Brock J.

LaMeres

Introduction to
Logic Circuits
& Logic Design
with Verilog
INTRODUCTION TO LOGIC CIRCUITS &
LOGIC DESIGN WITH VERILOG
INTRODUCTION TO LOGIC CIRCUITS &
LOGIC DESIGN WITH VERILOG
1 ST E DITION

Brock J. LaMeres
Brock J. LaMeres
Department of Electrical & Computer Engineering
Montana State University
Bozeman, MT, USA

ISBN 978-3-319-53882-2 ISBN 978-3-319-53883-9 (eBook)

DOI 10.1007/978-3-319-53883-9

Library of Congress Control Number: 2017932539

# Springer International Publishing AG 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. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

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
Preface
The purpose of this new book is to fill a void that has appeared in the instruction of digital circuits
over the past decade due to the rapid abstraction of system design. Up until the mid-1980s, digital
circuits were designed using classical techniques. Classical techniques relied heavily on manual design
practices for the synthesis, minimization, and interfacing of digital systems. Corresponding to this design
style, academic textbooks were developed that taught classical digital design techniques. Around 1990,
large-scale digital systems began being designed using hardware description languages (HDLs) and
automated synthesis tools. Broad-scale adoption of this modern design approach spread through the
industry during this decade. Around 2000, hardware description languages and the modern digital
design approach began to be taught in universities, mainly at the senior and graduate level. There
were a variety of reasons that the modern digital design approach did not penetrate the lower levels of
academia during this time. First, the design and simulation tools were difficult to use and overwhelmed
freshman and sophomore students. Second, the ability to implement the designs in a laboratory setting
was infeasible. The modern design tools at the time were targeted at custom integrated circuits, which
are cost and time prohibitive to implement in a university setting. Between 2000 and 2005, rapid
advances in programmable logic and design tools allowed the modern digital design approach to be
implemented in a university setting, even in lower level courses. This allowed students to learn the
modern design approach based on HDLs and prototype their designs in real hardware, mainly field
programmable gate arrays (FPGAs). This spurred an abundance of textbooks to be authored teaching
hardware description languages and higher levels of design abstraction. This trend has continued until
today. While abstraction is a critical tool for engineering design, the rapid movement toward teaching only
the modern digital design techniques has left a void for freshman and sophomore level courses in digital
circuitry. Legacy textbooks that teach the classical design approach are outdated and do not contain
sufficient coverage of HDLs to prepare the students for follow-on classes. Newer textbooks that teach
the modern digital design approach move immediately into high-level behavioral modeling with minimal
or no coverage of the underlying hardware used to implement the systems. As a result, students are not
being provided the resources to understand the fundamental hardware theory that lies beneath the
modern abstraction such as interfacing, gate level implementation, and technology optimization.
Students moving too rapidly into high levels of abstraction have little understanding of what is going
on when they click the “compile & synthesize” button of their design tool. This leads to graduates who can
model a breadth of different systems in an HDL, but have no depth into how the system is implemented in
hardware. This becomes problematic when an issue arises in a real design, and there is no foundational
knowledge for the students to fall back on in order to debug the problem.

This new book addresses the lower level foundational void by providing a comprehensive, bottoms-
up, coverage of digital systems. This book begins with a description of lower level hardware including
binary representations, gate-level implementation, interfacing, and simple combinational logic design.
Only after a foundation has been laid in the underlying hardware theory is the Verilog language
introduced. The Verilog introduction gives only the basic concepts of the language in order to model,
simulate, and synthesize combinational logic. This allows the students to gain familiarity with the
language and the modern design approach without getting overwhelmed by the full capability of the
language. This book then covers sequential logic and finite state machines at the structural level. Once
this secondary foundation has been laid, the remaining capabilities of Verilog are presented that allow
sophisticated, synchronous systems to be modeled. An entire chapter is then dedicated to examples of
sequential system modeling, which allows the students to learn by example. The second part of this
textbook introduces the details of programmable logic, semiconductor memory, and arithmetic circuits.
This book culminates with a discussion of computer system design, which incorporates all of the

v
vi • Preface

knowledge gained in the previous chapters. Each component of a computer system is described with an
accompanying Verilog implementation, all while continually reinforcing the underlying hardware beneath
the HDL abstraction.

Written the Way It Is Taught

The organization of this book is designed to follow the way in which the material is actually learned.
Topics are presented only once sufficient background has been provided by earlier chapters to fully
understand the material. An example of this learning-oriented organization is how the Verilog language is
broken into two chapters. Chapter 5 presents an introduction to Verilog and the basic constructs to model
combinational logic. This is an ideal location to introduce the language because the reader has just
learned about combinational logic theory in Chap. 4. This allows the student to begin gaining experience
using the Verilog simulation tools on basic combinational logic circuits. The more advanced constructs of
Verilog such as sequential modeling and test benches are presented in Chap. 8 only after a thorough
background in sequential logic is presented in Chap. 7. Another example of this learning-oriented
approach is how arithmetic circuits are not introduced until Chap. 12. While technically the arithmetic
circuits in Chap. 12 are combinational logic circuits and could be presented in Chap. 4, the student does
not have the necessary background in Chap. 4 to fully understand the operation of the arithmetic circuitry
so its introduction is postponed.

This incremental, just-in-time presentation of material allows the book to follow the way the material
is actually taught in the classroom. This design also avoids the need for the instructor to assign sections
that move back-and-forth through the text. This not only reduces course design effort for the instructor
but allows the student to know where they are in the sequence of learning. At any point, the student
should know the material in prior chapters and be moving toward understanding the material in
subsequent ones.
An additional advantage of this book’s organization is that it supports giving the student hands-on
experience with digital circuitry for courses with an accompanying laboratory component. The flow is
designed to support lab exercises that begin using discrete logic gates on a breadboard and then move
into HDL-based designs implemented on off-the-shelf FPGA boards. Using this approach to a laboratory
experience gives the student experience with the basic electrical operation of digital circuits, interfacing,
and HDL-based designs.

Learning Outcomes
Each chapter begins with an explanation of its learning objective followed by a brief preview of the
chapter topics. The specific learning outcomes are then presented for the chapter in the form of concise
statements about the measurable knowledge and/or skills the student will possess by the end of the
chapter. Each section addresses a single, specific learning outcome. This eases the process of
assessment and gives specific details on student performance. There are 600+ exercise problems
and concept check questions for each section tied directly to specific learning outcomes for both
formative and summative assessment.

Teaching by Example
With over 200 worked examples, concept checks for each section, 200+ supporting figures, and 600+
exercise problems, students are provided with multiple ways to learn. Each topic is described in a clear,
concise written form with accompanying figures as necessary. This is then followed by annotated worked
examples that match the form of the exercise problems at the end of each chapter. Additionally, concept
check questions are placed at the end of each section in this book to measure the student’s general
 Preface • vii

understanding of the material using a concept inventory assessment style. These features provide the
student multiple ways to learn the material and build an understanding of digital circuitry.

Course Design
This book can be used in multiple ways. The first is to use the book to cover two, semester-based
college courses in digital logic. The first course in this sequence is an introduction to logic circuits and
covers Chaps. 1, 2, 3, 4, 5, 6, and 7. This introductory course, which is found in nearly all accredited
electrical and computer engineering programs, gives students a basic foundation in digital hardware and
interfacing. Chapters 1, 2, 3, 4, 5, 6 and 7 only cover relevant topics in digital circuits to make room for a
thorough introduction to Verilog. At the end of this course, students have a solid foundation in digital
circuits and are able to design and simulate Verilog models of concurrent and hierarchical systems. The
second course in this sequence covers logic design using Chaps. 8, 9, 10, 11, 12, and 13. In this second
course, students learn the advanced features of Verilog such as procedural assignments, sequential
behavioral modeling, system tasks, and test benches. This provides the basis for building larger digital
systems such as registers, finite state machines, and arithmetic circuits. Chapter 13 brings all of the
concepts together through the design of a simple 8-bit computer system that can be simulated and
implemented using many off-the-shelf FPGA boards.
This book can also be used in a more accelerated digital logic course that reaches a higher level of
abstraction in a single semester. This is accomplished by skipping some chapters and moving quickly
through others. In this use model, it is likely that Chap. 2 on numbers systems and Chap. 3 on digital
circuits would be quickly referenced but not covered in detail. Chapters 4 and 7 could also be covered
quickly in order to move rapidly into Verilog modeling without spending significant time looking at the
underlying hardware implementation. This approach allows a higher level of abstraction to be taught but
provides the student with the reference material so that they can delve in the details of the hardware
implementation if interested.
All exercise and concept problems that do not involve a Verilog model are designed so that they can
be implemented as a multiple choice or numeric entry question in a standard course management
system. This allows the questions to be automatically graded. For the Verilog design questions, it is
expected that the students will upload their Verilog source files and screenshots of their simulation
waveforms to the course management system for manual grading by the instructor or teaching assistant.

Instructor Resources
Instructors adopting this book can request a solution manual that contains a graphic-rich description
of the solutions for each of the 600+ exercise problems. Instructors can also receive the Verilog solutions
and test benches for each Verilog design exercise. A complementary lab manual has also been
developed to provide additional learning activities based on both the 74HC discrete logic family and
an off-the-shelf FPGA board. This manual is provided separately from the book in order to support the
ever-changing technology options available for laboratory exercises.

Bozeman, MT, USA Brock J. LaMeres

Acknowledgments

Dr. LaMeres is eternally grateful to his family for their support of this project. To JoAnn, your love
and friendship makes everything possible. To Alexis, your kindness and caring brings joy to my
heart. To Kylie, your humor and spirit fills me with laughter and pride. Thank you so much.

Dr. LaMeres would also like to thank the 400+ engineering students at Montana State University
that helped proof read this book in preparation for the first edition.

ix
Contents
1: INTRODUCTION: ANALOG VS. DIGITAL ........................................................... 1
1.1 DIFFERENCES BETWEEN ANALOG AND DIGITAL SYSTEMS ............................................. 1
1.2 ADVANTAGES OF DIGITAL SYSTEMS OVER ANALOG SYSTEMS ........................................ 2
2: NUMBER SYSTEMS ............................................................................................ 7
2.1 POSITIONAL NUMBER SYSTEMS ................................................................................ 7
2.1.1 Generic Structure ........................................................................................ 8
2.1.2 Decimal Number System (Base 10) ........................................................... 9
2.1.3 Binary Number System (Base 2) ................................................................ 9
2.1.4 Octal Number System (Base 8) .................................................................. 10
2.1.5 Hexadecimal Number System (Base 16) ................................................... 10
2.2 BASE CONVERSION ................................................................................................. 11
2.2.1 Converting to Decimal ................................................................................. 11
2.2.2 Converting From Decimal ........................................................................... 14
2.2.3 Converting Between 2n Bases .................................................................... 17
2.3 BINARY ARITHMETIC ................................................................................................ 21
2.3.1 Addition (Carries) ........................................................................................ 21
2.3.2 Subtraction (Borrows) ................................................................................. 22
2.4 UNSIGNED AND SIGNED NUMBERS ............................................................................ 23
2.4.1 Unsigned Numbers ..................................................................................... 24
2.4.2 Signed Numbers ......................................................................................... 24
3: DIGITAL CIRCUITRY AND INTERFACING .......................................................... 37
3.1 BASIC GATES ......................................................................................................... 37
3.1.1 Describing the Operation of a Logic Circuit ................................................ 37
3.1.2 The Buffer .................................................................................................... 39
3.1.3 The Inverter ................................................................................................. 40
3.1.4 The AND Gate ............................................................................................. 40
3.1.5 The NAND Gate .......................................................................................... 41
3.1.6 The OR Gate ............................................................................................... 41
3.1.7 The NOR Gate ............................................................................................ 41
3.1.8 The XOR Gate ............................................................................................. 42
3.1.9 The XNOR Gate .......................................................................................... 43
3.2 DIGITAL CIRCUIT OPERATION .................................................................................... 44
3.2.1 Logic Levels ................................................................................................ 44
3.2.2 Output DC Specifications ............................................................................ 45
3.2.3 Input DC Specifications ............................................................................... 46
3.2.4 Noise Margins ............................................................................................. 47
3.2.5 Power Supplies ........................................................................................... 48
3.2.6 Switching Characteristics ............................................................................ 51
3.2.7 Data Sheets ................................................................................................. 51

xi
xii • Contents

3.3 LOGIC FAMILIES ...................................................................................................... 56

3.3.1 Complementary Metal Oxide Semiconductors (CMOS) ............................. 56
3.3.2 Transistor-Transistor Logic (TTL) ................................................................ 65
3.3.3 The 7400 Series Logic Families ................................................................. 67
3.4 DRIVING LOADS ...................................................................................................... 71
3.4.1 Driving Other Gates .................................................................................... 71
3.4.2 Driving Resistive Loads .............................................................................. 73
3.4.3 Driving LEDs ............................................................................................... 75
4: COMBINATIONAL LOGIC DESIGN ..................................................................... 81
4.1 BOOLEAN ALGEBRA ................................................................................................ 81
4.1.1 Operations ................................................................................................... 82
4.1.2 Axioms ......................................................................................................... 82
4.1.3 Theorems .................................................................................................... 83
4.1.4 Functionally Complete Operation Sets ....................................................... 98
4.2 COMBINATIONAL LOGIC ANALYSIS .............................................................................. 99
4.2.1 Finding the Logic Expression from a Logic Diagram .................................. 99
4.2.2 Finding the Truth Table from a Logic Diagram ............................................ 100
4.2.3 Timing Analysis of a Combinational Logic Circuit ...................................... 101
4.3 COMBINATIONAL LOGIC SYNTHESIS ............................................................................ 103
4.3.1 Canonical Sum of Products ........................................................................ 103
4.3.2 The Minterm List (Σ) .................................................................................... 104
4.3.3 Canonical Product of Sums (POS) ............................................................. 106
4.3.4 The Maxterm List (Π) .................................................................................. 108
4.3.5 Minterm and Maxterm List Equivalence ..................................................... 110
4.4 LOGIC MINIMIZATION ................................................................................................ 112
4.4.1 Algebraic Minimization ................................................................................ 112
4.4.2 Minimization Using Karnaugh Maps ........................................................... 113
4.4.3 Don’t Cares ................................................................................................. 125
4.4.4 Using XOR Gates ........................................................................................ 126
4.5 TIMING HAZARDS & GLITCHES .................................................................................. 129
5: VERILOG (PART 1) .............................................................................................. 141
5.1 HISTORY OF HARDWARE DESCRIPTION LANGUAGES ..................................................... 142
5.2 HDL ABSTRACTION ................................................................................................ 145
5.3 THE MODERN DIGITAL DESIGN FLOW ........................................................................ 149
5.4 VERILOG CONSTRUCTS ............................................................................................ 152
5.4.1 Data Types .................................................................................................. 153
5.4.2 The Module ................................................................................................. 156
5.4.3 Verilog Operators ........................................................................................ 159
5.5 MODELING CONCURRENT FUNCTIONALITY IN VERILOG .................................................. 164
5.5.1 Continuous Assignment .............................................................................. 164
5.5.2 Continuous Assignment with Logical Operators ......................................... 164
5.5.3 Continuous Assignment with Conditional Operators .................................. 165
5.5.4 Continuous Assignment with Delay ............................................................ 167
 Contents • xiii

5.6 STRUCTURAL DESIGN AND HIERARCHY ...................................................................... 170

5.6.1 Lower-Level Module Instantiation ............................................................... 170
5.6.2 Gate Level Primitives .................................................................................. 172
5.6.3 User-Defined Primitives .............................................................................. 173
5.6.4 Adding Delay to Primitives .......................................................................... 174
5.7 OVERVIEW OF SIMULATION TEST BENCHES ................................................................ 175
6: MSI LOGIC ........................................................................................................... 181
6.1 DECODERS ............................................................................................................ 181
6.1.1 Example: One-Hot Decoder ........................................................................ 181
6.1.2 Example: 7-Segment Display Decoder ....................................................... 184
6.2 ENCODERS ............................................................................................................ 188
6.2.1 Example: One-Hot Binary Encoder ............................................................. 188
6.3 MULTIPLEXERS ....................................................................................................... 190
6.4 DEMULTIPLEXERS .................................................................................................... 193
7: SEQUENTIAL LOGIC DESIGN ............................................................................ 199
7.1 SEQUENTIAL LOGIC STORAGE DEVICES ..................................................................... 199
7.1.1 The Cross-Coupled Inverter Pair ................................................................ 199
7.1.2 Metastability ................................................................................................ 200
7.1.3 The SR Latch .............................................................................................. 202
7.1.4 The S’R’ Latch ............................................................................................. 205
7.1.5 SR Latch with Enable .................................................................................. 208
7.1.6 The D-Latch ................................................................................................. 209
7.1.7 The D-Flip-Flop ........................................................................................... 211
7.2 SEQUENTIAL LOGIC TIMING CONSIDERATIONS ............................................................. 214
7.3 COMMON CIRCUITS BASED ON SEQUENTIAL STORAGE DEVICES .................................... 216
7.3.1 Toggle Flop Clock Divider ........................................................................... 216
7.3.2 Ripple Counter ............................................................................................ 217
7.3.3 Switch Debouncing ..................................................................................... 217
7.3.4 Shift Registers ............................................................................................. 221
7.4 FINITE STATE MACHINES .......................................................................................... 223
7.4.1 Describing the Functionality of a FSM ........................................................ 223
7.4.2 Logic Synthesis for a FSM .......................................................................... 225
7.4.3 FSM Design Process Overview .................................................................. 232
7.4.4 FSM Design Examples ............................................................................... 233
7.5 COUNTERS ............................................................................................................ 241
7.5.1 2-Bit Binary Up Counter .............................................................................. 241
7.5.2 2-Bit Binary Up/Down Counter .................................................................... 242
7.5.3 2-Bit Gray Code Up Counter ....................................................................... 245
7.5.4 2-Bit Gray Code Up/Down Counter ............................................................ 247
7.5.5 3-Bit One-Hot Up Counter ........................................................................... 249
7.5.6 3-Bit One-Hot Up/Down Counter ................................................................ 250
7.6 FINITE STATE MACHINE’S RESET CONDITION .............................................................. 254
7.7 SEQUENTIAL LOGIC ANALYSIS ................................................................................... 255
7.7.1 Finding the State Equations and Output Logic Expressions of a FSM ...... 255
xiv • Contents

7.7.2 Finding the State Transition Table of a FSM ............................................... 256

7.7.3 Finding the State Diagram of a FSM .......................................................... 257
7.7.4 Determining the Maximum Clock Frequency of a FSM .............................. 258
8: VERILOG (PART 2) .............................................................................................. 271
8.1 PROCEDURAL ASSIGNMENT ...................................................................................... 271
8.1.1 Procedural Blocks ....................................................................................... 271
8.1.2 Procedural Statements ................................................................................ 274
8.1.3 Statement Groups ....................................................................................... 279
8.1.4 Local Variables ............................................................................................ 279
8.2 CONDITIONAL PROGRAMMING CONSTRUCTS ................................................................ 280
8.2.1 if-else Statements ........................................................................................ 280
8.2.2 case Statements ......................................................................................... 281
8.2.3 casez and casex Statements ...................................................................... 283
8.2.4 forever Loops .............................................................................................. 283
8.2.5 while Loops ................................................................................................. 283
8.2.6 repeat Loops ............................................................................................... 284
8.2.7 for Loops ...................................................................................................... 284
8.2.8 disable ......................................................................................................... 285
8.3 SYSTEM TASKS ...................................................................................................... 286
8.3.1 Text Output .................................................................................................. 286
8.3.2 File Input/Output .......................................................................................... 287
8.3.3 Simulation Control and Monitoring .............................................................. 289
8.4 TEST BENCHES ...................................................................................................... 290
8.4.1 Common Stimulus Generation Techniques ................................................ 291
8.4.2 Printing Results to the Simulator Transcript ............................................... 292
8.4.3 Automatic Result Checking ......................................................................... 293
8.4.4 Using Loops to Generate Stimulus ............................................................. 295
8.4.5 Using External Files in Test Benches ......................................................... 296
9: BEHAVIORAL MODELING OF SEQUENTIAL LOGIC ........................................ 303
9.1 MODELING SEQUENTIAL STORAGE DEVICES IN VERILOG ............................................... 303
9.1.1 D-Latch ........................................................................................................ 303
9.1.2 D-Flip-Flop ................................................................................................... 304
9.1.3 D-Flip-Flop with Asynchronous Reset ........................................................ 304
9.1.4 D-Flip-Flop with Asynchronous Reset and Preset ...................................... 305
9.1.5 D-Flip-Flop with Synchronous Enable ........................................................ 306
9.2 MODELING FINITE STATE MACHINES IN VERILOG ......................................................... 307
9.2.1 Modeling the States .................................................................................... 309
9.2.2 The State Memory Block ............................................................................. 309
9.2.3 The Next State Logic Block ........................................................................ 309
9.2.4 The Output Logic Block ............................................................................... 310
9.2.5 Changing the State Encoding Approach .................................................... 312
9.3 FSM DESIGN EXAMPLES IN VERILOG ........................................................................ 313
9.3.1 Serial Bit Sequence Detector in Verilog ...................................................... 313
9.3.2 Vending Machine Controller in Verilog ........................................................ 315
9.3.3 2-Bit, Binary Up/Down Counter in Verilog ................................................... 317
 Contents • xv

9.4 MODELING COUNTERS IN VERILOG ............................................................................ 319

9.4.1 Counters in Verilog Using a Single Procedural Block ................................ 319
9.4.2 Counters with Range Checking .................................................................. 320
9.4.3 Counters with Enables in Verilog ................................................................ 320
9.4.4 Counters with Loads ................................................................................... 321
9.5 RTL MODELING ..................................................................................................... 322
9.5.1 Modeling Registers in Verilog ..................................................................... 322
9.5.2 Registers as Agents on a Data Bus ............................................................ 323
9.5.3 Shift Registers in Verilog ............................................................................. 325
10: MEMORY ............................................................................................................ 331
10.1 MEMORY ARCHITECTURE AND TERMINOLOGY .............................................................. 331
10.1.1 Memory Map Model .................................................................................... 331
10.1.2 Volatile Versus Non-volatile Memory .......................................................... 332
10.1.3 Read Only Versus Read/Write Memory ..................................................... 332
10.1.4 Random Access Versus Sequential Access .............................................. 332
10.2 NON-VOLATILE MEMORY TECHNOLOGY ...................................................................... 333
10.2.1 ROM Architecture ........................................................................................ 333
10.2.2 Mask Read Only Memory (MROM) ............................................................ 336
10.2.3 Programmable Read Only Memory (PROM) .............................................. 337
10.2.4 Erasable Programmable Read Only Memory (EPROM) ............................ 338
10.2.5 Electrically Erasable Programmable Read Only Memory (EEPROM) ....... 340
10.2.6 FLASH Memory ........................................................................................... 341
10.3 VOLATILE MEMORY TECHNOLOGY ............................................................................. 342
10.3.1 Static Random Access Memory (SRAM) ................................................... 342
10.3.2 Dynamic Random Access Memory (DRAM) .............................................. 345
10.4 MODELING MEMORY WITH VERILOG .......................................................................... 352
10.4.1 Read-Only Memory in Verilog ..................................................................... 352
10.4.2 Read/Write Memory in Verilog .................................................................... 353
11: PROGRAMMABLE LOGIC ................................................................................. 359
11.1 PROGRAMMABLE ARRAYS ......................................................................................... 359
11.1.1 Programmable Logic Array (PLA) ............................................................... 359
11.1.2 Programmable Array Logic (PAL) ............................................................... 360
11.1.3 Generic Array Logic (GAL) .......................................................................... 361
11.1.4 Hard Array Logic (HAL) ............................................................................... 362
11.1.5 Complex Programmable Logic Devices (CPLD) ........................................ 362
11.2 FIELD PROGRAMMABLE GATE ARRAYS (FPGAS) ........................................................ 363
11.2.1 Configurable Logic Block (or Logic Element) ............................................. 364
11.2.2 Look-Up Tables (LUTs) ............................................................................... 365
11.2.3 Programmable Interconnect Points (PIPs) ................................................. 368
11.2.4 Input/Output Block (IOBs) ........................................................................... 369
11.2.5 Configuration Memory ................................................................................. 370
xvi • Contents

12: ARITHMETIC CIRCUITS .................................................................................... 373

12.1 ADDITION .............................................................................................................. 373
12.1.1 Half Adders ................................................................................................. 373
12.1.2 Full Adders .................................................................................................. 374
12.1.3 Ripple Carry Adder (RCA) .......................................................................... 376
12.1.4 Carry Look Ahead Adder (CLA) .................................................................. 378
12.1.5 Adders in Verilog ......................................................................................... 381
12.2 SUBTRACTION ........................................................................................................ 386
12.3 MULTIPLICATION ...................................................................................................... 389
12.3.1 Unsigned Multiplication ............................................................................... 389
12.3.2 A Simple Circuit to Multiply by Powers of Two ........................................... 392
12.3.3 Signed Multiplication ................................................................................... 393
12.4 DIVISION ............................................................................................................... 395
12.4.1 Unsigned Division ....................................................................................... 395
12.4.2 A Simple Circuit to Divide by Powers of Two ............................................. 398
12.4.3 Signed Division ........................................................................................... 399
13: COMPUTER SYSTEM DESIGN ......................................................................... 403
13.1 COMPUTER HARDWARE ........................................................................................... 403
13.1.1 Program Memory ........................................................................................ 404
13.1.2 Data Memory ............................................................................................... 404
13.1.3 Input/Output Ports ....................................................................................... 404
13.1.4 Central Processing Unit .............................................................................. 405
13.1.5 A Memory Mapped System ........................................................................ 406
13.2 COMPUTER SOFTWARE ............................................................................................ 408
13.2.1 Opcodes and Operands .............................................................................. 409
13.2.2 Addressing Modes ...................................................................................... 409
13.2.3 Classes of Instructions ................................................................................ 410
13.3 COMPUTER IMPLEMENTATION – AN 8-BIT COMPUTER EXAMPLE .................................... 417
13.3.1 Top Level Block Diagram ............................................................................ 417
13.3.2 Instruction Set Design ................................................................................. 418
13.3.3 Memory System Implementation ................................................................ 419
13.3.4 CPU Implementation ................................................................................... 423
13.4 ARCHITECTURE CONSIDERATIONS .............................................................................. 444
13.4.1 Von Neumann Architecture ......................................................................... 444
13.4.2 Harvard Architecture ................................................................................... 444
APPENDIX A: LIST OF WORKED EXAMPLES ...................................................... 449
INDEX ....................................................................................................................... 455
Chapter 1: Introduction: Analog
vs. Digital
We often hear that we live in a digital age. This refers to the massive adoption of computer systems
within every aspect of our lives from smart phones to automobiles to household appliances. This statement
also refers to the transformation that has occurred to our telecommunications infrastructure that now
transmits voice, video and data using 1’s and 0’s. There are a variety of reasons that digital systems have
become so prevalent in our lives. In order to understand these reasons, it is good to start with an
understanding of what a digital system is and how it compares to its counterpart, the analog system.
The goal of this chapter is to provide an understanding of the basic principles of analog and digital systems.

Learning Outcomes—After completing this chapter, you will be able to:

1.1 Describe the fundamental differences between analog and digital systems.
1.2 Describe the advantages of digital systems compared to analog systems.

1.1 Differences Between Analog and Digital Systems

Let’s begin by looking at signaling. In electrical systems, signals represent information that is
transmitted between devices using an electrical quantity (voltage or current). An analog signal is defined
as a continuous, time-varying quantity that corresponds directly to the information it represents. An
example of this would be a barometric pressure sensor that outputs an electrical voltage corresponding
to the pressure being measured. As the pressure goes up, so does the voltage. While the range of the
input (pressure) and output (voltage) will have different spans, there is a direct mapping between the
pressure and voltage. Another example would be sound striking a traditional analog microphone. Sound
is a pressure wave that travels through a medium such as air. As the pressure wave strikes the
diaphragm in the microphone, the diaphragm moves back and forth. Through the process of inductive
coupling, this movement is converted to an electric current. The characteristics of the current signal
produced (e.g., frequency and magnitude) correspond directly to the characteristics of the incoming
sound wave. The current can travel down a wire and go through another system that works in the
opposite manner by inductively coupling the current onto another diaphragm, which in turn moves back
and forth forming a pressure wave and thus sound (i.e., a speaker). In both of these examples, the
electrical signal represents the actual information that is being transmitted and is considered analog.
Analog signals can be represented mathematically as a function with respect to time.
In digital signaling the electrical signal itself is not directly the information it represents, instead, the
information is encoded. The most common type of encoding is binary (1’s and 0’s). The 1’s and 0’s are
represented by the electrical signal. The simplest form of digital signaling is to define a threshold voltage
directly in the middle of the range of the electrical signal. If the signal is above this threshold, the signal is
representing a 1. If the signal is below this threshold, the signal is representing a 0. This type of signaling
is not considered continuous as in analog signaling, instead, it is considered to be discrete because the
information is transmitted as a series of distinct values. The signal transitions between a 1 to 0 or 0 to
1 are assumed to occur instantaneously. While this is obviously impossible, for the purposes of
information transmission, the values can be interpreted as a series of discrete values. This is a digital
signal and is not the actual information, but rather the binary encoded representation of the original
information. Digital signals are not represented using traditional mathematical functions, instead, the
digital values are typically held in tables of 1’s and 0’s.

# Springer International Publishing AG 2017 1

B.J. LaMeres, Introduction to Logic Circuits & Logic Design with Verilog,
DOI 10.1007/978-3-319-53883-9_1
2 • Chapter 1: Introduction: Analog vs. Digital

Figure 1.1 shows an example analog signal (left) and an example digital signal (right). While the
digital signal is in reality continuous, it represents a series of discrete 1 and 0 values.

Fig. 1.1
Analog (left) vs. digital (right) signals

CONCEPT CHECK

CC1.1 If a digital signal is only a discrete representation of real information, how is it possible to
produce high quality music without hearing “gaps” in the output due to the digitization
process?
(A) The gaps are present but they occur so quickly that the human ear can’t
detect them.
(B) When the digital music is converted back to analog sound the gaps are smoothed
out since an analog signal is by definition continuous.
(C) Digital information is a continuous, time-varying signal so there aren’t gaps.
(D) The gaps can be heard if the music is played slowly, but at normal speed, they
can’t be.

1.2 Advantages of Digital Systems over Analog Systems

There are a variety of reasons that digital systems are preferred over analog systems. First is their
ability to operate within the presence of noise. Since an analog signal is a direct representation of the
physical quantity it is transmitting, any noise that is coupled onto the electrical signal is interpreted as
noise on the original physical quantity. An example of this is when you are listening to an AM/FM radio
and you hear distortion of the sound coming out of the speaker. The distortion you hear is not due to
 1.2 Advantages of Digital Systems over Analog Systems • 3

actual distortion of the music as it was played at the radio station, but rather electrical noise that was
coupled onto the analog signal transmitted to your radio prior to being converted back into sound by the
speakers. Since the signal in this case is analog, the speaker simply converts it in its entirety (noise +
music) into sound. In the case of digital signaling, a significant amount of noise can be added to the
signal while still preserving the original 1’s and 0’s that are being transmitted. For example, if the signal is
representing a 0, the receiver will still interpret the signal as a 0 as long as the noise doesn’t cause the
level to exceed the threshold. Once the receiver interprets the signal as a 0, it stores the encoded value
as a 0 thus ignoring any noise present during the original transmission. Figure 1.2 shows the exact same
noise added to the analog and digital signals from Fig. 1.1. The analog signal is distorted; however, the
digital signal is still able to transmit the 0’s and 1’s that represent the information.

Fig. 1.2
Noise on analog (left) and digital (right) signals

Another reason that digital systems are preferred over analog ones is the simplicity of the circuitry. In
order to produce a 1 and 0, you simply need an electrical switch. If the switch connects the output to a
voltage below the threshold, then it produces a 0. If the switch connects the output to a voltage above the
threshold, then it produces a 1. It is relatively simple to create such a switching circuit using modern
transistors. Analog circuitry, however, needs to perform the conversion of the physical quantity it is
representing (e.g., pressure, sound) into an electrical signal all the while maintaining a direct correspon-
dence between the input and output. Since analog circuits produce a direct, continuous representation of
information, they require more complicated designs to achieve linearity in the presence of environmental
variations (e.g., power supply, temperature, fabrication differences). Since digital circuits only produce a
discrete representation of the information, they can be implemented with simple switches that are only
altered when information is produced or retrieved. Figure 1.3 shows an example comparison between an
analog inverting amplifier and a digital inverter. The analog amplifier uses dozens of transistors (inside
the triangle) and two resistors to perform the inversion of the input. The digital inverter uses two
transistors that act as switches to perform the inversion.
4 • Chapter 1: Introduction: Analog vs. Digital

Fig. 1.3
Analog (left) vs. digital (right) circuits

A final reason that digital systems are being widely adopted is their reduced power consumption.
With the advent of Complementary Metal Oxide Transistors (CMOS), electrical switches can be created
that consume very little power to turn on or off and consume relatively negligible amounts of power to
keep on or off. This has allowed large scale digital systems to be fabricated without excessive levels of
power consumption. For stationary digital systems such as servers and workstations, extremely large
and complicated systems can be constructed that consume reasonable amounts of power. For portable
digital systems such as smart phones and tablets, this means useful tools can be designed that are able
to run on portable power sources. Analog circuits, on the other hand, require continuous power to
accurately convert and transmit the electrical signal representing the physical quantity. Also, the circuit
techniques that are required to compensate for variances in power supply and fabrication processes in
analog systems require additional power consumption. For these reasons, analog systems are being
replaced with digital systems wherever possible to exploit their noise immunity, simplicity and low power
consumption. While analog systems will always be needed at the transition between the physical (e.g.,
microphones, camera lenses, sensors, video displays) and the electrical world, it is anticipated that the
push toward digitization of everything in between (e.g., processing, transmission, storage) will continue.

CONCEPT CHECK

CC1.2 When does the magnitude of electrical noise on a digital signal prevent the original informa-
tion from being determined?
(A) When it causes the system to draw too much power.
(B) When the shape of the noise makes the digital signal look smooth and continuous
like a sine wave.
(C) When the magnitude of the noise is large enough that it causes the signal to
inadvertently cross the threshold voltage.
(D) It doesn’t. A digital signal can withstand any magnitude of noise.
 Exercise Problems • 5

Summary
v An analog system uses a direct mapping v Using a discrete representation allows the dig-
between an electrical quantity and the infor- ital signals to be more immune to noise in
mation being processed. A digital system, on addition to requiring simple circuits that require
the other hand, uses a discrete representa- less power to perform the computations.
tion of the information.

Exercise Problems
Section 1.1: Differences Between Analog 1.1.9 Is the physical interface on the touch display of
an iPod analog or digital?
and Digital Systems
1.1.10 Take a look around right now and identify two
1.1.1 If an electrical signal is a direct function of a
digital technologies in use.
physical quantity, is it considered analog or
digital? 1.1.11 Take a look around right now and identify two
analog technologies in use.
1.1.2 If an electrical signal is a discrete representa-
tion of information, is it considered analog or
digital? Section 1.2: Advantages of Digital
1.1.3 What part of any system will always require an Systems over Analog Systems
analog component? 1.2.1 Give three advantages of using digital systems
1.1.4 Is the sound coming out of earbuds analog or over analog.
digital? 1.2.2 Name a technology or device that has evolved
1.1.5 Is the MP3 file stored on an iPod analog or from analog to digital in your lifetime.
digital? 1.2.3 Name an analog technology or device that has
1.1.6 Is the circuitry that reads the MP3 file from become obsolete in your lifetime.
memory in an iPod analog or digital? 1.2.4 Name an analog technology or device that has
1.1.7 Is the electrical signal that travels down ear- been replaced by digital technology but is still
phone wires analog or digital? in use due to nostalgia.
1.1.8 Is the voltage coming out of the battery in an 1.2.5 Name a technology or device invented in your
iPod analog or digital? lifetime that could not have been possible with-
out digital technology.
Chapter 2: Number Systems
Logic circuits are used to generate and transmit 1’s and 0’s to compute and convey information. This
two-valued number system is called binary. As presented earlier, there are many advantages of using a
binary system; however, the human brain has been taught to count, label and measure using the
decimal number system. The decimal number system contains 10 unique symbols (0 ! 9) commonly
referred to as the Arabic numerals. Each of these symbols is assigned a relative magnitude to the other
symbols. For example, 0 is less than 1, 1 is less than 2, etc. It is often conjectured that the 10 symbol
number system that we humans use is due to the availability of our 10 fingers (or digits) to visualize
counting up to 10. Regardless, our brains are trained to think of the real world in terms of a decimal
system. In order to bridge the gap between the way our brains think (decimal) and how we build our
computers (binary), we need to understand the basics of number systems. This includes the formal
definition of a positional number system and how it can be extended to accommodate any arbitrarily large
(or small) value. This also includes how to convert between different number systems that contain
different numbers of symbols. In this chapter, we cover 4 different number systems: decimal
(10 symbols), binary (2 symbols), octal (8 symbols), and hexadecimal (16 symbols). The study of
decimal and binary is obvious as they represent how our brains interpret the physical world (decimal)
and how our computers work (binary). Hexadecimal is studied because it is a useful means to represent
large sets of binary values using a manageable number of symbols. Octal is rarely used but is studied as
an example of how the formalization of the number systems can be applied to all systems regardless of
the number of symbols they contain. This chapter will also discuss how to perform basic arithmetic in the
binary number system and represent negative numbers. The goal of this chapter is to provide an
understanding of the basic principles of binary number systems.
Learning Outcomes—After completing this chapter, you will be able to:

2.1 Describe the formation and use of positional number systems.

2.2 Convert numbers between different bases.
2.3 Perform binary addition and subtraction by hand.
2.4 Use two’s complement numbers to represent negative numbers.

2.1 Positional Number Systems

A positional number system allows the expansion of the original set of symbols so that they can be
used to represent any arbitrarily large (or small) value. For example, if we use the 10 symbols in our
decimal system, we can count from 0 to 9. Using just the individual symbols we do not have enough
symbols to count beyond 9. To overcome this, we use the same set of symbols but assign a different
value to the symbol based on its position within the number. The position of the symbol with respect to
other symbols in the number allows an individual symbol to represent greater (or lesser) values. We can
use this approach to represent numbers larger than the original set of symbols. For example, let’s say we
want to count from 0 upward by 1. We begin counting 0, 1, 2, 3, 4, 5, 6, 7, 8 to 9. When we are out of
symbols and wish to go higher, we bring on a symbol in a different position with that position being valued
higher and then start counting over with our original symbols (e.g., . . ., 9, 10, 11,... 19, 20, 21,...). This is
repeated each time a position runs out of symbols (e.g., . . ., 99, 100, 101. . . 999, 1000, 1001,. . .).

First, let’s look at the formation of a number system. The first thing that is needed is a set of symbols.
The formal term for one of the symbols in a number system is a numeral. One or more numerals are used
to form a number. We define the number of numerals in the system using the terms radix or base.

# Springer International Publishing AG 2017 7

B.J. LaMeres, Introduction to Logic Circuits & Logic Design with Verilog,
DOI 10.1007/978-3-319-53883-9_2
8 • Chapter 2: Number Systems

For example, our decimal number system is said to be base 10, or have a radix of 10 because it consists
of 10 unique numerals or symbols.

Radix ¼ Base  the number of numerals in the number system

The next thing that is needed is the relative value of each numeral with respect to the other numerals
in the set. We can say 0 < 1 < 2 < 3 etc. to define the relative magnitudes of the numerals in this set. The
numerals are defined to be greater or less than their neighbors by a magnitude of 1. For example, in the
decimal number system each of the subsequent numerals is greater than its predecessor by exactly
1. When we define this relative magnitude we are defining that the numeral 1 is greater than the numeral
0 by a magnitude of 1; the numeral 2 is greater than the numeral 1 by a magnitude of 1, etc. At this point
we have the ability to count from 0 to 9 by 1’s. We also have the basic structure for mathematical
operations that have results that fall within the numeral set from 0 to 9 (e.g., 1 + 2 ¼ 3). In order to expand
the values that these numerals can represent, we need define the rules of a positional number system.

2.1.1 Generic Structure

In order to represent larger or smaller numbers than the lone numerals in a number system can
represent, we adopt a positional system. In a positional number system, the relative position of the
numeral within the overall number dictates its value. When we begin talking about the position of a
numeral, we need to define a location to which all of the numerals are positioned with respect to. We
define the radix point as the point within a number to which numerals to the left represent whole numbers
and numerals to the right represent fractional numbers. The radix point is denoted with a period (i.e., “.”).
A particular number system often renames this radix point to reflect its base. For example, in the base
10 number system (i.e., decimal), the radix point is commonly called the decimal point; however, the term
radix point can be used across all number systems as a generic term. If the radix point is not present in a
number, it is assumed to be to the right of number. Figure 2.1 shows an example number highlighting the
radix point and the relative positions of the whole and fractional numerals.

Fig. 2.1
Definition of radix point

Next, we need to define the position of each numeral with respect to the radix point. The position of
the numeral is assigned a whole number with the number to the left of the radix point having a position
value of 0. The position number increases by 1 as numerals are added to the left (2, 3, 4. . .) and
decreased by 1 as numerals are added to the right (1, 2, 3). We will use the variable p to represent
position. The position number will be used to calculate the value of each numeral in the number based on
its relative position to the radix point. Figure 2.2 shows the example number with the position value of
each numeral highlighted.

Fig. 2.2
Definition of position number (p) within the number
 2.1 Positional Number Systems • 9

In order to create a generalized format of a number, we assign the term digit (d) to each of the
numerals in the number. The term digit signifies that the numeral has a position. The position of the digit
within the number is denoted as a subscript. The term digit can be used as a generic term to describe a
numeral across all systems, although some number systems will use a unique term instead of digit which
indicates its base. For example, the binary system uses the term bit instead of digit; however, using the
term digit to describe a generic numeral in any system is still acceptable. Figure 2.3 shows the generic
subscript notation used to describe the position of each digit in the number.

Fig. 2.3
Digit notation

We write a number from left to right starting with the highest position digit that is greater than 0 and
end with the lowest position digit that is greater than 0. This reduces the amount of numerals that are
written; however, a number can be represented with an arbitrary number of 0’s to the left of the highest
position digit greater than 0 and an arbitrary number of 0’s to the right of the lowest position digit greater
than 0 without affecting the value of the number. For example, the number 132.654 could be written as
0132.6540 without affecting the value of the number. The 0’s to the left of the number are called leading
0’s and the 0’s to the right of the number are called trailing 0’s. The reason this is being stated is because
when a number is implemented in circuitry, the number of numerals is fixed and each numeral must have
a value. The variable n is used to represent the number of numerals in a number. If a number is defined
with n ¼ 4, that means 4 numerals are always used. The number 0 would be represented as 0000 with
both representations having an equal value.

2.1.2 Decimal Number System (Base 10)

As mentioned earlier, the decimal number system contains 10 unique numerals (0, 1, 2, 3, 4, 5, 6, 7,
8 and 9). This system is thus a base 10 or a radix 10 system. The relative magnitudes of the symbols are
0 < 1 < 2 < 3 < 4 < 5 < 6 < 7 < 8 < 9.

2.1.3 Binary Number System (Base 2)

The binary number system contains 2 unique numerals (0 and 1). This system is thus a base 2 or a
radix 2 system. The relative magnitudes of the symbols are 0 < 1. At first glance, this system looks very
limited in its ability to represent large numbers due to the small number of numerals. When counting up,
as soon as you count from 0 to 1, you are out of symbols and must increment the p + 1 position in order to
represent the next number (e.g., 0, 1, 10, 11, 100, 101, . . .); however, magnitudes of each position scale
quickly so that circuits with a reasonable amount of digits can represent very large numbers. The term bit
is used instead of digit in this system to describe the individual numerals and at the same time indicate
the base of the number.

Due to the need for multiple bits to represent meaningful information, there are terms dedicated to
describe the number of bits in a group. When 4 bits are grouped together, they are called a nibble. When
8 bits are grouped together, they are called a byte. Larger groupings of bits are called words. The size of
the word can be stated as either an n-bit word or omitted if the size of the word is inherently implied. For
example, if you were using a 32-bit microprocessor, using the term word would be interpreted as a 32-bit
word. For example, if there was a 32-bit grouping, it would be referred to as a 32-bit word. The leftmost bit
10 • Chapter 2: Number Systems

in a binary number is called the Most Significant Bit (MSB). The rightmost bit in a binary number is
called the Least Significant Bit (LSB).

2.1.4 Octal Number System (Base 8)

The octal number system contains 8 unique numerals (0, 1, 2, 3, 4, 5, 6, 7). This system is thus a
base 8 or a radix 8 system. The relative magnitudes of the symbols are 0 < 1 < 2 < 3 < 4 < 5 < 6 < 7.
We use the generic term digit to describe the numerals within an octal number.

2.1.5 Hexadecimal Number System (Base 16)

The hexadecimal number system contains 16 unique numerals. This system is most often referred
to in spoken word as “hex” for short. Since we only have 10 Arabic numerals in our familiar decimal
system, we need to use other symbols to represent the remaining 6 numerals. We use the alphabetic
characters A-F in order to expand the system to 16 numerals. The 16 numerals in the hexadecimal
system are 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, A, B, C, D, E and F. The relative magnitudes of the symbols are
0 < 1 < 2 < 3 < 4 < 5 < 6 < 7 < 8 < 9 < A < B < C < D < E < F. We use the generic term digit to
describe the numerals within a hexadecimal number.

At this point, it becomes necessary to indicate the base of a written number. The number 10 has an
entirely different value if it is a decimal number or binary number. In order to handle this, a subscript is
typically included at the end of the number to denote its base. For example, 1010 indicates that this
number is decimal “ten”. If the number was written as 102, this number would represent binary “one zero”.
Table 2.1 lists the equivalent values in each of the 4 number systems just described for counts from 010 to
1510. The left side of the table does not include leading 0’s. The right side of the table contains the same
information but includes the leading zeros. The equivalencies of decimal, binary and hexadecimal in this
table are typically committed to memory.

Table 2.1
Number system equivalency
 2.2 Base Conversion • 11

CONCEPT CHECK

CC2.1 The base of a number system is arbitrary and is commonly selected to match a particular
aspect of the physical system in which it is used (e.g., base 10 corresponds to our 10 fingers,
base 2 corresponds to the 2 states of a switch). If a physical system contained 3 unique
modes and a base of 3 was chosen for the number system, what is the base 3 equivalent of
the decimal number 3?
(A) 310 ¼ 113 (B) 310 ¼ 33 (C) 310 ¼ 103 (D) 310 ¼ 213

2.2 Base Conversion

Now we look at converting between bases. There are distinct techniques for converting to and
from decimal. There are also techniques for converting between bases that are powers of 2 (e.g., base
2, 4, 8, 16, etc.).

2.2.1 Converting to Decimal

The value of each digit within a number is based on the individual digit value and the digit’s position.
Each position in the number contains a different weight based on its relative location to the radix point.
The weight of each position is based on the radix of the number system that is being used. The weight of
each position in decimal is defined as:

Weight ¼ ðRadixÞp

This expression gives the number system the ability to represent fractional numbers since an
expression with a negative exponent (e.g., xy) is evaluated as one over the expression with the
exponent change to positive (e.g., 1/xy). Figure 2.4 shows the generic structure of a number with its
positional weight highlighted.

Fig. 2.4
Weight definition

In order to find the decimal value of each of the numerals in the number, its individual numeral value
is multiplied by its positional weight. In order to find the value of the entire number, each value of the
individual numeral-weight products is summed. The generalized format of this conversion is written as:

X
pmax
Total Decimal Value ¼ di  ðradixÞi
i¼pmin

In this expression, pmax represents the highest position number that contains a numeral greater
than 0. The variable pmin represents the lowest position number that contains a numeral greater than 0.
These limits are used to simplify the hand calculations; however, these terms theoretically could be +1
12 • Chapter 2: Number Systems

to 1 with no effect on the result since the summation of every leading 0 and every trailing 0 contributes
nothing to the result.
As an example, let’s evaluate this expression for a decimal number. The result will yield the original
number but will illustrate how positional weight is used. Let’s take the number 132.65410. To find the
decimal value of this number, each numeral is multiplied by its positional weight and then all of
the products are summed. The positional weight for the digit 1 is (radix)p or (10)2. In decimal this is
called the hundred’s position. The positional weight for the digit 3 is (10)1, referred to as the ten’s position.
The positional weight for digit 2 is (10)0, referred to as the one’s position. The positional weight for digit
6 is (10)1, referred to as the tenth’s position. The positional weight for digit 5 is (10)2, referred to as the
hundredth’s position. The positional weight for digit 4 is (10)3, referred to as the thousandth’s position.

When these weights are multiplied by their respective digits and summed, the result is the original
decimal number 132.65410. Example 2.1 shows this process step-by-step.

Example 2.1
Converting decimal to decimal

This process is used to convert between any other base to decimal.

2.2.1.1 Binary to Decimal

Let’s convert 101.112 to decimal. The same process is followed with the exception that the base in
the summation is changed to 2. Converting from binary to decimal can be accomplished quickly in your
head due to the fact that the bit values in the products are either 1 or 0. That means any bit that is a 0 has
no impact on the outcome and any bit that is a 1 simply yields the weight of its position. Example 2.2
shows the step-by-step process converting a binary number to decimal.
 2.2 Base Conversion • 13

Example 2.2
Converting binary to decimal

2.2.1.2 Octal to Decimal

When converting from octal to decimal, the same process is followed with the exception that the
base in the weight is changed to 8. Example 2.3 shows an example of converting an octal number to
decimal.

Example 2.3
Converting octal to decimal
14 • Chapter 2: Number Systems

2.2.1.3 Hexadecimal to Decimal

Let’s convert 1AB.EF16 to decimal. The same process is followed with the exception that the base is
changed to 16. When performing the conversion, the decimal equivalent of the numerals A-F need to be
used. Example 2.4 shows the step-by-step process converting a hexadecimal number to decimal.

Example 2.4
Converting hexadecimal to decimal

2.2.2 Converting From Decimal

The process of converting from decimal to another base consists of two separate algorithms. There
is one algorithm for converting the whole number portion of the number and another algorithm for
converting the fractional portion of the number. The process for converting the whole number portion
is to divide the decimal number by the base of the system you wish to convert to. The division will result in
a quotient and a whole number remainder. The remainder is recorded as the least significant numeral in
the converted number. The resulting quotient is then divided again by the base, which results in a new
quotient and new remainder. The remainder is recorded as the next higher order numeral in the new
number. This process is repeated until a quotient of 0 is achieved. At that point the conversion is
complete. The remainders will always be within the numeral set of the base being converted to.

The process for converting the fractional portion is to multiply just the fractional component of the
number by the base. This will result in a product that contains a whole number and a fraction. The whole
number is recorded as the most significant digit of the new converted number. The new fractional portion
is then multiplied again by the base with the whole number portion being recorded as the next lower order
numeral. This process is repeated until the product yields a fractional component equal to zero or the
desired level of accuracy has been achieved. The level of accuracy is specified by the number of
numerals in the new converted number. For example, the conversion would be stated as “convert this
decimal number to binary with a fractional accuracy of 4 bits”. This means the algorithm would stop once
4-bits of fraction had been achieved in the conversion.
 2.2 Base Conversion • 15

2.2.2.1 Decimal to Binary

Let’s convert 11.37510 to binary. Example 2.5 shows the step-by-step process converting a decimal
number to binary.

Example 2.5
Converting decimal to binary

2.2.2.2 Decimal to Octal

Let’s convert 10.410 to octal with an accuracy of 4 fractional digits. When converting the fractional
component of the number, the algorithm is continued until 4 digits worth of fractional numerals has been
achieved. Once the accuracy has been achieved, the conversion is finished even though a product with
a zero fractional value has not been obtained. Example 2.6 shows the step-by-step process converting a
decimal number to octal with a fractional accuracy of 4 digits.
16 • Chapter 2: Number Systems

Example 2.6
Converting decimal to octal

2.2.2.3 Decimal to Hexadecimal

Let’s convert 254.65510 to hexadecimal with an accuracy of 3 fractional digits. When doing this
conversion, all of the divisions and multiplications are done using decimal. If the results end up between
1010 and 1510, then the decimal numbers are substituted with their hex symbol equivalent (i.e., A to F).
Example 2.7 shows the step-by-step process of converting a decimal number to hex with a fractional
accuracy of 3 digits.
 2.2 Base Conversion • 17

Example 2.7
Converting decimal to hexadecimal

2.2.3 Converting Between 2n Bases

Converting between 2n bases (e.g., 2, 4, 8, 16, etc.) takes advantage of the direct mapping that each
of these bases has back to binary. Base 8 numbers take exactly 3 binary bits to represent all 8 symbols
(i.e., 08 ¼ 0002, 78 ¼ 1112). Base 16 numbers take exactly 4 binary bits to represent all 16 symbols (i.e.,
016 ¼ 00002, F16 ¼ 11112).
When converting from binary to any other 2n base, the whole number bits are grouped into the
appropriate-sized sets starting from the radix point and working left. If the final leftmost grouping does not
have enough symbols, it is simply padded on left with leading 0’s. Each of these groups is then directly
substituted with their 2n base symbol. The fractional number bits are also grouped into the appropriate-
sized sets starting from the radix point, but this time working right. Again, if the final rightmost grouping
does not have enough symbols, it is simply padded on the right with trailing 0’s. Each of these groups is
then directly substituted with their 2n base symbol.

2.2.3.1 Binary to Octal

Example 2.8 shows the step-by-step process of converting a binary number to octal.
18 • Chapter 2: Number Systems

Example 2.8
Converting binary to octal

2.2.3.2 Binary to Hexadecimal

Example 2.9 shows the step-by-step process of converting a binary number to hexadecimal.

Example 2.9
Converting binary to hexadecimal

2.2.3.3 Octal to Binary

When converting to binary from any 2n base, each of the symbols in the originating number are
replaced with the appropriate-sized number of bits. An octal symbol will be replaced with 3 binary bits
while a hexadecimal symbol will be replaced with 4 binary bits. Any leading or trailing 0’s can be removed
 2.2 Base Conversion • 19

from the converted number once complete. Example 2.10 shows the step-by-step process of converting
an octal number to binary.

Example 2.10
Converting octal to binary

2.2.3.4 Hexadecimal to Binary

Example 2.11 shows the step-by-step process of converting a hexadecimal number to binary.

Example 2.11
Converting hexadecimal to binary

2.2.3.5 Octal to Hexadecimal

When converting between 2n bases (excluding binary) the number is first converted into binary and
then converted from binary into the final 2n base using the algorithms described before. Example 2.12
shows the step-by-step process of converting an octal number to hexadecimal.
20 • Chapter 2: Number Systems

Example 2.12
Converting octal to hexadecimal

2.2.3.6 Hexadecimal to Octal

Example 2.13 shows the step-by-step process of converting a hexadecimal number to octal.

Example 2.13
Converting hexadecimal to octal
 2.3 Binary Arithmetic • 21

CONCEPT CHECK

CC2.2 A “googol” is the term for the decimal number 1e100. When written out manually this number
is a 1 with 100 zeros after it (e.g., 10,000,000,000,000,000,000,000,000,000,000,000,000,
000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000).
This term is more commonly associated with the search engine company Google, which uses a
different spelling but is pronounced the same. How many bits does it take to represent a googol in
binary?
(A) 100 bits (B) 256 bits (C) 332 bits (D) 333 bits

2.3 Binary Arithmetic

2.3.1 Addition (Carries)
Binary addition is a straightforward process that mirrors the approach we have learned for longhand
decimal addition. The two numbers (or terms) to be added are aligned at the radix point and addition
begins at the least significant bit. If the sum of the least significant position yields a value with two bits
(e.g., 102), then the least significant bit is recorded and the most significant bit is carried to the next higher
position. The sum of the next higher position is then performed including the potential carry bit from the
prior addition. This process continues from the least significant position to the most significant position.
Example 2.14 shows how addition is performed on two individual bits.

Example 2.14
Single bit binary addition

When performing binary addition, the width of the inputs and output is fixed (i.e., n-bits). Carries that
exist within the n-bits are treated in the normal fashion of including them in the next higher position sum;
however, if the highest position summation produces a carry, this is a uniquely named event. This event
is called a carry out or the sum is said to generate a carry. The reason this type of event is given special
terminology is because in real circuitry, the number of bits of the inputs and output is fixed in hardware
and the carry out is typically handled by a separate circuit. Example 2.15 shows this process when
adding two 4-bit numbers.
22 • Chapter 2: Number Systems

Example 2.15
Multiple bit binary addition

The largest decimal sum that can result from the addition of two binary numbers is given by
2(2n1). For example, two 8-bit numbers to be added could both represent their highest decimal value
of (2n1) or 25510 (i.e., 1111 11112). The sum of this number would result in 51010 or (1 1111 11102).
Notice that the largest sum achievable would only require one additional bit. This means that a single
carry bit is sufficient to handle all possible magnitudes for binary addition.

2.3.2 Subtraction (Borrows)

Binary subtraction also mirrors longhand decimal subtraction. In subtraction, the formal terms for the
two numbers being operated on are minuend and subtrahend. The subtrahend is subtracted from the
minuend to find the difference. In longhand subtraction, the minuend is the top number and the
subtrahend is the bottom number. For a given position if the minuend is less than the subtrahend, it
needs to borrow from the next higher order position to produce a difference that is positive. If the next
higher position does not have a value that can be borrowed from (i.e., 0), then it in turn needs to borrow
from the next higher position, and so forth. Example 2.16 shows how subtraction is performed on two
individual bits.

Example 2.16
Single bit binary subtraction

As with binary addition, binary subtraction is accomplished on fixed widths of inputs and output (i.e.,
n-bits). The minuend and subtrahend are aligned at the radix point and subtraction begins at the least
significant bit position. Borrows are used as necessary as the subtractions move from the least signifi-
cant position to the most significant position. If the most significant position requires a borrow, this is a
uniquely named event. This event is called a borrow in or the subtraction is said to require a borrow.
Again, the reason this event is uniquely named is because in real circuitry, the number of bits of the input
 2.4 Unsigned and Signed Numbers • 23

and output is fixed in hardware and the borrow in is typically handled by a separate circuit. Example 2.17
shows this process when subtracting two 4-bit numbers.

Example 2.17
Multiple bit binary subtraction

Notice that if the minuend is less than the subtrahend, then the difference will be negative. At this
point, we need a way to handle negative numbers.

CONCEPT CHECK

CC2.3 If an 8-bit computer system can only perform unsigned addition on 8-bit inputs and produce an
8-bit sum, how is it possible for this computer to perform addition on numbers that are larger than
what can be represented with 8-bits (e.g., 1,00010 + 1,00010 ¼ 2,00010)?
(A) There are multiple 8-bit adders in a computer to handle large numbers.
(B) The result is simply rounded to the nearest 8-bit number.
(C) The computer returns an error and requires smaller numbers to be entered.
(D) The computer keeps track of the carry out and uses it in a subsequent 8-bit
addition, which enables larger numbers to be handled.

2.4 Unsigned and Signed Numbers

All of the number systems presented in the prior sections were positive. We need to also have a
mechanism to indicate negative numbers. When looking at negative numbers, we only focus on the
mapping between decimal and binary since octal and hexadecimal are used as just another representa-
tion of a binary number. In decimal, we are able to use the negative sign in front of a number to indicate it
is negative (e.g., 3410). In binary, this notation works fine for writing numbers on paper (e.g., 10102),
but we need a mechanism that can be implemented using real circuitry. In a real digital circuit, the circuits
can only deal with 0’s and 1’s. There is no “” in a digital circuit. Since we only have 0’s and 1’s in the
hardware, we use a bit to represent whether a number is positive or negative. This is referred to as the
sign bit. If a binary number is not going to have any negative values, then it is called an unsigned number
and it can only represent positive numbers. If a binary number is going to allow negative numbers, it is
called a signed number. It is important to always keep track of the type of number we are using as the
same bit values can represent very different numbers depending on the coding mechanism that is
being used.
24 • Chapter 2: Number Systems

2.4.1 Unsigned Numbers

An unsigned number is one that does not allow negative numbers. When talking about this type of
code, the number of bits is fixed and stated up front. We use the variable n to represent the number of bits
in the number. For example, if we had an 8-bit number, we would say, “This is an 8-bit, unsigned number”.
The number of unique codes in an unsigned number is given by 2n. For example, if we had an 8-bit
number, we would have 28 or 256 unique codes (e.g., 0000 00002 to 1111 11112).
The range of an unsigned number refers to the decimal values that the binary code can represent. If
we use the notation Nunsigned to represent any possible value that an n-bit, unsigned number can take on,
the range would be defined as: 0 < Nunsigned < (2n  1)

Range of an UNSIGNED number ) 0  Nunsigned  ð2n 1Þ

For example, if we had an unsigned number with n ¼ 4, it could take on a range of values from +010
(00002) to +1510 (11112). Notice that while this number has 16 unique possible codes, the highest
decimal value it can represent is 1510. This is because one of the unique codes represents 010. This is
the reason that the highest decimal value that can be represented is given by (2n1). Example 2.18
shows this process for a 16-bit number.

Example 2.18
Finding the range of an unsigned number

2.4.2 Signed Numbers

Signed numbers are able to represent both positive and negative numbers. The most significant bit
of these numbers is always the sign bit, which represents whether the number is positive or negative.
The sign bit is defined to be a 0 if the number is positive and 1 if the number is negative. When using
signed numbers, the number of bits is fixed so that the sign bit is always in the same position. There are a
variety of ways to encode negative numbers using a sign bit. The encoding method used exclusively in
modern computers is called two’s complement. There are two other encoding techniques called signed
magnitude and one’s complement that are rarely used but are studied to motivate the power of two’s
complement. When talking about a signed number, the number of bits and the type of encoding is always
stated. For example, we would say, “This is an 8-bit, two’s complement number”.
 2.4 Unsigned and Signed Numbers • 25

2.4.2.1 Signed Magnitude

Signed Magnitude is the simplest way to encode a negative number. In this approach, the most
significant bit (i.e., leftmost bit) of the binary number is considered the sign bit (0 ¼ positive, 1 ¼ negative).
The rest of the bits to the right of the sign bit represent the magnitude or absolute value of the number. As
an example of this approach, let’s look at the decimal values that a 4-bit, signed magnitude number can
take on. These are shown in Example 2.19.

Example 2.19
Decimal values that a 4-bit, signed magnitude code can represent

There are drawbacks of signed magnitude encoding that are apparent from this example. First, the
value of 010 has two signed magnitude codes (00002 and 10002). This is an inefficient use of the
available codes and leads to complexity when building arithmetic circuitry since it must account for
two codes representing the same number.
The second drawback is that addition using the negative numbers does not directly map to how
decimal addition works. For example, in decimal if we added (5) + (1), the result would be 4. In signed
magnitude, adding these numbers using a traditional adder would produce (5) + (1) ¼ (6). This is
because the traditional addition would take place on the magnitude portion of the number. A 510 is
represented with 1012. Adding 1 to this number would result in the next higher binary code 1102 or 610.
Since the sign portion is separate, the addition is performed on |5|, thus yielding 6. Once the sign bit is
included, the resulting number is 6. It is certainly possible to build an addition circuit that works on
signed magnitude numbers, but it is more complex than a traditional adder because it must perform a
different addition operation for the negative numbers versus the positive numbers. It is advantageous to
have a single adder that works across the entire set of numbers.
Due to the duplicate codes for 0, the range of decimal numbers that signed magnitude can represent
is reduced by 1 compared to unsigned encoding. For an n-bit number, there are 2n unique binary codes
available but only 2n1 can be used to represent unique decimal numbers. If we use the notation NSM to
represent any possible value that an n-bit, signed magnitude number can take on, the range would be
defined as:
   
Range of a SIGNED MAGNITUDE number )  2n1  1  NSM  þ 2n1 1
26 • Chapter 2: Number Systems

Example 2.20 shows how to use this expression to find the range of decimal values that an 8-bit,
signed magnitude code can represent.

Example 2.20
Finding the range of a signed magnitude number

The process to determine the decimal value from a signed magnitude binary code involves treating
the sign bit separately from the rest of the code. The sign bit provides the polarity of the decimal number
(0 ¼ Positive, 1 ¼ Negative). The remaining bits in the code are treated as unsigned numbers and
converted to decimal using the standard conversion procedure described in the prior sections. This
conversion yields the magnitude of the decimal number. The final decimal value is found by applying the
sign. Example 2.21 shows an example of this process.

Example 2.21
Finding the decimal value of a signed magnitude number

2.4.2.2 One’s Complement

One’s complement is another simple way to encode negative numbers. In this approach, the
negative number is obtained by taking its positive equivalent and flipping all of the 1’s to 0’s and 0’s to
1’s. This procedure of flipping the bits is called a complement (notice the two e’s). In this way, the most
significant bit of the number is still the sign bit (0 ¼ positive, 1 ¼ negative). The rest of the bits represent
 2.4 Unsigned and Signed Numbers • 27

the value of the number, but in this encoding scheme the negative number values are less intuitive. As an
example of this approach, let’s look at the decimal values that a 4-bit, one’s complement number can
take on. These are shown in Example 2.22.

Example 2.22
Decimal values that a 4-bit, one’s complement code can represent

Again, we notice that there are two different codes for 010 (00002 and 11112). This is a drawback of
one’s complement because it reduces the possible range of numbers that can be represented from 2n to
(2n1) and requires arithmetic operations that take into account the gap in the number system. There are
advantages of one’s complement, however. First, the numbers are ordered such that traditional addition
works on both positive and negative numbers (excluding the double 0 gap). Taking the example of
(5) + (1) again, in one’s complement the result yields 4, just as in a traditional decimal system. Notice
in one’s complement, 510 is represented with 10102. Adding 1 to this entire binary code would result in
the next higher binary code 10112 or 410 from the above table. This makes addition circuitry less
complicated, but still not as simple as if the double 0 gap was eliminated. Another advantage of one’s
complement is that as the numbers are incremented beyond the largest value in the set, they roll over
and start counting at the lowest number. For example, if you increment the number 01112 (710), it goes to
the next higher binary code 10002, which is 710. The ability to have the numbers roll over is a useful
feature for computer systems.
If we use the notation N1comp to represent any possible value that an n-bit, one’s complement
number can take on, the range is defined as:
   
Range of a ONE’S COMPLEMENT number )  2n1 1  N1’s comp  þ 2n1 1

Example 2.23 shows how to use this expression to find the range of decimal values that a 24-bit,
one’s complement code can represent.
28 • Chapter 2: Number Systems

Example 2.23
Finding the range of a 1’s complement number

The process of finding the decimal value of a one’s complement number involves first identifying
whether the number is positive or negative by looking at the sign bit. If the number is positive (i.e., the sign
bit is 0), then the number is treated as an unsigned code and is converted to decimal using the standard
conversion procedure described in prior sections. If the number is negative (i.e., the sign bit is 1), then the
number sign is recorded separately and the code is complemented in order to convert it to its positive
magnitude equivalent. This new positive number is then converted to decimal using the standard conver-
sion procedure. As the final step, the sign is applied. Example 2.24 shows an example of this process.

Example 2.24
Finding the decimal value of a 1’s complement number

2.4.2.3 Two’s Complement

Two’s complement is an encoding scheme that addresses the double 0 issue in signed magnitude
and 1’s complement representations. In this approach, the negative number is obtained by subtracting its
positive equivalent from 2n. This is identical to performing a complement on the positive equivalent and
then adding one. If a carry is generated, it is discarded. This procedure is called “taking the two’s
complement of a number”. The procedure of complementing each bit and adding one is the most
common technique to perform a two’s complement. In this way, the most significant bit of the number
is still the sign bit (0 ¼ positive, 1 ¼ negative) but all of the negative numbers are in essence shifted up so
 2.4 Unsigned and Signed Numbers • 29

that the double 0 gap is eliminated. Taking the two’s complement of a positive number will give its
negative counterpart and vice versa. Let’s look at the decimal values that a 4-bit, two’s complement
number can take on. These are shown in Example 2.25.

Example 2.25
Decimal values that a 4-bit, two’s complement code can represent

There are many advantages of two’s complement encoding. First, there is no double 0 gap, which
means that all possible 2n unique codes that can exist in an n-bit number are used. This gives the largest
possible range of numbers that can be represented. Another advantage of two’s complement is that
addition with negative numbers works exactly the same as decimal. In our example of (5) + (1), the
result (4). Arithmetic circuitry can be built to mimic the way our decimal arithmetic works without the
need to consider the double 0 gap. Finally, the rollover characteristic is preserved from one’s comple-
ment. Incrementing +7 by +1 will result in 8.

If we use the notation N2comp to represent any possible value that an n-bit, two’s complement
number can take on, the range is defined as:
   
Range of a TWO’S COMPLEMENT number )  2n1  N2’s comp  þ 2n1 1

Example 2.26 shows how to use this expression to find the range of decimal values that a 32-bit,
two’s complement code can represent.
30 • Chapter 2: Number Systems

Example 2.26
Finding the range of a two’s complement number

The process of finding the decimal value of a two’s complement number involves first identifying
whether the number is positive or negative by looking at the sign bit. If the number is positive (i.e., the
sign bit is 0), then the number is treated as an unsigned code and is converted to decimal using
the standard conversion procedure described in prior sections. If the number is negative (i.e., the sign
bit is 1), then the number sign is recorded separately and a two’s complement is performed on the code in
order to convert it to its positive magnitude equivalent. This new positive number is then converted to
decimal using the standard conversion procedure. The final step is to apply the sign. Example 2.27
shows an example of this process.

Example 2.27
Finding the decimal value of a two’s complement number

To convert a decimal number into its two’s complement code, the range is first checked to determine
whether the number can be represented with the allocated number of bits. The next step is to convert the
decimal number into unsigned binary. The final step is to apply the sign bit. If the original decimal number
was positive, then the conversion is complete. If the original decimal number was negative, then the
two’s complement is taken on the unsigned binary code to find its negative equivalent. Example 2.28
shows this procedure when converting 9910 to its 8-bit, two’s complement code.
 2.4 Unsigned and Signed Numbers • 31

Example 2.28
Finding the two’s complement code of a decimal number

2.4.2.4 Arithmetic with Two’s Complement

Two’s complement has a variety of arithmetic advantages. First, the operations of addition, subtrac-
tion and multiplication are handled exactly the same as when using unsigned numbers. This means that
duplicate circuitry is not needed in a system that uses both number types. Second, the ability to convert
a number from positive to its negative representation by performing a two’s complement means
that an adder circuit can be used for subtraction. For example, if we wanted to perform the subtraction
1310 – 410 ¼ 910, this is the same as performing 1310 + (410) ¼ 910. This allows us to use a single adder
circuit to perform both addition and subtraction as long as we have the ability to take the two’s
complement of a number. Creating a circuit to perform two’s complement can be simpler and faster
than building a separate subtraction circuit, so this approach can sometimes be advantageous.

There are specific rules for performing two’s complement arithmetic that must be followed to ensure
proper results. First, any carry or borrow that is generated is ignored. The second rule that must be
followed is to always check if two’s complement overflow occurred. Two’s complement overflow refers
to when the result of the operation falls outside of the range of values that can be represented by the
number of bits being used. For example, if you are performing 8-bit, two’s complement addition,
the range of decimal values that can be represented is 12810 to +12710. Having two input terms of
12710 (0111 11112) is perfectly legal because they can be represented by the 8-bits of the two’s
32 • Chapter 2: Number Systems

complement number; however, the summation of 12710 + 12710 ¼ 25410 (1111 11102). This number does
not fit within the range of values that can be represented and is actually the two’s complement code for
210, which is obviously incorrect. Two’s complement overflow occurs if any of the following occurs:

• The sum of like signs results in an answer with opposite sign (i.e., Positive + Positive ¼ Neg-
ative or Negative + Negative ¼ Positive)
• The subtraction of a positive number from a negative number results in a positive number (i.e.,
Negative – Positive ¼ Positive)
• The subtraction of a negative number from a positive number results in a negative number (i.e.,
Positive – Negative ¼ Negative)

Computer systems that use two’s complement have a dedicated logic circuit that monitors for any of
these situations and lets the operator know that overflow has occurred. These circuits are straightforward
since they simply monitor the sign bits of the input and output codes. Example 2.29 shows how to use
two’s complement in order to perform subtraction using an addition operation.

Example 2.29
Two’s complement addition
 Exercise Problems • 33

CONCEPT CHECK

CC2.4 A 4-bit, two’s complement number has 16 unique codes and can represent decimal numbers
between 810 to +710. If the number of unique codes is even, why is it that the range of
integers it can represent is not symmetrical about zero?
(A) One of the positive codes is used to represent zero. This prevents the highest
positive number from reaching +810 and being symmetrical.
(B) It is asymmetrical because the system allows the numbers to roll over.
(C) It isn’t asymmetrical if zero is considered a positive integer. That way there are
eight positive numbers and eight negatives numbers.
(D) It is asymmetrical because there are duplicate codes for 0.

Summary
v The base, or radix, of a number system refers v Binary codes can represent both unsigned
to the number of unique symbols within its and signed numbers. For an arbitrary n-bit
set. The definition of a number system binary code, it is important to know the
includes both the symbols used and the rela- encoding technique and the range of values
tive values of each symbol within the set. that can be represented.
v The most common number systems are base v Signed numbers use the most significant
10 (decimal), base 2 (binary), and base position to represent whether the number is
16 (hexadecimal). Base 10 is used because negative (0 ¼ positive, 1 ¼ negative). The
it is how the human brain has been trained to width of a signed number is always fixed.
treat numbers. Base 2 is used because the v Two’s complement is the most common
two values are easily represented using elec- encoding technique for signed numbers. It
trical switches. Base 16 is a convenient way has an advantage that there are no duplicate
to describe large groups of bits. codes for zero and that the encoding
v A positional number system allows larger approach provides a monotonic progression
(or smaller) numbers to be represented of codes from the most negative number that
beyond the values within the original symbol can be represented to the most positive. This
set. This is accomplished by having each posi- allows addition and subtraction to work the
tion within a number have a different weight. same on two’s complement numbers as it
v There are specific algorithms that are used to does on unsigned numbers.
convert any base to or from decimal. There v When performing arithmetic using two’s com-
are also algorithms to convert between num- plement codes, the carry bit is ignored.
ber systems that contain a power-of-two v When performing arithmetic using two’s com-
symbols (e.g., binary to hexadecimal and plement codes, if the result lies outside of the
hexadecimal to binary). range that can be represented it is called
v Binary arithmetic is performed on a fixed width two’s complement overflow. Two’s comple-
of bits (n). When an n-bit addition results in a ment overflow can be determined by looking
sum that cannot fit within n-bits, it generates a at the sign bits of the input arguments and the
carry out bit. In an n-bit subtraction, if the min- sign bit of the result.
uend is smaller than the subtrahend, a borrow
in can be used to complete the operation.

Exercise Problems
Section 2.1: Positional Number Systems 2.1.3 What is the radix of the hexadecimal number
system?
2.1.1 What is the radix of the binary number system?
2.1.4 What is the radix of the octal number system?
2.1.2 What is the radix of the decimal number
system? 2.1.5 For the number 261.367, what position (p) is
the number 2 in?
34 • Chapter 2: Number Systems

2.1.6 For the number 261.367, what position (p) is 2.2.11 Convert 6710 to binary. Treat all numbers as
the number 1 in? unsigned.
2.1.7 For the number 261.367, what position (p) is 2.2.12 Convert 252.98710 to binary. Treat all numbers
the number 3 in? as unsigned. Use an accuracy of 4 fractional
2.1.8 For the number 261.367, what position (p) is bits and don’t round up.
the number 7 in? 2.2.13 Convert 6710 to octal. Treat all numbers as
2.1.9 What is the name of the number system unsigned.
containing 102? 2.2.14 Convert 252.98710 to octal. Treat all numbers
2.1.10 What is the name of the number system as unsigned. Use an accuracy of 4 fractional
containing 1010? digits and don’t round up.
2.1.11 What is the name of the number system 2.2.15 Convert 6710 to hexadecimal. Treat all num-
containing 1016? bers as unsigned.
2.1.12 What is the name of the number system 2.2.16 Convert 252.98710 to hexadecimal. Treat all
containing 108? numbers as unsigned. Use an accuracy of
4 fractional digits and don’t round up.
2.1.13 Which of the four number systems covered in
this chapter (i.e., binary, decimal, hexadecimal, 2.2.17 Convert 1 0000 11112 to octal. Treat all num-
and octal) could the number 22 be part of? bers as unsigned.
Give all that are possible. 2.2.18 Convert 1 0000 1111.0112 to hexadecimal.
2.1.14 Which of the four number systems covered in Treat all numbers as unsigned.
this chapter (i.e., binary, decimal, hexadecimal, 2.2.19 Convert 778 to binary. Treat all numbers as
and octal) could the number 99 be part of? unsigned.
Give all that are possible. 2.2.20 Convert F.A16 to binary. Treat all numbers as
2.1.15 Which of the four number systems covered in unsigned.
this chapter (i.e., binary, decimal, hexadecimal, 2.2.21 Convert 668 to hexadecimal. Treat all numbers
and octal) could the number 1F be part of? as unsigned.
Give all that are possible.
2.2.22 Convert AB.D16 to octal. Treat all numbers as
2.1.16 Which of the four number systems covered in unsigned.
this chapter (i.e., binary, decimal, hexadecimal,
and octal) could the number 88 be part of?
Give all that are possible. Section 2.3: Binary Arithmetic
2.3.1 Compute 10102 + 10112 by hand. Treat all
Section 2.2: Base Conversions numbers as unsigned. Provide the 4-bit sum
and indicate whether a carry out occurred.
2.2.1 If the number 101.111 has a radix of 2,
what is the weight of the position containing 2.3.2 Compute 1111 11112 + 0000 00012 by hand.
the bit 0? Treat all numbers as unsigned. Provide the 8-bit
sum and indicate whether a carry out occurred.
2.2.2 If the number 261.367 has a radix of 10, what is
the weight of the position containing the 2.3.3 Compute 1010.10102 + 1011.10112 by hand.
numeral 2? Treat all numbers as unsigned. Provide the 8-bit
sum and indicate whether a carry out occurred.
2.2.3 If the number 261.367 has a radix of 16, what is
the weight of the position containing the 2.3.4 Compute 1111 1111.10112 + 0000 0001.11002
numeral 1? by hand. Treat all numbers as unsigned. Pro-
vide the 12-bit sum and indicate whether a
2.2.4 If the number 261.367 has a radix of 8, what is carry out occurred.
the weight of the position containing the
numeral 3? 2.3.5 Compute 10102  10112 by hand. Treat all
numbers as unsigned. Provide the 4-bit differ-
2.2.5 Convert 1100 11002 to decimal. Treat all num- ence and indicate whether a borrow in
bers as unsigned. occurred.
2.2.6 Convert 1001.10012 to decimal. Treat all num- 2.3.6 Compute 1111 11112  0000 00012 by hand.
bers as unsigned. Treat all numbers as unsigned. Provide the
2.2.7 Convert 728 to decimal. Treat all numbers as 8-bit difference and indicate whether a borrow
unsigned. in occurred.
2.2.8 Convert 12.578 to decimal. Treat all numbers 2.3.7 Compute 1010.10102  1011.10112 by hand.
as unsigned. Treat all numbers as unsigned. Provide the
2.2.9 Convert F316 to decimal. Treat all numbers as 8-bit difference and indicate whether a borrow
unsigned. in occurred.
2.2.10 Convert 15B.CEF16 to decimal. Treat all num- 2.3.8 Compute 1111 1111.10112  0000 0001.11002
bers as unsigned. Use an accuracy of 7 frac- by hand. Treat all numbers as unsigned. Pro-
tional digits. vide the 12-bit difference and indicate whether
a borrow in occurred.
 Exercise Problems • 35

Section 2.4: Unsigned and Signed 2.4.17 Compute 410 – 510 using 4-bit two’s comple-
ment addition. You will need to first convert
Numbers
each number into its 4-bit two’s complement
2.4.1 What range of decimal numbers can be code and then perform binary addition (i.e.,
represented by 8-bit, two’s complement 410 + (510)). Provide the 4-bit result and indi-
numbers? cate whether two’s complement overflow
2.4.2 What range of decimal numbers can be occurred. Check your work by converting the
represented by 16-bit, two’s complement 4-bit result back to decimal.
numbers? 2.4.18 Compute 710  710 using 4-bit two’s comple-
2.4.3 What range of decimal numbers can be ment addition. You will need to first convert
represented by 32-bit, two’s complement each decimal number into its 4-bit two’s com-
numbers? plement code and then perform binary addition
(i.e., 710 + (710)). Provide the 4-bit result and
2.4.4 What range of decimal numbers can be
indicate whether two’s complement overflow
represented by 64-bit, two’s complement
occurred. Check your work by converting the
numbers?
4-bit result back to decimal.
2.4.5 What is the 8-bit, two’s complement code for
2.4.19 Compute 710 + 110 using 4-bit two’s comple-
+8810?
ment addition. You will need to first convert
2.4.6 What is the 8-bit, two’s complement code for each decimal number into its 4-bit two’s com-
8810? plement code and then perform binary addi-
2.4.7 What is the 8-bit, two’s complement code tion. Provide the 4-bit result and indicate
for 12810? whether two’s complement overflow occurred.
2.4.8 What is the 8-bit, two’s complement code Check your work by converting the 4-bit result
for 110? back to decimal.

2.4.9 What is the decimal value of the 4-bit, two’s 2.4.20 Compute 6410  10010 using 8-bit two’s com-
complement code 00102? plement addition. You will need to first convert
each number into its 8-bit two’s complement
2.4.10 What is the decimal value of the 4-bit, two’s code and then perform binary addition (i.e.,
complement code 10102? 6410 + (10010)). Provide the 8-bit result and
2.4.11 What is the decimal value of the 8-bit, two’s indicate whether two’s complement overflow
complement code 0111 11102? occurred. Check your work by converting the
2.4.12 What is the decimal value of the 8-bit, two’s 8-bit result back to decimal.
complement code 1111 11102? 2.4.21 Compute (99)10  1110 using 8-bit two’s com-
2.4.13 Compute 11102 + 10112 by hand. Treat all plement addition. You will need to first convert
numbers as 4-bit, two’s complement codes. each decimal number into its 8-bit two’s com-
Provide the 4-bit sum and indicate whether plement code and then perform binary addition
two’s complement overflow occurred. (i.e., (9910) + (1110)). Provide the 8-bit result
and indicate whether two’s complement over-
2.4.14 Compute 1101 11112 + 0000 00012 by hand. flow occurred. Check your work by converting
Treat all numbers as 8-bit, two’s complement the 8-bit result back to decimal.
codes. Provide the 8-bit sum and indicate
whether two’s complement overflow occurred. 2.4.22 Compute 5010 + 10010 using 8-bit two’s com-
plement addition. You will need to first convert
2.4.15 Compute 1010.10102 + 1000.10112 by hand. each decimal number into its 8-bit two’s com-
Treat all numbers as 8-bit, two’s complement plement code and then perform binary addi-
codes. Provide the 8-bit sum and indicate tion. Provide the 8-bit result and indicate
whether two’s complement overflow occurred. whether two’s complement overflow occurred.
2.4.16 Compute 1110 1011.10012 + 0010 0001.11012 Check your work by converting the 8-bit result
by hand. Treat all numbers as 12-bit, two’s back to decimal.
complement codes. Provide the 12-bit sum
and indicate whether two’s complement over-
flow occurred.
Chapter 3: Digital Circuitry
and Interfacing
Now we turn our attention to the physical circuitry and electrical quantities that are used to represent
and operate on the binary codes 1 and 0. In this chapter we begin by looking at how logic circuits are
described and introduce the basic set of gates used for all digital logic operations. We then look at the
underlying circuitry that implements the basic gates including digital signaling and how voltages are used
to represent 1’s and 0’s. We then look at interfacing between two digital circuits and how to ensure that
when one circuit sends a binary code, the receiving circuit is able to determine which code was sent.
Logic families are then introduced and the details of how basic gates are implemented at the switch level
are presented. Finally, interfacing considerations are covered for the most common types of digital loads
(i.e., other gates, resistors, and LEDs). The goal of this chapter is to provide an understanding of the
basic electrical operation of digital circuits.
Learning Outcomes—After completing this chapter, you will be able to:

3.1 Describe the functional operation of a basic logic gate using truth tables, logic expressions,
and logic waveforms.
3.2 Analyze the DC and AC behavior of a digital circuit to verify it is operating within
specification.
3.3 Describe the meaning of a logic family and the operation of the most common technologies
used today.
3.4 Determine the operating conditions of a logic circuit when driving various types of loads.

3.1 Basic Gates

The term gate is used to describe a digital circuit that implements the most basic functions possible
within the binary system. When discussing the operation of a logic gate, we ignore the details of how the
1’s and 0’s are represented with voltages and manipulated using transistors. We instead treat the inputs
and output as simply ideal 1’s and 0’s. This allows us to design more complex logic circuits without going
into the details of the underlying physical hardware.

3.1.1 Describing the Operation of a Logic Circuit

3.1.1.1 The Logic Symbol
A logic symbol is a graphical representation of the circuit that can be used in a schematic to show
how circuits in a system interface to one another. For the set of basic logic gates, there are uniquely
shaped symbols that graphically indicate their functionality. For more complex logic circuits that are
implemented with multiple basic gates, a simple rectangular symbol is used. Inputs of the logic circuit are
typically shown on the left of the symbol and outputs are on the right. Figure 3.1 shows two example logic
symbols.

# Springer International Publishing AG 2017 37

B.J. LaMeres, Introduction to Logic Circuits & Logic Design with Verilog,
DOI 10.1007/978-3-319-53883-9_3
38 • Chapter 3: Digital Circuitry and Interfacing

Fig. 3.1
Example logic symbols

3.1.1.2 The Truth Table

We formally define the functionality of a logic circuit using a truth table. In a truth table, each and
every possible input combination is listed and the corresponding output of the logic circuit is given. If a
logic circuit has n inputs, then it will have 2n possible input codes. The binary codes are listed in
ascending order within the truth table mimicking a binary count starting at 0. By always listing the input
codes in this way, we can assign a row number to each input that is the decimal equivalent of the binary
input code. Row numbers can be used to simplify the notation for describing the functionality of larger
circuits. Figure 3.2 shows the formation of an example 3-input truth table.

Fig. 3.2
Truth table formation

3.1.1.3 The Logic Function

A logic expression, (also called a logic function), is an equation that provides the functionality of
each output in the circuit as a function of the inputs. The logic operations for the basic gates are given a
L
symbolic set of operators (e.g., +, ∙, ), the details of which will be given in the next sections. The logic
function describes the operations that are necessary to produce the outputs listed in the truth table. A
logic function is used to describe a single output that can take on only the values 1 and 0. If a circuit
contains multiple outputs, then a logic function is needed for each output. The input variables can be
included in the expression description just as in an analog function. For example, “F(A, B, C) ¼ . . .” would
state that “F is a function of the inputs A, B and C”. This can also be written as “FA, B, C ¼ . . .”. The input
variables can also be excluded for brevity as in “F ¼ . . .”. Figure 3.3 shows the formation of an example
3-input logic expression.
 3.1 Basic Gates • 39

Fig. 3.3
Logic function formation

3.1.1.4 The Logic Waveform

A logic waveform is a graphical depiction of the relationship of the output to the inputs with respect to
time. This is often a useful description of behavior since it mimics the format that is typically observed
when measuring a real digital circuit using test equipment such as an oscilloscope. In the waveform,
each signal can only take on a value of 1 or 0. It is useful to write the logic values of the signal at each
transition in the waveform for readability. Figure 3.4 shows an example logic waveform.

Fig. 3.4
Example logic waveform

3.1.2 The Buffer

The first basic gate is the buffer. The output of a buffer is simply the input. The logic symbol, truth
table, logic function and logic waveform for the buffer are given in Fig. 3.5.

Fig. 3.5
Buffer symbol, truth table, logic function and logic waveform
40 • Chapter 3: Digital Circuitry and Interfacing

3.1.3 The Inverter

The next basic gate is the inverter. The output of an inverter is the complement of the input. Inversion
is also often called the not operation. In spoken word, we might say “A is equal to not B”. thus this gate is
also often called a not gate. The symbol for the inverter is the same as the buffer with the exception that
an inversion bubble (i.e., a circle) is placed on the output. The inversion bubble is a common way to show
inversions in schematics and will be used by many of the basic gates. In the logic function, there are two
common ways to show this operation. The first way is by placing a prime (‘) after the input variable (e.g.,
Out ¼ In’). This notation has the advantage that it is supported in all text editors but has the drawback that
it can sometimes be difficult to see. The second way to indicate inversion in a logic function is by placing
an inversion bar over the input variable (e.g., Out ¼ Ine ). The advantage of this notation is that it is easy to
see but has the drawback that it is not supported by many text editors. In this text, both conventions will
be used to provide exposure to each. The logic symbol, truth table, logic function and logic waveform for
the inverter are given in Fig. 3.6.

Fig. 3.6
Inverter symbol, truth table, logic function and logic waveform

3.1.4 The AND Gate

The next basic gate is the AND gate. The output of an AND gate will only be true (i.e., a logic 1) if all
of the inputs are true. This operation is also called a logical product because if the inputs were
multiplied together, the only time the output would be a 1 is if each and every input was a 1. As a result,
the logic operator is the dot (∙). Another notation that is often seen is the ampersand (&). The logic
symbol, truth table, logic function and logic waveform for a 2-input AND gate are given in Fig. 3.7.

Fig. 3.7
2-Input AND gate symbol, truth table, logic function and logic waveform

Ideal AND gates can have any number of inputs. The operation of an n-bit, AND gates still follows
the rule that the output will only be true when all of the inputs are true. Later sections will discuss the
limitations on expanding the number of inputs of these basic gates indefinitely.
 3.1 Basic Gates • 41

3.1.5 The NAND Gate

The NAND gate is identical to the AND gate with the exception that the output is inverted. The “N” in
NAND stands for “NOT”, which represents the inversion. The symbol for a NAND gate is an AND gate
with an inversion bubble on the output. The logic expression for a NAND gate is the same as an AND
gate but with an inversion bar over the entire operation. The logic symbol, truth table, logic function and
logic waveform for a 2-input NAND gate are given in Fig. 3.8. Ideal NAND gates can have any number of
inputs with the operation of an n-bit, NAND gate following the rule that the output is always the inversion
of an n-bit, AND operation.

Fig. 3.8
2-Input NAND gate symbol, truth table, logic function and logic waveform

3.1.6 The OR Gate

The next basic gate is the OR gate. The output of an OR gate will be true when any of the inputs
are true. This operation is also called a logical sum because of its similarity to logical disjunction in which
the output is true if at least one of the inputs is true. As a result, the logic operator is the plus sign (+). The
logic symbol, truth table, logic function and logic waveform for a 2-input OR gate are given in Fig. 3.9.
Ideal OR gates can have any number of inputs. The operation of an n-bit, OR gates still follows the rule
that the output will be true if any of the inputs are true.

Fig. 3.9
2-Input OR gate symbol, truth table, logic function and logic waveform

3.1.7 The NOR Gate

The NOR gate is identical to the OR gate with the exception that the output is inverted. The symbol
for a NOR gate is an OR gate with an inversion bubble on the output. The logic expression for a NOR
gate is the same as an OR gate but with an inversion bar over the entire operation. The logic symbol,
truth table, logic function and logic waveform for a 2-input NOR gate are given in Fig. 3.10. Ideal NOR
gates can have any number of inputs with the operation of an n-bit, NOR gate following the rule that the
output is always the inversion of an n-bit, OR operation.
42 • Chapter 3: Digital Circuitry and Interfacing

Fig. 3.10
2-Input NOR gate symbol, truth table, logic function and logic waveform

3.1.8 The XOR Gate

The next basic gate is the exclusive-OR gate, or XOR gate for short. This gate is also called a
difference gate because for the 2-input configuration, its output will be true when the input codes are
L
different from one another. The logic operator is a circle around a plus sign ( ). The logic symbol, truth
table, logic function and logic waveform for a 2-input XOR gate are given in Fig. 3.11.

Fig. 3.11
2-Input XOR gate symbol, truth table, logic function and logic waveform

Using the formal definition of an XOR gate (i.e., the output is true if any of the input codes are
different from one another), an XOR gate with more than two inputs can be built. The truth table for a
3-bit, XOR gate using this definition is shown in Fig. 3.12. In modern electronics, this type of gate has
found little use since it is much simpler to build this functionality using a combination of AND and OR
gates. As such, XOR gates with greater than two inputs do not implement the difference function.
Instead, a more useful functionality has been adopted in which the output of the n-bit, XOR gate is the
result of a cascade of 2-input XOR gates. This results in an ultimate output that is true when there is an
ODD number of 1’s on the inputs. This functionality is much more useful in modern electronics for error
correction codes and arithmetic. As such, this is the functionality that is seen in modern n-bit, XOR gates.
This functionality is also shown in Fig. 3.12.
 3.1 Basic Gates • 43

Fig. 3.12
3-Input XOR gate implementation

3.1.9 The XNOR Gate

The exclusive-NOR gate is identical to the XOR gate with the exception that the output is inverted.
This gate is also called an equivalence gate because for the 2-input configuration, its output will be true
when the input codes are equivalent to one another. The symbol for an XNOR gate is an XOR gate
with an inversion bubble on the output. The logic expression for an XNOR gate is the same as an XOR
gate but with an inversion bar over the entire operation. The logic symbol, truth table, logic function and
logic waveform for a 2-input XNOR gate are given in Fig. 3.13. XNOR gates can have any number of
inputs with the operation of an n-bit, XNOR gate following the rule that the output is always the inversion
of an n-bit, XOR operation (i.e., the output is true if there is an ODD number of 1’s on the inputs).

Fig. 3.13
2-Input XNOR gate symbol, truth table, logic function and logic waveform
44 • Chapter 3: Digital Circuitry and Interfacing

CONCEPT CHECK

CC3.1 Given the following logic diagram, which is the correct logic expression for F?

L
(A) F ¼ (AB)0 C
L
(B) F ¼ (A0 B0 ) C
0 0
L
(C) F ¼ (A B C)
L
(D) F ¼ AB0 C

3.2 Digital Circuit Operation

Now we turn our attention to the physical hardware that is used to build the basic gates just
described and how electrical quantities are used to represent and communicate the binary values
1 and 0. We begin by looking at digital signaling. Digital signaling refers to how binary codes are
generated and transmitted successfully between two digital circuits using electrical quantities (e.g.,
voltage and current). Consider the digital system shown in Fig. 3.14. In this system, the sending circuit
generates a binary code. The sending circuit is called either the transmitter (Tx) or driver. The transmitter
represents the binary code using an electrical quantity such as voltage. The receiving circuit
(Rx) observes this voltage and is able to determine the value of the binary code. In this way, 1’s and
0’s can be communicated between the two digital circuits. The transmitter and receiver are both
designed to use the same digital signaling scheme so that they are able to communicate with each
other. It should be noted that all digital circuits contain both inputs (Rx) and outputs (Tx) but are not
shown in this figure for simplicity.

Fig. 3.14
Generic digital transmitter/receiver circuit

3.2.1 Logic Levels

A logic level is the term to describe all possible states that a signal can have. We will focus explicitly
on circuits that represent binary values so these will only have two finite states (1 and 0). To begin, we
define a simplistic model of how to represent the binary codes using an electrical quantity. This model
uses a voltage threshold (Vth) to represent the switching point between the binary codes. If the voltage of
 3.2 Digital Circuit Operation • 45

the signal (Vsig) is above this threshold, it is considered a logic HIGH. If the voltage is below this
threshold, it is considered a logic LOW. A graphical depiction of this is shown in Fig. 3.15. The terms
HIGH and LOW are used to describe which logic level corresponds to the higher or lower voltage.

Fig. 3.15
Definition of logic HIGH and LOW

It is straightforward to have the HIGH level correspond to the binary code 1 and the LOW level
correspond to the binary code 0; however, it is equally valid to have the HIGH level correspond to the
binary code 0 and the LOW level correspond to the binary code 1. As such, we need to define how the
logic levels HIGH and LOW map to the binary codes 1 and 0. We define two types of digital assignments:
Positive Logic and Negative Logic. In Positive Logic, the logic HIGH level represents a binary 1 and the
logic LOW level represents a binary 0. In Negative Logic, the logic HIGH level represents a binary 0 and
the logic LOW level represents a binary 1. Table 3.1 shows the definition of positive and negative logic.
There are certain types of digital circuits that benefit from using negative logic; however, we will focus
specifically on systems that use positive logic since it is more intuitive when learning digital design for the
first time. The transformation between positive and negative logic is straightforward and will be covered
in Chap. 4.

Table 3.1
Definition of positive and negative logic

3.2.2 Output DC Specifications

Transmitting circuits provide specifications on the range of output voltages (VO) that they are
guaranteed to provide when outputting a logic 1 or 0. These are called the DC output specifications.
There are four DC voltage specifications that specify this range: VOH-max, VOH-min, VOL-max, and VOL-min.
The VOH-max and VOH-min specifications provide the range of voltages the transmitter is guaranteed to
provide when outputting a logic HIGH (or logic 1 when using positive logic). The VOL-max and VOL-min
specifications provide the range of voltages the transmitter is guaranteed to provide when outputting a
46 • Chapter 3: Digital Circuitry and Interfacing

logic LOW (or logic 0 when using positive logic). In the subscripts for these specifications, the “O”
signifies “output” and the “L” or “H” signifies “LOW” or “HIGH” respectively.
The maximum amount of current that can flow through the transmitter’s output (IO) is also specified.
The specification IOH-max is the maximum amount of current that can flow through the transmitter’s output
when sending a logic HIGH. The specification IOL-max is the maximum amount of current that can flow
through the transmitter’s output when sending a logic LOW. When the maximum output currents are
violated, it usually damages the part. Manufacturers will also provide a recommended amount of current
for IO that will guarantee the specified operating parameters throughout the life of the part. Figure 3.16
shows a graphical depiction of these DC specifications. When the transmitter output is providing current
to the receiving circuit (a.k.a., the load), it is said to be sourcing current. When the transmitter output is
drawing current from the receiving circuit, it is said to be sinking current. In most cases, the transmitter
sources current when driving a logic HIGH and sinks current when driving a logic LOW. Figure 3.16
shows a graphical depiction of these specifications.

Fig. 3.16
DC Specifications of a digital circuit

3.2.3 Input DC Specifications

Receiving circuits provide specifications on the range of input voltages (VI) that they will interpret as
either a logic HIGH or LOW. These are called the DC input specifications. There are four DC voltage
specifications that specify this range: VIH-max, VIH-min, VIL-max, and VIL-min. The VIH-max and VIH-min
specifications provide the range of voltages that the receiver will interpret as a logic HIGH (or logic
 3.2 Digital Circuit Operation • 47

1 when using positive logic). The VIL-max and VIL-min specifications provide the range of voltages that the
receiver will interpret as a logic LOW (or logic 0 when using positive logic). In the subscripts for these
specifications, the “I” signifies “input”.
The maximum amount of current that the receiver will draw, or take in, when connected is also
specified II). The specification IIH-max is the maximum amount of current that the receiver will draw when it
is being driven with a logic HIGH. The specification IIL-max is the maximum amount of current that the
receiver will draw when it is being driven with a logic LOW. Again, Fig. 3.16 shows a graphical depiction
of these DC specifications.

3.2.4 Noise Margins

For digital circuits that are designed to operate with each other, the VOH-max and VIH-max
specifications have equal voltages. Similarly, the VOL-min and VIL-min specifications have equal voltages.
The VOH-max and VOL-min output specifications represent the best case scenario for digital signaling as
the transmitter is sending the largest (or smallest) signal possible. If there is no loss in the interconnect
between the transmitter and receiver, the full voltage levels will arrive at the receiver and be interpreted
as the correct logic states (HIGH or LOW).
The worst-case scenario for digital signaling is when the transmitter outputs its levels at VOH-min and
VOL-max. These levels represent the furthest away from an ideal voltage level that the transmitter can
send to the receiver and are susceptible to loss and noise that may occur in the interconnect system. In
order to compensate for potential loss or noise, digital circuits have a predefined amount of margin built
into their worst-case specifications. Let’s take the worst-case example of a transmitter sending a logic
HIGH at the level VOH-min. If the receiver was designed to have VIH-min (i.e., the lowest voltage that would
still be interpreted as a logic 1) equal to VOH-min, then if even the smallest amount of the output signal was
attenuated as it traveled through the interconnect, it would arrive at the receiver below VIH-min and would
not be interpreted as a logic 1. Since there will always be some amount of loss in any interconnect
system, the specifications for VIH-min is always less than VOH-min. The difference between these two
quantities is called the Noise Margin. More specifically, it is called the Noise Margin HIGH (or NMH) to
signify how much margin is built into the Tx/Rx circuit when communicating a logic 1. Similarly, the VIL-
max specification is always higher than the VOL-max specification to account for any voltage added to the
signal in the interconnect. The difference between these two quantities is called the Noise Margin LOW
(or NML) to signify how much margin is built into the Tx/Rx circuit when communicating a logic 0. Noise
margins are always specified as positive quantities, thus the order of the subtrahend and minuend in
these differences.

NMH ¼ VOH-min VIH-min

NML ¼ VIL-max VOL-max

Figure 3.16 includes the graphical depiction of the noise margins. Notice in this figure that there is a
region of voltages that the receiver will not interpret as either a HIGH or LOW. This region lies between
the VIH-min and VIL-max specifications. This is the uncertainty region and should be avoided. Signals in
this region will cause the receiver’s output to go to an unknown voltage. Digital transmitters are designed
to transition between the LOW and HIGH states quickly enough so that the receiver does not have time to
react to the input being in the uncertainty region.
48 • Chapter 3: Digital Circuitry and Interfacing

3.2.5 Power Supplies

All digital circuits require a power supply voltage and a ground. There are some types of digital
circuits that may require multiple power supplies. For simplicity, we will focus on digital circuits that only
require a single power supply voltage and ground. The power supply voltage is commonly given the
abbreviations of either VCC or VDD. The “CC” or “DD” have to do with how the terminals of the transistors
inside of the digital circuit are connected (i.e., “collector to collector” or “drain to drain”). Digital circuits will
specify the required power supply voltage. Ground is considered an ideal 0v. Digital circuits will also
specify the maximum amount of DC current that can flow through the VCC (ICC) and GND (IGND) pins
before damaging the part.

There are two components of power supply current. The first is the current that is required for the
functional operation of the device. This is called the quiescent current (Iq). The second component of the
power supply current is the output currents (IO). Any current that flows out of a digital circuit must also
flow into it. When a transmitting circuit sources current to a load on its output pin, it must bring in that
same amount of current on another pin. This is accomplished using the power supply pin (VCC).
Conversely, when a transmitting circuit sinks current from a load on its output pin, an equal amount of
current must exit the circuit on a different pin. This is accomplished using the GND pin. This means that
the amount of current flowing through the VCC and GND pins will vary depending on the logic states that
are being driven on the outputs. Since a digital circuit may contain numerous output pins, the maximum
amount of current flowing through the VCC and GND pins can scale quickly and care must be taken not to
damage the device.

The quiescent current is often specified using the term ICC. This should not be confused with the
specification for the maximum amount of current that can flow through the VCC pin, which is often called
ICC-max. It is easy to tell the difference because ICC (or Iq) is much smaller than ICC-max for CMOS parts.
ICC (or Iq) is specified in the uA to nA range while the maximum current that can flow through the VCC pin
is specified in the mA range. Example 3.1 shows the process of calculating the ICC and IGND currents
when sourcing multiple loads.
 3.2 Digital Circuit Operation • 49

Example 3.1
Calculating ICC and IGND when sourcing multiple loads
50 • Chapter 3: Digital Circuitry and Interfacing

Example 3.2 shows the process of calculating the ICC and IGND currents when both sourcing and
sinking loads.

Example 3.2
Calculating ICC and IGND when both sourcing and sinking loads
 3.2 Digital Circuit Operation • 51

3.2.6 Switching Characteristics

Switching characteristics refer to the transient behavior of the logic circuits. The first group of
switching specifications characterize the propagation delay of the gate. The propagation delay is the
time it takes for the output to respond to a change on the input. The propagation delay is formally defined
as the time it takes from the point at which the input has transitioned to 50% of its final value to the point at
which the output has transitioned to 50% of its final value. The initial and final voltages for the input are
defined to be GND and VCC, while the output initial and final voltages are defined to be VOL and VOH.
Specifications are given for the propagation delay when transitioning from a LOW to HIGH (tPLH) and
from a HIGH to LOW (tPHL). When these specifications are equal, the values are often given as a single
specification of tpd. These specifications are shown graphically in Fig. 3.17.
The second group of switching specifications characterize how quickly the output switches between
states. The transition time is defined as the time it takes for the output to transition from 10% to 90% of
the output voltage range. The rise time (tr) is the time it takes for this transition when going from a LOW to
HIGH, and the fall time (tf) is the time it takes for this transition when going from a HIGH to LOW. When
these specifications are equal, the values are often given as a single specification of tt. These
specifications are shown graphically in Fig. 3.17.

Fig. 3.17
Switching characteristics of a digital circuit

3.2.7 Data Sheets

The specifications for a particular part are given in its data sheet. The data sheet contains all of the
operating characteristics for a part, in addition to functional information such as package geometries and
pin assignments. The data sheet is usually the first place a designer will look when selecting a part.
Figures 3.18, 3.19, and 3.20 show excerpts from an example data sheet highlighting some of the
specifications just covered.
52 • Chapter 3: Digital Circuitry and Interfacing

Fig. 3.18
Example data sheet excerpt (1)
 3.2 Digital Circuit Operation • 53

Fig. 3.19
Example data sheet excerpt (2)
54 • Chapter 3: Digital Circuitry and Interfacing

Fig. 3.20
Example data sheet excerpt (3)
 3.2 Digital Circuit Operation • 55

CONCEPT CHECK

CC3.2(a) Given the following DC specifications for a driver/receiver pair, in what situation may a logic
signal transmitted not be successfully received?

VOH-max ¼ þ3:4v VIH-max ¼ þ3:4v

VOH-min ¼ þ2:5v VIH-min ¼ þ2:5v
VOL-max ¼ þ1:5v VIL-max ¼ þ2:0v
VOL-min ¼ 0v VIL-min ¼ 0v

(A) Driving a HIGH with Vo ¼ +3.4v

(B) Driving a HIGH with Vo ¼ +2.5v
(C) Driving a LOW with Vo ¼ +1.5v
(D) Driving a LOW with Vo ¼ 0v

CC3.2(b) For the following driver configuration, which of the following is a valid constraint that could
be put in place to prevent a violation of the maximum power supply currents (ICC-max and
IGND-max)?

(A) Modify the driver transistors so that they can’t provide more than 5 mA on any
output.
(B) Apply a cooling system (e.g., a heat sink or fan) to the driver chip.
(C) Design the logic so that no more than half of the outputs are HIGH at any
given time.
(D) Drive multiple receivers with the same output pin.

CC3.2(c) Why is it desirable to have the output of a digital circuit transition quickly between the logic
LOW and logic HIGH levels?
(A) So that the outputs are not able to respond as the input transitions through the
uncertainty region. This avoids unwanted transitions.
(B) So that all signals look like square waves.
(C) To reduce power by minimizing the time spent switching.
(D) Because the system can only have two states, a LOW and a HIGH.
56 • Chapter 3: Digital Circuitry and Interfacing

3.3 Logic Families

It is apparent from the prior discussion of operating conditions that digital circuits need to have
comparable input and output specifications in order to successfully communicate with each other. If a
transmitter outputs a logic HIGH as +3.4v and the receiver needs a logic HIGH to be above +4v to be
successfully interpreted as a logic HIGH, then these two circuits will not be able to communicate. In order
to address this interoperability issue, digital circuits are grouped into Logic Families. A logic family is a
group of parts that all adhere to a common set of specifications so that they work together. The logic
family is given a specific name and once the specifications are agreed upon, different manufacturers
produce parts that work within the particular family. Within a logic family, parts will all have the same
power supply requirements and DC input/output specifications such that if connected directly, they will be
able to successfully communicate with each other. The phrase “connected directly” is emphasized
because it is very possible to insert an interconnect circuit between two circuits within the same logic
family and alter the output voltage enough so that the receiver will not be able to interpret the correct logic
level. Analyzing the effect of the interconnect circuit is part of the digital design process. There are many
logic families that exist (up to 100 different types!) and more emerge each year as improvements are
made to circuit fabrication processes that create smaller, faster and lower power circuits.

3.3.1 Complementary Metal Oxide Semiconductors (CMOS)

The first group of logic families we will discuss is called Complementary Metal Oxide
Semiconductors, or CMOS. This is currently the most popular group of logic families for digital circuits
implemented on the same integrated circuit (IC). An integrated circuit is where the entire circuit is
implemented on a single piece of semiconductor material (or chip). The IC can contain transistors,
resistors, capacitors, inductors, wires and insulators. Modern integrated circuits can contain billions of
devices and meters of interconnect. The opposite of implementing the circuit on an integrated circuit is to
use discrete components. Using discrete components refers to where every device (transistor, resistor,
etc.) is its own part and is wired together externally using either a printed circuit board (PCB) or jumper
wires as on a breadboard. The line between ICs and discrete parts has blurred in the past decades
because modern discrete parts are actually fabricated as an IC and regularly contain multiple devices
(e.g., 4 logic gates per chip). Regardless, the term discrete is still used to describe components that only
contain a few components where the term IC typically refers to a much larger system that is custom
designed.
The term CMOS comes from the use of particular types of transistors to implement the digital
circuits. The transistors are created using a Metal Oxide Semiconductor (MOS) structure. These
transistors are turned on or off based on an electric field, so they are given the name Metal Oxide
Semiconductor Field Effect Transistors, or MOSFETs. There are two transistors that can be built using
this approach that operate complementary to each other, thus the term Complementary Metal Oxide
Semiconductors. To understand the basic operation of CMOS logic, we begin by treating the MOSFETs
as ideal switches. This allows us to understand the basic functionality without diving into the detailed
electronic analysis of the transistors.

3.3.1.1 CMOS Operation

In CMOS, there is a single power supply (VCC or VDD) and a single ground (GND). The ground signal
is sometimes called VSS. The maximum input and output DC specifications are equal to the power supply
(VCC ¼ VOH-max ¼ VIH-max). The minimum input and output DC specification are equal to ground
(GND ¼ 0v ¼ VOL-min ¼ VIL-min). In this way, using CMOS simplifies many of the specifications. If you
state that you are using “CMOS with a +3.4v power supply”, you are inherently stating that
VCC ¼ VOH-max ¼ VIH-max ¼ +3.4v and that VOL-min ¼ VIL-min ¼ 0v. Many times the name of the logic
 3.3 Logic Families • 57

family will be associated with the power supply voltage. For example, a logic family may go by the name
“+3.3v CMOS” or “+2.5v CMOS”. These names give a first level description of the logic family operation,
but more details about the operation must be looked up in the data sheet.
There are two types of transistors used in CMOS. The transistors will be closed or open based on an
input logic level. The first transistor is called an N-type MOSFET, or NMOS. This transistor will turn on, or
close, when the voltage between the gate and source (VGS) is greater than its threshold voltage. The
threshold voltage (VT) is the amount of voltage needed to create a conduction path between the drain
and the source terminals. The threshold voltage of an NMOS transistor is typically between 0.2v to 1v
and much less than the VCC voltage in the system. The second transistor is called a P-type MOSFET, or
PMOS. This transistor turns on, or closes, when the voltage between the gate and the source (VGS) is
less than VT, where the VT for a PMOS is a negative value. This means that to turn on a PMOS transistor,
the gate terminal needs to be at a lower voltage than the source. The type of transistor (i.e., P-type or
N-type) has to do with the type of semiconductor material used to conduct current through the transistor.
An NMOS transistor uses negative charge to conduct current (i.e., Negative-Type) while a PMOS uses
positive charge (i.e., Positive-Type). Figure 3.21 shows the symbols for the PMOS and NMOS, the
fabrication cross-sections, and their switch level equivalents.

Fig. 3.21
CMOS transistors
58 • Chapter 3: Digital Circuitry and Interfacing

The basic operation of CMOS is that when driving a logic HIGH the switches are used to connect the
output to the power supply (VCC), and when driving a logic LOW the switches are used to connect
the output to GND. In CMOS, VCC is considered an ideal logic HIGH and GND is considered an ideal
logic LOW. VCC is typically much larger than VT so using these levels can easily turn on and off the
transistors. The design of the circuit must never connect the output to VCC and GND at the same time or
else the device itself will be damaged due to the current flowing directly from VCC to GND through the
transistors. Due to the device physics of the MOSFETS, PMOS transistors are used to form the network
that will connect the output to VCC (a.k.a., the pull-up network), and NMOS transistors are used to form
the network that will connect the output to GND (a.k.a., the pull-down network). Since PMOS transistors
are closed when the input is a 0 (thus providing a logic HIGH on the output) and NMOS transistors are
closed when the input is a 1 (thus providing a logic LOW on the output), CMOS implements negative
logic gates. This means CMOS can implement inverters, NAND and NOR gates but not buffers, AND
and OR gates directly. In order to create a CMOS AND gate, the circuit would implement a NAND gate
followed by an inverter and similarly for an OR gate and buffer.

3.3.1.2 CMOS Inverter

Let’s now look at how we can use these transistors to create a CMOS inverter. Consider the
transistor arrangement shown in Fig. 3.22.

Fig. 3.22
CMOS inverter schematic

The inputs of both the PMOS and NMOS are connected together. The PMOS is used to connect
the output to VCC and the NMOS is used to connect the output to GND. Since the inputs are
connected together and the switches operate in a complementary manner, this circuit ensures that
both transistors will never be on at the same time. When In ¼ 0, the PMOS switch is closed and the
NMOS switch is open. This connects the output directly to VCC, thus providing a logic HIGH on the
output. When In ¼ 1, the PMOS switch is open and the NMOS switch is closed. This connects the output
directly to GND, thus providing a logic LOW. This configuration yields an inverter. This operation is
shown graphically in Fig. 3.23.
 3.3 Logic Families • 59

Fig. 3.23
CMOS inverter operation

3.3.1.3 CMOS NAND Gate

Let’s now look at how we use a similar arrangement of transistors to implement a 2-input NAND
gate. Consider the transistor configuration shown in Fig. 3.24.

Fig. 3.24
CMOS 2-input NAND gate schematic
60 • Chapter 3: Digital Circuitry and Interfacing

The pull-down network consists of two NMOS transistors in series (M1 and M2) and the pull-up
network consists of two PMOS transistors in parallel (M3 and M4). Let’s go through each of the input
conditions and examine which transistors are on and which are off and how they impact the output. The
first input condition is when A ¼ 0 and B ¼ 0. This condition turns on both M3 and M4 creating two parallel
paths between the output and VCC. At the same time, it turns off both M1 and M2 preventing a path
between the output and GND. This input condition results in an output that is connected to VCC resulting
in a logic HIGH. The second input condition is when A ¼ 0 and B ¼ 1. This condition turns on M3 in the
pull-up network and M2 in the pull-down network. This condition also turns off M4 in the pull-up network
and M1 in the pull-down network. Since the pull-up network is a parallel combination of PMOS
transistors, there is still a path between the output and VCC through M3. Since the pull-down network
is a series combination of NMOS transistors, both M1 and M2 must be on in order to connect the output
to GND. This input condition results in an output that is connected to VCC resulting in a logic HIGH. The
third input condition is when A ¼ 1 and B ¼ 0. This condition again provides a path between the output
and VCC through M4 and prevents a path between the output and ground by having M2 open. This input
condition results in an output that is connected to VCC resulting in a logic HIGH. The final input condition
is when A ¼ 1 and B ¼ 1. In this input condition, both of the PMOS transistors in the pull-up network
(M3 and M4) are off preventing the output from being connected to VCC. At the same time, this input turns
on both M1 and M2 in the pull-down network connecting the output to GND. This input condition results in
an output that is connected to GND resulting in a logic LOW. Based on the resulting output values
corresponding to the four input codes, this circuit yields the logic operation of a 2-Input NAND gate. This
operation is shown graphically in Fig. 3.25.
 3.3 Logic Families • 61

Fig. 3.25
CMOS 2-input NAND gate operation

Creating a CMOS NAND gate with more than 2 inputs is accomplished by adding additional PMOS
transistors to the pull-up network in parallel and additional NMOS transistors to the pull-down network in
series. Figure 3.26 shows the schematic for a 3-Input NAND gate. This procedure is followed for creating
NAND gates with larger numbers of inputs.
62 • Chapter 3: Digital Circuitry and Interfacing

Fig. 3.26
CMOS 3-input NAND gate schematic

If the CMOS transistors were ideal switches, the approach of increasing the number of inputs could
be continued indefinitely. In reality, the transistors are not ideal switches and there is a limit on how many
transistors can be added in series and continue to operate. The limitation has to do with ensuring that
each transistor has enough voltage to properly turn on or off. This is a factor in the series network
because the drain terminals of the NMOS transistors are not all connected to GND. If a voltage develops
across one of the lower transistors (e.g., M3), then it takes more voltage on the input to turn on the next
transistor up (e.g., M2). If too many transistors are added in series, then the uppermost transistor in the
series may not be able to be turned on or off by the input signals. The number of inputs that a logic gate
can have within a particular logic family is called its fan-in specification. When a logic circuit requires a
number of inputs that exceeds the fan-in specification for a particular logic family, then additional logic
gates must be used. For example, if a circuit requires a 5-input NAND gate but the logic family has a
fan-in specification of 4, this means that the largest NAND gate available only has 4-inputs. The 5-input
NAND operation must be accomplished using additional circuit design techniques that use gates with
4 or less inputs. These design techniques will be covered in Chap. 4.

3.3.1.4 CMOS NOR Gate

A CMOS NOR gate is created using a similar topology as a NAND gate with the exception that the
pull-up network consists of PMOS transistors in series and the pull-down network that consists of NMOS
transistors in parallel. Consider the transistor configuration shown in Fig. 3.27.
 3.3 Logic Families • 63

Fig. 3.27
CMOS 2-input NOR gate schematic

The series configuration of the pull-up network will only connect the output to VCC when both inputs
are 0. Conversely, the pull-down network prevents connecting the output to GND when both inputs are
0. When either or both of the inputs are true, the pull-up network is off and the pull-down network is
on. This yields the logic function for a NOR gate. This operation is shown graphically in Fig. 3.28. As with
the NAND gate, the number of inputs can be increased by adding more PMOS transistors in series in the
pull-up network and more NMOS transistors in parallel in the pull-down network.
64 • Chapter 3: Digital Circuitry and Interfacing

Fig. 3.28
CMOS 2-input NOR gate operation

The schematic for a 3-input NOR gate is given in Fig. 3.29. This approach can be used to increase
the number of inputs up until the fan-in specification of the logic family is reached.
 3.3 Logic Families • 65

Fig. 3.29
CMOS 3-input NOR gate schematic

3.3.2 Transistor-Transistor Logic (TTL)

One of the first logic families that emerged after the invention of the integrated circuit was Transistor-
Transistor Logic (TTL). This logic family uses bipolar junction transistor (BJT) as its fundamental
switching item. This logic family defined a set of discrete parts that contained all of the basic gates in
addition to more complex building blocks. TTL was used to build the first computer systems in the 1960s.
TTL is not widely used today other than for specific applications because it consumes more power than
CMOS and cannot achieve the density required for today’s computer systems. TTL is discussed
because it was the original logic family based on integrated circuits so it provides a historical perspective
of digital logic. Furthermore, the discrete logic pin-outs and part-numbering schemes are still used today
for discrete CMOS parts.

3.3.2.1 TTL Operation

TTL logic uses BJT transistors and resistors to accomplish the logic operations. The operation of a
BJT transistor is more complicated than a MOSFET; however, it performs essentially the same switch
operation when used in a digital logic circuit. An input is used to turn the transistor on, which in turn allows
current to flow between two other terminals. Figure 3.30 shows the symbol for the two types of BJT
transistors. The PNP transistor is analogous to a PMOS and the NPN is analogous to an NMOS. Current
will flow between the Emitter and Collector terminals when there is a sufficient voltage on the Base
terminal. The amount of current that flows between the Emitter and Collector is related to the current
flowing into the Base. The primary difference in operation between BJTs and MOSFETs is that BJTs
require proper voltage biasing in order to turn on and also draws current through the BASE in order to
stay on. The detailed operation of BJTs is beyond the scope of this text, so an overly simplified model of
TTL logic gates is given.
66 • Chapter 3: Digital Circuitry and Interfacing

Fig. 3.30
PNP and NPN transistors

Figure 3.31 shows a simplified model of how TTL logic operates using BJTs and resistors. This
simplified model does not show all of the transistors that are used in modern TTL circuits but instead is
intended to provide a high-level overview of the operation. This gate is an inverter that is created with an
NPN transistor and a resistor. When the input is a logic HIGH, the NPN transistor turns on and conducts
current between its collector and emitter terminals. This in effect closes the switch and connects the
output to GND providing a logic LOW. During this state, current will also flow through the resistor to GND
through Q1 thus consuming more power than the equivalent gate in CMOS. When the input is a logic
LOW, the NPN transistor turns off and no current flows between its collector and emitter. This, in effect, is
an open circuit leaving only the resistor connected to the output. The resistor pulls the output up to VCC
providing a logic HIGH on the output. One drawback of this state is that there will be a voltage drop
across the resistor so the output is not pulled fully to VCC.
 3.3 Logic Families • 67

Fig. 3.31
TTL inverter

3.3.3 The 7400 Series Logic Families

The 7400 series of TTL circuits became popular in the 1960s and 1970s. This family was based on
TTL and contained hundreds of different digital circuits. The original circuits came in either plastic or
ceramic Dual-In-Line packages (DIP). The 7400 TTL logic family was powered off of a + 5v supply. As
mentioned before, this logic family set the pin-outs and part-numbering schemes for modern logic
families. There were many derivatives of the original TTL logic family that made modifications to improve
speed, reliability, decrease power and reduce power supplies. Today’s CMOS logic families within the
7400 series still use the same pin-outs and numbering schemes as the original TTL family. It is useful to
understand the history of this series because these parts are often used in introductory laboratory
exercises to learn how to interface digital logic circuits.
68 • Chapter 3: Digital Circuitry and Interfacing

3.3.3.1 Part Numbering Scheme

The part numbering scheme for the 7400 series and its derivatives contains five different fields:
(1) manufacturer, (2) temperature range, (3) logic family, (4) logic function and (5) package type. The
breakdown of these fields is shown in Fig. 3.32.

Fig. 3.32
7400 series part numbering scheme
 3.3 Logic Families • 69

3.3.3.2 DC Operating Conditions

Table 3.2 gives the DC operating conditions for a few of the logic families within the 7400 series.
Notice that the CMOS families consume much less power than the TTL families. Also notice that the TTL
output currents are asymmetrical. The differences between the IOH and IOL within the TTL families has to
do with the nature of the bipolar transistors and the resistors used to create the pull-up networks within
the devices. CMOS has symmetrical drive currents due to using complementary transistors for the pull-
up (PMOS) and pull-down networks (NMOS).

Table 3.2
DC operating conditions for a sample of 7400 series logic families

3.3.3.3 Pin-out Information for the DIP Packages

Figure 3.33 shows the pin-out assignments for a subset of the basic gates from the 74HC logic
family in the Dual-In-Line package form factor. Most of the basic gates within the 7400 series follow these
assignments. Notice that each of these basic gates comes in a 14-pin DIP package, each with a single
VCC and single GND pin. It is up to the designer to ensure that the maximum current flowing through the
VCC and GND pins does not exceed the maximum specification. This is particularly important for parts
that contain numerous gates. For example, the 74HC00 part contains four, 2-Input NAND gates. If each
of the NAND gates was driving a logic HIGH at its maximum allowable output current (i.e., 25 mA from
Fig. 3.19), then a total of 4∙25 mA + Iq ¼ ~100 mA would be flowing through its VCC pin. Since the VCC pin
can only tolerate a maximum of 50 mA of current (from Fig. 3.19), the part would be damaged since the
output current of ~100 mA would also flow through the VCC pin. The pin-outs in Fig. 3.33 are useful when
first learning to design logic circuits because the DIP packages plug directly into a standard breadboard.
70 • Chapter 3: Digital Circuitry and Interfacing

Fig. 3.33
Pin-outs for a subset of basic gates from the 74HC logic family in DIP packages
 3.4 Driving Loads • 71

CONCEPT CHECK

CC3.3 Why doesn’t the following CMOS transistor configuration yield a buffer?

(A) In order to turn on the NMOS transistor, VGS needs to be greater than zero. In the
given configuration, the gate terminal of the NMOS (G) needs to be driven above
the source terminal (S). If the source terminal was at +3.4v, then the input
(In) would never be able to provide a positive enough voltage to ensure the NMOS
is on because “In” doesn’t go above +3.4v.
(B) There is no way to turn on both transistors in this configuration.
(C) The power consumption will damage the device because both transistors will
potentially be on.
(D) The sources of the two devices can’t be connected together without causing a
short in the device.

3.4 Driving Loads

At this point we’ve discussed in depth how proper care must be taken to ensure that not only do the
output voltages of the driving gate meet the input specifications of the receiver in order to successfully
transmit 1’s and 0’s, but that the output current of the driver does not exceed the maximum specifications
so that the part is not damaged. The output voltage and current for a digital circuit depends greatly on the
load that is being driven. The following sections discuss the impact of driving some of the most common
digital loads.

3.4.1 Driving Other Gates

Within a logic family, all digital circuits are designed to operate with one another. If there is minimal
loss or noise in the interconnect system, then 1’s and 0’s will be successfully transmitted and no current
specifications will be exceeded. Consider the example in Example 3.3 for an inverter driving another
inverter from the same logic family.
72 • Chapter 3: Digital Circuitry and Interfacing

Example 3.3
Determining if specifications are violated when driving another gate as a load

From this example, it is clear that there are no issues when a gate is driving another gate from the
same family. This is as expected because that is the point of a logic family. In fact, gates are designed to
drive multiple gates from within their own family. Based solely on the DC specifications for input and
output current, it could be assumed that the number of other gates that can be driven is simply IO-max/II-
max. For the example in Example 3.3, this would result in a 74HC gate being able to drive 25,000 other
gates (i.e., 25 mA/1 uA ¼ 25,000). In reality, the maximum number of gates that can be driven is dictated
by the switching characteristics. This limit is called the fan-out specification. The fan-out specification
states the maximum number of other gates from within the same family that can be driven. As discussed
earlier, the output signal needs to transition quickly through the uncertainty region so that the receiver
does not have time to react and go to an unknown state. As more and more gates are driven, this
transition time is slowed down. The fan-out specification provides a limit to the maximum number of
gates from the same family that can be driven while still ensuring that the output signal transitions
between states fast enough to avoid the receivers from going to an unknown state. Example 3.4 shows
the process of determining the maximum output current that a driver will need to provide when driving the
maximum number of gates allowed by the fan-out specification.
 3.4 Driving Loads • 73

Example 3.4
Determining the output current when driving multiple gates as the load

3.4.2 Driving Resistive Loads

There are many situations where a resistor is the load in a digital circuit. A resistive load can be an
actual resistor that is present for some other purpose such as a pull-up, pull-down, or for impedance
matching. More complex loads such as buzzers, relays or other electronics can also be modeled as a
resistor. When a resistor is the load in a digital circuit, care must be taken to avoid violating the output
current specifications of the driver. The electrical circuit analysis technique that is used to evaluate how a
resistive load impacts a digital circuit is Ohm’s Law. Ohm’s Law is a very simple relationship between the
current and voltage in a resistor. Figure 3.34 gives a primer on Ohm’s Law. For use in digital circuits,
there are only a select few cases that this technique will be applied to, so no prior experience with Ohm’s
Law is required at this point.

Fig. 3.34
A primer on Ohm’s law
74 • Chapter 3: Digital Circuitry and Interfacing

Let’s see how we can use Ohm’s Law to analyze the impact of a resistive load in a digital circuit.
Consider the circuit configuration in Example 3.5 and how we can use Ohm’s Law to determine the
output current of the driver. The load in this case is a resistor connected between the output of the driver
and the power supply (+5v). When driving a logic HIGH, the output level will be approximately equal to
the power supply (i.e., +5v). Since in this situation both terminals of the resistor are at +5v, there is no
voltage difference present. That means when plugging into Ohm’s Law, the voltage component is 0v,
which gives 0 amps of current. In the case where the driver is outputting a logic LOW, the output will be
approximately GND. In this case, there is a voltage drop of +5v across the resistor (5v-0v). Plugging this
into Ohm’s Law yields a current of 50 mA flowing through the resistor. This can become problematic
because the current flows through the resistor and then into the output of the driver. For the 74HC logic
family, this would exceed the IO max specification of 25 mA and damage the part. Additionally, as more
current is drawn through the output, the output voltage becomes less and less ideal. In this example, the
first order analysis uses VO ¼ GND. In reality, as the output current increases, the output voltage will
move further away from its ideal value and may eventually reach a value within the uncertainty region.

Example 3.5
Determining the output current when driving a pull-up resistor as the load
 3.4 Driving Loads • 75

A similar process can be used to determine the output current when driving a resistive load between
the output and GND. This process is shown in Example 3.6.

Example 3.6
Determining the output current when driving a pull-down resistor as the load

3.4.3 Driving LEDs

A light emitting diode (LED) is a very common type of load that is driven using a digital circuit. The
behavior of diodes is typically covered in an analog electronics class. Since it is assumed that the reader
has not been exposed to the operation of diodes, the behavior of the LED will be described using a highly
simplified model. A diode has two terminals, the anode and cathode. Current that flows from the anode to
the cathode is called the forward current. A voltage that is developed across a diode from its anode to
cathode is called the forward voltage. A diode has a unique characteristic that when a forward voltage is
supplied across its terminal, it will only increase up to a certain point. The amount is specified as the
LED’s forward voltage (vf) and is typically between 1.5v and 2v in modern LEDs. When a power supply
circuit is connected to the LED, no current will flow until this forward voltage has been reached. Once it
has been reached, current will begin to flow and the LED will prevent any further voltage from developing
across it. Once current flows, the LED will begin emitting light. The more current that flows, the more light
that will be emitted up until the point that the maximum allowable current through the LED is reached and
then the device will be damaged. When using an LED, there are two specifications of interest: the
forward voltage and the recommended forward current. The symbols for a diode and an LED are given in
Fig. 3.35.
76 • Chapter 3: Digital Circuitry and Interfacing

Fig. 3.35
Symbols for a diode and a light emitting diode

When designing an LED driver circuit, a voltage must be supplied in order to develop the forward
voltage across the LED so that current will flow. A resistor is included in series with the LED for two
reasons. The first reason is to provide a place for any additional voltage provided by the driver to develop
in the situation that Vo > Vf, which is most often the case. The second reason for the resistor is to set the
output current. Since the voltage across the resistor will be a fixed amount (i.e., Vo-Vf), then the value of
the resistor can be chosen to set the current. This current is typically set to an optimum value that turns
on the LED to a desired luminosity while also ensuring that the maximum output current of the driver is
not violated. Consider the LED driver configuration shown in Example 3.7 where the LED will be turned
on when the driver outputs a HIGH.

Example 3.7
Determining the output current when driving an LED where HIGH ¼ ON
 3.4 Driving Loads • 77

Example 3.8 shows another example of driving an LED, but this time using a different configuration
where the LED will be on when the driver outputs a logic LOW.

Example 3.8
Determining the Output Current When Driving an LED where HIGH ¼ OFF

CONCEPT CHECK

CC3.4 A fan-out specification is typically around 6–12. If a logic family has a maximum output current
specification of IO-max ¼ 25 mA and a maximum input current specification of only II-max ¼ 1 uA,
a driver could conceivably source up to 25,000 gates (IO-max/II-max ¼ 25 mA/1 uA ¼ 25,000)
without violating its maximum output current specification. Why isn’t the fan-out specification
then closer to 25,000?
(A) The fan-out specification has significant margin built into it in order to protect the
driver.
(B) Connecting 25,000 loads to the driver would cause significant wiring congestion
and would be impractical.
(C) The fan-out specification is in place to reduce power, so keeping it small is
desirable.
(D) The fan-out specification is in place for AC behavior. It ensures that the AC loading
on the driver doesn’t slow down its output rise and fall times. If too many loads are
connected, the output transition will be too slow and it will reside in the uncertainty
region for too long leading to unwanted switching on the receivers.
78 • Chapter 3: Digital Circuitry and Interfacing

Summary
v The operation of a logic circuit can be exceeded, the circuit may not operate prop-
described using either a logic symbol, a truth erly or be damaged.
table, a logic expression, or a logic waveform. v The current exiting a logic circuit is equal to
v Logic gates represent the most basic the current entering.
operations that can be performed on binary v When a logic circuit sources current to a load,
numbers. They are BUF, INV, AND, NAND, an equivalent current is drawn into the circuit
OR, NOR, XOR, and XNOR. through its power supply pin.
v XOR gates that have a number of inputs v When a logic circuit sinks current from a load,
greater than two are created using a cascade an equivalent current flows out of the circuit
of 2-input XOR gates. This implementation through its ground pin.
has more practical applications such as arith- v The type of load that is connected to the
metic and error detection codes. output of a logic circuit dictates how much
v The logic level describes whether the electri- current will be drawn from the driver.
cal signal representing one of two states is v The quiescent current (Iq or Icc) is the current
above or below a switching threshold region. that the circuit always draws independent of
The two possible values that a logic level can the input/output currents.
be are HIGH or LOW. v Logic circuits have AC specifications that
v The logic value describes how the logic describe the delay from the input to the out-
levels are mapped into the two binary codes put (tPLH, tPHL) and also how fast the outputs
0 and 1. In positive logic a HIGH ¼ 1 and a transition between the HIGH and LOW levels
LOW ¼ 0. In negative logic a HIGH ¼ 0 and a (tr, tf).
LOW ¼ 1. v A logic family is a set of logic circuits that are
v Logic circuits have DC specifications that designed to operate with each other.
describe how input voltage levels are v The fan-in of a logic family describes the
interpreted as either HIGHs or LOWs maximum number of inputs that a gate
(VIH-max, VIH-min, VIL-max, and VIL-min). may have.
Specifications are also given on what output v The fan-out of a logic family describes the
voltages will be produced when driving a maximum number of other gates from within
HIGH or LOW (VOH-max, VOH-min, VOL-max, the same family that can be driven simulta-
and VOL-min). neously by one gate.
v In order to successfully transmit digital infor- v Complementary Metal Oxide Semiconductor
mation, the output voltages of the driver that (CMOS) logic is the most popular family
represent a HIGH and LOW must arrive at series in use today. CMOS logic use two
the receiver within the voltage ranges that transistors (NMOS and PMOS) that act as
are interpreted as a HIGH and LOW. If the complementary switches. CMOS transistors
voltage arrives at the receiver outside of draw very low quiescent current and can be
these specified input ranges, the receiver fabricated with extremely small feature sizes.
will not know whether a HIGH or LOW is v In CMOS, only inverters, NAND gates, and
being transmitted. NOR gates can be created directly. If it is
v Logic circuits also specify maximum current desired to create a buffer, AND gate, or OR
levels on the power supplies (IVCC, Ignd), gate, an inverter is placed on the output of
inputs (II-max), and outputs (IO-max) that may the original inverter, NAND, or NOR gate.
not be exceeded. If these levels are

Exercise Problems
Section 3.1: Basic Gates 3.1.4 Give the logic expression for a 3-input AND
gate with the input variables A, B, C and
3.1.1 Give the truth table for a 3-input AND gate with
output F.
the input variables A, B, C and output F.
3.1.5 Give the logic expression for a 3-input OR gate
3.1.2 Give the truth table for a 3-input OR gate with
with the input variables A, B, C and output F.
the input variables A, B, C and output F.
3.1.6 Give the logic expression for a 3-input XNOR
3.1.3 Give the truth table for a 3-input XNOR gate
gate with the input variables A, B, C and
with the input variables A, B, C and output F.
output F.
 Exercise Problems • 79

3.1.7 Give the logic waveform for a 3-input AND gate

with the input variables A, B, C and output F.
3.1.8 Give the logic waveform for a 3-input OR gate
with the input variables A, B, C and output F.
3.1.9 Give the logic waveform for a 3-input XNOR
gate with the input variables A, B, C and
output F.

Section 3.2: Digital Circuit Operation

3.2.1 Using the DC operating conditions from
Table 3.2, give the noise margin HIGH (NMH)
for the 74LS logic family.
3.2.2 Using the DC operating conditions from
Table 3.2, give the noise margin LOW (NML)
for the 74LS logic family.
Fig. 3.37
Driver configuration 2
3.2.3 Using the DC operating conditions from
Table 3.2, give the noise margin HIGH (NMH)
for the 74HC logic family with VCC ¼ +5v. 3.2.10 For the driver configuration in Fig. 3.37, give
the current flowing through the GND pin.
3.2.4 Using the DC operating conditions from
Table 3.2, give the noise margin LOW (NML) 3.2.11 Using the data sheet excerpt from Fig. 3.20,
for the 74HC logic family with VCC ¼ +5v. give the maximum propagation delay (tpd) for
the 74HC04 inverter when powered with
3.2.5 Using the DC operating conditions from VCC ¼ +2v.
Table 3.2, give the noise margin HIGH (NMH)
for the 74HC logic family with VCC ¼ +3.4v. 3.2.12 Using the data sheet excerpt from Fig. 3.20,
give the maximum propagation delay from low
3.2.6 Using the DC operating conditions from to high (tPLH) for the 74HC04 inverter when
Table 3.2, give the noise margin LOW (NML) powered with VCC ¼ +2v.
for the 74HC logic family with VCC ¼ +3.4v.
3.2.13 Using the data sheet excerpt from Fig. 3.20,
3.2.7 For the driver configuration in Fig. 3.36, give give the maximum propagation delay from high
the current flowing through the VCC pin. to low (tPHL) for the 74HC04 inverter when
powered with VCC ¼ +2v.
3.2.14 Using the data sheet excerpt from Fig. 3.20,
give the maximum transition time (tt) for the
74HC04 inverter when powered with VCC ¼ +2v.
3.2.15 Using the data sheet excerpt from Fig. 3.20,
give the maximum rise time (tr) for the 74HC04
inverter when powered with VCC ¼ +2v.
3.2.16 Using the data sheet excerpt from Fig. 3.20,
give the maximum fall time (tf) for the 74HC04
inverter when powered with VCC ¼ +2v.

Section 3.3: Logic Families

3.3.1 Provide the transistor-level schematic for a
4-Input NAND gate.
3.3.2 Provide the transistor-level schematic for a
4-Input NOR gate.
3.3.3 Provide the transistor-level schematic for a
2-Input AND gate.
Fig. 3.36 3.3.4 Provide the transistor-level schematic for
Driver configuration 1 a 2-Input OR gate.
3.3.5 Provide the transistor-level schematic for a
3.2.8 For the driver configuration in Fig. 3.36, give buffer.
the current flowing through the GND pin.
3.2.9 For the driver configuration in Fig. 3.37, give Section 3.4: Driving Loads
the current flowing through the VCC pin. 3.4.1 In the driver configuration shown in Fig. 3.38,
the buffer is driving its maximum fan-out speci-
fication of 6. The maximum input current for
this logic family is II ¼ 1 nA. What is the
80 • Chapter 3: Digital Circuitry and Interfacing

maximum output current (IO) that the driver will 3.4.4 For the LED driver configuration shown in
need to source? Fig. 3.41 where an output of HIGH on the driver
will turn on the LED, calculate the value of the
resistor (R) in order to set the LED forward
current to 5 mA. The LED has a forward volt-
age of 1.9v.

Fig. 3.41
Driver configuration 6
Fig. 3.38
Driver configuration 3
3.4.5 For the LED driver configuration shown in
Fig. 3.42 where an output of LOW on the driver
3.4.2 For the pull-down driver configuration shown in will turn on the LED, calculate the value of the
Fig. 3.39, calculate the value of the pull-down resistor (R) in order to set the LED forward
resistor (R) in order to ensure that the output current to 5 mA. The LED has a forward volt-
current does not exceed 20 mA. age of 1.9v.

Fig. 3.39
Driver configuration 4
Fig. 3.42
3.4.3 For the pull-up driver configuration shown in Driver configuration 7
Fig. 3.40, calculate the value of the pull-up
resistor (R) in order to ensure that the output
current does not exceed 20 mA.

Fig. 3.40
Driver configuration 5
Chapter 4: Combinational Logic
Design
In this chapter we cover the techniques to synthesize, analyze, and manipulate logic functions. The
purpose of these techniques is to ultimately create a logic circuit using the basic gates described in
Chap. 3 from a truth table or word description. This process is called combinational logic design.
Combinational logic refers to circuits where the output depends on the present value of the inputs.
This simple definition implies that there is no storage capability in the circuitry and a change on the input
immediately impacts the output. To begin, we first define the rules of Boolean algebra, which provide the
framework for the legal operations and manipulations that can be taken on a two-valued number system
(i.e., a binary system). We then explore a variety of logic design and manipulation techniques. These
techniques allow us to directly create a logic circuit from a truth table and then to manipulate it to either
reduce the number of gates necessary in the circuit or to convert the logic circuit into equivalent forms
using alternate gates. The goal of this chapter is to provide an understanding of the basic principles of
combinational logic design.
Learning Outcomes—After completing this chapter, you will be able to:
4.1 Describe the fundamental principles and theorems of Boolean algebra and how to use
them to manipulate logic expressions.
4.2 Analyze a combinational logic circuit to determine its logic expression, truth table, and
timing information.
4.3 Synthesis a logic circuit in canonical form (Sum of Products or Product of Sums) from a
functional description including a truth table, minterm list, or maxterm list.
4.4 Synthesize a logic circuit in minimized form (Sum of Products or Product of Sums) through
algebraic manipulation or with a Karnaugh map.
4.5 Describe the causes of timing hazards in digital logic circuits and the approaches to
mitigate them.

4.1 Boolean Algebra

The term Algebra refers to the rules of a number system. In Chap. 2 we discussed the number of
symbols and relative values of some of the common number systems. Algebra defines the operations
that are legal to perform on that system. Once we have defined the rules for a system, we can then use
the system for more powerful mathematics such as solving for unknowns and manipulating into equiva-
lent forms. The ability to manipulate into equivalent forms allows us to minimize the number of logic
operations necessary and also put into a form that can be directly synthesized using modern logic
circuits.
In 1854, English mathematician George Boole presented an abstract algebraic framework for a
system that contained only two states, true and false. This framework essentially launched the field of
computer science even before the existence of the modern integrated circuits that are used to implement
digital logic today. In 1930, American mathematician Claude Shannon applied Boole’s algebraic frame-
work to his work on switching circuits at Bell Labs, thus launching the field of digital circuit design and
information theory. Boole’s original framework is still used extensively in modern digital circuit design and
thus bears the name Boolean algebra. Today, the term Boolean algebra is often used to describe not only
George Boole’s original work, but all of those that contributed to the field after him.

# Springer International Publishing AG 2017 81

B.J. LaMeres, Introduction to Logic Circuits & Logic Design with Verilog,
DOI 10.1007/978-3-319-53883-9_4
82 • Chapter 4: Combinational Logic Design

4.1.1 Operations
In Boolean algebra there are two valid states (true and false) and three core operations. The
operations are conjunction (^, equivalent to the AND operation), disjunction (_, equivalent to the OR
operation), and negation (Ø, equivalent to the NOT operation). From these three operations, more
sophisticated operations can be created including other logic functions (i.e., BUF, NAND, NOR, XOR,
XNOR, etc.) and arithmetic. Engineers primarily use the terms AND, OR and NOT instead of conjunction,
disjunction and negation. Similarly, engineers primarily use the symbols for these operators described in
Chap. 3 (e.g., ∙, + and ‘) instead of ^, _, and Ø.

4.1.2 Axioms
An axiom is a statement of truth about a system that is accepted by the user. Axioms are very simple
statements about a system, but need to be established before more complicated theorems can be
proposed. Axioms are so basic that they do not need to be proved in order to be accepted. Axioms can
be thought of as the basic laws of the algebraic framework. The terms axiom and postulate are
synonymous and used interchangeably. In Boolean algebra there are five main axioms. These axioms
will appear redundant with the description of basic gates from Chap. 3, but must be defined in this
algebraic context so that more powerful theorems can be proposed.

4.1.2.1 Axiom #1 – Logical Values

This axiom states that in Boolean algebra, a variable A can only take on one of two values, 0 or 1. If
the variable A is not 0, then it must be a 1, and conversely, if it is not a 1, then it must be a 0.

Axiom #1 – Boolean Values: A ¼ 0 if A 6¼ 1, conversely A ¼ 1 if A 6¼ 0.

4.1.2.2 Axiom #2 – Definition of Logical Negation

This axiom defines logical negation. Negation is also called the NOT operation or taking the
complement. The negation operation is denoted using either a prime (‘), an inversion bar or the negation
symbol (Ø). If the complement is taken on a 0, it becomes a 1. If the complement is taken on a 1, it
becomes a 0.

Axiom #2 – Definition of Logical Negation: if A ¼ 0 then A0 ¼ 1, conversely, if A ¼ 1 then A0 ¼ 0.

4.1.2.3 Axiom #3 – Definition of a Logical Product

This axiom defines a logical product or multiplication. Logical multiplication is denoted using either a
dot (∙), an ampersand (&) or the conjunction symbol (^). The result of logical multiplication is true when
both inputs are true and false otherwise.

Axiom #3 – Definition of a Logical Product: A∙B ¼ 1 if A ¼ B ¼ 1 and A∙B ¼ 0 otherwise.

 4.1 Boolean Algebra • 83

4.1.2.4 Axiom #4 – Definition of a Logical Sum

This axiom defines a logical sum or addition. Logical addition is denoted using either a plus sign (+)
or the disjunction symbol (_). The result of logical addition is true when any of the inputs are true and
false otherwise.

Axiom #4 – Definition of a Logical Sum: A + B ¼ 1 if A ¼ 1 or B ¼ 1 and A + B ¼ 0 otherwise.

4.1.2.5 Axiom #5 – Logical Precedence

This axiom defines the order of precedence for the three operators. Unless the precedence is
explicitly stated using parentheses, negation takes precedence over a logical product and a logical
product takes precedence over a logical sum.

Axiom #5 – Definition of Logical Precedence: NOT precedes AND, AND precedes OR.

To illustrate Axiom #5, consider the logic function F ¼ A0 ∙B + C. In this function, the first operation that
would take place is the NOT operation on A. This would be followed by the AND operation of A0 with
B. Finally, the result would be OR’d with C. The precedence of any function can also be explicitly stated
using parentheses such as F ¼ (((A0 ) ∙ B) + C).

4.1.3 Theorems
A theorem is a more sophisticated truth about a system that is not intuitively obvious. Theorems are
proposed and then must be proved. Once proved, they can be accepted as a truth about the system
going forward. Proving a theorem in Boolean algebra is much simpler than in our traditional decimal
system due to the fact that variables can only take on one of two values, true or false. Since the number
of input possibilities is bounded, Boolean algebra theorems can be proved by simply testing the theorem
using every possible input code. This is called proof by exhaustion. The following theorems are used
widely in the manipulation of logic expressions and reduction of terms within an expression.

4.1.3.1 DeMorgan’s Theorem of Duality

Augustus DeMorgan was a British mathematician and logician who lived during the time of George
Boole. DeMorgan is best known for his contribution to the field of logic through the creation of what have
been later called the DeMorgan’s Theorems (often called DeMorgan’s Laws). There are two major
theorems that DeMorgan proposed that expanded Boolean algebra. The first theorem is named duality.
Duality states that an algebraic equality will remain true if all 0’s and 1’s are interchanged and all AND
and OR operations are interchanged. The new expression is called the dual of the original expression.
Example 4.1 shows the process of proving duality using proof by exhaustion.
84 • Chapter 4: Combinational Logic Design

Example 4.1
Proving DeMorgan’s theorem of duality using proof by exhaustion

Duality is important for two reasons. First, it doubles the impact of a theorem. If a theorem is proved
to be true, then the dual of that theorem is also proved to be true. This, in essence, gives twice the
theorem with the same amount of proving. Boolean algebra theorems are almost always given in pairs,
the original and the dual. That is why duality is covered as the first theorem.
The second reason that duality is important is because it can be used to convert between positive
and negative logic. Until now, we have used positive logic for all of our examples (i.e., a logic HIGH ¼
true ¼1 and a logic LOW ¼ false ¼0). As mentioned earlier, this convention is arbitrary and we could
have easily chosen a HIGH to be false and a LOW to be true (i.e., negative logic). Duality allows us to
take a logic expression that has been created using positive logic (F) and then convert it into an
equivalent expression that is valid for negative logic (FD). Example 4.2 shows the process for how this
works.
 4.1 Boolean Algebra • 85

Example 4.2
Converting between positive and negative logic using duality

One consideration when using duality is that the order of precedence follows the original function.
This means that in the original function, the axiom for precedence states the order as NOT-AND-OR;
however, this is not necessarily the correct precedence order in the dual. For example, if the original
function was F ¼ AB + C, the AND operation of A and B would take place first, and then the result would
be OR’d with C. The dual of this expression is FD ¼ A + BC. If the expression for FD was evaluated using
traditional Boolean precedence, it would show that FD does NOT give the correct result per the definition
of a dual function (i.e., converting a function from positive to negative logic). The order of precedence for
FD must correlate to the precedence in the original function. Since in the original function A and B were
operated on first, they must also be operated on first in the dual. In order to easily manage this issue,
parentheses can be used to track the order of operations from the original function to the dual. If we put
parentheses in the original function to explicitly state the precedence of the operations, it would take the
form F ¼ (AB) + C. These parentheses can be mapped directly to the dual yielding FD ¼ (A + B)C. This
order of precedence in the dual is now correct.
86 • Chapter 4: Combinational Logic Design

Now that we have covered the duality operation, its usefulness and its pitfalls, we can formally
define this theorem as:
DeMorgan’s Duality: An algebraic equality will remain true if all 0’s and 1’s are interchanged and all
AND and OR operations are interchanged. Furthermore, taking the dual of a positive logic function will
produce the equivalent function using negative logic if the original order of precedence is maintained.

4.1.3.2 Identity
An identity operation is one that when performed on a variable will yield itself regardless of the
variable’s value. The following is the formal definition of identity theorem. Figure 4.1 shows the gate level
depiction of this theorem.
Identity: OR’ing any variable with a logic 0 will yield the original variable. The dual: AND’ing any variable
with a logic 1 will yield the original variable.

Fig. 4.1
Gate level depiction of the identity theorem

The identity theorem is useful for reducing circuitry when it is discovered that a particular input will
never change values. When this is the case, the static input variable can simply be removed from the
logic expression making the entire circuit a simple wire from the remaining input variable to the output.

4.1.3.3 Null Element

A null element operation is one that, when performed on a constant value, will yield that same
constant value regardless of the values of any variables within the same operation. The following is the
formal definition of null element. Figure 4.2 shows the gate level depiction of this theorem.
Null Element: OR’ing any variable with a logic 1 will yield a logic 1 regardless of the value of the input
variable. The dual: AND’ing any variable with a logic 0 will yield a logic 0 regardless of the value of the
input variable.

Fig. 4.2
Gate level depiction of the null element theorem
 4.1 Boolean Algebra • 87

The null element theorem is also useful for reducing circuitry when it is discovered that a particular
input will never change values. It is also widely used in computer systems in order to set (i.e., force to a
logic 1) or clear (i.e., force to a logic 0) the value of a storage element.

4.1.3.4 Idempotent
An idempotent operation is one that has no effect on the input, regardless of the number of times the
operation is applied. The following is the formal definition of idempotence. Figure 4.3 shows the gate
level depiction of this theorem.

Idempotent: OR’ing a variable with itself results in itself. The dual: AND’ing a variable with itself results in
itself.

Fig. 4.3
Gate level depiction of the idempotent theorem

This theorem also holds true for any number of operations such as A + A + A + . . ... + A ¼ A and
AAA. . ...A ¼ A.

4.1.3.5 Complements
This theorem describes an operation of a variable with the variable’s own complement. The
following is the formal definition of complements. Figure 4.4 shows the gate level depiction of this
theorem.
Complements: OR’ing a variable with its complement will produce a logic 1. The dual: AND’ing a
variable with its complement will produce a logic 0.

Fig. 4.4
Gate level depiction of the complements theorem

The complement theorem is again useful for reducing circuitry when these types of logic
expressions are discovered.
88 • Chapter 4: Combinational Logic Design

4.1.3.6 Involution
An involution operation describes the result of double negation. The following is the formal definition
of involution. Figure 4.5 shows the gate level depiction of this theorem.
Involution: Taking the double complement of a variable will result in the original variable.

Fig. 4.5
Gate level depiction of the involution theorem

This theorem is not only used to eliminate inverters but also provides us a powerful tool for inserting
inverters in a circuit. We will see that this is used widely with the second of DeMorgan’s Laws that will be
introduced at the end of this section.

4.1.3.7 Commutative Property

The term commutative is used to describe an operation in which the order of the quantities or
variables in the operation have no impact on the result. The following is the formal definition of the
commutative property. Figure 4.6 shows the gate level depiction of this theorem.
Commutative Property: Changing the order of variables in an OR operation does not change the end
result. The dual: Changing the order of variables in an AND operation does not change the end result.

Fig. 4.6
Gate level depiction of commutative property

One practical use of the commutative property is when wiring or routing logic circuitry together.
Example 4.3 shows how the commutative property can be used to untangle crossed wires when
implementing a digital system.
 4.1 Boolean Algebra • 89

Example 4.3
Using the commutative property to untangle crossed wires

4.1.3.8 Associative Property

The term associative is used to describe an operation in which the grouping of the quantities or
variables in the operation have no impact on the result. The following is the formal definition of the
associative property. Figure 4.7 shows the gate level depiction of this theorem.
90 • Chapter 4: Combinational Logic Design

Associative Property: The grouping of variables doesn’t impact the result of an OR operation. The dual:
The grouping of variables doesn’t impact the result of an AND operation.

Fig. 4.7
Gate level depiction of the associative property

One practical use of the associative property is addressing fan-in limitations of a logic family. Since
the grouping of the input variables does not impact the result, we can accomplish operations with large
numbers of inputs using multiple gates with fewer inputs. Example 4.4 shows the process of using the
associative property to address a fan-in limitation.

Example 4.4
Using the associative property to address fan-in limitations
 4.1 Boolean Algebra • 91

4.1.3.9 Distributive Property

The term distributive describes how an operation on a parenthesized group of operations (or higher
precedence operations) can be distributed through each term. The following is the formal definition of the
distributive property. Figure 4.8 shows the gate level depiction of this theorem.
Distributive Property: An operation on a parenthesized operation(s), or higher precedence operator,
will distribute through each term.

Fig. 4.8
Gate level depiction of the distributive property

The distributive property is used as a logic manipulation technique. It can be used to put a logic
expression into a form more suitable for direct circuit synthesis, or to reduce the number of logic gates
necessary. Example 4.5 shows how to use the distributive property to reduce the number of gates in a
logic circuit.

Example 4.5
Using the distributive property to reduce the number of logic gates in a circuit
92 • Chapter 4: Combinational Logic Design

4.1.3.10 Absorption
The term absorption refers to when multiple logic terms within an expression produce the same
results. This allows one of the terms to be eliminated from the expression, thus reducing the number of
logic operations. The remaining terms essentially absorb the functionality of the eliminated term. This
theorem is also called covering because the remaining term essentially covers the functionality of both
itself and the eliminated term. The following is the formal definition of the absorption theorem. Figure 4.9
shows the gate level depiction of this theorem.
Absorption: When a term within a logic expression produces the same output(s) as another term, the
second term can be removed without affecting the result.

Fig. 4.9
Gate level depiction of absorption

This theorem is better understood by looking at the evaluation of each term with respect to the
original expression. Example 4.6 shows how the absorption theorem can be proven through proof by
exhaustion by evaluating each term in a logic expression.

Example 4.6
Proving the absorption theorem using proof by exhaustion

4.1.3.11 Uniting
The uniting theorem, also called combining or minimization, provides a way to remove variables
from an expression when they have no impact on the outcome. This theorem is one of the most widely
 4.1 Boolean Algebra • 93

used techniques for the reduction of the number of gates needed in a combinational logic circuit. The
following is the formal definition of the uniting theorem. Figure 4.10 shows the gate level depiction of this
theorem.
Uniting: When a variable (B) and its complement (B0 ) appear in multiple product terms with a common
variable (A) within a logical OR operation, the variable B does not have any effect on the result and can
be removed.

Fig. 4.10
Gate level depiction of uniting

This theorem can be proved using prior theorems. Example 4.7 shows how the uniting theorem can
be proved using a combination of the distributive property, the complements theorem, and the identity
theorem.

Example 4.7
Proving of the uniting theorem

4.1.3.12 DeMorgan’s Theorem

Now we look at the second of DeMorgan’s Laws. This second theorem is simply known as
DeMorgan’s Theorem. This theorem provides a technique to manipulate a logic expression that uses
AND gates into one that uses OR gates and vice-versa. It can also be used to manipulate traditional
Boolean logic expressions that use AND-OR-NOT operators, into equivalent forms that uses NAND and
NOR gates. The following is the formal definition of DeMorgan’s theorem. Figure 4.11 shows the gate
level depiction of this theorem.
94 • Chapter 4: Combinational Logic Design

DeMorgan’s Theorem: An OR operation with both inputs inverted is equivalent to an AND operation
with the output inverted. The dual: An AND operation with both inputs inverted is equivalent to an OR
operation with the output inverted.

Fig. 4.11
Gate level depiction of DeMorgan’s theorem

This theorem is used widely in modern logic design because it bridges the gap between the design
of logic circuitry using Boolean algebra and the physical implementation of the circuitry using CMOS.
Recall that Boolean algebra is defined for only three operations, the AND, the OR and inversion. CMOS,
on the other hand, can only directly implement negative-type gates such as NAND, NOR and NOT.
DeMorgan’s Theorem allows us to design logic circuitry using Boolean algebra and synthesize logic
diagrams with AND, OR, and NOT gates, and then directly convert the logic diagrams into an equivalent
form using NAND, NOR and NOT gates. As we’ll see in the next section, Boolean algebra produces logic
expressions in two common forms. These are the sum of products (SOP) and the product of sums
(POS) forms. Using a combination of involution and DeMorgan’s Theorem, SOP and POS forms can be
converted into equivalent logic circuits that use only NAND and NOR gates. Example 4.8 shows a
process to convert a sum of products form into one that uses only NAND gates.
 4.1 Boolean Algebra • 95

Example 4.8
Converting a sum of products form into one that uses only NAND gates
96 • Chapter 4: Combinational Logic Design

Example 4.9 shows a process to convert a product of sums form into one that uses only NOR gates.

Example 4.9
Converting a product of sums form into one that uses only NOR gates

DeMorgan’s Theorem can also be accomplished algebraically using a process known as breaking
the bar and flipping the operator. This process again takes advantage of the Involution Theorem, which
allows double negation without impacting the result. When using this technique in algebraic form,
involution takes the form of a double inversion bar. If an inversion bar is broken, the expression will
remain true as long as the operator directly below the break is flipped (AND to OR, OR to AND).
Example 4.10 shows how to use this technique when converting an OR gate with its inputs inverted
into an AND gate with its output inverted.
 4.1 Boolean Algebra • 97

Example 4.10
Using DeMorgan’s theorem in algebraic form (1)

Example 4.11 shows how to use this technique when converting an AND gate with its inputs inverted
into an OR gate with its output inverted.

Example 4.11
Using DeMorgan’s theorem in algebraic form (2)
98 • Chapter 4: Combinational Logic Design

Table 4.1 gives a summary of all the Boolean algebra theorems just covered. The theorems are
grouped in this table with respect to the number of variables that they contain. This grouping is the most
common way these theorems are presented.

Table 4.1
Summary of Boolean algebra theorems

4.1.4 Functionally Complete Operation Sets

A set of Boolean operators is said to be functionally complete when the set can implement all
possible logic functions. The set of operators {AND, OR, NOT} is functionally complete because every
other operation can be implemented using these three operators (i.e., NAND, NOR, BUF, XOR, XNOR).
The DeMorgan’s Theorem showed us that all AND and OR operations can be replaced with NAND and
NOR operators. This means that NAND and NOR operations could be by themselves functionally
complete if they could perform a NOT operation. Figure 4.12 shows how a NAND gate can be configured
to perform a NOT operation. This configuration allows a NAND gate to be considered functionally
complete because all other operations can be implemented.

Fig. 4.12
Configuration to use a NAND gate as an inverter
 4.2 Combinational Logic Analysis • 99

This approach can also be used on a NOR gate to implement an inverter. Figure 4.13 shows how a
NOR gate can be configured to perform a NOT operation, thus also making it functionally complete.

Fig. 4.13
Configuration to use a NOR gate as an inverter

CONCEPT CHECK

CC4.1 If the logic expression F ¼ ABCDEFGH is implemented with only 2-input AND gates, how
many levels of logic will the final implementation have? Hint: Consider using the associative
property to manipulate the logic expression to use only 2-input AND operations.

(A) 2 (B) 3 (C) 4 (D) 5

4.2 Combinational Logic Analysis

Combinational logic analysis refers to the act of deciphering the operation of a circuit from its final
logic diagram. This is a useful skill that can aid designers when debugging their circuits. This can also be
used to understand the timing performance of a circuit and to reverse-engineer an unknown design.

4.2.1 Finding the Logic Expression from a Logic Diagram

Combinational logic diagrams are typically written with their inputs on the left and their output on the
right. As the inputs change, the intermediate nodes, or connections, within the diagram hold the interim
computations that contribute to the ultimate circuit output. These computations propagate from left to
right until ultimately the final output of the system reaches its final steady state value. When analyzing the
behavior of a combinational logic circuit a similar left-to-right approach is used. The first step is to label
each intermediate node in the system. The second step is to write in the logic expression for each node
100 • Chapter 4: Combinational Logic Design

based on the preceding logic operation(s). The logic expressions are written working left-to-right until the
output of the system is reached and the final logic expression of the circuit has been found. Consider the
example of this analysis in Example 4.12.

Example 4.12
Determining the logic expression from a logic diagram

4.2.2 Finding the Truth Table from a Logic Diagram

The final truth table of a circuit can also be found in a similar manner as the logic expression. Each
internal node within the logic diagram can be evaluated working from the left to the right for each possible
input code. Each subsequent node can then be evaluated using the values of the preceding nodes.
Consider the example of this analysis is Example 4.13.
 4.2 Combinational Logic Analysis • 101

Example 4.13
Determining the truth table from a logic diagram

4.2.3 Timing Analysis of a Combinational Logic Circuit

Real logic gates have a propagation delay (tpd, tPHL, or tPLH) as presented in Chap. 3. Performing a
timing analysis on a combinational logic circuit refers to observing how long it takes for a change in the
inputs to propagate to the output. Different paths through the combinational logic circuit will take different
times to compute since they may use gates with different delays. When determining the delay of the
entire combinational logic circuit we always consider the longest delay path. This is because this delay
represents the worst case scenario. As long as we wait for the longest path to propagate through the
circuit, then we are ensured that the output will always be valid after this time. To determine which signal
path has the longest delay, we map out each and every path the inputs can take to the output of the
circuit. We then sum up the gate delay along each path. The path with the longest delay dictates the
delay of the entire combinational logic circuit. Consider this analysis shown in Example 4.14.
102 • Chapter 4: Combinational Logic Design

Example 4.14
Determining the delay of a combinational logic circuit

CONCEPT CHECK

CC4.2 Does the delay specification of a combinational logic circuit change based on the input values
that the circuit is evaluating?
(A) Yes. There are times when the inputs switch between inputs codes that use paths
through the circuit with different delays.
(B) No. The delay is always specified as the longest delay path.
(C) Yes. The delay can vary between the longest delay path and zero. A delay of zero
occurs when the inputs switch between two inputs codes that produce the same
output.
(D) No. The output is always produced at a time equal to the longest delay path.
 4.3 Combinational Logic Synthesis • 103

4.3 Combinational Logic Synthesis

4.3.1 Canonical Sum of Products
One technique to directly synthesize a logic circuit from a truth table is to use a canonical sum of
products topology based on minterms. The term canonical refers to this topology yielding potentially
unminimized logic. A minterm is a product term (i.e., an AND operation) that will be true for one and only
one input code. The minterm must contain every input variable in its expression. Complements are
applied to the input variables as necessary in order to produce a true output for the individual input code.
We define the word literal to describe an input variable which may or may not be complemented. This is a
more useful word because if we say that a minterm “must include all variables”, it implies that all variables
are included in the term uncomplemented. A more useful statement is that a minterm “must include all
literals”. This now implies that each variable must be included, but it can be in the form of itself or its
complement (e.g., A or A0 ). Figure 4.14 shows the definition and gate level depiction of a minterm
expression. Each minterm can be denoted using the lower case “m” with the row number as a subscript.

Fig. 4.14
Definition and gate level depiction of a minterm

For an arbitrary truth table, a minterm can be used for each row corresponding to a true output. If
each of these minterms’ outputs are fed into a single OR gate, then a sum of products logic circuit is
formed that will produce the logic listed in the truth table. In this topology, any input code that corresponds
to an output of 1 will cause its corresponding minterm to output a 1. Since a 1 on any input of an OR gate
will cause the output to go to a 1, the output of the minterm is passed to the final result. Example 4.15
shows this process. One important consideration of this approach is that no effort has been taken to
minimize the logic expression. This unminimized logic expression is also called the canonical sum. The
canonical sum is logically correct but uses the most amount of circuitry possible for a given truth table.
This canonical sum can be the starting point for minimization using Boolean algebra.
104 • Chapter 4: Combinational Logic Design

Example 4.15
Creating a canonical sum of products logic circuit using minterms

4.3.2 The Minterm List (Σ)

A minterm list is a compact way to describe the functionality of a logic circuit by simply listing the row
numbers that correspond to an output of 1 in the truth table. The ∑ symbol is used to denote a minterm
list. All input variables must be listed in the order they appear in the truth table. This is necessary because
since a minterm list uses only the row numbers to indicate which input codes result in an output of 1, the
minterm list must indicate how many variables comprise the row number, which variable is in the most
significant position and which is in the least significant position. After the ∑ symbol, the row numbers
corresponding to a true output are listed in a comma-delimited format within parentheses. Example 4.16
shows the process for creating a minterm list from a truth table.
 4.3 Combinational Logic Synthesis • 105

Example 4.16
Creating a minterm list from a truth table

A minterm list contains the same information as the truth table, the canonical sum and the canonical
sum of products logic diagram. Since the minterms themselves are formally defined for an input code, it
is trivial to go back and forth between the minterm list and these other forms. Example 4.17 shows how a
minterm list can be used to generate an equivalent truth table, canonical sum and canonical sum of
products logic diagram.
106 • Chapter 4: Combinational Logic Design

Example 4.17
Creating equivalent functional representations from a minterm list

4.3.3 Canonical Product of Sums (POS)

Another technique to directly synthesize a logic circuit from a truth table is to use a canonical product
of sums topology based on maxterms. A maxterm is a sum term (i.e., an OR operation) that will be false
for one and only one input code. The maxterm must contain every literal in its expression. Complements
are applied to the input variables as necessary in order to produce a false output for the individual input
code. Figure 4.15 shows the definition and gate level depiction of a maxterm expression. Each maxterm
can be denoted using the upper case “M” with the row number as a subscript.
 4.3 Combinational Logic Synthesis • 107

Fig. 4.15
Definition and gate level depiction of a maxterm

For an arbitrary truth table, a maxterm can be used for each row corresponding to a false output. If
each of these maxterms outputs are fed into a single AND gate, then a product of sums logic circuit is
formed that will produce the logic listed in the truth table. In this topology, any input code that corresponds
to an output of 0 will cause its corresponding maxterm to output a 0. Since a 0 on any input of an AND
gate will cause the output to go to a 0, the output of the maxterm is passed to the final result. Example
4.18 shows this process. This approach is complementary to the sum of products approach. In the sum
of products approach based on minterms, the circuit operates by producing 1’s that are passed to the
output for the rows that require a true output. For all other rows, the output is false. A product of sums
approach based on maxterms operates by producing 0’s that are passed to the output for the rows that
require a false output. For all other rows, the output is true. These two approaches produce the
equivalent logic functionality. Again, at this point no effort has been taken to minimize the logic expres-
sion. This unminimized form is called a canonical product. The canonical product is logically correct,
but uses the most amount of circuitry possible for a given truth table. This canonical product can be the
starting point for minimization using the Boolean algebra theorems.
108 • Chapter 4: Combinational Logic Design

Example 4.18
Creating a product of sums logic circuit using maxterms

4.3.4 The Maxterm List (Π)

A maxterm list is a compact way to describe the functionality of a logic circuit by simply listing the row
numbers that correspond to an output of 0 in the truth table. The Π symbol is used to denote a maxterm
list. All literals used in the logic expression must be listed in the order they appear in the truth table. After
the Π symbol, the row numbers corresponding to a false output are listed in a comma-delimited format
within parentheses. Example 4.19 shows the process for creating a maxterm list from a truth table.
 4.3 Combinational Logic Synthesis • 109

Example 4.19
Creating a maxterm list from a truth table

A maxterm list contains the same information as the truth table, the canonical product and the
canonical product of sums logic diagram. Example 4.20 shows how a maxterm list can be used to
generate these equivalent forms.
110 • Chapter 4: Combinational Logic Design

Example 4.20
Creating equivalent functional representations from a maxterm list

4.3.5 Minterm and Maxterm List Equivalence

The examples in Examples 4.17 and 4.20 illustrate how minterm and maxterm lists produce the
exact same logic functionality but in a complementary fashion. It is trivial to switch back and forth
between minterm lists and maxterm lists. This is accomplished by simply changing the list type (i.e.,
 4.3 Combinational Logic Synthesis • 111

min to max, max to min) and then switching the row numbers between those listed and those not
listed. Example 4.21 shows multiple techniques for representing equivalent logic functionality as a truth
table.

Example 4.21
Creating equivalent forms to represent logic functionality
112 • Chapter 4: Combinational Logic Design

CONCEPT CHECK

CC4.3 All logic functions can be implemented equivalently using either a canonical sum of products
(SOP) or canonical product of sums (POS) topology. Which of these statements is true with
respect to selecting a topology that requires the least amount of gates.
(A) Since a minterm list and a maxterm list can both be written to describe the same
logic functionality, the number of gates in an SOP and POS will always be
the same.
(B) If a minterm list has over half of its row numbers listed, an SOP topology will
require fewer gates than a POS.
(C) A POS topology always requires more gates because it needs additional logic to
convert the inputs from positive to negative logic.
(D) If a minterm list has over half of its row numbers listed, a POS topology will
require fewer gates than SOP.

4.4 Logic Minimization

We now look at how to reduce the canonical expressions into equivalent forms that use less logic.
This minimization is key to reducing the complexity of the logic prior to implementing in real circuitry.
This reduces the amount of gates needed, placement area, wiring and power consumption of the logic
circuit.

4.4.1 Algebraic Minimization

Canonical expressions can be reduced algebraically by applying the theorems covered in prior
sections. This process typically consists of a series of factoring based on the distributive property
followed by replacing variables with constants (i.e., 0’s and 1’s) using the Complements Theorem.
Finally, constants are removed using the Identity Theorem. Example 4.22 shows this process.
 4.4 Logic Minimization • 113

Example 4.22
Minimizing a logic expression algebraically

The primary drawback of this approach is that it requires recognition of where the theorems can be
applied. This can often lead to missed minimizations. Computer automation is often the best mechanism
to perform this minimization for large logic expressions.

4.4.2 Minimization Using Karnaugh Maps

A Karnaugh map is a graphical way to minimize logic expressions. This technique is named after
Maurice Karnaugh, American physicist, who introduced the map in its latest form in 1953 while working at
Bell Labs. The Karnaugh map (or K-map) is a way to put a truth table into a form that allows logic
minimization through a graphical process. This technique provides a graphical process that
accomplishes the same result as factoring variables via the distributive property and removing variables
via the Complements and Identity Theorems. K-maps present a truth table in a form that allows variables
to be removed from the final logic expression in a graphical manner.
114 • Chapter 4: Combinational Logic Design

4.4.2.1 Formation of a K-Map

A K-map is constructed as a two-dimensional grid. Each cell within the map corresponds to the
output for a specific input code. The cells are positioned such that neighboring cells only differ by one bit
in their input codes. Neighboring cells are defined as cells immediately adjacent horizontally and
immediately adjacent vertically. Two cells positioned diagonally next to each other are not considered
neighbors. The input codes for each variable are listed along the top and side of the K-map. Consider the
construction of a 2-input K-map shown in Fig. 4.16.

Fig. 4.16
Formation of a 2-input K-map

When constructing a 3-input K-map, it is important to remember that each input code can only differ
from its neighbor by one bit. For example, the two codes 01 and 10 differ by two bits (i.e., the MSB is
different and the LSB is different), thus they could not be neighbors; however, the codes 01-11 and 11-10
can be neighbors. As such, the input codes along the top of the 3-input K-map must be ordered
accordingly (i.e., 00-01-11-10). Consider the construction of a 3-input K-map shown in Fig. 4.17. The
rows and columns that correspond to the input literals can now span multiple rows and columns. Notice
how in this 3-input K-map, the literals A, A0 , B and B0 all correspond to two columns. Also, notice that B0
spans two columns, but the columns are on different edges of the K-map. The side edges of the 3-input
K-map are still considered neighbors because the input codes for these columns only differ by one bit.
This is an important attribute once we get to the minimization of variables because it allows us to
examine an input literal’s impact not only within the obvious adjacent cells but also when the variables
wrap around the edges of the K-map.
 4.4 Logic Minimization • 115

Fig. 4.17
Formation of a 3-input K-map

When constructing a 4-input K-map, the same rules apply that the input codes can only differ from
their neighbors by one bit. Consider the construction of a 4-input K-map in Fig. 4.18. In a 4-input K-map,
neighboring cells can wrap around both the top-to-bottom edges in addition to the side-to-side edges.
Notice that all 16 cells are positioned within the map so that their neighbors on the top, bottom and sides
only differ by one bit in their input codes.
116 • Chapter 4: Combinational Logic Design

Fig. 4.18
Formation of a 4-input K-map

4.4.2.2 Logic Minimization Using K-Maps (Sum of Products)

Now we look at using a K-map to create a minimized logic expression in a SOP form. Remember
that each cell with an output of 1 has a minterm associated with it, just as in the truth table. When two
neighboring cells have outputs of 1, it graphically indicates that the two minterms can be reduced into a
minimized product term that will cover both outputs. Consider the example given in Fig. 4.19.
 4.4 Logic Minimization • 117

Fig. 4.19
Observing how K-Maps visually highlight logic minimizations

These observations can be put into a formal process to produce a minimized SOP logic expression
using a K-map. The steps are as follows:
1. Circle groups of 1’s in the K-map following the rules:
• Each circle should contain the largest number of 1’s possible.
• The circles encompass only neighboring cells (i.e., side-to-side sides and/or top and
bottom).
• The circles must contain a number of 1’s that is a power of 2 (i.e., 1, 2, 4, 8 or 16).
• Enter as many circles as possible without having any circles fully cover another circle.
• Each circle is called a Prime Implicant.
2. Create a product term for each prime implicant following the rules:
• Each variable in the K-map is evaluated one-by-one.
• If the circle covers a region where the input variable is a 1, then include it in the product
term uncomplemented.
• If the circle covers a region where the input variable is a 0, then include it in the product
term complemented.
118 • Chapter 4: Combinational Logic Design

• If the circle covers a region where the input variable is both a 0 and 1, then the variable is
excluded from the product term.
3. Sum all of the product terms for each prime implicant.

Let’s apply this approach to our 2-input K-map example. Example 4.23 shows the process of finding
a minimized sum of products logic expression for a 2-input logic circuit using a K-map. This process
yielded the same SOP expression as the algebraic minimization and observations shown in Fig. 4.19,
but with a formalized process.

Example 4.23
Using a K-map to find a minimized sum of products expression (2-input)

Let’s now apply this process to our 3-input K-map example. Example 4.24 shows the process of
finding a minimized sum of products logic expression for a 3-input logic circuit using a K-map. This
example shows circles that overlap. This is legal as long as one circle does not fully encompass another.
Overlapping circles are common since the K-map process dictates that circles should be drawn that
group the largest number of ones possible as long as they are in powers of 2. Forming groups of ones
using ones that have already been circled is perfectly legal to accomplish larger groupings. The larger
the grouping of ones, the more chance there is for a variable to be excluded from the product term. This
results in better minimization of the logic.
 4.4 Logic Minimization • 119

Example 4.24
Using a K-map to find a minimized sum of products expression (3-input)

Let’s now apply this process to our 4-input K-map example. Example 4.25 shows the process of
finding a minimized sum of products logic expression for a 4-input logic circuit using a K-map.
120 • Chapter 4: Combinational Logic Design

Example 4.25
Using a K-map to find a minimized sum of products expression (4-input)

4.4.2.3 Logic Minimization Using K-Maps (Product of Sums)

K-maps can also be used to create minimized product of sums logic expressions. This is the same
concept as how a minterm list and maxterm list each produce the same logic function, but in comple-
mentary fashions. When creating a product of sums expression from a K-map, groups of 0’s are circled.
For each circle, a sum term is derived with a negation of variables similar to when forming a maxterm
(i.e., in the input variable is a 0, then it is included uncomplemented in the sum term and vice versa). The
 4.4 Logic Minimization • 121

final step in forming the minimized POS expression is to AND all of the sum terms together. The formal
process is as follows:
1. Circle groups of 0’s in the K-map following the rules:
• Each circle should contain the largest number of 0’s possible.
• The circles encompass only neighboring cells (i.e., side-to-side sides and/or top and
bottom).
• The circles must contain a number of 0’s that is a power of 2 (i.e., 1, 2, 4, 8 or 16).
• Enter as many circles as possible without having any circles fully cover another circle.
• Each circle is called a Prime Implicant.
2. Create a sum term for each prime implicant following the rules:
• Each variable in the K-map is evaluated one-by-one.
• If the circle covers a region where the input variable is a 1, then include it in the sum term
complemented.
• If the circle covers a region where the input variable is a 0, then include it in the sum term
uncomplemented.
• If the circles cover a region where the input variable is both a 0 and 1, then the variable is
excluded from the sum term.
3. Multiply all of the sum terms for each prime implicant.

Let’s apply this approach to our 2-input K-map example. Example 4.26 shows the process of finding
a minimized product of sums logic expression for a 2-input logic circuit using a K-map. Notice that this
process yielded the same logic expression as the SOP approach shown in Example 4.23. This illustrates
that both the POS and SOP expressions produce the correct logic for the circuit.

Example 4.26
Using a K-map to find a minimized product of sums expression (2-input)
122 • Chapter 4: Combinational Logic Design

Let’s now apply this process to our 3-input K-map example. Example 4.27 shows the process of
finding a minimized product of sums logic expression for a 3-input logic circuit using a K-map. Notice that
the logic expression in POS form is not identical to the SOP expression found in Example 4.24; however,
using a few steps of algebraic manipulation shows that the POS expression can be put into a form that is
identical to the prior SOP expression. This illustrates that both the POS and SOP produce equivalent
functionality for the circuit.

Example 4.27
Using a K-map to find a minimized product of sums expression (3-input)
 4.4 Logic Minimization • 123

Let’s now apply this process to our 4-input K-map example. Example 4.28 shows the process of
finding a minimized product of sums logic expression for a 4-input logic circuit using a K-map.

Example 4.28
Using a K-map to find a minimized product of sums expression (4-input)

4.4.2.4 Minimal Sum

One situation that arises when minimizing logic using a K-map is that some of the prime implicants
may be redundant. Consider the example in Fig. 4.20.
124 • Chapter 4: Combinational Logic Design

Fig. 4.20
Observing redundant prime implicants in a K-map

We need to define a formal process for identifying redundant prime implicants that can be removed
without impacting the result of the logic expression. Let’s start with examining the sum of products form.
First, we define the term essential prime implicant as a prime implicant that cannot be removed from
the logic expression without impacting its result. We then define the term minimal sum as a logic
expression that represents the most minimal set of logic operations to accomplish a sum of products
form. There may be multiple minimal sums for a given truth table, but each would have the same number
of logic operations. In order to determine if a prime implicant is essential, we first put in each and every
possible prime implicant into the K-map. This gives a logic expression known as the complete sum.
From this point we identify any cells that have only one prime implicant covering them. These cells are
 4.4 Logic Minimization • 125

called distinguished one cells. Any prime implicant that covers a distinguished one cell is defined as an
essential prime implicant. All prime implicants that are not essential are removed from the K-map. A
minimal sum is then simply the sum of all remaining product terms associated with the essential prime
implicants. Example 4.29 shows how to use this process.

Example 4.29
Deriving the minimal sum from a K-map

This process is identical for the product of sums form to produce the minimal product.

4.4.3 Don’t Cares

There are often times when framing a design problem that there are specific input codes that require
exact output values, but there are other codes where the output value doesn’t matter. This can occur for a
variety of reasons, such as knowing that certain input codes will never occur due to the nature of the
problem or that the output of the circuit will only be used under certain input codes. We can take
advantage of this situation to produce a more minimal logic circuit. We define an output as a don’t
care when it doesn’t matter whether it is a 1 or 0 for the particular input code. The symbol for a don’t care
is “X”. We take advantage of don’t cares when performing logic minimization by treating them as
whatever output value will produce a minimal logic expression. Example 4.30 shows how to use this
process.
126 • Chapter 4: Combinational Logic Design

Example 4.30
Using don’t cares to produce a minimal SOP logic expression

4.4.4 Using XOR Gates

While Boolean algebra does not include the exclusive-OR and exclusive-NOR operations, XOR and
XNOR gates do indeed exist in modern electronics. They can be a useful tool to provide logic circuitry
with less operations, sometimes even compared to a minimal sum or product synthesized using the
techniques just described. An XOR/XNOR operation can be identified by putting the values from a truth
table into a K-map. The XOR/XNOR operations will result in a characteristic checkerboard pattern in the
K-map. Consider the following patterns for XOR and XNOR gates in Figs. 4.21, 4.22, 4.23, and 4.24.
Anytime these patterns are observed, it indicates an XOR/XNOR gate.
 4.4 Logic Minimization • 127

Fig. 4.21
XOR and XNOR checkerboard patterns observed in K-maps (2-input)

Fig. 4.22
XOR and XNOR checkerboard patterns observed in K-maps (3-input)

Fig. 4.23
XOR checkerboard pattern observed in K-maps (4-input)
128 • Chapter 4: Combinational Logic Design

Fig. 4.24
XNOR checkerboard pattern observed in K-maps (4-input)

CONCEPT CHECK

CC4.4(a) Logic minimization is accomplished by removing variables from the original canonical logic
expression that don’t impact the result. How does a Karnaugh map graphically show what
variables can be removed?
(A) K-maps contain the same information as a truth table but the data is formatted
as a grid. This allows variables to be removed by inspection.
(B) K-maps rearrange a truth table so that adjacent cells have one and only one
input variable changing at a time. If adjacent cells have the same output value
when an input variable is both a 0 and a 1, that variable has no impact on the
interim result and can be eliminated.
(C) K-maps list both the rows with outputs of 1’s and 0’s simultaneously. This allows
minimization to occur for a SOP and POS topology that each have the same, but
minimal, number of gates.
(D) K-maps display the truth table information in a grid format, which is a more
compact way of presenting the behavior of a circuit.

CC4.4(b) A “Don’t Care” can be used to minimize a logic expression by assigning the output of a row
to either a 1 or a 0 in order to form larger groupings within a K-map. How does the output of
the circuit behave when it processes the input code for a row containing a don’t care?
(A) The output will be whatever value was needed to form the largest grouping in
the K-map.
(B) The output will go to either a 0 or a 1, but the final value is random.
(C) The output can toggle between a 0 and a 1 when this input code is present.
(D) The output will be driven to exactly halfway between a 0 and a 1.
 4.5 Timing Hazards & Glitches • 129

4.5 Timing Hazards & Glitches

Timing hazards, or glitches, refer to unwanted transitions on the output of a combinational logic
circuit. These are most commonly due to different delay paths through the gates in the circuit. In real
circuitry there is always a finite propagation delay through each gate. Consider the circuit shown in
Fig. 4.25 where gate delays are included and how they can produce unwanted transitions.

Fig. 4.25
Examining the source of a timing hazard (or glitch) in a combinational logic circuit

These timing hazards are given unique names based on the type of transition that occurs. A static
0 timing hazard is when the input switches between two input codes that both yield an output of 0 but the
output momentarily switches to a 1. A static 1 timing hazard is when the input switches between two
130 • Chapter 4: Combinational Logic Design

input codes that both yield an output of 1 but the output momentarily switches to a 0. A dynamic hazard
is when the input switches between two input codes that result in a real transition on the output (i.e., 0 to
1 or 1 to 0), but the output has a momentary glitch before reaching its final value. These definitions are
shown in Fig. 4.26.

Fig. 4.26
Timing hazard definitions

Timing hazards can be addressed in a variety of ways. One way is to try to match the propagation
delays through each path of the logic circuit. This can be difficult, particularly in modern logic families
such as CMOS. In the example in Fig. 4.25, the root cause of the different propagation delays was due to
an inverter on one of the variables. It seems obvious that this could be addressed by putting buffers on
the other inputs with equal delays as the inverter. This would create a situation where all input codes
would arrive at the first stage of AND gates at the same time regardless of whether they were inverted or
not and eliminate the hazards; however, CMOS implements a buffer as two inverters in series, so it is
difficult to insert a buffer in a circuit with an equal delay to an inverter. Addressing timing hazards in this
way is possible, but it involves a time-consuming and tedious process of adjusting the transistors used to
create the buffer and inverter to have equal delays.
Another technique to address timing hazards is to place additional circuitry in the system that will
ensure the correct output while the input codes switch. Consider how including a non-essential prime
implicant can eliminate a timing hazard in Example 4.31. In this approach, the minimal sum from Fig. 4.25
is instead replaced with the complete sum. The use of the complete sum instead of the minimal sum can
be shown to eliminate both static and dynamic timing hazards. The drawback of this approach is the
addition of extra circuitry in the combinational logic circuit (i.e., non-essential prime implicants).
 4.5 Timing Hazards & Glitches • 131

Example 4.31
Eliminating a timing hazard by including non-essential product terms

CONCEPT CHECK

CC4.5 How long do you need to wait for all hazards to settle out?
(A) The time equal to the delay through the non-essential prime implicants.
(B) The time equal to the delay through the essential prime implicants.
(C) The time equal to the shortest delay path in the circuit.
(D) The time equal to the longest delay path in the circuit.
132 • Chapter 4: Combinational Logic Design

Summary
v Boolean algebra defines the axioms and v A Karnaugh map (K-map) is a graphical
theorems that guide the operations that can approach to minimizing logic expressions.
be performed on a two-valued number A K-map arranges a truth table into a grid in
system. which the neighboring cells have input codes
v Boolean algebra theorems allow logic that differ by only one bit. This allows the
expressions to be manipulated to make cir- impact of an input variable on a group of
cuit synthesis simpler. They also allow logic outputs to be quickly identified.
expressions to be minimized. v A minimized sum of products expression can
v The delay of a combinational logic circuit is be found from a K-map by circling neighbor-
always dictated by the longest delay path ing ones to form groups that can be produced
from the inputs to the output. by a single product term. Each product term
v The canonical form of a logic expression is (aka prime implicant) is then summed
one that has not been minimized. together to form the circuit output.
v A canonical sum of products form is a logic v A minimized product of sums expression can
synthesis technique based on minterms. A be found from a K-map by circling neighbor-
minterm is a product term that will output a ing zeros to form groups that can be pro-
one for only one unique input code. A duced by a single sum term. Each sum term
minterm is used for each row of a truth table (aka prime implicant) is then multiplied
corresponding to an output of a one. Each of together to form the circuit output.
the minterms are then summed together to v A minimal sum or minimal product is a logic
create the final system output. expression that contains only essential prime
v A minterm list is a shorthand way of describ- implicants and represents the smallest num-
ing the information in a truth table. The sym- ber of logic operations possible to produce
bol “Σ” is used to denote a minterm list. the desired output.
Each of the input variables are added to this v A don’t care (X) can be used when the output
symbol as comma delimited subscripts. The of a truth table row can be either a zero or a
row number is then listed for each row one without affecting the system behavior.
corresponding to an output of a one. This typically occurs when some of the input
v A canonical product of sums form is a logic codes of a truth table will never occur. The
synthesis technique based on maxterms. A value for the row of a truth table containing a
maxterm is a sum term that will output a zero don’t care output can be chosen to give the
for only one unique input code. A maxterm most minimal logic expression. In a K-map,
is used for each row of a truth table don’t cares can be included to form the larg-
corresponding to an output of a zero. Each est groupings in order to give the least
of the maxterms are then multiplied together amount of logic.
to create the final system output. v While exclusive-OR gates are not used in
v A maxterm list is a shorthand way of describ- Boolean algebra, they can be visually
ing the information in a truth table. The sym- identified in K-maps by looking for checker-
bol “Π” is used to denote a maxterm list. board patterns.
Each of the input variables are added to this v Timing hazards are temporary glitches that
symbol as comma delimited subscripts. occur on the output of a combinational logic
The row number is then listed for each row circuit due to timing mismatches through dif-
corresponding to an output of a zero. ferent paths in the circuit. Hazards can be
v Canonical logic expressions can be minimized by including additional circuitry in
minimized through a repetitive process of the system or by matching the delay of all
factoring common variables using the distrib- signal paths.
utive property and then eliminating remaining
variables using a combination of the
complements and identity theorems.
 Exercise Problems • 133

Exercise Problems
Section 4.1: Boolean Algebra 4.1.17 Which Boolean algebra theorem describes the
situation where the order of variables in an
4.1.1 Which Boolean algebra theorem describes the
AND operation does not affect the result?
situation where any variable OR’d with itself
will yield itself? 4.1.18 Which Boolean algebra theorem describes the
situation where a variable OR’d with a 0 will
4.1.2 Which Boolean algebra theorem describes the
yield itself?
situation where any variable that is double
complemented will yield itself? 4.1.19 Use proof by exhaustion to prove that an OR
gate with its inputs inverted is equivalent to an
4.1.3 Which Boolean algebra theorem describes the
AND gate with its outputs inverted.
situation where any variable OR’d with a 1 will
yield a 1? 4.1.20 Use proof by exhaustion to prove that an AND
gate with its inputs inverted is equivalent to an
4.1.4 Which Boolean algebra theorem describes the
OR gate with its outputs inverted.
situation where a variable that exists in multiple
product terms can be factored out?
4.1.5 Which Boolean algebra theorem describes the
Section 4.2: Combinational Logic
situation where when output(s) corresponding Analysis
to a term within an expression are handled by 4.2.1 For the logic diagram given in Fig. 4.27, give
another term the original term can be the logic expression for the output F.
removed?
4.1.6 Which Boolean algebra theorem describes the
situation where any variable AND’d with its
complement will yield a 0?
4.1.7 Which Boolean algebra theorem describes the
situation where any variable AND’d with a 0 will
yield a 0?
4.1.8 Which Boolean algebra theorem describes the Fig. 4.27
situation where an AND gate with its inputs Combinational logic analysis 1
inverted is equivalent to an OR gate with its
outputs inverted? 4.2.2 For the logic diagram given in Fig. 4.27, give
4.1.9 Which Boolean algebra theorem describes the the truth table for the output F.
situation where a variable that exists in multiple 4.2.3 For the logic diagram given in Fig. 4.27, give
sum terms can be factored out? the delay.
4.1.10 Which Boolean algebra theorem describes the 4.2.4 For the logic diagram given in Fig. 4.28, give
situation where an OR gate with its inputs the logic expression for the output F.
inverted is equivalent to an AND gate with its
outputs inverted?
4.1.11 Which Boolean algebra theorem describes the
situation where the grouping of variables in an
OR operation does not affect the result?
4.1.12 Which Boolean algebra theorem describes the
situation where any variable AND’d with itself
will yield itself?
4.1.13 Which Boolean algebra theorem describes the
situation where the order of variables in an OR
operation does not affect the result?
4.1.14 Which Boolean algebra theorem describes the
Fig. 4.28
situation where any variable AND’d with a 1 will Combinational logic analysis 2
yield itself?
4.1.15 Which Boolean algebra theorem describes the 4.2.5 For the logic diagram given in Fig. 4.28, give
situation where the grouping of variables in an the truth table for the output F.
AND operation does not affect the result? 4.2.6 For the logic diagram given in Fig. 4.28, give
4.1.16 Which Boolean algebra theorem describes the the delay.
situation where any variable OR’d with its com- 4.2.7 For the logic diagram given in Fig. 4.29, give
plement will yield a 1? the logic expression for the output F.
134 • Chapter 4: Combinational Logic Design

4.3.8 For the 2-input minterm list in Fig. 4.31, give

the canonical sum of products (SOP) logic
diagram.
4.3.9 For the 2-input minterm list in Fig. 4.31, give
the truth Table.
4.3.10 For the 2-input minterm list in Fig. 4.31, give
the canonical product of sums (POS) logic
expression.
4.3.11 For the 2-input minterm list in Fig. 4.31, give
the canonical product of sums (POS) logic
Fig. 4.29 diagram.
Combinational logic analysis 3 4.3.12 For the 2-input minterm list in Fig. 4.31, give
the maxterm list.
4.2.8 For the logic diagram given in Fig. 4.29, give 4.3.13 For the 2-input maxterm list in Fig. 4.32, give
the truth table for the output F. the canonical sum of products (SOP) logic
4.2.9 For the logic diagram given in Fig. 4.29, give expression.
the delay.

Section 4.3: Combinational Logic

Synthesis Fig. 4.32
4.3.1 For the 2-input truth table in Fig. 4.30, give
Combinational logic synthesis 3
the canonical sum of products (SOP) logic
expression. 4.3.14 For the 2-input maxterm list in Fig. 4.32, give
the canonical sum of products (SOP) logic
diagram.
4.3.15 For the 2-input maxterm list in Fig. 4.32, give
the minterm list.
4.3.16 For the 2-input maxterm list in Fig. 4.32, give
the canonical product of sums (POS) logic
expression.
Fig. 4.30
Combinational logic synthesis 1 4.3.17 For the 2-input maxterm list in Fig. 4.32, give
the canonical product of sums (POS) logic
diagram.
4.3.2 For the 2-input truth table in Fig. 4.30, give the
canonical sum of products (SOP) logic 4.3.18 For the 2-input maxterm list in Fig. 4.32, give
diagram. the truth table.
4.3.3 For the 2-input truth table in Fig. 4.30, give the 4.3.19 For the 3-input truth table in Fig. 4.33, give
minterm list. the canonical sum of products (SOP) logic
expression.
4.3.4 For the 2-input truth table in Fig. 4.30, give the
canonical product of sums (POS) logic
expression.
4.3.5 For the 2-input truth table in Fig. 4.30, give the
canonical product of sums (POS) logic
diagram.
4.3.6 For the 2-input truth table in Fig. 4.30, give the
maxterm list.
4.3.7 For the 2-input minterm list in Fig. 4.31, give Fig. 4.33
the canonical sum of products (SOP) logic Combinational logic synthesis 4
expression.
4.3.20 For the 3-input truth table in Fig. 4.33, give the
canonical sum of products (SOP) logic
diagram.
Fig. 4.31 4.3.21 For the 3-input truth table in Fig. 4.33, give the
Combinational logic synthesis 2 minterm list.
 Exercise Problems • 135

4.3.22 For the 3-input truth table in Fig. 4.33, give the
canonical product of sums (POS) logic
expression.
4.3.23 For the 3-input truth table in Fig. 4.33, give the
canonical product of sums (POS) logic
diagram.
4.3.24 For the 3-input truth table in Fig. 4.33, give the
maxterm list.
4.3.25 For the 3-input minterm list in Fig. 4.34, give
the canonical sum of products (SOP) logic
expression.

Fig. 4.34 Fig. 4.36

Combinational logic synthesis 5 Combinational logic synthesis 7

4.3.26 For the 3-input minterm list in Fig. 4.34, give 4.3.38 For the 4-input truth table in Fig. 4.36, give the
the canonical sum of products (SOP) logic canonical sum of products (SOP) logic
diagram. diagram.
4.3.27 For the 3-input minterm list in Fig. 4.34, give 4.3.39 For the 4-input truth table in Fig. 4.36, give the
the truth table. minterm list.
4.3.28 For the 3-input minterm list in Fig. 4.34, give 4.3.40 For the 4-input truth table in Fig. 4.36, give the
the canonical product of sums (POS) logic canonical product of sums (POS) logic
expression. expression.
4.3.29 For the 3-input minterm list in Fig. 4.34, give 4.3.41 For the 4-input truth table in Fig. 4.36, give the
the canonical product of sums (POS) logic canonical product of sums (POS) logic
diagram. diagram.
4.3.30 For the 3-input minterm list in Fig. 4.34, give 4.3.42 For the 4-input truth table in Fig. 4.36, give the
the maxterm list. maxterm list.
4.3.31 For the 3-input maxterm list in Fig. 4.35, give 4.3.43 For the 4-input minterm list in Fig. 4.37, give
the canonical sum of products (SOP) logic the canonical sum of products (SOP) logic
expression. expression.

Fig. 4.35 Fig. 4.37

Combinational logic synthesis 6 Combinational logic synthesis 8

4.3.32 For the 3-input maxterm list in Fig. 4.35, give 4.3.44 For the 4-input minterm list in Fig. 4.37, give
the canonical sum of products (SOP) logic the canonical sum of products (SOP) logic
diagram. diagram.
4.3.33 For the 3-input maxterm list in Fig. 4.35, give 4.3.45 For the 4-input minterm list in Fig. 4.37, give
the minterm list. the truth Table.
4.3.34 For the 3-input maxterm list in Fig. 4.35, give 4.3.46 For the 4-input minterm list in Fig. 4.37, give
the canonical product of sums (POS) logic the canonical product of sums (POS) logic
expression. expression.
4.3.35 For the 3-input maxterm list in Fig. 4.35, give 4.3.47 For the 4-input minterm list in Fig. 4.37, give
the canonical product of sums (POS) logic the canonical product of sums (POS) logic
diagram. diagram.
4.3.36 For the 3-input maxterm list in Fig. 4.35, give 4.3.48 For the 4-input minterm list in Fig. 4.37, give
the truth table. the maxterm list.
4.3.37 For the 4-input truth table in Fig. 4.36, give 4.3.49 For the 4-input maxterm list in Fig. 4.38, give
the canonical sum of products (SOP) logic the canonical sum of products (SOP) logic
expression. expression.
136 • Chapter 4: Combinational Logic Design

Fig. 4.38
Combinational logic synthesis 9
Fig. 4.41
4.3.50 For the 4-input maxterm list in Fig. 4.38, give Logic minimization 3
the canonical sum of products (SOP) logic
diagram. 4.4.6 For the 2-input truth table in Fig. 4.41, use a
4.3.51 For the 4-input maxterm list in Fig. 4.38, give K-map to derive a minimized product of sums
the minterm list. (POS) logic expression.
4.3.52 For the 4-input maxterm list in Fig. 4.38, give 4.4.7 For the 2-input truth table in Fig. 4.42, use a
the canonical product of sums (POS) logic K-map to derive a minimized sum of products
expression. (SOP) logic expression.
4.3.53 For the 4-input maxterm list in Fig. 4.38, give
the canonical product of sums (POS) logic
diagram.
4.3.54 For the 4-input maxterm list in Fig. 4.38, give
the truth table.

Section 4.4: Logic Minimization Fig. 4.42

4.4.1 For the 2-input truth table in Fig. 4.39, use a Logic minimization 4
K-map to derive a minimized sum of products
(SOP) logic expression. 4.4.8 For the 2-input truth table in Fig. 4.42, use a
K-map to derive a minimized product of sums
(POS) logic expression.
4.4.9 For the 3-input truth table in Fig. 4.43, use a
K-map to derive a minimized sum of products
(SOP) logic expression.

Fig. 4.39
Logic minimization 1

4.4.2 For the 2-input truth table in Fig. 4.39, use a

K-map to derive a minimized product of sums
(POS) logic expression.
4.4.3 For the 2-input truth table in Fig. 4.40, use a Fig. 4.43
K-map to derive a minimized sum of products Logic minimization 5
(SOP) logic expression.
4.4.10 For the 3-input truth table in Fig. 4.43, use a
K-map to derive a minimized product of sums
(POS) logic expression.
4.4.11 For the 3-input truth table in Fig. 4.44, use a
K-map to derive a minimized sum of products
(SOP) logic expression.
Fig. 4.40
Logic minimization 2

4.4.4 For the 2-input truth table in Fig. 4.40, use a

K-map to derive a minimized product of sums
(POS) logic expression.
4.4.5 For the 2-input truth table in Fig. 4.41, use a
K-map to derive a minimized sum of products
(SOP) logic expression.

Fig. 4.44
Logic minimization 6
 Exercise Problems • 137

4.4.12 For the 3-input truth table in Fig. 4.44, use a 4.4.18 For the 4-input truth table in Fig. 4.47, use a
K-map to derive a minimized product of sums K-map to derive a minimized product of sums
(POS) logic expression. (POS) logic expression.
4.4.13 For the 3-input truth table in Fig. 4.45, use a 4.4.19 For the 4-input truth table in Fig. 4.48, use a
K-map to derive a minimized sum of products K-map to derive a minimized sum of products
(SOP) logic expression. (SOP) logic expression.

Fig. 4.45
Logic minimization 7

4.4.14 For the 3-input truth table in Fig. 4.45, use a

K-map to derive a minimized product of sums
(POS) logic expression.
4.4.15 For the 3-input truth table in Fig. 4.46, use a
K-map to derive a minimized sum of products
(SOP) logic expression.
Fig. 4.48
Logic minimization 10

4.4.20 For the 4-input truth table in Fig. 4.48, use a

K-map to derive a minimized product of sums
(POS) logic expression.
4.4.21 For the 4-input truth table in Fig. 4.49, use a
K-map to derive a minimized sum of products
(SOP) logic expression.

Fig. 4.46
Logic minimization 8

4.4.16 For the 3-input truth table in Fig. 4.46, use a

K-map to derive a minimized product of sums
(POS) logic expression.
4.4.17 For the 4-input truth table in Fig. 4.47, use a
K-map to derive a minimized sum of products
(SOP) logic expression.

Fig. 4.49
Logic minimization 11
Fig. 4.47
Logic minimization 9 4.4.22 For the 4-input truth table in Fig. 4.49, use a
K-map to derive a minimized product of sums
(POS) logic expression.
138 • Chapter 4: Combinational Logic Design

4.4.23 For the 4-input truth table in Fig. 4.50, use a

K-map to derive a minimized sum of products
(SOP) logic expression.

Fig. 4.52
Logic minimization 14

4.4.30 For the 4-input truth table and K-map in

Fig. 4.52, give the product terms for the essen-
tial prime implicants.
4.4.31 For the 4-input truth table and K-map in
Fig. 4.50 Fig. 4.52, give the minimal sum of products
Logic minimization 12 (SOP) logic expression.
4.4.32 For the 4-input truth table and K-map in
4.4.24 For the 4-input truth table in Fig. 4.50, use a Fig. 4.52, give the complete sum of products
K-map to derive a minimized product of sums (SOP) logic expression.
(POS) logic expression. 4.4.33 For the 4-input truth table and K-map in
4.4.25 For the 3-input truth table and K-map in Fig. 4.53, give the minimal sum of products
Fig. 4.51, provide the row number(s) of any (SOP) logic expression by exploiting “don’t
distinguished one-cells. cares”.

Fig. 4.51
Logic minimization 13

4.4.26 For the 3-input truth table and K-map in

Fig. 4.51, give the product terms for the essen-
tial prime implicants.
Fig. 4.53
4.4.27 For the 3-input truth table and K-map in Logic minimization 15
Fig. 4.51, give the minimal sum of products
logic expression.
4.4.34 For the 4-input truth table and K-map in
4.4.28 For the 3-input truth table and K-map in Fig. 4.53, give the minimal product of sums
Fig. 4.51, give the complete sum of products (POS) logic expression by exploiting “don’t
logic expression. cares”.
4.4.29 For the 4-input truth table and K-map in 4.4.35 For the 4-input truth table and K-map in
Fig. 4.52, provide the row number(s) of any Fig. 4.54, give the minimal sum of products
distinguished one-cells. (SOP) logic expression by exploiting “don’t
cares”.
 Exercise Problems • 139

Section 4.5: Timing Hazards & Glitches

4.5.1 Describe the situation in which a static-1 timing
hazard may occur.
4.5.2 Describe the situation in which a static-0 timing
hazard may occur.
4.5.3 In which topology will a static-1 timing hazard
occur (SOP, POS, or both)?
4.5.4 In which topology will a static-0 timing hazard
occur (SOP, POS, or both)?
4.5.5 For the 3-input truth table and K-map in
Fig. 4.51, give the product term that helps
eliminate static-1 timing hazards in this circuit.
4.5.6 For the 3-input truth table and K-map in
Fig. 4.51, give the sum term that helps elimi-
Fig. 4.54 nate static-0 timing hazards in this circuit.
Logic minimization 16 4.5.7 For the 4-input truth table and K-map in
Fig. 4.52, give the product term that helps
eliminate static-1 timing hazards in this circuit.
4.4.36 For the 4-input truth table and K-map in
Fig. 4.54, give the minimal product of sums 4.5.8 For the 4-input truth table and K-map in
(POS) logic expression by exploiting “don’t Fig. 4.52, give the sum term that helps elimi-
cares”. nate static-0 timing hazards in this circuit.
Chapter 5: Verilog (Part 1)
Based on the material presented in Chap. 4, there are a few observations about logic design that are
apparent. First, the size of logic circuitry can scale quickly to the point where it is difficult to design by
hand. Second, the process of moving from a high-level description of how a circuit works (e.g., a truth
table) to a form that is ready to be implemented with real circuitry (e.g., a minimized logic diagram) is
straightforward and well-defined. Both of these observations motivate the use of computer aided design
(CAD) tools to accomplish logic design. This chapter introduces hardware description languages (HDLs)
as a means to describe digital circuitry using a text-based language. HDLs provide a means to describe
large digital systems without the need for schematics, which can become impractical in very large
designs. HDLs have evolved to support logic simulation at different levels of abstraction. This provides
designers the ability to begin designing and verifying functionality of large systems at a high level of
abstraction and postpone the details of the circuit implementation until later in the design cycle. This
enables a top-down design approach that is scalable across different logic families. HDLs have also
evolved to support automated synthesis, which allows the CAD tools to take a functional description of a
system (e.g., a truth table) and automatically create the gate level circuitry to be implemented in real
hardware. This allows designers to focus their attention on designing the behavior of a system and not
spend as much time performing the formal logic synthesis steps that were presented in Chap. 4. The
intent of this chapter is to introduce HDLs and their use in the modern digital design flow. This chapter will
cover the basics of designing combinational logic in an HDL and also hierarchical design. The more
advanced concepts of HDLs such as sequential logic design, high level abstraction, and test benches
are covered later so that the reader can get started quickly using HDLs to gain experience with the
languages and design flow.
There are two dominant hardware description languages in use today. They are VHDL and Verilog.
VHDL stands for very high speed integrated circuit hardware description language. Verilog is not an
acronym but rather a trade name. The use of these two HDLs is split nearly equally within the digital
design industry. Once one language is learned it is simple to learn the other language, so the choice of
the HDL to learn first is somewhat arbitrary. In this text, we will use Verilog to learn the concepts of an
HDL. Verilog is more similar to the programming language C and less strict in its type casting than VHDL.
Verilog is also widely used in custom integrated circuit design so there is a great deal of documentation
and examples readily available online. The goal of this chapter is to provide an understanding of the
basic principles of hardware description languages.

Learning Outcomes—After completing this chapter, you will be able to:

5.1 Describe the role of hardware description languages in modern digital design.
5.2 Describe the fundamentals of design abstraction in modern digital design.
5.3 Describe the modern digital design flow based on hardware description languages.
5.4 Describe the fundamental constructs of Verilog.
5.5 Design a Verilog model for a combinational logic circuit using concurrent modeling
techniques (continuous signal assignment with logical operators and continuous signal
assignment with conditional operators).
5.6 Design a Verilog model for a combinational logic circuit using a structural design approach
(gate level primitives and user defined primitives).
5.7 Describe the role of a Verilog test bench.

# Springer International Publishing AG 2017 141

B.J. LaMeres, Introduction to Logic Circuits & Logic Design with Verilog,
DOI 10.1007/978-3-319-53883-9_5
142 • Chapter 5: Verilog (Part 1)

5.1 History of Hardware Description Languages

The invention of the integrated circuit is most commonly credited to two individuals who filed patents
on different variations of the same basic concept within six months of each other in 1959. Jack Kilby filed
the first patent on the integrated circuit in February of 1959 titled “Miniaturized Electronic Circuits” while
working for Texas Instruments. Robert Noyce was the second to file a patent on the integrated circuit in
July of 1959 titled “Semiconductor Device and Lead Structure” while at a company he cofounded called
Fairchild Semiconductor. Kilby went on to win the Nobel Prize in Physics in 2000 for his invention, while
Noyce went on to cofound Intel Corporation in 1968 with Gordon Moore. In 1971, Intel introduced the first
single-chip microprocessor using integrated circuit technology, the Intel 4004. This microprocessor IC
contained 2300 transistors. This series of inventions launched the semiconductor industry, which was
the driving force behind the growth of Silicon Valley, and led to 40 years of unprecedented advancement
in technology that has impacted every aspect of the modern world.

Gordon Moore, cofounder of Intel, predicted in 1965 that the number of transistors on an integrated
circuit would double every two years. This prediction, now known as Moore’s Law, has held true since the
invention of the integrated circuit. As the number of transistors on an integrated circuit grew, so did the
size of the design and the functionality that could be implemented. Once the first microprocessor was
invented in 1971, the capability of CAD tools increased rapidly enabling larger designs to be accom-
plished. These larger designs, including newer microprocessors, enabled the CAD tools to become even
more sophisticated and, in turn, yield even larger designs. The rapid expansion of electronic systems
based on digital integrated circuits required that different manufacturers needed to produce designs that
were compatible with each other. The adoption of logic family standards helped manufacturers ensure
their parts would be compatible with other manufacturers at the physical layer (e.g., voltage and current);
however, one challenge that was encountered by the industry was a way to document the complex
behavior of larger systems. The use of schematics to document large digital designs became too
cumbersome and difficult to understand by anyone besides the designer. Word descriptions of the
behavior were easier to understand, but even this form of documentation became too voluminous to
be effective for the size of designs that were emerging. Simultaneously there was a need to begin
simulating the functionality of these large systems prior to fabrication to verify accuracy. Due to the
complexity of these systems and the vast potential for design error, it became impractical to verify design
accuracy through prototyping.
In 1983, the US Department of Defense (DoD) sponsored a program to create a means to document
the behavior of digital systems that could be used across all of its suppliers. This program was motivated
by a lack of adequate documentation for the functionality of application specific integrated circuits
(ASICs) that were being supplied to the DoD. This lack of documentation was becoming a critical
issue as ASICs would come to the end of their life cycle and need to be replaced. With the lack of a
standardized documentation approach, suppliers had difficulty reproducing equivalent parts to those that
had become obsolete. The DoD contracted three companies (Texas Instruments, IBM, and Intermetrics)
to develop a standardized documentation tool that provided detailed information about both the interface
(i.e., inputs and outputs) and the behavior of digital systems. The new tool was to be implemented in a
format similar to a programming language. Due to the nature of this type of language-based tool, it was a
natural extension of the original project scope to include the ability to simulate the behavior of a digital
system. The simulation capability was desired to span multiple levels of abstraction to provide maximum
flexibility. In 1985, the first version of this tool, called VHDL, was released. In order to gain widespread
adoption and ensure consistency of use across the industry, VHDL was turned over to the Institute of
Electrical and Electronic Engineers (IEEE) for standardization. IEEE is a professional association that
defines a broad range of open technology standards. In 1987, IEEE released the first industry standard
version of VHDL. The release was titled IEEE 1076–1987. Feedback from the initial version resulted in a
major revision of the standard in 1993 titled IEEE 1076–1993. While many minor revisions have been
 5.1 History of Hardware Description Languages • 143

made to the 1993 release, the 1076–1993 standard contains the vast majority of VHDL functionality in
use today. The most recent VHDL standard is IEEE 1076–2008.
Also in 1983, the Verilog HDL was developed by Automated Integrated Design Systems as a logic
simulation language. The development of Verilog took place completely independent from the VHDL
project. Automated Integrated Design Systems (renamed Gateway Design Automation in 1985) was
acquired by CAD tool vendor Cadence Design Systems in 1990. In response to the popularity of Verilog’s
intuitive programming and superior simulation support, and also to stay competitive with the emerging
VHDL standard, Cadence made the Verilog HDL open to the public. IEEE once again developed the
open standard for this HDL, and in 1995 released the Verilog standard titled IEEE 1364–1995. This
release has undergone numerous revisions with the most significant occurring in 2001. It is common to
refer to the major releases as “Verilog 1995” and “Verilog 2001” instead of their official standard numbers.
The development of CAD tools to accomplish automated logic synthesis can be dated back to the
1970’s when IBM began developing a series of practical synthesis engines that were used in the design
of their mainframe computers; however, the main advancement in logic synthesis came with the founding
of a company called Synopsis in 1986. Synopsis was the first company to focus on logic synthesis
directly from HDLs. This was a major contribution because designers were already using HDLs to
describe and simulate their digital systems, and now logic synthesis became integrated in the same
design flow. Due to the complexity of synthesizing highly abstract functional descriptions, only lower-
levels of abstraction that were thoroughly elaborated were initially able to be synthesized. As CAD tool
capability evolved, synthesis of higher levels of abstraction became possible, but even today not all
functionality that can be described in an HDL can be synthesized.

The history of HDLs, their standardization, and the creation of the associated logic synthesis tools is
key to understanding the use and limitations of HDLs. HDLs were originally designed for documentation
and behavioral simulation. Logic synthesis tools were developed independently and modified later to
work with HDLs. This history provides some background into the most common pitfalls that beginning
digital designers encounter, that being that most any type of behavior can be described and simulated in
an HDL, but only a subset of well-described functionality can be synthesized. Beginning digital designers
are often plagued by issues related to designs that simulate perfectly but that will not synthesize
correctly. In this book, an effort is made to introduce Verilog at a level that provides a reasonable amount
of abstraction while preserving the ability to be synthesized. Figure 5.1 shows a timeline of some of the
major technology milestones that have occurred in the past 150 years in the field of digital logic and
HDLs.
144 • Chapter 5: Verilog (Part 1)

Fig. 5.1
Major milestones in the advancement of Digital Logic and HDLs

CONCEPT CHECK

CC5.1 Why does Verilog support modeling techniques that aren’t synthesizable?
(A) There wasn’t enough funding available to develop synthesis capability as it all went
to the VHDL project.
(B) At the time Verilog was created, synthesis was deemed too difficult to implement.
(C) To allow Verilog to be used as a generic programming language.
(D) Verilog needs to support all steps in the modern digital design flow, some of which
are unsynthesizable such as test pattern generation and timing verification.
 5.2 HDL Abstraction • 145

5.2 HDL Abstraction

HDLs were originally defined to be able to model behavior at multiple levels of abstraction.
Abstraction is an important concept in engineering design because it allows us to specify how systems
will operate without getting consumed prematurely with implementation details. Also, by removing the
details of the lower level implementation, simulations can be conducted in reasonable amounts of time to
model the higher-level functionality. If a full computer system was simulated using detailed models for
every MOSFET, it would take an impracticable amount of time to complete. Figure 5.2 shows a graphical
depiction of the different layers of abstraction in digital system design.

Fig. 5.2
Levels of design abstraction

The highest level of abstraction is the system level. At this level, behavior of a system is described
by stating a set of broad specifications. An example of a design at this level is a specification such as “the
computer system will perform 10 Tera Floating Point Operations per Second (10 TFLOPS) on double
146 • Chapter 5: Verilog (Part 1)

precision data and consume no more than 100 Watts of power”. Notice that these specifications do not
dictate the lower level details such as the type of logic family or the type of computer architecture to use.
One level down from the system level is the algorithmic level. At this level, the specifications begin to be
broken down into sub-systems, each with an associated behavior that will accomplish a part of the
primary task. At this level, the example computer specifications might be broken down into sub-systems
such as a central processing unit (CPU) to perform the computation and random access memory (RAM)
to hold the inputs and outputs of the computation. One level down from the algorithmic level is the
register transfer level (RTL). At this level, the details of how data is moved between and within
sub-systems are described in addition to how the data is manipulated based on system inputs. One
level down from the RTL level is the gate level. At this level, the design is described using basic gates and
registers (or storage elements). The gate level is essentially a schematic (either graphically or text-
based) that contains the components and connections that will implement the functionality from the
above levels of abstraction. One level down from the gate level is the circuit level. The circuit level
describes the operation of the basic gates and registers using transistors, wires and other electrical
components such as resistors and capacitors. Finally, the lowest level of design abstraction is the
material level. This level describes how different materials are combined and shaped in order to
implement the transistors, devices and wires from the circuit level.
HDLs are designed to model behavior at all of these levels with the exception of the material level.
While there is some capability to model circuit level behavior such as MOSFETs as ideal switches and
pull-up/pull-down resistors, HDLs are not typically used at the circuit level. Another graphical depiction of
design abstraction is known as the Gajski and Kuhn’s Y-chart. A Y-chart depicts abstraction across
three different design domains: behavioral, structural and physical. Each of these design domains
contains levels of abstraction (i.e., system, algorithm, RTL, gate, and circuit). An example Y-chart is
shown in Fig. 5.3.
 5.2 HDL Abstraction • 147

Fig. 5.3
Y-Chart of design abstraction

A Y-chart also depicts how the abstraction levels of different design domains are related to each
other. A top-down design flow can be visualized in a Y-chart by spiraling inward in a clockwise direction.
Moving from the behavioral domain to the structural domain is the process of synthesis. Whenever
synthesis is performed, the resulting system should be compared with the prior behavioral description.
This checking is called verification. The process of creating the physical circuitry corresponding to the
structural description is called implementation. The spiral continues down through the levels of abstrac-
tion until the design is implemented at a level that the geometries representing circuit elements
(transistors, wires, etc.) are ready to be fabricated in silicon. Figure 5.4 shows the top-down design
process depicted as an inward spiral on the Y-chart.
148 • Chapter 5: Verilog (Part 1)

Fig. 5.4
Y-Chart illustrating top-down design approach

The Y-chart represents a formal approach for large digital systems. For large systems that are
designed by teams of engineers, it is critical that a formal, top-down design process is followed to
eliminate potentially costly design errors as the implementation is carried out at lower levels of
abstraction.

CONCEPT CHECK

CC5.2 Why is abstraction an essential part of engineering design?

(A) Without abstraction all schematics would be drawn at the transistor-level.
(B) Abstraction allows computer programs to aid in the design process.
(C) Abstraction allows the details of the implementation to be hidden while the higher-
level systems are designed. Without abstraction, the details of the implementation
would overwhelm the designer.
(D) Abstraction allows analog circuit designers to include digital blocks in their systems.
 5.3 The Modern Digital Design Flow • 149

5.3 The Modern Digital Design Flow

When performing a smaller design or the design of fully-contained sub-systems, the process can be
broken down into individual steps. These steps are shown in Fig. 5.5. This process is given generically
and applies to both classical and modern digital design. The distinction between classical and modern is
that modern digital design uses HDLs and automated CAD tools for simulation, synthesis, place and
route, and verification.

Fig. 5.5
Generic digital design flow

This generic design process flow can be used across classical and modern digital design, although
modern digital design allows additional verification at each step using automated CAD tools. Figure 5.6
shows how this flow is used in the classical design approach of a combinational logic circuit.
150 • Chapter 5: Verilog (Part 1)

Fig. 5.6
Classical digital design flow

The modern design flow based on HDLs includes the ability to simulate functionality at each step of
the process. Functional simulations can be performed on the initial behavioral description of the system.
At each step of the design process the functionality is described in more detail, ultimately moving toward
the fabrication step. At each level, the detailed information can be included in the simulation to verify that
the functionality is still correct and that the design is still meeting the original specifications. Figure 5.7
shows the modern digital design flow with the inclusion of simulation capability at each step.
 5.3 The Modern Digital Design Flow • 151

Fig. 5.7
Modern digital design flow

CONCEPT CHECK

CC5.3 Why did digital designs move from schematic-entry to text-based HDLs?
(A) HDL models could be much larger by describing functionality in text similar to
traditional programming language.
(B) Schematics required sophisticated graphics hardware to display correctly.
(C) Schematics symbols became too small as designs became larger.
(D) Text was easier to understand by a broader range of engineers.
152 • Chapter 5: Verilog (Part 1)

5.4 Verilog Constructs

Now we begin looking at the details of Verilog. The original Verilog standard (IEEE 1364) has been
updated numerous times since its creation in 1995. The most significant update occurred in 2001, which
was titled IEEE 1394–2001. In 2005 minor corrections and improvements were added to the standard,
which resulted in IEEE 1394–2005. The constructs described in this book reflect the functionality in the
IEEE 1394–2005 standard. The functionality of Verilog (e.g., operators, signal types, functions, etc.) is
defined within the Verilog standard, thus it is not necessary to explicitly state that a design is using the
IEEE 1394 package because it is inherent in the use of Verilog. This chapter gives an overview of the
basic constructs of Verilog in order to model simple combinational logic circuits and begin gaining
experience with logic simulations. The more advanced constructs of Verilog are covered in Chap. 8
with examples given throughout Chaps. 9, 10, 11, 12, and 13.

A Verilog design describes a single system in a single file. The file has the suffix *.v. Within the file,
the system description is contained within a module. The module includes the interface to the system
(i.e., the inputs and outputs) and the description of the behavior. Figure 5.8 shows a graphical depiction
of a Verilog file.

Fig. 5.8
The anatomy of a Verilog file

Verilog is case sensitive. Also, each Verilog assignment, definition or declaration is terminated with
a semicolon (;). As such, line wraps are allowed and do not signify the end of an assignment, definition or
declaration. Line wraps can be used to make Verilog more readable. Comments in Verilog are supported
in two ways. The first way is called a line comment and is preceded with two slashes (i.e., //). Everything
after the slashes is considered a comment until the end of the line. The second comment approach is
called a block comment and begins with /* and ends with a */. Everything between /* and */ is considered
a comment. A block comment can span multiple lines. All user-defined names in Verilog must start with
an alphabetic letter, not a number. User-defined names are not allowed to be the same as any Verilog
keyword. This chapter contains many definitions of syntax in Verilog. The following notations will be used
throughout the chapter when introducing new constructs.

bold ¼ Verilog keyword, use as is, case sensitive.

italics ¼ User-defined name, case sensitive.
<> ¼ A required characteristic such as a data type, input/output, etc.
 5.4 Verilog Constructs • 153

5.4.1 Data Types

In Verilog, every signal, constant, variable and function must be assigned a data type. The IEEE
1394–2005 standard provides a variety of pre-defined data types. Some data types are synthesizable,
while others are only for modeling abstract behavior. The following are the most commonly used data
types in the Verilog language.

5.4.1.1 Value Set

Verilog supports four basic values that a signal can take on: 0, 1, X, and Z. Most of the pre-defined
data types in Verilog store these values. A description of each value supported is given below.

Value Description
0 A logic zero, or false condition.
1 A logic one, or true condition.
x or X Unknown or uninitialized.
z or Z High impedance, tri-stated, or floating.

In Verilog, these values also have an associated strength. The strengths are used to resolve the
value of a signal when it is driven by multiple sources. The names, syntax and relative strengths are
given below.

Strength Description Strength level

supply1 Supply drive for VCC 7
supply0 Supply drive for VSS, or GND 7
strong1 Strong drive to logic one 6
strong0 Strong drive to logic zero 6
pull1 Medium drive to logic one 5
pull0 Medium drive to logic zero 5
large Large capacitive 4
weak1 Weak drive to logic one 3
weak0 Weak drive to logic zero 3
medium Medium capacitive 2
small Small capacitive 1
highz1 High impedance with weak pull-up to logic one 0
highz0 High impedance with weak pull-down to logic zero 0

When a signal is driven by multiple drivers, it will take on the value of the driver with the highest
strength. If the two drivers have the same strength, the value will be unknown. If the strength is not
specified, it will default to strong drive, or level 6.

5.4.1.2 Net Data Types

Every signal within Verilog must be associated with a data type. A net data type is one that models
an interconnection (aka., a net) between components and can take on the values 0, 1, X, and Z. A signal
with a net data type must be driven at all times and updates its value when the driver value changes. The
most common synthesizable net data type in Verilog is the wire. The type wire will be used throughout
this text. There are also a variety of other more advanced net data types that model complex digital
systems with multiple drivers for the same net. The syntax and description for all Verilog net data types
are given below.
154 • Chapter 5: Verilog (Part 1)

Type Description
wire A simple connection between components.
wor Wired-OR. If multiple drivers, their values are OR’d together.
wand Wired-AND’d. If multiple drivers, their values are AND’d together.
supply0 Used to model the VSS, (GND), power supply (supply strength inherent).
supply1 Used to model the VCC power supply (supply strength inherent).
tri Identical to wire. Used for readability for a net driven by multiple sources.
trior Identical to wor. Used for readability for nets driven by multiple sources.
triand Identical to wand. Used for readability for nets driven by multiple sources.
tri1 Pulls up to logic one when tri-stated.
tri0 Pulls down to logic zero when tri-stated.
trireg Holds last value when tri-stated (capacitance strength inherent).

Each of these net types can also have an associated drive strength. The strength is used in
determining the final value of the net when it is connected to multiple drivers.

5.4.1.3 Variable Data Types

Verilog also contains data types that model storage. These are called variable data types. A variable
data type can take on the values 0, 1, X, and Z, but does not have an associated strength. Variable data
types will hold the value assigned to them until their next assignment. The syntax and description for the
Verilog variable data types are given below.

Type Description
reg A variable that models logic storage. Can take on values 0, 1, X, and Z.
integer A 32-bit, 2’s complement variable representing whole numbers between
2,147,483,64810 to +2,147,483,647.
real A 64-bit, floating point variable representing real numbers between –(2.2x10308)10
to +(2.2x10308)10.
time An unsigned, 64-bit variable taking on values from 010 to +(9.2x1018).
realtime Same as time. Just used for readability.

5.4.1.4 Vectors
In Verilog, a vector is a one-dimensional array of elements. All of the net data types, in addition to the
variable type reg, can be used to form vectors. The syntax for defining a vector is as follows:
<type> [<MSB_index>:<LSB_index>] vector_name

While any range of indices can be used, it is common practice to have the LSB index start at zero.
Example:

wire [7:0] Sum; // This defines an 8-bit vector called “Sum” of type wire. The
// MSB is given the index 7 while the LSB is given the index 0.

reg [15:0] Q; // This defines a 16-bit vector called “Q” of type reg.

Individual bits within the vector can be addressed using their index. Groups of bits can be accessed
using an index range.
Sum[0]; // This is the least significant bit of the vector “Sum” defined
above.
Q[15:8]; // This is the upper 8-bits of the 16-bit vector “Q” defined
above.
 5.4 Verilog Constructs • 155

5.4.1.5 Arrays
An array is a multi-dimensional array of elements. This can also be thought of as a “vector of
vectors”. Vectors within the array all have the same dimensions. To declare an array, the element type
and dimensions are defined first followed by the array name and its dimensions. It is common practice to
place the start index of the array on the left side of the “:” when defining its dimensions. The syntax for the
creation of an array is shown below.
<element_type> [<MSB_index>:<LSB_index>] array_name [<array_start_index>:
<array_end_index>];

Example:
reg[7:0] Mem[0:4095]; // Defines an array of 4096, 8-bit vectors of type reg.
integer A[1:100]; // Defines an array of 100 integers.

When accessing an array, the name of the array is given first, followed by the index of the element. It
is also possible to access an individual bit within an array by adding appending the index of element.

Example:
Mem[2]; // This is the 3rd element within the array named “Mem”.
// This syntax represents an 8-bit vector of type reg.

Mem[2][7]; // This is the MSB of the 3rd element within the array named “Mem”.
// This syntax represents a single bit of type reg.

A[2]; // This is the 2nd element within the array named “A”. Recall
// that A was declared with a starting index of 1.
// This syntax represents a 32-bit, signed integer.

5.4.1.6 Expressing Numbers Using Different Bases

If a number is simply entered into Verilog without identifying syntax, it is treated as an integer.
However, Verilog supports defining numbers in other bases. Verilog also supports an optional bit size
and sign of a number. When defining the value of arrays, the “_” can be inserted between numerals to
improve readability. The “_” is ignored by the Verilog compiler. Values of numbers can be entered in
either upper or lower case (i.e., z or Z, f or F, etc.). The syntax for specifying the base of a number is as
follows:

<size_in_bits>’<base><value>

Note that specifying the size is optional. If it is omitted, the number will default to a 32-bit vector with
leading zeros added as necessary. The supported bases are as follows:

Syntax Description
‘b Unsigned binary
‘o Unsigned octal
‘d Unsigned decimal
‘h Unsigned hexadecimal
‘sb Signed binary
‘so Signed octal
‘sd Signed decimal
‘sh Signed hexadecimal
156 • Chapter 5: Verilog (Part 1)

Example:
10 // This is treated as decimal 10, which is a 32-bit signed vector.
4’b1111 // A 4-bit number with the value 11112.
8’b1011_0000 // An 8-bit number with the value 101100002.
8’hFF // An 8-bit number with the value 111111112.
8’hff // An 8-bit number with the value 111111112.
6’hA // A 6-bit number with the value 0010102. Note that leading zeros
// were added to make the value 6-bits.
8’d7 // An 8-bit number with the value 000001112.
32’d0 // A 32-bit number with the value 0000_000016.
’b1111 // A 32-bit number with the value 0000_000F16.
8’bZ // An 8-bit number with the value ZZZZ_ZZZZ.

5.4.1.7 Assigning Between Different Types

Verilog is said to be a weakly-typed (or loosely typed) language, meaning that it permits
assignments between different data types. This is as opposed to a strongly-typed language (such as
VHDL) where signal assignments are only permitted between like types. The reason Verilog permits
assignment between different types is because it treats all of its types as just groups of bits. When
assigning between different types, Verilog will automatically truncate or add leading bits as necessary to
make the assignment work. The following examples illustrate how Verilog handles a few assignments
between different types. Assume that a variable called ABC_TB has been declared as type reg[2:0].
Example:

ABC_TB ¼ 2’b00; // ABC_TB will be assigned 3’b000. A leading bit is automatically

added.
ABC_TB ¼ 5; // ABC_TB will be assigned 3’b101. The integer is truncated to
3-bits.
ABC_TB ¼ 8; // ABC_TB will be assigned 3’b000. The integer is truncated to
3-bits.

5.4.2 The Module

All systems in Verilog are encapsulated inside of a module. Modules can include instantiations of
lower-level modules in order to support hierarchical designs. The keywords module and endmodule
signify the beginning and end of the system description. When working on large designs, it is common
practice to place each module in its own file with the same name.

module module_name (port_list); // Pre Verilog-2001

// port_definitions
// module_items
endmodule

or
module module_name (port_list and port_definitions); // Verilog-2001 and after
// module_items
endmodule

5.4.2.1 Port Definitions

The first item within a module is its definition of the inputs and outputs, or ports. Each port needs to
have a user-defined name, a direction, and a type. The user-defined port names are case sensitive and
must begin an alphabetic character. The port directions are declared to be one of the three types: input,
output, and inout. A port can take on any of the previously described data types, but only wires,
 5.4 Verilog Constructs • 157

registers, and integers are synthesizable. Port names with the same type and direction can be listed on
the same line separated by commas.
There are two different port definition styles supported in Verilog. Prior to the Verilog-2001 release,
the port names were listed within parentheses after the module name. Then within the module, the
directionality and type of the ports were listed. Starting with the Verilog-2001 release, the port directions
and types could be included alongside the port names within the parenthesis after the module name.
This approach mimicked more of an ANSCI-C approach to passing inputs/outputs to a system. In this
text, the newer approach to port definition will be used. Example 5.1 shows multiple approaches for
defining a module and its ports.

Example 5.1
Declaring Verilog module ports

5.4.2.2 Signal Declarations

A signal that is used for internal connections within a system is declared within the module before its
first use. Each signal must be declared by listing its type followed by a user-defined name. Signal names
of like type can be declared on the same line separated with a comma. All of the legal data types
described above can be used for signals; however, only types net, reg, and integer will synthesize
directly. The syntax for a signal declaration is as follows:
<type> name;
158 • Chapter 5: Verilog (Part 1)

Example:
wire node1; // declare a signal named “node1” of type wire
reg Q2, Q1, Q0; // declare three signals named “Q2”, “Q1”, and “Q0”, all of
type reg
wire [63:0] bus1; // declare a 64-bit vector named “bus1” with all bits of
type wire
integer i,j; // declare two integers called “i” and “j”

Verilog supports a hierarchical design approach, thus signal names can be the same within a
sub-system as those at a higher level without conflict. Figure 5.9 shows an example of legal signal
naming in a hierarchical design.

Fig. 5.9
Verilog signals and systems

5.4.2.3 Parameter Declarations

A parameter, or constant, is useful for representing a quantity that will be used multiple times in the
architecture. The syntax for declaring a parameter is as follows:
parameter <type> constant_name ¼ <value>;

Note that the type is optional and can only be integer, time, real, or realtime. If a type is provided,
the parameter will have the same properties as a variable of the same time. If the type is excluded, the
parameter will take on the type of the value assigned to it.
Example:

parameter BUS_WIDTH ¼ 64;

parameter NICKEL ¼ 8’b0000_0101;

Once declared, the constant name can be used throughout the module. The following example
illustrates how we can use a constant to define the size of a vector. Notice that since we defined the
constant to be the actual width of the vector (i.e., 32-bits), we need to subtract one from its value when
defining the indices (i.e., [31:0]).

Example:
wire [BUS_WIDTH-1:0] BUS_A; // It is acceptable to add a “space”
for readability
 5.4 Verilog Constructs • 159

5.4.2.4 Compiler Directives

A compiler directive provides additional information to the simulation tool on how to interpret the
Verilog model. A compiler directive is placed before the module definition and is preceded with a backtick
(i.e., `). Note that this is not an apostrophe. A few of the most commonly used compiler directives are as
follows:

Syntax Description
`timescale < unit>,<precision> Defines the timescale of the delay unit and its smallest precision
`include < filename> Includes additional files in the compilation
`define < macroname > <value> Declares a global constant

Example:
‘timescale 1ns/1ps // Declares the unit of time is 1 ns with a precision of 1ps.
// The precision is the smallest amount that the time can
// take on. For example, with this directive the number
// 0.001 would be interpreted as 0.001 ns, or 1 ps.
// However, the number 0.0001 would be interpreted as 0 since
// it is smaller than the minimum precision value.

5.4.3 Verilog Operators

There are a variety of pre-defined operators in the Verilog standard. It is important to note that
operators are defined to work on specific data types and that not all operators are synthesizable.

5.4.3.1 Assignment Operator

Verilog uses the equal sign (¼) to denote an assignment. The left-hand side (LHS) of the assign-
ment is the target signal. The right-hand side (RHS) contains the input arguments and can contain both
signals, constants, and operators.

Example:
F1 ¼ A; // F1 is assigned the signal A
F2 ¼ 4’hAA; // F2 is an 8-bit vector and is assigned the value 101010102

5.4.3.2 Bitwise Logical Operators

Bitwise operators perform logic functions on individual bits. The inputs to the operation are single
bits and the output is a single bit. In the case where the inputs are vectors, each bit in the first vector is
operated on by the bit in the same position from the second vector. If the vectors are not the same length,
the shorter vector is padded with leading zeros to make both lengths equal. Verilog contains the following
bitwise operators:

Syntax Operation
~ Negation
& AND
| OR
^ XOR
~^ or ^~ XNOR
<< Logical shift left (fill empty LSB location with zero)
>> Logical shift right (fill empty MSB location with zero)
160 • Chapter 5: Verilog (Part 1)

Example:
~X // invert each bit in X
X&Y // AND each bit of X with each bit of Y
X|Y // OR each bit of X with each bit of Y
X^Y // XOR each bit of X with each bit of Y
X ~^ Y // XNOR each bit of X with each bit of Y
X << 3 // Shift X left 3 times and fill with zeros
Y >> 2 // Shift Y right 2 times and fill with zeros

5.4.3.3 Reduction Logic Operators

A reduction operator is one that uses each bit of a vector as individual inputs into a logic operation
and produces a single bit output. Verilog contains the following reduction logic operators.

Syntax Operation
& AND all bits in the vector together (1-bit result)
~& NAND all bits in the vector together (1-bit result)
| OR all bits in the vector together (1-bit result)
~| NOR all bits in the vector together (1-bit result)
^ XOR all bits in the vector together (1-bit result)
~^ or ^~ XNOR all bits in the vector together (1-bit result)

Example:
&X // AND all bits in vector X together
~&X // NAND all bits in vector X together
|X // OR all bits in vector X together
~|X // NOR all bits in vector X together
^X // XOR all bits in vector X together
~^X // XNOR all bits in vector X together

5.4.3.4 Boolean Logic Operators

A Boolean logic operator is one that returns a value of TRUE (1) or FALSE (0) based on a logic
operation of the input operations. These operations are used in decision statements.

Syntax Operation
! Negation
&& AND
|| OR

Example:
!X // TRUE if all values in X are 0, FALSE otherwise
X && Y // TRUE if the bitwise AND of X and Y results in all ones, FALSE otherwise
X || Y // TRUE if the bitwise OR of X and Y results in all ones, FALSE otherwise

5.4.3.5 Relational Operators

A relational operator is one that returns a value of TRUE (1) or FALSE (0) based on a comparison of
two inputs.
 5.4 Verilog Constructs • 161

Syntax Description
¼¼ Equality
!¼ Inequality
< Less than
> Greater than
<¼ Less than or equal
>¼ Greater than or equal

Example:
X ¼¼ Y // TRUE if X is equal to Y, FALSE otherwise
X !¼ Y // TRUE if X is not equal to Y, FALSE otherwise
X<Y // TRUE if X is less than Y, FALSE otherwise
X>Y // TRUE if X is greater than Y, FALSE otherwise
X <¼ Y // TRUE if X is less than or equal to Y, FALSE otherwise
X >¼ Y // TRUE if X is greater than or equal to Y, FALSE otherwise

5.4.3.6 Conditional Operators

Verilog contains a conditional operator that can be used to provide a more intuitive approach to
modeling logic statements. The keyword for the conditional operator is ? with the following syntax:
<target_net> ¼ <Boolean_condition> ? <true_assignment> : <false_assignment>;

This operator specifies a Boolean condition in which if evaluated TRUE, the true_assignment will be
assigned to the target. If the Boolean condition is evaluated FALSE, the false_assignment portion of the
operator will be assigned to the target. The values in this assignment can be signals or logic values. The
Boolean condition can be any combination of the Boolean operators described above. Nested condi-
tional operators can also be implemented by inserting subsequent conditional operators in place of the
false_value.

Example:
F ¼ (A ¼¼ 1’b0) ? 1’b1 : 1’b0; // If A is a zero, F¼1, otherwise F¼0.
This models an inverter.

F ¼ (sel ¼¼ 1’b0) ? A : B; // If sel is a zero, F¼A, otherwise F¼B.

This models a selectable switch.

F ¼ ((A ¼¼ 1’b0) && (B ¼¼ 1’b0)) ? 1’b’0 : // Nested conditional statements.

((A ¼¼ 1’b0) && (B ¼¼ 1’b1)) ? 1’b’1 : // This models an XOR gate.
((A ¼¼ 1’b1) && (B ¼¼ 1’b0)) ? 1’b’1 :
((A ¼¼ 1’b1) && (B ¼¼ 1’b1)) ? 1’b’0;

F ¼ ( !C && (!A || B) ) ? 1’b1 : 1’b0; // This models the logic expression

// F ¼ C’(A’+B).

5.4.3.7 Concatenation Operator

In Verilog, the curly brackets (i.e., {}) are used to concatenate multiple signals. The target of this
operation must be the same size of the sum of the sizes of the input arguments.
162 • Chapter 5: Verilog (Part 1)

Example:
Bus1[7:0] ¼ {Bus2[7:4], Bus3[3:0]}; // Assuming Bus1, Bus2, and Bus3 are all 8-bit
// vectors, this operation takes the upper
4-bits of
// Bus2, concatenates them with the lower
4-bits of
// Bus3, and assigns the 8-bit combination
to Bus1.

BusC ¼ {BusA, BusB}; // If BusA and BusB are 4-bits, then BusC
// must be 8-bits.

BusC[7:0] ¼ {4’b0000, BusA}; // This pads the 4-bit vector BusA with 4x
leading
// zeros and assigns to the 8-bit vector
BusC.

5.4.3.8 Replication Operator

Verilog provides the ability to concatenate a vector with itself through the replication operator. This
operator uses double curly brackets (i.e., {{}}) and an integer indicating the number of replications to be
performed. The replication syntax is as follows:
{<number_of_replications>{<vector_name_to_be_replicated>}}

Example:
BusX ¼ {4{Bus1}}; // This is equivalent to: BusX ¼ {Bus1, Bus1, Bus1, Bus1};
BusY ¼ {2{A,B}}; // This is equivalent to: BusY ¼ {A, B, A, B};
BusZ ¼ {Bus1, {2{Bus2}}}; // This is equivalent to: BusZ ¼ {Bus1, Bus2, Bus2};

5.4.3.9 Numerical Operators

Verilog also provides a set of numerical operators as follows:

Syntax Operation
+ Addition
 Subtraction (when placed between arguments)
 2’s complement negation (when placed in front of an argument)
* Multiplication
/ Division
% Modulus
** Raise to the power
<<< Shift to the left, fill with zeros
<<< Shift to the right, fill with sign bit

Example:

X+Y // Add X to Y
X-Y // Subtract Y from X
-X // Take the two’s complement negation of X
X*Y // Multiply X by Y
X/Y // Divide X by Y
X%Y // Modulus X/Y
X ** Y // Raise X to the power of Y
X <<< 3 // Shift X left 3 times, fill with zeros
X >>> 2 // Shift X right 2 times, fill with sign bit
 5.4 Verilog Constructs • 163

Verilog will allow the use of these operators on arguments of different sizes, types and signs. The
rules of the operations are as follows:

• If two vectors are of different sizes, the smaller vector is expanded to the size of the larger
vector.
– If the smaller vector is unsigned, it is padded with zeros.
– If the smaller vector is signed, it is padded with the sign bit.
• If one of the arguments is real, then the arithmetic will take place using real numbers.
• If one of the arguments is unsigned, then all arguments will be treated as unsigned.

5.4.3.10 Operator Precedence

The following is the order of precedence of the Verilog operators:

Operators Precedence Notes

!~+ Highest Bitwise/Unary
{} {{}} Concatenation/Replication
() # No operation, just parenthesis
** Power
*/% Binary Multiply/Divide/Modulo
+ # Binary Addition/Subtraction
<< >> <<< >>> Shift Operators
< <¼ > > ¼ Greater/Less than Comparisons
¼¼! ¼ # Equality/Inequality Comparisons
&~& AND/NAND Operators
^ ~^ XOR/XNOR Operators
| ~| # OR/NOR Operators
&& Boolean AND
|| Boolean OR
?: Lowest Conditional Operator

CONCEPT CHECK

CC5.4(a) What revision of Verilog added the ability to list the port names, types, and directions just
once after the module name?
(A) Verilog-1995.
(B) Verilog-2001.
(C) Verilog-2005.
(D) SystemVerilog.
CC5.4(b) What is the difference between types wire and reg?
(A) They are the same.
(B) The type wire is a simple interconnection while reg will hold the value of its last
assignment.
(C) The type wire is for scalars while the type reg is for vectors.
(D) Only wire is synthesizable.
164 • Chapter 5: Verilog (Part 1)

5.5 Modeling Concurrent Functionality in Verilog

It is important to remember that Verilog is a hardware description language, not a programming
language. In a programming language, the lines of code are executed sequentially as they appear in the
source file. In Verilog, the lines of code represent the behavior of real hardware. Thus, the assignments
are executed concurrently unless specifically noted otherwise.

5.5.1 Continuous Assignment

Verilog uses the keyword assign to denote a continuous signal assignment. After this keyword, an
assignment is made using the ¼ symbol. The left-hand side (LHS) of the assignment is the target signal
and must be a net type. The right hand side (RHS) contains the input arguments and can contain nets,
regs, constants, and operators. A continuous assignment models combinational logic. Any change to the
RHS of the expression will result in an update to the LHS target net.
Example:

assign F1 ¼ A; // F1 is updated anytime A changes, where A is a signal

assign F2 ¼ 1’b0; // F2 is assigned the value 0
assign F3 ¼ 4’hAA; // F3 is an 8-bit vector and is assigned the value 101010102

Each individual assignment will be executed concurrently and synthesized as separate logic
circuits. Consider the following example.
Example:
assign X ¼ A;
assign Y ¼ B;
assign Z ¼ C;

When simulated, these three lines of Verilog will make three separate signal assignments at the
exact same time. This is different from a programming language that will first assign A to X, then B to Y
and finally C to Z. In Verilog this functionality is identical to three separate wires. This description will be
directly synthesized into three separate wires.

Below is another example of how continuous signal assignments in Verilog differ from a sequentially
executed programming language.
Example:
assign A ¼ B;
assign B ¼ C;

In a Verilog simulation, the signal assignments of C to B and B to A will take place at the same time.
This means during synthesis, the signal B will be eliminated from the design since this functionality
describes two wires in series. Automated synthesis tools will eliminate this unnecessary signal name.
This is not the same functionality that would result if this example was implemented as a sequentially
executed computer program. A computer program would execute the assignment of B to A first, then
assign the value of C to B second. In this way, B represents a storage element that is passed to A before
it is updated with C.

5.5.2 Continuous Assignment with Logical Operators

Each of the logical operators described in Sect. 5.4.3.2 can be used in conjunction with concurrent
signal assignments to create individual combinational logic circuits. Example 5.2 shows how to design a
Verilog model of a combinational logic circuit using this approach.
 5.5 Modeling Concurrent Functionality in Verilog • 165

Example 5.2
Modeling combinational logic using continuous assignment with logical operators

5.5.3 Continuous Assignment with Conditional Operators

Logical operators are good for describing the behavior of small circuits; however, in the prior
example we still needed to create the canonical sum of products logic expression by hand before
describing the functionality in Verilog. The true power of an HDL is when the behavior of the system
can be described fully without requiring any hand design. The conditional operator allows us to describe
a continuous assignment using Boolean conditions that effect the values of the result. In this approach,
we use the conditional operator (?) in conjunction with the continuous assignment keyword assign.
Example 5.3 shows how to design a Verilog model of a combinational logic circuit using continuous
assignment with conditional operators. Note that this example uses the same truth table as in
Example 5.2 to illustrate a comparison between approaches.
166 • Chapter 5: Verilog (Part 1)

Example 5.3
Modeling combinational logic using continuous assignment with conditional operators (1)

In the prior example, the conditional operator was based on a truth table. Conditional operators can
also be used to model logic expressions. Example 5.4 shows how to design a Verilog model of a
combinational logic circuit when the logic expression is already known. Note that this example again
uses the same truth table as in Example 5.2 and Example 5.3 to illustrate a comparison between
approaches.
 5.5 Modeling Concurrent Functionality in Verilog • 167

Example 5.4
Modeling combinational logic using continuous assignment with conditional operators (2)

5.5.4 Continuous Assignment with Delay

Verilog provides the ability to model gate delays when using a continuous assignment. The # is used
to indicate a delayed assignment. For combinational logic circuits, the delay can be specified for all
transitions, for rising and falling transitions separately, and for rising, falling, and transitions to the value
off separately. A transition to off refers to a transition to Z. If only one delay parameter is specified, it is
used to model all delays. If two delay parameters are specified, the first parameter is used for the rise
time delay while the second is used to model the fall time delay. If three parameters are specified, the
third parameter is used to model the transition to off. Parenthesis are optional but recommended when
using multiple delay parameters.

assign #(<del_all>) <target_net> ¼ <RHS_nets, operators, etc. . .>;

assign #(<del_rise, del_fall>) <target_net> ¼ <RHS_nets, operators, etc. . .>;
assign #(<del_rise, del_fall, del_off>) <target_net> ¼ <RHS_nets, operators, etc. . .>;

Example:
assign #1 F ¼ A; // Delay of 1 on all transitions.
assign #(2,3) F ¼ A; // Delay of 2 for rising transitions and 3 for falling.
assign #(2,3,4) F ¼ A; // Delay of 2 for rising, 3 for falling, and 4 for off
transitions.

When using delay, it is typical to include the `timescale directive to provide the units of the delay
being specified. Example 5.5 shows a graphical depiction of using delay with continuous assignments
when modeling combinational logic circuits.
168 • Chapter 5: Verilog (Part 1)

Example 5.5
Modeling delay in continuous assignments

Verilog also provides a mechanism to model a range of delays that are selected by a switch set in the
CAD compiler. There are three delays categories that can be specified: minimum, typical, and maximum.
The delays are separated by a “:”. The following is the syntax of how to use the delay range capability.
assign #(<min>:<typ>:<max>) <target_net> ¼ <RHS_nets, operators, etc.. . .>;

Example:

assign #(1:2:3) F ¼ A;
// Specifying a range of delays for all
transitions.
assign #(1:1:2, 2:2:3) F ¼ A; // Specifying a range of delays for rising/
falling.
assign #(1:1:2, 2:2:3, 4:4:5) F ¼ A; // Specifying a range of delays for each
transition.
 5.5 Modeling Concurrent Functionality in Verilog • 169

The delay modeling capability in continuous assignment is designed to model the behavior of real
combinational logic with respect to short duration pulses. When a pulse is shorter than the delay of the
combinational logic gate, the pulse is ignored. Ignoring brief input pulses on the input accurately models
the behavior of on-chip gates. When the input pulse is faster than the delay of the gate, the output of
the gate does not have time to respond. As a result, there will not be a logic change on the output. This
is called inertial delay modeling and is the default behavior when using continuous assignments.
Example 5.6 shows a graphical depiction of inertial delay behavior in Verilog.

Example 5.6
Inertial delay modeling when using continuous assignment

CONCEPT CHECK

CC5.5(a) Why is concurrency such an important concept in HDLs?

(A) Concurrency is a feature of HDLs that can’t be modeled using schematics.
(B) Concurrency allows automated synthesis to be performed.
(C) Concurrency allows logic simulators to display useful system information.
(D) Concurrency is necessary to model real systems that operate in parallel.
CC5.5(b) Why does modeling combinational logic in its canonical form with continuous assignment
with logical operators defeat the purpose of the modern digital design flow?
(A) It requires the designer to first create the circuit using the classical digital design
approach and then enter it into the HDL in a form that is essentially a text-based
netlist. This doesn’t take advantage of the abstraction capabilities and
automated synthesis in the modern flow.
(B) It cannot be synthesized because the order of precedence of the logical
operators in Verilog doesn’t match the precedence defined in Boolean algebra.
(C) The circuit is in its simplest form so there is no work for the synthesizer to do.
(D) It doesn’t allow an else clause to cover the outputs for any remaining input
codes not explicitly listed.
170 • Chapter 5: Verilog (Part 1)

5.6 Structural Design and Hierarchy

Structural design in Verilog refers to including lower-level sub-systems within a higher-level module
in order to produce the desired functionality. This is called hierarchy, and is a good design practice
because it enables design partitioning. A purely structural design will not contain any behavioral
constructs in the module such as signal assignments, but instead just contain the instantiation and
interconnections of other sub-systems. A sub-system in Verilog is simply another module that is called by
a higher-level module. Each lower-level module that is called is executed concurrently by the calling
module.

5.6.1 Lower-Level Module Instantiation

The term instantiation refers to the use or inclusion of a lower-level module within a system. In
Verilog, the syntax for instantiating a lower-level module is as follows.
module_name <instance_identifier> (port mapping. . .);

The first portion of the instantiation is the module name that is being called. This must match the
lower level module name exactly, including case. The second portion of the instantiation is an optional
instance identifier. Instance identifier are useful when instantiating multiple instances of the same lower-
level module. The final portion of the instantiation is the port mapping. There are two techniques to
connect signals to the ports of the lower-level module, explicit and positional.

5.6.1.1 Explicit Port Mapping

In explicit port mapping the names of the ports of the lower-level sub-system are provided along with
the signals they are being connected to. The lower-level port name is preceded with a period (.) while the
signal it is being connected is enclosed within parenthesis. The port connections can be listed in any
order since the details of the connection (i.e., port name to signal name) are explicit. Each connection is
separated by a comma. The syntax for explicit port mapping is as follows:
module_name <instance identifier> (.port_name1(signal1), .port_name2(signal2),
etc.);

Example 5.7 shows how to design a Verilog model of a hierarchical system that consists of two
lower-level modules.
 5.6 Structural Design and Hierarchy • 171

Example 5.7
Verilog structural design using explicit port mapping

5.6.1.2 Positional Port Mapping

In positional port mapping the names of the ports of the lower-level modules are not explicitly listed.
Instead, the signals to be connected to the lower-level system are listed in the same order in which the
ports were defined in the sub-system. Each signal name is separated by a comma. This approach
requires less text to describe the connection, but can also lead to misconnections due to inadvertent
mistakes in the signal order. The syntax for positional port mapping is as follows:

module_name : <instance_identifier> (signal1, signal2, etc.);

Example 5.8 shows how to create the same structural Verilog model as in Example 5.7, but using
positional port mapping instead.
172 • Chapter 5: Verilog (Part 1)

Example 5.8
Verilog structural design using positional port mapping

5.6.2 Gate Level Primitives

Verilog provides the ability to model basic logic functionality through the use of primitives. A primitive
is a logic operation that is simple enough that it doesn’t require explicit modeling. An example of this
behavior can be a basic logic gate or even a truth table. Verilog provides a set of gate level primitives to
model simple logic operations. These gate level primitives are not(), and(), nand(), or(), nor(), xor(), and
xnor(). Each of these primitives are instantiated as lower-level sub-systems with positional port
mapping. The port order for each primitive has the output listed first followed by the input(s). The output
and each of the inputs are scalars. Gate level primitives do not need to explicitly created as they are
provided as part of the Verilog standard. One of the benefits of using gate level primitives is that the
number of inputs is easily scaled as each primitive can accommodate an increasing number of inputs
automatically. Furthermore, modeling using this approach essentially provides a gate-level netlist, so it
represents a very low-level, detailed gate level implementation that is ready for technology mapping.
Example 5.9 shows how to use gate level primitives to model the behavior of a combinational logic
circuit.
 5.6 Structural Design and Hierarchy • 173

Example 5.9
Modeling combinational logic circuits using gate level primitives

5.6.3 User-Defined Primitives

A user-defined primitive (UDP) is a system that describes the behavior of a low-level component
using a logic table. This is very useful for creating combinational logic functionality that will be used
numerous times. UDPs are also useful for large truth tables where it is more convenient to list the
functionality in table form. UDPs are lower-level sub-systems that are intended to be instantiated in
higher-level modules just like gate-level primitives, with the exception that the UPD needs to be created
in its own file. The syntax for a UDP is as follows:

primitive primitive_name (output output_name,

input input_name1, input_name2, ...);
table
in1_val in2_val ... : out_val;
in1_val in2_val ... : out_val;
:
endtable
endprimitive

A UDP must list its output(s) first in the port definition. It also does not require types to be defined for
the ports. For combinational logic UDPs, all ports are assumed to be of type wire. Example 5.10 shows
how to design a user-defined primitive to implement a combinational logic circuit.
174 • Chapter 5: Verilog (Part 1)

Example 5.10
Modeling combinational logic circuits with a user-defined primitive

5.6.4 Adding Delay to Primitives

Delay can be added to primitives using the same approach as described in Sect. 5.5.4. The delay is
inserted after the primitive name but before the instance name.

Example:
not #2 U0 (An, A); // Gate level primitive for an inverter with delay
of 2.
and #3 U3 (m0, An, Bn, Cn); // Gate level primitive for an AND gate with delay
of 3.
SystemX_UDP #1 U0 (F, A, B, C); // UDP with a delay of 1.
 5.7 Overview of Simulation Test Benches • 175

CONCEPT CHECK

CC5.6 Does the use of lower-level sub-modules model concurrent functionality? Why?
(A) No. Since the lower-level behavior of the module being instantiated may contain
non-concurrent behavior, it is not known what functionality will be modeled.
(B) Yes. The modules are treated like independent sub-systems whose behavior runs
in parallel just as if separate parts were placed in a design.

5.7 Overview of Simulation Test Benches

One of the essential components of the modern digital design flow is verifying functionality through
simulation. This simulation takes place at many levels of abstraction. For a system to be tested, there
needs to be a mechanism to generate input patterns to drive the system and then observe the outputs to
verify correct operation. The mechanism to do this in Verilog is called a test bench. A test bench is a file in
Verilog that has no inputs or outputs. The test bench instantiates the system to be tested as a lower-level
module. The test bench generates the input conditions and drives them into the input ports of the system
being tested. Verilog contains numerous methods to generate stimulus patterns. Since a test bench will not
be synthesized, very abstract behavioral modeling can be used to generate the inputs. The output of the
system can be viewed as a waveform in a simulation tool. Verilog also has the ability to check the outputs
against expected results and notify the user if differences occur. Figure 5.10 gives an overview of how test
benches are used in Verilog. The techniques to generate the stimulus patterns are covered in Chap. 8.

Fig. 5.10
Overview of Verilog test benches
176 • Chapter 5: Verilog (Part 1)

CONCEPT CHECK

CC5.7 How can the output of a DUT be verified when it is connected to a signal that does not go
anywhere?
(A) It can’t. The output must be routed to an output port on the test bench.
(B) The values of any dangling signal are automatically written to a text file.
(C) It is viewed in the logic simulator as either a waveform or text listing.
(D) It can’t. A signal that does not go anywhere will cause an error when the Verilog file
is compiled.

Summary
v The modern digital design flow relies on com- v Instantiating other modules from within a
puter aided engineering (CAE) and computer higher-level module is how Verilog
aided design (CAD) tools to manage the size implements hierarchy. A lower-level module
and complexity of today’s digital designs. can be instantiated as many times as
v Hardware description languages (HDLs) desired. An instance identifier is useful is
allow the functionality of digital systems to keeping track of each instantiation. The
be entered using text. VHDL and Verilog are ports of the component can be connected
the two most common HDLs in use today. using either explicit or positional port
v In the 1980’s, two major HDLs emerged, mapping.
VHDL and Verilog. VHDL was sponsored by v Concurrency is the term that describes
the Department of Defense while Verilog was operations being performed in parallel. This
driven by the commercial industry. Both were allows real-world system behavior to be
later standardized by IEEE. modeled.
v The ability to automatically synthesize a logic v Verilog provides the continuous assignment
circuit from a Verilog behavioral description operator to support modeling concurrent
became possible approximately 10 years systems. Complex logic circuits can be
after the original definition of Verilog. As implemented by using continuous assign-
such, only a sub-set of the behavioral ment with logical operators or conditional
modeling techniques in Verilog can be auto- operators.
matically synthesized. v Verilog sub-systems are also treated as con-
v HDLs can model digital systems at different current sub-systems.
levels of design abstraction. These include v Delay can be modeled in Verilog for all
the system, algorithmic, RTL, gate, and cir- transitions, or for individual transitions (rise,
cuit levels. Designing at a higher level of fall, off). A range of delays can also be
abstraction allows more complex systems to provided (min:typ:max). Delay can be
be modeled without worrying about the added to continuous assignments and
details of the implementation. sub-system instantiations.
v In a Verilog source file, all functionality is v Gate level primitives are provided in Verilog
contained within a module. The first portion to implement basic logic functions (not, and,
of the module is the port definition. The sec- nand, or, nor, xor, xnor). These primitives are
ond portion contains declarations of internal instantiated just like any other lower-level
signals/constants/parameters. The third por- sub-system.
tion contains the description of the behavior. v User Defined Primitives are supported in
v A port is an input or output to a system that is Verilog that allow the functionality of a circuit
defined as part of the initial module state- to be described in table form.
ment. A signal, or net, is an internal connec- v A simulation test bench is a Verilog file that
tion within the system that is declared inside drives stimulus into a device under test
of the module. A signal is not visible outside (DUT). Test benches do not have inputs or
of the system. outputs and are not synthesizable.
 Exercise Problems • 177

Exercise Problems
Section 5.1: History of HDLs 5.3.2 Which step in the modern digital design flow
does the following statement relate to: after
5.1.1 What was the original purpose of Verilog?
realistic gate and wiring delays are determined,
5.1.2 Can all of the functionality that can be one last simulation should be performed to
described in Verilog be simulated? make sure the design meets the original timing
5.1.3 Can all of the functionality that can be requirements.
described in Verilog be synthesized? 5.3.3 Which step in the modern digital design flow
does the following statement relate to: if the
Section 5.2: HDL Abstraction memory is distributed around the perimeter of
the CPU, the wiring density will be minimized.
5.2.1 Give the level of design abstraction that the
following statement relates to: if there is ever 5.3.4 Which step in the modern digital design flow
an error in the system, it should return to the does the following statement relate to: the
reset state. design meets all requirements so now I’m
building the hardware that will be shipped.
5.2.2 Give the level of design abstraction that the
following statement relates to: once the design 5.3.5 Which step in the modern digital design flow
is implemented in a sum of products form, does the following statement relate to: the sys-
DeMorgan’s Theorem will be used to convert tem will be broken down into three
it to a NAND-gate only implementation. sub-systems with the following behaviors.
5.2.3 Give the level of design abstraction that the 5.3.6 Which step in the modern digital design flow
following statement relates to: the design will does the following statement relate to: this sys-
be broken down into two sub-systems, one that tem needs to have 10 Gbytes of memory.
will handle data collection and the other that 5.3.7 Which step in the modern digital design flow
will control data flow. does the following statement relate to: to meet
5.2.4 Give the level of design abstraction that the the power requirements, the gates will be
following statement relates to: the interconnect implemented in the 74HC logic family.
on the IC should be changed from aluminum to
copper to achieve the performance needed in Section 5.4: Verilog Constructs
this design.
5.4.1 What is the name of the main design unit in
5.2.5 Give the level of design abstraction that the Verilog?
following statement relates to: the MOSFETs
5.4.2 What portion of the Verilog module describes
need to be able to drive at least 8 other loads in
the inputs and outputs.
this design.
5.4.3 What step is necessary if a system requires
5.2.6 Give the level of design abstraction that the
internal connections?
following statement relates to: this system will
contain 1 host computer and support up to 5.4.4 What are all the possible values that a Verilog
1000 client computers. net type can take on?
5.2.7 Give the design domain that the following activ- 5.4.5 What is the highest strength that a value can
ity relates to: drawing the physical layout of the take on in Verilog.
CPU will require six months of engineering 5.4.6 What is the range of decimal numbers that can
time. be represented using the type integer in
5.2.8 Give the design domain that the following activ- Verilog?
ity relates to: the CPU will be connected to 5.4.7 What is the width of the vector defined using
4 banks of memory. the type [63:0] wire?
5.2.9 Give the design domain that the following activ- 5.4.8 What is the syntax for indexing the most signif-
ity relates to: the fan-in specifications for this icant bit in the type [31:0] wire? Assume the
logic family require excessive logic circuitry to vector is named example.
be used.
5.4.9 What is the syntax for indexing the least signif-
5.2.10 Give the design domain that the following activ- icant bit in the type [31:0] wire? Assume the
ity relates to: the performance specifications vector is named example.
for this system require 1 TFLOP at < 5W.
5.4.10 What is the difference between a wire and reg
type?
Section 5.3: The Modern Digital 5.4.11 How many bits is the type integer by default?
Design Flow 5.4.12 How many bits is the type real by default?
5.3.1 Which step in the modern digital design flow
does the following statement relate to: a CAD
tool will convert the behavioral model into a
gate-level description of functionality.
178 • Chapter 5: Verilog (Part 1)

Section 5.5: Modeling Concurrent Func-

tionality in Verilog
5.5.1 Design a Verilog model to implement the
behavior described by the 3-input minterm list
shown in Fig. 5.11. Use continuous assign-
ment with logical operators. Declare your mod-
ule and ports to match the block diagram
provided. Use the type wire for your ports.

Fig. 5.13
System G functionality

5.5.6 Design a Verilog model to implement the

behavior described by the 3-input truth table
Fig. 5.11 shown in Fig. 5.13. Use continuous assign-
ment with conditional operators. Declare your
System E functionality module and ports to match the block diagram
provided. Use the type wire for your ports.
5.5.2 Design a Verilog model to implement the
behavior described by the 3-input minterm list 5.5.7 Design a Verilog model to implement the
shown in Fig. 5.11. Use continuous assign- behavior described by the 4-input minterm list
ment with conditional operators. Declare your shown in Fig. 5.14. Use continuous assign-
module and ports to match the block diagram ment and logical operators. Declare your mod-
provided. Use the type wire for your ports. ule and ports to match the block diagram
provided. Use the type wire for your ports.
5.5.3 Design a Verilog model to implement the
behavior described by the 3-input maxterm
list shown in Fig. 5.12. Use continuous assign-
ment with logical operators. Declare your mod-
ule and ports to match the block diagram
provided. Use the type wire for your ports.

Fig. 5.14
System I functionality

5.5.8 Design a Verilog model to implement the

behavior described by the 4-input minterm list
Fig. 5.12
shown in Fig. 5.14. Use continuous assign-
System F functionality ment and conditional operators. Declare your
module and ports to match the block diagram
5.5.4 Design a Verilog model to implement the provided. Use the type wire for your ports.
behavior described by the 3-input maxterm 5.5.9 Design a Verilog model to implement the
list shown in Fig. 5.12. Use continuous assign- behavior described by the 4-input maxterm
ment with conditional operators. Declare your list shown in Fig. 5.15. Use continuous assign-
module and ports to match the block diagram ment and logical operators. Declare your mod-
provided. Use the type wire for your ports. ule and ports to match the block diagram
5.5.5 Design a Verilog model to implement the provided. Use the type wire for your ports.
behavior described by the 3-input truth table
shown in Fig. 5.13. Use continuous assign-
ment with logical operators. Declare your mod-
ule and ports to match the block diagram
provided. Use the type wire for your ports.

Fig. 5.15
System J functionality
 Exercise Problems • 179

5.5.10 Design a Verilog model to implement the This is considered structural because you will
behavior described by the 4-input maxterm need to instantiate the user defined primitive
list shown in Fig. 5.15. Use continuous assign- just like a traditional sub-system. You will need
ment and conditional operators. Declare your to create both the upper level module and the
module and ports to match the block diagram lower-level UDP. Declare your module and
provided. Use the type wire for your ports. ports to match the block diagram provided.
5.5.11 Design a Verilog model to implement the Use the type wire for your ports.
behavior described by the 4-input truth table 5.6.3 Design a Verilog model to implement the
shown in Fig. 5.16. Use continuous assign- behavior described by the 3-input maxterm
ment and logical operators. Declare your mod- list shown in Fig. 5.12. Use a structural design
ule and ports to match the block diagram approach based on gate level primitives. This
provided. Use the type wire for your ports. is considered structural because you will need
to instantiate the gate level primitives just like a
traditional sub-system; however, you don’t
need to create the gate level modules as they
are already built into the Verilog standard. You
will need to determine a logic expression for
the system prior to connecting the gate level
primitives. You can use whatever approach
you prefer to create the logic expression (i.e.,
canonical SOP/POS, minimized SOP/POS,
etc.). Declare your module and ports to match
the block diagram provided. Use the type wire
for your ports.
5.6.4 Design a Verilog model to implement the
behavior described by the 3-input maxterm
list shown in Fig. 5.12. Use a structural design
approach based on a user defined primitive.
This is considered structural because you will
need to instantiate the user defined primitive
Fig. 5.16 just like a traditional sub-system. You will need
System K functionality to create both the upper level module and the
lower-level UDP. Declare your module and
ports to match the block diagram provided.
5.5.12 Design a Verilog model to implement the Use the type wire for your ports.
behavior described by the 4-input truth table
shown in Fig. 5.16. Use continuous assign- 5.6.5 Design a Verilog model to implement the
ment and conditional operators. Declare your behavior described by the 3-input truth table
module and ports to match the block diagram shown in Fig. 5.13. Use a structural design
provided. Use the type wire for your ports. approach based on gate level primitives. This
is considered structural because you will need
to instantiate the gate level primitives just like a
Section 5.6: Structural Design in Verilog traditional sub-system; however, you don’t
5.6.1 Design a Verilog model to implement the need to create the gate level modules as they
behavior described by the 3-input minterm list are already built into the Verilog standard. You
shown in Fig. 5.11. Use a structural design will need to determine a logic expression for
approach based on gate level primitives. This the system prior to connecting the gate level
is considered structural because you will need primitives. You can use whatever approach
to instantiate the gate level primitives just like a you prefer to create the logic expression (i.e.,
traditional sub-system; however, you don’t canonical SOP/POS, minimized SOP/POS,
need to create the gate level modules as they etc.). Declare your module and ports to match
are already built into the Verilog standard. You the block diagram provided. Use the type wire
will need to determine a logic expression for for your ports.
the system prior to connecting the gate level 5.6.6 Design a Verilog model to implement the
primitives. You can use whatever approach behavior described by the 3-input truth table
you prefer to create the logic expression (i.e., shown in Fig. 5.13. Use a structural design
canonical SOP/POS, minimized SOP/POS, approach based on a user defined primitive.
etc.). Declare your module and ports to match This is considered structural because you will
the block diagram provided. Use the type wire need to instantiate the user defined primitive
for your ports. just like a traditional sub-system. You will need
5.6.2 Design a Verilog model to implement the to create both the upper level module and the
behavior described by the 3-input minterm list lower-level UDP. Declare your module and
shown in Fig. 5.11. Use a structural design ports to match the block diagram provided.
approach based on a user defined primitive. Use the type wire for your ports.
180 • Chapter 5: Verilog (Part 1)

5.6.7 Design a Verilog model to implement the list shown in Fig. 5.15. Use a structural design
behavior described by the 4-input minterm list approach based on a user defined primitive.
shown in Fig. 5.14. Use a structural design This is considered structural because you will
approach based on gate level primitives. This need to instantiate the user defined primitive
is considered structural because you will need just like a traditional sub-system. You will need
to instantiate the gate level primitives just like a to create both the upper level module and the
traditional sub-system; however, you don’t lower-level UDP. Declare your module and
need to create the gate level modules as they ports to match the block diagram provided.
are already built into the Verilog standard. You Use the type wire for your ports.
will need to determine a logic expression for 5.6.11 Design a Verilog model to implement the
the system prior to connecting the gate level behavior described by the 4-input truth table
primitives. You can use whatever approach shown in Fig. 5.16. Use a structural design
you prefer to create the logic expression (i.e., approach based on gate level primitives. This
canonical SOP/POS, minimized SOP/POS, is considered structural because you will need
etc.). Declare your module and ports to match to instantiate the gate level primitives just like a
the block diagram provided. Use the type wire traditional sub-system; however, you don’t
for your ports. need to create the gate level modules as they
5.6.8 Design a Verilog model to implement the are already built into the Verilog standard. You
behavior described by the 4-input minterm list will need to determine a logic expression for
shown in Fig. 5.14. Use a structural design the system prior to connecting the gate level
approach based on a user defined primitive. primitives. You can use whatever approach
This is considered structural because you will you prefer to create the logic expression (i.e.,
need to instantiate the user defined primitive canonical SOP/POS, minimized SOP/POS,
just like a traditional sub-system. You will need etc.). Declare your module and ports to match
to create both the upper level module and the the block diagram provided. Use the type wire
lower-level UDP. Declare your module and for your ports.
ports to match the block diagram provided. 5.6.12 Design a Verilog model to implement the
Use the type wire for your ports. behavior described by the 4-input truth table
5.6.9 Design a Verilog model to implement the shown in Fig. 5.16. Use a structural design
behavior described by the 4-input maxterm approach based on a user defined primitive.
list shown in Fig. 5.15. Use a structural design This is considered structural because you will
approach based on gate level primitives. This need to instantiate the user defined primitive
is considered structural because you will need just like a traditional sub-system. You will need
to instantiate the gate level primitives just like a to create both the upper level module and the
traditional sub-system; however, you don’t lower-level UDP. Declare your module and
need to create the gate level modules as they ports to match the block diagram provided.
are already built into the Verilog standard. You Use the type wire for your ports.
will need to determine a logic expression for
the system prior to connecting the gate level Section 5.7: Overview of Simulation Test
primitives. You can use whatever approach
you prefer to create the logic expression (i.e., Benches
canonical SOP/POS, minimized SOP/POS, 5.7.1 What is the purpose of a test bench?
etc.). Declare your module and ports to match 5.7.2 Does a test bench have input and output ports?
the block diagram provided. Use the type wire
for your ports. 5.7.3 Can a test bench be simulated?
5.6.10 Design a Verilog model to implement the 5.7.4 Can a test bench be synthesized?
behavior described by the 4-input maxterm
Chapter 6: MSI Logic
This chapter introduces a group of combinational logic building blocks that are commonly used in
digital design. As we move into systems that are larger than individual gates, there are naming
conventions that are used to describe the size of the logic. Table 6.1 gives these naming conventions.
In this chapter we will look at medium scale integrated circuit (MSI) logic. Each of these building blocks
can be implemented using the combinational logic design steps covered in Chaps. 4 and 5. The goal of
this chapter is to provide an understanding of the basic principles of MSI logic.

Table 6.1
Naming convention for the size of digital systems

Learning Outcomes—After completing this chapter, you will be able to:

6.1 Design a decoder circuit using both the classical digital design approach and the modern
HDL-based approach.
6.2 Design an encoder circuit using both the classical digital design approach and the modern
HDL-based approach.
6.3 Design a multiplexer circuit using both the classical digital design approach and the
modern HDL-based approach.
6.4 Design a demultiplexer circuit using both the classical digital design approach and the
modern HDL-based approach.

6.1 Decoders
A decoder is a circuit that takes in a binary code and has outputs that are asserted for specific values
of that code. The code can be of any type or size (e.g., unsigned, two’s complement, etc.). Each output
will assert for only specific input codes. Since combinational logic circuits only produce a single output,
this means that within a decoder, there will be a separate combinational logic circuit for each output.

6.1.1 Example: One-Hot Decoder

A one-hot decoder is a circuit that has n inputs and 2n outputs. Each output will assert for one and
only one input code. Since there are 2n outputs, there will always be one and only one output asserted at

# Springer International Publishing AG 2017 181

B.J. LaMeres, Introduction to Logic Circuits & Logic Design with Verilog,
DOI 10.1007/978-3-319-53883-9_6
182 • Chapter 6: MSI Logic

any given time. Example 6.1 shows the process of designing a 2-to-4 one-hot decoder by hand
(i.e., using the classical digital design approach).

Example 6.1
2-to-4 One-hot decoder – logic synthesis by hand

As decoders get larger, it is necessary to use hardware description languages to model their
behavior. Example 6.2 shows how to model a 3-to-8 one-hot decoder in Verilog with continuous
assignment and logic operators.
 6.1 Decoders • 183

Example 6.2
3-to-8 One-hot decoder – Verilog modeling using logical Operators

This description can be further simplified by using vector notation for the ports and describing the
functionality using conditional operators. Example 6.3 shows how to model the 3-to-8 one-hot decoder in
Verilog using continuous assignment with conditional operators.
184 • Chapter 6: MSI Logic

Example 6.3
3-to-8 One-hot decoder – Verilog modeling using conditional operators

6.1.2 Example: 7-Segment Display Decoder

A 7-segment display decoder is a circuit used to drive character displays that are commonly found in
applications such as digital clocks and household appliances. A character display is made up of
7 individual LEDs, typically labeled a-g. The input to the decoder is the binary equivalent of the decimal
or Hex character that is to be displayed. The output of the decoder is the arrangement of LEDs that will
form the character. Decoders with 2-inputs can drive characters “0” to “3”. Decoders with 3-inputs can
drive characters “0” to “7”. Decoders with 4-inputs can drive characters “0” to “F” with the case of the Hex
characters being “A, b, c or C, d, E and F”.
Let’s look at an example of how to design a 3-input, 7-segment decoder by hand. The first step in the
process is to create the truth table for the outputs that will drive the LEDs in the display. We’ll call these
outputs Fa, Fb, . . ., Fg. Example 6.4 shows how to construct the truth table for the 7-segment display
decoder. In this table, a logic 1 corresponds to the LED being ON.
 6.1 Decoders • 185

Example 6.4
7-Segment display decoder – truth table

If we wish to design this decoder by hand we need to create seven separate combinational logic
circuits. Each of the outputs (Fa – Fg) can be put into a 3-input K-map to find the minimized logic
expression. Example 6.5 shows the design of the decoder from the truth table in Example 6.4 by hand.
186 • Chapter 6: MSI Logic

Example 6.5
7-Segment display decoder – logic synthesis by hand

This same functionality can be implemented in Verilog using concurrent modeling techniques.
Example 6.6 shows how to model the 7-segment decoder in Verilog using continuous assignment with
logic operators.
 6.1 Decoders • 187

Example 6.6
7-Segment display decoder – Verilog modeling using logical operators

Again, a more compact description of the decoder can be accomplished if the ports are described
as vectors and a conditional operator is used. Example 6.7 shows how to model the 7-segment decoder
in Verilog using continuous assignment with conditional operators.

Example 6.7
7-Segment display decoder – Verilog modeling using conditional operators
188 • Chapter 6: MSI Logic

CONCEPT CHECK

CC6.1 In a decoder, a logic expression is created for each output. Once all of the output logic
expressions are found, how can the decoder logic be further minimized?
(A) By using K-maps to find the output logic expressions.
(B) By buffering the inputs so that they can drive a large number of other gates.
(C) By identifying any logic terms that are used in multiple locations (inversions,
product terms, and sum terms) and sharing the interim results among multiple
circuits in the decoder.
(D) By ignoring fan-out.

6.2 Encoders
An encoder works in the opposite manner as a decoder. An assertion on a specific input port
corresponds to a unique code on the output port.

6.2.1 Example: One-Hot Binary Encoder

A one-hot binary encoder has n outputs and 2n inputs. The output will be an n-bit, binary code which
corresponds to an assertion on one and only one of the inputs. Example 6.8 shows the process of
designing a 4-to-2 binary encoder by hand (i.e., using the classical digital design approach).
 6.2 Encoders • 189

Example 6.8
4-to-2 Binary encoder – logic synthesis by hand

In Verilog, an encoder can be implemented using continuous assignment with either logical or
conditional operators. Example 6.9 shows how to model the encoder in Verilog using these techniques.
190 • Chapter 6: MSI Logic

Example 6.9
4-to-2 Binary encoder – Verilog modeling using logical and conditional operators

CONCEPT CHECK

CC6.2 If it is desired to have the outputs of an encoder produce 0’s for all input codes not defined in
the truth table, can “don’t cares” be used when deriving the minimized logic expressions?
Why?
(A) No. Don’t cares aren’t used in encoders.
(B) Yes. Don’t cares can always be used in K-maps.
(C) Yes. All that needs to be done is to treat each X as a 0 when forming the most
minimal prime implicant.
(D) No. Each cell in the K-map corresponding to an undefined input code needs to
contain a 0 so don’t cares are not applicable.

6.3 Multiplexers
A multiplexer is a circuit that passes one of its multiple inputs to a single output based on a select
input. This can be thought of as a digital switch. The multiplexer has n select lines, 2n inputs, and one
output. Example 6.10 shows the process of designing a 2-to-1 multiplexer by hand (i.e., using the
classical digital design approach).
 6.3 Multiplexers • 191

Example 6.10
2-to-1 Multiplexer – logic synthesis by hand

In Verilog, a multiplexer can be implemented using continuous assignment with either logical or
conditional operators. Example 6.11 shows how to model the multiplexer in Verilog using these
techniques.
192 • Chapter 6: MSI Logic

Example 6.11
4-to-1 Multiplexer – Verilog modeling using logical and conditional operators

CONCEPT CHECK

CC6.3 How are the product terms in a multiplexer based on the identity theorem?
(A) Only the select product term will pass its input to the final sum term. Since all of the
unselected product terms output 0, the input will be passed through the sum term
because anything OR’d with a 0 is itself.
(B) The select lines are complemented such that they activate only one OR gate.
(C) The select line inputs will produce 1’s on the inputs of the selected product term.
This allows the input signal to pass through the selected AND gate because
anything AND’d with a 1 is itself.
(D) The select line inputs will produce 0’s on the inputs of the selected sum term. This
allows the input signal to pass through the selected OR gate because anything
OR’d with a 0 is itself.
 6.4 Demultiplexers • 193

6.4 Demultiplexers
A demultiplexer works in a complementary fashion to a multiplexer. A demultiplexer has one input
that is routed to one of its multiple outputs. The output that is active is dictated by a select input. A demux
has n select lines that chooses to route the input to one of its 2n outputs. When an output is not selected,
it outputs a logic 0. Example 6.12 shows the process of designing a 1-to-2 demultiplexer by hand
(i.e., using the classical digital design approach).

Example 6.12
1-to-2 Demultiplexer – logic synthesis by hand

In Verilog, a demultiplexer can be implemented using continuous assignment with either logical or
conditional operators. Example 6.13 shows how to model the demultiplexer in Verilog using these
techniques
194 • Chapter 6: MSI Logic

Example 6.13
1-to-4 Demultiplexer – Verilog modeling using logical and conditional operators

CONCEPT CHECK

CC6.4 How many select lines are needed in a 1-to-64 demultiplexer?

(A) 1 (B) 4 (C) 6 (D) 64

Summary
v The term medium scale integrated circuit v A decoder is a system that has a greater
(MSI) logic refers to a set of basic combina- number of outputs than inputs. The behavior
tional logic circuits that implement simple, of each output is based on each unique
commonly used functions such as decoders, input code.
encoders, multiplexers, and demultiplexers. v An encoder a system that has a greater num-
MSI logic can also include operations such ber of inputs than outputs. A compressed
as comparators and simple arithmetic output code is produced based on which
circuits. input(s) lines are asserted.
v While an MSI logic circuit may have multiple v A multiplexer is a system that has one output
outputs, each output requires its own unique and multiple inputs. At any given time, one
logic expression that is based on the system and only one input is routed to the output
inputs. based on the value on a set of select lines.
 Exercise Problems • 195

For n select lines, a multiplexer can support 2n v HDLs are particularly useful for describing
inputs. MSI logic due to their abstract modeling
v A demultiplexer is a system that has one capability. Through the use of Boolean
input and multiple outputs. The input is conditions and vector assignments, the
routed to one of the outputs depending on behavior of MSI logic can be modeled in a
the value on a set of select lines. For n select compact and intuitive manner.
lines, a demultiplexer can support 2n outputs.

Exercise Problems
Section 6.1: Decoders 6.1.2 Design a Verilog model for a 4-to-16 one-hot
decoder using continuous assignment and
6.1.1 Design a 4-to-16 one-hot decoder by hand.
gate level primitives. Use the module port defi-
The block diagram and truth table for the
nition given in Fig. 6.2.
decoder are given in Fig. 6.1. Give the
minimized logic expressions for each output
(i.e., F0, F1, . . ., F15) and the full logic diagram
for the system.

Fig. 6.2
4-to-16 one-hot module definition

6.1.3 Design a Verilog model for a 4-to-16 one-hot

decoder using continuous assignment and
logical operators. Use the module port defi-
nition given in Fig. 6.2.
6.1.4 Design a Verilog model for a 4-to-16 one-hot
decoder using continuous assignment and
conditional operators. Use the module port
definition given in Fig. 6.2.
6.1.5 Design a 4-input, 7-segment HEX character
decoder by hand. The system has four inputs
called A, B, C and D. The system has 7 outputs
called Fa, Fb, Fc, Fd, Fe, Ff, and Fg. These
outputs drive the individual LEDs within the
display. A logic 1 on an output corresponds to
the LED being ON. The display will show the
HEX characters 0–9, A, b, c, d, E, and F
corresponding to the 4-bit input code on A. A
template for creating the truth tables for this
system is provided in Fig. 6.3. Provide the
minimized logic expressions for each of the
seven outputs and the overall logic diagram
for the decoder.
Fig. 6.1
4-to-16 one-hot decoder functionality
196 • Chapter 6: MSI Logic

Fig. 6.3
7-segment display decoder truth table
 Exercise Problems • 197

6.1.6 Design a Verilog model for a 4-input, 7-seg- for each output and the full logic diagram for
ment HEX character decoder using continuous the system.
assignment and logical operators. Use the
module port definition given in Fig. 6.4 for
your design. The system has a 4-bit input
vector called ABCD and a 7-bit output vector
called F. The individual scalars within the out-
put vector (i.e., F[6:0]) correspond to the char-
acter display segments a, b, c, d, e, f, and g
respectively. A logic 1 on an output
corresponds to the LED being ON. The display
will show the HEX characters 0–9, A, b, c, d, E,
and F corresponding to the 4-bit input code on
A. A template for creating the truth table is
provided in Fig. 6.3. The signals in this table
correspond to the ports in this problem as
follows: Fa ¼ F(6), Fb ¼ F(5), Fc ¼ F(4), Fig. 6.5
Fd ¼ F(3), Fe ¼ F(2), Ff ¼ F(1), and Fg ¼ F(0). 8-to-3 one-hot encoder functionality

6.2.2 Design a Verilog model for an 8-to-3 binary

encoder using continuous assignment and
gate level primitives. Use the module port defi-
nition given in Fig. 6.6.
Fig. 6.4
7-segment display decoder module definition

6.1.7 Design a Verilog model for a 4-input, 7-seg-

ment HEX character decoder using continuous
assignment and conditional operators. Use the
module port definition given in Fig. 6.4 for
your design. The system has a 4-bit input Fig. 6.6
vector called ABCD and a 7-bit output vector 8-to-3 one-hot encoder module definition
called F. The individual scalars within the out-
put vector (i.e., F[6:0]) correspond to the char-
6.2.3 Design a Verilog model for an 8-to-3 binary
acter display segments a, b, c, d, e, f, and g
encoder using continuous assignment and log-
respectively. A logic 1 on an output
ical operators. Use the module port definition
corresponds to the LED being ON. The display
given in Fig. 6.6.
will show the HEX characters 0–9, A, b, c, d, E,
and F corresponding to the 4-bit input code on 6.2.4 Design a Verilog model for an 8-to-3 binary
A. A template for creating the truth table is encoder using continuous assignment and
provided in Fig. 6.3. The signals in this table conditional operators. Use the module port def-
correspond to the ports in this problem as inition given in Fig. 6.6.
follows: Fa ¼ F(6), Fb ¼ F(5), Fc ¼ F(4),
Fd ¼ F(3), Fe ¼ F(2), Ff ¼ F(1), and Fg ¼ F(0). Section 6.3: Multiplexers
6.3.1 Design an 8-to-1 multiplexer by hand. The
Section 6.2: Encoders block diagram and truth table for the multi-
6.2.1 Design an 8-to-3 binary encoder by hand. The plexer are given in Fig. 6.7. Give the minimized
block diagram and truth table for the encoder logic expressions for the output and the full
are given in Fig. 6.5. Give the logic expressions logic diagram for the system.
198 • Chapter 6: MSI Logic

Fig. 6.7
8-to-1 multiplexer functionality

6.3.2 Design a Verilog model for an 8-to-1 multi-

plexer using continuous assignment and gate
level primitives. Use the module port definition
given in Fig. 6.8.

Fig. 6.9
1-to-8 demultiplexer functionality

Fig. 6.8 6.4.2 Design a Verilog model for a 1-to-8 demulti-

plexer using continuous assignment and gate
8-to-1 multiplexer module definition
level primitives. Use the module port definition
given in Fig. 6.10 for your design.
6.3.3 Design a Verilog model for an 8-to-1 multi-
plexer using continuous assignment and logi-
cal operators. Use the module port definition
given in Fig. 6.8.
6.3.4 Design a Verilog model for an 8-to-1 multi-
plexer using continuous assignment and con-
ditional operators. Use the module port
definition given in Fig. 6.8.

Section 6.4: Demultiplexers Fig. 6.10

1-to-8 demultiplexer module definition
6.4.1 Design a 1-to-8 demultiplexer by hand. The
block diagram and truth table for the demulti-
plexer are given in Fig. 6.9. Give the minimized 6.4.3 Design a Verilog model for a 1-to-8 demulti-
logic expressions for each output and the full plexer using continuous assignment and logi-
logic diagram for the system. cal operators. Use the module port definition
given in Fig. 6.10 for your design.
6.4.4 Design a Verilog model for a 1-to-8 demulti-
plexer using continuous assignment and con-
ditional operators. Use the module port
definition given in Fig. 6.10 for your design.
Chapter 7: Sequential Logic Design
In this chapter we begin looking at sequential logic design. Sequential logic design differs from
combinational logic design in that the outputs of the circuit depend not only on the current values of the
inputs but also on the past values of the inputs. This is different from the combinational logic design
where the output of the circuitry depends only on the current values of the inputs. The ability of a
sequential logic circuit to base its outputs on both the current and past inputs allows more sophisticated
and intelligent systems to be created. We begin by looking at sequential logic storage devices, which are
used to hold the past values of a system. This is followed by an investigation of timing considerations of
sequential logic circuits. We then look at some useful circuits that can be created using only sequential
logic storage devices. Finally, we look at one of the most important logic circuits in digital systems, the
finite state machine. The goal of this chapter is to provide an understanding of the basic operation of
sequential logic circuits.

Learning Outcomes—After completing this chapter, you will be able to:

7.1 Describe the operation of a sequential logic storage device.
7.2 Describe sequential logic timing considerations.
7.3 Design a variety of common circuits based on sequential storage devices (toggle flops,
ripple counters, switch debouncers, and shift registers).
7.4 Design a finite state machine using the classical digital design approach.
7.5 Design a counter using the classical digital design approach and using an HDL-based,
structural approach.
7.6 Describe the finite state machine reset condition.
7.7 Analyze a finite state machine to determine its functional operation and maximum clock
frequency.

7.1 Sequential Logic Storage Devices

7.1.1 The Cross-Coupled Inverter Pair
The first thing that is needed in sequential logic is a storage device. The fundamental storage device
in sequential logic is based on a positive feedback configuration. Consider the circuit in Fig. 7.1. This
circuit configuration is called the cross-coupled inverter pair. In this circuit if the input of U1 starts with a
value of 1, it will produce an output of Q ¼ 0. This output is fed back to the input of U2, thus producing an
output of Qn ¼ 1. Qn is fed back to the original input of U1, thus reinforcing the initial condition. This
circuit will hold, or store, a logic 0 without being driven by any other inputs. This circuit operates in a
complementary manner when the initial value of U1 is a 0. With this input condition, the circuit will store a
logic 1 without being driven by any other inputs.

# Springer International Publishing AG 2017 199

B.J. LaMeres, Introduction to Logic Circuits & Logic Design with Verilog,
DOI 10.1007/978-3-319-53883-9_7
200 • Chapter 7: Sequential Logic Design

Fig. 7.1
Storage using a cross-coupled inverter pair

7.1.2 Metastability
The cross-coupled inverter pair in Fig. 7.1 exhibits what is called metastable behavior due to its
positive feedback configuration. Metastability refers to when a system can exist in a state of equilibrium
when undisturbed but can be moved to a different, more stable state of equilibrium when sufficiently
disturbed. Systems that exhibit high levels of metastability have an equilibrium state that is highly
unstable, meaning that if disturbed even slightly the system will move rapidly to a more stable point of
equilibrium. The cross-coupled inverter pair is a highly metastable system. This system actually contains
three equilibrium states. The first is when the input of U1 is exactly between a logic 0 and logic 1 (i.e.,
VCC/2). In this state, the output of U1 is also exactly VCC/2. This voltage is fed back to the input of U2, thus
producing an output of exactly VCC/2 on U2. This in turn is fed back to the original input on U1 reinforcing
the initial state. Despite this system being at equilibrium in this condition, this state is highly unstable.
With minimal disturbance to any of the nodes within the system, it will move rapidly to one of two more
stable states. The two stable states for this system are when Q ¼ 0 or when Q ¼ 1 (see Fig. 7.1). Once
the transition begins between the unstable equilibrium state toward one of the two more stable states, the
positive feedback in the system continually reinforces the transition until the system reaches its final
state. In electrical systems, this initial disturbance is caused by the presence of noise, or unwanted
voltage in the system. Noise can come from many sources including random thermal motion of charge
carriers in the semiconductor materials, electromagnetic energy, or naturally occurring ionizing particles.
Noise is present in every electrical system so the cross-coupled inverter pair will never be able to stay in
the unstable equilibrium state where all nodes are at VCC/2.
The cross-coupled inverter pair has two stable states, thus it is called a bistable element. In order to
understand the bistable behavior of this circuit, let’s look at its behavior when the initial input value on U1
is set directly between a logic 0 and logic 1 (i.e., VCC/2) and how a small amount of noise will cause the
system to move toward a stable state. Recall that an inverter is designed to have an output that quickly
transitions between a logic LOW and HIGH in order to minimize the time spent in the uncertainty region.
This is accomplished by designing the inverter to have what is called gain. Gain can be thought of as a
multiplying factor that is applied to the input of the circuit when producing the output (i.e., Vout ¼ gain∙Vin).
The gain for an inverter will be negative since the output moves in the opposite direction of the input. The
inverter is designed to have a very high gain such that even the smallest change on the input when in the
transition region will result in a large change on the output. Consider the behavior of this circuit shown in
Fig. 7.2. In this example, let’s represent the gain of the inverter as –g and see how the system responds
when a small positive voltage noise (Vn) is added to the VCC/2 input on U1.
 7.1 Sequential Logic Storage Devices • 201

Fig. 7.2
Examining metastability moving toward the state Q ¼ 0

Figure 7.3 shows how the system responds when a small negative voltage noise (Vn) is added to
the VCC/2 input on U1.
202 • Chapter 7: Sequential Logic Design

Fig. 7.3
Examining metastability moving toward the state Q ¼ 1

7.1.3 The SR Latch

While the cross-coupled inverter pair is the fundamental storage concept for sequential logic, there
is no mechanism to set the initial value of Q. All that is guaranteed is that the circuit will store a value in
one of two stable states (Q ¼ 0 or Q ¼ 1). The SR Latch provides a means to control the initial values in
this positive feedback configuration by replacing the inverters with NOR gates. In this circuit, S stands for
set and indicates when the output is forced to a logic 1 (Q ¼ 1), and R stands for reset and indicates when
the output is forced to a logic 0 (Q ¼ 0). When both S ¼ 0 and R ¼ 0, the SR Latch is put into a store mode
and it will hold the last value of Q. In all of these input conditions, Qn is the complement of Q. Consider
the behavior of the SR Latch during its store state shown in Fig. 7.4.
 7.1 Sequential Logic Storage Devices • 203

Fig. 7.4
SR Latch behavior – store state (S ¼ 0, R ¼ 0)

The SR Latch has two input conditions that will force the outputs to known values. The first condition
is called the set state. In this state, the inputs are configured as S ¼ 1 and R ¼ 0. This input condition will
force the outputs to Q ¼ 1 (e.g. setting Q) and Qn ¼ 0. The second input condition is called the reset
state. In this state the inputs are configured as S ¼ 0 and R ¼ 1. This input condition will force the outputs
to Q ¼ 0 (i.e., resetting Q) and Qn ¼ 1. Consider the behavior of the SR Latch during its set and reset
states shown in Fig. 7.5.
204 • Chapter 7: Sequential Logic Design

Fig. 7.5
SR Latch behavior – set (S ¼ 1, R ¼ 0) and reset (S ¼ 0, R ¼ 1) states

The final input condition for the SR Latch leads to potential metastability and should be avoided.
When S ¼ 1 and R ¼ 1, the outputs of the SR Latch will both go to logic 0’s. The problem with this state is
that if the inputs subsequently change to the store state (S ¼ 0, R ¼ 0), the outputs will go metastable and
then settle in one of the two stable states (Q ¼ 0 or Q ¼ 1). The reason this state is avoided is because
the final resting state of the SR Latch is random and unknown. Consider this operation shown in Fig. 7.6.

Fig. 7.6
SR Latch behavior – don’t use state (S ¼ 1 and R ¼ 1)
 7.1 Sequential Logic Storage Devices • 205

Figure 7.7 shows the final truth table for the SR Latch.

Fig. 7.7
SR Latch truth table

The SR Latch has some drawbacks when it comes to implementation with real circuitry. First, it takes
two independent inputs to control the outputs. Second, the state where S ¼ 1 and R ¼ 1 causes
problems when real propagation delays are considered through the gates. Since it is impossible to
match the delays exactly between U1 and U2, the SR Latch may occasionally enter this state and
experience momentary metastable behavior. In order to address these issues, a number of
improvements can be made to this circuit to create two of the most commonly used storage devices in
sequential logic, the D-Latch and the D-Flip-Flop. In order to understand the operation of these storage
devices, two incremental modifications are made to the SR Latch. The first is called the S’R’ Latch and
the second is the SR Latch with enable. These two circuits are rarely implemented and are only
explained to understand how the SR Latch is modified to create a D-Latch and ultimately a D-Flip-Flop.

7.1.4 The S’R’ Latch

The S’R’ Latch operates in a similar manner as the SR Latch with the exception that the input codes
corresponding to the store, set, and reset states are complemented. To accomplish this complementary
behavior, the S’R’ Latch is implemented with NAND gates configured in a positive feedback configura-
tion. In this configuration, the S’R’ Latch will store the last output when S0 ¼ 1, R’ ¼ 1. It will set the output
(Q ¼ 1) when S0 ¼ 0, R’ ¼ 1. Finally, it will reset the output (Q ¼ 0) when S0 ¼ 1, R’ ¼ 0. Consider the
behavior of the S’R’ Latch during its store state shown in Fig. 7.8.
206 • Chapter 7: Sequential Logic Design

Fig. 7.8
S’R’ Latch behavior – store state (S0 ¼ 1, R’ ¼ 1)

Just as with the SR Latch, the S’R’ Latch has two input configurations to control the values of the
outputs. Consider the behavior of the S’R’ Latch during its set and reset states shown in Fig. 7.9.
 7.1 Sequential Logic Storage Devices • 207

Fig. 7.9
S’R’ Latch behavior – set (S00 ¼ 0, R” ¼ 1) and Reset (S00 ¼ 1, R” ¼ 0) states

And finally, just as with the SR Latch, the S’R’ Latch has a state that leads to potential metastability
and should be avoided. Consider the operation of the S’R’ Latch when the inputs are configured as S0 ¼ 0
and R’ ¼ 0 shown in Fig. 7.10.

Fig. 7.10
S’R’ Latch behavior – don’t use state (S00 ¼ 0 and R” ¼ 0)
208 • Chapter 7: Sequential Logic Design

The final truth table for the S’R’ Latch is given in Fig. 7.11.

Fig. 7.11
S00 R” Latch truth table

7.1.5 SR Latch with Enable

The next modification that is made in order to move toward a D-Latch and ultimately a D-Flip-Flop is
to add an enable line to the S’R’ Latch. The enable is implemented by adding two NAND gates on the
input stage of the S’R’ Latch. The SR Latch with enable is shown in Fig. 7.12. In this topology, the use of
NAND gates changes the polarity of the inputs so this circuit once again has a set state where S ¼ 1,
R ¼ 0 and a reset state of S ¼ 0, R ¼ 1. The enable line is labeled C, which stands for clock. The rationale
for this will be demonstrated upon moving through the explanation of the D-Latch.

Fig. 7.12
SR Latch with enable schematic

Recall that any time a 0 is present on one of the inputs to a NAND gate, the output will always be a
1 regardless of the value of the other inputs. In the SR Latch with enable configuration, any time C ¼ 0,
the outputs of U3 and U4 will be 1’s and will be fed into the inputs of the cross-coupled NAND gate
configuration (U1 and U2). Recall that the cross-coupled configuration of U1 and U2 is an S’R’ Latch and
will be put into a store state when S0 ¼ 1 and R’ ¼ 1. This is the store state (C ¼ 0). When C ¼ 1, it has the
effect of inverting the values of the S and R inputs before they reach U1 and U2. This condition allows the
set state to be entered when S ¼ 1, R ¼ 0, C ¼ 1 and the reset state to be entered when S ¼ 0, R ¼ 1,
C ¼ 1. Consider this operation in Fig. 7.13.
 7.1 Sequential Logic Storage Devices • 209

Fig. 7.13
SR Latch with enable behavior – store, set, and reset

Again, there is a potential metastable state when S ¼ 1, R ¼ 1 and C ¼ 1 that should be avoided.
There is also a second store state when S ¼ 0, R ¼ 0 and C ¼ 1 that is not used because storage is to be
dictated by the C input.

7.1.6 The D-Latch

The SR Latch with enable can be modified to create a new storage device called a D-Latch. Instead
of having two separate input lines to control the outputs of the latch, the R input of the latch is instead
210 • Chapter 7: Sequential Logic Design

driven with an inverted version of the S input. This prevents the S and R inputs from ever being the same
value and removes the two “Don’t Use” states in the truth table shown in Fig. 7.12. The new, single input
is renamed D to stand for data. This new circuit still has the behavior that it will store the last value of Q
and Qn when C ¼ 0. When C ¼ 1, the output will be Q ¼ 1 when D ¼ 1 and will be Q ¼ 0 when D ¼ 0. The
behavior of the output when C ¼ 1 is called tracking the input. The D-Latch schematic, symbol and truth
table are given in Fig. 7.14.

Fig. 7.14
D-Latch schematic, symbol and truth table

The timing diagram for the D-Latch is shown in Fig. 7.15.

Fig. 7.15
D-Latch timing diagram
 7.1 Sequential Logic Storage Devices • 211

7.1.7 The D-Flip-Flop

The final and most widely used storage device in sequential logic is the D-Flip-Flop. The D-Flip-Flop
is similar in behavior to the D-Latch with the exception that the store mode is triggered by a transition, or
edge on the clock signal instead of a level. This allows the D-Flip-Flop to implement higher frequency
systems since the outputs are updated in a shorter amount of time. The schematic, symbol and truth
table are given in Fig. 7.16 for a rising edge triggered D-Flip-Flop. To indicate that the device is edge
sensitive, the input for the clock is designated with a “>”. The U3 inverter in this schematic creates the
rising edge behavior. If U3 is omitted, this circuit would be a negative edge triggered D-Flip-Flop.

Fig. 7.16
D-Flip-Flop (rising edge triggered) schematic, symbol, and truth table

The D-Flip-Flop schematic shown above is called a master/slave configuration because of how the
data is passed through the two D-Latches (U1 and U2). Due to the U4 inverter, the two D-Latches will
always be in complementary modes. When U1 is in hold mode, U2 will be in track mode and vice versa.
When the clock signal transitions HIGH, U1 will store the last value of data. During the time when the
clock is HIGH, U2 will enter track mode and pass this value to Q. In this way, the data is latched into the
storage device on the rising edge of the clock and is present on Q. This is the master operation of the
device because U1, or the first D-Latch, is holding the value, and the second D-Latch (the slave) is simply
passing this value to the output Q. When the clock transitions LOW, U2 will store the output of U1. Since
there is a finite delay through U1, the U2 D-Latch is able to store the value before U1 fully enters track
mode. U2 will drive Q for the duration of the time that the clock is LOW. This is the slave operation of the
device because U2, or the second D-Latch, is holding the value. During the time the clock is LOW, U1 is
in track mode, which passes the input data to the middle of the D-Flip-Flop preparing for the next rising
edge of the clock. The master / slave configuration creates a behavior where the Q output of the D-Flip-
Flop is only updated with the value of D on a rising edge of the clock. At all other times, Q holds the last
value of D. An example timing diagram for the operation of a rising edge D-Flip-Flop is given in Fig. 7.17.
212 • Chapter 7: Sequential Logic Design

Fig. 7.17
D-Flip-Flop (rising edge triggered) timing diagram

D-Flip-Flops often have additional signals that will set the initial conditions of the outputs that are
separate from the clock. A reset input is used to force the outputs to Q ¼ 0, Qn ¼ 1. A preset input is used
to force the outputs to Q ¼ 1, Qn ¼ 0. In most modern D-Flip-Flops, these inputs are active LOW,
meaning that the line is asserted when the input is a 0. Active LOW inputs are indicated by placing an
inversion bubble on the input pin of the symbol. These lines are typically asynchronous, meaning that
when they are asserted, action is immediately taken to alter the outputs. This is different from a
synchronous input in which action is only taken on the edge of the clock. Fig. 7.18 shows the symbols
and truth tables for two D-Flip-Flop variants, one with an active LOW reset and another with both an
active LOW reset and active LOW preset.
 7.1 Sequential Logic Storage Devices • 213

Fig. 7.18
D-Flip-Flop with asynchronous reset and preset

D-Flip-Flops can also be created with an enable line. An enable line controls whether or not the
output is updated. Enable lines are synchronous, meaning that when they are asserted, the outputs will
be updated on the rising edge of the clock. When de-asserted, the outputs are not updated. This
behavior in effect ignores the clock input when de-asserted. Fig. 7.19 shows the symbol and truth
table for a D-Flip-Flop with a synchronous enable.

Fig. 7.19
D-Flip-Flop with synchronous enable

The behavior of the D-Flip-Flop allows us to design systems that are synchronous to a clock signal.
A clock signal is a periodic square wave that dictates when events occur in a digital system. A
synchronous system based on D-Flip-Flops will allow the outputs of its storage devices to be updated
upon a rising edge of the clock. This is advantageous because when the Q outputs are storing values
they can be used as inputs for combinational logic circuits. Since combinational logic circuits contain a
certain amount of propagation delay before the final output is calculated, the D-Flip-Flop can hold the
inputs at a steady value while the output is generated. Since the input on a D-Flip-Flop is ignored during
all other times, the output of a combinational logic circuit can be fed back as an input to a D-Flip-Flop.
214 • Chapter 7: Sequential Logic Design

This gives a system the ability to generate outputs based on the current values of inputs in addition to
past values of the inputs that are being held on the outputs of D-Flip-Flops. This is the definition of
sequential logic. An example synchronous, sequential system is shown in Fig. 7.20.

Fig. 7.20
An example synchronous system based on a D-Flip-Flop

CONCEPT CHECK

CC7.1(a) What will always cause a digital storage device to come out of metastability and settle in
one of its two stable states? Why?
(A) The power supply. The power supply provides the necessary current for the
device to overcome metastability.
(B) Electrical noise. Noise will always push the storage device toward one state or
another. Once the storage device starts moving toward one of its stable states,
the positive feedback of the storage device will reinforce the transition until the
output eventually comes to rest in a stable state.
(C) A reset. A reset will put the device into a known stable state.
(D) A rising edge of clock. The clock also puts the device into a known stable
state.
CC7.1(b) What was the purpose of replacing the inverters in the cross-coupled inverter pair with
NOR gates to form the SR Latch?
(A) NOR gates are easier to implement in CMOS.
(B) To provide the additional output Qn.
(C) To provide more drive strength for storing.
(D) To provide inputs to explicitly set the value being stored.

7.2 Sequential Logic Timing Considerations

There are a variety of timing specifications that need to be met in order to successfully design
circuits using sequential storage devices. The first specification is called the setup time (tsetup or ts). The
setup time specifies how long the data input needs to be at a steady state before the clock event. The
second specification is called the hold time (thold or th). The hold time specifies how long the data input
 7.2 Sequential Logic Timing Considerations • 215

needs to be at a steady state after the clock event. If these specifications are violated (i.e., the input
transitions too close to the clock transition), the storage device will not be able to determine whether the
input was a 1 or 0 and will go metastable. The time a storage device will remain metastable is a
deterministic value and is specified by the part manufacturer (tmeta). In general, metastability should be
avoided; however, knowing the maximum duration of metastability for a storage device allows us to
design circuits to overcome potential metastable conditions. During the time the device is metastable,
the output will have random behavior. It may go to a steady state 1, a steady state 0, or toggle between a
0 and 1 uncontrollably. Once the device comes out of metastability, it will come to rest in one of its two
stable states (Q ¼ 0 or Q ¼ 1). The final resting state is random and unknown. Another specification for
sequential storage devices is the delay from the time a clock transition occurs to the point that the data is
present on the Q output. This specification is called the Clock-to-Q delay and is given the notation tCQ.
These specifications are shown in Fig. 7.21.

Fig. 7.21
Sequential storage device timing specifications

CONCEPT CHECK

CC7.2 Which D-flop-flop timing specification requires all of combinational logic circuits in the system
to settle on their final output before a triggering clock edge can occur?
(A) tsetup (B) thold (C) tCQ (D) tmeta
216 • Chapter 7: Sequential Logic Design

7.3 Common Circuits Based on Sequential Storage Devices

Sequential logic storage devices give us the ability to create sophisticated circuits that can make
decisions based on the current and past values of the inputs; however, there are a variety of simple, yet
useful circuits that can be created with only these storage devices. This section will introduce a few of
these circuits.

7.3.1 Toggle Flop Clock Divider

A Toggle Flop is a circuit that contains a D-Flip-Flop configured with its Qn output wired back to its D
input. This configuration is also commonly referred to as a T-Flip-Flop or T-Flop. In this circuit, the only
input is the clock signal. Let’s examine the behavior of this circuit when its outputs are initialized to Q ¼ 0,
Qn ¼ 1. Since Qn is wired to the D input, a logic 1 is present on the input before the first clock edge. Upon
a rising edge of the clock, Q is updated with the value of D. This puts the outputs at Q ¼ 1, Qn ¼ 0. With
these outputs, now a logic 0 is present on the input before the next clock edge. Upon the next rising edge
of the clock, Q is updated with the value of D. This time the outputs go to Q ¼ 0, Qn ¼ 1. This behavior
continues indefinitely. The circuit is called a toggle flop because the outputs simply toggle between a
0 and 1 every time there is a rising edge of the clock. This configuration produces outputs that are square
waves with exactly half the frequency of the incoming clock. As a result, this circuit is also called a clock
divider. This circuit can be given its own symbol with a label of “T” indicating it is a toggle flop. The
configuration of a Toggle Flop (T-Flop) and timing diagram are shown in Fig. 7.22.

Fig. 7.22
Toggle flop clock frequency divider
 7.3 Common Circuits Based on Sequential Storage Devices • 217

7.3.2 Ripple Counter

The toggle flop configuration can be used to create a simple binary counter called a ripple counter. In
this configuration, the Qn output of a toggle flop is used as the clock for a subsequent toggle flop. Since
the output of the first toggle flop is a square wave that is ½ the frequency of the incoming clock, this
configuration will produce an output on the second toggle flop that is ¼ the frequency of the incoming
clock. This is by nature the behavior of a binary counter. The output of this counter is present on the Q
pins of each toggle flop. Toggle flops are added until the desired width of the counter is achieved with
each toggle flop representing one bit of the counter. Since each toggle flop produces the clock for the
subsequent latch, the clock is said to ripple through the circuit, hence the name ripple counter. A 3-bit
ripple counter is shown in Fig. 7.23.

Fig. 7.23
3-bit ripple counter

7.3.3 Switch Debouncing

Another useful circuit based on sequential storage devices is a switch debouncer. Mechanical
switches have a well-known issue of not producing clean logic transitions on their outputs when pressed.
This becomes problematic when using a switch to create an input for a digital device because it will
cause unwanted logic level transitions on the output of the gate. In the case of a clock input, this
unwanted transition can cause a storage device to unintentionally latch incorrect data.
218 • Chapter 7: Sequential Logic Design

The primary cause of these unclean logic transitions is due to the physical vibrations of the metal
contacts when they collide with each other during a button press or switch actuation. Within a mechanical
switch, there is typically one contact that is fixed and another that is designed to move when the button is
pressed. The contact that is designed to move can be thought of as a beam that is fixed on one side and
free on the other. As the free side of the beam moves toward the fixed contact in order to close the circuit,
it will collide and then vibrate just as a tuning fork does when struck. The vibration will eventually diminish
and the contact will come to rest, thus making a clean electrical connection; however, during the vibration
period the moving contact will bounce up and down on the destination contact. This bouncing causes the
switch to open and close multiple times before coming to rest in the closed position. This phenomenon is
accurately referred to as switch bounce. Switch bounce is present in all mechanical switches and gets
progressively worse as the switches are used more and more.
Figure 7.24 shows some of the common types of switches found in digital systems. The term pole is
used to describe the number of separate circuits controlled by the switch. The term throw is used to
describe the number of separate closed positions the switch can be in.

Fig. 7.24
Common types of mechanical switches

Let’s look at switch bounce when using a SPST switch to provide an input to a logic gate. A SPST
requires a resistor and can be configured to provide either a logic HIGH or LOW when in the open
position and the opposite logic level when in the closed position. The example configuration in Fig. 7.25
provides a logic LOW when in the open position and a logic HIGH when in the closed position. In the
open position, the input to the gate (SW) is pulled to GND to create a logic LOW. In the closed position,
the input to the gate is pulled to VCC to create a logic HIGH. A resistor is necessary to prevent a short
circuit between VCC and GND when the switch is closed. Since the input current specification for a logic
gate is very small, the voltage developed across the resistor due to the gate input current is negligible.
This means that the resistor can be inserted in the pull-down network without developing a noticeable
voltage. When the switch closes, the free-moving contact will bounce off of the destination contact
numerous times before settling in the closed position. During the time while the switch is bouncing, the
switch will repeatedly toggle between the open (HIGH) and closed (LOW) positions.
 7.3 Common Circuits Based on Sequential Storage Devices • 219

Fig. 7.25
Switch bouncing in a single pole, single throw switch

A possible solution to eliminate this switch bounce is to instead use a SPDT switch in conjunction
with a sequential storage device. Before looking at this solution, we need to examine an additional
condition introduced by the SPDT switch. The SPDT switch has what is known as break-before-make
behavior. The term break is used to describe when a switch is open while the term make is used to
describe when the switch is closed. When a SPDT switch is pressed, the input will be floating during the
time when the free-moving contact is transitioning toward the destination contact. During this time, the
output of the switch is unknown and can cause unwanted logic transitions if it is being used to drive the
input of a logic gate.
Let’s look at switch bounce when using a SPDT switch without additional circuitry to handle
bouncing. A SPDT has two positions that the free-moving contact can make a connection to (i.e., double
throw). When using this switch to drive a logic level into a gate, one position is configured as a logic HIGH
and the other a logic LOW. Consider the SPDT switch configuration in Fig. 7.26. Position 1 of the SPDT
switch is connected to GND, while position 2 is connect to VCC. When unpressed the switch is in position
1. When pressed, the free-moving contact will transition from position 1 to 2. During the transition the
free-moving contact is floating. This creates a condition where the input to the gate (SW) is unknown.
This floating input will cause unpredictable behavior on the output of the gate. Upon reaching position
2, the free-moving contact will bounce off of the destination contact. This will cause the input of the logic
gate to toggle between a logic HIGH and floating repeatedly until the free-moving contact comes to rest
in position 2.
220 • Chapter 7: Sequential Logic Design

Fig. 7.26
Switch bouncing in a single pole, double throw switch

The SPDT switch is ideal for use with an S’R’ Latch in order to produce a clean logic transition. This
is because during the break portion of the transition, an S’R’ Latch can be used to hold the last value of
the switch. This is unique to the SPDT configuration. The SPST switch in comparison does not have the
break characteristic, rather it always drives a logic level in both of its possible positions. Consider the
debounce circuit for a SPDT switch in Fig. 7.27. This circuit is based on an S’R’ Latch with two pull-up
resistors. Since the S’R’ Latch is created using NAND gates, this circuit is commonly called a NAND-
Debounce circuit. In the unpressed configuration, the switch drives S0 ¼ 0 and the R2 pull-up resistor
drives R’ ¼ 1. This creates a logic 0 on the output of the circuit (Qn ¼ 0). During a switch press, the free-
moving contact is floating, thus it is not driving in a logic level into the S’R’ Latch. Instead, both pull-up
resistors pull S0 and R’ to 1’s. This puts the latch into its hold mode and the output will remain at a logic
0 (Qn ¼ 0). Once the free-moving contact reaches the destination contact, the switch will drive R’ ¼ 0.
Since at this point the R1 pull-up is driving S0 ¼ 1, the latch outputs a logic 1 (Qn ¼ 1). When the free-
moving contact bounces off of the destination contact, it will put the latch back into the hold mode;
however, this time the last value that will be held is Qn ¼ 1. As the switch continues to bounce, the latch
will move between the Qn ¼ 1 and Qn ¼ “Last Qn” states, both of which produce an output of 1. In this
way, the SPDT switch in conjunction with the S’R’ Latch produces a clean 0 to 1 logic transition despite
the break-before-make behavior of the switch and the contact bounce.
 7.3 Common Circuits Based on Sequential Storage Devices • 221

Fig. 7.27
NAND debounce circuit for a SPDT switch

7.3.4 Shift Registers

A shift register is a chain of D-Flip-Flops that each are connected to a common clock. The output of
the first D-Flip-Flop is connected to the input of the second D-Flip-flop. The output of the second D-Flip-
Flop is connected to the input of the third D-Flip-Flop, and so on. When data is present on the input to the
first D-Flip-Flop, it will be latched upon the first rising edge of the clock. On the second rising edge of the
clock, the same data will be latched into the second D-Flip-Flop. This continues on each rising edge of
the clock until the data has been shifted entirely through the chain of D-Flip-Flops. Shift registers are
222 • Chapter 7: Sequential Logic Design

commonly used to convert a serial string of data into a parallel format. If an n-bit, serial sequence of
information is clocked into the shift register, after n clocks the data will be held on each of the D-Flip-Flop
outputs. At this moment, the n-bits can be read as a parallel value. Consider the shift register configura-
tion shown in Fig. 7.28.

Fig. 7.28
4-bit shift register

CONCEPT CHECK

CC7.3 Which D-flip-flop timing specification is most responsible for the ripple delay in a ripple
counter?
(A) tsetup (B) thold (C) tCQ (D) tmeta
 7.4 Finite State Machines • 223

7.4 Finite State Machines

Now we turn our attention to one of the most powerful sequential logic circuits, the finite state
machine (FSM). A FSM, or state machine, is a circuit that contains a pre-defined number of states (i.e., a
finite number of states). The machine can exist in one and only one state at a time. The circuit transitions
between states based on a triggering event, most commonly the edge of a clock, in addition to the values
of any inputs of the machine. The number of states and all possible transitions are pre-defined. Through
the use of states and a pre-defined sequence of transitions, the circuit is able to make decisions on the
next state to transition to based on a history of past states. This allows the circuit to create outputs that
are more intelligent compared to a simple combinational logic circuit that has outputs based only on the
current values of the inputs.

7.4.1 Describing the Functionality of a FSM

The design of a state machine begins with an abstract word description of the desired circuit
behavior. We will use a design example of a push-button motor controller to describe all of the steps
involved in creating a finite state machine. Example 7.1 starts the FSM design process by stating the
word description of the system.

Example 7.1
Push-button window controller – word description

7.4.1.1 State Diagrams

A state diagram is a graphical way to describe the functionality of a finite state machine. A state
diagram is a form of a directed graph, in which each state (or vertex) within the system is denoted as a
circle and given a descriptive name. The names are written inside of the circles. The transitions between
states are denoted using arrows with the input conditions causing the transitions written next to them.
Transitions (or edges) can move to different states upon particular input conditions or remain in the same
state. For a state machine implemented using sequential logic storage, an evaluation of when to
transition states is triggered every time the storage devices update their outputs. For example, if the
system was implemented using rising edge triggered D-flip-Flops, then an evaluation would occur on
every rising edge of the clock.
224 • Chapter 7: Sequential Logic Design

There are two different types of output conditions for a state machine. The first is when the output
only depends on the current state of the machine. This type of system is called a Moore Machine. In this
case, the outputs of the system are written inside of the state circles. This indicates the output value that
will be generated for each specific state. The second output condition is when the outputs depend on
both the current state and the system inputs. This type of system is called a Mealy Machine. In this case,
the outputs of the system are written next to the state transitions corresponding to the appropriate input
values. Outputs in a state diagram are typically written inside of parentheses. Example 7.2 shows the
construction of the state diagram for our push-button window controller design.

Example 7.2
Push-button window controller – state diagram
 7.4 Finite State Machines • 225

7.4.1.2 State Transition Tables

The state diagram can now be described in a table format that is similar to a truth table. This puts the
state machine behavior in a form that makes logic synthesis straightforward. The table contains the
same information as in the state diagram. The state that the machine exists in is called the current state.
For each current state that the machine can reside in, every possible input condition is listed along with
the destination state of each transition. The destination state for a transition is called the next state. Also
listed in the table are the outputs corresponding to each current state and, in the case of a Mealy
Machine, the output corresponding to each input condition. Example 7.3 shows the construction of the
state transition table for the push-button window controller design. This information is identical to the
state diagram given in Example 7.2.

Example 7.3
Push-button window xontroller – state transition table

7.4.2 Logic Synthesis for a FSM

Once the behavior of the state machine has been described, it can be directly synthesized. There
are three main components of a state machine: the state memory; the next state logic; and the output
logic. Figure 7.29 shows a block diagram of a state machine highlighting these three components. The
next state logic block is a group of combinational logic that produces the next state signals based on the
current state and any system inputs. The state memory holds the current state of the system. The current
state is updated with next state on every rising edge of the clock, which is indicated with the “>” symbol
within the block. This behavior is created using D-Flip-Flops where the current state is held on the Q
outputs of the D-Flip-Flops, while the next state is present on the D inputs of the D-Flip-Flops. In this way,
every rising edge of the clock will trigger an evaluation of which state to move to next. This decision is
based on the current state and the current inputs. The output logic block is a group of combinational logic
that creates the outputs of the system. This block always uses the current state as an input and,
depending on the type of machine (Mealy vs. Moore), uses the system inputs. It is useful to keep this
block diagram in mind when synthesizing finite state machines as it will aid in keeping the individual
design steps separate and clear.
226 • Chapter 7: Sequential Logic Design

Fig. 7.29
Main components of a finite state machine

7.4.2.1 State Memory

The state memory is the circuitry that will hold the current state of the machine. Upon a rising edge of
a clock it will update the current state with the next state. At all other times, the next state input is ignored.
This gives time for the next state logic circuitry to compute the results for the next state. This behavior is
identical to that of a D-Flip-Flop, thus the state memory is simply one or more D-Flip-Flops. The number
of D-Flip-Flops required depends on how the states are encoded. State encoding is the process of
assigning a binary value to the descriptive names of the states from the state diagram and state transition
tables. Once the descriptive names have been converted into representative codes using 1’s and 0’s, the
states can be implemented in real circuitry. The assignment of codes is arbitrary and can be selected in
order to minimize the circuitry needed in the machine.
There are three main styles of state encoding. The first is straight binary encoding. In this approach
the state codes are simply a set of binary counts (i.e., 00, 01, 10, 11. . .). The binary counts are assigned
starting at the beginning of the state diagram and incrementally assigned toward the end. This type of
encoding has the advantage that it is very efficient in minimizing the number of D-Flip-Flops needed for
the state memory. With n D-Flip-Flops, 2n states can be encoded. When a large number of states is
required, the number of D-Flip-Flops can be calculated using the rules of logarithmic math. Example 7.4
shows how to solve for the number of bits needed in the binary state code based on the number of states
in the machine.
 7.4 Finite State Machines • 227

Example 7.4
Solving for the number of bits needed for binary state encoding

The second type of state encoding is called gray code encoding. A gray code is one in which the
value of a code differs by only one bit from any of its neighbors, (i.e., 00, 01, 11, 10. . .). A gray code is
useful for reducing the number of bit transitions on the state codes when the machine has a transition
sequence that is linear. Reducing the number of bit transitions can reduce the amount of power
consumption and noise generated by the circuit. When the state transitions of a machine are highly
non-linear, a gray code encoding approach does not provide any benefit. Gray code is also an efficient
coding approach. With n D-Flip-Flops, 2n states can be encoded just as in binary encoding. Figure 7.30
shows the process of creating n-bit, gray code patterns.
228 • Chapter 7: Sequential Logic Design

Fig. 7.30
Creating an n-bit gray code pattern

The third common technique to encode states is using one-hot encoding. In this approach, a
separate D-Flip-Flop is asserted for each state in the machine. For an n-state machine, this encoding
approach requires n D-Flip-Flops. For example, if a machine had three states, the one-hot state codes
would be “001”, “010” and “100”. This approach has the advantage that the next state logic circuitry is
very simple; further, there is less chance that the different propagation delays through the next state logic
will cause an inadvertent state to be entered. This approach is not as efficient as binary and gray code in
terms of minimizing the number of D-Flip-Flops because it requires one D-Flip-Flop for each state;
however, in modern digital integrated circuits that have abundant D-Flip-Flops, one-hot encoding is
commonly used.
Figure 7.31 shows the differences between these three state encoding approaches.

Fig. 7.31
Comparison of different state encoding approaches

Once the codes have been assigned to the state names, each of the bits within the code must be
given a unique signal name. The signal names are necessary because the individual bits within the state
 7.4 Finite State Machines • 229

code are going to be implemented with real circuitry so each signal name will correspond to an actual
node in the logic diagram. These individual signal names are called state variables. Unique variable
names are needed for both the current state and next state signals. The current state variables are driven
by the Q outputs of the D-Flip-Flops holding the state codes. The next state variables are driven by the
next state logic circuitry and are connected to the D inputs of the D-Flip-Flops. State variable names are
commonly chosen that are descriptive both in terms of their purpose and connection location. For
example, current state variables are often given the names Q, Q_cur or Q_current to indicate that they
come from the Q outputs of the D-Flip-Flops. Next state variables are given names such as Q*, Q_nxt or
Q_next to indicate that they are the next value of Q and are connected to the D input of the D-Flip-Flops.
Once state codes and state variable names are assigned, the state transition table is updated with the
detailed information.
Returning to our push-button window controller example, let’s encode our states in straight binary
and use the state variable names of Q_cur and Q_nxt. Example 7.5 shows the process of state encoding
and the new state transition table.

Example 7.5
Push-button window controller – state encoding

7.4.2.2 Next State Logic

The next step in the state machine design is to synthesize the next state logic. The next state logic
will compute the values of the next state variables based on the current state and the system inputs.
Recall that a combinational logic function drives one and only one output bit. This means that every bit
within the next state code needs to have a dedicated combinational logic circuit. The state transition table
contains all of the necessary information to synthesize the next state logic including the exact
output values of each next state variable for each and every input combination of state code and system
input(s).
230 • Chapter 7: Sequential Logic Design

In our push-button window controller example, we only need to create one combinational logic
circuit because there is only one next state variable (Q_nxt). The inputs to the combinational logic circuit
are Q_cur and Press. Notice that the state transition table was created such that the order of the input
values are listed in a binary count just as in a formal truth table formation. This makes synthesizing the
combinational logic circuit straightforward. Example 7.6 shows the steps to synthesize the next state
logic for this the push-button window controller.

Example 7.6
Push-button window controller – next state logic

7.4.2.3 Output Logic

The next step in the state machine design is to synthesize the output logic. The output logic will
compute the values of the system outputs based on the current state and, in the case of a Mealy
machine, the system inputs. Each of the output signals will require a dedicated combinational
logic circuit. Again, the state transition table contains all of the necessary information to synthesize the
output logic.
In our push-button window controller example, we need to create one circuit to compute the output
“Open_CW” and one circuit to compute the output “Close_CCW”. In this example, the inputs to these
circuits are the current state (Q_cur) and the system input (Press). Example 7.7 shows the steps to
synthesize the output logic for the push-button window controller.
 7.4 Finite State Machines • 231

Example 7.7
Push-button window controller – output logic

7.4.2.4 The Final Logic Diagram

The final step is the design of the state machine is to create the logic diagram. It is useful to recall the
block diagram for a state machine from Fig. 7.29. A logic diagram begins by entering the state memory.
Recall that the state memory consists of D-Flip-Flops that hold the current state code. One D-Flip-Flop is
needed for every current state variable. When entering the D-Flip-Flops, it is useful to label them with the
current state variable they will be holding. The next part of the logic diagram is the next state logic. Each
of the combinational logic circuits that compute the next state variables should be drawn to the left of D-
Flip-Flop holding the corresponding current state variable. The output of each next state logic circuit is
connected to the D input of the corresponding D-Flip-Flop. Finally, the output logic is entered with the
inputs to the logic coming from the current state and potentially from the system inputs.
232 • Chapter 7: Sequential Logic Design

Example 7.8 shows the process for creating the final logic diagram for our push-button window
controller. Notice that the state memory is implemented with one D-Flip-Flop since there is only 1-bit in
the current state code (Q_cur). The next state logic is a combinational logic circuit that computes Q_nxt
based on the values of Q_cur and Press. Finally, the output logic consists of two separate combinational
logic circuits to compute the system outputs Open_CW and Close_CCW based on Q_cur and Press. In
this diagram the Qn output of the D-Flip-Flop could have been used for the inverted versions of Q_cur;
however, inversion bubbles were used instead in order to make the diagram more readable.

Example 7.8
Push-button window controller – logic diagram

7.4.3 FSM Design Process Overview

The entire finite state machine design process is given in Fig. 7.32.
 7.4 Finite State Machines • 233

Fig. 7.32
Finite state machine design flow

7.4.4 FSM Design Examples

7.4.4.1 Serial Bit Sequence Detector
Let’s consider the design of a 3-bit serial sequence detector. Example 7.9 provides the word
description, state diagram, and state transition table for this finite state machine.
234 • Chapter 7: Sequential Logic Design

Example 7.9
Serial bit sequence detector (part 1)
 7.4 Finite State Machines • 235

Example 7.10 provides the state encoding and next state logic synthesis for the 3-bit serial bit
sequence detector.

Example 7.10
Serial bit sequence detector (part 2)
236 • Chapter 7: Sequential Logic Design

Example 7.11 shows the output logic synthesis and final logic diagram for the 3-bit serial bit
sequence detector.

Example 7.11
Serial bit sequence detector (part 3)

7.4.4.2 Vending Machine Controller

Let’s now look at the design of a simple vending machine controller. Example 7.12 provides the word
description, state diagram, and state transition table for this finite state machine.
 7.4 Finite State Machines • 237

Example 7.12
Vending machine controller (part 1)
238 • Chapter 7: Sequential Logic Design

Example 7.13 provides the state encoding and next state logic synthesis for the simple vending
machine controller.

Example 7.13
Vending machine controller (part 2)
 7.4 Finite State Machines • 239

Example 7.14 shows the output logic synthesis and final logic diagram for the vending machine
controller.

Example 7.14
Vending machine controller (part 3)
240 • Chapter 7: Sequential Logic Design

CONCEPT CHECK

CC7.4(a) What allows a finite state machine to make more intelligent decisions about the system
outputs compared to combinational logic alone?
(A) A finite state machine has knowledge about the past inputs.
(B) The D-flip-flops allow the outputs to be generated more rapidly.
(C) The next state and output logic allows the finite state machine to be more
complex and implement larger truth Tables.
(D) A synchronous system is always more intelligent.
CC7.4(b) When designing a finite state machine, many of the details of the implementation can be
abstracted. At what design step do the details of the implementation start being
considered?
(A) The state diagram step.
(B) The state transition table step.
(C) The state memory synthesis step.
(D) The word description.
CC7.4(c) What impact does adding an additional state have on the implementation of the state
memory logic in a finite state machine?
(A) It adds an additional D-flip-flop.
(B) It adds a new state code that must be supported.
(C) It adds more combinational logic to the logic diagram.
(D) It reduces the speed that the machine can run at.
CC7.4(d) Which of the following statements about the next state logic is FALSE?
(A) It is always combinational logic.
(B) It always uses the current state as one of its inputs.
(C) Its outputs are connected to the D inputs of the D-flip-flops in the state memory.
(D) It uses the results of the output logic as part of its inputs.
CC7.4(e) Why does the output logic stage of a finite state machine always use the current state as
one of its inputs?
(A) If it didn’t, it would simply be a separate combinational logic circuit and not be
part of the finite state machine.
(B) To make better decisions about what the system outputs should be.
(C) Because the next state logic is located too far away.
(D) Because the current state is produced on every triggering clock edge.
CC7.4(f) What impact does asserting a reset have on a finite state machine?
(A) It will cause the output logic to produce all zeros.
(B) It will cause the next state logic to produce all zeros.
(C) It will set the current state code to all zeros.
(D) It will start the system clock.
 7.5 Counters • 241

7.5 Counters
A counter is a special type of finite state machine. A counter will traverse the states within a state
diagram in a linear fashion continually circling around all states. This behavior allows a special type of
output topology called state-encoded outputs. Since each state in the counter represents a unique
counter output, the states can be encoded with the associated counter output value. In this way, the
current state code of the machine can be used as the output of the entire system.

7.5.1 2-Bit Binary Up Counter

Let’s consider the design of a 2-bit binary up counter. Example 7.15 provides the word description,
state diagram, state transition table, and state encoding for this counter.

Example 7.15
2-bit binary up counter (part 1)
242 • Chapter 7: Sequential Logic Design

Example 7.16 shows the next state and output logic synthesis, the final logic diagram, and resultant
representative timing diagram for the 2-bit binary up counter.

Example 7.16
2-bit binary up counter (part 2)

7.5.2 2-Bit Binary Up/Down Counter

Let’s now consider a 2-bit binary up/down counter. In this type of counter, there is an input that
dictates whether the counter increments or decrements. This counter can still be implemented as a
 7.5 Counters • 243

Moore machine and use state-encoded outputs. Example 7.17 provides the word description, state
diagram, state transition table, and state encoding for this counter.

Example 7.17
2-bit binary up/down counter (part 1)
244 • Chapter 7: Sequential Logic Design

Example 7.18 shows the next state and output logic synthesis, the final logic diagram, and resultant
representative timing diagram for the 2-bit binary up/down counter.

Example 7.18
2-bit binary up/down counter (part 2)
 7.5 Counters • 245

7.5.3 2-Bit Gray Code Up Counter

A gray code counter is one in which the output only differs by one bit from its prior value. This type of
counter can be implemented using state-encoded outputs by simply encoding the states in gray code.
Let’s consider the design of a 2-bit gray code up counter. Example 7.19 provides the word description,
state diagram, state transition table, and state encoding for this counter.

Example 7.19
2-bit gray code up counter (part 1)
246 • Chapter 7: Sequential Logic Design

Example 7.20 shows the next state and output logic synthesis, the final logic diagram, and resultant
representative timing diagram for the 2-bit gray code up counter.

Example 7.20
2-bit gray code up counter (part 2)
 7.5 Counters • 247

7.5.4 2-Bit Gray Code Up/Down Counter

Let’s now consider a 2-bit gray code up/down counter. In this type of counter, there is an input that
dictates whether the counter increments or decrements. This counter can still be implemented as a
Moore machine and use state-encoded outputs. Example 7.21 provides the word description, state
diagram, state transition table, and state encoding for this counter.

Example 7.21
2-bit gray code up/down counter (part 1)
248 • Chapter 7: Sequential Logic Design

Example 7.22 shows the next state and output logic synthesis, the final logic diagram, and resultant
representative timing diagram for the 2-bit gray code up/down counter.

Example 7.22
2-bit gray code up/down counter (part 2)
 7.5 Counters • 249

7.5.5 3-Bit One-Hot Up Counter

A one-hot counter creates an output in which one and only one bit is asserted at a time. In an up
counter configuration, the assertion is made on the least significant bit first, followed by the next higher
significant bit, and so on (i.e., 001, 010, 100, 001. . .). A one-hot counter can be created using state-
encoded outputs. For a n-bit counter, the machine will require n D-Flip-Flops. Let’s consider a 3-bit
one-hot up counter. Example 7.23 provides the word description, state diagram, state transition table,
and state encoding for this counter.

Example 7.23
3-bit one-hot up counter (part 1)
250 • Chapter 7: Sequential Logic Design

Example 7.24 shows the next state and output logic synthesis, the final logic diagram, and resultant
representative timing diagram for the 3-bit one-hot up counter.

Example 7.24
3-bit one-hot up counter (part 2)

7.5.6 3-Bit One-Hot Up/Down Counter

Let’s now consider a 3-bit one-hot up/down counter. In this type of counter, there is an input that
dictates whether the counter increments or decrements. This counter can still be implemented as a
Moore machine and use state-encoded outputs. Example 7.25 provides the word description, state
diagram, state transition table, and state encoding for this counter.
 7.5 Counters • 251

Example 7.25
3-bit one-hot up/down counter (part 1)
252 • Chapter 7: Sequential Logic Design

Example 7.26 shows the next state and output logic synthesis for the 3-bit one-hot up/down counter.

Example 7.26
3-bit one-hot up/down counter (part 2)
 7.5 Counters • 253

Finally, Example 7.27 shows the logic diagram and resultant representative timing diagram for the
counter.

Example 7.27
3-bit one-hot up/down counter (part 3)

CONCEPT CHECK

CC7.5 What characteristic of a counter makes it a special case of a finite state machine?
(A) The state transitions are mostly linear, which reduces the implementation
complexity.
(B) The outputs are always a gray code.
(C) The next state logic circuitry is typically just sum terms.
(D) There is never a situation where a counter could be a Mealy machine.
254 • Chapter 7: Sequential Logic Design

7.6 Finite State Machine’s Reset Condition

The one-hot counter designs in Examples 7.23 and 7.25 where the first FSM examples that had an
initial state that was not encoded with all 0’s. Notice that all of the other FSM examples had initial states
with state codes comprised of all 0’s (e.g., w_closed ¼ 0, S0 ¼ “00”, C0 ¼ “00”, GC_0 ¼ “00, etc.). When
the initial state is encoded with all 0’s, the FSM can be put into this state by asserting the reset line of all of
the D-Flip-Flops in the state memory. By asserting the reset line, the Q outputs of all of the D-Flip-Flips
are forced to 0’s. This sets the initial current state value to whatever state is encoded with all 0’s. The
initial state of a machine is often referred to as the reset state. The circuitry to initialize state machines is
often omitted from the logic diagram as it is assumed that the necessary circuitry will exist in order to put
the state machine into the reset state. If the reset state is encoded with all 0’s, then the reset line can be
used alone; however, if the reset state code contains 1’s, then both the reset and preset lines must be
used to put the machine into the reset state upon start up. Let’s look at the behavior of the one-hot up
counter again. Figure 7.33 shows how using the reset lines of the D-Flip-Flops alone will cause the circuit
to operate incorrectly. Instead, a combination of the reset and preset lines must be used to get the
one-hot counter into its initial state of Hot_0 ¼ “00100 .

Fig. 7.33
Finite state machine reset state
 7.7 Sequential Logic Analysis • 255

Resets are most often asynchronous so that they can immediately alter the state of the FSM. If a
reset was implemented in a synchronous manner and there was a clock failure, the system could not be
reset since there would be no more subsequent clock edges that would recognize that the reset line was
asserted. An asynchronous reset allows the system to be fully restarted even in the event of a clock
failure.’

CONCEPT CHECK

CC7.6 What is the downside of using D-flip-flops that do not have preset capability in a finite state
machine?
(A) The finite state machine will run slower.
(B) The next state logic will be more complex.
(C) The output logic will not be able to support both Mealy and Moore type machine
architectures.
(D) The start-up state can never have a 1 in its state code.

7.7 Sequential Logic Analysis

Sequential logic analysis refers to the act of deciphering the operation of a circuit from its final logic
diagram. This is similar to combinational logic analysis with the exception that the storage capability of
the D-flip-flops must be considered. This analysis is also used to understand the timing of a sequential
logic circuit and can be used to predict the maximum clock rate that can be used.

7.7.1 Finding the State Equations and Output Logic Expressions of a FSM
When given the logic diagram for a finite state machine and it is desired to reverse-engineer its
behavior, the first step is to determine the next state logic and output logic expressions. This can be
accomplished by first labeling the current and next state variables on the inputs and outputs of the D-flip-
flops that are implementing the state memory of the FSM. The outputs of the D-flip-flops are labeled with
arbitrary current state variable names (e.g., Q1_cur, Q0_cur, etc.) and the inputs are labeled with
arbitrary next state variable names (e.g., Q1_nxt, Q0_nxt, etc.). The numbering of the state variables
can be assigned to the D-flip-flops arbitrarily as long as the current and next state bit numbering is
matched. For example, if a D-flip-flop is labeled to hold bit 0 of the state code, its output should be labeled
Q0_cur and its input should be labeled Q0_nxt.
Once the current state variable nets are labeled in the logic diagram, the expressions for the next
state logic can be found by analyzing the combinational logic circuity driving the next state variables
(e.g., Q1_nxt, Q0_nxt). The next state logic expressions will be in terms of the current state variables
(e.g., Q1_cur, Q0_cur) and any inputs to the FSM.
The output logic expressions can also be found by analyzing the combinational logic driving the
outputs of the FSM. Again, these will be in terms of the current state variables and potentially the inputs
to the FSM. When analyzing the output logic, the type of machine can be determined. If the output logic
only depends on combinational logic that is driven by the current state variables, the FSM is a Moore
machine. If the output logic depends on both the current state variables and the FSM inputs, the FSM is a
Mealy machine. An example of this analysis approach is given in Example 7.28.
256 • Chapter 7: Sequential Logic Design

Example 7.28
Determining the next state logic and output logic expression of a FSM

7.7.2 Finding the State Transition Table of a FSM

Once the next state logic and output logic expressions are known, the state transition table can be
created. It is useful to assign more descriptive names to all possible state codes in the FSM. The number
of unique states possible depends on how many D-flip-flops are used in the state memory of the FSM.
For example, if the FSM uses two D-flip-flops there are four unique state codes (i.e., 00, 01, 10, 11). We
can assign descriptive names such as S0 ¼ 00, S1 ¼ 01, S2 ¼ 10, S3 ¼ 11. When first creating the
transition table, we assign labels and list each possible state code. If a particular code is not used, it can
be removed from the transition table at the end of the analysis. The state code that the machine will start
in can be found by analyzing its reset and preset connections. This code is typically listed first in the
table. The transition table is then populated with all possible combinations of current states and inputs.
 7.7 Sequential Logic Analysis • 257

The next state codes and output logic values can then be populated by evaluating the next state logic
and output logic expressions found earlier. An example of this analysis is shown in Example 7.29.

Example 7.29
Determining the state transition table of a FSM

7.7.3 Finding the State Diagram of a FSM

Once the state transition table is found, creating the state diagram becomes possible. We start the
diagram with the state corresponding to the reset state. We then draw how the FSM transitions between
each of its possible states based on the inputs to the machine and list the corresponding outputs. An
example of this analysis is shown in Example 7.30.
258 • Chapter 7: Sequential Logic Design

Example 7.30
Determining the state diagram of a FSM

7.7.4 Determining the Maximum Clock Frequency of a FSM

The maximum clock frequency is often one of the banner specifications for a digital system. The
clock frequency of a FSM depends on a variety of timing specifications within the sequential circuit
including the setup and hold time of the D-flip-flop, the clock-to-Q delay of the D-flip-flop, the combina-
tional logic delay driving the input of the D-flip-flop, the delay of the interconnect that wires the circuit
together, and the desired margin for the circuit. The basic concept of analyzing the timing of FSM is to
determine how long we must wait after a rising (assuming a rising edge triggered D-flip-flop) clock edge
occurs until the subsequent rising clock edge can occur. The amount of time that must be allowed
between rising clock edges depends on how much delay exists in the system. A sufficient amount of time
 7.7 Sequential Logic Analysis • 259

must exist between clock edges to allow the logic computations to settle so that on the next clock edge
the D-flip-flops can latch in a new value on their inputs.
Let’s examine all of the sources of delay in a FSM. Let’s begin by assuming that all logic values are
at a stable value and we experience a rising clock edge. The value present on the D input of the D-flip-
flop is latched into the storage device and will appear on the Q output after one clock-to-Q delay of the
device (tCQ). Once the new value is produced on the output of the D-flip-flop, it is then used by a variety of
combinational logic circuits to produce the next state codes and the outputs of the FSM. The next state
code computation is typically longer than the output computation so let’s examine that path. The new
value on Q propagates through the combinational logic circuitry and produces the next state code at the
D input of the D-flip-flop. The delay to produce this next state code includes wiring delay in addition to
gate delay. When analyzing the delay of the combinational logic circuitry (tcmb) and the delay of the
interconnect (tint), the worst case path is always considered. Once the new logic value is produced by the
next state logic circuitry, it must remain stable for a certain amount of time in order to meet the D-flip-flop’s
setup specification (tsetup). Once this specification is met, the D-flip-flop could be clocked with the next
clock edge; however, this represents a scenario without any margin in the timing. This means that if
anything in the system caused the delay to increase even slightly, the D-flip-flop could go metastable. To
avoid this situation, margin is included in the delay (tmargin). This provides some padding so that the
system can reliably operate. A margin of 10% is typical in digital systems. The time that must exist
between rising clock edges is then simply the sum of all of these sources of delay
(tCQ + tcmb + tint + tsetup + tmargin). Since the time between rising clock edges is defined as the period of
the signal (T), this value is also the definition of the period of the fastest clock. Since the frequency of a
signal is simply f ¼ 1/T, the maximum clock frequency for the FSM is the reciprocal of the sum of the
delay.
One specification that is not discussed in the above description is the hold time of the D-flip-flop
(thold). The hold specification is the amount of time that the input to the D-flip-flop must remain constant
after the clock edge. In modern storage devices, this time is typically very small and considerably less
than the tCQ specification. If the hold specification is less than tCQ it can be ignored because the output of
the D-flip-flop will not change until after one tCQ anyway. This means that the hold requirements are
inherently met. This is the situation with the majority of modern D-flip-flops. In the rare case that the hold
time is greater than tCQ, then it is used in place of tCQ in the summation of delays. Figure 7.34 gives the
summary of the maximum clock frequency calculation when analyzing a FSM.
260 • Chapter 7: Sequential Logic Design

Fig. 7.34
Timing analysis of a finite state machine

Let’s take a look at an example of how to use this analysis. Example 7.31 shows this analysis for the
FSM analyzed in prior sections but this time considering the delay specifications of each device.
 7.7 Sequential Logic Analysis • 261

Example 7.31
Determining the maximum clock frequency of a FSM

CONCEPT CHECK

CC7.7 What is the risk of running the clock above its maximum allowable frequency in a finite state
machine?
(A) The power consumption may drop below the recommended level.
(B) The setup and hold specifications of the D-flip-flops may be violated, which may
put the machine into an unwanted state.
(C) The states may transition too quickly to be usable.
(D) The crystal generating the clock may become unstable.
262 • Chapter 7: Sequential Logic Design

Summary
v Sequential logic refers to a circuit that bases v A finite state machine (FSM) is a system that
its outputs on both the present and past produces outputs based on the current value
values of the inputs. Past values are held in of the inputs and a history of past inputs. The
sequential logic storage device. history of inputs are recorded as states that the
v All sequential logic storage devices are machine has been in. As the machine
based on a cross-coupled feedback loop. responds to new inputs, it transitions between
The positive feedback loop formed in this states. This allows a finite state machine to
configuration will hold either a 1 or a 0. This make more sophisticated decisions about
is known as a bistable device. what outputs to produce by knowing its history.
v If the inputs of the feedback loop in a sequen- v A state diagram is a graphical way to
tial logic storage device are driven to exactly describe the behavior of a FSM. States are
between a 1 and a 0 (i.e., Vcc/2) and then represented using circles and transitions are
released, the device will go metastable. represented using arrows. Outputs are listed
Metastability refers to the behavior where either inside of the state circle or next to the
the device will ultimately be pushed toward transition arrow.
one of the two stable states in the system, v A state transition table contains the same
typically by electrical noise. Once the device information as a state diagram, but in tabular
begins moving toward one of the stable format. This allows the system to be more
states, the positive feedback will reinforce easily synthesized because the information
the transition until it reaches the stable is in a form similar to a truth table.
state. The stable state that the device will v The first step in FSM synthesis is creating the
move toward is random and unknown. state memory. The state memory consists of
v Cross-coupled inverters are the most basic a set of D-flip-flops that hold the current state
form of the positive feedback loop configura- of the FSM. Each state in the FSM must be
tion. To give the ability to drive the outputs of assigned a binary code. The type of
the storage device to known values, the encoding is arbitrary; however, there are cer-
inverters are replaced with NOR gates to tain encoding types that are commonly used
form the SR Latch. A variety of other such as binary, gray code, and one-hot.
modifications can be made to the loop con- Once the codes are assigned, state variables
figuration to ultimately produce a D-latch and need to be defined for each bit position for
D-flip-flop. both the current state and the next state
v A D-flip-flop will update its Q output with the codes. The state variables for the current
value on its D input on every triggering edge state represent the Q outputs of the D-flip-
of a clock. The amount of time that it takes for flops, which hold the current state code. The
the Q output to update after a triggering clock state variables for the next state code repre-
edge is called the “t-clock-to-Q” (tCQ) sent the D inputs of the D-flip-flops. A D-flip-
specification. flop is needed for each bit in the state code.
v The setup and hold times of a D-flip-flop On the triggering edge of a clock, the current
describe how long before (tsetup) and after state will be updated with the next state code.
(thold) the triggering clock edge that the data v The second step in FSM synthesis is creating
on the D input of the device must be stable. If the next state logic. The next state logic is
the D input transitions too close to the trigger- combinational logic circuitry that produces
ing clock edge (i.e., violating a setup or hold the next state codes based on the current
specification) then the device will go meta- state variables and any system inputs. The
stable and the ultimate value on Q is next state logic drives the D inputs of the D-
unknown. flip-flops in the state memory.
v A synchronous system is one in which all v The third step in FSM synthesis is creating
logic transitions occur based on a single the output logic. The output logic is combina-
timing event. The timing event is typically tional logic circuitry that produces the system
the triggering edge of a clock. outputs based on the current state, and
v There are a variety of common circuits that potentially, the system inputs.
can be accomplished using just sequential v The output logic always depends on the cur-
storage devices. Examples of these circuits rent state of a FSM. If the output logic also
include switch debouncing, toggle-flops, rip- depends on the system inputs, the machine
ple counters, and shift registers. is a Mealy machine. If the output logic does
 Exercise Problems • 263

not depend on the system inputs, the calculate the possible number of state codes
machine is a Moore machine. that the machine has. The state codes are
v A counter is a special type of finite state then used to calculate the next state logic
machine in which the states are traversed and output values. From this information a
linearly. The linear progression of states state transition table can be created and in
allows the next state logic to be simplified. turn, the state diagram.
The complexity of the output logic in a counter v The maximum frequency of a FSM is found
can also be reduced by encoding the states by summing all sources of time delay that
with the desired counter output for that state. must be accounted for before the next trig-
This technique, known as state-encoded gering edge of the clock can occur. These
outputs, allows the system outputs to simply sources include tCQ, the worst case combina-
be the current state of the FSM. tional logic path, the worst case interconnect
v The reset state of a FSM is the state that the delay path, the setup/hold time of the D-flip-
machine will go to when it begins operation. flops, and any margin that is to be included.
The state code for the reset state must be The sum of these timing delays represents
configured using the reset and/or preset lines the smallest period (T) that the clock can
of the D-flip-flops. If only reset lines are used have. This is then converted to frequency.
on the D-flip-flops, the reset state must be v If the tCQ time is greater than or equal to the
encoded using only zeros. hold time, the hold time can be ignored in the
v Given the logic diagram for a state machine, maximum frequency calculation. This is
the logic expression for the next state memory because the outputs of the D-flip-flops are
and the output logic can be determined by inherently held while the D-flip-flops are pro-
analyzing the combinational logic driving the ducing the next output value. The time it
D inputs of the state memory (i.e., the next takes to change the outputs after a triggering
state logic) and the combinational logic driving clock edge is defined as tCQ. This means as
the system outputs (i.e., the output logic). long as tCQ  thold, the hold time specification
v Given the logic diagram for a state diagram, is inherently met since the logic driving the
the state diagram can be determined by first next state codes uses the Q outputs of the
finding the logic expressions for the next D-flip-flops.
state and output logic. The number of D-flip-
flops in the logic diagram can then be used to

Exercise Problems
For some of the following exercise problems normal behavior. The details of how to create
you will be asked to design a Verilog model your own model of a D-Flip-Flop are provided
and perform a functional simulation. You will later in Chap. 8.
be provided with a test bench for each of
these problems. The details of how to create
your own Verilog test bench are provided later
in Chap. 8. For some of the following exercise
problems you will be asked to use D-Flip-
Flops as part of a Verilog design. You will be
provided with the model of the D-Flip-Flop
and can declare it as a component in your
design. The Verilog module port definitions
for a D-Flip-Flop is given in Fig. 7.35. Keep
in mind that this D-Flip-Flop has an active
LOW reset. This means that when the reset
line is pulled to a 0, the outputs will go to
Q ¼ 0, Qn ¼ 1. When the reset line is LOW, Fig. 7.35
the incoming clock is ignored. Once the reset D-Flip-Flop module definition
line goes HIGH, the D-Flip-Flop resumes
264 • Chapter 7: Sequential Logic Design

Section 7.1: Sequential Logic Storage 7.1.11 For the timing diagram shown in Fig. 7.38,
draw the outputs Q and Qn for a rising edge
Devices
triggered D-flip-flop with active LOW.
7.1.1 What does the term metastability refer to in a
sequential storage device?
7.1.2 What does the term bistable refer to in a
sequential storage device?
7.1.3 You are given a cross-coupled inverter pair in
which all nodes are set to Vcc/2. Why will this
configuration always move to a more stable
state?
7.1.4 An SR Latch essentially implements the same
cross-coupled feedback loop to store informa-
tion as in a cross-coupled inverter pair. What is
the purpose of using NOR gates instead of
inverters in the SR Latch configuration?
7.1.5 Why isn’t the input condition S ¼ R ¼ 1 used in Fig. 7.38
an SR Latch? D-Flip-Flop timing diagram exercise 3
7.1.6 How will the output Q behave in an SR Latch if
Section 7.2: Sequential Logic Timing
the inputs continuously switch between S ¼ 0,
R ¼ 1 and S ¼ 1, R ¼ 1 every 10 ns? Considerations
7.1.7 How do D-flip-flops enable synchronous 7.2.1 What timing specification is violated in a D-flip-
systems? flop when the data is not held long enough
before the triggering clock edge occurs?
7.1.8 What signal in the D-flip-flop in Fig. 7.35 has
the highest priority? 7.2.2 What timing specification is violated in a D-flip-
flop when the data is not held long enough after
7.1.9 For the timing diagram shown in Fig. 7.36,
the triggering clock edge occurs?
draw the outputs Q and Qn for a rising edge
triggered D-flip-flop with active LOW. 7.2.3 What is the timing specification for a D-flip-flop
that describes how long after the triggering
clock edge occurs that the new data will be
present on the Q output?
7.2.4 What is the timing specification for a D-flip-flop
that describes how long after the device goes
metastable that the outputs will settle to known
states.
7.2.5 If the Q output of a D-flip-flop is driving the D
input of another D-flip-flop from the same logic
family, can the hold time be ignored if it is less
than the clock-to-Q delay? Provide an expla-
nation as to why or why not.

Fig. 7.36 Section 7.3: Common Circuits Based on

D-Flip-Flop timing diagram exercise 1 Sequential Storage Devices
7.3.1 In a Toggle Flop (T-flop) configuration, the Qn
7.1.10 For the timing diagram shown in Fig. 7.37, output of the D-flip-flop is routed back to the D
draw the outputs Q and Qn for a rising edge input. This can lead to a hold time violation if
triggered D-flip-flop with active LOW. the output arrives on the input too quickly.
Under what condition(s) is a hold time violate
not an issue?
7.3.2 In a Toggle Flop (T-flop) configuration, what
timing specifications dictate how quickly the
next edge of the incoming clock can occur?
7.3.3 One drawback of a ripple counter is that the
delay through the cascade of D-flip-flops can
become considerable for large counters. At
what point does the delay of a ripple counter
prevent it from being useful?
7.3.4 A common use of a ripple counter is in the
creation of a 2n programmable clock divider.
In a ripple counter, bit(0) has a frequency that
Fig. 7.37 is exactly 1/2 of the incoming clock, bit(1) has a
D-Flip-Flop timing diagram exercise 2 frequency that is exactly 1/4 of the incoming
 Exercise Problems • 265

clock, bit(2) has a frequency that is exactly 1/8 (a) How many D-Flip-Flops will this machine
of the incoming clock, etc. This behavior can take if the states are encoded in binary?
be exploited to create a divided down output (b) How many D-Flip-Flops will this machine
clock that is divided by multiples of 2n by take if the states are encoded in gray
selecting a particular bit of the counter. The code?
typical configuration of this programmable
clock divider is to route each bit of the counter (c) How many D-Flip-Flops will this machine
to an input of a multiplexer. The select lines take if the states are encoded in one-
going to the multiplexer choose which bit of hot?
the counter are used as the divided down
clock output. This architecture is shown in
Fig. 7.39. Design a Verilog model to implement
the programmable clock divider shown in this
figure. Use the module port definition provided
in this figure for your design. Use a 4-bit ripple
counter to produce four divided versions of the
clock (1/2, 1/4, 1/8, and 1/16). Your system will
take in two select lines that will choose which
version of the clock is to be routed to the out-
put. Instantiate the D-flip-flop model provided
to implement the ripple counter. Implement the
4-to-1 multiplexer using continuous assign-
ment. The multiplexer does not need to be its
own sub-system.

Fig. 7.40
FSM 1 state diagram

7.4.2 For the state diagram in Fig. 7.40, is this a

Mealy or Moore machine?
7.4.3 Design the finite state machine circuitry by
hand to implement the behavior described by
the state diagram in Fig. 7.40. Name the cur-
rent state variables Q1_cur and Q0_cur and
name the next state variables Q1_nxt and
Q0_nxt. Use the following state codes:
Start ¼ “00”
Midway ¼ “01”
Done ¼ “10”
(a) What is the next state logic expression
for Q1_nxt?
(b) What is the next state logic expression
for Q0_nxt?
(c) What is the output logic expression for
Dout?
(d) Draw the final logic diagram for this
Fig. 7.39 machine.
Programmable clock module definition 7.4.4 Design a Verilog model to implement the
behavior described by the state diagram in
7.3.5 What phenomenon causes switch bounce in a Fig. 7.40. Use the module port definition
SPST switch? provided in Fig. 7.41 for your design. Name
the current state variables Q1_cur and
7.3.6 What two phenomena causes switch bounce in
Q0_cur and name the next state variables
a SPDT switch?
Q1_nxt and Q0_nxt. Instantiate the D-Flip-
Flop model provided to implement your state
Section 7.4: Finite State Machines memory. Use continuous assignment with logi-
7.4.1 For the state diagram in Fig. 7.40, answer the cal operators for the implementation of your
following questions regarding the number of next state and output logic.
D-Flip-Flops needed to implement the state (a) How many D-Flip-Flops will this machine
memory of the finite state machine. take if the states are encoded in binary?
266 • Chapter 7: Sequential Logic Design

(b) How many D-Flip-Flops will this machine

take if the states are encoded in gray
code?
(c) How many D-Flip-Flops will this machine
take if the states are encoded in one-
hot?

Fig. 7.41
FSM 1 module definition
Fig. 7.42
FSM 2 state diagram
7.4.5 Design a Verilog model to implement the
behavior described by the state diagram in 7.4.8 For the state diagram in Fig. 7.42, is this a
Fig. 7.40. Use the module port definition Mealy or Moore machine?
provided in Fig. 7.41 for your design. Name 7.4.9 Design the finite state machine circuitry by
the current state variables Q1_cur and hand to implement the behavior described by
Q0_cur and name the next state variables the state diagram in Fig. 7.42. Name the cur-
Q1_nxt and Q0_nxt. Instantiate the D-Flip- rent state variables Q1_cur and Q0_cur and
Flop model provided to implement your state name the next state variables Q1_nxt and
memory. Use continuous assignment with con- Q0_nxt. Also, use the following state codes:
ditional operators for the implementation of
S0 ¼ “00”
your next state and output logic.
S1 ¼ “01”
7.4.6 Design a Verilog model to implement the
behavior described by the state diagram in S2 ¼ “10”
Fig. 7.40. Use the module port definition S3 ¼ “11”
provided in Fig. 7.41 for your design. Name
(a) What is the next state logic expression
the current state variables Q1_cur and
for Q1_nxt?
Q0_cur and name the next state variables
Q1_nxt and Q0_nxt. Instantiate the D-Flip- (b) What is the next state logic expression
Flop model provided to implement your state for Q0_nxt?
memory. Use User-Defined Primitives for the (c) What is the output logic expression for
implementation of your next state and output Dout?
logic.
(d) Draw the final logic diagram for this
7.4.7 For the state diagram in Fig. 7.42, answer the machine.
following questions regarding the number of D-
7.4.10 Design a Verilog model to implement the
Flip-Flops needed to implement the state
behavior described by the state diagram in
memory of the finite state machine.
Fig. 7.42. Use the module port definition
(a) How many D-Flip-Flops will this machine provided in Fig. 7.43 for your design. Name
take if the states are encoded in binary? the current state variables Q1_cur and
(b) How many D-Flip-Flops will this machine Q0_cur and name the next state variables
take if the states are encoded in gray Q1_nxt and Q0_nxt. Instantiate the D-Flip-
code? Flop model provided to implement your state
memory. Use continuous assignment with logi-
(c) How many D-Flip-Flops will this machine
cal operators for the implementation of your
take if the states are encoded in one-
next state and output logic.
hot?
 Exercise Problems • 267

dimes and dispense a product anytime the

customer has entered 20 cents. Your FSM
has two inputs, Nin and Din. Nin is asserted
whenever the customer enters a nickel while
Din is asserted anytime the customer enters a
dime. Your FSM has two outputs, Dispense
and Change. Dispense is asserted anytime
the customer has entered at least 20 cents
and Change is asserted anytime the customer
has entered more than 20 cents and needs a
nickel in change.
Fig. 7.43
(a) Provide the state diagram for this FSM.
FSM 2 module definition
(b) Encode your states using binary
encoding. How many D-Flip-Flops does
7.4.11 Design a Verilog model to implement the
it take to implement the state memory for
behavior described by the state diagram in
this FSM?
Fig. 7.42. Use the module port definition
provided in Fig. 7.43 for your design. Name (c) Provide the state transition table for
the current state variables Q1_cur and this FSM.
Q0_cur and name the next state variables (d) Synthesize the combinational logic
Q1_nxt and Q0_nxt. Instantiate the D-Flip- expressions for the next state logic.
Flop model provided to implement your state
(e) Synthesize the combinational logic
memory. Use continuous assignment with logi-
expressions for the output logic.
cal operators for the implementation of your
next state and output logic. (f) Is this machine a Mealy or Moore
machine?
7.4.12 Design a Verilog model to implement the
behavior described by the state diagram in (g) Draw the logic diagram for this FSM.
Fig. 7.42. Use the module port definition 7.4.15 Design a finite state machine by hand that
provided in Fig. 7.43 for your design. Name controls a traffic light at the intersection of a
the current state variables Q1_cur and busy highway and a seldom used side road.
Q0_cur and name the next state variables You will be designing the control signals for just
Q1_nxt and Q0_nxt. Instantiate the D-Flip- the red, yellow, and green lights facing the
Flop model provided to implement your state highway. Under normal conditions, the high-
memory. Use User-Defined Primitives for the way has a green light. The side road has a
implementation of your next state and output car detector that indicates when a car pulls up
logic. by asserting a signal called CAR. When CAR is
7.4.13 Design a 4-bit serial bit sequence detector by asserted, you will change the highway traffic
hand similar to the one described in Example light from green to yellow. Once yellow, you will
7.9. The input to your state detector is called always go to red. Once in the red position, a
DIN and the output is called FOUND. Your built in timer will begin a countdown and pro-
detector will assert FOUND anytime there is a vide your controller a signal called TIMEOUT
4-bit sequence of “0101”. For all other input when 15 s has passed. Once TIMEOUT is
sequences the output is not asserted. asserted, you will change the highway traffic
light back to green. Your system will have three
(a) Provide the state diagram for this FSM.
outputs GRN, YLW, and RED that control when
(b) Encode your states using binary the highway facing traffic lights are on (1 ¼ ON,
encoding. How many D-Flip-Flops does 0 ¼ OFF).
it take to implement the state memory for
(a) Provide the state diagram for this FSM.
this FSM?
(b) Encode your states using binary
(c) Provide the state transition table for
encoding. How many D-Flip-Flops does
this FSM.
it take to implement the state memory for
(d) Synthesize the combinational logic this FSM?
expressions for the next state logic.
(c) Provide the state transition table for
(e) Synthesize the combinational logic this FSM.
expression for the output logic.
(d) Synthesize the combinational logic
(f) Is this machine a Mealy or Moore expressions for the next state logic.
machine?
(e) Synthesize the combinational logic
(g) Draw the logic diagram for this FSM. expressions for the output logic.
7.4.14 Design a 20 cent vending machine controller (f) Is this machine a Mealy or Moore
by hand similar to the one described in Exam- machine?
ple 7.12. Your controller will take in nickels and
(g) Draw the logic diagram for this FSM.
268 • Chapter 7: Sequential Logic Design

Section 7.5: Counters (f) What is the output logic expression for
Count(0)?
7.5.1 Design a 3-bit binary up counter by hand. This
state machine will need eight states and (g) Draw the logic diagram for this counter.
require three bits for the state variable codes. 7.5.4 Design a Verilog model for a 3-bit binary
Name the current state variables Q2_cur, up/down counter. Instantiate the D-Flip-Flop
Q1_cur, and Q0_cur and the next state model provided to implement your state mem-
variables Q2_nxt, Q1_nxt, and Q0_nxt. The ory. Use whatever concurrent modeling
output of your counter will be a 3-bit vector approach you wish to model the next state
called Count. and output logic. Use the module port definition
(a) What is the next state logic expression provided in Fig. 7.45 for your design.
for Q2_nxt?
(b) What is the next state logic expression
for Q1_nxt?
(c) What is the next state logic expression
for Q0_nxt?
(d) What is the output logic expression for
Count(2)?
(e) What is the output logic expression for
Count(1)?
Fig. 7.45
(f) What is the output logic expression for
Count(0)? 3-Bit binary up/down counter module definition
(g) Draw the logic diagram for this counter.
7.5.5 Design a 3-bit gray code up counter by hand.
7.5.2 Design a Verilog model for a 3-bit binary up This state machine will need eight states and
counter. Instantiate the D-Flip-Flop model require three bits for the state variable codes.
provided to implement your state memory. Name the current state variables Q2_cur,
Use whatever concurrent modeling approach Q1_cur, and Q0_cur and the next state variables
you wish to model the next state and output Q2_nxt, Q1_nxt, and Q0_nxt. The output of your
logic. Use the module port definition provided counter will be a 3-bit vector called Count.
in Fig. 7.44 for your design.
(a) What is the next state logic expression
for Q2_nxt?
(b) What is the next state logic expression
for Q1_nxt?
(c) What is the next state logic expression
for Q0_nxt?
(d) What is the output logic expression for
Count(2)?
Fig. 7.44 (e) What is the output logic expression for
3-Bit binary up counter module definition Count(1)?
(f) What is the output logic expression for
7.5.3 Design a 3-bit binary up/down counter by hand. Count(0)?
The counter will have an input called “Up” that (g) Draw the logic diagram for this counter.
will dictate the direction of the counter. When
Up ¼ 1, the counter should increment and when 7.5.6 Design a Verilog model for a 3-bit gray code up
Up ¼ 0 it should decrement. This state machine counter. Instantiate the D-Flip-Flop model
will need eight states and require three bits for provided to implement your state memory.
the state variable codes. Name the current state Use whatever concurrent modeling approach
variables Q2_cur, Q1_cur, and Q0_cur and the you wish to model the next state and output
next state variables Q2_nxt, Q1_nxt, and logic. Use the module port definition provided
Q0_nxt. The output of your counter will be a in Fig. 7.46 for your design.
3-bit vector called Count.
(a) What is the next state logic expression
for Q2_nxt?
(b) What is the next state logic expression
for Q1_nxt?
(c) What is the next state logic expression
for Q0_nxt?
(d) What is the output logic expression for Fig. 7.46
Count(2)? 3-Bit gray code up counter module definition
(e) What is the output logic expression for
Count(1)?
 Exercise Problems • 269

7.5.7 Design a 3-bit gray code up/down counter by 7.6.5 If a finite state machine has eight unique states
hand. The counter will have an input called that are encoded in binary and all D-flip-flops
“Up” that will dictate the direction of the used for the state memory use their reset lines,
counter. When Up ¼ 1, the counter should what is the state code that the machine will go
increment and when Up ¼ 0 it should decre- to upon reset?
ment. This state machine will need eight states
and require three bits for the state variable Section 7.7: Sequential Logic Analysis
codes. Name the current state variables
Q2_cur, Q1_cur, and Q0_cur and the next 7.7.1 For the finite state machine logic diagram in
state variables Q2_nxt, Q1_nxt, and Q0_nxt. Fig. 7.48, give the next state logic expression
The output of your counter will be a 3-bit vector for Q_nxt.
called Count.
(a) What is the next state logic expression
for Q2_nxt?
(b) What is the next state logic expression
for Q1_nxt?
(c) What is the next state logic expression
for Q0_nxt?
(d) What is the output logic expression for Fig. 7.48
Count(2)? Sequential logic analysis 1
(e) What is the output logic expression for
Count(1)? 7.7.2 For the finite state machine logic diagram in
(f) What is the output logic expression for Fig. 7.48, give the output logic expression
Count(0)? for Tout.
(g) Draw the logic diagram for this counter. 7.7.3 For the finite state machine logic diagram in
7.5.8 Design a Verilog model for a 3-bit gray code Fig. 7.48, give the state transition table.
up/down counter. Instantiate the D-Flip-Flop 7.7.4 For the finite state machine logic diagram in
model provided to implement your state mem- Fig. 7.48, give the state diagram.
ory. Use whatever concurrent modeling 7.7.5 For the finite state machine logic diagram in
approach you wish to model the next state Fig. 7.48, give the maximum clock frequency.
and output logic. Use the module port definition
provided in Fig. 7.47 for your design. 7.7.6 For the finite state machine logic diagram in
Fig. 7.49, give the next state logic expression
for Q_nxt.

Fig. 7.47
3-Bit gray code up/down counter module
definition Fig. 7.49
Sequential logic analysis 2
Section 7.6: Finite State Machine’s Reset
Condition 7.7.7 For the finite state machine logic diagram in
Fig. 7.49, give the output logic expression for F.
7.6.1 Are resets typically synchronous or
asynchronous? 7.7.8 For the finite state machine logic diagram in
Fig. 7.49, give the state transition table.
7.6.2 Why is it necessary to have a reset/preset
condition in a finite state machine? 7.7.9 For the finite state machine logic diagram in
Fig. 7.49, give the state diagram.
7.6.3 How does the reset/preset condition corre-
spond to the behavior described in the state 7.7.10 For the finite state machine logic diagram in
diagram? Fig. 7.49, give the maximum clock frequency.
7.6.4 When is it necessary to also use the preset line 7.7.11 For the finite state machine logic diagram in
(s) of a D-flip-flop instead of just the reset line Fig. 7.50, give the next state logic expressions
(s) when implementing the state memory of a for Q1_nxt and Q0_nxt.
finite state machine?
270 • Chapter 7: Sequential Logic Design

7.7.12 For the finite state machine logic diagram in

Fig. 7.50, give the output logic expression for
Return.
7.7.13 For the finite state machine logic diagram in
Fig. 7.50, give the state transition table.
7.7.14 For the finite state machine logic diagram in
Fig. 7.50, give the state diagram.
7.7.15 For the finite state machine logic diagram in
Fig. 7.50, give the maximum clock frequency.

Fig. 7.50
Sequential logic analysis 3
Chapter 8: Verilog (Part 2)
In Chap. 5 Verilog was presented as a way to describe the behavior of concurrent systems. The
modeling techniques presented were appropriate for combinational logic because these types of circuits
have outputs dependent only on the current values of their inputs. This means a model that continuously
performs signal assignments provides an accurate model of this circuit behavior. In Chap. 7 sequential
logic storage devices were presented that did not continuously update their outputs based on the
instantaneous values of their inputs. Instead, sequential storage devices only update their outputs
based upon an event, most often the edge of a clock signal. The modeling techniques presented in
Chap. 5 are unable to accurately describe this type of behavior. In this chapter, we describe the Verilog
constructs to model signal assignments that are triggered by an event in order to accurately model
sequential logic. We can then use these techniques to describe more complex sequential logic circuits
such as finite state machines and register transfer level systems. This chapter will also present how to
create test benches and look at more advanced features that are commonly used in Verilog to model
modern systems. The goal of this chapter is to give an understanding of the full capability of hardware
description languages.
Learning Outcomes—After completing this chapter, you will be able to:

8.1 Describe the behavior of Verilog procedural assignment and how they are used to model
sequential logic circuits.
8.2 Model combinational logic circuits using a Verilog procedural assignment and conditional
programming constructs.
8.3 Describe the functionality of common Verilog system tasks.
8.4 Design a Verilog test bench to verify the functional operation of a system.

8.1 Procedural Assignment

Verilog uses procedural assignment to model signal assignments that are based on an event. An
event is most commonly a transition of a signal. This provides the ability to model sequential logic circuits
such as D-flip-flops and finite state machines by triggering assignments off of a clock edge. Procedural
assignments can only drive variable data types (i.e., reg, integer, real, and time), thus they are ideal for
modeling storage devices. Procedural signal assignments can be evaluated in the order they are listed,
thus they are able to model sequential assignments.
A procedural assignment can also be used to model combinational logic circuits by making signal
assignments when any of the inputs to the model change. Despite the left-hand-side of the assignment
not being able to be of type wire in procedural assignment, modern synthesizers will recognize properly
designed combinational logic models and produce the correct circuit implementation. Procedural assign-
ment also supports standard programming constructs such as if-else decisions, case statements, and
loops. This makes procedural assignment a powerful modeling approach in Verilog and is the most
common technique for designing digital systems and creating test benches.

8.1.1 Procedural Blocks

All procedural signal assignments must be enclosed within a procedural block. Verilog has two types
of procedural blocks, initial and always.

# Springer International Publishing AG 2017 271

B.J. LaMeres, Introduction to Logic Circuits & Logic Design with Verilog,
DOI 10.1007/978-3-319-53883-9_8
272 • Chapter 8: Verilog (Part 2)

8.1.1.1 Initial Blocks

An initial block will execute all of the statements embedded within it one time at the beginning of the
simulation. An initial block is not used to model synthesizable behavior. It is instead used within test
benches to either set the initial values of repetitive signals or to model the behavior of a signal that only
has a single set of transitions. The following is the syntax for an initial block.

initial
begin // an optional “: name” can be added after the begin keyword
signal_assignment_1
signal_assignment_2
:
end

Let’s look at a simple model of how an initial block is used to model the reset line in a test bench. In
the following example, the signal “Reset_TB” is being driven into a DUT. At the beginning of the
simulation, the initial value of Reset_TB is set to a logic zero. The second assignment will take place
after a delay of 15 time units. The second assignment statement sets Reset_TB to a logic one. The
assignments in this example are evaluated in sequence in the order they are listed due to the delay
operator. Since the initial block executes only once, Reset_TB will stay at the value of its last assignment
for the remainder of the simulation.
Example:
initial
begin
Reset_TB ¼ 1’b0;
#15 Reset_TB ¼ 1’b1;
end

8.1.1.2 Always Blocks

An always block will execute forever, or for the duration of the simulation. An always block can be
used to model synthesizable circuits in addition to non-synthesizable behavior in test benches. The
following is the syntax for an always block.

always
begin
signal_assignment_1
signal_assignment_2
:
end

Let’s look at a simple model of how an always block can be used to model a clock line in a test
bench. In the following example, the value of the signal Clock_TB will continuously change its logic value
every 10 time units.
Example:

always
begin
#10 Clock_TB ¼ ~Clock_TB;
end

By itself, the above always block will not work because when the simulation begins, Clock_TB does
not have an initial value so the simulator will not know what the value of Clock_TB is at time zero. It
will also not know what the output of the negation operation (~) will be at time unit 10. The following
example shows the correct way of modeling a clock signal using a combination of initial and always
 8.1 Procedural Assignment • 273

blocks. Verilog allows assignments to the same variable from multiple procedural blocks, so the following
example is valid. Note that when the simulation begins, Clock_TB is assigned a logic zero. This provides
a known value for the signal at time zero and also allows the always block negation to have a
deterministic value. The example below will create a clock signal that will toggle every 10 time units.

Example:
initial
begin
Clock_TB ¼ 1’b0;
end

always
begin
#10 Clock_TB ¼ ~Clock_TB;
end

8.1.1.3 Sensitivity Lists

A sensitivity list is used in conjunction with a procedural block to trigger when the assignments within
the block are executed. The symbol @ is used to indicate a sensitivity list. Signals can then be listed
within parenthesis after the @ symbol that will trigger the procedural block. The following is the base
syntax for a sensitivity list.
always @ (signal1, signal2)
begin
signal_assignment_1
signal_assignment_2
:
end

In this syntax, any transition on any of the signals listed within the parenthesis will cause the always
block to trigger and all of its assignments to take place one time. After the always block ends, it will await
the next signal transition in the sensitivity list to trigger again. The following example shows how to model
a simple 3-input AND gate. In this example, any transition on inputs A, B, or C will cause the block to
trigger and the assignment to F to occur.

Example:
always @ (A, B, C)
begin
F ¼ A & B & C;
end

Verilog also supports keywords to limit triggering of the block to only rising edge or falling edge
transitions. The keywords are posedge and negedge. The following is the base syntax for an edge
sensitive block. In this syntax, only rising edge transitions on signal1 or falling edge transitions on signal2
will cause the block to trigger.
always @ (posedge signal1, negedge signal2)
begin
signal_assignment_1
signal_assignment_2
:
end

Sensitivity lists can also contain Boolean operators to more explicitly describe behavior. The
following syntax is identical to the syntax above.
274 • Chapter 8: Verilog (Part 2)

always @ (posedge signal1 or negedge signal2)

begin
signal_assignment_1
signal_assignment_2
:
end

The ability to model edge sensitivity allows us to model sequential circuits. The following example
shows how to model a simple D-flip-flop.
Example:

always @ (posedge Clock)

begin
Q ¼ D; // Note: This model does not include a reset.
end

In Verilog-2001, the syntax to support sensitivity lists that will trigger based on any signal listed on
the right-hand-side of any assignment within the block was added. This syntax is @*. The following
example how to use this modeling approach to model a 3-input AND gate.

Example:
always @*
begin
F ¼ A & B & C;
end

8.1.2 Procedural Statements

There are two kinds of signal assignments that can be used within a procedural block, blocking and
non-blocking.

8.1.2.1 Blocking Assignments

A blocking assignment is denoted with the = symbol and the evaluation and assignment of each
statement takes place immediately. Each assignment within the block is executed in parallel. When this
behavior is coupled with a sensitivity list that contains all of the inputs to the system, this approach can
model synthesizable combinational logic circuits. This approach provides the same functionality as
continuous assignments outside of a procedural block. The reason that designers use blocking
assignments instead of continuous assignment is that more advanced programming constructs are
supported within Verilog procedural blocks. These will be covered in the next section. Example 8.1
shows how to use blocking assignments within a procedural block to model a combinational logic circuit.
 8.1 Procedural Assignment • 275

Example 8.1
Using blocking assignments to model combinational logic

8.1.2.2 Non-blocking Assignments

A non-blocking assignment is denoted with the <= symbol. When using non-blocking assignments,
the assignment to the target signal is deferred until the end of the procedural block. This allows the
assignments to be executed in the order they are listed in the block without cascading interim
assignments through the list. When this behavior is coupled with triggering the block off of a clock signal,
this approach can model synthesizable sequential logic circuits. Example 8.2 shows an example of using
non-blocking assignments to model a sequential logic circuit.

Example 8.2
Using non-blocking assignments to model sequential logic

The difference between blocking and non-blocking assignments is subtle and is often one of the
most difficult concepts to grasp when first learning Verilog. One source of confusion comes from the fact
that blocking and non-blocking assignments can produce the same results when they either contains a
276 • Chapter 8: Verilog (Part 2)

single assignment or a list of assignments that don’t have any signal interdependencies. A signal
interdependency refers to when a signal that is the target of an assignment (i.e., on the LHS of an
assignment) is used as an argument (i.e., on the RHS of an assignment) in subsequent statements.
Example 8.3 shows two models that produce the same results regardless of whether a blocking or
non-blocking assignment is used.

Example 8.3
Identical behavior when using blocking vs. non-blocking assignments

When a list of statements within a procedural block does have signal interdependencies, blocking
and non-blocking assignments will have different behavior. Example 8.4 shows how signal
interdependencies will cause different behavior between blocking and non-blocking assignments. In
this example, all inputs are listed in the sensitivity list with the intent of modeling combinational logic.
 8.1 Procedural Assignment • 277

Example 8.4
Different behavior when using blocking vs. non-blocking assignments (1)

Example 8.5 shows another case where signal interdependencies will cause different behavior
between blocking and non-blocking assignments. In this example, the procedural block is triggered by
the rising edge of a clock signal with the intent of modeling two stages of sequential logic.
278 • Chapter 8: Verilog (Part 2)

Example 8.5
Different behavior when using blocking vs. non-blocking assignments (2)

While the behavior of these procedural assignments can be confusing, there are two design
guidelines that can make creating accurate, synthesizable models straightforward. They are:
1. When modeling combinational logic, use blocking assignments and list every input in the
sensitivity list.
2. When modeling sequential logic, use non-blocking assignments and only list the clock and
reset lines (if applicable) in the sensitivity list.
 8.1 Procedural Assignment • 279

8.1.3 Statement Groups

A statement group refers to how the statements in a block are processed. Verilog supports two types
of statement groups: begin/end and fork/join. When using begin/end, all statements enclosed within the
group will be evaluated in the order they are listed. When using a fork/join, all statements enclosed within
the group will be evaluated in parallel. When there is only one statement within procedural block, a
statement group is not needed. For multiple statements in a procedural block, a statement group is
required. Statement groups can contain an optional name that is appended after the first keyword
preceded by a “:”. Example 8.6 shows a graphical depiction of the difference between begin/end and
fork/join groups. Note that this example also shows the syntax for naming the statement groups.

Example 8.6
Behavior of statement groups begin/end vs. fork/join

8.1.4 Local Variables

Local variables can be declared within a procedural block. The statement group must be named and
the variables will not be visible outside of the block. Variables can only be of variable type.

Example:
initial
begin: stim_block // it is required to name the block when declaring
local variables
integer i; // local variables can only be of variable type
i¼2;
end
280 • Chapter 8: Verilog (Part 2)

CONCEPT CHECK

CC8.1 If a model of a combinational logic circuit excludes one of its inputs from the sensitivity list,
what is the implied behavior?
(A) A storage element because the output will be held at its last value when the
unlisted input transitions.
(B) An infinite loop.
(C) A don’t care will be used to form the minimal logic expression.
(D) Not applicable because this syntax will not compile.

8.2 Conditional Programming Constructs

One of the more powerful features that procedural blocks provide in Verilog is the ability to use
conditional programming constructs such as if-else decisions, case statements, and loops. These
constructs are only available within a procedural block and can be used to model both combinational
and sequential logic.

8.2.1 if-else Statements

An if-else statement provides a way to make conditional signal assignments based on Boolean
conditions. The if portion of statement is followed by a Boolean condition that if evaluated TRUE will
cause the signal assignment listed after it to be performed. If the Boolean condition is evaluated FALSE,
the statements listed after the else portion are executed. If multiple statements are to be executed in
either the if or else portion, then the statement group keywords begin/end need to be used. If only one
statement is to be executed, then the statement group keywords are not needed. The else portion of the
statement is not required and if omitted, no assignment will take place when the Boolean condition is
evaluated FALSE. The syntax for an if-else statement is as follows:

if (<boolean_condition>)
true_statement
else
false_statement

The syntax for an if-else statement with multiple true/false statements is as follows:

if (<boolean_condition>)
begin
true_statement_1
true_statement_2
end
else
begin
false_statement_1
false_statement_2
end

If more than one Boolean condition is required, additional if-else statements can be embedded
within the else clause of the preceding if statement. The following shows an example of if-else
statements implementing two Boolean conditions.
 8.2 Conditional Programming Constructs • 281

if (<boolean_condition_1>)
true_statement_1
else if (<boolean_condition_2>)
true_statement_2
else
false_statement

Let’s look at using an if-else statement to describe the behavior of a combinational logic circuit.
Recall that a combinational logic circuit is one in which the output depends on the instantaneous values
of the inputs. This behavior can be modeled by placing all of the inputs to the circuit in the sensitivity list of
an always block and using blocking assignments. Using this approach, a change on any of the inputs in
the sensitivity list will trigger the block and the assignments will take place immediately. Example 8.7
shows how to model a 3-input combinational logic circuit using if-else statements within a procedural
always block.

Example 8.7
Using if-else statements to model combinational logic

8.2.2 case Statements

A case statement is another technique to model signal assignments based on Boolean conditions.
As with the if-else statement, a case statement can only be used inside of a procedural block. The
statement begins with the keyword case followed by the input signal name that assignments will be
based off of enclosed within parenthesis. The case statement can be based on multiple input signal
names by concatenating the signals within the parenthesis. Then a series of input codes followed by the
corresponding assignment is listed. The keyword default can be used to provide the desired signal
assignment for any input codes not explicitly listed. When multiple input conditions have the same
assignment statement, they can be listed on the same line comma-delimited to save space. The keyword
endcase is used to denote the end of the case statement. The following is the syntax for a case
statement.
282 • Chapter 8: Verilog (Part 2)

case (<input_name>)
input_val_1 : statement_1
input_val_2 : statement_2
:
input_val_n : statement_n
default : default_statement
endcase

Example 8.8 shows how to model a 3-input combinational logic circuit using a case statement within
a procedural block. Note in this example the inputs are scalars so they must be concatenated so that the
input values can be listed as 3-bit vectors. In this example, there are three versions of the model
provided. The first explicitly lists out all binary input codes. This approach is more readable because it
mirrors a truth table form. The second approach only lists the input codes corresponding to an output of
one and uses the default clause to handle all other input codes. The third approach shows how to list
multiple input codes with the same assignment on the same line using a comma-delimited series.

Example 8.8
Using case statements to model combinational logic
 8.2 Conditional Programming Constructs • 283

If-else statements can be embedded within a case statement and, conversely, case statements can
be embedded within an if-else statement.

8.2.3 casez and casex Statements

Verilog provides two additional case statements that support don’t cares in the input conditions. The
casez statement allows the symbols? and Z to represent a don’t care. The casex statement extends the
casez statement by also interpreting X as a don’t care. Care should be taken when using the casez and
casex statement as it is easy to create unintended logic when using don’t cares in the input codes.

8.2.4 forever Loops

A loop within Verilog provides a mechanism to perform repetitive assignments infinitely. This is
useful in test benches for creating stimulus such as clocks or other periodic signals. We have already
covered a looping construct in the form of an always block. An always block provides a loop with a
starting condition. Verilog provides additional looping constructs to model more sophisticated behavior.
All looping constructs must reside with a procedural block.
The simplest looping construct is the forever loop. As with other conditional programming
constructs, if multiple statements are associated with the forever loop they must be enclosed within a
statement group. If only one statement is used the statement group is not needed. A forever loop within
an initial block provides identical behavior as an always loop without a sensitivity loop. It is important to
provide a time step event or delay within a forever loop or it will cause a simulation to hang. The following
is the syntax for a forever loop in Verilog.

forever
begin
statement_1
statement_2
:
statement_n
end

Consider the following example of a forever loop that generates a clock signal (CLK) with a period of
10 time units. In this example, the forever loop is embedded within an initial block. This allows the initial
value of CLK to be set to zero upon the beginning of the simulation. Once the forever loop is entered, it
will execute indefinitely. Notice that since there is only one statement after the forever keyword, a
statement group (i.e., begin/end) is not needed.

Example:
initial
begin
CLK ¼ 0;

forever
#10 CLK ¼ ~CLK;

end

8.2.5 while Loops

A while loop provides a looping structure with a Boolean condition that controls its execution. The
loop will only execute as long as the Boolean condition is evaluated true. The following is the syntax for a
Verilog while loop.
284 • Chapter 8: Verilog (Part 2)

while (<boolean_condition>)
begin
statement_1
statement_2
:
statement_n
end

Let’s implement the previous example of a loop that generates a clock signal (CLK) with a period of
10 time units as long as EN ¼ 1. The TRUE Boolean condition for the while loop is EN ¼ 1. When EN ¼ 0,
the while loop will be skipped. When the loop becomes inactive, CLK will hold its last assigned value.
Example:

initial
begin
CLK ¼ 0;

while (EN ¼¼ 1)
#10 CLK ¼ ~CLK;

end

8.2.6 repeat Loops

A repeat loop provides a looping structure that will execute a fixed number of times. The following is
the syntax for a Verilog repeat loop.

repeat (<number_of_loops>)
begin
statement_1
statement_2
:
statement_n
end

Let’s implement the previous example of a loop that generates a clock signal (CLK) with a period of
10 time units, except this time we’ll use a repeat loop to only produce 10 clock transitions, or 5 full periods
of CLK.
Example:
initial
begin
CLK ¼ 0;
repeat (10)
#10 CLK ¼ ~CLK;
end

8.2.7 for Loops

A for loop provides the ability to create a loop that can automatically update an internal variable.
A loop variable within a for loop is altered each time through the loop according to a step assignment. The
starting value of the loop variable is provided using an initial assignment. The loop will execute as long as
a Boolean condition associated with the loop variable is TRUE. The following is the syntax for a Verilog
for loop:
 8.2 Conditional Programming Constructs • 285

for (<initial_assignment>; <Boolean_condition>; <step_assignment>)

begin
statement_1
statement_2
:
statement_n
end

The following is an example of creating a simple counter using the loop variable. The loop variable
i was declared as an integer prior to this block. The signal Count is also of type integer. The loop variable
will start at 0 and increment by 1 each time through the loop. The loop will execute as long as i < 15, or
16 times total. For loops allow the loop variable to be used in signal assignments within the block.
Example:

initial
begin
for (i¼0; i<15; i¼i+1)
#10 Count ¼ i;
end

8.2.8 disable
Verilog provides the ability to stop a loop using the keyword disable. The disable function only works
on named statement groups. The disable function is typically used after a certain fixed amount of time or
within a conditional construct such as an if-else or case statement that is triggered by a control signal.
Consider the following forever loop example that will generate a clock signal (CLK), but only when an
enable (EN) is asserted. When EN ¼ 0, the loop will disable and the simulation will end.
Example:

initial
begin
CLK ¼ 0;
forever
begin: loop_ex
if (EN ¼¼ 1)
#10 CLK ¼ ~CLK;
else
disable loop_ex; // The group name to be disabled comes after the
keyword
end
end

CONCEPT CHECK

CC8.2 When using an if-else statement to model a combinational logic circuit, is using the else
clause the same as using don’t cares when minimizing a logic expression with a K-map?
(A) Yes. The else clause allows the synthesizer to assign whatever output values are
necessary in order to create the most minimal circuit.
(B) No. The else clause explicitly states the output values for all input codes not listed
in the if portion of the statement. This is the same as filling in the truth table with
specific values for all input codes covered by the else clause and the synthesizer
will create the logic expression accordingly.
286 • Chapter 8: Verilog (Part 2)

8.3 System Tasks

A system task in Verilog is one that is used to insert additional functionality into a model that is not
associated with real circuitry. There are three main groups of system tasks in Verilog: (1) text output;
(2) file input/output; and (3) simulation control. All system tasks begin with a $ and are only used during
simulation. These tasks are ignored by synthesizers so they can be included in real circuit models. All
system tasks must reside within procedural blocks.

8.3.1 Text Output

Text output system tasks are used to print strings and variable values to the console or transcript of a
simulation tool. The syntax follows ANSI C where double quotes (“”) are used denote the text string to be
printed. Standard text can be entered within the string in addition to variables. Variable can be printed in
two ways. The first is to simply list the variable in the system task function outside of the double quotes. In
this usage, the default format to be printed will be decimal unless a task is used with a different default
format. The second way to print a variable is within a text string. In this usage, a unique code is inserted
into the string indicating the format of how to print the value. After the string, a comma separated list of
the variable name(s) is listed that corresponds positionally to the codes within the string. The following
are the most commonly used text output system tasks.

Task Description
$display() Print text string when statement is encountered and append a newline.
$displayb() Same as $display, but default format of any arguments is binary.
$displayo() Same as $display, but default format of any arguments is octal.
$displayh() Same as $display, but default format of any arguments is hexadecimal.
$write() Same as $display, but the string is printed without a newline.
$writeb() Same as $write, but default format of any arguments is binary.
$writeo() Same as $write, but default format of any arguments is octal.
$writeh() Same as $write, but default format of any arguments is hexadecimal.
$strobe() Same as $display, but printing occurs after all simulation events are executed.
$strobeb() Same as $strobe, but default format of any arguments is binary.
$strobeo() Same as $strobe, but default format of any arguments is octal.
$strobeh() Same as $strobe, but default format of any arguments is hexadecimal.
$monitor() Same as $display, but printing occurs when the value of an argument changes.
$monitorb() Same as $monitor, but default format of any arguments is binary.
$monitoro() Same as $monitor, but default format of any arguments is octal.
$monitoron Begin tracking argument changes in subsequent $monitor tasks.
$monitoroff Stop tracking argument changes in subsequent $monitor tasks.

The following is a list of the most common text formatting codes for printing variables within a string.

Code Format
%b Binary values
%o Octal values
%d Decimal values
%h Hexadecimal values
%f Real values using decimal form
%e Real values using exponential form
%t Time values
%s Character strings
%m Hierarchical name of scope (no argument required when printing)
%l Configuration library binding (no argument required when printing)
 8.3 System Tasks • 287

The format letters in these codes are not case sensitive (i.e., %d and %D are equivalent). Each of
these formatting codes can also contain information about truncation of leading and trailing digits.
Rounding will take place when numbers are truncated. The formatting syntax is as follows:

% < number_of_leading_digits > . < number_of_trailing_digits > <format_code_

letter>

There are also a set of string formatting and character escapes that are supported for use with the
text output system tasks.

Code Description
\n Print a new line.
\t Print a tab.
\” Print a quote (“).
\\ Print a backslash (\).
%% Print a percent sign (%).

The following is a set of examples using common text output system tasks. For these examples,
assume two variables have been declared and initialized as follow: A ¼ 3 (integer) and B ¼ 45.6789
(real). Recall that Verilog uses 32-bit codes to represent type integer and real.

Example:
$display("Hello World"); // Will print: Hello World
$display("A ¼ %b", A); // This will print: A ¼ 00000000000000000000000000000011
$display("A ¼ %o", A); // This will print: A ¼ 00000000003
$display("A ¼ %d", A); // This will print: A ¼ 3
$display("A ¼ %h", A); // This will print: A ¼ 00000003
$display("A ¼ %4.0b", A); // This will print: A ¼ 0011

$display("B ¼ %f", B); // This will print: B ¼ 45.678900

$display("B ¼ %2.0f", B); // This will print: B ¼ 46
$display("B ¼ %2.1f", B); // This will print: B ¼ 45.7
$display("B ¼ %2.2f", B); // This will print: B ¼ 45.68

$display("B ¼ %e", B); // This will print: B ¼ 4.567890e+001

$display("B ¼ %1.0e", B); // This will print: B ¼ 5e+001
$display("B ¼ %1.1e", B); // This will print: B ¼ 4.6e+001
$display("B ¼ %2.2e", B); // This will print: B ¼ 4.57e+001

$write("A is ", A, "\n"); // This will print: A is 3

$writeb("A is ", A, "\n"); // This will print: A is 00000000000000000000000000000011
$writeo("A is ", A, "\n"); // Will print: A is 00000000003
$writeh("A is ", A, "\n"); // Will print: A is 00000003

8.3.2 File Input/Output

File I/O system tasks allow a Verilog module to create and/or access data files is the same way files
are handled in ANSI C. This is useful when the results of a simulation are a large and need to be stored in
a file as opposed to viewing in a waveform or transcript window. This is also useful when complex
stimulus vectors are to be read from an external file and driven into a device under test. Verilog supports
the following file I/O system task functions:
288 • Chapter 8: Verilog (Part 2)

Task Description
$fopen() Opens a file and returns a unique file descriptor.
$fclose() Closes the file associated with the descriptor.
$fdisplay() Same as $display but statements are directed to the file descriptor.
$fwrite() Same as $write but statements are directed to the file descriptor.
$fstrobe() Same as $strobe but statements are directed to the file descriptor.
$fmonitor() Same as $monitor but statements are directed to the file descriptor.
$readmemb() Read binary data from file and insert into previously defined memory array.
$readmemh() Read hexadecimal data from file and insert into previously defined memory array.

The $fopen() function will either create and open, or open an existing file. Each file that is opened is
given a unique integer called a file descriptor that is used to identify the file in other I/O functions. The
integer must be declared prior to the first use of $fopen. A file name argument is required and provided
within double quotes. By default, the file is opened for writing. If the file name doesn’t exist, it will be
created. If the file name does exist, it will be overwritten. An optional file_type can be provided that gives
specific action for the file opening including opening an existing file and appending to a file. The following
are the supported codes for $fopen().

$fopen types Description

“r” or “rb” Open file for reading.
“w” or “wb” Create for writing.
“a” or “ab” Open for writing and append to the end of file.
“r+” or “r + b” or “rb+” Open for update, reading or writing file.
“w+” or “w + b” or “wb+” Create for update.
“a+” or “a + b” or “ab+” Open or create for update, append to the end of file.

Once a fie is open, data can be written to it using the $fdisplay(), $fwrite(), $fstrobe(), and
$fmonitor() tasks. These functions require two arguments. The first argument is the file descriptor and
the second is the information to be written. The information follows the same syntax as the I/O system
tasks. The following example shows how to create a file and write data to it. This example will create a
new file called “Data_out.txt” and write two lines of text to it with the values of variables A and B.

Example:
integer A ¼ 3;
real B ¼ 45.6789;
integer FILE_1;

initial
begin
FILE_1 ¼ $fopen("Data_out.txt", "w");
$fdisplay(FILE_1, "A is %d", A);
$fdisplay(FILE_1, "B is %f", B);
$fclose(FILE_1);
end

When reading data from a file, the functions $readmemb() and $readmemh() can be used. These
tasks require that a storage array be declared that the contents of the file can be read into. These tasks
have two arguments, the first being the name of the file and the second being the name of the storage
array to store the file contents into. The following example shows how to read the contents of a file into a
storage array called “memory”. Assume the file contains eight lines, each containing a 3-bit vector. The
vectors start at 000 and increment to 111 and each symbol will be interpreted as binary using the
$readmemb() task. The storage array “memory” is declared to be an 8x3 array of type reg. The
 8.3 System Tasks • 289

$readmemb() task will insert each line of the file into each 3-bit vector location within “memory”. To
illustrate how the data is stored, this example also contains a second procedural block that will print the
contents of the storage element to the transcript.
Example:
reg[2:0] memory[7:0];

initial
begin: Read_Block
$readmemb("Data_in.txt", memory);
end
initial
begin: Print_Block
$display("printing memory %b", memory[0]); // This will print “000”
$display("printing memory %b", memory[1]); // This will print “001”
$display("printing memory %b", memory[2]); // This will print “010”
$display("printing memory %b", memory[3]); // This will print “011”
$display("printing memory %b", memory[4]); // This will print “100”
$display("printing memory %b", memory[5]); // This will print “101”
$display("printing memory %b", memory[6]); // This will print “110”
$display("printing memory %b", memory[7]); // This will print “111”
end

8.3.3 Simulation Control and Monitoring

Verilog also provides a set of simulation control and monitoring tasks. The following are the most
commonly used tasks in this group.

Task Description
$finish() Finishes simulation and exits.
$stop() Halts the simulation and enters an interactive debug mode.
$time() Returns the current simulation time as a 64-bit vector.
$stime() Returns the current simulation time as a 32-bit integer.
$realtime() Returns the current simulation time as a 32-bit real number.
$timeformat() Controls the format used by the %t code in print statements.
The arguments are: (<unit>, <precision>, <suffix>, <min_field_width>)
where:
<unit> 0 ¼ 1 sec
-1 ¼ 100 ms
-2 ¼ 10 ms
-3 ¼ 1 ms
-4 ¼ 100us
-5 ¼ 10us
-6 ¼ 1us
-7 ¼ 100 ns
-8 ¼ 10 ns
-9 ¼ 1 ns
-10 ¼ 100 ps
-11 ¼ 10 ps
-12 ¼ 1 ps
-13 ¼ 100 fs
-14 ¼ 10 fs
15 ¼ 1 fs
290 • Chapter 8: Verilog (Part 2)

<precision > ¼ The number of decimal points to display.

<suffix > ¼ A string to be appended to time to indicate units.
<min_field_width > ¼ The minimum number of characters to display.

The following shows an example of how these tasks can be used.

Example:
initial
begin

$timeformat (-9, 2, "ns", 10);

$display("Stimulus starting at time: %t", $time);

#10 A_TB¼0; B_TB¼0; C_TB¼0;

#10 A_TB¼0; B_TB¼0; C_TB¼1;
#10 A_TB¼0; B_TB¼1; C_TB¼0;
#10 A_TB¼0; B_TB¼1; C_TB¼1;
#10 A_TB¼1; B_TB¼0; C_TB¼0;
#10 A_TB¼1; B_TB¼0; C_TB¼1;
#10 A_TB¼1; B_TB¼1; C_TB¼0;
#10 A_TB¼1; B_TB¼1; C_TB¼1;

$display("Simulation stopping at time: %t", $time);

end

This example will result in the following statements printed to the simulator transcript:
Stimulus starting at time: 0.00 ns
Simulation stopping at time: 80.00 ns

CONCEPT CHECK

CC8.3 How can Verilog system tasks be included in synthesizable circuit models when they provide
inherently unsynthesizable functionality?
(A) They can’t. System tasks can only be used in test benches.
(B) The “$” symbol tells the CAD tool that the task can be ignored during synthesis.
(C) The designer must only use system tasks that model sequential logic.

8.4 Test Benches

The functional verification of Verilog designs is accomplished through simulation using a test bench.
A test bench is a Verilog model that instantiates the system to be tested as a sub-system, generates the
input patterns to drive into the sub-system, and observes the outputs. The system being tested is often
called a device under test (DUT) or unit under test (UUT). Test benches are only used for simulation so
they can use abstract modeling techniques that are unsynthesizable to generate the stimulus patterns.
Verilog conditional programming constructions and system tasks can also be used to report on the status
of a test and also automatically check that the outputs are correct.
 8.4 Test Benches • 291

8.4.1 Common Stimulus Generation Techniques

When creating stimulus for combinational logic circuits, it is common to use a procedural block to
generate all possible input patterns to drive the DUT and especially any transitions that may cause timing
errors. Example 8.9 shows a test bench for a combinational logic circuit where an initial block contains a
series of delayed assignments to provide the stimulus to the DUT. This block creates every possible input
pattern, delayed by a fixed amount. Note that the initial block will only execute once. If the patterns were
desired to repeat indefinitely, an always block without a sensitivity list could be used instead.

Example 8.9
Test bench for a combinational logic circuit

Multiple procedural blocks can be used within a Verilog module to provide parallel functionality.
Using both initial and always blocks allows the test bench to drive both repetitive and aperiodic signals.
Initial and always blocks can also be used to drive the same signal in order to provide a starting value and
a repetitive pattern. Example 8.10 shows a test bench for a rising edge triggered D-flip-flop with an
asynchronous, active LOW reset in which multiple procedural blocks are used to generate the stimulus
patterns for the DUT.
292 • Chapter 8: Verilog (Part 2)

Example 8.10
Test bench for a sequential logic circuit

8.4.2 Printing Results to the Simulator Transcript

In the past test bench examples, the input and output values are observed using either the
waveform or listing tool within the simulator tool. It is also useful to print the values of the simulation to
a transcript window to track the simulation as each statement is processed. Messages can be printed
that show the status of the simulation in addition to the inputs and outputs of the DUT using the text
output system tasks. Example 8.11 shows a test bench that prints the inputs and output to the transcript
of the simulation tool. Note that the test bench must wait some amount of delay before evaluating the
output, even if the DUT does not contain any delay.
 8.4 Test Benches • 293

Example 8.11
Printing test bench results to the transcript

8.4.3 Automatic Result Checking

Test benches can also perform automated checking of the results using the conditional program-
ming constructs described earlier in this chapter. Example 8.12 shows an example of a test bench that
uses if-else statements to check the output of the DUT and print a PASS/FAIL message to the transcript.
294 • Chapter 8: Verilog (Part 2)

Example 8.12
Test bench with automatic output checking
 8.4 Test Benches • 295

8.4.4 Using Loops to Generate Stimulus

When creating stimulus that follow regular patterns such as counting, loops can be an effective way
to produce the input vectors. A for loop is especially useful for generating exhaustive stimulus patterns
for combinational logic circuits. An integer loop variable can increment within the for loop and then be
assigned to the DUT inputs as type reg. Recall that in Verilog, when an integer is assigned to a variable of
type reg, it is truncated to matched the size of the reg. This allows a binary count to be created for an
input stimulus pattern by using an integer loop variable that increments within a for loop. Example 8.13
shows how the stimulus for a combinational logic circuit can be produced with a for loop.

Example 8.13
Using a loop to generate stimulus in a test bench
296 • Chapter 8: Verilog (Part 2)

8.4.5 Using External Files in Test Benches

There are often cases where the results of a test bench need to be written to an external file, either
because they are too verbose or because there needs to be a stored record. Verilog allows writing to
external files via the file I/O system tasks (i.e., $fdisplay(), $fwrite(), $fstrong(), and $fmonitor()). Example
8.14 shows a test bench in which the input vectors and the output of the DUT are written to an external file
using the $fdisplay() system task.

Example 8.14
Printing test bench results to an external file
 8.4 Test Benches • 297

It is often the case that the input vectors are either too large to enter manually or were created by a
separate program. In either case, a useful technique in test benches is to read input vectors from an
external file. Example 8.15 shows an example where the input stimulus vectors for a DUT are read from
an external file using the $readmemb() system task.

Example 8.15
Reading test bench stimulus vectors from an external file
298 • Chapter 8: Verilog (Part 2)

CONCEPT CHECK

CC8.4 Could a test bench ever use always blocks and sensitivity lists exclusively to create its
stimulus? Why or why not?
(A) Yes. The signal assignments will simply be made when the block ends.
(B) No. Since a sensitivity list triggers when there is a change on one or more of the
signals listed, the blocks in the test bench would never trigger because there is no
method to make the initial signal transition.

Summary
v To model sequential logic, an HDL needs to v Variables can be defined within a procedural
be able to trigger signal assignments based block as long as the block is named.
on an event. This is accomplished in Verilog v Procedural blocks allow more advanced
using procedural assignment. modeling constructs in Verilog. These
v There are two types of procedural blocks in include if-else statements, case statements,
Verilog, initial and always. An initial block and loops.
executes one time. An always block runs v Verilog provides numerous looping
continually. constructs including forever, while, repeat,
v A sensitivity list is a way to control when a and for. Loops can be terminated using the
Verilog procedural block is triggered. A sen- disable keyword.
sitivity list contains a list of signals. If any of v System Tasks provide additional functionality
the signals in the sensitivity list transitions it to Verilog models. Tasks begin with the $
will cause the block to trigger. If a sensitivity symbol and are omitted from synthesis. Sys-
list is omitted, the block will trigger immedi- tem tasks can be included in synthesizable
ately. Sensitivity lists are most commonly logic models.
used with always blocks. v There are three groups of system tasks: text
v Sensitivity lists and always blocks are used output, file input/output, and simulation con-
to model synthesizable logic. Initial blocks trol and monitoring.
are typically only used in test benches. v System tasks that perform printing functions
Always blocks are also used in test benches. can output strings in addition to variable
v There are two types of signal assignments values. Verilog provides a mechanism to
that can be used within a procedural block, print the variable values in a variety of format.
blocking and non-blocking. v A test bench is a way to simulate a device
v A blocking assignment is denoted with the ¼ under test (DUT) by instantiating it as a
symbol. All blocking assignments are made sub-system, driving in stimulus, and observ-
immediately within the procedural block. ing the outputs. Test benches do not have
Blocking assignments are used to model inputs or outputs and are unsynthesizable.
combinational logic. Combinational logic v Test benches for combinational logic typically
models list all input to the circuit in the exercise the DUT under an exhaustive set of
sensitivity list. stimulus vectors. These include all possible
v A non-blocking assignment is denoted with logic inputs in addition to critical transitions
the <¼ symbol. All non-blocking that could cause timing errors.
assignments are made when the procedural v Text I/O system tasks provide a way to print
block ends and are evaluated in the order the results of a test bench to the simulation
they appeared in the block. Blocking tool transcript.
assignments are used to model sequential v File I/O system tasks provide a way to print
logic. Sequential logic models list only the the results of a test bench to an external file
clock and reset in the sensitivity list. and also to read in stimulus vectors from an
external file.
 Exercise Problems • 299

v Conditional programming constructs can be v Assignment from an integer to a reg in a for

used within a test bench to perform automatic loop is allowed. The binary value of the inte-
checking of the outputs of a DUT within a test ger is truncated to fit the size of the reg
bench. vector.
v Loops can be used in test benches to auto-
matically generate stimulus patterns. A for
loop is a convenient technique to produce a
counting pattern.

Exercise Problems
Section 8.1: Procedural Assignment

8.1.1 When using a sensitivity list with a procedural

block, what will cause the block to trigger?
8.1.2 When a sensitivity list is not used with a proce-
dural block, when will the block trigger?
8.1.3 When are statements executed when using
blocking assignments?
8.1.4 When are statements executed when using
non-blocking assignments?
8.1.5 When is it possible to exclude statement groups
from a procedural block?
8.1.6 What is the difference between a begin/end and
fork/join group when each contain multiple
statements?
8.1.7 What is the difference between a begin/end and
fork/join group when each contain only a single
statements?
8.1.8 What type of procedural assignment is used Fig. 8.1
when modeling combinational logic? System I functionality
8.1.9 What type of procedural assignment is used
when modeling sequential logic?
8.2.2 Design a Verilog model to implement the behav-
8.1.10 What signals should be listed in the sensitivity ior described by the 4-input truth table in Fig. 8.1.
list when modeling combinational logic? Use procedural assignment and a case state-
8.1.11 What signals should be listed in the sensitivity ment. Declare the module to match the block
list when modeling sequential logic? diagram provided. Use the type wire for the
inputs and type reg for the output.
8.2.3 Design a Verilog model to implement the behav-
Section 8.2: Conditional Programming ior described by the 4-input minterm list in
Constructs Fig. 8.2. Use procedural assignment and an
if-else statement. Declare the module to match
8.2.1 Design a Verilog model to implement the behav- the block diagram provided. Use the type wire for
ior described by the 4-input truth table in Fig. 8.1. the inputs and type reg for the output.
Use procedural assignment and an if-else state-
ment. Declare the module to match the block
diagram provided. Use the type wire for the
inputs and type reg for the output. Hint: Notice
that there are far more input codes producing
F ¼ 0 than producing F ¼ 1. Can you use this
to your advantage to make your if-else statement Fig. 8.2
simpler? System J functionality
300 • Chapter 8: Verilog (Part 2)

8.2.4 Design a Verilog model to implement the behav- 8.2.8 Design a Verilog model to implement the behav-
ior described by the 4-input minterm list in ior described by the 4-input truth table in Fig. 8.4.
Fig. 8.2. Use procedural assignment and a Use procedural assignment and a case state-
case statement. Declare the module to match ment. Declare the module to match the block
the block diagram provided. Use the type wire diagram provided. Use the type wire for the
for the inputs and type reg for the output. inputs and type reg for the output.
8.2.5 Design a Verilog model to implement the behav- 8.2.9 Fig. 8.5 shows the topology of a 4-bit shift regis-
ior described by the 4-input maxterm list in ter when implemented structurally using D-Flip-
Fig. 8.3. Use procedural assignment and an Flops. Design a Verilog model to describe this
if-then statement. Declare the module to match functionality using a single procedural block and
the block diagram provided. Use the type wire for non-blocking assignments instead of
the inputs and type reg for the output. instantiating D-Flip-Flops. The figure also
provides the block diagram for the module port
definition. Use the type wire for the inputs and
type reg for the outputs.

Fig. 8.3
System K functionality

8.2.6 Design a Verilog model to implement the behav-

ior described by the 4-input maxterm list in
Fig. 8.3. Use procedural assignment and a
case statement. Declare the module to match
the block diagram provided. Use the type wire
for the inputs and type reg for the output.
8.2.7 Design a Verilog model to implement the behav-
ior described by the 4-input truth table in Fig. 8.4.
Use procedural assignment and an if-else state-
ment. Declare the module to match the block
diagram provided. Use the type wire for the
inputs and type reg for the output. Hint: Notice Fig. 8.5
that there are far more input codes producing 4-bit shift register functionality
F ¼ 1 than producing F ¼ 0. Can you use this
to your advantage to make your if-else statement
simpler? 8.2.10 Design a Verilog model for a counter using a for
loop with an output type of integer. Fig. 8.6
shows the block diagram for the module defini-
tion. The counter should increment from 0 to
31 and then start over. Use delay in your loop to
update the counter value every 10 ns. Consider
using the loop variable of the for loop to gener-
ate your counter value.

Fig. 8.6
Integer counter block diagram

8.2.11 Design a Verilog model for a counter using a for

loop with an output type of reg[4:0]. Fig. 8.7
shows the block diagram for the module defini-
tion. The counter should increment from
000002 to 111,112 and then start over. Use
Fig. 8.4 delay in your loop to update the counter value
System L functionality every 10 ns. Consider using the loop variable of
the for loop to generate an integer version of
 Exercise Problems • 301

your count value, and then assign it to the out- whether the output of the DUT is correct. For
put variable of type reg[4:0]. each input vector, print a message using $dis-
play() that indicates the current input vector
being tested, the resulting output of your DUT,
and whether the DUT output is correct.
8.4.5 Design a Verilog test bench to verify the func-
tional operation of the system in Fig. 8.3. Your
test bench should drive in every possible input
code for the vector ABCD (i.e., “0000”, “0001”,
Fig. 8.7 “0010”, . . ., “1111”). Have your test bench change
5-bit binary counter block diagram the input pattern every 10 ns using delay within
your procedural block.
8.4.6 Design a Verilog test bench to verify the func-
Section 8.3: System Tasks tional operation of the system in Fig. 8.3 with
automatic checking. Your test bench should
8.3.1 Are system tasks synthesizable? Why or why drive in every possible input code for the vector
not? ABCD (i.e., “0000”, “0001”, “0010”, . . ., “1111”).
8.3.2 What is the difference between the tasks $dis- Have your test bench change the input pattern
play() and $write()? every 10 ns using delay within your procedural
block. Use conditional statements to check
8.3.3 What is the difference between the tasks $dis- whether the output of the DUT is correct. For
play() and $monitor()? each input vector, print a message using $dis-
8.3.4 What is the data type returned by the task play() that indicates the current input vector
$fopen()? being tested, the resulting output of your DUT,
and whether the DUT output is correct.
8.4.7 Design a Verilog test bench to verify the func-
Section 8.4: Test Benches tional operation of the system in Fig. 8.4. Your
test bench should drive in every possible input
8.4.1 Design a Verilog test bench to verify the func- code for the vector ABCD (i.e., “0000”, “0001”,
tional operation of the system in Fig. 8.1. Your “0010”, . . ., “1111”). Have your test bench change
test bench should drive in every possible input the input pattern every 10 ns using delay within
code for the vector ABCD (i.e., “0000”, “0001”, your procedural block.
“0010”, . . ., “1111”). Have your test bench change 8.4.8 Design a Verilog test bench to verify the func-
the input pattern every 10 ns using delay within tional operation of the system in Fig. 8.4 with
your procedural block. automatic checking. Your test bench should
8.4.2 Design a Verilog test bench to verify the func- drive in every possible input code for the vector
tional operation of the system in Fig. 8.1 with ABCD (i.e., “0000”, “0001”, “0010”, . . ., “1111”).
automatic checking. Your test bench should Have your test bench change the input pattern
drive in every possible input code for the vector every 10 ns using delay within your procedural
ABCD (i.e., “0000”, “0001”, “0010”, . . ., “1111”). block. Use conditional statements to check
Have your test bench change the input pattern whether the output of the DUT is correct. For
every 10 ns using delay within your procedural each input vector, print a message using $dis-
block. Use conditional statements to check play() that indicates the current input vector
whether the output of the DUT is correct. For being tested, the resulting output of your DUT,
each input vector, print a message using $dis- and whether the DUT output is correct.
play() that indicates the current input vector 8.5.9 Design a Verilog test bench to verify the functional
being tested, the resulting output of your DUT, operation of the system in Fig. 8.4. Your test bench
and whether the DUT output is correct. should drive in every possible input code for the
8.4.3 Design a Verilog test bench to verify the func- vector ABCD (i.e., “0000”, “0001”, “0010”, . . .,
tional operation of the system in Fig. 8.2. Your “1111”). Have your test bench change the input
test bench should drive in every possible input pattern every 10 ns using delay within your proce-
code for the vector ABCD (i.e., “0000”, “0001”, dural block. Print the results to an external file
“0010”, . . ., “1111”). Have your test bench change named “output_vectors.txt” using $fdisplay().
the input pattern every 10 ns using delay within 8.5.10 Design a Verilog test bench that reads in test
your procedural block. vectors from an external file to verify the func-
8.4.4 Design a Verilog test bench to verify the func- tional operation of the system in Fig. 8.4. Create
tional operation of the system in Fig. 8.2 with an input text file called “input_vectors.txt” that
automatic checking. Your test bench should contains each input code for the vector ABCD
drive in every possible input code for the vector (i.e., “0000”, “0001”, “0010”, . . ., “1111”), each on
ABCD (i.e., “0000”, “0001”, “0010”, . . ., “1111”). a separate line in the file. Your test bench should
Have your test bench change the input pattern read in the vectors using $readmemb(), drive
every 10 ns using delay within your procedural each code into the DUT, and print the results to
block. Use conditional statements to check the transcript using $display().
Chapter 9: Behavioral Modeling
of Sequential Logic
In this chapter, we will look at modeling sequential logic using the more sophisticated behavioral
modeling techniques presented in Chap. 8. We will begin by looking at modeling sequential storage
devices. Next, we will look at the behavioral modeling of finite state machines. Finally, we will look at
register transfer level, or RTL modeling. The goal of this chapter is to provide an understanding of how
hardware description languages can be used to create behavioral models of synchronous digital
systems.
Learning Outcomes—After completing this chapter, you will be able to:
9.1 Design a Verilog behavioral model for a sequential logic storage device.
9.2 Describe the process for creating a Verilog behavioral model for a finite state
machine.
9.3 Design a Verilog behavioral model for a finite state machine.
9.4 Design a Verilog behavioral model for a counter.
9.5 Design a Verilog register transfer level (RTL) model of a synchronous digital system.

9.1 Modeling Sequential Storage Devices in Verilog

9.1.1 D-Latch
Let’s begin with the model of a simple D-Latch. Since the outputs of this sequential storage device
are not updated continuously, its behavior is modeled using a procedural assignment. Since we want to
create a synthesizable model of sequential logic, non-blocking assignments are used. In the sensitivity
list, we need to include the C input since it controls when the D-Latch is in track or store mode. We also
need to include the D input in the sensitivity list because during the track mode, the output Q will be
assigned the value of D so any change on D needs to trigger the procedural assignments. The use of an
if-else statement is used to model the behavior during track mode (C ¼ 1). Since the behavior is not
explicitly stated for when C ¼ 0, the outputs will hold their last value, which allows us to simply omit the
else portion of the if statement to complete the model. Example 9.1 shows the behavioral model for a
D-Latch.

Example 9.1
Behavioral model of a D-latch in Verilog

# Springer International Publishing AG 2017 303

B.J. LaMeres, Introduction to Logic Circuits & Logic Design with Verilog,
DOI 10.1007/978-3-319-53883-9_9
304 • Chapter 9: Behavioral Modeling of Sequential Logic

9.1.2 D-Flip-Flop
The rising edge behavior of a D-Flip-Flop is modeled using a (posedge Clock) Boolean condition in
the sensitivity list of a procedural block. Example 9.2 shows the behavioral model for a rising edge
triggered D-Flip-Flop with both Q and Qn outputs.

Example 9.2
Behavioral model of a D-flip-flop in Verilog

9.1.3 D-Flip-Flop with Asynchronous Reset

D-Flip-Flops typically have a reset line to initialize their outputs to known states (e.g., Q ¼ 0, Qn ¼ 1).
Resets are asynchronous, meaning whenever they are asserted, assignments to the outputs takes place
immediately. If a reset was synchronous, the outputs would only update on the next rising edge of the
clock. This behavior is undesirable because if there is a system failure, there is no guarantee that a clock
edge will ever occur. Thus, the reset may never take place. Asynchronous resets are more desirable not
only to put the D-Flip-Flops into a known state at startup, but also to recover from a system failure that
may have impacted the clock signal. In order to model this asynchronous behavior, the reset signal is
included in the sensitivity list. This allows both clock and the reset transitions to trigger the procedural
block. The edge sensitivity of the reset can be specified using posedge (active HIGH) or negedge (active
LOW). Within the block an if-else statement is used to determine whether the reset has been asserted or
a rising edge of the clock has occurred. The if-else statement first checks whether the reset input has
been asserted since it has the highest priority. If it has, it makes the appropriate assignments to the
outputs (Q ¼ 0, Qn ¼ 1). If the reset has not been asserted, the else clause is executed, which
corresponds to a rising edge of clock (Q < ¼ D, Qn < ¼ ~ D). No other assignments are listed in the
block, thus the outputs are only updated on a transition of the reset or clock. At all other times the outputs
remain at their current value, thus modeling the store behavior of the D-Flip-Flop. Example 9.3 shows the
behavioral model for a rising edge triggered D-Flip-Flop with an asynchronous, active LOW reset.
 9.1 Modeling Sequential Storage Devices in Verilog • 305

Example 9.3
Behavioral model of a D-flip-flop with asynchronous reset in Verilog

9.1.4 D-Flip-Flop with Asynchronous Reset and Preset

A D-Flip-Flop with both an asynchronous reset and asynchronous preset is handled in a similar
manner as the D-Flip-Flop in the prior section. The preset input is included in the sensitivity list in order to
trigger the block whenever a transition occurs on either the clock, reset, or preset inputs. The edge
sensitivity keywords are used to dictated whether the preset is active HIGH or LOW. Nested if-else
statements are used to first check whether a reset has occurred; then whether a preset has occurred;
and finally, whether a rising edge of the clock has occurred. Example 9.4 shows the model for a rising
edge triggered D-Flip-Flop with asynchronous, active LOW reset and preset.
306 • Chapter 9: Behavioral Modeling of Sequential Logic

Example 9.4
Behavioral model of a D-flip-flop with asynchronous reset and preset in Verilog

9.1.5 D-Flip-Flop with Synchronous Enable

An enable input is also a common feature of modern D-Flip-Flops. Enable inputs are synchronous,
meaning that when they are asserted, action is only taken on the rising edge of the clock. This means
that the enable input is not included in the sensitivity list of the always block. Since enable is only
considered when there is a rising edge of the clock, the logic for the enable is handled in a nested if-else
statement that is included in the section that models the behavior for when a rising edge of clock is
detected. Example 9.5 shows the model for a D-Flip-Flop with a synchronous enable (EN) input. When
EN ¼ 1, the D-Flip-Flop is enabled and assignments are made to the outputs only on the rising edge of
the clock. When EN ¼ 0, the D-Flip-Flop is disabled and assignments to the outputs are not made. When
disabled, the D-Flip-Flop effectively ignores rising edges on the clock and the outputs remain at their last
values.
 9.2 Modeling Finite State Machines in Verilog • 307

Example 9.5
Behavioral model of a D-flip-flop with synchronous enable in Verilog

CONCEPT CHECK

CC9.1 Why is the D input not listed in the sensitivity list of a D-flip-flop?
(A) To simplify the behavioral model.
(B) To avoid a setup time violation if D transitions too closely to the clock.
(C) Because a rising edge of clock is needed to make the assignment.
(D) Because the outputs of the D-flip-flop are not updated when D changes.

9.2 Modeling Finite State Machines in Verilog

Finite state machines can be easily modeled using the behavioral constructs from Chap. 8. The
most common modeling practice for FSMs is to declare two signals of type reg that are called
current_state and next_state. Then a parameter is declared for each descriptive state name in the
state diagram. A parameter also requires a value, so the state encoding can be accomplished during the
parameter declaration. Once the signals and parameters are created, all of the procedural assignments
in the state machine model can use the descriptive state names in their signal assignments. Within the
Verilog state machine model, three separate procedural blocks are used to describe each of the
functional blocks, state memory, next state logic, and output logic. In order to examine how to model a
308 • Chapter 9: Behavioral Modeling of Sequential Logic

finite state machine using this approach, let”s use the push-button window controller example from
Chap. 7. Example 9.6 gives the overview of the design objectives for this example and the state diagram
describing the behavior to be modeled in Verilog.

Example 9.6
Push-button window controller in Verilog – design description

Let”s begin by defining the ports of the module. The system has an input called Press and two
outputs called Open_CW and Close_CCW. The system also has clock and reset inputs. We will design
the system to update on the rising edge of the clock and have an asynchronous, active LOW, reset.
Example 9.7 shows the port definitions for this example. Note that outputs are declared as type reg while
inputs are declared as type wire.

Example 9.7
Push-button window controller in Verilog – port definition
 9.2 Modeling Finite State Machines in Verilog • 309

9.2.1 Modeling the States

Now we begin designing the finite state machine in Verilog using behavioral modeling constructs.
The first step is to create two signals that will be used for the state variables. In this text we will always
name these signals current_state and next_state. The signal current_state will represent the outputs of
the D-flip-flops forming the state memory and will hold the current state code. The signal next_state will
represent the D inputs to the D-flip-flops forming the state memory and will receive the value from the
next state logic circuitry. Since the FSM will be modeled using procedural assignment, both of these
signals will be declared of type reg. The width of the reg vector depends on the number of states in the
machine and the encoding technique chosen. The next step is to declare parameters for each of the
descriptive state names in the state diagram. The state encoding must be decided at this point. The
following syntax shows how to declare the current_state and next_state signals and the parameters.
Note that since this machine only has two states, the width of these signals is only 1-bit.
reg current_state, next_state;
parameter w_closed ¼ 1’b0,
w_open ¼ 1’b1;

9.2.2 The State Memory Block

Now that we have variables and parameters for the states of the FSM, we can create the model for
the state memory. State memory is modeled using its own procedural block. This block models the
behavior of the D-Flip-Flops in the FSM that are holding the current state on their Q outputs. Each time
there is a rising edge of the clock, the current state is updated with the next state value present on the D
inputs of the D-Flip-Flops. This block must also model the reset condition. For this example, we will have
the state machine go to the w_closed state when Reset is asserted. At all other times, the block will
simply update current_state with next_state on every rising edge of the clock. The block model is very
similar to the model of a D-Flip-Flop. This is as expected since this block will synthesize into one or more
D-Flip-Flops to hold the current state. The sensitivity list contains only Clock and Reset and assignments
are only made to the signal current_state. The following syntax shows how to model the state memory of
this FSM example.

always @ (posedge Clock or negedge Reset)

begin: STATE_MEMORY
if (!Reset)
current_state <¼ w_closed;
else
current_state <¼ next_state;
end

9.2.3 The Next State Logic Block

Now we model the next state logic of the FSM using a second procedural block. Recall that the next
state logic is combinational logic, thus we need to include all of the input signals that the circuit considers
in the next state calculation in the sensitivity list. The current_state signal will always be included in the
sensitivity list of the next state logic block in addition to any inputs to the system. For this example, the
system has one other input called Press. This block makes assignments to the next_state signal. It is
common to use a case statement to separate out the assignments that occur at each state. At each state
within the case statement, an if-else statement is used to model the assignments for different input
conditions on Press. The following syntax shows how to model the next state logic of this FSM example.
Notice that we include a default clause in the case statement to ensure that the state machine has a path
back to the reset state in the case of an unexpected fault.
310 • Chapter 9: Behavioral Modeling of Sequential Logic

always @ (current_state or Press)

begin: NEXT_STATE_LOGIC
case (current_state)
w_closed : if (Press ¼¼ 1’b1) next_state ¼ w_open; else next_state ¼
w_closed;
w_open : if (Press ¼¼ 1’b1) next_state ¼ w_closed; else next_state ¼
w_open;
default : next_state ¼ w_closed;
endcase
end

9.2.4 The Output Logic Block

Now we model the output logic of the FSM using a third procedural block. Recall that output logic is
combinational logic, thus we need to include all of the input signals that this circuit considers in the output
assignments. The current_state will always be included in the sensitivity list. If the FSM is a Mealy
machine, then the system inputs will also be included in the sensitivity list. If the machine is a Moore
machine, then only the current_state will be present in the sensitivity list. For this example, the FSM is a
Mealy machine so the input Press needs to be included in the sensitivity list. Note that this block only
makes assignments to the outputs of the machine (Open_CW and Close_CCW). The following syntax
shows how to model the output logic of this FSM example. Again, we include a default clause to ensure
that the state machine has explicit output behavior in the case of a fault.

always @ (current_state or Press)

begin: OUTPUT_LOGIC
case (current_state)
w_closed : if (Press ¼¼ 1’b1)
begin
Open_CW ¼ 1’b1;
Close_CCW ¼ 1’b0;
end
else
begin
Open_CW ¼ 1’b0;
Close_CCW ¼ 1’b0;
end
w_open : if (Press ¼¼ 1’b1)
begin
Open_CW ¼ 1’b0;
Close_CCW ¼ 1’b1;
end
else
begin
Open_CW ¼ 1’b0;
Close_CCW ¼ 1’b0;
end
default : begin
Open_CW ¼ 1’b0;
Close_CCW ¼ 1’b0;
end
endcase
end

Putting this all together yields a behavioral model for the FSM that can be simulated and
synthesized. Example 9.8 shows the entire model for this example.
 9.2 Modeling Finite State Machines in Verilog • 311

Example 9.8
Push-button window controller in Verilog – full model

Example 9.9 shows the simulation waveform for this state machine. This functional simulation was
performed using ModelSim-Altera Starter Edition 10.1d. A macro file was used to display the current and
next state variables using their parameter names instead of their state codes. This allows the functional-
ity of the FSM to be more easily observed. This approach will be used for the rest of the FSM examples in
this book.
312 • Chapter 9: Behavioral Modeling of Sequential Logic

Example 9.9
Push-button window controller in Verilog – simulation waveform

9.2.5 Changing the State Encoding Approach

In the prior example we only had two states and they were encoded as: w_closed ¼1’b0;
w_open_1’b1. This encoding technique is considered binary; however, a gray code approach would
yield the same codes since the width of the variables were only one bit. The way that state variables and
state codes are assigned in Verilog makes is straightforward to change the state codes. The only
consideration that must be made is expanding the size of the current_state and next_state variables
to accommodate the new state codes. The following example shows how the state encoding would look
if a one-hot approach was used (w_closed ¼2’b01; w_open_2’b10). Note that the state variables now
must be two bits wide. This means the state variables need to be declared as type reg[1:0]. Example 9.10
shows the resulting simulation waveforms. The simulation waveform shows the value of the state codes
instead of the state names.
reg [1:0] current_state, next_state;
parameter w_closed ¼ 2’b01,
w_open ¼ 2’b10;

Example 9.10
Push-button window controller in Verilog – changing state codes
 9.3 FSM Design Examples in Verilog • 313

CONCEPT CHECK

CC9.2 Why is it always a good design approach to model a generic finite state machine using three
processes?
(A) For readability.
(B) So that it is easy to identify whether the machine is a Mealy or Moore.
(C) So that the state memory process can be re-used in other FSMs.
(D) Because each of the three sub-systems of a FSM has unique inputs and outputs
that should be handled using dedicated processes.

9.3 FSM Design Examples in Verilog

This section presents a set of example finite state machine designs using the behavioral modeling
constructs of Verilog. These examples are the same state machines that were presented in Chap. 7.

9.3.1 Serial Bit Sequence Detector in Verilog

Let’s look at the design of the serial bit sequence detector finite state machine from Chap. 7 using
the behavioral modeling constructs of Verilog. Example 9.11 shows the design description and port
definition for this state machine.

Example 9.11
Serial bit sequence detector in Verilog – design description and port definition
314 • Chapter 9: Behavioral Modeling of Sequential Logic

Example 9.12 shows the full model for the serial bit sequence detector. Notice that the states are
encoded in binary, which requires three bits for the variables current_state and next_state.

Example 9.12
Serial bit sequence detector in Verilog – full model
 9.3 FSM Design Examples in Verilog • 315

Example 9.13 shows the functional simulation waveform for this design.

Example 9.13
Serial bit sequence detector in Verilog – simulation waveform

9.3.2 Vending Machine Controller in Verilog

Let’s now look at the design of the vending machine controller from Chap. 7 using the behavioral
modeling constructs of Verilog. Example 9.14 shows the design description and port definition.

Example 9.14
Vending machine controller in Verilog – design description and port definition
316 • Chapter 9: Behavioral Modeling of Sequential Logic

Example 9.15 shows the full model for the vending machine controller. In this model, the descriptive
state names Wait, 25¢, and 50¢ cannot be used directly. This is because Verilog user-defined names
cannot begin with a number. Instead, the letter “s” is placed in front of the state names in order to make
them legal Verilog names (i.e., sWait, s25, s50).

Example 9.15
Vending machine controller in Verilog – full model
 9.3 FSM Design Examples in Verilog • 317

Example 9.16 shows the resulting simulation waveform for this design.

Example 9.16
Vending machine controller in Verilog – simulation waveform

9.3.3 2-Bit, Binary Up/Down Counter in Verilog

Let’s now look at how a simple counter can be implemented using the three-block behavioral
modeling approach in Verilog. Example 9.17 shows the design description and port definition for the
2-bit, binary up/down counter FSM from Chap. 7.

Example 9.17
2-bit up/down counter in Verilog – design description and port definition

Example 9.18 shows the full model for the 2-bit up/down counter using the three-block modeling
approach. Since a counter’s outputs only depend on the current state, counters are Moore machines.
This simplifies the output logic block since it only needs to contain the current state in its sensitivity list.
318 • Chapter 9: Behavioral Modeling of Sequential Logic

Example 9.18
2-bit up/down counter in Verilog – full model (three block approach)

Example 9.19 shows the resulting simulation waveform for this counter finite state machine.

Example 9.19
2-bit up/down counter in Verilog – simulation waveform
 9.4 Modeling Counters in Verilog • 319

CONCEPT CHECK

CC9.3 The procedural block for the state memory is nearly identical for all finite state machines with
one exception. What is it?
(A) The sensitivity list may need to include a preset signal.
(B) Sometimes it is modeled using an SR latch storage approach instead of with D-flip-
flop behavior.
(C) The name of the reset state will be different.
(D) The current_state and next_state signals are often swapped.

9.4 Modeling Counters in Verilog

Counters are a special case of finite state machines because they move linearly through their
discrete states (either forward or backwards) and typically are implemented with state-encoded outputs.
Due to this simplified structure and wide spread use in digital systems, Verilog allows counters to be
modeled using a single procedural block with arithmetic operators (i.e., + and ). This enables a more
compact model and allows much wider counters to be implemented in a practical manner.

9.4.1 Counters in Verilog Using a Single Procedural Block

Let’s look at how we can model a 4-bit, binary up counter with an output called CNT. We want to
model this counter using the “+” operator to avoid having to explicitly define a state code for each state as
in the three-block modeling approach to FSMs. The “+” operator works on the type reg so the counting
behavior can simply be modeled using CNT < ¼ CNT + 1. The procedural block also needs to handle the
reset condition. Both the Clock and Reset signals are listed in the sensitivity list. Within the block, an
if-else statement is used to handle both the reset and increment behaviors. Example 9.20 shows the
Verilog model and simulation waveform for this counter. When the counter reaches its maximum value of
“1111”, it rolls over to “0000” and continues counting because it is declared to only contain 4-bits.

Example 9.20
Binary counter using a single procedural block in Verilog
320 • Chapter 9: Behavioral Modeling of Sequential Logic

9.4.2 Counters with Range Checking

When a counter needs to have a maximum range that is different from the maximum binary value of
the count vector (i.e., <2n  1), then the procedural block needs to contain range checking logic. This
can be modeled by inserting a nested if-else statement beneath of the else clause that handles the
behavior for when the counter receives a rising clock edge. This nested if-else first checks whether the
count has reached its maximum value. If it has, it is reset back to it minimum value. If it hasn’t, the counter
is incremented as usual. Example 9.21 shows the Verilog model and simulation waveform for a counter
with a minimum count value of 010 and a maximum count value of 1010. This counter still requires 4-bits to
be able to encode 1010.

Example 9.21
Binary counter with range checking in Verilog

9.4.3 Counters with Enables in Verilog

Including an enable in a counter is a common technique to prevent the counter from running
continuously. When the enable is asserted, the counter will increment on the rising edge of the clock
as usual. When the enable is de-asserted, the counter will simply hold its last value. Enable lines are
synchronous, meaning that they are only evaluated on the rising edge of the clock. As such, they are
modeled using a nested if-else statement within the main if-else statement checking for a rising edge of
the clock. Example 9.22 shows an example model for a 4-bit counter with enable.
 9.4 Modeling Counters in Verilog • 321

Example 9.22
Binary counter with enable in Verilog

9.4.4 Counters with Loads

A counter with a load has the ability to set the counter to a specified value. The specified value is
provided on an input port (i.e., CNT_in) with the same width as the counter output (CNT). A synchronous
load input signal (i.e., Load) is used to indicate when the counter should set its value to the value present
on CNT_in. Example 9.23 shows an example model for a 4-bit counter with load capability.
322 • Chapter 9: Behavioral Modeling of Sequential Logic

Example 9.23
Binary counter with load in Verilog

CONCEPT CHECK

CC9.4 If a counter is modeled using only one procedural block in Verilog, is it still a finite state
machine? Why or why not?
(A) Yes. It is just a special case of a FSM that can easily be modeled using one block.
Synthesizers will recognize the single block model as a FSM.
(B) No. Using only one block will synthesize into combinational logic. Without the
ability to store a state, it is not a finite state machine.

9.5 RTL Modeling

Register Transfer Level modeling refers to a level of design abstraction in which vector data is
moved and operated on in a synchronous manner. This design methodology is widely used in data path
modeling and computer system design.

9.5.1 Modeling Registers in Verilog

The term register describes a group of D-Flip-Flops running off of the same clock, reset, and enable
inputs. Data is moved in and out of the bank of D-flip-flops as a vector. Logic operations can be made on
the vectors and are latched into the register on a clock edge. A register is a higher level of abstraction that
 9.5 RTL Modeling • 323

allows vector data to be stored without getting into the details of the lower level implementation of the D-
Flip-Flops and combinational logic. Example 9.24 shows an RTL model of an 8-bit, synchronous register.
This circuit has an active LOW, asynchronous reset that will cause the 8-bit output Reg_Out to go to
0 when it is asserted. When the reset is not asserted, the output will be updated with the 8-bit input
Reg_In if the system is enabled (EN ¼ 1) and there is a rising edge on the clock. If the register is disabled
(EN ¼ 0), the input clock is ignored. At all other times, the output holds its last value.

Example 9.24
RTL model of an 8-bit register in Verilog

9.5.2 Registers as Agents on a Data Bus

One of the powerful topologies that registers can easily model is a multi-drop bus. In this topology,
multiple registers are connected to a data bus as receivers, or agents. Each agent has an enable line that
controls when it latches information from the data bus into its storage elements. This topology is
synchronous, meaning that each agent and the driver of the data bus is connected to the same clock
signal. Each agent has a dedicated, synchronous enable line that is provided by a system controller
elsewhere in the design. Example 9.25 shows this multi-drop bus topology. In this example system, three
registers (A, B, and C) are connected to a data bus as receivers. Each register is connected to the same
clock and reset signals. Each register has its own dedicated enable line (A_EN, B_EN, and C_EN).
324 • Chapter 9: Behavioral Modeling of Sequential Logic

Example 9.25
Registers as agents on a data bus – system topology

This topology can be modeled using RTL abstraction by treating each register as a separate
procedural block. Example 9.26 shows how to describe this topology with an RTL model in Verilog.
Notice that the three procedural blocks modeling the A, B, and C registers are nearly identical to each
other except for the signal names they use.

Example 9.26
Registers as agents on a data bus – RTL model in Verilog
 9.5 RTL Modeling • 325

Example 9.27 shows the resulting simulation waveform for this system. Each register is updated
with the value on the data bus whenever its dedicated enable line is asserted.

Example 9.27
Registers as agents on a data bus – simulation waveform

9.5.3 Shift Registers in Verilog

A shift register is a circuit which consists of multiple registers connected in series. Data is shifted
from one register to another on the rising edge of the clock. This type of circuit is often used in serial-to-
parallel data converters. Example 9.28 shows an RTL model for a 4-stage, 8-bit shift register.
326 • Chapter 9: Behavioral Modeling of Sequential Logic

Example 9.28
RTL model of a 4-stage, 8-bit shift register in Verilog

CONCEPT CHECK

CC9.5 Does RTL modeling synthesize as combinational logic, sequential logic, or both? Why?
(A) Combinational logic. Since only one process is used for each register, it will be
synthesized using basic gates.
(B) Sequential logic. Since the sensitivity list contains clock and reset, it will synthesize
into only D-flip-flops.
(C) Both. The model has a sensitivity list containing clock and reset and uses an if-else
statement indicative of a D-flip-flop. This will synthesize a D-flip-flop to hold the
value for each bit in the register. In addition, the ability to manipulate the inputs into
the register (using either logical operators, arithmetic operators, or choosing dif-
ferent signals to latch) will synthesize into combinational logic in front of the D input
to each D-flip-flop.
 Exercise Problems • 327

Summary
v A synchronous system is modeled with a signals are first declared called current_state
procedural block and a sensitivity list. The and next_state of type reg. Then a parameter
clock and reset signals are always listed by is defined for each unique state in the
themselves in the sensitivity list. Within the machine with the state name and desired
block is an if-else statement. The if clause of state code. Throughout the rest of the
the statement handles the asynchronous model, the unique state names can be used
reset condition while the else clause handles as both assignments to current_state/
the synchronous signal assignments. next_state and as inputs in case and if-else
v Edge sensitivity is modeled within a proce- statements. This approach allows the model
dural block using the (posedge Clock or to be designed using readable syntax while
negedge reset) syntax in the sensitivity lists. providing a synthesizable design.
v Most D-flip-flops and registers contain a syn- v Counters are a special type of finite state
chronous enable line. This is modeled using machine that can be modeled using a single
a nested if-else statement within the main procedural block. Only the clock and reset
procedural block’s if-else statement. The signals are listed in the sensitivity list of the
nested if-else goes beneath the clause for counter block.
the synchronous signal assignments. v Registers are modeled in Verilog in a similar
v Generic finite state machines are modeled manner to a D-flip-flop with a synchronous
using three separate procedural blocks that enable. The only difference is that the inputs
describe the behavior of the next state logic, and outputs are vectors.
the state memory, and the output logic. Sep- v Register Transfer Level, or RTL, modeling
arate blocks are used because each of the provides a higher level of abstraction for
three functions in a FSM are dependent on moving and manipulating vectors of data in
different input signals. a synchronous manner.
v In Verilog, descriptive state names can be
created for a FSM using parameters. Two

Exercise Problems
Section 9.1: Modeling Sequential Storage 9.2.3 When using the three-procedural block behav-
Devices in Verilog ioral modeling approach for finite state
machines, does the next state logic block
9.1.1 How does a Verilog model for a D-flip-flop han-
model combinational or sequential logic?
dle treating reset as the highest priority input?
9.2.4 When using the three-procedural block behav-
9.1.2 For a Verilog model of a D-flip-flop with a syn-
ioral modeling approach for finite state
chronous enable (EN), why isn’t EN listed in
machines, does the state memory block
the sensitivity list?
model combinational or sequential logic?
9.1.3 For a Verilog model of a D-flip-flop with a syn-
9.2.5 When using the three-procedural block behav-
chronous enable (EN), what is the impact of
ioral modeling approach for finite state
listing EN in the sensitivity list?
machines, does the output logic block model
9.1.4 For a Verilog model of a D-flip-flop with a syn- combinational or sequential logic?
chronous enable (EN), why is the behavior of
9.2.6 When using the three-procedural block behav-
the enable modeled using a nested if-else
ioral modeling approach for finite state
statement under the else clause handling the
machines, what inputs are listed in the sensi-
logic for the clock edge input?
tivity list of the next state logic block?
9.2.7 When using the three-procedural block behav-
Section 9.2: Modeling Finite State Machines
ioral modeling approach for finite state
in Verilog machines, what inputs are listed in the sensi-
9.2.1 What is the advantage of using parameters for tivity list of the state memory block?
the state when modeling a finite state 9.2.8 When using the three-procedural block behav-
machine? ioral modeling approach for finite state
9.2.2 What is the advantage of having to assign the machines, what inputs are listed in the sensi-
state codes during the parameter declaration tivity list of the output logic block?
for the state names when modeling a finite 9.2.9 When using the three-procedural block behav-
state machine? ioral modeling approach for finite state
328 • Chapter 9: Behavioral Modeling of Sequential Logic

machines, how can the signals listed in the

sensitivity list of the output logic block immedi-
ately indicate whether the FSM is a Mealy or a
Moore machine?
9.2.10 Why is it not a good design approach to com-
bine the next state logic and output logic
behavior into a single procedural block?

Section 9.3: FSM Design Examples in Verilog

9.3.1 Design a Verilog behavioral model to imple-
ment the finite state machine described by
the state diagram in Fig. 9.1. Use the port
definition provided in this figure for your design.
Use the three-block approach to modeling
FSMs described in this chapter for your design.
Model the state variables using parameters
and encode the states in binary using the fol-
lowing state codes: Start ¼ “00”, Midway ¼ “01”,
Done ¼ “10”.

Fig. 9.2
FSM 2 state diagram and module definition

9.3.4 Design a Verilog behavioral model to imple-

ment the finite state machine described by
the state diagram in Fig. 9.2. Use the port
definition provided in this figure for your design.
Use the three-block approach to modeling
FSMs described in this chapter for your design.
Model the state variables using parameters
and encode the states in one-hot using the
following state codes: S0 ¼ “0001”,
S1 ¼ “0010”, S2 ¼ “0100”, and S3 ¼ “1000”.
9.3.5 Design a Verilog behavioral model for a 4-bit
serial bit sequence detector similar to Example
Fig. 9.1 9.11. Use the port definition provided in
Fig. 9.3. Use the three-block approach to
FSM 1 state diagram and module definition
modeling FSMs described in this chapter for
your design. The input to your sequence detec-
9.3.2 Design a Verilog behavioral model to imple- tor is called DIN and the output is called
ment the finite state machine described by FOUND. Your detector will assert FOUND any-
the state diagram in Fig. 9.1. Use the port time there is a 4-bit sequence of “0101”. Model
definition provided in this figure for your design. the states in this machine with parameters.
Use the three-block approach to modeling Choose any state encoding approach
FSMs described in this chapter for your design. you wish.
Model the state variables using parameters
and encode the states in one-hot using the
following state codes: Start ¼ “001”, Mid-
way ¼ “010”, Done ¼ “100”.
9.3.3 Design a Verilog behavioral model to imple-
ment the finite state machine described by
the state diagram in Fig. 9.2. Use the port
definition provided in this figure for your design.
Use the three-block approach to modeling
Fig. 9.3
FSMs described in this chapter for your design.
Model the state variables using parameters Sequence detector module definition
and encode the states in binary using the fol-
lowing state codes: S0 ¼ “00”, S1 ¼ “01”, 9.3.6 Design a Verilog behavioral model for a
S2 ¼ “10”, and S3 ¼ “11”. 20-cent vending machine controller similar to
Example 9.14. Use the port definition provided
in Fig. 9.4. Use the three-block approach to
 Exercise Problems • 329

modeling FSMs described in this chapter for Section 9.4: Modeling Counters in Verilog
your design. Your controller will take in nickels 9.4.1 Design a Verilog behavioral model for a 16-bit,
and dimes and dispense a product anytime the binary up counter using a single procedural
customer has entered 20 cents. Your FSM has block. The block diagram for the port definition
two inputs, Nin and Din. Nin is asserted when- is shown in Fig. 9.6.
ever the customer enters a nickel while Din is
asserted anytime the customer enters a dime.
Your FSM has two outputs, Dispense and
Change. Dispense is asserted anytime the
customer has entered at least 20 cents and
Change is asserted anytime the customer has
entered more than 20 cents and needs a nickel
in change. Model the states in this machine
with parameters. Choose any state encoding Fig. 9.6
approach you wish. 16-bit binary up counter block diagram

9.4.2 Design a Verilog behavioral model for a 16-bit,

binary up counter with range checking using a
single procedural block. The block diagram for
the port definition is shown in Fig. 9.6. Your
counter should count up to 60,000 and then
start over at 0.
9.4.3 Design a Verilog behavioral model for a 16-bit,
Fig. 9.4 binary up counter with enable using a single
Vending machine module definition procedural block. The block diagram for the
port definition is shown in Fig. 9.7.
9.3.7 Design a Verilog behavioral model for a finite
state machine for a traffic light controller. Use
the port definition provided in Fig. 9.5. This is
the same problem description as in exercise
7.4.15. This time, you will implement the func-
tionality using the behavioral modeling
techniques presented in this chapter. Your
FSM will control a traffic light at the intersection
of a busy highway and a seldom used side Fig. 9.7
road. You will be designing the control signals
16-bit binary counter with enable block diagram
for just the red, yellow, and green lights facing
the highway. Under normal conditions, the
highway has a green light. The side road has 9.4.4 Design a Verilog behavioral model for a 16-bit,
car detector that indicates when car pulls up by binary up counter with enable and load using a
asserting a signal called CAR. When CAR is single procedural block. The block diagram for
asserted, you will change the highway traffic the port definition is shown in Fig. 9.8.
light from green to yellow, and then from yellow
to red. Once in the red position, a built-in timer
will begin a countdown and provide your con-
troller a signal called TIMEOUT when
15 seconds has passed. Once TIMEOUT is
asserted, you will change the highway traffic
light back to green. Your system will have three
outputs GRN, YLW, and RED, which control
when the highway facing traffic lights are on
(1 ¼ ON, 0 ¼ OFF). Model the states in this Fig. 9.8
machine with parameters. Choose any state 16-bit binary counter with load block diagram
encoding approach you wish.
9.4.5 Design a Verilog behavioral model for a 16-bit,
binary up/down counter using a single proce-
dural block. The block diagram for the port
definition is shown in Fig. 9.9. When Up ¼ 1,
the counter will increment. When Up ¼ 0, the
counter will decrement.

Fig. 9.5
Traffic light controller module definition
330 • Chapter 9: Behavioral Modeling of Sequential Logic

9.5.5 Design a Verilog RTL model of the multi-drop

bus topology in Fig. 9.12. Each of the 16-bit
registers (RegA, RegB, RegC, and RegD) will
latch the contents of the 16-bit data bus if their
enable line is asserted. Each register should
be modeled using an individual procedural
block.
Fig. 9.9
16-bit binary up/down counter block diagram
Section 9.5: RTL Modeling
9.5.1 In register transfer level modeling, how does
the width of the register relate to the number of
D-flip-flops that will be synthesized?
9.5.2 In register transfer level modeling, how is the
synchronous data movement managed if all
registers are using the same clock?
9.5.3 Design a Verilog RTL model of a 32-bit, syn-
chronous register. The block diagram for the
port definition is shown in Fig. 9.10. The register
has a synchronous enable. The register should
be modeled using a single procedural block.

Fig. 9.10
32-bit register block diagram

9.5.4 Design a Verilog RTL model of an 8-stage,

16-bit shift register. The block diagram for the
port definition is shown in Fig. 9.11. Each stage
of the shift register will be provided as an out-
put of the system (A, B, C, D, E, F, G, and H). Fig. 9.12
The shift register should be modeled using a
Agents on a bus block diagram
single procedural block.

Fig. 9.11
16-bit shift register block diagram
Chapter 10: Memory
This chapter introduces the basic concepts, terminology, and roles of memory in digital systems.
The material presented here will not delve into the details of the device physics or low-level theory of
operation. Instead, the intent of this chapter is to give a general overview of memory technology and its
use in computer systems in addition to how to model memory in Verilog. The goal of this chapter is to give
an understanding of the basic principles of semiconductor-based memory systems.
Learning Outcomes—After completing this chapter, you will be able to:

10.1 Describe the basic architecture and terminology for semiconductor-based memory
systems.
10.2 Describe the basic architecture of non-volatile memory systems.
10.3 Describe the basic architecture of volatile memory systems.
10.4 Design a Verilog behavioral model of a memory system.

10.1 Memory Architecture and Terminology

The term memory is used to describe a system with the ability to store digital information. The term
semiconductor memory refers to systems that are implemented using integrated circuit technology.
These types of systems store the digital information using transistors, fuses, and/or capacitors on a
single semiconductor substrate. Memory can also be implemented using technology other than
semiconductors. Disk drives store information by altering the polarity of magnetic fields on a circular
substrate. The two magnetic polarities (north and south) are used to represent different logic values
(i.e., 0 or 1). Optical disks use lasers to burn pits into reflective substrates. The binary information is
represented by light either being reflected (no pit) or not reflected (pit present). Semiconductor memory
does not have any moving parts, so it is called solid state memory and can hold more information per unit
area than disk memory. Regardless of the technology used to store the binary data, all memory has
common attributes and terminology that are discussed in this chapter.

10.1.1 Memory Map Model

The information stored in memory is called the data. When information is placed into memory, it is
called a write. When information is retrieved from memory, it is called a read. In order to access data in
memory, an address is used. While data can be accessed as individual bits, in order to reduce the
number of address locations needed, data is typically grouped into N-bit words. If a memory system has
N ¼ 8, this means that 8-bits of data are stored at each address. The number of address locations is
described using the variable M. The overall size of the memory is typically stated by saying “M  N”. For
example, if we had a 16  8 memory system, that means that there are 16 address locations, each
capable of storing a byte of data. This memory would have a capacity of 16  8 ¼ 128 bits. Since the
address is implemented as a binary code, the number of lines in the address bus (n) will dictate the
number of address locations that the memory system will have (M ¼ 2n). Figure 10.1 shows a graphical
depiction of how data resides in memory. This type of graphic is called a memory map model.

# Springer International Publishing AG 2017 331

B.J. LaMeres, Introduction to Logic Circuits & Logic Design with Verilog,
DOI 10.1007/978-3-319-53883-9_10
332 • Chapter 10: Memory

Fig. 10.1
Memory map model

10.1.2 Volatile Versus Non-volatile Memory

Memory is classified into two categories depending on whether it can store information when power
is removed or not. The term non-volatile is used to describe memory that holds information when the
power is removed, while the term volatile is used to describe memory that loses its information when
power is removed. Historically, volatile memory is able to run at faster speeds compared to non-volatile
memory, so it is used as the primary storage mechanism while a digital system is running. Non-volatile
memory is necessary in order to hold critical operation information for a digital system such as start-up
instructions, operations systems, and applications.

10.1.3 Read Only Versus Read/Write Memory

Memory can also be classified into two categories with respect to how data is accessed. Read Only
Memory (ROM) is a device that cannot be written to during normal operation. This type of memory is
useful for holding critical system information or programs that should not be altered while the system is
running. Read/Write memory refers to memory that can be read and written to during normal operation
and is used to hold temporary data and variables.

10.1.4 Random Access Versus Sequential Access

Random Access Memory (RAM) describes memory in which any location in the system can be
accessed at any time. The opposite of this is sequential access memory, in which not all address
locations are immediately available. An example of a sequential access memory system is a tape drive.
In order to access the desired address in this system, the tape spool must be spun until the address is in
a position that can be observed. Most semiconductor memory in modern systems is random access. The
terms RAM and ROM have been adopted, somewhat inaccurately, to also describe groups of memory
with particular behavior. While the term ROM technically describes a system that cannot be written to, it
has taken on the additional association of being the term to describe non-volatile memory. While the term
RAM technically describes how data is accessed, it has taken on the additional association of being the
term to describe volatile memory. When describing modern memory systems, the terms RAM and ROM
are used most commonly to describe the characteristics of the memory being used; however, modern
memory systems can be both read/write and non-volatile, and the majority of memory is random access.

CONCEPT CHECK

CC10.1 An 8-bit wide memory has eight address lines. What is its capacity in bits?
(A) 64 (B) 256 (C) 1024 (D) 2048
 10.2 Non-volatile Memory Technology • 333

10.2 Non-volatile Memory Technology

10.2.1 ROM Architecture
This section describes some of the most common non-volatile memory technologies used to store
digital information. An address decoder is used to access individual data words within the memory
system. The address decoder asserts one and only one word line (WL) for each unique binary address
that is present on its input. This operation is identical to a binary-to-one-hot decoder. For an n-bit
address, the decoder can access 2n, or M words in memory. The word lines historically run horizontally
across the memory array, thus they are often called row lines and the word line decoder is often called
the row decoder. Bit lines (BL) run perpendicular to the word lines in order to provide individual bit storage
access at the intersection of the bit and word lines. These lines typically run vertically through the
memory array, thus they are often called column lines. The output of the memory system (i.e., Data_Out)
is obtained by providing an address and then reading the word from buffered versions of the bit lines.
When a system provides individual bit access to a row, or access to multiple data words sharing a row
line, a column decoder is used to route the appropriate bit line(s) to the data out port.
In a traditional ROM array, each bit line contains a pull-up network to VCC. This provides the ability to
store a logic 1 at all locations within the array. If a logic 0 is desired at a particular location, an NMOS pull-
down transistor is inserted. The gate of the NMOS is connected to the appropriate word line and the drain
of the NMOS is connected to the bit line. When reading, the word line is asserted and turns on the NMOS
transistor. This pulls the bit line to GND and produces a logic 0 on the output. When the NMOS transistor
is excluded, the bit line remains at a logic 1 due to the pull-up network. Figure 10.2 shows the basic
architecture of a ROM.
334 • Chapter 10: Memory

Fig. 10.2
Basic architecture of read only memory (ROM)

Figure 10.3 shows the operation of a ROM when information is being read.
 10.2 Non-volatile Memory Technology • 335

Fig. 10.3
ROM operation during a read

Memory can be designed to be either asynchronous or synchronous. Asynchronous memory

updates its data outputs immediately upon receiving an address. Synchronous memory only updates
its data outputs on the rising edge of a clock. The term latency is used to describe the delay between
when a signal is sent to the memory (either the address in an asynchronous system or the clock in a
synchronous system) and when the data is available. Figure 10.4 shows a comparison of the timing
diagrams between asynchronous and synchronous ROM systems during a read cycle.
336 • Chapter 10: Memory

Fig. 10.4
Asynchronous vs. synchronous ROM operation during a read cycle

10.2.2 Mask Read Only Memory (MROM)

A Mask Read Only Memory (MROM) is a non-volatile device that is programmed during fabrication.
The term mask refers to a transparent plate that contains patterns to create the features of the devices on
an integrated circuit using a process called photolithography. An MROM is fabricated with all of the
features necessary for the memory device with the exception of the final connections between the NMOS
transistors and the word and bit lines. This allows the majority of the device to be created prior to knowing
what the final information to be stored is. Once the desired information to be stored is provided by the
customer, the fabrication process is completed by adding connections between certain NMOS
transistors and the word/bit lines in order to create logic 0’s. Figure 10.5 shows an overview of the
MROM programming process.
 10.2 Non-volatile Memory Technology • 337

Fig. 10.5
MROM overview

10.2.3 Programmable Read Only Memory (PROM)

A Programmable Read Only Memory (PROM) is created in a similar manner as an MROM except
that the programming is accomplished post-fabrication through the use of fuses or anti-fuses. A fuse is
an electrical connection that is normally conductive. When a certain amount of current is passed through
the fuse it will melt, or blow, and create an open circuit. The amount of current necessary to open the fuse
is much larger than the current the fuse would conduct during normal operation. An anti-fuse operates in
the opposite manner as a fuse. An anti-fuse is normally an open circuit. When a certain amount of current
is forced into the anti-fuse, the insulating material breaks down and creates a conduction path. This turns
the anti-fuse from an open circuit into a wire. Again, the amount of current necessary to close the anti-
fuse is much larger than the current the anti-fuse would experience during normal operation. A PROM
uses fuses or anti-fuses in order to connect/disconnect the NMOS transistors in the ROM array to the
word/bit lines. A PROM programmer is used to burn the fuses or anti-fuses. A PROM can only be
programmed once in this manner, thus it is a read only memory and non-volatile. A PROM has the
advantage that programming can take place quickly as opposed to an MROM that must be programmed
through device fabrication. Figure 10.6 shows an example PROM device based on fuses.
338 • Chapter 10: Memory

Fig. 10.6
PROM overview

10.2.4 Erasable Programmable Read Only Memory (EPROM)

As an improvement to the one-time programming characteristic of PROMs, an electrically program-
mable ROM with the ability to be erased with ultra-violet (UV) light was created. The Erasable Program-
mable Read Only Memory (EPROM) is based on a floating-gate transistor. In a floating-gate transistor,
an additional metal-oxide structure is added to the gate of an NMOS. This has the effect of increasing the
threshold voltage. The geometry of the second metal-oxide is designed such that the threshold voltage is
high enough that normal CMOS logic levels are not able to turn the transistor on (i.e., VT1  VCC). This
threshold can be changed by applying a large electric field across the two metal structures in the gate.
This causes charge to tunnel into the secondary oxide, ultimately changing it into a conductor. This
phenomenon is called Fowler–Nordheim tunneling. The new threshold voltage is low enough that normal
CMOS logic levels are not able to turn the transistors off (i.e., VT2  GND). This process is how the
device is programmed. This process is accomplished using a dedicated programmer, thus the EPROM
must be removed from its system to program. Figure 10.7 shows an overview of a floating-gate transistor
and how it is programmed.
 10.2 Non-volatile Memory Technology • 339

Fig. 10.7
Floating-gate transistor – programming

In order to change the floating-gate transistor back into its normal state, the device is exposed to a
strong ultra-violet light source. When the UV light strikes the trapped charge in the secondary oxide, it
transfers enough energy to the charge particles that they can move back into the metal plates in the gate.
This, in effect, erases the device and restores it back to a state with a high threshold voltage. EPROMs
contain a transparent window on the top of their package that allows the UV light to strike the devices.
The EPROM must be removed from its system to perform the erase procedure. When the UV light erase
procedure is performed, every device in the memory array is erased. EPROMs are a significant
improvement over PROMs because they can be programmed multiple times; however, the programming
and erase procedures are manually intensive and require an external programmer and external eraser.
Figure 10.8 shows the erase procedure for a floating-gate transistor using UV light.
340 • Chapter 10: Memory

Fig. 10.8
Floating-gate transistor – erasing with UV light

An EPROM array is created in the exact same manner as in a PROM array with the exception that
additional programming circuitry is placed on the IC and a transparent window is included on the
package to facilitate erasing. An EPROM is non-volatile and read only since the programming procedure
takes place outside of its destination system.

10.2.5 Electrically Erasable Programmable Read Only Memory (EEPROM)

In order to address the inconvenient programming and erasing procedures associated with
EPROMs, the Electrically Erasable Programmable ROM (EEPROM) was created. In this type of circuit,
the floating-gate transistor is erased by applying a large electric field across the secondary oxide. This
electric field provides the energy to move the trapped charge from the secondary oxide back into the
metal plates of the gate. The advantage of this approach is that the circuitry to provide the large electric
field can be generated using circuitry on the same substrate as the memory array, thus eliminating the
need for an external UV light eraser. In addition, since the circuitry exists to generate large on-chip
voltages, the device can also be programmed without the need for an external programmer. This allows
an EEPROM to be programmed and erased while it resides in its target environment. Figure 10.9 shows
the procedure for erasing a floating-gate transistor using an electric field.

Fig. 10.9
Floating-gate transistor – erasing with electricity
 10.2 Non-volatile Memory Technology • 341

Early EEPROMs were very slow and had a limited number of program/erase cycles, thus they were
classified into the category of non-volatile, read only memory. Modern floating-gate transistors are now
capable of access times on scale with other volatile memory systems, thus they have evolved into one of
the few non-volatile, read/write memory technologies used in computer systems today.

10.2.6 FLASH Memory

One of the early drawbacks of EEPROM was that the circuitry that provided the capability to
program and erase individual bits also added to the size of each individual storage element. FLASH
EEPROM was a technology that attempted to improve the density of floating-gate memory by program-
ming and erasing in large groups of data, known as blocks. This allowed the individual storage cells to
shrink and provided higher density memory parts. This new architecture was called NAND FLASH and
provided faster write and erase times coupled with higher density storage elements. The limitation of
NAND FLASH was that reading and writing could only be accomplished in a block-by-block basis. This
characteristic precluded the use of NAND FLASH for run-time variables and data storage, but was well
suited for streaming applications such as audio/video and program loading. As NAND FLASH technol-
ogy advanced, the block size began to shrink and software adapted to accommodate the block-by-block
data access. This expanded the applications that NAND FLASH could be deployed in. Today, NAND
FLASH memory is used in nearly all portable devices (e.g., smart phones, tablets, etc.) and its use in
solid state hard drives is on pace to replace hard disk drives and optical disks as the primary non-volatile
storage medium in modern computers.

In order to provide individual word access, NOR FLASH was introduced. In NOR FLASH, circuitry is
added to provide individual access to data words. This architecture provided faster read times than
NAND FLASH, but the additional circuitry causes the write and erase times to be slower and the
individual storage cell size to be larger. Due to NAND FLASH having faster write times and higher
density, it is seeing broader scale adoption compared to NOR FLASH despite only being able to access
information in blocks. NOR FLASH is considered random access memory while NAND FLASH is
typically not; however, as the block size of NAND FLASH is continually reduced, its use for variable
storage is becoming more attractive. All FLASH memory is non-volatile and read/write.

CONCEPT CHECK

CC10.2 Which of the following is suitable for implementation in a read only memory?
(A) Variables that a computer program needs to continuously update.
(B) Information captured by a digital camera.
(C) A computer program on a spacecraft.
(D) Incoming digitized sound from a microphone.
342 • Chapter 10: Memory

10.3 Volatile Memory Technology

This section describes some common volatile memory technologies used to store digital
information.

10.3.1 Static Random Access Memory (SRAM)

Static Random Access Memory (SRAM) is a semiconductor technology that stores information
using a cross-coupled inverter feedback loop. Figure 10.10 shows the schematic for the basic SRAM
storage cell. In this configuration, two access transistors (M1 and M2) are used to read and write from the
storage cell. The cell has two complementary ports called Bit Line (BL) and Bit Line’ (BLn). Due to the
inverting functionality of the feedback loop, these two ports will always be the complement of each other.
This behavior is advantageous because the two lines can be compared to each other to determine the
data value. This allows the voltage levels used in the cell to be lowered while still being able to detect the
stored data value. Word Lines are used to control the access transistors. This storage element takes six
CMOS transistors to implement and is often called a 6T configuration. The advantage of this memory cell
is that it has very fast performance compared to other sub-systems because of its underlying technology
being simple CMOS transistors. SRAM cells are commonly implemented on the same IC substrate as
the rest of the system, thus allowing a fully integrated system to be realized. SRAM cells are used for
cache memory in computer systems.

Fig. 10.10
SRAM storage element (6T)

To build an SRAM memory system, cells are arranged in an array pattern. Figure 10.11 shows a
4  4 SRAM array topology. In this configuration, word lines are shared horizontally across the array in
order to provide addressing capability. An address decoder is used to convert the binary encoded
address into the appropriate word line assertions. N storage cells are attached to the word line to provide
the desired data word width. Bit lines are shared vertically across the array in order to provide data
access (either read or write). A data line controller handles whether data is read from or written to the
cells based on an external write enable (WE) signal. When WE is asserted (WE ¼ 1), data will be written
to the cells. When WE is de-asserted (WE ¼ 0), data will be read from the cells. The data line controller
also handles determining the correct logic value read from the cells by comparing BL to BLn. As more
cells are added to the bit lines, the signal magnitude being driven by the storage cells diminishes due to
the additional loading of the other cells. This is where having complementary data signals (BL and BLn)
 10.3 Volatile Memory Technology • 343

is advantageous because this effectively doubles the magnitude of the storage cell outputs. The
comparison of BL to BLn is handled using a differential amplifier that produces a full logic level output
even when the incoming signals are very small.
SRAM is volatile memory because when the power is removed, the cross-coupled inverters are not
able to drive the feedback loop and the data is lost. SRAM is also read/write memory because the
storage cells can be easily read from or written to during normal operation.

Fig. 10.11
4  4 SRAM array topology

Let’s look at the operation of the SRAM array when writing the 4-bit word “0111” to address “01”.
Figure 10.12 shows a graphical depiction of this operation. In this write cycle, the row address decoder
observes the address input “01” and asserts WL1. Asserting this word line enables all of the access
transistors (i.e., M1 and M2 in Fig. 10.10) of the storage cells in this row. The line drivers are designed to
have a stronger drive strength than the inverters in the storage cells so that they can override their values
during a write. The information “0111” is present on the Data_In bus and the write enable control line is
344 • Chapter 10: Memory

asserted (WE ¼ 1) to indicate a write. The data line controller passes the information to be stored to the
line drivers, which in turn converts each input into complementary signals and drives the bit lines. This
overrides the information in each storage cell connected to WL1. The address decoder then de-asserts
WL1 and the information is stored.

Fig. 10.12
SRAM operation during a write cycle – storing “0111” to address “01”

Now let’s look at the operation of the SRAM array when reading a 4-bit word from address “10”. Let’s
assume that this row was storing the value “1010”. Figure 10.13 shows a graphical depiction of this
operation. In this read cycle, the row address decoder asserts WL2, which allows the SRAM cells to drive
their respective bit lines. Note that each cell drives a complementary version of its stored value. The input
control line is de-asserted (WE ¼ 0), which indicates that the sense amps will read the BL and BLn lines
in order to determine the full logic value stored in each cell. This logic value is then routed to the
Data_Out port of the array. In an SRAM array, reading from the cell does not impact the contents of
the cell. Once the read is complete, WL2 is de-asserted and the read cycle is complete.
 10.3 Volatile Memory Technology • 345

Fig. 10.13
SRAM operation during a read cycle – reading “0101” from address “10”

10.3.2 Dynamic Random Access Memory (DRAM)

Dynamic Random Access Memory (DRAM) is a semiconductor technology that stores information
using a capacitor. A capacitor is a fundamental electrical device that stores charge. Figure 10.14 shows
the schematic for the basic DRAM storage cell. The capacitor is accessed through a transistor (M1).
Since this storage element takes one transistor and one capacitor, it is often referred to as a 1T1C
configuration. Just as in SRAM memory, word lines are used to access the storage elements. The term
digit line is used to describe the vertical connection to the storage cells. DRAM has an advantage over
SRAM in that the storage element requires less area to implement. This allows DRAM memory to have
much higher density compared to SRAM.
346 • Chapter 10: Memory

Fig. 10.14
DRAM storage element (1T 1C)

There are a variety of considerations that must be accounted for when using DRAM. First, the
charge in the capacitor will slowly dissipate over time due to the capacitors being non-ideal. If left
unchecked, eventually the data held in the capacitor will be lost. In order to overcome this issue,
DRAM has a dedicated circuit to refresh the contents of the storage cell. A refresh cycle involves
periodically reading the value stored on the capacitor and then writing the same value back again at
full signal strength. This behavior also means that that DRAM is volatile because when the power is
removed and the refresh cycle cannot be performed, the stored data is lost. DRAM is also considered
read/write memory because the storage cells can be easily read from or written to during normal
operation.
Another consideration when using DRAM is that the voltage of the word line must be larger than VCC
in order to turn on the access transistor. In order to turn on an NMOS transistor, the gate terminal must be
larger than the source terminal by at least a threshold voltage (VT). In traditional CMOS circuit design, the
source terminal is typically connected to ground (0 v). This means the transistor can be easily turned on
by driving the gate with a logic 1 (i.e., VCC) since this creates a VGS voltage much larger than VT. This is
not always the case in DRAM. In DRAM, the source terminal is not connected to ground, but rather to the
storage capacitor. In the worst-case situation, the capacitor could be storing a logic 1 (i.e., VCC). This
means that in order for the word line to be able to turn on the access transistor, it must be equal to or
larger than (VCC + VT). This is an issue because the highest voltage that the DRAM device has access to
is VCC. In DRAM, a charge pump is used to create a voltage larger than VCC + VT that is driven on the
word lines. Once this voltage is used, the charge is lost so the line must be pumped up again before its
next use. The process of “pumping up” takes time that must be considered when calculating the
maximum speed of DRAM. Figure 10.15 shows a graphical depiction of this consideration.

Fig. 10.15
DRAM charge pumping of word lines
 10.3 Volatile Memory Technology • 347

Another consideration when using DRAM is how the charge in the capacitor develops into an actual
voltage on the digital line when the access transistor is closed. Consider the simple 4  4 array of DRAM
cells shown in Fig. 10.16. In this topology, the DRAM cells are accessed using the same approach as in
the SRAM array from Fig. 10.11.

Fig. 10.16
Simple 4  4 DRAM array topology

One of the limitations of this simple configuration is that the charge stored in the capacitors cannot
develop a full voltage level across the digit line when the access transistor is closed. This is because the
digit line itself has capacitance that impacts how much voltage will be developed. In practice, the
capacitance of the digit line (CDL) is much larger than the capacitance of the storage cell (CS) due to
having significantly more area and being connected to numerous other storage cells. This becomes an
issue because when the storage capacitor is connected to the digit line, the resulting voltage on the digit
line (VDL) is much less than the original voltage on the storage cell (VS). This behavior is known as
charge sharing because when the access transistor is closed, the charge on both capacitors is
distributed across both devices and results in a final voltage that depends on the initial charge in the
system and the values of the two capacitors. Example 10.1 shows an example of how to calculate the
final digit line voltage when the storage cell is connected.
348 • Chapter 10: Memory

Example 10.1
Calculating the final digit line voltage in a DRAM based on charge sharing

The issue with the charge sharing behavior of a DRAM cell is that the final voltage on the word line is
not large enough to be detected by a standard logic gate or latch. In order to overcome this issue, modern
DRAM arrays use complementary storage cells and sense amplifiers. The complementary cells store
the original data and its complement. Two digit lines (DL and DLn) are used to read the contents of the
storage cells. DL and DLn are initially pre-charged to exactly VCC/2. When the access transistors are
closed, the storage cells will share their charge with the digit lines and move them slightly away from
VCC/2 in different directions. This allows twice the voltage difference to be developed during a read.
 10.3 Volatile Memory Technology • 349

A sense amplifier is then used to boost this small voltage difference into a full logic level that can be read
by a standard logic gate or latch. Figure 10.17 shows the modern DRAM array topology based on
complementary storage cells.

Fig. 10.17
Modern DRAM array topology based on complementary storage cells

The sense amplifier is designed to boost small voltage deviations from VCC/2 on DL and DLn to full
logic levels. The sense amplifier sits in-between DL and DLn and has two complementary networks, the
N-sense amplifier and the P-sense amplifier. The N-sense amplifier is used to pull a signal that is below
VCC/2 (either DL or DLn) down to GND. A control signal (N-Latch or NLATn) is used to turn on this
network. The P-sense amplifier is used to pull a signal that is above VCC/2 (either DL or DLn) up to VCC. A
control signal (Active Pull-Up or ACT) is used to turn on this network. The two networks are activated in a
sequence with the N-sense network activating first. Figure 10.18 shows an overview of the operation of a
DRAM sense amplifier.
350 • Chapter 10: Memory

Fig. 10.18
DRAM sense amplifier

Let’s now put everything together and look at the operation of a DRAM system during a read
operation. Figure 10.19 shows a simplified timing diagram of a DRAM read cycle. This diagram shows
the critical signals and their values when reading a logic 1. Notice that there is a sequence of steps that
must be accomplished before the information in the storage cells can be retrieved.
 10.3 Volatile Memory Technology • 351

Fig. 10.19
DRAM operation during a read cycle – reading a 1 from a storage cell

A DRAM write operation is accomplished by opening the access transistors to the complementary
storage cells using WL, disabling the pre-charge drivers and then writing full logic level signals to the
storage cells using the Data_In line driver.

CONCEPT CHECK

CC10.3 Which of the following is suitable for implementation in a read/write memory?

(A) A look up table containing the values of sine.
(B) Information captured by a digital camera.
(C) The boot up code for a computer.
(D) A computer program on a spacecraft.
352 • Chapter 10: Memory

10.4 Modeling Memory with Verilog

10.4.1 Read-Only Memory in Verilog
A read-only memory in Verilog can be defined in two ways. The first is to simply use a case
statement to define the contents of each location in memory based on the incoming address. A second
approach is to declare an array and then initialize its contents. When using an array, a separate
procedural block handles assigning the contents of the array to the output based on the incoming
address. The array can be initialized using either an initial block or through the file I/O system tasks
$readmemb() or $readmemh(). Example 10.2 shows two approaches for modeling a 4  4 ROM
memory. In this example the memory is asynchronous, meaning that as soon as the address changes
the data from the ROM will appear immediately. To model this asynchronous behavior the procedural
blocks are sensitive to the incoming address. In the simulation, each possible address is provided (i.e.,
“00”, “01”, “10”, and “11”) to verify that the ROM was initialized correctly.

Example 10.2
Behavioral models of a 4  4 asynchronous read only memory in Verilog
 10.4 Modeling Memory with Verilog • 353

A synchronous ROM can be created in a similar manner as in the asynchronous approach. The
only difference is that in a synchronous ROM, a clock edge is used to trigger the procedural block
that updates data_out. A sensitivity list is used that contains the clock to trigger the assignment. Example
10.3 shows two Verilog models for a synchronous ROM. Notice that prior to the first clock edge, the
simulator does not know what to assign to data_out so it lists the value as unknown (X).

Example 10.3
Behavioral models of a 4  4 synchronous read only memory in Verilog

10.4.2 Read/Write Memory in Verilog

In a simple read/write memory model, there is an output port that provides data when reading
(data_out) and an input port that receives data when writing (data_in). Within the module, an array signal
is declared with elements of type reg. To write to the array, signal assignment are made from the data_in
port to the element within the array corresponding to the incoming address. To read from the array, the
data_out port is assigned the element within the array corresponding to the incoming address. A write
enable (WE) signal tells the system when to write to the array (WE ¼ 1) or when to read from the array
(WE ¼ 0). In an asynchronous R/W memory, data is immediately written to the array when WE ¼ 1 and
data is immediately read from the array when WE ¼ 0. This is modeled using a procedural block with a
354 • Chapter 10: Memory

sensitivity list containing every input to the system. Example 10.4 shows an asynchronous R/W 4  4
memory system and functional simulation results. In the simulation, each address is initially read from to
verify that it does not contain data. The data_out port produces unknown (X) for the initial set of read
operations. Each address in the array is then written to. Finally, the array is read from verifying that the
data that was written can be successfully retrieved.

Example 10.4
Behavioral model of a 4  4 asynchronous read/write memory in Verilog

A synchronous read/write memory is made in a similar manner with the exception that a clock is
used to trigger the procedural block managing the signal assignments. In this case, the WE signal acts
as a synchronous control signal indicating whether assignments are read from or written to the RW array.
Example 10.5 shows the Verilog model for a synchronous read/write memory and the simulation
waveform showing both read and write cycles.
 10.4 Modeling Memory with Verilog • 355

Example 10.5
Behavioral model of a 4  4 synchronous read/write memory in Verilog

CONCEPT CHECK

CC10.4 Explain the advantage of modeling memory in Verilog without going into the details of the
storage cell operation.
(A) It allows the details of the storage cell to be abstracted from the functional
operation of the memory system.
(B) It is too difficult to model the analog behavior of the storage cell.
(C) There are too many cells to model so the simulation would take too long.
(D) It lets both ROM and R/W memory to be modeled in a similar manner.
356 • Chapter 10: Memory

Summary
v The term memory refers to large arrays of v The basic architecture of a ROM consists of
digital storage. The technology used in mem- intersecting bit lines (vertical) and word lines
ory is typically optimized for storage density (horizontal) that contain storage cells at their
at the expense of control capability. This is crossing points. The data is read out of the
different from a D-flip-flop, which is optimized ROM array using the bit lines. Each bit line
for complete control at the bit level. contains a pull-up resistor to initially store a
v A memory device always contains an logic 1 at each location. If a logic 0 is desired
address bus input. The number of bits in the at a certain location, a pull-down transistor is
address bus dictates how many storage placed on a particular bit line with its gate
locations can be accessed. An n-bit address connected to the appropriate word line.
bus can access 2n (or M) storage locations. When the storage cell is addressed, the
v The width of each storage location (N) allows word line will assert and turn on the pull-
the density of the memory array to be down transistor producing a logic 0 on the
increased by reading and writing vectors of output.
data instead of individual bits. v There are a variety of technologies to imple-
v A memory map is a graphical depiction of a ment the pull-down transistor in a ROM. Dif-
memory array. A memory map is useful to ferent ROM architectures include MROMs,
give an overview of the capacity of the array PROMs, EPROMs, and EEPROMs. These
and how different address ranges of the array memory types are non-volatile.
are used. v A R/W memory requires a storage cell that
v A read is an operation in which data is can be both read from and written to during
retrieved from memory. A write is an opera- normal operation. A DRAM (dynamic RAM)
tion in which data is stored to memory. cell is a storage element that uses a capaci-
v An asynchronous memory array responds tor to hold charge corresponding to a logic
immediately to its control inputs. A synchro- value. An SRAM (static RAM) cell is a stor-
nous memory array only responds on the age element that uses a cross-coupled
triggering edge of clock. inverter pair to hold the value being stored
v Volatile memory will lose its data when the in the positive feedback loop formed by the
power is removed. Non-volatile memory will inverters. Both DRAM and SRAM and vola-
retain its data when the power is removed. tile and random access.
v Read Only Memory (ROM) is a memory type v The floating-gate transistor enables memory
that cannot be written to during normal oper- that is both non-volatile and R/W. Modern
ation. Read/Write (R/W) memory is a mem- memory systems based on floating-gate
ory type that can be written to during normal transistor technology allow writing to take
operation. Both ROM and R/W memory can place using the existing system power supply
be read from during normal operation. levels. This type of R/W memory is called
v Random Access Memory (RAM) is a memory FLASH. In FLASH memory, the information
type in which any location in memory can be is read out in blocks, thus it is not technically
accessed at any time. In Sequential Access random access.
Memory the data can only be retrieved in a v Memory can be modeled in Verilog using an
linear sequence. This means that in sequen- array data type consisting of elements of
tial memory the data cannot be accessed type reg.
arbitrarily.

Exercise Problems
Section 10.1: Memory Architecture 10.1.4 For a 512 k  32-bit memory system, what is
the capacity in bytes?
and Terminology
10.1.5 For a 512 k  32 memory system, how wide
10.1.1 For a 512 k  32 memory system, how many
does the incoming address bus need to be in
unique address locations are there? Give the
order to access every unique address
exact number.
location?
10.1.2 For a 512 k  32 memory system, what is the
10.1.6 Name the type of memory with the following
data width at each address location?
characteristic: when power is removed, the
10.1.3 For a 512 k  32 memory system, what is the data is lost.
capacity in bits?
 Exercise Problems • 357

10.1.7 Name the type of memory with the following system that will write a 0 to the cell, then read it
characteristic: when power is removed, the back to verify the 0 was stored and then repeat
memory still holds its information. the write/read cycles for a 1.
10.1.8 Name the type of memory with the following
characteristic: it can only be read from during
normal operation.
10.1.9 Name the type of memory with the following
characteristic: during normal operation, it can
be read and written to.
10.1.10 Name the type of memory with the following
characteristic: data can be accessed from any
address location at any time.
10.1.11 Name the type of memory with the following
characteristic: data can only be accessed in
consecutive order, thus not every location of
memory is available instantaneously.

Section 10.2: Non-volatile Memory

Technology
10.2.1 Name the type of memory with the following
characteristic: this memory is non-volatile,
read/write, and only provides data access in
blocks. Fig. 10.20
10.2.2 Name the type of memory with the following SRAM cell block diagram
characteristic: this memory uses a floating gate
transistor, can be erased with electricity, and 10.3.4 Why is a DRAM cell referred to as a 1T 1C
provides individual bit access. configuration?
10.2.3 Name the type of memory with the following 10.3.5 Why is a charge pump necessary on the word
characteristic: this memory is non-volatile, read/ lines of a DRAM array?
write, and provides word-level data access.
10.3.6 Why does a DRAM cell require a refresh cycle?
10.2.4 Name the type of memory with the following
characteristic: this memory uses a floating- 10.3.7 For the DRAM storage cell shown in Fig. 10.21,
gate transistor that is erased with UV light. solve for the final voltage on the digit line after
the access transistor (M1) closes if initially
10.2.5 Name the type of memory with the following VS ¼ VCC (i.e., the cell is storing a 1). In this
characteristic: this memory is programmed by system, CS ¼ 5 pF, CDL ¼ 10 pF, and
blowing fuses or anti-fuses. VCC ¼ +3.4 v. Prior to the access transistor
10.2.6 Name the type of memory with the following closing, the digit line is pre-charged to VCC/2.
characteristic: this memory is partially
fabricated prior to knowing the information to
be stored.

Section 10.3: Volatile Memory

Technology
10.3.1 How many transistors does it take to imple-
ment an SRAM cell?
10.3.2 Why doesn’t an SRAM cell require a refresh
cycle?
10.3.3 Design a Verilog model for the SRAM system
shown in Fig. 10.20. Your storage cell should
be designed such that its contents can be
overwritten by the line driver. Consider using
signal strengths for this behavior (e.g., strong1
will overwrite a weak0). You will need to create
a system for the differential line driver with Fig. 10.21
enable. This driver will need to contain a high DRAM charge sharing exercise
impedance state when disabled. Both your line
driver (Din) and receiver (Dout) are differential.
These systems can be modeled using simple 10.3.8 For the DRAM storage cell shown in Fig. 10.21,
if-else statements. Create a test bench for your solve for the final voltage on the digit line after
358 • Chapter 10: Memory

the access transistor (M1) closes if initially 10.4.3 Design a Verilog model for the 16  8, asyn-
VS ¼ GND (i.e., the cell is storing a 0). In this chronous, read/write memory system shown in
system, CS ¼ 5 pF, CDL ¼ 10 pF, and Fig. 10.24. Create a test bench to simulate
VCC ¼ +3.4 v. Prior to the access transistor your model. Your test bench should first read
closing, the digit line is pre-charged to VCC/2. from all of the address locations to verify they
are uninitialized. Next, your test bench should
Section 10.4: Modeling Memory write unique information to each of the address
locations. Finally, your test bench should read
with Verilog from each address location to verify that the
10.4.1 Design a Verilog model for the 16  8, asyn- information that was written was stored and
chronous, read only memory system shown in can be successfully retrieved.
Fig. 10.22. The system should contain the
information provided in the memory map. Cre-
ate a test bench to simulate your model by
reading from each of the 16 unique addresses
and observing data_out to verify it contains the
information in the memory map.

Fig. 10.24
16  8 asynchronous R/W memory block diagram

10.4.4 Design a Verilog model for the 16  8, synchro-

nous, read/write memory system shown in
Fig. 10.25. Create a test bench to simulate
your model. Your test bench should first read
from all of the address locations to verify they
are uninitialized. Next, your test bench should
write unique information to each of the address
locations. Finally, your test bench should read
from each address location to verify that the
Fig. 10.22 information that was written was stored and
16  8 asynchronous ROM block diagram can be successfully retrieved.

10.4.2 Design a Verilog model for the 16  8, synchro-

nous, read only memory system shown in
Fig. 10.23. The system should contain the
information provided in the memory map. Cre-
ate a test bench to simulate your model by
reading from each of the 16 unique addresses
and observing data_out to verify it contains the
information in the memory map.

Fig. 10.25
16  8 synchronous R/W memory block diagram

Fig. 10.23
16  8 synchronous ROM block diagram
Chapter 11: Programmable Logic
This chapter provides an overview of programmable logic devices (PLDs). The term PLD is used as
a generic description for any circuit that can be programmed to implement digital logic. The technology
and architectures of PLDs have advanced over time. A historical perspective is given on how the first
programmable devices evolved into the programmable technologies that are prevalent today. The goal of
this chapter is to provide a basic understanding of the principles of programmable logic devices.
Learning Outcomes—After completing this chapter, you will be able to:

11.1 Describe the basic architecture and evolution of programmable logic devices.
11.2 Describe the basic architecture of Field Programmable Gate Arrays (FPGAs).

11.1 Programmable Arrays

11.1.1 Programmable Logic Array (PLA)
One of the first commercial PLDs developed using modern integrated circuit technology was the
programmable logic array (PLA). In 1970, Texas Instrument introduced the PLA with an architecture
that supported the implementation of arbitrary, sum of products logic expressions. The PLA was
fabricated with a dense array of AND gates, called an AND plane, and a dense array of OR gates,
called an OR plane. Inputs to the PLA each had an inverter in order to provide the original variable and its
complement. Arbitrary SOP logic expressions could be implemented by creating connections between
the inputs, the AND plane, and the OR plane. The original PLAs were fabricated with all of the necessary
features except the final connections to implement the SOP functions. When a customer provided the
desired SOP expression, the connections were added as the final step of fabrication. This configuration
technique was similar to an MROM approach. Figure 11.1 shows the basic architecture of a PLA.

Fig. 11.1
Programmable logic array (PLA) architecture

# Springer International Publishing AG 2017 359

B.J. LaMeres, Introduction to Logic Circuits & Logic Design with Verilog,
DOI 10.1007/978-3-319-53883-9_11
360 • Chapter 11: Programmable Logic

A more compact schematic for the PLA is drawn by representing all of the inputs into the AND and
OR gates with a single wire. Connections are indicated by inserting X’s at the intersections of wires.
Figue 11.2 shows this simplified PLA schematic implementing two different SOP logic expressions.

Fig. 11.2
Simplified PLA schematic

11.1.2 Programmable Array Logic (PAL)

One of the drawbacks of the original PLA was that the programmability of the OR plane caused
significant propagation delays through the combinational logic circuits. In order to improve on the
performance of PLAs, the programmable array logic (PAL) was introduced in 1978 by the company
Monolithic Memories, Inc. The PAL contained a programmable AND-plane and a fixed-OR plane. The
fixed-OR plane improved the performance of this programmable architecture. While not having a
programmable OR-plane reduced the flexibility of the device, most SOP expressions could be
manipulated to work with a PAL. Another contribution of the PAL was that the AND-plane could be
programmed using fuses. Initially, all connections were present in the AND-plane. An external program-
mer was used to blow fuses in order to disconnect the inputs from the AND gates. While the fuse
approach provided one-time-only programming, the ability to configure the logic post-fabrication was a
significant advancement over the PLA, which had to be programmed at the manufacturer. Figure 11.3
shows the architecture of a PAL.
 11.1 Programmable Arrays • 361

Fig. 11.3
Programmable array logic (PAL) architecture

11.1.3 Generic Array Logic (GAL)

As the popularity of the PAL grew, additional functionality was implemented to support more
sophisticated designs. One of the most significant improvements was the addition of an output logic
macrocell (OLMC). An OLMC provided a D-Flip-Flop and a selectable mux so that the output of the SOP
circuit from the PAL could be used either as the system output or the input to a D-Flip-Flop. This enabled
the implementation of sequential logic and finite state machines. The OLMC also could be used to route
the I/O pin back into the PAL to increase the number of inputs possible in the SOP expressions. Finally,
the OLMC provided a multiplexer to allow feedback from either the PAL output or the output of the D-Flip-
Flop. This architecture was named a generic array logic (GAL) to distinguish its features from a
standard PAL. Figure 11.4 shows the architecture of a GAL consisting of a PAL and an OLMC.
362 • Chapter 11: Programmable Logic

Fig. 11.4
Generic array logic (GAL) architecture

11.1.4 Hard Array Logic (HAL)

For mature designs, PALs and GALs could be implemented as a hard array logic (HAL) device.
A HAL was a version of a PAL or GAL that had the AND plane connections implemented during
fabrication instead of through blowing fuses. This architecture was more efficient for high volume
applications as it eliminated the programming step post-fabrication and the device did not need to
contain the additional programming circuitry.

In 1983, Altera Inc., was founded as a programmable logic device company. In 1984, Altera
released its first version of a PAL with a unique feature that it could be programmed and erased multiple
times using a programmer and an UV light source similar to an EEPROM.

11.1.5 Complex Programmable Logic Devices (CPLD)

As the demand for larger programmable devices grew, the PAL”s architecture was not able to scale
efficiently due to a number of reasons: first, as the size of combinational logic circuits increased, the PAL
encountered fan-in issues in its AND plane; secondly, for each input that was added to the PAL, the
amount of circuitry needed on the chip grew geometrically due to requiring a connection to each AND
gate in addition to the area associated with the additional OLMC. This led to a new PLD architecture in
which the on-chip interconnect was partitioned across multiple PALs on a single chip. This partitioning
meant that not all inputs to the device could be used by each PAL so the design complexity increased;
however, the additional programmable resources outweighed this drawback and this architecture was
broadly adopted. This new architecture was called a complex programmable logic device (CPLD).
 11.2 Field Programmable Gate Arrays (FPGAs) • 363

The term simple programmable logic device (SPLD) was created to describe all of the previous PLD
architectures (i.e., PLA, PAL, GAL, and HAL). Figure 11.5 shows the architecture of the CPLD.

Fig. 11.5
Complex PLD (CPLD) architecture

CONCEPT CHECK

CC11.1 What is the only source of delay mismatch from the inputs to the outputs in a programmable
array?
(A) The AND gates will have different delays due to having different numbers of
inputs.
(B) The OR gates will have different delays due to having different numbers of inputs.
(C) An input may or may not go through an inverter before reaching the AND gates.
(D) None. All paths through the programmable array have identical delay.

11.2 Field Programmable Gate Arrays (FPGAs)

To address the need for even more programmable resources, a new architecture was developed by
Xilinx Inc. in 1985. This new architecture was called a field programmable gate array (FPGA). An
FPGA consists of an array of programmable logic blocks (or logic elements) and a network of program-
mable interconnect that can be used to connect any logic element to any other logic element. Each logic
block contained circuitry to implement arbitrary combinational logic circuits in addition to a D-Flip-Flop
and a multiplexer for signal steering. This architecture effectively implemented an OLMC within each
block, thus providing ultimate flexibility and providing significantly more resources for sequential logic.
364 • Chapter 11: Programmable Logic

Today, FPGAs are the most commonly used programmable logic device with Altera Inc. and Xilinx Inc.
being the two largest manufacturers. Figure 11.6 shows the generic architecture of an FPGA.

Fig. 11.6
Field programmable gate array (FPGA) architecture

11.2.1 Configurable Logic Block (or Logic Element)

The primary reconfigurable structure in the FPGA is the configurable logic block (CLB) or Logic
Element (LE). Xilinx Inc. uses the term CLB while Altera uses LE. Combinational logic is implemented
using a circuit called a Look-Up Table (LUT), which can implement any arbitrary truth table. The details
of a LUT are given in the next section. The CLB/LE also contains a D-Flip-Flop for sequential logic. A
signal steering multiplexer is used to select whether the output of the CLB/LE comes from the LUT or
from the D-Flip-Flop. The LUT can be used to drive a combinational logic expression into the D input of
 11.2 Field Programmable Gate Arrays (FPGAs) • 365

the D-Flip-Flop, thus creating a highly efficient topology for finite state machines. A global routing
network is used to provide common signals to the CLB/LE such as clock, reset and enable. This global
routing network can provide these common signals to the entire FPGA or local groups of CLB/LEs.
Figure 11.7 shows the topology of a simple CLB/LE.

Fig. 11.7
Simple FPGA configurable logic block (or logic element)

CLB/LEs have evolved to include numerous other features such as carry in/carry out signals so that
arithmetic operations can be cascaded between multiple blocks in addition to signal feedback and D-flip-
Flop initialization.

11.2.2 Look-Up Tables (LUTs)

A look-up table is the primary circuit used to implement combinational logic in FPGAs. This topology
has also been adopted in modern CPLDs. In a LUT, the desired outputs of a truth table are loaded into a
local configuration SRAM memory. The SRAM memory provides these values to the inputs of a
multiplexer. The inputs to the combinational logic circuit are then used as the select lines to the
multiplexer. For an arbitrary input to the combinational logic circuit, the multiplexer selects the appropri-
ate value held in the SRAM and routes it to the output of the circuit. In this way, the multiplexer looks up
the appropriate output value based on the input code. This architecture has the advantage that any logic
function can be created without creating a custom logic circuit. Also, the delay through the LUT is
identical regardless of what logic function is being implemented. Figure 11.8 shows a 2-input combina-
tional logic circuit implemented with a 4-input multiplexer.
366 • Chapter 11: Programmable Logic

Fig. 11.8
2-input LUT implemented with a 4-input multiplexer

Fan-in limitations can be encountered quickly in LUTs as the number of inputs of the combinational
logic circuit being implemented grows. Recall that multiplexers are implemented with an SOP topology in
which each product term in the first level of logic has a number of inputs equal to the number of select
lines plus one. Also recall that the sum term in the second level of logic in the SOP topology has a
number of inputs equal to the total number of inputs to the multiplexer. In the example circuit shown in
Fig. 11.8, each product term in the multiplexer will have three inputs and the sum term will have four
inputs. As an illustration of how quickly fan-in limitations are encountered, consider the implication of
increasing the number of inputs in Fig. 11.8 from two to three. In this new configuration, the number of
inputs in the product terms will increase from three to four and the number of inputs in the sum term will
increase to from four to eight. Eight inputs is often beyond the fan-in specifications of modern devices,
meaning that even a 3-input combinational logic circuit will encounter fan-in issues when implemented
using a LUT topology.
To address this issue, multiplexer functionality in LUTs is typically implemented as a series of
smaller, cascaded multiplexers. Each of the smaller multiplexers progressively choose which row of
the truth table to route to the output of the LUT. This eliminates fan-in issues at the expense of adding
additional levels of logic to the circuit. While cascading multiplexers increases the overall circuit delay,
this approach achieves a highly consistent delay because regardless of the truth table output value, the
number of levels of logic through the multiplexers is always the same. Figure 11.9 shows how the 2-input
truth table from Fig. 11.8 can be implemented using a 2-level cascade of 2-input multiplexers.
 11.2 Field Programmable Gate Arrays (FPGAs) • 367

Fig. 11.9
2-input LUT implemented with a 2-level cascade of 2-input multiplexers

If more inputs are needed in the LUT, additional MUX levels are added. Figure 11.10 shows the
architecture for a 3-input LUT implemented with a 3-level cascade of 2-input multiplexers.
368 • Chapter 11: Programmable Logic

Fig. 11.10
3-Input LUT implemented with a 3-level cascade of 2-input multiplexers

Modern FPGAs can have LUTs with up to 6 inputs. If even more inputs are needed in a combina-
tional logic expression, then multiple CLB/LEs are used that form even larger LUTs.

11.2.3 Programmable Interconnect Points (PIPs)

The configurable routing network on an FPGA is accomplished using programmable switches.
A simple model for these switches is to use an NMOS transistor. A configuration SRAM bit stores
whether the switch is opened or closed. On the FPGA, interconnect is routed vertically and horizontally
between the CLB/LEs with switching points placed throughout the FPGA to facilitate any arbitrary routing
configuration. Figure 11.11 shows how the routing can be configured into a full cross-point configuration
using programmable switches.
 11.2 Field Programmable Gate Arrays (FPGAs) • 369

Fig. 11.11
FPGA programmable interconnect

11.2.4 Input/Output Block (IOBs)

FPGAs also contain Input/Output Blocks (IOB) that provide programmable functionality for inter-
facing to external circuitry. The IOBs contain both driver and receiver circuitry so that they can be
programmed to be either inputs or outputs. D-Flip-Flops are included in both the input and output
circuitry to support synchronous logic. Figure 11.12 shows the architecture of an FPGA IOB.
370 • Chapter 11: Programmable Logic

Fig. 11.12
FPGA input/output block (IOB)

11.2.5 Configuration Memory

All of the programming information for an FPGA is contained within configuration SRAM that is
distributed across the IC. Since this memory is volatile, the FPGA will lose its configuration when power
is removed. Upon power-up, the FPGA must be programmed with its configuration data. This data is
typically held in a non-volatile memory such as FLASH. The “FP” in FPGA refers to the ability to program
the device in the field, or post-fabrication. The “GA” in FPGA refers to the array topology of the
programmable logic blocks or elements.

CONCEPT CHECK

CC11.2 What is the primary difference between an FPGA and a CPLD?

(A) The ability to create arbitrary SOP logic expressions.
(B) The abundance of configurable routing.
(C) The inclusion of D-flip-flops.
(D) The inclusion of programmable I/O pins.
 Exercise Problems • 371

Summary
v A programmable logic device (PLD) is a v FPGAs contain an array of programmable
generic term for a circuit that can be logic elements that each consist of combina-
configured to implement arbitrary logic tional logic capability and sequential logic
functions. storage. FPGAs also contain a programma-
v There are a variety of PLD architectures that ble interconnect network that provides the
have been used to implement combinational highest level of flexibility in programmable
logic. These include the PLA and PAL. These logic.
devices contain an AND-plane and an v A look-up-table (LUT) is a simple method to
OR-plane. The AND-plane is configured to create a programmable combinational logic
implement the product terms of a SOP circuit. A LUT is simply a multiplexer with the
expression. The OR-plane is configured to inputs to the circuit connected to the select
implement the sum term of a SOP lines of the MUX. The desired outputs of the
expression. truth table are connected to the MUX inputs.
v A GAL increases the complexity of logic As different input codes arrive on the select
arrays by adding sequential logic storage lines of the MUX, they select the
and programmable I/O capability. corresponding logic value to be routed to
v A CPLD significantly increases the density of the system output.
PLDs by connecting an array of PALs
together and surrounding the logic with I/O
drivers.

Exercise Problems
Section 11.1: Programmable Arrays
11.1.1 Name the type of programmable logic
described by the characteristic: this device
adds an output logic macrocell to a traditional
PAL.
11.1.2 Name the type of programmable logic
described by the characteristic: this device
combines multiple PALs on a single chip with
a partitioned interconnect system.
11.1.3 Name the type of programmable logic
described by the characteristic: this device
has a programmable AND-plane and program-
mable OR-plane.
11.1.4 Name the type of programmable logic
described by the characteristic: this device
has a programmable AND-plane and fixed
OR-plane.
11.1.5 Name the type of programmable logic Fig. 11.13
described by the characteristic: this device is Blank PAL Schematic
a PAL or GAL that is programmed during
manufacturing. Section 11.2: Field Programmable Gate
11.1.6 For the following unconfigured PAL schematic Arrays (FPGAs)
in Fig. 11.13, draw in the connection points 11.2.1 Give a general description of a Field Program-
(i.e., the X’s) to implement the two SOP logic mable Gate Array that differentiates it from
expressions shown on the outputs. other programmable logic devices.
372 • Chapter 11: Programmable Logic

11.2.2 Which part of an FPGA is described by the

following characteristic: this is used to interface
between the internal logic and external
circuitry.
11.2.3 Which part of an FPGA is described by the
following characteristic: this is used to config-
ure the on-chip routing.
11.2.4 Which part of an FPGA is described by the
following characteristic: this is the primary pro-
grammable element that makes up the array.
11.2.5 Which part of an FPGA is described by the
following characteristic: this part is used to
implement the combinational logic within the
array.
11.2.6 Draw the logic diagram of a 4-Input Look-Up
Table (LUT) to implement the truth table
provided in Fig. 11.14. Implement the LUT
with only 2-input multiplexers. Be sure to label
the exact location of the inputs (A, B, C, and D),
the desired value for each row of the truth
Fig. 11.14
table, and the output (F) in the diagram.
4-input LUT exercise
Chapter 12: Arithmetic Circuits
This chapter presents the design and timing considerations of circuits to perform basic arithmetic
operations including addition, subtraction, multiplication, and division. A discussion is also presented on
how to model arithmetic circuits in Verilog. The goal of this chapter is to provide an understanding of the
basic principles of binary arithmetic circuits.
Learning Outcomes—After completing this chapter, you will be able to:

12.1 Design a binary adder using both the classical digital design approach and the modern
HDL-based approach.
12.2 Design a binary subtractor using both the classical digital design approach and the modern
HDL-based approach.
12.3 Design a binary multiplier using both the classical digital design approach and the modern
HDL-based approach.
12.4 Design a binary divider using both the classical digital design approach and the modern
HDL-based approach.

12.1 Addition
Binary addition is performed in a similar manner to performing decimal addition by hand. The
addition begins in the least significant position of the number (p ¼ 0). The addition produces the sum
for this position. In the event that this positional sum cannot be represented by a single symbol, then the
higher order symbol is carried to the subsequent position (p ¼ 1). The addition in the next higher
position must include the number that was carried in from the lower positional sum. This process
continues until all of the symbols in the number have been operated on. The final positional sum can
also produce a carry, which needs to be accounted for in a separate system.
Designing a binary adder involves creating a combinational logic circuit to perform the positional
additions. Since a combinational logic circuit can only produce a scalar output, circuitry is needed to
produce the sum and the carry at each position. The binary adder size is pre-determined and fixed prior
to implementing the logic (i.e., an n-bit adder). Both inputs to the adder must adhere to the fixed size,
regardless of their value. Smaller numbers simply contain leading zeros in their higher order positions.
For an n-bit adder, the largest sum that can be produced will require n + 1 bits. To illustrate this, consider
a 4-bit adder. The largest numbers that the adder will operate on are 11112 + 11112. (or 1510 + 1510). The
result of this addition is 111102 (or 3010). Notice that the largest sum produced fits within 5 bits, or n + 1.
When constructing an adder circuit, the sum is always recorded using n-bits with a separate carry out bit.
In our 4-bit example, the sum would be expressed as “1110” with a carry out. The carry out bit can be
used in multiple word additions, used as part of the number when being decoded for a display, or simply
discarded as in the case when using two’s complement numbers.

12.1.1 Half Adders

When creating an adder, it is desirable to design incremental sub-systems that can be re-used. This
reduces design effort and minimizes troubleshooting complexity. The most basic component in the adder
is called a half adder. This circuit computes the sum and carry out on two input arguments. The reason it
is called a half adder instead of a full adder is because it does not accommodate a carry in during the
computation, thus it does not provide all of the necessary functionality required for the positional adder.
Example 12.1 shows the design of a half adder. Notice that two combinational logic circuits are required

# Springer International Publishing AG 2017 373

B.J. LaMeres, Introduction to Logic Circuits & Logic Design with Verilog,
DOI 10.1007/978-3-319-53883-9_12
374 • Chapter 12: Arithmetic Circuits

in order to produce the sum (the XOR gate) and the carry out (the AND gate). These two gates are in
parallel to each other, thus the delay through the half adder is due to only one level of logic.

Example 12.1
Design of a half adder

12.1.2 Full Adders

A full adder is a circuit that still produces a sum and carry out, but considers three inputs in the
computations (A, B, and Cin). Example 12.2 shows the design of a full adder.

Example 12.2
Design of a full adder
 12.1 Addition • 375

As mentioned before, it is desirable to re-use design components as we construct more complex

systems. One such design re-use approach is to create a full adder using two half adders. This is
straightforward for the sum output since the logic is simply two cascaded XOR gates (Sum ¼ A  B  Cin).
The carry out is not as straightforward. Notice that the expression for Cout derived in Example 12.2
contains the term (A + B). If this term could be manipulated to use an XOR gate instead, it would allow the
full adder to take advantage of existing circuitry in the system. Fig. 12.1 shows a derivation of an
equivalency that allows (A + B) to be replaced with (A  B) in the Cout logic expression.

Fig. 12.1
A useful logic equivalency that can be exploited in arithmetic circuits

The ability to implement the carry out logic using the expression Cout ¼ AB + (A  B)Cin allows us to
implement a full adder with two half adders and the addition of a single OR gate. Example 12.3 shows
this approach. In this new configuration, the sum is produced in two levels of logic while the carry out is
produced in three levels of logic.
376 • Chapter 12: Arithmetic Circuits

Example 12.3
Design of a full adder out of half adders

12.1.3 Ripple Carry Adder (RCA)

The full adder can now be used in the creation of multi-bit adders. The simplest topology exploiting
the full adder is called a ripple carry adder (RCA). In this approach, full adders are used to create the sum
and carry out of each bit position. The carry out of each full adder is used as the carry in for the next
higher position. Since each subsequent full adder needs to wait for the carry to be produced by the
preceding stage, the carry is said to ripple through the circuit, thus giving this approach its name.
Example 12.4 shows how to design a 4-bit ripple carry adder using a chain of full adders. Notice that
the carry in for the full adder in position 0 is tied to a logic 0. The 0 input has no impact on the result of the
sum but enables a full adder to be used in the 0th position.
 12.1 Addition • 377

Example 12.4
Design of a 4-bit ripple carry adder (RCA)

While the ripple carry adder provides a simple architecture based on design re-use, its delay can
become considerable when scaling to larger inputs sizes (e.g., n ¼ 32 or n ¼ 64). A simple analysis of the
timing can be stated such that if the time for a full adder to complete its positional sum is tFA, then the time
for an n-bit ripple carry adder to complete its computation is tRCA ¼ ntFA.
If we examine the RCA in more detail, we can break down the delay in terms of the levels of logic
necessary for the computation. Example 12.5 shows the timing analysis of the 4-bit RCA. This analysis
determines the number of logic levels in the adder. The actual gate delays can then be plugged in to find
the final delay. The inputs to the adder are A, B and Cin and are always assumed to update at the same
time. The first full adder requires two levels of logic to produce its sum and three levels to produce its
carry out. Since the timing of a circuit is always stated as its worst case delay, we say that the first full
adder takes three levels of logic. When the carry (C1) ripples to the next full adder (FA1), it must
propagate through two additional levels of logic in order to produce C2. Notice that the first half adder
in FA1 only depends on A1 and B1, thus it is able to perform this computation immediately. This half adder
can be considered as first level logic. More importantly, it means that when the carry in arrives (C1),
only two additional levels of logic are needed, not three. The levels of logic for the RCA can be expressed
as 3 + 2(n  1). If each level of logic has a delay of tgate, then a more accurate expression for the RCA
delay is tRCA ¼ (3 + 2(n  1))tgate.
378 • Chapter 12: Arithmetic Circuits

Example 12.5
Timing analysis of a 4-bit ripple carry adder

12.1.4 Carry Look Ahead Adder (CLA)

In order to address the potentially significant delay of a ripple carry adder, a carry look ahead (CLA)
adder was created. In this approach, additional circuitry is included that produces the intermediate carry
in signals immediately instead of waiting for them to be created by the preceding full adder stage. This
allows the adder to complete in a fixed amount of time instead of one that scales with the number of bits in
the adder. Example 12.6 shows an overview of the design approach for a CLA.
 12.1 Addition • 379

Example 12.6
Design of a 4-bit carry look ahead adder (CLA) – overview

For the CLA architecture to be effective, the look ahead circuitry needs to be dependent only on the
system inputs A, B, and Cin (i.e., C0). A secondary characteristic of the CLA is that it should exploit as
much design re-use as possible. In order to examine the design re-use aspects of a multi-bit adder, the
concepts of carry generation (g) and propagation (p) are used. A full adder is said to generate a carry if
its inputs A and B result in Cout ¼ 1 when Cin ¼ 0. A full adder is said to propagate a carry if its inputs A
and B result in Cout ¼ 1 when Cin ¼ 1. These simple statements can be used to derive logic expressions
for each stage of the adder that can take advantage of existing logic terms from prior stages. Example
12.7 shows the derivation of these terms and how algebraic substitutions can be exploited to create look
ahead circuitry for each full adder that is only dependent on the system inputs. In these derivations, the
variable i is used to represent position since p is used to represent the propagate term.
380 • Chapter 12: Arithmetic Circuits

Example 12.7
Design of a 4-bit carry look ahead adder (CLA) – algebraic formation

Example 12.8 shows a timing analysis of the 4-bit carry look ahead adder. Notice that the full adders
are modified to add the logic for the generate and propagate bits in addition to removing the unnecessary
gates associated with creating the carry out.
 12.1 Addition • 381

Example 12.8
Timing analysis of a 4-bit carry look ahead adder

The 4-bit CLA can produce the sum in four levels of logic as long as fan-in specifications are met. As
the CLA width increases, the look ahead circuitry will become fan-in limited and additional stages will be
required to address the fan-in. Regardless, the CLA has considerably less delay than a RCA as the width
of the adder is increased.

12.1.5 Adders in Verilog

12.1.5.1 Structural Model of a Ripple Carry Adder in Verilog
A structural model of a ripple carry adder is useful to visualize the propagation delay of the circuit in
addition to the impact of the carry rippling through the chain. Example 12.9 shows the structural model for
a full adder in Verilog consisting of two half adders. The half adders are created using two gate-level
primitives for the XOR and AND operations, each with a delay of 1 ns. The full adder is created by
instantiating two versions of the half adder as sub-systems plus one additional gate-level primitive for the
OR gate.
382 • Chapter 12: Arithmetic Circuits

Example 12.9
Structural model of a full adder using two half adders in Verilog

Example 12.10 shows the structural model of a 4-bit ripple carry adder in Verilog. The RCA is
created by instantiating four full adders. Notice that a logic 1’b0 can be directly inserted into the port map
of the first full adder to model the behavior of C0 ¼ 0.
 12.1 Addition • 383

Example 12.10
Structural model of a 4-bit ripple carry adder in Verilog

When creating arithmetic circuitry, testing under all input conditions is necessary to verify function-
ality. Testing under each and every input condition can require a large number of input conditions. To test
an n-bit adder under each and every numeric input condition will take (2n)2 test vectors. For our simple
4-bit adder example, this equates to 256 input patterns. The large number of input patterns precludes the
use of manual signal assignments in the test bench to stimulate the circuit. One approach to generating
the input test patterns is to use nested for loops. Example 12.11 shows a test bench that uses two nested
for loops to generate the 256 unique input conditions for the 4-bit ripple carry adder. Note that the loop
variables i and j are declared as type integer and then automatically incremented within the for loops.
Within the loops, the loop variables i and j are assigned to the DUT inputs A_TB and B_TB. The
truncation to 4-bits is automatically handled in Verilog. The simulation waveform illustrates how the
ripple carry adder has a noticeable delay before the output sum is produced. During the time the carry is
rippling through the adder chain, glitches can appear on each of the sum bits in addition to the carry out
signal. The values in this waveform are displayed as unsigned decimal symbols to make the results
easier to interpret.
384 • Chapter 12: Arithmetic Circuits

Example 12.11
Test bench for a 4-bit ripple carry adder using nested for loops in Verilog

12.1.5.2 Structural Model of a Carry Look Ahead Adder in Verilog

A carry look ahead adder can also be modeled using procedural assignments and modified full
adder sub-systems. Example 12.12 shows a structural model for a 4-bit CLA in Verilog. In this example,
the gate delay is modeled at 1 ns. The delay due to multiple levels of logic is entered manually to simplify
the model. The two cascaded XOR gates in the modified full adder are modeled using a single, 3-input
gate primitive with 2 ns of delay.
 12.1 Addition • 385

Example 12.12
Structural model of a 4-bit carry look ahead adder in Verilog

Example 12.13 shows the simulation waveform for the 4-bit carry look ahead adder. The outputs still
have intermediate transitions while the combinational logic is computing the results; however, the overall
delay of the adder is bound to 4*tgate.

Example 12.13
4-bit carry look ahead adder – simulation waveform
386 • Chapter 12: Arithmetic Circuits

12.1.5.3 Behavior Model of an Adder using Arithmetic Operators in Verilog

Verilog also supports adder models at a higher level of abstraction using the “+” operator. Note that
when adding two n-bit arguments the sum produced will be n + 1 bits. This can be handled in Verilog by
concatenating the Cout and Sum outputs on the LHS of the assignment. The entire add operation can be
accomplished in a single continuous assignment that contains both the concatenation and addition
operators. When using continuous assignment, the LHS must be a net data type. This means the outputs
Cout and Sum need to be declared as type wire. If it was desired to have the outputs declared of type reg,
a procedural assignment could be used instead. Example 12.14 shows the behavioral model for a 4-bit
adder in Verilog.

Example 12.14
Behavioral model of a 4-bit adder in Verilog

CONCEPT CHECK

CC12.1 Does a binary adder behave differently when it’s operating on unsigned vs. two’s comple-
ment numbers? Why or why not?
(A) Yes. The adder needs to keep track of the sign bit, thus extra circuitry is needed.
(B) No. The binary addition is identical. It is up to the designer to handle how the two’s
complement codes are interpreted and whether two’s complement overflow
occurred using a separate system.

12.2 Subtraction
Binary subtraction can be accomplished by building a dedicated circuit using a similar design
approach as just described for adders. A more effective approach is to take advantage of two’s
complement representation in order to re-use existing adder circuitry. Recall that taking the two’s
complement of a number will produce an equivalent magnitude number, but with the opposite sign
(i.e., positive to negative or negative to positive). This means that all that is required to create a subtractor
from an adder is to first take the two’s complement of the subtrahend input. Since the steps to take the
two’s complement of a number involve complementing each of the bits in the number and then adding
 12.2 Subtraction • 387

1, the logic required is relatively simple. Example 12.15 shows a 4-bit subtractor using full adders. The
subtrahend B is inverted prior to entering the full adders. Also, the carry in bit C0 is set to 1. This handles
the “adding 1” step of the two’s complement. All of the carries in the circuit are now treated as borrows
and the sum is now treated as the difference.

Example 12.15
Design of a 4-bit subtractor using full adders

A programmable adder/subtractor can be created with the use of a programmable inverter and a
control signal. The control signal will selectively invert B and also change the C0 bit between a 0 (for
adding) and a 1 (for subtracting). Example 12.16 shows how an XOR gate can be used to create a
programmable inverter for use in a programmable adder/subtractor circuit.

Example 12.16
Creating a programmable inverter using an XOR gate
388 • Chapter 12: Arithmetic Circuits

We can now define a control signal called (ADDn/SUB) that will control whether the circuit performs
addition or subtraction. Example 12.17 shows the architecture of a 4-bit programmable adder/subtractor.
It should be noted that this programmability adds another level of logic to the circuit, thus increasing its
delay. The programmable architecture in Example 12.17 is shown for a ripple carry adder; however, this
approach works equally well for a carry look ahead adder architecture.

Example 12.17
Design of a 4-bit programmable adder/subtractor

When using two’s complement representation in arithmetic, care must be taken to monitor for two’s
complement overflow. Recall that when using two’s complement representation, the number of bits of the
numbers is fixed (e.g., 4-bits) and if a carry/borrow out is generated, it is ignored. This means that the
Cout bit does not indicate whether two’s complement overflow occurred. Instead, we must construct
additional circuitry to monitor the arithmetic operations for overflow. Recall from Chap. 2 that two’s
complement overflow occurs in any of these situations:

• The sum of like signs results in an answer with opposite sign

(i.e., Positive + Positive ¼ Negative or Negative + Negative ¼ Positive).
• The subtraction of a positive number from a negative number results in a positive number
(i.e., Negative – Positive ¼ Positive).
• The subtraction of a negative number from a positive number results in a negative number
(i.e., Positive – Negative ¼ Negative).

The construction of circuitry for these conditions is straightforward since the sign bit of all numbers
involved in the operation indicates whether the number is positive or negative. The sign bits of the input
arguments and the output are fed into combinational logic circuitry that will assert for any of the above
conditions. These signals are then logically combined to create two’s complement overflow signal.
 12.3 Multiplication • 389

CONCEPT CHECK

CC12.2 What modifications can be made to the programmable adder/subtractor architecture so that
it can be used to take the 2’s complement of a number?
(A) Remove the input A.
(B) Add an additional control signal that will cause the circuit to ignore A and just
perform a complement on B and then add 1.
(C) Add an additional 1 to the original number using an OR gate on Cin.
(D) Set A to 0, put the number to be manipulated on B, and put the system into
subtraction mode. The system will then complement the bits on B and then add
1, thus performing two’s complement negation.

12.3 Multiplication
12.3.1 Unsigned Multiplication
Binary multiplication is performed in a similar manner to performing decimal multiplication by hand.
Recall the process for long multiplication. First, the two numbers are placed vertically over one another
with their least significant digits aligned. The upper number is called the multiplicand and the lower
number is called the multiplier. Next, we multiply each individual digit within multiplier with the entire
multiplicand, starting with the least position. The result of this interim multiplication is called the partial
product. The partial product is recorded with its least significant digit aligned with the corresponding
position of the multiplier digit. Finally, all partial products are summed to create the final product of the
multiplication. This process is often called the shift and add approach. Example 12.18 shows the process
for performing long multiplication on decimal numbers highlighting the individual steps.

Example 12.18
Performing long multiplication on decimal numbers
390 • Chapter 12: Arithmetic Circuits

Binary multiplication follows this same process. Example 12.19 shows the process for performing
long multiplication on binary numbers. Note that the inputs represent the largest unsigned numbers
possible using 4-bits, thus producing the largest possible product. The largest product will require 8-bits
to be represented. This means that for any multiplication of n-bit inputs, the product will require 2n bits for
the result.

Example 12.19
Performing long multiplication on binary numbers

The first step in designing a binary multiplier is to create circuitry that can compute the product on
individual bits. Example 12.20 shows the design of a single-bit multiplier.

Example 12.20
Design of a single-bit multiplier

We can create all of the partial products in one level of logic by placing an AND gate between each
bit pairing in the two input numbers. This will require n2 AND gates. The next step involves creating
adders that can perform the sum of the columns of bits within the partial products. This step is not as
straightforward. Notice that in our 4-bit example in Example 12.19 that the number of input bits in the
column addition can reach up to 6 (in position 3). It would be desirable to re-use the full adders previously
 12.3 Multiplication • 391

created; however, the existing full adders could only accommodate 3 inputs (A, B, Cin). We can take
advantage of the associative property of addition to form the final sum incrementally. Example 12.21
shows the architecture of this multiplier. This approach implements a shift and add process to compute
the product and is known as a combinational multiplier because it is implemented using only combina-
tional logic. Note that this multiplier only handles unsigned numbers.

Example 12.21
Design of a 4-bit unsigned multiplier

This multiplier can have a significant delay, which is caused by the cascaded full adders. Example
12.22 shows the timing analysis of the combinational multiplier highlighting the worst case path through
the circuit.
392 • Chapter 12: Arithmetic Circuits

Example 12.22
Timing analysis of a 4-bit unsigned multiplier

12.3.2 A Simple Circuit to Multiply by Powers of Two

In digital systems, a common operation is to multiply numbers by powers of two. For unsigned
numbers, multiplying by two can be accomplished by performing a logical shift left. In this operation, all
bits are moved to the next higher position (i.e., left) by one position and filling the 0th position with a zero.
This has the effect of doubling the value of the number. This can be repeated to achieve higher powers of
two. This process works as long as the resulting product fits within the number of bits available. Example
12.23 shows this procedure.

Example 12.23
Multiplying an unsigned binary number by two using a logical shift left
 12.3 Multiplication • 393

12.3.3 Signed Multiplication

When performing multiplication on signed numbers, it is desirable to re-use the unsigned multiplier
in Example 12.21. Let’s examine if this is possible. Recall in decimal multiplication that the inputs are
multiplied together independent of their sign. The sign of the product is handled separately following
these rules:

• A positive number times a positive number produces a positive number.

• A negative number times a negative number produces a positive number.
• A positive number times a negative number produces a negative number.

This process does not work properly in binary due to the way that negative numbers are represented
with two’s complement. Example 12.24 illustrates how an unsigned multiplier incorrectly handles signed
numbers.

Example 12.24
Illustrating how an unsigned multiplier incorrectly handles signed numbers

Instead of building a dedicated multiplier for signed numbers, we can add functionality to the
unsigned multiplier previously presented to handle negative numbers. The process involves first
identifying any negative numbers. If a negative number is present, the two’s complement is taken on it
to produce its equivalent magnitude, positive representation. The multiplication is then performed on the
positive values. The final step is to apply the correct sign to the product. If the product should be negative
due to one of the inputs being negative, the sign is applied by taking the two’s complement on the final
result. This creates a number that is now in 2n two’s complement format. Example 12.25 shows an
illustration of the process to correctly handle signed numbers using an unsigned multiplier.
394 • Chapter 12: Arithmetic Circuits

Example 12.25
Process to correctly handle signed numbers using an unsigned multiplier

CONCEPT CHECK

CC12.3 Will the AND gates used to compute the partial products in a binary multiplier ever experi-
ence an issue with fan-in as the size of the multiplier increases?
(A) Yes. When the number of bits of the multiplier arguments exceed the fan-in
specification of the AND gates used for the partial products, a fan-in issue has
occurred.
(B) No. The number of inputs of the AND gates performing the partial products will
always be two, regardless of the size of the input arguments to the multiplier.
 12.4 Division • 395

12.4 Division
12.4.1 Unsigned Division
There are a variety of methods to perform division, each with trade-offs between area, delay, and
accuracy. To understand the general approach to building a divider circuit, let’s focus on how a simple
iterative divider can be built. Basic division yields a quotient and a remainder. The process begins by
checking whether the divisor goes into the highest position digit in the dividend. The number of times this
dividend digit can be divided is recorded as the highest position value of the quotient. Note that when
performing division by hand, we typically skip over the condition when the result of these initial operations
are zero, but when breaking down the process into steps that can be built with logic circuits, each step
needs to be highlighted. The first quotient digit is then multiplied with the divisor and recorded below the
original dividend. The next lower position digit of the dividend is brought down and joined with the product
from the prior multiplication. This forms a new number to be divided by the divisor to create the next
quotient value. This process is repeated until each of the quotient digits have been created. Any value
that remains after the last subtraction is recorded as the remainder. Example 12.26 shows the long
division process on decimal numbers highlight each incremental step.

Example 12.26
Performing long division on decimal numbers

Long division in binary follows this same process. Example 12.27 shows the long division process
on two 4-bit, unsigned numbers. This division results in a 4-bit quotient and a 4-bit remainder.
396 • Chapter 12: Arithmetic Circuits

Example 12.27
Performing long multiplication on binary numbers

When building a divider circuit using combinational logic, we can accomplish the computation using
a series of iterative subtractors. Performing division is equivalent to subtracting the divisor from the
interim dividend. If the subtraction is positive, then the divisor went into the dividend and the quotient is a
1. If the subtraction yields a negative number, then the divisor did not go into the interim dividend and the
quotient is 0. We can use the borrow out of a subtraction chain to provide the quotient. This has the
advantage that the difference has already been calculated for the next subtraction. A multiplexer is used
to select whether the difference is used in the next subtraction (Q ¼ 0), or if the interim divisor is simply
brought down (Q ¼ 1). This inherently provides the functionality of the multiplication step in long division.
Example 12.28 shows the architecture of a 4-bit, unsigned divider based on the iterative subtraction
approach. Notice that when the borrow out of the 4-bit subtractor chain is a 0, it indicates that the
subtraction yielded a positive number. This means that the divisor went into the interim dividend once. In
this case, the quotient for this position is a 1. An inverter is required to produce the correct polarity of the
quotient. The borrow out is also fed into the multiplexer stage as the select line to pass the difference to
the next stage of subtractors. If the borrow out of the 4-bit subtractor chain is a 1, it indicates that the
subtraction yielded a negative number. In this case, the quotient is a 0. This also means that the
difference calculated is garbage and should not be used. The multiplexer stage instead selects the
interim dividend as the input to the next stage of subtractors.
 12.4 Division • 397

Example 12.28
Design of a 4-bit unsigned divider using a series of iterative subtractors

To illustrate how this architecture works, Example 12.29 walks through each step in the process
where 11112 (1510) is divided by 01112 (710). In this example, the calculations propagate through the logic
stages from top to bottom in the diagram.
398 • Chapter 12: Arithmetic Circuits

Example 12.29
Dividing 11112 (1510) by 01112 (710) using the iterative subtraction architecture

12.4.2 A Simple Circuit to Divide by Powers of Two

For unsigned numbers, dividing by two can be accomplished by performing a logical shift right. In
this operation, all bits are moved to the next lower position (i.e., right) by one position and then filling the
highest position with a zero. This has the effect of halving the value of the number. This can be repeated
to achieve higher powers of two. This process works until no more ones exist in the number and the result
is simply all zeros. Example 12.30 shows this process.
 12.4 Division • 399

Example 12.30
Dividing an unsigned binary numbers by two using a logical shift right

12.4.3 Signed Division

When performing division on signed numbers, a similar strategy as in signed multiplication is used.
The process involves first identifying any negative numbers. If a negative number is present, the two’s
complement is taken on it to produce its equivalent magnitude, positive representation. The division is
then performed on the positive values. The final step is to apply the correct sign to the divisor and
quotient. This is accomplished by taking the two’s complement if a negative number is required. The
rules governing the polarities of the quotient and remainders are:

• The quotient will be negative if the input signs are different (i.e., pos/neg or neg/pos).
• The remainder has the same sign as the dividend.

CONCEPT CHECK

CC12.4 Could a shift register help reduce the complexity of a combinational divider circuit? How?
(A) Yes. Instead of having redundant circuits holding the different shifted versions of
the divisor, a shift register could be used to hold and shift the divisor after each
subtraction.
(B) No. A state machine would then be needed to control the divisor shifting, which
would make the system even more complex.

Summary
v Binary arithmetic is accomplished using v A ripple carry adder performs addition by
combinational logic circuitry. These circuits reusing lower level components that each
tend to be the largest circuits in a system performs a small part of the computation. A
and have the longest delay. Arithmetic full adder is made from two half adders and a
circuits are often broken up into interim ripple carry adder is made from a chain of full
calculations in order to reduce the overall adders. This approach simplifies the design
delay of the computation. of the adder but leads to long delay times
since the carry from each sum must ripple
400 • Chapter 12: Arithmetic Circuits

to the next higher position’s addition before it accomplished by hand using the shift and
can complete. add approach. The partial products of the
v A carry look ahead adder attempts to elimi- multiplication can be performed using
nate the linear dependence of delay on the 2-input AND gates. The sum of the partial
number of bits that exists in a ripple carry products can have more inputs than the typi-
adder. The carry look ahead adder contains cal ripple carry adder can accommodate. To
dedicated circuitry that calculates the carry handle this, the additions are performed two
bits for each position of the addition. This bits at a time using a series of adders.
leads to a more constant delay as the width v Division can be accomplished using an itera-
of the adder increases. tive subtractor architecture.
v A binary multiplier can be created in a
similar manner to the way multiplication is

Exercise Problems
Section 12.1 – Addition 12.1.5 Design a Verilog model for an 8-bit Ripple
Carry Adder (RCA) using a structural design
12.1.1 Give the total delay of the full adder shown in
approach. This involves creating a half adder
Fig. 12.2 if all gates have a delay of 1 ns.
(half_adder.v), full adder (full_adder.v), and
then finally a top-level adder (rca.v) by
instantiating eight full adder sub-systems.
Model the logic operations using gate level
primitives. Give each gate primitive a delay of
1 ns. The general topology and module defini-
tion for the design are shown in Fig. 12.4. Cre-
ate a test bench to exhaustively verify your
design under all input conditions. The test
bench should drive in different values every
Fig. 12.2 30 ns in order to give sufficient time for the
Full Adder Timing Exercise results to ripple through the adder.

12.1.2 Give the total delay of the full adder shown in

Fig. 12.2 if the XOR gates have delays of 5 ns
while the AND and OR gates have delays of
1 ns.
12.1.3 Give the total delay of the 4-bit ripple carry
adder shown in Fig. 12.3 if all gates have a
delay of 2 ns.

Fig. 12.4
4-Bit RCA Module Definition

12.1.6 Give the total delay of the 4-bit carry look

ahead adder shown in Fig. 12.5 if all gates
have a delay of 2 ns.

Fig. 12.3
4-Bit RCA Timing Exercise

12.1.4 Give the total delay of the 4-bit ripple carry

adder shown in Fig. 12.3 if the XOR gates
have delays of 10 ns while the AND and OR
12.1.7 Give the total delay of the 4-bit carry look
gates have delays of 2 ns.
ahead adder shown in Fig. 12.5 if the XOR
gates have delays of 10 ns while the AND
and OR gates have delays of 2 ns.

Fig. 12.5
4-Bit CLA Timing Exercise
 Exercise Problems • 401

12.1.8 Design a Verilog model for an 8-bit Carry Look port called “Sum_Diff”. The model should
Ahead Adder (cla.v). The model should instan- assert the output “Cout_Bout” when an addi-
tiate eight instances of a modified full adder tion creates a carry or when a subtraction
(mod_full_adder.v), which is implemented creates a borrow. The circuit will also assert
with gate-level primitives. The carry look the output Vout when either operation results in
ahead logic should be implemented using con- two’s complement overflow. The port definition
tinuous assignment with logical operators and block diagram for the system is shown in
within the cla.v module. All logic operations Fig. 12.8. Create a test bench to exhaustively
should have 1 ns of delay. The topology and verify this design under all input conditions.
port definition for the design are shown in
Fig. 12.6. Create a test bench to exhaustively
verify this design under all input conditions.
The test bench should drive in different values
every 30 ns in order to give sufficient time for
the signals to propagate through the adder.

Fig. 12.8
Programmable Adder/Subtractor Module
Definition

Section 12.3 – Multiplication

Fig. 12.6 12.3.1 Give the total delay of the 4-bit unsigned multi-
4-Bit CLA Module Definition plier shown in Fig. 12.9 if all gates have a delay
of 1 ns. The addition is performed using a ripple
Section 12.2 – Subtraction carry adder.
12.2.1 How is the programmable add/subtract topol-
ogy shown in Fig. 12.7 analogous to 2’s com-
plement arithmetic?

Fig. 12.7
Programmable Adder/Subtractor Block Diagram Fig. 12.9
4-Bit Unsigned Multiplier Block Diagram
12.2.2 Will the programmable adder/subtractor archi-
tecture shown in Fig. 12.7 work for negative
12.3.2 For the 4-bit unsigned multiplier shown in
numbers encoded using signed magnitude or
Fig. 12.9, how many levels of logic does it
1’s complement?
take to compute all of the partial products?
12.2.3 When calculating the delay of the programma-
12.3.3 For the 4-bit unsigned multiplier shown in
ble adder/subtractor architecture shown in
Fig. 12.9, how many AND gates are needed
Fig. 12.7 does the delay of the XOR gate that
to compute the partial products?
acts as the programmable inverter need to be
considered? 12.3.4 For the 4-bit unsigned multiplier shown in
Fig. 12.9, how many total AND gates are
12.2.4 Design a Verilog model for an 8-bit, program-
used if the additions are implemented using
mable adder/subtractor. The design will have
full adders made of half adders?
an input called “ADDn_SUB” that will control
whether the system behaves as an adder (0) or 12.3.5 Based on the architecture of a unsigned multi-
as a subtractor (1). The design should operate plier in Fig. 12.9, how many AND gates are
on two’s complement signed numbers. The needed to compute the partial products if the
result of the operation(s) will appear on the inputs are increased to 8-bits?
402 • Chapter 12: Arithmetic Circuits

12.3.6 For an 8-bit multiplier, how many bits are 12.3.12 Design a Verilog model for an 8-bit signed
needed to represent the product? multiplier using whatever modeling approach
12.3.7 For an 8-bit unsigned multiplier, what is the you wish. Create a test bench to exhaustively
largest value that the product can ever take verify this design under all input conditions.
on? Give your answer in decimal. The port definition for this multiplier is given in
Fig. 12.11.
12.3.8 For an 8-bit signed multiplier, what is the larg-
est value that the product can ever take on?
Give your answer in decimal.
12.3.9 For an 8-bit signed multiplier, what is the
smallest value that the product can ever take
on? Give your answer in decimal. Fig. 12.11
12.3.10 What is the maximum number of times that a 8-Bit Signed Multiplier Module Definition
4-bit unsigned multiplicand can be multiplied
by two using the logical shift left approach
before the product is too large to be Section 12.4 – Division
represented by an 8-bit-product? Hint: The
12.4.1 For a 4-bit divider, how many bits are needed
maximum number of times this operation can
for the quotient?
be performed corresponds to when the multi-
plicand starts at its lowest possible non-zero 12.4.2 For a 4-bit divider, how many bits are needed
value (i.e., 1). for the remainder?
12.3.11 Design a Verilog model for an 8-bit unsigned 12.4.2 Explain the basic concept of the iterative-
multiplier using whatever modeling approach subtractor approach to division.
you wish. Create a test bench to exhaustively 12.4.4 For the 4-bit divider shown in Example 12.28,
verify this design under all input conditions. estimate the total delay assuming all gates
The port definition for this multiplier is given in have a delay of 1 ns.
Fig. 12.10.

Fig. 12.10
8-Bit Unsigned Multiplier Module Definition
Chapter 13: Computer System Design
One of the most common digital systems in use today is the computer. A computer accomplishes
tasks through an architecture that uses both hardware and software. The hardware in a computer
consists of many of the elements that we have covered so far. These include registers, arithmetic and
logic circuits, finite state machines, and memory. What makes a computer so useful is that the hardware
is designed to accomplish a predetermined set of instructions. These instructions are relatively simple,
such as moving data between memory and a register or performing arithmetic on two numbers. The
instructions are comprised of binary codes that are stored in a memory device and represent the
sequence of operations that the hardware will perform to accomplish a task. This sequence of
instructions is called a computer program. What makes this architecture so useful is that the preexisting
hardware can be programmed to perform an almost unlimited number of tasks by simply defining the
sequence of instructions to be executed. The process of designing the sequence of instructions, or
program, is called software development or software engineering.
The idea of a general purpose computing machine dates back to the 19th century. The first
computing machines were implemented with mechanical systems and were typically analog in nature.
As technology advanced, computer hardware evolved from electromechanical switches to vacuum
tubes and ultimately to integrated circuits. These newer technologies enabled switching circuits and
provided the capability to build binary computers. Today’s computers are built exclusively with semicon-
ductor materials and integrated circuit technology. The term microcomputer is used to describe a
computer that has its processing hardware implemented with integrated circuitry. Nearly all modern
computers are binary. Binary computers are designed to operate on a fixed set of bits. For example, an
8-bit computer would perform operations on 8-bits at a time. This means it moves data between registers
and memory and performs arithmetic and logic operations in groups of 8-bits.

This chapter will cover the basics of a simple computer system and present the design of an 8-bit
system to illustrate the details of instruction execution. The goal of this chapter is to provide an
understanding of the basic principles of computer systems.

Learning Outcomes—After completing this chapter, you will be able to:

13.1 Describe the basic components and operation of computer hardware.
13.2 Describe the basic components and operation of computer software.
13.3 Design a fully operational computer system using Verilog.
13.4 Describe the difference between the Von Neumann and Harvard computer architectures.

13.1 Computer Hardware

Computer hardware refers to all of the physical components within the system. This hardware
includes all circuit components in a computer such as the memory devices, registers, and finite state
machines. Figure 13.1 shows a block diagram of the basic hardware components in a computer.

# Springer International Publishing AG 2017 403

B.J. LaMeres, Introduction to Logic Circuits & Logic Design with Verilog,
DOI 10.1007/978-3-319-53883-9_13
404 • Chapter 13: Computer System Design

Fig. 13.1
Hardware components of a computer system

13.1.1 Program Memory

The instructions that are executed by a computer are held in program memory. Program memory is
treated as read only memory during execution in order to prevent the instructions from being overwritten
by the computer. Some computer systems will implement the program memory on a true ROM device
(MROM or PROM), while others will use a EEPROM that can be read from during normal operation but
can only be written to using a dedicated write procedure. Programs are typically held in non-volatile
memory so that the computer system does not lose its program when power is removed. Modern
computers will often copy a program from non-volatile memory (e.g., a hard disk drive) to volatile memory
after startup in order to speed up instruction execution. In this case, care must be taken that the program
does not overwrite itself.

13.1.2 Data Memory

Computers also require data memory, which can be written to and read from during normal
operation. This memory is used to hold temporary variables that are created by the software program.
This memory expands the capability of the computer system by allowing large amounts of information to
be created and stored by the program. Additionally, computations can be performed that are larger than
the width of the computer system by holding interim portions of the calculation (e.g., performing a 128-bit
addition on a 32-bit computer). Data memory is implemented with R/W memory, most often SRAM or
DRAM.

13.1.3 Input/Output Ports

The term port is used to describe the mechanism to get information from the output world into or out
of the computer. Ports can be input, output, or bidirectional. I/O ports can be designed to pass information
in a serial or parallel format.
 13.1 Computer Hardware • 405

13.1.4 Central Processing Unit

The central processing unit (CPU) is considered the brains of the computer. The CPU handles
reading instructions from memory, decoding them to understand which instruction is being performed,
and executing the necessary steps to complete the instruction. The CPU also contains a set of registers
that are used for general purpose data storage, operational information, and system status. Finally, the
CPU contains circuitry to perform arithmetic and logic operations on data.

13.1.4.1 Control Unit

The control unit is a finite state machine that controls the operation of the computer. This FSM has
states that perform fetching the instruction (i.e., reading it from program memory), decoding the instruc-
tion, and executing the appropriate steps to accomplish the instruction. This process is known as fetch,
decode, and execute and is repeated each time an instruction is performed by the CPU. As the control
unit state machine traverses through its states, it asserts control signals that move and manipulate data
in order to achieve the desired functionality of the instruction.

13.1.4.2 Data Path – Registers

The CPU groups its registers and ALU into a sub-system called the data path. The data path refers
to the fast storage and data manipulations within the CPU. All of these operations are initiated and
managed by the control unit state machine. The CPU contains a variety of registers that are necessary to
execute instructions and hold status information about the system. Basic computers have the following
registers in their CPU:

• Instruction Register (IR) – The instruction register holds the current binary code of the
instruction being executed. This code is read from program memory as the first part of
instruction execution. The IR is used by the control unit to decide which states in its FSM to
traverse in order to execute the instruction.
• Memory Address Register (MAR) – The memory address register is used to hold the current
address being used to access memory. The MAR can be loaded with addresses in order to
fetch instructions from program memory or with addresses to access data memory and/or I/O
ports.
• Program Counter (PC) – The program counter holds the address of the current instruction
being executed in program memory. The program counter will increment sequentially through
the program memory reading instructions until a dedicated instruction is used to set it to a new
location.
• General Purpose Registers – These registers are available for temporary storage by the
program. Instructions exist to move information from memory into these registers and to move
information from these registers into memory. Instructions also exist to perform arithmetic and
logic operations on the information held in these registers.
• Condition Code Register (CCR) – The condition code register holds status flags that provide
information about the arithmetic and logic operations performed in the CPU. The most common
flags are negative (N), zero (Z), two’s complement overflow (V), and carry (C). This register can
also contain flags that indicate the status of the computer, such as if an interrupt has occurred
or if the computer has been put into a low-power mode.

13.1.4.3 Data Path – Arithmetic Logic Unit (ALU)

The arithmetic logic unit is the system that performs all mathematical (i.e., addition, subtraction,
multiplication, and division) and logic operations (i.e., and, or, not, shifts, etc.). This system operates on
data being held in CPU registers. The ALU has a unique symbol associated with it to distinguish it from
other functional units in the CPU.
406 • Chapter 13: Computer System Design

Figure 13.2 shows the typical organization of a CPU. The registers and ALU are grouped into the
data path. In this example, the computer system has two general purpose registers called A and B. This
CPU organization will be used throughout this chapter to illustrate the detailed execution of instructions.

Fig. 13.2
Typical CPU organization

13.1.5 A Memory Mapped System

A common way to simplify moving data in or out of the CPU is to assign a unique address to all
hardware components in the memory system. Each input/output port and each location in both program
and data memory are assigned a unique address. This allows the CPU to access everything in the
memory system with a dedicated address. This reduces the number of lines that must pass into the CPU.
A bus system facilitates transferring information within the computer system. An address bus is driven by
the CPU to identify which location in the memory system is being accessed. A data bus is used to
transfer information to/from the CPU and the memory system. Finally, a control bus is used to provide
other required information about the transactions such as read or write lines. Figure 13.3 shows the
computer hardware in a memory mapped architecture.
 13.1 Computer Hardware • 407

Fig. 13.3
Computer hardware in a memory mapped configuration

To help visualize how the memory addresses are assigned, a memory map is used. This is a
graphical depiction of the memory system. In the memory map, the ranges of addresses are provided for
each of the main subsections of memory. This gives the programmer a quick overview of the available
resources in the computer system. Example 13.1 shows a representative memory map for a computer
system with an address bus with a width of 8-bits. This address bus can provide 256 unique locations.
For this example, the memory system is also 8-bits wide, thus the entire memory system is 256
8 in
size. In this example 128 bytes are allocated for program memory; 96 bytes are allocated for data
memory; 16 bytes are allocated for output ports; and 16 bytes are allocated for input ports.
408 • Chapter 13: Computer System Design

CONCEPT CHECK

CC13.1 Is the hardware of a computer programmed in a similar way to a programmable logic

device?
(A) Yes. The control unit is reconfigured to produce the correct logic for each
unique instruction just like a logic element in an FPGA is reconfigured to
produce the desired logic expression.
(B) No. The instruction code from program memory simply tells the state
machine in the control unit which path to traverse in order to accomplish the
desired task.

Example 13.1
Memory map for a 256
8 memory system

13.2 Computer Software

Computer software refers to the instructions that the computer can execute and how they are
designed to accomplish various tasks. The specific group of instructions that a computer can execute
is known as its instruction set. The instruction set of a computer needs to be defined first before the
computer hardware can be implemented. Some computer systems have a very small number of
instructions in order to reduce the physical size of the circuitry needed in the CPU. This allows the
CPU to execute the instructions very quickly, but requires a large number of operations to accomplish a
 13.2 Computer Software • 409

given task. This architectural approach is called a reduced instruction set computer (RISC). The
alternative to this approach is to make an instruction set with a large number of dedicated instructions
that can accomplish a given task in fewer CPU operations. The drawback of this approach is that the
physical size of the CPU must be larger in order to accommodate the various instructions. This
architectural approach is called a complex instruction set computer (CISC).

13.2.1 Opcodes and Operands

A computer instruction consists of two fields, an opcode and an operand. The opcode is a unique
binary code given to each instruction in the set. The CPU decodes the opcode in order to know which
instruction is being executed and then takes the appropriate steps to complete the instruction. Each
opcode is assigned a mnemonic, which is a descriptive name for the opcode that can be used when
discussing the instruction functionally. An operand is additional information for the instruction that may be
required. An instruction may have any number of operands including zero. Figure 13.4 shows an
example of how the instruction opcodes and operands are placed into program memory.

Fig. 13.4
Anatomy of a computer instruction

13.2.2 Addressing Modes

An addressing mode describes the way in which the operand of an instruction is used. While modern
computer systems may contain numerous addressing modes with varying complexities, we will focus on
just a subset of basic addressing modes. These modes are immediate, direct, inherent, and indexed.

13.2.2.1 Immediate Addressing (IMM)

Immediate addressing is when the operand of an instruction is the information to be used by the
instruction. For example, if an instruction existed to put a constant into a register within the CPU using
410 • Chapter 13: Computer System Design

immediate addressing, the operand would be the constant. When the CPU reads the operand, it simply
inserts the contents into the CPU register and the instruction is complete.

13.2.2.2 Direct Addressing (DIR)

Direct addressing is when the operand of an instruction contains the address of where the informa-
tion to be used is located. For example, if an instruction existed to put a constant into a register within the
CPU using direct addressing, the operand would contain the address of where the constant was located
in memory. When the CPU reads the operand, it puts this value out on the address bus and performs an
additional read to retrieve the contents located at that address. The value read is then put into the CPU
register and the instruction is complete.

13.2.2.3 Inherent Addressing (INH)

Inherent addressing refers to an instruction that does not require an operand because the opcode
itself contains all of the necessary information for the instruction to complete. This type of addressing is
used on instructions that perform manipulations on data held in CPU registers without the need to access
the memory system. For example, if an instruction existed to increment the contents of a register (A),
then once the opcode is read by the CPU, it knows everything it needs to know in order to accomplish the
task. The CPU simply asserts a series of control signals in order to increment the contents of A and then
the instruction is complete. Notice that no operand is needed for this task. Instead, the location of the
register to be manipulated (i.e., A) is inherent within the opcode.

13.2.2.4 Indexed Addressing (IND)

Indexed addressing refers to instructions that will access information at an address in memory to
complete the instruction, but the address to be accessed is held in another CPU register. In this type of
addressing, the operand of the instruction is used as an offset that can be applied to the address located
in the CPU register. For example, let’s say an instruction existed to put a constant into a register
(A) within the CPU using indexed addressing. Let’s also say that the instruction was designed to use
the contents of another register (B) as part of the address of where the constant was located. When the
CPU reads the opcode, it understands what the instruction is and that B holds part of the address to be
accessed. It also knows that the operand is applied to B to form the actual address to be accessed. When
the CPU reads the operand, it adds the value to the contents of B and then puts this new value out on the
address bus and performs an additional read. The value read is then put into the CPU register A and the
instruction is complete.

13.2.3 Classes of Instructions

There are three general classes of instructions: (1) loads and stores; (2) data manipulations; and
(3) branches. To illustrate how these instructions are executed, examples will be given based on the
computer architecture shown in Fig. 13.3.

13.2.3.1 Loads and Stores

This class of instructions accomplishes moving information between the CPU and memory. A load
is an instruction that moves information from memory into a CPU register. When a load instruction uses
immediate addressing, the operand of the instruction is the data to be loaded into the CPU register. As an
example, let’s look at an instruction to load the general purpose register A using immediate addressing.
Let’s say that the opcode of the instruction is x00 8600 , has a mnemonic LDA_IMM, and is inserted into
program memory starting at x00 0000 . Example 13.2 shows the steps involved in executing the LDA_IMM
instruction.
 13.2 Computer Software • 411

Example 13.2
Execution of an instruction to “Load Register A Using Immediate Addressing”

Now let’s look at a load instruction using direct addressing. In direct addressing, the operand of the
instruction is the address of where the data to be loaded resides. As an example, let’s look at an
instruction to load the general purpose register A. Let’s say that the opcode of the instruction is x00 8700 ,
has a mnemonic LDA_DIR, and is inserted into program memory starting at x00 0800 . The value to be
loaded into A resides at address x00 8000 , which has already been initialized with x00 AA00 before this
instruction. Example 13.3 shows the steps involved in executing the LDA_DIR instruction.
412 • Chapter 13: Computer System Design

Example 13.3
Execution of an instruction to “Load Register A Using Direct Addressing”

A store is an instruction that moves information from a CPU register into memory. The operand of a
store instruction indicates the address of where the contents of the CPU register will be written. As an
example, let’s look at an instruction to store the general purpose register A into memory address x00 E000 .
Let’s say that the opcode of the instruction is x00 9600 , has a mnemonic STA_DIR, and is inserted into
program memory starting at x00 0400 . The initial value of A is x00 CC00 before the instruction is executed.
Example 13.4 shows the steps involved in executing the STA_DIR instruction.
 13.2 Computer Software • 413

Example 13.4
Execution of an instruction to “Store Register A Using Direct Addressing”

13.2.3.2 Data Manipulations

This class of instructions refers to ALU operations. These operations take action on data that
resides in the CPU registers. These instructions include arithmetic, logic operators, shifts and rotates,
and tests and compares. Data manipulation instructions typically use inherent addressing because the
operations are conducted on the contents of CPU registers and don’t require additional memory access.
As an example, let’s look at an instruction to perform addition on registers A and B. The sum will be
placed back in A. Let’s say that the opcode of the instruction is x00 4200 , has a mnemonic ADD_AB, and is
inserted into program memory starting at x00 0400 . Example 13.5 shows the steps involved in executing the
ADD_AB instruction.
414 • Chapter 13: Computer System Design

Example 13.5
Execution of an instruction to “Add Registers A and B”

13.2.3.3 Branches
In the previous examples the program counter was always incremented to point to the address of
the next instruction in program memory. This behavior only supports a linear execution of instructions. To
provide the ability to specifically set the value of the program counter, instructions called branches are
used. There are two types of branches: unconditional and conditional. In an unconditional branch, the
program counter is always loaded with the value provided in the operand. As an example, let’s look at an
instruction to branch always to a specific address. This allows the program to perform loops. Let’s say
that the opcode of the instruction is x00 2000 , has a mnemonic BRA, and is inserted into program memory
starting at x00 0600 . Example 13.6 shows the steps involved in executing the BRA instruction.
 13.2 Computer Software • 415

Example 13.6
Execution of an instruction to “Branch Always”

In a conditional branch, the program counter is only updated if a particular condition is true. The
conditions come from the status flags in the condition code register (NZVC). This allows a program to
selectively execute instructions based on the result of a prior operation. Let’s look at an example
instruction that will branch only if the Z flag is asserted. This instruction is called a branch if equal to
zero. Let’s say that the opcode of the instruction is x00 2300 , has a mnemonic BEQ, and is inserted into
program memory starting at x00 0500 . Example 13.7 shows the steps involved in executing the BEQ
instruction.
416 • Chapter 13: Computer System Design

Example 13.7
Execution of an instruction to “Branch if Equal to Zero”

Conditional branches allow computer programs to make decisions about which instructions to
execute based on the results of previous instructions. This gives computers the ability to react to input
signals or take action based on the results of arithmetic or logic operations. Computer instruction sets
typically contain conditional branches based on the NZVC flags in the condition code registers. The
following instructions are based on the values of the NZVC flags.

• BMI – Branch if minus (N ¼ 1)

• BPL – Branch if plus (N ¼ 0)
• BEQ – Branch if equal to Zero (Z ¼ 1)
 13.3 Computer Implementation – An 8-Bit Computer Example • 417

• BNE – Branch if not equal to Zero (Z ¼ 0)

• BVS – Branch if two’s complement overflow occurred, or V is set (V ¼ 1)
• BVC – Branch if two’s complement overflow did not occur, or V is clear (V ¼ 0)
• BCS – Branch if a carry occurred, or C is set (C ¼ 1)
• BCC – Branch if a carry did not occur, or C is clear (C ¼ 0)

Combinations of these flags can be used to create more conditional branches.

• BHI – Branch if higher (C ¼ 1 and Z ¼ 0)

• BLS – Branch if lower or the same (C ¼ 0 and Z ¼ 1)
• BGE – Branch if greater than or equal ((N ¼ 0 and V ¼ 0) or (N ¼ 1 and V ¼ 1)), only valid for
signed numbers
• BLT – Branch if less than ((N ¼ 1 and V ¼ 0) or (N ¼ 0 and V ¼ 1)), only valid for signed
numbers
• BGT – Branch if greater than ((N ¼ 0 and V ¼ 0 and Z ¼ 0) or (N ¼ 1 and V ¼ 1 and Z ¼ 0)), only
valid for signed numbers
• BLE – Branch if less than or equal ((N ¼ 1 and V ¼ 0) or (N ¼ 0 and V ¼ 1) or (Z ¼ 1)), only valid
for signed numbers

CONCEPT CHECK

CC13.2 Software development consists of choosing which instructions, and in what order, will
be executed to accomplish a certain task. The group of instructions is called the
program and is inserted into program memory. Which of the following might a software
developer care about?
(A) Minimizing the number of instructions that need to be executed to accom-
plish the task in order to increase the computation rate.
(B) Minimizing the number of registers used in the CPU to save power.
(C) Minimizing the overall size of the program to reduce the amount of program
memory needed.
(D) Both A and C.

13.3 Computer Implementation – An 8-Bit Computer Example

13.3.1 Top Level Block Diagram
Let’s now look at the detailed implementation and instruction execution of a computer system.
In order to illustrate the detailed operation, we will use a simple 8-bit computer system design.
Example 13.8 shows the block diagram for the 8-bit computer system. This block diagram also contains
the Verilog file and module names, which will be used when the behavioral model is implemented.
418 • Chapter 13: Computer System Design

Example 13.8
Top level block diagram for the 8-Bit computer system

We will use the memory map shown in Example 13.1 for our example computer system. This
mapping provides 128 bytes of program memory, 96 bytes of data memory, 16x output ports, and 16x
input ports. To simplify the operation of this example computer, the address bus is limited to 8-bits. This
only provides 256 locations of memory access, but allows an entire address to be loaded into the CPU as
a single operand of an instruction.

13.3.2 Instruction Set Design

Example 13.9 shows a basic instruction set for our example computer system. This set provides a
variety of loads and stores, data manipulations, and branch instructions that will allow the computer to be
programmed to perform more complex tasks through software development. These instructions are
sufficient to provide a baseline of functionality in order to get the computer system operational. Additional
instructions can be added as desired to increase the complexity of the system.
 13.3 Computer Implementation – An 8-Bit Computer Example • 419

Example 13.9
Instruction set for the 8-Bit computer system

13.3.3 Memory System Implementation

Let’s now look at the memory system details. The memory system contains program memory, data
memory, and input/output ports. Example 13.10 shows the block diagram of the memory system. The
program and data memory will be implemented using lower level components (rom_128x8_sync.v and
rw_96x8_sync.v), while the input and output ports can be modeled using a combination of RTL blocks
and combinational logic. The program and data memory sub-systems contain dedicated circuitry to
handle their addressing ranges. Each output port also contains dedicated circuitry to handle its unique
address. A multiplexer is used to handle the signal routing back to the CPU based on the address
provided.
420 • Chapter 13: Computer System Design

Example 13.10
Memory system block diagram for the 8-Bit computer system

13.3.3.1 Program Memory Implementation in Verilog

The program memory can be implemented in Verilog using the modeling techniques presented in
Chapter 12. To make the Verilog more readable, the instruction mnemonics can be declared as
parameters. This allows the mnemonic to be used when populating the program memory array.
The following Verilog shows how the mnemonics for our basic instruction set can be defined as
parameters.
 13.3 Computer Implementation – An 8-Bit Computer Example • 421

parameter LDA_IMM ¼ 8’h86; //-- Load Register A with Immediate Addressing

parameter LDA_DIR ¼ 8’h87; //-- Load Register A with Direct Addressing
parameter LDB_IMM ¼ 8’h88; //-- Load Register B with Immediate Addressing
parameter LDB_DIR ¼ 8’h89; //-- Load Register B with Direct Addressing
parameter STA_DIR ¼ 8’h96; //-- Store Register A to memory (RAM or IO)
parameter STB_DIR ¼ 8’h97; //-- Store Register B to memory (RAM or IO)
parameter ADD_AB ¼ 8’h42; //-- A <¼ A + B
parameter SUB_AB ¼ 8’h43; //-- A <¼ A - B
parameter AND_AB ¼ 8’h44; //-- A <¼ A and B
parameter OR_AB ¼ 8’h45; //-- A <¼ A or B
parameter INCA ¼ 8’h46; //-- A <¼ A + 1
parameter INCB ¼ 8’h47; //-- B <¼ B + 1
parameter DECA ¼ 8’h48; //-- A <¼ A - 1
parameter DECB ¼ 8’h49; //-- B <¼ B - 1
parameter BRA ¼ 8’h20; //-- Branch Always
parameter BMI ¼ 8’h21; //-- Branch if N¼1
parameter BPL ¼ 8’h22; //-- Branch if N¼0
parameter BEQ ¼ 8’h23; //-- Branch if Z¼1
parameter BNE ¼ 8’h24; //-- Branch if Z¼0
parameter BVS ¼ 8’h25; //-- Branch if V¼1
parameter BVC ¼ 8’h26; //-- Branch if V¼0
parameter BCS ¼ 8’h27; //-- Branch if C¼1
parameter BCC ¼ 8’h28; //-- Branch if C¼0

Now the program memory can be declared as an array type with initial values to define the program.
The following Verilog shows how to declare the program memory and an example program to perform a
load, store, and a branch always. This program will continually write x00 AA” to port_out_00.
reg[7:0] ROM[0:127];

initial
begin
ROM[0] ¼ LDA_IMM;
ROM[1] ¼ 8’hAA;
ROM[2] ¼ STA_DIR;
ROM[3] ¼ 8’hE0;
ROM[4] ¼ BRA;
ROM[5] ¼ 8’h00;
end

The address mapping for the program memory is handled in two ways. First, notice that the array
type defined above uses indices from 0 to 127. This provides the appropriate addresses for each location
in the memory. The second step is to create an internal enable line that will only allow assignments from
ROM to data_out when a valid address is entered. Consider the following Verilog to create an internal
enable (EN) that will only be asserted when the address falls within the valid program memory range of
0 to 127.
always @ (address)
begin
if ( (address >¼ 0) && (address <¼ 127) )
EN ¼ 1’b1;
else
EN ¼ 1’b0;
end

If this enable signal is not created, the simulation and synthesis will fail because data_out
assignments will be attempted for addresses outside of the defined range of the ROM array. This enable
line can now be used in the behavioral model for the ROM as follows:
always @ (posedge clock)
begin
if (EN)
data_out ¼ ROM[address];
end
422 • Chapter 13: Computer System Design

13.3.3.2 Data Memory Implementation in Verilog

The data memory is created using a similar strategy as the program memory. An array signal is
declared with an address range corresponding to the memory map for the computer system (i.e., 128 to
223). An internal enable is again created that will prevent data_out assignments for addresses outside
of this valid range. The following is the Verilog to declare the R/W memory array:

reg[7:0] RW[128:223];

The following is the Verilog to model the local enable and signal assignments for the R/W memory:
always @ (address)
begin
if ( (address >¼ 128) && (address <¼ 223) )
EN ¼ 1’b1;
else
EN ¼ 1’b0;
end

always @ (posedge clock)

begin
if (write && EN)
RW[address] ¼ data_in;
else if (!write && EN)
data_out ¼ RW[address];
end

13.3.3.3 Implementation of Output Ports in Verilog

Each output port in the computer system is assigned a unique address. Each output port also
contains storage capability. This allows the CPU to update an output port by writing to its specific
address. Once the CPU is done storing to the output port address and moves to the next instruction in
the program, the output port holds its information until it is written to again. This behavior can be modeled
using an RTL procedural block that uses the address bus and the write signal to create a synchronous
enable condition. Each output port is modeled with its own block. The following Verilog shows how the
output ports at x00 E000 and x00 E100 are modeled using address specific procedural blocks.
//-- port_out_00 (address E0)
always @ (posedge clock or negedge reset)
begin
if (!reset)
port_out_00 <¼ 8’h00;
else
if ((address ¼¼ 8’hE0) && (write))
port_out_00 <¼ data_in;
end

//-- port_out_01 (address E1)

always @ (posedge clock or negedge reset)
begin
if (!reset)
port_out_01 <¼ 8’h00;
else
if ((address ¼¼ 8’hE1) && (write))
port_out_01 <¼ data_in;
end

:
“the rest of the output port models go here. . .”
:
 13.3 Computer Implementation – An 8-Bit Computer Example • 423

13.3.3.4 Implementation of Input Ports in Verilog

The input ports do not contain storage, but do require a mechanism to selectively route their
information to the data_out port of the memory system. This is accomplished using the multiplexer
shown in Example 13.10. The only functionality that is required for the input ports is connecting their
ports to the multiplexer.

13.3.3.5 Memory data_out Bus Implementation in Verilog

Now that all of the memory functionality has been designed, the final step is to implement the
multiplexer that handles routing the appropriate information to the CPU on the data_out bus based on the
incoming address. The following Verilog provides a model for this behavior. Recall that a multiplexer is
combinational logic, so if the behavior is to be modeled using a procedural block, all inputs must be listed
in the sensitivity list and blocking assignments are used. These inputs include the outputs from the
program and data memory in addition to all of the input ports. The sensitivity list must also include the
address bus as it acts as the select input to the multiplexer. Within the block, an if-else statement is used
to determine which sub-system drives data_out. Program memory will drive data_out when the incoming
address is in the range of 0 to 127 (x00 0000 to x00 7F00 ). Data memory will drive data_out when the address is
in the range of 128 to 223 (x00 8000 to x00 DF00 ). An input port will drive data_out when the address is in the
range of 240 to 255 (x00 F000 to x00 FF00 ). Each input port has a unique address so the specific addresses are
listed as nested if-else clauses.

always @ (address, rom_data_out, rw_data_out,

port_in_00, port_in_01, port_in_02, port_in_03,
port_in_04, port_in_05, port_in_06, port_in_07,
port_in_08, port_in_09, port_in_10, port_in_11,
port_in_12, port_in_13, port_in_14, port_in_15)

begin: MUX1

if ( (address >¼ 0) && (address <¼ 127) )

data_out ¼ rom_data_out;
else if ( (address >¼ 128) && (address <¼ 223) )
data_out ¼ rw_data_out;
else if (address ¼¼ 8’hF0) data_out ¼ port_in_00;
else if (address ¼¼ 8’hF1) data_out ¼ port_in_01;
else if (address ¼¼ 8’hF2) data_out ¼ port_in_02;
else if (address ¼¼ 8’hF3) data_out ¼ port_in_03;
else if (address ¼¼ 8’hF4) data_out ¼ port_in_04;
else if (address ¼¼ 8’hF5) data_out ¼ port_in_05;
else if (address ¼¼ 8’hF6) data_out ¼ port_in_06;
else if (address ¼¼ 8’hF7) data_out ¼ port_in_07;
else if (address ¼¼ 8’hF8) data_out ¼ port_in_08;
else if (address ¼¼ 8’hF9) data_out ¼ port_in_09;
else if (address ¼¼ 8’hFA) data_out ¼ port_in_10;
else if (address ¼¼ 8’hFB) data_out ¼ port_in_11;
else if (address ¼¼ 8’hFC) data_out ¼ port_in_12;
else if (address ¼¼ 8’hFD) data_out ¼ port_in_13;
else if (address ¼¼ 8’hFE) data_out ¼ port_in_14;
else if (address ¼¼ 8’hFF) data_out ¼ port_in_15;

end

13.3.4 CPU Implementation

Let’s now look at the central processing unit details. The CPU contains two components, the control
unit (control_unit.v) and the data path (data_path.v). The data path contains all of the registers and the
ALU. The ALU is implemented as a sub-system within the data path (alu.v). The data path also contains
424 • Chapter 13: Computer System Design

a bus system in order to facilitate data movement between the registers and memory. The bus system is
implemented with two multiplexers that are controlled by the control unit. The control unit contains the
finite state machine that generates all control signals for the data path as it performs the fetch-decode-
execute steps of each instruction. Example 13.11 shows the block diagram of the CPU in our 8-bit
microcomputer example.

Example 13.11
CPU block diagram for the 8-Bit computer system

13.3.4.1 Data Path Implementation in Verilog

Let’s first look at the data path bus system that handles internal signal routing. The system consists
of two 8-bit busses (Bus1 and Bus2) and two multiplexers. Bus1 is used as the destination of the PC, A,
and B register outputs, while Bus2 is used as the input to the IR, MAR, PC, A, and B registers. Bus1 is
 13.3 Computer Implementation – An 8-Bit Computer Example • 425

connected directly to the to_memory port of the CPU to allow registers to write data to the memory
system. Bus2 can be driven by the from_memory port of the CPU to allow the memory system to provide
data for the CPU registers. The two multiplexers handle all signal routing and have their select lines
(Bus1_Sel and Bus2_Sel) driven by the control unit. The following Verilog shows how the multiplexers
are implemented. Again, a multiplexer is combinational logic so all inputs must be listed in the sensitivity
list of its procedural block and blocking assignments are used. Two additional signal assignments are
also required to connect the MAR to the address port and to connect Bus1 to the to_memory port.
always @ (Bus1_Sel, PC, A, B)
begin: MUX_BUS1
case (Bus1_Sel)
2’b00 : Bus1 ¼ PC;
2’b01 : Bus1 ¼ A;
2’b10 : Bus1 ¼ B;
default : Bus1 ¼ 8’hXX;
endcase
end

always @ (Bus2_Sel, ALU_Result, Bus1, from_memory)

begin: MUX_BUS2
case (Bus2_Sel)
2’b00 : Bus2 ¼ ALU_Result;
2’b01 : Bus2 ¼ Bus1;
2’b10 : Bus2 ¼ from_memory;
default : Bus1 ¼ 8’hXX;
endcase
end

always @ (Bus1, MAR)

begin
to_memory ¼ Bus1;
address ¼ MAR;
end

Next, let’s look at implementing the registers in the data path. Each register is implemented using a
dedicated procedural block that is sensitive to clock and reset. This models the behavior of synchronous
latches, or registers. Each register has a synchronous enable line that dictates when the register is
updated. The register output is only updated when the enable line is asserted and a rising edge of the
clock is detected. The following Verilog shows how to model the instruction register (IR). Notice that the
signal IR is only updated if IR_Load is asserted and there is a rising edge of the clock. In this case, IR is
loaded with the value that resides on Bus2.
always @ (posedge clock or negedge reset)
begin: INSTRUCTION_REGISTER
if (!reset)
IR <¼ 8’h00;
else
if (IR_Load)
IR <¼ Bus2;
end

A nearly identical block is used to model the memory address register. A unique signal is declared
called MAR in order to make the Verilog more readable. MAR is always assigned to address in this
system.
always @ (posedge clock or negedge reset)
begin: MEMORY_ADDRESS_REGISTER
if (!reset)
MAR <¼ 8’h00;
else
if (MAR_Load)
MAR <¼ Bus2;
end
426 • Chapter 13: Computer System Design

Now let’s look at the program counter block. This register contains additional functionality beyond
simply latching in the value of Bus2. The program counter also has an increment feature that will take
place synchronously when the signal PC_Inc coming from the control unit is asserted. This is handled
using an additional nested if-else clause under the portion of the block handling the rising edge of clock
condition.
always @ (posedge clock or negedge reset)
begin: PROGRAM_COUNTER
if (!reset)
PC <¼ 8’h00;
else
if (PC_Load)
PC <¼ Bus2;
else if (PC_Inc)
PC <¼ MAR + 1;
end

The two general purpose registers A and B are modeled using individual procedural blocks as
follows:
always @ (posedge clock or negedge reset)
begin: A_REGISTER
if (!reset)
A <¼ 8’h00;
else
if (A_Load)
A <¼ Bus2;
end

always @ (posedge clock or negedge reset)

begin: B_REGISTER
if (!reset)
B <¼ 8’h00;
else
if (B_Load)
B <¼ Bus2;
end

The condition code register latches in the status flags from the ALU (NZVC) when the CCR_Load
line is asserted. This behavior is modeled using a similar approach as follows:
always @ (posedge clock or negedge reset)
begin: CONDITION_CODE_REGISTER
if (!reset)
CCR_Result <¼ 8’h00;
else
if (CCR_Load)
CCR_Result <¼ NZVC;
end

13.3.4.2 ALU Implementation in Verilog

The ALU is a set of combinational logic circuitry that performs arithmetic and logic operations. The
output of the ALU operation is called Result. The ALU also outputs 4 status flags as a 4-bit bus called
NZVC. The ALU behavior can be modeled using case and if-else statements that decide which operation
to perform based on the input control signal ALU_Sel. The following Verilog shows an example of how to
implement the ALU addition functionality. A case statement is used to decide which operation is being
performed based on the ALU_Sel input. Under each operation clause, a series of procedural statements
are used to compute the result and update the NZVC flags. Each of these flags is updated individually.
The N flag can be simply driven with position 7 of the ALU result since this bit is the sign bit for signed
 13.3 Computer Implementation – An 8-Bit Computer Example • 427

numbers. The Z flag can be driven using an if-else condition that checks whether the result was x00 0000 .
The V flag is updated based on the type of the operation. For the addition operation, the V flag will be
asserted if a POS + POS ¼ NEG or a NEG + NEG ¼ POS. These conditions can be checked by looking
at the sign bits of the inputs and the sign bit of the result. Finally, the C flag can be computed as the 8th bit
in the addition of A + B.

always @ (A, B, ALU_Sel)

begin
case (ALU_Sel)
3’b000 : begin //-- Addition

//-- Sum and Carry Flag

{NZVC[0], Result} ¼ A + B;

//-- Negative Flag

NZVC[3] ¼ Result[7];

//-- Zero Flag

if (Result ¼¼ 0)
NZVC[2] ¼ 1;
else
NZVC[2] ¼ 0;

//-- Two’s Comp Overflow Flag

if ( ((A[7]¼¼0) && (B[7]¼¼0) && (Result[7] ¼¼ 1)) ||
((A[7]¼¼1) && (B[7]¼¼1) && (Result[7] ¼¼ 0)) )
NZVC[1] ¼ 1;
else
NZVC[1] ¼ 0;

end

:
//-- other ALU operations go here...
:

default : begin
Result ¼ 8’hXX;
NZVC ¼ 4’hX;
end
endcase

end

13.3.4.3 Control Unit Implementation in Verilog

Let’s now look at how to implement the control unit state machine. We’ll first look at the formation of
the Verilog to model the FSM and then turn to the detailed state transitions in order to accomplish a
variety of the most common instructions. The control unit sends signals to the data path in order to move
data in and out of registers and into the ALU to perform data manipulations. The finite state machine is
implemented with the behavioral modeling techniques presented in Chapter 9. The model contains three
processes in order to implement the state memory, next state logic, and output logic of the FSM.
Parameters are created for each of the states defined in the state diagram of the FSM. The states
associated with fetching (S_FETCH_0, S_FETCH_1, S_FETCH_2) and decoding the opcode
(S_DECODE_3) are performed each time an instruction is executed. A unique path is then added
after the decode state to perform the steps associated with executing each individual instruction. The
FSM can be created one instruction at a time by adding additional state paths after the decode state. The
428 • Chapter 13: Computer System Design

following Verilog code shows how the user-defined state names are created for nine basic instructions
(LDA_IMM, LDA_DIR, STA_DIR, LDB_IMM, LDB_DIR, STB_DIR, ADD_AB, BRA and BEQ). Eight bit
state variables are created for current_state and next_state to accommodate future state codes. The
state codes are assigned in binary using integer format to allow additional states to be easily added.

reg [7:0] current_state, next_state;

parameter S_FETCH_0 ¼ 0, //-- Opcode fetch states
S_FETCH_1 ¼ 1,
S_FETCH_2 ¼ 2,

S_DECODE_3 ¼ 3, //-- Opcode decode state

S_LDA_IMM_4 ¼ 4, //-- Load A (Immediate) states

S_LDA_IMM_5 ¼ 5,
S_LDA_IMM_6 ¼ 6,

S_LDA_DIR_4 ¼ 7, //-- Load A (Direct) states

S_LDA_DIR_5 ¼ 8,
S_LDA_DIR_6 ¼ 9,
S_LDA_DIR_7 ¼ 10,
S_LDA_DIR_8 ¼ 11,

S_STA_DIR_4 ¼ 12, //-- Store A (Direct) States

S_STA_DIR_5 ¼ 13,
S_STA_DIR_6 ¼ 14,
S_STA_DIR_7 ¼ 15,

S_LDB_IMM_4 ¼ 16, //-- Load B (Immediate) states

S_LDB_IMM_5 ¼ 17,
S_LDB_IMM_6 ¼ 18,

S_LDB_DIR_4 ¼ 19, //-- Load B (Direct) states

S_LDB_DIR_5 ¼ 20,
S_LDB_DIR_6 ¼ 21,
S_LDB_DIR_7 ¼ 22,
S_LDB_DIR_8 ¼ 23,

S_STB_DIR_4 ¼ 24, //-- Store B (Direct) States

S_STB_DIR_5 ¼ 25,
S_STB_DIR_6 ¼ 26,
S_STB_DIR_7 ¼ 27,

S_BRA_4 ¼ 28, //-- Branch Always States

S_BRA_5 ¼ 29,
S_BRA_6 ¼ 30,

S_BEQ_4 ¼ 31, //-- Branch if Equal States

S_BEQ_5 ¼ 32,
S_BEQ_6 ¼ 33,
S_BEQ_7 ¼ 34,

S_ADD_AB_4 ¼ 35; //-- Addition States

Within the control unit module, the state memory is implemented as a separate procedural block that
will update the current state with the next state on each rising edge of the clock. The reset state will be the
first fetch state in the FSM (i.e., S_FETCH_0). The following Verilog shows how the state memory in the
control unit can be modeled. Note that this block models sequential logic so non-blocking assignments
are used.
 13.3 Computer Implementation – An 8-Bit Computer Example • 429

always @ (posedge clock or negedge reset)

begin: STATE_MEMORY
if (!reset)
current_state <¼ S_FETCH_0;
else
current_state <¼ next_state;
end

The next state logic is also implemented as a separate procedural block. The next state logic depends
on the current state, instruction register (IR), and the condition code register (CCR_Result). The following
Verilog gives a portion of the next state logic process showing how the state transitions can be modeled.
always @ (current_state, IR, CCR_Result)
begin: NEXT_STATE_LOGIC
case (current_state)
S_FETCH_0 : next_state ¼ S_FETCH_1; //-- Path for FETCH instruction
S_FETCH_1 : next_state ¼ S_FETCH_2;
S_FETCH_2 : next_state ¼ S_DECODE_3;

S_DECODE_3 : if (IR ¼¼ LDA_IMM) next_state ¼ S_LDA_IMM_4;

//-- Register A
else if (IR ¼¼ LDA_DIR) next_state ¼ S_LDA_DIR_4;
else if (IR ¼¼ STA_DIR) next_state ¼ S_STA_DIR_4;
else if (IR ¼¼ LDB_IMM) next_state ¼ S_LDB_IMM_4;
//-- Register B
else if (IR ¼¼ LDB_DIR) next_state ¼ S_LDB_DIR_4;
else if (IR ¼¼ STB_DIR) next_state ¼ S_STB_DIR_4;
else if (IR ¼¼ BRA) next_state ¼ S_BRA_4;
//-- Branch Always
else if (IR ¼¼ ADD_AB) next_state ¼ S_ADD_AB_4; //-- ADD
else next_state ¼ S_FETCH_0;
//-- others go here

S_LDA_IMM_4 : next_state ¼ S_LDA_IMM_5; //-- Path for LDA_IMM instruction

S_LDA_IMM_5 : next_state ¼ S_LDA_IMM_6;
S_LDA_IMM_6 : next_state ¼ S_FETCH_0;

:
Next state logic for other states goes here. . .
:
endcase
end

Finally, the output logic is modeled as a third, separate procedural block. It is useful to explicitly state
the outputs of the control unit for each state in the machine to allow easy debugging and avoid
synthesizing latches. Our example computer system has Moore type outputs so the process only
depends on the current state. The following Verilog shows a portion of the output logic process.
always @ (current_state)
begin: OUTPUT_LOGIC
case (current_state)

S_FETCH_0 : begin //-- Put PC onto MAR to provide address

of Opcode
IR_Load ¼ 0;
MAR_Load ¼ 1;
PC_Load ¼ 0;
PC_Inc ¼ 0;
A_Load ¼ 0;
B_Load ¼ 0;
ALU_Sel ¼ 3’b000;
CCR_Load ¼ 0;
Bus1_Sel ¼ 2’b00; //-- "00"¼PC, "01"¼A, "10"¼B
Bus2_Sel ¼ 2’b01; //-- "00"¼ALU, "01"¼Bus1,
"10"¼from_memory
write ¼ 0;
end
430 • Chapter 13: Computer System Design

S_FETCH_1 : begin //-- Increment PC, Opcode will be available next state
IR_Load ¼ 0;
MAR_Load ¼ 0;
PC_Load ¼ 0;
PC_Inc ¼ 1;
A_Load ¼ 0;
B_Load ¼ 0;
ALU_Sel ¼ 3’b000;
CCR_Load ¼ 0;
Bus1_Sel ¼ 2’b00; //-- "00"¼PC, "01"¼A, "10"¼B
Bus2_Sel ¼ 2’b00; //-- "00"¼ALU, "01"¼Bus1, "10"¼from_memory
write ¼ 0;
end;

:
Output logic for other states goes here. . .
:

endcase
end

13.3.4.3.1 Detailed Execution of LDA_IMM

Now let’s look at the details of the state transitions and output signals in the control unit FSM when
executing a few of the most common instructions. Let’s begin with the instruction to load register A using
immediate addressing (LDA_IMM). Example 13.12 shows the state diagram for this instruction. The first
three states (S_FETCH_0, S_FETCH_1, S_FETCH_2) handle fetching the opcode. The purpose of
these states is to read the opcode from the address being held by the program counter and put it into the
instruction register. Multiple states are needed to handle putting PC into MAR to provide the address of
the opcode, waiting for the memory system to provide the opcode, latching the opcode into IR, and
incrementing PC to the next location in program memory. Another state is used to decode the opcode
(S_DECODE_3) in order to decide which path to take in the state diagram based on the instruction being
executed. After the decode state, a series of three more states are needed (S_LDA_IMM_4,
S_LDA_IMM_5, S_LDA_IMM_6) to execute the instruction. The purpose of these states is to read the
operand from the address being held by the program counter and put it into A. Multiple states are needed
to handle putting PC into MAR to provide the address of the operand, waiting for the memory system to
provide the operand, latching the operand into A, and incrementing PC to the next location in program
memory. When the instruction completes, the value of the operand resides in A and PC is pointing to the
next location in program memory, which is the opcode of the next instruction to be executed.
 13.3 Computer Implementation – An 8-Bit Computer Example • 431

Example 13.12
State diagram for LDA_IMM
432 • Chapter 13: Computer System Design

Example 13.13 shows the simulation waveform for executing LDA_IMM. In this example, register A
is loaded with the operand of the instruction, which holds the value x00 AA00 .

Example 13.13
Simulation waveform for LDA_IMM

13.3.4.3.2 Detailed Execution of LDA_DIR

Now let’s look at the details of the instruction to load register A using direct addressing (LDA_DIR).
Example 13.14 shows the state diagram for this instruction. The first four states to fetch and decode the
opcode are the same states as in the previous instruction and are performed each time a new instruction
is executed. Once the opcode is decoded, the state machine traverses five new states to execute the
instruction (S_LDA_DIR_4, S_LDA_DIR_5, S_LDA_DIR_6, S_LDA_DIR_7, S_LDA_DIR_8). The pur-
pose of these states is to read the operand and then use it as the address of where to read the contents to
put into A.
 13.3 Computer Implementation – An 8-Bit Computer Example • 433

Example 13.14
State diagram for LDA_DIR

Example 13.15 shows the simulation waveform for executing LDA_DIR. In this example, register A
is loaded with the contents located at address x00 8000 , which has already been initialized to x00 AA00 .
434 • Chapter 13: Computer System Design

Example 13.15
Simulation waveform for LDA_DIR

13.3.4.3.3 Detailed Execution of STA_DIR

Now let’s look at the details of the instruction to store register A to memory using direct addressing
(STA_DIR). Example 13.16 shows the state diagram for this instruction. The first four states are again the
same as prior instructions in order to fetch and decode the opcode. Once the opcode is decoded, the
state machine traverses four new states to execute the instruction (S_STA_DIR_4, S_STA_DIR_5,
S_STA_DIR_6, S_STA_DIR_7). The purpose of these states is to read the operand and then use it as
the address of where to write the contents of A to.
 13.3 Computer Implementation – An 8-Bit Computer Example • 435

Example 13.16
State diagram for STA_DIR

Example 13.17 shows the simulation waveform for executing STA_DIR. In this example, register A
already contains the value x00 CC00 and will be stored to address x00 E000 . The address x00 E000 is an output
port (port_out_00) in our example computer system.
436 • Chapter 13: Computer System Design

Example 13.17
Simulation waveform for STA_DIR

13.3.4.3.4 Detailed Execution of ADD_AB

Now let’s look at the details of the instruction to add A to B and store the sum back in A (ADD_AB).
Example 13.18 shows the state diagram for this instruction. The first four states are again the same as
prior instructions in order to fetch and decode the opcode. Once the opcode is decoded, the state
machine only requires one more state to complete the operation (S_ADD_AB_4). The ALU is combina-
tional logic so it will begin to compute the sum immediately as soon as the inputs are updated. The inputs
to the ALU are Bus1 and register B. Since B is directly connected to the ALU, all that is required to start
the addition is to put A onto Bus1. The output of the ALU is put on Bus2 so that it can be latched into A on
the next clock edge. The ALU also outputs the status flags NZVC, which are directly connected to the
condition code register. A_Load and CCR_Load are asserted in this state. A and CCR_Result will be
updated in the next state (i.e., S_FETCH_0).
 13.3 Computer Implementation – An 8-Bit Computer Example • 437

Example 13.18
State diagram for ADD_AB

Example 13.19 shows the simulation waveform for executing ADD_AB. In this example, two load
immediate instructions were used to initialize the general purpose registers to A ¼ x00 FF00 and B ¼ x00 0100
prior to the addition. The addition of these values will result in a sum of x00 0000 and assert the carry (C) and
zero (Z) flags in the condition code register.
438 • Chapter 13: Computer System Design

Example 13.19
Simulation waveform for ADD_AB

13.3.4.3.5 Detailed Execution of BRA

Now let’s look at the details of the instruction to branch always (BRA). Example 13.20 shows the
state diagram for this instruction. The first four states are again the same as prior instructions in order to
fetch and decode the opcode. Once the opcode is decoded, the state machine traverses four new states
to execute the instruction (S_BRA_4, S_BRA_5, S_BRA_6). The purpose of these states is to read the
operand and put its value into PC to set the new location in program memory to execute instructions.
 13.3 Computer Implementation – An 8-Bit Computer Example • 439

Example 13.20
State diagram for BRA
440 • Chapter 13: Computer System Design

Example 13.21 shows the simulation waveform for executing BRA. In this example, PC is set back
to address x00 0000 .

Example 13.21
Simulation waveform for BRA

13.3.4.3.6 Detailed Execution of BEQ

Now let’s look at the branch if equal to zero (BEQ) instruction. Example 13.22 shows the state
diagram for this instruction. Notice that in this conditional branch, the path that is taken through the FSM
depends on both IR and CCR. In the case that Z ¼ 1, the branch is taken, meaning that the operand is
loaded into PC. In the case that Z ¼ 0, the branch is not taken, meaning that PC is simply incremented to
bypass the operand and point to the beginning of the next instruction in program memory.
 13.3 Computer Implementation – An 8-Bit Computer Example • 441

Example 13.22
State diagram for BEQ

Example 13.23 shows the simulation waveform for executing BEQ when the branch is taken. Prior to
this instruction, an addition was performed on x00 FF00 and x00 0100 . This resulted in a sum of x00 0000 , which
asserted the Z and C flags in the condition code register. Since Z ¼ 1 when BEQ is executed, the branch
is taken.
442 • Chapter 13: Computer System Design

Example 13.23
Simulation waveform for BEQ when taking the branch (Z ¼ 1)

Example 13.24 shows the simulation waveform for executing BEQ when the branch is not taken.
Prior to this instruction, an addition was performed on x00 FE00 and x00 0100 . This resulted in a sum of x00 FF00,
which did not assert the Z flag. Since Z ¼ 0 when BEQ is executed, the branch is not taken. When not
taking the branch, PC must be incremented again in order to bypass the operand and point to the next
location in program memory.
 13.3 Computer Implementation – An 8-Bit Computer Example • 443

Example 13.24
Simulation waveform for BEQ when the branch is not taken (Z ¼ 0)
444 • Chapter 13: Computer System Design

CONCEPT CHECK

CC13.3 The 8-bit microcomputer example presented in this section is a very simple architec-
ture used to illustrate the basic concepts of a computer. If we wanted to keep this
computer an 8-bit system but increase the depth of the memory, it would require
adding more address lines to the address bus. What changes to the computer system
would need to be made to accommodate the wider address bus?
(A) The width of the program counter would need to be increased to support the
wider address bus.
(B) The size of the memory address register would need to be increased to
support the wider address bus.
(C) Instructions that use direct addressing would need additional bytes of
operand to pass the wider address into the CPU 8-bits at a time.
(D) All of the above.

13.4 Architecture Considerations

13.4.1 Von Neumann Architecture
The computer system just presented represents a very simple architecture in which all memory
devices (i.e., program, data, and I/O) are grouped into a single memory map. This approach is known as
the Von Neumann architecture, named after the 19th century mathematician that first described this
structure in 1945. The advantage of this approach is in the simplicity of the CPU interface. The CPU can
be constructed based on a single bus system that executes everything in a linear progression of states,
regardless of whether memory is being accessed for an instruction or a variable. One of the drawbacks of
this approach is that an instruction and variable data cannot be read at the same time. This creates a
latency in data manipulation since the system needed to be constantly switching between reading
instructions and accessing data. This latency became known as the Von Neumann bottleneck.

13.4.2 Harvard Architecture

As computer systems evolved and larger data sets in memory were being manipulated, it became
apparent that it was advantageous to be able to access data in parallel with reading the next instruction.
The Harvard architecture was proposed to address the Von Neumann bottleneck by separating the
program and data memory and using two distinct bus systems for the CPU interface. This approach
allows data and program information to be accessed in parallel and leads to performance improvement
when large numbers of data manipulations in memory need to be performed. Figure 13.5 shows a
comparison between the two architectures.
 13.4 Architecture Considerations • 445

Fig. 13.5
Von Neumann vs. Harvard Architecture

CONCEPT CHECK

CC13.4 Does a computer with a Harvard architecture require two control unit state machines?
(A) Yes. It has two bus systems that need to be managed separately so two finite
state machines are required.
(B) No. A single state machine is still used to fetch, decode, and execute the
instruction. The only difference is that if data is required for the execute
stage, it can be retrieved from data memory at the same time the state
machine fetches the opcode of the next instruction from program memory.
446 • Chapter 13: Computer System Design

Summary
v A computer is a collection of hardware v A computer also contains data memory to
components that are constructed to perform hold temporary variables during run time.
a specific set of instructions to process and v A computer also contains input and output
store data. The main hardware components ports to interface with the outside world.
of a computer are the central processing unit v A memory mapped system is one in which
(CPU), program memory, data memory, and the program memory, data memory, and I/O
input/output ports. ports are all assigned a unique address. This
v The CPU consists of registers for fast storage, allows the CPU to simply process information
an arithmetic logic unit (ALU) for data manip- as data and addresses and allows the pro-
ulation, and a control state machine that gram to handle where the information is
directs all activity to execute an instruction. being sent to. A memory map is a graphical
v A CPU is typically organized into a data path representation of what address ranges vari-
and a control unit. The data path contains all ous components are mapped to.
circuitry used to store and process information. v There are three primary classes of
The data path includes the registers and the instructions. These are loads and stores,
ALU. The control unit is a large state machine data manipulations, and branches.
that sends control signals to the data path in v Load instructions move information from
order to facilitate instruction execution. memory into a CPU register. A load instruc-
v The control unit continuously performs a tion takes multiple read cycles. Store
fetch-decode-execute cycle in order to com- instructions move information from a CPU
plete instructions. register into memory. A store instruction
v The instructions that a computer is designed takes multiple read cycles and at least one
to execute is called its instruction set. write cycle.
v Instructions are inserted into program mem- v Data manipulation instructions operate on
ory in a sequence that when executed will information being held in CPU registers.
accomplish a particular task. This sequence Data manipulation instructions often use
of instructions is called a computer program. inherent addressing.
v An instruction consists of an opcode and a v Branch instructions alter the flow of instruc-
potential operand. The opcode is the unique tion execution. Unconditional branches
binary code that tells the control state always change the location in memory of
machine which instruction is being executed. where the CPU is executing instructions.
An operand is additional information that may Conditional branches only change the loca-
be needed for the instruction. tion of instruction execution if a status flag is
v An addressing mode refers to the way that asserted.
the operand is treated. In immediate v Status flags are held in the condition code
addressing the operand is the actual data to register and are updated by certain
be used. In direct addressing the operand is instructions. The most commonly used flags
the address of where the data is to be are the negative flag (N), zero flag (Z), two’s
retrieved or stored. In inherent addressing complement overflow flag (V), and carry
all of the information needed to complete flag (C).
the instruction is contained within the opcode
so no operand is needed.

Exercise Problems
Section 13.1: Computer Hardware 13.1.4 What computer hardware sub-system contains
the state machine that orchestrates the fetch-
13.1.1 What computer hardware sub-system holds
decode-execute process?
the temporary variables used by the program?
13.1.5 What computer hardware sub-system contains
13.1.2 What computer hardware sub-system contains
the circuitry that performs mathematical and
fast storage for holding and/or manipulating
logic operations?
data and addresses?
13.1.6 What computer hardware sub-system holds
13.1.3 What computer hardware sub-system allows
the instructions being executed?
the computer to interface to the outside world?
 Exercise Problems • 447

Section 13.2: Computer Software ROM[1] ¼ 8’hF0;

ROM[2] ¼ STA_DIR;
13.2.1 In computer software, what are the names of
ROM[3] ¼ 8’hE0;
the most basic operations that a computer can ROM[4] ¼ BRA;
perform? ROM[5] ¼ 8’h00;
13.2.2 Which element of computer software is the End
binary code that tells the CPU which instruction
13.3.3 Add the functionality to the computer model
is being executed?
from 13.3.2 the ability to perform the
13.2.3 Which element of computer software is a col- instructions LDB_IMM, LDB_DIR, and
lection of instructions that perform a desired STB_DIR. Modify the example programs
task? given in exercise 13.3.1 and 13.3.2 to use
13.2.4 Which element of computer software is the register B in order to simulate your
supplementary information required by an implementation.
instruction such as constants or which 13.3.4 Add the functionality to the computer model
registers to use? from 13.3.3 the ability to perform the addition
13.2.5 Which class of instructions handles moving instruction ADD_AB. Test your addition instruc-
information between memory and CPU tion by simulating the following program. The
registers? first addition instruction will perform
x00 FE00 + x00 0100 ¼ x00 FF00 and assert the negative
13.2.6 Which class of instructions alters the flow of
(N) flag. The second addition instruction will
program execution?
perform x00 0100 + x00 FF00 ¼ x00 0000 and assert
13.2.7 Which class of instructions alters data using the carry (C) and zero (Z) flags. The third addi-
either arithmetic or logical operations? tion instruction will perform
x00 7F00 + x00 7F00 ¼ x00 FE00 and assert the two’s
Section 13.3: Computer Implementation – complement overflow (V) and negative
An 8-bit Computer Example (N) flags.

13.3.1 Design the example 8-bit computer system initial

presented in this chapter in Verilog with the begin
ability to execute the three instructions ROM[0] ¼ LDA_IMM; //-- test 1
LDA_IMM, STA_DIR, and BRA. Simulate your ROM[1] ¼ 8’hFE;
computer system using the following program ROM[2] ¼ LDB_IMM;
that will continually write the patterns x00 AA” ROM[3] ¼ 8’h01;
and x00 BB00 to output ports port_out_00 and ROM[4] ¼ ADD_AB;
port_out_01: ROM[5] ¼ LDA_IMM; //-- test 2
ROM[6] ¼ 8’h01;
initial ROM[7] ¼ LDB_IMM;
begin ROM[8] ¼ 8’hFF;
ROM[0] ¼ LDA_IMM; ROM[9] ¼ ADD_AB;
ROM[1] ¼ 8’hAA; ROM[10] ¼ LDA_IMM; //-- test 3
ROM[2] ¼ STA_DIR; ROM[11] ¼ 8’h7F;
ROM[3] ¼ 8’hE0; ROM[12] ¼ LDB_IMM;
ROM[4] ¼ STA_DIR; ROM[13] ¼ 8’h7F;
ROM[5] ¼ 8’hE1; ROM[14] ¼ ADD_AB;
ROM[6] ¼ LDB_IMM; ROM[15] ¼ BRA;
ROM[7] ¼ 8’hBB; ROM[16] ¼ 8’h00;
ROM[8] ¼ STB_DIR; end
ROM[9] ¼ 8’hE0;
13.3.5 Add the functionality to the computer model
ROM[10] ¼ STB_DIR;
ROM[11] ¼ 8’hE1; from 13.3.4 the ability to perform the branch if
ROM[12] ¼ BRA; equal to zero instruction BEQ. Simulate your
ROM[13] ¼ 8’h00; implementation using the following program.
end The first addition in this program will perform
x00 FE00 + x00 0100 ¼ x00 FF00 (Z ¼ 0). The
13.3.2 Add the functionality to the computer model subsequent BEQ instruction should NOT take
from 13.3.1 the ability to perform the LDA_DIR the branch. The second addition in this pro-
instruction. Simulate your computer system gram will perform x00 FF00 + x00 0100 ¼ x00 0000
using the following program that will continually (Z ¼ 1) and SHOULD take the branch. The
read from port_in_00 and write its contents to final instruction in this program is a BRA that
port_out_00: is inserted for safety. In the event that the BEQ
is not operating properly, the BRA will set the
initial
program counter back to x00 0000 and prevent the
begin
program from running away.
ROM[0] ¼ LDA_DIR;
448 • Chapter 13: Computer System Design

initial Section 13.4: Architectural

begin
ROM[0] ¼ LDA_IMM; //-- test 1
Considerations
ROM[1] ¼ 8’hFE; 13.4.1 Would the instruction set need to be different
ROM[2] ¼ LDB_IMM; between a Von Neumann versus a Harvard
ROM[3] ¼ 8’h01; architecture? Why or why not?
ROM[4] ¼ ADD_AB; 13.4.2 Which of the three classes of computer
ROM[5] ¼ BEQ; //--NO branch instructions (loads/stores, data manipulations,
ROM[6] ¼ 8’h00; and branches) are sped up by moving from the
Von Neumann architecture to the Harvard
ROM[7] ¼ LDA_IMM; //-- test 2 architecture.
ROM[8] ¼ 8’h01; 13.4.3 In a memory mapped, Harvard architecture,
ROM[9] ¼ LDB_IMM;
would the I/O system be placed in the program
ROM[10] ¼ 8’hFF;
memory or data memory block?
ROM[11] ¼ ADD_AB;
ROM[12] ¼ BEQ; //-- Branch 13.4.4 A Harvard architecture requires two memory
ROM[13] ¼ 8’h00; address registers to handle two separate mem-
ory systems. Does it also require two instruc-
ROM[14] ¼ BRA; tion registers? Why or why not?
ROM[15] ¼ 8’h00; 13.4.5 A Harvard architecture requires two memory
end address registers to handle two separate mem-
ory systems. Does it also require two program
13.3.6 Add the functionality to the computer model
counters? Why or why not?
from 13.3.4 all of the remaining instructions in
the set shown in Example 13.9. You will need
to create test programs to verify the execution
of each instruction.
Appendix A: List of Worked Examples
EXAMPLE 2.1 CONVERTING DECIMAL TO DECIMAL ........................................................ ....................................................... 12
EXAMPLE 2.2 CONVERTING BINARY TO DECIMAL ........................................................ ........................................................ 13
EXAMPLE 2.3 CONVERTING OCTAL TO DECIMAL ......................................................... ........................................................ 13
EXAMPLE 2.4 CONVERTING HEXADECIMAL TO DECIMAL .................................................... ................................................... 14
EXAMPLE 2.5 CONVERTING DECIMAL TO BINARY ........................................................ ........................................................ 15
EXAMPLE 2.6 CONVERTING DECIMAL TO OCTAL ......................................................... ........................................................ 16
EXAMPLE 2.7 CONVERTING DECIMAL TO HEXADECIMAL .................................................... ................................................... 17
EXAMPLE 2.8 CONVERTING BINARY TO OCTAL .......................................................... ......................................................... 18
EXAMPLE 2.9 CONVERTING BINARY TO HEXADECIMAL ..................................................... .................................................... 18
EXAMPLE 2.10 CONVERTING OCTAL TO BINARY ......................................................... ........................................................ 19
EXAMPLE 2.11 CONVERTING HEXADECIMAL TO BINARY .................................................... ................................................... 19
EXAMPLE 2.12 CONVERTING OCTAL TO HEXADECIMAL .................................................... .................................................... 20
EXAMPLE 2.13 CONVERTING HEXADECIMAL TO OCTAL .................................................... .................................................... 20
EXAMPLE 2.14 SINGLE BIT BINARY ADDITION .......................................................... .......................................................... 21
EXAMPLE 2.15 MULTIPLE BIT BINARY ADDITION ......................................................... ........................................................ 22
EXAMPLE 2.16 SINGLE BIT BINARY SUBTRACTION ....................................................... ....................................................... 22
EXAMPLE 2.17 MULTIPLE BIT BINARY SUBTRACTION ...................................................... ..................................................... 23
EXAMPLE 2.18 FINDING THE RANGE OF AN UNSIGNED NUMBER .............................................. ............................................. 24
EXAMPLE 2.19 DECIMAL VALUES THAT A 4-BIT, SIGNED MAGNITUDE CODE CAN REPRESENT ........................... .......................... 25
EXAMPLE 2.20 FINDING THE RANGE OF A SIGNED MAGNITUDE NUMBER ........................................ ........................................ 26
EXAMPLE 2.21 FINDING THE DECIMAL VALUE OF A SIGNED MAGNITUDE NUMBER ................................... .................................. 26
EXAMPLE 2.22 DECIMAL VALUES THAT A 4-BIT, ONE'S COMPLEMENT CODE CAN REPRESENT .......................... .......................... 27
EXAMPLE 2.23 FINDING THE RANGE OF A 1’S COMPLEMENT NUMBER .......................................... ......................................... 28
EXAMPLE 2.24 FINDING THE DECIMAL VALUE OF A 1’S COMPLEMENT NUMBER .................................... ................................... 28
EXAMPLE 2.25 DECIMAL VALUES THAT A 4-BIT, TWO'S COMPLEMENT CODE CAN REPRESENT .......................... .......................... 29
EXAMPLE 2.26 FINDING THE RANGE OF A TWO'S COMPLEMENT NUMBER ........................................ ....................................... 30
EXAMPLE 2.27 FINDING THE DECIMAL VALUE OF A TWO'S COMPLEMENT NUMBER .................................. .................................. 30
EXAMPLE 2.28 FINDING THE TWO'S COMPLEMENT CODE OF A DECIMAL NUMBER .................................. .................................. 31
EXAMPLE 2.29 TWO'S COMPLEMENT ADDITION ......................................................... ......................................................... 32
EXAMPLE 3.1 CALCULATING ICC AND IGND WHEN SOURCING MULTIPLE LOADS ................................... ................................... 49
EXAMPLE 3.2 CALCULATING ICC AND IGND WHEN BOTH SOURCING AND SINKING LOADS ............................. ............................ 50
EXAMPLE 3.3 DETERMINING IF SPECIFICATIONS ARE VIOLATED WHEN DRIVING ANOTHER GATE AS A LOAD ................. ................. 72
EXAMPLE 3.4 DETERMINING THE OUTPUT CURRENT WHEN DRIVING MULTIPLE GATES AS THE LOAD ...................... ..................... 73
EXAMPLE 3.5 DETERMINING THE OUTPUT CURRENT WHEN DRIVING A PULL-UP RESISTOR AS THE LOAD ................... .................. 74
EXAMPLE 3.6 DETERMINING THE OUTPUT CURRENT WHEN DRIVING A PULL-DOWN RESISTOR AS THE LOAD ................ ................ 75
EXAMPLE 3.7 DETERMINING THE OUTPUT CURRENT WHEN DRIVING AN LED WHERE HIGH ¼ ON ..................... .................... 76
EXAMPLE 3.8 DETERMINING THE OUTPUT CURRENT WHEN DRIVING AN LED WHERE HIGH ¼ OFF .................... ................... 77
EXAMPLE 4.1 PROVING DEMORGAN’S THEOREM OF DUALITY USING PROOF BY EXHAUSTION .......................... ......................... 84
EXAMPLE 4.2 CONVERTING BETWEEN POSITIVE AND NEGATIVE LOGIC USING DUALITY ............................... ............................... 85
EXAMPLE 4.3 USING THE COMMUTATIVE PROPERTY TO UNTANGLE CROSSED WIRES ................................ ................................ 89
EXAMPLE 4.4 USING THE ASSOCIATIVE PROPERTY TO ADDRESS FAN-IN LIMITATIONS ................................ ................................ 90
EXAMPLE 4.5 USING THE DISTRIBUTIVE PROPERTY TO REDUCE THE NUMBER OF LOGIC GATES IN A CIRCUIT ............... ............... 91
EXAMPLE 4.6 PROVING THE ABSORPTION THEOREM USING PROOF BY EXHAUSTION ................................ ................................ 92
EXAMPLE 4.7 PROVING OF THE UNITING THEOREM ...................................................... ...................................................... 93
EXAMPLE 4.8 CONVERTING A SUM OF PRODUCTS FORM INTO ONE THAT USES ONLY NAND GATES ..................... .................... 95
EXAMPLE 4.9 CONVERTING A PRODUCT OF SUMS FORM INTO ONE THAT USES ONLY NOR GATES ...................... ...................... 96
EXAMPLE 4.10 USING DEMORGAN’S THEOREM IN ALGEBRAIC FORM (1) ....................................... ....................................... 97
EXAMPLE 4.11 USING DEMORGAN’S THEOREM IN ALGEBRAIC FORM (2) ....................................... ....................................... 97
EXAMPLE 4.12 DETERMINING THE LOGIC EXPRESSION FROM A LOGIC DIAGRAM .................................. ................................. 100
EXAMPLE 4.13 DETERMINING THE TRUTH TABLE FROM A LOGIC DIAGRAM ...................................... ...................................... 101
EXAMPLE 4.14 DETERMINING THE DELAY OF A COMBINATIONAL LOGIC CIRCUIT .................................. .................................. 102

# Springer International Publishing AG 2017 449

B.J. LaMeres, Introduction to Logic Circuits & Logic Design with Verilog,
DOI 10.1007/978-3-319-53883-9
450 • Appendix A: List of Worked Examples

EXAMPLE 4.15 CREATING A CANONICAL SUM OF PRODUCTS LOGIC CIRCUIT USING MINTERMS ........................ ........................ 104
EXAMPLE 4.16 CREATING A MINTERM LIST FROM A TRUTH TABLE ............................................ ........................................... 105
EXAMPLE 4.17 CREATING EQUIVALENT FUNCTIONAL REPRESENTATIONS FROM A MINTERM LIST ........................ ....................... 106
EXAMPLE 4.18 CREATING A PRODUCT OF SUMS LOGIC CIRCUIT USING MAXTERMS ................................ ................................ 108
EXAMPLE 4.19 CREATING A MAXTERM LIST FROM A TRUTH TABLE ........................................... ........................................... 109
EXAMPLE 4.20 CREATING EQUIVALENT FUNCTIONAL REPRESENTATIONS FROM A MAXTERM LIST ................................................ 110
EXAMPLE 4.21 CREATING EQUIVALENT FORMS TO REPRESENT LOGIC FUNCTIONALITY ............................................................. 111
EXAMPLE 4.22 MINIMIZING A LOGIC EXPRESSION ALGEBRAICALLY ........................................................................................ 113
EXAMPLE 4.23 USING A K-MAP TO FIND A MINIMIZED SUM OF PRODUCTS EXPRESSION (2-INPUT) ............................................ 118
EXAMPLE 4.24 USING A K-MAP TO FIND A MINIMIZED SUM OF PRODUCTS EXPRESSION (3-INPUT) ............................................ 119
EXAMPLE 4.25 USING A K-MAP TO FIND A MINIMIZED SUM OF PRODUCTS EXPRESSION (4-INPUT) ..................... ..................... 120
EXAMPLE 4.26 USING A K-MAP TO FIND A MINIMIZED PRODUCT OF SUMS EXPRESSION (2-INPUT) ..................... ..................... 121
EXAMPLE 4.27 USING A K-MAP TO FIND A MINIMIZED PRODUCT OF SUMS EXPRESSION (3-INPUT) ..................... ..................... 122
EXAMPLE 4.28 USING A K-MAP TO FIND A MINIMIZED PRODUCT OF SUMS EXPRESSION (4-INPUT) ..................... ..................... 123
EXAMPLE 4.29 DERIVING THE MINIMAL SUM FROM A K-MAP ............................................... .............................................. 125
EXAMPLE 4.30 USING DON’T CARES TO PRODUCE A MINIMAL SOP LOGIC EXPRESSION ............................ ............................ 126
EXAMPLE 4.31 ELIMINATING A TIMING HAZARD BY INCLUDING NON-ESSENTIAL PRODUCT TERMS ....................... ....................... 131
EXAMPLE 5.1 DECLARING VERILOG MODULE PORTS .................................................... .................................................... 157
EXAMPLE 5.2 MODELING COMBINATIONAL LOGIC USING CONTINUOUS ASSIGNMENT WITH LOGICAL OPERATORS ............. ............. 165
EXAMPLE 5.3 MODELING COMBINATIONAL LOGIC USING CONTINUOUS ASSIGNMENT WITH CONDITIONAL OPERATORS (1) ....... ....... 166
EXAMPLE 5.4 MODELING COMBINATIONAL LOGIC USING CONTINUOUS ASSIGNMENT WITH CONDITIONAL OPERATORS (2) ....... ....... 167
EXAMPLE 5.5 MODELING DELAY IN CONTINUOUS ASSIGNMENTS ............................................. ............................................. 168
EXAMPLE 5.6 INERTIAL DELAY MODELING WHEN USING CONTINUOUS ASSIGNMENT. ................................ ................................ 169
EXAMPLE 5.7 VERILOG STRUCTURAL DESIGN USING EXPLICIT PORT MAPPING .................................... ................................... 171
EXAMPLE 5.8 VERILOG STRUCTURAL DESIGN USING POSITIONAL PORT MAPPING .................................. ................................. 172
EXAMPLE 5.9 MODELING COMBINATIONAL LOGIC CIRCUITS USING GATE LEVEL PRIMITIVES ........................... ........................... 173
EXAMPLE 5.10 MODELING COMBINATIONAL LOGIC CIRCUITS WITH A USER-DEFINED PRIMITIVE ......................... ........................ 174
EXAMPLE 6.1 2-TO-4 ONE-HOT DECODER – LOGIC DYNTHESIS BY HAND ....................................... ...................................... 182
EXAMPLE 6.2 3-TO-8 ONE-HOT DECODER – VERILOG MODELING USING LOGICAL OPERATORS ........................ ........................ 183
EXAMPLE 6.3 3-TO-8 ONE-HOT DECODER – VERILOG MODELING USING CONDITIONAL OPERATORS ..................... ..................... 184
EXAMPLE 6.4 7-SEGMENT DISPLAY DECODER – TRUTH TABLE .............................................. .............................................. 185
EXAMPLE 6.5 7-SEGMENT DISPLAY DECODER – LOGIC SYNTHESIS BY HAND .................................... .................................... 186
EXAMPLE 6.6 7-SEGMENT DISPLAY DECODER – VERILOG MODELING USING LOGICAL OPERATORS ...................... ...................... 187
EXAMPLE 6.7 7-SEGMENT DISPLAY DECODER – VERILOG MODELING USING CONDITIONAL OPERATORS ................... .................. 187
EXAMPLE 6.8 4-TO-2 BINARY ENCODER – LOGIC SYNTHESIS BY HAND ....................................... ........................................ 189
EXAMPLE 6.9 4-TO-2 BINARY ENCODER – VERILOG MODELING USING LOGICAL AND CONDITIONAL OPERATORS ............. ............ 190
EXAMPLE 6.10 2-TO-1 MULTIPLEXER – LOGIC SYNTHESIS BY HAND .......................................... .......................................... 191
EXAMPLE 6.11 4-TO-1 MULTIPLEXER – VERILOG MODELING USING LOGICAL AND CONDITIONAL OPERATORS ............... .............. 192
EXAMPLE 6.12 1-TO-2 DEMULTIPLEXER – LOGIC SYNTHESIS BY HAND ........................................ ........................................ 193
EXAMPLE 6.13 1-TO-4 DEMULTIPLEXER – VERILOG MODELING USING LOGICAL AND CONDITIONAL OPERATORS ............. ............ 194
EXAMPLE 7.1 PUSH-BUTTON WINDOW CONTROLLER – WORD DESCRIPTION ..................................... ..................................... 223
EXAMPLE 7.2 PUSH-BUTTON WINDOW CONTROLLER – STATE DIAGRAM ........................................ ........................................ 224
EXAMPLE 7.3 PUSH-BUTTON WINDOW CONTROLLER – STATE TRANSITION TABLE .................................. ................................. 225
EXAMPLE 7.4 SOLVING FOR THE NUMBER OF BITS NEEDED FOR BINARY STATE ENCODING ........................... .......................... 227
EXAMPLE 7.5 PUSH-BUTTON WINDOW CONTROLLER – STATE ENCODING ....................................... ....................................... 229
EXAMPLE 7.6 PUSH-BUTTON WINDOW CONTROLLER – NEXT STATE LOGIC ...................................... ..................................... 230
EXAMPLE 7.7 PUSH-BUTTON WINDOW CONTROLLER – OUTPUT LOGIC ......................................... ........................................ 231
EXAMPLE 7.8 PUSH-BUTTON WINDOW CONTROLLER – LOGIC DIAGRAM ........................................ ....................................... 232
EXAMPLE 7.9 SERIAL BIT SEQUENCE DETECTOR (PART 1) ................................................ ................................................ 234
EXAMPLE 7.10 SERIAL BIT SEQUENCE DETECTOR (PART 2) ............................................... ............................................... 235
EXAMPLE 7.11 SERIAL BIT SEQUENCE DETECTOR (PART 3) ............................................... ............................................... 236
EXAMPLE 7.12 VENDING MACHINE CONTROLLER (PART 1) ................................................ ............................................... 237
EXAMPLE 7.13 VENDING MACHINE CONTROLLER (PART 2) ................................................ ............................................... 238
EXAMPLE 7.14 VENDING MACHINE CONTROLLER (PART 3) ................................................ ............................................... 239
EXAMPLE 7.15 2-BIT BINARY UP COUNTER (PART 1) .................................................... ................................................... 241
 Appendix A: List of Worked Examples • 451

EXAMPLE 7.16 2-BIT BINARY UP COUNTER (PART 2) .................................................... ................................................... 242

EXAMPLE 7.17 2-BIT BINARY UP/DOWN COUNTER (PART 1) ............................................... ............................................... 243
EXAMPLE 7.18 2-BIT BINARY UP/DOWN COUNTER (PART 2) ............................................... ............................................... 244
EXAMPLE 7.19 2-BIT GRAY CODE UP COUNTER (PART 1) ................................................. ................................................ 245
EXAMPLE 7.20 2-BIT GRAY CODE UP COUNTER (PART 2) ................................................. ................................................ 246
EXAMPLE 7.21 2-BIT GRAY CODE UP/DOWN COUNTER (PART 1) ........................................... ............................................ 247
EXAMPLE 7.22 2-BIT GRAY CODE UP/DOWN COUNTER (PART 2) ........................................... ............................................ 248
EXAMPLE 7.23 3-BIT ONE-HOT UP COUNTER (PART 1) .................................................. .................................................. 249
EXAMPLE 7.24 3-BIT ONE-HOT UP COUNTER (PART 2) .................................................. .................................................. 250
EXAMPLE 7.25 3-BIT ONE-HOT UP/DOWN COUNTER (PART 1) .............................................. ............................................. 251
EXAMPLE 7.26 3-BIT ONE-HOT UP/DOWN COUNTER (PART 2) .............................................. ............................................. 252
EXAMPLE 7.27 3-BIT ONE-HOT UP/DOWN COUNTER (PART 3) .............................................. ............................................. 253
EXAMPLE 7.28 DETERMINING THE NEXT STATE LOGIC AND OUTPUT LOGIC EXPRESSION OF A FSM .................... .................... 256
EXAMPLE 7.29 DETERMINING THE STATE TRANSITION TABLE OF A FSM ....................................... ...................................... 257
EXAMPLE 7.30 DETERMINING THE STATE DIAGRAM OF A FSM ............................................. ............................................. 258
EXAMPLE 7.31 DETERMINING THE MAXIMUM CLOCK FREQUENCY OF A FSM .................................... ................................... 261
EXAMPLE 8.1 USING BLOCKING ASSIGNMENTS TO MODEL COMBINATIONAL LOGIC ................................. ................................. 275
EXAMPLE 8.2 USING NON-BLOCKING ASSIGNMENTS TO MODEL SEQUENTIAL LOGIC ................................ ................................ 275
EXAMPLE 8.3 IDENTICAL BEHAVIOR WHEN USING BLOCKING VS. NON-BLOCKING ASSIGNMENTS ........................ ........................ 276
EXAMPLE 8.4 DIFFERENT BEHAVIOR WHEN USING BLOCKING VS. NON-BLOCKING ASSIGNMENTS (1) ..................... .................... 277
EXAMPLE 8.5 DIFFERENT BEHAVIOR WHEN USING BLOCKING VS. NON-BLOCKING ASSIGNMENTS (2) ..................... .................... 278
EXAMPLE 8.6 BEHAVIOR OF STATEMENT GROUPS BEGIN/END VS. FORK/JOIN .................................... .................................... 279
EXAMPLE 8.7 USING IF-ELSE STATEMENTS TO MODEL COMBINATIONAL LOGIC .................................... ................................... 281
EXAMPLE 8.8 USING CASE STATEMENTS TO MODEL COMBINATIONAL LOGIC ..................................... ..................................... 282
EXAMPLE 8.9 TEST BENCH FOR A COMBINATIONAL LOGIC CIRCUIT ........................................... ........................................... 291
EXAMPLE 8.10 TEST BENCH FOR A SEQUENTIAL LOGIC CIRCUIT ............................................ ............................................ 292
EXAMPLE 8.11 PRINTING TEST BENCH RESULTS TO THE TRANSCRIPT ......................................... ......................................... 293
EXAMPLE 8.12 TEST BENCH WITH AUTOMATIC OUTPUT CHECKING ........................................... ........................................... 294
EXAMPLE 8.13 USING A LOOP TO GENERATE STIMULUS IN A TEST BENCH ...................................... ..................................... 295
EXAMPLE 8.14 PRINTING TEST BENCH RESULTS TO AN EXTERNAL FILE ........................................ ....................................... 296
EXAMPLE 8.15 READING TEST BENCH STIMULUS VECTORS FROM AN EXTERNAL FILE .............................. .............................. 297
EXAMPLE 9.1 BEHAVIORAL MODEL OF A D-LATCH IN VERILOG ............................................. ............................................. 303
EXAMPLE 9.2 BEHAVIORAL MODEL OF A D-FLIP-FLOP IN VERILOG .......................................... .......................................... 304
EXAMPLE 9.3 BEHAVIORAL MODEL OF A D-FLIP-FLOP WITH ASYNCHRONOUS RESET IN VERILOG ...................... ..................... 305
EXAMPLE 9.4 BEHAVIORAL MODEL OF A D-FLIP-FLOP WITH ASYNCHRONOUS RESET AND PRESET IN VERILOG ............. ............ 306
EXAMPLE 9.5 BEHAVIORAL MODEL OF A D-FLIP-FLOP WITH SYNCHRONOUS ENABLE IN VERILOG ...................... ..................... 307
EXAMPLE 9.6 PUSH-BUTTON WINDOW CONTROLLER IN VERILOG – DESIGN DESCRIPTION ............................ ........................... 308
EXAMPLE 9.7 PUSH-BUTTON WINDOW CONTROLLER IN VERILOG – PORT DEFINITION .............................. .............................. 308
EXAMPLE 9.8 PUSH-BUTTON WINDOW CONTROLLER IN VERILOG – FULL MODEL ..................................................................... 311
EXAMPLE 9.9 PUSH-BUTTON WINDOW CONTROLLER IN VERILOG – SIMULATION WAVEFORM .......................... .......................... 312
EXAMPLE 9.10 PUSH-BUTTON WINDOW CONTROLLER IN VERILOG – CHANGING STATE CODES ........................ ........................ 312
EXAMPLE 9.11 SERIAL BIT SEQUENCE DETECTOR IN VERILOG – DESIGN DESCRIPTION AND PORT DEFINITION .............. ............. 313
EXAMPLE 9.12 SERIAL BIT SEQUENCE DETECTOR IN VERILOG – FULL MODEL ................................... ................................... 314
EXAMPLE 9.13 SERIAL BIT SEQUENCE DETECTOR IN VERILOG – SIMULATION WAVEFORM ........................... ........................... 315
EXAMPLE 9.14 VENDING MACHINE CONTROLLER IN VERILOG – DESIGN DESCRIPTION AND PORT DEFINITION .............. .............. 315
EXAMPLE 9.15 VENDING MACHINE CONTROLLER IN VERILOG – FULL MODEL .................................... ................................... 316
EXAMPLE 9.16 VENDING MACHINE CONTROLLER IN VERILOG – SIMULATION WAVEFORM ............................ ............................ 317
EXAMPLE 9.17 2-BIT UP/DOWN COUNTER IN VERILOG – DESIGN DESCRIPTION AND PORT DEFINITION ................... ................... 317
EXAMPLE 9.18 2-BIT UP/DOWN COUNTER IN VERILOG – FULL MODEL (THREE BLOCK APPROACH) ..................... ..................... 318
EXAMPLE 9.19 2-BIT UP/DOWN COUNTER IN VERILOG – SIMULATION WAVEFORM ................................. ................................. 318
EXAMPLE 9.20 BINARY COUNTER USING A SINGLE PROCEDURAL BLOCK IN VERILOG .............................. .............................. 319
EXAMPLE 9.21 BINARY COUNTER WITH RANGE CHECKING IN VERILOG ........................................ ........................................ 320
EXAMPLE 9.22 BINARY COUNTER WITH ENABLE IN VERILOG ............................................... .............................................. 321
EXAMPLE 9.23 BINARY COUNTER WITH LOAD IN VERILOG ................................................ ................................................ 322
EXAMPLE 9.24 RTL MODEL OF AN 8-BIT REGISTER IN VERILOG ............................................ ........................................... 323
452 • Appendix A: List of Worked Examples

EXAMPLE 9.25 REGISTERS AS AGENTS ON A DATA BUS – SYSTEM TOPOLOGY ................................... .................................. 324
EXAMPLE 9.26 REGISTERS AS AGENTS ON A DATA BUS – RTL MODEL IN VERILOG .............................. .............................. 324
EXAMPLE 9.27 REGISTERS AS AGENTS ON A DATA BUS – SIMULATION WAVEFORM ................................ ............................... 325
EXAMPLE 9.28 RTL MODEL OF A 4-STAGE, 8-BIT SHIFT REGISTER IN VERILOG ................................. ................................. 326
EXAMPLE 10.1 CALCULATING THE FINAL DIGIT LINE VOLTAGE IN A DRAM BASED ON CHARGE SHARING ................. ................ 348
EXAMPLE 10.2 BEHAVIORAL MODELS OF A 4  4 ASYNCHRONOUS READ ONLY MEMORY IN VERILOG ................... .................. 352
EXAMPLE 10.3 BEHAVIORAL MODELS OF A 4  4 SYNCHRONOUS READ ONLY MEMORY IN VERILOG .................... ................... 353
EXAMPLE 10.4 BEHAVIORAL MODEL OF A 4  4 ASYNCHRONOUS READ/WRITE MEMORY IN VERILOG ................... .................. 354
EXAMPLE 10.5 BEHAVIORAL MODEL OF A 4  4 SYNCHRONOUS READ/WRITE MEMORY IN VERILOG .................... ................... 355
EXAMPLE 12.1 DESIGN OF A HALF ADDER ........................................................... ........................................................... 374
EXAMPLE 12.2 DESIGN OF A FULL ADDER ........................................................... ........................................................... 374
EXAMPLE 12.3 DESIGN OF A FULL ADDER OUT OF HALF ADDERS ............................................ ........................................... 376
EXAMPLE 12.4 DESIGN OF A 4-BIT RIPPLE CARRY ADDER (RCA) ........................................... .......................................... 377
EXAMPLE 12.5 TIMING ANALYSIS OF A 4-BIT RIPPLE CARRY ADDER .......................................... .......................................... 378
EXAMPLE 12.6 DESIGN OF A 4-BIT CARRY LOOK AHEAD ADDER (CLA) – OVERVIEW .............................. ............................. 379
EXAMPLE 12.7 DESIGN OF A 4-BIT CARRY LOOK AHEAD ADDER (CLA) – ALGEBRAIC FORMATION ..................... ..................... 380
EXAMPLE 12.8 TIMING ANALYSIS OF A 4-BIT CARRY LOOK AHEAD ADDER ...................................... ...................................... 381
EXAMPLE 12.9 STRUCTURAL MODEL OF A FULL ADDER USING TWO HALF ADDERS IN VERILOG ........................ ....................... 382
EXAMPLE 12.10 STRUCTURAL MODEL OF A 4-BIT RIPPLE CARRY ADDER IN VERILOG .............................. .............................. 383
EXAMPLE 12.11 TEST BENCH FOR A 4-BIT RIPPLE CARRY ADDER USING NESTED FOR LOOPS IN VERILOG ................ ............... 384
EXAMPLE 12.12 STRUCTURAL MODEL OF A 4-BIT CARRY LOOK AHEAD ADDER IN VERILOG .......................... ......................... 385
EXAMPLE 12.13 4-BIT CARRY LOOK AHEAD ADDER – SIMULATION WAVEFORM .................................. ................................. 385
EXAMPLE 12.14 BEHAVIORAL MODEL OF A 4-BIT ADDER IN VERILOG ......................................... ........................................ 386
EXAMPLE 12.15 DESIGN OF A 4-BIT SUBTRACTOR USING FULL ADDERS ....................................... ...................................... 387
EXAMPLE 12.16 CREATING A PROGRAMMABLE INVERTER USING AN XOR GATE ................................. ................................. 387
EXAMPLE 12.17 DESIGN OF A 4-BIT PROGRAMMABLE ADDER/SUBTRACTOR ..................................... .................................... 388
EXAMPLE 12.18 PERFORMING LONG MULTIPLICATION ON DECIMAL NUMBERS .................................... ................................... 389
EXAMPLE 12.19 PERFORMING LONG MULTIPLICATION ON BINARY NUMBERS ..................................... .................................... 390
EXAMPLE 12.20 DESIGN OF A SINGLE-BIT MULTIPLIER ................................................... ................................................... 390
EXAMPLE 12.21 DESIGN OF A 4-BIT UNSIGNED MULTIPLIER ............................................... ............................................... 391
EXAMPLE 12.22 TIMING ANALYSIS OF A 4-BIT UNSIGNED MULTIPLIER ......................................... ......................................... 392
EXAMPLE 12.23 MULTIPLYING AN UNSIGNED BINARY NUMBER BY TWO USING A LOGICAL SHIFT LEFT .................... ................... 392
EXAMPLE 12.24 ILLUSTRATING HOW AN UNSIGNED MULTIPLIER INCORRECTLY HANDLES SIGNED NUMBERS ................ ................ 393
EXAMPLE 12.25 PROCESS TO CORRECTLY HANDLE SIGNED NUMBERS USING AN UNSIGNED MULTIPLIER .................. ................. 394
EXAMPLE 12.26 PERFORMING LONG DIVISION ON DECIMAL NUMBERS ......................................... ........................................ 395
EXAMPLE 12.27 PERFORMING LONG MULTIPLICATION ON BINARY NUMBERS ..................................... .................................... 396
EXAMPLE 12.28 DESIGN OF A 4-BIT UNSIGNED DIVIDER USING A SERIES OF ITERATIVE SUBTRACTORS .................. .................. 397
EXAMPLE 12.29 DIVIDING 11112 (1510) BY 01112 (710) USING THE ITERATIVE SUBTRACTION ARCHITECTURE ............ ............ 398
EXAMPLE 12.30 DIVIDING AN UNSIGNED BINARY NUMBERS BY TWO USING A LOGICAL SHIFT RIGHT ..................... .................... 399
EXAMPLE 13.1 MEMORY MAP FOR A 256  8 MEMORY SYSTEM ........................................... ........................................... 408
EXAMPLE 13.2 EXECUTION OF AN INSTRUCTION TO “LOAD REGISTER A USING IMMEDIATE ADDRESSING” ................................... 411
EXAMPLE 13.3 EXECUTION OF AN INSTRUCTION TO “LOAD REGISTER A USING DIRECT ADDRESSING” ................... ................... 412
EXAMPLE 13.4 EXECUTION OF AN INSTRUCTION TO “STORE REGISTER A USING DIRECT ADDRESSING” .................. .................. 413
EXAMPLE 13.5 EXECUTION OF AN INSTRUCTION TO “ADD REGISTERS A AND B” ................................. ................................. 414
EXAMPLE 13.6 EXECUTION OF AN INSTRUCTION TO “BRANCH ALWAYS” ........................................ ....................................... 415
EXAMPLE 13.7 EXECUTION OF AN INSTRUCTION TO “BRANCH IF EQUAL TO ZERO” ................................ ................................ 416
EXAMPLE 13.8 TOP LEVEL BLOCK DIAGRAM FOR THE 8-BIT COMPUTER SYSTEM ................................. ................................. 418
EXAMPLE 13.9 INSTRUCTION SET FOR THE 8-BIT COMPUTER SYSTEM ......................................... ........................................ 419
EXAMPLE 13.10 MEMORY SYSTEM BLOCK DIAGRAM FOR THE 8-BIT COMPUTER SYSTEM ............................ ........................... 420
EXAMPLE 13.11 CPU BLOCK DIAGRAM FOR THE 8-BIT COMPUTER SYSTEM .................................... .................................... 424
EXAMPLE 13.12 STATE DIAGRAM FOR LDA_IMM ..................................................... ..................................................... 431
EXAMPLE 13.13 SIMULATION WAVEFORM FOR LDA_IMM ................................................ ................................................ 432
EXAMPLE 13.14 STATE DIAGRAM FOR LDA_DIR ...................................................... ..................................................... 433
EXAMPLE 13.15 SIMULATION WAVEFORM FOR LDA_DIR ................................................ ................................................ 434
EXAMPLE 13.16 STATE DIAGRAM FOR STA_DIR ...................................................... ..................................................... 435
 Appendix A: List of Worked Examples • 453

EXAMPLE 13.17 SIMULATION WAVEFORM FOR STA_DIR ................................................ ................................................ 436

EXAMPLE 13.18 STATE DIAGRAM FOR ADD_AB ...................................................... ...................................................... 437
EXAMPLE 13.19 SIMULATION WAVEFORM FOR ADD_AB ................................................. ................................................ 438
EXAMPLE 13.20 STATE DIAGRAM FOR BRA .......................................................... ......................................................... 439
EXAMPLE 13.21 SIMULATION WAVEFORM FOR BRA .................................................... .................................................... 440
EXAMPLE 13.22 STATE DIAGRAM FOR BEQ ......................................................... ......................................................... 441
EXAMPLE 13.23 SIMULATION WAVEFORM FOR BEQ WHEN TAKING THE BRANCH (Z ¼ 1) .......................... .......................... 442
EXAMPLE 13.24 SIMULATION WAVEFORM FOR BEQ WHEN THE BRANCH IS NOT TAKEN (Z ¼ 0) ...................... ..................... 443
Index
A Capacity, 331
Carry, 21
Absorption, 92 Carry look ahead adders (CLA), 378
Abstraction, 145 Charge sharing, 347
AC specifications. See Switching characteristics Classical digital design flow, 149
Adder/subtractor circuit, 387 CMOS. See Complementary metal oxide
Adders semiconductor (CMOS)
in Verilog, 381 Combinational logic analysis, 99
Addition, 21, 373 Combining, 93
AND gate, 40 Commutative property, 88
Anti-fuse, 337 Complementary metal oxide semiconductor
Associative property, 89 (CMOS), 4, 56
Asynchronous memory, 335 gates, 58
Axioms, 82 CMOS inverter, 58
logical negation, 82 CMOS NAND gate, 59
logical precedence, 83 CMOS NOR gate, 62
logical product, 82 operation, 57
logical sum, 83 Complements, 87
logical values, 82 Complete sum, 125
Complex programmable logic device (CPLD), 363
B Computer system design, 403, 410, 413, 414, 418,
Base, 7 419, 424, 427, 430
Base conversions, 11 addressing modes, 409
binary to decimal, 12 arithmetic logic unit (ALU), 405
binary to hexadecimal, 18 central processing unit, 405
binary to octal, 17 condition code register, 405
decimal to binary, 15 control unit, 405
decimal to decimal, 11 data memory, 404
decimal to hexadecimal, 16 data path, 405
decimal to octal, 15 direct addressing, 410
hexadecimal to binary, 19 example 8-bit system, 417
hexadecimal to decimal, 14 control unit, 427
CPU, 424
hexadecimal to octal, 20
detailed instruction execution, 430
octal to binary, 19
instruction set, 418
octal to decimal, 13
memory system, 419
octal to hexadecimal, 19
general purpose registers, 405
Binary addition. See Addition
hardware, 403
Binary number system, 9
immediate addressing, 410
Binary subtraction. See Subtraction
indexed addressing, 410
Bipolar junction transistor (BJT), 65
inherent addressing, 410
Bistable, 200
input output ports, 404
Boolean algebra, 81
instructions, 403
Boolean algebra theorems, 83 branches, 414
Borrows, 22 data manipulations, 413
Break-before-make switch behavior, 219 loads and stores, 410
Buffer, 39 register, 405
Byte, 10 memory address register, 405
memory map, 408
C
memory mapped system, 406
Canonical product of sums, 106 opcodes, 409
Canonical sum of products, 103 operands, 409

# Springer International Publishing AG 2017 455

B.J. LaMeres, Introduction to Logic Circuits & Logic Design with Verilog,
DOI 10.1007/978-3-319-53883-9
456 • Index

Computer system design (cont.) Digital design flow, 149

program, 403 Diodes, 75
counter, 405 7400 DIP pin-out, 69
memory, 404 Discrete components, 56
registers, 405 Disjunction (_), 82
software, 403, 409 Distinguished one cells, 125
Configurable logic block (CLB), 365 Distributive property, 91
Conjunction (^), 82 Division, 395
Converting between bases. See Base conversions by powers of 2, 398
Converting between positive and negative logic, 84 signed, 399
Counters, 241 unsigned, 395
designing by hand, 241 using iterative subtractions, 396
Covering, 92 D latch, 210
Cross-coupled inverter pair, 199 Don’t cares (X), 125
Double pole, double throw (DPDT) switch, 218
D Double pole, single throw (DPST) switch, 218
Data sheet, 51 Driving loads, 71
7400 DC operating conditions, 69 LEDs, 76
DC specifications, 46 resistive loads, 73
II-max, 47 Dual in-line package (DIP), 69
IIH-max, 47 Duality, 83
IIL-max, 47 Dynamic hazard, 130
IO-max, 46 Dynamic random access memory (DRAM), 345
IOH-max, 46
E
IOL-max, 46
Iq (quiescent current), 48 Electrical signaling, 1
NMH, 47 Electrically erasable programmable read only
NML, 47 memory (EEPROM), 340
VIH-max, 47 Encoders, 188
VIH-min, 47 Erasable programmable read only memory
VIL-max, 47 (EPROM), 338
VIL-min, 47 Essential prime implicant, 125
VOH-max, 46
VOH-min, 46 F
VOL-max, 46
Field programmable gate array (FPGA), 364
VOL-min, 46
Finite state machines (FSM), 223
Decimal number system, 9
binary state encoding, 226
Decoders, 181
design examples by hand, 233
DeMorgan’s theorem, 93
design process, 232
DeMorgan’s theorem of duality, 83
final logic diagram, 231
Demultiplexer design by hand, 193
gray code state encoding, 227
Demultiplexers, 193
introduction, 223
Design abstraction, 145
next state logic, 229
Design domains, 146
one-hot state encoding, 228
behavioral, 146
output logic, 230
physical, 146
reset condition, 254
structural, 146
state diagram, 223
Design levels, 146
state memory, 226
algorithmic, 146
state transition table, 225
circuit, 146
state variables, 229
gate, 146
synthesis by hand, 225
register transfer, 146
FLASH memory, 341
system, 146
NAND-FLASH, 341
Design simplicity, 3
NOR-FLASH, 341
D-flip-flop, 211
Floating-gate transistor, 338
Digit, 9
Forward current (IF), 75
notation, 9
 Index • 457

Forward voltage (VF), 75 Metal oxide semiconductor field effect transistor

Full adders, 374 (MOSFET), 56
Functionally complete sets, 98 Metastability, 200
Fuse, 337 Minimal sum, 123, 125
Minimization, 112
G of logic algebraically, 112
Gajski and Kuhn’s Y-chart, 146 of logic using K-maps, 116
Gates, 37 Minterm list (∑), 104
Generic array logic (GAL), 361 Minterms, 103
Glitches, 129 Modern digital design flow, 149
Moore machine, 224
H MOSFET. See Metal oxide semiconductor field
effect transistor (MOSFET)
Half adders, 374 Most significant bit (MSB), 10
Hard array logic (HAL), 362 Multiplexer design by hand, 190
Hazards, 129 Multiplexer modeling in Verilog, 191
Hexadecimal number system, 10 Multiplexers, 190
History of HDLs, 142 Multiplication, 389
by powers of 2, 392
I
combinational multiplier, 391
Idempotent, 87 shift and add approach, 389
Identity theorem, 86 signed, 393
Input/output blocks (IOBs), 369 unsigned, 389
Integrated circuit, 56
Inverter, 40
N
Involution, 88 NAND-debounce circuit, 220
NAND gate, 41
K
Negation (Ø), 82
Karnaugh map (K-map), 113 Negative logic, 45
Nibble, 10
L NMOS, 57
Noise, 2
Large scale integrated circuit (LSI) logic, 181
Noise margin HIGH (MNH), 47
Leading zero, 9
Noise margin LOW (MNL), 47
Least significant bit (LSB), 10
Non-volatile memory, 332
Light emitting diodes (LEDs), 75
NOR gate, 41
Logic block (LE), 365
NPN, 65
Logic expression, 38
Null element, 86
Logic families, 56
Numerals, 7
Logic function, 38
Logic HIGH, 45 O
Logic levels, 45
Logic LOW, 45 Octal number system, 10
Logic minimization, 112 Ohm's law, 73
Logic symbol, 37 One-hot binary encoder design by hand, 188
Logic synthesis, 103 One-hot binary encoder modeling in Verilog, 188
Logic value, 45 One-hot decoder design by hand, 182
Logic waveform, 39 One-hot decoder modeling in Verilog, 182
Look-up table (LUT), 365 One’s complement numbers, 27
OR gate, 41
M Output DC specifications. See DC specifications
Output logic macrocell (OLMC), 361
Mask read only memory (MROM), 336
Maxterm list (Π), 108 P
Maxterms, 106
Mealy machine, 224 7400 Part numbering scheme, 68
Medium scale integrated circuit (MSI) logic, 181 Place and route, 149
Memory map model, 331 PMOS, 57
458 • Index

PNP, 65 Sourcing multiple loads, 50

Positional number system, 7 SR latch, 202, 205
Positional weight, 11 SR latch with enable, 208
Positive logic, 45 Static 0 hazard, 130
Postulates, 82 Static 1 hazard, 130
Power consumption, 4 Static random access memory (SRAM), 342
Power supplies, 48 Subtraction, 22, 387
ICC, 48 Sum of products (SOP) form, 94
IGND, 48 Switch debouncing, 217
VCC, 48 Switching characteristics, 51
Prime implicant, 117 tf (fall time), 51
Product of sums (POS) form, 94 tPHL (propagation delay HIGH to LOW), 51
Programmable array logic (PAL), 360 tPLH (propagation delay LOW to HIGH), 51
Programmable interconnect points (PIPs), 368 tr (rise time), 51
Programmable logic array (PLA), 359 tt (transition time), 51
Programmable read only memory (PROM), 337 Synchronous memory, 335
Proof by exhaustion, 83
Pull-down network, 58 T
Pull-up network, 58
Technology mapping, 149
Q Timing hazards, 129
Toggle flop (T-flop), 216
Quiescent current (Iq), 48 Trailing zero, 9
Transistor-Transistor Logic (TTL), 65
R Transmitter/receiver circuit, 44
Radix, 7 Truth table formation, 38
point, 8 TTL operation, 65
Random access memory (RAM), 333 Two’s complement arithmetic, 31
Two’s complement numbers, 29
Range
one’s complement numbers, 27 U
signed magnitude numbers, 25
two’s complement numbers, 29 Uniting, 93
Read cycle, 331 Unsigned numbers, 24
Read only memory (ROM), 332, 333
Read/write (RW) memory, 332 V
Ripple carry adders (RCA), 376 Verification, 147
Ripple counter, 217 Verilog, 141, 155, 159–163, 170–173, 286, 287, 289,
S 303–306, 308, 319, 322, 323, 325, 381
always blocks, 272
7-Segment decoder design by hand, 184 arrays, 155
7-Segment decoder modeling in Verilog, 186 behavioral modeling techniques, 309, 310,
Semiconductor memory, 331 312, 319–321
Sequential access memory, 333 adders, 381
Sequential logic analysis, 255 agents on a bus, 323
Sequential logic timing, 215 counters, 319
7400 Series logic families, 67 D-flip-flops, 304
Shift register, 222 D-flip-flop with enable, 306
Signaling, 1 D-flip-flop with preset, 305
Signed magnitude numbers, 25 D-flip-flop with reset, 304
Signed numbers, 24
D-latches, 303
finite state machines, 308
Simple programmable logic device (SPLD), 363
encoding styles, 312
Single pole, double throw (SPDT) switch, 218
next state logic, 309
Single pole, single throw (SPST) switch, 218
output logic, 310
Sinking current, 46, 47 state memory, 309
Small scale integrated circuit (SSI) logic, 181 state variables, 309
Sourcing and sinking multiple loads, 50 registers, 323
Sourcing current, 46 RTL modeling, 322
 Index • 459

shift registers, 325 parameters, 158

up counter, 319 procedural assignment, 271
up counter with range checking, 320 procedural blocks, 271
up counters with enables, 320 repeat loops, 284
up counters with loads, 321 resolution, 153
casex statements, 283 RTL modeling, 322
casez statements, 283 sensitivity lists, 273
compiler directives, 159 signal declaration, 157
include, 159 statement groups, 279
timescale, 159 structural design and hierarchy, 170
continuous assignment, 164 explicity port mapping, 170
with conditional operators, 165 gate level primitives, 172
with delay, 167 instantiation, 170
with logical operators, 164 positional port mapping, 171
counters, 319 user defined primitives, 173
data types, 153 system tasks, 286
disable, 285 file I/O, 287
drive strength, 153 simulation control, 289
finite state machines, 308 text I/O, 286
for loops, 284 user defined primitives, 173
forever loops, 283 value set, 153
gate level primitives, 172 variable data types, 154
history, 143 vectors, 154
if-else statements, 280, 281 while loops, 283
initial blocks, 272 Very large scale integrated circuit (VLSI) logic, 181
net data types, 154 Volatile memory, 332
number formatting
binary, 155 W
decimal, 155
Weight, 11
hex, 155
Word, 10
octal, 155
Write cycle, 331
operators, 159
assignment, 159 X
bitwise logical, 159
bitwise replication, 162 X - don’t cares, 125
Boolean logic, 160 XNOR gate, 43
concatenation, 161 XOR gate, 42
conditional, 161 XOR/XNOR gates in K-maps, 126
numerical, 162
precedence, 163 Y
reduction, 160
relational, 160 Y-chart, 146