commonly used, involves an explanation which is
probably often wrong, and certainly never proved to be right.” No
doubt such cases are occasionally cited as examples of recovery
from tubercular meningitis.
12 “Tubercular Meningitis,” by Samuel Jones Gee, M.D., in Reynolds's System of
Medicine, Philada., 1879, vol. i. p. 832.

PROGNOSIS.—Although there are on record undoubted instances of

recovery from tubercular meningitis, yet their number is so small that
practically the prognosis is fatal. It is safe to say that in almost all the
reported cases of recovery the diagnosis was erroneous.13 Even
should the patient survive the attack, he is usually left with paralyzed
limbs and impaired mental faculties, and dies not long afterward from
a recurrence of the disease or from tuberculosis of the lungs or other
organs.
13 Hahn, “Recherches sur la Méningite tuberculeuse et sur le Traitement de cette
Maladie” (Arch. gén. de méd., 4e Série, vols. xx. and xxi.), claims to have cured 7
cases, but of 5 of them there is no evidence that they were examples of tubercular
meningitis at all. The subject of the curability of tubercular meningitis is ably treated
by Cadet de Gassicourt (Traité clinique des Maladies de l'Enfance, vol. iii., Paris,
1884, p. 553 et seq.). His conclusion is that most of the alleged cures are cases of
 meningitis of limited extent, arising from the presence of tubercular tumors, syphilitic
gummata, cerebral scleroses, and neoplasms of various kinds.

TREATMENT.—In view of the fatality of the disease, and of its frequent

occurrence in childhood, the prophylactic treatment is of great
importance. Every effort should be made to protect children whose
parents or other near relatives are tuberculous or scrofulous, and
who are themselves delicate, puny, or affected with any
constitutional disorder, from tubercular meningitis, by placing them in
the best possible hygienic conditions. Pure air, suitable clothing,
wholesome and sufficient food, and plenty of out-of-door exercise
are indispensable. Sedentary amusements and occupations should
be sparingly allowed. Especial pains should be taken to prevent
fatigue by much study, and school-hours should be of short duration.
The hygiene of the school-room is of paramount importance, and if
its ventilation, temperature, and light are not satisfactory, the child
should not be permitted to enter it. The bed-chamber should be well
ventilated night and day. A sponge-bath, cold or tepid according to
the season or to the effect on the patient, should be given daily,
followed by friction with a towel. The bowels must be kept regular by
appropriate diet if possible, or by simple laxatives, such as magnesia
or rhubarb. For delicate, pale children some preparation of iron will
be useful. The choice must be left to the practitioner, but one of the
best in such cases is the tartrate of iron and potassium, of which
from two to six grains, according to the age, may be given three
times daily after meals. Cod-liver oil is invaluable for scrofulous
patients or where there is a lack of nutrition. A teaspoonful, given
after meals, is a sufficient dose, and it is usually taken without
difficulty by children, or if there be much repugnance to it some one
of the various emulsions may be tried in proportionate dose. Along
with this, iodide of iron will in many cases be found useful or as a
substitute for the oil when the latter cannot be borne. It is best given
in the form of the officinal syrup, in the dose of from five to twenty
drops. Change of air is useful in stimulating the nutritive functions,
and a visit to the seashore or mountains during warm weather will
often be followed by general improvement.
Since it is not possible to arrest the disease when once begun, the
efforts of the physician must be directed toward relieving the
sufferings of the patient as far as possible. In the early period the
restlessness at night and inability to sleep will call for sedatives, such
as the bromide of sodium or of potassium, in the dose of ten or
fifteen grains at bedtime or oftener. This should be well diluted with
water, sweetened if necessary. The addition of five to twenty drops of
the tincture of hyoscyamus increases the effect. Sometimes chloral
hydrate, either alone or combined with the bromide when the latter
fails, will procure quiet sleep. From five to ten grains may be given at
a dose, according to the age. Compresses wet with spirit and water
or an ice-cap may be applied to the head if there be much pain in
that region, or it may be necessary to give opium in some of its forms
by the mouth, such as the tincture or fluid extract, in doses of from
one to five drops. Constipation is best overcome by means of
calomel in three- to five-grain doses, to which may be added, when
necessary, an equal amount of jalap powder, or an enema of
soapsuds may be administered. Active purging should be avoided.
Liquid nourishment, such as milk, gruel of oatmeal, farina, or barley,
beef-tea, broths, etc., must be given in moderate quantities at
intervals of a few hours so long as the patient is able to swallow.
Occasional sponging of the whole surface with warm or cool water,
and scrupulous attention to cleanliness after defecation, especially
when control of the sphincters is lost, will add to his comfort. He
should occupy a large and well-ventilated chamber, from which all
persons whose presence is not necessary for his care and comfort
should be excluded. He should be protected from noise and from
bright light, and should lie on a bed of moderate width for
convenience of tending.

There is no specific treatment at present known which is likely to be

of any benefit in this disease, any more than in tuberculosis of other
organs than the brain. Common experience has shown that mercury,
which formerly had so high a reputation in the treatment of cerebral
diseases of early life, not only fails completely, but adds to the
sufferings of the patient when pushed to salivation. The iodide of
potassium is recommended by almost all writers, but, so far as I