Shoji (digital research team, AT&T Bell Laboratories) reveals the most recent research findings from AT&T, and the technology behind digital circuit design and implementation in telecommunications, semiconductor, and network industries. His discussions include: overview of theory, models of devices and low frequency circuits, digital circuit theory including inductance, microstates, submicrostates, local time, and complexity of interconnects. Annotation c. by Book News, Inc., Portland, Or.
"synopsis" may belong to another edition of this title.
To execute the complex signal processing of a modern computer system, a huge number of transistors and large silicon chips are required. State-of-the-art MOS integrated circuits contain several million field-effect transistors within an area of about 1 cm x 1 cm. Active devices such as MOSFETs and BJTs scale down to the micron and to the submicron sizes, and their switching speed has increased to the limit where many new physical limits emerge. Logic gates switch from the high to the low logic states in less than 100 ps. In that time an electromagnetic signal in free space propagates only 3 cm. A signal excited on a lossy transmission line fabricated on a dielectric propagates slower than that, and much effort has been concentrated on reducing the delay to the theoretical limit. Even in the limit, however, a signal sent out from a corner of the chip does not reach another corner in zero time. The fact that an electrical signal cannot propagate instantly from the source to the destination is crucial in determining the chip performance. It means that the signal propagation must be treated accurately according to Maxwell's electromagnetic theory. This theoretical development is carried out by incorporating inductance into the conventional integrated circuit theory. When that is done, a circuit that has size Lambda (cm) has signal delay Lambda / c, where c = 3 x 10 to the 10th cm/sec is the speed of light. The unrealistic feature of the conventional theory that the signal propagates at infinite speed is removed. Thus the effect of including inductance is to build a relativistic circuit theory. This impressive adjective heightens our interest in investigating these new issues.
Inductance is peculiar in a circuit. There are two kinds of inductances: a shielded and an unshielded, or naked, inductance. The value of the shielded inductance is not related to the size of the circuit. The framework of circuit theory which includes shielded inductances is well established. Here we study the properties of a circuit that include naked inductance, which creates a magnetic field in the region of the circuit. The magnitude of the naked inductance is related to the physical dimensions of the circuit. Inclusion of naked inductance adds many new features to circuit theory.
In the operational regime where the signal propagation delay becomes the key issue, the transient phenomena within electron devices are unaffected by the relativistic complications. Owing to the extremely small device size achieved by the scaledown, electron devices can be understood using classical physics, even though device operation is based on such a slow phenomenon as current-carrier diffusion. In this book I intend to develop an integrated circuit theory of classical electron devices in a relativistic interconnect environment.
This text presents an opportunity to study many practical circuit problems in great depth, in terms of basic physics. Many existing circuit design aids have been produced without critically looking into the basic electrical phenomena of integrated circuits at high speeds. One issue raised in this book is quite fundamental: How far is the use of electrostatic potential,or voltage, justified in describing circuit operation? As Maxwell's theory teaches, a conservative electrostatic potential does not exist if the electromotive force created by a time-varying magnetic field becomes significant. An integrated circuit in the limit of high frequencies is immersed in the time-dependent magnetic field the circuit itself makes, and therefore the definition of the circuit's node voltage becomes ambiguous. Conventional integrated circuit theory, that is, the description of the charge and discharge capacitance through resistance, assumes the universal existence of voltage. A new high-speed integrated circuit theory should be built by including the effects of naked inductance into conventional integrated circuit theory, and I intend to show that self-inductance can be included consistently into integrated circuit theory. In this new theory some node voltages lose their physical meaning and we suffer from that inconvenience, but we find a pleasing general consistency, in that integration of self-inductance into the theory can be carried out quite neatly using the many peculiar features of scaled-down integrated circuits. The extension of conventional circuit theory is limited, however, by a fundamental difficulty of including mutual inductance in the theory. A study of this difficulty shows that it constitutes the natural and fundamental limit of the equivalent circuit model.
Integrated circuits are very complex devices. On an integrated circuit chip, various interactions between the circuits occur, which affect circuit performance fundamentally. This complexity becomes significant even at the lower switching speeds at which the relativistic effects are still not significant. This effect originates from nature's inability to maintain the separate identity of excitations, expressed as the thermodynamic law of irreversibility. The structures that exist on an integrated circuit chip tend to mix signals, creating complex and often undesirable effects. For our theory to be practical this complexity must be included at high frequencies. Although these complex effects are fundamental to high-speed integrated circuit theory, they have not been studied enough. Some material related to this subject is presented in this book.
Fundamental to this new high-speed integrated circuit theory is the equivalent circuit. Equivalent circuits are theoretical models of the physical objects of our present research or development. Modeling an integrated circuit by an equivalent circuit is not a trivial task, since many devices and passive components are densely packed, and no circuit is totally independent of the others. One issue addressed in this book is how to draw simple and physically accurate equivalent circuits. This issue deserves attention, since equivalent circuits are the most powerful construct of electronic circuit theory. Equivalent circuits enable us to model any physical phenomena that exist in nature, and in some cases maybe even those that are merely imagined by human minds.
AcknowledgmentsThe two basic ideas on which this work was founded came from two giants of modern electronics by whom I was educated in the 1960s, Professor Aldert van der Ziel and Dr. William Shockley. The first idea, that an electronic circuit contains a lot of physics, that must be clarified and interpreted by a complete mathematical analysis, is due to Professor van der Ziel. The origin of his viewpoint will be clear if we remember how he discovered the mechanisms of parametric amplification. The second idea, that all the semiconductor devices are variations on one fundamental device, the electron triode, and that semiconductor circuits using different devices are all very much alike, is my interpretation of what I learned from Dr. Shockley. I am surprised to find how these great mentors have influenced their student so deeply and so fundamentally for so long a time. I am eternally grateful to them.
Professor O. Wing agreed to review the manuscript at a very early phase. He gave me a lot of moral support in carrying out the work. Professor E. Friedman of the University of Rochester, Professor S. Kang of the University of Illinois, and Professor A. Willson of the University of California, Los Angeles, reviewed the original manuscript and gave valuable comments, suggestions for improvement, and encouragement.I am grateful to Dr. M. D. Alston, who read the entire manuscript, corrected English, improved technical interpretations, and gave valuable technical comments. Ms. H. Khorramabadi and Messrs. M. Pinto, T. Wik, V. Califano, and R. Marley reviewed an early version of the manuscript and gave me helpful comments. Ms. Khorramabadi, Mr. Pinto, and Mr. Wik have been my co-workers for many years; we share many happy memories of technical successes. While I was writing this book, I had many enlightening discussions with Mr. Kwok Ng, who was also writing a book. His book was recently published, and is referred to in this one.
My co-workers in AT&T Bell Laboratories, Cedarcrest, Pennsylvania, especially Mr. R. Krambeck and Ms. J. Sabnis, have supplied a lot of technical problems that became the nucleus of this book. The problem of Section 3.5 was one of the design problems Ms. Sabnis and I had to solve in the late 1980s. The work was supported by the management of the Computing Science Research Center of the AT&T Bell Laboratories, especially by T. G. Szymanski, R. Sethi, P. J. Weinberger, and A. G. Fraser. They have supported the fundamental theoretical study of electrical phenomena in integrated circuits for many years now. Their faith in this work was essential to its completion. I am grateful to the editors of the Addison-Wesley Publishing Company, Messrs. John Wait and Simon Yates, who kindly offered to publish this work, and especially to Ms. Avanda Peters, who worked out the difficult problems of printing a book that contains many complicated mathematical formulas and illustrations.
02016348From the Back Cover:
telecommunications, semiconductor, and network industries. It is essential for engineers involved in VLSI and integrated circuit design to become informed about this emerging technology.
Written by one of the leaders of the digital circuit development team at AT&T Bell Laboratories, this book reveals much of their groundbreaking work. It covers all facets of the technology, from basic theories of physics to a practical guide for designing and implementing digital circuits.
You will come to understand how to include interconnect inductance into the conventional integrated circuit theory. This book also provides an in-depth examination of how interactions between circuits affect performance, a significant factor in high-speed switching. The book also demonstrates how to draw simple and physically accurate equivalent circuits, a skill fundamental to the study of digital circuit technology. Covering these and numerous other important topics, High Speed Digital Circuits will bring you up to date on the most current digital circuit research and will enable you to remain at the forefront of circuit design.
"About this title" may belong to another edition of this title.
Book Description Hardcover. Book Condition: New. BRAND NEW US edition / [Hardcover] / FREE UPGRADE to FedEx, UPS or Priority Mail / receive your book within 1-4 business days! / [clean wrapped, well protected] [Authentic edition exactly the same as the Official Listing] [ships within 1 business day] / Free tracking number / Genuine US Bookstore! Get your book in perfect condition! We also ship international via FedEx, UPS, or USPS Express!. Bookseller Inventory # 423520171211000505
Book Description Addison Wesley Publishing Company, 1996. Hardcover. Book Condition: New. Bookseller Inventory # DADAX020163483X
Book Description Prentice Hall, 1996. Hardcover. Book Condition: New. book. Bookseller Inventory # M020163483X